Firms' participation in the Swiss COVID-19 loan programme

Fuhrer, Lucas Marc; Ramelet, Marc-Antoine; Tenhofen, Jörn Article Firms' participation in the Swiss COVID-19 loan programme Swiss Journal of Economics and Statistics Provided in Cooperation with: Swiss Society of Economics and Statistics, Zurich Suggested Citation: Fuhrer, Lucas Marc; Ramelet, Marc-Antoine; Tenhofen, Jörn (2021) : Firms' participation in the Swiss COVID-19 loan programme, Swiss Journal of Economics and Statistics, ISSN 2235-6282, Springer, Heidelberg, Vol. 157, Iss. 1, pp. 1-22, https://doi.org/10.1186/s41937-021-00070-4 This Version is available at: https://hdl.handle.net/10419/259766 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Swiss Journa l o f Economics and Statistics Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 https://doi.org/10.1186/s41937-021-00070-4 ORIGINAL ARTICLE Open Access Firms’ participation in the Swiss COVID-19 loan programme Lucas Marc Fuhrer, Marc-Antoine Ramelet and Jörn Tenhofen* Abstract This paper analyses the determinants of firm participation in the Swiss COVID-19 loan programme, which aims to bridge firms’ liquidity shortfalls that have resulted from the pandemic. State-guaranteed COVID-19 loans are widely used by Swiss firms, with 20% of all firms participating, resulting in a sizeable programme of 2.4% of GDP. We use a comprehensive dataset to study the determinants of firm participation. Our results can be summarised as follows. First, participation was largely driven by the exposure of a firm to lockdown restrictions and to the intensity of the virus in the specific region. Second, we show that firms associated with lower liquidity ratios had a significantly higher probability of participating in the programme. Third, we find no clear evidence that firm indebtedness affected participation in the programme and no evidence that pre-existing potential “zombie firms” participated more strongly in the loan programme. Fourth, we show that the programme reached younger and smaller firms, which could be financially more vulnerable as they are less likely to obtain outside finance during a crisis. Overall, we conclude that given its objective, the programme appears to be successful. Keywords: COVID-19, Loan programme, Guarantees, Firm behaviour JEL classification: D22; H81 1 Introduction Aside from its impact on public health, the COVID-19 pandemic caused a major economic shock. Governments reacted with a series of large-scale economic measures, ranging from short-time work schemes to credit support facilities. In Switzerland, the COVID-19 emergency loan programme was one of the key measures used to address the economic fallout of the pandemic. The Federal Council announced the programme on 25 March 2020 and stated the following objective: “Last Friday, 20 March 2020, the Federal Council presented a comprehensive package of measures to cushion the economic impact of the coronavirus pandemic. Bridging credit facilities should provide companies with sufficient liquidity to cover their current overheads despite turnover reductions associated with the new coronavirus.” *Correspondence: [email protected] Swiss National Bank, Börsenstrasse 15, 8022 Zürich, Switzerland This paper studies the key determinants of firm participation in the COVID-19 loan programme. The aim of our analysis is to assess whether the loan programme can be considered successful given the objective stated by the government. Additionally, we evaluate whether the programme comes with potential negative side effects. Understanding why firms chose to participate in the programme is important for at least two reasons. First, the success of the programme can be evaluated. Second, lessons can be learned for potential future loan programmes. Participation in the COVID-19 loan programme was sizeable, as 20% of all firms participated in this programme comprising a guaranteed loan volume of 2.4% of annual GDP. Participation is even more sizeable when considering the fact that approximately 60% of all small and medium-sized enterprises (SMEs) in Switzerland were debt-free prior to the crisis. The COVID-19 © The Author(s). 2021 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/. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 2 of 22 loan programme enabled firms to receive a governmentguaranteed bank loan quickly (usually within one business day) and with a minimum of bureaucracy. Firms could easily obtain the loan, and the requirements were minimal. As loans are guaranteed by the government and banks can refinance the loans at the central bank, loan supply was perfectly elastic. Because of this, whether a firm took a COVID-19 loan purely reflects loan demand. Loan conditions were also favourable and included a 100% guarantee and an attractive interest rate of 0% in the first year for small loans, as well as a rather long loan maturity of at least 5years. 1 We analyse firm participation in the COVID-19 programme by estimating a binary response model.2Our analysis focuses on the following questions: first, we evaluate whether a firm’s exposure to lockdown restrictions and to the virus intensity in the specific region can explain its participation in the loan programme. Second, we assess whether firms associated with lower liquidity ratios had a higher participation rate. Likewise, we analyse whether participation is higher for more indebted firms and whether it is particularly more so for firms in a pre-existing potentially precarious financial situation, i.e. firms with a relatively low profitability and high indebtedness before the pandemic hit (“zombie firms”). Finally, we document whether the loan programme reached potentially more vulnerable firms, such as younger and smaller companies. To address these questions, we build a comprehensive dataset combining various data sources. In particular, we match the complete set of firms in Switzerland from the register of commerce (BUR database) to the list of firms participating in the loan programme (JANUS database). Our findings can be summarised as follows. First, we show that participation in the loan programme is positively related to the exposure of a firm’s activity to lockdown restrictions as well as to the regional virus intensity, which we use as a proxy for households becoming more cautious. Second, we show that firms with an ex ante weaker liquidity position had a higher probability to participate in the programme. Importantly, these effects are economically meaningful; we can explain a wide range of firm participation rates. Hence, we find supporting evidence for the loan programme’s success in reaching its objective. Third, we find no clear evidence that firm indebtedness affected participation and no evidence that participation was higher for firms with an ex ante relatively low profitability and high indebtedness, i.e. what 1These are the conditions as originally announced and effective until the end of the application period. 