COVID-19 Financial Aid and Productivity: Has Support Been Well Spent? — Bruegel, November 2021

Summary

A Bruegel Policy Contribution, Issue 21/21 dated November 2021, by Carlo Altomonte, Maria Demertzis, Lionel Fontagné and Steffen Müller, asking whether COVID-19 financial aid to firms has been well spent. The executive summary says most European Union countries have progressed with vaccination and are seeing a rebound, and that regions dependent on services needing physical proximity were hit hardest and began with the highest shares of zombie firms. It finds the main goal of protecting employment has been achieved, though how efficiently is varied, and that an analysis of France, Germany and Italy shows potential for cleansing effects.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Issuer: Source documents
Document type: Covid 19 Financial Aid And Productivity Has Support Been Well Spent Pb
Case: Covid 19 Financial Aid And Productivity Has Support Been Well Spent Pb

Cite this document

“COVID-19 Financial Aid and Productivity: Has Support Been Well Spent? — Bruegel, November 2021.” Original at www.bruegel.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/covid-19-financial-aid-and-productivity-has-support-been-well-spent-pb/, SHA-256 141233f386160c30cc1a8ca9c7a0671769a552d0cb4998c16faa1ce7cdfee276.

Full text

Policy Contribution
Issue n˚21/21 | November 2021
COVID-19 financial aid and
productivity: has support
been well spent?
Carlo Altomonte, Maria Demertzis, Lionel Fontagné
and Steffen Müller

Executive summary
Most European Union countries have made good progress with vaccinating their populations
Carlo Altomonte (carlo. against COVID-19 and are now seeing a rebound in economic activity. While the scarring
altomonte@bruegel.org) effects of the crisis and the long-term implications of the pandemic are only partially
is a Non-resident Fellow at understood, the effects of support given to firms can be evaluated in order to help plan the
Bruegel and a Professor at removal of crisis support.
Bocconi University
European regions and countries that depend heavily on services requiring physical
proximity have been hit hardest by COVID-19-related measures. But these services sectors
Maria Demertzis (maria. tend also to be the smallest and least-productive in any economy, implying that, coming
demertzis@bruegel.org) is into the crisis, the highest shares of zombie firms were in these sectors. Reliance on physical
Deputy Director at Bruegel proximity and the higher incidence of zombies to start with have combined to make those
services-dependent economies particularly vulnerable to any attempt to remove the support
put in place during the pandemic.
Lionel Fontagné (lionel.
fontagne@univ-paris1.fr) is The evidence shows that the main goal of the provision of support during the COVID-19
a Professor at Paris School of crisis, namely to protect employment, has been achieved. However, the evidence is varied on
Economics how efficiently this was done, in terms of helping firms that have a good chance of surviving,
while not supporting those that will in any case exit.

Steffen Müller (steffen. An analysis of France, Germany and Italy shows the potential for ‘cleansing effects’ in that
mueller@iwh-halle.de) is it was the least-productive firms that have been affected most by the crisis. While support
Head of Structural Change was generally not targeted at protecting good firms only, financial support went by and large
and Productivity at Halle to those with the capacity to survive and succeed. Labour schemes have been effective in
Institute for Economic protecting employment.
Research

This Policy Contribution is an output from the MICROPROD project (http://www.microprod.
eu/), which received funding from the European Union’s Horizon 2020 research and innovation
programme under grant agreement no. 822390. The authors thank colleagues at Bruegel for
valuable comments and suggestions. Lionel Jeanrenaud and Maddalena Conte provided
valuable research assistance. Opinions are those of the authors.

Recommended citation
Altomonte, C., M. Demertzis, L. Fontagné and S. Mueller (2021) ‘COVID-19 financial aid and
productivity: has support been well spent?’Policy Contribution 21/2021, Bruegel
1 Introduction
The support given to firms in European Union countries and the United States in the wake of
COVID-19 to protect economic activity and employment has been unprecedented. In the EU,
this support has prevented the emergence of unemployment and has kept average employ-
ment high. Employment in the EU has remained broadly constant while in the US, employ-
ment fell by more than 10 percentage points at the start of the pandemic, with subsequent
recovery (Figure 1).

Figure 1: Employment levels, 2019Q1 = 100
105

100 European Union

95
United States
90

85
2019Q1 2019Q2 2019Q3 2019Q4 2020Q1 2020Q2 2020Q3 2020Q4

Source: Bruegel based on Eurostat. Note: values are seasonally adjusted. The EU and US panels have different scales.

However, support given to firms has not been uniformly successful. The very generous
support has also meant that the level of bankruptcies has remained far below pre-pandemic
levels. The natural process of exits and entries has been affected without a clear indication
of what this will mean for the future of productivity. Has productive value been protected,
or have resources been spent to sustain otherwise unproductive firms? While new business
registrations appear to have returned to pre-pandemic levels, bankruptcies remain artificially
low (Figure 2)1. This could indicate that the support provided is keeping unproductive firms
alive. On the other hand, it is important to establish whether the recovery in the rate of new
business registrations (Figure 2) marks a return to market dynamism or ‘forced’ self-employ-
ment resulting from the loss of employment.

Figure 2: Firm entries and exits, average EU (index Q4 2019 = 100)
140

120
Business bankruptcies
100

80
Business registrations
60

2015Q1 2015Q2 2015Q3 2015Q4 2016Q1 2016Q2 2016Q3 2016Q4 2017Q1 2017Q2 2017Q3 2017Q4 2018Q1 2018Q2 2018Q3 2018Q4 2019Q1 2019Q2 2019Q3 2019Q4 2020Q1 2020Q2 2020Q3 2020Q4 2021Q1 2021Q2

Source: Bruegel based on Eurostat. Notes: Seasonally adjusted, EU data based on available countries.

