6HVVLRQ
7+(3526$1'&2162)),6&$/58/(6
(08),6&$/58/(6$1(:$16:(572$12/'48(67,21"
)DEUL]LR%DODVVRQHDQG'DQLHOH)UDQFR
,QWURGXFWLRQ
Fiscal sustainability is a central tenet of European Monetary Union
(EMU); it is a precondition for financial and monetary stability. Budgetary
flexibility is needed for stabilisation policy; it has become more important
in EMU as member states can no longer rely on a monetary policy tailored
on national needs nor on exchange rate adjustments. EMU fiscal rules have
been designed with the goal to ensure that national policies keep a sound
fiscal stance while allowing sufficient margins for budgetary flexibility in
bad times
1
.
The Stability and Growth Pact commits EMU member states to a
medium term objective of budgetary position close to balance or in surplus.
The main rationale for such a target is that its attainment will allow
member states to deal with normal cyclical fluctuations while keeping the
government deficit within the value of 3% of GDP set in the Treaty of
Maastricht
2
. Compliance with this threshold, and with the 60 per cent
ceiling for the debt to GDP ratio, will prevent the public finances of EMU
member states from taking unsustainable paths.
In this paper we try to assess to what extent the issue facing the
founders of EMU was a new one in the field of public finance and to what
extent the solution chosen can be regarded as innovative. To this end, we
review the literature on budgetary rules from its very beginning to the years
immediately before the Treaty of Maastricht (section 2). The review is
largely based on quotations drawn from economists and policy makers. On
the basis of this review, in section 3 we argue that the bulk of EMU fiscal
__________
*
Research Department, Banca d’Italia. The views expressed in this paper are those of the authors
and do not commit the Banca d’Italia. The authors wish to thank Prof. Sergio Steve for his
comments and suggestions.
1
The economic policy framework of EMU is extensively examined in Buti and Sapir (1998) and in
Buti, Brunila and Franco (2001). The theory of fiscal sustainability and its links with EMU fiscal
rules are reviewed in Balassone and Franco (2000a); see also the papers in Banca d’Italia (2000).
On the flexibility allowed by EMU fiscal rules see Buti
. (1997) and Balassone and Monacelli
(2000).
2
The actual definition of this medium term objective requires several factors to be taken into
account; see section 3.1.
regulation does not qualify as innovative; however, the interaction between
the multinational nature of EMU and the lack of a federal political
authority (a truly innovative feature) shaped the solution chosen. The
highly decentralised setting of fiscal policy in EMU gave prominence to
moral hazard issues and EMU fiscal rules, while drawing heavily on ideas
that are central to the long lasting debate on fiscal rules, are innovative in
the way in which different approaches are blended and complemented by
innovative and pragmatic choices.
7KHEDODQFHGEXGJHWUXOHDQGLWVDPHQGPHQWV
Mankind has always displayed a certain degree of awareness of the
potential negative effects of excessive borrowing. Exhortation to sound
fiscal behaviour can be found as early as in the Bible: “And thou shalt lend
unto many nations, but thou shalt not borrow” (Deuteronomy 15:6).
Several centuries later, as Hansen (1941) reminds us: “Scholastic
theologicians, like Thomas Aquinas, were bitterly opposed to loans…
Political philosophers of the early modern period continued to regard the
prior accumulation of treasures as superior to borrowing… [For] Jean
Bodin … emergencies should be met by accumulated hoards, and only war
provided justification for extraordinary levies or loans. Thomas Hobbes
was more realistic … [allowing] the monarch [to resort] occasionally even
to the public credit… [but] Adam Smith reverted to the older tradition …
Hume likewise wrote … [that] to mortgage the public revenues … [is] a
practice that appears ruinous” (p. 110).
Burkhead (1954) notes that there is a common body of doctrine that
may be characterised as the classical view of debt and deficits that goes
from Smith to Mill. These writers recognised that there are productive uses
to which borrowed resources may be put; however, they feared that
unproductive use was more likely and strongly opposed deficit finance
when giving policy advice. They noted that interest payments would pose a
burden on future taxpayers, but their main concern was the loss of wealth
borne when the deficit was incurred in the first place.
Smith opposed unbalanced budget on the ground that government
borrowing would deprive society of resources which could be invested
more productively. He also noted that beyond a certain threshold debt
inevitably leads to national bankruptcy.
Say argued that the possibility of borrowing allows governments “…
to conceive gigantic projects that lead sometimes to disgrace, sometimes to
glory, but always to a state of financial exhaustion; to make war
themselves and stir up others to do the like; to subsidize every mercenary
agent and deal in the blood and the consciences of mankind; making capital
which should be the fruit of industry and virtue, the prize of ambition,
pride, and wickedness
3
”.
Ricardo refers to the debt as “… one of the most terrible scourges
which was ever invented to afflict a nation
4
”, as “… a system which tends
to make us less thrifty, to blind us to our real situation”. He feared that the
citizen initially “deludes himself with the belief, that he is as rich as
before” and then, faced with the taxes levied to pay for the debt, is tempted
“… to remove himself and his capital to another country, where he will be
exempted from such burthens
5
”.
In short, for a long time the only budgetary rule was that of a
balanced budget. This rule was probably based on an analogy between
government and family finance drawn when the budget of the State was the
budget of a monarch and it was separate from the finances of his subjects.
The precept was therefore to avoid living beyond one’s means. “As Adam
Smith put it, ‘what is prudence in the conduct of every private family, can
scarcely be folly in that of a great kingdom’”
6
or, following Dickens’ Mr.
Micawber: “… if a man had twenty pounds a year for his income, and
spent nineteen pound nineteen shillings and sixpence, he would be happy,
but … if he spent twenty pounds one he would be miserable
7
”.
In the second part of the XIX century, the precept of a balanced
budget still found a widespread endorsement. Ursula Hicks notes that
“Gladstonian budgeting is inextricably bound up with the theory of the
ever-balanced (or even over-balanced) budget” (1953, p. 25) and quotes the
following statement by Lowe, a disciple of Gladstone, “I would define a
Chancellor of the Exchequer as an animal who ought to have a surplus; if
__________
3
Say (1853), p. 483.
4
The Works and Correspondence of David Ricardo, P. Sraffa, ed., (1952-73), vol. IV, p. 197.
5
The Works and Correspondence of David Ricardo, P. Sraffa, ed., (1952-73), vol. IV, p. 247-8.
6
Premchand (1983), p.4.
7
Charles Dickens, .
under extraordinary conditions he has not a surplus he fails to fulfil the
very end and object of his being” (p. 25)
8
.
Deficit and debt drew less attention from economists. For instance,
as Burkhead (1954) notes, Marshall’s 3ULQFLSOHV devote no attention to
these issues. Noticeably, Puviani (1903) devotes most of its analysis of
fiscal illusion to public expenditure and revenue. While he notes that
politicians may prefer borrowing to extraordinary levies because citizens
underestimate future interest burdens, this argument remains relatively
unimportant in his analysis of the methods employed by governments to
influence citizens’ perception of fiscal policy.
The consensus on the balanced budget is witnessed by Pigou’s 1929
writing: “in normal times the main part of a government’s revenue is
required to meet regular expenditure that recurs year after year. There can
be no question that in a well-ordered State all such expenditure will be
provided for out of taxation, and not by borrowing. To meet it by
borrowing … would involve an ever-growing government debt and a
corresponding ever-growing obligation of interest. … The national credit
would suffer heavy damage; ... This thesis is universally accepted” (1929,
p. 233).
Even after the keynesian revolution the virtue of a balanced budget
kept being praised. Truman’s 1951 (FRQRPLF 5HSRUW RI WKH 3UHVLGHQW
stated that “… we should make it the first principle of economic and fiscal
policy in these times to maintain a balanced budget, and to finance the cost
of national defence on a ‘pay-as-we-go’ basis”.
As Schumacher noted in 1946, the precepts of sound public finance
were grounded in the opinion that the economy is self-equilibrating
9
. “The
logical corollary of orthodox economics is orthodox finance. If it is
believed that all factors of production are normally and inevitably utilised
by private business, it follows that the State can obtain the use of such
factors only by preventing private business from using them. … From this
it follows that the first principle of ‘sound’ public finance is that the budget
should be balanced” (p. 86).
__________
8
U. Hicks relates this view to the objective of reducing the debt and taxation, to the prevailing
favourable economic conditions and also to some difficulties in managing the budget. She notes the
growth of administrative expertise in budgeting contributed to the development of a different
approach in the 1930s.
9
Schumacher, while reporting these views, did not share them.
Almost seventy years later than Pigou, Buchanan echoes his words:
“the first century and one-half of our national political history did, indeed,
embody a norm of budget balance. This rule was not written in the
constitution document, as such, but rather it was part of an accepted set of
attitudes about how government should, and must, carry on its fiscal
affairs” (1997, p. 119).
However, even in family finance borrowing is not necessarily evil.
Even classical advocates of the balanced budget were aware of the
necessity of allowing borrowing in certain circumstances and of its
usefulness in others. Therefore economists have had a hard job in trying to
specify under what circumstances exception to the balanced budget rule
were to be allowed, caught between the Scylla of missed opportunities as a
consequence of the constraint and the Charybdis of waste and instability
caused by its removal.
The need for exceptions as well as the need for tight rules to deal
with them was clearly recognised by Pigou (1929). He deemed it to be
plain that when “non-remunerative government expenditures RQ D ZKROO\
DEQRUPDOVFDOH have to be undertaken, as in combating the consequences
of an earthquake or to meet an imminent threat of war … to collect what is
required, and required at a very short notice in these conditions, through
the machinery of taxation is politically and administratively impracticable”
(p. 39; italics ours). He also argued that concerning “government
expenditure devoted to producing capital equipment … WKHIUXLWVRIZKLFK
ZLOOVXEVHTXHQWO\ EH VROGWR SXUFKDVHUVIRU IHHV … it is generally agreed
that the required funds ought to be raised by loans. …Upon this matter …
there is no room for controversy” (p. 36; italics ours). Finally, he notes that
“…since changes in taxation always involve disturbance, to keep the rates
of taxation as nearly as possible constant from year to year … it may be
desirable … to arrange a budget so that good and bad years make up for
one another, DGHILFLWLQRQHEDODQFLQJDVXUSOXVLQDQRWKHU" (p. 35; italics
ours).
2UGLQDU\YVH[WUDRUGLQDU\ILQDQFH
One first exception was thus found in the distinction between
ordinary and extraordinary finance: the former dealt with recurrent
expenditures, to be financed by recurrent revenues so as to avoid the
depletion of non-renewable assets; the latter dealt with one-off outlays to
be backed also by borrowed funds. Also the rationale for this exception
was found by way of analogy to family finance. De Viti de Marco (1953)
points out that “…if an individual has to face an expense which he reckons
to exceed his annual income … he will either have to sell his assets or raise
a loan” (p. 390, our translation) and applies the same line of reasoning to
public finances
10
.
De Viti de Marco is very much aware of classification problems as
the extraordinariness of an outlay is a matter for subjective assessment both
at the individual and at the collective level: “this subjective element does
not allow to define a rigorous and objective rule that draws the line …
between ordinary and extraordinary finance” (p. 390, our translation).
Margins for moral hazard and opportunistic behaviour arise as “the
distinction between ‘ordinary’ and ‘extraordinary’ receipts and expenditure
is admittedly not clear-cut, depending ultimately on the judgement of the
classifying authority as to whether the receipts and expenditure in question
are to continue indefinitely in the future” (United Nations, 1951, p. 61).
While extreme cases were easily identified (on the one hand, interest
outlays and salaries; on the other, the cost of a war), in some cases it is not
straightforward to see what is ordinary and what is not. “It is impossible to
define ex ante what is an extraordinary outlay. Building a school may be an
extraordinary effort for a small town, an ordinary one for a big city”
(Einaudi, 1948, p. 318; our translation). Ultimately, “there is no great
technical difficulty in producing for a series of years budgets which are
balanced at the end of the year to the nearest penny … Perhaps half a
dozen financial writers in the country would understand from the published
accounts what was happening, but I doubt if any one of the half dozen is
capable of making the position clear to the public
11
”.
National experiences did not differ much. “In the case of France, the
extraordinary budget was proverbially the dumping place for all
expenditures which could not be balanced by tax receipts” (Hansen, 1941,
p. 199). In 1945 Keynes notes that in the United Kingdom “the present
criterion leads to meaningless anomalies. A new G.P.O. is charged
‘below’, a new Somerset House ‘above’. A Capital contribution to school
buildings is ‘above’ in the Exchequer Accounts and is paid for out of
Revenues, and is ‘below’ in the Local Authority Accounts and is paid for
__________
10
De Viti de Marco goes on to justify deficit finance as a less painful alternative to extraordinary
taxation which may penalise liquidity constrained taxpayers.
11
Sir F. Phillips, writing in 1936, quoted in Middleton (1985), p. 82.
out of loans. The cost of a road is ‘above’, of a railway ‘below’. And so
on
12
”. “In Canada, although not always realised even by Canadians, a
budgetary distinction between ordinary and capital expenditures has been
made ever since the confederation in 1867. The official reports show
surpluses in fifty of the sixty-six years following 1867; but if the
accounting were made on the United States basis, surpluses would appear
in only fifteen of the sixty-six years” (Hansen, 1941, p. 199)
13
.
7KHGRXEOHEXGJHW
The double budget is a refinement of the ordinary/extraordinary
distinction which reduces the degree of arbitrariness of the decision
concerning which expenditures can be deficit-financed. The budget is split
into a current and a capital account. While the former must be balanced or
in surplus, the latter can run a deficit (the so-called JROGHQUXOH) and thus
allows to spread the cost of durables over all the financial years in which
they will be in use rather than charging it entirely on one year. It can be a
powerful instrument in overcoming liquidity constraints and fostering
economic development structurally.
Arguments along these lines can be found earlier than the dual
budget debate per se. The productive character of a large part of public
outlays was noted by German scholars in the second part of the XIX
century. They also argued that government can borrow to finance
undertakings that are expected to improve the income of future generations
(Cohn, 1895). Bastable (1927) argues that “non-economic (i.e. non-
remunerative) expenditure is primarily to be met out of income, and, unless
it can be so dealt with, ought not to be incurred…[and] … that borrowing
__________
12
Memorandum by Keynes for the National Debt Enquiry, 21 June 1945, in D. Moggridge and A.
Robinson, eds., (1971-89), vol. XXVII, pp. 406-7. On the UK experience, Clarke (1998) also notes
that “in the best Gladstonian tradition …. On the expenditure side, what mattered was expenditure
above the famous ‘line’ in the Exchequer accounts, dating from the Sinking Fund Act of 1875,
broadly … distinguishing a revenue account from a capital account – but by no means
unambiguously … Only an old Treasury hand could be expected to know the difference within this
hybrid accounting framework …. therefore, the simple moral imperative of balancing the budget
was in practice wrapped in the esoteric conventions of the public accounts” (p. 64).
13
In Italy, in the late 19
th
Century and the early 20
th
Century revenues and expenditures related to the
construction of railways were included in a special balance sheet and separated from other ordinary
and extra-ordinary items. Revenues were represented by the proceeds of the sale of bonds,
expenditures by the outlays for investment projects (see Nitti, 1903). De facto, an item specific
golden rule was implemented.
should hardly ever be adopted except for strictly economic expenditure,
and then only when the extension of the State domain is clearly advisable”
(pp. 670-1).
The usefulness of a dual budget has been long debated
14
. It is still an
unsettled issue, which has been tackled in different ways in different
countries and at different times. Sweden, introduced the dual budget in
1937 and suppressed it in 1980.
First of all, the distinction between current and capital items retains a
certain degree of ambiguity which can be used opportunistically. “The
classification procedures which are to be followed in separating “current”
and “capital” transactions are among the most controversial and difficult
questions in budgetary procedure, especially in view of the frequent abuses
of so-called “capital budgets” in hiding deficits which otherwise would
have become apparent” (United Nations, 1951, p. 11).
According to Lindbeck (1968), this distinction “facilitated tactical
political manoeuvres and hampered the fiscal policy debate for many years
[in Sweden] by focusing it on complicated bookkeeping issues understood
by very few and of very little economic relevance” (p. 34).
In principle, one can distinguish between durable goods producing a
direct revenue, durable goods producing an indirect revenue as,
respectively, investments by publicly owned enterprises and public
infrastructures that reduce the costs borne by private producers and/or
consumers and durable goods with pure consumption functions. It may be
argued that the latter should be excluded from the capital account as they
do not affect growth and thus do not imply a future financial benefit for the
public sector; therefore they worsen the sector’s net worth.
In practice, however, the divide between the second and the third
category is very unclear. In the case of infrastructures, for example, there is
the issue of the treatment of expenditures determined by the attempt to
reduce the impact on the environment. If the overall costs increase should
these expenditures be considered as producing an indirect revenue or as
pure consumption? If for this reason we include in the capital account all
durables, we end up creating a distortion in allocation only based on
duration, rather than on contribution to growth, thus "the analogy with
__________
14
See the accounts in Premchand (1983) and Poterba (1995). For a recent discussion in the context of
EMU, see Balassone and Franco (2000b).
private accounting may be conductive to an irrational preference for capital
expenditures over current expenditures" (Goode and Birnbaum, 1955,
p. 1)
15
.
Clearly there are current expenditures, such as those increasing
human capital, that can give a relevant contribution to growth as "indirect
revenue need not come through a durable good" (Steve, 1972, p. 164; our
translation). If one is not careful about the expenditures to be included in
the capital section, the dual budget may result “...in a preference for
expenditures on physical assets rather than greater spending for intangibles
such as health or education” (Colm and Wagner, 1963, p. 125). Thus, "the
need for a return, either in the limited financial sense or in the broader
context of the social return, is a view that needs to be applied over a wider
spectrum of public expenditures and not confined to capital budget only"
(Premchand, 1983, p. 296).
However, the inclusion in the capital account (which can be financed
through debt) of all expenditures contributing to human capital would
imply high levels of deficits and pose serious problems of classification
16
.
One should also take into account that a part of expenditures replaces
existing capital.
Furthermore, the possibility to borrow, without strict limits, in order
to finance investments can lower the attention paid when evaluating the
costs and benefits of each project. In a way with the double budget the
analogy between government and private finance moves from the
household to the business sector where the distinction between current and
capital budget is customary. But the analogy between public sector
accounts and those of private enterprises overlooks the absence of
mechanisms that would penalise the public body investing in low revenue
projects.
__________
15
For a discussion along these lines see also Steve (1972; pp. 163-5). Steve also notes that drawing
the line between durable goods with direct and indirect revenue would pose similar problems.
16
Bastable (1927) already acknowledged the usefulness of non-remunerative expenditures such as
those on education, improved housing and the like, however he also pointed out that there is a “…
difficulty of application. The results of expenditure of the kind are hard to trace or measure, and
any of statement respecting them must rest in a great degree of conjecture”. (pp. 621-2).
6WDELOLVDWLRQSROLF\
Another attempt at justifying deviations from the balanced budget
rule came from Keynesian theory where the budget plays a crucial role in
cushioning the effects of cyclical downswings in the economy
compensating for insufficient private demand. Therefore a balanced budget
was no longer to be achieved in each financial year but to be attained over
the whole length of the economic cycle.
On April the 5
th
of 1933 Keynes wrote on The Times: “The next
budget should be divided into two parts, one of which shall include those
items of expenditure which it would be proper to treat as loan-expenditure
in the present circumstances”. Later he sharpens the distinction between
the government’s own current expenditure and a capital budget to provide
for sufficient national investment. In 1942 he writes: “I should aim at
having a surplus on the ordinary budget, which would be transferred to the
Capital Budget, thus gradually replacing dead-weight debt by productive or
semi-productive debt… I should not aim at attempting to compensate
cyclical fluctuations by means of the ordinary budget, I should leave this
duty to the capital budget
17
”.
Fiscal policy in Sweden and in the USA moved along these lines
18
.
In 1937 Sweden reformed its budget rules and abandoned the annual
balancing. In Lindbeck’s account, the Swedish reform was based on the
idea that “in normal times the capital budget should be financed by loans
whereas the current budget should be financed by taxes. In boom periods
the current budget should, however, be overbalanced, hence part of the
capital budget would be financed by taxes; in recession the current budget
should be underbalanced, hence partly financed by loans” (1968, p. 33).
Hansen explains how in the USA, “President Roosevelt … divided
federal expenditures into ‘ordinary’ and ‘extraordinary’. The former relate
to the ‘operating expenditure for the normal and continuing functions of
government’ … [and] … should be met out of current revenues’… He
expressed the hope that in times of prosperity current revenues would so
__________
17
‘Budgetary Policy’, 15 May 1942, in D. Moggridge and A. Robinson, eds., (1971-89), vol. XXVII,
pp. 277-8.
18
These developments reflected common problems but were to a large extent unrelated. On the
relationship between Swedish fiscal policy and Keynesian theories see Lundberg (1996). He recalls
that in 1929 Lindhal considered the use of fiscal policies to affect the level and composition of
demand and that Myrdal was asked to write an appendix to the government budget proposal of
January 1933 on the issue of the feasibility of active fiscal policies.
far exceed ordinary expenditures as to produce ‘a surplus that can be
applied against the public debt’… The extraordinary expenditures, which
are concerned with loans, capital expenditure and relief of need, he deemed
to be sufficiently flexible in character as to permit their contraction and
expansion as a ‘partial offset for the rise and fall in the national income”
(1941, p. 219).
However, the idea of balancing the government accounts over the
course of the business cycle had an exceptionally brief life span. Blinder
and Solow (1974) point out that while it “… had considerable appeal… in
the immediate post-Keynesian years, when the balanced budget was [still]
influential, it is almost never discussed nowadays” (p. 37). It was the turn
of functional finance to take the lead.
“Functional finance rejects completely the traditional doctrines of
‘sound finance’ and the principle of trying to balance the budget over the
solar year or any other arbitrary period … government fiscal policy …
shall all be undertaken with an eye only to the results of these actions on
the economy” (Lerner, 1943, p. 41).
Hansen noted that “if one adopts wholeheartedly the principle that
government financial operations should be regarded exclusively as
instruments of economic and public policy, the concept of a balanced
budget, however defined, can play no role in the determination of that
policy” (1941, p. 188).
The way in which these ideas were first met is exemplified in the
following passage by Chamberlain in 1933
19
: “If I were to pretend I could
lay out a programme under which what I borrowed this year would be met
by a surplus at the end of three years, everyone would soon perceive that I
was only resorting to the rather transparent device of making an
unbalanced budget look respectable
20
”.
__________
19
Middleton (1985) reviews the debate about budgetary policy in the United Kingdom in the 1930s.
In 1933 the Treasury stressed the risks related to unbalanced budgets: “Would not the ordinary
taxpayer and the business man very soon begin to have a feeling of uneasiness and apprehension?
After all people will realise that the bill must be paid if not this year next year or the year after.
Uncertainty and apprehension about the future would very quickly cancel out any immediate
psychological benefit which the reduction of taxation by unbalancing the Budget would promote”
(1985, p. 88).
20
Neville Chamberlain (Chancellor of the Exchequer), quoted in Sabine (1970, p. 15).
It was also pointed out that “the requirement of a balanced budget
was and still is the simplest and clearest rule to impose ‘fiscal discipline’
and to hold government functions and expenditure to a minimum… Even
an avowedly counter cyclical policy is believed to give rise to an upward
trend in expenditures that might not otherwise occur. The expenditures
undertaken to counteract a depression are unlikely to be discounted in the
succeeding boom. If the boom is countered at all, the measures taken will
be credit restriction or increased taxation” (Smithies, 1960).
The obstacles posed by politics to a symmetric and timely reaction
of the budget to cyclical developments were stressed. Agreement over the
appropriate budgetary items to use may take too long; it may prove
difficult to reduce expenditures once they have been increased. Drees
(1955) and Steve (1972) provide early discussions of the relevance of the
balance of powers between the Parliament and the Government and of the
relationship between the Government and its Parliamentary majority:
budgetary rules cannot be evaluated per se but need to be set in the overall
institutional context.
Among the remedies suggested to the political problem described, an
enhanced reliance on automatic stabilisers and the so-called formula
flexibility were suggested. The latter consisted in the introduction of a
predetermined relationship between tax rates (or benefits levels) and the
level of economic activity
21
.
But support in favour of functional finance was strong. “Even if
stability in the budget has something to recommend it, stability in the
economy is surely better… Who makes the rule? Who decides when to
abide to it and when to countermand it? Furthermore, within the framework
of a political democracy, the case for taking stabilisation policy out of the
hands of politicians is an uneasy one: into whose hands shall it be
placed?… No budgetary rule can be provided with a solid intellectual
foundation. This will hardly be new to economists. The best that can be
said for rules is that some of them may be better than incompetently
__________
21
Biehl, in summarising several papers on fiscal policy issues, notes that “It is strange to see that,
e.g., the old-fashioned concept of the simple budget balance rule is still widely used in many
countries and that …. the full employment budget concept, the structural margin of fiscal impact
concept, and the concept of the cyclically adjusted neutral budget … are only known to a small
circle of specialists.” (1973, p. 6).
managed discretionary policy…” (Blinder and Solow, 1974, pp. 43 and
45). This view was broadly accepted in some public finance textbooks
22
.
Along the same lines, though less aggressively, Steve (1972) notes
that “budgetary policy cannot be reduced to simple rules, it should take
into account the overall effects of the budget on private demand
components and national income” (p. 170; our translation)
23
.
The stagflation in the 70s; the difficulties concerning the estimate of
the actual impact of budget changes on the economy; the risks of fine
tuning given the lags between the decision to change the budget and its
implementation; the development of theoretical models questioning the
possibility for the Government to influence the level of government
activity all contributed to a decline of interest in the theory of functional
finance.
Advocates of the balanced budget regained the fore. “The balanced-
budget principle played a crucial role in holding the pre-keynesian fiscal
constitution together, and constraining the otherwise inherent biases of that
system to over-expenditure and deficit finance. Once the balanced-budget
had been bowled over by the Keynesian revolution, those biases were
unleashed” (Buchanan, Wagner and Burton, 1978, p. 47)
24
.
Politicians praised again the virtues of balanced budgets: “At one
time, it was regarded as the hallmark of good government to maintain a
balanced-budget; to ensure that, in time of peace, Government spending
was fully financed by revenues from taxation, with no need for
Government borrowing. Over the years, this simple and beneficent rule
was increasingly disregarded … And I have balanced the budget” (Nigel
Lawson, Budget Statement, 1988).
__________
22
See, for instance, Johansen (1965), in which the use of budgetary items for stabilisation policy is
unquestioned, the focus of the analysis being on the choice of the most appropriate instruments.
23
Steve stresses that the budget balance cannot be considered in isolation: the level and composition
of public revenue and expenditure are extremely important.
24
Keynes’ own views about active fiscal policy were rather prudent. He stressed the need to control
inflation and retain appropriate market incentives. In evaluating the UK budget in 1940, he noted
“The importance of a war budget is not because it will ‘finance’ the war. The goods ordered by the
supply department will be financed anyway. Its importance is
: to prevent the social evils of
inflation now and later; to do this in a way which satisfies the popular sense of social justice; whilst
maintaining adequate incentives to work and economy”. (‘Notes on the Budget’, 21 September
1940, in Moggridge and Robinson, eds., 1971-89, vol. XXII, p. 218).
The recent policy debate has largely recognised that in normal
circumstances automatic stabilisers ought to be allowed to operate freely
25
.
On the contrary, discretionary fiscal action is generally considered
problematic in view of irreversibility and timing problems and of the
uncertainty about its effects
26
.
(08ILVFDOUXOHVDQHZDQVZHUWRDQROGTXHVWLRQ"
European Monetary Union represents a new historical development.
For the first time a number of sovereign countries adopt a common
currency while retaining independent fiscal policies. The need for fiscal
rules complementing monetary union has been at the core of the debate on
EMU since the early nineties
27
.
Some arguments were put forward against the introduction of fiscal
rules at the European level. It was noted that fiscal rules may have costs in
terms of stabilisation policies and may hamper the achievement of
allocative and distributive objectives. It was also noted that excessively
stringent rules may be counter-productive. If the Pact leads to an unduly
tight fiscal stance in one or more countries, pressure may mount on the
ECB to deliver a monetary offsetting
28
. Otherwise, the credibility of the
Pact may be endangered
29
.
However, the prevailing view in the policy debate was clearly in
favour of the introduction of formal rules. It was argued that procedural or
fiscal rules are necessary because the factors that in recent decades have
__________
25
The issue is extensively discussed in OECD (1999). OECD notes that “in the future governments
should guard against the asymmetric use of automatic stabilisers, although this obviously does not
preclude all discretionary action, particularly for structural reasons.” (p. 145).
26
See European Commission (2001), Kilpatrick (2001), Taylor (2000) and Wren-Lewis (2000).
27
For a review of the justifications put forward for the Pact and for an analysis of its potential
macroeconomic implications see European Commission (1997), Artis and Winkler (1997) and
Eichengreen and Wyplosz (1998).
28
Canzoneri and Diba (2001).
29
It was also noted that the multiplicity of fiscal authorities does not provide strong arguments in
favour of permanent constraints on the deficit as it may actually dilute the pressure on the central
bank. According to Canzoneri and Diba (2001), a more relevant reason to have fiscal rules is to
underpin the ‘functional’, as opposed to the ‘legal’, independence of the central bank. Without a
credible deficit criterion ensuring government fiscal solvency, the central bank would not be able
to keep control of the price level.
determined fiscal profligacy in several countries have not disappeared.
Moreover, the multinational dimension of EMU is likely to increase the
need for such rules.
Stark (2001) describes the genesis and the rationale of the Stability
and Growth Pact
30
; he stresses how in Europe, up to the early nineties, lax
fiscal policy “… occurred although it is indisputable that unsound fiscal
policy practices have adverse effects on price stability, growth and
employment: large deficits and large public debt place constraints on the
ability of a country … to act during different stages of the business
cycle…; the State’s absorption of resources which would otherwise have
found their way into private investments results in higher long term interest
rates …; … a stifling government debt ratio impair(s) the overall efficiency
of an economy and create(s) risks to price stability…; these problems are
especially pronounced in monetary union since … the policy of a single
country might have adverse consequences for all the other participating
countries”. These arguments combine the two main strands of opinion
about the budgetary balance: the one stressing the importance of
stabilisation policies and budgetary flexibility and the other maintaining
that unbalanced budgets imply distortions in the allocation of resources
31
.
It was also pointed out that, without strong rules, the legal
independence of the European Central Bank (ECB) may turn out to be an
empty shell because of pressure by high-debt countries for ex ante bail-out
(refraining from raising interest rates in conditions of inflationary tensions)
or ex post bail-out (debt relief through unanticipated inflation).
EMU can
induce unilateral fiscal expansions since governments may feel less
inclined to preserve fiscal rectitude, as they individually face a less steep
interest rate schedule in a monetary union than under flexible exchange
rates.
The debate on fiscal rules in EMU was grounded on the wider
debate that took place in the nineties about the role of fiscal institutions and
procedures in shaping budgetary outcomes
32
. While certain political
configurations, such as weak coalition governments, have been recognised
as conducive to budgetary misbehaviour or to hampering attempts to
__________
30
See also Costello (2001).
31
See Buti (1998).
32
See Kopits and Symansky (1998).
redress the budgetary situation
33
, inadequate budgetary institutions and
procedures may also contribute to a lack of fiscal discipline
34
.
In this context, institutional reforms in the fiscal domain have been
discussed and introduced in several countries. As noted by Beetsma (2001),
these reforms come in two main categories: (a) the introduction of
procedural rules conducive to a responsible fiscal behaviour and (b) the
introduction of a fiscal rule, i.e. a permanent constraint on domestic fiscal
policy in terms of an indicator of the overall fiscal performance (budget
balance, borrowing, debt, reserves) of central and/or local government.
In national experiences, both types of measures have proved to be
effective tools in containing political biases in fiscal policy-making and in
achieving and sustaining fiscal discipline. In a multinational context, the
adoption of harmonised tight budgetary procedures may lead to
fundamental problems from the point of view of national sovereignty
(Beetsma, 2001). Moreover, institutional reforms are more difficult to
monitor centrally, compared to numerical targets. The latter are also
simpler to evaluate and easier to grasp by public opinion and policy-
makers. In the end a clear consensus emerged about the introduction of
common numerical rules and an elaborated multilateral surveillance
mechanism
35
.
The fiscal framework of EMU was developed gradually. The Treaty
of Maastricht in 1992 set the fiscal criteria to be met for joining Monetary
Union. The Stability and Growth Pact (SGP), adopted by the European
Council in Amsterdam in June 1997, developed these criteria with a view
to permanently restraining deficit and debt levels while allowing room for
fiscal stabilisation. The Pact also strengthened the monitoring procedures
complementing the quantitative rules.
__________
33
See, e.g., Roubini and Sachs (1989), Alesina and Drazen (1991), Alt and Lowry (1994), Alesina
and Perotti (1995) and Balassone and Giordano (2001).
34
See, e.g., von Hagen and Harden (1994) and the essays in Strauch and von Hagen (2000).
35
See Buti and Sapir (1998) and Stark (2001).
$GHVFULSWLRQ
36
As we have anticipated in the introduction, EMU fiscal rules have
been designed with the goal to ensure that national policies keep a sound
fiscal stance while allowing sufficient margins for budgetary flexibility in
bad times.
The Treaty of Maastricht stated that budget deficits cannot be larger
than 3 per cent of GDP unless (a) under exceptional circumstances, such as
deep recessions, (b) they remain close to 3 per cent, (c) the excess only
lasts for a limited period of time
37
. If the deficit exceeds the 3 per cent limit
when the above three conditions are not met, the deficit is deemed
“excessive” and it sets off a procedure intended to force corrective
measures by the deviating country. If such measures are not taken the
Treaty foresees monetary sanctions which increase as situations of
excessive deficit persist
38
.
The Stability and Growth Pact specified what is meant by
“exceptional” and “limited period” in the clauses allowing a deficit greater
than 3 per cent of GDP not to be considered “excessive”. A recession is
considered exceptional if real GDP diminishes by 2 per cent. Milder
recession (where the reduction in real GDP is of at least 0.75 per cent) may
also be considered exceptional if, for example, they are abrupt. The excess
above 3 per cent must be reabsorbed as soon as the “exceptional
circumstances” allowing it are over.
The Pact also specified that each country should aim for a medium
term objective of a budgetary position “close to balance or in surplus”.
According to the guidelines provided by the European Council
39
, the choice
of the medium term target should take into account both the budgetary
risks of recessions and those linked to fluctuations of other economic
factors (e.g. interest rates). Countries with debt ratios above 60 per cent of
GDP should also take into account the need to decrease such ratio, at a
satisfactory pace, towards the threshold. Moreover, an increase in the debt
__________
36
A more detailed description of the rules is provided in Buti and Sapir (1997) and in Cabral (2001).
37
The three conditions make the 3 per cent threshold extremely binding (see Buti ., 1997).
38
See Cabral (2001) for a description.
39
Council Resolution on the Stabilty and Growth Pact, 17 June, 1997; Council Regulation
n. 1466/97, 7 July 1997; Council Declaration, 1 May 1998; Opinion of the Monetary Committee,
12 October 1998 as approved by the Council.
ratio during recessions should be avoided
40
. Finally, other risk factors, such
as the effects of demographic trends, ought to be taken into account
41
.
According to the European Council, compliance with the Pact
should be assessed considering the cyclical position of the economy. In
practice, EMU fiscal rules require that each member state choose a
budgetary target in cyclically adjusted terms and let automatic stabilisers or
discretionary action operate symmetrically around it. The lower this budget
balance with respect to the 3 per cent threshold, the wider the margins for
counter cyclical policy without running the risk of an excessive deficit.
Each member state must submit its budgetary targets officially in
multi-year budgetary documents (Stability Programmes); these documents
are updated annually and are subject to a review by the European
Commission aimed at assessing their consistency with EMU fiscal rules.
Overall, the approach taken by the EU can be characterised as less
flexible than the solutions adopted in some federally structured countries
42
:
a) the rules are defined on the basis of established numerical parameters;
b) H[ SRVW compliance with the parameters is required each year;
overshoots must be rapidly dealt with;
c) margins of flexibility are envisaged only in connection with exceptional
cyclical events (established H[DQWH as a decline in GDP) or in any case
events beyond the governments’ control;
d) no special provision is made for investment expenditure
43
;
e) monitoring procedures are envisaged, whereby peer pressure is
strengthened by the European Council’s power to make formal
representations to governments of the need to adopt corrective measures
during the year and by the application of pre-established monetary
sanctions.
__________
40
Art. 104C of the Treaty says that when the ratio is above 60 per cent of GDP it must “diminish
sufficiently” and approach 60 per cent “at a satisfactory pace”. If the ratio increases, the excessive
deficit procedure begins. It should be noted that, while the Treaty allows exceptions to the 3 per
cent deficit criterion, it does not for the criterion concerning the debt ratio See Balassone and
Monacelli (2000).
41
The choice of the medium term fiscal target is examined in Artis and Buti (2001, Dalgaard and de
Serres (2001) and Barrel and Dury (2001).
42
See Balassone and Franco (1999).
43
No distinction is made in the Treaty between current and capital expenditure for the purposes of
determining the deficit. The volume of capital expenditure is included only among the relevant
factors to be borne in mind when deciding whether there is excessive debt.
$QHZDQVZHUWRDQROGTXHVWLRQ"
With European Monetary Union for the first time the need for fiscal
rules arises in a multinational context. The review in the previous section
shows how the arrangements adopted are deeply embedded in the long-
lasting debate on budgetary rules. The novel features of EMU guided the
choice between alternative solutions and required the introduction of some
innovations.
The need to reconcile fiscal soundness and budgetary flexibility led
to combine different approaches:
a) setting a predetermined upper bound for the deficit is a new pragmatic
solution
44
;
b) balancing the budget over the cycle is a precept derived from the
keynesian approach. In 1951 a report by the United Nations,
commenting the 1937 Swedish reform, points out that “while counter –
cyclical budgeting introduced an element of flexibility in the fiscal
policy of Government, the concept of ‘financial soundness’ has been
retained” (p. 69);
c) prudence when fixing the average target to be achieved over the cycle
(“close to balance or in surplus”) has a classical flavour. In 1927
Bastable argued that “the safest rule for practice is that which lays down
the expedience of estimating for a moderate surplus, by which the
possibility of a deficit will be reduced to a minimum” (p. 611).
The stress on fiscal soundness motivates the rejection of a dual
budget approach and of any distinction between ordinary and extraordinary
finance. However, pragmatism called for the allowance of margins for
exceptional circumstances, this rests on the idea that “in some
circumstances, indeed, a balanced budget is a pedantic luxury, which a
community, hard pressed by sudden and exceptional misfortune, can ill
afford” Dalton (1934, p. 12).
A broadly balanced budget, like that required by the SGP, may
negatively affect the public investment level; this effect can be especially
relevant during the transition to the low debt levels consistent with the
chosen structural balance. The double burden determined by this transition
__________
44
The deficit ceiling, although arbitrary, is reminiscent of the results obtained by Domar (1944) in
the analysis of fiscal sustainability assuming a constant deficit. Perhaps conscious of the partial
equilibrium nature of Domar’s results, the introduction of a debt ceiling as well avoids
convergence at high levels of debt. See Balassone and Franco (2000a).
can be assimilated to that arising from the transition from a pay-as-you-go
to a funded pension system. However, besides the criticism to the double
budget system examined in the previous sections, in the context of EMU
the golden rule would be an obstacle to deficit and debt reduction. In
particular, given the ratio of public investment as a percentage of GDP, the
long-run equilibrium level of government debt could be very high,
especially in an environment of low inflation. This could imply that the
debt ratio would rise in low-debt countries, while in high-debt countries
there would be a very slow pace of debt re-absorption. The golden rule
would also meet with practical difficulties, such as the evaluation of
amortisation, and would make the multilateral surveillance process more
complex, by providing leeway for opportunistic behaviour. Governments
would have an incentive to classify current expenditure as capital
spending
45
.
The asymmetry in EMU between the monetary regime, with the
single currency and a single monetary authority, and the political
landscape, lacking an authority of federal rank, gave prominence to moral
hazard issues. It is probably at the roots of the rejection of both the dual
budget and the distinction between ordinary and extra-ordinary finance. It
motivated the adoption of a detailed multilateral surveillance procedure
and the introduction of a predetermined limit for the annual deficit in a
framework that envisages the targeting of a balanced budget over the
cycle
46
.
EMU may be termed a “radical federation”, where in the absence of
fiscal rules member states enjoy absolute autonomy in matters of public
expenditure and taxation and recourse to debt. In this context, the stability
of monetary and financial conditions represents a public good to which all
local governments contribute by maintaining sustainable budget positions.
There is an incentive for each local government to exploit the benefits
accruing from the discipline of others without itself complying with the
rules. This creates a double cost for the other entities: the free-rider’s
__________
45
See Balassone and Franco (2000b).
46
EMU fiscal rules are targeted at national governments while many EMU member states have a
federal or highly decentralised structure. A free riding problem can re-emerge at national level.
This problem is analysed in Balassone and Franco (1999).
excessive indebtedness can put pressure on interest rates to rise; it can also
result in bankruptcies requiring bail-outs
47
.
The need for monitoring was felt also in earlier days. For instance,
Durrell (1917) argues that: “the public and the Parliament should be
satisfied that … there is some authority which … will give timely warning
if that expenditure or those obligations are either outrunning
the revenue
provided for the year or engaging the nation too deeply in the future”
(p. 242). However the monitoring procedure adopted for EMU is novel
with respect to its scale, complexity and tightness.
Until now the chosen mix of approaches has been successful in
securing a reduction in budget deficits and debt across EMU member
states. It remains to be seen whether it will also be successful in
maintaining fiscal discipline once at regime. Unfavourable economic
developments will put to test EMU’s fiscal constitution and the issue of
legitimacy of rules in a democracy pointed out by Blinder and Solow may
come to the fore again.
Fiscal rules can be successfully implemented over a long period of
time only if public opinion considers them a valuable contribution to policy
making. In the words of Bastable (1927): “it but remains to again lay
emphasis on the fact that good finance cannot be attained without
intelligent care on the part of the citizens. The rules of budgetary
legislation are serviceable in keeping administration within limits; but
prudent expenditure, productive and equitable taxation, and due
equilibrium between income and outlay will only be found where
responsibility is enforced by the public opinion of an active and
enlightened community (p. 761).
__________
47
The risks clearly increase if member states are asymmetric in some relevant respect (e.g.
accumulated public debt). These considerations are likely to have motivated the inclusion of a
“rule” concerning not only deficits but also debt.
5()(5(1&(6
Alesina, A. and A. Drazen (1991), “Why are Stabilizations delayed?”,
$PHULFDQ(FRQRPLF5HYLHZ, vol. 82, pp. 1170-88.
Alesina, A. and R. Perotti (1995), “The Political economy of budget
Deficits”, ,0)6WDII3DSHUV, vol. 42, pp. 1-31.
Alt, J.E. and R.C. Lowry (1994), “Divided Government and Budget
Deficits: Evidence from the States”, $PHULFDQ 3ROLWLFDO 6FLHQFH
5HYLHZ, vol. 88, pp. 811-28.
Artis, M.J., and M. Buti (2001), “Setting Medium-Term Fiscal Targets in
EMU”, in A. Brunila, M. Buti e D. Franco (eds.) (2001).
Artis, M.J., and B. Winkler (1997), “The Stability Pact: safeguarding the
credibility of the European Central Bank”, CEPR Discussion Paper,
No. 1688, August.
Balassone, F. and D. Franco (1999), “Fiscal Federalism and the Stability
and Growth Pact: a Difficult Union”, -RXUQDORI3XEOLF)LQDQFHDQG
3XEOLF&KRLFH, vol. XVII, n. 2-3, pp. 137-66.
_______________________________ (2000a), “Assessing Fiscal
Sustainability: a Review of Methods with a view to EMU”, Banca
d’Italia.
_______________________________ (2000b), “Public Investment, the
Stability Pact and the Golden Rule”, )LVFDO 6WXGLHV, vol. 21, n. 2,
pp. 207-29.
Balassone, F. and R. Giordano (2001), “Budget Deficits and Coalition
Governments”, 3XEOLF&KRLFHvol. 106, pp. 327-49.
Balassone, F. and D. Monacelli (2000), “EMU Fiscal Rules: is there a
Gap?”, Temi di discussione, n. 375, Banca d’Italia.
Banca d’Italia (2000), )LVFDO 6XVWDLQDELOLW\, Proceedings of the Banca
d’Italia workshop on Public Finance held in Perugia, 20-22 January
2000, Rome.
Barrel, R. and K. Dury (2001), “Will the SGP Ever Be Breached?”, in A.
Brunila, M. Buti e D. Franco (eds.) (2001).
Bastable, C.F. (1927), 3XEOLF)LQDQFH, Macmillan, London.
Beetsma, R. (2001), “Does EMU Need a Stability Pact?”, in A. Brunila, M.
Buti e D. Franco (eds.) (2001).
Biehl, D. (1973), “Introduction of Round Table Discussion”, in H. Giersch,
)LVFDO3ROLF\DQG'HPDQG0DQDJHPHQW, J.C.B. Mohr, Tubingen.
Blinder, A. and R. M. Solow (1974), 7KH (FRQRPLFV RI 3XEOLF )LQDQFH,
The Brookings Institution, Washington.
Brunila, A., M. Buti e D. Franco (eds.) (2001), 7KH6WDELOLW\DQG*URZWK
3DFW±7KH)LVFDO$UFKLWHFWXUHRI(08, Palgrave, forthcoming.
Buchanan, J.M. (1997), “The balanced budget amendment: Clarifying the
arguments”, 3XEOLF&KRLFH, 90, pp. 117-38.
Buchanan, J.M., R.E. Wagner and J. Burton (1978), “The consequences of
Mr. Keynes”, Hobart Paper 78, Institute of Economic Affairs,
London.
Burkhead, J. (1954), “The balanced Budget”, 4XDUWHUO\ -RXUQDO RI
(FRQRPLFV, LXVIII, May, pp. 191-216.
Buti, M., D. Franco and H. Ongena (1997), “Budgetary Policies During
Recessions – Retrospective Application of the Stability and Growth
Pact to the Post-War Period”, 5HFKHUFKHV(FRQRPLTXHVGH/RXYDLQ,
Vol. 63, No. 4, pp. 321-66.
_________________________________ (1998), “Fiscal discipline and
flexibility in EMU: the implementation of the Stability and Growth
Pact”, 2[IRUGUHYLHZRI(FRQRPLF3ROLF\, Vol. 14, No. 3, pp. 81-97.
Brunila, A., M. Buti and D. Franco (eds.) (2001),7KH6WDELOLW\DQG*URZWK
3DFW±7KH)LVFDO$UFKLWHFWXUHRI(08, Palgrave, Basingstoke.
Buti, M. and A. Sapir (1998), (FRQRPLF3ROLF\LQ(08±$6WXG\E\WKH
(XURSHDQ&RPPLVVLRQ6HUYLFHV, Clarendon Press, Oxford.
Cabral, A.J. (2001), Main Aspects of the Working of the Stability and
Growth Pact, in A. Brunila, M. Buti and D. Franco (eds.)
(2001).Canzoneri, M.B. and B.T. Diba (2001), “The Stability and
Growth Pact: a Delicate Balance or an Albatross?”, in A. Brunila, M.
Buti e D. Franco (eds.) (2001).
Clarke, P. (1998), “Keynes, Buchanan and the Balanced Budget doctrine”,
in J. Maloney (1998), 'HEWDQG'HILFLWV, Cheltenham, Edward Elgar.
Cohn, G. (1895), 7KHVFLHQFHRI)LQDQFH, Chicago, University of Chicago
Press.
Colm, G. and P. Wagner (1963), “Some observation on the budget
concept”, 5HYLHZRI(FRQRPLF6WXGLHV, Vol. 45, pp. 122-6.
Costello, D. (2001), “The SGP: How Did We Get There?”, in A. Brunila,
M. Buti e D. Franco (eds.) (2001).
Dalsgaard, T. and A. de Serres (2001), “Estimating Prudent Budgetary
Margins”, in A. Brunila, M. Buti e D. Franco (eds.) (2001).
Dalton, H., T. Brinley, J.N. Reednam, T.J. Hughes and W.J. Leaning
(1934), 8QEDODQFHG %XGJHWV ± $ 6WXG\ RI WKH )LQDQFLDO &ULVLV LQ
)LIWHHQ&RXQWULHV, London, George Routledge & Sons.
De Viti de Marco, A. (1953), 3ULQFLSL GL HFRQRPLD ILQDQ]LDULD, Torino,
Einaudi.
Drees, W.Jr. (1955), 2Q WKH OHYHO RI *RYHUQPHQW H[SHQGLWXUH LQ WKH
1HWKHUODQGVDIWHUWKH:DU, Leiden.
Durrel, A.J.V. (1917), 7KHSULQFLSOHDQGSUDFWLFHRIWKHV\VWHPRIFRQWURO
RYHU3DUOLDPHQWDU\JUDQWV, London, Gieves Publishing House.
Einaudi, L. (1948), 3ULQFuSLGLVFLHQ]DGHOODILQDQ]D, Torino, Boringhieri.
Eichengreen, B. and C. Wyplosz (1998), “Stability Pact: more than a minor
nuisance?”, (FRQRPLF3ROLF\, April, pp. 67-104.
European Commission (2001), 3XEOLF)LQDQFHVLQ(08±, Brussels.
Goode, R. and E.A. Birnbaum (1955), *RYHUQPHQW &DSLWDO %XGJHWV,
Washington D.C., International Monetary Fund.
Hansen, A.A. (1941), )LVFDO 3ROLF\ DQG %XVLQHVV &\FOHV, New York,
Northon & C.
Hicks, U. K. (1953), “The Budget as an Instrument of Policy, 1837-1953”,
7KH7KUHH%DQNV5HYLHZ, pp. 16-34.
Johansen, L. (1965), 3XEOLFHFRQRPLFV, Amsterdam, North-Holland.
Keynes, J. M. (1971-1989), 7KH &ROOHFWHG :ULWLQJV RI -RKQ 0D\QDUG
.H\QHV, Macmillan, London.
Kilpatrick, A. (2001), “Transparent Frameworks, Fiscal Rules and Policy-
making under Uncertainty”, paper included in this volume.
Kopits, G. and S. Symansky (1998), “Fiscal Policy Rules”, IMF
Occasional Paper, No. 162.
Lerner, A.P. (1943), “Functional Finance and the Federal Debt”, 6RFLDO
5HVHDUFK, 10, pp. 38-51.
Lindbeck, A. (1968), “Theories and Problems in Swedish Economic Policy
in the Post-War Period”, $PHULFDQ (FRQRPLF 5HYLHZ, June
Supplement, pp. 1-80.
Lundberg, E. F. (1996), 7KH 'HYHORSPHQW RI 6ZHGLVK DQG .H\QHVLDQ
0DFURHFRQRPLF 7KHRU\ DQG LWV ,PSDFW RQ (FRQRPLF 3ROLF\,
Raffaele Mattioli Foundation, Cambridge University Press,
Cambridge.
Maloney, J. (1998), 'HEWDQG'HILFLWV, Cheltenham, Edward Elgar.
Middleton, R. (1985), 7RZDUGV WKH 0DQDJHG (FRQRP\ .H\QHV WKH
7UHDVXU\DQGWKHILVFDOSROLF\GHEDWHRIWKHV, London,
Methuen,
Moggridge, D. and A. Robinson (eds.) (1971-89), 7KH&ROOHFWHG:ULWLQJV
RI-RKQ0D\QDUG.H\QHV, London, Royal Economic Society.
Nitti, F. S. (1903), 3ULQFLSLGL6FLHQ]DGHOOH)LQDQ]H, Napoli, Luigi Pierro.
OECD (1999), (FRQRPLF2XWORRN, No. 66, Paris.
Pigou, A.C. (1929), $ 6WXG\ LQ 3XEOLF )LQDQFH, 2
nd
revised edition,
London, Macmillan.
Poterba, J.M. (1995), “Capital budgets, borrowing rules and state capital
spending”, -RXUQDORI3XEOLF(FRQRPLFV, Vol. 56, pp. 165-87.
Premchand, A. (1983), *RYHUQPHQW EXGJHWLQJ DQG H[SHQGLWXUH FRQWUROV
WKHRU\DQGSUDFWLFH, Washington, IMF.
Puviani, A. (1903), Teoria dell’illusione fiscale, reprinted by F. Volpi,
Milano, ISEDI, 1973.
Roubini, N. and J.D. Sachs (1989), “Political and Economic Determinants
of Budget Deficits in the Industrial Democracies”, (XURSHDQ
(FRQRPLF5HYLHZ, Vol. 33, pp. 903-38.
Sabine, B.E.V. (1970), %ULWLVK %XGJHWV LQ 3HDFH DQG :DU ,
London, Allen and Unwin.
Say J.B. (1853), $7UHDWLVHRQ3ROLWLFDO(FRQRP\, translation from the 4
th
French edition, Philadelphia, Lippincott, Grambo & Co.
Schumacher, E. F. (1944), “Public Finance – Its Relation to Full
Employment”, in F.A. Burchardt (ed.), 7KH (FRQRPLFV RI )XOO
(PSOR\PHQW, Basil Blackwell, Oxford.
Sraffa, P. (ed.) (1952-73), 7KH :RUNV DQG &RUUHVSRQGHQFH RI 'DYLG
5LFDUGR, Cambridge, Cambridge University Press.
Stark, J. (2001), “Where did the Pact come from? Genesis of a Pact”, in A.
Brunila, M. Buti e D. Franco (eds.) (2001).
Steve, S. (1972), /H]LRQLGLVFLHQ]DGHOOHILQDQ]H, Padova, Cedam.
