Quantitative Strategies and Liquidity Stress across Financial Crises: Evidence from U.S. Equities

Abstract

To reconstruct the monthly S&P 500 membership from 2007 to 2025, this paper employs public records. This work is trying to minimize universe look-ahead bias of a sample limited to the constituents of a universe. The membership snapshot at the end of each month is used for the construction of the portfolio. It is not necessary that stocks have return histories for the whole sample. Licensed archives are more trustworthy than public records. To get a sense of the impact of this data constraint, this work compares point-in-time sample to a survivor panel using ex-post membership lists. Fixed weight assumptions for handling missing returns are also tested. This reconstruction affords an opportunity to focus on two components of the same universe: the performance of four equity strategies during normal months and crisis months, and the evolution of market illiquidity during the two periods. Regime differences in strategy performance. This work calculates the changes in log illiquidity by using a regression fit with HAC standard errors with six lags. Forecasts have moderate predictive power, but the rankings change, so it is not possible to rank the three models. Some sources of selection bias remain unresolved in the case of public reconstruction.