Here is a scenario that I’m confident has happened.
You are a naïve but intelligent investor, with a penchant for researching a subject on your own; a young microbiologist, let us say, recently promoted to associate professor. You have a very small but growing nest egg in a bank savings account. You expect to add to it regularly, but the interest rate is pitiful.
It’s time to think seriously about investing. You want to learn for yourself what investing is all about because you suspect that there’s a lot of bad advice out there. You take to the internet to start exploring.
You come across an article that won the Financial Analyst Journal’s award for best paper of 2013. You tracked it down from something you saw recently in The Economist. It is written by two Yale professors and two researchers at investment management firms – a good mix of academics and practitioners. You decide this article must provide reliable information. It is as good a starting point as any.
After reading the article you conclude that:
There are easily identifiable factors that enable you to assemble superior stock portfolios.
The difference between the best stock portfolio and the worst is huge; $10,000 invested in the best portfolio grows to $8.5 million over 40 years but less than $1 million in an average portfolio and only $17,000 in the worst.
A great deal of academic research into these factors has been performed over more than 30 years.
One of the most potent factors has been discovered only recently.
In short, there is a lot of data, there has been a lot of research and the conclusions it reaches are powerful. You are starting to feel already that solid evidence shows you how to invest. You’re almost ready to go out and invest in the stock portfolio that the article showed was the best.
Wait a minute…
But first you continue your search. You start to discover some surprising information. You learn that most professional investment managers get results that are worse than the stock market’s average. Furthermore, the percentage of those managers that perform better than average consistently is less than you would expect if each were randomly assigned by a coin-toss to do better or worse than average each year.
Didn’t these professional investment managers know about the research, you wonder? They must have gone to inferior schools. But no, you find that isn’t the case.
This is a conundrum. Imagine how baffled our microbiology professor must be. It’s as if years of research found out how to cure polio, but when doctors treat polio, fewer patients recover than if they hadn’t been treated.
Our microbiologist also wonders whether the Yale professors know that practitioners are apparently not paying any attention to their research. Don’t the investment manager coauthors know that their ideas are not being employed? If they do, their article doesn’t say a word about it. Wouldn’t they at least lament that the research is not being put into practice, or speculate about why it isn’t? Or if practitioners are trying to put the research into practice, why isn’t it working?
What the article actually says
The 2013 Financial Analysts Journal article I referred to is “Liquidity as an Investment Style” by Roger G. Ibbotson, Zhiwu Chen, Daniel Y.-J. Kim and Wendy Y. Hu. The Economist does not directly reference this article but one written by Ibbotson and Thomas M. Idzorek in The Journal of Portfolio Management titled “Dimensions of Popularity.”
The information that The Economist highlights1 is that over the 42 years between 1972 and 2013, a portfolio comprised of the stocks with the lowest-quartile share turnover in the previous year returned an annualized 15.51%, while a portfolio composed of the stocks with the highest-quartile turnover returned 8.27%. Hence, $1,000 invested in the low-turnover portfolio would have grown to more than $426,000 over the 42 years while the high-turnover portfolio would have grown to only $28,000. This is a very big difference. The second and third quartile portfolios were intermediate between the two. In other words, the higher the turnover in one year the lower the next year’s return.
Is it popularity, liquidity or turnover?
In the article in The Economist the quartiles are labeled quartiles of trading volume; in Ibbotson et al.’s 2013 Financial Analysts Journal article they are labeled quartiles of liquidity; and in Ibbotson and Idzorek’s Journal of Portfolio Management article they are labeled quartiles of popularity.
Which is it? All are measured as the share turnover in a given year. Ibbotson et al. define that as the sum of the 12 monthly trading volumes divided by each month’s shares outstanding.
Is this liquidity? Not necessarily. Ibbotson et al. do not claim it is the best measure of liquidity, just a simple one that “works well.” Hence in their 2013 article they choose to label it as a measure of the liquidity “style,” or “factor,” just as book-to-market ratio is associated with the value investing style or factor.
In their 2013 article, Ibbotson et al. compare the power of the liquidity factor with the power of the vaunted three factors studied by Fama and French: size, value and momentum. Liquidity has greater power to divide the high-performing portfolios from the low-performing ones than any other factor except value, and even there it’s a close call.
