Marius Brülhart et al., COVID-19 financial support to small businesses in Switzerland: evaluation and outlook (2020)
Summary
Article by Marius Brülhart, Rafael Lalive, Tobias Lehmann and Michael Siegenthaler in the Swiss Journal of Economics and Statistics (2020) on how small businesses in Switzerland used COVID-19 support. A survey of 1011 self-employed workers and small business owners shows that lockdown exposure and economic structure explain use of support for labour costs fairly well. Take-up of government-backed corona loans followed firm fundamentals less closely: previously indebted businesses borrowed more readily than debt-free ones. The authors propose making loan repayment contingent on future profits. The paper notes the partial lockdown began on 16 March, 2020 and that demand fell by more than 30%.
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
Cite this document
“Marius Brülhart et al., COVID-19 financial support to small businesses in Switzerland: evaluation and outlook (2020).”
Original at api.unil.ch. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/marius-br-lhart-et-al-covid-19-financial-support-to-small-businesses-i-9e4230449c/
, SHA-256 287ccc5025c3820f934716936f5dac940e4f9ae15c016164621711e2bb350195.Full text
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Swiss Journal of
https://doi.org/10.1186/s41937-020-00060-y
Economics and Statistics
ORIGINAL ARTICLE Open Access
COVID-19 financial support to small
businesses in Switzerland: evaluation and
outlook
Marius Brülhart1, Rafael Lalive1*, Tobias Lehmann1 and Michael Siegenthaler2
Abstract
We analyse small businesses’ recourse to public support measures during the COVID-19 crisis using a survey
of 1011 self-employed workers and small business owners in Switzerland. We find that “objective” measures of
lockdown affectedness and economic structure explain fairly well how businesses availed of support measures
to cover labour costs. Recourse to government-backed corona loans, however, appears to be driven to a
larger extent by behavioural idiosyncrasies across firms. Specifically, previously indebted businesses took out
corona loans more readily than those who had been debt-free before the pandemic. Since uptake is not well
in line with firm fundamentals, we propose making loan repayments contingent on future profits. This will
more effectively target and sustain businesses that are in trouble today but would be viable in the absence
COVID-19.
1 Introduction estimated that total demand fell by more than 30% dur-
Faced with rapidly increasing Covid-19 infections, the ing the Swiss lockdown, and GDP is projected to con-
Swiss government decreed a partial lockdown on 16 tract by more than 5% over the whole year2. Demand
March, 2020. Non-essential shops, restaurants, enter- will likely remain subdued at least in certain sectors for
tainment venues, schools, nurseries and universities were as long as social distancing is warranted and foreign
closed, and non-essential workers were asked to stay at markets are depressed. This state of affairs could poten-
home and work from there if possible. The lockdown tially take several years until a vaccine or an effective
was gradually relaxed after 27 April, with most restric- cure is found.
tions lifted by 8 June. For many businesses, the pan- The combination of a time-limited forced shutdown of
demic meant a dramatic or even complete loss of large parts of the economy followed by a potentially long
turnover. Recent research shows that overall spending lasting drag on contact-intensive activities presents eco-
would have been severely reduced even in the absence of nomic policy-makers with monumental challenges. The
a decreed lockdown, because of consumers’ fear of infec- main goals in the initial lockdown period were to avoid
tion and because of a collapse in export demand1. It is excessive income losses by workers in affected sectors and
to avoid damage to the productive structure of the econ-
1
See e.g. Brzezinski et al. (2020). omy through large-scale layoffs and bankruptcies.
Switzerland has been widely credited for meeting these
* Correspondence: Rafael.Lalive@unil.ch aims successfully, through three essential policy tools: wage
We thank Cédric Tille (the editor), an anonymous referee, Jonathan Chanel
and Guillaume Rais for helpful comments. Marius Brülhart and Rafael Lalive compensation for furloughed employees (short-term work,
are members of Swiss National COVID-19 Science Task Force. “Kurzarbeit”), income support for independent workers and
1
Department of Economics, HEC Lausanne, University of Lausanne, Lausanne, small-business owners, and state-backed loans to cover
Switzerland
Full list of author information is available at the end of the article
2
See SECO (2020) and KOF (2020).
© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 2 of 13
businesses’ liquidity needs (corona loans). Rapid and un- respondents are based in German-speaking Switzerland,
bureaucratic intervention was possible thanks to Switzer- 26% in French-speaking Switzerland and 11% in Italian-
land’s exceptionally healthy public finances, the prior exist- speaking Switzerland. Fifty-nine percent of respondents
ence of some of the relevant policy tools, a functioning are male, and the average age is 57. Fully 31% of sur-
banking sector, and a consensual, inclusive system of veyed businesses had availed of public subsidies for
government. short-term work of employees, and 33% had availed of
In this paper, we shed light on the effectiveness of income replacement for independent workers and small
these tools in the early phase of the lockdown. To this business owners. Eighteen percent of sample businesses
end, we analyse data from a survey of 1011 independent had availed or intended to avail of corona loans4.
workers and small- and medium-size firms conducted in The aim of our data analysis is to explore the determi-
mid-April 2020. nants of small businesses’ recourse to those different
From this experience, we derive some implications for sources of financial support. We consider three sets of
how economic policy measures could be adapted if the explanatory factors. All but one of the regressors are
virus were to continue to constrain contact-intensive eco- converted into binary variables for ease of interpretation:
nomic activities or even necessitate further lockdowns.