1 According to the latest available bankruptcy update from the Halle Institute for Economic Research, the number of
corporate bankruptcies in Germany was still extremely low in August 2021.

2 Policy Contribution | Issue n˚21/21 | November 2021
As the rate of COVID-19 vaccination increases and the EU economy is steadily revived
from its artificially comatose state, we ask three questions:

• Which EU regions are likely to see a wave of bankruptcies? We look at the levels of zombie
firms pre-COVID-19 in European regions and how each of these regions has been affected
by the pandemic, in terms of falling trade or closures of businesses that relied on close
physical contact, such as services. We find that those countries/regions most likely to be
affected by COVID-19-related measures were also those that entered the pandemic with
higher shares of zombie firms.
• Has COVID-19 had a ‘cleansing effect’? Typically, less-productive firms are harder hit
during a recession. We look at preliminary data for a number of countries to understand
whether there has been a cleansing effect and observe that the least productive firms were
indeed hit hardest.
• Has financial aid been productive? Because financial aid had to be given out quickly,
it was given indiscriminately. In the trade-off between quickly reaching those in need,
compared to reaching only those that can survive and be productive, speed was opted for,
running the risk of inhibiting cleansing effects. The rationale was very much the preser-
vation of employment, not necessarily the preservation of productive capacity. Now the
recovery phase is underway, we can begin to evaluate whether financial aid ended up
with the most productive firms and if not, what the extent is of misallocation of funds. This
serves a dual purpose: evaluate the effectiveness of policies while helping to inform how
measures should be phased out, at what speed and with what sequence.

2 More than one year of financial support:
what next?
2.1 Which EU regions are most likely to see waves of bankruptcies?
To gauge the potential exposure of European regions to bankruptcies in the post-pandemic
context, we looked at two variables: first, the share of ‘zombie’ firms by region before the pan-
demic, and second an index of exposure of local sectors to economic disruptions stemming
from measures applied to reduce contagion risks. Those regions with both a larger share of
zombie firms and sectors relatively more affected by COVID-19 restrictions (such as restau-
rants affected by lockdowns, to give but one example), are more likely to face a larger shares
of bankruptcies as pandemic-related support measures are phased out and more normal
market conditions return.
In line with CompNet (2020), we defined a zombie firm as a company that reports negative
profits for three consecutive years, is more than 10 years old and is not characterised as ‘high
growth’, ie its workforce has not grown by more than 20 percent in the past three years. We
used Orbis company-level data to compute the share of zombie firms within each country
(region), sector and year. Zombie shares at the country (region) level were then computed as
averages across sectors, for the period 2015-20192.
We measured the exposure of sectors to pandemic-related contagion risks using the

2 For robustness, we also computed shares of zombie firms using an alternative definition: companies who report
interest coverage ratios below 1 for three consecutive years. Again, we restrict to companies that are 10 years old or
more and that are not ‘high growth’ (20 percent increase in employment over the past three years). We also look at
an alternative country (region) measure of zombie firms, computing the share only in specific industries that have
been negatively affected by the pandemic (accommodation and food service activities, administrative and support
service activities, education, human health and social work activities, arts, entertainment and recreation and other
service activities), with similar results.

3 Policy Contribution | Issue n˚21/21 | November 2021
Internal Disruption Risk Indicator (IDRI), which is an indicator of physical proximity of
persons in production and distribution processes3. The higher this indicator, the higher the
percentage of firms in an area in sectors affected by COVID-19-related restrictions and which
are therefore likely to experience difficulties.
For the largest EU economies (Germany, France, Italy, Poland and Spain) we looked at the
zombie shares and IDRI at regional level4. For the other EU countries, we looked at coun-
try-level data.
Figure 3 shows IDRI for each geographical area plotted against the share of zombie firms
operating in each area on average from 2015 to 2019. The size of each bubble is proportional
to each region’s GDP. Vertical and horizontal lines represent the median EU IDRI and share of
zombie firms, respectively (further charts with alternative definitions of zombie firms to check
robustness are available on request).

Figure 3: IDRI versus share of zombie firms (2015-2019 average)

Source: Bruegel based on ORBIS and IDRI.

A first striking result is the positive relationship between the IDRI score and the share of
pre-pandemic zombie firms (Figure 3). In other words, many of the regions strong in sec-
tors that were likely to be affected by COVID-19-related restrictions also had high shares of
zombie firms before the pandemic. Figure 4 summarises this: red for regions/countries with
high zombie shares/high prevalence of affected sectors, green for low zombie shares/low
prevalence of affected sectors, yellow for low zombie shares but high prevalence of affected
sectors, and orange for high zombie shares but low prevalence of affected sectors.

3 The IDRI uses an official classification of economic activity in terms of firms’ exposure to potential contagion
risk, as a function of the physical proximity required in the production and distribution process. The indicator
thus measures the extent to which local areas are affected by disruptions arising from the introduction of safety
regulations in the workplace, depending on their regional-level sectoral specialisation as quantified by labour
shares in each sector. For more details, and an application to all EU regions, see Altomonte et al (2020). The
Appendix details the formal construction of the index.
4 NUTS for Germany and NUTS2 for the other four. NUTS1 corresponds to German Länder, NUTS2 to regions,
départements, comunidades autónomas and voivodships in Italy, France, Spain and Poland, respectively.

4 Policy Contribution | Issue n˚21/21 | November 2021
Figure 4: A summary of vulnerabilities

Source: Bruegel. Note: Regions/countries in red have high zombie shares/high prevalence of affected sectors. Green = low zombie shares/
low prevalence of affected sectors. Yellow = low zombie shares/high prevalence of affected sectors. Orange = high zombie shares/low
prevalence of affected sectors.

