December 8, 2021

gurqui

Only The Finest Women

Firm-level policy support during the Covid-19 crisis

Peter Harasztosi, Laurent Maurin, Rozália Pál, Debora Revoltella, Wouter van der Wielen 18 November 2021

Massive policy support sheltered many European firms from the unprecedented shock of the Covid-19 pandemic. Data from a 2021 vintage of the European Investment Bank Investment Survey (EIBIS) reveal that 56% of European firms got support via at least one specific policy in response to the crisis.1 Figure 1 reports the share of companies that have benefitted from policy support across EU countries. The majority of them received only one type of support; around a third of companies that received support benefitted from two types of policies; and 4% of firms benefitted from three types of support. Among the types of policy support, subsidies or financial support (type 3) is the most common, used by 36% of the firms, a ratio somewhat higher in the Northern and Western economies and in the Central and Eastern economies. A similar share of firms (16–17%), benefitted from the deferral of payments (type 2) or credit support to be paid back (type 1).

Figure 1 Intensity of the policy support across European economies (% of recipient firm)

Source: Authors’ calculations based on the European Investment Bank Investment Survey (EIBIS) 2021. 
Note: Percentage of firms having benefitted from at least one type of support. The colour reflects the region in which the economy is located: Red indicates Central and Eastern economies, Green indicates Southern economies, and Orange indicates Northern and Western economies.

Firms that recorded larger sales losses are more likely to have been supported. Figure 2 considers sector asymmetries using two sectoral breakdowns. In the top panel, we look at the share of firms with policy support in relationship to the share of firms with large sales losses (of more than 25%) for 12 broad sectors.2 The panel confirms that services comprise some of the most affected sectors, such as hotels and restaurants, as well as some not or positively affected, such as IT and telecommunications. With an R2 of 76%, the positive relationship with the prevalence of policy support suggests that, across sectors, the support was strongly connected to the change in activity: the stronger the decline in turnover in the sector, the higher the intensity of the policy support. When the types of policy support are investigated separately, subsidies or other financial support (type 3) shows the strongest relationship with sales losses. This type of policy includes temporary unemployment schemes, which are most directly linked to activity. 

Smaller firms recorded larger sales losses and therefore got more support. The EIB Investment Survey (EIBIS) four-sector breakdown is used together with a breakdown by firm size categories in the bottom panel of Figure 2. For each of the four sectors considered separately, smaller firms were more affected, a result confirmed in the literature (Gourinchas et al. 2021). They are positioned at the right of their peers in the same sector as they are more likely to have recorded larger sales losses compared to large firms: respectively, 29% versus 9% in the manufacturing sector, 35% versus 29% in the services sector, 18% versus 1% in the construction sector, and 26% versus 16% in the infrastructure sector. Consequently, smaller firms are also more likely to be supported: they are positioned above peers in the same sector. 

Figure 2 Determinant of the allotment of policy support (% of firms)

           

Source: Authors’ calculations based on EIBIS 2021. 
Note: Any type of policy support. 

Pre-crisis productivity levels and liquidity ratios appear as major discriminant factors to obtaining policy support. We estimate the discriminatory power of firm characteristics in obtaining policy support by estimating separate probit models, each time controlling for country, sector, size, and sales decline and then adding a firm characteristic one by one.3 Figure 3 plots the results, i.e. the changes in the predicted probability of getting the policy support (of any type). When the characteristic is binary, the presence or absence is reported as ‘yes’ or ‘no’. When it is continuous, ‘high’ relates to being in the top decile and ‘low’ in the first decile. Firms with low pre-Covid productivity are significantly more likely to be supported than firms with high productivity. Being an exporter also matters, albeit to a lesser extent. The other real characteristics do not seem to have a predictive impact.

Turning to the financial indicators, firms with low liquidity ratios and financially constrained firms are more likely to get policy support. While firms in distress, firms with low return on assets, firms recording losses, and highly indebted firms are more likely to get support, the difference is not statistically significant. Conversely, firms with a lower liquidity ratio, with fewer cash buffers, or financially constrained firms, are significantly more likely to get policy support. This suggests that the primary goal of the policy support – avoiding a liquidity dry-out and sharp rise in insolvencies – was achieved (Hadjibeyli et al. 2021). Overall, we do not find evidence that the support was tilted towards firms already weak before the crisis, such as financially distressed or zombie firms. Focusing on firms located in Croatia, Finland, Slovenia, and Slovakia, Bighelli et al. (2021) also conclude that employment subsidies and direct subsidies have only been marginally distributed towards ‘zombies’. 

Figure 3 Predicted probability of getting the policy support across firm characteristics pre-Covid (% of respondents)

Source: Authors’ estimations based on the EIBIS 2021 matched with the ORBIS database. 
Note: The vertical line reports the 95% confidence interval of the conditional probability of getting the support. Two overlapping lines indicate that the factor does not alter the probability significantly. Red bars indicate when the characteristic is significantly discriminant.

The support catalysed recapitalisation and supported investment

The support dampened the impact of sales losses on investment. Figure 4 plots the percentage of firms planning to invest more in the current financial year, depending on the sales losses they recorded during the first year of the Covid-19 crisis and distinguishing among firms that have benefitted from the support and those that have not. The share of firms planning to invest more increases when sales losses decrease. Moreover, for the same level of sales losses, firms that were supported plan to raise investment by more. The difference is especially pronounced for large sales losses. Estimations confirm that large sales losses lead to lower investment and that policy support partly compensated for this impact. We estimate a probit model to explain the probability that a firm increases investment in the current financial year. The estimations confirm that large sales losses (above 25%) reduce the probability to increase investment by 5 to 8 percentage points. At the same time, obtaining policy support, of any type, significantly increases the probability of raising overall investment by 2 to 3 percentage points. 