Strauch, R. and J. von Hagen (eds.) (2000), ,QVWLWXWLRQV3ROLWLFVDQG)LVFDO
3ROLF\, Dordrecht, Kluwer Academic Publishers.
Taylor, J. B. (2000), “Reassessing Discretionary Fiscal Policy”, -RXUQDORI
(FRQRPLF3HUVSHFWLYHV, Vol. 14, No. 3, pp. 21-36.
United Nations (1951), %XGJHWDU\ 6WUXFWXUH DQG &ODVVLILFDWLRQ RI
*RYHUQPHQW$FFRXQWV, New York.
von Hagen, J. and I. Harden (1994), National Budget Processes and Fiscal
Performance, (XURSHDQ (FRQRP\ 5HSRUWV DQG 6WXGLHV, No. 3,
pp. 311-418.
Wren-Lewis, S. (2000), “The Limits to Discretionary Fiscal Stabilisation
Policy”, 2[IRUG 5HYLHZ RI (FRQRPLF 3ROLF\, Vol. 16, No. 4,
pp. 92-105.
),6&$/58/(686()8/32/,&<)5$0(:25.
25811(&(66$5<251$0(17"
*HRUJH.RSLWV
7KHEXGJHWVKRXOGEHEDODQFHG
WKHWUHDVXU\VKRXOGEHUHILOOHG
SXEOLFGHEWVKRXOGEHUHGXFHG«
M. T. CICERO (63 B.C.)
,QWURGXFWLRQ
Rules-based macroeconomic policies are in fashion. In the monetary
area, since the early nineties, an increasing number of countries have
adopted inflation targeting. The latter has displaced the targeting of
monetary aggregates, or of the exchange rate, as the rule of choice in
advanced economies. In the fiscal area, a parallel trend is under way, as
rules to eliminate or to contain budget deficits and to reduce the public debt
are gaining considerable popularity in various parts of the world
1
.
All these rules share at least one feature in common: they seek to
confer credibility to the conduct of macroeconomic policies by removing
discretionary intervention. Their goal is to achieve trust by guaranteeing
that fundamentals will remain predictable and robust regardless of the
government in charge. There are, however, obvious differences; for one
thing, credibility is not built at a uniform speed. Whereas an exchange rate
rule may provide immediate credibility following its introduction—and
equally, may be vulnerable to a sharp and sudden loss in credibility in the
event of an erosion in competitiveness or perception of misaligned
fundamentals—inflation targeting may take longer to establish credibility,
and balanced-budget rules usually become credible only after an extended
track record.
__________
* International Monetary Fund. Robert Hagemann, Geert Langenus, Ludger Schuknecht, and other
workshop participants provided useful comments. The author alone is responsible for the views
expressed, which do not necessarily reflect those of the International Monetary Fund.
1
Following the definition in Kopits and Symansky (1998), a fiscal policy rule is a permanent
constraint on fiscal policy, expressed in terms of a summary indicator of fiscal performance, such
as the government budget deficit, borrowing, debt, or a major component thereof.
Perhaps partly because of the long gestation period and partly
because of the particular experience of some countries, occasionally fiscal
rules are characterized as a fig leaf. According to this view, governments
either do not need rules since they apply discipline on a discretionary basis
anyway, or alternatively, if they adopt rules, they are not likely to follow
them seriously.
The purpose of this paper is to examine the merits of these
arguments, though without attempting to refute the truism that some
governments do in fact follow prudent countercyclical fiscal policy on a
discretionary basis, in a manner that is observationally-equivalent to a
well-designed set of fiscal rules. This point applies equally to prudent
discretionary monetary policy that obviates reliance on inflation targeting.
Specifically, an attempt is made here to find support for a rules-based fiscal
policy framework. As part of this endeavour, much of the paper is
devoted—drawing on international experience—to a discussion of the
attributes that such a framework must have in order to maximize its
usefulness.
In weighing the pros and cons of fiscal policy rules, the paper
ventures beyond the mainly Eurocentric focus of this workshop and takes a
broader view, since much of the recent popularity of fiscal rules can be
found in emerging market economies, as they seek to establish credibility
in financial markets. Also, wherever relevant, the discussion is cast in the
broader setting of rules-based macroeconomic policies, that is, including
references to monetary rules as well.
(YROXWLRQRIILVFDOSROLF\UXOHV
The virtue of fiscal discipline has been heralded for a long time—
during at least two millennia, as attested by the opening citation. However,
occasionally, departures from discipline have been justified politically and
conferred analytical respectability—most notably, in the aftermath of the
Great Depression. In many advanced economies, discretionary demand
management, instead of remaining broadly neutral or of offsetting the
effect of the cycle, has led to a nearly continuous increase in government
spending that outpaced revenue capacity. In other words, fiscal policy
exhibited a procyclical stance and a deficit bias
2
. A similar process can be
detected in some less developed countries, where the deficit bias emerged
with the pursuit of developmental objectives, against the background of
swings in capital flows and primary commodity prices
3
. Starting in
the 1980s, recognition of this bias and its contribution to public
indebtedness, as well as of its potential adverse repercussions on private
investment, prompted some governments to introduce medium-term fiscal
consolidation programs to restore macroeconomic stability and fiscal
sustainability. More recently, this was increasingly followed by a shift to
fiscal policy rules.
Formal attempts at casting the virtue of fiscal discipline into
permanent rules, through constitutional or legal provisions, at various
levels of government, span over a century and a half. During this period,
we can identify three fairly distinct waves. In the first wave, subnational
governments in some federal systems adopted autonomously the golden
rule. Under this rule, most states in the U.S. since the mid-19
th
century and
several cantons in Switzerland since the 1920s assumed an obligation to
maintain current budget balance. In essence, their goal was to gain access
to market-based financing of capital expenditure, absent a precedent of
bailouts by the national government.
In the second wave, after World War II, several industrial countries
(Germany, Italy, Japan, Netherlands) introduced balanced-budget rules that
underpinned their stabilization programs, following monetary reform. Most
of these were of the golden rule type. Other rules, limiting or prohibiting
the financing of budget deficits from specified domestic sources (mainly
central banks), were assumed in the 1960s, including in some developing
countries (Indonesia, CFA franc zone). Under all these rules, considerable
scope remained for creative accounting and other nontransparent practices
that could undermine compliance.
The current wave, starting with New Zealand’s Fiscal Responsibility
Act of 1994—shortly after the pioneering introduction of inflation
targeting in that country—has seen an increasing number of industrial and
emerging market economies introduce fiscal rules (Table 1). These rules
__________
2
For evidence of a procyclical fiscal stance since the 1970s in the euro area, see European
Commission (2000). Similarly, Taylor (2000) found that during much of the last four decades the
U.S. has followed a procyclical (or at best ineffective) discretionary fiscal policy.
3
Procyclical fiscal policy has been documented for Latin America in 1970-95, in Gavin and others
(1996).
encompass a range of balanced-budget obligations, debt limits, and
expenditure limits, at various levels of government. In contrast to the
previous waves, a common denominator of the recent rules is that they are
supported by more or less strict transparency standards consisting of
generally accepted accounting conventions, timely and regular reporting
requirements, and a medium-term macro-budgetary framework. Generally,
all these elements are enshrined in broad legislation or international treaty,
with carefully spelled out accountability obligations. By analogy, inflation
targeting is usually set in an institutional context characterized by
transparency, central bank independence, and accountability.
Present fiscal policy rules are fairly diverse in both design and
implementation. Whereas Anglo-Saxon countries place primary emphasis
on transparency (Australia, Canadian provinces, New Zealand, United
Kingdom), in continental Europe (EMU Stability and Growth Pact,
Switzerland’s proposal) and emerging market economies (Argentina,
Brazil, Colombia, Peru, India’s proposal) rely far more on a set of
numerical reference values (targets, limits) on performance indicators. In
federal systems with strong subnational autonomy, the rules are assumed
only by the central government (Argentina, India’s proposal); in other
federal systems with concern about potential bailouts and external
spillovers of fiscal misbehaviour across jurisdictions, the rules are imposed
on each government level in a coordinated fashion (Brazil, EMU).
Most rules allow for escape clauses in the event of unforeseen
exogenous shocks. Objectively determined escape clauses may take
various forms: simply a medium-term target balance or surplus, without
explicit margins around it (New Zealand); explicit margins around a target
balanced-budget or surplus requirement, calibrated on cyclical deviations
in output growth (EMU, Swiss proposal); or alternatively, operation of a
contingency fund (Argentina, Peru). In other cases, the escape clause is to
be invoked in a discretionary manner, in the event of an international crisis,
a national calamity, or other loosely defined shocks (Brazil, India’s
proposal, U.S. proposal).
An independent arbitration authority is clearly defined in some
countries (Brazil, EMU), while at most a monitoring agency has been
appointed in others (Argentina). In some instances, the government is
subject to financial or judicial sanctions for noncompliance with the rules
(Brazil, Canadian provinces, EMU, CFA franc zone). For the most part, the
authorities are exposed to loss of reputation upon noncompliance.
8QQHFHVVDU\RUQDPHQW"
Skepticism about the usefulness or effectiveness of fiscal rules is
grounded on several arguments, ranging from theoretical to practical ones.
From a theoretical perspective, neither traditional macroeconomic analysis,
nor any principles of public finance are predicated on a rules-based fiscal
policy. Indeed, a discretionary approach has been widely viewed as
instrumental for the achievement of conventional fiscal goals or
functions—namely, stabilization, distributional fairness, and allocative
efficiency. Likewise, monetary rules were not deemed to be superior to
discretionary monetary policy. In all, the main virtue of discretionary
demand management was that it afforded short-run flexibility to offset
large exogenous disturbances, especially those that could lead to a
prolonged and significant unemployment. In the postwar period very few
authors (Friedman, 1948) questioned this conventional wisdom
4
.
Another source of skepticism (or at least agnosticism) about rules is
that a government can commit credibly to fiscal discipline without any
permanent rules. This observation finds support in a few practical
illustrations. In this regard, U.S. fiscal and monetary discipline in recent
years has been viewed as an example of prudent discretionary
policymaking. Since the mid-1990s, high growth and low inflation,
accompanied by budget surpluses, can be taken as evidence of the
redundance of formal balanced-budget requirements and inflation
targeting
5
. In a similar vein, it has been argued that rules do not really
matter in the conduct of fiscal policy, and further, that policy credibility is
formed regardless of actual adherence to rules. The example of Germany
suggests that public confidence (at home or abroad) in policy management
has not been altered by the authorities’ more than occasional failure to
meet, since the 1970s, the golden rule or the M3 target. Likewise, in Japan,
suspension of the rule since 1975 has had no effect in this regard. An
alternative interpretation of these examples is that a reputation of prudent
macroeconomic management acquired through a prolonged period of good
__________
4
In a departure from the mainstream, Friedman (1948) recommended a cyclically-adjusted
balanced-budget rule as a long-run policy guideline, with the purpose of eliminating the
uncertainty and undesirable political implications of discretionary action, including a procyclical
fiscal stance. However, he qualified the proposal with the caveat that such a rule may be
insufficient to offset stubborn and strong cyclical fluctuations, which would warrant discretionary
intervention.
5
Incidentally, the extended application of the Budget Enforcement Act of 1990 can be characterized
as a procedural rule—to support the discretionary approach—rather than a policy rule.
performance, often in a rules-based context, obviates further adherence to
fiscal rules. Of course, conversely, absence of such reputation would argue
for the adoption of rules
6
.
Occasionally, rules are criticized for imposing unnecessary
bureaucratic requirements. Why not, instead, just let the market forces
exert discipline on misbehaving governments
7
? There is, however,
considerable evidence that financial markets—as typified by credit rating
agencies—tend to react with considerable lag to either a deterioration or an
improvement in fundamentals. It can be argued that, if well designed, fiscal
rules can mimic market pressures in a more rapid and efficient manner,
and, above all, without the heavy penalty (namely, sudden capital outflow,
high risk premium) imposed by perceptions of fiscal misbehaviour.
A much more common objection to fiscal rules is that, by their very
nature, they invite abuse and are doomed to be ineffective. Typically, they
induce nontransparent behaviour, largely through creative accounting
practices to circumvent the rules. Perhaps the most graphic illustration is
close to home, namely, the rule embodied in Article 81 of the Italian
Constitution, which is wide open to interpretation to the point of rendering
it meaningless
8
. Also, creative accounting and other forms of opaque
application of fiscal rules have been found, for example, in some
U.S. states and the Netherlands
9
. A similar criticism has been leveled at
medium-term fiscal adjustment plans adopted in a number of industrial
countries in the 1980s, most notably, under the Gramm-Rudman-Hollings
Act in the United States. These cases simply demonstrate that fiscal targets,
whether set in the context of permanent rules or of medium-term
adjustment programs (including Fund-supported programs) heighten the
__________
6
Drazen (2000) indicates that policy credibility is built through two alternative routes: reputation or
rules.
7
Most recently, this criticism was leveled at Brazil’s rules by the outgoing Finance Secretary of the
State of Sao Paulo.
8
According to Article 81, no new taxes or expenditures can be introduced through the annual budget
law, and legislation on new or increased outlays must indicate their sources of financing. It is
probably against the background of this experience that Italian economists (Alberto Alesina,
Franco Reviglio, Vito Tanzi, to name a few) tend to be particularly critical of fiscal rules.
9
Nontransparent application of fiscal rules has been documented in Suits and Fisher (1985) for the
states of Michigan and New York, and in Wellink (1996) for the Netherlands.
temptation to resort to nontransparent practices, much like with monetary
targeting
10
.
However, far from being an inherent flaw that invalidates rules, the
proliferation of loopholes must be recognized and dealt with through
appropriate design and implementation—as discussed below. These
episodes underscore the overarching importance of strict transparency
requirements not only for discretionary policymaking, but more important,
as an integral component of any set of rules. It is for this reason, that,
unlike in the previous historical waves, practically all recently established
rules include standards of transparency and other features—often under so-
called fiscal responsibility legislation—intended to strengthen the
effectiveness of the rules.
8VHIXOSROLF\IUDPHZRUN"
The strongest case for rules is rooted in political economy. In a
democratic society, rules are necessary to restrain politically rational
policymakers who conduct discretionary policies with a deficit bias when
facing an electorate that fails to understand, or is indifferent to, the
intertemporal budget constraint (Buchanan and Wagner, 1977). More
formally, it has been demonstrated that rules-based policies are superior to
a discretionary approach, since the latter is time inconsistent, given a
democratic government’s tendency to abandon previously announced
policy commitments (Kydland and Prescott, 1977).
From a somewhat different angle, it can be shown that rational
governments are prone to use suboptimal discretionary policies to enhance
their chances for re-election, rather than maximize social welfare, by
exploiting an information advantage vis-à-vis the electorate (Cukierman
and Meltzer, 1986). This observation underscores that rules can prevent
such an outcome if they are accompanied by transparency requirements to
reduce or eliminate information asymmetry.
Following these arguments, the primary usefulness of a well-
designed and appropriately implemented set of permanent fiscal rules, that
prevents a deficit bias, consists of establishing a depoliticized framework
__________
10
Charles Goodhart’s observation (made in connection with targeting monetary aggregates) that a
statistical or accounting measure ceases to be a reliable performance indicator once it is declared an
official target provides a strong argument for transparency in the application of fiscal rules.
for fiscal policy—much like the depoliticization of monetary policy under
inflation targeting. Accordingly, with widely available information about
macroeconomic developments and prospects, only the relative spending
priorities and the tax structure that are subject to legislative and public
debate, but not the budget balance or the level of expenditures which are
predetermined by rules.
At a practical level, the above case for rules-FXP-transparency is
probably strongest for emerging market economies; at the other end of the
spectrum, the argument tends to vanish for advanced economies with a
solid reputation of fiscal rectitude, as noted above. This spectrum can be
viewed in a dynamic sense: over a prolonged period of time, as a country
successfully applies fiscal rules, it gradually gains an ever stronger
reputation that eventually permits abandonment of the rules, without loss
of credibility (Germany and Japan).
However, even governments enjoying a solid reputation may want to
refrain from pursuing discretionary countercyclical fiscal policy in view of
the associated implementation lags, irreversibility, and political constraints.
Accumulated evidence on the ineffectiveness of discretionary activism
suggests adoption of a simple budget balance rule that allows for the
operation of automatic stabilizers; discretionary action should be applied,
though much less frequently, only for longer-term structural objectives—
such as social security reform or tax reform, aimed at fiscal sustainability,
intergenerational equity, or efficiency (Taylor, 2000).
For emerging market economies, including for those that have
reached the last stage of post-socialist transition, the potential usefulnesss
of rules cannot be overstated. As they open up, while experiencing fiscal
stress, these economies are exposed to considerable and rapid shifts in
capital movements that can result in a currency crisis. This is reflected in a
relatively high risk premium that raises the cost of capital, with a
depressing effect on much-needed investment. Thus, as part of an effort to
reduce vulnerability to speculative attacks (including from contagion or
other exogenous shocks) and to promote stability and growth, these
countries are well advised in considering the adoption of fiscal rules
(Kopits, 2000).
For similar reasons, in a federal system, rules can be usefully applied
at the subnational level of government. Largely because of the need to
build good reputation in financial markets, subnational governments may
choose to adopt fiscal rules in the absence of a potential bailout (in the
U.S., Switzerland, Canada). Otherwise, these governments may be subject
to centrally imposed rules to prevent moral hazard (as in Brazil, or in EU
member countries under EMU), particularly given a relatively small
portion of fiscal activity under central (or supranational) control. In this
case, adherence to rules generally helps lower the individual government’s
default risk premium, as well as the country risk premium faced by the
entire federal (or supranational) government
11
.
In what follows, we shall examine the design and implementation
characteristics that make fiscal rules a useful policy framework—thereby
countering the above argument that fiscal rules are inherently flawed and
thus doomed to failure. These attributes involve both technical
infrastructure and institutional infrastructure. On the technical side, fiscal
rules must be designed taking into account, for example, the interaction
between the public sector and the economy, including estimates of the
response of the fiscal position to exogenous shocks. In addition, the rules
must be based on a set of institutional building blocks, including
transparency standards, an arbitration authority to oversee compliance, and
sanctions for noncompliance.
In the monetary area, these considerations apply in an analogous
manner to inflation targeting Technically, inflation targeting must be
supported by sufficient information on the transmission mechanism and by
reliable inflation forecasts. On the institutional side, it is necessary to
establish an independent central bank that operates transparently, including
through publication of inflation reports, and is accountable to the
government and the public at large.
'HVLJQLQJDXVHIXOIUDPHZRUN
5XOHVDWWKHQDWLRQDOOHYHO
In countries that face a large public debt burden, a major objective of
fiscal policy rules is to reduce the public debt ratio and then to stabilize it
at a prudent level. Whereas in Brazil and New Zealand the government of
__________
11
The decline in interest rates experienced by highly indebted EMU participant countries cannot,
however, be ascribed unambiguously to the adoption of the rules. Given the identification problem
arising from having simultaneously joined the currency union and adopted the fiscal rules, it is
difficult to determine what proportion of the interest rate decline in Italy represents a fall in default
risk or the disappearance of the currency risk.
the day is required to set a target or ceiling for the debt ratio, under EMU,
governments are obliged to reduce the gross debt ratio to 60 percent of
GDP.
In general, the outstanding liabilities of the consolidated public
sector are seen as a summary measure (among many others) of a country’s
vulnerability. Financial markets tend to assess default risk on the
outstanding debt of the public sector as a whole, rather than just the central
government, given the implicit guarantee provided by the central
government to the rest of the public sector
12
. Also, a measure of gross debt,
rather than net debt or net worth, is preferred since the marketability and
valuation of government assets—with the notable exception of foreign
exchange reserves—usually are open to question.
While a medium-term limit on the gross debt-to-GDP ratio can be
interpreted as a broad gauge of fiscal rectitude and sustainability, year-to-
year debt ceilings are less likely to be credible or operationally effective.
Indeed, as measures of public indebtedness (especially as a proportion of
GDP) may be exposed to valuation changes and other factors beyond the
control of the authorities, they are difficult to treat as an annual operational
target.
A more common rule is defined in reference to a comprehensive
flow indicator of fiscal performance, such as the budget balance or
government borrowing. To enhance its effectiveness, the indicator needs to
be operationally simple, flexible, and growth-oriented
with obvious
tradeoffs among these criteria. 2SHUDWLRQDO VLPSOLFLW\ requires that the
indicator, while possibly consistent with a medium-term debt limit, be
amenable to monitoring and control during budget execution. This criterion
is met by the overall balanced-budget requirement in Argentina and by the
overall deficit limit in Peru. It can be argued that, operationally, even more
useful would be an obligation to maintain a minimum primary surplus
(excluding interest expenses, beyond the immediate control of the
authorities) that could be calibrated to the desired reduction in the debt
ratio
13
.
__________
12
Again, possible exceptions are countries without the precedent of bailouts of defaulting subnational
governments by the central government. In such cases, credit rating agencies assess risk separately
for each borrowing government jurisdiction.
13
See the relationship between a primary surplus rule and the debt target in the Annex.
Ideally, the IOH[LELOLW\ criterion can be realized with carefully
designed escape clauses that are triggered objectively by exogenous
shocks. The preferred option—along the lines envisaged initially by
Friedman—would be based on an indicator of cyclically-adjusted balance
that accommodates the effect of automatic stabilizers around a trend GDP
growth rate, DOORZLQJ for overall budget deficits during below-trend
growth, butUHTXLULQJ surpluses during above-trend growth, as had been
proposed for the federal government in Switzerland
14
. A more practical
solution (albeit not necessarily neutral with respect to the cycle) consists of
targeting overall balance or surplus over the cycle but subject to a preset
deficit limit, as required under the Stability and Growth Pact, sufficient to
accommodate the impact of a significant recession
15
. An alternative
approach is to require balance or surplus over the cycle, as in New
Zealand, without any limits, thus allowing not only for the operation of
automatic stabilizers, but also for discretionary countercyclical action. An
advantage of this approach is that it provides flexibility even with low
output elasticities of tax revenue.
Another escape clause for mitigating the effect of exogenous shocks
consists of accumulation (drawdown) of reserves in (from) a contingency
fund in the event of an upturn (downturn) in activity, as envisaged under
the fiscal rules recently promulgated in Argentina and Peru, much like with
the “rainy day funds” in the case of some state governments in the United
States. The least desirable approach would be simply to leave to the
authorities discretion to interpret events, such as a national calamity or a
threat to national security (as proposed, for instance, in India), for invoking
the escape clause.
A JURZWKRULHQWHGindicator seeks to avoid placing an undue burden
of compliance with the rule on cuts in government investment spending—a
damaging outcome, given generally high expected social rates of return on
infrastructure projects. This can be accomplished by requiring current
__________
14
Assuming an annual trend growth rate of 2 percent, the Swiss proposal would require the
government to generate excess revenue when actual GDP growth exceeds 1.8 percent and would
allow for excess expenditure when it declines below 0.5 percent a year. The excess revenue
(expenditure) was specified in reference to increments in excess (shortfall) in GDP growth.
Although not strictly speaking a permanent rule, Chile’s structural surplus target for the central
government is a comparable approach.
15
On the basis of historically estimated fiscal parameters, a 1 percent decline in output is estimated to
result, on average, in a 0.6 percent budget deficit in the EU. Therefore, the 3 percent deficit
reference value under EMU is compatible with a 5 percent below-trend deviation in GDP—which
should be quite sufficient for countries on a 2 to 3 percent trend growth path.
balance, under the so-called golden rule, applied in the majority of
U.S. states, and in Germany and Brazil at both the federal and state levels.
More appropriately, consistent with the golden rule, New Zealand
prescribes observance of operating balance
16
.
Whereas cash-based current balance permits borrowing to finance
gross investment expenditures, accrual-based operating balance allows
borrowing only for investment net of depreciation. An additional rule that
can buttress the growth objective is a limit on the proportion of a major
component of current expenditures in total expenditures. Along these lines,
Brazil has introduced a limit on the share of the government wage bill
(including government pension payments).
However, caution is needed to prevent the leakages (by financing
camouflaged current expenditure) associated with the golden rule, which
have been so prevalent in Germany and in some U.S. states. Specifically, it
would be necessary to follow a transparent and unambiguous, yet
operationally sensible, definition of what constitutes capital expenditure.
The operating balance requirement, followed in New Zealand, obviates
measurement refinements and has a smoothing effect on the fiscal
outcome.
Subject to these caveats, fiscal rules should preclude
overidentification. Accordingly, the balanced budget rule should operate
when the limit on the debt ratio has been met. Otherwise, as suggested
earlier, in periods when the actual debt ratio exceeds the limit, the
government would be expected to generate a primary surplus consistent
with convergence to the prescribed debt ratio limit.
5XOHVDWWKHVXEQDWLRQDOOHYHO
A key issue to be addressed in a decentralized system is the
application of fiscal rules at subnational levels of government. The case for
subnational rules is particularly strong when a country, such as Argentina
or Brazil, is confronted with a major fiscal adjustment task that cannot be
met by the central government alone. In fact, the smaller the share of the
__________
16
From the perspective of intergenerational equity, the golden rule should be defined in terms of the
operating balance, so that taxpayers in each time period pay for the costs (depreciation plus
interest), of existing capital assets from which they derive benefits in that period; see Robinson
(1998). Furthermore, these costs could be reduced by capital gains accrued in that period.
central or supranational government, as in the EU, greater is the need for
applying subnational rules to counter the moral hazard that may arise
among subnational governments (or national governments in the EU) to
incur fiscal imbalances with repercussions on the borrowing costs of the
rest of the federal system. The fundamental principle underlying these
arguments is that rules—and more broadly, fiscal responsibility
legislation—need to be imposed on the corresponding government level,
that is, the ORFXV RI DFFRXQWDELOLW\ for policymaking. Stated differently,
whereas in a unitary system policy formulation and decisions take place
only at the national or central level, in a federal system they are dispersed
among the national and subnational levels.
An additional critical condition for well-functioning subnational
rules is that underlying vertical (regional) imbalances be broadly offset
through an adequate mechanism of intergovernmental compensatory
transfers. These transfers should be determined, if possible, by objective
indicators of expenditure needs and taxing capacity in each subnational
jurisdiction
17
, consistent with a clear assignment of spending functions and
revenue sources. Budget or debt rules should be viewed as complementary
rather than as substitutes for such a mechanism.
In federal systems, there are two basic approaches to fiscal
responsibility, and in particular, to designing fiscal rules
18
. Although
usually only one approach is present, in a few countries (Germany) both
are followed. Under the DXWRQRPRXV DSSURDFK, the initiative for
establishing rules arises from individual subnational governments.
Following this bottom-up approach, in Canada, Switzerland and the United
States, many subnational governments have adopted the golden rule,
enforced with varying degrees of stringency
19
, while others retained
discretionary policymaking. By and large, in these countries, subnational
__________
17
The determination of intergovernmental transfers in several Scandinavian countries can be
regarded as exemplary in this respect; see Rattso (1998).
18
For a review of the international experience with fiscal policy rules at subnational levels of
government under each approach, and lessons for Argentina and Brazil, see Kopits, Jiménez, and
Manoel (2000).
19
For example, while in some U.S. states the golden rule is applied only , in others it is
applied on an
basis as well; a number of states do not permit carryover of unspent
appropriations from year to year; some states have contingency funds; and the scope for creative
accounting varies among states. In Canada, there are differences in the design of rules across
provinces, including in the nature of the penalties for noncompliance; for example, in one province,
the penalty consists of salary cuts for cabinet members unless the overrun in the budget deficit is
caused by exogenous shocks.
governments have direct access to financial markets to meet their
borrowing requirements, and there is rarely a precedent of bailouts of
insolvent subnational governments by the national government; hence,
their desire to maintain a favourable credit rating in the markets. More
recently, in deference to subnational autonomy, Argentina sought to follow
this approach—notwithstanding a trail of bailout operations—by adopting
rules at the federal level and inviting provinces to follow suit on a
voluntary basis.
By contrast, under the FRRUGLQDWHG DSSURDFK, all subnational
governments are subject to uniform rules to ensure a degree of fiscal
discipline under the surveillance of a central authority. For the most part,
this top-down approach is introduced against the background of past
bailouts or under some form of implicit or explicit guarantees to rescue
subnational governments in distress. Coordination also becomes necessary
in federations (or confederations) where lower levels of government are
responsible for the bulk of fiscal activity, with considerable potential
spillovers from the misbehaviour of one government on the risk premium
of another government within the federal system. Perhaps the strongest
argument for this approach is the need to bring about a lasting fiscal
adjustment encompassing the entire general government or consolidated
public sector, in the face of a possible sustainability problem with likely
repercussions on countrywide risk premium.
An early example of this approach—that resulted from a major
bailout episode—was the informal agreement among the Australian states,
later formalized under the authority of the Loan Council, setting borrowing
limits on the states
20
. Other examples where lower-level governments are
subject to statutory debt limits include Brazil, Colombia, EU members, and
CFA franc zone members(the limit being set as a proportion of government
revenue, or GDP, of the jurisdiction). In an interesting variant of this
practice, in Brazil, consistent with an overall target debt-GDP ratio (set by
the Senate, upon recommendation of the President) for the public sector as
a whole, each level of government is assigned a uniform limit for its debt-
revenue ratio, implying a fiscal adjustment—to be completed over a
specified number of years—for state governments whose ratio exceeds the
limit set for all state governments. In addition, all Brazilian states and
__________
20
This arrangement, which operated in different forms during 1923-92, was increasingly
circumvented through ingenious financing techniques (sales, leaseback operations, etc.); hence, it
was finally replaced with transparent reporting requirements on the states’ fiscal policy intentions
and performance—in line with the requirements of the Charter of Budget Honesty Act of 1998.
German /lQGHU are required to follow the golden rule, and EU member
countries (as well as implicitly lower-level governments within the EU) are
committed to maintain overall balance, subject to the deficit limit
21
.
An important distinction between the two approaches—arising
mainly from the implicit or explicit bailout provision, present in the
coordinated approach
22
—is that, while under the autonomous approach
each subnational government seeks to gain credibility for its own fiscal
policy, under the coordinated approach the goal is to establish collective
credibility for overall macroeconomic policy—that is, also the monetary
policy stance of the federation. Given the diffusion of effort among
subnational jurisdictions to achieve collective credibility, as opposed to
individual credibility for each jurisdiction, the incentive for free-rider
behaviour by circumventing the rules has been far stronger under the
coordinated approach. Therefore, under the latter, there is greater need to
introduce sanctions for noncompliance (see below) and to create a
mechanism for enforcing corrective action by the delinquent government—
in exchange for assistance or waiver of fines by the central or supranational
authority—as envisaged in Brazil, Colombia, the EU, or the CFA franc
zone.
The flexibility criterion is also relevant for the design of subnational
fiscal rules, especially as regards the treatment of asymmetric shocks.
Shocks that are concentrated in certain regions could be compensated with
cyclically-adjusted rules and contingency funds at the subnational level, or
with intergovernmental transfers. For instance, within the EU, besides the
provision of waivers from the deficit reference value in case of a
significant recession, Structural and Cohesion Funds are made available to
member contries on the basis of regional need as well as vulnerability to
shocks.
__________
21
Within the EU, federal governments (Austria, Germany, Italy, Spain) have endeavored to design
derivative EMU rules for subnational levels of government. For an analysis of the Italian case, see
Balassone and Franco (1999).
22
See the discussion of the relevance of an implicit bailout clause with regard to the EMU fiscal
reference values, in Eichengreen and von Hagen (1995) and McKinnon (1996).
,PSOHPHQWLQJDXVHIXOIUDPHZRUN
7UDQVSDUHQF\
It is widely recognized that transparency is conducive to successful
fiscal policy
23
whether in the context of rules-based or of discretionary
policymaking
24
. But, as indicated, the need for transparency is exacerbated
in the application of fiscal policy rules in the face of mounting pressures
for engaging in creative accounting and operating procedures to comply
formally, but not in fact, with preset performance indicators.
Specifically, the usefulness of fiscal rules hinges onWUDQVSDUHQF\LQ
LQVWLWXWLRQDO VWUXFWXUH DQG IXQFWLRQV, that is, in the relations within the
public sector, as well as the relations between the government and private
sector entities. Transparency serves to contain or reduce quasi-fiscal
activities through covert subsidies at below-cost pricing or government
guarantees
often used as a substitute for explicit budgetary operations.
Equally important is WUDQVSDUHQF\ LQ ILVFDO UHSRUWLQJ through
comprehensive, timely, frequent, and detailed government reporting (based
on appropriate accounting standards), as mandated for compliance with
fiscal rules in New Zealand, Brazil, and the EU
25
.
%XGJHWSURFHVV
Over the past decade, an increasing number of countries, especially
developed ones, have been preparing a multiyear macro-budgetary
framework as part of the annual budget exercise. Although procedures (in
terms of the degree of detail, realism of underlying macroeconomic
forecasts and policy assumptions, etc.) tend to vary among countries, such
__________
23
See Kopits and Craig (1998), which forms the basis of the International Monetary Fund’s Code of
Good Practices in Fiscal Transparency.
24
Much like in New Zealand (though without a balanced-budget rule) Australia’s Charter of Budget
Honesty Act 1998 requires the national authorities to publish: fiscal strategy statements; annual
reports on budget and fiscal outlook (including mid-year reports), and final budget outcome;
intergenerational reports; and pre-election economic and fiscal outlook reports.
25
This is illustrated, for example, by the requirements under EMU to follow accrual-based
accounting; to classify privatization receipts as financing in the calculation of the budget balance;
to measure debt on a gross basis; and to expand coverage to the general government.
a medium-term process is an important prerequisite for a well-informed
policy debate
26
.
In particular, a rolling multiyear macro-budgetary process is an
essential ingredient of effective fiscal rules, since it alerts the authorities
and financial markets alike as to the policy adjustments or reform measures
that may be necessary for compliance with the rule. More generally, it
disciplines policymakers and ensures that they are accountable for adhering
to budget targets. For these reasons, the preparation of medium-term
budget forecasts is an integral part of fiscal policy rules and of associated
reporting requirements in Argentina, Brazil, New Zealand, Peru, and EU
members.
In addition, to ensure compliance in the near term, it is useful to
establish a mechanism to enforce a mid-course correction for unanticipated
deviations from target, unless they stem from cyclical fluctuations covered
by escape clauses or are offset with recourse to a contingency fund.
Revenue shortfalls or expenditure overruns that are generated by the
executive or legislative branches would have to be met with automatic
measures specified to offset the budgetary effect of the deviation (Brazil).
6WDWXWHVVXUYHLOODQFHDQGVDQFWLRQV
Largely dictated by judicial precedent or tradition, the statutory basis
of fiscal policy rules differs from country to country. In general, the rules
are enshrined in a law or in the constitution, and in rare cases they are
contained in an administrative or policy guideline. If the rules affect a
group of countries (as in a currency union), they are prescribed as an
international treaty obligation (Table 1).
Another key institutional element is the authority responsible for the
surveillance and enforcement of the rules, as well as of the associated
transparency requirements. In most cases, this responsibility is exercised
by the audit office that reports to the legislature, while ultimate arbitration
and judgement usually rests with the courts. The question remains,
however, as to the technical competence of these entities in assessing
compliance with the rules (including accounting procedures, multiyear
framework, etc.). For example, in Peru, the central bank provides some
__________
26
For an overview of multiyear budgets and fiscal targets in OECD countries, see OECD (1995).
technical support for this purpose. Under a more comprehensive approach,
in the EU, the Council of Ministers responsible for Economy and Finance
(ECOFIN) exercises the surveillance authority, with the support of the
Commission and with specialized monitoring (of compliance with
accounting standards) by Eurostat.
In a related matter, it is necessary to determine the nature and the
extent of sanctions for noncompliance with the rules. At the national level,
sanctions usually consist of loss in reputation or adverse judicial
decision—in some countries, including penalties borne by the responsible
elected or appointed officials. However, in federal systems, financial
sanctions are levied on the delinquent government (noninterest-earning
deposits under the EMU or the CFA franc zone, outright fines in Canada
and Colombia, suspension of transfers in Brazil) or personal sanctions are
imposed on chief financial officials (criminal proceedings in Brazil, salary
cuts for cabinet members in a Canadian province).
3UHFRQGLWLRQVDQGFRQYHUJHQFH
From the very outset, successful implementation of fiscal policy
rules is predicated on three preconditions. The first is a FRQFHUWHGRXWUHDFK
FDPSDLJQ, including education and media coverage—which may take a
couple of years—to generate sufficient public understanding of the need
for rules and eventually support for their implementation (Argentina,
Brazil, New Zealand, EU). Second, this campaign must be accompanied by
a political debate that will lead to a EURDG OHJLVODWLYH FRQVHQVXV for the
introduction of fiscal policy rules. Such consensus has been fundamental
particularly where fiscal rules need passage of a constitutional amendment
(Germany, Switzerland, and U.S.). Third, it is necessary to map out
carefully a FRQYHUJHQFHSDWK. This, in essence, calls for an initial medium-
term adjustment program (Argentina, New Zealand, Peru, Switzerland,
EU) that includes a preannounced path for key performance indicators
(overall balance, current balance, etc.) in the run-up to the effective date of
implementation
27
. The corresponding annual budgets are then bound by the
preannounced values. The convergence path must provide, insofar as
__________
27
In the EU compliance with the EMU deficit reference value was to be achieved over a five-year
period (ending in 1997) albeit without specifying the profile of the convergence. In Argentina and
Peru, declining annual budget deficit limits have been specified in the fiscal responsibility laws,
over a three- and two-year period, respectively, prior to full compliance with the rule. By contrast,
in the Brazilian legislation, the rule entered into effect immediately following enactment.
possible, for the explicit treatment of deviations from these values in the
event of exogenous shocks during that period
28
. Overall, adequate
preparation and convergence are likely to be more important in the
implementation of fiscal rules than of monetary rules.
These preconditions, to be met in the first place with respect to fiscal
rules at the central level, should be accompanied or followed by a similar
effort at the subnational level. Overall, the introduction of fiscal rules is
greatly facilitated by progress in a number of areas that are equally relevant
for effective discretionary fiscal management. Indeed, transparency in
institutional arrangements and in accounting and reporting requirements
should be expedited, even without legislative action on rules. Of course,
these steps need to be followed up with further technical work and an
active public dialogue at all levels.
Finally, in most countries an important prerequisite for successful
implementation of fiscal rules is the phase-in of structural reforms that
ensure sustainability of the rules—in the face of fragility in the financial
system, rigidities in the public sector employment, demographic pressures,
or regional imbalances. These reform measures often encompass a number
of areas such as intergovernmental fiscal relations, tax structure, and public
pensions
29
.
&RQFOXGLQJUHPDUNV
With the primary objective of conferring credibility on
macroeconomic policies, while correcting the public sector deficit bias and
containing public indebtedness, an increasing number of advanced and
emerging market economies have adopted various forms of fiscal rules. In
contrast to previous types of fiscal rules, which were characterized by
ambiguities and by overall lack of transparency, recently introduced rules
have the potential of serving as a useful depoliticized policy framework.
However, an examination of the arguments for and against fiscal rules, as
well as the accumulated experience, confirms that rules are by no means a
universal panacea.
__________
28
For an analysis of the importance of starting with a strong initial budgetary position, with
simulations for EMU, see Eichengreen and Wyplosz (1998).
29
See Kopits (1997) on the need for social security reform for compliance with EMU fiscal reference
values.
To conclude, three broad lessons emerge. First, governments with a
strong reputation of fiscal prudence do not need to be constrained by rules.
Second, in countries where such a reputation is lacking, fiscal rules can
provide a useful policy framework and, over time, contribute to stability
and growth. Third, to enhance their usefulness, fiscal rules need to be well
designed at national and subnational levels of government, combining
simplicity, flexibility, and growth-oriented criteria; furthermore, they must
be implemented in a transparent manner, with the support of an appropriate
institutional infrastructure (especially as regards the budgetary process and
surveillance mechanism), and following careful preparation and
convergence.
7DEOH
6HOHFWHGFRXQWULHV6XPPDU\RIILVFDOSROLF\UXOHV
1
Rule/
Country
Effective
Date
Coverage
2
Basic
Rules
3
Escape
Clause
3
Additional
Rule
3
Statute
4
Sanction
5
Argentina 2000 NG OB/DL CF EL L J
Brazil 2001 NG, SG CB WL L J
Canada various SG CB L J
EU
members
1997 GG OB/DL MY T F
Germany 1969
6
NG, SG CB C J
New
Zealand
1994 GG PB MY L R
Peru 2000 NG OB/DL CF EL L J
Switzerland various SG CB C J
United
States
various SG CB CF C, J
Brazil 2001 NG, SG SL L J
Colombia 1997 SG PL L J
EU
members
1997 GG PL T J
New
Zealand
1994 GG SL L R
1) Excluding prohibition or limits on financing from specific sources.
2) General government (GG), national (central, federal) government (NG) or subnational (including
local) government (SG).
3) Budget rules consist of overall balance (OB), operating balance (PB), or current balance (CB),
subject to a prescribed limit on deficit (DL) as a proportion of GDP, applied on an annual basis,
except if specified on a multiyear (MY) basis. Also, a contingency fund (CF) is provided in some
cases. Additional rules consist of limits on primary expenditure (EL) or wage bill (WL). Debt rules
are specified as a limit for a given year (SL) or permanently (PL), as a proportion of GDP or of
government revenue.
4) Constitution (C), legal provision (L), or international treaty (T).
5) Sanctions for noncompliance: reputational (R), judicial (J), or financial (F).
6) The origins of the present rule can be traced to the Constitution of 1871, subject to modifications in
1919, 1949, and 1969.
$11(;
6,03/($5,7+0(7,&2)),6&$/58/(6
A fiscal policy rule can be specified in terms of a gradual reduction
in the public sector debt to (or maintenance at) a prudent level or ratio to
GDP. At the same time, this objective may be sufficiently flexible to
accommodate the effect of automatic stabilizers.
The intertemporal determination of public debt can be expressed as:
G
[(1L)/(1J)]G ±E
where (as a proportion of GDP, unless otherwise indicated):
d = stock of public sector debt
i = average nominal interest rate on public debt
g = nominal GDP growth rate
b = primary budget surplus.
In a highly indebted country, the authorities will target:
G
G
which is to be met within Q years, with a PLQLPXP annual reduction of [in
the debt ratio, by means of an operational rule expressed in terms of the
structural primary surplus:
E
(L±J)G [ (1)
Further, the operational target is defined in reference to trend
growth:
E
≡ U (1 *$3
) ±F (1 ± *$3
) ±N
where:
U = government revenue
F = primary current expenditure
N = capital expenditure
= revenue elasticity with respect to GAP
= expenditure elasticity with respect to GAP
*$3 = difference between trend GDP and actual GDP.
Therefore,
E
E is DOORZHG when *$3 !0
and
E
E is UHTXLUHGwhen*$3 0.
Compliance with rule (1) may be accompanied by variations in the
debt ratio that reflect deviations from trend growth rate: the debt ratio falls
(increases) with positive (negative) deviations and remains unchanged
when the economy is on the trend growth path.
Rule (1) implies that if the targeted reduction in the debt ratio is set
equal to the growth rate, [ JG
, then the target primary surplus becomes
E
LG (2)
which implies structural RYHUDOOEDODQFH. In the event, the balanced-budget
rule (2) leads to a fall in the debt ratio equivalent to the growth rate.
As an alternative, of particular relevance for a country in need of
infrastructure expenditure with a high expected social rate of return, the
target may be reset according to the golden rule, requiring structural
FXUUHQWEDODQFH,
E
N LG (3)
Rule (3) should be, of course, easier to meet than either (1) or (2), though it
still results in a fall in the debt ratio to the extent that N
<JG
However, a preferable approach would be to redefine the golden rule
in terms of an RSHUDWLQJ EDODQFH requirement (i.e., equivalence between
current revenue and current expenditure, including depreciation allowances
δ
), following accrual-based accounting,
E
N ±
δ
LG (4)
In addition, the balanced-budget rule may be supplemented with an
expenditure limit, set on primary spending or a major component thereof,
such as the wage bill. To safeguard it from cyclical fluctuations in output
or prices, this limit can be set in proportion to trend GDP.
5()(5(1&(6
Balassone, F. and D. Franco (2001), “Fiscal Federalism and the Stability
and Growth Pact: A Difficult Union”, paper included in this volume.
Buchanan, J.M. and R.E. Wagner (1877), 'HPRFUDF\ LQ 'HILFLW 7KH
3ROLWLFDO/HJDF\RI/RUG.H\QHV, New York, Academic Press.
Cukierman, A. and A.H. Meltzer (1986), “A Positive Theory of
Discretionary Policy, the Cost of Democratic Government, and the
Benefits of a Constitution”, (FRQRPLF ,QTXLU\, Vol. 24, July,
pp. 367-88.
Drazen, A. (2000), 3ROLWLFDO (FRQRP\ LQ 0DFURHFRQRPLFV Princeton,
Princeton University Press.
Eichengreen, B. and J. von Hagen (1995), “Fiscal Policy and Monetary
Union: Federalism, Fiscal Restrictions and the No-Bailout Rule”,
CEPR Discussion Paper, No. 1247, London, Centre for Economic
Policy Research.
Eichengreen, B. and C. Wyplosz (1998), “Stability Pact: More than a
Minor Nuisance?”, (FRQRPLF3ROLF\, Vol. 26, pp. 67-104.
European Commission, Directorate-General for Economic and Financial
Affairs (2000), (XURSHDQ (FRQRP\ 3XEOLF )LQDQFHV LQ (08 ±
, No. 3, Brussels, European Communities.
Friedman, M. (1948), “A Monetary and Fiscal Framework for Economic
Stability”, $PHULFDQ(FRQRPLF5HYLHZ, Vol. 38, June, pp. 245-64.
Gavin, M., R. Hausmann, R. Perotti and E. Talvi (1996), “Managing Fiscal
Policy in Latin America and the Caribbean: Volatility,
Procyclicality, and Limited Creditworthiness”, IDB Working Paper,
No. 13, Washington, Inter-American Development Bank.
Kopits, G. (1997), “Are Social Security Finances Compatible with EMU?”,
IMF Paper on Policy Analysis and Assessment PPAA/97/3,
Washington, February.
Kopits, G. (2000), “How Can Fiscal Policy Help Avert Currency Crises?”,
in )LQDQFLDO&ULVLV$1HYHU(QGLQJ6WRU\Vienna, Oesterreichische
Nationalbank, pp. 35–45.
Kopits, G. and J. Craig (1998), 7UDQVSDUHQF\LQ*RYHUQPHQW2SHUDWLRQV,
IMF Occasional Paper, No. 158, Washington, International
Monetary Fund.
Kopits, G., J.P. Jiménez and A. Manoel (2000), “Responsabilidad Fiscal a
Nivel Subnacional: Argentina y Brasil”, ;,,6HPLQDULR5HJLRQDOGH
3ROtWLFD)LVFDO&RPSHQGLRGH'RFXPHQWRVSantiago, UNECLAC,
(January 24 –26), pp. 25–57.
Kopits, G. and S. Symansky (1998), )LVFDO3ROLF\5XOHVIMF Occasional
Paper No. 162, Washington, International Monetary Fund.
Kydland, F.E. and E.C. Prescott (1977), “Rules Rather than Discretion:
The Inconsistency of Optimal Plans”, -RXUQDORI3ROLWLFDO(FRQRP\,
Vol. 85, pp. 473-91.
McKinnon, R.I. (1996), “Monetary Regimes, Government Borrowing
Constraints, and Market-Preserving Federalism: Implications for
EMU”, paper presented at the Conference on the Nation State in a
Global Information Era: Policy Changes, Queen’s University,
Kingston, November 13 –15 (unpublished).
OECD (1995), %XGJHWLQJIRU5HVXOWV3HUVSHFWLYHVRQ3XEOLF([SHQGLWXUH
0DQDJHPHQWParis, OECD.
Rattso, J. (ed.) (1998), )LVFDO )HGHUDOLVP DQG 6WDWH/RFDO )LQDQFH WKH
6FDQGLQDYLDQ3HUVSHFWLYH Cheltenham, Edward Elgar
Robinson, M. (1998), “Measuring Compliance with the Golden Rule”,
)LVFDO6WXGLHV, Vol. 19, No. 4, pp. 447-62.
Suits, D.B. and R.C. Fisher (1985), “A Balanced Budget Constitutional
Amendment: Economic Complexities and Uncertainties”, 1DWLRQDO
7D[-RXUQDO, Vol. 38, December, pp. 467-77.
Taylor, J.B. (2000), “Reassessing Discretionary Fiscal Policy”,-RXUQDORI
(FRQRPLF3HUVSHFWLYHV, Vol. 14, Summer, pp. 21-36.
Wellink, A.H. (1996), “Budgetary Control: Goodhart’s Law in
Government Finances?”, in C. Kool, J. Muysken and T. van Veen
(eds.), (VVD\VLQ0RQH\%DQNLQJDQG5HJXODWLRQ(VVD\VLQ+RQRU
RI&-2RUW, Dordrecht, Kluwer Academic Publishers.
0$1$*,1*38%/,&(;3(1',785(620((0(5*,1*
32/,&<,668(6$1'$)5$0(:25.)25$1$/<6,6
3DXO$WNLQVRQDQG3DXOYDQGHQ1RRUG
,QWURGXFWLRQ
Most OECD countries have experienced improvements in overall
fiscal positions in recent years. In terms of the general government finance
balance, the OECD area as a whole has achieved a surplus in 2000 for the
first time since 1969, which is projected to be maintained in 2001-02 (see
(FRQRPLF 2XWORRN 68, December 2000). As a result, there has been a
widespread tendency to reduce taxes in many countries, including in
Europe, and there are also signs that restraint in public expenditure is
being relaxed. At this stage these trends can be easily accommodated
without seriously weakening underlying fiscal positions in many countries,
as potential economic growth is considered to have picked up and interest
rates are low. However, uncertainty about the true underlying strength of
fiscal positions remains, as the surprisingly strong revenue growth in
recent years may incorporate a larger cyclical component (and a
correspondingly smaller structural component) than assumed
1
. Therefore,
while tax reductions are welcome in view of the scope they provide for
improving incentive structures in the economy, in a longer-term
perspective, with population ageing, public expenditure restraint to match
these tax cuts is called for.
Obviously, in a fiscal surplus environment expenditure restraint is a
delicate issue and not easily achieved. After a relatively long episode of
fiscal rigour, pent-up demand for public goods and services in many
countries may result into more calls for increased government spending.
With fiscal positions strong, these calls may be met without a sufficiently
careful trade-off between alternatives or without a sufficient evaluation of
__________
*
OECD. This paper is based on documentation originally prepared for the semi-annual meeting of
Working Party No. 1 of the OECD’s Economic Policy Committee on 16 and 17 October 2000.
However, the authors are writing in a personal capacity and it does not necessarily reflect the view
of the Organisation or its Member countries. They are indebted to Thomas Liebig for his
contribution, to Jon Blöndal and several colleagues in the Economics Department for comments
and to Anne Eggimann and Chantal Nicq for technical assistance.
1
See Van den Noord (2000).
the possible consequences (economic, social, environmental or other) of
spending choices. To the extent an assessment of such choices involves
normative judgements, economic analysis has little to say. However, where
there is scope for a given set of policy objectives to be achieved in more
cost-effective ways, there is a role for economic analysis. Public
expenditure is often examined in three dimensions. The first dimension
refers to the macroeconomic costs of public expenditure, which include the
economic distortions stemming from the tax burden and fiscal
sustainability risks associated with a growing debt burden. The second
dimension refers to allocative efficiency, or the outcomes achieved for a
marginal unit of public expenditure, and the third one to technical
efficiency, or the resource inputs for a marginal unit of output of public
goods and services.
The purpose of this paper is to highlight the main policy issues
related to public expenditure in OECD countries and to provide an
analytical framework for its assessment. After a brief review of public
expenditure developments in Section 2, Section 3 discusses the
three-pronged analytical concept referred to above. Section 4 examines the
various policy options that might be considered and that have been
experimented with in some countries with a view to raising the
performance of public expenditure. Section 5 concludes with an inventory
of assessment criteria that might serve to facilitate further analysis for
individual countries.
7UHQGVLQSXEOLFH[SHQGLWXUHDQGIRUFHVVKDSLQJWKHP
This section identifies recent trends in public expenditure in
aggregate and by economic or functional category, and highlights their
distinguishing features, as well as the factors that have shaped these trends
and that are likely to operate in the future
2
. These factors include
demographics, macroeconomic conditions and policy requirements,
regulatory reform, the design of entitlement programmes, income effects
and cost developments.
__________
2
For a review covering a longer time span, see Tanzi and Schuknecht (2000).
7UHQGVLQJHQHUDOJRYHUQPHQWH[SHQGLWXUH
Although institutional arrangements and the boundaries of the
public sector vary both over time and widely across countries, national
accounts data for the general government provide a reasonable basis for
examining the main trends in public expenditure on a cross-country basis.
The totals, as well as a breakdown by economic category for most OECD
countries since 1965, are reported in Figure 1, Table 1 and Table 2. A
more detailed breakdown can be found in the Annex. The main features of
these trends are:
- Total spending as a share of GDP rose rapidly nearly everywhere until
the early 1980s. Since then most countries have given greater weight to
expenditure restraint, often in the context of medium-term fiscal
strategies, and growth of spending generally slowed. Since the early
1990s most countries, with Japan being the most notable exception,
have achieved reductions to levels below those of the early 1980s.
- Government spending displays a clear counter-cyclical pattern in most
OECD countries, rising sharply at the time of recessions around 1975,
1982 and the early 1990s. Given that each successive cyclical spending
peak has exceeded previous peaks it is probably too early to be certain
that the long-term upward spending trend has been broken. The test will
come during a future downturn.
