We try to distinguish good active strategies from flashes in the pan.
By Laurence B. Siegel1
Laurence B. Siegel is the Gary P. Brinson director of research at the CFA Institute Research Foundation.
Which currently popular investment strategies are flashes in the pan, and which are actually worthy innovations? Which are somewhere in between?
Because the theoretical or academic pedigree of an investment strategy helps mightily to sell it, marketers often represent whatever they’re selling as “real science,” with roots in the work of Nobel Prize-winning financial economists. Some of these claims are justified. Some are almost entirely hype. More often, they’re in between— good investment ideas that, sooner or later, take on the shape of fads and become crowded trades that lose their effectiveness.
Let’s look, through this lens, at some strategies that are receiving attention from investors. Before starting, I’ll state my biases: Indexing is usually a very good way for investors to obtain access to an asset class, because it is hard to pick successful active managers and active strategies— but active management is not useless.2Far from it! Many strategies, especially value strategies, have delivered superior returns over long periods of time. Given the behavioral biases and information asymmetries that we observe almost everywhere, value investing can be expected to continue to perform well on average over time (although not all the time!). The effort to distinguish good active strategies from doubtful ones is well worth one’s while.
Because most strategies start out as plausibly good ideas, at least in a backtest, and because no active strategy can “work” everywhere and always, most of the investment ideas covered in this article are in-betweeners. They are neither exact science nor pure baloney. But, as practiced and marketed, they sometimes lean to one side or the other.
For each strategy, several questions should be asked:
Cap-weighted indexing is the base case to which all active strategies must be compared. For a strategy to be considered active, it is necessary to define what is not active. A number of passive strategies can be imagined, including buy-and-hold-forever-without-rebalancing. But the only passive strategy—in fact, the only strategy of any kind—that everyone can follow without any stocks or bonds or other assets being “left over” is cap-weighted indexing. Cap-weighted indexing is macroconsistent.3
Cap-weighted indexing has been criticized mightily by proponents of the active, non-cap-weighted strategies discussed below. One author said it was worse than Marxism.4 (It’s not.) The primary criticism is that one can do better by observing the fundamental values of securities. Cap-weighted indexing is said to overweight the stocks that are the most overpriced, to bet on momentum or the “greater fool theory,” or to be demonstrably inefficient.
Of course, it’s always possible to find an investment better than the cap-weighted benchmark with the benefit of hindsight. But we do not select securities for future holding periods with that kind of foreknowledge. So, the cap-weighted index, demonstrably efficient when we use neutral return, risk, correlation forecasts that do not embody any special foreknowledge, is the benchmark for evaluating all other (that is, all active) strategies.
But indexing has recently become massively popular, not just as a benchmark but as a portfolio to be held for long-term investment. Is this trend hazardous? Someone must set the prices of securities on the margin, and that “someone” must be the community of active managers. They need to be paid for their work and risk. Are we at the point where there’s so much indexing that no one is setting the prices?
Not nearly. If not enough people are setting security prices, active management should be easy and alphas should be huge.5 The difficulty that both traditional active managers and hedge funds in the 21st century are having in extracting alpha is testimony to the fact that indexing is at least fairly efficient and active management is not easy. (We’d also note that we do not know the optimal amount of effort to devote to alpha production, if the objective is allocating resources to maximize long-term economic growth. This is a worthy topic for future research.)
Cliff Asness argues that index fund investors, free-riding on the price discovery efforts of their active brethren, are taking advantage of capitalism’s greatest gift—the information in prices.6 This is exactly what they should be doing. The ability to use this information without paying for it, Asness concludes, is not a defect of capitalism, but one of its most valuable features. That’s where I come out, too.7
In extremis, the world doesn’t need conventional active managers at all. As Rex Sinquefield pointed out a generation ago, corporations can set the prices of their own securities by deciding whether to issue more stock, buy some back, or leave well enough alone. (They know more about the value of their stock than managers do.)
But don’t worry, this is not about to happen. There is no shortage of active managers or of active management ideas.
