Like stocks that have low price-to-earnings ratios? How about ones that have outpaced the market? Or shares of small companies? Those are known as factors: quantifiable characteristics that some money managers use to identify stocks associated with above-market returns. But factor investing is tricky. Sometimes it pays; other times it doesn’t. Bloomberg Opinion columnists Nir Kaissar and Noah Smith recently met online to debate whether factor investing is worth the effort. They previously discussed corporate debt.
Nir Kaissar: It’s widely acknowledged that some factors have historically outpaced the broad market.
For example, companies that are cheap relative to earnings, cash flow or book value have beaten the market during the past six decades. The same is true of small companies and highly profitable ones.
In a 2017 paper titled “Replicating Anomalies,” economists Kewei Hou, Chen Xue and Lu Zhang identified 67 factors that have produced statistically significant outperformance from 1967 to 2014. In other words, the success of those factors most likely isn’t attributable to chance.
Seeing an opportunity, fund companies have rolled out a dizzying variety of factor funds in recent years. Investors have poured $762 billion into exchange-traded funds that track factor indexes, according to Bloomberg Intelligence. That’s up from $98 billion at the end of 2007.
But the question is whether factor investing will continue to pay. Many investors are skeptical. Returns for value investing, arguably the best-known factor, have lagged the market for more than a decade. Meanwhile, broad market indexes such as the Standard & Poor’s 500 Index, which have no meaningful factor exposure, have been among the best performers.
The answer may depend on why factor investing has been profitable in the first place: Is it compensation for taking additional risk or an opportunity to exploit other investors’ mistakes? It’s a hotly debated question, and it relates not only to factor investing, but to how the markets work more generally.
Noah Smith: I think there are two main questions about factor investing, and you’ve already touched on both.
The first question is what these factors are. Why did things like value, size and momentum show outsized returns for so many decades? Efficient-markets theory says that these outsized returns represent compensation for taking risk — for example, that small stocks sometimes crash even when the market as a whole is not crashing.
As asset manager Cliff Asness has pointed out, that interpretation sort of makes sense for factors like size and value that represent long-term characteristics of companies. But for momentum, it doesn’t really make sense — companies that have high momentum one year often have low momentum the next. It looks like the momentum premium is simply free money, the product of some enduring market inefficiency. This question is important because investors deserve to know whether factor investing is actually increasing their risk, or whether they’re beating the market.
The second question is how long factors persist. You’ve already noted that the value premium has been shrinking over time. But a lot of factors decay even faster. A 2015 paper by economists R. David McLean and Jeffrey Pontiff found that when academics publish a paper about a factor, it tends to shrink or disappear shortly afterward. But a factor tends to hold up between the time they’re discovered and the time the paper is published, implying that the disappearance isn’t a result of publication bias. Instead, this suggests that the market is full of small inefficiencies, which academics and investors are constantly discovering and correcting, and which temporarily manifest themselves as factors.
NK: The two questions are closely related. If factors are compensation for risk, as the efficient-market hypothesis suggests, then they should endure. There will always be investors who don’t want the incremental risk and are happy to let others pursue the premiums, the same way bond investors forgo the equity-risk premium from stocks.
On the other hand, if factors are inefficiencies in the market, or free money, you would expect them to quickly disappear after discovery, unless there are hurdles to capturing them. The factors in McLean and Pontiff’s study fit that description. And the fact that they disappeared suggests they were easily exploited.
But what about the well-known factors that are behind many of the index funds available to investors, such as value, momentum, size and profitability? I find risk-based arguments more compelling than behavioral ones.
For one, those factors have been widely exploited by money managers for decades, and yet they’ve endured. There’s roughly $4 trillion in actively managed U.S. equity mutual funds, many of them chasing one or more of those factors. Granted, active managers have a horrible habit of underperforming, but the fault likely lies more with their fees than the factors.
Factor investing has also been more volatile than the market, as measured by annualized standard deviation, which reinforces intuition about the risks involved. It makes sense that beaten-down value companies and small companies are riskier than the broad market, which is dominated by blue-chip stocks. Profitable companies tend to attract competition and regulation, both of which are bad for future profits. And a 2013 paper by economists Kent Daniel and Tobias Moskowitz found that momentum has a tendency to crash.