- There are marked differences in spending levels across major OECD
regions, and these differences have changed substantially over time. In
1965, spending in the United States was around 26 per cent of GDP,
just below the OECD average, some 7 percentage points higher than in
Japan and some 7 percentage points lower than in the euro area. During
the 35 years since then, the rise in spending in the United States, around
4 percentage points, has been far less than in the euro area
(12 percentage points) and in Japan (19 percentage points). The result
is that it is now Japan where spending, at 38 per cent of GDP, is close
to the OECD average (37 per cent of GDP), while in the United States
it is some 7 percentage points below average, and in the euro area it is
nearly 9 percentage points above average.
- The major factor that has put upward pressure on spending over this
period in nearly all countries has been the establishment and expansion
−
Figure 1. Trends in general government total outlays by economic category
Per cent of GDP
Income transfers Interest payments Consumption Subsidies Net capital outlays
0
10
20
30
40
50
60
1965 1970 1975 1980 1985 1990 1995 2000
0
10
20
30
40
50
60
1965 1970 1975 1980 1985 1990 1995 2000
0
10
20
30
40
50
60
1965 1970 1975 1980 1985 1990 1995 2000
0
10
20
30
40
50
60
1965 1970 1975 1980 1985 1990 1995 2000
)LJXUH
7UHQGVLQJHQHUDOJRYHUQPHQWWRWDORXWOD\V
E\HFRQRPLFFDWHJRU\
SHUFHQWRI*'3
Table 1.* H Q H UD OJ R Y H UQ PH Q WR X WOD \V E\F R X Q WU\
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995
2000
1
Australia 24.6 25.2 31.3 32.3 37.8 33.0 35.4 31.4
Austria 36.6 38.0 44.4 47.2 50.1 48.5 52.4 48.8
Belgium 35.0 39.7 47.6 53.4 57.3 50.8 50.3 46.7
Canada 27.8 33.8 38.9 39.1 45.4 46.0 45.3 37.8
Denmark
2
31.8 40.1 47.1 55.0 58.0 53.6 56.6 51.3
Finland 30.3 29.7 37.0 37.1 42.3 44.4 54.3 44.8
France 37.6 37.6 42.3 45.4 51.9 49.6 53.6 51.2
Germany 35.3 37.2 47.1 46.5 45.6 43.8 46.3 43.0
Greece 22.0 23.3 27.1 29.6 42.3 47.8 46.6 43.7
Ireland 36.0 37.7 40.7 47.6 50.5 39.5 37.6 27.7
Italy 32.8 32.7 41.0 41.8 50.6 53.1 52.3 46.7
Japan 19.0 19.0 26.8 32.0 31.6 31.3 35.6 38.2
Korea 14.5 14.8 16.9 19.2 17.6 18.3 19.3 23.4
Mexico .. .. .. .. .. .. 21.4 ..
Netherlands 34.7 37.0 45.7 50.9 51.9 49.4 47.7 41.5
Norway 29.1 34.9 39.8 43.9 41.5 49.7 47.6 40.6
Por tugal
2
18.1 18.0 25.2 28.1 42.9 44.2 41.2 42.1
Spain 19.5 21.7 24.1 31.3 39.4 41.4 44.0 38.5
Sweden 33.5 41.7 47.3 56.9 59.9 55.8 62.1 53.9
United Kingdom
2
33.5 36.7 44.4 43.0 44.0 41.9 44.4 38.4
United States 25.6 29.6 32.3 31.3 33.8 33.6 32.9 29.3
Euro area 33.1 33.9 40.9 43.0 47.2 46.3 49.1 45.1
OECD 26.9 29.2 34.4 35.5 38.1 38.0 39.4 36.5
1. Estimates
2. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD 68, December 2000, OECD National Accounts and OECD calculations.
1. Estimates.
2. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data
are backward extrapolations based on earlier National Accounts series.
: OECD Economic Outlook 68, December 2000, OECD National Accounts and OECD
calculations.
7DEOH
*HQHUDOJRYHUQPHQWRXWOD\VE\FRXQWU\
SHUFHQWRI*'3
Table 2.
Per cent of GDP
Income transfers Subsidies Interest payments Consumption
Net capital outlays
1
Total outlays
Australia 8.3 1.2 2.0 18.5 1.4 31.4
Austria 18.3 2.5 3.5 19.4 5.1 48.8
Belgium 14.4 1.5 6.7 21.0 3.1 46.7
Canada 10.9 1.1 7.4 18.4 0.0 37.8
Denmark 17.2 2.3 4.5 25.3 2.0 51.3
Finland 12.6 1.5 3.1 20.8 6.7 44.8
France 18.1 1.3 3.3 23.4 5.1 51.2
Germany 18.6 1.7 3.4 18.8 0.4 43.0
Greece 16.1 0.2 7.2 15.0 5.2 43.7
Ireland 9.7 0.7 2.2 11.8 3.2 27.7
Italy 17.3 1.2 6.5 17.9 3.7 46.7
Japan 15.7 0.6 4.0 10.1 7.8 38.2
Korea 3.3 0.3 1.6 9.7 8.6 23.4
Netherlands 11.8 1.6 3.9 22.6 1.5 41.5
Norway 13.7 2.5 1.6 18.8 4.1 40.6
Portugal 12.5 1.2 3.2 21.0 4.2 42.1
Spain 12.4 1.0 3.6 16.9 4.6 38.5
Sweden 18.3 1.8 4.1 26.5 3.2 53.9
United Kingdom 13.1 0.5 2.7 18.3 3.8 38.4
United States 10.5 0.2 3.6 14.1 0.9 29.3
Euro area 16.7 1.4 4.2 19.7 3.0 45.1
OECD 12.8 0.8 3.8 15.7 3.4 36.5
Source: OECD Analytical Database figures underlying OECD Economic Outlook 68, December 20
00.
1. Net fixed investment plus net capital transfers.
7DEOH(VWLPDWHGJHQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\IRUWKH\HDU
SHUFHQWRI*'3
1. Net fixed investment plus net capital transfers.
: OECD Analytical Database figures underlying OECD Economic Outlook 68, December 2000.
−
of programmes and provision of services in the social policy domain
(public pensions, income support, health care, education and other
public services)
3
. The income support element of these entitlements is
reflected in a persistent rise in income transfer payments until the
mid-1990s. While these payments are no longer rising at the area-wide
level, they have not fallen much and have made little contribution to the
overall spending decline since the early 1990s.
- A second factor that contributed importantly to upward pressure on
spending until the mid-1990s was debt interest. This reflected a
combination of rising public indebtedness, as large and sustained
budget deficits became common after the first oil price shock in 1974,
and rising interest rates. As interest rates have declined and budget
positions have improved during the 1990s, these forces have reversed
and debt interest payments have declined, accounting for half of the
overall decline in spending at the area-wide level since 1995
4
.
- Other major categories of spending, i.e. subsidies, government
consumption and net government capital outlays, have displayed few
general patterns and little overall trend. Net capital outlays and
subsidies have made modest contributions to the recent declines in total
spending in many countries (capital spending in Japan is the major
exception, although even there net capital outlays are now only at their
1980 level as a share of GDP). While only a few countries have
significantly reduced government consumption (the United States, Italy
and, especially, Canada stand out during the past decade), such
spending has not been a source of significant pressure in most countries
since the trend toward spending restraint began in the early 1980s.
Upward pressure on spending is likely to re-emerge in the decades
ahead. The major force behind this pressure would be the ageing of
populations and consequent demands this implies on social spending,
notably on pensions, health care and associated personal services. Other
forces may include the need to restore spending in areas where restraint
__________
3
See for example Oxley and Martin (1991), MacFarlan and Oxley (1996) and OECD (1998 ).
4
It should be noted that recorded data overstate the importance of debt interest payments because
the part that reflects the inflation compensation component in nominal interest rates has a
counterpart in the erosion of the real value of outstanding debt,
an inflation tax, which is never
recorded in the budgetary accounts. This was significant when inflation was high, but at this stage,
with inflation low in most countries, the bulk of interest payments imply a real burden on
taxpayers.
has been applied and could prove to have gone too far, and the likelihood
of rising interest rates in countries where these have fallen to low levels
but public indebtedness remains high.
%UHDNLQJGRZQSXEOLFH[SHQGLWXUHE\IXQFWLRQ
For policy purposes it is important to focus on expenditures in terms
of their functions, each of which may involve a mix of economic
categories, since measures to affect spending must be justified in terms of
their concrete purpose. The breakdown in Table 3 attempts to group
government spending in line with basic concepts of public economics.
Four major types of government functions which call for expenditure (over
and above debt servicing, which is obligatory) are distinguished, each
referring to different cases where markets and prices will fail to result in
efficient outcomes, and therefore call for government intervention
5
:
3XEOLF JRRGV DQG VHUYLFHV This category comprises the provision of
essential “pure” public goods and services that cannot be rationed by
the price mechanism and therefore would not be supplied in efficient
amounts if markets were used to make them available. Examples are
national defence and general public services such as administration,
legislation and regulation.
0HULW JRRGV DQG VHUYLFHV These are public goods that in principle
could be (and in most countries to some extent are) made available
through markets. In many cases, government provision of such goods
and services is justified because of a conviction that they would
otherwise be provided in less than the efficient amount, because a
significant number of consumers lack the required purchasing power,
while externalities give these goods and services a public goods
element. For example, government provision of education is common
because citizens may ignore the social return of human capital
investment, or are unable to fund it. Usually informational asymmetry
is mentioned as an important additional economic motive for the
government to be engaged in the delivery or provision of merit goods
and services. These asymmetries limit the ability of the consumer to
identify the quality of the goods and services fully and therefore distort
__________
5
This breakdown has been introduced by Oxley and Martin (1991).
prices and the quantities delivered. Health care is an important example
in this regard.
(FRQRPLFVHUYLFHV. This refers to the provision or co-funding of private
goods or services by the government. Intervention has often been felt to
be desirable in markets for goods and services that are prone to natural
monopolies, where externalities are judged to result in inefficient
supply if provision is left to the market, or where particular groups of
providers are felt to warrant assistance. Prominent examples include
public utilities (where entry barriers are associated with the sunk cost
of distribution networks) and financial support for specific activities
such as research and development, small and medium-sized enterprises
and agriculture. It should be noted that where these services are
provided by public enterprises their cost is not consolidated with the
general government accounts. Hence their operations will only be
reflected in public expenditure to the extent that the government
subsidises them.
6RFLDO WUDQVIHUV. These are transfers that provide support for income
and living standards. Beneficiaries may include those whose market
income is low or has declined sharply, or who face exceptional
expenses due to old age, disability, sickness, unemployment, etc.
6
.
Unfortunately, the functional breakdown in Table 3 covers a
narrower range of countries and a shorter period than the breakdown by
economic categories due to data constraints. In particular, the series only
start in 1980 and, for most countries, the latest year for which data are
available is 1995, due to problems associated with the adoption of the new
national accounting standards, SNA93 and ESA95. Moreover, as the data
are drawn from a range of sources aside from the national accounts, they
are not always comparable across countries. Nevertheless, a few broad
patterns emerge from Table 3. First, the share of “pure” public goods in
GDP has remained fairly stable in most countries in the sample during
both the 1980s and the first half of the 1990s. Major exceptions are the
United Kingdom, where the expenditure share of public goods sharply
declined during the 1980s, and the United States, where a marked drop in
__________
6
Obviously, this functional category largely overlaps with the economic category of income
transfers. However, there are differences; the latter category includes income transfers to other
countries (for example contributions to international institutions and development aid), whereas
the former includes both cash transfers and imputed transfers in kind.
defence spending after the end of the cold war led to a fall in public goods
expenditure in the 1990s. In most countries the share of economic services
in GDP has remained broadly constant as well, although significant falls
were recorded in Japan and Norway in the 1980s and in Germany, Italy,
the United Kingdom and Australia in the 1990s. By contrast, the main
spending hikes have been registered in the social policy area (merit goods
and social transfers), both in the 1980s and 1990s.
Although functional spending patterns have thus been subject to
change in the past two decades, the overall picture has remained that in
those countries with large amounts of government spending relative to
GDP, much of that spending is on social transfer and merit goods
(Figure 2). Most European countries are in the upper range of total
expenditures, as well as merit goods and social transfer expenditures,
whereas the United States, Japan, Australia, Korea and New Zealand are in
the lower range. On average public spending on social transfers and merit
goods in the countries in the sample in 1995 amounted to nearly 30 per
cent of their GDP. Moreover, the range from 7½ per cent of GDP in Korea
to nearly 40 per cent of GDP in Sweden was wide (but see below for a
comparison which includes private spending). Meanwhile, public goods on
average represented around 7 per cent of GDP in 1995, with the
Netherlands, France and Spain being at the upper end of the scale.
Economic services spending is relatively small and varies little across
countries.
6RPHPHDVXUHPHQWLVVXHVWKHOLPLWVWRJHQHUDOJRYHUQPHQWGDWD
Notwithstanding the advantages of general government data drawn
from national accounts in terms of availability and cross-country
comparability, classification and measurement issues are likely to make it
necessary to draw on other data sources. Institutional arrangements and the
borders of the public sector do not always correspond well to the general
government. In particular, financial relationships with state-owned
enterprises are an important element of public finances in some countries.
Moreover, the assessment of policies that motivate government spending
may not be possible without reference to any private spending that
supplements or accompanies it. Two issues stand out.
7DEOH
6WUXFWXUHRIJRYHUQPHQWRXWOD\VE\IXQFWLRQ
SHUFHQWRI*'3
Total
expenditure Total Defence
General
public
services
Other
functions Total Education Health
Other
social
services Total Pensions Disability Sickness
Family
cash
benefits
Unemploy-
ment
Housing
and other
benefits
Economic
services
Public debt
interest
1980 34.8 7.8 2.5 3.0 2.3 10.5 5.9 4.4 0.2 6.8 3.8 0.8 0.1 0.9 0.7 0.4 5.9 3.1
1990 36.4 8.1 2.0 2.9 3.2 10.4 4.3 5.3 0.8 7.8 3.2 1.2 0.2 1.3 1.4 0.6 6.4 3.7
1995 37.7 8.2 1.9 3.0 3.3 10.5 4.5 5.5 0.5 9.0 3.2 1.2 0.1 2.2 2.0 0.3 5.6 4.1
1990 48.3 3.9 1.0 2.9 0.0 11.0 5.2 5.2 0.6 17.9 12.3 1.7 0.2 2.1 1.2 0.4 3.1 4.0
1995 52.2 4.5 0.9 3.6 0.0 11.9 5.3 5.7 0.8 19.2 13.1 1.7 0.2 1.9 1.8 0.4 3.1 4.4
1980 39.6 4.0 1.6 2.4 0.0 10.5 5.0 5.4 0.1 8.1 3.1 0.7 0.0 0.7 1.6 2.0 2.6 5.4
1990 46.7 3.2 1.7 1.5 0.0 12.2 5.4 6.7 0.1 10.8 4.3 1.0 0.1 0.5 2.4 2.5 2.6 9.5
1995 46.3 2.9 1.4 1.5 0.0 12.3 5.8 6.5 0.0 11.5 4.8 1.0 0.1 0.8 1.8 3.1 2.4 9.6
1980 55.6 7.4 2.5 4.1 0.8 17.8 7.7 5.7 4.5 16.8 6.0 1.9 2.0 1.1 5.3 0.4 6.0 3.9
1990 57.4 6.1 2.0 4.1 0.0 15.7 6.2 5.1 4.4 17.7 6.3 1.8 1.2 1.5 5.4 1.5 5.9 7.3
1995 59.9 6.0 1.7 4.3 0.0 16.5 6.5 5.1 4.9 20.8 7.4 2.2 0.6 1.9 6.3 2.3 5.6 6.4
1980 35.4 3.3 1.4 1.8 0.1 11.5 4.8 5.0 1.7 11.7 5.5 2.9 0.2 1.1 1.6 0.4 0.9 1.0
1990 44.4 3.4 1.4 1.9 0.1 15.2 6.4 6.4 2.5 15.9 7.4 3.3 0.6 1.9 2.1 0.6 1.0 1.4
1995 54.3 3.3 1.6 1.6 0.1 15.2 6.6 5.6 3.0 22.5 8.9 3.9 0.5 2.7 5.5 1.1 1.1 4.0
1990 50.2 10.1 3.1 3.9 3.1 12.6 5.1 6.5 1.0 18.7 10.9 1.5 0.5 2.1 2.6 1.1 3.3 2.9
1993 55.4 9.2 2.9 4.5 1.8 14.1 5.9 7.1 1.1 20.9 12.0 1.5 0.6 2.2 3.3 1.3 3.1 3.5
1991 47.4 6.4 1.9 4.6 0.0 12.7 4.4 7.1 1.2 16.3 9.7 1.2 0.4 1.3 3.0 0.6 5.3 2.9
1995 49.7 5.2 1.4 3.9 0.0 13.9 4.5 8.0 1.3 18.2 10.7 1.4 0.5 1.2 3.7 0.8 4.5 3.7
Public goods Merit goods Income transfers
1. Expenditure by function may not add up to total expenditure as these are derived from different sources. In particular, expenditures by
function refers to fiscal years and total expenditure to calendar year; moreover, total expenditure is net of capital transfers received.
: OECD National Accounts, Social Expenditure Database, , OECD; and .
Total
expenditure Total Defence
General
public
services
Other
functions Total Education Health
Other
social
services Total Pensions Disability Sickness
Family
cash
benefits
Unemploy-
ment
Housing
and other
benefits
Economic
services
Public debt
interest
1980 42.3 5.3 1.7 3.4 0.2 10.7 4.8 5.6 0.3 12.4 9.0 1.5 0.3 1.0 0.6 0.0 6.4 5.0
1990 53.2 6.5 1.9 4.2 0.4 12.4 5.8 6.3 0.3 16.3 11.9 1.9 0.2 0.6 1.5 0.0 6.2 9.4
1995 52.2 6.5 1.7 4.5 0.3 10.2 4.5 5.3 0.3 17.9 13.5 1.8 0.1 0.4 2.0 0.0 4.6 11.5
1980 32.8 4.1 0.9 3.3 0.0 9.9 4.9 4.6 0.4 9.6 4.0 0.4 0.1 0.2 0.0 2.4 6.0 3.2
1990 31.6 4.4 0.9 3.4 0.0 8.9 3.7 4.7 0.5 10.0 5.0 0.5 0.1 0.2 0.3 2.2 4.3 3.9
1995 36.5 4.5 0.9 3.6 0.0 10.1 3.8 5.6 0.6 12.8 6.2 0.5 0.1 0.2 0.5 2.9 5.3 3.8
1990 18.3 6.7 3.9 2.0 0.8 1.8 m 1.7 0.1 1.3 0.8 0.3 .. 0.0 0.1 0.2 2.8 0.5
1995 19.2 5.7 2.9 2.0 0.8 5.6 3.6 1.8 0.2 1.8 1.3 0.3 .. 0.0 0.1 0.1 3.7 0.5
1980 57.6 12.6 2.9 9.7 .. 13.0 6.3 5.7 1.1 20.6 7.6 4.4 3.3 2.0 2.3 1.0 6.1 3.8
1990 54.9 11.7 2.4 9.2 .. 11.4 4.6 5.8 1.0 21.7 8.4 4.6 2.9 1.2 3.4 1.2 6.3 5.9
1995 52.2 11.6 1.8 9.8 .. 12.0 4.6 6.5 1.0 19.2 7.5 3.9 1.9 1.0 4.0 1.0 6.4 5.9
1990 46.2 6.5 1.8 4.7 0.0 11.4 5.5 5.8 0.1 16.3 7.6 2.0 0.9 2.6 2.8 0.3 0.0 8.5
1994 38.9 5.3 1.1 4.2 0.0 10.6 5.2 5.3 0.1 13.5 6.0 1.7 1.1 2.0 2.0 0.7 0.0 4.8
1980 45.0 4.9 2.6 2.2 0.1 13.5 5.9 5.9 1.7 10.9 5.1 1.9 1.5 1.3 0.4 0.8 9.1 3.1
1990 50.6 6.2 3.1 3.0 0.2 17.1 6.4 6.5 4.2 15.8 6.3 2.8 1.6 1.9 2.1 1.1 7.6 3.6
1993 52.0 6.3 2.6 3.1 0.6 18.4 6.8 6.6 5.1 15.9 6.2 2.7 1.2 2.3 2.4 1.1 7.2 2.8
1990 45.0 8.0 2.3 2.7 3.1 8.8 4.3 4.2 0.3 9.5 5.3 1.9 0.6 0.7 1.0 0.1 5.7 8.1
1995 49.9 8.3 2.2 2.0 4.1 10.6 5.4 4.7 0.5 12.1 7.3 1.7 0.6 0.7 1.7 0.1 6.3 6.3
Table 3.
(continued)
Per cent of GDP
1
Public goods Merit goods Income transfers
7DEOHFRQWLQXHG
6WUXFWXUHRIJRYHUQPHQWRXWOD\VE\IXQFWLRQ
SHUFHQWRI*'3
1. Expenditure by function may not add up to total expenditure as these are derived from different sources. In particular, expenditures by
function refers to fiscal years and total expenditure to calendar year; moreover, total expenditure is net of capital transfers received.
: OECD National Accounts, Social Expenditure Database, , OECD; and .
Total
expenditure Total Defence
General
public
services
Other
functions Total Education Health
Ot her
social
services Total Pensi ons Disabil it y Sickness
Fami ly
cash
benefits
Unemploy-
ment
Housing
and other
benefits
Economic
services
Public debt
interest
1990 41.8 8.6 1.5 1.7 5.3 9.7 4.2 5.2 0.3 13.4 7.8 1.3 0.9 0.2 3.0 0.2 5.5 3.8
1995 45.2 9.9 1.4 1.8 6.7 10.6 4.8 5.5 0.3 14.9 8.9 1.3 1.1 0.3 3.2 0.2 5.9 5.2
1980 60.0 6.9 3.3 2.9 0.7 20.0 7.6 8.4 3.9 16.5 7.2 2.1 2.3 1.7 1.6 1.5 na 3.9
1990 59.1 5.6 2.6 2.9 0.1 18.8 6.8 7.6 4.4 19.2 7.9 2.8 2.5 2.2 2.6 1.2 2.8 4.9
1995 64.5 5.4 2.3 3.0 0.1 17.2 6.6 5.7 4.9 21.2 8.6 2.7 1.1 2.0 4.5 2.1 3.4 6.8
1980 45.3 8.1 5.0 1.9 1.3 10.8 4.6 5.1 1.1 12.8 7.1 1.1 0.3 1.8 1.7 1.0 4.0 4.7
1990 41.4 6.5 4.1 1.9 0.5 10.2 4.3 5.0 0.9 13.7 7.1 1.8 0.4 1.6 1.3 1.5 4.1 3.4
1995 43.6 5.4 3.2 1.9 0.2 11.5 4.6 5.7 1.2 15.6 7.3 2.8 0.2 1.9 1.3 2.1 3.3 3.6
1980 32.5 10.6 7.1 2.5 1.0 9.7 5.3 4.0 0.5 9.3 6.3 0.9 0.3 0.5 0.9 0.5 3.7 3.2
1990 34.9 10.9 7.0 2.8 1.0 10.8 5.3 5.2 0.3 8.5 6.2 0.8 0.3 0.2 0.7 0.4 3.1 5.1
1995 34.3 9.2 5.2 2.9 1.0 11.9 5.0 6.5 0.4 9.4 6.5 1.0 0.3 0.3 0.6 0.6 2.8 4.8
1. Expenditure by function may not add up to total expenditure as these are derived from different sources. In particular, expenditures by function refers to fiscal years and total expenditure to calendar year;
moreover, total expenditure is net of capital transfers received.
OECD National Accounts, Social Expenditure Database; , OECD; and .
Table 3. (continued)
Per cent of GDP
1
Public goods Merit goods Income transfers
7DEOHFRQWLQXHG
6WUXFWXUHRIJRYHUQPHQWRXWOD\VE\IXQFWLRQ
SHUFHQWRI*'3
1. Expenditure by function may not add up to total expenditure as these are derived from different sources. In particular, expenditures by
function refers to fiscal years and total expenditure to calendar year; moreover, total expenditure is net of capital transfers received.
: OECD National Accounts, Social Expenditure Database, , OECD; and .
Figure 2.
Per cent of GDP
1. Data are for 1993.
2. Data are for 1994.
Source: OECD National Accounts, Social Expenditure Database; Education at a Glance , OECD; and Survey of Current
Business
.
0
10
20
30
40
50
60
Denmark
Netherlands
Sweden
France (1)
Norway (1)
Italy
Spain
Finland
Germany
Portugal
United Kingdom
Austria
Canada
United States
Australia
Japan
New Zealand (2)
Korea
Public goods and Economic services Merit goods and Income transfers Public debt interest
)LJXUH
6WUXFWXUHRIJRYHUQPHQWRXWOD\VE\IXQFWLRQ
SHUFHQWRI*'3
First, obtaining comparative data on public employment, an item
which may importantly influence public spending, proves to be
problematic. Given the range of employment statuses, activities in which
the public sector is engaged as provider, and institutional arrangements for
doing so within and across countries, not to mention the changes that have
taken place over time, there are many obstacles to constructing a consistent
data set. In particular, the new national accounts system (SNA93/ESA95)
classifies employment by activity and not by employer, and does not (at
least at this stage) allow large parts of public employment in sectors such
as health and social work, education and other community social and
personal services to be identified. The Public Management Service
(PUMA) in OECD has devoted considerable effort to addressing these
problems and has constructed a database for 21 countries on the basis of
responses to a regular questionnaire. The figures reported in Table 4,
which are in terms of full-time equivalents for all levels of government,
LH the general government sector as defined in the national accounts but
excluding social security administrations, indicate that a wide range exists
across countries. Trends over time during the past decade have in most
cases been either steady or slightly declining, suggesting that public
employment restraint has made some contribution to the success most
countries have had in holding down spending
7
.
Second, public expenditure may not be the only way to deliver
certain services or to achieve particular objectives. Private spending may
have a role to play and, if government intervention is felt to be warranted,
a mix of regulatory arrangements, mandates and tax incentives may be
used to encourage such spending. Social policy areas in particular are
managed in ways that differ substantially across countries so that
international comparisons of resources devoted to achieving policy
objectives in these areas will be highly misleading if no account is taken of
private spending. Furthermore, the extent to which social benefits are
taxed varies across countries, distorting comparisons. The OECD has
recently addressed these problems by estimating the amount of social
expenditure covering social transfers (net of taxation of benefits) and merit
good expenditures but excluding those on education, by the private sector
in response to regulations, mandates and tax incentives (Table 5). Overall
__________
7
PUMA has also on occasion collected data for a wider definition of the public sector, including
state-owned enterprises (whose financial performance impinges on the government’s overall
financial position), but these are not maintained on a continuing basis.
the data suggest that while public social expenditure as a share of GDP, as
reported in national accounts, varies widely across countries (in the range
of 15 to 40 per cent) differences in total social expenditure, including
policy-induced private expenditure, are much smaller (they range from 18
to 28 per cent)
8
.
$VVHVVLQJSXEOLFH[SHQGLWXUH
The purpose of this section is to provide broad criteria for the
assessment of public expenditure in OECD countries. Since government
expenditure reflects collective choices that emerge from the political
process and vary across countries, there will be limits to what economic
analysis alone can provide. However, it should be possible to evaluate the
economic consequences of the way these choices interact with institutional
arrangements and other elements of the economic environment and to
make judgements about the extent to which the apparent objectives which
underlie these choices are in fact being achieved in a cost-effective way.
This can be done at three levels Lits macroeconomic consequences;
LL the allocation of resources within the economy; and LLL the technical
or operational efficiency with which it is carried out.
7KHPDFURHFRQRPLFFRQVHTXHQFHV
High or rising public expenditure raises two major issues from a
macroeconomic perspective:
- First, it poses financing problems that make it difficult to ensure fiscal
discipline and thus tends to make macroeconomic policy management
difficult, depending on how well the processes of budget formulation
and implementation operate. A lack of adequate planning and
evaluation procedures incorporated in the formulation process, such as
safeguards against the use of unrealistic economic assumptions, have
often led countries to overestimate how much spending could be
afforded. Insufficient controls at the implementation stage have had
__________
8
These estimates also correct for the fact that in some countries ( Germany) social transfers are
not subject to taxation whereas in others they are. To make numbers comparable, tax payments by
social security recipients have been netted out in the latter group of countries.
Table 4.
1990 1991 1992 1993 1994 1995 1996 1997 1998 199
Australia 14.815.415.314.914.614.314.214.314.015.
Austria 10.610.510.610.811.011.011.010.110.0 ..
Canada 20.3 21.1 21.4 21.1 20.4 19.8 19.3 18.5 17.9 17.
Czech Republic .. .. .. .. 13.8 14.6 14.7 .. .. ..
Denmark 26.626.426.627.327.827.526.526.6 .. ..
Finland 23.224.325.325.525.724.725.025.024.3 ..
France 20.4 20.7 21.0 21.5 21.6 21.6 21.7 21.4 21.7 ..
Germany 15.1 .. 14.1 13.8 13.5 13.2 13.0 12.9 12.6 12.
Hungary 0.0 0.0 0.0 0.0 24.8 24.6 23.4 22.8 22.5 ..
Iceland 14.6 15.2 15.4 15.6 .. .. .. .. .. ..
Ireland 17.417.717.617.717.316.816.415.914.614.
Italy 17.317.217.518.017.517.517.317.0 .. ..
Korea 4.5 4.6 4.7 4.7 4.6 4.4 4.5 4.4 4.5 ..
Luxembourg 8.1 8.1 8.0 6.9 .. .. .. .. .. ..
Netherlands 12.9 12.2 12.9 12.7 12.4 11.8 .. .. .. ..
New Zealand 14.214.514.113.913.412.412.211.912.2 ..
Portugal 0.0 11.9 15.1 15.0 14.8 15.2 15.3 15.5 15.2 ..
Spain 14.014.314.915.616.016.216.015.7 .. ..
Sweden 28.4 28.9 28.8 28.6 26.8 26.2 .. .. .. ..
United Kingdom 16.1 16.2 15.9 14.0 12.5 11.6 10.9 10.4 10.2 10.
United States 14.9 15.2 15.2 15.1 15.0 14.9 14.8 14.4 14.4 ..
Public employment in general covers all individuals paid by government funds at all levels of government, and corresponds to the general government
excluding public enterprises and social security administrations.
Australia
Excludes financial and trading government enterprises. Full-time + Part-time
Austria
Excludes public corporations. Full-time Equivalent
Canada
Does not include government business enterprises. Full-time + Part-time + Casual
Does not include First Nations and Inuits Government.
Czech Republic
Full-time Equivalent
Denmark
Full-time Equivalent
Finland
Excluding state enterprises. Full-time + Part-time
France
Excluding public operators of the Posts and telecommunication since 1991. Full-time Equivalent
Germany
Includes military. This total does not match with the summation of the 3 levels (federa + länder + municipalities), but this is
the total sent by the country. The total may include the indirect public sector. Full-time + Part-time
Hungary
Excludes military. Headcount
Ireland
The public service comprises Civil Service, Garda Siochana (Police Force), Education Sector,
Defence Forces, Health Sector, non-commercial State-sponsored Bodies and Local Authorities. Actual Members
Italy
Post and telecommunication services have been excluded since 1994. Full-time Equivalent
Korea
Provisional data.
Netherlands
Data are low in comparison to other countries as there are many individuals working part-time. Full-time Equivalent
New Zealand
Excludes public enterprises. Full-time Equivalent
Portugal
Includes public and other employees in central administration and only public employees in Local and Regional
administrations. Full-time + Part-time Excludes Social security. Before 1994, data do not include Regional autonomous administrations.
Spain
Includes social security employment managed at the central level.
Sweden
Full-time Equivalent
United Kingdom
Excludes NHS Trusts and public corporations. Full-time Equivalent
United States
Annual averages. Includes part-time and season workers. Actual Members
OECD/PUMA PSPE- Public Sector Pay and Employment Database (2000).
7DEOH
3XEOLFHPSOR\PHQWDVDSHUFHQWDJHRIWRWDOHPSOR\PHQW
Table 5.
Per cent of GDP
Private
Total
2
Net Public
3
Mandatory Voluntary
Gross public social
expenditure
4
Australia 21.6 18.7 0.3 2.7 20.3
Belgium .. .. .. .. 30.1
Canada 21.2 17.9 .. 3.5 20.8
Denmark 24.4 23.6 0.3 0.5 37.6
Finland 25.7 25.1 0.0 0.7 35.7
Germany 27.7 25.9 1.0 0.8 30.4
Ireland 18.7 17.4 .. 1.5 21.8
Italy 22.3 20.9 .. 1.4 21.8
Netherlands 25.0 21.2 0.5 3.4 30.1
Norway .. 21.9 0.6 .. 31.5
Sweden 27.0 25.4 0.2 1.4 36.4
United Kingdom 26.0 22.3 0.3 3.6 25.9
United States 24.5 17.5 0.5 7.8 17.1
1. Social expenditure covers: cash-benefits for old age, disability, occupational injury and disease and sickness;
services for the elderly and disabled; survivors’ pensions; family cash benefits; family services; active labour
market programmes; unemployment benefits; health care expenditure; housing benefits.
2. The total is a consolidated figure and may be less than the sum of the components.
3. Calculated as gross public social expenditure less direct taxes and social security contributions levied on
social transfers and benefit income claimed back through taxes on consumption, plus tax breaks for social
purposes.
4. General government social expenditure (for definition, see note 1).
Adema (2000).
similar effects. Another element is processes that work to encourage the
reversal of any spending increases designed to mitigate cyclical
downswings, either via automatic stabilisers or of a discretionary nature,
once the cycle turns up. But it is also an issue of relationships between
various government entities. These include both horizontal relationships,
LHbetween finance ministries and sectoral, or spending, ministries and
agencies, and vertical ones, LHbetween central and lower levels of
government.
7DEOH
1HWVRFLDOH[SHQGLWXUHE\VRXUFH
SHUFHQWRI*'3
- Second, the disincentives and distortions created by the tax burden
required to meet the government’s financing needs may carry high
economic cost. To some degree this can be seen as an issue of resource
allocation and income distribution. But since the tax burden is largely
driven by the overall level of government spending and often impinges
on overall economic performance, it also has a macroeconomic
dimension.
At this stage, the fiscal situation and outlook in most OECD
countries is better than for many years, implying that financing
expenditure without heavy recourse to borrowing has generally been
achieved. This has been facilitated by the expenditure restraint that most
countries have been able to exercise during the 1990s and contrasts with
the persistent budgetary problems which emerged during the 1970s, at a
time when public expenditure was rising rapidly. Nonetheless, tax burdens
are now very high in many countries, especially in Europe, and have led to
concerns about the fairness of their incidence, their impact on economic
behaviour (particularly in labour markets) and the sustainability of
potentially mobile tax bases. Importantly, public expenditure control will
become more challenging now that fiscal positions are improving
throughout the area. In a surplus environment political pressures could
lead to uncoordinated tax cuts and spending increases, which might
eliminate options of financing structural reforms or retiring debt. Indeed,
countries with budget surpluses are already showing signs that fiscal
management can be difficult in these situations. If government expenditure
starts to rise from current levels, the tax implications and the associated
distortions further down the road may prove very problematic, the more so
since demands from spending and pressures on the revenue base may
continue to increase in the future as populations in most OECD countries
age.
Against this backdrop the assessment of the macroeconomic impact
of public expenditure needs to consider the following questions:
- Are processes of evaluation and planning in place to ensure that public
expenditure decisions are based on a realistic view of their cost and
overall affordability? Do these processes work to encourage the
reversal of spending increases to mitigate cyclical downturns, whether
discretionary or arising from the operation of automatic stabilisers,
once cyclical conditions improve?
- Is public expenditure sufficiently well controlled so that the
implementation of budgetary plans is not frequently undermined by
unpleasant “surprises” on the spending side?
- Is spending by lower levels of government either adequately overseen
and controlled by the Finance Ministry or dependent on their ability to
finance it without recourse to the central government?
- Is the tax burden needed to finance expenditure likely to be (1)
acceptable in terms of its consequences for economic behaviour and (2)
sustainable in terms of the ability to avoid the erosion of major tax
bases?
- Do periods of buoyant revenues and strong fiscal positions encourage
rises in expenditure that are difficult to reverse?
$OORFDWLYHHIILFLHQF\
Government spending is an important vehicle for implementing
collective choices about resource allocation and income distribution that
emerge from the political process. Several objectives behind decisions to
intervene in the market economy and their rationale can be identified.
These include the need to provide public goods, a view that merit goods
should be made more widely available than would result without
intervention, concern to influence income distribution in some way,
environmental considerations, the desire to limit the exercise of monopoly
power or to address other forms of market failure.
However, where government intervention of some kind is warranted,
reliance does not necessarily have to fall exclusively on public
expenditure, which should be reserved for cases in which it has advantages
in terms of simplicity, transparency, fairness or cost-effectiveness. As
noted earlier, intervention often involves a mix of expenditure, regulatory
arrangements, mandates and tax incentives. In addition, the government
engages in bilateral or tri-partite agreements and acts through the provision
of information and moral suasion. Such instruments may, in turn, be
reinforced by or replaced with self-regulatory codes of conduct and
standards in the private sector
9
. For example, whereas in many countries
__________
9
See for examples and case studies OECD (1997 ) and OECD (1997 ).
government agencies (often converted into state-owned enterprises) have
been widely created to provide services that are prone to abuse of
monopoly power or to ensure universal service, in others that objective has
been pursued more through regulatory mandates or administrative
guidance. Moreover, in some countries, notably the United States,
regulatory mandates and tax incentives have been designed to prompt the
private sector to provide social protection such as pensions and health care
coverage in the government’s place. The estimates described earlier
(Table 5) suggest that, at least in the area of social protection, the degree to
which public and private spending act as substitutes is substantial: while
total, LHpublic and private, policy-induced social expenditure in
Denmark, the Netherlands and the United States is essentially identical, at
around 25 per cent of GDP, the private share ranges from less than 1 per
cent of GDP in Denmark to 4 per cent in the Netherlands and more than
8 per cent in the United States.
Intervention is not always necessary or desirable and simply
ensuring that private markets work well is often the best way to pursue
objectives. Spending and other policy measures are too often undertaken
without an adequate and objective assessment of their costs and impact
10
.
In some cases this may reflect inadequate evaluation systems embedded in
the policy formulation process. But it may also reflect the tendency for
benefits of policy action to create significant political constituencies in
their support while the costs -- in terms of, say, higher taxes, interest rates
or regulatory compliance burdens -- are spread thinly over a large number
of people without impinging enough on any to generate real opposition.
Moreover, even if the balance of costs and benefits shifts over time as
economic and social conditions change, beneficiaries may still be able to
protect their interests through political action policies and programmes
11
.
On the other hand, in assessing the mix of public expenditure,
regulatory arrangements, mandates and tax incentives, it should be
recognised that regulation and tax measures are not costless alternatives to
public expenditure since these affect economic incentives and behaviour.
Indeed, a risk associated with tightened spending control is that more of
the policy agenda shifts onto off-budget mandates and other instruments
__________
10
See Martin (2000).
11
See Persson and Tabellini (2000) for a survey of normative theories of the determinants of public
expenditure.
whose incidence and effects are difficult to identify and to assess. This
shifting reliance away from spending risks reducing the transparency of
the overall set of policy interventions. Tax incentive schemes are less
transparent than expenditures and may give rise to tax planning activities,
especially if the overall tax burden is high. Government regulations that
are poorly designed impede innovation or create unnecessary barriers to
trade, investment and innovation and may have considerable costs in terms
of the capability of markets to adjust to changing circumstances.
Where public expenditure appears to be warranted as a way of
achieving objectives, evidence needs to be sought of not only
over-provision but also evidence of failure to deliver due to
under-provision. Some forces which encourage over-provision in the form
of programmes that are unnecessary or that fail to adapt to changing
circumstances are noted above. But there are also reasons why public
expenditure may be insufficient even in essential areas. The adoption of
top-down cash limits or failure to prioritise in the face of fiscal austerity
may have led to unintended rationing. In less mature market economies tax
bases may not be sufficiently exploited, or may be underdeveloped, due to
a large informal economy; the economy may have experienced a major
financing crisis with important and long-lasting social effects; or it may be
in transition from a centrally planned system. In some cases, improvements
in the framework conditions in which the private sector operates, tax
incentives, mandates or regulatory changes may be helpful. But in areas
where the government has assumed responsibility for certain activities
-- HJ publicly-run education and health care, or public goods such as
police protection and administration of justice -- there may be no
alternative to an adequate level of public spending.
In view of these considerations the following questions are relevant
for the assessment of allocational efficiency of public expenditure:
- Is government intervention warranted in all areas where public
expenditure is taking place? Is there significant scope for increasing the
role of the private sector?
- Where it is warranted, is the mix of public expenditure, regulation and
tax incentives appropriate?
- Is a country’s performance in achieving public policy goals such as an
appropriate level and equitable distribution of income, health status,
school enrolment, quality of the environment and safety commensurate
with the resources allocated to them across government functions and
programmes?
- Are there domains of economic activity or social policy where
performance is clearly below par, and is a lack of public expenditure at
the root of this problem? If so, what is the reason for any under-funding
and how can such problems be eased?
7HFKQLFDOHIILFLHQF\
The discussion so far has focused on the extent to which public
expenditure is consistent with satisfactory macroeconomic management
and performance and whether its role has been properly identified.
However, countries also have a clear interest in ensuring that public
expenditure is “technically efficient”, LH avoids waste. While
conceptually different from allocational efficiency, technical efficiency
has important implications for many of the issues raised above: avoiding
technical inefficiencies will free up available resources to help achieve
public policy goals by promoting the efficient allocation of resources
across programmes and items. At the same time, it will facilitate
macroeconomic policy management by making expenditure easier to
control. With fiscal discipline heightening, technical efficiency in the
production of public goods and services has received growing attention
among budget officials in OECD Member countries.
Unfortunately, there are many obstacles to achieving higher levels
of operational efficiency. Particularly where bureaucratic structures are
complex or responsibility for decisions is highly centralised, managers
may lack the authority to take measures that would improve performance.
In addition, entrenched work and management habits, rigid seniority-based
pay scales and strong union power in the public service may operate to
limit flexibility, for example to make the most cost-effective use of new
technology. Furthermore, the incentives for managers in the public
administration to enhance efficiency may be weak since efficiency gains
risk leading to less, rather than more, resources being available to them or
may not translate into improved pay or other advantages.
The following questions are relevant for the assessment of technical
efficiency of public expenditure:
- What evidence is available on the technical efficiency of public
spending? Is comparative information concerning technical efficiency
(benchmarking) available and is it used as an input to policy changes?
- Can areas be identified where there is significant scope for efficiency
gains?
0DLQDUHDVRIUHIRUP
This part of the paper sets out a number of broad areas in which
reforms designed to improve the cost-effectiveness of public expenditure
have been considered or implemented in several countries. It must be
recognised that many policy initiatives are unlikely to be easily
transferable across countries in view of differing political contexts. But
countries’ experience may offer a useful starting point for international
benchmarking and peer reviews. The various reform areas considered here
have been grouped under four main headings: L budgetary processes and
control; LL fiscal relations between central and lower levels of
government; LLL market-based provision and other allocation mechanisms;
and LY flexible incentive and control mechanisms.
%XGJHWDU\SURFHVVHVDQGFRQWURO
Three main aspects of budgetary processes and control are being
considered here: fiscal transparency, the adoption of medium term
frameworks and fiscal rules, and fiscal risks of financial transactions and
the wider public sector.
)LVFDOWUDQVSDUHQF\
A high level of transparency in budgetary matters -- involving
accurate and objective information on how government money is used, the
cost of government programmes and, to the extent possible, their
benefits -- provides a basis for informed debate about budgetary policy
among the public and within the government resources. By increasing the
chances that sound policy options will be identified and strengthening
political support for them, fiscal transparency is therefore likely to
encourage fiscal discipline and a more satisfactory allocation.
Furthermore, it also improves the basis for households, business and
financial market participants to make wise consumption, saving and
investment decisions. To encourage transparency, the IMF has developed
its &RGH RQ *RRG 3UDFWLFHV RQ )LVFDO 7UDQVSDUHQF\, and the OECD’s
Working Party of Senior Budget Officials is now in the process of
finalising a reference document to assist governments in making
improvements, 2(&' %HVW 3UDFWLFHV IRU %XGJHW 7UDQVSDUHQF\ (See
Box 1). Three OECD countries which have undertaken significant reforms
in many aspects of their budgetary process and management systems -- the
United Kingdom, Australia and New Zealand -- have published
self-evaluation reports under the IMF Code
12
. These reports, together with
the guidance offered by the 2(&' %HVW 3UDFWLFHV will be used as
benchmarks in assessing practices in the country under review and in
motivating proposals for improvement.
A number of countries have moved to increase the amount and
quality of information they make widely available, facilitating better
public analysis and debate. Fiscal policy statements have been introduced
in some countries in order to prompt the legislature to discuss aggregate
government finances (often in a medium-term framework, see below) prior
to the presentation of the budget itself, while Annual Reports or
Performance Reports that are separate from the Budget offer improved
outcome and output information by reporting expected and actual
performance. For example, “Value for money” reports like those prepared
by the National Audit Office in the United Kingdom have an important
role in increasing the transparency of public spending and enhancing
parliamentary oversight of the budget process. Generational accounts,
which typically show that older generations benefit at the expense of
younger ones, were introduced in the United States in 1993, followed by
Germany, New Zealand, Norway, Sweden and the United Kingdom. These
provide important supplementary information, even if many of the
assumptions which underpin them raise questions about their overall value.
Finally, greater resources (e.g. in the form of supporting secretariats) are
__________
12
Greece has also completed a report and is expected to publish it soon.
%R[%HVW3UDFWLFHVRQ)LVFDO7UDQVSDUHQF\
The IMF’s &RGH RQ *RRG 3UDFWLFHV RQ )LVFDO 7UDQVSDUHQF\ is
based on four principles:
– &ODULW\ RI UROHV DQG UHVSRQVLELOLWLHV establishing clear boundaries
between the public and private sectors; and within the public sector
between fiscal, monetary and government business enterprise
activities.
– 3XEOLF DYDLODELOLW\ RI LQIRUPDWLRQ -- LH a commitment to publish
comprehensive financial information at clearly specified intervals.
– 2SHQ EXGJHW SUHSDUDWLRQ, execution and reporting, according to
published statistical and accounting standards for government
reporting.
– ,QGHSHQGHQWDVVXUDQFHVRILQWHJULW\-- HJ. through external audit and
statistical independence.
The draft list of 2(&'EHVWSUDFWLFHV includes the following items:
– Governments should publish a pre-budget statement outlining the
aggregate levels of revenues, expenditure, surplus or deficit and debt
several months prior to the release of the government’s budget
proposal. The objective is to cast budget policy in a macroeconomic
and medium-term setting, thereby establishing a top-down fiscal
policy anchor.
– The budget should contain explicit detail on the economic
assumptions used and statements of tax expenditures, financial
liabilities and financial assets, non-financial assets, employee pension
obligations and contingent liabilities. Several tracking and update
reports should be available. These could include monthly out-turn
reports and mid-year updates.
– The annual financial statements (or government accounts) serves as
a compliance report for parliamentary and wider accountability
purposes and should be certified by the auditor. Transparent financial
statements should include information on the budgetary out-turn, debt
structure and borrowing, commitments, contingent liabilities, trust
moneys held by the government and accounting policies.
7DEOH
$FFRXQWLQJV\VWHPVLQ2(&'FRXQWULHV¶FHQWUDOJRYHUQPHQWV
Full cash basis Full accrual basis
Accrual basis except
for capital expenditure
Both full cash basis
and full accrual basis
Cash basis except
for certain transactions
Budgeting Reporting Budgeting Reporting Budgeting Reporting Budgeting Reporting Budgeting Reporting
Austria Austria Australia Australia Canada
1
Canada Italy Italy United States
2
Finland
3
Czech Rep. Czech Rep. New
Zealand
Greece Denmark
4
Denmark
4
Sweden
5
Fr ance
6
France Germany New
Zealand
Iceland Iceland Finland
3
Poland
5
Germany Hungary Sweden
Greece Ireland United
States
7
Hungary Japan
Ireland Korea
Japan Mexico
Korea Netherlands
Mexico Norway
Netherlands Portugal
Norway Spain
Poland Switzerland
Portugal
1
Turkey
Spain United Kingdom
Switzerland
Turkey
United Kingdom
1
Moving to full accrual budgeting.
Interest on government debt, certain civil service pension plans, loan and guarantee programmes are on an accrual basis.
Salaries and wages are on an accrual basis.
And certain other minor exceptions to full accruals.
Civil service pensions are on an accrual basis.
Interest on government debt is on an accrual basis.
“Non-exchange revenue”, most of which is taxes, is recognised on a modified cash basis.
OECD, based on country submissions.
Figure 3.
1XPEHURI2(&'FRXQWULHVWKDWVSHOORXWWKHLUPHGLXPWHUPILVFDOREMHFWLYHV
Source: OECD, based on country submissions.
13
10
13
8
9
12
12
16
17
other
Debt in nominal terms
Revenue in nominal terms
Deficit/surplus in nominal terms
Expenditure in nominal terms
Revenue as a % of GDP
Expenditure as a % of GDP
Debt as % of GDP
Deficit/surplus as a % of GDP
Number of countries
)LJXUH
being devoted by the legislature to evaluation of the budget (e.g. as France,
Italy, Mexico and Sweden; the United States has for many years devoted
large resources in this area). Other aspects of transparency are the
publication of rules on the granting of subsidies and transfers (e.g. as
required by EU rules), and the disclosure of public procurement practices
(e.g. as in the United Kingdom).
An important element of fiscal transparency is an accounting system
that delivers as fair and accurate a picture, on a consistent basis at both
budgeting and reporting stages, of the impact of the government’s
activities on its overall financial positions as is possible. This has led
several countries -- Australia and New Zealand have moved furthest
(Table 6) -- to make increased use of accrual accounting methods
13
.
Accrual accounting recognises the financial implications of transactions
when they occur, irrespective of when cash is paid or received. Traditional
cash accounting, in contrast, can more easily lead to a misleading picture
of commitments undertaken when payments can be accelerated or
deferred. This makes it an unsatisfactory basis, at least by itself, for
monitoring recent developments or for the assessment of long-term
sustainability of public finances. Important differences include the
recognition under accruals systems of L capital costs through charging for
depreciation; LL accruing interest obligations on discounted or zero
coupon debt instruments; and LLL future commitments accrued under
pay-as-you-go civil service pension plans.
0HGLXPWHUPIUDPHZRUNVDQGILVFDOUXOHV
Many OECD member countries (Japan is an important exception)
have adopted medium-term frameworks for aggregate government
spending, usually covering three to five years, and support this with
medium-term objectives for one or more fiscal variables (Figure 3). This
development stems from the recognition that annual budgeting may
exacerbate the natural short-term focus of political decision-makers and
cause authorities to lose sight of future costs of decisions, the best
__________
13
As budgetary management philosophy has shifted towards encouraging decentralisation of
day-to-day decision-making, accrual accounting has also served as a management tool by
providing a better basis for accountability than cash accounting. More is said about this in Part 4.3
below. Greater budget transparency therefore is likely to support improved technical efficiency in
addition to its benefits in terms of fiscal discipline and allocative efficiency.
allocation mix and the appropriate timing of expenditures. Medium-term
frameworks, moreover, aim to anchor annual expenditure appropriations in
medium-term projections. They oblige governments to recognise the
implications of current budgetary decisions for government finances in the
future and to take account of changes in structural and demographic
factors and rising government debt levels, as well as the evolving cyclical
situation. At the same time, they limit inefficiencies that arise from annual
appropriations for multi-year capital projects. To be successful in
facilitating expenditure control and fiscal discipline, it is important that
these frameworks be supported by systems for evaluating spending
programmes objectively and they are carried out on the basis of realistic
economic assumptions, as these are a major determinant of the overall
“affordability” envelope.
In several countries rules have been adopted that automatically
trigger sanctions when certain targets or ceilings set by the medium-term
framework are breached. The best-known example is the Budget
Enforcement Act (BEA) of 1990 in the United States, which formulates
caps on spending which, once they have been accepted by elected officials,
are enforced by requiring any extra spending to be offset by spending cuts
without reference to the overall fiscal position. This is widely seen as
having contributed significantly to improved fiscal discipline, although
spending caps are proving more difficult to enforce now that surpluses are
mounting. An earlier, and somewhat different, rules-based approach in the
United States was incorporated in the Gramm-Rudman-Hollings (GRH)
laws of 1985 and 1987. These specified deficit targets, rather than
expenditure targets, which were to be enforced by “sequestration”
(uniform percentage reductions) in selected spending programmes. GRH
was ultimately discredited because the objective of declining deficits was
repeatedly deferred, as it was not reinforced by agreements about where
the necessary adjustments should take place and the violations of deficits
targets were substantially influenced by factors outside the control or
influence of the political process (HJrecession). As a result sequesters
required to eliminate deficit increases were very large and politically
unfeasible. Another example of a rules-based fiscal framework, also
formulated in terms of budget balances and not expenditure alone, is the
Stability and Growth Pact (SGP) in the European Union. This has a
distinguishing feature in that it uniformly applies to several countries at
once, which may encourage compliance since failure to meet an
international commitment is more difficult than just announcing a change
in domestic policies. Like the GRH laws, the SGP’s focus on budget
balances makes compliance vulnerable to unexpected changes in the
cyclical position of the economy, so its success requires that policies in
most times be designed to achieve positions significantly better than the
deficit ceilings that lead to sanctions (ultimately, in this case, fines).
