Would it be possible to create a universe of stocks with negative ROE and then apply these rules to get a usable list?

I assume so, though I haven't done so.

I have created screens for negative-book-positive-earnings firms in the past. But again, with the VL database since that's what I have handy.


I did a simple screen simply looking for these (book<0, ttm eps>0, current eps>0, proj eps growth>0), then did a final sort on absolute size of cash pile, yet another way to try to avoid true duds.
Top 5 monthly 2007-2024 CAGR 18 vs S&P 11. Beat the market before,during, and after the credit crunch: risk 3/4 of SPY based on my DDD3 metric. This demonstrates that they seem to be, on average, good picks.

Some of them are considered famously reliable businesses. Moody's, Coke, YUM, PM, Dell. Domino's and Papa John, not sure why.

The problem is that there aren't that many, especially more than a decade ago. So it isn't a statistically huge ocean from which to go fishing.

Jim