The search is on for alpha, or excess return, without compromising the risk/return profile.
The critical trade-off between risk and return has been well understood for over five decades now – Markowitz first demonstrated that the investor’s objective is to maximise the portfolio’s expected return, subject to an acceptable level of risk. He also noted that as stocks are added to a portfolio, the expected return and risk change in very specific ways, based on the way in which the additional securities co-vary with the securities already in the portfolio.
The source of investment ideas tends to fall into one of two categories: (i) ideas that are related to the natural biases that investors exhibit as humans, and (ii) those that are grounded in rational theories about the way in which the capital markets work. The latter serve to offer explanations for how securities are priced, relating incremental return opportunities to the additional risk that must be borne for incurring them. The former, however, augment the risk-based theories that relates higher average returns to sensitivity to a common risk factor. Instead the outperformance is linked to behavioural explanations related to themes that are firmly grounded in psychology.
Whatever the motivation, it seems fairly clear that pricing anomalies or market inefficiencies do exist and generally persist over reasonable amounts of time, long enough at any rate for careful fund managers to be able to exploit them. The key to mining this seam of investment ideas is how they are combined in a portfolio. There is little additional benefit to be gained from tilting a portfolio towards stocks with lower book value to price, if that portfolio already holds stocks that on average offer greater than benchmark exposure to cash flow to price, for example. These two characteristics of stocks, both value-related will be highly correlated, offering in combination little added return for the additional risk borne. While the implications of exposure to two such similar value measures is readily understood, the power of computers must be harnessed to understand the more complex relationships between certain factors.
Quantitative techniques can be used to provide portfolio managers with greater understanding of the way in which security characteristics or factors are inter-related. At the base level, a simple table of correlations will offer a historical perspective. Beyond that, cluster analysis is a statistical technique that can be used to group data and to establish a hierarchy of relationships.
This type of analysis gives additional insight on relationships between commonly used factors. The notion is simply that an astute fund manager would probably only chose to select one factor among those that are in its grouping lower in the hierarchy. In combination, the benefits of achieving this kind of factor diversification will show in the performance results and the risk/return trade-off enjoyed by the portfolio.
The unhappy truth is that many fund managers still shy away from employing statistical reasoning when making their investment decisions. The evidence suggests that market inefficiencies exist, but are more readily exploited by model-driven techniques than heuristics. Disciplined investors using quantitative methods or extremely astute traditional, fundamental managers will enjoy alpha without compromising the risk/return profile, generating low risk profits typically at the expense of the unwary investor.
Eoin Murray is European head of quantitative management at Northern Trust Global Investments in London