Has AI Destroyed Value Investing?

Maybe. Maybe Not.

For years, people have been talking about the underperformance (or even death) of value investing.

They aren't wrong.

From the years-long underperformance of small cap value stocks to the relentless momentum of unprofitable meme stocks (some with zero revenue), buying optically cheap companies based on price-to-earnings or price-to-book has not been a winning strategy.

It's why I was intrigued when Guy Spier, one of the most prominent value investors out there, wrote a Bloomberg article arguing that the golden age of value investing is over.

It's an interesting read. It raises many questions about how old-school value investors will survive in an era where AI powers an increasing amount of investment research.

Perhaps it was an unpleasant truth. But the days of getting an edge by reading 10-Ks are fading. The argument is even firmer if the SEC abandons quarterly reporting requirements (which seems more probable by the day). As the models get better, the edge in spending hours reading 10-Ks will likely diminish even more.

So what are the consequences?

For starters, stockpicking isn't going away anytime soon. As long as humans are involved, there will be a desire to actively manage and outperform the market. This is particularly true among younger individuals, who have more hands-on experience managing their portfolios through brokerage apps like Robinhood.

But with AI investment tools democratizing all kinds of information, knowledge curation becomes a critical source of edge. Perhaps in this context, Orwell would say something like some signals are more equal than others.

And with the increasingly sophisticated power of AI, we may find newer, more sophisticated signals that have direct correlations with stock movements. It's something I'm working on now (stay tuned to learn more!).

Then there is time arbitrage. In the mid-1970s, the average holding price for a stock was five years. In 2022, it decreased to 5.5 months. In today's market environment, this figure is probably even less. I agree with Guy that time arbitrage is still an edge. That said, when stocks are trading at multiples not seen since the Dot Com bubble, forward returns (even if you buy and hold) may not be as attractive as they appear.

At its core, value investing isn't about buying stocks that trade with single-digit PEs. It's about buying things that are less than they're worth. It's a higher-level definition, but it's one with an important distinction.

On one level, LLMs and AI-powered investment research platforms will make it much easier to understand a company's financials. But on the other hand, I think their sheer power will help technologically-savvy investors find unique fundamental and technical signals for individual companies. It'll be fascinating to see how this shakes out!

Prompt of the Week

I've always liked Jeff Bezos's regret minimization framework. Basically, the idea is that you pretend to fast-forward to when you are 80 years old. Then, sitting as an 80-year-old, you look back on your life, identify the regrets you may have, and work today to minimize them.

Instead of fast-forwarding to 80, I like the prompt below because it is more short-term. Try it and see for yourself!

"Fast forward 3 years. Based on how I think and act now—what's the regret I'm most likely to have if nothing changes? Then, please design a 7-day self-recalibration challenge to help me break that trajectory."