Working Paper

Conditional Asset Pricing Models with Instruments Closer to Agentís Information Sets

Co-author: Eric Ghysels. My Job Market Paper. Asset pricing models are often estimated via GMM with a set of instruments that are ad hoc. Even the optimal instruments, following Hansen (1985), Chamberlain (1987) and Nagel and Singleton (2011), only use a very small subset of financial market data. In this paper we introduce instruments that are more closely related to the day-to-day available financial market signals observed by agents in the economy. We call these MIDAS instruments as we borrow insights from MIDAS regressions, following Ghysels, Santa-Clara, and Valkanov (2005) and Ghysels, Santa-Clara, and Valkanov (2006). We show that MIDAS instruments largely increase the efficiency of estimation and evaluation of conditional asset pricing models.