Bad Times, Good Credit

by Bo Becker, Stockholm School of Economics and the Swedish House of Finance, Marieke Bos, Swedish House of Finance, and Kasper Roszbach, Norges Bank.

The world is currently facing what for many countries is likely to be the biggest economic downturn in nearly a century. Many governments have responded to the downturn by introducing policies to help both firms and employees. Some measures support households, others cover variable or fixed costs of firms or inject capital into businesses. One of the most broadly applied policies involves liquidity support through partially state-guaranteed loans intermediated by banks, such as the loan guarantee scheme under which the Norwegian state guarantees 90 per cent of new bank loans to businesses.[1]

When providing loans, particularly in a period of scarce public resources, an important question is whether banks, and governments, are able to select the firms with good long-term economic prospects in the midst of an economic downturn and avoid supporting businesses that would not have had a viable future even without a Corona crisis. If banks have little or a reduced ability to select “winners” from “losers”, then government-funded liquidity aid could prove a very costly way of supporting the economy.

A recent study we conducted suggests that banks are in fact better able to identify which firms are safer or riskier during bad times than they are during good times.[2]

Earlier research has shown that banks’ limited knowledge about borrowers’ creditworthiness constitutes an important friction for an efficient allocation of credit.

Credit is the main form of financing of most firms but the flow of credit is highly cyclical: in recessions, the volume of new credit is usually low and loan spreads are high. Economists and policymakers have long been concerned that the scarcity of credit in recessions reflects a drop in the supply of credit by financial institutions – not a reduction in the demand by firms – that exacerbates the intensity of the business cycle. Current crisis policies to encourage banks to lend without having to bear full responsibility for the risks involved can be seen as a – crude but easy to implement – way to counteract banks’ tendency to contract credit supply when credit is actually needed. 

In our research we analyzed if the accuracy of the ratings that bank staff assign to their borrowers declines in recessions (procyclicality) or whether bad economic times enhanced banks’ ability to separate the wheat from the chaff (countercyclicality).

Research had so far proposed a range of explanations for why either of these two effects could occur in practice. Some economists have argued that information quality is countercyclical because banks in a recession exert more effort to monitor and inspect their customers (Ruckes, 2004) and have fewer new clients – about whom they know less and that are thus more difficult to classify (Dell’Ariccia and Marquez, 2006). An alternative explanation brought forward is that banks see the skills of their staff deteriorate in economic upturns because fewer firms fail in better economic times (Berger and Udell 2004). Other economists have argued we should expect procyclical information quality, for example because a reduction in investment opportunities during a downturn reduces the quality of information.

By analyzing the accuracy of internally produced risk ratings of corporate customers and the customer monitoring activity by loan officers in a large Nordic bank over time, we find that the bank’s ability to classify borrowers by credit quality is greater during bad times and worse during good times. The bank thus has a greater ability to predict future loan defaults in economic downturns because its internal ratings – quality scores made by the bank’s loan officers – enable a better sorting of firms according to their default risk during bad economic times than during good times.

Our analysis also enables us to exclude a number of possible explanations for this countercyclicality of information frictions. We find, for example, that shifts in the mix of new and old borrowers over time cannot explain our findings, nor do we find evidence of increased monitoring activity by loan officers in recessions. By analyzing alternative credit ratings of the same firms that are made by an external agency and do not rely on human information collection, we can also reject that changes in human skills are responsible for the countercyclical pattern.

As an auxiliary finding, we establish that soft information, which often is proprietary to a lender, is a more powerful predictor of defaults than hard information – that is typically captured by annual reports and other numeric or registered data –  during bad times. Although soft and hard information display the same countercyclical patterns, this does suggest that the collection and processing of softer information, such as business plans and information collected during business visits can greatly improve the accuracy of a bank’s assessments of firms in bad times.

Our study does not investigate to what extent the countercyclicality of information quality has affected bank lending in the past. Our findings do imply that banks during the current downturn should be well positioned to separate the better firms from the worse, and to efficiently allocate (more) credit during bad economic times, such as the current crisis.

Bankplassen er en fagblogg av ansatte i Norges Bank. Synspunktene som uttrykkes her representerer forfatternes syn og kan ikke nødvendigvis tillegges Norges Bank. Har du spørsmål eller innspill, kontakt oss gjerne på bankplassen@norges-bank.no.


[1] See https://www.giek.no/lanegarantiordningen/ for details about “Lånegarantiordningen”. 

[2] Bo Becker, Marieke Bos and Kasper Roszbach, Bad Times, Good Credit, forthcoming in Journal of Money. Credit and Banking (2020). Becker is at the Stockholm School of Economics and the Swedish House of Finance, Bos is at the Swedish House of Finance and Roszbach is at Norges Bank and the University of Groningen.

Leave a Reply