All of that should comfort investors who are worried that well-known factors are fleeting.
NS: It seems likely that there are some factors that represent risk and others that represent inefficiencies in the market. In fact, both are probably worth investing in for some managers. If you have the skill to spot an inefficiency early, you can be one of the people to profit from trading away the mispricing (after all, someone has to!). Knowledge of true risk factors, meanwhile, allows a manager to construct a portfolio with a better Sharpe ratio, which is unambiguously good for clients.
Note, by the way, that there’s a third class of factors — inefficiencies that can’t get traded away. The best theoretical example of this that I know of is the famous 1990 paper on noise traders by economists Brad De Long, Andrei Shleifer, Larry Summers and Robert Waldmann. The theory is basically that there occasionally are stampeding herds of irrational investors (noise traders) who buy or sell some stocks for no good reason, and that smart investors and money managers don’t have enough firepower to trade against these herds — so they instead try to ride the mini-bubbles and mini-crashes, which ends up exacerbating the market inefficiency. The noise traders create both risk and return, but their actions make the market less efficient because they aren’t tied to anything fundamental. I think the momentum factor, and the momentum crashes outlined by Daniel and Moskowitz, could come from noise traders.
So there’s a good case to be made for factor investing whether or not factors represent inefficiencies, risk or both. But there’s also a danger. If money managers incorrectly identify risk factors, they’re going to invest badly. People who didn’t realize how much the value premium had decayed would have been putting too much money into value stocks over the past decade.
That brings us to the famous debate over factor timing. Should managers or investors assume that factors are permanent, that they decay or that they move in cycles?
NK: Now we’re into real controversy! Here, too, the question of factor timing is closely related to the one about whether factors are free money or compensation for risk.
Investors should assume that inefficiencies will be quickly corrected, if they can be, but not by them. Few investors have the tools to compete with high-powered traders scouring markets for a free lunch. And because those factors disappear quickly, they’re obviously not permanent or cyclical or subject to decay.
It’s also safer to assume that inefficiencies that can’t be corrected, such as noise traders, involve greater risk, or at least more volatility. That’s been true of most factors that have endured. It should also help investors ward off dubious factor funds, which I suspect will come in greater numbers. If a fund promises premiums without a discernable associated risk, it’s probably too good to be true.
So I would put risk-based factors and inefficiencies that can’t be corrected in the same category, which brings me back to factor timing. The well-known factors have been cyclical. Think about value’s recent experience, for example, which is common. It lagged during the dot-com craze of the late 1990s. It then outperformed in the years leading up to the 2008 financial crisis. And it has lagged ever since.
In hindsight, it’s tempting to think that factor timing is possible. We can draw from our experience around market timing for guidance. We know it’s exceedingly difficult to profit from binary, all-in-all-out moves. There’s also some evidence that tilting away from the market when it’s expensive, and vice versa, can add modest value.
I suspect the same lessons apply to factor timing, but most investors would be better off diversifying their factors and not attempting to time them.
NS: It seems to me that the takeaway for investors is that factor investing can be a good way to boost their performance, but that it comes with several caveats:
1) You need a manager who can differentiate between long-lasting risk factors and short-lived market inefficiencies,
2) You need to understand whether factor investing has made your portfolio riskier, and whether you’re OK with that,
3) You need a manager who knows when factors such as value have decayed in importance over time, and
4) You need to decide whether you want a manager who tries to time the factors.
Factor investing is good, but it’s not simple or easy. You can’t just call up any money manager who does factor investing and give them your money. You have to understand the complexities of the issue, and make sure that your manager does as well. And most importantly, you have to make sure that any extra fees you pay for factor investing are justified by the potential added risk-adjusted returns.
To contact the authors of this story: Nir Kaissar at firstname.lastname@example.orgNoah Smith at email@example.com
To contact the editor responsible for this story: James Greiff at firstname.lastname@example.org
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Nir Kaissar is a Bloomberg Opinion columnist covering the markets. He is the founder of Unison Advisors, an asset management firm. He has worked as a lawyer at Sullivan & Cromwell and a consultant at Ernst & Young.
Noah Smith is a Bloomberg Opinion columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion.
©2018 Bloomberg L.P.