o provide a thorough overview of my topic, I have divided this presentation into three parts: First, I will present the rationale for short-extension strategies, which go by many other names, including 120/20 and 130/30 strategies. Second, I will discuss the optimum ratio for short extension to have in a portfolio. Although no absolute optimal ratio exists, I can describe what investors should look for. Third, I will discuss the practical implications of these strategies-what has worked, what has not worked, and what to watch out for.
Rationale for Short-Extension StrategiesWhen my colleagues and I first considered shortextension strategies about five years ago, we conceptualized a relative performance triangle, shown in Figure 1, that describes the determinants of an active manager's portfolio performance. Starting at the top, active managers begin by forecasting security returns. Whether dealing with stocks, bonds, or real estate, active managers begin by forecasting alphas. They then use those alphas to form active weights, relative to their benchmark or-in a market-neutral strategy-an absolute level. Such active weights, in turn, determine the relative return of the portfolio, which arises from the realized security returns. To measure actual performance, managers monitor signal quality or the information coefficient, which is basically the correlation between the forecasts and the actual outcomes. Ronald Kahn and Richard Grinold have done excellent work on the relationship between the information coefficient (IC) and the information ratio (IR).1 When we considered this relationship about six years ago, we realized there was a gap between signal strength and the way in which it translated into active weights. This is what we call the "transfer coefficient" (TC). Under full covariance matrix parameter definitions, exact equations for the expected active portfolio return can be derived. For example, the following equation shows how the expected active return on a portfolio, E(R A ), is achieved:( 1) where N = number of securities A = active risk target The information coefficient measures the quality of the manager's forecasting, and the number of securities is a crude indicator of the breadth of the portfolio. Higher returns will be achieved as the following increase: transfer coefficient, information coefficient, and risk for a given level of skill. The formula is highly intuitive and provides a nice linear logic. The key insight is that the transfer coefficient plays a significant role.Short-extension strategies can increase a portfolio's information ratio and thus its efficiency and can be used with a range of security types and strategies. Because no single ratio is ideal for short extensions, managers should adjust their ratios according to several variables-including tracking error, costs of shorting, and security risk. Therefore, managers using these strategies should target overall portfolio risk, not the shortextension ratio.