During our recent webinar with the International Association of Credit Portfolio Managers (IACPM), we discussed how assessing the credit risk of small- and medium-sized enterprises (SMEs) can be one of the most challenging tasks for a creditor when there is a lack of available financial information. The need for robust methods to assess the potential credit risk deterioration and outlook of SMEs has become even more important during the uncertain period caused by the COVID-19 pandemic. Arsene Lui, Senior Analyst, Quantitative Modelling, shares his answers to some of the top questions posed by your industry peers during this session.
How do we expect statistical models that were trained outside COVID-19 time period to work during a pandemic period?
Although the COVID-19 global pandemic per se is not included in the training samples of our analytical models, we did use samples from previous economic cycles that covered crisis periods. In addition, the models include important drivers of default risk that can be experienced at any point in time, such as the profitability, liquidity, and capital structure of companies. Since obtaining the latest financial information for SMEs can be difficult, we also look at other signals that are updated on a more frequent basis. This can include the Point-in-Time (PiT) probability of default (PD) based on equity market information, and historical trade payable data, such as the change in outstanding balances or the change in payment delay, to obtain a clearer picture of what is happening.
What is the difference between credit deterioration due to COVID-19 and the previous global financial crisis?
The reasons behind the respective downturns are quite different. In 2008, risk-taking by banks and the bursting of the housing bubble impacted the liquidity of financial institutions, leading to a significant drop in credit supply to corporates. Over the past 10 years, companies in the banking industry have taken many steps to improve and solidify their balance sheets, and now the industry has lower leverage and better asset quality. Additional regulatory requirements designed to mitigate risks at banks, such as Basel III, and the stress-testing exercises done by central banks have also helped to reduce the chance of any recurring problems with financial institutions.
The shutdown of economies due to COVID-19 was responsible for this latest downturn and the deterioration of credit. Central banks have taken a proactive approach by injecting liquidity into the market through the lowering of interest rates and the use of asset purchase programs. Empirically, the median PD of corporates in the U.S. and Europe, according to our PD Model Market Signal, have greatly improved since June. Although they are not entirely getting back to the pre-COVID level, it suggests that the credit deterioration due to COVID-19 will be less long-lasting than the previous global financial crisis.
Are U.S. trends comparable to Canadian trends with regards to COVID-19’s impact on SMEs?
From the number of new cases and the number of deaths due to COVID-19, we can see that Canada has done a better job than the U.S. in controlling the spread of the virus. A rebound in the economy in Canada compared with the U.S. was seen during Q3 2020. However, the unemployment rates are still high in both countries, and SMEs continue to face pressure to increase their operating cash flow to meet their debt obligations. Most of the analytical models developed by S&P Global Market Intelligence are globally applicable. The analysis shown in the webinar can be repeated for companies in Canada (or other countries) to estimate the impact of COVID-19 on each industry, or on the economy as a whole.
How do you forecast the credit risk of corporates under potential recovery trajectories?
During the webinar, I used the PD Model Fundamental and the Macro-scenario Model to forecast the credit risk of corporates under three scenarios for Q4 2020. One scenario was prepared by S&P Global Ratings’ economists, and I prepared two additional scenarios to capture more pessimistic and optimistic alternatives. Users of our Macro-scenario Model can supply their own forecasts based on different time horizons, or borrow the stress-testing scenarios created by regulators, to assess credit scores throughout the potential recovery trajectories.
What data is included in the training of the models? What is the in-sample/out-of-sample performance of the models?
We followed a comprehensive approach in the development of each model. Information on training data, the results of the tests, and insights on model performance are all documented in our detailed validation reports. We also regularly carry out backtesting exercises on each model. If you would like to learn more about the models, please visit us online to find out more.
 “Banks nearly took down the economy in 2008. Now the industry hopes to redeem itself”, CNBC, March 17, 2020.