Tensor Factor Asset Pricing Models

Work in progress

This paper provides two additional applications of tensor factor models estimated by TPCA: 1) international asset pricing that involves time, industry and country dimensions, and 2) conditional asset pricing of equities based on (a) the cross-section of individual firm returns, (b) the time series of monthly returns and (c) data encoding characteristics across firms and time.

The international asset pricing application suggests that there is one world factor and one segmentation factor that together explain the international asset returns.

The conditional asset pricing application aims to estimate time varying loadings based on the time series behavior of a firm’s characteristics, relative to the entire universe of stocks.