@macaron Equally weighted portfolios have a value and size tilt, i.e., they hold a higher percentage of small companies and a higher percentage of value companies than their parent indices. This probably explains a lot of the difference in performance.
MSCI has some equal weighted indices, I can run a regression to compare it to the standard MSCI World index, which is value-weighted (i.e., market-cap weighted). The MSCI index data website seems to be down right now unfortunately.
edit: I managed to download the data, here are the results for the Fama-French three-factor (market, size (SMB), and value (HML)) model with momentum (WML). Asterisks denote statistically signifant results (at p < 0.0001) and α is annualized.
Market
SMB
HML
WML
α
R[sup]2[/sup]
MSCI World
0.99*
-0.19*
-0.01
-0.01
0.02%
0.99
MSCI World Equal-Weighted
1.00*
0.24*
0.21*
-0.13*
0.43%
0.97
The returns are explained very well by the four factors in both cases (very high R[sup]2[/sup]). The equal-weighted index has quite strong size and value tilts, as I expected. It also has a negative momentum load, which also makes sense since the index has to sell stocks that recently did well or buy stocks that recently did poorly to bring them back to equal weighting, so it's trading against momentum.
There's no ETF following this index as far as I'm aware, so this is purely academic
. There's nothing particularly special about equal-weighted indices though, their performance is generally explained quite well by exposure to factors.