One reason for this correlation is that firms operating in services sectors have been particularly
exposed to the economic disruptions stemming from the pandemic, as both production and dis-
tribution in these sectors typically require high levels of personal contact. At the same time, firms
in services sectors tend to be smaller and less productive, and hence more likely, all things being
equal, to become zombies.
Countries including Spain, Greece, Portugal and Cyprus are especially exposed, with high IDRI
scores and large shares of zombie firms. More than 4 percent of firms were zombies in virtually all
regions of Spain from 2015 to 2019. Although many Italian regions have high IDRI scores (because
of the importance of tourism and services sectors), the relative share of zombie firms is lower,
because of the significant restructuring that had already happened as a result of the debt crisis.

2.2 The impact of financial support on productivity: country results
Based on data for Italy, Germany and France, we analysed next whether COVID-19 affected
companies differently depending on their level of productivity. We also aimed to identify whether
support given was uniform across firms, and therefore indiscriminate, or in line with their levels
of productivity. We used the level of pre-crisis productivity, starting from the viewpoint that sup-
port should have gone primarily to firms that: i) were heavily affected by the pandemic but, ii) are
also worth rescuing with taxpayers’ money. Policies that direct support to good (ie productive)
firms are then preferable to indiscriminate policies.
Evidence shows that in all three countries, there have been cleansing effects. In Italy, firms
that were less productive before COVID-19 faced higher risk of default during the pandemic. A
few firms that were classified as productive became almost zombies, or ‘zombie-lite’ during the
pandemic. Similar results occurred in France and Germany. In France insolvencies caused by
the crisis have been concentrated in the least-productive firms. Nevertheless, some firms with
higher-than-average levels of productivity were also hit hard.
When it comes to the way public support was allocated, however, the picture is not entirely

5 Policy Contribution | Issue n˚21/21 | November 2021
the same in all countries. In France (section 2.2.3), public support measures appear to have had
a neutral impact on the distribution of insolvencies across all levels of firm productivity. In Italy
(section 2.2.1), public support appears to have been in line with productivity for medium and
large firms. However, support for the smallest firms was more strongly misallocated as, among
those firms, more support was given to relatively low-productivity firms. Last, in Germany
(section 2.2.2) the use of short-time work schemes appears to have been productivity-neutral,
while less-productive firms have been much more likely to receive financial support than highly
productive firms.

2.2.1 Italy
We merged firm balance-sheet information with data on the allocation of financial support
received during the COVID-19 crisis in 20205. We ended up with a sample of 1,030,899 SMEs
that received a total of €115 billion in guaranteed loans (the average public guarantee on the
loan was 85 percent). The sample represents 96 percent of all loans granted to SMEs in Italy,
covering firms employing 5.8 million people.
We also merged information on each firms’ credit risk score6 and divided firms into six
categories: safe; financially vulnerable but not impacted by the COVID-19 shock; finan-
cially vulnerable and impacted by the COVID-19 shock; financially risky and impacted by
the COVID-19 shock; ‘zombie-lite’, or firms brought to the brink of default by the COVID-19
shock; and zombie firms. Based on 2018 data, we calculated (median/average) productivity
both as sales per employee and value added per employee7.
Table 1 shows evidence of a ‘cleansing’ effect. As the credit risk increases, from ‘safe’ to
‘zombie’ firms, the level of productivity decreases. Firms that were less productive in 2018
turned out to be more at risk of default at the end of 2020. We show this result for both median
and average levels of productivity.

Table 1: Productivity and risk of default
Median productivity Average productivity
Value Value
Sales per Sales per
Credit risk score post-COVID-19 added per added per
employee employee
employee employee
Safe 134.5 39.3 264.1 53.5
Vulnerable non-impacted 117 30.8 218.1 41.6
Vulnerable impacted 90 29.7 158.2 41.6
Risky impacted 60.8 26.5 90.6 35.8
Zombie-lite 84 26 143.8 36.9
Zombie 87.7 23 143.1 33.2
Total sample of firms 108 31.7 213.2 45.3
Source: Bruegel based on CERVED data available in M&M (2021). Note: Productivity calculated based on 2018 data.

We then used information on the allocation of loans to check whether firms that were
more productive in 2018 received relatively higher intensities of loans (defined as the ratio of
loans to sales). We divided firms into five quintiles of productivity and four size categories,
based on turnover (Table 2).
Table 2 shows that loan intensity appears to be broadly in line with productivity (higher
intensity for higher quintiles of productivity), for all firm size categories, with the exception
of the smallest firms (turnover below €1 million). In this group, the least-productive firms

5 Working with CERVED, a data-analysis company.
6 An index calculated by CERVED and used by banks to assess credit risk.
7 For more details, see Altomonte (2021).

6 Policy Contribution | Issue n˚21/21 | November 2021
received relatively greater loan intensities. Thus, while support has gone to unproductive
larger firms, the degree of potential misallocation of funds was most significant within the
group of smallest firms.

Table 2: Loans and productivity (median of intensity of loan)
Size class (turnover)
Productivity (from least Average
< €1m €1m-€5m €5m-€10m > €10m
to most)* total
1st quintile 15.6% 6.9% 8.9% 6.3% 14.9%
2nd quintile 11.6% 8.4% 8.9% 8.8% 11.1%
3rd quintile 10.4% 10.2% 11.1% 9.4% 10.3%
4rth quintile 10.2% 11.6% 12.3% 10.1% 10.8%
5th quintile 10.9% 12.6% 13.0% 9.5% 11.3%
Total sample of firms 12.2% 10.9% 12.0% 9.5%
Source: Bruegel based on CERVED data available in M&M (2021). Notes: *Productivity from least to most, measured as valued-added per
employee (2018). Intensity of loan defined as the ratio of loan to sales.