Science or baloney?
Cap-weighted indexing is firmly rooted in science.
Fundamental indexing (of equities) is just like cap-weighted indexing except that you don’t use cap weights; you use other data about the sizes of companies as the weights. These other data can be sales, earnings, dividends, number of employees, book value, or a combination of these. What is not used is the market’s own assessment of the stock’s value, in the hope that stocks that are overpriced by the market (relative to their fundamental value) get a smaller allocation than they do in a cap-weighted index.
What an appealing idea for beating benchmarks! We all know that markets are not perfect at pricing securities, and avoiding those that are overpriced, while giving generous weights to those that are underpriced, would seem to be an almost surefire formula for success.
It’s not a bad strategy, but it’s just value investing. As Paul Kaplan of Morningstar shows, a fundamental index is just a yield tilt applied to a cap-weighted index: “The fundamentally weighted index underweights stocks that have lower yields than the market-cap-weighted portfolio and overweights stocks that have higher yields than the market-cap-weighted portfolio.”8
What is good about fundamental indexing, relative to most other value indexes, is that it does not segment the market into two halves (value and growth) and throw away the growth stocks. It reweights them just like it does the value stocks. This practice reflects an appropriate degree of humility about which stocks are growth and which are value—it is impossible to know for sure.
Science or baloney?
Fundamental indexing is science in an elegant marketing wrapper. But the wrapper doesn’t tell the full story that it is just a tilt to value and, like any other value strategy, it won’t work in every period.
In 1972, the great finance professor Fischer Black said that if investors cannot leverage or sell short, the security market line (the line expressing the relationship between beta and expected return) should be flatter than predicted by the capital asset pricing model, or CAPM, so that low-beta stocks will generate a positive alpha.9(The reason is that investors wanting to increase their returns will buy high-beta stocks as a substitute for leveraging up their portfolios, leaving low-beta stocks undersubscribed and underpriced.) Robert Haugen, another professor, made a career out of studying this phenomenon. Many firms now manage low-volatility portfolios, and the low-volatility or low-beta anomaly has produced good results up through the present day.
But why hasn’t it been arbitraged away? Now that hedge funds, exchange-traded-fund managers, and other investors can use leverage and sell short, why don’t investors just hold portfolios that are efficiently balanced between low- and high-volatility stocks, and leverage them up if they want higher risk and higher expected returns?
One reason may be that the leveraged fund industry manages “only” a couple of trillion dollars, tiny compared to the overall size of markets and not enough to bring the prices of low-volatility stocks into parity with the rest of the market. A careful analysis by Ilmanen et al. (2015) shows that as recently as two years ago there was little sign of overpricing or crowding in low-volatility U.S. equities. The limits of arbitrage—the fact that smart investors rarely have enough capital to correct large market mispricings—make it possible for factors that everybody knows about, such as value and low volatility, to deliver positive excess returns for a long time.10
But another possibility is that the anomaly is being arbitraged away; we just can’t see it yet in the data. When Bloomberg Markets calls low-volatility investing a “craze,” it just might be too late.11
Science or baloney?
Low volatility has been a winning factor bet for a very long time. There is a scientific basis for it. Still, with large capital flows into the strategy, there is legitimate concern that it can’t work indefinitely.
Risk parity is just “low volatility” for asset classes. But consider how it’s typically implemented: You borrow short (that is, at short-term interest rates) and lend long (by buying bonds and other long-dated assets). What could possibly go wrong?
Joking aside, risk parity has a bit of a scientific pedigree. With risk parity, each asset class is held in a weight that causes it to make the same contribution to total portfolio risk (standard deviation) as every other asset. The portfolio may then be—but does not have to be—leveraged up or down to the volatility level that the client prefers.
According to Kaplan (2015), the risk-parity portfolio “sit[s]…between the minimum variance portfolio and the equally weighted portfolio in that it is the solution to an optimization program that assigns equal weight to…two diversification measures…[namely] the standard deviation [and the] average log-weight.”12
If you’re still following this, I can simplify by saying that the risk-parity portfolio is optimal under a carefully defined set of conditions.
