VoxEU Column COVID-19

Support for small businesses amid COVID-19

A sizeable proportion of enterprises, especially SMEs, in receipt of financial assistance from the government will fail to repay. This column asks whether, and to what extent, it may be beneficial to apply a screening mechanism to deter those mostly likely to fail to repay from seeking financial assistance in the first place. The answer largely turns on the relative weights attached for the objectives of stabilisation as compared with allocative efficiency.

The combination of the Coronavirus pandemic and the policy measures of lockdown and quarantine introduced in response has had a drastic effect on the cash flows and solvency of businesses in the many countries affected, and in particular on small, and medium-sized enterprises (SMEs), which generally have had less own resources and no access to external finance except through their banks. Fairlie (2020), for example, estimated that “the number of active business owners in the United States plummeted by 3.3 million or 22% over the crucial two-month window from February to April 2020”. The drop in business owners was the largest on record, and losses were felt across nearly all industries and even for incorporated businesses (see also Foroohar 2020).

So, if a country was to avert an economic collapse, with a large proportion of its SME population being forced to shut up shop, the need was to get external financial assistance to them, and quickly. Given the massive numbers of SMEs in any country (approximately 5.9 million in the UK, making up over 99% of all businesses, according to UK Parliament business statistics in 2019), this could hardly be done, at least not quickly enough, directly from a government office, since they were not set up to do this; rather it had to be done via the existing relationship between SMEs and their main bank.

But a problem is that SMEs have a relatively high failure rate and are notoriously liable to fail to repay the due amount on their borrowing, principal and interest, thus becoming non-performing loans (NPLs) on bank books. The likelihood of credit extension to SMEs transforming into NPLs in the aftermath of the Coronavirus pandemic will obviously be even greater given the manifold uncertainties and changes to conditions and behaviour that the pandemic, and the policy response, have generated.  Moreover, banks’ losses from other sources in these conditions are already massive (Morris and Walker 2020).

So if the banks themselves were to be left carrying the can for any significant share of the losses arising from the NPLs on such loans, they would have wished to be extremely careful in monitoring and checking which SMEs would be provided with such emergency credit, and which would be refused. Such monitoring, however, takes time and effort. Moreover, the banks might be more conservative in their own interests than would be socially, or politically, desirable.

For all these reasons, in the UK the government then decided that such emergency loans to SMEs, known as ‘bounce-back loans’ (BBLs), would henceforth be 100% guaranteed by the government, i.e. they would not count as NPLs or cause losses to the banks (although the scheme requires lenders to pursue defaulters before calling in the government guarantee, it is difficult to believe that they will spend much time or effort in doing so).  As a result, there was no reason for the banks not to provide all-comers, who were demonstrably prior SMEs on their books, with loans immediately on the occasion of being asked. Nor was there any real likelihood that such credit expansion would be constrained by (regulatory) capital or liquidity requirements. Quantitative easing (QE) had made, and would increasingly make, liquidity in fulsome supply. Since credit expansion in the face of the pandemic emergency was the government’s intention, any shortage of bank capital that might occur would be offset perhaps by relaxing the rigour of the regulations and/or by restricting bank pay-outs in buybacks, dividends, and, even perhaps, in executive remuneration, as the ECB has just done.

There is an upper limit on the BBLs of £50,000 per claimant. The loans carry a fixed interest rate of 2.5% (though zero in the first year). Given these attractive terms there has, not surprisingly, been a rapid and huge take-up, amounting to £28 billion as of end June (Evans 2020). Apart from the normal costs of bankruptcy, there are no special, or additional, sanctions on borrowers who then default. In view of the massive uncertainties of the current situation, many might think it worth taking the gamble of borrowing the money in order to see whether they might have a chance of being successful in the eventual recovery, even when they have private information that the chances are slim and that they are more likely than not to fail. There may even be some who intend to divert such funds entirely to their own consumption, hoping to evade attention in the confusion and mass defaults that will ensue. But this is fraud, and should be pursued and constrained by the usual processes.

The government guarantees and favourable contractual terms have triggered a series of public debate on the likely massive defaults on the UK’s emergency loans (e.g.  Bounds and Thomas 2020), Morris et al. 2020, Treanor 2020).  But why does it matter that many of these BBL loans will never get repaid, and that attempts to chase up defaulters will involve much cost and effort? One reason, obviously, is that it will add to future public sector debt and deficit, though one has to net off the increased revenue that such extra expenditure will provide in the short run (Thomas et al. 2020).

But there is another reason for concern, which is the likely misallocation of resources that will come from keeping alive enterprises that are sub-par, even zombie companies (e.g. Thomas and Parker 2020). Similarly, Elliott (2020) wrote, “[r]ight now, governments are making, guaranteeing, or encouraging very large sums of loans with weak or non-existent credit underwriting. That may be the only choice at the moment, but it is not a long-term solution. Propping up companies that cannot survive in the longer run without continuing support becomes a real drag on an economy.”

Assuming that the government puts some weight on allocative efficiency in order to raise productivity and output over future years (as well as through a desire to maintain current employment and output in the face of the pandemic and the lockdown), it will need to try to screen out unprofitable, and less profitable, potential borrowers. Recall that it has withdrawn this role from banks in pursuit of a swifter disbursement of funds, via 100% guarantees.

