OPEN ISSUES ON THE IMPLEMENTATION OF THE STABILITY AND GROWTH PACT.

Buti, Marco ; Martinot, Bertrand

Bertrand Martinot [*]

Now that the budget deficits in the Euro Area are approaching balance, the Stability and Growth Pact (SGP) looks like a largely non-constraining institutional framework with little impact on national fiscal policies. This article challenges this view and argues that the implementation of the SGP 'at cruising speed' is faced with a number of outstanding issues: safeguarding the automatic stabilisers under the SGP; coping with the consequences of the asymmetric nature of the SGP for the co-ordination of macroeconomic policies; and ensuring the long-run sustainability of public finances. It concludes that enlarging the scope and enhancing the credibility of the stability and convergence programmes to become a true instrument of fiscal policy coordination in the Euro Area would be a first step in lifting the uncertainties surrounding the implementation of the SGP

Introduction

European integration and the formation of Monetary Union in Europe (EMU) have involved a number of agreements on the design of fiscal policy. The Stability and Growth Pact (SGP) is the most recent of these, having been agreed in Amsterdam and formally adopted in the summer of 1997. This Pact complements and tightens the fiscal provisions laid down in the Maastricht Treaty and aims to make budgetary discipline a permanent feature of EMU. [1] It was born in the context of both historically high public debt and renewed theoretical interest in rules-based fiscal policies. In addition to the general arguments in favour of institutional budgetary settings designed to curb the so-called 'deficit bias', strong fiscal discipline was deemed necessary to enhance the credibility of the single monetary policy. The SGP is probably the most stringent 'commitment technology' ever adopted by a group of governments in an attempt to establish and maintain sound public finances.

The Maastricht criterion that budget deficits should not exceed 3 per cent of GDP and the subsequent strengthening of fiscal constraints in the SGP were major driving forces behind the remarkable fiscal adjustment that took place in Europe in the transition period to EMU. [2] At that time, most European policymakers considered the SGP mainly as a prerequisite to participating in EMU. The economic debate, accordingly, mainly focussed on the short-term impact of the budgetary adjustment in a sluggish economic environment. [3] Now that the transition to EMU is over, fiscal policies in the European Union and in the smaller group of Euro Area economies seem basically on track. Fiscal deficits are low, public debt ratios are decreasing, the high tax burden is being reduced, and the cyclical position of Euro Area countries has improved. In this situation the SGP looks like a benign, largely 'harmless' institutional framework.

This article challenges the view that the implementation of the SGP is no longer an issue. We argue that the SGP, if properly applied, will have important implications for the behaviour of budgetary authorities in the short-term evolution of the budget balance over the cycle, and influence the interplay with the single monetary authority, and in the long term because of its impact on the sustainability of public finances. Recent theoretical and empirical research suggests that the practical implementation of the SGP 'at cruising speed', for which we do not have any experience, is faced with a number of open issues which are just emerging.

More precisely, three questions deserve particular attention:

* How can we safeguard automatic stabilisers while complying with the fiscal discipline requirements of the Treaty?

* What are the consequences of the SGP for the coordination of macroeconomic policies?

* To what extent is the long-run sustainability of public finances guaranteed by the SGP?

The following section addresses the issue of the choice of the appropriate medium-term fiscal targets which would safeguard the 3 per cent ceiling while allowing automatic stabilisers to come into play freely. The third section focuses on fiscal policy coordination under the Pact. The fourth section is devoted to the implications of the SGP for long-term sustainability in the light of the expected budgetary consequences of ageing populations. The fifth section reviews the experience of the first year of the SGP and points to the short-term challenges facing the implementation of the Pact. A conclusion follows.

Fiscal discipline and stabilisation under the SGP

Economies can have self-stabilising mechanisms, such as automatic stabilisers, built into them to deal with shocks. If tax rates are progressive or receipts rise more than in line with output, and spending on items such as social security depend on the level of employment, then, as the economy slows down, spending should automatically rise and taxes should automatically fall more than in proportion to output. This should help boost demand and will help cushion the effects of any shock to an economy. Ensuring that automatic stabilisers were safeguarded under the SGP was one of the earliest concerns about the Pact raised by economists. [4] It was argued that given the 3 per cent of GDP 'hard ceiling', automatic stabilisers would run the risk of being curtailed precisely when countries, having lost monetary independence in EMU, would need them most to cope with large or asymmetric shocks. There is some flexibility in the Pact. In the case of a 'severe' recession, the SGP allows individual countries to breach the 3 per cent threshold. However, such recessions are by their nature relatively infrequent. [5] Furthermore, this escape clause only holds when the deviation is temporary and the deficit remains close to 3 per cent of GDP.

There is much discussion of the SGP and automatic stabilisers in the literature. A first strand aims at quantifying the loss in stabilisation capacity resulting from a constrained fiscal policy. A second strand attempts to estimate appropriate medium-term targets through descriptive statistics of past budgetary behaviour or econometric analysis (using structural VAR models or simulations with large macro-econometric models). There are a number of general conclusions from this debate. In particular, it is agreed that it would be possible to fully safeguard automatic stabilisers within the spirit of the SGP by reaching a suitable medium term target for the budget deficit. This would be such as to guarantee a safety margin that would allow enough breathing space for the automatic stabilisers to work freely without breaching the 3 per cent ceiling. [6] If deficits were set in this way then the operation of automatic stabilisers on their own would result in smoothing the business cycle by between 20 and 30 per cen t for most EU countries (Buti and Sapir, 1998; Van den Noord, 2000). [7]