A limitation of rules-based approaches is that unless a Parliament is
constrained by constitutional limits, a government cannot commit either
itself or a successor to a future course of action. The persistent deferral of
action under the Gramm-Rudman-Hollings laws, without bringing the
sequestration provisions into effect, is a case in point. This has led some
countries to a less ambitious alternative which legislates principles rather
than rules, emphasises transparency and relies on these to frame public
debate and encourage market discipline in a way that pushes government
to respect legislated principles. The most comprehensive effort in this
direction has been made in New Zealand, where the Fiscal Responsibility
Act sets out “principles of responsible fiscal management”. No specific
targets are set in the Act, but it obliges the government to explain any
departures from legislated principles, how long they will persist and how it
intends to return to these principles. Similar legislation has been
introduced in the United Kingdom (Code for Fiscal Stability) and in
Australia (Charter of Budget Honesty).
)LVFDOULVNVRIILQDQFLDOWUDQVDFWLRQVDQGWKHZLGHUSXEOLFVHFWRU
The operations of public sector entities not subject to the constraints
that arise in the normal budget processes affect public finances and
resource allocation in the economy more widely. Many countries, at one
time or another, have had bad experiences with these as poor performance
has led problems to build up over time whose consequences eventually had
to be recognised on the budget, if only in terms of higher debt servicing
costs. Prudent management of public finances therefore requires
comprehensive attention to the whole public sector.
There are a number of operations which usually fall outside the
budget process that entail risks to public finances
14
, but five in particular
stand out:
__________
14
See Blejer and Cheasty (1991) for an overview.
-First, H[WUDEXGJHWDU\IXQGV may be created to circumvent the ordinary
budget process, say to implement financial support quickly. These tend
to reduce fiscal transparency and it may be difficult to exercise
oversight over their expenditures. A well-known example is the
complex build up of unification-related funds in Germany before they
were taken on-budget in 1995.
- Second, VWDWHRZQHG HQWHUSULVHV (SOEs) have often performed badly
and proved to be a drain on public finances. Sometimes the problems
have been reflected merely in poor returns on capital but there have
been many instances of SOEs requiring subsidies (often disguised as a
“capital injection”), or the need to take over debts as part of financial
restructuring (the assumption in Japan of the debt of the Japan Railway
Settlement Corporation and the National Forest Special Account in
1998, amounting to 5.4 per cent of GDP, is an example). Many
countries have made progress --the United Kingdom and New Zealand
stand out but Mexico, Australia and many euro area countries have also
made important advances -- reorganising and changing governance
arrangements to improve the operations of SOEs by subjecting them
more fully to market disciplines by privatising them wholly or partially.
Nevertheless, improving the management of SOEs remains an
important challenge in Turkey, the European transition countries and
some EU countries, while much of the restructuring that has occurred,
for example in the airline industry, has yet to be tested by a recession.
- Third, VWDWHRZQHG ILQDQFLDO LQVWLWXWLRQV are normally off-budget but
generally come under government direction. This may result in lending
at reduced interest rates or investment in assets not selected on the basis
of sound market considerations. Even where the government avoids
applying direct pressure, disciplines on management are often weak. In
a number of countries state ownership has often led to fiscal problems.
In countries as diverse a France and New Zealand, the insulation from
market disciplines that arises with state ownership and difficulties
ensuring effective supervision led to the collapse of major banks
(Crédit Lyonnais, Bank of New Zealand) with ultimately large fiscal
consequences. The trend in OECD countries has been almost
universally toward greater reliance on market forces and disciplines in
financial markets, which should limit the risks in the future, but state
ownership is still significant in some countries and, in some, provides a
vehicle for circumventing normal budgetary processes. In Mexico,
notably during the run-up to the 1994 election and the subsequent
financial crisis, and in Turkey state-owned financial institutions have
effectively been used as instruments for fiscal expansion. In Japan, the
channelling of retail savings deposits from the Post Office through the
Trust Fund Bureau to state-owned enterprises -- although subject to
Parliamentary approval -- is not transparent in terms of its impact on
public finances and questions exist as to the quality of the resulting
allocation of resources.
- Fourth, FRQWLQJHQW OLDELOLWLHV DVVRFLDWHG ZLWK JXDUDQWHHV normally
affect the budget only when cash payments are required. The variety of
these liabilities is wide. Examples include: insolvent or under-funded
deposit insurance systems (the Savings and Loan collapse in the United
States); the programme of guarantees in Japan for lending to small and
medium-sized enterprises; the commitment in the United Kingdom in
the late 1960s and 1970s to guarantee the dollar value of certain
sterling liabilities; and New Zealand government guarantees on
international loans to finance a series of major projects designed to
cope with high oil prices that were expected to rise further in the early
1980s. Some have proved very costly.
- Fifth, the PDQDJHPHQW RI RIILFLDO ILQDQFLDO DVVHWV DQG OLDELOLWLHV may
result in capital losses or gains. This primarily concerns foreign
exchange reserves and government debt. However, financial asset
portfolios may also be established to cover specific obligations such as
public employee pension reserves (Canada) or catastrophic losses (New
Zealand), and various lending programmes may also generate financial
assets of considerable value. Since the amounts involved are large, the
financial risks they entail -- particularly in the areas of foreign currency
exposure for countries with large foreign reserves or substantial
government debt raised in foreign currencies -- are also large. Two
general principles have been proposed in this area. First, financial
management should be determined in the context of the government’s
overall financial position. This should involve taking account of
reasonably expected future cash flows, and requiring that any subsidy
element in government lending programmes be identified and taken
into account. Second, the government should construct its overall
financial asset and liability portfolio to hedge permanent shocks to its
financial position, LHpursue an “insurance” objective rather than an
independent return objective. Implementing such a strategy require a
centralised policy-setting function, although centralising operational
functions is not likely to be appropriate.
)LVFDOUHODWLRQVEHWZHHQFHQWUDODQGORZHUOHYHOVRIJRYHUQPHQW
The public sector of nearly all countries includes more than one
level of government and there is a wide variation across countries in
relations between the different levels, both in terms of allocation of
responsibilities and financing arrangements. Achieving effective
management of total public expenditure is greatly facilitated if these
responsibilities ensure that decision-making authority rests where it can
best be exercised and if these financing arrangements ensure that spending
decisions take account of the full costs that they entail.
A number of OECD countries have either confronted the need for
change in their arrangements in the recent past or are now doing so. In the
United States longstanding federal entitlement programmes entailing
detailed rules and matching grants have recently been replaced with a
system of block grants to enable the states to provide a number of social
services and develop social transfer programmes on a local basis. While
there are risks of migration and a “race to the bottom” undermining this
policy, it represents an effort to deal with the widely-recognised failure of
previous federal programmes
15
. Considerable experimentation is involved
here, but it is hoped that innovative programmes at local levels may prove
to offer models that operate more effectively, and as such contribute to
enhance the efficiency of social policy. In the United Kingdom, Scotland
and Wales have recently opted for their own local parliaments; and in
Spain the regional financing system was revised in 1997 with a view to
better matching spending responsibilities with revenue raising powers of
the regions. In both cases, it is too early to assess the effects. The
European Union incorporated the “principle of subsidiarity”, LHthat
public policy and its implementation should be assigned to the lowest level
with the capacity to achieve objectives, in the Maastricht Treaty as a guide
for future integration efforts. Nevertheless, debate has continued on the
extent to which defence and social policy should be moved from national
governments to the EU, or federal, level.
__________
15
Oates (1999).
According to the basic principles of fiscal federalism, the central
governments should have the responsibility for the macroeconomic
stabilisation and income redistribution functions. Local governments
simply lack the means for macroeconomic control while the spatial
mobility of economic units limits the scope for regions to redistribute
income -- as greater ambition in this field would risk encouraging an
exodus of wealthy citizens and an influx of poorer ones. In addition to the
stabilisation and redistribution functions, it is natural for the central
government to provide certain “national” public goods (like national
defence) that provide services to the entire population of the country. In
contrast, local governments are often well placed to ensure the provision
of certain merit goods and services, particularly where consumption is
limited to their own jurisdictions. By matching the supply of such goods
and services with the particular preferences and circumstances of their
constituencies, local provision may raise economic welfare above that
which results from more uniform levels of such services that are likely
under national provision. The empirical record broadly confirms these
patterns of vertical distribution of tasks (Table 7).
So long as any significant decision-making responsibility for
expenditure is devolved to lower levels of government it is important that
they face a hard budget constraint. Otherwise, the incentives could well be
for them to spend excessively, and overall fiscal discipline may be difficult
to ensure. Since they do not have access to monetary instruments of public
financing, in principle they do face a hard budget constraint
16
. However,
for this constraint to be binding, lower level governments should not be
able rely on transfers from above to bail them out of fiscal difficulties and,
at the margin, they should be required to fund their own expenditures fully
through local taxes or by borrowings whose debt servicing they have to
ensure themselves. Where it is not deemed possible to make lower
governments responsible for their own relations with financial markets or
to allow them to suffer the consequences of mismanagement without being
rescued, there is little alternative to retaining tight control at the central
level over their spending and borrowing
17
. Efficient exercise of such tight
control at the central level, in turn, requires good provision of information
to the central authorities and strong financial reporting systems.
__________
16
See Eichengreen and Von Haagen (1996).
17
See for this argument OECD (1996), pp. 12-8.
The desirability of ensuring that lower levels of government face a
hard budget constraint to the extent that they have decision-making
authority over expenditures does not preclude providing central
government financial support for activities carried out by other levels of
government. Several rationales for such support exist:
7KH LQWHUQDOLVDWLRQ RI VSLOORYHU EHQHILWV WR RWKHU MXULVGLFWLRQV.
Conditional, or “matching”, grants are best employed to fund the
provision of local services, which generate benefits for residents of
other jurisdictions. It is important that these be structured with clear
limits in order that they not turn into entitlements that undermine the
hard budget constraint. A possible alternative way to deal with spillover
effects, but which may be politically difficult, is to enlarge the
geographical extent of local jurisdictions to internalise all the benefits
and costs, HJby bringing central cities and suburbs into a single
jurisdiction (Toronto provides a recent example).
)LVFDO HTXDOLVDWLRQ DFURVV MXULVGLFWLRQV. Unconditional or “block”
grants are typically the appropriate vehicle for purposes of fiscal
equalisation -- LHto channel funds from relatively wealthy
jurisdictions to poorer ones. Such transfers, which are often based on
an equalisation formula that measures the “fiscal need” or “fiscal
capacity” of each jurisdiction, play a major role in countries such as
Germany, Canada and Australia, and can be justified by equity
considerations. From an efficiency perspective they raise questions,
however, since they may impede changes in cost differentials and flows
of resources that regional adjustment requires.
$PRUHHTXLWDEOHDQGHIILFLHQWRYHUDOOWD[V\VWHP. Central government
general taxes with a single (progressive) rate structure applying to the
whole nation are less likely to create fiscal incentives for relocation.
This would thus argue for “tax sharing”, as in countries such as
Germany, Austria, Mexico and Norway, under which tax bases and rate
schedules are defined on a nation-wide basis while the proceeds are
split between the central government and local constituencies.
7DEOH
Table 7.
Per cent of functional category
General
administration
Law and order Security Education Health Welfare Housing Leisure
Transport and
communication
Central Local Central Local Central Local Central Local Central Local Central Local Central Local Central Local Central Local
Federal countries:
Australia 52 48 15 85 100 0 28 72 51 49 91 9 32 68 29 71 22 78
Canada 59 41 0 100 100 0 8 92 17 83 66 34 19 81 16 84 31 69
Germany 46 54 0 100 100 0 5 95 71 29 77 23 5 95 4 96 50 50
United States 70 30 16 84 100 0 6 94 55 45 73 27 67 33 16 84 29 71
Switzerland 40 60 0 100 84 16 10 90 43 57 82 18 8 92 7 93 37 63
Average
2
53 47 6 96 99 1 11 89 47 53 78 28 26 74 14 86 34 66
Unitary Countries:
2
Denmark 6436861499 15149 89247534159425852 48
France 70307624100 0633797 392 82080277360 40
Norway 67338218100 04951445681194060386264 36
Netherlands 69 31 67 33 100 0 80 20 85 15 82 18 38 62 14 86 55 45
United Kingdom 78 22 58 42 100 0 0 100 100 0 91 9 0 100 0 100 67 33
Average
3
70307426100 04951673378222773247660 40
1. 1992, or latest available. In all countries, central government includes the social secutiry system while local expenditure shares do no always reflect local decision making power.
2.
Local includes sub-central (state) levels.
3. Unweighted.
Pola (1999).
0DUNHWEDVHGSURYLVLRQDQGRWKHUDOORFDWLRQPHFKDQLVPV
One way to support efforts to keep aggregate expenditure in, is to
enhance the allocative efficiency of public expenditure. This raises a
number of issues that are covered below, notably the choice between
targeting or universal provision, the greater use of market mechanisms for
provision and funding, and a market-based approach towards public
infrastructure investment.
7DUJHWLQJYHUVXVXQLYHUVDOSURYLVLRQ
As noted, social spending in the form of transfer payments and
provision of merit goods, mainly in the areas of income support, health and
education, is a very large share of government expenditure in most
countries. Issues of the desirability of targeting versus wider free or
subsidised provision of benefits or services arise in virtually all domains of
social policy. Targeting has the advantage of ensuring that the funds used
are provided cost-effectively since it allows the most serious problems
arising from low income or lack of access to be addressed at reasonably
modest cost. In countries such as Australia and New Zealand, where social
security systems give priority to assistance schemes designed to protect the
poorest groups, targeting plays an important role in an effort to reconcile
fiscal discipline and equity objectives. Targeting, however, is not without
problems. The means testing that it often involves implies high marginal
effective tax rates, especially on labour income, as the withdrawal of
benefits may proceed rapidly and limit the gains from work once income
rises beyond the means threshold. Where targeting is applied, therefore,
care must be taken to minimise the extent to which it discourages work
effort and creates poverty traps.
While there is considerable variation across programmes and across
countries, most income transfer programmes and other forms of social
services are based more along general insurance lines, with eligibility for
benefits widespread or even universal. When building programmes along
insurance lines and widespread access is an important objective, reducing
intended population coverage is not a feasible approach to controlling
costs. However, there may be scope for reducing moral hazard associated
with the programmes and for better monitoring of beneficiaries and their
fulfilment of eligibility criteria
18
.
Although the general case for public intervention in the financing of
merit goods is clear, the boundary between where this intervention should
end and where users should bear the full costs themselves is not. The
issues may vary across different types of merit goods (Box 2 illustrates
some of the issues that arise in the case of tertiary education), and different
countries will draw the boundaries differently. Many of the issues involved
come down to the extent to which users can pay for these services and
appropriate the benefits but attitudes towards different concepts of equity
are also important. In this regard, many of the considerations are similar to
those influencing attitudes to the targeting of benefits. The demand for
goods and services which fall close to the boundary in many countries,
such as cosmetic and lifestyle-enhancing medical care, child care,
long-term care for the disabled and personal services for the elderly may
prove to be highly elastic as societies age, labour force participation rises
and technology advances. Assessments of where the limits to public
financing of merit goods should be will have to be founded on hard
evaluations of their costs and their impacts on government finances.
Fourth, where merit goods and services are to be publicly financed,
governments must address the issue of how the delivery of the goods and
services can be achieved most cost-effectively. The major areas are health
and education, but similar issues arise with other merit goods. One general
consideration is the need to ensure that the design of the eligibility
conditions corresponds to equity concepts influencing the decision to
provide public financing. However broadly or narrowly these conditions
are defined they need to be maintained and enforced. Beyond this the main
issues are similar to many of those raised elsewhere in Part 4 of this paper.
(QKDQFLQJWKHUROHRIPDUNHWPHFKDQLVPV
Many governments have found it increasingly useful to test the
boundaries of government provision. They question whether certain goods
and services are so distinctive that their provision must remain in the
public sector and, where provision remains public, have developed
rationing mechanisms which allow targeting the provision of such goods
__________
18
For a thorough discussion of possible reforms of social transfer programmes see MacFarlan and
Oxley (1996).
and services within a commercial setting without compromising
fundamental policy goals. A number of examples follow.
&RQWUDFWLQJRXW has been tried both in the provision of local
consumer goods and services (HJ operating city bus services, waste
collection and child-care services) and in the purchase of inputs for public
sector agencies (HJ maintenance and cleaning of public buildings,
information technology and financial services). It is not always easy to
write the contract and manage it effectively, but where these difficulties
can be overcome contracting it has often led to substantial savings to the
public purse and an improvement in the quality of the services provided.
19
Even when production ends up being retained in-house, the efficiency of
public sector agencies is likely to improve through an effective “threat” of
competition outside.
OECD Member countries are coming under increased pressure to
liberalise SXEOLF SURFXUHPHQW markets
20
. While some countries have
become more receptive, many others are still restrictive in opening their
public procurement market to foreign suppliers. Moreover, since goods
and services purchased by the government often cannot be delivered
“off-the-shelf”, cost-plus contracts tend to prevail. This weakens the
incentives for producers to prevent cost over-runs and delays. In the key
area of defence procurement there has been a movement away from
cost-plus contracts towards competition among a selected number of
suppliers and contracts where suppliers take some part of the risk of cost
over-runs. There has also been some move away from preferential
purchasing arrangements. However, complicated and opaque procedures
for tendering persist, which give rise to serious entry barriers and raise the
bargaining position of “insider” suppliers. In order to level the playing
field both among (potential) suppliers and between suppliers and the
__________
19
Examples surveyed in OECD (1997e) range from social policy functions such as residential
treatment homes for children with behavioural and emotional problems (Iceland) and case
management services for the unemployed (Australia) to skilled professional services such as audit
functions (New Zealand Audit Office), information technology functions (Inland Revenue
Department, in the United Kingdom) and airport management (City of Indianapolis), to low
technology operations such as cleaning services (National Hospital, Copenhagen) and catering
operation (Turkish Ministry of Finance). Since evidence is accumulating that contracting out can
lead to efficiency gains while service quality levels are maintained or improved, its use is generally
increasing.
20
For example, the European Union has already issued directives that have formally liberalised
public procurement and made tendering transparent, although import penetration of publicly
procured goods generally remains low. In addition, the international framework regarding public
procurement has also been strengthened in the WTO.
government, best practice principles as those for contracting out also apply
to public procurement. In particular, the team responsible for the purchase
should maintain careful scrutiny of cost, possess the technical skills for
overseeing the quality of goods or services delivered, and be held
accountable
21
.
8VHU FKDUJLQJ has become widespread with the objective of
reducing excess demand and improving public services through the
introduction of market signals
22
. User charges aim to create a sustainable
basis for revenue raising to finance certain services, while relieving the
general taxpayer of costs properly born by the users who benefit directly
from them. The discipline this imposes on users promotes allocational
efficiency and, by subjecting the government organisations providing
services to a market test user charging is expected to encourage
customer-oriented management and improve the financial and service
performance of the public supplier. However, social considerations may
limit the extent to which setting user charges in line with costs is
acceptable and user charging will be viable only if the transaction costs of
collection of charges are lower than the efficiency gains that result from
market-type provision.
9RXFKHUV constitute an emerging instrument for the distribution of
merit goods and services in a number of OECD countries (Box 3). They
aim to remove undesirable distributional effects associated with user
charging and/or private provision. Through vouchers individuals receive
entitlements to a good or service which they may “cash in” at some
specified set of suppliers, which redeem them for cash from a funding
body. The value of vouchers can be varied in order to pursue distributional
objectives and/or to target the aid to specific groups. Designation of the
recipient ensures that they are not tradable across consumers and
designation of the services that they are not equivalent to cash.
__________
21
OECD (1994).
22
The range of government services which can be subject to user charging that covers all or part of
the costs of providing them is wide, and several case studies are reported in OECD (1998
). The
United States Nuclear Regulatory Commission is now fully financed by user charges. In
Barcelona, the Fire Department charges for its services. Its motivation is to increase public
awareness of the need to maintain facilities and buildings properly, and in fact it only levies
charges when there is evidence of negligence. When the Attorney-General’s legal practice in
Australia moved to a user charging regime client service improved dramatically. Other examples
are numerous, usually for services which for one reason or another full commercialisation of the
activity, say by creating a state-owned enterprise or by privatising it altogether, is not feasible.
%R[:KRVKRXOGSD\IRUWHUWLDU\HGXFDWLRQ"
The part of educational expenses on tertiary education
covered by individuals or other private sources varies widely
throughout the OECD -- from negligible amounts in many European
countries (HJ in Denmark, Sweden and Austria) to more than
50 per cent in Korea and Japan.
A priori, those who benefit from higher education
1
should
and, financial markets permitting, could pay for it. Therefore, a high
share of private financing does not necessarily lead to low
investment in tertiary education. On the contrary, in some OECD
countries with high spending on tertiary education relative to GDP,
the share of private finance is among the highest (HJ in the United
States and in Korea; see OECD, 1998E). Tertiary education does not
have the characteristics of a public good, as there is some rivalry in
consumption and consumers are excludable. The fact that the bulk
of tertiary education is nevertheless publicly financed in most
OECD countries therefore seems to be motivated by the existence of
substantial externalities, (other) market imperfections or certain
policy objectives associated with higher education.
3RVLWLYH H[WHUQDOLWLHV RI WHUWLDU\ HGXFDWLRQ" If tertiary
education leads to positive externalities, the market would provide
less than efficient amounts. However, the degree of these
externalities is controversial. Some have emphasised a positive
impact on productivity. Johnson (1984) argued, for example, that
even the low-skilled might benefit from subsidising higher
education if high- and low-skilled work are complementary
2
. Others
have stressed externalities beyond potential increases in GDP, such
as greater social cohesion, reduced crime rates and more
appreciation for cultural goods. Furthermore, it is often suggested
that investment in tertiary education would alleviate employment
problems and contribute to a necessary increase in the qualification
of the workforce. Nevertheless, it is relatively difficult to identify
(or even measure) “pure” externalities that would neither accrue to
the individual nor to its current or future employers and that could
therefore not be reflected in present or prospective wages.
&RUUHFWLQJ RWKHU PDUNHW IDLOXUH" While wage negotiations
could in principle internalise the benefits -- and thus achieve
efficient levels of provision if no pure externalities are present --
market imperfections might inhibit these outcomes
3
. Among these
obstacles are the risk and uncertainty that surround human capital
investment, liquidity constraints for low-income households
4
and
information asymmetries (HJ hidden knowledge). Many of these
impediments are particularly pronounced for high-cost studies
(HJ sciences, engineering, medicine). Furthermore, certain forms of
non-university education that require co-operation with employers
might not be provided in an optimal amount, as employers could
refrain from training in skills that would benefit the employee or
future employers in case of a workplace change. Labour market
restrictions (HJ limited scope for productivity-based pay or for
individual contracts) could also prevent efficient levels in the
absence of public support
5
.
&RQWULEXWLRQWRDFKLHYLQJHTXLW\JRDOV" The impact of public
support for tertiary education on various concepts of equity is
ambiguous
6
. On the one hand, it facilitates the access of low-income
households, which might otherwise not be able to afford higher
education, thereby promoting social mobility. On the other hand,
those who tend to profit most from post-secondary education have
already a relatively high level of education. Furthermore, the
average recipient can expect an above average lifetime income. In
addition, low social and income groups tend to participate
under-proportionally
7
. However, empirical evidence indicates that
increases in net private post-secondary education cost lead to
decreases in enrolment rates for lower income students (for the
United States, see HJ McPherson/Schapiro, 2000). Finally,
financing of public expenditure on education through general
taxation leads to a transfer from those having no children to
families. It could be argued that this contributes to equity if the cost
of raising children is not fully compensated by other family/children
support, though not all families profit from tertiary education.
7HUWLDU\HGXFDWLRQDVDPHULWJRRG" The merit good argument
is often seen as justifying the generally free and compulsory
provision of primary and secondary education. While it seems
reasonable to argue that children might not be fully aware of the
benefits, it is much less clear why this argument should apply to
higher education as well. After all, the choice of whether or not to
pursue post-secondary education is SHUVH only open to people with
considerable prior education, some of which might have already
experienced the benefits of education in general. Furthermore, even
if higher education is considered to be a merit good, it is not clear
whether high levels of government support lead to an increase in the
aggregate consumption of this good
8
.
7KH LPSDFW RI WD[HV DQG XQHPSOR\PHQW EHQHILWV Taxes in
general, and progressive income taxation in particular, introduce a
wedge between private and fiscal rates of return
9
. Higher education
leads to higher income and, thus, net additional tax revenue. This
might, even when discounted, outweigh the public costs of financing
tertiary education. Thus, financing tertiary education could be an
acceptable investment from a pure budgetary point of view, even in
the absence of externalities
10
. As higher education enhances the
opportunity cost of not working, it could furthermore lower the
adverse effects associated with unemployment benefits.
The above analysis indicates that some government
expenditure on tertiary education seems to be warranted, though it is
questionable whether the current high levels in many OECD
countries can be justified by these arguments. In addition, public
support is often not targeted to the above-mentioned failures that
could justify intervention. Furthermore, public financing also entails
inefficiencies, particularly in the case of generally free provision. In
particular, the signalling function of the tertiary education market is
severely undermined and neither the study combinations desired by
students nor those demanded by the labour market nor those that
provide the most externalities are reflected in the price to the
students. This also entails the danger that part of the support for
tertiary education might be captured by the education institutions
and that flexibility in the tertiary education market is severely
hampered.
OECD member countries have responded to these challenges
by introducing new financing approaches that move from free
provision and grants towards tuition and loans, while trying to
influence student behaviour to make tertiary education more
cost-effective and ensuring that the participation of low-income
households is not discouraged. These approaches include LQWHUDOLD
the time-limiting of student aid (HJ the Netherlands, Finland),
means tested tuition fees (HJthe United Kingdom),
income-contingent student loan repayment (HJ New Zealand) and
differentiated student contributions by field (HJ Australia)
11
.
Nevertheless, most of these approaches also entail inefficiencies
(HJ they might lead to high marginal effective tax rates), and may
not be equally apt for other countries. In any case, the policy
implications vary with the type of imperfection that public support
is intended to correct. If, for example, capital market access of
low-income households is the primary problem, providing student
loans would be a better remedy than free provision for all groups
(see HJ Creedy, 1995).
________________________
1. The terms “higher education”, “tertiary education” and “post-secondary education” are used
interchangeably in this box. For a definition and distinction of the latter two terms see OECD
(2000).
2. The increase in productivity would then normally lead to higher wages for both high- and
low-skilled workers.
3. For an overview of the discussion, see e.g. Stern and Ritzen (1991).
4. Financial institutions might be particularly reluctant to lend to these households since the
building of human capital is not separable from the effort of the individual and not well
observable.
5. Furthermore, minimum wage restrictions might inhibit contracts in which low-skilled workers
pay for the benefits of their training through reduced pay.
6. Tsakloglou and Antoninis (1999), for example, provide empirical evidence for Greece, where
tertiary education is provided free of charge (according to the Greek constitution). They
conclude that the distributional impact of providing free tertiary education is negligible and
could even be regressive.
7. See OECD (1997d) and OECD (1999d) for a discussion of access and equity issues in tertiary
education.
8. Testing whether or not public support meets the merit objective, Becker (1974) assumes a
reciprocal interdependence between taxpayers and tertiary education recipients (i.e. the
behaviour of each actor influences the decision of the other) and concludes that public
spending could even lead to a decrease in overall spending and consumption of higher
education. However, Arcelus and Levine (1986) assume reciprocal interdependence and arrive
at the opposite conclusion.
9. Fiscal rates of return are calculated on the “life-time value of additional income-tax receipts
and employee social-security contributions less social transfers, for those who complete
university education, compared with the public costs of educating of a university student and
the taxes lost on earnings forgone during the time of study” (OECD, 1998c).
10. This however would not be socially optimal. OECD (1998c) provides estimates of fiscal rates
of return to university level education for males and females in seven OECD countries, ranging
from four per cent for women in Sweden to 13 per cent for their Belgian counterparts. Still,
government spending on tertiary education does not per se lead to higher net fiscal revenue, as
at least a part of tertiary education might have been undertaken as well in the absence of
financial support.
11. For an overview, see OECD (1998 ).
%R[9RXFKHUVDVDPHDQVRIGLVWULEXWLQJSXEOLFVHUYLFHV
Voucher systems are regimes in which individuals receive
entitlements to a good or service which they “cash in” at some specified set
of suppliers, which then redeem them from a funding body. They may be
explicit or implicit, but must provide a margin of choice to some or all
consumers. They potentially have considerable flexibility as a device for
allocating public services as they can be made universally available,
means-tested, or structured to permit or prohibit top-ups for particular
recipients. Moreover, by providing targeted groups of consumers with
purchasing power, vouchers facilitate and complement the introduction of
user charges. However, concerns exist that they may work against the
objective of co-ordinated provision since they encourage competitive
behaviour by suppliers. Furthermore, because they increase the choice they
allow users of public services, vouchers often elicit resistance from
established providers. Given these considerations, voucher schemes are
often controversial and experience with them has been limited. Most of the
concrete examples come from the United States and the United Kingdom.
These cover areas such as primary and secondary schooling (several
experiments in US municipalities; widespread reforms in England and
Wales culminating in the 1988 Education Reform Act), non-compulsory
education and training (“Youth Credits” in England and Wales, later
replaced by “Learning Credits” in the United Kingdom), higher education
(the introduction of competitive tendering into the allocation of block grants
in the United Kingdom in the 1980s), food stamps (United States) and
social care (the Independent Living Fund for the severely disabled in the
United Kingdom). Overall the results have been mixed and further
experimentation will be needed to separate what works well from what does
not. But some conditions for vouchers to realise efficiency gains appear to
be the following:
- Some degree of competition between providers should exist. If not, the
rationale for vouchers disappears. Therefore, vouchers are less suited for
public services that by their nature must be provided in the geographical
vicinity of the consumer as this entails a risk of local monopoly.
- Since capacity constraints create a “sellers market” that may prompt
providers to adopt or maintain non-price-rationing mechanisms, the
authorities need to make sure that the provision capacity is adequate.
- It may be useful in some cases to allow vouchers to be topped up with
out of pocket payments. This may be efficient to the extent it allows
price differentiation and competition among providers and facilitates a
mix of private and public provision.
0DQDJLQJSXEOLFLQIUDVWUXFWXUHSURMHFWV
Significant amounts of resources, both in budgetary and
economy-wide terms, are devoted to capital expenditure by governments
(Figure 4) and state-owned enterprises dependent on government financial
support. Much of this investment consists of large projects which generate
construction and procurement contracts involving large sums of money, in
turn creating vested interests in such investment. The benefits, or returns,
on such investment are often hard to measure -- indeed the impossibility or
inappropriateness of simply applying a market test may explain why the
activity is in the public sector -- making objective assessment of proposals
for government investment very difficult. An influential paper by
Aschauer (1989) argued that benefits from public investment (based on US
data) are very high and, by implication, that more spending would be
desirable. However, the empirical literature that this paper stimulated is
not uniformly supportive of this view and many studies call attention to the
fact that investments are costs whose impact is negative unless there is an
adequate return.
In view of the difficulties involved, the OECD in 1998 reviewed the
main issues involved on the basis of submissions from thirteen countries in
response to a questionnaire which led to the following policy conclusions:
- Large-scale, diffuse programmes of infrastructure development cannot
be relied upon to increase output or welfare in the long run.
- The key to effective public investment lies in which infrastructure
projects are chosen. Proper targeting of public investment requires
effective institutions.
- The sectoral policy environment in which physical investment
decisions are made is crucial to the effectiveness of government
investment.
- Cost-benefit analysis can provide a useful indicative input to the public
investment process. Many countries use it in one form or another and
its role could be usefully strengthened.
- Effective public investment requires an environment of fiscal
discipline.
- Many countries are exploring various forms of public-private
partnership for investment projects.
These conclusions point to the need for persuasive evidence that
projects are needed, rather than presuming it; the need to find ways to
discriminate among projects in order to single out those that will have a
high return; the usefulness of cost-benefit analysis in this regard (Table 8);
and the desirability of involving the private sector (Table 9).
)OH[LEOHLQFHQWLYHVDQGFRQWUROPHFKDQLVPVLQWRPDQDJHPHQWRIWKH
JRYHUQPHQW
Maintaining fiscal discipline requires enforcement mechanisms to
ensure that budgets are implemented along intended lines. Most countries
have traditionally relied on highly centralised financial and personnel
control to achieve this and some (HJ Japan and Germany) continue to
regard this as the best way to proceed. Such centralised control can be a
source of inefficiency by limiting the authority of managers outside the
control ministries to exercise their judgement, however, and many OECD
countries have sought improvements by allowing these managers more
autonomy and flexibility in their day-to-day operations (Box 4 describes
several case studies). By empowering and motivating managers to improve
performance, this offers scope for efficiency gains reflected in lower
staffing levels and reduced operating expenditures, as well as improved
public services. But achieving these gains requires strategic controls and
the ability to define clear objectives to enable performance assessment that
ensures that accountability goes hand-and-hand with greater autonomy.
This is substantially more easy where activities can be subjected to a
market test than with core government activities such as provision of
public goods and in the social policy domain.
Reforms designed to introduce more flexible management systems
imply important changes in the relationship between central budget offices
and sectoral, or spending, ministries and agencies
23
and in their operations.
Initiatives to date have had several main elements. One is the introduction
−
__________
23
The new role of the central budget office is reflected in such activities as: devising a more
effective budget system to control the budget total and establish priorities among programmes;
integrating budgeting with other management processes; require spending agencies to
measure performance and evaluate results;
developing new guidelines and methods for holding
managers accountable; and
promoting new information and reporting systems.
Figure 4.
*HQ HUDOJRYHUQ PHQWLQY HVWPHQWLQ2 (& ' FRX QWULHV
1. Data for government outlays are not available.
Source: OECD.
0
5
10
15
20
25
Aus
t
r
a
li
a
A
ustri
a
Be
l
gium
C
an
a
da
Denma
rk
Finl
and
F
ranc
e
G
e
rmany
G
re
e
ce
Iceland
Ire
l
and
Italy
J
a
pa
n
Kore
a
M
exi
co (1)
Net
herl
ands
New Zea
l
and
Norw
a
y
Portugal
Spain
Sw
ede
n
Sw
it
z
erl
and
T
urkey
Unit
ed K
i
ngd
om
Unite
d
States
Per cent
Per cent of government outlays Per cent of GDP
)LJXUH
*HQHUDO*RYHUQPHQW,QYHVWPHQWLQ2(&'FRXQWULHV
7DEOH
8VHRIFRVWEHQHILWDQDO\VLVIRUSXEOLFLQYHVWPHQW
&RXQWU\
,VFRVWEHQHILW
DQDO\VLV
XVHG"
'HVFULSWLRQRIFRVWEHQHILWDQDO\VLV
Australia yes, in some
cases
Public enterprises are mainly involved in selling goods
and services in markets in order to earn a commercial
return. Reflecting the predominantly commercial nature
of their operations, investment decisions are made on the
basis of commercial viability and expected rate of return.
In contrast, investment decisions in the general
government sector also take into account the
Government’s broader expenditure priorities and budget
objectives.
Austria yes, but use is
variable
Most ministries have created a number of boards and
commissions whose function is to advice on investment
decisions. Usually, these advisory commissions are
composed of representatives of the employers’ and
employees’ associations and university or research
institute experts. The planning procedures are of different
quality and depth, ranging from relatively detailed project
planning to generally worded declarations of intent.
Finland yes, for public
transport
A socio-economic impact study is the basis for
investment evaluation in transport. It includes a
cost-benefit calculation as well as an assessment of
impacts which cannot be valued in monetary terms.
Decisions on transport investment are taken by the
Parliament, where these studies provide one input into a
long and complicated decision process that is largely
political in nature. Cost-benefit analysis is used mainly in
order to eliminate poor projects from the selection
process.
Greece yes, in some
cases
The EU co-financed projects are subject to special
evaluation rules that include an assessment of the
socio-economic significance of the project and its
compatibility with EU policies. With the exception of
subsidies for private investment in less-developed
regions, cost-benefit analysis is not required.
Japan no Although the Japanese submission recognises a need for
cost/benefit analysis, it notes that such assessments pose
theoretical and practical difficulties. Recently various
levels of governments have initiated studies to evaluate
different types of public investment analysis.
Norway yes Cost-benefit analysis most extensively used for
investment in roads, but is also used elsewhere. In
general, ministries are reluctant to quantify benefits from
projects on the grounds that the estimations are
incomplete. The use of cost-benefit analysis by ministries
appears to be increasing. Actual investment decision are
made in Parliament; there is some evidence that
Parliament uses cost-benefit analyses as a screening
device to determine which projects should be considered,
but generally the influence of such studies was variable
among members of Parliament.
Spain yes Cost-benefit analysis is widely used to decide which
investment projects are the most appropriate. For
instance, for the large public investment projects
receiving aid (such as from the European Structural
Funds) a cost-benefit analysis is always performed with a
view to assessing the socio-economic returns. However,
this type of analysis is not done in order to compare
public investment projects in different sectors.
Turkey yes Cost-benefit analysis is undertaken. Other analytical
techniques can be used depending on the nature and the
characteristics of the project handled (e.g. technical
feasibility, environmental impact analysis, social
benefits). Any project that is feasible according to
economic or social criteria has to be consistent with the
development plans and annual programmes too. The
sectoral priorities are based on the results of the project
analyses, but political choices may also affect outcomes.
United
Kingdom
yes Under new investment control arrangements, departments
receive a set amount of money for investment purposes.
They will have to set out in detail how these resources are
to be managed so as to provide “best value for money and
ensure positive social returns”. HM Treasury has shared
responsibility (with other departments) for monitoring
these plans.
United
States
yes, as part of
a larger
process
Capital assets are not selected on the basis of rate of
return. Recently, a process called “capital programming”
has been used. This involves the planning, budgeting,
procurement and management of an asset. Departments
may use analytical procedures that resemble cost-benefit
analysis as part of the first two phases, but are not
required to do so. More generally, investments are
analysed with respect to how they contribute to meeting
the agency’s “mission, goals and objectives”.
: OECD, compiled from country submissions.
7DEOH
5HGHILQLQJWKHERXQGDULHVRISXEOLFDQGSULYDWHLQYHVWPHQW
DFWLYLW\
&RXQWU\ 'HILQLWLRQRISULYDWHDQGSXEOLFUROHVLQLQYHVWPHQW
Australia There is a growing realisation in Australia that the private
sector is able to play a larger and more effective role in many
areas such as electricity generation, telecommunications, the
provision of education and hospital services, prisons and road
funding. In many cases, this has resulted in the Government’s
ceasing activity in areas that can be more efficiently
undertaken by the private sector and introducing measures to
improve efficiency in those activities remaining in the public
sector.
Austria Government intervention is taking on new forms. Less
emphasis is placed on financial flows and more on regulation
and on providing frameworks and incentive systems. The
volume of government activity is shrinking and what is
considered to be a government responsibility is changing.
This is especially true for infrastructure projects in
telecommunications and energy supply, which until recently
were an exclusive domain of a public-sector monopoly.
Finland A debate is taking place about the Government’s role in
Finnish society. For the time being, this has focused more on
the transfer system and not much on public investment
(which is comparatively small). There have been some
attempts to “privatise” public investment. However, this
approach does not extend to decision making: the
Government is still responsible for making the investment
decision.
Germany It is customary practice in Germany for private enterprises to
be involved in the planning and construction of public
infrastructure projects. Over the past few years, greater scope
has been made for private enterprise in performing
public-sector tasks. This includes greater involvement of
private finance and privatisation of telecommunications and
postal services.
Japan Co-operation between public and private actors in
implementing investment projects (called “third sector” in the
submissions”) is currently being looked at by the Japanese
Government. The submission notes that Japan has already
used these arrangements during the latter half of the 1980s,
but encountered serious problems. Both the private and
public sectors found it difficult to co-ordinate their different
objectives and responsibilities were often unclear in third
sector entities.
Spain Several new financing techniques have been introduced
recently. Private-sector participation in the construction of
motorways, rail tracks and hydraulic works have been
encouraged through long-term concession contracts
(Build-operate-transfer). The turnkey method of payment has
been used for several large projects, this shifts the risk of cost
over-runs onto private partners and also pushes the public
costs onto future budgets. Other public-private partnerships
have been created for several new railways and water works.
Turkey Recent investment projects and plans have encouraged
participation of the private sector, even in sectors
traditionally dominated by the public investment. In
particular, the Build Operate Transfer model has been
extensively used in Turkey.
United
Kingdom
The Private Finance Initiative (PFI) transforms government
departments from being owners and operators of assets into
purchasers of services from the private sector, while private
firms become long-term providers of services rather than
simply up-front asset builders. Privatisation has also re-set
the boundaries between private and public investment,
leaving postal services and London Transport as the only
two state-owned firms with large investment programmes.
United
States
The line dividing public and private investment tends to be
rather sharply drawn, at least as far as investment by the
Federal Government is concerned. Investment partnerships
between Federal and private entities are not common in the
United States. However, the “capital programming process”
asks departments to consider whether private entities can
better undertake an activity and, if so, to forego investments
related to that activity.
: OECD, compiled from country submissions.
%R[(QKDQFLQJIOH[LELOLW\LQSXEOLFPDQDJHPHQW
In 1997, the Public Management Service reviewed the efforts of five
OECD countries -- Australia, France, New Zealand, Sweden and the United
Kingdom -- to reform the way they organise and manage the public sector
1
. The
review found that in all five there is professed consensus within government
that the centralised model no longer suits the needs and conditions of public
management. Reform has been centred around accountability frameworks in
which the government entrusts spending agencies with flexibility in using
resources, in exchange for holding them responsible for results. The repertoire
of devices for enforcing managerial accountability includes strategic and
operational plans, performance measures and targets, contracts for personal and
organisational performance, de-coupling service delivery from policy making,
new accounting rules and annual reports, more active use of evaluation and
auditing, and financial inducements and sanctions.
The five countries have different governing traditions and have
approached reform differently. France has a long tradition of detailed
supervision by financial controllers, and it has moved cautiously to enlarge the
operational discretion of local managers. Sweden is at the other end of the
spectrum, for it has a long history of small ministries and relatively autonomous
agencies. Sweden gives managers more latitude than is found in some other
countries, so that innovations have been less dramatic than elsewhere. By the
early 1980s, the United Kingdom had already retreated from the doctrine of
Treasury Control that it has practised for more than a century. Its financial
management initiative launched in 1982, the Next Steps initiative commenced
half a dozen years later, and more recent fundamental expenditure reviews have
been spurred by political support at the top of the government for re-shaping
the public sector. Australia entered the reform era with highly centralised
controls, but it has discarded many personnel and financial restrictions and
adopted a variety of political and administrative arrangements to stimulate
management improvement. As a small country with an open economy, New
Zealand felt its future well-being threatened by powerful international forces,
and it responded by creatively adapting commercial practices to public
management.
The five countries face similar problems in restructuring national
administration. All must establish new relationships between the centre, which
is politically accountable for governmental performance, and operating units,
where services are provided and most resources are spent. Defining the new
relationship has been difficult because strategic controls must be devised in
place of the discredited H[DQWH controls.
All of the countries must motivate managers to take initiative and
responsibility over what they spend and produce and to accept that the
performance of their organisation depends on their personal performance.
There has been an enormous turnover of senior and middle managers in New
Zealand and the United Kingdom, as many officials discomfited by the new
managerialism have left on their own accord or have been encouraged to
depart. The importation of new managers appear to be inconsequential in
France and Australia. Each government must determine what is acceptable risk,
as operating agencies are given discretion to spend resources and take other
actions that may have important political of financial ramifications. This issue
is least troublesome in Sweden, where the line between ministries and agencies
is well marked, and most pressing in New Zealand and the United Kingdom,
where the independence of agencies has called into question the Westminster
doctrine of ministerial accountability.
Each government has devised an instrument of choice to assure that
performance information influences organisational behaviour. Australia relies
on programme evaluation both before policies have been initiated and after they
have been funded; France is emphasising responsibility centres as a means of
imbuing civil servants with awareness that their actions can make a difference
in the quality of service; Sweden has placed increasing reliance on annual
reports that are audited for reliability of financial and performance statements;
the United Kingdom looks to framework documents and performance targets to
concentrate managerial attention on key objectives and results; New Zealand
invests considerable resources in negotiating performance agreements for chief
executives and purchase agreements for agencies. Every country faces the
problem that no matter how much it generates by way of performance
information, decisions may be taken and resources allocated in disregard of
objectives and results.
Because of the difficulty of implanting a performance culture, every
country has had a spate of disappointments; none has accomplished everything
it set out to do. The United Kingdom found that the financial management
initiative had produced better information, but had done little to liberate
managers at operating levels; it subsequently appeared that Next Steps had
energised the newly established agencies but had not yet transformed the
central departments. Australia has been vexed by the problem of packaging
performance information into a useful format, and it has also been disappointed
by the less than optimal use of the programme structure. New Zealand has
made relatively little headway in measuring outcomes, and the relationship
between ministers purchasing services and agencies supplying them has not
been sufficiently clarified. Sweden has been disappointed by the failure of the
multi-year budget frames to deepen the quality of budget work. France has
found that, despite government guidelines, some important ministries have
dragged their feet in devolving responsibility to local agencies.
________________________
1. This box is a condensed version of the Executive Summary that appears in OECD
(1997 ).
of top-down spending ceilings, consistent with the medium-term
expenditure frameworks (see above), with the elected officials retaining an
important role in designing the overall budget and stating spending
priorities. Several countries, including the United Kingdom, Australia,
New Zealand, the Netherlands and all Nordic countries allow almost
complete discretion in spending within cash limits on running costs.
Operating units are allowed to shift funds among items of expenditure and
between fiscal years. In several countries (Australia, Denmark, the
Netherlands, New Zealand and Sweden) agencies notionally earn or pay
interest on carried-forward or pre-spent funds, although typically
carry-over and pre-spending is limited to a certain percentage of
appropriated funds
24
. This allows increased flexibility to shift funds
between fiscal years which reduces incentives for end-of-year spending
hikes and poor resource use. An important element of this set-up is that
future funding is not reduced by under-expenditure in a previous year.
Second, financial information systems similar to those required to
ensure fiscal transparency (FI Part 4.1) are also needed to allow control
ministries and elected officials to monitor performance and to ensure
overall financial control. These should include: reliance on accrual-based
accounting to the extent feasible, in order to identify when, where and how
many resources are being used; cost data that are complete; budgets
(LHexact spending plans) and financial reporting (LH H[SRVW
expenditures) that are on the same basis; and auditing to ensure the
integrity of the accounts.
Third, several countries have made efforts to reinforce the use of
market mechanisms, such as privatisation, contracting out, and exposing
activities to private competition, by the development of “internal markets”
as a device for enhancing accountability. New Zealand has gone furthest in
this direction, making the split between the government’s role as an owner
and a purchaser explicit; levying a capital charge for the government’s
investment; contracting for the services of chief executives of ministries
and agencies along the lines of contractual relationships in the private
sector; and negotiating purchase agreements for the “sale” of output from
agencies to ministries. The difficulty of defining and measuring the
“output” of ministries and government agencies has been a major
challenge for this approach, and it has proved difficult in a simulated
__________
24
OECD (1997 ).
market to clarify relationships between purchasers and suppliers.
Furthermore, performance contracts between public sector agencies are
unlike arms-length agreements between unrelated parties. A major
dilemma relates to the efficiency dividend. Unlike in real markets, there is
no incentive to lower prices when efficiency increases. Taking away the
dividend would penalise managers or agencies for being efficient, while
allowing them to keep the dividend would enable them to spend on
services that were not contracted for in the budget. Overall, developing
market-type discipline through contracting and development of
quasi-markets remains at an experimental stage.
Finally, some countries have sought more market oriented and
flexible approaches to public sector pay determination, conditions of
employment and staffing levels (notably the United Kingdom, Australia,
New Zealand and the Nordic countries). Growing use is being made of
workers under contract rather than permanent civil servants while salary
scales and job classifications have been revised to allow greater use of
promotions as an incentive and to link pay to performance indicators. In
some cases (HJNew Zealand and Sweden) broader reforms include
giving each government department autonomy to bargain with its own
employees over pay rates, working conditions and other matters. In some
instances managers have been given flexibility in staffing -- selection,
hiring, deployment and performance management. But in many countries
considerable rigidities persist in public sector wage structures and staff
management is less flexible than in the private sector. Such reforms,
coupled with top-down expenditure ceilings, may be more durable means
of increasing productivity and controlling costs than centrally imposed
wage restraint
25
.
0RQLWRULQJSURJUHVVDFKHFNOLVW
Based on the above discussion of main reform areas, a
comprehensive checklist of criteria has been compiled to assess the policy
efforts to enhance the cost-effectiveness of public expenditure in
individual countries. The range of criteria put forward is probably much
__________
25
See for a recent comparison of wage determination in the public sector in two countries, France
and Italy, OECD (1998
).
too wide to be considered for every single OECD country, and, dependent
on the situation, their importance will vary from country to country.
With regard to EXGJHWDU\ SURFHVVHV DQG FRQWURO, the following
questions look relevant:
- How well does the country adhere to principles of transparency put
forward in the IMF’s Code on Good Practices on Fiscal Transparency
and the OECD Best Practices for Budget Transparency?
- What is the scope for improving fiscal transparency? Are budget
documents clear and unambiguous? Are contingent liabilities reported?
To what degree has accrual accounting been adopted?
- Have medium-term expenditure frameworks been adopted and what has
been the experience to date? Have budget targets or spending ceilings
been adhered to? If not, what have been the main causes? Have cyclical
factors importantly affected the ability to achieve targets?
- Are systems in place to ensure that the economic assumptions that
underlie budget plans are realistic and that the future implications of
spending programmes are objectively evaluated and reflected in
medium-term planning?
- What enforcement mechanisms have been adopted in circumstances
where rules-based approaches are in place? Are these deemed to be
credible? Do they exert a genuine impact on expenditure discipline?
- What are the risks to public finances from financial operations and
from the activities of public sector entities outside the usual budget
process?
- Should the mandating and reporting of off budget expenditure and the
operations of public entities outside the general government become
more transparent and comprehensive? Do they include estimates of any
subsidy element that is not appropriated from the budget and gains or
losses arising from market risks such as from foreign currency
exposures?
- Are the governance arrangements that apply to public entities outside
the general government well designed to encourage good performance
and to trigger changes when such performance is not delivered? What
improvements should be made?
On fiscal relations between central and lower levels of government
relevant questions are the following:
- What mechanisms are in place to enforce local governments’ spending
discipline?
- If local governments have discretion to raise money in financial
markets are they required to do so on the basis of their own
credit-worthiness? To what extent does such borrowing enjoy an
explicit or implicit guarantee from the central government?
- How are intergovernmental grants or transfers determined? Are they
designed to ensure that they do not operate as an entitlement which
encourages low quality expenditure?
To assess the role of PDUNHWEDVHG SURYLVLRQ DQG RWKHU UDWLRQLQJ
PHFKDQLVPV the following questions are important:
- Does the design of social programmes in the country under review take
account of their interaction with the tax system and their economy-wide
impact, notably on the labour market?
- Is there scope for gains to be achieved by more careful targeting of
social policies involving income transfers and financing the provision
of merit goods?
- Where is the boundary in the country under review between these
benefits and social services financed publicly and those left to
individuals to cover from their personal resources?
- Are the longer-term financial implications of the balance between
targeted assistance and more universally available programmes and
services been carefully evaluated and factored into budget planning?
- Are eligibility conditions for social benefit programmes enforced well?
- Are systems for provision and financing health care, education and
other social services delivering satisfactory outcomes on a
cost-effective basis?
- To what extent are market-type mechanisms of provision, such as
contracting-out, user fees and vouchers employed in the country under
review? What has been the experience to date and to what extent have
the above basic conditions for success been met?
- Has the scope for using market mechanisms without compromising
social policy goals been fully exploited? For which areas could the
introduction of market mechanisms be considered?
- Is there evidence of large pent-up demand for public infrastructure or
important areas where under-provision is hurting economic
performance? Alternatively, is there evidence of over-investment in
public capital goods?
- Is the decision making process behind public infrastructure investment
transparent? How extensive is the use of cost-benefit analysis? To the
extent that formal cost-benefit methods are not used, are there ways of
ensuring that the full costs of projects are taken into account? Are
external effects sufficiently taken into consideration?
- What is the private sector’s involvement in public investment policy? Is
extensive use made of public-private partnerships and how are these
being governed? Is there scope for increased reliance on the private
sector and market forces? Is public procurement open to competition
and transparent?
Finally, progress in the introduction of IOH[LEOH LQFHQWLYHV DQG FRQWURO
PHFKDQLVPV into management of the government could be assessed on
the basis of the following questions:
- How much progress has been made with devolution of day-to-day
decision making to the operational levels of the public administration?
Is there scope for further moves in this direction?
- Is there scope for greater use of market mechanisms, contracting or
development of simulated markets which might enhance technical
efficiency?
- Have modern accounting and reporting systems been introduced to
facilitate performance assessment?
- To what degree has market-oriented human resource management been
adopted? What are the obstacles for reforms in this area?
$11(;
7DEOHV
A1. General government outlays by economic category: income transfers
A2. General government outlays by economic category: subsidies
A3. General government outlays by economic category: interest
payments
A4. General government outlays by economic category: consumption
A5. General government outlays by economic category: net capital
outlays
Table A1.
