If you were to look at the TOTAL performance of all small stocks over the 31 years versus all large stocks over the 31 years, perhaps the large stock performance would look a little better than the small stocks ... mainly because a substantial percentage of the small stocks disappeared over the period (and essentially went to zero or were acquired at some level that stops its compounding at that point). I don't have sufficient data to say one way or another, and it remains a thought experiment for me. If anyone has data, and the wherewithal to massage it, it might be interesting to see the results and the methodology.

This surprisingly hard to test, even when you have the data : )

The reason is that you can't simply (say) divide stocks into deciles by market cap and see how each decile did on average over time. Each interval that you reconstitute, some of the stocks will have moved between bands, which causes them no longer to be counted in the same band...their move counts against a different band in the next interval. If those moves are caused by mean reverting random noise, you'll be pumping non-existent returns from large caps to small caps. (a small stock does well and that performance counted in its small cap band; it's promoted to the next band in the next period; it falls again, and that fall is counted against the higher band, repeat...)

If there is only one division, between small caps and large caps, that is much less of a problem, as there is only one boundary being crossed. You can add a "slush" band as well, I think some index providers do this: a small crossing of the dividing line isn't acted upon.

In the end, there isn't really much of a small cap effect to speak of, and to a large extent there never was. Mainly there is an illiquidity premium, compounded by the fact that illiquid stocks and very old prices have slightly more errors than liquid stocks and more recent periods. If some easily-investable small cap group clearly outperformed (say) some large cap group over very long periods, it would get bid up to prices that the effect disappeared. Instead, we see short and long cycles of one or the other being in the lead as fashions change.

A much more interesting study is to look at the differences in returns based on share turnover ratio (ratio of average daily volume to shares outstanding). But that's for another day.

Jim