Heightened Volatility – Is low net exposure the solution?

22 Dec Heightened Volatility – Is low net exposure the solution?

Rarely. In this post, I continue to evaluate whether a zero net equity exposure portfolio (long equity exposure – short equity exposure) has less volatility than the same portfolio with non-zero net equity exposure. The first part of this study showed that applying constant positive net exposures to select portfolio styles produced portfolios with lower realized volatilities measured from 1954 to 2014. Rebalancing a portfolio monthly to the same non-zero net exposure produced less return volatility than a portfolio rebalanced to zero net exposure.

However, this does not guarantee that a non-zero net exposure portfolio would produce the lowest volatility portfolio when measured over a shorter period of time. In this post, I use rolling 3-year periods from 1954 to 2014 to identify net exposures for the lowest volatility versions of the Earnings Yield, Book Value-to-Price, and Momentum portfolio styles. While the net exposures vary over time, few of the lowest volatility portfolios had low net exposure (greater than –20% and less than 20%).

Low net exposure is rare among minimum volatility portfolios

Using a scenario analysis where I vary the net exposure from -100% to 100%, I identify the lowest volatility portfolio for each portfolio style for each month and present the net exposure. For each month, the portfolio volatility is the standard deviation of the portfolio return over the prior 36 months including the current month.

The lowest volatility portfolios did not frequently have low net exposures. The results differed by portfolio style with the Momentum style having the highest frequency of low net exposures at 25.17%. The net exposure varies over time between positive and negative values. Below are the net exposures for the lowest volatility portfolios for each portfolio style.

There is a wide range of exposures with the average and median exposures beyond the low net exposure range.

How often an exposure occurs for the lowest volatility portfolios:

Note:  Net = ±20% excludes instances of Net = 0%.  Net = 0% exposures ranges between ±1%.

There are a few points worth noting. First, over the full period, the net exposures are rarely close to zero (i.e., between -1% and 1%). Second, the net exposures vary 10 – 20% around their averages during the period. Third, the net exposure varies depending on the portfolio style.

Consistent results in various market environments

To evaluate if and how the net exposure results vary by market environment, I analyzed the results over time by dividing the full 60 year period into sub-periods of 20 years. The first subsample is shortened due to the beginning date for the data and most recent subsample is further subdivided to include the “Tech boom and bust” (1991 – 2002), the market increase leading up to the financial crisis (2003 – 2008) and the period following the crisis (2009 – 2014). This analysis includes periods of market stress, when investors usually seek out low net exposure equity long/short strategies. The table shows the average net exposures and the Sharpe ratios for the lowest volatility portfolios (“Variable Net”) by style. In addition, I have included similar metrics for the lowest volatility portfolios with constant net exposure (“Constant Net”) and the zero net exposure portfolios (“Base Case”) from the first part of the study.

Note:  Risk-free rate used in the Sharpe Ratio is the 3-month t-bill.

The average net exposure for the lowest volatility portfolios was different from zero for all sub periods. In addition, both flavors of the lowest volatility portfolios, constant and variable net exposure, have risk-adjusted returns that exceed the base case risk-adjusted returns. The only exception is the Earnings Yield variable net portfolio, which underperformed in the “1971 – 1990” sub-period due to poor returns rather than higher risk; the variable net portfolio failed to capture the peak return of the base case and constant net exposure portfolios in 1986.

What about high volatility periods?

How do the portfolios perform during periods of above-average market volatility? I compared the risk profiles of the lowest volatility portfolios with the base case portfolios for each portfolio style. Using the S&P 500, monthly volatility is the standard deviation of the monthly return over the prior 36-months. The periods of comparison include the months when the volatility exceeds the long-term average of the monthly volatilities from 1954 to 2014.

When are the high volatility periods?

Note: Average annualized S&P 500 3-year volatility is 14.11% from June 1954 to December 2014.

How do the portfolios perform?

As expected, the lowest volatility portfolios have lower risk profiles than the base case portfolios. This is consistent across the three portfolio styles. The worst case is that the risks are approximately the same. During the August 1986 – October 1990 period, there was little to no risk reduction for all portfolio styles. Considering all portfolio styles, the Book Value-to-Price saw the smallest difference between lowest volatility portfolio risk and that of the base case. The lesson from these observations is that each portfolio style has unique characteristics that govern the appropriate net exposure, but that non-zero net exposure is no worse than zero net exposure.

Below are the annualized risks over the above-average volatility periods.

Earnings Yield Annualized Risk (in %)

^* Duration is less than 30 months

Book Value-to-Price Annualized Risk (in %)

* Duration is less than 30 months

Momentum Annualized Risk (in %)

* Duration is less than 30 months

Conclusion

Using 3-year rolling averages, I showed that the appropriate net exposures for the lowest volatility portfolios vary but rarely settle in the low net range. Based on this study of these 3 styles, low net exposure does not consistently lower portfolio volatility in the face of heightened market volatility.

The next post will continue to evaluate the portfolios’ risk profiles during the above-average volatility periods focusing on the time after market volatility exceeds its long-term average and approaches the peak volatility.