Trailing with a margin: Official macro forecasts in EU fiscal surveillance

Budgetary planning in the euro area involves a codified interaction between national governments and the European Commission where economic and fiscal forecasts play a crucial role. This column marks the first extension of the European Fiscal Board’s Forecast Tracker. Next to the European Commission’s biannual forecast of key macro variables, the extension adds those underpinning the projections of the draft budgetary plans of euro area member states, and allows for a more detailed analysis of fiscal policymaking in the EU.

Macroeconomic forecasts serve a multitude of purposes. The most obvious but also naïve one is to predict as accurately as possible the future course of a given variable of interest (say, GDP or government revenues). Another one is more strategic, namely, to contribute to a collective process where forecasts are political statements and straying from the pack can draw heavy criticism, including that of being an armchair economist especially when vindicated in hindsight.

Forecasts underpinning budgetary plans of euro area member states are a clear case in point. While fiscal policy remains in the hands of national authorities, member states are expected to take economic policy as a matter of common concern and to subject their budgetary plans to outside scrutiny. Specifically, euro area member states are legally required to send their draft budgetary plans (DBPs) to the European Commission for an assessment, who, to that end, produces its own economic forecasts as benchmark. The underlying objective of this exercise is to expose potentially unrealistic national projections, which may affect fiscal outcomes and produce undesired spillover effects (e.g. Jonung and Larch 2006). However, since no one has a crystal ball, how far can and will an assessor depart from the ‘original’?

Any reader who at some point of his or her professional career was involved in forecasting will be familiar with the predicament: no matter how convinced one may be of one’s own forecast, it is likely to be rejected in real time if it departs too much from the prevailing consensus. And to be proven right once outcomes are revealed is of little consolation as the forecasting cycle has moved on to the next year. In short, when in action, looking back is not an option.

However, from time to time, it is worth to take a step back and reflect on past forecast performance. This column summarises the results of a basic analysis that contrasts two sets of projections – those underpinning the DBPs of euro area member states and the ones of the European Commission. It builds on and extends the Forecast Tracker, a tool launched by the Secretariat of the European Fiscal Board (EFB) back in June this year (Larch and Barbera Mazzola 2025). The original version covered one-year-ahead projections of key macro variables of the biannual economic forecast of the European Commission – spring and autumn.The extension incorporates the same macro variables from the budgetary plans of member states who adopted the single currency.

The DBPs are highly templated, and their content is defined by law. They include a macroeconomic scenario that serves as basis for the government’s fiscal outlook and policy. The added value of extending the Forecast Tracker to the projections underpinning DBPs is straightforward: it allows users to explore the relative forecast performance of national authorities and the European Commission. The two parties usually dialogue while developing their respective views of the future, and the resulting forecasts are shaped by this process.

A first look at the interaction between national and European Commission forecasts

We start our analysis by focusing on euro area member states that have submitted DBPs every year since 2013. The DBP series are shorter than those of the European Commission forecast, because the requirement to submit DBPs to Brussels prior to the adoption by national parliaments was only introduced at a later stage when, in the wake of the global financial and economic crisis, the EU decided to step up the coordination of member states’ fiscal policies. Nonetheless, a comparison of the two series for the overlapping years illustrates the potential of the extended Forecast Tracker inviting further research.

A priori, errors on economic activity and government revenues should move together: the latter are typically projected as a function of the former, since forecasters make assumptions regarding the elasticity of revenue to nominal GDP. On the other hand, errors on expenditure can be driven by different dynamics, in particular the political process. While real GDP growth might attract more attention in the public debate when forecasts are published, nominal GDP growth is more relevant for fiscal surveillance, as budgetary components and the debt ratio very much depend on price developments.

The European Commission tends to trail national authorities

Visual inspection (Figure 1) already offers an interesting insight: the European Commission tends to follow the inclination of euro-area governments, i.e. when the governments’ forecasts of economic activity are optimistic (or pessimistic) those of the European Commission will err in the same direction. In terms of magnitude though, the European Commission tends to be more moderate. Germany seems to be an exception: on average, the European Commission seems to be more optimistic on nominal GDP growth than the German authorities.

Apart from testifying to the European Commission’s surveillance role, these results may also be affected by the timeline of the surveillance processes. Both sets of forecasts are published in the autumn of each year. The national authorities start their preparation in the summer and are likely to have closed their forecasting process by the mid-October DBP deadline, while the European Commission’s cut-off date is a bit later – usually at the end of the same month, i.e. its information set can be more complete.

