In an excellent recent interview with The Investors Podcast, quantitative guru Tobias Carlisle talks about the powerful force of mean reversion in the stock market and economy, and offers a number of insights about quantitative investment strategies.
Carlisle says that, while many investors focus on companies with the best recent growth, the nature of business cycles often means that those companies underperform going forward, while businesses with the worst recent growth often do the best going forward. “Most businesses and stock markets are cyclical where they have periods of both good and bad returns,” Carlisle explained. “Mean reversion can sometimes fool you when you’re looking at the trend of earnings, so the point of the book [his Deep Value] was to show that most businesses experience the fundamental mean reversion. Few businesses can sustain high growth in the long run. If you can anticipate its implications and get something that’s cheap and is at the bottom of its business cycle with earnings that are falling, then you arrive at the difference between the market price and the intrinsic value and as you see the intrinsic value increase, you stand to make good returns.”
Carlisle also talks about his approach to selling. “Often, there will be a catalyst or you will be buying another security that is even cheaper,” he says. “We find that it is actually the discount to intrinsic value, which is the main driver for the stock to increase in value. Holding a stock for a year will give you the bulk of the return when exercising deep value strategies. In the US where you pay significantly less capital gains tax after a year, the best combination of profiting from the discount and limited taxation is often by holding the stock for one year and one day.”
When it comes to using quantitative strategies, Carlisle says that it’s crucial that you don’t try to pick and choose the stocks that the strategy selects. “Once you find a good ratio, you have to apply that to the stocks without a fear of failure although it may not seem very profitable, but you will find that it if you try to cherry pick those ideas, they underperform the screen and the reason for that is because the experts tend to underperform these simple statistical models and use their discretion to override these models too often,” he said. “Through various unrelated fields, they found that the statistical model acts as a ceiling on performance and anything you do [to] the screens actually underperforms. When I’m in the quantitative mode, I just run whatever is in the model and put them in the portfolio.”