Here is the analogy to low-volatility investing within an asset class: It works if assets (or asset classes) are mispriced relative to one another because investors can’t or don’t want to use leverage. In the absence of leverage, investors overweight high-risk asset classes to increase their expected return. Investors who can leverage take the other side of this trade through risk parity, which overweights low-risk assets such as bonds, then leverages the whole portfolio.
So, is risk parity a sensible strategy? Yes, for the same reason as low volatility within equities— safer assets have tended to outperform on a risk-adjusted basis—but like low volatility within equities, it has a peso problem.
A peso problem has nothing to do with the feeling you get when you’ve drunk too much tequila, or spent too many pesos. It is the economist’s term for a risk that is present in an asset but that cannot be found in the historical data. It comes from the Mexican peso’s behavior around the time of the 1982 crisis: The yield on peso-denominated assets was very high, but the peso had never crashed, so the high yield appeared to deliver a free lunch. After the peso had provided a high yield for a while, however, it did crash—just as the market had predicted.
With risk parity, the peso problem occurs if long rates spike up and one’s fat position in bonds loses value, or if the yield curve inverts, raising borrowing costs. (Risk parity can perform well when rates rise slowly, because the gain from leveraging higher-yielding assets outweighs the capital loss from the overall rise in yields. However, when yields rise quickly, risk parity can be expected to lose.13) To this “peso” risk one must add leverage, liquidity, and counterparty risks. These are the risks for which risk parity’s apparent return premium is compensation.
Leverage only presents a serious risk if used carelessly. Many well-designed investment products include leverage. But we have heard of risk-parity products that had an unleveraged volatility of 3.4% being leveraged to a volatility target of 15%—that’s between four- and five-to-one, a disaster waiting to happen. Careful risk management is essential to the success of risk parity.
Science or baloney?
Using leverage when other investors cannot is science, but it’s risky unless risks are managed very carefully. Do it if you’re confident that bond yields won’t spike up during your intended holding period. However, claims that risk parity portfolios are structurally or inherently better than other portfolios are mostly baloney. You also need to decide whether the use of leverage presents a risk that you cannot tolerate.
Factor Investing and Smart Beta
Quant is dead, long live the quant.
When quantitative investing—as then practiced— hit a very rough patch starting around 2007, quants didn’t die; they became factor investors, assembling factor-based index funds and ETFs (fundamental indexing is one) and marketing them, in some firms, as “smart beta.” The factors include the usual suspects—size, value, and momentum— as well as some new ones, such as “carry” (yield), quality, and low volatility.
Old fogeys will remember when yield was the only factor anyone knew about—as John Burr Williams said in his 1938 attempt at writing poetry, “A cow for her milk / A hen for her eggs / And a stock, by heck, for her dividends.” Everything old is new again.
As practiced by the more naïve managers, factor investing is chasing your tail, betting on past factor performance; it is data mining. But if you can make at least partially accurate forecasts of factor returns, it’s much better. At bottom, factor investing is just a claim that you can add alpha, but at the factor level rather than the security level. Stephen Sexauer, the CIO of the San Diego County pension fund, said that most quantitative active strategies can be thought of as a highly intelligent “bot” wandering through the securities markets.14There is also a mother bot, or perhaps a committee of human beings, monitoring the bot to see if it is doing anything useful. The criteria for usefulness are whether it’s making a profit and whether its success is attracting a swarm of competing bots big enough to take the profits away. If the latter, or if it’s not producing a profit, the bot gets retired. This isn’t a bad way to try to make money; it just has its limits. Managers— being human—tend to pursue each quant active opportunity identified by the bot well past its limits, as we saw from the “quant crash” of August 2007.
Factors and “quant” investing, then, are just active management, subject to Barton Waring’s Two Conditions needed for justifying active management: 1) You must believe that managers skillful at adding alpha exist, in the future and not just in past performance data; and 2) you must, yourself, have the skill to select them from among a population of managers that underperforms on average.15
Science or baloney?
