|
Asset Pricing with Many Assets
The lectures will discuss what progress we can make in asset pricing once we can exploit the availability of many assets to conduct asset pricing tests and estimate risk factors and related premia. The lectures will consist of four segments and the content will be (please click on the title to view the respective paper):
1) Conducting asset pricing tests with many assets
This segment studies how standard asset pricing methods (e.g., cross-sectional regressions, mimicking portfolio projections) behave when the number of available assets grows to infinity.
2) Estimating risk premia with omitted factors using many assets
The availability of many test assets allows us to address the potential for omitted risk factors in asset pricing models. We discuss the three-step risk premium estimator of Giglio and Xiu (2021).
3) Weak factors
Another fundamental issue in asset pricing is the presence of weak factors. We discuss how modern machine-learning techniques, and the availability of many test assets, can be combined to yield a solution to the weak factor issue.
4) Factor zoo
The many test assets we have come from a long literature searching for anomalies. While these represent a strength in terms of testing specific models, they can be a liability when trying to learn the correct model. In this segment we discuss how machine-learning methods can help discriminate useful vs. redundant and useless assets.
|