Quantifying the loss in out put resulting from inadequate safety margins

Numerical simulations of small calibrated models or large econometric models have been carried out in order to quantify the loss in stabilisation arising from a constrained fiscal policy. Along these lines, Eichengreen and Wyplosz (1998) provide a counterfactual analysis of the SGP by running simplified econometric models for the four biggest EU members. They simulate the output developments that would have resulted from the application of the 3 per cent ceiling on past fiscal policies during the 1974-95 period. Their findings suggest that the loss in output stabilisation arising from past (and often imprudent) fiscal policies would have been substantial, though unequally shared among member states. [8]

This study, however, assumes somewhat unrealistically that no alteration in agents' behaviour occurs after the advent of the SGP. First, liquidity constraints and expectations on future fiscal policy matter for the consumers' behaviour. Against this background, the SGP, provided it is credible, will be internalised by economic agents and is likely to affect private consumption, hence the effectiveness of automatic stabilisers. Second, the 'spirit' of the SGP requires that, at 'cruising speed', governments should alter their fiscal behaviour, probably adopting some kind of tax smoothing subject to the 3 per cent ceiling.

It is important to take account of the impact of rule guided fiscal policy on individual behaviour when assessing the SGP. Martinot (2000) produces a first attempt to provide a quantification of the impact of the SGP on the variability of private consumption taking account of these structural changes. Referring to the existing literature, a highly general formulation of the consumption function can be written as follows:

[C.sub.t] = f[[E.sub.t]([Y.sub.t+i]-[T.sub.t+i]), [Y.sub.t] - [T.sub.t] (1)

where Y - T stands for disposable income and E is the expectation operator. Introducing an ad hoc exogenous output gap, based on past data, of the form In [Y.sub.t] = trend+A(L)ln [Y.sub.t] + [[epsilon].sub.t], one can simulate the consumption function for various choices of the parameters (budget deficit, interest rates, liquidity constraints, variance of output gap). The degree of forward-lookingness on the part of consumers is the most important assumption for the structure of the results. In Martinot (2000), the consumption function is partly of the Blanchard (1985) type with rational expectations and individuals facing finite horizon, and partly myopic with some individuals facing liquidity constraints. [9]

In any study such as that of Martinot (2000) some measure of the change in variability in the economy has to be constructed. One relevant criterion for measuring the loss in stabilisation capacity of the fiscal policy, as a function of the government's medium-term budgetary target [[d.sup.*].sub.s] (expressed in cyclically-adjusted terms) is the ratio

r([[d.sup.*].sub.s])= var C (under the SGP) - var C (without the SGP)/var C (wihtout the SGP) (2)

The findings in Martinot (2000), based on calibrating this model for France, suggest that a deficit target close to the SGP's 3 per cent ceiling could result in an extremely high variance in consumption. Conversely, the implementation of the SGP would barely affect the variance in consumption for a structural deficit around 1 per cent. [10] Therefore, it would appear that the choice of an appropriate safety margin is crucial in assessing the impact of the SGP on consumption and output volatility.

Estimating the appropriate safety margin (I): descriptive analyses

If we assume that there is no role for discretionary fiscal policy, the total safety margin under the 3 per cent threshold can be decomposed into a cyclical component and a safety margin allowing for 'erratic' budgetary developments not linked to the operation of the built-in stabilisers. The cyclical component can be seen as depending on the variability of the business cycle and on the sensitivity of tax and expenditure to economic activity. Hence the budget deficit as a percentage of GDP, d, can be written as:

d = [d.sub.s] - [alpha]G+[epsilon] (3)

where [d.sub.s] is the cyclically adjusted budget deficit, [alpha] accounts for the (average) sensitivity of the deficit to the output gap G, and [epsilon] accounts for unforeseen variability of the budget.

The medium term target [[d.sup.*].sub.s] needed to avoid breaching the 3 per cent of GDP ceiling for any (potentially large and negative) values of G and [varepsilon] can be computed relatively easily using (4):

[[d.sup.*].sub.s] = 3+[alpha][G.sub.max] - [[epsilon].sub.max] (4)

A number of recent studies have attempted to estimate the cyclical and the erratic components of the deficit target on the basis of past business cycle history. This requires that we can calibrate both the sensitivity of the deficit to output and also estimate the potential scale of the output gap in severe recessions.

The sensitivity of the deficit ratio to the output gap depends on the elasticities of various taxes and expenditure to GDP. According to widely used estimates by the OECD and the European Commission, the average sensitivity of the budget balance to the cycle is around 0.5. The Nordic countries and the Netherlands tend to have a higher sensitivity (0.8-0.9) because of their more extensive tax and welfare systems (Van den Noord, 2000; European Commission, 2000).

The 'worst possible' output gap was estimated in European Commission (1999) on the basis of three alternative measures. This study calculated:

* the unweighted average of the largest negative output gaps in EU Member States which, over the period 1960-98, was 4 per cent of GDP;

* the largest negative output gap that has been recorded in each Member State over the period 1960-97;

* the average volatility of the output gap in each Member State, as measured by twice its standard deviation.

The 'representative' negative output gap for each country was then assumed to be well gauged by the average of the two extreme cases of these three alternatives. The cyclical safety margin for the difference between the target budget deficit and the SGP ceiling was estimated to be of the order of 1.5-2 per cent of GDP for the large majority of EU countries and above 2.5-3.5 per cent of GDP for the Nordic countries and the Netherlands. Using a similar methodology, the IMF (1998) and the OECD (1997) find that a structural deficit in the range of 0.5-1.5 per cent of GDP and below 1.5 per cent of GDP, respectively, would be enough to allow the automatic stabilisers to operate without breaching the 3 per cent of GDP deficit threshold even in periods of pronounced cyclical slowdown.

