Market Sensitivity – continuing discussion on net exposure

02 Dec Market Sensitivity – continuing discussion on net exposure

In previous posts, I demonstrated that an equity long/short portfolio with a non-zero net exposure could have less return volatility than the same portfolio with net exposure equal to 0 (i.e., dollar-neutral). While lower volatility is a valuable outcome, the resulting portfolio’s sensitivity to market returns is equally important to many investors. In fact, for an equity long/short portfolio, most prefer a portfolio return profile that has low or no sensitivity to market returns.

The post on October 29^th showed that the low volatility portfolios were more sensitive than the base portfolios to the market index. This was evident from common factor analyses using the market return from the Kenneth R. French data library (“Broad Market”). This post presents summary results from the common factor analyses using the S&P 500 return. These results show that the low volatility portfolios are not more sensitive than the base case portfolios to the market return.

Why is this important? In the balancing act that is investing, it is not self-evident which variable (beyond performance) matters more – less return volatility or less sensitivity to market returns. Intuition suggests that varying net exposure from 0 increases sensitivity to market returns. The following results show that the varying the net exposure from zero does not increase sensitivity to S&P 500 market return and produces a less volatile return series. In addition, the observations regarding the size effect and momentum are consistent with the previous analyses using the broad market.

Background on Market Indices as Benchmarks

Investors and analysts use benchmarks, usually broad market indices, to evaluate portfolio performance including risk. The benchmark reflects the universe of securities available for investment as well as the appropriate risk profile.

The net exposure study used the total universe of U.S. equity securities, capitalization weighted, as the market return because it represented the investable universe for constructing the long/short portfolios. However, given the prevalence of the S&P 500 as the U.S. equity market representative, the S&P 500 provides as a “practical” benchmark for analysis.

Less sensitive to the S&P 500 

The low volatility and base case sensitivities to S&P 500 returns are similar. This result differs from the previous results where the low volatility portfolios tended to be more sensitive to the broad market. The other observations are the same – less sensitivity to large firm returns and high momentum returns. As a reminder, the market sensitivities result from regressing the portfolio returns against the market return and the three Fama/French factors – Value/Growth, Size and Momentum. A value close to 0 implies little to no sensitivity to the factor – a change in the value of the factor implies no change in the portfolio return.

Note: R_m is the S&P 500 monthly return. R_f is the 3-month t-bill monthly rate.

Varying results for other equity indices

To confirm the results using the S&P 500, I performed the same common factor analysis using other common market benchmarks – Russell 3000, Wilshire 5000 and the MSCI US All Cap. The large increases in the market sensitivities are apparent for the MSCI US All Cap and the Broad Market.

Market Sensitivities for Earnings Yield Portfolios

Market Sensitivities for Book Value-to-Price Portfolios

Market Sensitivities for Momentum Portfolios

S&P 500 return is similar to other common equity indices

To determine if the index returns were similar, I calculated the correlations among them.  The broad market is closely related to the MSCI US All Cap index (0.999); it has little relationship to the S&P 500, Russell 3000 and the Wilshire 5000.  This is interesting because the MSCI US All Cap index represents 99% of the U.S. equity universe while the S&P 500 is believed by many to best represent the U.S equity market.

Correlation Matrix

Note: The period end date is Dec. 2014. The start dates are as follows: (1) S&P 500 – July 1951; (2) Russell 3000 – Sept. 1987; (3) Wilshire 5000 – April 2009; (4) MSCI US All Cap – Dec. 2007.

The S&P 500 is unique among the common indices in that it is not rules-based;  instead, a committee selects the companies that make up the index.  In addition, the S&P 500 is limited to large and mid-cap equities while the other indices extend to small and micro-cap equities.

Note: Large Cap is > $10 billion. Mid-cap < $10 billion and > $2 billion. Small cap < $2 billion. Micro Cap < $500 million. Conclusion

Even though the low volatility portfolios are sensitive to the broad market index, they are not sensitive the S&P 500.  Therefore, the concern about non-zero net exposures increasing sensitivity to market returns must account for the benchmark used to evaluate portfolio performance.  In this case, varying the net exposure does not increase market sensitivity and it reduces the return volatility.