Why, one wonders, hasn’t this been pointed out before? Fama and French’s identification of the factors value and size in 1992 and 1993 and their addition of the momentum factor have defined a three-factor model for years. How did they miss liquidity?
In their Journal of Portfolio Management article Ibbotson and Idzorek posit that all of these factors are measures of popularity. A stock is unpopular if it has low value, is small in capitalization or has low turnover. Any way you look at it, it’s the popular stocks that are overvalued and therefore will have low returns, while the unpopular ones are undervalued and will have high returns.
However, they also show convincingly that their liquidity factor is a new factor. It is not already captured in the other three factors, but takes on values independent of them. For example, among the lowest-quartile size portfolios, the low-liquidity (i.e., low-turnover) portfolios had an annualized return of 15.36% while the high-liquidity portfolios had an annualized return of 1.32%. Among the highest-quartile value portfolios, the low-liquidity portfolios returned 18.43% while the high-liquidity portfolios returned 9.98%.
The more things change the more they remain the same
Is this popularity idea new? No, not at all. Contrarian strategy – going against “the prevailing beliefs in the marketplace” – has been a vogue investing style since at least the 1970s, when David Dreman drew media attention for practicing it. Some say that Benjamin Graham and David Dodd were the original contrarians, in their book on security analysis in 1934. But it’s hard to classify their style as contrarian when their book, the classic security analysis text, now virtually defines “prevailing belief.” Like so many other such strategies it has had its ups and its downs. It would be difficult to conclude definitively whether it beat the market consistently or not. And like many others, Dreman suffered disaster in the 2007-2009 crash and was fired from his investment management job. Ibbotson and Idzorek present the idea that popularity subsumes all the styles that have anticipated top-performing stocks as if it were something new. But it’s not at all.
How does one measure popularity? Ibbotson and Idzorek equate it with high turnover, but that doesn’t necessarily make sense unless they’re defining popularity to mean both popularity and unpopularity combined – that is, merely attention. High-turnover stocks, after all, have to have traders on both the buy and sell sides. How can you tell from its trading volume whether investors like a stock or don’t like it? Ibbotson et al. remark in their 2013 article that “During the recent financial liquidity crises … stock liquidity increased.” Was that because stocks were popular or unpopular?
What about momentum – how does that square with the popularity hypothesis?
Momentum doesn’t shoehorn into Ibbotson and Idzorek’s popularity hypothesis. You could argue that value stocks are unpopular and that small-cap stocks are unpopular, but you can’t argue that momentum stocks are unpopular – quite the reverse, it’s their popularity that causes the momentum. So how do Ibbotson and Idzorek explain the momentum effect in terms of popularity?
They explain it in the obvious way: there is no momentum effect. They say that, “stocks that have performed well in the previous twelve months, appear to do better than those that performed relatively poorly. The momentum premium has been more erratic over more recent periods, with much of the research attempting to understand how and why it seems to have worked. Xiong and Ibbotson [2014] show that stocks that have accelerating prices are more likely to crash and have very poor returns.”2
The latter observation doesn’t need an academic reference. It’s self-evident. If a stock’s price is climbing on sheer momentum alone it is in a bubble and will crash. If that crash will occur at an unpredictable time, momentum can turn into mean-reversion at any instant, and the momentum strategy won’t work. It’s a mystery why this simple observation hasn’t scotched momentum as a factor before. We will come back to that shortly.
What is our microbiologist to think now?
After following these strands of research and doing a little more digging our microbiologist is yet more confused. She discovers that many researchers say the “small-cap effect” doesn’t work anymore – or that its reported outperformance was the result of incorrect measurements. And now the award-winning Financial Analysts Journal article questions momentum. What’s left of these well-researched factors?
She also does a little thinking for herself. She wonders whether once practitioners read this latest article (if they even do read the research) and learn that lower-turnover stocks perform much better, they won’t all run out and buy them. Won’t that make them popular high-turnover stocks? And then won’t they perform poorly?
Also, the three factors -- size, value and momentum -- appear to have become the standard factors said to “explain” all returns. How could another one that’s just as good or better, and so seemingly obvious, pop up so easily?
How can it be that none of these questions are raised in the article?
Are Ibbotson et al.’s findings interesting?
The question is, are Ibbotson et al.’s findings interesting, and does their article deserve to win an award?