The estimation of the erratic term [varepsilon] is less straightforward. A first attempt to provide a measure of this component was made by Artis and Buti (2000). In order to disentangle erratic budgetary fluctuations they compare budget deficit forecasts made in spring each year by the European Commission with actual outturns for the same year. Once the estimated budgetary effects of forecast errors on GDP growth are netted out, what is left can be taken as an approximation of the 'pure' risk of erratic budgetary developments. By retaining only the positive value of this component, the additional margin to cover for this risk is estimated to be of the order of 1/2--1 per cent of GDP.

All in all, these analyses show that broadly balanced budgets seem an adequate target for most EU countries, but some should aim for a surplus. This method is appealing because of its simplicity and its intuitive feature. Its main shortcomings may be its lack of theoretical foundations, its (inevitable) backward-looking character and the ignorance of any feedback from the choice of the target to the cyclicality of the economy.

Estimating the appropriate safety margin (2): econometric analyses

In order to overcome, at least partly, the limitations of the simple descriptive statistical analyses, more sophisticated techniques have been applied. These have involved the estimation of small structural VARs and the use of large econometric models, and both are discussed in this section.

Dalsgaard and de Serres (1999) estimated a structural VAR model for 11 EU countries in order to assess the effect of four independent economic disturbances on the government deficit: supply, fiscal, real private demand and monetary shocks. Stochastic simulations were then performed over a given time horizon. Non-fiscal shocks were simulated according to their past distribution, the fiscal shocks being truncated in order to capture only the fiscal developments due to the functioning of automatic stabilisers. For a given initial fiscal position (equal to a medium-term target), the probability of the fiscal deficit breaching the 3 per cent ceiling over the chosen time horizon was calculated. This experiment was replicated for various time horizons and initial budget positions. The paper considers as 'safe' a medium-term target, which does not lead the current deficit to breach the 3 per cent ceiling with a 90 per cent confidence over a three-year horizon. The simulation results suggested that, for the majority o f countries, the appropriate medium-term target would be of the order of 1-1.5 per cent of GDP. For Finland, the United Kingdom, Denmark and Sweden, moderate surpluses would be required.

A more sophisticated approach is to use a large-scale macroeconometric model and simulate how the automatic stabilisers operate under the SGP. Barrell and Pina (2000) and Dury and Pina (2000), for instance, use an explicit model of the European economies (namely NiGEM, the National Institute Global Model) to assess the working of automatic stabilisers under the SGP with stochastic shocks applied in the future but derived from estimated past macroeconomic relationships. In Barrell and Pina (2000), the automatic stabilisers are allowed to operate freely over the 1999-2005 period. The probability of breaching the 3 per cent ceiling is then calculated. Their findings suggest that compliance with the targets announced in the current stability and convergence programmes would make the SGP and the unconstrained operation of automatic stabilisers broadly compatible. In Dury and Pina (2000), the fiscal policy fully internalises the provisions embodied in the SGP. [11] Again, the results support the view that complianc e with the stability and convergence programmes would not affect the stabilisation capacity of the fiscal policies.

The main advantage of these approaches is that they capture the interplay between output fluctuations and fiscal policy developments. Moreover, in the case of the NiGEM simulations, another advantage is that the model tries to capture some structural developments that might be expected to take place with the establishment of EMU. The main shortcomings of these approaches is that they use large, complicated models that are not transparent, and they do not quantify the loss in the effectiveness of the SGP that would result from the choice of an inappropriate medium-term target for the budget deficit.

In spite of the diversity of the above methods, there exists a large consensus about the medium term targets to be reached by each Member State. Typically, Nordic countries should aim at a surplus (1-2 per cent of GDP) while the other Member States could safely comply with the SGP by targeting a roughly balanced budget over the cycle. All these studies are subject to serious qualifications. Above all, the way the automatic stabilisers operate within the SGP is partly unresolved because it is highly sensitive to assumptions on their nature and on the nature of cycles.

The size, origin and synchronicity of the shocks affect the size and volatility of cyclical fluctuations in output. While EMU will entail changes in the potential distribution of shocks it is not clear in what direction these will take. In the longer run, it might be expected that cyclical variations will become more synchronised between EMU members. Given the stability-oriented macroeconomic framework of EMU, country-specific, policy-induced shocks are likely to decrease in the Euro Area. Furthermore, trade integration may spread shocks more uniformly across frontiers. Sound macroeconomic conditions will also increase the effectiveness of stabilisation policies designed to smooth the business cycle. The extent of these convergence-promoting effects is not known and in the shorter run, different economic structures as well as constraints on the use of stabilisation policies may work in the opposite direction.

It is also unclear whether the sensitivity of budget balances to the cycle will change significantly once in EMU. Ongoing reforms of the tax system -- in particular of corporate taxes and social security contributions -- and of transfer payments to the unemployed entail a reduction of the budget's cyclical sensitivity but such an effect will take time to materialise.

Finally, as argued by Artis and Buti (2000), the nature of the EGB's monetary strategy and its degree of 'conservatism' affect the degree of ambition of the fiscal target. In relation to the pre-ERM era, of course, countries give up the possibility of using their own monetary policy actions to stabilise their economies; however, compared to the ERM period, (non-German) countries can expect a proportionate weight in ECB decisions which they did not have in the Bundesbank's policy.

All in all, EMU is 'too young' to allow us to conclude that the current consensus on the safety margins needed to allow the SGP to operate without costs will stand the test of time. In addition, as the cyclical behaviour of the Euro Area economy adapts to the new EMU environment, the issue of the appropriate medium-term targets for budget deficits will need to be addressed again.

