Smart Beta Investing-Testing the low-volatility effect on the NSE

Smart beta strategies are at the intersection of active and passive investing, and their goal is to beat the market using a rules based, transparent, and low-cost approach. Smart beta investment strategies are designed to add value by systematically exploiting known sources of excess returns.

Among the long-standing anomalies in modern investment theory, perhaps none is as puzzling and compelling as the low-volatility effect. It challenges the traditional equilibrium asset pricing theory that an asset’s expected return is directly proportional to its beta or systematic risk, or, in other words, higher-risk securities should be rewarded with higher expected returns while lower-risk assets receive lower expected returns. Contrary to that theory, the empirical evidence of numerous academic studies has illustrated that low-volatility or low-risk investing outperforms the broad market as well as high-risk strategies over a long-term investment horizon with much less realized volatility.

Low-volatility investing is not a new concept, but the financial crisis of 2008/2009 and the high volatility that characterized the second half of 2011 brought it back to the fore in the global investment community as a useful risk management tool. The resurgence of low-volatility investing has also reignited the theoretical debate on the properties of a true market portfolio. In this piece, we analyze the low-volatility effect in the Nigerian equity market and empirically investigate whether stocks with low volatility earn higher risk-adjusted returns. This work draws from the methodology employed by Blitz, D and P. Vliet (2007).

We employ weekly data of all the listed equities on the Nigerian Stock Exchange (NSE) from 2003 to June 2014, and the 3-months treasury bills rate as a proxy for the risk-free rate. The risk-free rate is the barometer for our excess returns. The NSE-ASI is our benchmark proxy. Returns are log-transformed in order to make them additive over time. The log-transformed excess returns are used throughout our analysis for all return calculations. At the end of each quarter, we construct equally weighted quartile portfolios by ranking stocks on the past 3-year volatility of weekly returns. Stocks with the lowest volatility scores are assigned to the top quartile. Portfolios are rebalanced with a quarterly frequency and transaction costs of 1% are employed throughout our analysis.

For the four different quartiles’ portfolios, we calculate the return (in excess of the risk-free rate) over the quarter following portfolio formation. For the resulting time series of returns we calculate both the average, standard deviation and Sharpe ratio. We also test for the statistical significance of the difference between two Sharpe ratios. We employ both a regression based methodology and a double-sorting methodology in order to disentangle the volatility effect from other known effects. We also controlled for possible systematic exposures to size and value effects and capped sector weights.

Our results are quite interesting. On average, the low-volatility strategy outperformed the market benchmark in 72% of the quarters studied in our analysis when the straight returns are compared. The results show asymmetry in the performance, with the low-volatility portfolio outperforming the benchmark more frequently (93% of the time) when market returns were negative. This asymmetric response to market movements highlights the ability of low-volatility strategies to provide downside protection in uncertain times. On an average, the low-volatility portfolio outperformed the benchmark by about 9.7 percentage points during market downturns.

The results become more interesting when we shift to a risk-adjusted performance perspective instead of looking at straight returns. Ex-post standard deviations can be seen to progressively increase for the consecutive quartile of portfolios. The volatility of the top quartile portfolio is only about 45% that of the market portfolio. At the other end we have the bottom quartile portfolio, with a standard deviation which is almost twice that of the market portfolio. Combined with its low return, this results in a very low Sharpe ratio for the high risk stock portfolio. Because the other volatility quartile portfolios exhibit relatively small differences in average returns, their Sharpe ratios are driven primarily by the standard deviation in the denominator. One of our key findings is that the top quartile of low risk stocks achieves a Sharpe ratio of 0.86, compared to a Sharpe ratio of only 0.48 for the market portfolio. This difference in Sharpe ratios is statistically significant at the 1% level. The Sharpe ratios show a steadily declining pattern across the volatility sorted portfolios.

Our estimated beta and alpha from a CAPM style regression of quarterly quartile of portfolio returns on quarterly returns of the market shows that the low risk portfolio combines a very low beta of 0.42 with a positive alpha of 4.6% per annum, which is statistically significantly different from zero at the 5% significance level. The betas increase progressively for the consecutive quartile of portfolios, suggesting that volatility and beta are related risk measures. The bottom quartile portfolio consisting of the highest risk stocks exhibits an estimated beta of 1.32 and a negative alpha of 6.8% per annum. This finding implies a negative relation between risk and return. The combined alpha spread for the low risk minus high risk portfolio amounts to 11.4%. The risk/return characteristics of the volatility sorted portfolios is in clear violation of the theoretical (CAPM) security market line.

Benchmark driven equity investors may not easily benefit from the low volatility effect as they are facing a relative return objective and are either not allowed or willing to apply leverage. However, for investors interested in high Sharpe ratio investment opportunities such as pension funds, it may be much easier to benefit from the volatility effect, by applying leverage within their asset mix. These investors could include the decision to invest in low risk stocks in the strategic asset allocation process. A recommendation for absolute return investors is to distinguish between low risk, high risk and traditional stocks as separate asset classes, just like they distinguish between value versus growth stocks in their strategic asset allocation decision making.

Olugbenga A. Olufeagba