I believe that engaging in market timing is essential to achieving long-term stock investing success. That belief follows from Robert Shiller’s Nobel-prize-winning research finding that valuations affect long-term returns. If valuations affect long-term returns, then irrational exuberance is a real thing and stocks are more risky at times when the level of irrational exuberance is high than they are when the level of irrational exuberance is low. So the investor seeking to keep her risk profile roughly stable over time must engage in market timing to have any hope of doing so. Not only does market timing always work, it is always required.
Market Timing Always Works
There is a huge amount of resistance in this field to the idea that market timing always works and is always required. Before Shiller came along, the dominant belief among academics was that the market is efficient and that investors are 100 percent rational in the choices they make. If that were so, irrational exuberance would not exist and market timing would not work. So today there are two schools of thought as to how stock investing works, each starting from fundamentally different core premises.
Seth Klarman On Margin Of Safety Investing
This is part nine of a ten-part series on some of the most important and educational literature for investors with a focus on value. Across this ten-part series, I’m taking a look at ten academic studies and research papers from some of the world’s most prominent value investors and fund managers. All of the material Read More
How can we know which model is legitimate?
The usual way of analyzing the question is to determine the strength of the correlation between today’s CAPE value and the return earned on stocks over the next 10 or 15 or 20 years. If the market is efficient, there should not be any correlation. If irrational exuberance is a real thing, the correlation should be strong. Long-term stock returns should be highly predictable.
The correlation is quite strong, strong enough to persuade me that Shiller’s model for understanding how stock investing works is the right one. But the correlation is not perfect for two reasons.
One, it can take a good bit of time for stock prices to return to fair-value levels. Short-term timing really doesn’t work and any assessment of the correlation between valuation levels and returns produced is based on a mix of short-term and long-term results. The correlation is always weaker than it would be if it were possible to isolate long-term returns.
Two, stock returns play out in cycles in which long bull markets are followed by long bear markets. Prices go steadily up for a long stretch of time (with occasional, temporary price drops mixed in) and then steadily down for a long stretch of time. It can take 40 years or even a bit longer for a bull/bear cycle to come to completion. That means that we only have four bull/bear cycles to look at (we have good records of stock prices going back to 1870). It’s hard to have great confidence in a correlation that has held true for only four bull/bear cycles. The case for Shillers’ model will be stronger when more data is in the books (presuming that his model really is the better model).
The Correlation Between Valuations And Long-Term Returns
There has been a good bit of argument over the years between those who find the correlation between valuations and long-term returns to be very strong and those who find it to be not be terribly strong. It has been hard to reach a consensus on the matter because the correlation is less strong in recent decades than it was in all the years of stock market history available for review from 2006 backwards. Were we to see a price crash within the next year or two or three, the case for the Shiller model would be strengthened. But in the event that the Shiller model is not legitimate, there is no reason to believe that a price crash is more likely today than it would be at any other time. Price crashes are always equally likely in a world in which the market is efficient.
I believe that a factor that is not given sufficient attention is the way in which prices have always headed upward for a long stretch of time and then downward for a long stretch of time. If the market were efficient, prices would play out in the form of a random walk. But that is not at all what we see when we look at the historical return data. It’s not just that the data shows a strong correlation between valuations and long-term returns. The pattern in which prices play out supports Shiller’s idea that it is shifts in investor emotion that drive stock price changes rather than economic developments. Investor emotions are slow to change. Once the idea gets locked in that prices will he headed upward, irrational exuberance rules the day for a large number of days Then, once the idea gets locked in that prices will be headed downward, irrational depression rules the day,
Stock prices do not reflect the current day’s economic realities. They reflect the stubbornly held emotions of investors, positive for a long stretch of time and then negative for a long stretch of time following a price drop painful enough to break the optimistic outlook. The fact that prices have always played out in this manner is strong evidence that Shiller is right, evidence every bit as strong as the more frequently referenced evidence that today’s valuation level tells us much about the stock return that will apply for the next 10 and 15 and 20 years.
Rob’s bio is here.