2In that sense, we focus on the extensive margin and not the intensive margin of the programme. We focus on the former, as we can better separate demand and supply effects for the participation decision, as explained in Section 3. we identify as zombie firms.3Fourth, we show that the programme reached younger and smaller firms. Hence, the loan programme reached firms for which access to outside finance is typically more challenging—particularly during a crisis. Overall, our results are robust to different specifications and rely on several measures that exploit variation across sectors, regions and firm sizes.4 Our contribution to the literature is twofold. First, our paper contributes to the growing literature that studies the COVID-19 loan programme in Switzerland. For instance, firm participation in the programme is analysed by Brülhart, Lalive, Lehmann, and Siegenthaler (2020)and Zoller-Rydzek and Keller (2020). Our paper complements these studies, which are based on surveys, by instead using a comprehensive dataset combining various data sources. Moreover, we use what we believe to be exogenous measures of lockdown restrictions at a relatively granular level. Additionally, we explicitly account for firms’ liquidity position, which seems to be an economically important driver for participation in the loan programme. Second, we contribute to the more general literature that studies government-guaranteed loan programmes and their implications for the real economy. The existing literature points overall to the usefulness of such programmes in reducing informational costs and in dampening the effects of adverse aggregate shocks. Section 2describes the related literature, while Section 3 describes the COVID-19 loan programme, makes an international comparison and provides an overview of firms in Switzerland. Section 4presents the data that are used in the empirical analysis in Section 5. Finally, Section 6provides the conclusion. 2 Literature review Research on the Swiss COVID-19 loan programme is at thisstageonlynascent.Weareawareofthreecontributions. Similar to our paper, two studies (Brülhart et al., 2020; Zoller-Rydzek and Keller, 2020) investigate the determinants of participation in government support programmes during the pandemic. In addition to the loan programme, both papers also consider other support programmes, such as short-time work. In contrast to our analysis, which is based on a comprehensive dataset of all eligible firms, these two contributions are based on 3In the literature, there are various definitions of zombie firms. We had to resort to a definition, which can actually be operationalised based on the data we have available. In particular, we only have a cross-section of group-wise (headcounts within sectors) indicators of indebtedness and profitability. The lack of a time-series dimension, for instance, excludes the possibility to apply the OECD-definition of a zombie firm, which is as follows: “Zombie firms are defined as firms aged ≥10 years and with an interest coverage ratio <1over three consecutive years.” (see, e.g. OECD Economic Policy Paper No. 21, December 2017: “Confronting the zombies: policies for productivity revival”). 4In general, participation could result from a precautionary motive, where the financing is not actually needed to make current payments but just held for precautionary reasons. However, the trigger for that precautionary behaviour should still be the determining factors explained in the main text. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 3 of 22 surveys. More specifically, Brülhart et al. (2020)usea survey of 1011 self-employed workers and SMEs conducted in mid-April 2020.5The participants in that survey were asked about the importance and their participation in three government programmes, namely (1) short-time work, (2) income replacement for self-employed workers and small business owners and (3) COVID-19 loans. Programme participation is then related to different variables measuring the extent of the lockdown as well as firmspecific economic (e.g. employment), financial (e.g. debt and profit ratio) and other (e.g. linguistic region, age and education of respondent) variables. Brülhart et al. (2020) find that lockdown restrictions are positively related with the usage of both short-time work and COVID-19 loans. However, they find that lockdown restrictions are less important for explaining the participation in the loan programme than for explaining the participation in other government support programmes. Moreover, they find that previously indebted firms are more likely to take up COVID-19 loans. Another study based on survey evidence is the one by Zoller-Rydzek and Keller (2020), who build a theoretical model and test the resulting empirical implications by using data from the ZHAW managers barometer survey.6 In line with their theoretical model, they find that the prepandemic business situation seems to be an important driver of programme participation. In particular, firms in a worse ex ante situation are less likely to take out a COVID- 19 loan. Zoller-Rydzek and Keller (2020) conclude that there seems to be no evidence that the programme creates zombie firms. In their model, a zombie firm is a firm that survives the crisis thanks to the programme but cannot repay the debt. The third contribution by Kaufmann (2020)doesnot study the determinants of programme participation but investigates its effect on the macroeconomy. In particular, he analyses the impact of the COVID-19 loan programme on unemployment. He finds that higher loan supply due to the programme indeed reduces unemployment, with approximately CHF 400,000 of loan volume needed to save one job. Apart from the aforementioned more specific literature on the Swiss COVID-19 loan programme, our paper relates to different strands of the literature relevant for government credit guarantee programmes.7The unifying questions in this regard are why such a programme might be needed, which firms should be targeted and whether these programmes have been effective. 5The respondents were taken from an online pool of LINK , which provides a nationwide, representative sample of 115,000 individuals. 6In the first half of April 2020, 205 managers of Swiss companies were asked about their business situation and their response to the pandemic. 7Public credit guarantee programmes have existed at least since the beginning of the twentieth century. According to Green (2003), more than 2000 such schemes existed in almost 100 countries. First, why might a government-guarantee loan programme be needed? There is a broad literature on financial frictions, where informational asymmetries or moral hazard and thus agency problems potentially lead to a more difficult access to credit.8For instance, in the financial accelerator literature in the spirit of Bernanke and Gertler (1989) and Bernanke et al. (1999), agency costs lead to a premium on external finance and deadweight losses, while models along the line of Stiglitz and Weiss (1981) feature equilibrium rationing. Crisis situations such as the COVID-19 pandemic could lead to a sudden increase in uncertainty and informational problems, in turn increasing the difficulty to access credit or even leading to rationing. In such a situation, there might be a welfare-improving role for state guarantees as an insurance mechanism.9By overcoming informational problems, the state as an entity with “deep pockets” basically acts as insurance for the entire economy. Second, for whom might a government-guarantee loan programme be set up? The findings in the literature indicate that SMEs are particularly affected by informational issues and hence face problems in obtaining external finance. Gertler and Gilchrist (1993,1994)studythe impact of a cash squeeze on firms of different sizes and find that small firms, in contrast to larger ones, cannot use borrowing as easily to smooth cash-flow shocks. Small firms typically have less outside options of external finance. Chodorow-Reich (2014) use the more recent episode of the 2008/09 financial crisis to show that SMEs have more difficulty accessing credit during credit crunches, which in turn has negative implications for the real economy. Thus, it is not surprising that credit guarantee programmes are among the most common forms of government support for SMEs, as indicated, for instance, in Beck, Klapper, and Mendoza (2010).10 Third, have credit guarantee programmes worked? Overall, previous governmental loan guarantee programmes are typically found to be successful. This evaluation is carried out along several dimensions. For example, Cowling (2010) finds that small firms in the UK are indeed affected by credit rationing and that this situation can be addressed by a guarantee programme. In another study on the UK, Gonzalez-Uribe and Wang (2020)investigate the effect of a loan guarantee introduced during the 2008/09 financial crisis on firm outcomes. They find that the economic benefits significantly outweigh the costs of 8Seminal contributions in this literature are, for example, Bernanke and Gertler (1989), Bernanke, Gertler, and Gilchrist (1999), Carlstrom and Fuerst (1997), Holmström and Tirole (1997) and Kiyotaki and Moore (1997). 