This potential misallocation to the smallest firms can be explained by a particular feature
of the public guarantee in Italy, by which loans up to €25,000 received a 100 percent public
guarantee, without any bank scrutiny (compared to a standard guarantee of 80 percent of public
coverage). Standard loans (80 percent guarantee with bank screening) made up roughly 85
percent of loans, and appear to have been allocated in line with productivity (also controlling for
firm size). Micro-loans (<€25,000) with a 100 percent guarantee (no bank screening), amounted
to €21.7 billion, or 15 percent of total loans, and mostly went to very small companies with turn-
overs below €1 million (90 percent of these loans went to micro companies).
Italian firms with more than 10 employees, constituting around 11 percent of total firms in
the sample, are relatively productive and have received 64 percent of the total funds. Firms with
fewer than five employees, some relatively unproductive, constitute 76 percent of the sample,
and have received 23 percent of total funds.
There are two ways to evaluate the economic impact of possible defaults as public support
measures are phased out. The first is to look at the financial loss equal to the value of the loan
given to the company that defaults. Table 3 shows that the value of all loans at risk is estimated
to be just below €9 billion, 7.8 percent of total loans in the sample. Zombie-lite and zombie
firms account for just over €5 billion. There is therefore a limited risk to financial stability, as a
result of phasing out financial support.

Table 3: Risk of default and its economic cost
Loans (€ Loans at risk Employees
Employees No. of firms
millions) (€ millions) at risk
Safe 45,969 776 2,239,115 38,803 308,172
Vulnerable non-
31,726 2,211 1,361,157 96,200 284,516
impacted
Vulnerable
13,901 759 864,230 48,231 181,884
impacted
Risky impacted 727 137 55,410 9,970 9,347
Zombie-lite 14,008 2,619 789,239 157,989 166,071
Zombie 8,731 2,486 477,471 149,784 80,909
Total sample of
115,062 8,988 5,786,622 500,976 1,030,899
firms
Source: CERVED Group estimates, as reported in M&M (2021). Note: Loans and employees at risk calculated applying to each firms its probability of
default based on the credit risk score.

7 Policy Contribution | Issue n˚21/21 | November 2021
The second way to assess the impact is to calculate the jobs at risk, should a certain
number of firms default. This is potentially much more serious than the financial impact, with
500,000 people at risk from corporate defaults. Most of these jobs are in zombie or zombie-lite
firms that also the least productive. More specifically, 23 percent of total jobs at risk come
from the hospitality industry (food service and hotels), which represents only 12 percent of
firms and only 5 percent of loans, but 28 percent of all zombie-lite firms. Also note that the
employment estimate is a lower bound, as it does not include those employees who might
be laid off by surviving firms in order to restore profitability, as short-time work schemes are
phased out.

2.2.2 Germany
A quarter of German firms claimed to have been hit hard by the first wave of the pandemic8.
Sectors similar to those in other countries, in particular hospitality, have been affected by
social distancing measures (Figure 5), although these sectors are not as central to Germany’s
economy as they are for other countries, such as Austria and Italy. What did matter for Ger-
many however, was the initial breakdown of international value chains, particularly for the
strongly international integrated manufacturing sector.

Figure 5: Percentage of firms hit hard by the crisis (by sector)

Hospitality 65%
Other personal services 48%
Education 46%
Agriculture, foresty, fishing 10%
Construction 8%
Mining 2%

0% 10% 20% 30% 40% 50% 60% 70% 80%
Source: Bruegel based on IAB establishment panel.

Initially, about one in eight firms self-assessed as close to market exit because of the pan-
demic, with large variations by German Länder. Urban centres and the west of the country
were more affected in terms of firms coming close to exiting the market (Figure 6).

8 All calculations for Germany are based on the IAB establishment panel 2020 run by the Institute for Employment
Research (IAB) and refer to July-September 2020 (see https://fdz.iab.de/en/FDZ_Establishment_Data/IAB_
Establishment_Panel.aspx). We weight the data using cross sectional weights provided by the IAB. A general
description of the data set is provided by Ellguth et al (2014).

8 Policy Contribution | Issue n˚21/21 | November 2021
Figure 6: Percentage of firms close to market exit (by region)

16.0 15.1
10.1
9.6
12.8
15.1
7.4 11.1
13.2
8.2
11.6 9.1

9.4

13.9 12.5
10.6

Source: Bruegel based on IAB establishment panel.

As in Italy, less-productive German firms were hit hardest. Figure 7 ranks firms by their
pre-pandemic labour productivity levels. One third of the least-productive firms (lowest
quartile) reported being hit hard, whereas this was only the case for about one fifth of the
most-productive firms. Similarly, 19 percent of the least-productive firms were close to exit,
compared to just 7 percent of the initially most-productive firms9. While this shows the
potential for cleansing effects, it also tells a cautious story about the potential of losing very
productive firms. A careful policy would certainly try to avoid losing 7 percent of the country’s
most-productive firms.

Figure 7: Impact of the COVID-19 pandemic by 2019 labour productivity quartile
40%
Strongly affected
30% Close to exit

20%

10%

0%
Bottom 25% 2nd quartile 3rd quartile Top 25%
Source: Bruegel based on IAB establishment panel.

Figure 8 shows how government aid was distributed with respect to initial firm
productivity. In the figure, we distinguish between financial aid and the use of the generous
short-time work scheme because the motives for take-up differ. While financial aid is more
like emergency help, short-time work schemes are attractive for firms that try to maintain
their well-trained workforces. Hence, short-time work schemes will more likely help relatively
productive firms with high hiring (training) costs. Figure 8 supports these conjectures by

9 The link between productivity and both outcomes still holds when plant size and sector/region fixed effects are
netted out in a regression setting.