1. Lockdown variables: These variables are included to
2 Data and empirical model capture the impact of the pandemic on the activity
Our empirical analysis builds on a survey conducted of the sample businesses. They should therefore
with LINK Institute, an established polling organisation correlate strongly with those businesses’ use of
in Switzerland. LINK have a nationwide, representative support measures. We consider the following
online participant sample of some 115,000 individuals, lockdown variables:
from which subsamples can be drawn for specific sur- 1.1 Whether a business was locked down de jure
veys. One attractive feature of this sampling environ- (i.e. by government decree),
ment is that a lot of baseline information is known 1.2 Whether a business was locked down de facto
about the participant population, which can then be cor- (i.e. in principle allowed to operate and open,
related with the sample of individuals answering a spe- but severely restricted in its activity because of a
cific survey3. lack of custom and/or because of sanitary
The survey sample underlying this paper was drawn regulations),
from three sub-populations, as self-reported by the par- 1.3 The share of April 2020 turnover that
ticipants: the self-employed, the liberal professions and respondents estimate to have lost due to the
firm owners. The common denominator across these pandemic (the only non-binary variable),
three sub-populations is that they can be considered as 1.4 Whether physical proximity amongst workers is
“job creators”—be it only for themselves or also for important for the activity of the business, and
others. We shall refer to this population as “small busi- 1.5 Whether physical proximity to customers is
nesses” or “businesses” in the remainder of this paper. important for the activity of the business.
The survey was conducted between 14 and 20 April 2. Economic variables: These variables are included to
2020. We had a response rate of 36%, which is satisfac- capture pre-crisis characteristics of the sample busi-
tory in this context. An analysis of non-response shows, nesses that can be expected to affect the extent to
perhaps surprisingly, no statistically significant differ- which they avail of different support measures.
ences in response rates across economic sectors and They are mainly designed to capture the financial
only minor differences in terms of individual characteris- capacity of small businesses and are therefore also
tics. We find some signs of lower non-response rates by expected to correlate with businesses’ use of sup-
older individuals and by people with higher educational port measures. We consider the following economic
qualifications, but overall we are confident of the repre- variables:
sentativeness of our data sample (see also Brülhart, 2.1 Employment (1, between 2 and 9, or more than
Klaeui, Lalive, Lehmann, & Siegenthaler, 2020b). 10 full-time equivalent jobs, (FTE)),
In total, we have answers from 1011 small businesses. 2.2 Debt ratio in 2019 (0, between 1% and 25% or
Table 1 shows summary statistics. Sixty-three percent of more than 25% of own capital),
3
This is a methodological advantage compared to one-off surveys with
retrospective questions, such as that used in a related study by Zoller-
4
Rydzek and Keller (2020). Their survey covered 205 managers of Swiss Respondents had four options. The percentages shown here are
firms in the second week of April 2020. Due to differences in data col- shares of respondents who declared the respective policy measure to
lection and selection of explanatory variables, our findings are difficult be of average or high importance, as opposed to being of low or no
to compare. importance (our “important” definition of takeup, see Table 1).
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 3 of 13
Table 1 Summary statistics
Variables Obs Mean Std. Dev. Min. Max.
Dependent variables
Short-time work: used 1011 0.389 0.488 0 1
Short-time work: important 890 0.306 0.461 0 1
Income replacement: used 1011 0.411 0.492 0 1
Income replacement: important 892 0.333 0.472 0 1
Corona loans: used 1011 0.291 0.454 0 1
Corona loans: important 875 0.181 0.385 0 1
Lockdown variables
Closed de jure 1011 0.405 0.491 0 1
Closed de facto 1011 0.256 0.437 0 1
Expected share of April turnover lost 935 − 0.436 0.459 −1 1
Physical closeness important: workers 1011 0.428 0.495 0 1
Physical closeness important: clients 1011 0.641 0.480 0 1
Economic variables
Employment
1 FTE 1011 0.456 0.498 0 1
2–9 FTE 1011 0.396 0.489 0 1
> 10 FTE 1011 0.148 0.356 0 1
Debt ratio 2019
0 904 0.676 0.468 0 1
0.01–0.25 904 0.140 0.348 0 1
0.26–1 904 0.079 0.270 0 1
Profit ratio 2019
<0 987 0.023 0.149 0 1
0–0.25 987 0.721 0.449 0 1
> 0.25 987 0.232 0.423 0 1
Owner private wealth 2019
< 50k CHF 832 0.246 0.431 0 1
50k–500k 832 0.388 0.487 0 1
> 500k CHF 832 0.189 0.392 0 1
Labour cost share 2019
< 0.33 935 0.465 0.499 0 1
0.33–0.66 935 0.257 0.437 0 1
0.66–1 935 0.203 0.402 0 1
Behavioural variables
German-speaking Switzerland 1011 0.630 0.483 0 1
French-speaking Switzerland 1011 0.261 0.439 0 1
Italian-speaking Switzerland 1011 0.109 0.312 0 1
Age
< 45 989 0.138 0.345 0 1
45 ≤ age ≤ 65 989 0.694 0.461 0 1
> 65 989 0.169 0.375 0 1
Education
Compulsory school only 1006 0.017 0.129 0 1
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 4 of 13
Table 1 Summary statistics (Continued)
Variables Obs Mean Std. Dev. Min. Max.