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995 2000
Australia 4.23.95.96.87.46.98.58.3
Austria 12.9 14.1 14.9 16.2 17.9 17.7 19.5 18.3
Belgium 11.9 11.0 14.5 16.1 17.2 15.1 15.5 14.4
Canada 5.2 6.5 8.6 8.3 10.5 11.2 12.6 10.9
Denm ark
1
6.8 10.5 13.5 16.2 16.1 17.8 20.4 17.2
Finland 7.6 5.9 8.7 9.5 11.8 12.6 16.1 12.6
France 11.5 12.0 14.1 15.5 17.7 16.9 18.5 18.1
Germ any 13.0 13.0 17.2 16.6 16.0 15.2 18.1 18.6
Greece 6.8 7.6 7.1 8.9 14.4 14.4 15.1 16.1
Ireland 10.0 10.0 10.0 10.7 12.9 11.9 12.6 9.7
Italy 11.9 11.8 14.4 14.2 17.1 18.1 16.7 17.3
Ja pan 4.7 4.6 7 .7 10.1 10.9 11.4 13.4 15.7
K orea 0.9 0.6 0 .7 1.3 1.5 2.0 2 .1 3.3
M e xico .. .. .. .. .. .. 2.6 ..
Netherlands 10.0 10.8 14.3 16.4 15.5 15.5 15.3 11.8
Norway 6.6 9.0 10.0 11.3 11.8 16.0 15.8 13.7
Portugal
1
2.3 2.1 5.2 7.0 7.8 8.5 11.8 12.5
S pa in 4.5 5.9 7 .4 10.9 12.7 12.7 13.9 12.4
Sweden 7.7 10.1 13.8 17.1 17.7 19.3 21.3 18.3
United Kingdom
1
6.9 8.0 10.2 11.6 13.7 11.9 15.4 13.1
United States 5.0 7.1 10.2 9.8 9.8 10.0 11.8 10.5
Euro area 10.9 11.2 14.0 14.7 15.8 15.5 17.0 16.7
OECD 6.5 7.5 10.2 10.7 11.4 11.5 13.2 12.8
1. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD Analytical Database figures underlying 68, December 2000 and
OECD National Accounts.
7DEOH$
*HQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\,QFRPHWUDQVIHUV
SHUFHQWRI*'3
Table A2.
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995 2000
Australia 0.70.91.11.51.71.31.31.2
A ustria 2.3 1.8 3 .0 3.1 3.2 3.1 2 .9 2.5
B elgium 2.3 2.3 2.6 2.8 2.4 1.7 1 .5 1.5
Canada 0.9 0.9 2.5 2.7 2.5 1.5 1.1 1.1
Denm ark
1
1.82.62.73.12.92.22.52.3
F inla nd 3.2 2.8 3 .8 3.3 3.1 2.9 2.8 1.5
France 2.52.22.22.12.61.81.51.3
G e rm any 1.2 1.7 1 .9 2.0 2.0 2.0 2.1 1.7
Greece 1.41.03.23.03.71.20.40.2
Ire land 2.7 3.4 2 .5 2.6 2.3 1.1 1.0 0.7
Italy 1.71.92.82.92.62.01.51.2
Ja pan 0.7 1.1 1 .5 1.5 1.1 1.1 0.8 0.6
K orea 0.3 0.3 1 .4 0.9 0.6 0.6 0.7 0.3
M e xico .. .. .. .. .. .. 0.7 ..
N etherlan ds 0.9 1.0 1.2 1.7 2.0 1.7 1 .1 1.6
Norway 3.43.84.65.24.24.53.72.5
Portugal
1
1.01.31.76.04.21.81.41.2
S pa in 0.5 0.5 0 .7 1.1 1.3 1.1 1.1 1.0
Sweden 1.1 1.3 2.5 3.3 3.9 3.6 3.8 1.8
United Kingdom
1
1.61.73.62.52.00.90.70.5
United States 0.2 0.5 0.5 0.5 0.5 0.4 0.3 0.2
Euro area 1.6 1.7 2.0 2.1 2.2 1.8 1.7 1.4
OECD 0.81.01.41.41.41.10.90.8
1. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD Analytical Database figures underlying 68, December 2000 and
7DEOH$
*HQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\6XEVLGLHV
SHUFHQWRI*'3
Table A3.
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995 2000
Australia 2.42.42.33.24.84.04.22.0
A ustria 0.8 1.1 1.3 2.4 3.5 4.0 4.3 3.5
Belgium 2.8 3.6 4.2 6.6 11.1 11.9 9.3 6.7
Canada 2.9 3.7 3.8 5.4 8.4 9.5 9.6 7.4
Denmark
1
1.11.31.23.99.67.36.44.5
Finland 1.0 1.0 0.6 1.0 1.8 1.4 4.0 3.1
France 0.01.11.21.42.82.93.73.3
G erm an y 0.7 0.9 1.3 1.9 2.9 2.5 3.7 3.4
Greece 0.6 0.8 1.1 2.0 4.4 8.7 11.1 7.2
Ireland 5.0 5.0 5.0 6.4 10.0 7.9 5.4 2.2
Italy 1.1 1.5 3.3 5.0 7.8 9.4 11.5 6.5
Japa n 0.4 0.6 1.2 3.2 4.5 3.9 3.8 4.0
K orea 0.0 0.3 0.4 0.6 0.7 0.5 0.4 1.6
M exico .. .. .. .. .. .. 4 .9 ..
N etherlands 2.7 2.9 3.1 3.8 6.3 5.9 5.9 3.9
Norway 1.41.61.53.13.23.62.81.6
Portugal
1
0.60.50.62.26.76.65.23.2
S pain 0.4 0.3 0.3 0.4 1.9 3.8 5.2 3.6
Sweden 1.8 2.0 2.3 4.1 8.4 5.0 7.1 4.1
United Kingdom
1
3.83.93.94.74.93.43.62.7
United States 1.9 2.2 2.4 3.2 5.0 5.1 4.8 3.6
Euro area 0.8 1.3 1.8 2.6 4.5 4.9 5.9 4.2
OECD 1.41.82.13.14.74.85.03.8
1. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD Analytical Database figures underlying 68, Decem ber 2000 and
7DEOH$
*HQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\,QWHUHVWSD\PHQWV
SHUFHQWRI*'3
Table A4.
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995 2000
Australia 13.4 14.4 18.9 18.6 20.0 18.5 18.6 18.5
Austria 14.6 16.1 18.2 18.4 19.5 18.8 20.4 19.4
Belgium 16.7 17.6 21.4 23.0 23.0 20.3 21.5 21.0
Canada 15.6 20.5 21.8 21.3 21.9 22.4 21.4 18.4
Denmark
1
16.7 20.4 25.1 27.2 25.8 25.6 25.8 25.3
Finland 14.2 15.1 17.8 18.7 20.6 21.6 22.8 20.8
France 16.9 17.4 19.5 21.5 23.7 22.3 23.9 23.4
Germany 15.0 15.5 20.1 19.9 19.7 18.0 19.8 18.8
Greece 8.2 8.8 10.6 11.4 14.2 15.1 15.3 15.0
Ireland 13.3 14.3 18.2 19.4 18.1 15.1 14.9 11.8
Italy 16.2 14.9 16.1 16.8 18.6 20.2 17.9 17.9
Japan 8.2 7.4 10.0 9.8 9.6 9.0 9.8 10.1
Korea 9.5 9.7 11.3 11.9 10.4 10.5 9.7 9.7
M exico .. .. .. .. .. .. 10.5 ..
Netherlands 23.6 24.9 28.2 29.1 26.4 24.3 24.0 22.6
Norway 14.6 16.4 18.7 18.7 18.1 20.8 20.9 18.8
Portugal
1
11.5 13.3 14.4 14.0 15.0 16.4 18.6 21.0
Spain 9.1 10.2 11.3 14.3 15.9 16.9 18.1 16.9
Sweden 17.9 22.5 25.2 29.6 28.2 27.7 26.3 26.5
United Kingdom
1
17.2 18.0 22.4 21.6 20.9 19.9 19.8 18.3
United States 16.4 18.5 18.1 16.8 17.1 16.6 15.3 14.1
Euro area 15.4 15.8 18.5 19.5 20.3 19.7 20.4 19.7
OECD 14.6 15.7 17.3 17.0 17.3 16.8 16.6 15.7
1. Prior to 1988 in the case of Denmark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD Analytical Database figures underlying 68, Decem ber 2000 and
7DEOH$
*HQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\&RQVXPSWLRQ
SHUFHQWRI*'3
Table A5.
Per cent of GDP
1965 1970 1975 1980 1985 1990 1995 2000
Australia 3.83.63.12.33.82.42.81.4
A ustria 6 .0 4 .8 7 .0 7.0 6.1 4.9 5.3 5.1
B elgiu m 1.3 5 .3 4 .9 4 .9 3 .5 1 .8 2 .5 3 .1
Canada 3.1 2.3 2.2 1.4 2.1 1.4 0.6 0.0
Denm ark
2
5.45.44.74.63.60.71.52.0
F inland 4.3 4.9 6.1 4.7 5.0 5.8 8.6 6.7
France 6.75.05.34.95.15.76.05.1
G erm any 5.4 6.0 6.4 6.1 4.9 6.1 2.6 0.4
Greece 5.15.15.14.35.58.34.65.2
Ire land 5 .0 5 .0 5 .0 8.5 7.2 3.6 3.7 3.2
Italy 2.02.54.42.94.53.54.73.7
Ja p a n 5 .0 5 .2 6 .3 7.5 5.6 6 .0 7.9 7.8
K orea 3.8 3.8 2.9 4.6 4.4 4 .8 6 .4 8 .6
M exico .. .. .. .. .. .. 2 .7 ..
Netherlands -2.5 -2.7 -1.0 -0.1 1.8 1.9 1.4 1.5
Norway 3.24.14.95.64.24.94.44.1
Portugal
2
2.7 0.8 3.4 -1.2 9.3 10.9 4.3 4.2
S pain 5.0 4.7 4.6 4.6 7.6 6 .9 5 .7 4.6
Sweden 4.9 5.9 3.6 2.7 1.6 0.2 3.6 3.2
United Kingdom
2
4.15.14.42.62.45.94.93.8
United States 2.0 1.3 1.1 1.0 1.4 1.4 0.7 0.9
Euro area 4.3 3.9 4.6 4.1 4.3 4.4 4.2 3.0
OECD 3.53.23.43.33.33.83.83.4
1. Net fixed investment plus net capital transfers.
2. Prior to 1988 in the case of Denm ark, 1995 for Portugal and 1987 for the United Kingdom data are backward
extrapolations based on earlier National Accounts series.
OECD Analytical Database figures underlying 68, Decem ber 2000 and
7DEOH$
*HQHUDOJRYHUQPHQWRXWOD\VE\HFRQRPLFFDWHJRU\1HWFDSLWDORXWOD\V
SHUFHQWRI*'3
5()(5(1&(6
Adema, W. (2000), “Revisiting real social spending across countries: a
brief note”, 2(&'(FRQRPLF6WXGLHV, No. 30, pp.191-7.
Arcelus, F.J. and A.L. Levine (1986), “Merit goods and public choice: the
case of higher education”, 3XEOLF)LQDQFH 41/3, pp. 303-15.
Aschauer, D.A. (1984), “Is public expenditure productive?”, -RXUQDO RI
0RQHWDU\(FRQRPLFV 23, pp. 177-200.
Becker, G.S. (1974), “A Theory of Social Interactions”, -RXUQDO RI
3ROLWLFDO(FRQRP\, 82/6, pp. 1063-93.
Blejer, M.I. and A. Cheasty (1991), “The measurement of fiscal deficits:
analytical and methodological issues”, -RXUQDO RI (FRQRPLF
/LWHUDWXUH, Vol. 29, pp. 1644-78.
Creedy, J. (1995), 7KH(FRQRPLFVRI+LJKHU(GXFDWLRQ, Edgar Elgard.
Eichengreen, B. and J. von Hagen (1996), “Fiscal policy and monetary
union: is there a trade-off between federalism and budgetary
restrictions?”, NBER Working Paper, No. W5517.
Fossati, A. and G. Panella (eds.) (1999), )LVFDOIHGHUDOLVPLQWKH(XURSHDQ
8QLRQ, London/New York, Routledge.
Johnson, G.E. (1984), “Subsidies for Higher Education”, -RXUQDORI/DERU
(FRQRPLFV, No. 2/3, pp. 303-18.
MacFarlan, M. and H. Oxley (1996), “Social transfers: spending patterns,
institutional arrangements and policy responses”, 2(&' (FRQRPLF
6WXGLHV, No. 27, pp. 147-94.
McPherson, M.S. and M.O. Shapiro (2000), “Financing Lifelong Learning,
Trends and Patterns of Participation and Financing in US Higher
Education”, +LJKHU (GXFDWLRQ 0DQDJHPHQW, No. 12/2, pp. 131-56,
OECD.
Martin, J.P. (2000), “What works among active labour market policies:
evidence from OECD countries’ experiences”, 2(&' (FRQRPLF
6WXGLHV, No. 30, pp. 79-113.
Oates, W.E. (1999), “An essay on fiscal federalism”, -RXUQDORI(FRQRPLF
/LWHUDWXUH, Vol. 37, pp. 1120-249.
OECD (1994), 1HZ ZD\V RI PDQDJLQJ LQIUDVWUXFWXUH SURYLVLRQ, PUMA
Occasional Papers, No. 6.
_____ (1996), 2(&'(FRQRPLF2XWORRN Vol. 60, December.
_____ (1997D), 5HJXODWRU\ LPSDFW DQDO\VLV EHVW SUDFWLFHV LQ 2(&'
FRXQWULHV.
_____ (1997E), &RRSHUDWLYH DSSURDFKHV WR UHJXODWLRQ, PUMA
Occasional Papers, No. 18.
_____ (1997F), 0DNLQJ:RUN3D\.
_____ (1997G), (GXFDWLRQDQGHTXLW\LQ2(&'FRXQWULHV.
_____ (1997H), &RQWUDFWLQJ RXW JRYHUQPHQW VHUYLFHV, PUMA Occasional
Papers, No. 20.
_____ (1997I), 0RGHUQEXGJHWLQJ
_____ (1998D), 0DLQWDLQLQJSURVSHULW\LQDQ$JHLQJ6RFLHW\.
_____ (1998E), (GXFDWLRQ3ROLF\$QDO\VLV.
_____ (1998F), (GXFDWLRQDWDJODQFH2(&'LQGLFDWRUV.
_____ (1998G), 8VHU FKDUJLQJ IRU JRYHUQPHQW VHUYLFHV EHVW SUDFWLFH
JXLGHOLQHVDQGFDVHVWXGLHV, PUMA Occasional Papers, No. 22.
_____ (1998H),:DJH GHWHUPLQDWLRQ LQ WKH SXEOLF VHFWRU D )UDQFH,WDO\
FRPSDULVRQ, PUMA Occasional Papers, No. 21.
_____ (1999), (GXFDWLRQ3ROLF\$QDO\VLV.
_____ (2000), (GXFDWLRQDWDJODQFH2(&'LQGLFDWRUVHGLWLRQ.
Oxley, H. and J. P. Martin (1991), “Controlling government spending and
deficits: trends in the 1980s and prospects for the 1990s”, 2(&'
(FRQRPLF6WXGLHV, No. 17, pp. 145-89.
Persson, T. and G. Tabellini (2000), “Political economics and public
finance”, CEPR Discussion Papers, No. 2235.
Pola, G. (1999), “A comparative view of local finances in EU member
countries: are any lessons to be drawn?”, in A. Fossati and G.
Panella (eds.) (1999), )LVFDO IHGHUDOLVP LQ WKH (XURSHDQ 8QLRQ,
Routledge, London/New York, pp. 15-62.
Stern, D. and J.M.M. Ritzen (1991), “Introduction and overview”, in D.
Stern and J.M.M. Ritzen (eds.) (1991), 0DUNHWIDLOXUHLQ7UDLQLQJ"
1HZ (FRQRP\ $QDO\VLV DQG (YLGHQFH RQ 7UDLQLQJ RI $GXOW
(PSOR\HHV, Springer, pp. 1/14.
Tanzi, V. and L. Schuknecht (2000), 3XEOLFVSHQGLQJLQWKH
FHQWXU\,
Cambridge University Press.
Tsakloglou, P. and M. Antoninis (1999), “On the distributional impact of
public education: evidence from Greece”, (FRQRPLFV RI (GXFDWLRQ
5HYLHZ, Vol. 18, pp. 439-52.
Van den Noord, P. (2000), “The size and role of automatic stabilisers in
the 1990s and beyond”, 2(&' (FRQRPLFV 'HSDUWPHQW :RUNLQJ
3DSHUV No. 230, Paris.
'(6,*1,1*02'(/%$6('),6&$/32/,&<58/(6
-DYLHU-3pUH]DQG3DXO+LHEHUW
,QWURGXFWLRQ
In order to generate solvency for the fiscal sector, leading
macroeconomic forecasting models employ a fiscal rule. The rule is
designed to guarantee that the intertemporal budget constraint of the
government is satisfied – thereby generating model closure. In addition to
effectively ruling out the possibility of an explosive path for fiscal
variables such as the government debt ratio, the adopted rule can strongly
influence the adjustments of fiscal variables against shocks or policy
changes. The choice of fiscal rule thus entails potentially significant
consequences for the intertemporal behaviour of fiscal variables, as well as
effects on macroeconomic and financial variables in the model. More
importantly perhaps, policy choices can be influenced by the specification
of the fiscal rule, as the impact of policy changes and reforms is often
assessed on the basis of a macroeconomic model that can be influenced by
the specification of the fiscal rule.
Existing fiscal rules employed in leading macroeconomic forecasting
models are generally imposed H[RJHQRXVO\, and involve backward-looking
behaviour on the part of the fiscal sector, despite the widespread use of a
forward-looking framework in modelling households, firms and the
monetary authority. This modelling strategy can, in principle, lead to
inconsistencies in the fiscal sector with other sectors of the model as the
functional form and calibration of the rule is largely determined outside the
auspices of the model.
In this paper, we describe a recent methodological proposal put forth
by Perez & Hiebert (2001) to identify the appropriate fiscal rule
HQGRJHQRXVO\in a stochastic model, based solely on the properties of the
model itself. Specifically, we discuss how a state-contingent policy rule
can be obtained, relating fiscal instruments to different shocks affecting the
__________
*
European Central Bank. The views expressed in this paper are those of the authors and do not
necessarily reflect those of the European Central Bank (ECB). The authors would like to thank F.
Orlandi, R. Johnson, J.-P. Vidal, M. Catenaro, M. Tujula, P. Cour-Thiman, R. Strauch, along with
seminar participants at the
and the Fiscal Policies
Division of the ECB for helpful discussions and comments. Any remaining errors are of course the
sole responsibility of the authors.
economy and expectations of future developments of the economy – all on
the basis of the existing set-up of a the model at hand. This endogenous
fiscal policy rule is inherently consistent with the fundamental structure of
the model on which it is based, wholly integrated with all agents and
sectors in the economy, and with the structural parameters of the model. In
addition, the proposed fiscal rule, with a fiscal authority systematically
reacting to disturbances to the economy is consistent with the standard role
of government of fiscal stabilisation in the face of economic fluctuations.
The paper is organised as follows. In Section 2, we discuss the
rationale for fiscal policy rules, in the context of uncertainty in public
finance – and how to capture these concepts in a macroeconomic model. In
Section 3, we explore the need for fiscal rules in macroeconomic models,
and on the diversity of existing rules. In Section 4 we offer the rationale for
an alternative specification in the form of an endogenous fiscal rule.
Finally, in Section 5 we present some concluding remarks.
0DLQIHDWXUHVRIILVFDOSROLF\UXOHV
6RXUFHVRIYDULDELOLW\LQSXEOLFILQDQFHV
Government budgets are subject to considerable uncertainty, given
various shocks which have an impact on public finances, both of a
permanent and temporary nature. The effects of on public finances of
uncertainty may first come from non-budgetary sources. Shocks may be of
the standard macroeconomic type – involving, for instance, shocks to
technology, demand, energy, and labour supply. They may also be due to
changes in key financial variables, such as interest rates and exchange
rates. The effects of on public finances of uncertainty may also come from
budgetary sources. This could include uncertainty which often is inherent
in developing public spending plans, particularly relating to unforeseen
developments in military spending (HJ wars) and agricultural spending
(HJ compensation for droughts, epidemics). In addition to these sources of
temporary unpredictability, shocks of a more permanent nature to public
finances can include demographic shifts (HJ pressure on public
expenditure expected to accompany population ageing) and underlying
changes in tax collection (HJ the development of “underground”
economies).
Given the numerous sources of unpredictability, there is
considerable scope for forecasting errors in developing and executing fiscal
plans. Kilpatrick (2001) postulates that in a stochastic world, in order to
maintain budgetary stability in any given period, the fiscal authority should
react to any shocks that affect public spending or the tax base. This implies
that either implicitly or explicitly, governments should have a type of
contingency fund to cater for these sources of uncertainty. Although in
reality, the identification of shocks is not straightforward – see, for
example, Blanchard and Perotti (1999), in a macroeconomic model, the
source of all shocks hitting public finances or the economy is clearly
identifiable.
0RGHOOLQJXQFHUWDLQW\LQSXEOLFILQDQFHV
In theory, the government reaction to shocks affecting the budget in
a model should be shock-specific. As such, in a model characterised by
optimising forward-looking agents, the intertemporal fiscal rule should
include policy reactions to the different shocks affecting the economy. In
the absence of an active monetary authority monetising shocks to debt, as
is in the case of most industrialised countries, the fiscal authority should
react to any innovation affecting the fiscal sector through the adjustment of
budgetary items.
For instance, we may expect that different unforeseen shocks, such
as those to output (HJ an oil shock) or those to government spending (HJ
a war or drought) would have differing impacts on the government budget
and therefore elicit a nuanced reaction based on the source of the shock.
Moreover, the rule should, in principle, be introduced in a forward-looking
manner consistent with the other sectors of the model. As such, the rule
should be based on expectations of future values of relevant variables.
Thus, the rule would amount to some combination (either linear or
nonlinear) of endogenous and exogenous variables in the model. We
import and discuss these a rule which satisfies the above criteria in
Section 4.
2Q WKH QHHG IRU ILVFDO UXOHV LQ PDFURHFRQRPLF PRGHOV DQG RQ
WKHLUGLYHUVLW\
In establishing the basis for including a fiscal rule in a
macroeconomic model, one must look first to the government budget
constraint, which takes the following standard form in discrete time:
[1]
3
%
,J
3
%
1
1
−
−
+=+
τ
where %
t
stands for time-W nominal debt, J
t
is real primary spending,
τ
tax
collection in real terms, and ,
t
≡1+L
t
, the nominal interest rate on bonds .
Simply put, this condition entails that the government has to issue debt to
pay for spending in excess of tax collection. The aggregate variables
defined above may of course be broken down into their subcategories in
macroeconomic models used in practice. Solving this equation forward we
have that:
[2]
[]
1
1
0
1
1
11
0
0
1
1
lim
++
++
=
++
++
∞→
++++
∞
=
=
++
++
+−
=+
∏
∑
∏
3
%
,
J
,3
%
π
τ
π
where π is the time-W inflation rate. For the government to be solvent, the
second term of the right-hand side of the previous expression has to be
equal to zero. In other words, any shock affecting spending or real debt
should be covered by tax changes
. In any standard model with optimising
debt-holders, this is the condition close to the exact form of one of the
optimality conditions (the transversality condition attached to bond-
holdings) that has to be verified.
*HQHUDOIRUPXODWLRQRIDILVFDOUXOH
The fiscal rules used in existing macroeconomic models are based
on maintaining budgetary solvency required by (1). As the rules
__________
1
Seigniorage revenue is neglected for the sake of simplicity.
2
Effectively, this can be considered as a no-Ponzi game condition, whereby in order to guarantee
solvency, the government must be able to back all debt through its tax and spending system.
traditionally involve adjustment on the revenue side of the government
budget, we can write any generic tax system as:
[3]
()
\F
τττ
+= ,...,
where the first part of the equation embodies the normal tax system of the
economy (income taxes, consumption taxes, etc.), while the second
component represents the revenue adjustment by the government to
guarantee solvency, and react to shocks. Permanent shocks to the economy
would show up in the first part of the equation, while transitory shocks
would be catered for by the second. In theory, fiscal closure rules –
captured here by the
τ
t
– can take various functional forms, including
several types of variables, not only lagged values of certain state variables.
These rules for model economies approximate the actual reaction to shocks
by the fiscal authority.
7UDGLWLRQDOVSHFLILFDWLRQRIILVFDOFORVXUHUXOHVLQPDFURPRGHOV
Budgetary adjustment is generally either in the form of either a tax-
difference rule – as in MULTIMOD (IMF) and NIGEM (National
Research Institute) – whereby the change in taxation is a function of the
objective variable; or tax level rule – as in MSG2 (a model developed by
McKibbin and Sachs) – whereby the tax rate itself is adjusted in reference
to the objective variable. To illustrate, a tax-difference rule would take
some variant of the following generic form:
[4]
)()(
*
11
*
111 −−−−−
−∆+−=− [[E[[D
ττ
where [ is the objective variable (LH government debt or deficit), with an
asterisk denoting the steady state value, and a and b the speed of
adjustment parameters, which are calibrated.
∆≡
/ stands for the first
difference operator.
The calibration of the exogenous rules currently used in practice
requires the calibration of the so-called VSHHG RI DGMXVWPHQW parameters,
controlling the behaviour of the adjustment variable to deviations of target
values in the rule (HJdeficit, debt) from their steady-state values. Mitchell
HWDO (2000) find that this calibration may be somewhat informal or ad-hoc,
although some modellers have pursued more formal exercises in the
derivation of their fiscal rule – see, for example, the derivation based on a
quadratic loss function in Barrell HWDO (1994). In any case, when designing
the rules and calibrating the key parameters, these existing rules do not
consider explicitly who the debt holders are in the models they are
analysing, and tend to focus the stability analysis on the system formed by
the budget constraint, (1), while the calibration of the speed of adjustment
parameters is then done on the basis of some advocated properties of the
model solution and responses to shocks. Although this is a practical and
partially valid approximation, it does not guarantee that the resulting fiscal
rule is fully consistent with the properties of the model it is trying to close.
The rules used in practice are quite diverse in specification, and
some recent studies have found through standardised simulations that the
various specifications of these rules can lead to widely divergent results.
For instance, Mitchell HW DO (2000) compare the response of standardised
version of the fiscal rules of leading macroeconomic forecasting models,
including NIGEM, MULTIMOD and MSG. They find that the impulse
response function to a shock in government expenditure differs widely,
ranging from a relatively monotonic adjustment to a nonlinear adjustment
path. Bryant and Zhang (1996) also find that the response of variables can
differ quite substantially on the basis of alternative standardised fiscal
rules. They conclude on the basis of this evidence that generally, there is a
particularly imprecise understanding of how economies respond to fiscal
policy actions. Lastly, Barrell HW DO1994) also find that the
implementation of the fiscal policy rule has a significant effect on model
properties in comparing the tax rules of NIGEM, MULTIMOD and MSG.
Despite the variation in results, little consensus exists on the proper
formulation in terms of the dynamic adjustment component of the fiscal
rules in the literature. As noted in Mitchell HW DO (2000) and
Johnson (2000), this wide variety of fiscal rules found in the literature
highlights the lack of agreement amongst modellers regarding the
appropriate functional form for these rules. This at least can partially be
attributed to their formulation which, to a certain extent, may lack rigorous
theoretical underpinnings fully consistent with the model in which they are
used. Their formulation is imposed outside the confines of the model, and
can involve the considerable use of judgement in some cases. In this sense,
their derivation cannot be entirely consistent with all of the other economic
variables in models by design. This type of lack of internal consistency in
modelling has been criticised by many for its lack of microfoundations
starting with Lucas (1976). A more fundamental criticism of exogenously
imposed fiscal rules is their inherent vulnerability to the points raised in the
Lucas paper, as changes in the baseline parameters of the model may not
directly lead to a change in the form or calibration of the fiscal rule.
$SURSRVDOWRLGHQWLI\PRGHOEDVHGILVFDOFORVXUHUXOHV
Based on the principles outlined in Section 2, it is sensible to
postulate that a fiscal closure rule code for a government could take the
form of a given reaction to transitory shocks affecting their budgets,
τ
t
= I(VKRFNV
t
). This rule would imply that, on average over the
simulation horizon, any increase in tax collection due to shocks of one sign
would be offset by decreasing tax collection (transfers) coming from
shocks of the opposite sign. A policy reacting to innovations would be
countercyclical by nature, and non-distortionary, as it should be that
(
t-1
[
τ
t
] = (
t-1
[I(VKRFNV
t
)] = 0. If the fiscal authority were to react to time-
-W shocks, this would be enough to ensure stability of the model economy.
'HULYDWLRQRI5DPVH\W\SHRSWLPDOSROLFLHV
In order to endogenously calibrate the form of
τ
t
, and as an
alternative to the exogenous imposition of a fiscal rule, one solution would
be to derive a fiscal rule based entirely on the design of the markets in the
model, in a fully-fledged optimising framework entirely consistent with the
microeconomic foundations of the model, so that one could obtain the
coefficients in
τ
t
optimally and also the optimal form of
τ
(F
t
, \
t
,...) in (3).
For the development and implementation of such a rule in simple models
see, for example, Chari HWDO (1994) or Manzano and Ruiz (2000).
Although the strategy pursued in this literature would entail many
desirable characteristics, it is generally limited to the analysis of fairly
simplified economies – and would be impractical for large-scale macro
models, given the level of complexity of the economy in these models and
their level of disaggregation. In fact, solving a dynamic optimisation model
in which the government maximises agents’ utility subject to all Euler
conditions in agents’ problems would be cumbersome, if not impossible,
with the level of disaggregation in large-scale macroeconomic forecasting
models.
)LVFDOWKHRU\RIWKHSULFHOHYHO
Another alternative that goes beyond the standard practice outlined
in the previous Section uses the stability properties of the model under
analysis to FRQVWUDLQ the coefficients of rules of the form (4). In other
words, the coefficients a and b are chosen on the basis of agents’ decisions
ensuring stability of real debt. Nonetheless, these coefficients cannot be
uniquely calibrated using this strategy, and the formulation of the
functional form of the rule itself remains DGKRF. This shortcoming is
natural, as this approach has been used for alternative purposes of
macroeconomic modelling, mainly to stress the close link that the
government budget constraint imposes between monetary and fiscal
policies – see, for example, Leeper (1991), Sims (1994), Woodford (1995),
Mc Callum (2001) or Andrés HWDO. (2001).
$Q LGHQWLILFDWLRQ PHWKRGRORJ\ EDVHG RQ WKH DQDO\VLV RI IRUHFDVW
HUURUV
This methodology to retrieve HQGRJHQRXV fiscal policy rules in
models with imperfect foresight on the part of agents generates rules which
are formulated entirely on the structural parameters and framework of the
model. In this way, the rule adjusts automatically in response to any
changes to structural parameters of the model, thereby reducing the
susceptibility of this sector to the Lucas Critique.
The rule is derived using standard stability analysis theory for
rational expectations models, based on Blanchard and Khan (1981),
Sims (2000) and Novales HW DO (1999). It is constructed based on the
expectations errors of agents within the model. It can be expressed in
implicit form or explicit form, whereby the fiscal authority systematically
reacts to individual shocks to the economy via a state-contingent lump sum
tax on households
3
.
__________
3
See Perez and Hiebert (2001) for an illustration of this on the basis of a simplistic standard model.
Nevertheless, it should be stressed that the proposed identification methodology is, in principle,
general enough to be applied to any given large-scale macroeconomic model with optimising
agents.
%R[
7KHIRXUVWHSVUHTXLUHGWRFRQVWUXFWDPRGHOEDVHGILVFDOUXOH
1. Perform the stability analysis of the model at hand to determine the
VWDELOLW\QHHGV of the system, assuming
τ
t
=0;
2. Identify the expectation error relevant for debt stability, from the
relevant row(s) of the stability conditions, and then use them to
guess a first tentative relation between
τ
t
and the relevant
expectation error(s).
3. Determine the value of the coefficient(s) of that relation on the basis
of the stability analysis of the system, and it has to compute the
stability conditions of the system including the guess (initialised, for
example, to one); then the parameter(s) are calibrated in such a way
that the stability conditions are exactly the initial set (computed in
the first step).
4. Compute the fiscal closure rule by using the calibrated coefficient
and the relevant condition relating expectation errors and shocks.
*RYHUQPHQWVROYHQF\DQGVWDELOLW\DQDO\VLVLQUDWLRQDOH[SHFWDWLRQV
PRGHOV
Any given dynamic stochastic rational expectations model can be
written, without lack of generality, in the following implicit form:
[5] )(X
t+1
, X
t
, ε , η
+1
) = 0
where the vector X
t
contains the endogenous and exogenous variables in the
model, as well as the conditional expectations in the model; they may be
decision variables of the economic agents, such as consumption or real
debt holdings, or variables obtained as functions of decisions, such as real
interest rates, or exogenous variables like random shocks or policy
variables decided by the government. The vector ε
contains the
innovations in the laws of motion of the exogenous states, and η
is the
vector of expectational errors, satisfying (
t
(η
+1
) = 0, where the operator
(
t
(•) denotes the expected value of the argument given the information set
available up to time-W.
Proceeding to conduct a stability analysis of the above system, we
express the linearised/ log-linearised version of the system around the
deterministic steady-state can as:
[6] Γ
0
X
+1
= Γ
1
X + ψ ε
+1
+ Π η
+1
plus a set of transversality conditions:
[7]
[
]
0lim
0
=
+
→
X
ϕ
ZKHUH LV WKH DSSURSULDWH GLVFRXQW UDWH IRU WKLV PRGHO )RU WKH
transversality conditions to hold, we need to add a set of VWDELOLW\
FRQGLWLRQV to the system described in (6). For this model, these stability
conditions are defined by the eigenvectors associated with the unstable
eigenvalues of the system (6). Assuming invertibility, the stability analysis
is based on Γ
0
-1
Γ
1
.
4
The key to obtaining the stability conditions is
obtaining the transversality conditions attached to the unstable eigenvalues,
given that those attached to the stable eigenvalues are always satisfied.
Expressing the stability conditions as a linear (or log-linear) relationship
between the expectational errors and the structural shocks affecting the
economy, a unique stationary equilibrium satisfies the condition:
[8] 3
Γ
0
-1
(ψ ε
+1
+ Π η
+1
) = 0 for all W
where P
s
denotes the rows of the decomposed matrix of Γ
0
-1
Γ
1
which
amounts to a particular linear (or log-linear) combination of the
endogenous and exogenous variables in the model. The above closing
condition would be needed to solve for all expectational errors in the
model. As discussed in Sims (2000), for the equilibrium to be uniquely
determined, one such condition should be present for each expectational
error in the model. The stable paths of the approximated model economy
can be simulated given (6) just by appending (8).
In order to construct the fiscal rule, we would need to detect the
relevant stability condition for debt given by (8), and then identify the
necessary fiscal policy reaction, linking the tax instruments to the structural
__________
4
Note that invertibility is not strictly required – a more generalised solution algorithm is available in
the form of
-decompositions (see, for instance, Sims (2000) or Novales (1999)).
innovations in the economy in order to endogenously determine policy
responses by identifying the coefficients γ
η
s
in:
τ
= γ
η
s
η
t
s
and then identify an implied relation:
τ
= γ
ε
s
ε
t
s
2EWDLQLQJDQHQGRJHQRXVILVFDOUXOHLQLPSOLFLWIRUP
Consider a standard neoclassical growth model with agents
maximising their discounted utility derived from consumption, and with
debt, in which the transversality condition associated with debt takes the
form:
0lim =
+
+
+
∞→
3
%
(
ϕ
Using the standard analysis outlined in Perez and Hiebert (2001)
would imply that for this transversality condition to hold, one equation
summarising the relevant stability conditions should be added to the system
of first order conditions and constraints. This condition would take the
form of a linear/non-linear combination of real debt with other variables in
the economy, such as consumption, (F
t
) or the capital stock (N
t
):
function
debt
(%
t
/3
t
, c
t
, (
t
[I
debt
(F
t+1
, N
t+1
, ...)], ..., shocks
t
) =0
and should hold in each single period of time. This condition would be
unique, and would replace the transversality condition in the set of
optimality conditions used to solve for all the variables in the model. One
way to give some economic interpretation to this type of condition would
be as follows. Once agents internalise that the government commits itself
to be solvent, they behave in such a way that indeed the resulting
equilibrium is stationary and the government debt is valued and held by the
agents.
From an economic point of view, and for the purposes of policy
analysis, the intuition behind a pure analysis of the stability conditions
might be considered a bit obscure. Although imposing such conditions to
solve for the variables in the model is technically correct, it is somewhat
more difficult to give some economic meaning in the framework of the
model being analysed. Specifically, when imposing the transversality
conditions for bonds, one may wonder which instrument the government
would be moving on the event of, for instance, a recession. The implicit
formulation above reflects how agents internalise the commitment from the
part of the government to be solvent. When this commitment is
internalised, solvency is automatic. Indeed, a way to rewrite the stability
condition for bonds would be:
(
t
[I
debt
(F
t+1
, N
t+1
, ...)] = function
debt
-1
(%
t
/3
t
, F
t
, ..., shocks
t
)
so that imposing solvency implies a certain form to solve for agents’
beliefs. This is why for the solution to be unique there should be one such
condition per expectation error or expectation in the model.
([SUHVVLQJWKHHQGRJHQRXVILVFDOUXOHLQH[SOLFLWIRUP
From the fiscal policy point of view we would be interested in
knowing what amount of revenue given by
τ
t
would stabilise debt and
make the transversality condition hold. To do so, first note that the
condition including debt of the type outlined above has a counterpart
involving either a linear or non-linear relationship between the
expectational error associated with agents’ interest rate forecasts,
t
I
, a
subset of
t
– the structural shocks. Using this as an example, by
construction,
t
I
≡
I
debt
(F
t
, N
t
, …) - (
t-1
[I
debt
(F
t
, N
t
, ...)]
we can then postulate that the government should raise or decrease revenue
in line with agents’ relevant expectational error:
ξγτ
ξ
−=
and we can identify the first coefficient on the right hand side of the above
expression out of the stability analysis of the system at hand. Combining
the coefficient identified by the means outlined above with the implied
relation between
t
I
and the structural innovations would give us a
relationship of the form:
τ
= -
t
that can be identified as the fiscal policy closure rule. Taking the set of first
order conditions and constraint corresponding to the model under analysis,
and appending this rule would produce stable outcomes fully consistent
with the model solution, where all of the behavioural elements that the
literature on fiscal rules normally imposes on the fiscal rule would be
transferred back to the properties of the model itself.
&RQFOXVLRQV
In principle, the methodology outlined in this paper is applicable to a
a wide range of macroeconomic forecasting models, given its requirement
only of rational expectations frameworks in stochastic imperfect foresight
models – and, accordingly, the presence of expectations errors.
Specifically, we explain how the presence of forward-looking agents,
combined with some other mild conditions, is sufficient to generate model
closure and intertemporal behaviour consistent with the foundations of the
model at hand.
The model-based rules which would result from an application of the
methodology proposed in this paper would share many of the desirable
features of exogenously-imposed rules. Most importantly, they guarantee
solvency on the part of the government and rule out instrument instability.
In addition, the proposal presented here possesses some additional
appealing properties not shared by exogenously imposed fiscal rules.
Firstly, the rules are forward-looking and in a manner consistent with the
specification of other sectors in the economy. More generally, the rules are
consistent with the setup of the model in which they are implemented, by
design, meaning that a change in structural parameters will automatically
be reflected in the fiscal rule. Secondly, they are state-contingent.
Exogenously imposed fiscal rules may involve acyclical features, where,
for example, adjustment of taxes is dependent solely on the observed
deviation of the deficit or debt from its target value. The endogenous fiscal
rules derived here, on the other hand, are exhibit shock-specific fiscal
policy responses, which is a desirable property from an economic point of
view. Thirdly, the rules, in principle, ought to produce relatively smooth
adjustment processes for taxes consistent with the behaviour of households
with concave utility functions - which gives the result that households
smooth consumption. Unlike many exogenously imposed fiscal rules, the
impulse response of variables is consistent with the optimal time path of
adjustment of agents within the model, and adjustment is not dependent on
calibrated parameters. Lastly, the rules would embody only counter-
cyclical automatic adjustments on the part of the fiscal authority unless a
discretionary component is assumed – and in this way can be considered as
mimicking automatic stabilisation properties of government budgets.
5()(5(1&(6
Barrell, R., J. Sefton and J. in’t Veld (1994), “Fiscal solvency and fiscal
policy”, National Institute of Economic and Social Research,
Discussion Paper, No. 67.
Blanchard, O. and R. Perotti (1999), “An Empirical Characterization of the
Dynamic Effects of Changes in Government Spending and Taxes on
Output”, NBER Working Paper, No. W7269.
Bryant, R.C. and L. Zhang (1996a), “Intertemporal fiscal policy in
macroeconomic models: introduction and major alternatives”,
Discussion Paper in International Economics, No. 123, Washington
D.C., Brookings Institution.
___________________________ (1996b), “Alternative specifications of
intertemporal fiscal policy in a small theoretical model”, Discussion
Paper in International Economics, No. 124, Washington D.C.,
Brookings Institution.
Chari, V.V., L. Christiano and P. Kehoe (1994), “Optimal Fiscal Policy in
a Business Cycle Model”, -RXUQDO RI 3ROLWLFDO (FRQRP\, 102,
pp. 616-52.
Johnson, R. (2000), “An implementation of the constant-tax rule in a
numerical macro model”, mimeo, Federal Reserve Bank of Kansas
City, November.
Kilpatrick, A. (2001), "Transparent frameworks, fiscal rules and policy-
making under uncertainty", paper included in this volume.
Laxton, D., P. Isard, H. Faruqee, E. Prasad and B. Turtelboom (1998),
"MULTIMOD Mark III: The core dynamic and steady-state
models", Occasional Paper No. 164, Washington DC, International
Monetary Fund.
Leeper, E.M. (1991), “Equilibria under 'active' and 'passive' monetary and
fiscal policies”, -RXUQDO RI 0RQHWDU\ (FRQRPLFV 27, November,
pp. 129-47.
Lucas, R. (1976), “Econometric Policy Evaluation: A Critique”, -RXUQDORI
0RQHWDU\(FRQRPLFV, 1(2), Supplementary Series, pp. 19-46.
Manzano, B. and J. Ruiz (2000), “Optimal Contingent Fiscal Policy in a
Business Cycle Model”, Paper presented at the 15th Annual
Congress of the European Economic Association held in Bolzano,
Italy, September.
McCallum, B. T. (2001), "Indeterminacy, bubbles, and the fiscal theory of
price level determination", -RXUQDORI 0RQHWDU\ (FRQRPLFV 47(1),
pp. 19-30.
McKibbin, W.J. and J.D. Sachs (1991), "Global Linkages: Macroeconomic
Interdependence and Cooperation in the World Economy",
Washington DC, Brookings Institution.
Mitchell, P.R., J.E. Sault and K.F. Wallis (2000), “Fiscal policy rules in
macroeconomic models: principles and practice”, (FRQRPLF
0RGHOOLQJ 17, pp. 171-93.
Mitchell, P.R., J.E. Sault, P.N. Smith and K.F. Wallis (1998), “Comparing
global economic models”, (FRQRPLF0RGHOOLQJ 15, pp. 1-48.
Novales, A., E. Dominguez, J.J. Pérez, and J. Ruiz (1999), “Solving
nonlinear rational expectations models by eigenvalue-eigenvector
decompositions”, in R. Marimon and A. Scott (eds.),&RPSXWDWLRQDO
PHWKRGV IRU WKH VWXG\ RI G\QDPLF HFRQRPLHV, Oxford University
Press.
Perez, J. and P. Hiebert (2001), “Forecast-error based fiscal policy rules for
macroeconomic models”, mimeo, March.
Sims, C. A. (2000), “Solving linear rational expectations models”, mimeo,
Yale University.
Sims, C. A. (1994), “A simple model for study of the determination of the
price level and the interaction of monetary and fiscal policies”,
(FRQRPLF7KHRU\ 4, pp. 489-94.
Andrés J., F. Ballabriga and J. Vallés (2001), “Non-Ricardian Fiscal
Policies in an Open Monetary Union”, mimeo, January.
Woodford, M. (1995), “Price level determinacy without the control of a
monetary aggregate”, Carnegie-Rochester Conference Series on
Public Policy, 42, pp. 1-46.
75$163$5(17)5$0(:25.6),6&$/58/(6
$1'32/,&<0$.,1*81'(581&(57$,17<
$QGUHZ.LOSDWULFN
,QWURGXFWLRQ
1. Over the past 30 years increasingly rapid flows of information have
radically altered the power and role of the state. The ability of governments
to influence their economies directly has gradually diminished. Electorates
have become more sophisticated and better informed; labour markets have
become more atomised; financial markets are more open and integrated
and exert increasing power across national boundaries; and ever closer
relationships between European States and among the G7 countries, for
example, have strengthened fiscal surveillance and peer pressure. All these
factors have contrived to limit the scope and impact of national fiscal
policies.
2. Furthermore, fiscal policy itself failed more often than not to deliver
stability. This undermined belief in the power of the authorities to deliver
desirable macro outcomes. It is true that the task of fiscal policy was
complicated by exogenous shocks, such as the oil crises in the 1970s, and
persistent trends like the increase in social security spending through the
1980s. And monetary policy has been at least as much to blame.
Nonetheless there have been clear episodes – at least in the UK – when the
operation of fiscal policy has been a destabilising rather than stabilising
force.
3. During this period developments in economic thinking also
contributed to a gradual change of approach – shifting away from activist
or interventionist fiscal policies at the macro level. In the macroeconomic
context, academic attention has for the most part focused on monetary
policy. This has been associated with many worthwhile policy changes,
most notably the focus on credibility and the move towards independent
central banks and greater transparency. There are parallel gains to be made
to fiscal policy, although to date the academic community has paid
relatively little attention to this.
__________
*
HM Treasury – United Kingdom. The views expressed in this paper are those of the author and do
not necessarily represent those of HM Treasury.
4. In a dynamic and information hungry society transparency can
promote a better understanding of policy and greater fiscal discipline,
particularly when combined with well-defined fiscal rules and mature and
respected institutions. This paper explores how a more transparent
macroeconomic framework in the UK has helped put policy on a sounder
footing. In a wider context, the recent IMF Code for Fiscal Transparency
should similarly help improve the conduct of policy in other countries.
5. Although transparency appears to be a necessary condition for a
successful fiscal policy it is unlikely to be sufficient, at least in the first
instance. In particular, fiscal rules can play an important additional role.
The paper describes the way in which fiscal rules have operated within a
transparent framework in the UK since 1997 to produce better outcomes
than experienced in previous decades.
6. Fiscal policy – perhaps in contrast to monetary policy – is a complex
and multi-faceted business. The practical realities of fiscal policy-making
are much more complicated than is typically portrayed in macro textbooks.
All fiscal policy-makers face a number of potentially conflicting pressures
beyond the decision to set the balance between spending and receipts or the
short-term fiscal stance. For example, there will be a need to ensure that:
• public spending plans and the services they deliver meet the needs and
expectations of the electorate;
• the tax burden remains low and the right incentives are in place to
encourage work, enterprise and savings;
• ambitions for redistribution are met; and that
• policy is sustainable over the longer term and that the Government
remains solvent.
7. In the UK the control of fiscal policy (including tax and spending
policy) rests largely in the hands of the Chancellor and the Treasury. The
ex ante decision in setting the balance between spending and taxation is
thus relatively unencumbered by the need to take account of balances
between regions or departments. The legal and parliamentary processes are
also relatively straightforward and well understood. In this respect the
institutional framework is sound and supports the policy-making process.
8. General uncertainty about the state and direction of the economy is
an important influence on the way in which decisions are made. Even
where a high degree of control is exerted from the centre, there remains a
measure of manoeuvre in practice. Data and other lags from administrative,
implementation, legal or other factors often mean that little is known about
the consequences of policy actions until some time after the event.
Forecasting errors can be large but are not often analysed. The paper looks
at the issue of uncertainty in the context of policy-making where the
objectives of fiscal policy are clear and explores the trade-offs that must be
made when making fiscal decisions.
9. In the light of these circumstances it is perhaps not surprising that
many Governments have failed to manage fiscal policy well. The paper
argues for there is a need for a cautious approach to policy-making and that
a transparent macroeconomic framework underpinned by clear fiscal rules
can play an important role in clarifying the purpose and conduct of fiscal
policy and in producing more successful outcomes than in the past.
/HVVRQVRIWKHSDVW
0DFURHFRQRPLFLQVWDELOLW\
10. The primary objective of macroeconomic policy, and thus fiscal
policy, should always be economic stability. Economic stability provides
suitable conditions for the achievement of the high levels of growth and
employment that governments and electorates desire. For this reason the
new macroeconomic framework that has been put in place in the UK since
1997 has been geared towards delivering economic stability.
11. Until recently, however, the UK’s macroeconomic experience was
one of instability. Charts 1 and 2 show the volatile paths for the overall
fiscal deficit, public debt and net worth that the UK experienced from the
1970s on. There were periods of substantial deficit, rapid build-up of debt
and decline in net worth. Similar volatile patterns can be found for growth
and inflation. This instability frequently translated into the uncomfortable
policy choices.
12. Between 1979 and 1996 the overall deficit averaged 3 per cent of
GDP. Moreover, as Chart 3 shows, there was a substantial decline in public
sector net investment, from around 6 per cent of GDP to less than 1 per
cent. Part of this change reflects the impact of privatisation but the general
picture holds if adjustment is made for this factor. Compared with other G7
countries over this period the UK had a more volatile economy, invested
less and grew more slowly (and had higher inflation).
&KDUW
3XEOLFVHFWRUQHWERUURZLQJ
SHUFHQWRI*'3
-4
-2
0
2
4
6
8
10
1970/71 1972/73 1974/75 1976/77 1978/79 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99
1
Excluding windfall tax and associated spending.
&KDUW
3XEOLFVHFWRUQHWGHEWDQGQHWZRUWK
SHUFHQWRI*'3
0
10
20
30
40
50
60
70
80
90
1974/75 1976/77 1978/79 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99
Public sector net worth
Public sector net debt
,PSUHFLVHREMHFWLYHV
13. Looking back, one of the lessons of this period has been the need to
identify – and stick to – clear objectives. Too often in the past the purpose
and precise objectives of fiscal policy were left unspecified or vague,
allowing policy-makers an inappropriate degree of discretion. Policy could
thus be changed in the light of circumstances, both economic and political,
but without great risk of being called to account, at least in the short-term.
3RRUFRRUGLQDWLRQRISROLF\
14. A notably serious failing during this time was improper coordination
of fiscal and monetary policy. This seems paradoxical, given that the
decision to set both the fiscal stance and the interest rate rested in the hands
of one person, the Chancellor of the Exchequer. But as Table 1 shows,
more often than not short-term interest rates were cut within a few days of
the announced Budget package. Between 1979 and 1996, interest rates
were cut on 14 such occasions, and raised in only 2.
0
1
2
3
4
5
6
7
1970/71 1972/73 1974/75 1976/77 1978/79 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99
Public sector net Investment"
&KDUW
3XEOLFVHFWRUQHW,QYHVWPHQW
7DEOH
3RVW%XGJHW,QWHUHVWUDWHFKDQJHV
Budget date
Interest rate
movements
Rate change Decision
12 June 1979 15 June up 2 pps
26 March 1980
10 March 1981 11 March Down 2 pps
9 March 1982 12 March Down ½ pp
15 March 1983 15 March Down ½ pp
13 March 1984 7 March Down ¼ pp
15 March Down ¼ pp
19 March 1985 20 March Down ½ pp
29 March Down ½ pp
18 March 1986 19 March Down 1 pp
17 March 1987 10 March Down ½ pp
19 March Down ½ pp
15 March 1988 17 March Down ½ pp
14 March 1989
20 March 1990
19 March 1991 22 March Down ½ pp
10 March 1992
16 March 1993
30 November 1993 23 November Down ½ pp
29 November 1994 7 December up ½ pp
28 November 1995 13 December Down ¼ pp
26 November 1996
15. It also appears that the stance of fiscal policy did not work fully in
the interests of stability. For example, during the 1980s and early 1990s the
underlying stance of policy seems to have been relatively tight when the
economy was weak and loose when the economy was overheating, as Chart
4 indicates.
-3
-2
-1
0
1
2
3
4
5
6
-8 -6 -4 -2 0 2 4 6
&\FOLFDOHIIHFWV
16. Care was also needed in taking account of the effects of the
economic cycle on the public finances. A particularly instructive episode
occurred between 1986 and 1996. During this economic cycle output is
estimated to have ranged from some 4 per cent above trend in the late
1980s to a similar amount below trend in the recession of the early 1990s.
The overall fiscal balance also fluctuated significantly, from a surplus of
just over 1 per cent at the peak in 1988 to a deficit approaching 8 per cent
of GDP by 1993.
17. When the fiscal position moved into surplus (1988-89) after a long
period in deficit there was a belief that the economy’s potential had
increased - there were few visible signs of inflation for example - and that
the economy would remain strong and the surplus would last.
Commentators and politicians began to talk of ‘supply-side’ miracles and
even in terms of ‘repaying the national debt’, and policy was loosened.
18. The underlying position was not so rosy however. Potential output
had not risen as far as people thought. And when it became clear output
was well above potential it was too late: inflation accelerated, policy was
tightened and the economy went into reverse.
&KDUW
2YHUDOO6WUXFWXUDO%DODQFHDQGWKH&\FOHWR
19. One lesson to be drawn from this episode is the importance of trying
to take the impact of the economic cycle on the public finances into
account. Such estimates are necessarily imprecise but it is far better to
make some attempt to adjust for the cycle than none at all.
