by Rob Bennett
The Bogleheads Forum has been hosting for several weeks a super thread on the merits (or lack thereof) of Valuation-Informed Indexing. A poster going by the screen-name “Grayfox” brought up a great point that I want to highlight and respond to here.
The thread was launched to explore research by Wade Pfau (Associate Professor of Economics at the National Graduate Institute for Policy Studies in Tokyo, Japan) showing that Valuation-Informed Indexing beat Buy-and-Hold in 102 of the 110 rolling 30-year time-periods now in the historical record and that long-term timing provides comparable risk and the same average asset allocation as a 50/50 fixed allocation strategy but with much higher returns.
Prescience Partners returned 6.75% for the second quarter, underperforming the S&P 500's 8.55% return but coming out ahead of the Barclay Equity Long/ Short Index's 2.62% return. However, for the first six months of the year, Prescience is up 30.66%, doubling the S&P's 15.25% return and smashing the Barclay Equity Long/ Short Index's 9.27% return. Read More
As the poster using the screen-name “Fred Flintstone” notes: “The paper refutes a central tenet of the Boglehead investing philosophy.”
That it does.
But Greyfox brings up a point that has only rarely been examined in much depth.
He says: “Without some sound theoretical basis, there is really no way to know that some superior results of a strategy aren’t just random results that won’t repeat in the future.”
I strongly agree.
The fact that the entire historical record shows both that Buy-and-Hold can never work well for for the long-term investor and that Valuation-Informed Indexing will always offer higher risk-adjusted returns is powerful evidence in support of my favorite investing strategy. But Grayfox is right. Even 140 years of data is not enough. Without a theory that makes sense of that data, making changes in your stock allocation pursuant to the consistent message of the data is a dangerous business.
Because the data could be the product of data-mining.
There are all sorts of things that we can “learn” about stock investing that are simply not true simply by arranging data in ways that make it appear to support our biases. The test of a data-based strategy is the theory behind it. If the theory makes sense and if the data supports the theory, there’s a good chance you have founds something od great value. If all you have is data, you have come up with a good reason to begin serious investigations of what is really going on, nothing more.
Say that someone discovered that stock prices always go up dramatically during months with an “r” in them (January, etc.) and always go down dramatically during months with no “r” in them. Would that finding constitute good cause to increase your allocation in all the r months and lower it in all the non-r months?
Not in my book.
It could just be a coincidence.
But if the person making the finding could offer a convincing explanation of why stocks would perform better in r months, I would take the data-based finding far more seriously. Data alone doesn’t cut it and theory alone doesn’t cut it. The combination of the two makes for a powerful case.
Why is it that stocks always do well in the long term starting from times of low or moderate prices and always perform poorly in the long term starting from times of insanely high prices?
It’s because insanely high prices are — insane!
Do you see?
I recall my father and mother arguing at the kitchen table one morning when I was a boy as to whether stock investing is gambling or not. My father said that stocks are shares in businesses and thus investing in stocks is not gambling. My mother said that no one can make sense of stock investing and thus investing is akin to putting money down on a horse at the race track.
I’ve always been a peacemaker. So my belief (and the historical data backs me up on this) is that stock investing is two, two, two, types of money transaction in one. It really is in part an investment in real businesses. And it really is in part gambling.
What determines which trait of their stock investing experience is dominant at any given time? Valuations!
When stocks are priced fairly, each $1,000 that you put into an index fund is all going to the purchase of businesses that can be realistically expected to generate profits for many years to come. When stocks are priced at the times fair value (as they were at the top of the bubble), only $350 of each $1,000 you put into an index fund is going to the purchase of shares of businesses. The other $650 is going to the purchase of cotton-candy nothingness fated to be blown away in the wind sometime over the following 10 years or so.
Those who like the idea of investing their retirement money in real businesses and don’t at all like the idea of gambling with their retirement money should be putting a larger portion of their money in stocks at times when the investment component of an index fund purchase is high and a smaller portion of their money in stocks art times when the gambling component of an index fund purchase is high.
That’s the theory.