9 Policy Contribution | Issue n˚21/21 | November 2021
showing that financial aid was taken up predominantly by unproductive firms. By contrast
there is no strong relationship between the uptake of short-time work schemes and the level
of firm productivity10.

Figure 8: State aid in Germany by 2019 labour productivity quartile

80%
Short-time work
60% Financial government support

40%

20%

0%
Bottom 25% 2nd quartile 3rd quartile Top 25%

Source: Bruegel based on IAB establishment panel.

We do not have micro data on the amount of German government loans so are unable to
evaluate the potential financial implications of company defaults. However, we can evaluate
the risk using employment data. Although a higher fraction of unproductive firms was hit
hard, they typically employ fewer people. Table 4 confirms this by showing that the relatively
few high-productivity firms hit hard by the pandemic employed 50 percent more workers
than the many low-productivity firms that were also hit hard (Table 4, left panel; 2.26 million
compared to 1.4 million). In each of the productivity quartiles, about 500,000 workers work
in firms that are close to exit (right side of Table 4). This means that if expressed as number
of jobs instead of number of employers, the pandemic primarily hit good jobs in productive
firms. As these firms are more resilient to crises, the number of jobs in firms that are close to
exit is however more evenly spread over the firm productivity distribution.

Table 4: Affected jobs by firm productivity (in million workers)
Hit hard by pandemic Close to exit due to pandemic
No Yes No Yes
Labour productivity 2019 Mln workers Mln workers Mln workers Mln workers
Bottom 25% 2.76 1.40 3.62 0.54
2nd quartile 3.02 1.03 3.64 0.41
3rd quartile 3.90 1.37 4.85 0.42
Top 25% 7.88 2.26 9.60 0.54
Total 17.60 6.05 21.70 1.90
Source: Bruegel based on IAB establishment panel.

10 In fact, when controlling for establishment size and sector/region fixed effects affiliation in a regression
framework, the productivity gradient prevails for financial aid and is non-existent for take-up of short-time work.
This implies that financial aid is more likely than short-time work schemes to generate zombie firms.

10 Policy Contribution | Issue n˚21/21 | November 2021
2.2.3 France
The distribution of public support and its effectiveness in France were scrutinised in a July
2021 report prepared by the Committee on the Monitoring and Evaluation of Financial Sup-
port Measures for Companies Confronted by the COVID-19 Epidemic (Cœuré, 2021; Cœuré
committee hereafter)11. A combination of four main emergency measures was implemented:
a short-time work scheme, a ‘Solidarity Fund’, state-guaranteed loans and deferral of social
security contributions. Commercial courts were also closed temporarily, which de facto im-
pacted bankruptcies. Finally, additional measures were introduced in the wake of the second
virus wave, including provision of compensation for companies’ fixed costs, corporate tax loss
carry-backs, and, even more specifically, compensation for the stopping of ski lifts in winter
resorts. In order to maximise eligibility, this support is not making it conditional on various
criteria. This has the consequence of potential windfall effects, a choice that was acknowl-
edged in implementing the measures.
To evaluate the measures, the Cœuré committee matched firms’ balance sheets, VAT data
and other administrative information for the 3.5 million businesses that received support, for
the period from 1 April 2020 to 30 March 2021. The benefitting companies received €53 billion
in subsidies and €148 billion in loans.
In France as in other European countries, notwithstanding differences in the details
of implementation of public support, the combination of state-guaranteed loans and the
temporary halt in investment led to an increase in the cash position of the businesses that
received support. Business insolvencies reduced by 39 percent in 2020 compared to 2019.
Bénassy-Quéré et al (2021) showed that while the percentage of all firms entering insolvency
increased from 3.6 percent to 6.6 percent from March to December 2020, this increase would
have been to 11.9 percent without the measures.
Overall, the measures managed to cushion the impact of the pandemic and the lockdown
on firms’ gross operating surpluses. There was a €45 billion drop in the operating profits
in 2020, compared to 2019. A drop in their gross operating surplus was experienced by 59
percent of French companies in 2020. Estimates from the Cœuré committee suggest that
this number would have been 72 percent without the support schemes. These measures
prevented any drop in the surplus during the second wave. The hospitality sector received
the most support (worth up to 99 percent of its pre-crisis gross operating profits). Three
sectors actually saw their surpluses increase during the crisis (IT, agriculture and household
services).
A good indicator of the potential impact of the short-term work scheme is the ratio of
hours compensated for under the scheme to total paid hours. This ratio peaked at 29 percent
in the private sector in April 2020 and went down to 5 percent in May 2021.
On the characteristics of firms that received support, small firms in France contribute an
18 percent share of private employment but have received a much larger share of support.
Small businesses benefitted from 63 percent of the payments from the solidarity fund, 49
percent of the deferrals of social contributions, 33 percent of payments under the short-time
work scheme, and 29 percent of state-guaranteed loans.
In terms of financial health, the evidence is mixed. Firms that were low and high profit-
ability before the pandemic are underrepresented in the sample of supported businesses.
However, the lowest profitability firms received relatively larger amounts (support relative to
turnover before the pandemic). This mixed evidence suggests low-productivity firms had lim-
ited access to the schemes, but those that did receive support were relatively more supported.
Zombie firms made up 7.2 percent of French firms in the private sector (excluding finance
and agriculture) in 2018, but only half of zombie firms (3.4 percent of supported firms) benefit-
ted from at least one measure during the first wave of the pandemic (during the second wave,
the figure was 3 percent). But using a different criterion more representative of firm size – the

11 Comité de suivi et d’évaluation des mesures de soutien financier aux entreprises confrontées à l’épidémie de Covid-19.
See Cœuré (2021) and Bénassy-Quéré et al (2021).