Factor investing is science if there is fair reason to believe the manager can identify factors with a positive expected alpha—until the factor becomes a crowded trade and the expected excess return flips to negative. Then, unless you change your factor bets, it’s baloney. It’s hard to see how any factor can work forever.
Unconstrained Investing: Hedge Funds, Absolute Return, Market Neutral, Long/Short, and Portable Alpha
Thomas Idzorek (2014) beautifully characterizes these strategies as ways to “enable managers to maximize the benefit of their skill.” He also discusses the differences among them in some detail, so I don’t have to. Please refer to Idzorek (2014).
Most of these strategies increase the impact of manager skill by relaxing the long-only constraint and the no-leverage constraint (where security positions must add to 100% and no more). In the case of portable alpha, the convention is relaxed that says alpha must be delivered to the investor in the asset class in which it was produced by the manager. Note that these strategies aren’t fully unconstrained, but the relaxing of traditional constraints is the distinguishing feature that they have in common.
The catch, as always, is that you must have skill in the first place. Managers who are lousy at alpha production in traditional (long-only, unleveraged) investing are unlikely to become better at it when the constraints are removed. Given the small number of traditional active managers who consistently beat their benchmarks over time, too many nontraditional active managers claim that they’ll do so now that they’ve been liberated from the long-only constraint. After all, they’re the same people. Removing a constraint does not make you smarter.
But constraints do reduce the potential for earning return. So, why are they there? For risk control. They protect the investor from the manager’s mistakes, some of which, as enough time accumulates, are likely to be big enough to destroy a completely unconstrained portfolio.
If investors didn’t care about risk, unconstrained investing would be the bee’s knees. But, as Harry Markowitz said more than 60 years ago, “That afternoon in the library, I was struck by the notion that you should be interested in risk as well as return.”16 Thank goodness for Markowitz’s afternoon in the library, and for constraints. Science or baloney?
If you truly can add alpha after transaction costs and fees, in the future when you are managing real money and not just in a backtest, then unconstrained investing is science. If your active process is really just a random number generator, then it’s baloney. But the fees can be spectacular…for the manager.
The Endowment Model
In the 1990s, the great investment managers Jack Meyer at the Harvard Management Company and David Swensen at the Yale University investment office became famous for earning very high returns for their institutions. Their secret was large allocations to unconventional investments such as venture capital, private equity (buyout) funds, hedge funds, timber, and commodities. In Meyer’s case, these were supplemented by aggressive internal management of conventional asset classes, conducted by highly paid specialists.
The word quickly got out that top universities had discovered a new model of investing, focused on alternatives. Many smaller endowments and foundations, and some pension funds, tried to copy the Harvard and Yale models. This first met with some success, but later—especially after the crash of 2008—these institutions found that the new strategies consumed a lot of liquidity (due to capital calls from venture and private equity firms), required hiring expensive and finicky staffs, and sometimes lost dramatic amounts of money.
By June 2008, about the peak of the alternatives craze, 60% of the total assets of a group of eight top universities were invested in hedge funds and nonmarketable securities (mostly private equity). Including expected capital calls as well as hedge funds and nonmarketable securities, the number rises to a stunning 91%.
This could not end well. As I wrote in 2013, “We don’t have data on exactly how these endowment funds performed in the crash or how hard they had to scramble for liquidity in 2009. The anecdotal evidence is that the results weren’t pretty. A number of universities had to go to the capital markets to borrow, while others made drastic adjustments to spending. Liquidity is more than just a theoretical concern!”17
Crashes and liquidity crises aside, we can’t all earn huge alphas at the expense of each other. If the endowment model is implemented successfully on a large scale by many investors, that means there must also be a large population of investors willing to earn negative alphas relative to the world’s overall capital market benchmark. In this age of accurate performance measurement and hypercompetitive investment organizations, it is more of a stretch than it used to be to assume that such a population of willing losers will always be there. For all but the most skilled investors, endowment-model investing may be a good idea whose time has passed.