The SGP and policy coordination

In EMU, a centralised body, the European Central Bank (ECB), conducts monetary policy, while fiscal policies rest with Member States. This situation may call for coordination between national fiscal policies in order to avoid negative spillovers and coordination between monetary and fiscal policies, so as to ensure an adequate policy mix at the national and Euro Area level.

The SGP provides a seemingly clear-cut answer to the quest for macroeconomic coordination. To the extent that EMU members select an appropriate medium-term target and just let automatic stabilisers play around that target, no need for explicit coordination arises. The ECB will preserve price stability and, provided that this is not endangered, will cushion symmetric shocks; automatic stabilisers at the national level will smooth country-specific shocks. Hence, what we will see in place will be essentially a rules-based, 'negative' coordination, that also aims, through close monitoring, to prevent budgetary misbehaviour that might disrupt the functioning of EMU.

According to this 'philosophy', it is unnecessary to coordinate national fiscal policies if one rules out discretionary policy. In that case, the strict compliance with the quantitative targets laid down in the SGP (i.e. the 'close-to-balance' rule combined with the 3 per cent ceiling) should suffice to optimise the stabilisation properties of fiscal policies (and to avoid spillovers, exert pressure on monetary policy, etc). The fiscal stance of the Euro Area would therefore be considered a simple ex post outcome.

This 'automatic pilot' view of the conduct of fiscal policies in EMU is however oversimplified. Automatic stabilisers may not be sufficient in the event of a particularly severe symmetric shock that monetary policy alone is unable to counter. If a discretionary fiscal impulse was needed, each member country might refrain from taking the initiative, hoping to free ride on other countries. This wait-and-see attitude would be compounded by the threat of sanctions under the SGP should the fiscal loosening bring the budget deficit above the 3 per cent ceiling. The fiscal response at EU-level, therefore, could be suboptimal if no coordination occurred amongst fiscal authorities. [2]

In addition, an appropriate medium-term target and the sanctions foreseen by the SGP [13] would prevent the deficit from breaching the 3 per cent ceiling in 'bad' times. However, the incentives to let automatic stabilisers work fully over the whole economic cycle may be waning. While the founders of the SGP probably had this 'philosophy' in mind, actual behaviour may be quite different. In the case of a booming economy the logic of the SGP implies that the government should aim at a budgetary surplus. The SGP, however, does not prevent national policymakers from targeting a balanced budget when the output gap is in positive territory instead of letting the automatic stabilisers work.

Finally, the fact that the SGP is "all sticks and no carrots" (Bean, 1998) may entail a pro-cyclical bias in the conduct of budgetary policy. Chart 1 illustrates this case. It plots the total and the cyclically adjusted budget balance against the output gap. The dotted lines represent budgetary behaviour consistent with the SGP 'philosophy', while the bold lines represent 'surplus-resistant' fiscal behaviour. If governments remain trapped in the 'no-surplus' culture of the 1970s and 1980s, they will tend to offset the working of the automatic stabilisers for sufficiently large, positive output gaps. [14] While the actual budget balance would remain broadly stable, the cyclically adjusted budget balance would worsen, adding to the risk of overheating. The result would be an overburdening of monetary policy and an unbalanced policy mix. Here, the coordination failure would be between fiscal policies and the single monetary authority.

An effective, coordinated monitoring would be required in order to make sure that coherent fiscal behaviour takes place both in bad and good times. A key issue in this respect is to shift the attention from actual to cyclically adjusted developments of the budget. However, the difficulty in computing structural balances may hinder the effectiveness of such monitoring. [15]

SGP, ageing and the long-term sustainability of public finances

The arithmetic of debt, deficit and age-related spending

Now that the positive drift in general government expenditure experienced in the 1970s and 1980s seems to be over in most Member States, ageing represents the main structural challenge for public finances. This long-run dimension to fiscal policy is still not fully incorporated in budgetary behaviour and institutions. Several official documents already point to the need to address the ageing issue in the context of the SGP. [16]

Strictly speaking, long-term sustainability is ensured if there is no long-term upward drift in debt accumulation (equivalently the ratio of debt to GDP is bounded). If applied at all times, the Maastricht cum SGP framework obviously ensures sustainability. For instance, d is the deficit as a percentage of GDP, y is the rate of growth of real income, [pi] is the rate of inflation, b is the stock of debt, from the familiar government budget constraint, and b is the change in the debt stock as a percentage of GDP then

b = d - (y + [pi])b (5)

And hence the debt stock is sustainable with b([infinity]), the debt stock as time goes to infinity, bounded if

d[less than]d = 3%, y+[pi] = 4% then b([infinity])[less than]d/y+[pi] = 75% (6)

Hence, compliance with the SGP is sufficient to ensure long term sustainability. Besides, a 'close-to-balance' structural deficit would result in a de facto long-term cancellation of public debt, with the debt stock as a percentage of GDP converging toward zero. These simple calculations imply that any rise in spending, such as that due to ageing, would be offset by moves in the opposite direction so as to keep the structural balance constant.

Clearly, a downward trend in the debt stock, by entailing a reduction in the interest burden, would help to accommodate the budgetary consequences of ageing.

It is often argued that EMU members should strive for more ambitious targets than those necessary to safeguard the 3 per cent ceiling in order to pre-empt the effect of ageing. A useful benchmark is the level of the budget balance that would allow them to offset fully the expected impact of ageing on public spending via lower interest payments. Starting again from the debt accumulation identity, under the usual assumptions of a constant interest rate and growth rate, it is easy to show [17] that the medium-term fiscal balance d[degrees] required to fully offset the rise in expenditure by a saving on interest payments is given by:

d[degrees] = A(y + [pi])/i[T - 1 - [e.sup.-(y+[pi])T]/y + [pi]] - b(0)(y + [pi]) (7)

where A is the cumulated impact of ageing on public spending (percentage of GDP) over the period 0-T, and b(0) is the initial stock of debt. Table 1 presents this calculation under various assumptions concerning the initial debt level, the impact of ageing and the nominal interest rate and growth rate.

