Unnoticed potential output revisions and their impact on the “stimulus/austerity debate”

Jérémie Cohen-Setton, Natacha Valla, 17 August 2010

a

A

The austerity-versus-stimulus debate is the summer’s hot topic – see for instance the recent Reinhart and Rogoff (2010) column on this site and the response from Paul Krugman (2010) on his New York Times blog. But do we really know what “fiscal austerity” means? How confident are we about its quantitative assessment?

Headline fiscal numbers are fairly uncontroversial; they are factual accounting numbers, and we know what they mean. But, we are treading on thinner ice when it comes to more structural measures of a nation’s fiscal stance.

In particular, empirical measures of so-called “potential output” have become absolutely instrumental in making macroeconomic and policy assessments. Although considerable progress has been made to estimate it, there is still a great deal of uncertainty as to the robustness of these estimates – especially during large business cycle turning points.

Our concerns are that:

  • massive revisions in potential output estimates have been applied for the 2008-2011 period;
  • these revisions have gone ahead without being thoroughly discussed in policy circles and made clear to markets and the general public;
  • the economic rationale for applying such revisions is questionable, and
  • yet these revisions are key for assessing the macroeconomic stance of different countries and whether they should engage in more or less policy stimulus.

We show that these revisions have a considerable impact on the implied fiscal stance – i.e. whether the policy measures taken imply a loosening or a tightening of budgetary policy – of European countries. We also show that behind these swings in potential GDP estimates lie swings in the time-varying non-accelerating inflation rate of unemployment (NAIRU). While we would have expected a lot of controversy around the NAIRU revisions given their importance in making macroeconomic and policy assessments in the current environment, instead we have been amazed by how little discussion there has been on this issue.

Massive revisions: Has the crisis really shaved 1% off potential growth?

“Potential output” (often defined as the non-inflationary output growth rate, and put more simply, a more-or-less sophisticated proxy for underlying “trend” growth) is a concept that has become instrumental in assessing macroeconomic policies and trends in the EU and elsewhere. By definition, unobservable potential output needs to be inferred or estimated. Nowadays, estimation methods are plenty, and considerable progress has been made in the field (see for example the various task forces put in place by the Ecofin under the auspices of the Economic Policy Committee).

This is all a lot more than a mere conceptual exercise, as this theoretical and empirical effort is subsequently put at work in a wealth of policy evaluation exercises, including the implementation of the Stability and Growth Pact provisions and the European fiscal surveillance framework.

In this process, one would expect to see measures of potential output displaying a fair amount of smoothness over time. On the one hand, the underlying strength of an economy can arguably be hit by, say, a war, or a prolonged period of negative investment – hence a possible time-variation in potential output estimates. On the other hand, it is difficult to see why, from one year to the other, views on potential output could be revised up and down to significant amounts.

Yet, revisions to the “potential” rate of GDP growth that underlies recent vintages of macroeconomic forecasts have been massive, and went more or less unnoticed beyond expert circles. The substantial downward revisions applied by the European Commission to potential growth can be seen in Figure 1. Between 2008 and 2011, potential growth forecasts for the Eurozone as a whole have been reduced from 2.0% to 1.3% for 2008, 2.0% to 0.8% for 2009, 1.8% to 0.8% for 2010, and 1.9% to 1.0% for 2011. In other words, according to these estimates, the recession would have shaved 1% off potential growth. The same pattern can be observed for core and periphery countries, although with significant differences across countries.

Figure 1. Substantial downward revisions to 2005-2011 potential output growth

Source: European Commission, Spring 2008 and 2010 forecasts. The chart measure the difference between potential output as estimated by the EC in 2010 and 2008. A negative reading corresponds to a negative revision.

Behind it all: Diverging views on unemployment

The factors behind these dramatic potential output adjustments can be many since the production function approach relates potential output to the capital stock, the potential labour supply – itself a function of the working age population and the NAIRU – and the permanent component of Total Factor Productivity as explained in more detail in the methodology box at the end.

What Figures 2 and 3 show however, is that for the countries most affected by potential GDP revisions, such as Greece, Ireland, and Spain, the decrease in potential GDP estimates comes almost entirely from a decrease in the labour contribution to potential growth due to a huge increase in the NAIRU.

Figure 2. Decomposition of the revisions to potential GDP estimates: Spain

Figure 3. Substantial revisions in NAWRU are driving the change in potential GDP estimates

Source: European Commission, Spring 2008 and 2010 forecasts; our calculations. NAWRU stands for “Non-accelerating wage rate of unemployment”.

What should we think of these results? The least that we can say is that they should not remain unnoticed as we expect broad disagreements as to whether it is reasonable to qualify the bulk of the increase in the headline unemployment rate as a fundamental increase in the NAIRU.

Indeed, while we expect a broad consensus on the fear that some elements of joblessness could become entrenched (see the latest OECD Economic Outlook) or that the crisis might impact medium term potential growth even through other channels (see the EC 2009 QREA), the economic rationale for significantly revising NAIRU estimates during large business-cycle turning points appears disputable.

Policy assessment is not robust to volatile views on potential output and unemployment

Staying agnostic about whether or not these revisions are warranted, we now highlight their implications for policy assessment. In particular, we look at the impact of the revisions of potential GDP estimates on the implied fiscal stance for key European countries.