11 Policy Contribution | Issue n˚21/21 | November 2021
number of employees – support has been aligned with the share of zombie firms in total employ-
ment. The same conclusion pertains to the share of zombie firms in the capital stock or in value
added.

Table 5: Share of zombie firms in the French economy and support received
Share of all firms
Share of all firms that
receiving support that
are zombies (2018)
are zombies
First wave 3.4 7.2
Number of firms
Second wave 3.0
First wave 6.8 7.5
Number of employees
Second wave 6.2
First wave 7.2 8.1
Capital (incl. intangibles)
Second wave 6.7
First wave 3.7 4.0
Value added
Second wave 3.4
Source: adapted from Cœuré (2021), p. 282.

On the whole, evidence from France, covering more than 6 million firms, 3.5 million of
which received COVID-19 aid, shows that the supported firms were also those that needed
it most. The evidence also showed that not all firms requested the help to which they were
entitled and zombie firms were not disproportionately supported. Support helped prevent
insolvencies, but there is not enough evidence to conclude whether or not support boosted
firms’ subsequent performances.
Evidence from smaller countries indicates similar outcomes to France. On Croatia, Fin-
land, Slovakia and Slovenia, Bighelli et al (2021, 2021a) reported that government support was
provided efficiently. Most of the help given reached medium-sized productive firms, and only
marginal funds were ‘wasted’ on unproductive (zombie) firms. Bighelli et al (2021, 2021a)
argued however that funding was not sufficient to reverse the big productivity decline caused
by the pandemic. While initial help was there to maintain employment, this support, they
argued, will have to be phased out as early as possible, so that new support can help restore
and increase firm productivity.

3 Literature on phasing out support
measures and financial stability
One of the main sectors offered support during the pandemic has been the banking sector.
Specific measures included moratoria on loan payments, public guarantees and capital relief
measures. When and how should these measures be unwound and what risks does that
entail?
A number of papers have attempted to answer these questions. Haselmann and Troger
(2021) used German micro-level data and predicted significant capital shortfalls, which could
jeopardise investor confidence and fuel financial instability. They argued in favour of ending
accounting practices that conceal loan losses and suggested that while banks with no realistic
prospect of meeting regulatory capital requirements should be forced to exit the market,
others should be recapitalised (taking the US Troubled Asset Relief Program as an example).
Lehmann (2021) noted that, by the third quarter of 2020, payment moratoria covered 6.4

12 Policy Contribution | Issue n˚21/21 | November 2021
percent of euro-area corporate loan stock, which will inevitably be associated with substan-
tial credit risk. He argued that, among other things, the supportive measures put the scrutiny
of banks’ management of non-performing loans at risk and the ECB should monitor banks’
practices more closely in terms of the credit risk of individual borrowers. Finally, Beck et al
(2021) agreed that restoring banks’ balance sheet transparency is a priority. Before unwinding
capital relief measures, authorities should first phase out borrower relief measures and relax
loan classification. Importantly, phasing out of measures should be communicated clearly to
enable banks to adjust balance sheets (Beck et al, 2021).
The European Commission’s Directorate-General for Financial Stability, Financial Ser-
vices and Capital Markets Union (European Commission, 2020) argued in favour of reforming
insolvency frameworks and further developing secondary markets for distressed assets in
order to move non-performing loans off banks’ balance sheets. The European Commission
further believes that the EU Recovery and Resilience Facility can support reforms that aim to
reduce non-performing loans, while stressing that market-based solutions should remain the
first and primary tool.
The Financial Stability Board (FSB, 2021) noted that, regarding financial stability in gen-
eral, there is a trade-off between procyclical risks in case of sudden withdrawal of support
measures and the progressive development of financial stability risks linked to these support
measures (distortion of resource allocation and asset prices, moral hazard, deterioration of
credit quality). Authorities should ensure that measures are targeted by, for instance, requir-
ing beneficiaries to opt in. Generally, terms should be made progressively less generous and
authorities need to clearly communicate about the sequence of withdrawal of support.
While Laeven et al (2020) noted the risk of zombification, they argued that the COVID-19
crisis is different to previous crises because it also hit viable sectors. They argued that in order
to avoid the zombification of the economy, credit guarantees need to be fine-tuned, gov-
ernments should incentivise financing through equity instead of debt, and that supervisory
authorities need to ensure that banks maintain sound capital positions. More generally, they
argued in favour of improving the efficiency of the insolvency framework and bankruptcy
laws. Helmersson et al (2021) looked at firm-level and loan-level data and found that zombies
only benefitted from loan schemes to a modest degree. Helmersson et al (2021) agreed that
insolvency frameworks should be reformed.
Finally, the OECD (2021) argued that countries need to provide clear roadmaps on path-
ways to recovery. Among other things, they suggested that countries should provide support
for search and career guidance and training activities for employees (and more specifically in
the sector of digitalisation).