Inspection of the table shows that countries with high debt ratios, who have a lot of potential for interest savings, require a higher total deficit or a lower surplus to generate the interest savings needed to offset a given rise in age-related spending. For instance, if the expected increase in spending over the next 30 years is 4 per cent of GDP -- as in the case of Belgium and Italy, according to Franco and Munzi (1997) -- a broadly balanced budget (-0.2 per cent of GDP) would be required if the initial stock of public debt is 100 per cent of GDP, while a surplus of 1.4 per cent of GDP is needed if the debt ratio is 60 per cent of GDP. As is well known from debt arithmetic, however, these less ambitious budgetary targets do not imply that high debt countries have an 'easier job'. It is quite the opposite, as it can be easily shown that the primary surplus corresponding to the required overall budget balance is higher in the case of high debt countries. If the primary surplus, not the total deficit, is tak en as a measure of the policy effort, pre-empting ageing requires a tougher adjustment in countries with higher initial debt. The same conclusion is attained if the discretionary policy effort is proxied by the 'tax gap' or other similar indicators.

Ageing and the SGP: some political-economy considerations

Striving for more ambitious structural targets can make a significant contribution to pre-empting the budgetary implications of ageing via lower interest payments. However, as clearly stated in European Commission (2000), reducing the stock of public debt should not be seen as a substitute for tackling the problem at source, namely through reforms of age-sensitive spending. Furthermore, one may ask whether a constant deficit target, even if ambitious, entails an 'optimal' public debt path taking into considerations the expected time profile of ageing-related spending.

We look at this complex issue from the point of view of intergenerational equity. The close-to-balance rule implies a regular decline in public debt while the increase in spending due to ageing populations is expected to occur gradually in most countries until 2010 and accelerate sharply thereafter. For pay-as-you-go (PAYG) pension systems, balancing the books will require a sharp rise in social security contributions and/or a decrease in replacement ratios. Public debt, by transferring interest payments from one generation to another, can be used to correct potential inequities arising from these developments. Therefore, intergenerational equity considerations might lead governments to favour large swings in the evolution of the public debt. This conclusion can be illustrated through a small accounting model calibrated on EU data. Under the assumptions that the replacement rate remains unchanged, [18] there is no shift towards funding pension schemes and the very long-term debt ratio (in 2100) is given (see appendix for details). Chart 2 illustrates various developments in contribution rates (underlying various intergenerational transfers) resulting from three (exogenously determined) deficit paths:

* In the constant deficit (1 per cent of GDP) scenario, the intergenerational transfers resulting from the debt dynamics are more or less neutral since the steady state debt level would be reached before ageing translates into significant increases in contributions.

* In the second scenario, the government decides to run surpluses in the early stage of the ageing process (3-4 per cent of GDP over ten years), thereby significantly decreasing the public debt ratio, which, in turn, would allow them to relieve the tax burden on the generations living in the subsequent period when the budgetary impact of ageing will be fully felt.

* In the third scenario, the government is myopic and runs a deficit of 2-2.5 per cent in the early stage of the ageing process. While probably gaining the approval of current generations, this deficit path would be detrimental to future generations, as they would be faced both with unfavourable demographic developments and higher interest payments. This effect, however, would remain small as the SGP prohibits deficits that could be seen as too large. The SGP, therefore, would 'protect' the interest of those generations.

As shown in Chart 2, future generations will have to face increasing contributions but the rise will be smaller the higher the retrenchment effort made by current generations. 'Front loading' the adjustment (initial surplus scenario) would allow public debt to drift temporarily upwards in the course of the following decades, thereby smoothing the subsequent burdening of future generations. In addition, this scenario would make baby boomers, who are comparatively well off, contribute for their disadvantaged children. Hence, compared to a constant deficit and, even more, to myopic government behaviour, running surpluses in an initial period would intuitively improve intergenerational equity. [19]

This analysis points to the benefits of accelerating budgetary consolidation now, beyond the targets of the SGP. Such a conclusion, however, hinges upon maintaining the current PAYG pensions systems. If governments were to envisage a drastic shift to funding, the conclusion could be reversed. Typically, it could be desirable to allow for a temporary rise in public debt while shifting to partly funded systems in order to relieve the so-called 'double burden' falling on current generations.

In both cases, therefore, reconciling long-term compliance with the SGP and intergenerational equity remains an open issue. Whether EU members should add a 'long-run safety margin' to the cyclical safety margin is unclear and depends also on the type of pension reforms envisaged (trimming the generosity of PAYG systems or shifting to funding).

First lessons from the fiscal policy under the SGP

A good start

Since the beginning of 1999, fiscal policies in the EU have been conducted within the framework of the SGP. All in all, countries have lived up to the commitments set by the SGP. The 3 per cent reference value has de facto become a 'hard ceiling' that no Member State has breached or even approached. Much progress has been made towards reaching a 'safe' medium-term target. It had often been alleged that 'Maastricht fatigue' would prevail once the member countries were in EMU. So far, on the contrary, the need to continue the structural budgetary adjustment has been widely recognised and clearly reiterated in the stability and convergence programmes in 1999 and 2000. As shown in Table 2, by 2003 the fiscal targets are broadly in line with the medium-term targets as estimated in part 2. Indeed, deficits are expected to be reduced and practically eliminated by the end of the planning period for the Euro Area as a whole.

