Dividend yields and expected stock returns: alternative procedures for inference and measurement

Alternative ways of conducting inference and measurement forlong-horizon forecasting are explored with an application todividend yields as predictors of stock returns. Monte Carloanalysis indicates that the Hansen and Hodrick (1980) procedureis biased at long horizons, but the alternatives perform better.These include an estimator derived under the null hypothesisas in Richardson and Smith (1991), a reformulation of the regressionas Jegadeesh (1990), and a vector autoregression (VAR) as inCampbell and Shiller (1988), Kandel and Stambaugh (1988), andCampbell (1991). The statistical properties of long-horizonstatistics generated from the VAR indicate interesting patternsin expected stock returns.