Errors on the budget balance show a less straightforward relationship between the two sets of forecasts than in the case of economic activity. For the cluster of member states where forecasts tend to overestimate deficits, the difference between DBPs and the European Commission is, on average, smaller compared to the group of countries for which forecasts underestimate the size of the budgetary shortfall (see Figure 1, Panel b). This could be intuitive as optimism is arguably easier to sell than pessimism.

Zooming in on the drivers of the budget balance errors, the forecasts of both the European Commission and national authorities show optimism on expenditure, i.e. expenditure tends to turn out higher than projected. Again, the degree of optimism is more pronounced in the DBPs. For government revenues, instead, there are two clusters where both forecasts are either optimistic or pessimistic. Italy and Spain are notable exceptions: national authorities tend towards moderate optimism while the European Commission tends to be on the cautious side.

Figure 1 Forecast errors of the European Commission relative to members states’ draft budgetary plans

Notes: Each panel shows the forecast error of a specific variable by the European Commission (x-axis) and in the DBP (y-axis). A positive forecast error is associated with pessimism (the outturn is better than expected) and a negative forecast error with optimism (the outturn falls short of projections), except for expenditure where the opposite holds. A country code above (below) the 45° line means that, for a given error of the national authorities, the European Commission was more pessimistic (optimistic) – again, the opposite for expenditure.
Source: European Commission, DBPs.

Beyond visual inspection

Table 1 shows the results of basic statistical tests. Starting with economic activity, the numbers confirm that, on average, both national authorities and the European Commission show a tendency towards optimism in forecasting GDP growth, although differences are generally not statistically significant and the European Commission’s projections tend to be comparatively more accurate. Among the countries that submitted DBPs since 2014, only Italy and Malta are significantly more optimistic than the European Commission in terms of both real and nominal GDP growth.  A few other countries – Estonia, France, Luxembourg, the Netherlands, and Slovenia – exhibit statistically significant differences in forecasting real GDP growth.

The statistical tests also confirm how national authorities tend to be more optimistic than the European Commission about the nominal budget balance in percent of GDP, forecasting – on average – smaller deficits. For Belgium, Italy, Malta, and Spain the optimism of the DBPs seems to be driven by more sanguine revenue projections compared to the European Commission, while both sets of forecast underestimate growth in expenditure. In the case of France, however, the authorities’ optimism on the budget balance seems to originate in expenditure projections that systematically underestimate outturns. 

Despite Germany’s higher degree of pessimism in forecasting revenues, its average forecast errors for the budget balance are not significantly different from those of the European Commission – a result shared with Slovakia. The attitude in Germany towards forecasting government revenues is also discussed in EFB (2020). The Working Party on Tax Revenue Estimates (AKS), one of several independent assessors in the German fiscal framework, responsible for producing tax revenue projections, has been observed to have “systematically underestimated tax revenue since 2009”. One driver for this persistent underestimation could be positive surprises in growth, though the differences reported in Table 1 are not statistically significant at standard levels of confidence. Another possible explanation are conservative elasticities of revenue to GDP, possibly motivated by the political economy of national budget negotiations between the German ministry of finance and line ministries.

A final point: to check our results are not affected by the impact of the COVID pandemic, an exogenous and truly unpredictable event, we repeated the statistical tests removing the years 2020 and 2021, from the sample. The results do not change materially. 

Table 1 Comparison of forecast errors between the European Commission and DBPs

Notes: The table shows mean forecast errors of the Commission minus mean forecast errors of the DBPs for each member state over the period 2014 2024. Positive (negative) values show higher (lower) optimism by the national authorities, except for expenditure where the signs are the opposite. *, **, *** indicate a level of statistical significance of 0.1, 0.05, 0.01.
Source: European Commission, DBPs.

Conclusions

The extension of the Forecast Tracker to economic and fiscal projections of euro-area member states sheds light on the relative forecast performance of key actors in EU fiscal surveillance. Our preliminary findings reassert the importance of monitoring and regularly assessing budgetary forecasts. Our analysis shows how the European Commission and national authorities tend to err in the same direction. At the same time, the errors of the European Commission are, on average, smaller.

While the quality of forecasts always plays a role, the wide prevalence of optimistic expenditure projections – i.e. expenditure turning out higher than projected – raises concerns for the implementation of the revised EU economic governance framework. The revised framework revolves around a multiannual net expenditure path over four years or more predicated on member states’ economic growth projections for the medium term. Our findings also speak to the need to enhance national frameworks geared towards instilling prudence in budgetary planning, such as a wider involvement of independent fiscal institutions in the budgetary process.

To add more perspective over medium term budgeting, the next evolution of the Forecast Tracker will incorporate the projections underpinning the stability and convergence programmes of all EU member states. Required under the old EU fiscal rules, these documents include information to assess whether member states were on track towards reaching their medium-term budgetary objectives. 

Source : VOXeu