The impact of the SGP on national fiscal strategies in the first year of the euro can be illustrated by plotting, as in Chart 3, the fiscal consolidation in 1999 against the remaining adjustment effort to attain the close-to-balance targets. [20] There seems to be a positive correlation between the deficit reduction and the initial gap between the cyclically adjusted deficit and the 'safe' benchmark. This would indicate that prima fade the adjustment still to be accomplished might have been one of the factors shaping fiscal strategies in the first year of the euro. Indeed, the budget balance improved even in countries with a negative output gap and in spite of the activity slowdown.

The continuation of fiscal consolidation in 1999 may have been instrumental in facilitating a growth-friendly monetary stance. The potential usefulness of the SGP as a coordination device to bring about a balanced policy mix at the outset of EMU was first highlighted by Allsopp and Vines (1996, 1998). From a short-term perspective, a common drive towards further fiscal adjustment would create the conditions to allow the European Central Bank to deliver a desirable offsetting monetary response: "only if all (countries) act together will the monetary offset to fiscal tightening be likely to eventuate. Thus, participating governments will not only want to commit themselves, they will want to impose commitment on others as well" (Allsopp and Vines, 1996, p. 99). Without such a common undertaking, the likelihood of an over-restrictive monetary stance would increase, particularly because of the credibility-building strategy that should be being adopted by the newly created ECB.

Chart 4 illustrates the policy mix in the first year of the euro. The graph pictures the fiscal stance -- as measured by the change in the cyclically adjusted primary balance (CAPB) -- and the change in the monetary conditions captured by the variation in the Monetary Conditions Index (MCI). [21]

Chart 4 shows that the Euro Area and most countries fall in the top-left quadrant, where a fiscal tightening is accompanied by easier monetary conditions. Hence, contrary to the fears of many observers during the political controversies in early 1999, the cyclical downturn due to the Asian crisis did not trigger a general move towards expansionary fiscal policies, which could have implied a loss of credibility of the commitment to budgetary discipline. This contributed to the credibility of the whole EMU stability-oriented policy framework and permitted an easier stance of monetary policy than may have emerged otherwise.

Immediate challenges

The debate on the Pact has so far focussed on the possible constraint it can put on national fiscal policies in a context of sluggish economic activity. Now that European economies are experiencing a cyclical upturn, the application of the SGP has to meet new challenges. For the first time, Member States have to cope with the asymmetric nature of the SGP highlighted in section 3. Even if they comply formally with the SGP, the current development of national fiscal stance may destabilise the policy mix.

A stylised description of the risk of an unbalanced policy mix is provided in Chart 5. As was discussed above, a damaging policy conflict in the first year of the euro between national fiscal authorities and the newly created ECB was avoided. On the contrary, the outturn was a combination of fiscal tightening and monetary loosening which helped to consolidate the credibility of the stability-oriented framework of EMU and put in place the foundations of the current cyclical upturn.

The risk of 'coordination failure' may arise again in the coming years even though it was avoided in 1999. In the present context, it is widely recognised that the appropriate policy mix should once again involve a combination of further fiscal consolidation with a relatively accommodating monetary policy. This view is conditioned on the fact that Member States' debt levels are hardly sustainable if due account is taken to the impact of ageing. In addition, in most cases, the cyclically adjusted deficit still lies above the level that might be regarded as 'safe'. Finally, moderate monetary and financial conditions would stimulate business investment, which has been the weakest component of activity in most countries in the past decade while being an important component of US growth.

However, as highlighted in the second section, a 'surplus-resistant' budgetary behaviour is likely to prove politically tempting. Failure to compensate for the effects of the planned reductions in the tax revenue via spending retrenchment would result in a pro-cyclical shift in the fiscal stance. This would put an increased burden on the single monetary authority and result in a strong monetary tightening. As pictured in Chart 5, once again, though for different reasons than in 1999, the risk would be an unbalanced policy mix, which could jeopardise the current macroeconomic framework.

Concluding remarks

This paper has argued that, far from being a simple accounting constraint aiming at keeping budget deficits under control, the SGP will deeply affect the behaviour of fiscal authorities. Provided that countries choose an appropriate medium-term target for the structural budget deficit and let automatic stabilisers play freely, a high degree of cyclical smoothing will be attained without endangering the 3 per cent deficit ceiling set by the Treaty. Ambitious fiscal targets, by entailing a rapid fall in the stock of public debt, will help to pre-empt, at least partly, the budgetary pressure arising from ageing populations.

However, other important aspects related to the implementation of the SGP remain open: first, given its asymmetric nature, the SGP is unlikely to curb the tendency to run pro-cyclical fiscal policies in good times which would amplify the business cycle and overburden the single monetary policy; second, 'passive' coordination of national fiscal policies may not be sufficient to ensure an adequate fiscal stance at the national and Euro Area level; third, long-run sustainability issues are not adequately covered by the SGP and some of its provisions -- namely a constant deficit target over time -- may not be optimal from the point of view of intergenerational equity.

So far, the stability and convergence programmes which set the medium-term budgetary strategy of EU countries have provided useful guidelines for the assessment of the medium-term fiscal stance and the monitoring of budgetary developments. Nevertheless, they could be improved to address some of the open issues mentioned above. In particular, programmes could be broadened to cover the anticipated budgetary consequences of ageing under different reform scenarios and they could also make explicit the 'reaction function' of fiscal authorities to unexpected cyclical developments or windfall gains (such as the receipts due to the third generation mobile phone (UMTS) licences). Enhancing the credibility of the medium-term programmes in these ways would be a first step in lifting the uncertainties surrounding the implementation of the SGP.

