The chart below shows the lead up to the well-known bankruptcy of Kodak from the perspective of two quantitative credit risk models (PD Model Market Signals and PD Model Fundamentals). Looking at this public company data, one can gain a good perspective on the rapidly rising risk levels of that company prior to the bankruptcy filing.
In Part 1 of this blog series we introduced the concept of the Credit Risk Iceberg. The Credit Risk Iceberg breaks down counterparty exposures into three levels based on how difficult it is to garner a full perspective on the exposures. The three levels are essentially public rated, public unrated and private unrated companies. In this post, we dig into public company data and credit analytics for understanding public companies.
For the Iceberg, the first several meters below the surface are still in relatively visible waters. Here you can still see the shape and extent of the iceberg if you have the vantage and time to look. In our analogy, this is the universe of public companies. Public companies are generally required to file financials regularly by their country’s regulatory body. In the case of the United States, for example, the Securities and Exchange Commission (SEC) requires all public companies to file financials quarterly. In other jurisdictions, public company filing regulations may differ, but they are still required to file at least once a year or more. Equity price and volume levels are also available for public companies on a daily basis from exchanges globally. Additional information can be obtained from news, transcripts, research reports and more. These data points are available for almost all public companies and can be found easily on the comprehensive financial information platform offered by S&P Capital IQ.
Though data points are publicly available for roughly 50,000 global public companies, challenges still exist with making full use of the data. Given that filings are found in dozens of different languages, the first challenge is translating and making sense of them all. Additionally, data must be normalized across accounting regimes so that commonly used financial terms mean the same thing for each company (E.g. does revenue include discontinued revenues or not?).
Once we have sufficient historical financials, we can run regression analysis with the financials against historical defaults to build models that assess credit levels and provide forward looking probability of default scores. Some models rely on financials for a longer term view of creditworthiness, while others rely on market pricing to provide daily updates. For public company counterparties, we can use either of these types of models along with the available data to generate quantitative evaluations of creditworthiness.
Probability of default scores produced by models that utilize primarily fundamental data tend to be less volatile. The scores are generally considered to be more “through the cycle” and hence have a longer duration. Models that produce probability of default scores using market data tend to be much more volatile. These market signal models can be built off of equity, bond, or credit default swap (CDS) data or a combination thereof. The output from these market signal models has a much shorter window of validity and is often used as an early warning signal. You can learn more about how to interpret multiple credit risk indicators here.
In part three of the Credit Risk Iceberg, we will take a look at private company data and credit analytics.