Closed chenandrewy closed 1 month ago
Section III states that their return regressions use value-weighted portfolios. Updated in https://github.com/OpenSourceAP/CrossSection/commit/0af9fee61b55f704f163d434d8b43059c63d888b
I assigned sweight = VW based on the wrong paper (Litzenberger and Ramaswamy, 1982, JF). The 1979 paper is not super explicit about stock weights but from the description around equation (55) and Table 1, it sounds like they use equal-weighting. The correct paper is here: https://doi.org/10.1016/0304-405X(79)90012-6
And sweight is fixed here: https://github.com/OpenSourceAP/CrossSection/commit/f9b4823224f19d30ba6fa2e8c52e6cfefe88a778
all the other predictors are non NA