Forecasting Stock Returns

Abstract

One of the key tasks in seeking to generate attractive returns is producing realistic and reasonable return expectations and forecasts. In the Markowitz mean-variance framework, an investor's objective is to choose a portfolio of securities that has the largest expected return for a given level of risk (as measured by the portfolio's variance). In the case of common stock, by return (or expected return) of a stock, we mean the change (or expected change) in the stock price over the period, plus any dividends paid, divided by the starting price. Of course, since we do not know the true values of the securities’ expected returns and covariances, these must be estimated or forecasted. Equity portfolio managers have used various statistical models for forecasting returns and risk. These models, referred to as predictive return models, make conditional forecasts of expected returns using the current information set. Predictive return models include regressive models, linear autoregressive models, dynamic factor models, and hidden-variable models. Keywords: Theorie de la Speculation; The Theory of Speculation; information set; predictability; multivariate random walks; martingales; martingale; predictive return models; Regressive model; Linear autoregressive model; Dynamic factor model; Hidden-variable model; autoregressive model; approximate; limited