The Covid-19 crisis was characterised in most industrialised countries by an unprecedented drop in economic activity for many firms as well as exceptional government financial aid programmes to help these firms. Fears that a large number of businesses, especially small ones, would fail were prevalent, especially in the first phase of the crisis. This fear motivated researchers to estimate the impact of the Covid-19 shock on firms’ cash flows (taking into account government support). An example of this is the paper by Gourinchas et al. (2021), who construct a model of firm cost minimisation with rich firm-level financial data. They estimate the impact of the crisis on business failures among small and medium-sized enterprises (SMEs). Their simulations suggest that many SMEs would have failed in the absence of policy support. This is also the conclusion of Barnes et al. (2021) in the British case. For France, Coeuré (2021) offers a very detailed firm-level analysis of public support.
Having avoided a massive wave of firm failures during the crisis, today many countries are at a key juncture as the activity of firms – especially those mostly hit by restrictions – is quickly rebounding at the same time as the policy support to these firms is reduced or eliminated. A key question – symmetric to the early concern that public support may not be enough to counter the economic shock to SMEs – is whether the rebound in economic activity will be enough to compensate the reduction in public support. The policy discussion is also complicated by limited real-time information on the financial situation of firms. Existing analysis is based either on simulations (in addition to papers cited above, see Bureau et al. 2021) or on precise but dated (i.e. December 2020) balance sheets information of firms (Doucinet et al. 2021). This lack of real-time information makes it difficult to predict whether and how quickly firm failures will rebound in the near future, an issue which has key consequences for the labour market dynamics or banks losses.
One source of real time data on the financial situation of firms is bank account data. This source has barely been exploited.1 An exception is the Bank of England (Hurley et al. 2021), but the latest data point for which is December 2020.
In a new paper (Epaulard et al. 2021), we exploit a unique data et on corporate accounts of a major French Bank – Crédit Mutuel Alliance Fédérale – to study the evolution of the financial position of French micro-enterprises (MEs) (firms with fewer than ten employees) and SMEs (between 10 and 250 employees). We use a sample of approximately 100,000 anonymised MEs/SMEs throughout the Covid-19 crisis. Specifically, our data track inflows and outflows into checking and savings account, as well as balances on these accounts, and liabilities at the bank. While the dataset does not have any information on non-financial assets and liabilities, it still allows us to track, in near real time and at high frequency, firms’ liquidity and net financial position. Hence, we focus on two dimensions of the financial situation of firms that are known to be good predictors of their risk of failure: liquidity and solvency.
In the future, as the government progressively withdraws its support to firms, monthly updates will allow us to monitor (1) whether the economic recovery is sufficient to offset the drop in financial aid, and (2) whether firms can repay the debt contracted during the crisis. The latest data point we present here is September 2021.
More cash in all sectors
The first striking observation is a significant increase, at the aggregate level, of cash balances held on corporate accounts. This can be attributed to financial support provided by the government, in part through state-guaranteed loans (known as ‘PGE’ in French), which were heavily used by firms in France. All sectors experience an improvement in liquidity, as illustrated in Figure 1. There are fewer companies in August 2021 in a ‘very weak’ or ‘weak’ liquidity position (defined as the share of firms below the 5th and 25th percentile of pre-crisis cash balances) than in the four months preceding the pandemic.
Figure 1 Liquidity situation of all businesses
Note: The dark red part ("Very weak") corresponds to the share of firms whose liquidity situation at the end of the month is below the threshold that corresponded to the worst 5% of firms before the crisis. The other thresholds are the 25%, 75% and 95% thresholds. The thresholds are calculated within each type of company (SME or ME) as well as within companies in the same turnover bracket.
Source: Crédit Mutuel Alliance Fédérale data.
Net financial position: More heterogeneous situations in September 2021
Beyond liquidity, we also investigate firms’ insolvency risk through their net debt. Net debt corresponds to the difference between the remaining principal balance on all debts held at the bank and the balance on checking and saving accounts. The former includes term loans and short-term debts, including state-guaranteed loans.
