1. Average postbeta each month by size-beta decile group;Table,1963
2. Our average portfolio returns are typically close but can dier by as much as 60 basis points in absolute value for some combinations of size and beta. Two possible explanations for the dierence are (a) CRSP and Compustat have updated and backlled their databases, so we are working with slightly dierent data; and (b) we have included delisting returns in our returns calculations, which Fama and French might not have done. The former cause is dicult to investigate without access to the CRSP and;Our average postbeta and size results are very close to those of FF92
3. Average robust postbeta each month by size-robust beta decile group;Table,1963
4. Table 46 shows the dierence in the average beta estimates (LS -robust). The average returns on the portfolios formed using robust pre-ranking betas are not very dierent from those computed using the LS betas. The average post-ranking robust betas are generally not too dierent from their LS counterparts. Overall, this suggests the beta decile breakpoints are not being driven by extreme outliers;time-series average returns within each size-robust beta group, while Table 45 shows the time-series average post-ranking betas