My answer is yes to the first question, emphatically no to the second. They have discovered – or at least brought to our attention – an interesting and surprising phenomenon. However, they display startlingly little curiosity as to what caused it. The microbiologist of our little parable would think so too. As a scientist, she wouldn’t discover that a parasite attached itself to a particular host, and decide it was because that host was popular and leave it at that; she would dig deeper. Ibbotson et al., in their paper on the liquidity factor, don’t even break down their results year by year to let us know if there is any consistency over time – though it would have been very easy to do and would satisfy the most superficial curiosity. They only give the 20-year results.
The problem is that they don’t need to do any of that to get published and even to win an award. The field suffers from a subtle corruption. All that an article needs to do is to show that some pattern of investing would have beaten the market, and it will get attention from the media, and it will get attention from those who package and sell investment management – and potentially bring great financial reward to the authors. As enlightened as we may have become about efficient markets, or at least unpredictable markets, and about passive investing, the attention is still focused on beating the market.
A similar corruption afflicts other academic fields, but I don’t think it’s as serious. For example I would assume it is easier for a nutrition researcher to get attention for discovering a correlation of a particular diet with weight loss than for another kind of study, no matter how inexplicable or statistically spurious the correlation may be. Weight-loss diets may even have something in common with market-beating strategies; they might look like they work in correlations, but when put into practice, they don't. I don’t know whether unexplained weight loss findings are prominent in or dominate nutrition research, but I doubt it. In the financial journals however, hard-to-explain findings of market outperformance do have a tendency to dominate.
The pretense of “data-driven”
Furthermore these it-beat-the-market studies get away without much depth of examination because of a pretense that has gripped not just finance but all of economics: that you do the most objective and therefore professional research by standing back and letting the data speak for itself.
The same pretense allowed – indeed, required – ratings agencies to rate very low-quality collateralized debt obligations (CDOs) as being of the highest quality. The data in their extensive database said that default rates for home mortgages were very low, so the ratings agencies, being data-driven – in fact required to be data-driven – used low default rates in their formulas, even if they knew that circumstances had changed and future default rates would be much higher. In an April 27, 2008 article in The New York Times journalist Roger Lowenstein wrote:
“Moody’s did not have access to the individual loan files, much less did it communicate with the borrowers or try to verify the information they provided in their loan applications. ‘We aren’t loan officers,’ Claire Robinson, a 20-year veteran who is in charge of asset-backed finance for Moody’s, told me. ‘Our expertise is as statisticians on an aggregate basis. We want to know, of 1,000 individuals, based on historical performance, what percent will pay their loans?’”
Historical performance was obviously the wrong input. The statisticians of Moody’s were practicing a form of deliberate ignorance. But their data-driven straitjacket left the ratings agencies no choice even if they had wanted it. If they were curious about who these home-mortgage borrowers being packaged into CDOs really were, they had to kill that curiosity and stick to the data. They were getting paid big bucks by the CDO manufacturers.
It is because of this professionally enforced lack of curiosity about underlying mechanisms that a “factor” like momentum can be regarded as meaningful for so long. Formal statistical analyses of the type that financial researchers do automatically, without much thought – almost always simple regressions – have shown that momentum is a “factor” of performance. A little more thought would make one realize that momentum can’t be relied on, as Ibbotson et al. concluded. It would perhaps make a researcher devise a more well-thought-out method of analysis to explore the issue. But since the data is assumed to speak for itself, and the professional researcher is deemed to only inject a too-subjective element by speculating overly about what it means, any deeper analysis is contraindicated.
Unfortunately our microbiologist, too confused to understand what is going on in this strange investment finance field, may decide it speaks a different language and she is not going to understand it. She may wash her hands of it and engage an expensive but incompetent advisor or investment manager or both to do it all for her. She’ll be uncomfortable every time she talks to the advisor because nothing the advisor says makes any sense to her. This is what many investors are living with now. As a field, we have failed them.
Michael Edesess, a mathematician and economist, is a visiting fellow with the Centre for Systems Informatics Engineering at City University of Hong Kong, a principal and chief strategist of Compendium Finance and a research associate at EDHEC-Risk Institute. 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, has just been published by Berrett-Koehler.
Read more articles by Michael Edesess
