Equity Valuation and Portfolio Management

KEY POINTS

• Ordered systems are definable and predictable by relatively simple rules; random systems cannot be modeled and are inherently unpredictable; complex systems can be at least partly comprehended and modeled, but only with difficulty.
• Stock price behavior is permeated by a complex web of interrelated return effects, and it requires a complex approach to stock selection, portfolio construction, and performance evaluation to capture this complexity.
• A complex approach combines the breadth of coverage and the depth of analysis needed to maximize investment opportunity and potential reward.
• Simple methods of measuring return effects (such as quintiling or univariate, single-variable regression) are naive because they assume that prices are responding only to the single variable under consideration.
• Simultaneous analysis of all relevant variables via multivariate regression takes into account and adjusts for interrelationships between effects, giving the return to each variable separately.
• Disentangling distinguishes real effects from mere proxies and thereby distinguishes between real and spurious investment opportunities.
• Because disentangling controls for proxy effects, pure return effects are additive, each having the potential to improve portfolio performance.
• In general, disentangling enhances the predictive power of estimated returns by providing a clearer picture of the relationships between investor behavior, fundamental variables, and macroeconomic conditions.
• To ...