9See, for example, Honohan (2010). Offering a public loan guarantee programme can be undertaken for other reasons as well, for instance, to kick-start a lending process by allowing participants to gain experience, which might be important for the economic development of a particular sector, region or country. 10Beck et al. (2010) survey 76 credit guarantee schemes in 46 developed and developing countries. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 4 of 22 the programme. Riding and Haines (2001) and Riding, Madill, and Haines (2007), using the case of Canada, study the question of “additionality” of a credit guarantee programme. They ask whether such a programme leads to the extension of additional loans, which otherwise would not have been granted, or whether there is just a substitution of private loans by publicly guaranteed ones. Using credit scoring, they show that firms that otherwise would not have obtained a loan (based on the credit score) are able to secure a loan via the programme. Finally, Saito and Tsuruta (2014) analyse the costs in terms of adverse selection and moral hazard of these programmes. Their findings indicate the presence of both costs. Based on the rich public credit guarantee landscape in Japan, they show that banks with more risky customers offer more guaranteed loans. Moreover, they find that firms with guaranteed loans are more likely to default. This finding is more prevalent for guarantee programmes covering 100% than for programmes covering 80%. 3 COVID-19 loan programme On 26 March 2020, the Swiss federal government launched the COVID-19 loan guarantee programme to provide firms quick access to loans that could be used to bridge potential liquidity shortfalls resulting from the pandemic.11 The programme was open to the vast majority of firms residing in Switzerland; only firms with an annual turnover of more than CHF 500 million and firms founded after February 2020 could not participate.12 Under the programme, companies could receive from their bank government-guaranteed loans for an amount up to 10% of their annual turnover (up to a maximum of CHF 20 million) and with a maturity of five years.13 Afirst loan tranche of up to CHF 500,000 is fully guaranteed by the government. Larger companies could apply for a second tranche (called COVID-19 plus loan), of which the federal government would guarantee 85%. The pricing of the loan programme was attractive, as the first (second) tranche has an interest rate of 0% (0.5% for the guaranteed part) in the first year.14 Access to the loans was quick and easy since lending took place via existing client-bank relationships; the money was typically disbursed within a day. Firms did not need to have a pre-existing credit history or credit relationship—a bank account was sufficient. The 11Apart from the federal COVID-19 loan programme, there have been loan support programmes set up by individual cantons and joint programmes, such as the one offered for start-ups. Quantitatively, the federal programme is by far the largest programme. 12There are approximately 300 firms (out of a total of more than 600,000) with an annual turnover of more than CHF 500 million. 13In case of hardship, the original emergency decree establishing the programme allowed for an extension of another 2 years. 14For COVID-19 plus loans, each bank can fix its own interest rate for the remaining 15% of the loan. For both tranches, the emergency decree stipulated that the interest rate for the subsequent years would be determined by the government and would reflect market conditions. period for submission of applications for the programme lasted from 26 March to 31 July 2020. Nonetheless, there are a couple of requirements that may reduce the attractiveness of COVID-19 loans for some firms. For instance, the loan cannot be used to finance investments (other than replacement investments). Participating firms are not allowed to reimburse capital contributions or pay dividends. Moreover, COVID-19 loans cannot be used to refinance private or shareholder loans or repay intra-group loans. Likewise, there are restrictions on internal (potentially international) transactions.15 Firm participation was hence not obvious ex ante, particularly for larger and more complex firms. Normally, credit creation reflects both loan supply and demand. However, we exploit the fact that due to the structure of the programme as well as the coordinated and complementary policy measures taken, participation exclusively reflects firms’ demand for emergency loans. Loan supply—in terms of programme participation—was almost perfectly elastic.16 Indeed, banks had basically no incentive to reject loan applications: (i) credit risk was small or even non-existent due to the government guarantee;17 (ii) liquidity risk was also absent due to the SNB’s COVID-19 refinancing facility (CRF), by which banks can refinance the guaranteed part of the loan at the SNB policy rate by posting the guaranteed part as collateral;18 (iii) regulatory constraints on banks’ balance sheets were also relaxed via the Swiss financial market supervisory authority’s (FINMA) temporary adjustment of the leverage ratio calculation and at the request of the SNB, the deactivation of the countercyclical capital buffer by the federal government.19 The Swiss programme has not been the only loan guarantee programme established in the face of the pandemic. Tables 13–15 in the Appendix give an overview of loan guarantee programmes set up internationally at the same time as the Swiss programme. Most programmes focus on SMEs as the most relevant beneficiaries. Similar to the maturity of the loans in Switzerland, a maturity of 5 years is quite typical. The Swiss programme closes, however, at an unusually early date. Most programmes 15See Art. 6, COVID-19-Solidarbürgschaftsverordnung. 16Loan supply in terms of the loan volume, however, is also determined by supply-side factors, such as the parameters of the programme. For instance, the loan volume is capped at 10% of a firm’s revenue. The potential total loan volume hence varies across cantons. This fact is exploited by Kaufmann (2020) in his analysis of the macroeconomic employment effects of the programme. 17Few COVID-19 standard loan applications were initially rejected. Anecdotal evidence points to the fact that rejections were due to incorrectly filling in the application. Credit risk considerations may be more relevant for COVID-19 plus loans, but rejection rates seem to have been low also in this segment. 18The CRF was established simultaneously and in coordination with the COVID-19 loan programme. By mid-June, banks drew via the CRF liquidity amounting to almost two-thirds of the entire COVID-19 credit limits. 19The FINMA temporarily excluded central bank reserves from the leverage ratio calculation. