Europe’s fiscal policy credibility problem

Roel Beetsma, Benjamin Bluhm, Massimo Giuliodori, Peter Wierts, 1 July 2011



On 29 September 2010 the European Commission presented six proposals for reinforcing economic governance and the European Council reached a political agreement on the package in March 2011. Negotiations with the European Parliament have started. The Council has set June 2011 as the target date for adoption. In principle, we should be approaching the moment that the final package will be made public. How should we evaluate the outcome?

Fiscal revisions

The answer depends on whether the reform addresses the root causes of Europe’s fiscal problems. Fiscal positions have deteriorated through:

  • The unwinding of macroeconomic imbalances and the outfall from the crisis;
  • Misreporting and weaknesses in national budgeting processes; and
  • Insufficient incentives to comply with the EU fiscal rules.

The relevance of each of these factors differs across countries.

Research by De Castro et al. (2011) provides useful insights into the effects of the last two on fiscal outcomes. They explore the properties of subsequent revisions in the budget balances of a given year. The quality of the first-release figures is important, because these figures are an input for the next budget. Moreover, fiscal surveillance is based on these figures. For example, they may indicate that fiscal policy is on an unsustainable course and, hence, enable policymakers to undertake timely corrections. Results show that preliminary data releases are biased estimators of final data, with later vintages showing larger deficits. Eurostat’s decisions on data revisions explain a significant share of the bias.

In Beetsma et al. (2011) we confirm and complement these findings. We analyse the credibility of the whole budget process from planning to implementation and ex-post control. An implementation error arises when fiscal outcomes at the implementation stage (in real time) differ from fiscal plans. A revision error arises when ex-post data differ significantly from real-time data. Our conceptual framework stresses that ministries of finance control the production of first-release figures, and may have an incentive to be over-optimistic at this stage. In most countries, the national statistical office is responsible for compiling the budget balance data once the fiscal year is over. Given the independent position of statistical offices in most countries, we expect ex-post data to be free from political distortions. We expect that fiscal transparency reduces the leeway for misreporting and, therefore, reduces the size of fiscal slippages. Moreover, tighter national fiscal rules may serve as a self-enforced commitment device implying that the Ministry of Finance takes more responsibility for “prudent” fiscal outcomes, and, hence, smaller revision errors.

Figure 1 shows planned budgets and first-release and ex-post budgetary outcomes for each country and each year in our sample.1 Both the implementation errors and the revision errors are often substantial. Moreover, there is no obvious visible difference in their average size. Some specific cases stand out from the graphs. First, we see that for Greece, in all but two years the first-release balance falls short of the planned balance, while the ex-post balance is always lower and sometimes substantially lower than the first-release balance. Second, large negative spikes occur for Austria and Belgium in 2004 and 2005, respectively. In both cases, transactions with the national railway company were reclassified afterwards by Eurostat as deficit-increasing measures.

Figure 1. Planned, first-release and ex-post balances

Table 1 shows the average sizes of the implementation and revision errors in our dataset. The last column (total error) indicates that ex-post figures for the budget balance are on average 0.5% GDP lower than originally planned. While plans are too optimistic relative to the first-release outcomes (average implementation error is -0.17% GDP), the latter in turn are too optimistic relative to the eventual outcomes (average revision error is -0.34% GDP). Moreover, the largest part of the revision error is on the revenue side of the budget. This is in line with our conjecture that the margin for discretion in real-time fiscal data may be larger for revenues than for expenditure. During the fiscal year, the Ministry of Finance has a direct control over revenue projections, while expenditure estimations also depend on input from the spending ministries. Its margin for the strategic use of revenue projections may also be larger since revenue developments are endogenous to the economic cycle, and depend on seasonal patterns. Expenditure, on the other hand, is more under the direct control of the spending ministries. Finally, the use of accrual, rather than cash accounting provides further room for deliberate over-optimism in the first-release revenues data.

Table 1. Implementation and revision errors


Implementation Error

Revision Error

Total Error













Primary exp.




Notes: Errors are expressed in percent of GDP; Implementation error is based on first-release data minus plans; revision error is based on ex-post data minus first-release data; total error is based on ex-post data minus plans. * = significance at the 10% level; ** = significance at the 5% level; *** = significance at the 1% level.

