1/9 Thread: Value investing's El Farol Bar problem
Much has been talked about the apparent death of value investing. What I perceive though is value investing is having its El Farol Bar problem.
What is El Farol Bar problem?
Much has been talked about the apparent death of value investing. What I perceive though is value investing is having its El Farol Bar problem.
What is El Farol Bar problem?
2/9 El Farol is the happening bar in town. There are 100 people in the town interested in going to the bar on a Thursday night.
If the number of people in the bar crosses 60, people don't want to go since it's too crowded. They would rather stay home instead.
If the number of people in the bar crosses 60, people don't want to go since it's too crowded. They would rather stay home instead.
3/9 Now imagine El Farol publishes attendance data for Thursday night every week.
Voila, you can build your model now!
People start extrapolating based on last week's data. Since other people are doing the same, it's not very helpful.
Voila, you can build your model now!
People start extrapolating based on last week's data. Since other people are doing the same, it's not very helpful.
4/9 Then people start looking at average of last few weeks. Some look at weather, other potential activities in the town on the same night etc etc.
Whatever your model is, you will see the attendance data the next day and will know whether your model was correct.
Whatever your model is, you will see the attendance data the next day and will know whether your model was correct.
5/9 Model iterations are never ending. Bad models will simply not survive over time.
At some point, the effective model of yesteryear's will also not work. The iterations are the only constant.
I believe markets act in a similar fashion.
At some point, the effective model of yesteryear's will also not work. The iterations are the only constant.
I believe markets act in a similar fashion.
6/9 In the book "Investing: The Last Liberal Art", Robert Hagstrom mentioned about major strategies that worked in each decade:
In the 1930 and 40s, Discount-to-book value was in vogue, thanks to Graham and Dodd.
Then dividend yield reigned the 50s.
In the 1930 and 40s, Discount-to-book value was in vogue, thanks to Graham and Dodd.
Then dividend yield reigned the 50s.
7/9 60s was all about which companies can grow earnings faster.
By 80s, Buffett's "owner-earnings" or cash flows, and more recently, it's the ROIC that seems to do the trick.
Perhaps, even more recently, it's the TAMs and LTV/CACs that have produced eye popping returns!
By 80s, Buffett's "owner-earnings" or cash flows, and more recently, it's the ROIC that seems to do the trick.
Perhaps, even more recently, it's the TAMs and LTV/CACs that have produced eye popping returns!
8/9 Just like the forecasters in El Farol Bar, "value investors" in the market approached *value* differently than the past because the models needed iterations for it to work.
This is coming from a person who owns zero SaaS companies.
This is coming from a person who owns zero SaaS companies.
End/ Markets are complex, adaptive, and evolutionary!
That's how it has always been, and that's how it is always going to be.
That's how it has always been, and that's how it is always going to be.
