The Trial of Barry Ritholtz

The Full Implications of No Predictability

However, while many people now accept the premise of no predictability in the stock market, most of them don’t understand its full implications.

If there is no predictability, you must rule out conclusions like “investors underperform their investments because they time the market badly” (Morningstar’s “mind the gap” studies notwithstanding — see the Appendix for why those studies don’t support this conclusion). If the market is unpredictable, then any trade has an unpredictable result. It cannot be a “badly timed” trade with a bad result or a “well-timed” trade with a good result. It can only be a trade with an unpredictable, i.e. random, result.¹

You can’t argue that rebalancing a portfolio is beneficial because it sells high and buys low. You can only say it sells high relative to past price and buys low relative to past price. You have no idea if it buys low relative to future price or sells low relative to future price.

You can’t use a fund’s or a manager’s record of past investment performance to decide whether to invest in that fund or that manager in the future.

A manager who is paid more cannot be expected to achieve better investment results than one who is paid less, since the future is equally unpredictable for both.

All of these things are well supported by empirical evidence. Past investment performance (net of fees) has no clear correlation with future investment performance. Buying “high” after stock market rises and selling “low” after drops does not necessarily negatively impact performance. (The latter experiment can be replicated by anyone who chooses to do it using their own parameters.)

Investment managers who are paid more do not outperform those who are paid less; in fact, the opposite — the highest paid managers, hedge fund managers, achieve the worst investment performance after fees. And as we shall see, rebalancing does not necessarily boost performance in the long run.

I know that many people will not believe that these claims are well-supported by empirical evidence. Nevertheless, they are. I will debate anybody on this point who agrees to send me the evidence they believe refutes it in advance.

Does Ritholtz understand these implications? This is what was not clear to me.

The Trial

Before finishing Ritholtz’s book and before writing my review, I asked Ritholtz for a Zoom interview to discuss it. He accepted cheerfully and with alacrity. He also suggested that if I shared my questions in advance, he could “come better prepared with specific citations and research.”

I went him one better. I wrote him a four page tract of comments, questions, and arguments. But despite providing him in advance with my critique, I felt that while he defended against it adequately on several key points, he was unable to mount a successful defense on others.

Which is why I am taking the liberty of calling it a trial — if only for dramatic effect.

Spoiler alert: not guilty, except possibly on one minor count.

What Were the Charges and Why Was He Not Guilty?

Ritholtz’s recommendations for investing in his book were wholly in tune with his no-predictability stance. But some statements in the book weren’t. That’s why I put him on trial.

Let’s take a few points on which some of the statements in his book were out of tune with his recommendations, and with the theme of no predictability. These are the charges that arose during the trial.

The charge: He says, at location 4,045 on my Kindle, “Rebalancing back to your original weightings has been shown to add 75 to 150 basis points of additional returns over the long term.”

Huh? “Shown to add 75 to 150 basis point over the long term” compared to what? Presumably compared to not rebalancing — what else? So let’s take the long term. Imagine you’re 100 years old and when you were 3 years old you were bequeathed $1,000, which was then invested for you 60% in stocks and 40% in bonds. If you rebalanced until the end of 2024 you would have $4.2 million now. But if you never rebalanced you would have $6.1 million.

And what about risk? Your worst white knuckle period would have occurred if you rebalanced in 1929-1932, as William Bernstein showed in his 2013 book, “Deep Risk.” Each time you rebalanced, stocks did worse, leaving you with very little. There was nothing quite as extreme as that extraordinarily severe white knuckle period if you practiced buy-and-hold. And what is risk except for the risk of white knuckles? The investment field’s choice of volatility as a proxy for risk shows that that is what it thinks risk is: white knuckles.

The defense: That is not what Ritholtz actually recommends to investors

Ritholtz says in one place (location 1,062 in my Kindle version) “Rebalance every few years.” Then he says it again, “rebalance once every few years,” at location 7,167.

I am in complete agreement. Because by “rebalance” he surely doesn’t mean “mindlessly restore your asset mix to the one you held five years ago.” Rather, he must mean, “take stock of your financial and life situation now, five years later, and consider how or whether to restructure your portfolio to one that suits your new situation.”

Not guilty on this count. Though one must wonder why, in one or two places in the book, he sings the praises of some unspecified version of rebalancing.

The charge: The behavior gap

Ritholtz says, “The behavior gap is the underperformance between what each asset class returns and how investors in those assets did.” It is presumed to be due to investors selling low in a panic and buying high in a FOMO frenzy. This is now the virtually universally accepted conventional wisdom. But it is wrong — see Appendix.

The defense: That presumption is not actually what Ritholtz warns of

Ritholtz does not say outright, as so many others do, that this “gap” is due to investors timing the market wrong by panicking and selling after it has gone down and then buying after it has gone up. Instead, he rightly invokes the fact that “30.9% of the investors who panic sell never return to reinvest in risky assets.”

Yes, this is the problem with investors panicking and getting out of the market — that they might never get back in again. The Morningstar mind-the-gap measure does not measure this. To measure it would take an entirely different kind of study — one that I wish Morningstar had done. Being in the market all or most of the time is the best way to achieve superior investment returns.

Not guilty on this count, though I do hold him responsible for coasting along with the conventional wisdom about the “gap” without pointing out that the gap measure is wrong.

