VoxEU Column Financial Regulation and Banking Global crisis Productivity and Innovation

The cleansing effect of banking crises

Recessions are periods of low opportunity costs for time and resources, and hence can facilitate a productivity-enhancing reallocation of resources and improve productivity growth. However, recessions can also slow productivity growth by intensifying credit frictions, for instance, through the accumulation of legacy assets in the banking sector. This column investigates the interaction between these two channels in the recent banking crisis and shows that US regions with more restructuring of inefficient banks during the post-Global Crisis recession experienced higher productivity growth in the real sector in subsequent years.

It is known that inefficient banks exert a strain on growth by misallocating capital in the real sector (Peek and Rosengren 2005, Caballero et al. 2008). Sunk costs, soft budget constraints, and gambling for resurrection can motivate bank equity holders to postpone the realisation of losses and continue lending to non-performing borrowers. Therefore, in good times and in markets characterised by asymmetric information, there will exist a set of inefficient banks and firms that ensure each other’s survival. In a recent paper (Gropp et al. 2020), we examine whether banking crises under some circumstances give rise to an opportunity to eliminate these persistent, inefficient lending relationships and facilitate redistribution of funds to borrowers with a higher marginal product of capital. In particular, we examine the effect of tough versus lenient supervisory intervention during the crisis and its subsequent effects on productivity.

For banks close to the minimum capital requirement, loan loss provisioning is costly in good times since their capital held may fall below the minimum. Hence, these banks have an incentive to reduce their restructuring activity concerning their nonperforming assets. This process in turn can result in evergreening of unproductive firms. These unproductive firms, rather than exiting, stay in the market and can reduce productivity growth. In addition, competition can be distorted: given that loans to such firms are essentially a subsidy to an inefficient firm, new, more efficient firms have a more difficult time entering the market or increasing their market share. This channel further reduces productivity.

The occurrence of a financial crisis can cleanse the market of such inefficient banks and firms. Marginal banks fall below capital requirements and supervisors intervene, closing the bank, restructuring it, or placing their bad assets in a bad bank. This choice implies that funds are now redistributed to better-performing existing or new firms, which in addition no longer have subsidised competitors and hence can grow more quickly. We test these hypotheses, using data from the US.

Our identification relies on regional differences in supervisory forbearance in the US. There are reasons why a supervisor might choose to be forbearing on a distressed bank instead of closing and restructuring its assets. Forbearance implies that legacy assets would remain on the balance sheet, and the bank itself continues to have incentives to avoid realising losses. Hence, its low-quality borrower firms continue to be funded, and presumably none of the positive effects of them disappearing from the market will be realised. Hence, we posit that the long-term effects of the financial crisis should, ceteris paribus, depend on the degree of forbearance by the supervisor.

We create a cross-sectional sample of Metropolitan Statistical Areas (MSAs) using averages from the two periods from 2007 to 2010 and from 2011 to 2015 to test our ideas. Our findings show that, during the crisis, MSAs with more supervisory forbearance on distressed banks experience lower exits at the establishment and firm levels and, similarly, fewer job losses in their real sector. One standard deviation higher supervisory forbearance during the crisis would lead to approximately a 2.9 percentage point lower rate of establishment exits and job destructions. We further show that higher forbearance on banks results in a weaker banking sector in the second period (2011-2015), reflected in relatively lower bank capital and higher non-performing assets (as shown in Figure 1). Consequently, post-crisis new job creation and job reallocation, as well as wages, employment, patents, and output growth, are lower for MSAs in which supervisory forbearance was higher during the crisis. One standard deviation higher forbearance during the crisis leads to a 3.6 percentage point lower post-crisis rate of establishment entry and job creation. Overall, our results show that, for every job lost due to lower supervisory forbearance during the crisis, there will be 1.05 new jobs created after the crisis.

Figure 1 Non-performing assets and supervisory forbearance

Notes: The figure plots the average MSA-level banks’ non-performing assets across quartiles of average supervisory forbearance during the crisis period.

