The article analyzes how to predict future stock returns when there is uncertainty about which model to use. By studying data from the euro area, the US, and the UK, the researchers found that using a Bayesian Model Averaging approach can improve stock return forecasts. In the euro area, various economic and financial factors consistently affect risk premiums, while in the US, only a few predictors are important. In the UK, financial variables are better at forecasting stock returns. The BMA framework is most effective for longer time periods and performs better than other forecasting models.