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 5 of 22 Fig. 1 Firms in Switzerland and financing sources. Sources: FSO (BUR) and SECO. The graph on the left uses our set of firms, discussed in Section 4. The graph on the right uses data from a study commissioned by SECO and conducted by the Lucerne University of Applied Sciences and Arts in the fall of 2016. The survey covers 1922 SMEs in Switzerland (with less than 250 employees) and assesses their financing forms, sources and conditions. “Bank debt” typically consists of mortgages and credit lines, whereas “Other debt” contains, for example, bonds and trade credit were initially intended to be open until at least the end of 2020. The Swiss conditions in terms of the share of the loan guaranteed and interest rate are more on the generous side. An interest rate of 0% without a guaranteefeeforthefirsttrancheisatthelowerendofthe range and the guarantee of 100% is of course at the upper end. However, there are a couple of other countries that also offer such a comprehensive guarantee. Given these attractive terms, it is probably not surprising that the usage of the Swiss programme is considerable relative to Table 1 Descriptive statistics Groups Firms Mean Std. dev. Min Max Participation (yes/no) – 675,111 0.15 0.36 0.00 1.00 Lockdown index (sectors within cantons) 469 674,423 0.30 0.17 0.00 0.97 Home office index (sectors within cantons) 469 674,423 0.51 0.27 0.00 1.00 Short-time work (sectors within cantons) 1118 671,713 0.20 0.15 0.00 3.50 Retail payments (sectors within cantons) 540 344,859 −0.42 0.67 −1.00 4.22 Virus cases (in canton) 26 675,111 0.40 0.28 0.11 1.07 Fatality cases (in canton) 26 675,111 21.01 22.08 0.00 88.90 Cash ratio, mean (headcount groups within sectors) 45 234,067 0.28 0.08 0.11 0.41 Liquidity ratio, mean (sectors) 63 530,351 1.29 0.35 0.19 2.57 Liquidity ratio, mean (sectors within cantons) 560 471,257 2.70 1.51 0.72 51.08 Liquidity ratio, median (sectors within cantons) 560 471,257 1.66 0.54 0.51 5.43 External financing (headcount groups within sectors) 18 214,489 0.39 0.11 0.28 0.67 Debt ratio, mean (sectors) 44 483,976 0.66 0.14 0.29 1.04 Debt ratio, mean (headcount groups within sectors) 54 230,420 0.41 0.17 0.21 0.87 Debt ratio, mean (sectors within cantons) 561 471,728 0.72 0.26 0.35 6.73 Debt ratio, median (sectors within cantons) 561 471,728 0.68 0.10 0.31 0.97 Profit margin, mean (headcount groups within sectors) 40 218,682 0.09 0.06 0.02 0.32 Profit to int. ratio, mean (headcount groups within sectors) 31 160,133 0.36 0.31 0.07 1.71 Sources: FSO (BUR), JANUS, Faber et al. (2020), SECO, SNB, FOPH, FSO, CompNet. See main text for details. The table shows the number of groups available for the variable, and the corresponding number of firms to which the group variable can be matched. The mean, standard deviation, minimum and maximum are computed for the matched firms. See main text for the variable definitions Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 6 of 22 Fig. 2 Firm participation, by economic sector and canton. Sources: FSO (BUR), JANUS and own calculations. The economic sectors are obtained by aggregating the more granular NOGA two-digit codes. The sector Others contains the activities of households as employers, the production activities of households for their own use and the activities of extra-territorial organisations and bodies GDP in international comparison. The Swiss programme is similar in magnitude to the US programme and to the two programmes in the UK combined. Only the programmes in Hong Kong and Italy are larger in relation to GDP. The COVID-19 loan programme focuses on SMEs and aims to provide quick access to bank financing. Both of those aspects are motivated by the structure of firms in Switzerland and their financing sources. Figure 1presents the distribution of firms’ size in terms of the number of full-time equivalent employees (graph on the left) and their financing (graph on the right). The distribution of firms’ size illustrates the importance of small firms for the Swiss economy. More than 92% of firms have less than 10 employees, and over 99% have less than 250 employees, thereby fitting the definition of an SME used by the Swiss State Secretariat for Economic Affairs (SECO).20 Given the importance of smaller firms, it is not surprising that the programme focuses on SMEs. Moreover, an examination of the typical financing structure of firms in Switzerland indicates that a majority of SMEs do not have debt: 62% of all SMEs in Switzerland were debt-free before the pandemic. This phenomenon is most pronounced for the smallest firms with 2–10 employees: two out of three of those firms are exclusively equity financed. This share drops with increasing firm size: 50% of SMEs with 50–250 employees have some form of debt outstanding. Across all firm sizes, the dominating type of outside financing is bank debt. 20Those firms in that 99% of firms employ two out of three employees in Switzerland. For more information, please visit the following website: https:// www.kmu.admin.ch/kmu/en/home.html. Overall, the data indicate that a significant share of Swiss SMEs do not have an established credit relationship. This might be a problem if firms suddenly have to bridge liquidity shortfalls by outside finance (e.g. bank debt) and could be particularly problematic for young firms that have existed for only a couple of years. A government loan guarantee programme is a potential solution to this problem, as it eliminates credit risk and solves potential informational problems between borrowers and lenders, which otherwise might impede the extension of credit. 4Data 4.1 Construction of dependent variable Our analysis is based on data comprising all firms in Switzerland. We bring together two datasets: on the one hand, data from the entire registry of commerce are used (Betriebs- und Unternehmensregister, short BUR); on the other hand, data from the registry of all the COVID- 19 loans, recorded by the guaranteeing organisations, are used (called the JANUS database). The entries in the two datasets are matched through a unique firm identifier, which is available in both registries. We work with an anonymised version of the matched dataset, but we do know which firms have a COVID-19 loan and which firms do not. Both datasets are cross-sectional and correspond to a snapshot at the end of the COVID-19 loan programme.21 Table 1provides descriptive statistics. Our cleaned dataset contains 675,111 active firms in Switzerland that were eligible for a COVID-19 loan. This 21The September version of JANUS that we use reflects the loans outstanding as of 31 July 2020 (when the COVID-19 loan programme ended). The BUR reflects Swiss firms as of 17 August 2020. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 7 of 22 set of firms is obtained by selecting entities from an initial 1.87 million entries available in the BUR registry. We excluded entries that are not active, currently in liquidation, entities without economic activity (such as investment and legal purpose vehicles) as well as domestic and foreign government entities (such as public administrations). We kept data on financial companies as well as companies operating in the primary sector (i.e. agriculture) as those companies were also eligible for a COVID- 19 loan. The exclusion criteria are further detailed in Appendix “Data construction”. The BUR data also provide other information, such as the economic sector,22 firm size (in terms of full-time equivalent employees), firm age (via the entry date in the registry), or the canton (there are 26 cantons in Switzerland) in which the firm is legally registered.23 Our firm count is close to the available count of 656,364 active firms reported by the Federal Statistical Office (FSO) in January 2020. Similarly, our set of firms replicates well the economic sector, region, headcount and legal form distributions that are made available by the FSO (see Appendix “Data construction”). According to the latest estimate at the time of writing, there were 135,261 standard COVID-19 loans outstanding when the programme ended. This corresponds to about 20% of our firm count. We are able to match 103,605 loans to the BUR, as some loans were already paid back in full and not all loans have a unique firm identifier (see Appendix “Data construction”). Hence, we obtain by construction a lower participation rate of approximately 15%. Our data set shows that firms participated in the loan programme across sectors and cantons. Figure 2shows the participation rates by broad economic sectors and cantons. Overall, participation across cantons is characterised by a considerable heterogeneity. By sector, the dispersion is even larger. The sector with the highest participation rate is accommodation and food services,with 43%. The lowest sectoral participation rates, below 3%, are found in agriculture, mining and utilities and in others (consisting of household-related production and extraterritorial organisations). Across cantons, the participation rate ranges between 7% (Appenzell Innerrhoden) and 25% (Ticino). 