Enhancing forecasted total risk models

Why are ESG ratings useful in quant strategies?

To determine whether or not ESG practices are an indicator of future risk, i.e. high future price volatility, we investigated whether ESG metrics provide insights that are not captured via traditional fundamental risk models. If they do, ESG ratings can be used to identify unknown risks found in companies that are considered low risk by traditional risk models.

Figure 1 categorises the MSCI World stocks into five ‘bins’ based on total forecasted risk. Given that all quality of ESG stocks are found in broadly similar quantities in each bin, the chart suggests that ESG ratings are uncorrelated with the forecasted total risk that traditional fundamental risk models calculate.

Further supporting the idea that ESG ratings provide insights not found in traditional fundamental risk forecasts, figure 2 exams the residual volatility2 of MSCI World securities, which is the volatility that is not forecasted by the fundamental risk model. As shown, stocks ranked poorly on ESG practices have more residual volatility and therefore possess larger levels of unknown risks than better ESGranked stocks, leading us to conclude that ESG ratings can help reveal companies with high future price volatility.

How we use ESG ratings in quant strategies

We devised a systematic way to incorporate the heightened risk profile of companies rated poorly on ESG practices into portfolio construction.

We use an optimiser to build a portfolio that has the appropriate level of risk, subject to stock-specific maximum position limits that are based on the risk of each individual stock. We use multiple lenses to identify these risks and a proprietary risk-scaling process to ratchet down our stockspecific maximum position limit based on the risks that we find. This process is applied systematically across all stocks.

We integrate our ESG research findings as a lens in our riskscaling process. As a company’s ESG rating falls, our model scales down the maximum allowable position to reflect the increasing uncertainly around volatility and the heightened probability of an outlier event.

Impact on investment decisions

Offshore drilling company Seadrill Limited is considered very high risk from both a fundamental and statistical perspective, and accordingly, our maximum allowable position size would only normally be around 1.5%.

Seadrill was involved and ultimately found partially guilty for Australia’s largest marine oil spill. The Montara Oil Spill in August 2009 lasted for 74 days and spilled about 150,000 barrels of oil into Australian and Indonesian waters. Seadrill faces extremely high financial and reputational risks related to health and safety accidents and hydrocarbon spills, and has shown no evidence of diminishing those risks since 2009. Not surprisingly, Seadrill’s ESG rating is CCC. Even if we think the stock is attractively priced, the additional ESG risk has reduced the maximum allowable position by two thirds – to about 0.5% of a portfolio.

While the Seadrill example is interesting, what about a stock that fundamental and statistical models do not show to be risky?

Healthcare company Baxter International exhibits low fundamental risk but has encountered numerous product safety issues. In December 2013, lawsuits were filed due to deaths potentially related to a blood thinner, Heparin. The next month, they received a letter regarding repeated violations of manufacturing standards at a plant in Illinois and in May of last year, the company recalled SIGMA Spectrum Infusion Pump tied to severe adverse events.

Fundamental and statistical models generally see the stock as low-risk and in fact, depending on how you measure volatility, the risks are approximately 20% lower than the average stock. So while the balance sheet and income statement appear to make the company seem low-risk and warrant a position size of 3%, the weak ESG evaluations resulted in us shrinking our maximum allowable position size to 1.9%.