Vocational qualifications 1006 0.322 0.468 0 1
Academic qualifications 1006 0.656 0.475 0 1
Female 1011 0.411 0.492 0 1
Swiss national 1011 0.812 0.391 0 1
Profit ratio in 2019 (less than 0, between 0 and variables with their associated coefficient vectors β esti-
25%, or more than 25% of total revenue), mated with OLS, γj is a sector fixed effect, and ϵij is a
2.3 Owner’s private net wealth in 2019 (less than random error.
CHF 50,000, between CHF 50,000 and CHF
500,000, more than CHF 500,000), and 3 Results and discussion
2.4 Share of labour costs in total costs in 2019 (less 3.1 Results
than 33%, between 33% and 66%, more than Figure 1 shows that businesses closed by government de-
66%). cree (de jure) on average expected to lose 70% of their
3. Behavioural variables: These variables are included revenue in April, whereas businesses that could remain
to capture characteristics of the sample businesses open without major restrictions reported revenue losses
and their owners that cannot, on standard of 10% on average. Closed businesses also availed much
economic grounds, be expected to affect the extent more of all three policy measures, “short-time work”,
to which those businesses avail of different support “income replacement” and “corona loans”. The contrast
measures. They are therefore not expected to is largest for the “income replacement” measure, which
correlate strongly with businesses’ use of support was used by 60% of businesses closed de jure and by
measures, conditional on controlling for the only 6% of businesses that could remain open without
lockdown and economic variables. We consider the major restrictions.
following behavioural variables: These unconditional statistics show that the take-up of
3.1 Linguistic region (German, French or Italian- government support measures was clearly driven by
speaking Switzerland), firms’ lockdown affectedness. But affectedness explains
3.2 Age of the respondent (below 45, between 45 only part of the take-up: not all closed businesses availed
and 65, over 65), of the measures, and many businesses that remained in
3.3 Educational attainment of the respondent operation did. We therefore turn to multiple regression
(compulsory school only, vocational analysis in order to explore alternative determinants of
qualifications, academic qualifications), firms’ recourse to public support.
3.4 Gender of the respondent (1 = female), and In Tables 2, 3 and 4, we report our estimates of differ-
3.5 Nationality of the respondent (1 = Swiss). ent variants of Eq. (1). We consider three dependent var-
iables, one for each of the policy measures we study:
In addition, we also consider the inclusion of fixed ef- “short-time work” (Table 2), “income replacement”
fects for 13 broad sectors5. (Table 3) and “corona loans” (Table 4). A graphical illus-
Using these variables, we estimate variants of the fol- tration of our main coefficient estimates is provided in
lowing multiple regression model: Fig. 2.
The results of the regressions are shown in Tables 2, 3
and 4. For each policy measure, the dependent variable
Y measure
ij ¼ β0 þ βmeasure
ldown Xij
ldown
þ βmeasure
econ Xecon
ij is constructed as follows: we code as 1 all businesses
þ βbehav Xij
measure behav
þ γ j þ ϵij
measure
; ð1Þ who answered with either “of intermediate importance”
or “highly important”, and as 0 all businesses who an-
where Y measure
ij is a binary variable capturing whether swered with “not used” (thus dropping businesses who
small business i in sector j has availed of the relevant answered “weakly important”)6.
policy measure, the three Xij vectors contain, respect- For each dependent variable, we show estimates of six
ively, lockdown, economic and behavioural independent specifications: only lockdown explanatory variables,
5
These sectors are: construction, education, art and culture, hotels and 6
This is the version “[Dependent variable]: important” as of Table 1.
tourism, restaurants and bars, health, agriculture and gardening, We ran the same regressions on the version “[Dependent variable]:
information technology, non-food retail, food retail, hairdressing and used”, where the response category “weakly important” was coded as 1
beauty, legal and financial services, advertising and graphic design, and instead of being dropped. The results are very similar (see the working
others (including small manufacturing). paper version of this paper, Brülhart et al., 2020c).
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 5 of 13
Fig. 1 Reported loss of turnover and recourse to support measures
lockdown + economic explanatory variables and lock- therefore, “objective” affectedness was a much stronger
down + economic + behavioural explanatory variables, determinant of support uptake than “subjective” behav-
each time without and with inclusion of sector fixed ef- ioural aspects.
fects (coefficients not shown). Result 2: Amongst the lockdown variables, physical
Rather than discussing each table and each of the 18 closeness of workers has a strikingly large effect on
estimated regression models individually, we present re- small businesses’ recourse to public support
sults by focusing on what we consider to be the most Amongst the lockdown measures, de jure closure has
important findings. the strongest effect in the majority of regressions, and de
Result 1: Overall, lockdown and economic variables facto closure also has a strong effect throughout. This is
have considerably more explanatory power than be- as expected. More surprising, perhaps, is the finding that
havioural variables physical closeness amongst workers has a large and sta-
If we focus on specifications with industry fixed effects, tistically significant effect throughout, whereas physical
adjusted R2 values of regressions that include only lock- closeness to clients appears as a weaker determinant of
down and economic variables range between 0.22 (corona firms’ recourse to policy measures7. The result that busi-
loans) and 0.31 (income replacement). These are satisfac- nesses were more likely to avail of short-time work,
tory values for a linear probability model, especially given ceteris paribus, if closeness amongst workers is import-
the unavoidable imprecisions of survey responses. More- ant also suggests that the short-time work and income
over, the variables that are expected to correlate strongly replacement programmes may have contributed to con-
with recourse to policy measures do correlate strongly. taining the virus by reducing contacts in the workplace.
