People write for many different reasons, but typically there is a financial motive. This nudges the author in the direction of seeking instant popularity. This includes scaring people witless, confirming existing biases, and striving for instant winners.
But what happens if you have an important idea that does not fit these categories? My own success has come from an unusual contrarian approach: I seek themes that the conventional wisdom has unwisely rejected. My approach is based upon data and evidence. Surprisingly, that often leads to buying markets and sectors that are unloved by the market. This approach can identify the very best opportunities and, also keep investors on the right side of the market.
One of the attractions of writing at FATrader, beyond the obvious opportunity for interaction with a group of investment experts, is the audience we are seeking. That audience emphasizes those who want to learn and to understand what they are doing. My hope is that it will be a more patient and thoughtful group than I find at sites where people want to speed read and find a tip.
Right or wrong, that is my hope and I am giving it a try. My priority is to write about what is most important and valuable to the investor. This means challenging conventional wisdom that is simply wrong. Anyone who challenges orthodoxy needs to be persuasive and to provide evidence. The entire subject is more like a mini-book. I cannot do it in a post or two, and most would not read it anyway. I am splitting my topic into digestible chunks, each worthy of discussion. The investment relevance will not always be obvious. That is by design. Convincing people to be open-minded is more readily done with an oblique approach.
Please play along and give this idea a chance. Be willing to place yourself in challenging, non-investment circumstances.
The Baseball Challenge
Let’s suppose that we had a simple forecasting problem. How many home runs will be hit this year in MLB? As data scientists, we would turn to history to provide the basis for a projection, while realizing that “past performance does not guarantee future success.”
Fortunately for us, we have a lot of data available. Thanks to recent research on historical results, the data extends back to 1871. A good approach is to consider how many plate appearances it takes to get a home run.
In the early days of baseball, the same ball was used for a long time – sometimes most of the game. That was less costly for owners, but the ball became mushy and “dead” as the game progressed. The lively ball era came when balls were changed more frequently, and the needed plate appearances for a home run decreased. When baseball’s color line was broken, there were many more great hitters, moving the averages even lower. Let’s take a closer look at the post-war era.
We can readily see changes from events. When there were more expansion teams, talent was thin. Free agency also seemed to increase home runs. And finally, we have the steroid era. When it ended, plate appearances per home run increased, at least until recently. Some credit the uppercut swing, going for a home run and willing to accept a strikeout, as the reason for the recent decline.
Statistical and Forecasting Conclusion
This is what statisticians call a non-stationary series. The mean and standard deviation change over time. We would be foolish to use the dead ball data to predict the number of modern era home runs.
If many others were doing so, we could profitably take their bets. They would be the dumb money.
By coincidence, the most popular stock valuation measure, Prof. Robert Shiller’s CAPE ratio, also uses data as far back as 1871. This approach is widely appreciated for several reasons:
- It uses known data – no forecasting. No one can quarrel with the calculated result.
- It was developed by the top minds in behavioral economics.
- It correctly signaled danger in the 2000 era.
- It supports a popular meme about stocks, bubbles, and danger.
There have been various criticisms over the years, mostly based on technical changes in measuring variables. My colleague covered the current debate nicely in this post.
Let’s move beyond the standard criticisms, beginning with a long-term chart of the ratio.
Even if you are not a statistician, the problem of different eras leaps from the page. If we wanted to make a solid prediction related to current times, we would not invoke the “dead ball” era of CAPE.
Should we be surprised at these changes? Certainly not. Out of thousands of possible candidates over the years, let’s focus on improvements in analysis since 1980 – improved accounting rules, professional analysts with regular estimates, computers, information transfer, and changing regulations. Some of the most important factors occurred post-2000.
We can speculate on the reasons, but the evidence itself is clear. This is a non-stationary series. Many people are betting on it, allowing their entire world view to be shaped by 19th century data. We can take the other side of this bet, identifying conventional wisdom as “dumb money.” (BTW, the real pros do not use CAPE, but that is a story for another day).
More – much more – to come on this theme. So much Wall Street “research” and conventional wisdom now uses data stretching to 1871.
Just because irrelevant data are available does not mean you must use it!
Here is Chicago in 1869.
And thanks to the excellent work of Vanderbilt student Eric Lo, my intern last year. He has a very bright future.