Though it’s true the great majority of screens have not worked very well in the last several years, you might cut yourself at least a bit of slack.
The S&P 500 has been an extremely difficult target in the last few years because of the outstanding performance of a few giant firms that dominate it.
As an old-school example, take the Value Line 1700 universe, equally weighted, as a proxy for the set of typical stocks an MI screen might be picking from.
The S&P 500 beat the VL equal weight by 6.3%/year in the six years to the pandemic bottom (!).
Those using MI screens were likely to feel bad if using the S&P as their benchmark, even if their stock picking was quite a bit above random.
Then, of course, we saw one of the biggest deep-V bungee bottoms of all time, which reliably breaks any screen using momentum.
So, for this stretch especially, perhaps a better test of one’s success: have the absolute returns been pleasant enough?
And do you have sufficient reason to believe that will continue to be the case?
As for how to do MI successfully, or more successfully, it remains hard.
My approaches have been to concentrate on a deeper, simpler screens with very modest goals.
Aim for 20% outperformance and you’ll underperform; aim for 4% outperformance and you might actually get it.
Pick a few factors that seem to continue to hold up, even if they don’t offer huge advantage.
Each person’s style and research will lead to different conclusions, but a couple of factors I’ve found to have some utility through the recent hard stretch:
A universe that excludes at least some stocks furthest from 52 week highs.
High ROE, though not necessarily just the highest few.
Five year sales growth has been outstanding as a predictor. That may fade somewhat as the cycle turns, but it has always been pretty good.
Cash is never a bad thing.
And some quirky things–Medical devices continue to be good bets, and very high P/B firms do much better than you might expect.
The LargeCapCash screen is new, but has beat the S&P by a bit in its first 8 months since the post.
That’s after friction, 40 stocks deep, 2 month cycle, Top 40 HTD 45 as suggested.
It uses little except ROE and cash. Long run backtest S&P plus about 4-6%/year.
Like most of the things I look at these days, it aims to be boring, with just a small and safe long run advantage.
Truthfully I generally have little or no money in MI most of the time these days.
I create and test and track and rank and select screens continually, but I don’t have much money in it.
I have done so well for so long with Berkshire that it has dimmed my interest in other approaches.
(over 20% IRR for over 20 years)
When it’s cheap I own a whole lot, when it’s not I own less. Rinse and repeat.
When it’s really cheap I go in with both feet, with an embarrassing amount of my portfolio in deep in the money call options.
Some years are truly ghastly, but I know it’ll come back a year or two later.
The attraction is that the financial results are remarkably steady, and it’s not too hard to create a yardstick for the value of a share.
Because of the way I invest, I find those are more important than a high rate of growth.