4.2 Explanatory variables The figures described above reflect how firms that operate in different sectors and regions were affected by the crisis. However, an unanswered question is what drove 22We build sectoral categorisations based on the General Classification of Economic Activities (NOGA) two-digit codes. 23Unfortunately, the BUR registry does not contain financial information such as cash holdings or leverage. Hence, we resort to group-level financial variables in the analysis (see Table 1). Similarly, we do not have information on non-COVID-19 loans that a firm might have secured. This lack of information is important unless the decision to request a non-COVID-19 or a COVID-19 loan occurs at random, which is most likely not the case for larger firms. Section 5.5 shows that our results are robust when excluding larger firms. participation in the loan programme? We bring answers by considering three broad dimensions of loan demand, namely, a firm’s sensitivity to the lockdown, its exposure to the virus intensity, and the firm’s initial financial conditions. The different measures that we use are summarised in Table 1. Because these measures are not available at the firm level, we use group variables from various data sources; each firm is then matched to its corresponding group. 4.2.1 Sensitivity to the lockdown Loan demand may reflect the abrupt fall in revenue implied by lockdown restrictions. Assessing a firm’s sensitivity to the lockdown is not straightforward: hence, we resort to four different measures. To ensure exogeneity, our preferred measures are a lockdown index, which relies on physical proximity, and a home office index, which relies on the possibility to perform tasks at home. Faber, Ghisletta, and Schmidheiny (2020) built the index by using the Occupational Information Network (ONET)survey, which asks workers questions about the level of physical proximity that is required in their occupation. Individual survey answers are translated into an index that is available for economic sectors within cantons, yielding a total of 469 groups. The index ranges between zero and one. A value of zero corresponds to little physical proximity needed, whereas a value of one indicates that physical proximity is essential to the worker’s tasks. The lowest index values are found in sectors, such as financial and insurance activities, or agriculture, whereas the highest values are found in sectors such as accommodation or construction.24 Faber et al. (2020) also compute a home office index with the ONET survey. The home office index can be used as an alternative measure of lockdown restrictions.25 In contrast to the lockdown index, the home office index captures the possibility for a worker to perform tasks at home. A value of zero indicates that tasks cannot be operated remotely (for instance, a machine is needed), whereas a value of one implies that the worker can readily perform tasks from home. The two indices are exogenous in the sense that a firm cannot easily (or rapidly) alter the work conditions that require physical proximity for production or that allow workers producing from their homes. We complement the indices by using two indicators of business activity. First, the proportion of firms that use the Swiss short-time work scheme (or Kurzarbeit)ina given sector within a canton is obtained from SECO numbers relative to the firm counts in our cleaned dataset.26 24During the lockdown, essential sectors (such as food stores, pharmacies, petrol stations, banks or railway stations) were allowed to operate freely. These sectors are assigned an index value of zero by Faber et al. (2020). The index value is also set to zero for workers who report working in the public sector. 25Both indices capture the lockdown restrictions similarly. The correlation coefficient between the two indices is -0.65 when using household observations for which the indices can be mapped. 26The SECO short-time work numbers are as of April 2020. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 8 of 22 Fig. 3 Main explanatory variables by firm participation. Sources: FSO (BUR), JANUS, Faber et al. (2020), SECO, SNB, FOPH, FSO, CompNet. See main text for details. Notes: The distribution of the variables’ values by firm participation is shown. The liquidity and debt ratios are at the sector-canton level. The median is boxed around the 25th and 75th percentiles. The whiskers are the 10th and 90th percentiles, respectively This gives a total of 1,118 groups. Second, we obtained data on retail card payments in Switzerland from Kraenzlin, Meyer, and Nellen (2020). Based on this data, we compute the year-on-year percentage change in transaction values for April 2020. Comprising 540 groups, the data are available for sectors within cantons. 4.2.2 Exposure to virus intensity Loan demand may also reflect the severity of the pandemic situation per se. The cautious behaviour of households (i.e. going less to shops or buying more online) may increase with the severity of the pandemic. In particular, the degree of behavioural adjustment is likely to be regional. To measure the intensity of the virus spread, we use the cumulative cases (as a percentage of the cantonal population) in the canton in which the firm is legally registered. Additionally, we use the cumulative number of fatalities due to the virus (expressed per 100,000 inhabitants). Both measures are as of 13 July for the 26 Swiss cantons and are obtained from the Federal Office of Public Health (FOPH). 4.2.3 Initial financial conditions Loan demand may depend on a firm’s initial financial conditions. We measure liquidity and debt conditions via several group-level variables. The broader group-level liquidity variable that we have is a cash to assets ratio obtained from the CompNet survey of Swiss firms.27 The average ratio in 2017 is available across five headcount groups within nine sectors, representing a total of 45 groups. Additionally, we use a more granular liquidity ratio made available for the year 2018 by the FSO. In particular, the liquid asset to short-term debt ratio provided by the FSO is not only available for 63 sectors but also for sectors within cantons (totalling 560 groups). The corresponding mean and median ratios were computed by the FSO for groups that contained a minimum of five surveyed firms.28 External financing is measured in two ways. First, we use the proportion of firms with external financing (both bank and non-bank debt) in 2016. The data was made available for broad headcount groups within sectors and totals 18 groups.29 Second, we measure indebtedness via the debt to asset ratio. We use the average ratio in 2017 from CompNet; this ratio is available for 44 of the 45 headcount-sector groups mentioned above. Additionally, we use the more granular debt ratios made available for the year 2018 by the FSO. The FSO debt ratio is available for 54 sectors but also for sectors within 27CompNet documents individual firms across several countries. The survey consists of a sample of non-financial firms with at least one employee. For Switzerland, the CompNet 2017 sample covers around 20% of the country’s total firm revenue; see CompNet (2020) for further details. 28These more granular data on firm liquidity are not publicly available. 29See SECO (2016), Studie zur Finanzierung der KMU in der Schweiz . Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 15 of 22 Table 7 Results: Firm age interactions (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) (4) Part.