Attributing performance to ESG factors

Our attribution methodology allows an investor to determine the impact of ESG integration on their fund’s returns, and to assess the importance of ESG factors on performance through a sensitivity analysis.

Our U.S. Large Cap Sustainable Alpha fund is based on the premise that ESG characteristics are systematically mispriced in the market and that excess returns can be earned by combining ESG with other fundamental data. To assess the impact of ESG factors on investment returns, we run a performance attribution analysis including the fund’s monthly returns and the monthly returns of its benchmark Russell 1,000 from January 2010 to June 2015. In addition, we retrieve the US risk-free rate and the Russell 1,000’s size factor and value factor from the Kenneth French database. These data sets allow us to do performance attribution using the Fama-French model, which can explain a portfolio’s return by: the market’s excess return, the size factor (the excess returns of small-cap over large-cap) and the value factor (the excess returns of value stocks over growth stocks).

To make it an ESG attribution analysis, we also retrieve ESG ratings of Russell 1,000 firms during the sample period. In line with the 30% cut-off that Fama and French use to build their size and value factors, we build default factors of the returns delivered by the (i) top 30% environmentally rated, (ii) top 30% socially rated and (iii) top 30% corporate governance rated firms. Finally, we ensure that all our factors are uncorrelated to the market benchmark, which is by far the largest driver of a long only equity fund.

In the case of our Sustainable Alpha fund, market benchmark swings are responsible for 92.0% of the return variation, as would be expected from an active strategy benchmarked to a market index. Figure 1 shows the breakdown of the remaining 8%.

The social, environmental and corporate governance factors explain 1.6%, 2.4% and 2.7% of positive excess returns respectively. While our fund’s returns are not explainable by the value factor, 0.9% of its returns can be explained by a size factor.

When performing a sensitivity analysis, we measured the actual response of our fund’s return to one unit change in factor return. We do this in a similar manner to having beta represent the response of our fund’s return to one unit in market benchmark return change, which in our case is 1.1, or 110% (figure 2). When the size and value factor returns were increased by 1%, the increase in the fund’s return was in absolute terms negligible at less than 0.10%. When the environmental, social and governance factor returns were increased by 1% independently, the fund’s returns increased by 0.47%, 0.44% and 0.52%, respectively.