Lessons from Tough Times
Our guest, Eva Xu, provides insight about macro investing in turbulent
markets. Her understanding of this topic evolved during the dramatic rise, fall
and revival of Bayswater Asset Management LLC.
Dr. Xu and Daniel Schuessler founded the firm in May 2004 with $25 million from
Man Global Strategies, part of Man Group. Man’s investment grew to over $900
million and Bayswater’s total assets reached a peak of $1.14 billion in August
Returns were strong and the business appeared to be an exceptional success
story. In May 2007, Man introduced a new capital-guaranteed product that
combined its AHL, Bayswater and four Man Global hedge fund portfolios. The
combination was attractive because Man’s AHL and Bayswater had low correlation
with each other. Bayswater’s longer-term macro approach complemented AHL’s
short- and medium-term trading.
Then came the quant meltdown of 2007, which in retrospect appears to have
presaged worse to come in 2008. Bayswater did not lose as much as some hedge
funds, but nevertheless its loss was in the double digits. The partners returned
money to investors and rethought their system, as Dr. Xu explains.
Bayswater launched a new global macro fund in July 2009, backed by former
Man Group chief Harvey McGrath’s Revere Capital Advisors. Some of the founders
of Revere knew the Bayswater team from the association with Man. They’re taking
an equity stake in the firm.
Prior to starting Bayswater, Ms. Xu and Mr. Schuessler were directors at
Mellon Capital Management, where they worked together for seven years. By the
time they left, she was a director of research and he director of hedge funds.
Subsequently she became vice president at Aetos Capital’s fund of hedge funds
division, responsible for portfolio construction. She has a PhD in economics and
has served as a consultant for the World Bank.
For more about risk control takeaways, see Futures Lab, the next section.
“From a risk/return tradeoff perspective, it is much better if you can
make money from cross-sectional trades with market neutral bets.”
Opalesque Futures Intelligence: How did you get into futures trading?
Eva Xu: My partner, Daniel Schuessler, and I were part of the original
team that developed Mellon’s Global Tactical Asset Allocation business, which I
joined in July 1995 and he joined some months later. We revamped the system
there so as to add countries, currencies and asset classes. This was a big
business, with over around $20 billion in the GTAA fund, managed accounts, hedge
fund, currency overlay and other investments when we left.
OFI: What is Bayswater’s strategy?
EXu: Our strategy is bottom-up fundamental, with relatively long-term,
big-picture, economic fundamentals-driven bets. It is not trend following and
has low correlation to a trend-following program like Man’s AHL, so Man Global
Strategies originally seeded us as a way to diversify their offerings. We trade
global equity indexes, global government bond indexes, currencies and
commodities, with a focus on developed countries.
OFI: Have you changed your strategy in recent years?
EXu: It is the same basic strategy that we previously traded for Man
Global Strategies, but we added a lot of downside risk management in the past
OFI: Does that mean you have become more discretionary?
EXu: Our core strategy remains systematic. For instance, to trade stock
indexes, we collect all earnings information for the companies in the indexes.
That data and country-level inflation and GDP growth forecasts all go into a
valuation model that gives us expected returns at company level. Those are
aggregated at index level and combined with risk and correlation measures for
countries. This systematic process led us to go long Europe and short Japan in
OFI: Why short Japan when there’s a recession in Europe as well?
EXu: Earnings are down in very country because of the recession and
analysts have downgraded earnings prospects across the world. But in Japan the
downgrade was much larger. Our model indicated a great opportunity.
OFI: How do you factor in expectations of future earnings?
EXu: We use analysts’ consensus earnings growth forecasts for next year
and beyond. That is forward looking data and under normal conditions that’s what
the market prices. What’s interesting is when bottom-up and top-down forecasts
diverge significantly, as happened in late 2007. Top-down analysts priced in
recession, but that did not show up in the forecasts of bottom-up analysts,
whose skill set is to look at individual companies, not whole economies.
OFI: What happened then?
EXu: This is the hazard we came up against. Going into the recession,
analysts’ forecasts were behind the curve. If the input going into a model
pretty is bad, so is the output. It leads you to double down and double down
again. After that experience, we added a risk overlay that kicks in when
top-down and bottom-up disagree widely.
OFI: What do you mean by risk overlay?
EXu: The overlay is not an input for the regular model but an indication
that things are going wrong. It indicates the probability of recession and/or
crisis. There are times when analysts miss the boat, so we need another
indicator that will move us to a different model. We don’t expect this to happen
frequently, only once every several years, depending on business cycles. It took
work to construct this indicator because data services’ top-down coverage is
very thin. Top-down forecasts were popular for a while after the 2002 recession
but afterwards people lost interest. Continuous series on top-down consensus
don’t exist and we have to collect the data ourselves.
OFI: Is the overlay itself computer-driven?
EXu: We have not turned into a discretionary manager. When the computer
indicates that the overlay is needed, we switch to a model that can deal with
the different environment. What we may do manually is to clean the data. We have
a lot of experience with that. We run all the models every day at the end of US
market closing. Any data that comes out in the previous 24 hours is collected by
the end of day. A lot of data collection is automated, but you have to make sure
the integrity of data is maintained at all times. We keep in mind the adage,
“Garbage in, Garbage out”!
OFI: What other risk indicators did you add since 2007?
EXu: We use long-term credit measurement as a signal of the equity market
cycle. This is not to be confused with the business cycle— the equity market
cycle leads the business cycle by six to eight months. The data we use –
high-yield spread – is easy to get. In 2009, the spread was dropping like a
stone, which told us to move into equities. Timing is very difficult in
recovery. Our new model performed well. We launched the fund in July and through
Nov were up 11.5%.
OFI: Have you made other significant changes?
EXu: We have made our trades faster. This is something we realized we can
do to improve the return distribution on a daily basis. In cross-sectional
equity trading, for instance, we added short term mean-reversion components to
the valuation-based strategy. That shortened the half-life of cross-sectional
trades. But we’re not a short term trader. Our equity cross-sectional positions
average six to eight months. Directional bets take longer, because they’re tied
to the market cycle.
OFI: Would you give an example of a cross-sectional trade?
EXu: A cross-sectional is a relative bet. For instance, in part of 2009
we were long equity, short bonds. After the 60% stock rally, this bet has run
its course, so we’ve dramatically reduced the position. Once it is off, we may
put on commodities. In most of 2009 we did not trade commodities. A lot of the
time we do not make directional bets, because cross-sectional strategies are
more reliable. They have higher Sharpe ratios and lower risk than directional
bets. From a risk/return tradeoff perspective, it is much better if you can make
money from cross-sectional trades with market neutral bets. We were long the
Australian dollar, the yen and the euro, while being short the US dollar, during
most of 2009. We got out of the AUD briefly because the valuation became too
OFI: Is this a carry trade?
EXu: It is similar to a carry trade but not exactly because we adjust the
picture for expected inflation. For instance, many traders were perplexed by the
yen going up. But that was due to fact that Japan has deflation, so even with
zero nominal rates, it has positive real interest rates! By contrast, the US is
subject to inflation, so it has zero or negative real interest rates. That theme
made money for us in ‘09.
OFI: Did you run into bottlenecks when your AUM went over $1 billion in
EXu: We don’t have the same capacity constraints that can affect CTAs.
We’re not a high-frequency trader and we do not trade single names. Anything we
do is a little drop in the market bucket.
OFI: What do you see in the future?
EXu: There will be downturns again, you have to be prepared. We want to
diversify within our own portfolio and are adding more volatility-loving
components, in particular commodity strategies.