(y/n) Age <1×Interacted measure 2.33*** 0.23* −0.16*** −0.21 Age [ 1, 5)×Interacted measure 2.50*** 0.46*** −0.15*** −0.32* Age [ 5, 10)×Interacted measure 1.74*** 0.75*** −0.15*** −0.29 Age 10+×Interacted measure 1.24*** 0.92*** −0.07** −0.09 Headcount dummies Yes Yes Yes Yes Age dummies Yes Yes Yes Yes Other demand determinants Yes Yes Yes Yes Constant Yes Yes Yes Yes Observations 471211 471211 471211 471211 Log-likelihood −212094.74 −212153.96 −212195.25 −212268.08 Interacted measure Lockdown index Virus cases Liquidity ratio Debt ratio Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. The interacted measures are listed in the last line of the table. Age is measured in years since the firm entered the registry of commerce. The first age group (Age <1) is the reference group for the coefficient of the chosen interacted variable. The coefficients of the other age groups consist of this reference coefficient plus the interaction term of the given age group. The other demand determinants comprise the Table 2variables (lockdown index, virus cases, liquidity ratio, debt ratio) excluding the chosen interacted variable shown in the respective columns. Standard errors are clustered at the level of the grouped variable of interest. The number of observations varies depending on data availability of the grouped variables. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively. 5.5 Robustness We provide three robustness checks. First, our findings are robust in smaller subsamples. Tables 9and 10 show the regression results for subsamples based on firm age and firm size groups. Estimating with subsamples is more restrictive than estimating with the interaction terms used in the previous subsection. With only a few exceptions, the variables of interest remain significant and the corresponding coefficient signs are unchanged for the subgroups. Second, our estimates are not affected when controlling via fixed effects for cantons and sectors. While adding interacted canton-sector dummies would by construction remove the variations exploited in our explanatory variables, we show in Table 11 that canton dummies do not affect the estimates that rely on sectoral variations (that is, the lockdown index and financial conditions). Likewise, adding dummy variables for sectors does not affect the estimate of virus intensity, which uses cantonal Table 8 Results: Firm size interactions (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) (4) Part.(y/n) FTE [ 0, 10)×Interacted measure 1.68*** 0.70*** −0.11*** −0.25 FTE [ 10, 50)×Interacted measure 2.01*** 1.08*** −0.11*** 0.06 FTE [ 50, 250)×Interacted measure 2.77*** 0.71** −0.12** 0.14 FTE 250+×Interacted measure 1.70* 0.99*** −0.23** 0.41 Headcount dummies Yes Yes Yes Yes Age dummies Yes Yes Yes Yes Other demand determinants Yes Yes Yes Yes Constant Yes Yes Yes Yes Observations 471211 471211 471211 471211 Log-likelihood −212260.43 −212245.70 −212283.71 −212267.28 Interacted measure Lockdown index Virus cases Liquidity ratio Debt ratio Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. The interacted measures are listed in the last line of the table. Firm size is measured in FTE employees. The first headcount group (FTE[0,10)) is the reference group for the coefficient of the chosen interacted variable. The coefficients of the other headcount groups (in FTE) consist of this reference coefficient plus the interaction term of the given headcount group. The other demand determinants comprise the Table 2variables (lockdown index, virus cases, liquidity ratio, debt ratio) excluding the chosen interacted variable shown in the respective columns. Standard errors are clustered at the level of the grouped variable of interest. The number of observations varies depending on data availability of the grouped variables. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 16 of 22 Fig. 7 Margins of key measures, by firm age. Predictive margins resulting from Table 7are shown. Whiskers indicate 95% confidence intervals Fig. 8 Margins of key measures, by firm size. Predictive margins resulting from Table 8are shown. Whiskers indicate 95% confidence intervals Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 17 of 22 Table 9 Robustness in subsamples, by firm age (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) (4) Part.(y/n) Lockdown index (sectors within cantons) 2.20*** 2.42*** 1.73*** 1.29*** Virus cases (in canton) 0.22 0.49*** 0.75*** 0.89*** Liquidity ratio, mean (sectors within cantons) −0.16*** −0.14*** −0.14*** −0.07** Debt ratio, mean (sectors within cantons) −0.30 −0.40** −0.33 −0.04 Constant Yes Yes Yes Yes Headcount dummies Yes Yes Yes Yes Age dummies Yes Yes Yes Yes Observations 12023 128230 130435 200517 Log-likelihood −5087.42 −56872.18 −53028.92 −96746.35 Sample Age <1Age[1,5)Age [ 5, 10)Age 10+ Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. Standard errors are clustered at the sector-canton level of the FSO financial variables. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively Table 10 Robustness in subsamples, by firm size (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) (4) Part.(y/n) (5) Part.(y/n) Lockdown index (sectors within cantons) 1.75*** 1.67*** 2.03*** 2.86*** 1.51* Virus cases (in canton) 0.73*** 0.69*** 1.08*** 0.75*** 0.95*** Liquidity ratio, mean (sectors within cantons) −0.11*** −0.11*** −0.10*** −0.07* −0.22* Debt ratio, mean (sectors within cantons) −0.21 −0.24 0.03 0.11 0.15 Constant Yes Yes Yes Yes Yes Headcount dummies Yes Yes Yes Yes Yes Age dummies Yes Yes Yes Yes Yes Observations 471211 432440 30809 6828 1128 Log-likelihood −212285.10 −188085.77 −19509.86 −3951.71 −502.12 Sample All firms FTE [ 0, 10)FTE [ 10, 50)FTE [ 50, 250)FTE 250+ Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. Standard errors are clustered at the sector-canton level of the FSO financial variables. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively Table 11 Robustness with canton and sector dummies (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) Lockdown index (sectors within cantons) 1.75*** 1.82*** −0.26 Virus cases (in canton) 0.73*** 1.17 0.66*** Liquidity ratio, mean (sectors within cantons) −0.11*** −0.12*** 0.01 Debt ratio, mean (sectors within cantons) −0.21 −0.19 −0.10 Constant Yes Yes Yes Headcount dummies Yes Yes Yes Age dummies Yes Yes Yes Canton dummies No Yes No Sector dummies No No Yes Observations 471211 471211 471211 Log-likelihood −212285.10 −211706.62 −204289.58 Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. Standard errors are clustered at the sector-canton level of the FSO financial variables. Sector dummies use the sectoral breakdown of the lockdown index. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 18 of 22 Fig. 9 Participation rate, by legal form. Sources: FSO (BUR), JANUS and own calculations. The legal forms are those prevailing in Switzerland variations. Third, as participating firms are not allowed to pay dividends, the participation rate of limited companies might be lower than that for firms with other legal forms. However, as shown by Fig. 9,thisisnotthe case. 6Conclusions We analyse the determinants of firm participation in the Swiss COVID-19 loan programme by using a comprehensive dataset. Overall, 20% of all firms applied for a COVID-19 loan, resulting in a sizeable programme of 2.4% of GDP. Our key findings for firm participation are as follows. First, the exposure of the firm to lockdown restrictions and the intensity of the virus in the specific region are important determinants of participation. Second, we show that firms associated with lower liquidity ratios had a significantly higher probability of participating in the programme. Third, we find no clear evidence that firm indebtedness affects participation in the programme and no evidence that pre-existing zombie firms participated more strongly in the