This month saw the release of Q2 GDP data for most eurozone countries as well as the UK and the US. The picture was one of record contractions in output in most countries. Only in a few cases shown in Figure 1, such as Lithuania and Finland, could the annual fall in Q2 GDP be kept at or below 5%. Double-digit falls were the norm with 13 countries out of the sample of 22 showing annual contractions of more than -10% in the second quarter. Spain and the UK recorded the strongest declines on this measure with contractions of -22.1% and -21.7%, respectively.
The y-axis in Figure 1 measures the total confirmed deaths due to COVID-19 per million inhabitants as of 9 August. It is quite clear that countries which suffered the strongest falls in GDP have also seen some of the highest death rates from COVID-19. A simple correlation analysis between the annual rate of change in Q2 GDP and the death rate yields a moderately strong, negative relationship of -62%, suggesting that the higher the death rate of a country the worse its Q2 economic performance.
Countries to the top of the trend line have a relatively high death rate for a given level of economic contraction. Countries below the line, on the other hand, have relatively lower death rates given the extent of the fall in their GDP.
Much has been written in the press about the UK’s rather bad performance on both dimensions – the country has the second-highest death rate behind only Belgium and the second-worst economic contraction behind Spain. Out of necessity and after a bit of back and forth, health workers and officials have come to some internationally recognised standards about how to count Covid-19 victims in order to make figures comparable across countries. But how sure can we be that international measures of GDP are equally comparable?
Measuring economic output in a sensible, consistent and internationally comparable way is actually no small feat. And things get even more complicated when we are trying to strip out the effects of price movements. An interesting bit of the UK’s Q2 GDP release was the fact that according to the ONS the “the volume of government activity fell while at the same time government expenditure increased in nominal terms.” In other words, real government consumption, e.g. in the form of services delivered in schools and hospitals, fell while nominal expenditure rose – leading to a jump in the government consumption deflator by 32.7%.
How credible are these estimates? A look across the pond might put things into perspective. In the US, the federal government contributed positively to growth in Q2 more than offsetting negative contributions from state and local governments. In the UK meanwhile, government consumption fell by a whopping 14% over the second quarter as a result of widespread school closures and people drastically reducing their consumption of non-COVID-19 related health services. Of course, US public sector consumption is not entirely comparable to the UK not least because of the role of the NHS, but is such a large discrepancy in the growth contribution of the public sector – slightly positive in the US, highly negative in the UK – plausible during a pandemic that affected countries in quite comparable ways? The treatment of public sector output and productivity in the national accounts has been subject to much debate over the years with the Atkinson Review from 2005 still being one of the most influential pieces of work on the subject, arguing for better measuring quality improvements in public services.
At the very least, this gap between countries in terms of public sector growth should serve as a reminder that GDP estimates should be taken with a grain of salt. And if schools reopen as planned and people start to see their GPs again for non-COVID-19 reasons, don’t be surprised when the UK public sector emerges as an unlikely driver of the economic recovery in Q3.
For more information, please contact:
Contact: Kay Daniel Neufeld firstname.lastname@example.org 020 7324 2841