pro-gold in commodities. So it really hasn’t been dependent on
market liquidity in a particular asset class. I think the important
thing is whether the next bear market turns out to be of a very different variety than we’ve seen in the past: a one-off, out-of-the-blue
event, where trend-following would not help.
Geoffrey Gerber: In using the term “trend,” are you referring to
a six-month window, a twelve-month window, or shorter-term
trends? Are you referring to a momentum-oriented strategy?
Antti Ilmanen: It definitely is momentum-oriented but in a directional sense. And the windows that are good typically range from
one month to twelve months. Financial markets exhibit positive
trending tendencies in most assets up to twelve months, and thereafter some mean-reversal patterns start to dominate. So investors get
broadly similar results for both long-run returns and tail-risk hedging, whether they rely on three-month or twelve-month windows.
Geoffrey Gerber: In an article you wrote in 2012, you argue
that in general accepting small risks has been well-rewarded,
but taking large risks has been poorly rewarded. Another way
to put this idea is that leveraging up low-volatility opportunities
tends to boost long-term returns. Given this framework, the strategy of buying twelve-month momentum stocks tends to outperform the market, but it also tends to outperform with greater risk.
In other words, taking less risk beats the market, but following the
trend also beats the market with more risk. So how do you equate
these two strategies?
Antti Ilmanen: Let me take a step back and say that there is good
evidence in many asset classes, and actually outside the field of
investing—for example, in racetrack betting—that boring assets
provide better risk-adjusted returns than their speculative peers.
One explanation is the lottery ticket story: We overpay for lottery
tickets and underpay for the boring stuff. Another explanation is
common leverage aversion: We basically overpay for speculative
assets because they give us a big bang for the buck without taking
direct leverage. This low-risk asset outperformance has worked historically in pretty much all asset classes.
But to your question about momentum: If you think of a long–
short portfolio, it’s not quite right that momentum and low risk are
negatively correlated. The high-risk stocks will be present both as
the largest longs and the largest shorts of the momentum portfolio
(and, by the way, you can use risk adjustment to take away that
bias). But in this kind of market-neutral application, there is no
reason for momentum and risk to have the negative link that you
suggested. Even if you think of a long-only portfolio, you should
like an asset that has two helpful characteristics even if these are
negatively related; it just means that you don’t find these opportunities too often. Still, these situations are useful high-conviction
signals. This is true for momentum and value. It also is true for
value and quality. Again, when you see these positive characteristics together, that’s a strong signal to buy.
Edward Baker: I’d like to go back to the issue of liquidity. At the
beginning of your book, you discuss returns from broad asset
classes as related to liquidity, and you use a subjective scale as your
assessment. I’m wondering if you’ve tried to assess liquidity more
systematically and measure it in a way that’s common across asset
classes. Or is that a futile exercise?
Antti Ilmanen: What you’re asking is a bigger task. There are so
many dimensions to liquidity that even if you just think of equities,
a systematic assessment would be hard when only a partial picture is
available. You would have the bid-ask spread angle. Separately, you’d
have to assess the market impact measured in many ways—depth or
volume and resilience of liquidity. Then you’d need to consider the
time dimension or lock-up period for private equity or hedge fund
investments. There’s market liquidity versus funding liquidity and
liquidity as a characteristic versus co-variance with bad times,
meaning that you lose liquidity when crises happen in bad times.
Academics have tried to study all of these different influences, but
packaging them together is difficult. Academics don’t tend to be
interested in issues that are as messy as this multidimensional
As practitioners, we can try, and some have tried. I referred to the
Citibank liquidity index in my book, and there have been other
efforts to develop a scoring system that’s roughly right. But I don’t
think there will ever be a method that achieves consensus about
what is a good liquidity metric, especially when asset classes as different as, say, government bonds and private equity, are compared.
Edward Baker: Measuring liquidity is particularly difficult for over-the-counter and private markets. Are there ways in which you’ve
approached these types of markets?
Antti Ilmanen: I’ve done less work on the private asset side, but
here is why I like illiquid investments less than many investors.
Think of illiquidity premiums in private assets. Basically, most
investors think that if they buy illiquid investments, they should
earn an illiquidity premium. That’s a fair normative statement, but
descriptively the empirical evidence is surprisingly stingy when it
comes to estimates of illiquidity premiums in real estate or private
equity. To get the longest history of illiquidity premiums among
any private assets, you would compare direct U.S. real estate investments versus the real estate investment trust (REIT) market. And if
you study data going back to the 1970s, you’d see that REITs actually had higher average returns than private real estate, so there’s an
inverse illiquidity premium. There have been some refinements to
this type of analysis—for example, adjustments for leverage and
sector composition. But even after these adjustments, there is a
tendency for REITs to have done a bit better, so the historical data
point to an inverse illiquidity premium.
Another important asset class for many investors is private equity.
I wrote in my book that the consensus among academics is that