Building
Efficient Portfolios: Managing Risk and Return Dynamics using Factors Building
Efficient Portfolios: Managing Risk and Return Dynamics using Factors The Capital Asset Pricing Model predicts expected return for any risky asset to equal the sum of the risk free rate plus beta times the equity premium (excess of general market return over the risk free rate). This relationship is described by the Security Market Line (SML). In this asset pricing model beta is the only firm specific term and it captures the responsiveness of a stock's return with the market as a whole. As a result, the SML is predicted to provide a complete description of how returns differ among stocks.
Empirical tests designed to examine the relationship between beta and returns have created a lot of controversy. For example, Fama and French, when examining the actual behavior of the SML from 1963 to 1990 using NYSE, AMEX and NASDAQ stocks, observed a flat to negative relation between returns and beta. A strictly positive relationship was predicted by CAPM. Results of this nature have fueled an active debate between the "is beta dead" and "is beta alive" camps.
Some conclusions that can be drawn from this debate are the importance of letting beta vary with time and the inclusion of additional factors to explain observed return behavior. In this lesson we explore these issues through application. We first work through the problem of building efficient portfolios in the presence of risk and return dynamics. A factor approach is adopted to reduce the dynamics of risk and return behavior by controlling the drivers of the dynamics. In particular, we work with a set of factors that provide some control over shifts in the yield curve, shifts in the equity risk premium plus the Fama and French Factors that are designed to control for the differential effects of firm size and value versus growth firms. We then conduct backtesting experiments to contrast the traditional Markowitz approach to a generalized factor approach.
The sample of firms used in this lesson is a subset of the current S&P100 index with data available from 1970 to 2001. This permits long term return behavior to be considered, in a setting where potential survivorship biases will affect both tests simultaneously.
Lesson Plan