This lends plausibility to our data. It is also interesting to
note that considering sector-specific dummies adds rela- 7
The two sectors in which the largest shares of sample businesses
tively little explanatory variation. This suggests that our
indicated physical closeness of workers to be important are hairdressing
basic set of control variables account well for the deter- and beauty (67%) and agriculture and gardening (52%). The two
ministic part of survey responses with respect to the take- sectors with the smallest shares of sample business declaring physical
up of policy measures. closeness of workers to be important are information technology (21%)
and graphic design and advertising (32%). The corresponding sectors
Adding the behavioural variables in some instances
for physical closeness to clients are hairdressing and beauty (100%) and
does not increase the adjusted R2. The largest observed healthcare (93%) at the upper end, and agriculture and gardening
increase in adjusted R2 is by 0.02 (corona loans). Overall, (33%) and information technology (34%) at the lower end.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 6 of 13
Table 2 Determinants of the use of the short-term work
Independent variables: (1) (2) (3) (4) (5) (6)
Closed de jure 0.27*** 0.25*** 0.32*** 0.29*** 0.31*** 0.28***
(0.05) (0.05) (0.04) (0.04) (0.04) (0.04)
Closed de facto 0.19*** 0.18*** 0.21*** 0.19*** 0.21*** 0.18***
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Expected share of April turnover lost 0.06 0.06 0.07* 0.07* 0.08** 0.07**
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Physical closeness important: workers 0.19*** 0.18*** 0.09*** 0.10*** 0.09*** 0.10***
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Physical closeness important: clients − 0.03 − 0.02 − 0.02 − 0.04 − 0.02 − 0.04
(0.04) (0.04) (0.03) (0.03) (0.03) (0.03)
Employment: 2–9 FTE 0.23*** 0.22*** 0.23*** 0.23***
(0.03) (0.03) (0.03) (0.03)
Employment: > 10 FTE 0.32*** 0.31*** 0.32*** 0.31***
(0.05) (0.05) (0.05) (0.05)
Debt ratio 2019: [0.01–0.25] 0.12*** 0.12*** 0.12*** 0.13***
(0.04) (0.04) (0.04) (0.04)
Debt ratio 2019: [0.26–1] 0.06 0.09 0.05 0.08
(0.05) (0.06) (0.05) (0.06)
Profit to sales ratio 2019: [0–0.25] − 0.05 − 0.04 − 0.04 − 0.03
(0.10) (0.10) (0.10) (0.10)
Profit to sales ratio 2019: > 0.25 − 0.13 − 0.12 − 0.12 − 0.11
(0.10) (0.11) (0.10) (0.10)
Owner private wealth 2019: [50k–500k CHF] 0.06 0.07* 0.07* 0.08**
(0.04) (0.04) (0.04) (0.04)
Owner private wealth 2019: > 500k CHF 0.01 0.00 0.03 0.03
(0.04) (0.04) (0.04) (0.04)
Labour cost share 2019: < 0.33 − 0.04 − 0.04 − 0.04 − 0.04
(0.04) (0.04) (0.04) (0.04)
Labour cost share 2019: [0.33–0.66] 0.14*** 0.13*** 0.13*** 0.13***
(0.04) (0.04) (0.04) (0.04)
German-speaking Switzerland 0.01 − 0.00
(0.03) (0.03)
Italian-speaking Switzerland 0.06 0.04
(0.05) (0.05)
45 ≤ age ≤ 65 − 0.05 − 0.06
(0.04) (0.04)
Age > 65 − 0.10* − 0.11**
(0.05) (0.05)
Education: vocational qualifications 0.08 0.04
(0.12) (0.12)
Education: academic qualifications 0.07 0.03
(0.12) (0.12)
Female − 0.03 − 0.02
(0.03) (0.03)
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 7 of 13
Table 2 Determinants of the use of the short-term work (Continued)
Independent variables: (1) (2) (3) (4) (5) (6)
Swiss national − 0.11*** − 0.10***
(0.04) (0.04)
Observations 822 822 822 822 822 822
Adjusted R2 0.12 0.16 0.28 0.29 0.28 0.30
Industry FE No Yes No Yes No Yes
All dependent and independent variables are binary (0/1), except for “Expected share of April turnover lost”. The dependent variable equals 1 for businesses that
describe short-term work as “highly important” or “of intermediate importance”, and it equals 0 if or businesses that describe short-term work as “not used”
(businesses describing short-term work as “weakly important” are omitted). Parameters are estimated by means of a linear probability model (OLS), with robust
standard errors (in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1)
Result 3: Larger and indebted businesses are more ratios, owner wealth), the second and third interpreta-
likely to claim public support tions appear particularly plausible.
Amongst the economic variables, firm size and indebt- Result 4: The take-up of corona loans is statistically
edness have the largest impact overall. Businesses with less well explained with lockdown and economic vari-
more than one full-time employee have a higher prob- ables than the take-up of measures that target labour
ability of recourse to public support, ceteris paribus, than incomes
one-person firms. This pattern is stronger for short-term The explanatory power of the lockdown and economic
work than for income replacement, which adds further factors as measured by adjusted R2 values (not consider-
plausibility to the data, as the main policy measure avail- ing fixed effects) is 0.19 for corona loans, whereas for
able to one-person firms is income replacement. Simi- short-term work it is 0.28 and for income replacement it
larly, the significant effect we observe for labour cost is 0.31. Hence, recourse to corona loans appears to be
shares on short-time work conforms with expectations, more random, and less driven by “rational” lockdown
as this is the policy measure designed to cover labour and economic factors than the other two policy
costs of crisis-hit firms. measures9.
