Introduction by the Director General of INSEE

Ten days ago, INSEE put out a news release announcing that the publication of "Point de conjoncture" for March was postponed and that the Institute intended to produce an evaluation of the economic situation every two weeks.

This first evaluation is being published today, Thursday 26 March, at the same time as the usual monthly business tendency survey results.

These survey results show a sharp fall, but they are not in themselves sufficient to give a true measure of the decline in activity, as for the most part, companies’ responses were collected before the start of the lockdown period.

We therefore had to use other sources of information, some quantitative, some only qualitative, and to adopt some unusual methods.

In this, we have received support and assistance from a wide variety of bodies which I would like to take this opportunity to thank: Banque de France, CB Bank Card Group, all the companies and professional federations that provided information via the DGE (Directorate General for Enterprise), France Industrie, Medef and Rexecode, as well as OFCE (French Economic Observatory). As a result, we believe that we can give an immediate order of magnitude of the loss of economic activity, as compared with a normal week, and as compared with what we might have expected.

However, before giving the go-ahead for publication, I did hesitate, for two reasons.

First, it may seem somewhat hollow to be measuring economic activity when our primary concern must be to avoid a health disaster. This is a time when we should be monitoring the epidemic and relying on public health expertise, and some people may be shocked by the amount of attention that is also being paid to economic figures. However, it seems essential to me to measure the shock that the economy is undergoing, as this information is indispensable for decision-makers and vital for economic stakeholders. And, even if we at the statistical institute did remain silent, others would in any case put forward their own assessments.

The second reason is that statisticians are always loath to come up with results of which they are not entirely sure. And what we are putting forward today is fragile to say the least, and subject to revision. First, because our methods have never before been tested in a situation such as this: this is unprecedented in the history of INSEE. And second because the situation itself is very changeable: in certain industrial sectors and in civil engineering, activity is recovering after being interrupted; in other sectors, such as services to businesses, for instance, the bottom has probably not yet been reached. Be that as it may, and however uncertain and imprecise these first figures may be, it seemed to me that it was better to publish them than to say nothing at all.

At the present time, we estimate that economic activity stands at around 65% of its normal level. I did not want us to give an estimate of quarterly GDP growth, much less for the year. This will depend mainly on the duration of this period of lockdown, which we have neither the legitimacy nor the competence to predict. To give an idea of possible scenarios, this publication puts forward several hypotheses. Regarding consumption, we estimate that total household consumption in France also currently stands at 65% of its normal level, with obviously some very strong sectoral differences. In this initial analysis, we have not attempted to make any other assessments.

INSEE’s Press Office is operational. Several of us will be available today to answer questions, but no additional figures should be expected.

Finally, I would like to thank everyone who contributed to this exceptional analysis, especially since they completed it in record time and in unusual conditions as we were all working from home.

This publication will return, if possible in greater depth, in two weeks’ time.

Jean-Luc Tavernier

In a rapidly evolving situation, the results of the business tendency surveys are very closely dependent on the collection period

Since the beginning of 2020, after first hitting China, the coronavirus epidemic has now spread to the rest of the world. The measures taken to stem this health crisis have serious consequences for the functioning of the world’s economies. Announcements on the subject have been made in rapid succession in the various countries concerned.

The business tendency surveys in France for January and February were relatively unaffected by this crisis. Results collected for March, however, directly reflect the concerns of businesses in this unprecedented context. But they probably paint only a partial picture.

Collection of the March business tendency surveys started on 26 February 2020 and analysis of the responses was finalised on 23 March 2020. However, from 14 March, paper format responses received by post could not be processed: from this date, only new online responses were included. In addition, contrary to our usual practice, given the exceptional situation, no reminders were sent out this month to non-responding companies.

For these reasons, the majority of the March surveys reflect business leaders’ opinions for the first half of March, i.e. before the announcement of the lockdown of the population in order to contain the epidemic.

Business tendency surveys for March 2020 indicate a substantial decline in the business climate, directly linked to the current health crisis

Business climates are presented at both sectoral and global levels in the various issues of Informations Rapides published at the same time as this "Point de conjoncture".

At global level, the business climate has lost 10 points, and stands at 95. This is the largest monthly decline in the indicator since the series began (1980). In October 2008, after the collapse of Lehman Brothers, the indicator fell by 9 points. The climate indicator for employment has also experienced its largest decline since the start of the series (1991). It has lost 9 points, and stands at 96.

At sectoral level, business climate indicators have deteriorated considerably in services (-14 points) and retail trade (-13 points). They are also deteriorating in wholesale trade (-5 points) and industry (-3 points). Most notably, business expectations in each sector are down sharply: the corresponding balances of opinion have lost 15 points in services, 25 points in retail trade, and 33 points in industry and wholesale trade. The business climate in the building construction industry, however, seems stable for March: this provides confirmation that these indicators must be read with caution this month; for the reasons mentioned above, they probably reflect business leaders’ opinions at the beginning of the month rather than the end. For the same reasons, it would seem difficult to use these indicators for calibrations intended to predict activity, given the very sudden nature of the slump throughout the month of March, a slump that is probably only partly reflected in the business tendency surveys.