20. Had such an attempt been made an unwarranted loosening of fiscal
policy at the top of the cycle might not have occurred. In retrospect, the
output gap was strongly positive by the late 1980s and implied a large
component of the fiscal surplus was due to cyclical, and thus temporary,
rather than structural factors. Estimates now show that rather than being in
surplus the underlying fiscal position was one of deficit of the order of
1 per cent of GDP. The loosening of policy merely acted to compound the
difficulties of an overheating economy and forced a greater retrenchment
subsequently than otherwise necessary.
7UHQGJURZWK
21. A further dimension to this experience concerned the assessment of
trend growth, rather than the position of the economy in relation to trend.
Trend growth was thought to have increased substantially. In fact,
throughout the 1980s the economy had confounded the critics and
recovered strongly with little evidence of inflation.
22. Published estimates of the medium-term growth path set out in
successive Budget reports were thus moved up from an average of 2¼ per
cent annual growth (over 4-5 years ahead) in the early part of the period to
3 per cent by 1989. Based on an assumption of a medium-term growth path
of 3 per cent the fiscal projections were duly flattered, making tax cuts
seemingly viable and consistent with a stable fiscal and debt position. It
was not long however before it became clear trend growth had been
significantly overestimated and had to be revised down.
&RQVHTXHQFHVIRUSXEOLFVSHQGLQJ
23. By creating uncertainty, mistakes over medium term assumptions
also carried important consequences for public spending settings. In
parallel with the often short-term and expedient approach to fiscal policy a
ritual of bid, counterbid, escalation and deadlock characterised much of
spending policy during this time. Annual jousting between spending
departments and the Treasury would typically continue until time ran out
and a special council of Ministers (eg Star Chamber) or even the Prime
Minister was required to declare in favour of Department X or Y.
24. In effect the planning horizon was a year ahead. This led to
inefficiencies of various kinds; including, for example, poor value for
money spending at year-end under a ‘use it or lose it’ mentality. While
indicative plans were set for years 2 and 3, little attention was paid to them,
not least because the economic and fiscal projections on which they were
based constantly changed. Institutional short-termism prevailed and
investment in particular suffered as a result.
6RPHFRQFOXVLRQV
25. Some broad conclusions and lessons may be drawn from the
experiences of this period. In particular, fiscal policy should benefit from:
• clear principles and targets;
• transparency and accountability;
• well-defined roles and responsibilities among the key actors;
• adequate mutual support between fiscal and monetary policy in
promoting stability;
• caution against over-optimistic and insufficiently forward-looking
assessments;
• taking account of the effects of the cycle on the public finances;
• a more stable and long-term framework for public spending
decisions; and
• fiscal (and monetary) decisions should always made in the best
interests of the economy.
5HIRUPVWRWKHIUDPHZRUN
/HDUQLQJWKHOHVVRQVRIWKHSDVW
26. The design of the UK’s new macroeconomic framework has taken
account of the lessons of the past. In addition, policy since 1997 has been
set to restore the health of the public finances. The Government’s reforms
have been based on sound principles and are supported by clear objectives
and firm rules. By aiming to create a climate of low inflation and strong
public finances not only is the basis for economic growth enhanced but it is
possible to take a much longer term view in plotting a course for public
spending and, in particular, to create a firmer path for public investment.
27. Rather than leaving too much scope for policy to be decided on the
basis of expediency, fiscal policy now operates under three significant
types of constraint:
• first, there is greater WUDQVSDUHQF\DQGDFFRXQWDELOLW\;
• second, LQGHSHQGHQWDQGWUDQVSDUHQWPRQHWDU\SROLF\;
• third, there are firm ILVFDOUXOHV
28. In all this the need to restore and subsequently maintain the
credibility of policy has been regarded as of paramount importance.
7UDQVSDUHQF\DQGDFFRXQWDELOLW\
29. Given the history of UK fiscal policy, and the fact that the public
sector deficit was over 4 per cent of GDP in 1997, it was clear when the
new Government entered office that considerable changes were necessary
to achieve a greater degree of credibility in policy making.
30. The need for transparency and the advantages of policy credibility
are well documented for monetary policy. Credibility and transparency are,
however, just as important when it comes to setting fiscal policy. For this
reason, a similar philosophy to that adopted for monetary policy was
pursued in the case of fiscal policy.
/HJLVODWLYHIUDPHZRUN
31. The Code for Fiscal Stability (HM Treasury (1998f)) was thus
created to provide a basic structure for the framework and was given
statutory backing in the Finance Act 1998. The Code strengthened the
openness, transparency and accountability of fiscal policy, features that
also characterised the framework for monetary policy following the
introduction of the 1998 Bank of England Act. It also improved the quality
of information given to the public, the lack of which in the past was an
important factor behind policy mistakes.
32. Legislating to make the Code a formal requirement of this
Government’s fiscal policy and those of future governments was an
important step forward. For good reason a number of details were left
outside the statute book. But what the legislation did do for the first time
was formally require any government to:
• specify its principles of fiscal management and state the objectives of
fiscal policy;
• set out key annual reporting requirements, in particular a consultative
Pre-Budget Report, an Economic and Fiscal Strategy Report as well as
the traditional Financial Statement and Budget Report, and a Debt
Management Report (see box 1); and
• adopt best practice accounting standards.
33. The Code draws together and makes clear the framework within
which fiscal policy must operate. It demonstrates the Government’s
commitment to well-based policy, helping to improve the time consistency
of fiscal policy. In other words, it supports the idea that the optimal policy
for the Government remains the same over time, helping to ensure that
short-term expediency does not take precedence over long-term planning.
3ULQFLSOHVRIILVFDOPDQDJHPHQW
34. Five principles of fiscal management are at the heart of the
framework:
• WUDQVSDUHQF\ in the setting of fiscal policy objectives, the
implementation of fiscal policy and in the publication of the public
accounts;
• VWDELOLW\ in the fiscal policy making process and in the way fiscal
policy impacts on the economy;
• UHVSRQVLELOLW\ in the management of public finances;
• IDLUQHVV, including between generations; and
• HIILFLHQF\ in the design and implementation of fiscal policy and in
managing both sides of the public sector balance sheet
35. These principles help to make clear what policy-makers should have
uppermost in their minds when setting policy.
5ROHVDQGUHVSRQVLELOLWLHV
36. An integral part of any successful management framework is a
%R[
.(<3529,6,2162)7+(&2'()25),6&$/67$%,/,7<
8QGHU WKH &RGH WKH *RYHUQPHQW XQGHUWDNHV D QXPEHU RI
FRPPLWPHQWV,WPXVW
• FRQGXFWILVFDODQGGHEWPDQDJHPHQWSROLF\LQDFFRUGDQFHZLWKD
VHWRIVSHFLILFSULQFLSOHV
• VWDWHH[SOLFLWO\LWVILVFDOSROLF\REMHFWLYHVDQGRSHUDWLQJUXOHV
DQGMXVWLI\DQ\FKDQJHVWRWKHP
• RSHUDWH GHEW PDQDJHPHQW SROLF\ WR DFKLHYH D VSHFLILF SULPDU\
REMHFWLYH
• GLVFORVH DQG TXDQWLI\ ZKHUH SRVVLEOH DOO GHFLVLRQV DQG
FLUFXPVWDQFHV ZKLFK PD\ KDYH D PDWHULDO LPSDFW RQ WKH
HFRQRPLFDQGILVFDORXWORRN
• HQVXUH WKDW EHVW SUDFWLFH DFFRXQWLQJ PHWKRGV DUH XVHG WR
FRQVWUXFWSXEOLFDFFRXQWV
• SXEOLVK D SUH%XGJHW UHSRUW WR HQFRXUDJH GHEDWH RQ WKH
SURSRVDOVXQGHUFRQVLGHUDWLRQIRUWKH%XGJHW
• SXEOLVKD)LQDQFLDO6WDWHPHQWDQG%XGJHW5HSRUWWRGLVFORVHWKH
NH\%XGJHWGHFLVLRQVDQGWKHVKRUWWHUPHFRQRPLFDQGILVFDO
RXWORRN
• SXEOLVK DQ (FRQRPLF DQG )LVFDO 6WUDWHJ\ 5HSRUW RXWOLQLQJ WKH
*RYHUQPHQW¶VORQJWHUPJRDOVVWUDWHJ\IRUWKHIXWXUHDQGKRZLW
LVSURJUHVVLQJLQPHHWLQJLWVILVFDOSROLF\REMHFWLYHV
• SXEOLVKDVSHFLILFUDQJHRILQIRUPDWLRQLQLWVHFRQRPLFDQGILVFDO
SURMHFWLRQV LQFOXGLQJHVWLPDWHVRI WKH F\FOLFDOO\DGMXVWHG ILVFDO
SRVLWLRQDQGORQJWHUPSURMHFWLRQVWRDVVHVVWKHVXVWDLQDEOLOW\RI
SRLF\
• LQYLWH WKH 1DWLRQDO $XGLW 2IILFH WR DXGLW FKDQJHV LQ WKH NH\
DVVXPSWLRQV DQG FRQYHQWLRQV XQGHUSLQQLQJ WKH ILVFDO
SURMHFWLRQV
• SURGXFHD'HEW0DQDJHPHQW5HSRUWRXWOLQLQJWKH*RYHUQPHQW¶V
GHEWPDQDJHPHQWSODQV
• UHIHUDOOUHSRUWVLVVXHGXQGHUWKH&RGHWRWKH+RXVHRI&RPPRQV
7UHDVXU\6HOHFW&RPPLWWHH
• HQVXUHWKDWWKHSXEOLFKDVIXOODFFHVVWRWKHUHSRUWVLVVXHGXQGHU
WKH&RGH
proper assignment of roles and responsibilities among the main actors. In
the case of the Monetary Policy Committee (MPC), established in 1997, its
job is to set interest rates to achieve the Government’s inflation target.
They are thus held to account for their performance in achieving low
inflation.
37. Responsibility and accountability for fiscal policy is equally clear.
The Chancellor has set clear targets against which commentators, the
public and Parliament alike can assess performance. Furthermore, under
the new public spending regime departments and their Ministers have been
allocated funds over three years in return for agreed pledges to achieve a
range of specific results (known as Public Service Agreements), all of
which are published and subject to regular appraisal.
&RQVHQVXV
38. Compared with the past there is a reasonable consensus over the
broad parameters of what constitutes sensible fiscal policy, and perhaps
even more agreement over what represents bad policy. Transparency helps
to build up the constituency for stability-oriented policies and encourages
people and businesses to plan for the long term, rather than basing
decisions only on what makes sense in the short term. In this way it helps
to allocate both public and private resources efficiently.
&\FOLFDODGMXVWPHQW
39. One aspect of the fiscal framework that builds on past experience is
the view that prominence should be given to the underlying fiscal position
through the use of cyclically-adjusted indicators. Estimates of structural
fiscal balances were thus published for the first time by a UK Government
in 1997.
40. There are uncertainties and differences of view over methodologies
used for calculations of structural balances so it follows that a high degree
of transparency is appropriate here too. The UK published a paper (HM
Treasury (1999b)) explaining how the cyclically-adjusted estimates
presented in Budget reports are constructed. It is thereby open to others to
make their own calculations for the purposes of assessing the
Government’s plans and comparing them against the Government's view;
and for example major international organisations such as the IMF, OECD
and European Commission do this.
$XWRPDWLFVWDELOLVHUV
41. Focusing on the cyclically-adjusted position also gives appropriate
prominence to the automatic stabilisers and the role they play in smoothing
the path of output by boosting aggregate demand when the economy is
below trend and curbing demand when the economy is above trend. The
strength of the automatic stabilisers will depend on particular
characteristics of the taxation and spending regimes, for example the
progressivity of taxes. But when considering the extent to which the fiscal
policy framework is supporting monetary policy, the strength of the
automatic stabilisers should be taken into account.
.H\DVVXPSWLRQV
42. Transparency over the key assumptions which form the basis of the
fiscal projections is important. It allows commentators and others scope to
assess the realism of the projections on which policy is based, and to
explore variants. In the UK these assumptions are independently audited by
the National Audit Office (NAO) who consider whether they are realistic
and cautious. Their reports are presented to Parliament and form part of the
Budget documentation (see, for example, NAO (2001)).
43. The Government also sets store in focusing on the longer term and
avoiding the risk of short-term reversals in policy. Thus policy is assessed
annually against longer-term developments, such as the consequences of
the ageing of the population.
6FUXWLQ\DQGDFFRXQWDELOLW\
44. The ability to conduct high quality scrutiny also depends on
transparency in a different direction; namely on timely, accurate and
relevant statistics and high quality accounting standards. The UK’s track
record on statistics is good and has been enhanced in a number of ways to
allow a better assessment of the fiscal aggregates. Moreover, all relevant
figures are presented in line with internationally accepted statistical
definitions.
45. Accounting standards in the public sector are undergoing a process
of change as the UK moves onto a resource and budgeting basis, as has
occurred in New Zealand and Australia for example (HM Treasury
(1999c)). They are thus becoming increasingly in line with best practice in
the private sector.
46. Although it is still a little way off, developments are well under way
to produce a set of Whole of Government Accounts by 2005-06. This will
be the first time a full consolidated set of accounts for the government
sector will be available in the UK. Consolidated accounts for central
government are likely to be available in 2003-04 and will help improve the
quality of public spending decisions in forthcoming Spending Reviews.
47. Published Budget and other reports such as the UK Convergence
Programme (HM Treasury (2000d)) contain a considerable amount of
information on the progress of and outlook for fiscal policy. The ability
and willingness to assess performance in a transparent way – for example
by including details of fiscal aggregates on a cyclically-adjusted basis and
by providing long-term fiscal projections - is an important feature of sound
policy. So too is a strong commitment to long-term goals and a willingness
to make policy-adjustments (in either direction) so that policy remains on
track.
2EMHFWLYHV
48. Above all, there must be transparency over the objectives of policy.
As set out in Analysing UK Fiscal Policy (HM Treasury (1999f)) the key
objectives of the Government’s fiscal policy are:
• over the PHGLXP WHUP, to ensure sound public finances and that
spending and taxation impact fairly both within and across generations.
In practice, this requires that:
½ the Government meets its key taxation and spending priorities
while avoiding an unsustainable and damaging rise in the burden
of debt; and
½ those generations who benefit from public spending also meet, as
far as possible, the costs of the services they consume;
• over the VKRUWWHUP, to support monetary policy, by:
½ allowing the automatic stabilisers to play their role in smoothing
the path of the economy in the face of variations in demand; and
½ to provide further support to monetary policy through changes in
the fiscal stance, where prudent and sensible.
49. The Government’s specific fiscal rules provide the operational basis
for achieving the medium term goals while the independent central bank
conditions the scope of fiscal policy over the shorter term. We turn to these
issues.
,QGHSHQGHQWDQGWUDQVSDUHQWPRQHWDU\SROLF\
50. Proper coordination of fiscal and monetary policy is an important
feature of the new macroeconomic framework. A high degree of
transparency, as well as clear roles and responsibilities for all parties,
ensures that the monetary policy authority is aware of fiscal policy
objectives and performance, including performance against those
objectives.
51. Monetary policy and fiscal policy each have the same aim of
underpinning long-term growth through economic stability so it is
appropriate for policy to be properly coordinated. Clearly defined
objectives and transparent procedures enhance this. The MPC and the
Government are each aware of what the other is trying to achieve. The
arrangements allow the Treasury (non-voting) representative to participate
fully in MPC deliberations, for example helping to provide MPC members
with a good understanding of tax and spending developments during the
year. In this way the main players are able to become aware of the likely
reaction to each other's policy decisions.
52. Awareness of the policy reaction function by key players is one
aspect of the system. However, the transparency and independence of the
monetary policy framework imposes a particularly important discipline on
fiscal policy. Its real impact in this context comes through the risk that
interest rates are raised in reaction to fiscal policy, and through comments
on the policy setting in the published minutes of the MPC meetings and the
Inflation Report. In effect, these provide a high profile judgement on the
Budget and other aspects of fiscal policy, which need to be taken into
account by the fiscal policy-maker if credibility is to be maintained.
53. There can be no clearer contrast with the past. The present
arrangements provide a credible threat of an interest rate rise in the event
of inappropriate fiscal policy. This constrains budgetary policy in a more
time consistent way than previously where a Chancellor remained free to
choose the option of declaring his economic policy, and the Budget in
particular, a success and worthy of an interest rate cut.
)LVFDOUXOHV
54. A transparent macroeconomic framework along these lines goes a
long way in holding policy to a more reasonable set of outcomes.
Nonetheless, in the context of a new framework, which takes time to be
fully understood, a simple guide to the operation of policy is also
necessary. Hence the need for fiscal rules.
55. Ideally, such rules should be realistic and relevant, capable of being
understood by all, as well as being measurable and achievable. They
should also be applied consistently. Different rules might share these
characteristics but it is unlikely that universal rules exist. For this reason it
was decided not to incorporate the specific fiscal rules within the UK
legislation.
56. Different rules should reflect, for example, the different
circumstances countries face at any particular time. Nonetheless, for a
group of countries such as those in the European Union a common set of
fiscal rules may be appropriate provided the diversity of Member States’
circumstances can be taken into account in the application of those rules.
57. A fiscal rule, or set of rules, is necessarily somewhat arbitrary.
However, rules can serve as a guide to better behaviour and for this reason
alone it is likely to make sense to introduce some rules. Once this has been
done it is important for the credibility of policy to stick with them.
58. Two fiscal rules apply in the UK:
• 8. ILVFDO UXOHV WKH JROGHQ UXOH: over the economic cycle, the
Government will borrow only to invest and not fund current
spending; and
• WKH VXVWDLQDEOH LQYHVWPHQW UXOH: public sector net debt as a
proportion of GDP will be held over the economic cycle at a stable
and prudent level.
59. The fiscal rules provide benchmarks against which the performance
of fiscal policy can be judged. The Government will meet the golden rule
if, on average over a complete economic cycle, the current budget is in
balance or surplus. The Government has also stated that, other things being
equal, a modest reduction in public sector net debt to below 40 per cent of
GDP over the economic cycle is desirable.
5DWLRQDOH
60. Economic theory sheds light on the UK’s past experience and
strengthens the rationale for stating explicit fiscal rules. The benefits of
establishing a sound and stable fiscal framework will be maximised if the
framework is credible - that is, if households and firms believe firmly that
the Government will deliver its commitments.
61. If the policy framework lacks credibility, households and firms will
continue to base their decisions on previous experience. Savings and
investment decisions would continue to anticipate poor fiscal management
and a return to volatile output, high inflation and low growth. The benefits
of a new framework would thereby be delayed until the Government was
able to establish a convincing track record of favourable policy outcomes.
62. Firm fiscal rules also modify the tendency for fiscal policy to deviate
from sound economic principles to provide short-term gains to certain
interest groups. Such tendencies often occur, not surprisingly, close to
elections. Indeed, as Keech (1985) suggests, even if a fiscal rule is not
optimal in a perfect world, it may well be the best economic response in a
situation where the unconstrained political process produces outcomes that
are even less desirable.
)OH[LELOLW\F\FOHVDQGRWKHUVKRFNV
63. Rules, by their very nature, are intended to impose restrictions on
behaviour. Fiscal rules can ensure that the public finances are managed
prudently and are maintained within sensible boundaries so that
Governments meet their spending commitments without jeopardising
economic stability or running up an unfair bill for future generations.
64. It is important, however, that the chosen rules allow sufficient
flexibility to react sensibly to economic developments. The right balance
needs to be struck between a rigid mechanical approach and one based on
unfettered discretion. In particular, there must be scope to accommodate
the impact of the economic cycle and room to act in the event of
exceptional economic shocks.
65. Fiscal policy can help to stabilise the economy through the operation
of the automatic stabilisers. These movements support monetary policy by
dampening economic cycles without putting at risk the long-term
sustainability of fiscal policy. It is essential that the chosen rules do not
override this inbuilt capacity to respond to changing economic
circumstances.
66. The chosen fiscal rules should also make room for sensible
discretionary adjustments to fiscal policy. For example, in the first years of
this Government’s term of office fiscal policy was additionally tightened to
support monetary policy and to restore the structural integrity of the public
finances as quickly as possible.
67. Fiscal rules should similarly incorporate a measure of flexibility to
accommodate exceptional shocks, not associated with the usual economic
cycle. The Code for Fiscal Stability thus permits a Government to deviate
from its fiscal rules in exceptional circumstances, such as wars or natural
disasters for example. The Code requires the reason for any such change to
be explained in public, with guidance on how policy will operate in the
meantime and for how long (if known).
6FRSH
68. A further consideration is the ambit of the rules. Ideally, they should
relate to the whole public sector, ie general, central and local government
and public corporations. The liabilities of public corporations could fall
ultimately on the taxpayer so it is appropriate that the fiscal rules extend
beyond the general government sector. Moreover, if the rules were applied
to these activities alone it could lead to perverse incentives to reclassify
spending in an attempt to get around the fiscal rules.
*ROGHQUXOH
69. The previous fiscal policy regime made no formal distinction
between current and capital spending and concentrated on a cash
aggregate, the public sector net cash requirement. A significant
shortcoming of this approach was that it created a bias against capital
spending. It also gave misleading signals when public assets were sold off.
Current and capital spending could be offset against each other, making
capital projects - where returns appear only in the future - an easy target
when it became necessary to tighten the overall fiscal policy stance. The
bias against capital contributed to a considerable under-investment in
public assets (HM Treasury (1998c)).
70. The golden rule draws a distinction between current and capital
spending and is designed to remove the bias against capital spending. The
rule also gives consideration to fairness, and in particular, fairness between
generations. Government decisions on spending and revenue may have
important implications across generations. For example, large-scale
investments, such as roads, produce benefits not just at the beginning but
over the whole of the investment’s effective life, which may be in excess of
40 years. It is fair that those generations who benefit from this spending
also meet some of the costs.
71. It is not practical, of course, to match the timing of the streams of
costs and benefits for each and every spending proposal. But, in aggregate,
the Government takes the view that current spending, which mainly
provides benefits to existing taxpayers, should be paid for by the current
generation of taxpayers. Similarly, because capital spending produces a
stream of services over time, it is appropriate that this form of spending is
financed initially through borrowing. As far as possible each generation
should pay for the benefits of the public services that it consumes.
72. It follows naturally that the definitions of current and capital
spending are important to the application of the rule. For the purpose of the
fiscal rules, the Government considers the best measure of capital currently
available is that used in the national accounts (Office of National Statistics
(1998)).
73. Another key feature of the golden rule, as the Government has
adopted it, is that it is defined over the economic cycle. As mentioned
earlier this allows room for the automatic stabilisers to operate freely. This
characteristic is shared with the sustainable investment rule.
6XVWDLQDEOHLQYHVWPHQWUXOH
74. The Government’s motive for borrowing reflects considerations
related to fairness between generations and factors related to the economic
cycle. Borrowing allows the government to spread the upfront costs
associated with capital projects across generations, so that the costs and
benefits are matched more evenly. Even in the absence of major
catastrophes such as war, most countries have positive levels of public net
debt. However, in many cases this is symptomatic of poor control of public
spending rather than high investment.
75. As noted above, the golden rule allows governments to borrow for
the purposes of investment. If left unconstrained, however, it is
conceivable that borrowing could reach levels that are too high,
notwithstanding the specific merits of the underlying investment. This
possibility motivates the Government’s sustainable investment rule, that net
debt should be held at a stable and prudent level.
/LPLWVWRSXEOLFVHFWRUGHEW
76. The concept of sustainability involves analysing the conditions
required to stabilise public debt at a given proportion of GDP. On this
basis, a fiscal policy is usually defined as sustainable if, given reasonable
assumptions, the government can maintain its current policies indefinitely
while continuing to meet its debt obligations.
77. If the real interest rate exceeds the real growth rate - as has been the
case for most of the last two decades - a primary surplus is generally
required to prevent the debt ratio from rising. The extent of this surplus
depends on the size of the interest rate-growth gap and the target public
debt to GDP ratio.
78. The risks that are faced by a country with high levels of public debt
are well known and readily apparent from calculations of primary balances
required to stabilise debt under different conditions. A seemingly
sustainable fiscal policy can quickly become unsustainable when real
interest rates outstrip growth. And the costs of fiscal policy becoming
unsustainable are likely to be high. Indeed, the corrective action needed to
avert a fiscal crisis or the debt servicing obligations created by rising debt
can threaten economic and political stability. Therefore, a disciplined and
prudent approach to fiscal policy is sensible.
79. A prudent fiscal policy can be defined as one that is likely to be
sustainable even in the event of adverse shocks. Thus, a prudent fiscal
policy is likely to lead governments to select a lower level of public debt.
80. It is important to avoid high levels of debt as well as unsustainable
levels. Although there is no clear consensus on the optimal level of public
debt, it is clear that high levels of public debt can limit the effectiveness of
policy. For example, high levels of public debt can:
• PDNHWKHSXEOLFILQDQFHVPRUHYXOQHUDEOHWRLQFUHDVHVLQLQWHUHVWUDWHV
DQGHFRQRPLFVKRFNV at high levels of public debt a sustainable fiscal
policy can quickly become unsustainable through adverse movements
in interest rates and/or growth rates;
• erode the ability of fiscal policy to buffer the economy against major
shocks: if debt is not maintained at low levels during favourable
economic times, there will be reduced scope for supporting monetary
policy and cushioning the economy when faced with unfavourable
shocks;
• OHDGWRDKLJKHUULVNSUHPLXPLQLQWHUHVWUDWHV: high public debt levels
increase default risk which leads to greater risk premia, higher interest
rates and potentially ’crowding out’ effects;
• OHDG WR D ORZ OHYHO RI *RYHUQPHQW VHUYLFHV SHUXQLW RI WD[ FROOHFWHG
ORZHU OHYHOV RI HFRQRPLF ZHOIDUH DQG KLJKHU OHYHOV RI VWUXFWXUDO
XQHPSOR\PHQW: high debt levels imply high levels of debt servicing -
resources that would otherwise be available for spending programmes
or to be distributed as tax cuts.
• WKUHDWHQLQWHUJHQHUDWLRQDOIDLUQHVV : high initial levels of debt can put
such intergenerational equity at risk, especially where pensions
systems are unfunded.
81. Even if fiscal policy is sustainable, the public debt ratio may not be
at an optimal level. Some level of public debt is clearly justified. However,
as noted above, high levels of public debt make the economy vulnerable to
the need for large adjustments in fiscal policy with negative consequences
for long-term growth and employment. This suggests that there may be a
middle ground: a level of debt that represents an optimal trade-off between
the need to undertake public investment (and funding this in an equitable
way) and the economic costs associated with higher levels of public debt.
82. A small number of academic studies have tried to identify the
optimal public debt ratio using empirical means. Three approaches may be
noted:
• inferring the optimal debt ratio by observing debt/equity ratios
prevailing in the private sector. The assumption implicit in this
approach is that whatever the optimal debt ratio may be, the private
sector has solved this to its own satisfaction. Thus given the
Government’s estimated assets, one could argue that the optimal debt
ratio for the UK may lie somewhere in the range of 30-50 per cent of
GDP (based on private sector gearing ratios). However, given the
differing risk characteristics of activities in the public sector the use of
private sector benchmarks is questionable, and even more so when
applied at an aggregate level.
• LQIHUULQJWKHRSWLPDOGHEWUDWLRIURPWHVWVRIG\QDPLFHIILFLHQF\. This
approach stems from economic theory and involves analysing
differentials between investment and profit levels or, alternatively,
economic growth rates and interest rates. One US study by Zee (1988)
suggested that the optimal public debt level is less than 20 per cent of
GDP, although the results are conditional on the parameters and
assumptions made in the model.
• HVWLPDWLRQRIWKHRSWLPDOSXEOLFGHEWUDWLRXVLQJVWDWLVWLFDOWHFKQLTXHV.
One study by Smyth and Hsing (1995) using US data suggested that
economic growth is maximised when public debt levels are around 50
per cent of GDP. Robson and Scarth (1997) argue for a target of 20 per
cent of GDP in the Canadian context. By contrast, another US study by
Asilis (1994) suggested that the costs of being away from the optimal
level are quite small: public debt levels need to rise substantially before
serious damage to the economy will occur. More generally, it is
important to note that the public debt that maximises growth need not
correspond to that which maximises welfare.
83. The methods and assumptions underpinning each of these
approaches are open to criticism; and the range of results illustrates the
difficulty encountered in arriving at a precise answer to the optimal debt
ratio question.
84. While it may not be possible to make definitive statements about
optimal ratios of debt to GDP, it is possible to say that lower debt ratios
allow more room for manoeuvre to redress shortfalls in public investment
or to undertake tax reforms that might enhance potential output and
employment. Lower debt thus conveys some advantages and over time
may provide room for more flexibility in the interpretation of shorter term
rules, such as overall budgetary balance.
&RQFOXVLRQ
85. The golden rule is particularly appropriate in the UK context, given a
history of public sector underinvestment and run down of net worth. The
sustainable investment rule ensures borrowing to finance investment is not
excessive and remains consistent with fiscal policy sustainability. By
setting the rules over the economic cycle appropriate scope is given to the
operation of the automatic stabilisers. The choice of holding public net debt
below 40 per cent of GDP (roughly equivalent to less than 50 per cent on
the Maastricht definition) is somewhat arbitrary, though not unreasonable.
It means that the public debt ratio is amongst the lowest in the EU and G7
and provides room to cope with unforeseen shocks.
8QFHUWDLQW\DQGFDXWLRQ
86. In considering the operation of fiscal policy, and Budget decisions in
particular, it is important to recognise the high degree of uncertainty that
surrounds any fiscal judgement. As noted earlier, the costs of making
faulty judgements can be severe. It is thus important to make every effort
to meet a set of fiscal rules so as to achieve credibility. The strength of this
commitment will be tested by the process through which judgements are
made in trying to meet the rules.
&UHGLELOLW\DQGFDXWLRQ
87. In order to have a good chance of meeting the fiscal rules there is a
need to guard against future uncertainties. If the costs of policy reversals
are high, and if a set of fiscal rules should be met to generate credibility, it
becomes all the more important to take a cautious approach.
88. But it is also the case that there are dangers in excessive caution,
since that might imply a sub-optimal tax or spending path, or
intergenerational imbalance, with associated welfare costs. A suitable
balance therefore needs to be struck between credibility and caution on the
one hand and efficiency, growth and fairness on the other.
89. If the sole aim of policy was to meet a particular fiscal rule, there is
no reason in theory why it should not be possible to achieve a very high
probability of doing so if the fiscal policy-maker is prepared to build in a
sufficiently large margin for error and does not mind systematically
overachieving the rule.
90. However, while there are potential benefits from budget surpluses,
such as lower risk premia, lower debt service costs and reduced
vulnerability to shocks, there are also potential costs, arising for example
from the distortionary effects of taxation, political costs associated with
ignoring calls for tax cuts/more spending
1
and costs associated with
possible distortions to intergenerational fairness.
91. This is shown conceptually in Chart 5. The first line shows the
expected costs associated with failing to meet the rule. These decrease as
the certainty of meeting the rule rises. The second line illustrates the
implicit expected costs of overachievement. As the probability of success
moves towards certainty the margin required to raise that probability – and
hence the distortionary taxes needed – rises at a disproportionate rate. The
cost functions are assumed to be non-linear because the cost of missing the
rule is likely to depend increasingly on the amount by which the rule is
missed, in both directions. The chart illustrates an assumed asymmetry in
that the costs of missing the rule on the downside (risking credibility) are
greater than those associated with overachieving the objective. A total cost
curve can be found by summing the two cost functions. Assuming the aim
is to minimise the total expected costs, a margin sufficient to raise the
probability of meeting the rule consistent with the lowest point of the total
cost curve is required.
92. It is conceivable that such a trade-off may change over time. For
example, the costs of missing the rule may diminish as a fiscal framework
becomes more established, and its credibility grows.
3UHFDXWLRQDU\VDYLQJ
93. The problem facing a responsible government in trying to smooth
taxes and/or spending in the face of future uncertainties is not dissimilar to
that facing consumers who want to smooth the path of consumption/utility.
__________
1
A recent study by the US General Accounting Office suggests that public support for fiscal
discipline quickly dissipates when there is excess money at the end of the year, see “Budget
Surpluses: Experiences of other nations and implications for the US” General Accounting Office
(1999).
&KDUW
&RVWVRIPHHWLQJDILVFDOUXOH
Being aware of the discomfort that unfavourable future shocks may bring
consumers build up precautionary savings in an attempt to cover such risks.
The greater the future uncertainties or the tighter the liquidity constraint
faced by the consumer, the greater the degree of precautionary saving.
94. Governments in industrialised countries are unlikely to face
problems of solvency; and they of course retain the power to tax. In this
respect they face less severe constraints than many consumers.
Nonetheless, the desire to be re-elected may operate in a similar way to a
liquidity constraint. In particular, Governments in some instances may
wish to insure themselves against the consequences of unexpected adverse
shocks in the run-up to an election. Moreover, at the margin a less than
cautious approach could force up risk premia on debt; while attempts to
raise taxes may be counter-productive. A responsible approach to fiscal
policy is thus likely to imply a certain degree of precautionary saving.
X %
Expected cost of
missing the fiscal rule
Cost of
overachieving
the fiscal rule
100%
50%
Probability of meeting rule
Cost
7RWDO&RVW
&RQWUDVWVZLWKPRQHWDU\SROLF\
95. In the monetary policy context, outcomes above or below the
inflation target are regarded as equally undesirable. The independent
central bank is able to raise or lower interest rates, at frequent intervals if
necessary, without fear of asymmetrical costs or penalties, provided
inflation is already low. Policy can be conducted in a time consistent
manner. In fiscal policy, however, the position is rather different.
96. Once a set of fiscal rules is in place it will generally be better for
credibility to meet the rules by a small margin rather than to miss by an
equivalent amount, particularly when the rules are new. The Government
in effect faces a reputational cost function which is asymmetric. Its form
PLJKWEHWKRXJKWRIDV
' PD[^ '`ZKHUH'LVWKHDFFXPXODWHG
deficit since the Government entered office, or when the rule was set.
97. An example is shown in Cart 6. This illustrates a quadratic
UHSXWDWLRQDO ORVV IXQFWLRQ RI WKH IRUP
' ' ò ' +HUH
reputational losses increase with the size of the deficit; and there is an
increasing marginal cost associated with a large accumulated deficit (or
surplus). Once in deficit, the Government faces reputational losses and to
stem the flow would require a surplus to meet the rule. If surpluses are
excessive, the Government also suffers a reputational loss, on the basis the
fiscal rule(s) is significantly overachieved and the electorate are
unimpressed by the failure to hand back the excess in the form of tax cuts
or better public services. The Government thus prefers to achieve a modest
VXUSOXVDQGDLPVWREHLQWKHUHJLRQ$%DW
The government’s reputational loss is:
Φ (D) = φ D + ½ ψ D
2
where , ψ > 0
98. It is also generally much easier to reduce taxes than to raise them;
and to increase spending rather than cut it. There are several reasons for
this:
• political factors: governments face elections and the instruments of
fiscal policy are powerful and tempting tools in the electoral armoury;
• administrative, legal, parliamentary and other lags make reversals of
policy difficult to achieve;
• losers, ie current taxpayers or recipients of public spending, are likely
to press for compensation whereas gainers (such as future taxpayers)
will be less vociferous.
$
&KDUW
7KH*RYHUQPHQW¶VUHSXWDWLRQDOORVVIXQFWLRQ
99. Fiscal decisions in Budgets are normally conducted annually and this
is a further reason why fiscal policy cannot operate like monetary policy.
But it is also the case that the principle of tax (and spending) smoothing
implies an asymmetric effect in the face of meeting a fiscal rule. The
deadweight costs
2
of distortionary taxes are an increasing function of the
tax rate so any requirement to raise taxes to meet a fiscal rule, eg at the end
of a cycle or a Parliament, will be more costly than an equivalent reduction
in taxes.
__________
2
A deadweight cost reflects the efficiency loss incurred when the imposition of non-lump sum taxes
drives the economy away from its free market equilibrium.
Φ'
A$ A%
A'
A
D
100. For these reasons time inconsistency remains a problem in the fiscal
policy context. The external constraints imposed by transparency and fiscal
rules are an important and necessary counter-balancing force. But the
principles of tax smoothing and gains from credibility can be enhanced by
a cautious approach to policy.
7LPHKRUL]RQ
101. The time horizon in the application of any fiscal rule is also relevant
since it will determine the speed of adjustment to shocks. For example, if
balance had to be achieved in each and every year, the Government would
need to change its spending plans or tax rates on a frequent basis. The
administrative costs of such a system - both to the Government and
taxpayers - makes it highly impractical. More fundamentally, frequent tax
rate changes could be destabilising, raise uncertainty and run counter to the
principle of tax smoothing.
102. The UK approach moves in the direction of tax smoothing in a
number of ways. First, allowance is made for transitory shocks related to
the economic cycle. The time horizon is thus likely to be several years,
though this is not precisely known. Second, capital and current budgets are
treated separately. There is no strong reason why current taxpayers should
finance capital investment in full so this potentially ’lumpy’ constraint is
effectively smoothed out, subject to a debt ratio criterion.
103. In principle, the time horizon might be stretched even more, leaving
only a fiscal sustainability constraint in place (in addition to the constraint
deriving from an independent monetary authority). However, it is not clear
that fiscal discipline is sufficiently entrenched or credible to make this
feasible.
)RUHFDVWHUURUV
104. While a Budget is an annual event, and taxes and spending can in
principle be adjusted each year, frequent and unpredictable adjustments to
taxes and public expenditure are far from ideal. The effects of fiscal
measures take time to work through, and are sometimes set to change only
some years ahead. Moreover, it is often not clear what impact a Budget has
had by the time of the subsequent Budget.
105. Fine tuning of fiscal policy is rarely feasible or sensible. It is
therefore appropriate to try to plan as smooth a path for taxes and spending
as possible. Given this aim – and consistent with the principle of stability –
a medium-term approach is necessary.
106. The emphasis on the medium term puts a great deal of weight on
forecasts. The fiscal projections are a crucial component in deciding which
Budget options for tax or spending appear feasible. But forecasts are
fallible, so care needs to be taken in assessing the risks involved.
107. These risks may be thought of as deriving from two sources:
mis-specification of general economic conditions, notably GDP, and more
specific fiscal risks, such as sudden shortfalls in taxes or overspends
unrelated to cyclical factors. In the UK, fiscal projections have suffered
from both sources of error. Charts 7 and 8 show successive projections of
GDP and borrowing against outcomes illustrate some of the problems.
108. It is instructive to look back at the size and sources of forecast errors
in the context of forecasts on the basis of which fiscal judgements were
made. For this purpose, we have looked at the period from 1986 to 1996,
which comprised a full economic cycle. It is difficult to make an
assessment earlier than this; and we have only limited information on
forecast errors since the new macroeconomic framework was put in place.
109. Table 2 sets out average one year ahead forecast errors for public
sector net borrowing (PSNB) for the period 1986-87 to 1996-97. It
distinguishes between actual errors and absolute errors, ie errors where the
signs are ignored. The errors refer to outturn less forecast; and the one year
ahead period relates to the immediate year, eg the one year ahead forecast
error for the March 1986 Budget refers to the outturn and forecast for the
fiscal year 1986-87. The table also shows the errors which occurred from
factors other than from GDP forecast errors, for convenience described as
‘fiscal’ errors. An unexpected shortfall in underlying income tax receipts
would be an example of such an error.
110. Over this period, the average error in forecasts of borrowing
amounted to 0.1 per cent of GDP. One year ahead forecasts of borrowing
were thus slightly optimistic on average. However, this hides considerable
variation: the average absolute error was more than 1 per cent of GDP, a
large error just one year ahead.
&KDUW
6XFFHVVLYHLOOXVWUDWLYHSURMHFWLRQVRIUHDO*'3
&KDUW
%RUURZLQJ361%0HGLXPWHUPSURMHFWLRQVDQGRXWWXUQV
98
100
102
104
106
108
110
112
1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96
1988-89 = 100
-4
-2
0
2
4
6
8
10
1986-87 1988-89 1990-91 1992-93 1994-95 1996-97 1998-99
7DEOH
&RPSDULVRQRIRQH\HDUDKHDGIRUHFDVWHUURUVERUURZLQJ361%
WR
7RWDOHUURU
*'3
µ)LVFDO¶HUURU
*'3
Actual 0.1 0.7
Absolute 1.1 1.1
111. Within the total, errors in forecasting GDP were an important
influence. But even if GDP had been forecast correctly, significant fiscal
errors remained, also amounting to around 1 per cent of GDP in absolute
terms. Although this is not shown in this table the errors grow significantly
as the time period is extended to two-year, three-year ahead errors, etc.
These errors are strongly influenced by a growing GDP error component.
For example, the two-year ahead absolute error in PSNB forecasts is
estimated to be 2 per cent of GDP, of which the fiscal error is 1.4 per cent
of GDP; three-years ahead the absolute average error is 3 per cent of GDP,
with the fiscal component some 2 per cent.
,PSOLFDWLRQVIRUWKHFRQGXFWRIILVFDOSROLF\
112. These results carry important implications for fiscal policy. First,
they indicate that even if GDP is forecast correctly, or if the economy
remains on trend, considerable errors can still arise from other sources.
Second, the considerable uncertainties implicit in forecasting mean in
practice that there is leeway for the policy-maker to fudge results if the
process is not wholly transparent.
113. Above all, the scale of the errors emphasises the need for caution in
policy-making. It is not unusual for a surge in spending, a large bonus or a
sudden shortfall in receipts to appear “out of the blue”, sometimes almost
before the ink has dried on a Budget. Political and other pressures can
dictate that a surplus or windfall should be given away or spent on
priorities. (Consider for example recent the clamour to spend UTMS
licence windfall receipts.) Given the size, frequency and direction of the
errors it is quite probable that a windfall, if fully spent for example, will
have to be clawed back before long.
%XLOGLQJDPDUJLQ
114. Given the potentially large errors in any forecast of the public
finances, aiming for exact balance, say, one year ahead would result in a
significant chance that the target was not met and might require drastic or
inappropriate remedial action to meet the objective. This could be costly in
economic terms and might compromise the ability of fiscal policy to
support monetary policy. For these reasons, it makes sense to build in a
margin to achieve a good probability of meeting the fiscal rules.
115. Suppose there is no uncertainty with respect to GDP on the basis that
the monetary authority is presumed to be successful in keeping the
economy on its trend path. Fiscal errors are likely to remain, however.
Using the mean and the variance of the fiscal errors over the past, Chart 9
shows the margin that would be required to meet a cyclically-adjusted
balance rule one year ahead. A surplus of 0.7 per cent of GDP is needed
simply to offset the previous bias to have a 50 per cent chance of being on
target. A surplus of well over 1 per cent of GDP would be required to
improve the probability of meeting the objective to, say, 75 per cent. The
chart illustrates the trade-off that exists between the need for certainty and
the margin required, which rises at a disproportionate rate to the degree of
certainty achieved.
7KUHHVRXUFHVRIHUURUV
116. There are a number of uncertainties that can be identified as key
sources error of error. We look at three here:
• where the economy is in relation to trend at the start of the forecast
(‘output gap’ error);
• specific fiscal and economic assumptions, such as VAT ratios,
unemployment, equity prices and turnover (which drive capital taxes),
interest rates (which impact on debt interest projections), oil prices,
and so on;
• how trend output is likely to evolve (trend growth assumption).
&KDUW
0HHWLQJWKH5XOHV$GGLWLRQDO0DUJLQ5HTXLUHG
0.0
0.5
1.0
1.5
2.0
2.5
40% 50% 60% 70% 80% 90%
µ2XWSXWJDS¶
117. One of the key uncertainties for the policy maker is knowing the
position of the economy in the economic cycle at any given point in time.
This might be called ‘output gap’ error, ie knowing how far the economy is
above potential (or otherwise). Important information, such as GDP, is
only available with a lag and can be subject to substantial revisions. What
appears plausible at Budget time can often look different several months or
years later. Nonetheless, a view must be taken in framing the Budget.
118. Taking account of the experience of the late 1980s, when
assumptions about potential and the output gap proved to be seriously
wrong, a more cautious approach has been followed since 1997. Because
of the inherent uncertainties involved the projections are ‘stress tested’ by
considering a more cautious case than the main projection. In this case, the
question is posed whether the fiscal rules would still be met in the event
that the level of trend output was 1 percentage point lower than assumed.
119. By way of illustrating the implications, consider the following
example. Assume the economy starts on trend and in fiscal balance and the
aim is to reach the next Budget a year later in the same state, ie again in
structural balance. Policy is set accordingly. Suppose further that it is
realised later in the year that the economy had in fact been running more
strongly than earlier supposed, say at 1 per cent above potential. With the
economy above trend the fiscal position would in fact be in structural
deficit of around 0.7 per cent of GDP, based on ready reckoners. So the
previous policy measures would have turned out to have been inappropriate
and remedial action, with associated costs, would have to be taken to get
back to the target.
120. In order to counter this risk and avoid potential disruptions to policy
from this source of error, it makes sense to aim for a surplus of at least 0.7
per cent of GDP in this example.
)LVFDOHUURU
121. As noted earlier, optimism in the past allowed policy to be looser
than warranted. Ex-ante projections turned out to be significantly worse
ex-post. Identifying the exact causes of this bias is not easy. However, by
adopting a cautious set of assumptions, ex-ante projections may turn out to
be acceptable ex-post, as indeed seems to have been the case recently in the
UK.
122. An example of adopting a more cautious approach can be seen from
the behaviour of the underlying VAT ratio which was expected to rise in
the early 1990s as the economy recovered from recession. However, this
did not occur and in fact the ratio fell. Had a cautious assumption been
adopted, as now with a gently falling ratio assumed, it is estimated that the
average error would have been reduced by 0.3 per cent of GDP.
123. There are other areas where a prudent approach has been adopted.
For instance instead of scoring all expected returns from Spend to Save
programmes (programmes designed to reduce fraud and raise tax) only
direct effects are taken into account. Unemployment is no longer assumed
to fall; and market interest rates - rather than internal, undisclosed
assumptions – are used explicitly to drive the debt interest forecasts. Had
these and other similarly cautious assumptions been used much of the
previous optimism bias would have been removed.
124. Assumptions in these areas are now independently audited by the
National Audit Office (NAO) to ensure that they remain both reasonable
and cautious. The full set of assumptions audited is shown in the box 2.
7UHQGJURZWKDVVXPSWLRQV
125. A particularly important assumption concerns trend growth, which
has a strong influence on the medium-term profile of taxes and, to a lesser
extent, spending. Under the current forecasting arrangements trend growth
for the public finance projections is taken to be 2¼ per cent a year in
contrast to the ‘neutral’ view of 2½ per cent used in the economic forecast.
126. The effect of additional annual growth of a quarter percentage point
on the public finances builds up over time. For example, after 3-4 years it
improves the budget balance by an amount approaching half a per cent of
GDP, equivalent to more than £4 billion extra spending a year.
127. It is also worth noting that a cautious trend growth assumption helps
to constrain policy choices that extend into the medium term. This is
especially important where a multi-year public spending regime is in place,
as is now the case in the UK. It also applies to tax where announcements
are often made in advance but may be implemented later and extend into
the future. By constraining the fiscal aggregates to a more moderate path
over the medium term, policy is less likely to be over-stretched. The short-
term outcome for the fiscal balances, and for policy overall, may be better
as a result.
,PSURYLQJWKHFKDQFHVRIPHHWLQJWKHUXOHV
128. The adoption of cautious assumptions raises the probability of
meeting the fiscal rules. Using the assessment of errors presented earlier, it
is possible to deduce the chances of meeting the rules under this approach.
To simplify, we assume the Bank meets the inflation target and the
economy remains close to trend. It is also assumed that there is no longer
any optimistic bias in the fiscal component of the projections but that the
absolute errors are similar to those previously. This appears broadly
appropriate on the basis of the limited evidence since 1997. Again, for
exposition we start from a position of balance, and the objective is to
achieve structural balance in a year’s time.
%R[
.H\DVVXPSWLRQVDXGLWHGE\WKH1$2
• Privatisation proceeds
1,6
Credit is taken only for proceeds from sales that have been
announced.
• Trend GDP growth
1,6
2
1
/
4
per cent a year.
• UK claimant unemployment
1,4,7, 8
Rising slowly to 1.06 million.
• Interest rates
1,6,7
3 month market rates change in line with market expectations.
• Equity prices
2,7
FT-All share index rises in line with money GDP.
• VAT
2,7
Ratio of VAT to consumption falls by 0.05 percentage points a
year.
• GDP deflator and RPI
2,7
Projections of price indices used to plan public expenditure are
consistent with RPIX.
• Composition of GDP
3
Shares of labour income and profits in national income are
broadly constant in the medium term.
• Funding
3
Funding assumptions used to project debt interest are consistent
with the public finances forecast and with financing policy.
• Oil prices
5, 8
$24.40 a barrel in 2001, the average of independent forecasts,
and then constant in real terms.
• Anti-tobacco smuggling measures
6
Only direct effects, including deterrent effects of fiscal marks,
are allowed for.
___________________________________________
1
NAO (1997a)
5
NAO (1999b)
2
NAO (1997b)
6
NAO (2000a)
3
NAO (1998a)
7
NAO (2000b)
4
NAO (1999a)
8
NAO (2001)
129. Policy is set to achieve this objective. The mean expectation for the
fiscal position is balance, ie there is a 50 per cent chance of achieving this
particular rule in one year ahead. This is unlikely to be sufficient to ensure
policy credibility since, if the pattern of past errors were repeated, there is a
good chance of a deficit of 1 per cent of GDP within one year.
130. In using the cautious case and cautious assumptions a structural
surplus equivalent to around ¾ per cent of GDP is being aimed for. This
improves the chances of achieving balance or surplus from 50 per cent to
over 70 per cent.
131. Over the medium term fiscal errors grow larger, reflecting greater
uncertainty. However, a cautious trend growth assumption builds up a
further margin over time. In principle, one might set the expected
additional margin from lower than central trend growth to match the
additional risk arising from future projections. On the other hand, since
some policy adjustments can be made in future Budgets this may not be
necessary.
&RQFOXVLRQV
132. On the basis of UK evidence, it appears that the probability of
meeting the golden rule (starting from a position of balance) over the
medium term, other things equal, remains better than evens and is perhaps
in the region of 70 per cent as a result of the adoption of cautious
assumptions. Like all such figures, they are only illustrative and depend on
several conditions that may not hold in practice. But they make the point
that in trying to meet a set of fiscal rules, some assessment needs to be
made of how likely it is they will be met and whether additional action is
required to improve the chances of doing so.
133. Good rules should produce good outcomes. But rules should not be
pursued for their own sake or simply because they are there. It is the spirit
of the law that should count not the letter of the law. However, fiscal rules
must deliver for most of the time and especially when first set.
134. Ideally a fiscal framework and associated rules should operate
through constrained discretion. Constrained first by the transparency of the
overall macroeconomic - monetary and fiscal - framework and second by
the particular fiscal rules. The operating rules should provide room for
some discretion:
• to allow scope for sensible adjustment to severe shocks;
• by leaving room for cyclical and other transitory shocks and by
allowing full scope for the automatic stabilisers to exert their
smoothing effect on the economy; and
• to ensure that sensible investments with appropriate rates of return are
not unduly held back.
135. Two fundamental objectives for fiscal policy seem to hold. First, to
ensure that policy remains sustainable over the longer term. Second, to
support monetary policy in the short-term to achieve a sensible policy mix.
136. The specific fiscal rules set need to take account of circumstances. In
the UK redressing the considerable underinvestment in public sector
infrastructure, and improving generational balance, have motivated the
focus on the golden rule. Setting the rules over the economic cycle has
general applicability where it is desired that automatic stabilisers should
have a free rein.
137. In practice, meeting a set of fiscal rules is important for credibility.
Over time, once credibility of fiscal policy has been regained and has
become entrenched it may be possible to relax fiscal rules further. In the
monetary policy context, for example, the Bundesbank achieved sufficient
credibility that whether or not its monetary targets were met was not of
vital importance, since it was more or less universally believed that the
central goal of low inflation would be achieved.
138. The path for fiscal policy is more difficult, not least because
Governments will always face re-election. Transparent frameworks and a
credible set of fiscal rules, however, currently offer the best route forward
within a political democracy.
139. The credibility of the monetary authority depends on output gap
smoothing and hitting an inflationary target. The credibility of the fiscal
authority depends on achieving a sustainable fiscal path and meeting its
fiscal rules. But it also depends critically on not disrupting the monetary
authority’s task, ie ensuring that there is good co-ordination of policy. The
independence and transparency of the monetary authority serves to ensure
a more disciplined and efficient fiscal policy.
140. Two trade-offs are of particular importance in calibrating a fiscal
system. First, between transparency and the rigidity of fiscal rules: a
greater degree of transparency may imply less need for rigid rules. Second,
between the rigidity of the fiscal rules and the extent of tax and/or spending
smoothing. The more rigid the rules the less scope there will be for
smoothing tax or spending in the face of shocks.
141. Uncertainty over the state of the economy and specific fiscal risks
interfere with the conduct of policy in practice. In particular, they can
reduce the transparency of policy and create risks over meeting fiscal rules.
Uncertainty thus can undermine the credibility of policy. Moreover, since
reversals of policy are costly – both in political terms and through
economic inefficiency – governments should favour approaches that reduce
these risks.
142. Countering uncertainty requires a fiscal margin, a form of
precautionary saving. The larger the margin, the higher the probability of
meeting the rules, thus strengthening the reputation of the fiscal policy-
maker and fiscal credibility. But a larger margin also involves potential
costs in terms of lost output (or reduced welfare) via higher taxes or lower
spending paths than would otherwise be the case.
143. In the UK the Government has introduced a transparent
macroeconomic framework, backed by legislation, and supported by a
clear set of fiscal rules. It has supplemented this by a comprehensive
outcome-based multi-year regime for public spending, which is now being
run on resource accounting and budgeting lines and more in keeping with
best-practice in the private sector.
