To what extent does hitting the fiscal brakes slow down the economic car? This debate, which has been going on for many years, has come back to the forefront with the new agreement between Greece and its creditors because it targets a significant swing in the primary balance (the government budget balance excluding interest charges), from a deficit of about 1% of GDP to a surplus of 3,5% by 2018 (European Commission data).
The following chart shows the relationship for Eurozone countries between the cumulative change in the primary balance and the cumulative change in real GDP. The latter, represented on the vertical axis, shows the ratio of real GDP for 2015 (based on IMF forecasts) and its level in 2008. A number above 1 implies that on average growth has been positive, though it doesn’t show the profile of GDP over this period. The primary balance is shown on the horizontal axis. To be more precise it shows the cumulative change in the cyclically adjusted primary balance since the year in which the low point was reached (so when the deficit reached a peak) and 2015 (for which the IMF forecasts are used). For all countries, the number is positive which means that the primary balance has improved. The cyclically adjusted balance is used to show the discretionary effort undertaken by governments to rein in their budget deficit.
Different comments can be made:
- As a caveat, the number of observations is rather limited, so this influences the quality of the regression. Moreover one may wonder whether Greece should be treated as a special case (this would influence the slope of the regression line)
- The intercept of the regression line shows that in the absence of any change in the cyclically adjusted primary balance, the cumulative GDP growth since 2008 would have been 5,6%
- In many countries the cumulative adjustment effort was below 4% of GDP, implying an, on average, limited annual effort (although the cumulative effect makes any additional effort harder and harder to achieve)
- On the question whether fiscal adjustment weighs on growth, the answer seems clear. The coefficient for the change in the cyclically adjusted primary balance and the change in GDP is 0,0104. This means that an improvement of the primary balance of 1 percent of GDP lowers the cumulative GDP growth with about 1 percentage point. The t-statistic for this coefficient is -2,77, so it’s quite significant. The drag of fiscal adjustment on growth may temporarily even be bigger considering that in this chart period averages are used. This result does not mean that fiscal adjustment should be avoided. It does suggest that the negative impact on growth should be taken into account in the design of fiscal, monetary and even structural policy.
One response to “Fiscal austerity and growth”
Dear William de Vijlder,
thank you for this. Unfortunately, you paint a rather simplistic picture and there are at least several problems with it. First, The correlation, which you are referring to, may be caused by a reversed relationship. It may be – and is indeed likely – that decreases in GDP growth led to austerity measures and not the other way around. For instance, in Spain the origin of the crisis was a unsustainably high growth rate that led to a subsequent drop in growth and an adoption of austerity measures afterwards. Second and related, a problem with the data is a regression the mean effect. The problem is that the start point of 2008 was characterised by artificially high GDP growth levels driven by e.g. a housing bubbles as in Spain. Thus the drop in GDP growth is a correction of this and not necessarily due to austerity. Third, the time variable is not only censored to the left but also to the right. Taking 2015 as a cut-off obscures long term effects. Interestingly, Ireland – the first country to adopt austerity measures – is back in the growth zone. Finally, this is only a binary regression. In order to have more confidence in the results we would need some control variable of the pre-crisis economic structure.
In sum, I am not convinced. Sorry.
Best
JB