Policy discussion

What can be done to avoid systematic negative fiscal surprises in the future? Our regressions shed some light on the variables that cause fiscal slippages. We find that an improvement in the quality of institutions, whether measured by the tightness of national fiscal rules, the medium-term budgetary framework or the degree of budgetary transparency, reduces the degree of optimism at the first-release stage and makes first-release figures more informative about the eventual outcomes. These results support the European Commission’s (2010) proposal to specify minimum requirements for national budgetary frameworks. Also our findings on the role of enhanced transparency support the European Commission (2010). Its proposal to require member states to integrate all extra-budgetary government transactions into the regular budgetary process and to report information on contingent liabilities with potentially large fiscal impact could further help to limit the scope for over-optimism in the implementation stage. Moreover, amendment proposals by the European Parliament (2010, p. 19-20, and p. 35) would provide a more general legal basis for the role of national budgetary frameworks in improving the implementation of fiscal policy at the national level. Its proposals on national ownership require Eurozone countries to incorporate the objectives of the Stability and Growth Pact into national law and national budgetary frameworks to ensure compliance with these objectives. Moreover, the European Council (March 2011 Conclusions, p. 10) concluded that: “Euro area Member States commit to translating EU fiscal rules as set out in the Stability and Growth Pact into national legislation. Member States will retain the choice of the specific national legal vehicle to be used, but will make sure that it has a sufficiently strong binding and durable nature (e.g. constitution or framework law).”

While the changes proposed by the European Commission serve a wider purpose than improving the accuracy of first-release macro and fiscal data, a more direct way to achieve the latter may be to transfer the responsibility for producing these data to an independent institution. However, to achieve this, both political and practical obstacles may have to be overcome. The main practical complication is that the Ministry of Finance always needs to be relied upon to provide relevant real-time data.

Our analysis also points to some recommendations regarding the conduct of fiscal surveillance at the European level. First, surveillance should focus more on systematic patterns in errors and their components. With first-release and ex-post data becoming available over longer horizons, the scope for such an approach is increasing. Second, by comparing fiscal data across countries, we can extract more accurate signals whether implementation and revision errors may be justified or not. Third, with Stability and Growth Pact surveillance based on first-release figures, there is an incentive for governments to bias these figures, which makes them less useful for fiscal surveillance. Our results suggest that this trade-off can be ameliorated by combining surveillance at the European level with enhanced fiscal transparency at the national level. In particular, judgement of first-release figures should be on a sufficiently comprehensive basis taking proper account of stock-flow adjustments and the risks associated with off-balance items. Finally, as our results show, revision errors in the budget balance may mask substantial and partially offsetting revision errors in revenues and spending. Therefore, it is important for fiscal surveillance to also focus on the individual components of the budget balance.


De Castro, F. JJ Pérez, and M Rodríguez (2011), Fiscal Data Revisions in Europe, mimeo, Banco de España / European Central Bank.

Beetsma, R, M Giuliodori, and P Wierts (2009), “Planning to Cheat: EU Fiscal Policy in Real Time”, Economic Policy,24(60):753-804.

Beetsma, R, B Bluhm, M Giuliodori, and P Wierts (2011), “From First-Release to Ex-Post Fiscal Data: Exploring the Sources of Revision Errors in the EU”, CEPR Discussion Paper  8413.

European Commission (2010). “Proposal for a Council Directive on requirements for budgetary frameworks of the Member States”.

European Parliament (2010/0280), DRAFT REPORT on the Proposal for a Regulation of the European Parliament and of the Council Amending Regulation (EC) No 1466/97 on the Strengthening of the Surveillance of Budgetary Positions and the Surveillance and Coordination of Economic Policies, Committee on Economic and Monetary Affairs. Rapporteur: C. Wortmann-Kool.

European Council (2011), “Conclusions”, 24/25 March.

1 Planning and first-release data are from the EU Stability and Convergence Programs (SCPs) submitted in the years 1998-2008; ex-post data are from the November 2010 AMECO dataset of the European Commission.

Topics: EU policies, Macroeconomic policy
Tags: Eurozone crisis, Greece, Ireland, Portugal, Spain

Professor of Pension Economics and Vice-Dean of the Faculty of Economics and Business, University of Amsterdam

Benjamin Bluhm

PhD student, GSEFM, Goethe-University Frankfurt

Massimo Giuliodori

Associate Professor, University of Amsterdam

Peter Wierts

Senior Economist in the Dutch Central Bank

Most Read