Conditional Asset Pricing in Emerging Stock Markets

Stock returns in emerging markets are to some extent predictable on the basis of selected instrument variables. We show that local information is more important than global information to capture emerging stock market returns. This is an indication for at least partial segmentation of emerging stock markets. Our empirical results further demonstrate that predictability can be explained by time-variation in economic risk premiums. Instead of testing a traditional beta pricing model, we test a fully conditional asset pricing model in a stochastic discount factor framework. Scaling the vector of returns incorporates conditioning information, and scaling the economic risk factors captures time-variation in risk premiums. This technique allows testing some conditional implications of stochastic discount factor models, estimating the fixed weights of scaled factors as if the model was unconditional.