loan programme. Fourth, we show that the programme reached all firms, including younger and smaller firms, which could be financially more vulnerable, as they are less likely to obtain outside finance during a crisis. In light of these findings, we conclude that given its stated objective, the programme seems to have been successful. Our analysis is intended to contribute to a broader understanding of the economic measures that were taken by governments during the COVID-19 crisis. Given the potentially far-reaching implications of such largescale policy measures, further empirical and theoretical research in this area is essential. For example, the impact of the programme on firm (e.g. profits, employment, and survival) and macroeconomic outcomes could be studied after some time has passed and more reliable data on actual outcomes become available. As another example, the role of firm networks (supply chains etc.) might be analysed with regard to participation. We leave these and further questions for future research. Appendix Data construction Registry of commerce. Firms are selected from the 1.87 million entries in the registry of commerce (Betriebs- und Unternehmensregister, short BUR).35 This is done by excluding the following entries: •Administratively and statistically non-active entries, •Entries with a non-definitive unique identifier (UID), •Non-relevant firm types (investment vehicles, legal purpose entities, foreign and domestic government entities, government companies), •Non-relevant enterprise types (administrative link or VAT units, public sector, public enterprise without personnel, public sector administrative unit, enterprise owned by foreign state, foreign units without employment, errors), •Entities currently in liquidation or in bankruptcy, •Entities without information on the canton, economic sector (NOGA two-digit code), headcount (groups), or entry date in the registry (hence firm age). Figure 10 in Appendix shows that the resulting sample replicates well the firm distributions by region, sector, headcount group and legal form; these distributions are available for 2018 from the FSO. Register of COVID-19 loans. Firms that participated in the COVID-19 loan programme are contained in the JANUS database developed by the Swiss guaranteeing institutions. Most firms can be matched to the registry of commerce via a unique identifier. Additionally, we match firms that obtained a COVID-19 plus loan, as these larger firms must take the standard COVID-19 loan as a first tranche. Out of the 135,261 firms that took a COVID- 19 loan by the end of the programme (end of July 2020), 103,605 firms can be matched to the register of commerce. The discrepancy is due to two reasons. First, not all unique identifiers can be reconciled. Second, the JANUS database continues to be updated; some (3990) loans were either already paid back in full or not yet entered in the database September version that we use (which contains 131,271 loans). 35For more details, see: https://www.bfs.admin.ch/bfs/en/home/registers/ enterprise-register/business-enterprise-register.html. Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 19 of 22 Fig. 10 Comparison with available firm distributions. The charts compare the final sample (BUR) to the distributions made available by the FSO in 2018. Firm counts for which data are available in the categories defined by the FSO are used. Headcounts use full-time equivalent (FTE) employees Main results with all coefficients displayed Table 12 Main results, all coefficients displayed (1) Part.(y/n) (2) Part.(y/n) (3) Part.(y/n) (4) Part.(y/n) (5) Part.(y/n) Lockdown index (sectors within cantons) 2.16*** 1.75*** Virus cases (in canton) 0.74*** 0.73*** Liquidity ratio, mean (sectors within cantons) −0.13*** −0.11*** Debt ratio, mean (sectors within cantons) −0.06 −0.21 FTE [10,50) 1.14*** 1.23*** 1.06*** 1.07*** 1.06*** FTE [50,250) 0.89*** 0.88*** 0.72*** 0.72*** 0.80*** FTE 250+ 0.23* 0.20** 0.00 0.02 0.06 Age [1,5) 0.20*** 0.20*** 0.13*** 0.13*** 0.15*** Age [5,10) 0.02 −0.00 −0.05 −0.07* −0.02 Age 10+ 0.17*** 0.18*** 0.15*** 0.14*** 0.18*** Constant −2.63*** −2.26*** −1.38*** −1.67*** −2.27*** Observations 674423 675111 471257 471728 471211 Log-likelihood −277189.69 −281379.54 −215653.56 −216776.51 −212285.10 Logit model. The dependent variable is a firm-level binary variable that indicates firm participation in the loan programme. Standard errors are clustered at the level of the grouped variable of interest; in column (5), clustering is at the sector-canton level of the FSO financial variables. The number of observations varies depending on data availability of the grouped variables. ***, ** and * denote statistical significance (two-tailed) at the 1%, 5% and 10% significance levels, respectively. Headcount is measured in full-time equivalent (FTE) employees, and firm age is in years. The base group has FTE [ 0, 10)and age <1 Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 20 of 22 Table 13 Loan guarantee programmes internationally Jurisdiction Beneficiary Guarantee/ maximum loan size Closing date Interest rate Loan maturity Usage (CHF bn) Usage (% of GDP) Switzerland Firms with turnover below CHF 500m 100% up to CHF 500’000, 85% up to CHF 20m; maximum of 10% of annual turnover 31 Jul 2020 0% interest rate up to CHF 500’000; part over CHF 500’000: 0.5% plus a bank specific rate on the remaining 15% of the loan 5 (+2) years 16.9 2.4% Australia (Coronavirus SME Guarantee Scheme) SME 50%/ AUD 250’000 30 Sep 2020 Initial 6-month interest holiday; rate decided by lender Up to 3 years N/A N/A Canada (Canada Emergency Business Account, CEBA) Small businesses and non-profits 100%/ CAD 40’000 N/A 0% interest rate, no fees or principal repayments until end-2022, then 5% interest rate Up to 5 years 20.6 1.3% France (Bpifrance) All firms 70–90% (higher for smaller firms); maximum of 25% of 2019 revenue or two years of payrolls 31 Dec 2020 No payment in the first year; interest rate set by the bank, guarantee cost ranging b/w 25–200 bp Repay by end-2020, or extended by maximum of 5 year 6.5 0.3% Source: Baudino (2020) and national sources Table 14 Loan guarantee programmes internationally (cont.) Jurisdiction Beneficiary Guarantee/ maximum loan size Closing date Interest rate Loan maturity Usage (CHF bn) Usage (% of GDP) Germany (Bundesregelung Kleinbeihilfen 2020) SME 100% for loans up to: EUR 500’000 for firms with 50 employees; EUR 800’000 for others 31 Dec 2020 Individual loan rate determined by bank N/A 5.0 (after 100 days) 0.1% Germany (Kreditanstalt für Wiederaufbau, KfW) All firms 90% for SME, 80% for others; EUR 1bn per company 31 Dec 2020 Subsidised loan rate (lower for SME) Up to 5 years 58.1 1.5% Hong Kong SAR (Special Financing Guarantee Scheme, SFGS) SME 100%; up to total amount of employee wages and rents for six months or HKD 4m Avaiable for 12 months Optional principal moratorium for 1 year; rate is Prime Rate minus 2.5%; no guarantee fees Up to 3 years 10.9 3.2% Italy (Fondo di Garanzia PMI) SME 80–90%: loans up to EUR 1.5m; 100%: loans up to EUR 800’000 17 Dec 2020 Guarantee cost waived; loan rates set by lenders N/A 87.9 4.6% Italy (Cassa Depositi e Prestiti) All firms (SME must first apply for SME plan) 70–90%; maximum of 25% of 2019 revenues or twice payroll costs 31 Dec 2020 Individual loan rate determined by bank 6 years N/A N/A Source: Baudino (2020) and national sources Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 21 of 22 Table 15 Loan guarantee programmes internationally (cont.) Jurisdiction Beneficiary Guarantee/ maximum loan size Closing date Interest rate Loan maturity Usage (CHF bn) Usage (% of GDP) Spain (Instituto de Credito Oficial) All firms 60–80% depending on company size and new/renewed loan); no explicit maximum 30 Sep 2020 Guarantee fees of 20–120 bp (to be borne by the bank) Up to 5 years N/A N/A United Kingdom (Coronavirus Business Interruption