Misspecification and Weak Identification in the Nontraded Factor Zoo

with Amedeo Andriollo and Xinyi Zhang, 2024, working paper.

[Paper] To explain the cross-section of asset returns, a “zoo” of economic factors that are not
portfolios (nontradables) have been proposed. In contrast to traded factors, the nontraded
factors tend to have smaller correlations with asset returns, risk premia
inference tends to be more fragile, and the issue of weak identification might be exacerbated
by the degree of model misspecification. Yet, robust inference has often
been overlooked in many empirical works, and limited efforts have been devoted to
“domesticate” such factors. With respect to the most commonly used test asset returns,
this paper aims at providing a comprehensive re-examination of the non-traded factor
zoo, with a focus on the aforementioned fragilities. We confirm that the vast majority of the original model specifications suffer from such
problems, and robust inference leads to strong evidence that most of the proposed
nontraded factors are unpriced. The findings are even more drastic when considering
the market factor as a control. When considering the entirety of the nontradables
with PCA, we conclude however that the zoo carries some nonzero pricing ability.