Don’t be fooled by public consumption figures

As the old adage goes, there are three types of lies: lies, damn lies, and statistics. In Covid-19 times, this is all the more true, because it has become a lot more difficult to collect economic data. GDP accounting is complicated anyway and one of the more contentious areas concerns the accounting for public consumption (think of expenditures for defence, education, public administration etc). Normally, GDP is computed at market prices. But for plenty of the services comprised in public consumption, there are no market prices. That is the reason why traditionally, they are accounted at cost. This brings some other problems. e.g. does a wage increase of public workers imply an increase in real GDP or is that just a price increase? Traditionally, statistical offices devised some rules of thumb to repartition cost increases between real growth (as some of the wage increase could reflect productivity gains) and inflation.

But how to account for lockdowns, when a number of public workers are actually paid, but unable to work? Should it be considered that the production of public services has followed its normal course because public wages have been paid? Or should it take into account the decline in activity and the fact that some public workers were unable to work? Enter Eurostat with an elaborate note. To cut a long story short: “When hours worked are used as the indicator for labour input, changes in working time because of COVID-19 should result in volume changes in the output of non-market services. But if indicators like hours paid or full time equivalent employees are used, short term changes in activity may be not be captured. In this case, suitable adjustments to the indicators should be made, in order to better reflect the hours actually worked in the period concerned”.

The question is, which countries have bothered to make precise estimate of the change in the volume of public services. To be sure, there have been various types of lockdown and some governments may have decided to spend more or less on public services. But the interesting thing to focus on is the difference between nominal and real growth of public consumption. Barring a strong wage increase, growth rates should be of the same magnitude, the difference being the public consumption deflator. Looking at nominal and real growth of public consumption, some countries stand out: France, the UK and to a lesser extent Portugal. In France, the statistical office Insee estimated that 25% of public workers, health workers excluded, were unable to work during the lockdown. This has led to a 13.2% reduction in public spending in the first half of 2020. In the UK, the Office of National Statistics made output estimates for public services, resulting in a very big real decline in public consumption (-17.5%). Conversely, it seems that in other countries, such as Germany or Spain, the output of public services was considered to have been maintained as public wages continued to be paid.

It is interesting to note that Belgium and the Netherlands have also experienced a sharp decline in public spending. But for these two countries, the decline is both nominal and real, so it has nothing to do with the output of public workers. In fact, both countries have accounted in their public expenditure data for the sharp decline in the number of non-emergency medical treatments performed during lockdown.

Of course, these differences in accounting for public services has also had an impact on GDP growth, with the difference between nominal and real GDP growth clearly much bigger in the UK, France and Portugal than in other European countries. The real GDP of these three countries is therefore calculated less favourably than the GDP of other European countries, and the recession in the first half of 2020 appears to be artificially deeper in France, the United Kingdom and Portugal than in the rest of Europe.

To measure the impact of these methodological differences, one can recalculate the evolution of GDP in the first half of 2020 by maintaining public expenditure at its level at the end of 2019 and compare this adjusted GDP with the observed GDP (see table). It appears, for example, that the French recession was 3.2 percentage points deeper than it would have been if INSEE had calculated the contribution of public expenditure to GDP as in Germany, Spain or Italy.

The upshot is that with the end of the confinement and public services gradually returning to normal, the effect will be reversed and their GDP will be boosted by strong real growth in public consumption in 2H20. France could therefore see its GDP boosted by 3.2ppt in the second part of the year, while the boost will be 3.4ppt for the United Kingdom and 0.7ppt for Portugal. Belgium and the Netherlands could also see their GDP topped up by an increase in public spending as non-emergency medical treatment resumes at a similar pace to that before the crisis. Two of the other laggards, namely Italy and Spain, won’t get this relief in terms of growth figures. This is especially true for Spain, which has been hit by another breakout of Covid-19 infections and has therefore suffered from a very bad tourist season, putting the country at risk of being one of the worst performing economies in the second half of the year.