The estimate of direct loss of activity is based on various sources

The aim of the analysis given below is to evaluate the loss of activity linked directly to the measures taken to curb the health crisis, and in particular the lockdown imposed on the population. The goal is thus to compare the estimated situation regarding activity in the current week with what would be expected in a “normal” week.

To do this, hypotheses had to be formulated for each branch about the share of activity that was maintained, based on various sources: direct feedback from companies (this was especially the case for transport) and professional federations via information collected by the DGE (Directorate General for Enterprise), France Industrie, Medef, Rexecode, etc. “High-frequency” data were also used, relating, for instance, to electricity consumption, rail transport and statistics on bank card transactions obtained via the CB Bank Card Group.

The loss of economic activity is currently estimated at 35% compared with a “normal” situation

The impact of the current health crisis may vary greatly depending on the activity branch (Table 1).

- Agricultural activity is likely to continue only slightly below “normal”, and agrifood industries are likely to be less affected than the rest of industry. Overall, activity should weaken only slightly.

- However, taking into account feedback from the different industrial branches, only half of activity in the rest of industry is expected to be maintained.

- Among market services, some branches have been very severely affected (transport, accommodation, food services, leisure, etc.) while others are probably affected much less (telecommunications, insurance, etc.). Elsewhere, property rental payments are inert and therefore for the most part not much affected in the short term. It is therefore expected that two-thirds of market service activity will be maintained.

- Concerning non-market services, in the national accounts they are usually calculated according to production costs, mainly wages, which makes them intrinsically very inert. The decline in activity is therefore conventionally less pronounced. Nevertheless, some non-market service activities are likely to dip sharply (e.g. day nurseries, child-care centres, libraries, sports activities).

All in all, the “immediate” loss of activity is estimated at about one third. This figure seems consistent with the initial information that has become available on the situation of employees, about a third of whom are expected to be active in their usual workplace, one third teleworking and the last third on the short-time working scheme. These figures also seem compatible with the observed decrease in electricity consumption, currently down by around one fifth compared with the “normal” functioning of the economy. Indeed, only part of this consumption has reacted to the drop in activity, as households in particular are continuing to consume electricity at home.

Household consumption is also expected to be 35% lower than “normal”

Based on the lockdown measures put in place (list of businesses that can remain open), feedback from professional federations and daily bank card transaction data, it is also possible to give an immediate estimate of the change in household consumption compared with a “normal” period.

Household consumption looks set to fall back very significantly, by around -35% (Table 2).

The largest contribution to this drop is the collapse in the consumption of numerous goods produced by manufacturing industry (-60%), accounting for 18 points of this decline. Some consumption expenditure has been cut back to the bare minimum, falling by between -100% and -90% (transport equipment, textiles, clothing). Other expenditure has remained stable (electricity, water), or has actually increased slightly (pharmaceutical industry, +5%).

Consumption of market services has also declined, by about 33%, contributing around 15 points to the total drop in consumption. However, excluding housing costs, which by their very nature are little affected in the short term, the decline stands at around 55% and is particularly significant in the catering, accommodation and transport sectors.

Non-market services consumption is likely to fall by 34%, contributing 2 points to the overall decline. Due to the suspension of renovation work, household consumption in the construction branch is expected to fall by 90%, contributing one point to the overall drop in household consumption.

Conversely, as a result of the lockdown and the drop in consumption of catering services, household consumption of agricultural and agrifood products is expected to increase by 6%, thereby reducing the total fall in consumption by 1 point.

In accounting terms, the impact of a one-month lockdown is likely to be a reduction of around twelve percentage points of quarterly GDP (or 3 annual GDP points)

Based on the estimate of these “immediate” effects, it is no easy task to calculate quarterly or annual forecasts of changes in French GDP:

- Change in GDP growth is very closely linked to the exit scenario at the end of the health crisis. INSEE is unable to make any forecasts on this subject.

- Support measures for households and businesses are aimed at preventing this sudden, but theoretically temporary, shock from leaving deep and permanent scars. However, it is possible that the return to normal will not be immediate.

- Lastly, the behaviour of economic agents will depend on their level of confidence in the improvement in the health situation and therefore remains very uncertain.

Nonetheless, by way of illustration, Table 3 provides an accounting conversion into quarterly and annual GDP points of the loss of activity caused by either one or two months of lockdown. One month would be equivalent to a loss of about 12 quarterly GDP points, or 3 annual GDP points.

In the coming weeks, and if possible in a fortnight’s time, INSEE will attempt to examine how these estimates are evolving and to shed further light on the economic situation in France.