04 March 2021

BIS: How much stress could Covid put on corporate credit? Evidence using sectoral data

Despite substantial losses in the sectors most affected by the pandemic, total corporate credit loss rates (ie losses in relation to the stock of debt) could fall short of those during the GFC because these sectors account for a smaller share of corporate borrowing than at that time.

Introduction

The Covid pandemic triggered the largest economic downturn since the Great Depression. Although the macroeconomic outlook is currently more favourable than it was at the peak of the crisis in the spring of 2020, the second wave of the pandemic is placing an additional strain on the recovery and reinforcing existing vulnerabilities, at least in some of the major advanced economies.

Non-financial corporate (NFC) bankruptcy rates remain fairly low in most countries, despite the sharp decline in economic activity (Banerjee, Cornelli and Zakrajšek (2020), IMF (2020a)). However, they are expected to rise as measures to support credit are wound back, new consumption habits and business practices accelerate the downsizing of specific sectors, and some firms run out of liquidity buffers (eg Banerjee, Illes, Kharroubi and Serena (2020)). The looming increase in corporate bankruptcies will generate credit losses that will need to be absorbed, either by the financial system or by taxpayers.

This article assesses potential corporate credit losses at the sectoral level for the G7 countries, China and Australia. On average, corporate credit accounts for slightly more than half of total private non-financial credit in these countries (ranging from 31% of total credit in Australia to 73% in China) and typically incurs larger credit losses during recessions than household credit.¹  As such, the outlook for corporate credit has a significant bearing on overall assessments of the health of the financial system. We project credit losses, defined as recognised impairments on bank and non-bank debt, until the end of 2022, assuming that the pandemic will have played out by then and its impact on credit losses will have materialised.²

We proceed in three steps. First, we construct sectoral economic projections for each of the nine economies in our sample following the approach in Rees (2020) (Box A).³ Specifically, we use a macroeconomic model with a rich industry structure in both demand and production to estimate the economic disturbances ("exogenous shocks") that explain the path of activity since the start of the pandemic.⁴ Conditional on assumptions of how these disturbances play out, we then use the model to project the evolution of sectoral output for each country up to 2023.

In the second step, we combine data on bonds and bank loans to derive corporate debt by sector for each of the G7 countries, China and Australia (Box B). The construction of sectoral corporate debt fills a data gap in the public domain.

In the third step, we draw on existing estimates from the literature on the GDP sensitivity of credit loss rates (ie losses in relation to the stock of corporate debt) for banks (Hardy and Schmieder (2013); see also Box C) to translate our sectoral output projections into projected credit loss rates. In doing so, we assume that the historical sensitivity of bank credit loss rates to aggregate GDP is the same across sectors as well as across bonds and bank loans. We then scale these sectoral credit loss rates using our estimates of sectoral debt to project total credit losses by sector and country.

We reach three key conclusions. First, corporate credit loss rates could rise substantially in sectors most affected by the pandemic. The sectoral dispersion in credit loss rates is likely to be wider than during the Great Financial Crisis (GFC) of 2007–09 because of unevenness in sectoral economic conditions as well as the tendency for credit losses to rise more than proportionally with output shortfalls. Second, aggregate corporate credit loss rates are likely to fall short of those sustained during the GFC, in large part because the sectors most affected by the Covid pandemic account for a comparably small share of total credit. Third, projected credit losses based on sectoral growth paths are larger than those based on aggregate GDP data alone. This highlights the importance of taking account of sectoral differences in economic conditions and credit exposures when estimating the implications of an uneven recession for corporate credit losses.....

much more at BIS

© BIS - Bank for International Settlements