The charge: The strange case of Jack Meyer

Ritholtz’s coverage of this topic really baffles me. He seemed to lament that, because some “politically correct” members of the Harvard board objected to endowment manager Jack Meyer’s $8 million salary when compared to the much lower salaries of tenured academics (who are not in general poorly paid), his expertise at producing high returns for Harvard’s endowment was lost.

This seems to imply that the more an investment manager is paid, the better his performance. Yet this isn’t true at all. If it were, how come David Swensen performed so well for more than 20 years managing the Yale endowment despite being paid one-eighth Meyer’s salary? And incidentally, after Swensen was lionized for years of outperformance, his performance slumped to average or below average — yet another case of a track record being no guide to the future. Hence, why would Ritholtz lament the loss of Meyer, when under the no predictability hypothesis –—and the example of Swensen — there was no reason to believe his investment returns would continue to be good?

The defense: ?

I’m afraid on this one there is no defense, because we didn’t get to talk about it much in our interview — at least I don’t think we talked about it much. Therefore, tentatively guilty as charged, but pending appeal.

Endnotes

¹ Some definitions distinguish random from unpredictable, or uncertain. Random means predictable but only according to a specifiable probability distribution. Uncertain means even the probability distribution cannot be specified.

APPENDIX

Why Morningstar’s Mind-the-Gap Measure Does Not Show That Investors Underperform Their Investments

I have not seen anyone question the eminently questionable inference that because Morningstar’s mind-the-gap measure usually shows a difference between investors’ time-weighted investment return and dollar-weighted return, it means that investors underperform their investments by buying high and selling low. It does not show this at all.

First, let’s consider the claim, that “investors” underperform their investments. What investors? All investors? But all investors put together can’t underperform the market (except by paying too much in fees); they are the market. If Morningstar means only some investors underperform their investments they must say which ones and why.

But beyond that question, there is something seriously wrong with interpreting Morningstar’s mind-the-gap measure that way.

The measure is the difference between two rates of return. Let’s say in both cases the investor(s) are invested in the total stock market, or a proxy for it. The first rate of return is the time-weighted rate of return, which measures the return an investor would have gotten if she invested an initial amount, let’s say $10,000, in, for example, a total stock market index fund and kept it there for a specific period of time without withdrawing from or adding to it.

The second rate of return is the internal rate of return (IRR) that the investor(s) got, taking into account their dollar inflows and outflows during the time period.

In the first case, the time-weighted rate of return, we know where the investor’s dollars were the whole time. They were in the index fund. But in the second case, we know only that the investor’s dollars were in the index fund some of the time, while at other times they were not in the index fund. They had been withdrawn, and then perhaps contributed back again later. But we do not know what those dollars were doing during the period after they had been withdrawn. Therefore, we do not have a complete measure of the investor’s performance, nor an appropriate comparison.

Let’s take the simplest possible example of how this measure can get it seriously wrong.

Investor A and investor B invest in the same index fund during a three-year period. Both of them initially invest $10,000. In year one the fund’s return is 5%, year two it is 10%, and year three it is 5%. Compounding those and annualizing, an investor who invested in the fund for the three years would have an annualized, time-weighted rate of return of 6.64%.

Now let’s consider investor A’s IRR. Investor A’s cash flows, which is all we know — remember we don’t know what the investor does with the money after it is withdrawn — are as follows:

Beginning of year 1: inflow of $10,000
End of year 1 / beginning of year 2: outflow of $10,500
End of year 2 / beginning of year 3: inflow of $10,500
End of year 3: outflow of $11,025

These figures are sufficient to calculate investor A’s IRR, which is 5%. Subtracting the investor return of 5% from the investment return of 6.64%, investor A had a 1.64% “behavior gap.”

Now let’s move to investor B. Investor B’s cash flows are as follows:

Beginning of year 1: inflow of $10,000
End of year 1 / beginning of year 2: outflow of $10,500
End of year 2 / beginning of year 3: inflow of $12,600
End of year 3: outflow of $13,230

Calculating investor B’s IRR, we again get 5%. Investor B’s “behavior gap” is again 1.64%.

If both of them, according to Morningstar’s mind-the-gap measure, had the same behavior gap, how come investor B wound up at the end of year 3 with $13,230 while investor A had only $11,025.

The answer is simple. After investor A withdrew their $10,500 they just sat on it for year 2 – it earned no interest. That’s why investor A’s cash inflow at the beginning of year 3 was the same as what they withdrew at the end of year 1.

But investor B apparently had some better alternative (maybe they bought and flipped a property), one that earned investor B a 20% return for year 2. That’s why investor B withdrew $10,500 at the end of year 1 but contributed back $12,600 at the beginning of year 3, 20% more than they had withdrawn at the end of year 1.

Why does investor B get no credit for finding a better alternative to the index fund during year 2? Both investors, according to the Morningstar measure’s implication, were bad at timing the market because they both got the same “gap.”

This measure does not measure investor “bad behavior.” It doesn’t really measure anything.

Economist and mathematician Michael Edesess is an adjunct professor and visiting faculty at the Hong Kong University of Science and Technology. In 2007, he authored a book about the investment services industry titled The Big Investment Lie, published by Berrett-Koehler. His new book, The Three Simple Rules of Investing, co-authored with Kwok L. Tsui, Carol Fabbri and George Peacock, was published by Berrett-Koehler in June 2014.

A message from Advisor Perspectives and VettaFi: To learn more about this and other topics, check out some of our webcasts.

Read more articles by Michael Edesess