Leverage increased for 17% of firms and supported firms strengthened their equity base by more. Figure 5 plots the share of firms that have increased their recourse 3 external finance, debt, or equity, depending on whether they have benefitted from policy support. The balance sheet expansion is stronger for those that have been supported. In the case of debt, part of the difference reflects the recourse to subsidised loans or guarantees and is therefore accounting the support, but this is not the case for equity – 7% of supported firms increased their equity base. Sales losses also raise the probability of increasing the equity base.4 The conjunction of these two effects suggests that the recapitalisation needs resulting from large sales losses are more likely to be filled when the company benefits from the policy support. Obtaining policy support increases the likelihood of crowding-in equity investors. Such an interpretation is somewhat supported by the estimated impact of firm characteristics. The higher the financial leverage and the lower the pre-Covid capital ratio, the more likely the increase in the equity base. Hence, part of the changes in the financial structure corrects pre-crisis balance sheet weaknesses. 

Figure 4 Investment plans conditional on Covid-19 sales losses and policy support (% respondent)   

             

Source: Computations based on the EIBIS 2021. 
Note: The x-axis reflects the sale losses reported by the company. The y-axis reports the percentage of firms surveyed that plan to raise investment in the current financial year.    

Figure 5 Policy support and balance sheet expansion (% of respondents)  

 

Source: Computations based on the EIBIS 2021.

The support was especially pronounced for digitalisation. Harasztosi et al. (2021) analyse the determinants of digitalisation, focusing on the impact of the policy support and the change in the liability structure.5 Digitalisation spending increases with pre-crisis productivity. This result comes after conditioning on the sector and country in which the firms operate. Hence, in the same country and sector, the more productive firms are more likely to digitalise further. This dynamic may contribute to widening the productivity gap as digitalisation fosters productivity. Sales losses reduce the likelihood of further digitalisation by five to ten percentage points. However, the effect is compensated by the policy allotment. On the one hand, firms that got policy support are unconditionally more likely to digitalise more, by four to five percentage points. On the other hand, firms that got the policy support while suffering large sales losses are more likely to digitalise than those that suffered sale losses but did not get the policy support, by five percentage points. 

Conclusion

We do not find evidence of abnormal misallocation of policy support as pre-crisis firm weaknesses do not appear to explain the allotment. It is true that firms recording larger sales losses and with low liquidity buffers got more support. But this suggests that the support, albeit untargeted and widely distributed, successfully reached the firms that suffered the most in terms of pandemic-induced revenue reductions. The first goal of the policy, avoiding a liquidity dry-out and freezing the corporate ecosystem, was achieved. Moreover, supported firms are more positive about their investment outlook and more likely to digitalise faster. Evidence suggests that they are in a better position to crowd-in investors and recapitalise. The conjunction of policy support and stronger equity bases accelerates the digital transformation of European corporates, a transformation that the crisis makes even more necessary. 

Authors’ note: The views expressed are those of the authors and do not necessarily reflect the position of the EIB. 

References

Bighelli, T, T Lalinksy and F di Mauro (2021), “Covid-19 government support may have not been as unproductively distributed as feared”, VoxEU.org, 19 August.

Gourinchas, P O, S Kalemli-Ozcan, V Penciakova and N Sander (2020), “COVID-19 and SME Failures”, CEPR Discussion Paper 15323.

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”, Direction générale du Trésor, Working Papers 2021/2, Ministère de l´Économie, des Finances et de la Relance.

Harasztosi, P, L Maurin, R Pál, D Revoltella and W van der Wielen (2021), “Policy support during the crisis: So far, so good?“, EC Annual Research Conference, 15 November. Forthcoming in the EIB Working Paper Series.

Revoltella, D, L Maurin and R Pal (2020), “EU firms in the post-COVID-19 environment: Investment-debt trade-offs and the optimal sequencing of policy responses”, VoxEU.org, 23 June.

Endnotes

1 Three types of support are considered in the analysis: (1) New subsidised or guaranteed credits that will need to be paid back in the future but have preferential or reduced interest rates and/or an extended repayment plan, (2) Deferral of payments which still leave a liability to be paid by the company in the future, and (3) subsidies or any other type of financial support that the company will not have to pay back in the future, a type of support that comprises job retention schemes.

2 While the EIBIS sampling is not designed to be representative of these 12 sectors, each is populated by 350 firms at least at the EU level. 

3 See Harasztosi et al. (2021) for more details.

4 See Revoltella et al. (2020) for the need to increase the capital base of corporates after the sharp fall in profits during the Covid-19 crisis.

5 Harasztosi et al. (2021) estimate the following equation: 

Where Fin relates to financial expansion, whether the firm has raised equity and/or debt. Sales is the dummy variable indicating if the firm reported a decline of more than 25% in its sales. Pol indicates that the firm has benefitted from at least one policy support measure. Each dummy takes the value one when the answer is positive and zero otherwise. Z is a set of firm characteristics, related to its balance sheet structure or P&L. Labour productivity is always incorporated in the equations.