The significantly lower take-up of corona loans by in- It is particularly surprising that firms with a low
dependent workers compared to businesses with more labour cost share seem if anything to have below-
than one worker might also be related to legal liability: average take-up of corona loans. Although this esti-
owners of sole proprietorships are liable for debt repay- mate could be biased towards zero because of impre-
ment with their private assets, whereas owners of incor- cise measurement, it is concerning that the measure
porated businesses are liable only with their business that comes closest to capturing the main target of the
assets. Hence, the latter may be more ready, other things policy—i.e. sustaining businesses for which non-wage
equal, to take on corona loans8. fixed costs account for a large share of outlays—has
The strong effects of indebtedness on the likelihood so little explanatory power.
of claiming income replacement and/or corona loans There are moreover indications that certain unex-
could reflect three patterns. First, it could capture the pected business characteristics determine the take-up of
fact that these measures benefit in particular finan- corona loans. The strongest effect of all variables, stron-
cially fragile businesses—i.e. those who need the mea- ger even than that of de jure closure, is observed on the
sures most. Second, it could be the manifestation of variable debt ratio. This is suggestive of a sort of “habit
indebted businesses having lower informational or at- effect”: firms already used to taking on debt are less re-
titudinal thresholds for accessing public support and luctant towards taking on additional debt. It might also
for incurring further debt. And third, this result could
arise from indebted firms replacing exiting debts with
cheaper corona loans. Given that our regressions con- 9
This finding is consistent with recent research on U.S. data,
trol for a range of variables related to financial fragil-
suggesting that the “Paycheck Protection Program” did not channel
ity in the corona crisis (the lockdown variables, profit support to the hardest hit firms and had little effect on employment at
small businesses (Granja et al., 2020; Chetty et al., 2020). The main
problems in the U.S. case seem to be uneven access to banking
services by small businesses and inefficient targeting of the policy at
8
We thank Cédric Tille for this insight. We cannot test the proposition the most affected sectors. Given the different institutional settings,
directly, as the legal form of our sample businesses is not recorded in these do not appear to be plausible concerns in the context of the
the data. Swiss small-business support policies.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 8 of 13
Table 3 Determinants of the use of income replacement
Independent variables: (1) (2) (3) (4) (5) (6)
Closed de jure 0.41*** 0.39*** 0.42*** 0.40*** 0.41*** 0.39***
(0.04) (0.05) (0.04) (0.05) (0.04) (0.05)
Closed de facto 0.11*** 0.11*** 0.13*** 0.12*** 0.13*** 0.12***
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Expected share of April turnover lost 0.15*** 0.14*** 0.15*** 0.14*** 0.16*** 0.15***
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Physical closeness important: workers 0.09*** 0.08** 0.06* 0.06* 0.05 0.05
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Physical closeness important: clients 0.06* 0.06* 0.06* 0.05 0.06* 0.04
(0.03) (0.04) (0.03) (0.03) (0.03) (0.03)
Employment: 2–9 FTE 0.09*** 0.10*** 0.10*** 0.11***
(0.03) (0.03) (0.03) (0.03)
Employment: > 10 FTE 0.10** 0.09* 0.10** 0.09*
(0.05) (0.05) (0.05) (0.05)
Debt ratio 2019: [0.01–0.25] 0.08* 0.09** 0.08* 0.10**
(0.04) (0.05) (0.04) (0.04)
Debt ratio 2019: [0.26–1] 0.10** 0.13** 0.10** 0.14**
(0.05) (0.06) (0.05) (0.05)
Profit to sales ratio 2019: [0–0.25] 0.06 0.06 0.07 0.07
(0.08) (0.08) (0.09) (0.08)
Profit to sales ratio 2019: > 0.25 0.01 0.01 0.03 0.03
(0.09) (0.09) (0.09) (0.09)
Owner private wealth 2019: [50k–500k CHF] 0.00 0.00 0.03 0.03
(0.04) (0.04) (0.04) (0.04)
Owner private wealth 2019: > 500k CHF − 0.12*** − 0.11*** − 0.07 − 0.06
(0.04) (0.04) (0.04) (0.04)
Labour cost share 2019: < 0.33 0.03 0.03 0.01 0.01
(0.04) (0.04) (0.04) (0.04)
Labour cost share 2019: [0.33–0.66] 0.08* 0.08* 0.06 0.06
(0.04) (0.04) (0.04) (0.04)
German-speaking Switzerland 0.05 0.05
(0.04) (0.04)
Italian-speaking Switzerland 0.08 0.08
(0.06) (0.06)
45 ≤ age ≤ 65 − 0.06 − 0.06
(0.04) (0.04)
Age > 65 − 0.13** − 0.14***
(0.05) (0.05)
Education: vocational qualifications 0.10 0.09
(0.13) (0.13)
Education: academic qualifications 0.04 0.03
(0.13) (0.13)
Female 0.03 0.03
(0.03) (0.03)
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 9 of 13
Table 3 Determinants of the use of income replacement (Continued)
Independent variables: (1) (2) (3) (4) (5) (6)
Swiss national − 0.07* − 0.07*
(0.04) (0.04)
Observations 823 823 823 823 823 823
Adjusted R2 0.28 0.28 0.31 0.31 0.32 0.32
Industry FE No Yes No Yes No Yes
All dependent and independent variables are binary (0/1), except for “Expected share of April turnover lost”. The dependent variable equals 1 for businesses that
describe income replacement as “highly important” or “of intermediate importance”, and it equals 0 if or businesses that describe income replacement as “not
used” (businesses describing income replacement as “weakly important” are omitted). Parameters are estimated by means of a linear probability model (OLS),
with robust standard errors (in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1)
reflect some firms replacing existing debt with cheaper repayment period. Moreover, the discounted cost to the
corona loans. public purse of such a policy could well be close to zero10.