4 Conclusions and way forward
European regions and countries that depend heavily on services requiring physical proximity
have been hit hardest by COVID-19-related measures. But these services sectors tend also to
be the smallest and least-productive in any economy, implying that, coming into the crisis,
the highest shares of zombie firms were in these sectors. Reliance on physical proximity and
the higher incidence of zombies to start with combine to make those services-dependent
economies particularly vulnerable to any attempt to remove the support put in place during
the pandemic.
Nevertheless, as countries recover, there is a need to remove support and allow markets to
return to picking natural winners and losers.
The evidence shows that the main goal of the provision of support during the COVID-19
crisis, namely to protect employment, was achieved. However, the evidence is varied on how
efficiently this was done, in terms of helping firms that have a good chance of surviving, while

13 Policy Contribution | Issue n˚21/21 | November 2021
not supporting those that will in any case exit.
The design of schemes matters, as shown in both Italy and Germany where support was allo-
cated in line with firm productivity.
Partial, not full, public guarantees for loans gave banks an incentive to carry out due diligence
and identify those firms most likely to survive. In Italy, this meant that the bulk of loans given
were in line with productivity. This also implies that the probability of default is lower for large
firms and as only small loans were potentially misallocated, a wave of defaults among small firms
will not pose a financial-stability problem.
Similarly, in Germany, short-time work schemes reached the most-productive firms. The big
and productive firms with the best-trained workforces had an incentive to use these schemes to
try to maintain their workforces.
In thinking of how to move forward, there are broader issues to consider. The
in-line-with-productivity argument implies that it is advisable to support large rather than small
firms. Big firms are more likely to be productive. Also wage losses associated with being made
unemployed are greater for workers leaving large firms (Jacobson et al, 1993; Fackler et al, 2021).
However, small firms also employ collectively a large number of people. Phasing out support
and letting small firms default may imply, as our evidence has shown, substantial employment
losses. Moreover, this would go against the rationale of the indiscriminate support for employ-
ment given in the initial phase of the crisis. On top of economic efficiency, other issues must
thus be considered, given that small firms play a very important role in terms of social cohesion.
Minimising the cost to the taxpayer needs to be balanced carefully against the importance of
maintaining employment.
Tax incentives could be one concrete way of balancing the return to functioning markets
while protecting viable firms as support is phased out. Firms are usually entitled to count current
losses against previous profits when filing corporate tax returns. Making this scheme much more
generous would help precisely those firms that should be protected – those that incurred losses
because of the pandemic but were profitable before the crisis. In contrast, all undirected support
schemes, including cash injections and exemptions from the obligation to file for bankruptcy,
can be distortionary and should be removed early on. Exceptions for young firms, which by defi-
nition cannot count today’s losses against previous profits, should be part of the policy toolbox.
The French data shows that support given during the pandemic crisis has been productiv-
ity-neutral: it did not favour any specific type of firm and therefore did not alter the dynamism
of the French economy. But the COVID-19 shock may have been transformative in terms of the
long-term impact on productivity. Advances in digital technologies and their accelerated take-up
during the pandemic are behind the observation of Brynjolfsson et al (2021), who showed that
most OECD countries still have to gain from digitalisation in terms of productivity. Building on
this argument, Brynjolfsson and Petropoulos (2021) were very optimistic that the combination
of accelerated take-up in digital technologies during the pandemic and massive macroeconomic
support will boost productivity.
Arguably, COVID-19 has forced firms to become more efficient. Firms caught in sudden and
prolonged shutdowns of the economy have had to optimise processes, cut costs and become
more efficient. This has meant that firms have had to become more innovative, and to digitise
and automate as much as possible. Those who managed it thus increased their overall agility. In
a survey reported by Maqui and Morris (2020), 75 percent of firms agreed that the pandemic has
helped make their businesses more efficient and resilient. Nine out of 10 firms had sped up the
adoption of digital technologies and automation.
If this gives rise to new and more productive firms, then a higher degree of employment
churning and re-skilling should perhaps be encouraged, as a way of adapting to new techno-
logical needs. Coupled with short-term unemployment support, this might be a better method
of creating new jobs sustainably but not at the expense of welfare in the short run. It may be too
soon to conclude that COVID-19-induced changes will necessarily lead to higher productivity
but longer-term changes will also need to guide short-term measures in terms of phasing out
support.

14 Policy Contribution | Issue n˚21/21 | November 2021
References
Altomonte, C. (2021) ‘The impact of COVID-19 on productivity: preliminary firm-level evidence from
Italy’, mimeo

Altomonte, C., A. Coali and G. Ottaviano (2020) ‘EU recovery plans should fund the COVID-19 battles
to come; not be used to nurse old wounds’, Bruegel Blog, 6 July, available at https://www.bruegel.
org/2020/07/eu-recovery-plans-should-fund-the-covid-19-battles-to-come-not-be-used-to-nurse-
old-wounds/

Arriola, C., P. Kowalski and F. van Tongeren (2020), ‘Localising value chains in the post-COVID world
would add to the economic losses and make domestic economies more vulnerable’, VOXEU, 15
November, available at https://voxeu.org/article/localising-value-chains-after-covid-would-add-
economic-losses-and-make-domestic-economies-more-vulnerable

Beck, T., B. Bruno and E. Carletti (2021) ‘When and how to unwind COVID-support measures to the
banking system?’ In-Depth Analysis Requested by the ECON committee, European Parliament,
available at https://www.europarl.europa.eu/RegData/etudes/IDAN/2021/659646/IPOL_
IDA(2021)659646_EN.pdf

Bénassy-Quéré, A., B. Hadjibeyli and G. Roulleau (2021) ‘French firms through the COVID storm:
Evidence from firm-level data’, VOXEU, 27 April, available at https://voxeu.org/article/french-firms-
through-covid-storm

Bighelli, T., T. Lalinsky and CompNet Data Providers (2021) ‘COVID-19 government support and
productivity: Micro-based cross-country evidence’, CompNet Policy Brief No. 14

Bighelli, T., T. Lalinksky and F. di Mauro (2021a) ‘Covid-19 government support may have not been as
unproductively distributed as feared’, VOXEU, 19 August, available at https://voxeu.org/article/covid-
19-government-support-may-have-not-been-unproductively-distributed-feared

Brynjolfsson, E. and G. Petropoulos (2021) ‘The coming productivity boom’, Bruegel Blog, 10 June,
available at https://www.bruegel.org/2021/06/the-coming-productivity-boom/

