LSE - Small Logo
LSE - Small Logo

Juanita González-Uribe

Su Wang

May 4th, 2020

Large economic benefits justify small-firm loan guarantees in the Covid-19 crisis

0 comments | 1 shares

Estimated reading time: 5 minutes

Juanita González-Uribe

Su Wang

May 4th, 2020

Large economic benefits justify small-firm loan guarantees in the Covid-19 crisis

0 comments | 1 shares

Estimated reading time: 5 minutes

Loan guarantees are a main policy response during the Covid-19 crisis. These guarantees were also popular during the Great Recession, but their demand from businesses in this crisis is unprecedented. The relatively worse economic conditions undoubtedly explain much of current demand, but also possibly highlight the differences in the guarantees’ design for each crisis. Will the new guarantees be cost effective, and what will be their impact?

Loan guarantees provide credit access to businesses that have no requisite collateral to access market loans. The question of what their effects are is contentious, however. Supporters claim that guarantees alleviate financial constraints. Critics argue that they increase participants’ risk-taking by allowing firms to borrow without pledging collateral and by providing lenders with a guarantee.

In responding to critics, the loan guarantee programs implemented during the Great Recession included several design features to curtail participants’ risk-taking incentives.

For example, in the UK scheme that started in 2009, lenders are incentivised by the partial guarantees on individual loans (75% of outstanding balance) and by the lender-level caps on the overall amount of guarantees sought (9.75% of the scheme’s size). Borrowers are incentivised because they remain fully liable, and because banks can request additional personal guarantees. Borrowers are also charged a premium of 2% in addition to the charges by lenders (on average, 5.8%) in order to fund the scheme. Perhaps as a result of this premium, take up relative to the target population was low during the Great Recession and has remained low since. Less than £800M in loans were issued by the scheme in 2009 to fewer than 7,000 companies, which corresponds to less than 5% of eligible firms.

By contrast, the loan guarantees currently offered by the UK government — the Coronavirus Business Interruption Scheme — charges no premium to borrowers, and in addition provides 12 months free of interest payments and of any lender-levied fees, although businesses remain fully liable. Lenders cannot take personal guarantees for low-value loans (below £250K), but the overall cap for lenders remains, and so does a slightly higher (80%) loan-level guarantee. Perhaps partly as a result of these new features, the demand has been unprecedented; in the three weeks since the scheme’s launch, more than 36,000 applications have been completed, and 15,000 businesses have been approved for a total of £2.82B.

What will be the effects of the new guarantees? Despite the increasing prevalence of loan guarantees, evidence for the success of such schemes is still sparse. This is due, in large part, to difficulties in accessing detailed data for small firms. But it is also because constructing meaningful counterfactual scenarios is challenging: What would have been the performance of firms absent the guarantees?

Our evidence from González-Uribe and Wang (2020) provides novel and useful insights for the Covid-19 crisis. In our paper, we measure the effects of the Enterprise Finance Guarantee, the UK loan guarantee program that started in 2009 as part of the UK’s business policy response to the Great Recession. Our results are consistent with the guarantees enabling a small group of financially constrained firms to retain workers during the Great Recession who otherwise would have been laid off, and whose retention was fundamental in rebuilding the businesses post-recession.

The estimation uses variation in participation from the program’s firm-size unexpected eligibility threshold. For eligible firms near the threshold, the guarantees increased average four-year profits, productivity, survival, and employment growth but not investment, relative to non-eligible firms. The relative increases in performance and employment occurred in lockstep with debt issuances, were absent prior to 2009, did not revert during 2010–2013, and mask large heterogeneity. The results are entirely driven by industries with high costs of employee training.

Additional evidence suggests that these results are mainly driven by effects on the minority of eligible firms that take up the scheme. Under this assumption, annual returns to guaranteed debt range between 16% and 20%, which comfortably exceed the above market scheme rates, and are below the cost of outside funding options.

There are three lessons from the results in González-Uribe and Wang (2020) for the Covid-19 crisis.

First, our results show that loan guarantees can have large economic benefits that justify their use as policy responses during the Covid-19 crisis. In our paper, we estimate that despite the low take-up, the economic benefits of guarantees during the Great Recession were 1.5 times their cost for firms near the eligibility threshold.

Second, the findings suggest that loan guarantee programs alone are not enough to incentivise the retention of all type of employees. During the Great Recession, firms used the guarantees only to retain the types of workers that could justify the guarantees’ costs. As we have also argued in Gonzalez-Uribe, Wang and Djankov (2020), this result implies that stimulus programs based on guarantees alone can be regressive because the poorer workers are also the more likely to have jobs with low training costs. Other schemes that also target workers in low training jobs who are easier to replace, such as the Job Retention Scheme, are therefore warranted to mitigate job losses during the Covid-19 crisis.

Finally, a word of caution. While lower costs to borrowers and the elimination of personal guarantees will certainly help increase the demand of guaranteed loans, the downside of these new features is potential risk-taking by borrowers and/or lenders (as shown by Lelarge, Sraer and Thesmar (2008) for the French loan guarantees implemented in the 90s), as well as the possibility of directing funds towards the wrong businesses and preventing efficient labour reallocation. Only time will tell whether the benefits from these new loan guarantees will compensate for the potential long-term difficulties when the loans come due for repayment in the future.

♣♣♣

Notes:


Juanita González-Uribe is an assistant professor at LSE. She holds a PhD in finance and economics from Columbia University. Her research focuses on entrepreneurship, private equity, innovation and policy. Her work has been published in prestigious journals, and has won several prizes, including the Jaime Fernandez de Araoz Award (JFA, 2017).

 

Su Wang is currently an assistant professor of finance at Amsterdam Business School, University of Amsterdam. Her research interests mainly lie in empirical corporate finance and entrepreneurship, with a focus on private and small and medium enterprises (SME). Su obtained her PhD in finance from LSE.

 

 

About the author

Juanita González-Uribe

Juanita González-Uribe is an assistant professor at LSE. She holds a PhD in finance and economics from Columbia University. Her research focuses on entrepreneurship, private equity, innovation and policy. Her work has been published in prestigious journals, and has won several prizes, including the Jaime Fernandez de Araoz Award (JFA, 2017).

Su Wang

Su Wang is currently an assistant professor of finance at Amsterdam Business School, University of Amsterdam. Her research interests mainly lie in empirical corporate finance and entrepreneurship, with a focus on private and small and medium enterprises (SME). Su obtained her PhD in finance from LSE.

Posted In: Economics and Finance | LSE alumni | LSE Authors

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.