Industry based equity premium forecasts

In this paper we used industry indexes to predict the equity premium in the US. We considered several types of predictive models: i) constant coefficients and constant volatility, ii) drifting coefficients and constant volatility, iii) constant coefficients and stochastic volatility and iv) drifting coefficients and stochastic volatility. The models were estimated through the particle learning algorithm, which is suitable for dealing with the problem that an investor faces in practice, given that it allows the investor to revise the parameters as new information arrives. All the models exhibit similar statistical predictive ability, but stochastic volatility models generate slightly higher utility gains.