For both the average and median firm in our sample, net debt of MEs and SMEs decreased significantly between February 2020 (just before the first lockdown) and September 2021. But this improvement in firms’ financial position was not uniform. Econometric analysis suggests that although the average net debt was reduced over the period, the evolution was less favourable for SMEs than for MEs. This is likely due to the design of the main cash subsidy (Fonds de Solidarité), which was initially more generous for smaller firms. The financial situation has also evolved less favourably in the Paris region. This confirms that in regions where a large share of economic activity relies on tourism, the economic shock triggered by the pandemic was stronger. Our analyses also show that in all geographic areas micro-enterprises in so-called ‘S1 industries’ (i.e. industries classified by the government as the most affected by health restrictions and therefore more subsidised) have experienced a larger reduction in net debt. The fact that the financial situation of small firms in sectors most affected by sanitary restrictions but also most supported by public policy improved more (apart from the Paris region) than firms in less affected sectors suggests that public support may have overcompensated the Covid-19 shock for many firms.
Within each sector there is an increase in heterogeneity of financial situations
Within each industry, there is an increase in the heterogeneity of net financial positions in September 2021 relative to the pre-pandemic situation. In almost all industries, we see an increase not only in the share of firms in a difficult financial situation, but also in the share of those in a strong financial position. The crisis has generated a more polarised distribution of firms within each sector, which may in the future lead to reallocation dynamics with potentially important consequences for the structure of sectors and aggregate productivity. Figure 2 illustrates this increased heterogeneity for the manufacturing, transport, and construction sectors.
Figure 2 Net financial situation of four sectors
Source: Crédit Mutuel Alliance Fédérale data.
There are two notable surprising dynamics. First, the accommodation and food service activities have experienced an overall decrease in the share of firms with ‘weak’ financial positions. Second, the construction sector appears to be the most fragile, with a sharp increase in the share of firms in a ‘very weak’ financial situation. A possible explanation lies in the generous support received by firms in the first sector relative to the second.
What do these results tell us about bankruptcies in the months to come? The number of bankruptcies fell dramatically in 2020 relative to pre-pandemic levels (by around 38% compared with 2019). This drop implies that a number of businesses that have survived in 2020 and 2021 would probably have failed in ‘normal’ times, although firm bankruptcies have responded to the same factors during the crisis as they usually do ( Cros et al. 2021). These firms likely belong to the category of firms we describe as having a ‘weak’ financial position. In several sectors (notably accommodation and food), an increase in bankruptcies in the coming months would thus mostly be a ‘catch-up’ process. But in other industries (notably construction and manufacturing), the rise in the number of firms in a ‘very weak’ financial position is larger than the ‘missing bankruptcies’. In these industries, the question whether weak firms will be able to repay their debt is still open. This calls for vigilance and close monitoring in the coming months. We believe the use of real-time bank data will be particularly useful.
Barnes, S, R Hillman, G Wharf and D MacDonald (2021), “How businesses are surviving Covid-19: The resilience of firms and the role of government support”, VoxEU.org, 16 July.
Bounie, D, Y Camara, É Fize, J Galbraith, C Landais, C Lavest, T Pazem and B Savatier (2020), “Consumption Dynamics in the COVID Crisis: Real Time Insights from French Transaction & Bank Data”, CAE focus, October.
Bureau, B, A Duquerroy, J Giorgi, M Lé, S Scott and F Vinas (2021), “Health crisis: (very) heterogeneous cash flow shocks”, Banque de France.
Coeuré, B (2021), “What 3.5 million French firms can tell us about the efficiency of Covid-19 support measures”, VoxEU.org, 8 September.
Doucinet, V, D Ly and G Torre (2021), “The differentiated impact of the crisis on companies’ financial situation”, Banque de France.
Epaulard, A, É Fize, T Le Calvé, P Martin, H Paris, K Parra Ramirez and D Sraer (2021), “The financial situation of French small firms based on their bank accounts in August 2021”, Focus Conseil d’Analyse Economique.
Gourinchas, P O, Ş Kalemli-Özcan, V Penciakova and N Sander (2021), “COVID-19 and SME Failures”, May.
Hurley, J, S Karmakar, E Markoska, E Walczak and D Walker (2021), “Impacts of the Covid‐19 Crisis: Evidence from 2 Million UK SMEs”, Bank of England, Staff Working Paper 924.
1 This is not the case for households’ financial situation during the crisis, as bank accounts have been analysed by several studies (e.g. Bounie et al. 2020).