We observe that firms in Italian-speaking Switzerland At the intensive margin, the main economic risk is
resort to corona loans more than those in the rest of the that corona debt will lead to suboptimally low private
country (controlling for lockdown severity, which was investment. One measure towards reducing that risk
somewhat greater in the Italian part). There are also in- is to lift the condition that corona loans may only be
dications that businesses headed by men and businesses used to finance current spending but not capital
headed by non-Swiss nationals are somewhat more spending or R&D spending (Gersbach, Mikosch, &
likely, ceteris paribus, to resort to corona loans. Sturm, 2020). Moreover, for particularly hard-hit and
difficult-to-transform businesses (e.g. hotels in certain
tourist resorts), debt forgiveness based on sector/re-
3.2 Discussion gion-specific parameters could be envisaged (Bonardi
Our results suggest that the take-up of corona loans across et al., 2020), as well as public co-payments towards
small businesses was more diffuse than that of labour sup- agreements between property owners and commercial
port measures. Some measurable behavioural factors (espe- renters (Brülhart et al., 2020a).
cially prior history of indebtedness) seem to have played a
role, and there appears to be considerable randomness in
the process. This has implications both for “life-or-death” 4 Summary and outlook
extensive-margin decisions and for “growth-vs-stagnation” The COVID-19 crisis has hit small businesses hard. Two
intensive-margin decisions by private businesses. At the ex- thirds of our 1011 sample firms had to fully or partially
tensive margin, some firms that would be viable post-crisis cease operations during the height of the lockdown in
and should thus avail of the subsidised loans may refrain April—40% because they had to close by government de-
from doing so, whilst some other firms take on cheap debt cree and 26% because remaining open was no longer
without needing to. At the intensive margin, firms saddled commercially viable.
with corona debt may be prevented during or after the cri- Moreover, the crisis afflicted small businesses ran-
sis from expanding in the way they would have in the ab- domly, in the sense that lockdown affectedness is not
sence of the pandemic. causally related to any measure of prior economic per-
The extensive-margin risk mainly consists of businesses formance. This can be seen in the data. If we regress our
folding because they are unwilling to take on extra debt main lockdown variables (closed de jure, closed de facto
and/or to burden future revenue streams with legacy costs and expected share of April turnover lost) on the eco-
from the COVID crisis. This risk could be mitigated by nomic variables, the adjusted R2 values range between
making loan repayment terms contingent on future 0.00 and 0.02. The impact of the corona shock was
profits, akin to student loans that only have to be repaid clearly orthogonal to the prior financial viability of indi-
once graduates earn a certain income (Bonardi, Brülhart, vidual businesses.
Danthine, Jondeau, & Rohner, 2020). This idea has been
further developed by Danthine, Fahlenbrach, and Morellec 10
There may also be a case for reviewing the harsh Swiss insolvency
(2020). They propose that corona loans should be option- laws. According to the World Bank’s Resolving Insolvency index,
ally convertible into preferred but non-voting equity Switzerland ranks 49th in the world, well behind Germany (4th), Italy
stakes (“Vorzugsaktien”). This would turn fixed debt- (21st), Austria (22nd) and France (26th). As shown e.g. by Favara et al.
(2016), bankruptcy codes that favour debt enforcement have a
repayment into flexible dividend payments, and it would constraining effect on investment and growth of financially distressed
free debtor firms from the constraints of a finite firms.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 10 of 13
Table 4 Determinants of the use of corona loans
Independent variables: (1) (2) (3) (4) (5) (6)
Closed de jure 0.15*** 0.15*** 0.18*** 0.17*** 0.17*** 0.15***
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Closed de facto 0.07** 0.08** 0.09*** 0.09** 0.09*** 0.08**
(0.04) (0.04) (0.03) (0.03) (0.03) (0.03)
Expected share of April turnover lost 0.06 0.07** 0.05 0.05 0.06 0.05
(0.04) (0.03) (0.03) (0.03) (0.04) (0.03)
Physical closeness important: workers 0.13*** 0.11*** 0.07** 0.07** 0.06* 0.06**
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Physical closeness important: clients − 0.02 0.01 − 0.02 − 0.00 − 0.02 − 0.01
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Employment: 2–9 FTE 0.09*** 0.06** 0.08*** 0.07**
(0.03) (0.03) (0.03) (0.03)
Employment: > 10 FTE 0.13*** 0.12*** 0.13*** 0.12***
(0.04) (0.04) (0.04) (0.04)
Debt ratio 2019: [0.01–0.25] 0.22*** 0.22*** 0.22*** 0.22***
(0.05) (0.05) (0.05) (0.05)
Debt ratio 2019: [0.26–1] 0.25*** 0.26*** 0.24*** 0.27***
(0.06) (0.06) (0.06) (0.06)
Profit to sales ratio 2019: [0–0.25] 0.05 0.05 0.05 0.04
(0.07) (0.07) (0.07) (0.07)
Profit to sales ratio 2019: > 0.25 0.01 0.01 0.00 0.00
(0.07) (0.07) (0.08) (0.08)
Owner private wealth 2019: [50k–500k CHF] − 0.07** − 0.06* − 0.05 − 0.04
(0.03) (0.03) (0.03) (0.03)
Owner private wealth 2019: > 500k CHF − 0.08** − 0.07* − 0.05 − 0.04
(0.04) (0.04) (0.04) (0.04)
Labour cost share 2019: < 0.33 − 0.04 − 0.04 − 0.04 − 0.04
(0.03) (0.03) (0.03) (0.03)
Labour cost share 2019: [0.33–0.66] 0.07* 0.07* 0.06 0.06
(0.04) (0.04) (0.04) (0.04)
German-speaking Switzerland 0.02 0.03
(0.03) (0.03)
Italian-speaking Switzerland 0.19*** 0.19***
(0.06) (0.06)
45 ≤ age ≤ 65 − 0.00 0.00
(0.04) (0.04)
Age > 65 − 0.06 − 0.06
(0.05) (0.05)
Education: vocational qualifications 0.03 − 0.01
(0.12) (0.11)
Education: academic qualifications 0.00 − 0.01
(0.12) (0.11)
Female − 0.05* − 0.01
(0.03) (0.03)
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 11 of 13
Table 4 Determinants of the use of corona loans (Continued)
Independent variables: (1) (2) (3) (4) (5) (6)
Swiss national − 0.07** − 0.06*
(0.04) (0.04)
Observations 806 806 806 806 806 806
Adjusted R2 0.07 0.12 0.19 0.22 0.21 0.24
Industry FE No Yes No Yes No Yes
All dependent and independent variables are binary (0/1), except for “Expected share of April turnover lost”. The dependent variable equals 1 for businesses that
describe corona loans as “highly important” or “of intermediate importance”, and it equals 0 if or businesses that describe corona loans as “not used” (businesses
describing corona loans as “weakly important” are omitted). Parameters are estimated by means of a linear probability model (OLS), with robust standard errors
(in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1)
When a historically large, entirely exogenous negative whilst some other firms will take on those subsidised
shock randomly affects large swathes of the private sec- loans without material need.