Appendix. Budget balance and pension contributions: a simple partial equilibrium analysis

This appendix is intended to trace the evolution of the contribution rates under various fiscal policies under the assumption that the PAYG pension scheme remains unchanged excepted 'parametric' measures regarding the replacement and the contribution rates. These simulations illustrate the burden, which is to fall on future generations:

The budget deficit is defined as follows:

[D.sub.t] = -[S.sub.t] + [[beta].sub.t][w.sub.t][N.sub.t] - [[alpha].sub.t][L.sub.t][w.sub.t] + i[B.sub.t] (A1)

where:

i: nominal interest rate on public debt

[B.sub.t]: stock of public debt

[S.sub.t]: primary surplus excepted pension contribution and expenditure

[[alpha].sub.t]: contribution rate associated with the PAYG

[[beta].sub.t]: replacement rate of pension

[L.sub.t]: employment

[N.sub.t]: retired population (assumed equal to the population aged over 65)

[w.sub.t]: gross wage (per head)

Using lower case letters to indicate variables in proportion of GDP, the deficit becomes :

[d.sub.t] = -[s.sub.t] + [[omega].sub.t] [[lambda].sub.t]/[e.sub.t] [[beta].sub.t] - [[omega].sub.t][[alpha].sub.t] + i[b.sub.t] (A2)

where:

[e.sub.t] = [L.sub.t]/[WAP.sub.t]: employment rate

[WAP.sub.t]: working age population

[[lambda].sub.t]: old age dependency ratio = [N.sup.t]/[WAP.sub.t]=[e.sub.t][N.sub.t]/[L.sub.t]

[[omega].sub.t] (share of labour income in GDP) = [w.sub.t][L.sub.t]/[Y.sub.t]

Let us simply assume that s = i[b.sub.0] - [d.sub.0] since the PAYG system is roughly balanced at the beginning of the simulation period (2000). Then, when [d.sub.t] and [[beta].sub.t] are given, [[alpha].sub.t] can be calculated from (A2):

[[alpha].sub.t]=1/[[omega].sub.t][[-s.sub.t] + [[omega].sub.t][[lambda].sub.t]/[e.sub.t][[beta].sub.t]-[d.sub.t] + i[b.sub.t]] (A3)

In order to stimulate the future behaviour of the contribution rate [[alpha].sub.t], a number of assumptions on the value of the parameters have to be made: we assume i = 5.5 per cent, the share of labour income to GDP remains fixed at two-thirds over the projection period ([[omega].sub.t] - 2/3); the debt ratio evolves according to the familiar debt accumulation identity under a nominal growth rate of 3.5 per cent.

As a benchmark scenario, [beta] is held constant (unchanged policies scenario) and equal to 100 per cent, [22] the employment rate (e) varies and is taken from aggregated national projections. b(0) = 70 per cent (stock of public debt in 2000) and the long-term debt ratio (i.e. in 2100) is given in order to facilitate comparison between the various scenarios (this level is 30 per cent of GDP, consistent with 1 per cent of GDP long-term deficit). Finally, [[lambda].sub.t] is taken from Eurostat projections (Chart 6).

(*.) Directorate General for Economic and Financial Affairs, European Commission. The authors would like to thank Andre Sapir for useful discussions. The opinions expressed in this article belong to the authors and should not be attributed to the European Commission or its services.

NOTES

(1.) Buti and Sapir (1998) discuss the evolution of European budgetary policy as it moved toward the Pact.

(2.) As shown by von Hagen et al. (2000), the Maastricht criteria brought about a radical shift in budgetary behaviour over the 1990s, with a substantial weakening of the reaction of budget balances to the cyclical conditions of the economy.

(3.) 'Representative' examples are Hughes Hallett and McAdam (1999), Allsopp et al. (1999).

(4.) See, for example, Eichengreen (1996).

(5.) The SGP defines a 'severe' economic downturn as a decline in output by at least 2 per cent or, on an ad hoc basis, when the fall is by at least 0.75 per cent. Buti, Franco and Ongena (1997) find 24 such episodes for EU countries over the period 1960-96.

(6.) This implies that, as a benchmark, discretionary fiscal policy is not considered. This may be too harsh a requirement, especially in the case of sharp cyclical downturns. See third section.

(7.) Barrell and Pina (2000), who estimate a lower impact of automatic stabilisation using the NiGEM model, have challenged these results. They arrive at increases in output volatility in the range of only 5 to 18 per cent (II per cent for the Euro Area as a whole).

(8.) The increase in the standard deviation of the output gap due to the application of the SGP ranges between zero (Germany) and 23 per cent (France).

(9.) Interestingly, these results are not very sensitive to the share of myopic households.

(10.) The main caveat of this method is obviously its incompleteness: there is no feedback from consumption to output.

(11.) The analysis includes the 'escape clause' and the penalties associated with the violation of the SGP.

(12.) Allsopp and Vines (1996) stress the need for budgetary coordination in the case of severe symmetric shocks.

(13.) On the sanctions under the SGP, see Buti et al. (1998).

(14.) A small reaction of the budget balance to the cycle is found in a number of recent studies (Melitz, 1997,2000; Wyplosz, 1999). Typical bureaucratic reactions (spend when the money comes in, tighten when it runs out) lead to pro-cyclical behaviour in public spending. These authors find that the budgetary swings due to the business cycle may be as low as 10 per cent, compared to an average of 50 per cent as commonly estimated.

(15.) The SGP includes the so-called 'significant divergence clause' which requires the Commission to identify and draw the Council's attention to "actual or expected significant divergence of the budgetary position from the medium-term budgetary objective, or the adjustment path towards it, as set in the programme for the government surplus/deficit." In the event of a 'significant divergence' being identified, the Council can request the Member State concerned to take corrective action. The political will to apply such a clause remains to be tested. For a discussion, see European Commission (2000).