144. Because of risks and future uncertainties the fiscal projections are
based on a cautious approach, with key assumptions audited by an
independent authority (the NAO), and are ‘stress tested’ to try to avoid
repetition of past mistakes. The effect is to improve the chances of meeting
the fiscal rules and thereby provide a smoother path for taxes and spending.
This has helped to improve policy credibility, for instance (along with
monetary policy) helping to reduce the risk premium on debt and has
strengthened efficiency in spending on public services.
145. Through a combination of a transparent framework, sensible fiscal
rules, cautious assumptions and a well-informed and equally transparent
independent monetary authority it is possible to arrive at desirable
outcomes for macroeconomic policy: economic stability based on low
inflation, sound public finances and an appropriate policy mix.
5()(5(1&(6
Aiyagari, S.R. and E.R. MacGrattan (1998), “The Optimum Quantity of
Debt”, -RXUQDORI0RQHWDU\(FRQRPLFV, Vol. 42, pp. 447-69.
Alesina, A. and R. Perotti (1995), “The Political Economy of Budget
Deficits”, 6WDII 3DSHUV, International Monetary Fund, Vol. 42,
March, pp. 1-31.
Alesina, A. and G. Tabellini (1990), “Positive Theory of Fiscal Deficits
and Government Debt”, 5HYLHZRI(FRQRPLF6WXGLHV, Vol. 57, July,
pp. 403–14.
Alesina, A. and R. Perotti (1999), “Budget Deficits and Budget
Institutions”, in J. Poterba and J. von Hagen (eds.), )LVFDO
,QVWLWXWLRQVDQG)LVFDO3HUIRUPDQFH, Chicago University Press.
Artis, M. and M. Buti (2000), “Close to Balance or in Surplus: A Policy-
Maker’s Guide to the Implementation of the Stability and Growth
Pact”, Centre for Economic Policy Research Working Paper,
No. 2515, London, CEPR.
Asilis, C.M. (1994), “The U.S. Public Debt: Implications for Growth”,
International Monetary Fund Working Paper, WP/94/4.
Barro, R.J. (1974), “Are Government Bonds Net Wealth?”, -RXUQDO RI
3ROLWLFDO(FRQRP\, 75, pp. 1095-117.
__________ (1979), “On the Determination of the Public Debt”, -RXUQDORI
3ROLWLFDO(FRQRP\, October.
__________ (1984), “The Behavior of United States Deficits”, NBER
Working Paper, No. 1309.
__________ (1986), “U.S. Deficits since World War I”, 6FDQGLQDYLDQ
-RXUQDORI(FRQRPLFV, No. 1, 1986.
__________ (1990), “Government Spending in a simple model of
endogenous growth”, -RXUQDO RI 3ROLWLFDO (FRQRP\ 98 (1),
pp. 103-17.
__________ (1995), “Optimal Debt Management”, NBER Working Paper,
No. 5327.
Blanchard, O., R. Dornbusch and W. Buiter (1986), “Public Debt and
Fiscal Responsibility”, in O. Blanchard, R. Dornbusch and R.
Layard (eds.), 5HVWRULQJ (XURSH¶V 3URVSHULW\, Macroeconomic
papers from the Centre for European Policy Studies, London, MIT
Press, pp. 125-53.
Bleaney, M., N. Gemmell and R. Kneller (2001), “Testing the endogenous
growth model: public expenditure, taxation, and growth over the
long run”, &DQDGLDQ-RXUQDORI(FRQRPLFV, Vol. 34, No. 1.
Bowden, R. J. (1995), “Government Debt and the Value of Government”
1HZ=HDODQG(FRQRPLF3DSHUV, 29(2).
Buiter, W.H. (1985), “A Guide to Public Sector Debt and Deficits”,
(FRQRPLF3ROLF\, November.
___________ (2001), “Notes on ‘A Code for Fiscal Stability’”, Oxford
Economic Papers, 53, pp. 1-19.
Cardarelli, R., S. James and L.J. Kotlikoff (2000), “Generational
Accounting in the United Kingdom´(FRQRPLF-RXUQDO, Vol. 110,
No. 467, November, pp. 547–74.
Chalk, N. (2000), “The sustainability of bond-financed deficits: an
overlapping generations approach”, -RXUQDORI0RQHWDU\(FRQRPLFV,
Vol. 45, No. 2.
Chalk, N. and R. Hemming (2000), “Assessing Fiscal sustainability in
Theory and Practice”, IMF Working Paper, WP/00/81.
Chouraqui, J.-C., B. Jones and R.B. Montador (1986), “Public Debt in a
Medium-term Perspective”, OECD Economic Studies, Vol. 0, No. 7,
pp. 103-53.
Feldstein, M. (1984a), “Debt and Taxes in the Theory of Public Finance”,
NBER Working Paper, No. 1433.
Hamilton, J. and F. Marjorie (1986), “On the Limitations of Government
Borrowing: A Framework for Empirical Testing”, $PHULFDQ
(FRQRPLF5HYLHZ, Vol. 76, No. 4.
HM Treasury (1997a), “Financial Statement and Budget Report”, March.
___________ (1997b), “Pre-Budget Report”, November.
___________ (1997c), “Fiscal Policy: Lessons from the last economic
cycle”, November.
___________ (1998a), “Financial Statement and Budget Report”, March.
___________ (1998b), “Economic and Fiscal Strategy Report”, June.
___________ (1998c), “Fiscal Policy: Current and capital spending”,
June.
___________ (1998d), “Comprehensive Spending Review”, July.
___________ (1998e), “Pre-Budget Report”, November.
___________ (1998f), “The Code for Fiscal Stability”, November.
___________ (1998g), “Delivering Economic Stability: Lessons from
Macro-economic Policy Experience”, November.
___________ (1999a), “Economic and Fiscal Strategy Report and
Financial Statement and Budget Report”, March.
___________ (1999b), “Fiscal Policy: Public Finances and the cycle”,
March.
___________ (1999c), “Resource Accounting and Budgeting”, April.
___________ (1999d), “The New Monetary Policy Framework”, October.
___________ (1999e), “Pre-Budget Report”, November.
___________ (1999f), “Analysing U.K. Fiscal Policy”, November.
___________ (1999g), “Trend Growth: prospects and implications for
policy”, November.
___________ (2000a), “Economic and Fiscal Strategy Report and
Financial Statement and Budget Report”, March.
___________ (2000b), “Spending Review 2000”, July.
___________ (2000c), “Pre-Budget Report”, November.
___________ (2000d), “UK Convergence Programme”, December.
___________ (2001a), “Economic and Fiscal Strategy Report and
Financial Statement and Budget Report”, March.
___________ (2001b), “Managing Resources”, April.
Holder, A. (1997), “Tax Smoothing: An Introduction and Reflections on
Policy”, HM Treasury Academic Panel Paper, No. 97(3).
Keech, W.R. (1985), “A Theoretical Analysis of the Case for a Balanced
Budget Amendment”, 3ROLF\6FLHQFHV, 18(2), pp. 157-68.
Kitterer, W. (1994), “Tax Versus Debt-Financing of Public Investment: A
Dynamic Simulation Analysis”, .UHGLWDQG.DSLWDO, Vol. 27, No. 2,
pp. 163-87.
Kopits, G. and J. Craig (1998), 7UDQVSDUHQF\LQ*RYHUQPHQW2SHUDWLRQV,
IMF Occasional Paper, No. 158.
Kopits, G. and S. Symansky (1998), )LVFDO3ROLF\5XOHV, IMF Occasional
Paper, No. 162, Washington, International Monetary Fund.
Kneller, R. (2000), “The Implications of the Comprehensive Spending
Review for the Long-Run Growth Rate: A View from the
Literature”, 1DWLRQDO ,QVWLWXWH (FRQRPLF 5HYLHZ, No. 171,
pp. 94-105.
MacKenzie, G.A. (1989), “Are All Summary Indicators of the Stance of
Fiscal Policy Misleading?”, IMF Staff Papers, Vol. 36, No. 4,
pp. 743-70.
Milesi-Ferretti, G.M. (2000), “Good, Bad or Ugly? On The Effects of
Fiscal Rules with Creative Accounting”, IMF Working Paper,
WP 00/172.
Milesi-Ferretti, G.M. and N. Roubini (1998), “On the taxation of Human
and Physical Capital in Models of Endogenous Growth”, -RXUQDORI
3XEOLF(FRQRPLFV, Vol. 70, No. 2., pp. 237-54.
Musgrave, R.A. (1939), “The Nature of Budgetary Balance and the Case
for a Capital Budget,” $PHULFDQ (FRQRPLF 5HYLHZ Vol 29,
pp. 260–71.
_____________ (1988), “Public Debt and Intergenerational Equity”, in
K.J. Arrow and M.J. Boskin (eds.), 7KH(FRQRPLFVRI3XEOLF'HEW
Proceedings of a conference held by the International Economic
Association at Stanford, New York, St. Martin’s Press, pp. 13-146.
Musgrave, R.A. and P.B. Musgrave (1989), 3XEOLF)LQDQFHLQ7KHRU\DQG
3UDFWLFH New York, McGraw Hill, Inc., 5
th
ed.
NAO (National Audit Office) (1997a), “Audit of Assumptions for the
July 1997 Budget Projections”, HC3693.
________________________ (1997b), “Audit of Assumptions for the
Pre-Budget Report”, November, HC361.
________________________ (1998a), “Audit of Assumptions for the
Budget”, March, HC616.
________________________ (1999a), “Audit of the Unemployment
Assumption for the March 1999 Budget Projections”, March,
HC294.
________________________ (1999b), “Audit of the Oil Price
Assumption for the Pre-Budget Report”, November, HC873.
________________________ (2000a), “Audit of Assumptions for the
March 2000 Budget”, March, HC348.
________________________ (2000b), “Audit of Assumptions for the Pre-
Budget 2000 Report”, HC959.
________________________ (2001), “Audit of Assumptions for the
March 2000 Budget”, March, HC.
Office of National Statistics (ONS) (1998), “UK National Accounts
Concepts, Sources and Methods.”
Phelps, E.S. (1965), “Second Essay on the Golden Rule of Accumulation”,
in R. Becker and E. Burmeister (eds.) (1991), *URZWK 7KHRU\
9RO 'HVFULSWLYH *URZWK 7KHRULHV ,QWHUQDWLRQDO /LEUDU\ RI
&ULWLFDO:ULWLQJVLQ(FRQRPLFV, No. 10, pp. 33-54, Elgar.
Robson, W.B.P. and W.M. Scarth (1997), “Out Front on Federal Debt
Reduction: Programs and Payoffs”, CD Howe Institute.
Robinson, M. (1998), “Measuring Compliance with the Golden Rule”,
)LVFDO6WXGLHV Vol. 19, November, pp. 447–62.
Rogoff, K. and A. Sibert (1988), “Elections and Macroeconomic Policy
Cycles”, 5HYLHZRI(FRQRPLF6WXGLHV, Vol. 55, January, pp. 1–16.
Smyth, D.J. and Y. Hsing (1995), In search of an optimal debt ratio for
Economic Growth, &RQWHPSRUDU\(FRQRPLF3ROLF\ Vol. 13, No. 4,
October.
Spaventa, L. (1987), “The Growth of Public Debt: Sustainability, Fiscal
Rules, and Monetary Rules”, ,QWHUQDWLRQDO 0RQHWDU\ )XQG 6WDII
3DSHUV, 34(2), pp. 374-99.
Tanzi, V. and D. Fanizza (1995), “Fiscal Deficit and Public Finance in
Industrial Countries: 1970-1994”, IMF Working Paper, No. 95/49.
Zee, H. (1988), The Sustainability and Optimality of Government Debt,
,QWHUQDWLRQDO0RQHWDU\)XQG6WDII3DSHUV, 35(4), pp. 658-85.
7+((92/87,212)7+()('(5$/%8'*(7$1'),6&$/
58/(6
5LFKDUG:3HDFK
,QWURGXFWLRQ
Following roughly two decades of large federal deficits and rising
federal debt as a share of GDP, the U.S. has enjoyed a dramatic improvement
in its fiscal balance over the past several years. From a deficit of $290 billion
or 4.7% of GDP in Fiscal 1992, the federal balance moved into surplus in
FY1998 and that surplus rose to $237 billion or 2.4 percent of GDP in Fiscal
2000
1
. More recently, the improvement from deficits to significant surpluses
has led to both faster growth of spending and a significant tax cut.
Nonetheless, under current tax and spending policies and reasonable economic
assumptions, the federal budget is projected to remain in surplus over the next
decade. Of course, it is widely recognized that under current policies, by
around 2015 the federal balance is likely to revert back to deficit as the baby-
boom generation begins to retire and draw on federal retirement and health
benefits.
Beginning in the mid 1980s, the federal government enacted a series of
fiscal rules intended to reduce the deficits and put federal fiscal policy on a
sustainable path. The first generation of fiscal rules established specific
numerical targets for the deficit that proved to be politically impossible to meet
as economic growth slowed in 1989 and 1990. Latter attempts were a blend of
numerical targets on some categories of spending and budget process rules that
were largely adhered to through most of the 1990s.
The fact that these rules were in place over the period when the federal
balance moved from deficit to surplus naturally leads some to conclude it was
__________
*
Federal Reserve Bank of New York.
I would like to thank Donald Rissmiller and A.J. Glusman for their diligent research assistance. All
errors and omissions are my own responsibility. The opinions expressed in this paper do not necessarily
reflect the views of the Federal Reserve Bank of New York nor the Federal Reserve System.
1
In the U.S., the federal fiscal year runs from October through September.
the rules per se that led to the improvement in the fiscal balance. This paper
will argue that the rules certainly played a useful role. First, the rules were
reasonably transparent; much more so than the traditional budget process. The
rules significantly changed the debate on budget issues, keeping attention
focused on the big picture of eliminating the deficits and so on the need to
make tradeoffs. Gradually, as the rules were adhered to for a longer and longer
period, the government’s credibility was enhanced in financial markets,
helping to reduce interest rates and the deficit.
But while helpful, rules are just rules. Ultimately it was the political will
to first enact and then adhere to the rules that is the main explanation for the
improvement in the federal balance. This political will stems from the fact that,
after a recognition delay, voters had come to recognize that the stance of fiscal
policy from the early 1980s to the early 1990s was unsustainable. What drove
that home most powerfully was the rising federal debt and the periodic need to
raise the statutory ceiling on that debt, a law that dates back to 1917. Indeed,
the need to raise the debt ceiling was such an unambiguous symbol of failure
on the part of the government that it often provided the catalyst for enactment
of meaningful legislation to slow the growth of outlays and/or raise the growth
rate of revenues.
That the fiscal rules of the 1980s and 1990s were “political symbolism”
of sorts is made all the clearer by the fact that while those rules are still the law
of the land they have been routinely ignored in recent years. The movement to
surpluses has once again changed the fiscal debate from whether the deficit
should be reduced by raising taxes or cutting spending to how much of the
surplus should be returned to taxpayers in the form of tax cuts, how much
should be used for additional spending, and how much should be used to pay
down the existing stock of federal debt.
The outline of this paper is as follows. Section 2 provides a broad
overview of trends in federal revenues, outlays, and debt over the post World
War II period. Section 3 provides a chronology and brief description of the
fiscal rules that have been enacted. Section 4 concludes.
(YROXWLRQRIWKH)HGHUDO%XGJHW
Over the 1950s and 1960s, despite the cold war with the former Soviet
Union and armed conflict in Korea and Southeast Asia, the federal government
ran relatively modest deficits. Over those two decades, the federal balance was
in deficit by an average of $3.7 billion (0.6 percent of GDP) per year. While in
nominal terms the stock of federal debt continued to grow, it fell steadily
expressed as a percent of GDP (figure 1).
However, conditions began to change by the mid 1970s as the growth of
outlays increased due primarily to the expansion of federal entitlement
programs or transfers (figure 2). The Social Security program was made more
generous in terms of benefits and coverage over the 1950s, 1960s, and early
1970s. Medicare Part A and Part B as well as Medicaid were enacted in 1965.
In contrast, defense spending was declining rapidly as a share of GDP.
Over the second half of the 1970s revenues also rose as a share of GDP,
limiting the size of the deficits such that debt held by the public remained
fairly stable as a share of GDP. The rapid growth of revenues was due in large
part to the rapid inflation of that period, which led to rapid growth of nominal
incomes which in turn pushed individuals into higher marginal tax rate
brackets (their were considerably more brackets and higher marginal rates at
that time). The enactment of the Economic Recovery Tax Act of 1981
significantly reduced the level of receipts as a share of GDP while also limiting
their future growth rate by lowering individual income tax rates and by
introducing inflation indexing into the individual income tax code. Even
though outlay growth began to slow in the 1980s, despite a relatively moderate
rise in defense spending, the size of the gap between outlays and revenues was
quite large. It narrowed a bit in the late 1980s in response to robust economic
growth, but then widened again in the early 1990s as the economy experienced
a relatively mild and brief recession followed by several years of sluggish
economic growth. Over the period from 1983 to 1992 the deficit averaged 4.3
percent of GDP. These large and sustained deficits caused the stock of debt
held by the public to rise from around 25 percent of GDP in the mid 1970s to
around 50 percent of GDP in the early to mid 1990s. Expressed as a share of
total household sector financial assets, the stock of federal debt held by the
public also roughly doubled, from around 10 percent in the mid 1970s to
around 20 percent by the early 1990s (figure 3).
)LJXUH
)HGHUDO5HFHLSWVDQG2XWOD\VDQG'HEW+HOGE\WKH3XEOLF
SHUFHQWRI*'3
Source: Department of the Treasury.
Fiscal Year
14
16
18
20
22
24
14
16
18
20
22
24
20
30
40
50
60
70
80
90
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
20
30
40
50
60
70
80
90
D eb t Held b y the P ub lic
Debt held by the public
Outlays
Receipts
)LJXUH
)HGHUDO2XWOD\VE\W\SH
SHUFHQWRI*'3
Source: Budget of the United States Government.
Since 1992, the U.S. has experienced a dramatic improvement in its
federal fiscal situation. From a deficit of $290 billion or 4.7% of GDP in Fiscal
1992, the federal budget moved into surplus in Fiscal 1998 and that surplus
rose to $237 billion or 2.4 percent of GDP in Fiscal 2000. The transition from
deficits to significant surpluses led to faster growth of spending in FY2000 and
FY2001 and a significant tax cut enacted in mid 2001. Nonetheless, under
current tax and spending policies and reasonable economic assumptions, the
federal budget is projected to remain in surplus over the next decade
(cfr. figure 4 and table 1). However, it is widely recognized that, due to the
aging of the population, the current policy fiscal projections over the next 50
years are quite gloomy.
Fiscal Year
0
5
10
15
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
0
5
10
15
Nondefense
National Defense
Entitlements / Transfers
Net Interest
)LJXUH
)HGHUDO'HEW+HOGE\WKH3XEOLF
SHUFHQWRI+RXVHKROG)LQDQFLDO$VVHWV
Source: Council of Economic Advisors, Department of the Treasury, and Board of Governors of the Federal
Reserve System.
The source of the 7 percent of GDP swing in the federal fiscal balance
over the period from Fiscal 1992 to Fiscal 2000 was a 3.3 percentage point of
GDP rise of total revenues and a 3.8 percentage point decline of total outlays.
Most of the improvement in receipts was in individual income taxes (up 2.7
percentage points), while corporate income tax receipts also rose as a share of
GDP (up 0.5 percentage points) due to a rise of corporate profits as a share of
national income (cfr. table 2). The sharp rise of individual income taxes as a
share of GDP has been the largest surprise and the largest source of error in the
underestimation of federal receipts in recent years. There are three main
reasons for this surprising growth: (1) growth of taxable personal income in
excess of growth of GDP, (2) growth of adjusted gross income in excess of
growth of taxable personal income, and (3) increases in effective tax rates on
adjusted gross income (CBO, 2000).
Fiscal Year
0
5
10
15
20
25
30
35
1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
0
5
10
15
20
25
30
35
)LJXUH
Over the period from Fiscal 1994 to Fiscal 2000, taxable personal
income, which consists of wages and salaries, personal interest income,
personal dividend income, rental income of persons, and proprietors’ income,
rose from 68.2 percent of GDP to 71.3 percent of GDP. Similarly, corporate
profits before tax rose from under 7 percent of GDP in Fiscal 1991 to 9.5
percent of GDP in Fiscal 1997. A major reason that all of these income shares
rose is that the statistical discrepancy between the income side and product
side of the U.S. National Income and Product Accounts (NIPAs) widened
substantially over the past several years. The statistical discrepancy, an entry
on the income side of the ledger that forces GDP as measured by total income
to equal GDP as measured by total expenditures, went from +0.3 percent of
GDP in 1991 to –1.3 percent of GDP in 2000. As a result, national income rose
at a compound annual rate of 5.9% over that nine-year period while GDP rose
at a 5.7% compound annual rate.
CBO “Current Policy” Baseline
)HGHUDO8QLILHG%DODQFH
ELOOLRQVRI86'ROODUV
-300
-100
100
300
500
700
900
-300
-100
100
300
500
700
900
July 2000
Feb 2001
July 1999
August 1998
10 Year Cumulative Surplus
(US $ Trillions)
August 1998: 1.6
July 1999: 2.9
July 2000: 5.8
February 2001:
5.6
Fiscal Year
CBO “Current Policy” Baseline
7DEOH
&RPSDULVRQRI&%2&XUUHQW3ROLF\0DFURHFRQRPLF$VVXPSWLRQV)LVFDO<HDUV
ACTUAL AVERAGE
Date FY1999 FY2000 FY2001 FY2002 FY2003 FY2004-2009/20
Real GDP 8/98 2,3 1,9 1,7 2,2 2,5 2,3
7/99 4,1 2,8 2,3 2,3 2,3 2,5
7/2000 4,1 5,1 3,4 2,8 2,8 2,7
2/2001 4,2 4,6 2,7 3,2 3,4 3,0
CPI-U 8/98 2,4 2,8 2,7 2,5 2,5 2,5
7/99 1,9 2,5 2,5 2,5 2,5 2,5
7/2000 1,9 3,0 2,7 2,7 2,7 2,5
2/2001 1,9 3,2 2,9 2,8 2,8 2,5
3-month T-Bill 8/98 5,3 4,9 4,6 4,5 4,4 4,4
7/99 4,5 5,0 4,8 4,5 4,5 4,5
7/2000 4,4 5,6 6,6 5,6 5,1 4,8
2/2001 4,6 5,7
5,2
4,8 5,1 4,9
FORECAST PROJECTIONS
7DEOH
)HGHUDO5HFHLSWVE\7\SH
Percent of GDP
Percent of Total FY 1992 FY 2000 Change
Total 100.0% 17.5% 20.8% 3.3
Individual Income 49.6% 7.7% 10.4% 2.7
Individual Withheld 34.0% 5.6% 7.1% 1.5
Individual Nonwithheld 15.6% 2.0% 3.3% 1.3
Capital Gains 5.8% 0.4% 1.2% 0.8
Social Insurance 32.2% 6.6% 6.7% 0.1
Corporate 10.2% 1.6% 2.1% 0.5
Other* 8.0% 1.6% 1.7% 0.1
* Includes excise taxes (3.4%), estate and gift taxes (1.4%), customs duties (1.0%) and miscellaneous (2.1%).
Note: Individual components may not sum to totals due to rounding.
Adjusted gross income (AGI), the income concept used as the basis for
income taxation, rose faster than taxable personal income due to sharply
increased capital gains realizations, larger withdrawals from tax-deferred
accounts such as IRAs and 401ks, and the fact that an increasing share of
Social Security benefits became subject to taxation. The final, and most
important, reason for the rapid rise of individual income taxes over this period
is the rise of average effective individual income tax rates (see figure 5). This
is due to increases in real income, which push taxpayers into higher marginal
rate brackets (real bracket creep), and shifts in the distribution of income
toward higher-income taxpayers, as shown in table 3.
)LJXUH
* Percent of wage and salary income.
** Percent of proprietors’ income, rental income of persons, personal divided income, and
personal interest income.
7UHQGVLQ(IIHFWLYH7D[5DWHV
SHUFHQW
Fiscal Year
7
9
11
13
15
17
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011
7
9
11
13
15
17
Individual
Withheld*
Baseline Projection
Feb 2001
Individual Non-withheld**
Jan 2000
7DEOH
3HUFHQWDJHRI5HWXUQV$GMXVWHG*URVV,QFRPH
DQG7D[/LDELOLWLHVIRU+LJK,QFRPH7D[SD\HUV7D[<HDUV
Source: Congressional Budget Office.
Note: AGI = Adjusted Gross Income.
* Data are based on tax returns processed through November, 1999.
Turning to outlays, Table 4 presents the major categories of outlays
expressed as a percent of GDP for Fiscal 1992 and Fiscal 2000. In total,
outlays fell from 22.2 % of GDP in Fiscal 1992 to 18.4% in Fiscal 2000, a
decline of 3.8 percentage points to the lowest level since the mid 1960s. About
half of the total decline was in defense outlays, which fell 1.9 percentage
points of GDP over the period. Entitlement outlays also fell as a percent of
GDP, due to relatively slow growth of new retirees over the period, concerted
efforts to slow the growth of health care outlays in both the public and private
sectors, and the combination of strong economic growth and policy changes
which slowed the growth of means-tested entitlement programs. Finally, with
both lower interest rates and a declining stock of debt, net interest payments
also fell as a share of GDP.
AGI (1998 dollars) 1993 1998*
Percentage of Returns
200,000 and above 1,1 1,6
500,000 and above 0,2 0,4
1 m illion and above 0,1 0,1
Percentage of A GI
200,000 and above 14,5 21,6
500,000 and above 7,7 13,1
1 m illion and above 4,9 9,3
P ercentage of T ax L iabilities
200,000 and above 29,8 39,8
500,000 and above 17,5 25,6
1 m illion and above 11,4 18,3
7DEOH
%XGJHW(QIRUFHPHQW$FW3UHVHQWDWLRQRI)HGHUDO2XWOD\V
* Family Support (AFDC), SSI, EITC, Student Loans, Farm Price Supports, Federal Civilian and Military
Retirement, etc.
Note: Individual components may not sum to totals due to rounding.
2YHUYLHZRI)LVFDO5XOHV
Federal law contains numerous provisions that can be labeled fiscal
rules. Some of these provisions are quite old, such as the debt ceiling, while
most were enacted in an attempt to reduce the large deficits of the 1980s and
the first half of the 1990s. This section provides an overview of the main fiscal
rules. Note that this review does not attempt to explain all of the legislative
detail of these laws, which is often quite complex.
% of
Total
FY1992 FY2000 Change
Outlays 100,0% 22,2% 18,4% -3,8
Discretionary 34,5% 8,6% 6,3% -2,3
Defense 16,5% 4,9% 3,0% -1,9
Nondefense 18,0% 3,7% 3,3% -0,4
Entitlements/ Transfers 57,5% 11,5% 10,6% -0,9
Social Security 22,7% 4,6% 4,2% -0,4
Medicare 12,1% 2,1% 2,2% 0,1
Medicaid 6,6% 1,1% 1,2% 0,1
Other * 16,1% 3,8% 3,0% -0,8
Net Interest 12,5% 3,2% 2,3% -0,9
Other
-4,5%
-1,1% -0,9% 0,2
% of GDP
$ 6WDWXWRU\'HEW&HLOLQJ
The Liberty Bond Act of 1917 established a statutory limit on the gross
indebtedness of the federal government. This was introduced as a simplifying
procedure during World War I, as prior to the passage of this law each
individual bond issue had to be approved by Congress. In the 1980s and 1990s
the need to raise the federal debt ceiling often provided the catalyst for
enactment of deficit reduction legislation.
% &RQJUHVVLRQDO%XGJHWDQG,PSRXQGPHQW&RQWURO$FWRI
The Budget Act, as it is called, established the current congressional
budget process and in so doing introduced some discipline to that process. The
Act established the House and Senate Budget Committees, the adoption of a
Congressional Budget Resolution, and the reconciliation process. It also
created the Congressional Budget Office, which gave the Congress an
analytical capability comparable to that of the Executive Branch.
& %DODQFHG%XGJHWDQG(PHUJHQF\'HILFLW&RQWURO$FWRI*UDPP
5XGPDQ+ROOLQJV,
This act established declining deficit targets for each year through Fiscal
1991, at which point the budget was to be balanced. It also established a
sequestration process under which previously approved spending was canceled
if the projected deficit exceeded the target. However, the actual deficit did not
need to meet the target, even if Congress changed policy over the course of the
year. The sequestration process was to be controlled by the General
Accounting Office, a branch of Congress.
' %DODQFHG %XGJHW DQG (PHUJHQF\ 'HILFLW &RQWURO 5HDIILUPDWLRQ $FW RI
*UDPP5XGPDQ+ROOLQJV,,
The Supreme Court had ruled that having the sequestration trigger in the
hands of the General Accounting Office was a violation of the Constitutional
principle of separation of powers. Under this legislation, the sequestration
trigger was placed in the hands of the Director of the Office of Management
and Budget (OMB), an office of the Executive Branch. In addition, the deficit
targets were revised and extended to FY1993, at which time the budget was to
be balanced.
( %XGJHW(QIRUFHPHQW$FWRI
As part of a major deficit reduction package that included increases in
taxes and reductions in spending, the Budget Enforcement Act (BEA)
significantly altered the budget enforcement mechanisms. First, it effectively
eliminated deficit targets. In their place it established nominal ceilings or
“caps” on discretionary budget authority and outlays over the period from
Fiscal 1991 to Fiscal 1995. Discretionary outlays, which currently represent
about one-third of total outlays, are those which are controlled by the annual
appropriations process. The caps are adjustable to accommodate
“emergencies” as well as other extenuating circumstances such as funding for
the International Monetary Fund. When initially enacted in 1990, there were
three separate caps for defense, international, and nondefense domestic
spending for the period from Fiscal 1991 through Fiscal 1993. Then from
Fiscal 1994 to Fiscal 1995 there was a single cap covering the total
discretionary category.
In addition, the BEA established the “pay-as-you-go” or “PAYGO”
process for revenues and direct spending. Direct spending is for the most part
entitlements or transfer programs that are established through authorizing
legislation. Outlays are determined not by annual appropriations but rather by
the eligibility criteria and benefit formulas of the specific programs in
combination with underlying economic and demographic conditions. Under
PAYGO, all changes in the tax code and in direct spending enacted in a
session of Congress must be “deficit neutral” over one year and five year
horizons. The Senate subsequently enacted a 10-year horizon test as well
2
.
PAYGO does not require congressional action if revenues fall or outlays grow
__________
2
The PAYGO provision made the scoring or pricing of legislation a critical exercise and introduced a
new form of lobbying. Interest groups would higher consultants to influence the way in which their
favored provisions were scored so as to minimize their cost.
due to changing economic or technical assumptions such as demographic
conditions and endogenous changes in effective tax rates. PAYGO is enforced
one year at a time, meaning that an overage in one year can not be offset by an
underage in another year. Moreover, an underage or overage in the PAYGO
section of the budget can not be transferred to the discretionary section of the
budget and vice versa.
Both the discretionary spending caps and PAYGO are enforced through
the threat of sequestration, an executive order permanently canceling some
previously approved spending. As mentioned above, the 1985 act (Gramm-
Rudman-Hollings I) established the sequestration process. The BEA of 1990
established new, more elaborate sequestration rules and procedures.
) )HGHUDO&UHGLW5HIRUP$FWRI
This Act changed the budget treatment of direct loans and loan
guarantees by the federal government. Prior to this legislation, direct loans
were treated as outlays when funds were dispersed. The outlays were netted
against repayments of principal and interest. Loan guarantees were recorded as
outlays only when there was a default that prompted the federal government to
make good on its guarantee. Under the Credit Reform Act, the estimated
discounted present value of the subsidy inherent in a direct loan or guarantee is
scored as an outlay in the year in which the loan or guarantee is made.
* 2PQLEXV%XGJHW5HFRQFLOLDWLRQ$FWRI
In addition to raising tax rates on upper-income taxpayers and reducing
spending, the Act extended the discretionary spending caps and PAYGO
through FY1998.
+ /LQH,WHP9HWR$FWRI
This law granted the President authority to cancel selected categories of
spending and tax provisions over the period from FY1997 to FY2004. It was
ultimately struck down by the Supreme Court.
, %XGJHW(QIRUFHPHQW$FWRI
This legislation extended the discretionary spending caps and PAYGO
through Fiscal 2002. Separate caps were again imposed on defense and
nondefense discretionary spending for the period from Fiscal 1997 to Fiscal
1999. In addition, a separate cap was introduced for a new category of outlays
called Violent Crime Reduction for the years Fiscal 1997 to Fiscal 2000. After
that, all discretionary spending was again covered under a single cap. As part
of the Transportation Equity Act for the 21
st
Century, a separate cap for
highway and mass transit spending was enacted. In reality, these special caps
were not ceilings but floors on categories of spending with particularly strong
constituencies.
)LVFDO5XOHVDQG&KDQJHVLQWKH3ULPDU\6WUXFWXUDO'HILFLW
While the fiscal rules of the 1980s and 1990s became more elaborate,
they were easily circumvented or simply overridden when political pressure to
restrain spending eased. For example, in the development of the budgets for
Fiscal 1999 and Fiscal 2000, the discretionary spending caps were effectively
overridden by declaring large increases in outlays to be “emergency” spending.
For Fiscal 2001, the Congress and the President simply enacted a very large
increase in the discretionary spending caps to accommodate the level of
spending they desired. Similarly, the tax cut enacted in 2001 should prompt a
sequester under the PAYGO rules, but it is widely expected that those rules
will be ignored.
One objective way of evaluating the effect of the enactment of fiscal
rules on the stance of fiscal policy is to look at the change in the primary
structural balance in the years immediately after enactment. The structural
deficit or balance is what the federal balance would be if the economy was
operating at potential. Accordingly, this measure of the balance is not affected
by short-run variations in growth of GDP. The primary structural balance is the
structural balance excluding interest payments on the stock of outstanding
debt. By excluding the interest cost of debt accumulated in the past we are able
to focus on the policy decisions of the present government. Changes in the
primary structural balance indicate whether fiscal policy is tightening or
easing.
)LJXUH
&KDQJHLQ3ULPDU\6WUXFWXUDO%DODQFH
DVDSHUFHQWRI3RWHQWLDO*'3
Source: Congressional Budget Office.
Figure 6 presents changes in the primary structural balance as a percent
of potential GDP, both estimated by the Congressional Budget Office, for
fiscal years from 1980 to 2000. Included in the chart are the dates of passage or
extension of the major fiscal rules of the past two decades. One of the first
things to note is that enactment of fiscal rules was not always associated with a
shift toward contractionary fiscal policy. For example, enactment of the
Budget Enforcement Act of 1990 was followed by expansionary fiscal policy
until enactment of the deficit reduction package of the Omnibus Budget
Reconciliation Act of 1993. The main reasons are two fold. First, while
Fiscal Year
-2
-1
0
1
1980 1985 1990 1995 2000
-2
-1
0
1
2
3
4
GRH-1
85
GRH-2
87
BEA
90
OBRA
93
BEA
97
BEA90 introduced the discretionary spending caps for the period from Fiscal
1990 to Fiscal 1995, the cap for Fiscal 1991 permitted a 6.6% increase in
discretionary spending for that year, well above the 3.8% compound annual
rate of growth over the preceding five years when Gramm-Rudman-Hollings I
and II were in effect. Second, the late 1980s and early 1990s witnessed an
explosion in the rate of growth of the federal health care programs (Medicare
and Medicaid) due to expansion of eligibility and rapid increases in health care
usage and prices. OBRA93 proved to be more successful in shifting the stance
of fiscal policy to contractionary by reining in the growth of the these health
care programs, in addition to some significant tax increases.
)LJXUH
)HGHUDO'HEW&HLOLQJVDQG*URVV)HGHUDO'HEW
ELOOLRQVRI86'ROODUV
Source: Council of Economic Advisors.
Fiscal Year
GRH-1
85
GRH-2
87
BEA
90
OBRA
93
BEA 97
0
1000
2000
3000
4000
5000
6000
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
0
1000
2000
3000
4000
5000
6000
Figure 7 presents the total amount of federal debt subject to limit and the
statutory debt ceiling for the period from fiscal 1980 to fiscal 2000. Note that
the dates of enactment of fiscal rules and deficit reduction packages always
correspond to periods when it was necessary to raise the debt ceiling. Note also
that over time, when the debt ceiling needed to be increased, it tended to be
increased by larger amounts. This demonstrates the power of this relatively
simple rule that quite by accident became instrumental in the improvement of
the federal fiscal balance.
&RQFOXVLRQ
Fiscal rules can be effective in restraining fiscal policy as long as they
are adhered to. They are more likely to be adhered to if compliance can be
easily observed, such as whether the stock of debt is rising. But most important
of all, a majority of voters must be convinced that adherence to the rules is in
their interest.
5()(5(1&(6
Congress of the United States, Congressional Budget Office (2000), 7KH
%XGJHWDQG(FRQRPLF2XWORRN)LVFDO<HDUV, January.
Keith, R. and A. Schick (1998), “Manual on the Federal Budget Process”,
Congressional Research Service (98-720 GOV).
Moon, M. and J. Mulvey (1995), (QWLWOHPHQWV DQG WKH (OGHUO\, Washington
DC, The Urban Institute Press.
THE ROLE OF FISCAL RULES
IN DETERMINING FISCAL PERFORMANCE
Suzanne Kennedy, Janine Robbins and François Delorme
*
1. Introduction
1
The topic of fiscal rules has attracted significant attention over the
last decade, as several countries have adopted fiscal rules in an attempt to
eliminate large deficits. More recently, fiscal rules have been the subject of
renewed interest as countries consider how to adapt fiscal policy for times
of surplus and how to ensure long-term sustainability of fiscal policy,
particularly in light of pressures related to population ageing.
The purpose of this paper is to examine the importance of fiscal
rules in determining fiscal performance. By necessity, the focus is on the
role for rules in fiscal consolidation; most countries under consideration
have not yet achieved surpluses or have only done so recently. As such, it
is too early to evaluate the effectiveness of rules in maintaining surplus
positions. The paper begins with a brief summary of the rationale for fiscal
rules and concerns related to their implementation. Section 3 compares
fiscal rules in Canada with similar practices in the United States, the
European Economic and Monetary Union, Germany, Japan, New Zealand
and Sweden. Fiscal rules at the subnational level in the United States and
Canada are also reviewed. General observations about the role of fiscal
rules are drawn from a comparison of the evolution of total government
structural balances in the above-mentioned countries and in countries
without legislated fiscal rules during the fiscal consolidation of the mid-
1990s. Section 4 reviews a selection of recent empirical studies addressing
fiscal rules and section 5 concludes.
Before proceeding, it is necessary to clarify this paper’s definition of
“fiscal rule”, as there are many possible definitions. A broad definition
__________
*
Ministry of Finance / Ministère des Finances – Canada.
The views expressed in this paper are those of the authors and do not reflect those of the
Department of Finance. The authors would like to thank Richard Hemming, Ernesto Stein and the
other participants at the Committee on Hemispheric Financial Issues’ Seminar on Debt and Fiscal
Management for their helpful comments and discussion. The authors are also grateful to Peter
DeVries and Gaétan Pilon for their suggestions. Any and all errors are the responsibility of the
authors.
1
This paper builds on Brunnen and Pilon (1996).
238 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
could encompass the set of rules and regulations according to which
budgets are drafted, approved and implemented (as used to define
“budgetary institutions” in Alesina and Perotti (1999)). For the purpose of
this paper, however, a much narrower definition is chosen: a “fiscal rule” is
defined as a statutory or constitutional restriction on fiscal policy that sets a
specific limit on a fiscal indicator such as the budgetary balance, debt,
spending, or taxation. In other words, the focus is restricted to rules that
impose a specific, binding constraint on the government’s range of policy
options. Policy rules or guidelines that are not legislated are not considered
to be fiscal rules in this analysis, because although they may influence the
decisions of the government, they do not impose binding constraints on
present or future governments.
This paper focuses on the experiences of industrialized countries
with fiscal rules. However, it should be noted that other countries,
particularly Latin American countries, have had interesting experiences
with fiscal rules as well. Appendix A provides a brief summary of fiscal
rules currently in place in three major Latin American countries.
2. Why Legislated Fiscal Rules?
In theory, discretionary policy can achieve the same outcomes as
fiscal rules, and should in fact be superior because it allows greater
flexibility. However, many have suggested that this is not the case in
practice. The literature identifies a number of potential problems that fiscal
rules may be used to address
2
. For example, numerous political economy
studies describe how electoral pressures may lead politicians to adopt a
short time horizon, resulting in socially suboptimal policy choices (see, for
example, Cukierman and Meltzer (1986) or Nordhaus (1975)).
Tufte (1978) and Rogoff (1990) demonstrate how government spending or
taxes may be influenced by a political budget cycle. Alesina and Tabellini
(1990) show that when successive governments have different policy
preferences, public debt may be used strategically to influence the choices
of successors, in which case the equilibrium level of debt will be higher
than is socially optimal. Weingast, Shepsle and Johnsen (1981)
demonstrate how political institutions may systematically bias public
decisions toward larger than efficient projects as legislators fail to
__________
2
For a more extensive analysis of the political economy of budget deficits, see Alesina and Perotti
(1995).
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 239
internalize the full cost of programs that benefit their geographical
constituencies but that are funded at the national level. In light of these
various types of potential distortions, fiscal rules may be viewed as the best
available replacement for a benevolent social planner.
In practice, fiscal rules have been adopted for a wide variety of
reasons, for example: (a) to ensure macroeconomic stability, as in post-war
Japan; (b) to enhance the credibility of the government’s fiscal policy and
aid in deficit elimination, as in some Canadian provinces; (c) to ensure
long-term sustainability of fiscal policy, especially in light of population
ageing, as in New Zealand; or (d) to minimize negative externalities within
a federation or international arrangement, as in the European Economic
and Monetary Union. Underlying most fiscal rules is a sense that present or
future governments, for any number of reasons, may not be willing or able
to implement optimal fiscal policy measures without external pressure.
However, this line of argument may be refuted by the fact that several
countries, such as Canada, have implemented successful fiscal adjustments
in recent years without legislated fiscal rules, as discussed in section 3.
The enactment of fiscal rules raises a number of issues concerning
flexibility, credibility, and transparency. One of the main concerns about
fiscal rules is that they may be overly restrictive and limit a government’s
ability to engage in legitimate countercyclical fiscal policy when required.
As such, legislation must be written in such a way that it provides some
flexibility, in order to be functional, yet not be so flexible that it becomes a
non-binding constraint. This concern is addressed in greater detail in
section 4.
In addition, rules should be transparent. As such, they should not be
overly complicated, and should be easy to monitor and defined in terms of
fiscal indicators that cannot be easily manipulated. As for credibility, the
rule should be viewed as permanent. This leads to the question of whether
the fiscal rule should be implemented by statutory or constitutional law, the
latter being far more difficult to change or revoke. Due to the costs of
changing a constitutional rule of law, it is likely that a constitutional fiscal
rule will be less explicit in its policy specification. At the same time, it is
likely that it will stand the test of time. By comparison, a statutory fiscal
rule has the advantage of increased clarity, yet is more likely to be altered
over time. Thus, there is a trade-off between longevity and clarity. An
additional concern with respect to constitutional rules is that they may
transfer the interpretation of an economic target or rule from policymakers
240 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
to constitutional court judges. A final issue related to credibility is that of
enforcement – there must be some mechanism to enforce the rule.
However, fiscal discipline is not guaranteed even in the presence of
the most effective fiscal rules; political commitment is also necessary if
rules are not to be circumvented. For example, Milesi-Ferretti (2000)
demonstrates that when a government has a margin for “creative
accounting”, the imposition of a fiscal rule may entail a trade-off between
costly “window-dressing” and real fiscal adjustment.
3. An Assessment of Fiscal Rules in Practice
This section provides an overview of fiscal rules in various
countries, under the Maastricht Treaty, and at the subnational level in the
U.S. and Canada, as summarized in Table 1. The most common type of
fiscal rule is a restriction on the budgetary balance. These often take the
form of balanced budget requirements, as in many of the U.S. states and
Canadian provinces. Inasmuch as these rules often apply only to the current
budget, they are equivalent to the “golden rule”, which specifies that
deficits may only be run in order to fund investment. Restrictions on the
budgetary balance may also be expressed in terms of specific target levels,
as under the Maastricht Treaty. Another common type of fiscal rule
consists of debt targets or restrictions, as implemented in the Maastricht
Treaty, most American states and some Canadian provinces. Alternatively,
there may be tax or expenditure limits, such as in the U.S., Sweden and
several American states. In addition, jurisdictions may require a general
referendum to be called to approve major taxation initiatives, as in some
American states and certain Canadian provinces. Among the jurisdictions
under consideration, the most comprehensive series of enacted fiscal rules
were found at the subnational level, in several Canadian provinces and
American states.
By their nature, legislated fiscal rules are intended to be permanent.
As such, they should be designed to apply over the economic cycle.
Indeed, most of the fiscal rules that are currently in force may be useful for
maintaining surpluses as well as for eliminating deficits. One possible
exception is the system of expenditure limits in the U.S., the value of
which has been questioned since the emergence of large and growing
surpluses.
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 241
It should be noted that both the United Kingdom and Australia
enacted legislation in 1998 setting out broad fiscal policy guidelines to
increase transparency and accountability in the conduct of fiscal policy.
The legislation in both countries established a new set of reporting
requirements and principles to guide the conduct of fiscal policy. However,
neither framework includes legislated numerical targets. Rather, the
emphasis is placed on requiring government to clearly set out its fiscal
strategy and targets. Because these frameworks do not impose binding
constraints on present and future governments in terms of numerical rules,
they are not considered legislated fiscal rules for the purpose of this paper.
Nevertheless, both countries provide useful examples of non-legislated
rules and targets, so their fiscal frameworks are discussed in the following
overview.
Table 1
Fiscal Rules at a Glance
Budgetary
Balance
Controls
Debt
Restrictions
Tax or
Expenditure
Controls
Referendum
for New
Taxes
US: Gramm-Rudman-Hollings
Legislation, 1985
X
US: Budget Enforcement Act
1990 onward
X
Maastricht Treaty
X X
Germany
X
Japan, 1997-98
X X X
New Zealand
X
Sweden
X
Federal Canada, 1991/92-
95/96
X
American States
48 40 27 3
Canadian Provinces
83 2 4
3.1 International Overview of Fiscal Rules
3.1.1 United States
At the federal government level in the United States, deficit controls
were introduced in 1985 through the Gramm-Rudman-Hollings (GRH)
legislation (the Balanced Budget and Emergency Deficit Control Act). It
242 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
imposed an annual deficit reduction schedule for a five-year period, with a
balanced budget set for 1991, which was later revised to 1993. The Act
covered on-budget items (i.e., excluding Social Security trust funds) and
deficit objectives were to be accomplished mainly through spending cuts.
Without an agreement on cuts to achieve the targeted deficit, automatic
across-the-board cuts in spending for most discretionary and some
mandatory programs had to take place.
Ultimately, both the 1991 and 1993 targets were missed by
significant amounts. One of the main reasons for this outcome was that the
targets were applied to the projected, rather than the actual deficits. In
using overly optimistic economic and fiscal forecasts, budgetary
projections could easily meet the targets, while the actual deficit exceeded
the limit every year.
The legislation was replaced by the Budget Enforcement Act of
1990 (BEA), which shifted the focus away from deficit targets toward
expenditure and revenue controls. Similar to the GRH legislation, the BEA
applies only to the on-budget accounts. It sets annual dollar limits on
discretionary spending, which are adjusted annually for revisions to
technical assumptions, emergency appropriations, and certain other
reasons. Pay-as-you-go rules apply to any new legislation affecting
mandatory spending and revenue, meaning that new legislation may not
impose net costs. A sequestration procedure is triggered if aggregate
discretionary appropriations enacted for a fiscal year exceed that year’s
spending caps or if a fiscal year’s aggregate mandatory spending and
receipts legislation is considered to entail a net cost. As an improvement
over the GRH legislation, the BEA applies only to parts of the budget that
are under the direct control of lawmakers, not to fluctuations rooted in the
economy nor changes in the cost of existing entitlement programs. The
1990 BEA applied to fiscal years 1990 to 1995, although it was extended to
fiscal year 1998 by the Omnibus Budget Reconciliation Act of 1993. The
Balanced Budget Act of 1997 extended the provisions through fiscal year
2002.
While the original BEA limits have sometimes been surpassed, the
legislation has been credited with improving fiscal discipline (Schick
(2000)) and limiting the costs associated with new legislation (OECD
(1999)). The spending caps were designed with a view to balancing the
budget by 2002, but this goal was attained much earlier, in 1998, mainly as
a result of strong economic growth. In light of the rising surpluses, the
spending restrictions have increasingly been viewed as unnecessarily tight.
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 243
The administration’s FY2001 budget plan proposed that spending caps be
raised and extended to 2010, such that government operations could be
maintained at currently enacted levels, with discretionary spending rising
in line with inflation.
3.1.2 The Fiscal Criteria of the Maastricht Treaty
The 1992 Maastricht Treaty set out convergence criteria that must be
satisfied in order for countries to participate in the European Economic and
Monetary Union (EMU). The main goal of the agreement in terms of fiscal
policy is to ensure fiscal discipline in member countries in order to prevent
fiscal crises that would negatively affect other countries. Under the Treaty,
fiscal discipline is to be judged on the basis of two main criteria:
(1) whether the government deficit as a percentage of GDP exceeds the
reference value of 3 per cent of GDP; and (2) whether the ratio of gross
government debt to GDP exceeds the reference value of 60 per cent of
GDP. Exceptions may be made with respect to the deficit criterion if the
ratio of the deficit to GDP has declined significantly and is close to the
reference value, or if the excess is only temporary and the ratio remains
close to the reference value. Exceptions may be made with respect to the
debt criterion if the debt-to-GDP ratio is diminishing at an acceptable pace.
The Maastricht Treaty provisions were strengthened by the Stability
and Growth Pact (SGP), which ensures that countries sustain their
commitment to fiscal prudence once they have joined the EMU. The SGP
was adopted in 1997 and took effect when the euro was launched on
January 1, 1999. In addition to the Treaty’s debt and deficit rules, the SGP
requires that member states set medium-term objectives of budgetary
positions close to balance or in surplus, in order to provide sufficient
flexibility to allow the operation of automatic fiscal stabilizers while
remaining within the 3 per cent deficit limit. This last point is considered to
be especially important since member countries can no longer rely on the
exchange rate instrument to dampen economic shocks.
The SGP also provides for increased monitoring, with an annual
review of the stability programmes of countries participating in the euro
area (and convergence programmes of those not participating in the euro
area). The programmes set out medium-term targets and the adjustment
path toward the targets. The Council of Ministers assesses and delivers an
opinion on each programme, based on the recommendation of the
European Commission. In addition, the Council and Commission regularly
244 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
monitor the implementation of the programmes and can recommend
corrective action if a significant divergence from the medium-term
budgetary objective or the adjustment path is identified.
In the case of an excessive deficit in a country participating in the
euro area, a course of remedial action will be proposed, which must be
implemented within ten months. Otherwise, the country may be subject to
sanctions in the form of a mandatory non-interest bearing deposit, which
varies in size with the magnitude of the excessive deficit, up to a maximum
of 0.5 per cent of GDP. If the excessive deficit is eliminated within two
years, the deposit will be returned to the country. If it is not eliminated
within that time frame, the deposit will become a fine.
By 1998, eleven of the fifteen EU member states had met the
convergence criteria and agreed to participate in EMU. As of the European
Commission’s Autumn 2000 forecast, all eleven EMU participants were
expected to comply with the deficit criterion in 2000. Seven countries were
expected to have gross debt levels at or below the 60 per cent reference
level, while the others had decreasing debt ratios and were thus considered
to be in compliance with the criteria.
The Maastricht Treaty fiscal criteria are generally credited with
having accelerated fiscal consolidation in the EU countries. For example,
France faced a fiscal crisis in the early 1990s, with the total government
deficit peaking at 6 per cent of GDP in 1993. However, the fiscal situation
improved significantly over the following years, largely due to
discretionary policy undertaken by the government in order to comply with
the Maastricht Treaty. By 1998, the deficit was under 3 per cent, consistent
with the convergence criteria.
3.1.3 Germany
Germany has a history of fiscal rules dating back to 1969. In that
year, a constitutional rule was introduced which requires a balanced
budget, but allows borrowing for investment expenditure (i.e., the golden
rule). This rule applies to the federal government and the entirety of its
budget – including consolidated federal enterprises and special funds. In
addition, some states’ constitutions include the golden rule.
The constitution specifies exceptions from a balanced budget during
times of macroeconomic disequilibrium or war, and an important German
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 245
policy mandate is that restrictive fiscal policy should not destabilize the
economy or restrict growth and prosperity. On several occasions, the
Constitutional Court has ruled that the need to stabilize the economy
warranted borrowing in excess of investment. Overall, the constitutional
fiscal rule poses only a minor constraint to the government and has not
prevented deficits.
More importantly, Germany is subject to the Maastricht Treaty and
the Stability and Growth Pact. As discussed earlier, these have imposed
effective constraints and fiscal discipline on European governments. Since
the Maastricht Treaty applies at the general government level, the German
federal government has proposed to determine legally binding allocations
of the Maastricht deficit limit between the federal government and the
states, as well as across states.
3.1.4 Japan
Japan has had a legislated fiscal rule since 1947, which prescribes
that bond issuance be limited to raising funds for financing public works.
The rule covers only the general account budget of the central government,
which represents only about 25 per cent of the central government’s total
budget. However, since 1975, deficit-financing bonds have been issued on
a regular basis in addition to construction bonds, which are exclusively for
public works. Moreover, the distinction between construction and deficit-
financing bonds became less clear as constructions bonds were issued to
cover more and more spending categories. As such, the fiscal rule has not
proven to be a binding constraint since 1975.
