French firms through the COVID storm: Evidence from firm-level data

Agnès Bénassy-Quéré, Benjamin Hadjibeyli, Guillaume Roulleau 27 April 2021

a

A

While vaccination is accelerating in advanced economies and restrictions are being lifted, the question of how to deal with the legacy of the COVID-19 crisis in the corporate sector is moving to the centre stage of policymaking. Based on Orbis data for 20 countries and on the methodology introduced by Gourinchas et al. (2020), Díez et al. (2021) estimate that the proportion of insolvent small and medium-sized enterprises (i.e. SMEs with negative equity) may rise by six percentage points over 2020-21, with lower figures in Asia and higher ones in Southern Europe. By allowing firms to optimize their payroll, these calculations implicitly account for emergency support provided by governments through generous short-time work schemes. Also based on Orbis data but with a different methodology, Demmou et al. (2021) stress the importance of alleviating labour costs for reducing the proportion of newly insolvent firms.

Relying on a large sample of countries allows to draw consistent cross-country comparisons. However, this is at the cost of not accounting for the within-sector heterogeneity of the shocks or for the variety of grants and tax reliefs received. For instance, Orbis data do not provide information on firm-level shocks, adjustments and take-ups of public support schemes, which have to be simulated. 

These cross-country estimations can thus be viewed as complements to more detailed work carried out at the national level, based on administrative datasets. Hadjibeyli et al. (2021) offer such an exercise for 1.8 million French firms from March to December 2020.1 Accounting for short-time work, direct subsidies and tax reliefs, the rise in the proportion of newly insolvent French firms due to the crisis is ‘only’ three percentage points. Without public support, though, the proportion would have risen by a whopping eight percentage points (Figure 1).

Figure 1 Microsimulation results, French firms over March to December 2020

Notes: Amongst all firms, 11.9% would have become insolvent over March-December 2020 without public support, a proportion that shrinks to 6.6% thanks to public support received, to be compared to the 3.6% in the counterfactual of no crisis; 71.1% of all firms remain solvent in all scenarios.
Source: Adapted from Hadjibeyli et al. (2021).

Importantly, insolvency is a signal of financial distress but does not necessarily involve bankruptcy. As a matter of fact, 17% of firms (mostly young companies) were already insolvent before March 2020. The distinction between illiquidity, insolvency, and viability will be crucial for the phasing out of emergency support, which will need to avoid both the risk of excess liquidations and that of a ‘zombification’ of the economy.

Incorporating observed shocks, adjustments and take-ups in microsimulations

These new results are derived by simulating the financial statements of 1.8 million non-financial firms (82% of French firms’ value-added) over the course of 2020. Turnovers are measured through individual monthly value-added tax (VAT) records. Firms react to negative shocks by reducing their payroll, through short-time work, non-renewal of short-term contracts or layoffs. Observed data are used up to June 2020 (turnover) or September 2020 (payroll). For the rest of the year, the simulations rely on sector-level data and on a number of assumptions. Specifically, firms are supposed to partially adjust their variable costs, whereas they cannot adjust their fixed costs. 

Public support consists in short-time work (compensation for the wages of workers who cannot work –– 84% of the worker net wage with a floor at minimum wage, with firm-level take-up data up to September), direct subsidies through the ‘solidarity fund’ (a subsidy scheme for SMEs particularly affected by the crisis, with firm-level data up to November), tax deferrals (only employer’s social contributions in the simulation, firm-level data up to October) and tax reliefs (employer’s social contributions for firms particularly affected by the crisis, fully simulated). When individual take-up data are missing, the grants are simulated based on the eligibility criteria.

When costs exceed revenues, a firm is supposed to first exhaust its cash, and then to borrow. It is assumed that, thanks in particular to state-guaranteed loans, firms are not subject to credit constraints. Hence, they accumulate debts when they run out of cash. When debts exceed assets, a firm is deemed insolvent. 

In the simulations, it is assumed that, without public support, firms would have had similar behaviour in terms of layoffs. In particular, they would have kept on their payroll the employees that in fact were put in partial activity. This assumption is probably extreme, since layoffs would have been much larger without the short-time work scheme. But, in the absence of any reliable guideline to estimate firms’ counterfactual behaviours, the choice was made to rely on an overly simple but transparent assumption. Hence, over March-December 2020, the 12% figure probably somewhat overestimates the number of French firms that would have become insolvent without public support; some of them could have survived by laying off employees. It nevertheless provides an order of magnitude of the proportion of firms at risk of massive layoffs, restructuring or (ultimately) liquidation.

Heterogeneity

Which form of public support has made the difference? According to Figure 2, large firms and mid-caps seem to have benefited mainly from short-time work: the proportion of newly insolvent firms is similar with the whole set of public support schemes (blue bars) as with short-time work only (hatched blue bars). In contrast, very small firms have benefited relatively more from the solidarity fund and from tax reliefs.  