tor, public intervention is necessary in order to prevent Moreover, some firms might have availed of corona
both individual hardship and a macroeconomic slump. loans to refinance their debts on more advantageous
The fact that entire sectors were hit through no fault of terms. To the extent that this could help some of those
their own by an exceptionally rare event also implied firms to weather the COVID storm, that effect is desir-
that moral hazard problems usually afflicting public sup- able. But there are likely to be firms for whom this rep-
port to businesses are much less of an issue in the resents a windfall that does not affect their survival
COVID-19 crisis (Bonardi et al., 2020). prospects.
After it had announced the lockdown, the Swiss gov- Looking ahead, the case for public generosity towards
ernment moved swiftly to compensate affected workers affected firms is growing weaker the longer the crisis con-
and, to a lesser extent, businesses. Of our sample firms, tinues. For a time-limited, government-imposed lockdown
42% availed of at least one of the three measures (short- spanning a few weeks or months, the consensus view is
time work and income replacement to cover labour in- that economic policy should aim at “freezing” productive
come, and corona loans to fill remaining liquidity needs), capacity by plugging most of the private-sector revenue
and 8% availed of all three11. shortfalls through the public purse (e.g. Alós-Ferrer et al.,
Our analysis suggests that the labour-income sup- 2020; Bonardi et al., 2020). Our estimates suggest that the
port measures reached their intended targets better policy measures taken by the Swiss government were ef-
than the corona loans: variables capturing the direct fective in that respect.
effects of the lockdown and the economic circum- Freezing cannot continue indefinitely, however, es-
stances of individual businesses have greater explana- pecially not once the economy is released from lock-
tory power in the determination of who resorted to down and consumer spending is picking up again
short-time work and income replacement than of who (Brown et al. 2020). The longer some businesses’
took out a corona loan. Take-up of corona loans was pandemic-related problems persist, the higher is the
driven to a larger extent by firms’ prior history of in- likelihood that closing or restructuring them would
debtedness as well as by unrelated demographic and be less costly in societal terms than keeping them
geographic variables. afloat with public subsidies. Moral hazard issues re-
Our evidence is consistent with cultural and behav- turn to the fore when the question becomes whether
ioural factors playing a larger role in the take-up of labour and capital should be allocated to more pro-
loans than for the other measures. The data do not ductive uses rather than remaining partly or fully idle
allow us to identify specific mechanisms, but it seems in a “frozen” business.
plausible to assume that deep-held attitudes and inhi- This in turn implies that the compensation rate of pol-
bitions to taking on debt vary across business owners. icy measures should gradually be ratcheted down. The
Differences in businesses’ acceptance of higher debt replacement rate of labour income measures could be
levels could mean that some viable businesses will lowered from the 80% currently applied (up to a ceiling),
shrink or exit because of an intrinsic aversion to debt, e.g. by 5% every quarter after the expiry of the current
guaranteed period. Similarly, the part of any new corona
loans that enjoys a state guarantee could be gradually
The percentages are based on the “important” variant of our takeup
11 lowered, and/or the interest rate charged could be
measures. increased.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 12 of 13
Fig. 2 Determinants of the use of support measures (illustration of estimated coefficients).The graphs show coefficient estimates and 95%
confidence intervals as reported in column (5) of Tables 2, 3 and 4
Another approach could be to modulate the maximum The main challenge for economic policy will be to find
duration with which wage support is granted. Kopp and the right balance between supporting labour incomes
Siegenthaler (2020) show that, during the financial crisis, and long-term viable businesses without preventing
most Swiss firms tended to leave the short-time work meaningful adaptation of the economy to the changed
scheme before they reached the maximum benefit dur- circumstances. It will be a delicate balancing act.
ation. However, the few firms that drew short-time work
Acknowledgements
subsidies until the legal maximum laid off a significant Not applicable.
share of previously subsidised workers. In these cases,
the subsidies only postponed but did not prevent dis- Authors’ contributions
MB and MS conceived the analysis, TL and MS analysed the data, RL
missals. These findings imply that phasing out govern- provided feedback. All authors read and approved the final manuscript.
ment support can reduce the risk of funding ‘zombie
jobs’. Short-time work support could be withdrawn dif- Funding
This study was funded by the Faculty of Business and Economics at
ferentially by economic sector, depending on affected- University of Lausanne.
ness by the pandemic.
