Skip to main content

From Papers to Portfolios

Recent empirical asset pricing research, tested with code, data and realistic constraints.

Two panels sharing a logarithmic axis for the ridge penalty z, with the ridgeless estimate at z = 0 on the left. In the top panel the in-sample Sharpe ratio falls from 7.8 for the ridgeless estimate to 2.3 at z = 10. In the bottom panel the out-of-sample Sharpe ratio rises from 1.05 for the ridgeless estimate to a peak of 1.97 at z = 0.32 and falls to 1.4 at z = 10.

Sharpe ratio of the stochastic discount factor of 153 factors against the ridge penalty z, estimated each month from the previous 240 months. Top: within the estimation window. Bottom: the following month, 1992 to 2025. The squares at the left are the unshrunk (ridgeless) estimate. Shrinkage gives up fit in sample and nearly doubles the Sharpe ratio out of sample.

Each notebook on this site starts from an idea in the empirical asset pricing literature and turns it into a portfolio an investor could hold. The notebooks do not claim new results. They show every step: the theory behind the idea, the data, the code, and how the portfolio performs once realistic constraints are imposed. All notebooks draw on the same data sources and report the same performance measures, so their results can be compared directly.

Stock returns from CRSP, intraday spreads from TAQ and the stock characteristics of Jensen, Kelly and Pedersen come from WRDS, which requires a subscription to reproduce. Factor and portfolio returns come from public sources: Ken French’s data library and the factor returns of Jensen, Kelly and Pedersen. The rendered results are stored with the site, so the pages read the same either way. The notebooks page lists them by topic.

Note

This site is a work in progress: notebooks are added and revised as they are developed.