How should the economy be measured in times of crisis? Since the start of lockdown, INSEE has regularly published analyses of the economic outlook in light of the uniqueness of the current crisis: the statistical information usually analysed is sometimes not available; and when it is, it may be less robust or less relevant than usual. This is why our analyses are also based on other types of data, [1] whether direct feedback from businesses and professional federations or high-frequency data available almost in real time.

What, for example, can be learned from recent trends in electricity consumption in France, or those of aggregated bank card transaction amounts? During lockdown, due to a fall in production, enterprises consumed 25% less electricity overall than normal, with a marked drop in transport and to a lesser extent in industry. On the household consumption side, the latest available data on bank card transactions suggest a very gradual recovery for certain expenditures on manufactured goods (household equipment, clothing and footwear), while spending on fuel and accommodation and catering remains extremely low.

Overall, according to the information available on 7 May, our estimate of the loss of economic activity due to the health crisis remains in the order of one third (–33%), quite similar to our previous estimates (–35% on 23 April, –36% on 9 April). As 11 May approaches – the date announced for the gradual lifting of the lockdown – a slight upturn appears to be taking shape, most notably in industry and construction, most likely thanks to the introduction of health measures allowing employees to return to their workplace. Household consumption overall is estimated to be down by 32% from its “normal” level (after –33% on 23 April and –35 % on 9 April).

Today, INSEE is also publishing its “Flash” estimate of private-sector salaried employment for Q1 2020; it is doing so after an earlier-than-customary analysis of the nominative social declarations filed by enterprises for March, as well as of the usual surveys. The drop in private-sector salaried employment is estimated at –2.3%, i.e. more than 450,000 net job destructions in one quarter, including almost 300,000 temporary jobs.

Along with the question of how to measure during times of crisis, there is also that of how to compare. This Point de conjoncture looks, among other things, at comparisons between the French regions and departments, as well as between countries.

The first exercise consists in applying homogenous hypotheses of lost activity by sector to different territories. A local-level breakdown of national loss-of-activity hypotheses pinpoints the regions that are structurally more exposed than average (Corsica, Île-de-France, Auvergne-Rhône-Alpes, Provence-Alpes-Côte d’Azur), due to the weight in these regions of accommodation and catering and of certain services to businesses. To complete this analysis, certain high-frequency data can be used at department level: this is the case for aggregated bank card transaction amounts, which after study show that the most highly populated departments are those that have seen their transaction amounts slump most heavily. These departments are probably the most affected in relative terms by the enforced closure of establishments open to the public; in parallel, they may also have seen sections of their population leaving the area when lockdown was announced.

Regarding inter-country comparisons, in normal times the national accounting standards ensure the comparability of the main macroeconomic aggregates. During the current health crisis, this comparability is probably not as sound. [2] In particular, the “Flash” estimates from the Q1 national accounts had to be calculated using incomplete information on March, which was the month that saw the start of lockdown in many countries and which therefore conditioned much of the quarterly trend. Most national statistical institutes have been using less conventional indicators and methods, which are thus less immediately comparable than usual. As regards the Eurozone countries, most notably Italy (decline in GDP of –4.7% in Q1), Spain (–5.2%) and France (–5.8%), at this stage it appears difficult to distinguish between genuine differences in activity levels and the effects of any methodological differences.

Whatever the case may be, these three countries are among those with the strictest lockdowns, in both regulatory and practical terms, as attested to empirically by the high-frequency relative indicators such as human mobility. As things stand, among the main Eurozone countries, only Germany seems to stand apart, by its slightly less strict lockdown measures and a recovery that will probably come a little earlier. These high-frequency data should obviously be closely monitored in the coming weeks: in the forthcoming Points de conjoncture, we will be attempting to measure the initial economic effects of the exit from lockdown.

 

1. We would like to thank all the partners who have continued to support us in this exercise. For this issue in particular: the CB Bank Card Group, the Directorate-General for Enterprises, SNCF Réseau, RTE, Enedis, France Industrie, Rexecode, and the Medef. The estimates in this document are the sole responsibility of INSEE.

2. The problem is one that more generally affects all statistical production: see the blog published on 6 May by the Director-General of INSEE: “Official statistics put to the test of the health crisis”.