(16.) See, for example, the 1998 Code of conduct on the practical implementation of the SGP (European Commission, 1999), the report on economic policy coordination to the Helsinki European Council in December 1999 and the 1999 and 2000 Council Opinions on the stability and convergence programmes. Moreover, in their updated stability and convergence programmes, several Member States explicitly mention the need to pursue fiscal consolidation and reduce debt levels at a fast pace as a means for pre-empting the budgetary impact of ageing populations.

(17.) See Buti (2000) and, for a similar approach, Artis and Buti (2000).

(18.) Analogous conclusions would be reached had pension contributions been maintained fixed and the adjustment made on the replacement rates.

(19.) Reaching quantitative conclusions regarding the effects of various fiscal strategies on intergenerational equity would require using generational accounting models, which is well beyond the scope of this article.

(20.) The budgetary targets used in the graph are the so-called 'minimal benchmarks' computed by the European Commission. This only covers the cyclical safety margin but not the erratic component in the budget. See European Commission (1999).

(21.) In the calculation of the Member States' MCI, weights of 1 and 1/3, respectively have been used for the real interest rate and the real exchange rate. For the Euro Area weights of 1 and 1/6 have been used, reflecting its lower share of external trade.

(22.) The choice of a more realistic replacement ratio would only lead to homothetic results.

REFERENCES

Allsopp, C., McKibbin, W. and Vines, D. (1999), 'Fiscal consolidation in Europe: some empirical issues', in Hughes Hallett, A., Hutchison. M.M. and Hougaard Jensen, S.E. (eds), Fiscal Aspects of European Monetary Integration, Cambridge, Cambridge University Press.

Allsopp, C. and Vines, D. (1996), 'Fiscal policy and EMU', National Institute Economic Review, 158, pp. 91-107.

--(1998), 'The assessment macroeconomic policy after EMU', Oxford Review of Economic Policy, 14 (3), Autumn, pp. 1-23.

Artis, M.J. and Buti, M. (2000), "'Close to balance or in surplus" -- a policy maker's guide to the implementation of the Stability and Growth Pact', Journal of Common Market Studies, forthcoming.

Barrell, R. and Pina, A. (2000), 'How important are automatic stabilisers in Europe?', EUI Working Papers, ECO. No 2000/2.

Blanchard, O.J. (1985), 'Debts, deficits and finite horizons', Journal of Political Economy, 93 (2), April, pp. 223-47.

Bean, C. (1998), 'Discussion', Economic Policy, 26, pp. 104-7.

Buti, M. (2000), 'Policy issues -- links with the Stability and Growth Pact', proceedings of the Banca d'Italia seminar on Long-Term Sustainability of Public Finances, forthcoming.

Buti, M., Franco, D. and Ongena, H. (1997). 'Budgetary policies during recessions: retrospective application of the "Stability and Growth Pact" to the post-war period', Recherches Economiques de Louvain, 63 (4), pp. 321-66.

--(1998), 'Fiscal discipline and flexibility in EMU: the implementation of the Stability and Growth Pact', Oxford Review of Economic Policy, 14 (3), pp. 81-97.

Buti, M. and Sapir, A. (eds) (1998), Economic Policy in EMU -- A Study by the European Commission Services, Oxford University Press.

Dalsgaard, T. and de Serres, A. (1999), 'Estimating prudent budgetary margins for II EU countries: a simulated SVAR model approach', OECD Economics Department Working Papers no. 216.

Dury, K. and Pina, A, (2000), 'European fiscal policy after EMU: simulating the operation of the Stability Pact', EUI Working Papers, ECO. No 2000/3.

Eichengreen, B. (1996), 'Saving Europe's automatic stabilisers', National Institute Economic Review, 159, pp. 92-8.

Eichengreen, B. and Wyplosz, C. (1998), 'Stability pact - more than a minor nuisance?', Economic Policy, 26, pp. 65-114.

European Commission (1999), 'Budgetary surveillance in EMU: the new stability and convergence programmes', European Economy, Supplement A. no 2/3.

--(2000), 'Public Finances in EMU - 2000', European Economy Reports and Studies, 3.

Franco, D. and Munzi, T. (1997), 'Ageing and fiscal policies in the European Union', European Economy Reports and Studies, 4, reprinted in Buti, M., Franco, D. and Pench, L.R. (eds), The Welfare State in Europe - Challenges and Reforms, London, Edward Elgar.

Hughes Hallett and McAdam, P. (1999), 'Implications of the Stability and Growth Pact: why the "growth" element is important', in Hughes Hallett, A., Hutchison, M.M. and Hougaard Jensen, S.E. (eds), Fiscal Aspects of European Monetary Integration, Cambridge University Press.

IMF (1998), World Economic Outlook, October.

Martinot, B. (2000), 'Pacte de stabilite et efficacite de la politique budgetaire', Document de travail, Ministere de Economie, des Finances et de I'Industrie.

Melitz J. (1997), 'Some cross-country evidence about debt, deficits and the behaviour of monetary and fiscal authorities', Centre for Economic Policy Research discussion paper no. 1653.

--(2000), 'Some cross-country evidence about fiscal policy behaviour and consequences for EMU', mimeo.

OECD (1997), OECD Economic Outlook, December.

Van den Noord, P. (2000), 'The size and role of automatic fiscal stabilisers in the 1990s and beyond', OECD Economics Department Working Papers, no. 230.

von Hagen, J., Hughes Hallett, A. and Strauch, R. (2000), Budgetary Consolidation in EMU, a study for the European Commission, forthcoming.

Wyplosz, C. (1999), 'Economic policy coordination in EMU: strategies and institutions', ZEI Policy Paper, B11.