I'd like to analyse the data as well. Can you open source your methods and sourced data? I'm very in favour of spreading this analysis, but I'm afraid that there might be some sort of bias going on (for example, I really want to adjust for industry, because I suspect that tech and media companies are leading this trend, while materials are not).

Edit: Just saw that this is an interactive notebook! Awesome! Going through it now.

Edit2: It looks like you've already done what I planned to do! Awesome work!

You also may want to try betting on the spread between male-only companies and ones with at least one woman on the board. If the investment thesis is sound (board gender diversity is beneficial), this would be a good way to test it. It would also bring you into market-neutral territory.

I wanted originally to go for board diversity, but haven't found that data set yet. I believe Morningstar has some of it, and am hoping to get my hands on it sooner rather than later.

Ok, but you're extracting the price of SPY, so you're not counting SPY's dividend. Use a portfolio that simply buys SPY at the beginning of the period and holds it.

The volatility thing it the biggest effect though.