Brynjolfsson, E., D. Rock and C. Syverson (2021) ‘The Productivity J-Curve: How Intangibles Complement
General Purpose Technologies’, American Economic Journal: Macroeconomics Vol 13, No. 1: 333-72

Cœuré B. (2021) Report of the Committee on the Monitoring and Evaluation of Financial Support
Measures for Companies Confronted with the Covid-19 Epidemic, Inspection Générale des Finances &
France Stratégie

Comptnet (2020) Firm Productivity Report, May, available at https://www.comp-net.org/fileadmin/_
compnet/user_upload/Documents/Productivity_Report_FINAL-.pdf

Ellguth, P., S. Kohaut and I. Möller (2014) ‘The IAB Establishment Panel—methodological essentials and
data quality’, Journal for Labour Market Research 47(1-2): 27-41

European Commission (2020) ‘Action Plan: Tackling non-performing loans in the aftermath of the
COVID-19 pandemic’, COM/2020/822 final, available at https://eur-lex.europa.eu/legal-content/EN/
TXT/?uri=CELEX:52020DC0822

Fackler, D., S. Müller and J. Stegmaier (2021) ‘Explaining Wage Losses after Job Displacement: Employer
Size and Lost Firm Wage Premiums’, Journal of the European Economic Association 19(5): 2695-2736

FSB (2021) COVID-19 support measures. Extending, amending and ending, 6 April, Financial Stability
Board

Haselmann, R. and T. Troger (2021) ‘When and how to unwind COVID-support measures to the banking
system?’ In-Depth Analysis Requested by the ECON committee, European Parliament, available at
http://www.europarl.europa.eu/RegData/etudes/IDAN/2021/659649/IPOL_IDA(2021)659649_
EN.pdf

15 Policy Contribution | Issue n˚21/21 | November 2021
Helmersson, T., L. Mingarelli, B. Mosk, A. Pietsch, B. Ravanetti, T. Shakir and J. Wendelborn (2021)
‘Corporate zombification: post-pandemic risks in the euro area’, Financial Stability Review, May 2021,
European Central Bank

INAIL (2020) Dossier Speciale COVID-19, Istituto Nazionale per l'Assicurazione contro gli Infortuni sul
Lavoro, available at https://www.inail.it/cs/internet/docs/alg-pubbli-dossier-speciali-covid-19_2021.
pdf

Jacobson, L.S., R.J. LaLonde and D.G. Sullivan (1993) ‘Earnings losses of displaced workers’, The American
Economic Review 83(4): 685-709

Laeven, L., G. Schepens and I. Schnabel (2020) ‘Zombification in Europe in times of pandemic’, VOXEU,
11 October, available at https://voxeu.org/article/zombification-europe-times-pandemic

Lehmann, A. (2021) ‘When and how to unwind COVID-support measures to the banking system?’
In-Depth Analysis Requested by the ECON committee, European Parliament, available at https://
www.bruegel.org/wp-content/uploads/2021/03/IPOL_IDA2021659636_EN.pdf

Maqui, E. and R. Morris (2020) ‘The long-term effects of the pandemic: insights from a survey of leading
companies’, ECB Economic Bulletin 8/2020, European Central Bank, available at https://www.ecb.
europa.eu/pub/economic-bulletin/focus/2021/html/ecb.ebbox202008_06~bad87fcf9b.en.html

OECD (2021) One year of SME and entrepreneurship policy responses to COVID-19: Lessons learned to
“build back better”, Organisation for Economic Co-operation and Development

M&M (2021) Dall’emergenza al rilancio: Misure per le imprese, le filiere e il lavoro, M&M - Minima
Moralia, available at http://www.associazionemandm.org/storage/media/2021/06/186/
dall%27emergenza-al-rilancio---imprese-filiere-lavoro-%28no_fil%29.pdf

Appendix
Technical note on the computation of the Internal Disruption Risk Indicator (IDRI)
IDRI is calculated for a given area by leveraging a detailed measure of risk exposure devel-
oped at sector-level by Inail (2020). Inail classified industries at the ISIC/NACE 2-digit level
according to three characteristics:

• Contact: the probability of social contact while at work premises (the index has a value
from 0, denoting low probability, eg in the case of agriculture, to 4, denoting high-proba-
bility, eg in the case of nurses);
• Proximity: the intrinsic characteristics of the workflow that do not allow for sufficient so-
cial distancing (the index value runs from 0, denoting working alone, to 4, denoting work
in close contact with other people);
• Aggregation: the level of contact with subjects other than the firm’s workers (eg restau-
rants, education, also with values from 1 to 4).

Publicly available data include a combined index that considers the two measures of
contact and proximity (riskclass), plus the aggregation index (weightaggregation). We use the aggre-
gation index (weightaggregation) to weight the first risk measure (riskclass). In doing so, we assign a
value ranging from 1 to 1.5 to each social-aggregation risk class (i.e. to weightaggregation), follow-
ing Inail (2020). The final risk measure at the NACE 2-digit level is thus:

riskweighted = riskclass * weightaggregation

We then built a regional-level measure by using the labour share for each sector within the region as
weights. Denoting by a the region and by j the sector, IDRI is then given by:

16 Policy Contribution | Issue n˚21/21 | November 2021

File and source

File: covid-19-financial-aid-and-productivity-has-support-been-well-spent-pb.pdf
Size: 792,024 bytes
SHA-256: 141233f386160c30cc1a8ca9c7a0671769a552d0cb4998c16faa1ce7cdfee276
Our copy: covid-19-financial-aid-and-productivity-has-support-been-well-spent-pb.pdf
Original: www.bruegel.org