NYSE Breakpoints and Weighting Conventions

Building a size or value factor portfolio requires first deciding what counts as "small" versus "big," or "cheap" versus "expensive" — some threshold, or breakpoint, that splits the universe into groups. The obvious approach is to rank all stocks in the investable universe and split at, say, the median market cap. Fama and French deliberately did not do this, and the alternative they chose became the field's standard convention for a specific reason.

Nasdaq has historically listed thousands of very small, thinly traded companies, far outnumbering NYSE-listed names. If breakpoints were computed using the whole universe including Nasdaq, the median market cap would be pulled down dramatically by that huge population of micro-caps, meaning stocks that are only modestly small by NYSE standards would get miscategorized as "big." The standard fix is to compute breakpoints using NYSE-listed stocks only — ranking just the NYSE universe to find the median size and the value-metric percentiles — and then apply those NYSE-derived cutoffs to sort the entire universe, NYSE plus AMEX plus Nasdaq, into small/big and value/growth buckets.

This convention matters for replication: two researchers using the same raw data but different breakpoint methodology (NYSE-only versus all-exchange) can get meaningfully different factor portfolios and different measured factor premia, even though both call it "the size factor." Any strategy or paper reporting factor returns should specify which breakpoint convention it used before its numbers are compared to another study's.