Figure 2 Share of newly insolvent firms by size (% of all firms)

Source: Hadjibeyli et al. (2021).

Unsurprisingly, accommodation & food is the most severely hit sector (Figure 3). Without public support, almost 30% of firms in this sector could have become insolvent between March and December 2020. With public support, this share falls dramatically. Information & communication is in a very different situation, with relatively high insolvency rate in normal times and not much more as a result of the crisis.

Figure 3 Share of newly insolvent firms by sector (% of all firms)

Source: Hadjibeyli et al. (2021).

Cleansing

The simulations also provide an insight into potential cleansing effects of firm exit. Figure 4 plots the distribution of labour productivity across initially solvent firms, after controlling for firms’ sector and size. In normal times, newly insolvent firms are on average less productive than the average firm. The red line picks at a lower level of productivity than the black line. With the crisis, insolvent firms are still less productive than the whole sample, but the gap is reduced: the blue lines pick in-between the red and the black ones. Hence, the cleansing effect is still there, but it is less effective than in normal times. Note that public support does not distort the distribution of newly insolvent firms: it does not systematically protect more productive nor less productive firms (solid and dotted blue lines are almost coincident), most likely because the support was provided very broadly.

Firms have been affected differently by the crisis depending on their sector and size. For instance, firms in the accommodation & food sector are typically less productive than manufacturing companies, and smaller firms are generally less productive than larger firms. This composition effect could add to the cleansing effect and raise aggregate productivity. However, this is only a short-term effect. The distribution of activity across sectors after the crisis is still extremely uncertain. Furthermore, aggregate productivity will also depend on new entrants. Replacing food services by healthcare may not raise aggregate productivity.

Figure 4 Distribution of labour productivity (in log) across initially solvent firms (controlling for sector and size)

Note: The sample is restricted to firms with at least one employee and excludes initially insolvent firms. The black line represents the distribution of labour productivity for the whole economy. The red line represents the same distribution for the subsample of firms becoming insolvent between March and December 2020 without a crisis. The solid blue line shows the distribution for the subsample of firms becoming insolvent with the crisis but without public policies. The dashed blue line shows the same distribution but accounting for public policies (short-time work, payroll tax deferral, tax relief, SMEs Solidarity Fund).
Source: Hadjibeyli et al. (2021).

Debt overhang

Even though corporate investment is mostly driven by demand, which is expected to rebound, a debt overhang can weigh on investment (Kalemli-Özcan et al. 2019, Demmou et al. 2021). Based on an econometric modelling of the reaction of corporate investment to firms’ indebtedness, the simulations suggest that, even if demand and profits get back to their pre-crisis level, the new debt caused by the crisis could translate into a 2% reduction in tangible investment compared to its pre-crisis trend.

In order to counteract these effects, the French recovery plan includes permanent cuts of production taxes to support firms’ margins, and quasi-equity injections (prêts participatifs and obligations Relance) to strengthen their balance sheets. 

All in all, it is somewhat reassuring that the various simulations and policy recommendations that are available at this stage tend to converge. The challenge now is one of implementation.

References

Demmou, L, S Calligaris, G Franco, D Dlugosch, M McGowan, and S Sakha (2021), “Insolvency and debt overhang following the COVID-19 outbreak: Assessment of risks and policy responses”, VoxEU.org, 22 January.

Díez, F, R Duval, J Fan, J Garrido, S Kalemli-Özcan, C Maggi, S Martinez-Peria, and N Pierri (2021), “Insolvency Prospects Among Small and Medium Enterprises in Advanced Economies: Assessment and Policy Options”, IMF Staff Discussion Note 002.

Gourinchas, P, S Kalemli-Özcan, V Penciakova, and N Sander (2020), ”COVID-19 and SME failure?” NBER Working Paper No. 27877, National Bureau of Economic Research, Cambridge, MA. 

Hadjibeyli, B, G Roulleau and A Bauer (2021), “Live and (don’t) let die: The impact of Covid-19 and public support on French firms”, French Treasury working paper, 2021-2, April.

Kalemli-Özcan, S, L Laeven, and D Moreno (2019) “Debt Overhang, rollover risk, and corporate investment: evidence from the European crisis”, ECB Working Paper Series No. 2241.

Endnotes

1 This work was carried out for and presented to the Committee for the Monitoring and Evaluation of Support Measures for Companies Confronted with the Covid-19 Epidemic chaired by Benoit Coeuré

a

A

Topics:  Covid-19

Professor, Paris School of Economics - University of Paris 1 Panthéon-Sorbonne

Economist, Financial Sector Economic Analysis unit, French Treasury

Economist, Economics of Knowledge, Innovation and Industrial Policies Unit, French Treasury

Events

CEPR Policy Research