In order to address the deficit which had persisted through the early
to mid-1990s, especially in light of future ageing-related pressures, the
government engaged in fiscal tightening and passed the Fiscal Structural
Reform Law in 1997. The legislation provided that the sum of the central
and local government deficits as a percentage of GDP should be reduced to
3 per cent or less by FY2003 (from around 6 per cent in FY1997).
Furthermore, it provided that the amount of deficit-financing bonds should
be reduced every fiscal year and issuance of such bonds should cease by
FY2003. The legislation also required that numerical limits be set for
expenditures in each major programme from FY1998 to FY2000. Finally,
it specified that the sum of taxes, payroll contributions and the deficit
should not exceed 50 per cent of GDP.
246 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
However, the fiscal tightening initiated in 1997 was too much for the
economy to bear. Under pressure from the Asian economic crisis and the
failure of some major Japanese financial institutions, the economy fell into
recession. In response, the government revised the Fiscal Structural
Reform Law in May 1998 to introduce more flexibility, and then formally
suspended its application in November 1998. Since that time, the
government has followed expansionary policies and the general
government gross debt-to-GDP ratio has skyrocketed, reaching 105.3 per
cent in 1999 (OECD (2000)).
3.1.5 New Zealand
In 1994, the Fiscal Responsibility Act (FRA) was enacted in New
Zealand to improve the conduct of fiscal policy by setting out principles of
responsible fiscal management and by promoting accountability and a
long-term focus in fiscal planning. In contrast to the Maastricht Treaty, the
FRA places more emphasis on transparency than on numerical targets. It
requires that the government run annual operating surpluses in order to
achieve unspecified “prudent” levels of Crown debt. Once these levels
have been achieved, they must be maintained by, on average, avoiding an
operating deficit. It also provides that sufficient levels of Crown net worth
(total Crown assets less total Crown liabilities) be maintained in case of
future adverse conditions. Temporary departures from these principles are
allowed as long as the government specifies the reasons for the departure
and sets out how and when it will return to the principles. The FRA does
not include any sanctions, but sets out detailed reporting requirements,
including a requirement to publish the government’s long-term fiscal
policy objectives.
Although the Act does not specify numerical debt targets, the
government has defined its targets for fiscally prudent levels of debt. The
present goal is to reduce gross debt to below 30 per cent of GDP and net
debt to below 20 per cent of GDP. The government has generally been
successful in meeting its goals. The target of running operating surpluses
has been achieved since the FRA came into force. Moreover, net debt fell
from around 50 per cent of GDP in the early 1990s to below 30 per cent in
1996-97, and it is expected to fall below 20 per cent in 2001-02. However,
it is difficult to assess the contribution of the fiscal rules to the
improvement in New Zealand’s fiscal situation; the success was likely due
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 247
to a combination of the fiscal rules, improved reporting requirements,
better economic conditions and political commitment.
3.1.6 Sweden
Sweden’s Fiscal Budget Act of 1996 requires Parliament to set
nominal expenditure limits for 27 expenditure areas of the central
government, including transfers to other levels of government, for a three-
year period. Each year, Parliament sets new limits for the third year, and
the ceilings are set so as to ensure that outlays fall as a proportion of GDP.
The measures were strengthened in 1999 through a prohibition on using
allocations transferred from previous years. Although the spending caps are
not accompanied by sanctions, to date they are considered to have been
effective controls. Overall, Sweden has achieved significant improvements
since it undertook its fiscal consolidation programme in 1994-95.
3.1.7 Canada
At the federal level in Canada, the Federal Spending Control Act set
limits on program spending from 1991-92 to 1995-96. It covered all
program spending, with the exception of that under major self-financing
programs, such as expenditures under the Unemployment Insurance Act.
The Act permitted overspending in one year if offset in the following
two years. If spending were under the limit for a fiscal year, the difference
could be allocated to a subsequent fiscal year. In addition, upward
adjustments in annual spending were allowed if associated with an
equivalent increase in revenues. Finally, the Auditor General was required
to express an opinion as to compliance with the Act.
Spending levels were lower than the set limits in every year except
1992-93 (part of underspending in 1991-92 was allocated to 1992-93 to
cover the excess spending). Moreover, actual spending from 1991-92 to
1995-96 was $23.4 billion under the aggregate spending control limits.
Overall, the legislation did not prove to be necessary for controlling
spending and therefore was not extended beyond 1995-96.
More importantly, the government introduced a number of non-
legislated policy rules that contributed significantly to the dramatic
improvement in Canada’s federal finances in the 1990s. In 1994, the
government adopted the practice of basing budget planning on economic
248 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
assumptions near the low end of the range of private sector forecasts, in
order to minimize the risk of taking inappropriate policy actions as a result
of overly-optimistic economic assumptions. In addition, the government
began setting two-year rolling deficit targets, with an ultimate goal of a
balanced budget. In 1995, the government introduced a Contingency
Reserve in its budget planning, to protect against adverse changes in the
economy or forecasting errors. If not needed, the reserves were applied to
deficit reduction. Through prudent economic planning assumptions and an
emphasis on credible, short-term fiscal targets, with a firm commitment to
meet these targets, the federal government was able to move from a deficit
to a surplus position.
In 1998, the federal government committed to follow a non-
legislated Debt Repayment Plan, under which a $3-billion Contingency
Reserve is set aside each year and is devoted to debt reduction if not
needed. In the 2000 budget plan, the government announced that the extra
economic prudence, which was previously included in revenue and
expenditure projections, will now be explicitly shown, in order to facilitate
evaluation of the credibility of the fiscal projections. When the extra
prudence is not required, it will become part of future planning surpluses.
In addition, the government recently stated that in the future, it will
announce each fall whether a greater amount than the $3-billion
Contingency Reserve should be devoted to that year’s debt paydown,
depending on the economic and fiscal circumstances at the time. As a
result of the Debt Repayment Plan and the growing economy, the ratio of
net public debt to GDP has been reduced from a peak of 71.2 per cent in
1995-96 to 58.9 per cent in 1999-2000.
3.2 Overview of Non-Legislated Numerical Rules
3.2.1 Australia
The Charter of Budget Honesty, passed in 1998, introduced a fiscal
framework in Australia similar to that in New Zealand. The Charter
requires governments to set out their medium-term fiscal strategy in each
budget as well as their short-term fiscal objectives and targets, although it
does not place any constraints on the nature of the targets.
The government’s original debt target was to reduce the
Commonwealth general government net debt-to-GDP ratio to half of its
1995-96 level by the turn of the century. This target has been comfortably
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 249
met, with net debt falling from a peak of almost 20 per cent of GDP in
1995-96 to around 7 per cent in 2000-01. About two-thirds of the reduction
reflects privatization proceeds, with the remaining third coming from
budget surpluses.
The government’s current medium-term objective is to balance the
budget over the economic cycle. Consistent with this goal, the government
aims to continue running surpluses over the short term. As a supplementary
objective, the government also aims to improve its net worth (a measure
that includes physical as well as financial assets). Recent projections
indicate that positive net worth could be achieved by 2003-04. It would
seem that the strategy of requiring government to set out clear objectives,
without dictating what these objectives should be, has served Australia
well.
3.2.2 United Kingdom
The United Kingdom adopted similar legislation in 1998, which set
out principles to guide the conduct of fiscal policy (the key principles are
transparency, stability, responsibility, fairness and efficiency). In addition,
the legislation requires that the government table in Parliament a code for
fiscal stability setting out its fiscal strategy in accordance with these
principles.
The current government’s fiscal code is guided by two rules: The
“golden rule”, under which borrowing should be used only to finance
investment; and the “sustainable investment rule”, which states that public
sector net debt is to be held at a stable and prudent level, which the
government currently defines as below 40 per cent of GDP. Both rules are
designed to apply over the economic cycle.
To date, the government has been successful in meeting its goals.
Public sector net debt was brought down from 44 per cent of GDP in 1996-
97 to 36.8 per cent in 1999-00.
250 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
3.3 Overview of Subnational Fiscal Rules in Canada and the U.S.
3.3.1 American States
3
All but two American states have provisions requiring a balanced
budget. Most states are subject to the relatively loose requirement that the
governor submit a balanced budget. In addition, 40 states require the
legislature to pass a balanced budget, and 34 require the governor to sign a
balanced budget. Most states have constitutional requirements or a
combination of constitutional and statutory requirements; only five rely
solely on statutory requirements. A majority of states are subject to more
stringent provisions that prevent them from carrying deficits over into the
next fiscal year or the next two-year budget period. Typically, state fiscal
rules carry no sanctions and apply only to general funds, excluding
separate accounts such as the capital account and accounts for social
insurance and employee retirement.
In addition to balanced budget rules, 27 states have tax and
expenditure limitations, which set limits on annual revenue or expenditure
increases. Of those with expenditure limits, most limit appropriations to
some index of inflation, often state personal income growth or a certain
percentage of state personal income. Three states require voter approval to
increase revenues.
Furthermore, most states have some form of constitutional or
statutory limits on the issuance of general obligation debt (debt which is
guaranteed by all government funds and the government’s ability to raise
taxes). Most limits are based on a formula involving state revenues or
appropriations, while some states impose maximum dollar limits. Fourteen
states allow general obligation debt to be overridden by a referendum or
supermajority vote, and a few states prohibit the issuance of general
obligation debt altogether.
Despite the balanced budget provisions, states generally can run
small, temporary deficits. However, strong economic growth over recent
years has enabled most states to run surpluses and build up reserves (most
have “rainy day” funds) in case of economic slowdown.
__________
3
All statistics taken from National Asociation of State Budget Officers (1999).
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 251
3.3.2 Canadian Provinces and Territories
Nine provinces and territories have enacted or tabled fiscal rules.
Each fiscal rule requires balanced budgets, except in the Yukon, where
deficits are permitted as long as no net debt is accumulated. Fiscal rules
cover the consolidated budget in every jurisdiction, except Saskatchewan
and the Yukon, where it is the general revenue fund in the former case and
the non-consolidated Public Accounts in the latter case.
Most provinces require a balanced budget on an annual basis.
However, New Brunswick and Saskatchewan are required to balance their
budgets over a four-year period. In Quebec and Ontario, deficits may be
offset by previously accumulated surpluses. Similar to Quebec’s
legislation, British Columbia requires that deficits be gradually eliminated
before achieving balanced budgets. Deficits are permitted in Nova Scotia,
Quebec and Ontario as long as they are offset in the next fiscal year.
Manitoba’s legislation includes a Fiscal Stabilization Fund representing up
to 5 per cent of annual expenditure to offset unforeseen fluctuations in
revenue. Many jurisdictions also have provisions for exceptional events
such as a major disaster. In terms of using surpluses, Alberta’s legislation
specifies how excess revenues can be used and Nova Scotia limits
spending on new programs to existing budgets.
Several provinces have chosen to also target debt reduction and
elimination. Manitoba has a debt repayment schedule incorporated into the
law, whereby a minimum of $75 million ($96 million in 2000-01) is
deposited every year into a Debt Repayment Fund to be applied against the
general purpose debt and/or pension liability at least every 5 years. In
Saskatchewan, the legislation states that a four-year debt management plan
must be tabled, although there are no specific requirements except that
surpluses must go towards debt reduction. After eliminating its net debt
(excluding pension obligations), Alberta legislated a 25-year debt
repayment schedule, to eliminate the accumulated debt by 2025. Finally, in
the Yukon, deficits are permitted, but the legislation prohibits net debt
accumulation. Although Nova Scotia does not have a debt repayment plan,
legislation requires a reduction in the amount of foreign currency exposure.
Ontario, Manitoba, Alberta and the Yukon have
taxation-by-referendum approval rules. Ontario’s legislation is the most
comprehensive, applying to personal income, corporate, retail sales,
employer health, gasoline and fuel, provincial land and education taxes. In
Manitoba, a referendum must be called for any major tax increases,
252 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
whereas in Alberta, it applies to the implementation of a retail sales tax. In
the Yukon, a referendum is required for increases or implementation of
new taxes covered by the Income Tax Act and Fuel Oil Tax Act.
Ontario, Manitoba, British Columbia and the Yukon have legislated
penalties for not achieving the fiscal targets. Ontario’s legislation applies to
members of the Executive Council, whereby salaries are reduced by 25 per
cent in the first year of a deficit and 50 per cent for each year thereafter.
For Manitoba, ministerial salaries are cut by 20 per cent in the first year of
a deficit and by 40 per cent if there are two or more consecutive deficits.
British Columbia’s legislation reduces salaries of the Executive Council by
20 per cent if balance targets are not met. The Yukon has the strictest of all
legislation as an election is triggered if any net debt is accumulated.
From discussions with the provinces, one of the main advantages of
legislated fiscal restrictions is that they increase the Finance Ministers'
bargaining power to promote unpopular fiscal measures within the Cabinet.
Essentially, policymakers can quote the rules as an external constraint in
reference to internal allocations of limited funds.
3.4 International Comparison of Fiscal Outcomes
In order to give a general indication of the success in fiscal
consolidation in countries with and without fiscal rules, this section
examines the change from 1995 to 1999 in the structural balances of the
countries discussed earlier. Structural balances are used in order to control
for the effects of the business cycle. Data are presented for the total
government sector, although for the purpose of this comparison, countries
are classified on the basis of whether they have fiscal rules at the central
government level. In Canada’s case, it should be noted that many of the
provinces had fiscal rules in place over this period. However, the deficit in
the mid-1990s was much greater at the federal level; in 1995, the federal
deficit accounted for approximately 73 per cent of the total government
deficit. Therefore, most of the consolidation was achieved at the federal
level, where there were no fiscal rules. Moreover, roughly three quarters of
the total provincial deficit in 1995 was attributable to the province of
Ontario, which did not adopt fiscal rules until 1999.
The period of 1995 to 1999 was chosen because it represents a time
of fiscal retrenchment in most countries and because it allows for a logical
separation of countries according to whether they have legislated fiscal
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 253
rules (Canada’s limits on program spending ended in 1995, and Sweden
adopted expenditure limits in 1996). EU countries are classified according
to whether they joined the EMU when it came into being (this approach
assumes that the countries that joined the EMU considered the convergence
criteria as a binding constraint).
The evolution of structural balances relative to GDP from 1995 to
1999 indicates that most countries achieved some degree of fiscal
consolidation over this period (Table 2). Notable exceptions are Japan,
which was engaging in deficit spending in an effort to combat a major
economic downturn, and New Zealand, which had already attained a
financial surplus in 1995. The improvement in the structural balance was
particularly strong in Sweden, Italy, the U.K., Canada and Australia. In
other words, major improvements were made in countries both with and
without fiscal rules.
Table 2
Total Government Structural Balance
(percent of GDP)
1995 1996 1997 1998 1999
Change
from
1995 to
1999
Countries with legislated fiscal rules
United States
-3.0 -2.3 -1.1 0.0 0.7 3.7
Germany
-2.7 -2.4 -1.6 -1.2 -0.5 2.2
France
-4.6 -2.8 -1.7 -1.8 -1.3 3.3
Italy
-7.2 -6.5 -2.0 -2.0 -0.8 6.4
New Zealand
2.5 2.4 1.7 2.5 0.7 -1.8
Sweden
-6.9 -1.9 -0.5 2.9 2.1 9.0
EU11 weighted average
-4.1 -3.0 -1.4 -1.3 -0.6 3.5
Countries without legislated fiscal rules
Canada
-4.4 -1.6 1.0 0.9 2.3 6.7
Japan
-3.1 -4.4 -3.6 -4.2 -6.0 -2.9
United Kingdom*
-5.0 -3.8 -2.1 0.2 1.1 6.1
Australia*
-3.5 -2.0 -0.4 0.3 1.6 5.1
Source: OECD Economic Outlook, December 2000.
Note: The United Kingdom and Australia have legislated fiscal frameworks with non-legislated
numerical rules, as discussed in section 4.2.
254 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
Naturally, such comparisons do not permit strong conclusions
regarding the role of fiscal rules, since other important factors, such as
political considerations, are not held constant. For example, we cannot go
so far as to conclude that fiscal rules are not necessary for fiscal success; in
some of the above countries, such as some of the euro area countries, such
consolidation might not have been possible without the imposition of strict
rules. However, we can conclude from the above evidence that having
legislated fiscal rules is not a necessary condition for successful fiscal
consolidation in all countries.
In addition, it is important to note that industrialized countries’
experience with fiscal rules at the national level is relatively short, and the
time frame studied above does not encompass any major recession in
developed countries. In other words, fiscal rules have not yet been
seriously tested. The real test of whether countries will respect their fiscal
rules when they become binding and whether adherence to such rules will
be harmful or beneficial to these countries will come with the next major
recession.
4. Selected Recent Empirical Studies
This section reviews a selection of recent empirical studies on fiscal
rules, which address the following themes: (a) whether fiscal rules are
effective; (b) how the characteristics of fiscal rules are related to their
effectiveness; (c) whether fiscal rules limit a government’s ability to
engage in countercyclical fiscal policy; and (d) the relationships among
political and budgetary institutions and fiscal consolidation.
4.1 Effectiveness
The fundamental question addressed by empirical research on fiscal
rules is whether they are effective. Poterba (1996, 1997) reviews the nature
of balanced budget requirements at the U.S. state level and considers the
empirical evidence in the current literature. His findings suggest that
changes in budget rules and, more broadly, fiscal institutions can affect
fiscal policy outcomes.
In a study on the effectiveness of tax and expenditure limits, Stansel
(1994) shows that the relative rate of growth of spending in states with tax
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 255
and expenditure limits declined significantly within five years of the
implementation of the limits. Moreover, the relative decline in the growth
of state taxes was also significant in the five years immediately following
the tax and expenditure limit enactment. It is difficult to determine if the
relative declines in the rates of spending and taxation growth are due to the
enactment of the tax and expenditure limits, the determination of the
government in power to reduce the relative growth rates or some other
unspecified variable. Given this correlation, however, the introduction of a
tax and expenditure limit could potentially be used as a signal of
commitment to reduce tax and expenditure growth on the part of
policymakers.
A Canadian econometric study by Kneebone and McKenzie (1997)
examines the Alberta government’s past reactions to unanticipated shocks
to expenditures and revenues. They find evidence of asymmetric behaviour
– unexpected losses in revenue did not affect the current budget, whereas
unexpected increases in revenue were built into current revenue plans.
They suggest that Alberta’s legislation at the time may have been a
response to this asymmetry, as it required higher than expected revenues to
be used for debt reduction and prohibited the government from budgeting
the entirety of forecasted corporate income tax and natural resource
revenues.
Eichengreen and Bayoumi (1994) explore the impact of fiscal rules
on state general obligation bond yields. Their evidence indicates that, for
average levels of debt-to-gross state product, tax rates, and lagged state
unemployment, if a state without a tax and expenditure limit enacts one of
the more strict tax and expenditure limits, interest costs would decline by
nearly 50 basis points. Eichengreen and Bayoumi argue that a tax and
expenditure limit reduces the likelihood of future surges of borrowing and
hence the likelihood of default.
Poterba and Rueben (1999), on the other hand, find that tax and
expenditure limits lead to different outcomes in terms of borrowing costs.
Similar to Bayoumi and Eichengreen, they find that states with expenditure
limits face lower borrowing rates than states without such limits. However,
they find that states with tax limitation legislation face higher borrowing
costs than states without similar laws, presumably because tax restrictions
may limit a government’s ability to pay interest on its bonds. As for laws
that restrict deficits, they confirm that states with weak laws face higher
borrowing costs than those with strict laws. Along the same lines,
256 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
Goldstein and Woglom (1992) find that states with limitations on
borrowing face a lower cost of borrowing.
However, evidence from Mattina and Delorme (1996) highlights
factors apart from fiscal rules that can also ensure fiscal discipline; their
research indicates that discipline imposed by market mechanisms can be
effective in encouraging fiscally responsible policies. They use an
approach based on methodology similar to that of Bayoumi, Goldstein and
Woglom (1995) to estimate the supply of credit available to three Canadian
provinces as a function of the yield spread
4
. The spread may be regarded as
the default risk associated with a particular province. The underlying
hypothesis contends that the financial community demands a risk premium
that rises at an increasing rate with respect to protracted debt accumulation.
The Canadian evidence supports the existence of a non-linear supply curve
consistent with the hypothesis of a “market-based fiscal discipline”. The
non-linear supply curves suggest that the yield spread begins to accelerate
rapidly once the debt-to-GDP ratios of two provinces diverge by 35 to 50
percentage points.
4.2 Nature of Fiscal Rules and Effectiveness
Research on fiscal rules has also attempted to determine how the
nature of fiscal rules is related to their effectiveness. Poterba (1994)
explores the dynamics of U.S. state taxes and expenditures in the late
1980s and early 1990s and finds that more restrictive U.S. state fiscal
institutions, particularly annual balanced budget requirements and tax and
expenditure limits, are correlated with more rapid fiscal adjustment to
unexpected deficits.
A study by von Hagen (1991) examines the effectiveness of debt
limits and balanced budget requirements in U.S. states by comparing fiscal
performance indicators in states with and without debt limits and in states
with varying degrees of strictness in terms of their balanced budget
requirements. His analysis suggests that the presence of debt limits or
stringent balanced budget requirements affects the distributions of per
capita state debt, the debt-to-income ratio, and the ratio of nonguaranteed
to guaranteed debt. He finds that there is a higher percentage of states with
high ratios of nonguaranteed to guaranteed debt in the state groups with
__________
4
The study estimated the supply functions for Ontario, Québec and Nova Scotia. The yields spreads
and debt-to-GDP ratios are expressed in terms of the respective British Columbian figure.
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 257
debt limits or with more stringent balanced budget requirements. Given
that state debt limits routinely restrict only guaranteed debt and that
balanced budget requirements only target on-budget or general account
activities, this would suggest that states endeavour to avoid the full impact
of fiscal rules through accounting measures.
However, this assertion is refuted by research by Bohn and Inman
(1996). They use data from the U.S. states from 1970-1991 to explore the
effectiveness of different types of rules. They conclude that there is little to
suggest that balanced budget rules shift deficits into other fiscal accounts.
They also find that tighter balanced budget rules are associated with higher
state surpluses. Their analysis indicates that the most effective rules are
constitutional (as opposed to statutory) requirements that apply to the end-
of-year balance, rather than ex-ante budget requirements, and are enforced
by an independently elected state supreme court.
Alesina et al. (1999) use data from twenty Latin American and
Caribbean countries from 1980 to 1992 to examine the effectiveness of
their budget institutions. They create an index of budgetary institutions that
takes account of various forms of borrowing constraints, the role of a
macroeconomic plan in constraining the budget process and rules regarding
modification of the budget. Their findings indicate that there is a
significant negative relation between the stringency of the constraints and
the size of the primary deficit.
4.3 Fiscal Rules and Stabilization
One of the main areas of research regarding fiscal rules focuses on
determining whether such rules constrain the government’s ability to use
fiscal policy to smooth business cycle fluctuations. Research in this regard
has been limited to the U.S. states and has produced somewhat mixed
results.
Using state level data from 1971 to 1990, Bayoumi and Eichengreen
(1995) find that states with relatively smaller cyclical offsets tend to have
more stringent fiscal constraints. Specifically, they find that moving from
no fiscal restraints to the most stringent restrictions lowers the fiscal offset
to income fluctuations by around 40 per cent (the fiscal offset is a measure
of the sensitivity of the level of the fiscal balance to real output). In
addition, they conduct simulations that indicate that such a reduction in
258 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
fiscal stabilizers could lead to a significant increase in the variance of
aggregate output.
However, Alesina and Bayoumi (1996) show that although tighter
fiscal rules are associated with lower cyclical variability of the fiscal
balance, this does not lead to increased state output variability. Using U.S.
state data from 1965 to 1992, they find no statistically significant relation
between the variability of real state output and the stringency of fiscal
controls. They speculate that this may arise simply because stabilization at
the state level may not be very important, or because tighter controls limit
politically motivated and potentially destabilizing fluctuations in the
surplus as well as limiting countercyclical policies, leading to an uncertain
impact on output variability.
Conversely, Levinson (1998) points out that Alesina and Bayoumi
(1996) do not control for unobserved state characteristics that may be
correlated with business cycle fluctuations and the existence of state fiscal
controls. He suggests that the size of the state is correlated with its ability
to affect business cycle fluctuations through countercyclical fiscal policy,
and posits that if state fiscal policy matters more in large states than small
states, then the difference between business cycle fluctuations in states
with lenient versus strict fiscal controls should be greater for large states
than for small states. He then shows that although states with stricter rules
do not have higher volatility on average (from 1969 to 1995), the
difference in volatility between states with lenient and strict rules is indeed
greater among large states than among small states. From this he concludes
that there is evidence that strict balanced budget rules do exacerbate
business cycle fluctuations.
4.4 Political Economy Aspects
Yet another interesting line of research explores the relationships
among the political system, budgetary institutions and fiscal consolidation.
Hallerberg and von Hagen (1999) use pooled time series data from the EU
states from 1981-94 to contradict earlier studies, which contend that
proportional representation systems are more deficit-prone than pluralist
systems. Instead, they show that the presence of either negotiated spending
targets or delegation of power to a strong finance minister is key in limiting
deficit growth. Moreover, they conclude that one-party majority
governments, most common in pluralist systems, are most suited to
delegation to a strong finance minister, while multi-party coalitions, most
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 259
common in proportional representation systems, would generally do better
with commitment to negotiated fiscal targets.
Stein, Talvi and Grisanti (1999) reach a somewhat different
conclusion with respect to Latin American countries. First, they find that
countries whose electoral systems exhibit a large degree of proportionality
tend to have more procyclical fiscal policies, larger governments and larger
deficits. Next, using an index of budgetary institutions similar to that used
in Alesina et al. (1999), but for the period from 1990 to 1995, they find that
constraints on the deficit, a greater concentration of power in the finance
minister and in the executive, and greater transparency in budget
procedures all tend to lower deficits and debt. Contrary to Hallerberg and
von Hagen’s (1999) conclusions for European countries, they do not find
evidence that strong budgetary institutions neutralize the potentially
negative effect of a large degree of proportionality in electoral systems on
government deficits in Latin American countries.
Arreaza, Sørensen and Yosha (1999) expand on Hallerberg and von
Hagen’s (1999) conclusions and find that the government’s ability to
smooth consumption through government consumption and transfers is
much higher in countries with either delegation of power or negotiated
fiscal targets. In addition, they find that there is no statistical relation
between the deficit and the amount of consumption smoothing in a given
country. Thus, they conclude that the presence of effective budgetary
institutions, as defined above, can lead to lower average deficits as well as
efficient consumption smoothing via government deficit spending.
An interesting non-empirical analysis by Corsetti and Roubini
(1996) addresses the relationship between fiscal rules and the level of
government, and contends that fiscal rules are more suited to subnational
governments than to national governments. They point out that states are
aided in balancing their budgets by countercyclical transfers from the
federal government, and that states’ efforts to balance their budgets would
otherwise have a much larger negative impact on their residents’ income.
They also note the supply- and demand-side macroeconomic effects of any
action on the part of the federal government to balance the budget during a
recession would be much greater than similar actions at the state level,
since state revenues and expenditures represent a much smaller proportion
of state income than do federal revenues and expenditures. Finally, they
argue that insofar as individual states’ business cycles are not perfectly
synchronized, the actions of any given state trying to balance its budget do
not have a national impact. Conversely, an attempt by the federal
260 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
government to balance its budget during a recession could affect the whole
country.
Corsetti and Roubini’s arguments are complemented by Bayoumi
and Eichengreen’s (1995) findings, which emphasize the importance of
central governments in providing fiscal stabilization. They find that the
U.S. federal budget and social security provided about six-sevenths of the
total fiscal offset to income fluctuations in the 1970s and 1980s, while state
budgets provided only around one-seventh. Moreover, they find that
central governments (including social security funds) in the U.S.,
Germany, Canada, Japan, France and the Netherlands provided similar
degrees of fiscal stabilization from 1970 to 1989. Similarly, as mentioned
in the section on fiscal stabilization, Alesina and Bayoumi (1996) suggest
that their finding that the stringency of fiscal rules does not affect state
output variability might reflect the fact that the state’s role in stabilization
is not very important. From this they conclude that, if this is the case,
balanced budget rules may be effective for subnational governments, but
not for national governments.
5. Conclusions
This paper attempts to shed light on the role of legislated fiscal rules
in determining fiscal performance. The focus is placed on evaluating the
role played by fiscal rules in the fiscal consolidation of the 1990s. Some of
the conclusions reached in this respect may be extended to the role for
rules in maintaining surpluses, but a thorough analysis of this issue is left
for future research.
A review of the experiences of various countries as well as
subnational levels of government suggests that fiscal rules can be useful
tools for fiscal retrenchment, if properly designed. However, an
examination of the structural balances of countries both with and without
fiscal rules during the fiscal consolidation of the mid-1990s shows that
fiscal rules are not necessary for successful fiscal consolidation in all cases.
In addition, before making any judgements about the value of fiscal rules,
it is important to note that fiscal rules at the national level have not yet
been seriously tested; the real test will come with the next major recession.
The evidence from empirical studies generally supports these
conclusions. Many studies find that fiscal rules do indeed have an impact
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 261
on fiscal outcomes. In addition, research has attempted to identify certain
characteristics of fiscal rules that tend to be associated with greater success,
with some studies finding that stricter rules, rules that apply to the actual
budgetary outcome, rather than forecasts, and constitutional rules that are
enforced by an independent body seem to be the most effective. As for
whether fiscal rules impede a government’s ability to engage in
countercyclical fiscal policy, empirical results have been mixed. However,
recent research suggests that fiscal rules exacerbate business cycle
fluctuations, although empirical work has yielded mixed results in this
regard.
Finally, researchers have addressed how a jurisdiction’s political
institutions determine the appropriate type of budgetary institutions needed
to ensure fiscal discipline. In this respect, some researchers have found that
countries governed by multi-party coalitions, usually countries with
proportional representation systems, can best control deficit growth
through negotiated fiscal targets, while countries with pluralist systems are
more likely to be able to achieve fiscal discipline through delegation of
power to a strong finance minister. Similarly, some have suggested that
fiscal rules are more appropriate at the subnational level than at the
national level.
Overall, it would seem that there may be a role for legislated fiscal
rules in certain cases, but that legislated rules are by no means necessary
for achieving fiscal consolidation in all jurisdictions. Determining the
conditions under which legislated fiscal rules are indeed necessary to
ensure fiscal discipline, or determining when political commitment, non-
legislated rules or a commitment to transparency would be sufficient, is an
area for further research.
262 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
APPENDIX
FISCAL RULES IN LATIN AMERICA
This section provides a brief summary of fiscal rules currently in
place in three major Latin American countries.
i) Argentina
In September 1999, the Argentine Congress passed the Fiscal
Responsibility Law. The law sets a ceiling for the deficit and requires that
it decline such that balance will be achieved in 2003. It also established a
Fiscal Stabilization Fund, financed through tax revenues, to dampen the
impact of cyclical fluctuations and external shocks on government
revenues. In addition, the law prohibits the creation of off-budget items and
sets out new reporting requirements. Finally, it provides for penalties for
civil servants who do not implement the budget.
ii) Peru
Peru’s Congress approved the Fiscal Transparency Law in
December 1999, which sets limits on the deficit, the growth of government
expenditure and the increase in public debt. Similar to Argentina’s
legislation, Peru’s also established a fiscal stabilization fund to ensure
savings in peak years that may be used in times of recession. Furthermore,
it contains measures to encourage transparency and requires that the budget
be prepared within a three-year macroeconomic framework.
iii) Brazil
In Brazil, the Fiscal Responsibility Law was enacted in May 2000.
In contrast to the legislation in Argentina and Peru, Brazil’s law applies to
all levels of government. The law prohibits financial support operations
among different levels of government, sets limits on personnel
expenditures and requires that limits on the indebtedness of each level of
government be set by the senate. It also includes measures to improve
transparency and accountability. Separate legislation imposes penalties for
violations of the Fiscal Responsibility Law by public officials.
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 263
REFERENCES
Alesina, A. and T. Bayoumi (1996), “The Costs and Benefits of Fiscal
Rules: Evidence from U.S. States”, NBER Working Paper,
No. 5614.
Alesina, A., R. Hausmann, R. Hommes and E. Stein (1999), “Budget
Institutions and Fiscal Performance in Latin America”, Journal of
Development Economics, Vol. 59, pp. 253-73.
Alesina, A. and R. Perotti (1995), “Political Economy of Budget Deficits”,
IMF Staff Papers, Vol. 42, No. 1, pp. 1-31.
______________________ (1999), “Budget Deficits and Budget
Institutions”, in J.M. Poterba and J. von Hagen (eds.), Fiscal
Institutions and Fiscal Performance, Chicago, The University of
Chicago Press, pp. 13-36.
Alesina, A. and G. Tabellini (1990), “A Positive Theory of Fiscal Deficits
and Government Debt”, Review of Economic Studies, Vol. 57, No. 3,
pp. 403-14.
Alt, J.E. and R.C. Lowry (1994), “Divided Government and Budget
Deficits: Evidence from the States”, American Political Science
Review, December 1994, Vol. 88, pp. 811-28.
Arreaza, A., B.E. Sørensen and O. Yosha (1999), “Consumption smoothing
through Fiscal Policy in OECD and EU Countries”, in J.M. Poterba
and J. von Hagen (eds.), Fiscal Institutions and Fiscal Performance,
Chicago, The University of Chicago Press, pp. 59-80.
Bayoumi, T. and B. Eichengreen (1994), “The Political Economy of Fiscal
Restrictions: Implications for Europe from the United States”,
European Economic Review, Vol. 38, No. 3-4, pp. 783-91.
________________________________ (1995), “Restraining Yourself: The
Implications of Fiscal Rules for Economic Stabilization”, IMF Staff
Papers, Vol. 42, No.1, pp. 32-48.
Bayoumi, T., M. Goldstein and G. Woglom (1995), “Do Credit Markets
Discipline Sovereign Borrowers? Evidence from US States”, Centre
for Economic Policy Research, Working Paper No. 1088.
Bohn, H. and R.P. Inman (1996), “Balanced-Budget Rules and Public
Deficits: Evidence from the U.S. States”, Carnegie-Rochester
Conference Series on Public Policy, Vol. 45, No. 1-4, pp. 13-76.
264 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
Brunnen, P. and G. Pilon (1996), “Are Legislated and Constitutional Fiscal
Rules Essential for Fiscal Consolidation?”, Department of Finance,
mimeo.
Corsetti, G. and N. Roubini (1996), “European versus American
Perspectives on Balanced-Budget Rules”, The American Economic
Review, pp. 408-13.
Cukierman, A. and A.H. Meltzer (1986), “A Positive Theory of
Discretionary Policy, the Cost of Democratic Government and the
Benefits of a Constitution”, Economic Inquiry, Vol. 24, No. 3,
pp. 367-88.
Finance Representatives from Canadian Provinces and Territories, Various
provincial draft bills and legislation.
Goldstein, M. and G. Woglom (1992), “Market-Based Fiscal Discipline in
Monetary Unions: Evidence from the U.S. Municipal Bond Market”,
in M. Canzoneri, V. Grilli, and P. Masson (eds.), Establishing a
Central Bank, Cambridge, Cambridge University Press.
Hallerberg, M. and J. von Hagen (1999) “Electoral Institutions, Cabinet
Negotiations, and Budget Deficits in the European Union”, in J.M.
Poterba and J. von Hagen (eds.), Fiscal Institutions and Fiscal
Performance, Chicago, The University of Chicago Press, pp. 209-32.
Kneebone, R.D. and K.J. McKenzie (1997), “Alberta’s Budgetary
Response to Fiscal Shocks”, in C.J. Bruce, R.D. Kneebone, K.J.
McKenzie (eds.), A Government Reinvented: A Study of Alberta’s
Deficit Elimination Program, Oxford University Press, pp. 231-47.
Levinson, A. (1998), “Balanced Budgets and Business Cycles: Evidence
from the States”, National Tax Journal, Vol. 51, No. 4, pp. 715-32.
Mattina, T. and F. Delorme (1996), “The Impact of Fiscal Policy on the
Risk Premium of Government Long-term Debt: Some Canadian
Evidence”, Canadian Department of Finance, Working Paper.
Milesi-Ferretti, G. (2000), “Good, Bad or Ugly? On the Effects of Fiscal
Rules with Creative Accounting”, IMF, Working Paper 00/172.
National Association of State Budget Officers (1999), Budget Processes in
the States.
Nordhaus, W.D. (1975), “The Political Business Cycle”, Review of
Economic Studies, Vol. 42, No. 2, pp. 169-90.
THE ROLE OF FISCAL RULES IN DETERMINING FISCAL PERFORMANCE 265
Organisation for Economic Co-operation and Development (OECD)
(1999), Budgeting in a Surplus Environment,
PUMA/SBO(99)3/FINAL.
_____________________________________________________ (2000),
OECD Economic Outlook, No. 67.
Poterba, J.M. (1994), “State Responses to Fiscal Crises: The Effects of
Budgetary Institutions and Politics”, Journal of Political Economy,
Vol. 102, No.4, pp. 799-821.
____________ (1996), “Budget Institutions and Fiscal Policy in the U.S.
States”, American Economic Review, Vol. 86, No. 2, pp. 395-400.
____________ (1997), “Do Budget Rules Work?” in A. Auerbach (ed.),
Fiscal Policy: Lessons from Empirical Research, Cambridge, MIT
Press, pp. 53-86.
Poterba, J.M. and K. Rueben (1999), “State Fiscal Institutions and the U.S.
Municipal Bond Market”, in J.M. Poterba and J. von Hagen (eds.),
Fiscal Institutions and Fiscal Performance, Chicago, The University
of Chicago Press, pp. 181-207.
Rogoff, K. (1990), “Equilibrium Political Budget Cycles”, American
Economic Review, Vol. 80, No. 1, pp. 21-36.
Schick, A. (2000), “A Surplus, If We Can Keep It”, The Brookings Review,
Vol. 18, No. 1, pp. 36-9.
Stansel, D. (1994), “Taming Leviathan: Are Tax and Spending Limits the
Answer?”, Policy Analysis, CATO Institute, No. 213.
Stein, E., E. Talvi and A. Grisanti (1999), “Institutional Arrangements and
Fiscal Performance: The Latin American Experience”, in J.M.
Poterba and J. von Hagen (eds.), Fiscal Institutions and Fiscal
Performance, Chicago, The University of Chicago Press, pp. 103-33.
Tufte, E.R. (1978), Political Control of the Economy, Princeton, Princeton
University Press.
United States General Accounting Office (GAO) (1993), Balanced Budget
Requirements: State Experiences and Implications for the Federal
Government, GAO/AFMD-93-58BR, General Accounting Office.
___________________________________________ (1994), Deficit
Reduction: Experiences of Other Nations, GAO/AIMD-95-30,
General Accounting Office.
266 SUZANNE KENNEDY, JANINE ROBBINS AND FRANÇOIS DELORME
___________________________________________ (1999), Budget
Surpluses: Experiences of Other Nations and Implications for the
United States, GAO/AIMD-00-23, General Accounting Office.
Vanheukelen, M. (1994), “Combating 'Excessive Deficits' in the European
Union: Context, Substance and Procedure”, Department of Finance
Canada, mimeo.
von Hagen, J. (1991), “A Note on the Empirical Effectiveness of Formal
Fiscal Restraints”, Journal of Public Economics, Vol. 44, No. 2,
pp. 199-210.
Weingast, B.R., K.A. Shepsle and C. Johnsen (1981), “The Political
Economy of Benefits and Costs: A Neoclassical Approach to
Distributive Politics”, Journal of Political Economy, Vol. 89, No. 4,
pp. 642-64.
COMMENTS ON SESSION I:
THE PROS AND CONS OF FISCAL RULES
Ludger Schuknecht
*
1. Introduction
The papers presented in this session provide very important insights
into the history, role and conditions for the proper functioning of rules.
Two types of rules were identified – numerical rules, including e.g. the
SGP deficit rule, and procedural rules on transparency, enforcement etc.
An important question raised in these papers was when do rules
work? It was argued that rules and their objectives must be clear, simple
and the outcome measurable as compared to the target. Transparency,
monitoring and enforcement must be secured, and rules must be hard to
change. The institutional framework in which rules are imbedded is crucial
to insure their “success”. The following comments try to pick up some of
these themes as discussed in individual papers.
2. Reputation versus rules for fiscal discipline
George Kopits paper provides an excellent discussion of the “pros
and cons” of fiscal rules. I only have one small quibble regarding the
possible substitution between rules and reputation to safeguard fiscal
discipline. George argues that rules should come first, but that over time
reputation may make rules unnecessary. Germany and Japan are mentioned
as examples for this. I would disagree. In the political market, reputation is
not necessarily an equally strong incentive for “good” behaviour as in
private markets. In the private market, principals (share holders) can
dismiss agents (managers) at any time and reputational capital may be
protected that way. In politics, an election victory provides a four year
franchise and agents can not do anything if they feel cheated until the next
election.
__________
*
European Central Bank. The Comment reflects the personal views of the author and not those of
the European Central Bank.
LUDGER SCHUKNECHT
268
Governments may squander a reputation of tight fiscal policies if
they think that short term fiscal profligacy would help to win the next
election and another four year franchise. Germany is perhaps an example
of this. In Germany, the 1970s witnessed relatively high fiscal deficits.
This followed (amongst other reasons) the erosion of the golden rule. In the
1990s, the golden rule was circumvented through special funds. In both
periods, reputation was no reason to keep fiscal discipline. This potential
failure of reputation as a disciplining device is an important argument why
rules should be hard to amend/circumvent.
3. Clarity of rules in practice
As to Andrew Kilpatrick's paper, I am intrigued by its upbeat
rhetoric. Only time will tell whether the combination of numerical and
procedural rules applied in the UK warrants such optimism and I have my
doubts. The paper stresses the importance of clear objectives and
transparency. But reading carefully, there is more vagueness in the rules
than it is claimed. The paper talks about the requirement of “prudent” debt
levels but what prudence is seems to be decided by government. There is
more vagueness when it is argued that government is to provide "support to
monetary policy through changes in the fiscal stance where prudent and
suitable”. “Sensible discretionary policies” are also mentioned as part of
the government’s fiscal strategy elsewhere in the paper. Moreover the
formal rules (e.g. the golden rule) are not ambitious and have proven to be
quite soft in other countries that applied them.
4. Rules beyond macro targets
I do not have specific suggestions of improvements for the Paul
Atkinsen and Paul Van Den Noord's study. But as a more general
comment, so far we have focussed mostly on rules which take away macro
discretion from policy makers. The Atkinsen/Van Den Noord paper
focuses on public expenditure and how the consolidation framework could
also help to obtain leaner and more efficient government. This is very
interesting and important. It raises the question whether we should also
have rules for other aspects of public finances. Should we have rules
specifying certain functional expenditure levels, distributional or social
objectives, even employment rates or minimum growth rates? There is
COMMENTS ON SESSION I: THE PROS AND CONS OF FISCAL RULES
269
certainly a tendency in this respect in the policy arena. This topic is
worthwhile discussing but moving in this direction also bears risks:
• the objectives may not be easy to specify and may not be clearly linked
to fiscal policies alone
• the objectives may be unrealistic and politicised and may even discredit
rules more generally.
The paper also does an excellent job in discussing the importance of
different ways by which expenditure policies can achieve more efficient
government. Let me just mention vouchers in education where the debate
focuses on how best to achieve high education standards. Government
could be the provider or just the financier of “public” goods and services
such as education. Here the question arises whether certain activities
should be done by the private or the public sector “as a rule”. The question
may be easy to answer for airlines (private) or the military (public) but the
“right” approach in education is not obvious and probably depends very
much on the country circumstances.
5. Implicit and contingent liabilities
My final comment once more refers to the tendency towards rule
erosion and circumvention. If you have a deficit limit as your only
constraint on fiscal policies why not move activities off budget? This could
take the form of financing items or contingent/implicit liabilities. The
Atkinsen/Van Den Noord paper briefly discusses this issue and there is an
emerging literature elsewhere. However, there is still very limited
understanding of this domain. This is made worse by lack of transparency
in government financial and off-budget accounts. Contingent and implicit
liabilities are often not accounted and provisioned for.
Here procedural rules seem important again, including the
application of modern accounting rules and high transparency standards for
the government. Numerical targets, such as a prohibition of government
guarantees and off-budget accounts, could be imagined.
The growing importance of implicit liabilities is well recognised,
e.g. in the debate on ageing and implicit liabilities from the financial
system. The costs to government, if such liabilities have to be covered by
the fiscus, can be very high. A better understanding of implicit and
LUDGER SCHUKNECHT
270
contingent liabilities and how to apply rules to control such liabilities
seems a research area warranting more attention from the fiscal perspective
in the future.
&200(176216(66,21,
7+(3526$1'&2162)),6&$/58/(6
*HHUW/DQJHQXV
I would like to start this discussion by congratulating the
contributors to this session for their excellent papers and presentations. At
the same time, I want to apologise to them for not discussing the individual
papers in great detail - time being probably too short for that - and
therefore not giving them the attention they deserve. What I would like to
do instead, is focus on a few general issues and try to summarise and put
into perspective what we have heard in the various presentations.
In my opinion, the two main questions in this session are the
following: 1) do we need fiscal rules (or are they, to quote George Kopits,
an unnecessary ornament)?; and 2) if so, how should these rules be
designed and implemented?
Starting with the first question, I understand that the majority view
of the different speakers is that fiscal rules are indeed useful although there
are some qualifications. Perez Garcia and Hiebert look at rules as a way to
simply understand policy in a model-based approach rather than to actually
govern it. Kopits argues that rules can be an unnecessary nuisance for
governments which already enjoy a sound fiscal reputation. Finally, Peach
is rather sceptical about the role played by the regulatory framework in
consolidating the US federal budget.
With respect to the usefulness of fiscal rules, one could also turn to
the empirical literature, an aspect which was a bit neglected in the
presentations. In the Kilpatrick paper it is argued that this issue has
received surprisingly little attention in the empirical literature, the rules-
vs.-discretion debate having almost exclusively focused on monetary rather
than fiscal policy - although there are obvious parallels to be drawn. I fully
agree with Kilpatrick: the existing empirical literature on this issue is
probably neither very rich nor convincing. Most studies look at State
finances in the US and try to establish a link between fiscal rules and fiscal
outcomes (usually a simple numerical indicator such as the overall budget
balance or the evolution of public debt). On average, they tend to find a
__________
*
Nationale Bank van België / Banque Nationale de la Belgique.
strong positive relationship which leads them to conclude that the existence
of fiscal rules is some sort of guarantee for a satisfying fiscal outcome.
Obviously, nearly all of these studies suffer from the endogeneity problem
described by Poterba (1996) which I believe can be quite damaging.
Perceived causality might very well be mere coexistence because of voter
preferences. Indeed, if the electorate is averse to sloppy fiscal policy, it will
probably tend to elect a government favouring strict fiscal discipline while
at the same time additionally constraining the government by adopting
clear fiscal rules. Both the rule and the outcome can simply reflect voter
preferences: there needn’t be any causal relationship between them.
Bearing this caveat in mind, the existing literature nevertheless
VXJJHVWV that fiscal rules do seem to have a favourable impact on fiscal
outcomes. This is not sufficient, however, to advocate their use. Even if
they are shown to be a good drug against fiscal diseases, one should make
sure that they do not have any harmful side-effects. They do put constraints
on fiscal flexibility and hence limit the scope for tax smoothing or active
anti-cyclical policy. In the latter respect, the backward-looking sections in
the Kilpatrick paper are very enlightening. The UK fiscal stance EHIRUH the
adoption of the new regulatory framework is clearly shown to be
procyclical rather than the opposite. In the year 2000 Annual Report of the
Belgian National Bank similar evidence is presented for Belgium and the
euro area. This suggests that, if governments face no constraints
whatsoever, they seem to use this freedom in potentially destabilising ways
which might enhance cyclical fluctuations. On this issue one can also refer
to the work by Alesina and Bayoumi (1996) analysing the link between
fiscal rules and output variability at the State level in the US and
concluding that the latter is actually smaller in States with stronger fiscal
rules. Obviously, State finances might only have a very weak impact on
State output and the correlation reported by Alesina and Bayoumi might to
some extent be spurious but, on the other hand, their results could confirm
that fiscal rules effctively limit the scope for destabilising fiscal policy. All
in all, there seems to be some evidence - albeit not entirely convincing -that
fiscal rules not only provide for better fiscal outcomes in the purely
accounting sense (higher budget balances, lower public debt) but also for
better fiscal policy in the broader macroeconomic sense.
As to ZK\ fiscal rules are appropriate, the basic explanation that was
echoed in the presentations is related to the dynamic inconsistency of voter
preferences: voters always tend to prefer a larger deficit in the current year
than they had preferred for this current year earlier. In this respect, I am not
fully convinced that Kopits’s view of rules not being necessary for
governments that already enjoy a sound reputation is correct. In my view,
reputation is a very asymmetrical feature: it takes a lot of time to build but
can be lost quite rapidly. I would also like to add a simpler argument of my
own concerning the reasons why rules can be useful: the mere existence of
fiscal rules can lead to an increased media attention and coverage of fiscal
policy. Indeed, the media generally like to report more extensively on
fiscal outcomes when they can compare them to pre-fixed targets. In
Europe, for instance, public finances have received unprecedented media
attention since the Maastricht criteria have been agreed upon. Because of
this, public awareness about public finance issues grows and this is
obviously beneficial.
If the answer to first question - do we need fiscal rules? - is by and
large yes, there is far less unanimity on the second question with respect to
the exact design and implementation of these rules. Kilpatrick describes
two lines of thinking although the distinction is probably more a matter of
emphasis than of anything else. One approach emphasises accountability
and transparency (the typical example being the UK budgetary framework)
whereas the other one is blunter in a way and relies more on simple
yardsticks for numerical budgetary indicators (the deficit criterion of the
Treaty on European Union is one of the best-known examples). This
distinction between procedural rules and numerical ones was also alluded
to by the Chairman in his opening remarks.
If one focuses for the time being on the numerical rules, the
literature, e.g. the papers by Inman (1996) and Bohn and Inman (1996),
provide a number of characteristics of ’good’ rules. They should be simple,
concern ex-post government accounts, be enforced by a non-partisan
agency which can effectively inflict penalties and be costly to amend. If I
stop here, then this list of desired characteristics reads to a certain extent as
a blueprint for the Maastricht deficit criterion for instance. In several
presentations a few other characteristics have been highlighted however.
The rules should ideally be growth-oriented, take into account the
generational balance - if not actually target generational equity - and
explicitly refer to trend economic growth. I very much agree with the
importance of the latter criteria and what I would like to stress is that, in
my view, they conflict to a certain extent with the more traditional criteria
of ’good’ fiscal rules, most notably the call for simplicity and the need for
penalties.
Obviously, from an optimality point of view, rules should target
actual fiscal policies rather than fiscal outcomes. In practice, one generally
assumes, however, that the latter are adequate proxies for the former and
rules typically apply to outcomes. Nevertheless, one should be aware of the
fact that these outcomes are co-determined by a number of exogenous
elements that fall outside the government’s direct control. Predominant
among them is obviously the business cycle. Thus, ideally, the influence of
the cycle should be wiped out and the rule should be based on a cyclically-
adjusted indicator. In addition, to a certain extent, the government can
always shift revenue and expenditure from one period to another so an
ideal rule should consider some indicator of long-term sustainability.
Currently, generational accounting measures are already routinely used in
the budgeting process in a number of countries. Finally, if rules should
ideally support economic growth, then at the very least one should
distinguish between current and capital spending and consider rules of the
golden type. We all know, however, that this distinction is not sufficient:
investment in human capital in the form of expenditure on education, tax
incentives for private investment, etc. can enhance growth in the same way
as government investment in physical capital.
If rules take into account all of these refinements, however, - as I
believe they should in order to allow for a richer policy analysis - then one
obviously loses in terms of simplicity and, probably, enforceability. In
addition, these issues, cyclical adjustment, generational accounting and the
impact of public finances on growth, are very controversial and rules that
try to take them into account might prove difficult to sell.
Summing up and coming back to the Chairman’s distinction between
procedural and numerical rules, I believe that the former might in the end
be more important. I would argue that we should to some extent move
away from what Kopits dubbed the ’eurocentric’ view of bluntly comparing
actual fiscal balances with simple numerical yardsticks and focus attention
rather on issues such as transparency. Meanwhile, the profession should
spend even more time and energy on developing true or at least better
indicators of actual fiscal policy - taking into account, as Van den Noord
and Atkinson argue, the impact of that policy on private-sector behaviour.
Once a consensus about methodological issues has been reached and we
feel confident enough to use these indicators, we might move one step
further and consider numerical rules actually targeting these indicators
rather than simple accounting balances. Even if a lot of progress has been
made already (e.g. concerning cyclically-adjusted balances) this final goal
is probably still a (large) number of workshops away.
5()(5(1&(6
Alesina, A. and T. Bayoumi (1996), "The costs and benefits of fiscal rules:
evidence from U.S. States", NBER, Working Paper, No. 5614.
Bohn, H. and R. P. Inman (1996), "Balanced-budget rules and public
deficits: evidence from the U.S. States", &DUQHJLH5RFKHVWHU
&RQIHUHQFH6HULHVRQ3XEOLF3ROLF\, Vol. 45, pp. 13-76.
National Bank of Belgium (2001), “2000 Report”
Inman, R. P. (1996), "Do balanced budget rules work? U.S. experience and
possible lessons for the EMU", NBER, Working Paper No. 5838.
Poterba, J. (1996), "Do budget rules work?", NBER, Working Paper,
No. 5550.
