dividend yield strategies are closely correlated with some value strategies based on, say, the book value-to-price ratio. Also in corporate
credits, carry and value can be similar, unless you create a fast-moving fair-value model. Even when you consider government bonds,
curve steepness and real yield are somewhat positively correlated,
but with currencies, purchasing power parity and carry strategies are
quite uncorrelated. The same is true for commodities, if you consider
some value or reversal signals versus carry/roll signals.
So yes, an overlap exists between carry and value, and I agree you
shouldn’t double-count it. But I think carry is worth keeping as
part of the menu of most important strategies—partly because anywhere it’s been studied, the evidence shows it has worked. If you
can find ways of investing carry in a way that’s somewhat uncorrelated with value, that’s a useful thing to do.
Edward Baker: In your world, it seems that a dynamic allocation
approach is important to successfully managing the desired factor
Antti Ilmanen: Yes, you dynamically allocate across assets if an
asset moves from one style group to another. But my core belief is
in strategic diversification across long-run rewarded factors I
believe in. Such factors include a few asset class risk premiums,
some style risk premiums with the most persistent and pervasive
supporting evidence (such as value, momentum, carry, and defensive styles), and illiquidity risk premiums, also through private
investments—these are all good sources of long-run return. I say
invest strategically in these factors and execute cost-effectively.
Then try to stick with your decisions rather than attempting to
tactically time these things.
The nuance here is that the above styles require you to move your
actual holdings, but strategic allocation into the styles is a good
idea rather than aggressively shifting the allocation, for example,
based on recent valuations. At AQR we’ve studied these things a
lot, and we find that it’s difficult to do better than strategic holdings
in these styles, say, by doing some contrarian style rotation.
I have changed my tune about this approach to some extent since
my book was published. At that point, I had more hope for various
contrarian strategies such as style timing or market timing. But our
recent work on these matters has shown surprisingly disappointing
performance whenever those contrarian ideas are translated into
actual trading rules. Style timing hasn’t worked, and with equity
timing the message is pretty much that if you had used Shiller PEs6
for timing for the past fifty years, you wouldn’t have beat the buy-
and-hold strategy; you would have done a little bit worse.
Geoffrey Gerber: So your suggestion is to allocate strategically
among the various risk premiums, and the only turnover in the
portfolio would be in the individual stocks that get you those
Antti Ilmanen: Yes, but doing these things not only in stocks. The
beautiful historical result is that the four style premia I mentioned
(value, momentum, carry, and defensive) have provided long-run
tailwinds pretty much in any asset class we’ve looked at, and we’ve
looked far and wide in equities, bonds, currencies, and commodities. Using the same ideas in many different places allows you to
diversify even better. We like to have consistent positive allocations
because these styles have yielded strategic, long-run edges. Let’s
take advantage of that.
There are counter stories about investors who got interested in
one particular style at some point in the past and then after some
good years, the inevitable bad years occurred, and after two or
three bad years, they decided to bail out. I agree that a strategy
works poorly if investors can’t stay with it. So you have to come up
with a solution that helps them stick with the strategy they chose.
Packaging these strategies together, diversifying among several
styles, and applying them in many different asset classes is very
important for a real-world approach that can help investors avoid
the bad habit of chasing multi-year returns up and down.
Edward Baker: But clearly there are some allocation tools that can
be used. Within currencies, for example, we know that carry works
inversely to volatility. So when volatility is increasing, investors
benefit from moving out of carry and into other strategies that
benefit from volatility, such as trend-following.
Antti Ilmanen: Let me tell some old stories here. After the 1998
Long-Term Capital Management crisis, I started working on a
carry-timing model. 7 Over the years, I was always adjusting things
a little bit based on the last crisis, and overall, the experience wasn’t
particularly happy. Several false alarms in the mid-2000s prompted
me to step out of the carry strategy just after its fall and before a
recovery. This is when I was doing hedge fund trading at Brevan
Howard. Then in the summer of 2007, a lasting signal occurred,
indicating this would have been a good time to be out of carry for
the next year and a half. Unfortunately, we were told to cut our
systematic strategies at that point. So it often happens that false
alarms make you lose faith.
Since 2009, we’ve been in a risk-on–risk-off world in which we’ve
had pretty short-lived risk-off environments. So carry timing
“Packaging these strategies together, diversifying among several styles,
and applying them in many different
asset classes is very important for a
real-world approach that can help
investors avoid the bad habit of chasing
multi-year returns up and down. ”