Measuring the impact of malfunctioning credit markets on productivity
Since the Global Crisis, there has been a renewed awareness of how frictions in credit markets can damage economic efficiency due to a higher cost of capital and/or capital being misallocated away from its most productive uses. This column presents a new methodological approach for calculating the cost of credit frictions which can be implemented with relatively simple data in multiple contexts. It finds that credit market frictions explain half of the fall in UK productivity in the Great Recession and depress output by 28% on average.
Since the 2008-09 global financial crisis, there has been a renewed awareness of how frictions in credit markets can damage economic efficiency, due to a higher cost of capital and/or capital being misallocated away from its most productive uses. The recovery from the Great Recession has been painfully slow, with lacklustre productivity growth around the world. The UK is no exception, and had one of the largest financial sectors as a share of the economy in 2007.
Figure 1 shows the level of productivity (GDP per hour worked) since 1979. Had the UK followed the three-decade trend prior to the global financial crisis of 2008/09, productivity would have been almost a quarter (24%) higher in 2020. Given this abysmal performance, it is unsurprising that average UK real wages have barely recovered to the level they were prior to the financial crisis after more than ten years.
Figure 1 GDP per hour worked in the UK
Notes: Whole economy GDP per hour worked, seasonally adjusted. ONS Statistical bulletin, Labour Productivity Q2 2019, release date 8 October 2019 (Q2 2008=100). Predicted value after Q2 2008 is the dashed line calculated assuming a historical average growth rate of 2.2%.
Many studies have argued that imperfections in financial markets play an important role in holding back investment and productivity. A recent micro literature uses bank-firm relations prior to crises to show that firms with a closer relationship with distressed banks suffered particularly badly during the crisis (e.g. Chodorow-Reich 2014 for the US; Bentolila et al. 2018 for Spain). More generally, geographical areas with more firms exposed to distressed banks suffer over and above the firm itself (Huber 2018 on Germany), suggesting important general equilibrium effects. These studies offer stronger causal credibility to the view that financial constraints matter compared to the earlier literature, which mostly did not have such compelling natural experiments (see Bond and Van Reenen 2007 for a survey). However, these micro studies cannot really address the magnitude of the losses in aggregate for the economy. To do this, an explicit equilibrium model combined with micro data is needed.
To investigate the role of credit frictions in affecting output, in a new paper (Besley et al. 2020) we develop such a model, which shows that a sufficient statistic for some credit frictions is the expected default probability. Most existing models do not have equilibrium default as a possibility even though this is important in the real world and is a crucial factor in determining the rate of interest people and firms really face – or whether they can get a loan at all. Our theoretical model of credit contracts shows that a suitably weighted default probability across all firms is key to calibrating the output loss from credit frictions.
However, a potential issue in applying the model is that the market’s expected probability of firm default is not directly observed. To address this issue, we use S&P Global’s PD algorithm, which is a piece of software algorithm typically used by lenders. This gives a credit rating which maps into the chance that the firm will not repay its loan. We use historical values of this default risk combined with the population of firm accounts to calculate a default probability for just about every UK firm since the mid-2000s. We then match this to administrative panel data on firm size and productivity.
The bottom line is that credit frictions appear to depress productivity by about 28% over our sample period – a surprisingly large amount for a financially developed country like the UK. The magnitude is greater for smaller firms than larger firms, consistent with the idea that SMEs face tougher financial constraints. The main cause of lower productivity is that investment is held back, and this lower capital intensity reduces output per worker. The contribution of misallocation: productive firms getting ‘too little’ capital only accounts for a tiny proportion of the losses.
Table 1 gives our calculations of how credit frictions depressed output over time during the Great Recession period. Column (2) has the loss in percentage per year, which averages at 28% as noted above. However, it is clear that there was a sharp deterioration in financial conditions in 2008-2009, causing a 4.8% fall in productivity which still had not been made up by the end of our sample period. The actual fall in productivity in these years was 9.3%, so this implies that credit frictions accounted for about half of the loss in productivity (4.8/9.3) in the crisis years. Obviously, this still leaves ample room for non-financial factors such as demand shortages, but it does highlight the fact that credit frictions really matter quantitatively as well as qualitatively.
Table 1 The effect of credit frictions on aggregate output (baseline results, full sample)
Source: Besley et al. (2019) Table 3
Notes: This table uses the full IDBR sample. Output loss is the proportionate fall in output as a result of credit frictions.
Our paper develops a new methodological approach for calculating the cost of credit frictions which can be implemented with relatively simple data in multiple contexts. It shows that credit frictions cause significant output losses and were particularly damaging during the Great Recession. Policies that help to improve the functioning of financial markets, especially for SMEs, should remain high on the policy agenda.
Bentolila, S, M Jansen and H Jimenez (2018) “When Credit Dries Up: Job Losses in the Great Recession”, Journal of the European Economic Association 16(3): 650-695.
Besley, T J, I Roland annd J Van Reenen (2020), “The Aggregate Consequences of Default Risk: Evidence from Firm-level Data”, CEPR Discussion Paper 14327.
Bond, S and J Van Reenen (2007), “Microeconometric Models of Investment and Employment”, Chapter 65 in E Leamer and J Heckman (eds), Handbook of Econometrics, Volume 6A.
Chodorow-Reich, G (2014) “The Employment Effects of Credit Market Disruptions: Firm-level Evidence from the 2008-09 Financial Crisis”, Quarterly Journal of Economics 129(1): 1-59.
Huber, K (2018) “Disentangling the Effects of a Banking Crisis: Evidence from German Firms and Counties”, American Economic Review 108(3): 868-98.