As noted by Larch and Turrin (2009), the prominence and use of structural measures of fiscal imbalances – i.e., deficits adjusted for cyclical effects – in policymaking has gained ground. Before the Stability and Growth Pact was revised in 2005, these measures had mostly been used as an analytical tool to better analyse the fiscal situation of the EU Member States. With the reform of the Pact, they have moved to centre stage of the EU fiscal surveillance framework.

There is a clear rationale for using cyclically-adjusted measures of public finances. If an economy is characterised by a potential output growth of 2%, an actual output growth of 4% implies that nominal headline fiscal balances look much better than they would otherwise – had actual GDP growth been in line with potential. As a result, it makes sense to use this gap between actual and potential output to filter out the “fake” cyclical improvements of public accounts in good times and their deterioration in bad times. Yet, if the 2% potential estimate turns out to be erroneous, and the truth were closer to 1%, fiscal balances adjusted for the cycle on the basis of the overly optimistic “2%” estimate would understate perceived surpluses in good times and the deficits in bad times.

But the kind of revisions in potential GDP introduced by the European Commission has made the assessment of fiscal policies especially challenging. Reconstructing counterfactuals of structural fiscal balances using 2008 beliefs on potential growth instead of the new vintage of potential GDP estimates, we show that potential GDP revisions can flip around fiscal policy assessment and be misleading about the actual degree of fiscal tightening.

Figure 4. “Structural” fiscal stance of core European countries
a) 2010

b) 2011

Source: European Commission's assessment of Stability Programmes, Tables from Annex 2 of the code of conduct; GS Global ECS Research. Note: The GS counterfactuals are computed using the EC Spring 2008 vintage forecasts for potential GDP and the current EC Spring 2010 forecasts for real GDP.  Note: These measures correspond to the “structural” fiscal stance as estimated based on the fiscal plans presented by EU countries in their Stability Programmes, usually around January 2010. As a consequence, they do not account for the new measures announced following the euro-zone crisis. For an evaluation of the new austerity measures and their potential impact on GDP growth, see our views published in Goldman Sachs’ European Weekly Analysts 10/20 and 10/24.

Germany and the Netherlands tougher; France, Italy, UK milder than we think

As a matter of fact, we end up with fiscal stance measures that sound closer to what we know about the policies that these countries were actually planning at that time. Assuming that the Commission’s Spring 2008 potential output estimates were correct, the 2010 fiscal stance of both Germany and the Netherlands would be less (almost by half!) expansionary than reported by the Stability Programmes. By contrast, the fiscal hand would look much friendlier in France, Italy and the UK than implied by the Stability Programmes. Likewise for 2011 – and this is perhaps more crucial to the current policy debate – France, Italy and the UK look as though they will “consolidate less than we think”.

Conclusion

What does it all tell us? First, revisions to potential output or NAIRU estimates are not benign. They translate into substantial judgement shifts about macroeconomic policies. Second, there should be more transparency on when, why, and how potential GDP growth revisions have been applied. After all, these revisions are put at work in a wealth of policy evaluation exercises, including the implementation of the Stability and Growth Pact provisions and the European fiscal surveillance framework. As a result, citizens have the right to be educated on these issues. Third, we believe that a confidence band – or, why not, a “fan chart” – should be highlighted not only around input variables, but also around fiscal structural balances themselves. This way we would avoid “bad surprises” in the future, when markets and the public at large suddenly discover that our beliefs around actual “austerity” are surrounded by massive uncertainly. The recently created Office for Budget Responsibility in the UK is a lot more advanced in this matter – why not equip European authorities with fans too?

Appendix

Box 1. How is potential GDP estimated under the production function approach?

Instead of taking a purely statistical methodology exploiting time series properties, the production function approach used to estimate potential output makes assumptions based on economic theory.

In particular, it relates potential output to the capital stock, hours worked, the Non-Accelerating Inflation (Wage) Rate of Unemployment (NAIRU or NAWRU) and the permanent component of Total Factor Productivity (TFP) as specified below.

Measuring potential output using a production function approach

Source: Denis and al. (2006)

References

Denis, Cécile, Daniel Grenouilleau, Kieran Mc Morrow and Röger Werner (2006), “Calculating potential growth rates and output gaps - A revised production function approach”, Economic Papers 247, DG ECFIN, European Commission

European Commission (2009), “Focus: The impact of the economic and financial crisis on potential growth”, Quarterly Report on the Euro Area, 8(2).

Goldman Sachs European Weekly Analyst 10/20, “Austerity and reforms: It would take a lot to derail the recovery”.

Goldman Sachs European Weekly Analyst 10/22, “Measuring fiscal adjustments: Behind the scenes”.

Goldman Sachs European Weekly Analyst 10/24, “The European policy response: A summary overview”.

Krugman, Paul (2010), “Reinhart and Rogoff are confusing me”, nytimes.com, 11 August.

Larch, Martin and Alessandro Turrini (2009), “The cyclically-adjusted budget balance in EU fiscal policy making: A love at first sight turned into a mature relationship”, Economic Papers 374, DG ECFIN, European Commission

OECD Employment Outlook (2010).

Reinhart, Carmen M and Kenneth Rogoff (2010), “Debt and growth revisited”, VoxEU.org, 11 August.

Topics: Macroeconomic policy
Tags: Eurozone crisis, Fiscal crisis, fiscal policy

Jérémie Cohen-Setton

PhD candidate in Economics at U.C. Berkeley

Natacha Valla

Executive Director at Goldman Sachs Global Economic Research

Vox Talks