Non-linear effects of bank restructuring

While bank restructuring, on average, improves the efficiency of financial intermediation, it is natural to believe that too much restructuring in the banking sector can have detrimental long-term effects on economic growth by impairing the process of financial intermediation. It is important to know whether there is an optimal level of restructuring, below which there are marginal gains from more restructuring and above which there are marginal losses. Consistent with this view, we find an inverse-U shape for the marginal effect of bank restructuring on real outcomes in the longer run. As shown in Figure 2, in the lowest tercile, the effect of restructuring on all measures of post-crisis productivity growth is close to zero. The estimates for the middle tercile show positive marginal effects on post-crisis productivity growth. Finally, moving beyond the mid-range of restructuring reduces the positive outcomes through the negative marginal effect estimated for the top tercile. This outcome can be interpreted as a negative consequence of too much restructuring. This finding complements our main results by showing that, although on average there are gains to make from bank restructuring, too much of it can also be detrimental.

Figure 2 Non-linear effects of bank restructuring

Notes: The figure presents marginal effects of bank restructuring on the outcome variables, in a piecewise linear framework. The models are estimated across three linear splines of the instrumented bank restructuring with knots at the terciles of its distribution, creating the three intervals denoted by ‘Low’, ‘Medium’, and ‘High’ on the horizontal axis, respectively. The coefficient estimates represent marginal effects, i.e. the change in the slope from the preceding interval.

Bank recapitalisation and restructuring

Bank bailouts are widespread in times of high stress when regulators identify distressed banks and recapitalise them under certain restrictive conditions with the aim of releasing these restrictions and exiting as soon as the bank returns to a healthy state. Berger et al. (2020) conceptualised this process of ‘catch, restrict, and release’ and provided empirical evidence of its ramifications. Whether recapitalisation of distressed banks incentivises them to realise losses and to cut funding to their unprofitable borrowers is still debated in the literature (Giannetti and Simonov 2014, Homar and Van Wijnbergen 2017, and Acharya et al. 2019).

We use the information on recipients and disbursements of the US government’s bank recapitalisation programme, known as the Capital Purchase Program (CPP), to test whether and how recapitalisation of distressed banks affects post-crisis productivity growth. CPP is the main component of TARP and consists of a range of bank-preferred stock and equity warrant purchase programmes aimed at stabilising the financial system starting in October 2008 (Berger and Roman 2017). Our results show that regions that received more CPP money experienced fewer establishment exits and less job destruction during the crisis. This finding is in sharp contrast to the idea that recapitalisation of distressed banks helps them to realise losses and consequently favours the process of restructuring in the real sector. An alternative interpretation of the CPP in our setting is to view it as one form of supervisory forbearance, rather than a solution to it. It certainly saved some distressed banks (supervisory forbearance), but it did not lead to productivity-enhancing cleansing forces in the real sector. Consistent with this view, we in fact find that post-crisis productivity is negatively affected by exposure to CPP funds.


Our findings show that restructuring of distressed banks during a crisis has positive long-term effects on productivity. We emphasise the importance of long-term productivity considerations in the design of optimal bank resolution mechanisms. Our results indicate that the challenge is the inherent trade-off between the short- and the long-term effects, which can complicate the political economy of the problem. For instance, in the short term, bailouts can look appealing to government officials, especially if the long-term costs bear less weight in their decision-making processes.


Acharya, V, T Eisert, C Eufinger and C Hirsch (2019), “Whatever it takes: The real effects of unconventional monetary policy”, Review of Financial Studies, 32(9): 3366-3411.

Berger, A N and R A Roman (2017), “Did Saving Wall Street Really Save Main Street? The Real Effects of TARP on Local Economic Conditions”, Journal of Financial and Quantitative Analysis, 52(5): 1827-1867.

Berger, A, S Nistor, S Ongena, and S Tsyplakov (2020), “Catch, Restrict, Release: The Real Story of Bank Bailouts”, Swiss Finance Institute Research Paper No. 20-45.

Caballero, R J, T Hoshi, and A K Kashyap (2008), “Zombie Lending and Depressed Restructuring in Japan”, American Economic Review 98(5): 1943–77.

Giannetti, M and A Simonov (2013), “On the Real Effects of Bank Bailouts: Micro-evidence from Japan”, American Economic Journal: Macroeconomics 5(1): 135-67.

Gropp, R, S Ongena, J Rocholl, and V Saadi (2020), “The Cleansing Effect of Banking Crises”, CEPR Discussion Paper No. 15025.

Homar, T and S van Wijnbergen (2017), “Bank Recapitalization and Economic Recovery after Financial Crises”, Journal of Financial Intermediation 32: 16-28.

Peek, J and E S Rosengren (2005), “Unnatural Selection: Perverse Incentives and the Misallocation of Credit in Japan.” American Economic Review 95(4): 1144–66.

2,835 Reads