Such a policy approach could help preserve viable Availability of data and materials
Our data are available from the authors.
businesses without preventing economically desirable
firm closures or reorganisations12. Ethics approval and consent to participate
Not applicable
Consent for publication
12
A staggered policy drawdown would be comparable to “benefit
Not applicable
frontloading” in unemployment insurance systems, whereby the
replacement ratio is lowered during the unemployment spell Competing interests
(Hopenhayn and Nicolini, 1997; Lindner and Reizer, 2020). In the Not applicable
unemployment insurance context, benefit frontloading can have
adverse moral hazard effects (Kolsrud et al., 2018). This is not a Author details
concern in the case of Covid support measures, where initial 1
Department of Economics, HEC Lausanne, University of Lausanne, Lausanne,
affectedness was exogenously determined, which further strengthens Switzerland. 2Labour Market Section, KOF Swiss Economic Institute, ETH
the case for a gradual lowering over time. Zurich, Zurich, Switzerland.
Brülhart et al. Swiss Journal of Economics and Statistics (2020) 156:15 Page 13 of 13
Received: 2 July 2020 Accepted: 18 August 2020
References
Alós-Ferrer, Carlos, et al. (2020) Coronavirus: testing and freezing – a survival
strategy for the Swiss economy. Mimeo: Department of Economics, University
of Zurich.
Bonardi, Jean-Philippe, Marius Brülhart, Jean-Pierre Danthine, Eric Jondeau and
Dominic Rohner (2020) The economics of wage compensation and corona
loans: why and how the state should bear most of the economic cost of the
COVID lockdown. VoxEU.org.
Brown, Martin, Matthias Fengler, Rafael Lalive, Robert Rohrkemper and Thomas
Spycher (2020) Monitoring consumption in Switzerland. Online Dashboard
monitoringconsumption.org.
Brülhart, Marius, Monika Bütler, Luca Crivelli, David Dorn, Jan-Egbert Sturm and
Beatrice Weder (2020a) Public matching payments for commercial rent
abatements during the COVID crisis. Policy Brief, Swiss National Covid-19
Science Task Force.
Brülhart, Marius, Jeremias Klaeui, Rafael Lalive, Tobias Lehmann and Michael
Siegenthaler (2020b). Covid Survey: Die Schweizer Selbständigerwerbenden
in der Covid19-Pandemie. Mimeo: University of Lausanne and ETH Zürich.
Brülhart, Marius, Rafael Lalive, Tobias Lehmann and Michael Siegenthaler (2020c)
COVID-19 financial support to small businesses in Switzerland: evaluation and
outlook. KOF Working Paper 483#, ETH Zurich.
Brzezinski, Adam, Valentin Kecht and David Van Dijcke (2020) The cost of staying
open – voluntary social distancing and lockdowns in the US. Economics
Working Paper #910, University of Oxford.
Chetty, Raj, John N. Friedman, Nathaniel Hendren and Michael Stepner (2020)
How did COVID-19 and stabilization policies affect spending and
employment? Mimeo: Harvard University.
Danthine, Jean-Pierre, Rüdiger Fahlenbrach and Erwan Morellec (2020) Comment
soutenir financièrement les PMEs suisses? Une solution durable visant à
augmenter la resilience de l’économie. Mimeo, E4S (UNIL, EPFL, IMD),
Lausanne.
Gersbach, Hans, Heiner Mikosch and Jan-Egbert Sturm (2020) Mit einer
Anpassung des COVID-19-Kreditprogramms gegen die Investitionsschwäche.
Oekonomenstimme.org.
Granja, Joao, Christos Makridis, Constantine Yannelis and Eric Zwick (2020) Did
the paycheck protection program hit the target? NBER Working Paper #27095.
Hopenhayn, Hugo A. and Juan Pablo Nicolini (1997) Optimal unemployment
insurance. Journal of Political Economy, 105(2): 412-438.
KOF (2020) Konjukturprognose, Mai 2020. Report, KOF Swiss Economic Institute,
ETH Zurich.
Kolsrud, Jonas, Camille Landais, Peter Nilsson and Johannes Spinnewijn (2018)
The optimal timing of unemployment benefits: theory and evidence from
Sweden. American Economic Review, 108(4-5), 985–1033.
Kopp, Daniel and Michael Siegenthaler (2020) Short-time work and
unemployment in and after the great recession. Mimeo: KOF Swiss Economic
Institute, ETH Zurich.
Lindner, Attila and Balázs Reizer (2020) Frontloading the unemployment benefit:
an empirical assessment. American Economic Journal: Applied Economics,
12(3): 140-174.
SECO (2020) Konjunkturprognose der Expertengruppe des Bundes – Juni 2020.
Report, Staatssekretariat für Wirtschaft, Bern.
Zoller-Rydzek, Benedikt and Florian Keller (2020) COVID-19: Guaranteed Loans
and Zombie Firms. Mimeo: ZHAW School of Management and Law,
Winterthur.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
File and source
- File
- brulhart-2020-covid-19-financial-support-to-small-businesses-in-switzerland-evaluation-and.pdf
- Size
- 891,662 bytes
- SHA-256
- 287ccc5025c3820f934716936f5dac940e4f9ae15c016164621711e2bb350195
- Our copy
- brulhart-2020-covid-19-financial-support-to-small-businesses-in-switzerland-evaluation-and.pdf
- Original
- api.unil.ch