Loan Scheme, CBILS) SME 100% up to GBP 250’000; 80% above GBP 250’000; up to GBP 5m N/A Interest holiday in first 12 months; guarantee fee waived, lenders pay a fee; loan terms set by each lender Up to 6 years 21.3 0.8% United Kingdom (Bounce Back Loan Scheme, BBLS) SME 100%; GBP 2’000– 50’000 but maximum of 25% of turnover N/A no fees, interest or repayment of principal in the first 12 months; after 12 months: interest rate of 2.5% Up to 6 years 42.4 1.6% USA (Paycheck Protection Program, PPP - CARES Act) SME 100% to end- 2020; up to the lesser of USD 10m or a payroll-based amount 30 Jun 2020 (extended to 8 Aug 2020) 1% interest rate; optional interest payment holiday for first 6 months 2 (5) years 477.8 2.5% Source: Baudino (2020) and national sources Fuhrer et al. Swiss Journal of Economics and Statistics (2021) 157:2 Page 22 of 22 Abbreviations AUD: Australian dollar; BBLS: Bounce Back Loan Scheme; BUR: Betriebs- und Unternehmensregister (registry of commerce); CAD: Canadian dollar; CARES Act: Coronavirus Aid Relief and Economic Security Act; CBILS: Coronavirus Business Interruption Loan Scheme; CEBA: Canada Emergency Business Account; CHF: Swiss franc; CRF: COVID-19 refinancing facility; EUR: Euro; FINMA: Eidgenössische Finanzmarktaufsicht (Swiss financial market supervisory authority); FOPH: Federal Office of Public Health; FSO: Federal Statistical Office; FTE: Full-time equivalent; GBP: Pound sterling; GDP: Gross domestic product; HKD: Hong Kong dollar; KfW: Kreditanstalt für Wiederaufbau (Credit Institute for Reconstruction); NOGA: Nomenclature Générale des Activités économiques (General Classification of Economic Activities); OECD: Organisation for Economic Co-operation and Development; ONET: Occupational Information Network; PPP: Paycheck Protection Program; SECO: Secrétariat d’Etat à l’économie (Swiss State Secretariat for Economic Affairs); SFGS: Special Financing Guarantee Scheme; SME: Small and medium-sized enterprises; SNB: Swiss National Bank; UID: Unique identifier; UK: United Kingdom; US: United States; USD: US dollar; VAT: Value-added tax; ZHAW: Zürcher Hochschule für Angewandte Wissenschaften (Zurich University of Applied Sciences) Acknowledgements We thank Romain Baeriswyl, Roman Baumann, Petra Gerlach, Oliver Gloede, Carlos Lenz, Cédric Tille (the editor), the participants of the research seminar at the Swiss National Bank and the Federal Finance Administration as well as an anonymous referee for very helpful and constructive comments. Moreover, we are grateful to Marius Faber, Maïlys Korber, Christoph Meyer, Christoph Odermatt, Kejo Starosta, Markus von Allmen and Reto Wernli for help with the data. The various datasets used in this paper were kindly provided by the State Secretariat for Economic Affairs, the Federal Statistical Office, the Institute of Financial Services of the Lucerne University of Applied Sciences and Arts, and the Faculty of Business and Economics of the University of Basel. The views, opinions, findings and conclusions or recommendations expressed in this paper are strictly those of the authors. They do not necessarily reflect the views of the Swiss National Bank. The Swiss National Bank takes no responsibility for any errors or omissions in, or for the correctness of, the information contained in this paper. Authors’ contributions All three authors have contributed to this paper on an equal footing. All authors have read and approved the final manuscript. Funding Not applicable. Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request and only with permission of the Swiss National Bank, the State Secretariat for Economic Affairs, the Federal Statistical Office, the Institute of Financial Services of the Lucerne University of Applied Sciences and Arts and the Faculty of Business and Economics of the University of Basel, as restrictions apply to the availability of these data and so are not publicly available. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Received: 21 January 2021 Accepted: 27 March 2021 References Andrews, D., & Petroulakis, F. (2019). Breaking the shackles: Zombie firms, weak banks and depressed restructuring in europe. ECB Working Paper,No 2240. Baudino, P. (2020). Public guarantees for bank lending in response to the COVID-19 pandemic. FSI Briefs,No 5. Beck, T., Klapper, L.F., & Mendoza, J.C. (2010). The typology of partial credit guarantee funds around the world. Journal of Financial Stability,6, 10–25. Bernanke, B., & Gertler, M. (1989). Agency costs, net worth, and business fluctuations. The American Economic Review,79(1), 14–31. Bernanke, B.S., Gertler, M., & Gilchrist, S. (1999). The financial accelerator in a quantitative business cycle framework. In J.B. Taylor & M. Woodford (Eds.) Handbook of Macroeconomics, vol. 1 (pp. 1341–1393): Elsevier. Chap. 21. Brülhart, M., Lalive, R., Lehmann, T., & Siegenthaler, M. (2020). COVID-19 financial support to small businesses in Switzerland: Evaluation and outlook. Swiss Journal of Economics and Statistics,156(15), 1–13. Carlstrom, C.T., & Fuerst, T.S. (1997). Agency costs, net worth, and business fluctutations: A computable general equilibrium analysis. The American Economic Review,87(5), 893–910. Chodorow-Reich, G. (2014). The employment effects of credit market disruptions: Firm-level evidence from the 2009-9 financial crisis. The Quarterly Journal of Economics,129(1), 1–59. CompNet (2020). User guide for the 7th Vintage of the CompNet dataset. Retrieved from: https://www.comp-net.org/fileadmin/_compnet/ user_upload/Documents/7th_Vintage/7th_Vintage_User_Guide.pdf. Cowling, M. (2010). The role of loan guarantee schemes in alleviating credit rationing in the UK. Journal of Financial Stability,6, 36–44. Faber, M., Ghisletta, A., & Schmidheiny, K. (2020). A lockdown index to assess the economic impact of the coronavirus. Swiss Journal of Economics and Statistics,156(11), 1–23. Gertler, M., & Gilchrist, S. (1993). The role of credit market imperfections in the monetary transmission mechanism: Arguments and evidence. Scandinavian Journal of Eonomics,95, 43–64. Gertles, M., & Gilchrist, S. (1994). Monetary policy, business cycles, and the behavior of small manufacturing firms. The Quarterly Journal of Economics, 109(2), 309–340. Gonzalez-Uribe, J., & Wang, S. (2020). The effects of small-firm loan guarantees in the uk: Insights for the COVID-19 pandemic crisis. Mimeo. Green, A. (2003). Credit guarantee schemes for small enterprises: An effective instrument to promote private sector-led growth?. SME Technical Working Papers Series,No. 10. Holmström, B., & Tirole, J. (1997). Financial intermediation, loanable funds, and the real sector. The Quarterly Journal of Economics,62(3), 663–691. Honohan, P. (2010). Partial credit guarantees: Principles and practice. Journal of Financial Stability,6, 1–9. Kaufmann, D. (2020). Does government-backed lending prevent unemployment? An assessment of the Swiss COVID-19 lending program. University of Neuchâtel, Mimeo. Kiyotaki, N., & Moore, J. (1997). Credit cycles. Journal of Political Economy,105(2), 211–248. Kraenzlin, S.P., Meyer, C., & Nellen, T. (2020). COVID-19 and regional shifts in Swiss retail payments. Swiss Journal of Economics and Statistics,156(14), 1–20. Riding, A., & Haines, G. (2001). Loan guarantees: Cost of default and benefits to small firms. Journal of Business Venturing,16(6), 565–612. Riding, A., Madill, J., & Haines, G. (2007). Incrementality of SME loan guarantees. Small Business Economics,29, 47–61. Saito, K., & Tsuruta, D. (2014). Information asymmetry in SME credit guarantee schemes: Evidence from Japan. RIETI Discussion Paper Series,14-E-042. Stiglitz, J.E., & Weiss, A. (1981). Credit rationing in markets with imperfect information. The American Economic Review,71(3), 393–410. Zoller-Rydzek, B., & Keller, F. (2020). COVID-19: Guaranteed loans and zombie firms. MPRA Paper,No. 100897. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.