Science or baloney?
The endowment model is science if you’re Jack Meyer or David Swensen; otherwise, beware. (Just ask Harvard after the past few years of underperformance.) It also helps to have friendly markets, with low prices and glaring inefficiencies. If you can start 25 years ago, you’ll be way ahead.
Distinguishing Science From Baloney
The investment management industry has made a great many managers rich beyond imagining. Many of them have earned their riches, but investors would be well served to understand better the difference between science and baloney in investment management. Science, applied to markets to generate superior returns, is worth its weight in gold. Baloney is worth its weight in baloney.
By distinguishing science from baloney, investors can deal better with the vast asymmetry between the interests of the manager, who needs to get it right once to become very rich, and those of the investor, who needs sustained and repeatable successes to build wealth over time.18Where are the customers’ yachts? Laurence B. Siegel is the Gary P. Brinson director of research at the CFA Institute Research Foundation
Asness, Clifford S. 2015. “How Can a Strategy Still Work if Everyone Knows About It?” Cliff’s Perspective. Aug. 31, AQR Capital Management.
Bernstein, Peter L. 1993. Capital Ideas: The Improbable Origins of Modern Wall Street. New York: Simon & Schuster.
Black, Fischer. 1972. “Capital Market Equilibrium with Restricted Borrowing.” Journal of Business, Vol. 45, No. 3 (July), pp. 444–454.
Fraser-Jenkins, Inigo, et al. “The Silent Road to Serfdom: Why Passive Investing is Worse Than Marxism,” Sanford C.
Bernstein & Co., LLC. Privately circulated.
Hurst, Brian, Michael Mendelson, and Yao Hua Ooi. 2013.
“Can Risk Parity Outperform If Yields Rise?” July, AQR Capital Management.
Idzorek, Thomas. 2014. “Caveat Emptor!” Morningstar magazine (April/May), pp. 4.
Ilmanen, Antti, Nielsen, Lars N., and Chandra, Swati. 2015.
“Are Defensive Stocks Expensive? A Closer Look at Value Spreads” (November), AQR Capital Management.
Kaplan, Paul D. 2008. “Why Fundamental Indexation Might—or Might Not—Work.” Financial Analysts Journal, Vol. 64, Issue 1 (January/February).
Kaplan, Paul D. 2015. “Which Portfolio Is the Most Diversified?” Morningstar magazine (June/July), pp. 64–66. Maillard, Sébastien, Thierry Roncalli, and Jérôme Teiletche.
2010. “The Properties of Equally Weighted Risk Contribution Portfolios,” Journal of Portfolio Management (Summer).
Sharpe, William F. 1991. “The Arithmetic of Active Management.” Financial Analysts’ Journal, Vol. 47, No. 1 (January/February), pp. 7–9.
Siegel, Laurence B. 2003. Benchmarks and Investment Management. Charlottesville, VA: CFA Institute Research Foundation.
Siegel, Laurence B. 2013. “The ‘New Finance’: Illiquidity, the Liquidity Premium, and Liquidity-Preserving Strategies.”
Ounavarra Review (now OCP Capital Review) (Winter).
Privately circulated; available from the author.
Siegel, Laurence B., and M. Barton Waring. “TIPS, the Dual Duration, and the Pension Plan.” Financial Analysts Journal, Vol. 60, Issue 5 (September/October).
Waring, M. Barton. 2004. “Liability-Relative Investing.” Journal of Portfolio Management, Vol. 30, No. 4 (Summer).
Waring, M. Barton. 2011. Pension Finance: Putting the Risks and Costs of Defined Benefit Plans Back Under Your Control. Hoboken, N.J.: John Wiley & Sons.
Waring, M. Barton, Duane Whitney, John Pirone, and Charles Castille. 2000. “Optimizing Manager Structure and Budgeting Manager Risk.” Journal of Portfolio Management, Vol. 26, No. 3 (Spring), pp. 90–104.
Williams, John Burr. 1938. The Theory of Investment Value. Cambridge, MA: Harvard University Press.