There are two ways of doing such screening, ex ante and ex post. Under ex ante screening, the government will only lend to SMEs with a proven record of past profitability. This works well if the future is going to be like the past; but one general assessment is that the pandemic has greatly changed the prospective conjuncture, so that future patterns of behaviour may differ a lot from those in the past. A second objection to ex ante screening is that it would generate many hard cases and, therefore, could evoke considerable political and social opposition initially. For example, the criterion that an SME has to be able to show profits in each of its last two accounting years would exclude an enterprise with huge profits in one of those years and a tiny loss in the other.

The second, ex post, form of screening involves the government imposing an additional (pecuniary) penalty on those failing to be sufficiently profitable to pay back the loan in full. This would work better if the potential borrowers had reasonably good (private) information on whether they were likely to succeed in the changed conjuncture of the recovery from the pandemic. This is the condition that we assume in our model in Goodhart et al. (2020). Even if this condition holds, it does, of course, have some further disadvantages. Unsuccessful borrowers could try to avoid the extra penalty by moving abroad or hiding their income, as with student loan repayment, where the pay-back is only a fraction of the outlay. In a sense, penalising the unsuccessful is akin to kicking a person when they are already down; so while ex post screening/sanctions would generate less political opposition initially, it would probably generate more afterwards.

Both ex ante, and ex post, screening have disadvantages. Possibly partly for these reasons, the UK government decided not to do any such screening on its BBL scheme. It is now, almost certainly, too late to reverse that decision, since that would represent a retroactive adjustment to the scheme’s conditions. But our concern has been rather to assess the normative issue of whether, and what degree of, (ex post) screening would have been socially optimal, rather than to propose any positive change to current policies. Bygones are bygones, and policy had to be made under extreme pressure in the heat of the moment.

To provide a normative perspective on the social optimality of screening for the government loan schemes to support SMEs, we developed an infinite horizon model with two sectors of oligopolistic small businesses in the presence of a pandemic shock. The COVID-19 pandemic shock forces the adversely shocked sector to close, while the other sector remains open. The government provides bounce-back loans to the adversely shocked sector to reopen after the pandemic. Potential entrepreneurs that apply for government loans to reopen businesses have private information about their profitability. Those with lower profitability are likely to default on government loans. The government can choose to implement a default sanction with the aim of lessening adverse selection.

The default penalty which we consider is modelled as a monetary deduction from the defaulter’s residual income, and the defaulter is still allowed to continue her business should she wish. We interpret this default sanction as the government requiring the borrowers to provide a certain level of personal guarantee. The exact choice of default sanction is (politically) delicate and difficult.

If the government gives great weight to future allocation concerns, we can show that the government would choose to implement a relatively harsh default sanction to deter unprofitable potential entrepreneurs from applying for loans and reopening businesses, but this pro-allocation policy leads to persistent unemployment and demand shortage. If the government is primarily concerned with stabilising the economy, we show that it would choose to implement a lenient default sanction, or even no sanctions (i.e. 100% guarantees), to restore full employment as soon as the pandemic has passed. In this case, the demand shortage is short-lived, but the economy is shifted to a lower long-run equilibrium path due to misallocation of resources. We establish which are the key parameters that will determine whether the government would want to be lenient, or tough, in trying to screen out the potentially less successful borrowers. These parameters include:

a) the degree to which the government’s monitoring/verification scheme works well;

b) the extent of market power of the entrepreneurs;

c) the project return or tail of very low-profitability potential borrowers; and

d) the discount rate to future outcomes being applied

The higher are (a), (c) and (d) and the lower is (b), the more the government would give a higher weight to current stabilisation, and the less – even none at all – to the efficient allocation objective.

We provide a numerical calibration to assess the social optimum, and show how the optimal default sanction varies as these parameters change. First, we show that the optimal default sanction is generally intermediate between the extreme versions of leniency and harshness.

Then we vary the discount rate and show how the optimal default sanction changes accordingly. What we conclude is that society would have to be extremely myopic, with a very high rate of time preference, to eschew entirely the option of screening out borrowers with poor profitability prospects. In a crisis, governments tend to be myopic. Act in haste, repent at leisure. But extreme pressure does cause extreme myopia.


Bounds, A and D Thomas (2020), “UK’s emergency loans for small companies likely to bring rash of defaults, say bankers”, Financial Times, 10 May.

Elliott, D (2020), “Top 5 Concerns about Policy in the New Era”, Oliver Wyman.

Evans, P (2020), “Coronavirus drags factories and services to record slump, British Chambers of Commerce to report”, Sunday Times, 28 June.

Foroohar, R (2020), “Small business: a canary in the US economic coal mine”, Financial Times, 28 June.

Goodhart, C A, D P Tsomocos and X Wang (2020), “Support for Small Businesses amid COVID-19”, CEPR Discussion Paper 15055. 

Morris, S and O Walker (2020), “Banks across Europe braced for further heavy-loss charges”, Financial Times, 26 July.

Morris, S, G Parker and D Thomas (2020), “UK banks warn 40%-50% of ‘bounce back’ borrowers will default”, Financial Times, 31 May.

Thomas, D and G Parker (2020), “UK bailout schemes could create coronavirus debt trap, warn banks”, Financial Timess, 25 May.

Thomas, D, S Morris and G Parker (2020), “Treasury and banks in talks to tackle coming wave of bad Covid debt”, Financial Timess, 26 July.

Treanor, J (2020), “RBS boss Sir Howard Davies calls for toxic coronavirus loans fund-Chairman warns of mass defaults by stricken small firms”, The Sunday Times, 3 May. 

2,100 Reads