There are not many strategies that have gone unscathed by the global financial crisis, indeed few people, places or things have gone unchallenged by this crisis and its effects. But one strategy that pre-dates the global financial crisis has proved its robustness by surviving, so far. Minimum variance (or managed volatility) strategies target volatility or risk rather than return.

One of the challenges in building a strategy that concentrates on return is that everyone else is trying to do the same thing. It is easier to build portfolios with consistently lower risk than the rest of the market than to build portfolios with consistently higher returns. Academic research has put this concept to the test and found that minimum variance strategies exhibit less volatility than traditional capitalisationweighted indices. Surprisingly, they have also provided returns that compete, and sometimes even outperform, more traditional approaches. Institutional investors around the world have demonstrated a growing interest in these strategies as reflected by the launch of the MSCI World Minimum Volatility Index in April 2007.

At State Street Global Advisors (SSGA) we first started investigating these strategies in 2006, and we have since developed investment strategies for both global and Australian equities. The historic simulations for minimum variance strategies have been sufficiently different from capitalisationbased indices to suggest that the target allocation to equities in a diversified portfolio could change. In fact, a paper by Thomas Poullaouec, SSGA vice president of asset allocation and currency, looked at the impact of including a minimum variance strategy on the risk of a diversified portfolio. Poullaouec’s paper compared the returns of the MSCI World Minimum Volatility Index and the MSCI World Index from June 1988 to May 2008, and found the Minimum Volatility Index outperformed its counterpart by 0.3 per cent per annum over that period and the variability of returns was significantly lower (11.6 per cent per annum versus 15.6 per cent per annum).

In fact, since the paper was prepared, the MSCI World Minimum Volatility Index has continued to outperform the MSCI World Index. Over the year to end 31 January 2008, the MSCI World Minimum Volatility Index fell by 30.4 per cent in USD terms, compared to 42.8 per cent for the MSCI World Index . At first glance this would suggest that minimum volatility portfolios should be used as a partial replacement for capitalisation-weighted equity portfolios. However, the paper also found that the correlation between the MSCI World Minimum Volatility Index and global sovereign bonds was higher than the corresponding correlation between the MSCI World Index and global sovereign bonds. This suggests the optimal allocation to minimum volatility strategies may involve a reduction in bond holdings as well as in capitalisation weighted equity holdings.

Poullaouec concluded that by looking at historical data, minimum volatility strategies have achieved lower downside risks than capitalisation indices over the past 20 years, and that this makes a compelling case for investors to consider a minimum volatility allocation in their portfolios as a potential substitute for equities and bonds.Too good to be tr ue?So what’s the catch? A too-good-tobe- true feeling about the strategy comes from the conventional belief that higher risk is inevitably linked with higher expected returns. In academic circles this is most famously expressed in the capital asset pricing model (CAPM). Under CAPM the market portfolio represents the best compromise between the risk and return available to investors; it is meanvariance efficient. CAPM predicts that expected returns should increase as beta, or risk, increases and that no other variable has this ability to influence expected returns.

While a market capitalisation weighted equity index may not be quite the “market portfolio” envisaged by CAPM, it is common practice in the industry to assume that it is close enough. The evidence, however, suggests that beta, or risk, is not associated with higher returns in equity portfolios, and that there are indeed factors that can better predict very long term returns in equity markets. Price-to-book value is probably the most famous of these other factors.Some conseq uencesThis is much more than just an academic debate about a choice of indices for passive equity mandates. Many active equity mandates have risk controls that are based on market capitalisation indices; ranges for stock weights, sector weights or country weights are often tied to the relevant market capitalisation index.

But if there were broad portfolios that provided better combinations of risk and return than market capitalisation weighted indices, and if these portfolios could be determined in advance, you might want to limit or avoid market capitalisation-based risk controls. The strategies we have constructed certainly contain risk controls; however those controls are not focussed on market capitalisation metrics. Our research suggests that loosening market capitalisation- based constraints increases the value-add from our quantitative stock selection models. Evaluating performance is another challenge for portfolios constructed using minimum volatility principles. For specialist equity mandates, it is common practice to use tracking error as the key risk metric. Tracking error is a measure of how synchronised a portfolio’s returns are with the market capitalisation index. In the minimum volatility world, where the focus is on absolute risk, not relative risk, a high tracking error may be an indication of a very low risk portfolio. And for an investor seeking to match liabilities, it may be low absolute risk that counts, not low risk against a market index. Such portfolios created to maximise total return while controlling for total volatility, are likely to gain greater acceptance over time, especially as institutional investors acknowledge that their true benchmark is their own liability stream, and not a cap-weighted benchmark.

Leave a comment