The 2008 sub-prime crisis, made the phrase "too big to fail" famous. A business that is likely to take down multiple other businesses if it collapses is defined as "too big to fail". Actually that is a somewhat inaccurate description of businesses that could cause systemic risk. It isn't the size of the business that is the problem; it is connections to other businesses. Of course, size and connectivity tend to go together.
Financial businesses like banks, NBFCs, investment banks, etc, generally have high levels of inter-connectivity. Even a relatively small financial business can cause contagion because of the high inter-connectivity.
However, non-financial businesses can also cause problems if large ecosystems revolve around them. For example, when US auto-majors like General Motors and Chrysler edged into bankrupt territory, they took down hundreds of component suppliers along with them.
After the 2008 crisis, there was a serious focus on identifying and managing risks associated with systematically sensitive firms. In India, the Financial Stability and Development Council (FSDC) has the task of monitoring such businesses. This involves identifying systemic risky businesses, then setting specific prudential norms for these firms, and also setting up bail-out systems in case things go wrong.
In October, the Indira Gandhi Institute of Development Research (IGIDR) released a working paper (WP-2013-021) with multiple authors, entitled A systematic approach to identify systemically important firms. This lays down a methodology to identify such companies.
The researchers have developed a unified Systemic Risk Index (SRI) for firms by combining three well-known statistical measures. They took the Conditional Value-At Risk (CoVAR), the Marginal Expected Shortfall (MES) and the Granger Causality (GC) measures for the 50 largest Indian businesses, quarter by quarter. One useful factor is that this data is publicly available and the research can be easily expanded or duplicated.
CoVaR and MES are measures of long-tail risk. These offer an estimate of losses in worst case scenarios. In equity markets, one way to use MES is to measure equity returns for a given listed company on the days when the overall market records its worst performances. The higher the MES, the greater the systemic risk of the given firm.
CoVar can be measured by looking at the Marketwide VaR when a given firm is distressed, versus when the firm is in an average or median state of financial health. Price-book value ratios are used for this. Again, high CoVaR implies greater chances of negative fallout if a given business collapses. The Granger Causality measure was invented by the Nobel prize winner, Clive Granger, to measure the degree and direction of inter-connection between firms and it can help predict how changes in one could cause changes in another.
The researchers ranked firms on percentile basis on each of these three measures and then averaged off ranks, to derive a single SRI rank. In addition, the values of these three measures were also normalised as GDP percentage. One interesting angle is that the research did not focus exclusively on banks and financials - it also examined other industries for potential systemic risk.
Some interesting conclusions are presented. Banks and other financials do indeed pose the highest level of systemic risk. But certain players from other sectors also show up consistently as highly systemic risky. Within the bank sector itself, SRI rank has changed over time. At the time of the 2008 crisis, ICICI Bank was rated the most likely to cause systemic risk but SBI is now reckoned more risky, for example.
Real-estate major DLF repeatedly shows up as posing high levels of systemic risk. So do highly indebted infrastructure developers like GMR Infrastructure. So do businesses like Tata Steel, Grasim, Sail.
In contrast, the IT majors generally show up at the lower end of the SRI and so does Asian Paints and some pharma companies. Consumer-driven majors like ITC, Hindustan Unilever have seen their SRI climbing in recent quarters and so has Airtel. This is a negative signal.
The methodology of the study is not only useful for policy-makers seeking to ring fence companies with high systemic risks. Given the methodology of calculation, businesses that consistently show up as low SRI are also low risk on standalone basis.
The paper is well worth studying not only for current and historic rankings, but also because the methodology and logic could be useful for a long-term investor who duplicates it. A reduction in the SRI for any given company may correlate to better stock market performance in the long run, while an increase in SRI could correlate to lower returns.
