(perceived) mispricing between two sectors or securities, with extraneous sources of risk
such as general market exposure hedged away. Moreover, because the funds often operate
930
P A R T V I I
Applied Portfolio Management
Statistical Arbitrage
Statistical arbitrage is a version of a market-neutral strategy, but one that merits its own
discussion. It differs from pure arbitrage in that it does not exploit risk-free positions
based on unambiguous mispricing (such as index arbitrage). Instead, it uses quantitative
and often automated trading systems that seek out many temporary and modest misalign-
ments in prices among securities. By taking relatively small positions in many of these
opportunities, the law of averages would make the probability of profiting from the col-
lection of ostensibly positive-value bets very high, ideally almost a “statistical certainty.”
Of course, this strategy presumes that the fund’s modeling techniques can actually identify
reliable, if small, market inefficiencies. The law of averages will work for the fund only if
the expected return is positive!
Statistical arbitrage often involves trading in hundreds of securities a day with holding
periods that can be measured in minutes or less. Such rapid and heavy trading requires
extensive use of quantitative tools such as automated trading and mathematical algorithms
to identify profit opportunities and efficient diversification across positions. These strategies
try to profit from the smallest of perceived mispricing opportunities, and require the fastest
trading technology and the lowest possible trading costs. They would not be possible with-
out the electronic communication networks discussed in Chapter 3.
A particular form of statistical arbitrage is pairs trading, in which stocks are paired up
based on an analysis of either fundamental similarities or market exposures (betas). The gen-
eral approach is to pair up similar companies whose returns are highly correlated but where
one company seems to be priced more aggressively than the other.
2
Market-neutral positions
can be formed by buying the relatively cheap firm and selling the expensive one. Many such
pairs comprise the hedge fund’s overall portfolio. Each pair may have an uncertain outcome,
but with many such matched pairs, the presumption is that the large number of long-short
bets will provide a very high probability of a positive abnormal return. More general versions
of pairs trading allow for positions in clusters of stocks that may be relatively mispriced.
Statistical arbitrage is commonly associated with data mining, which refers to sort-
ing through huge amounts of historical data to uncover systematic patterns in returns that
can be exploited by traders. The risk of data mining, and statistical arbitrage in general,
is that historical relationships may break down when fundamental economic conditions
change or, indeed, that the apparent patterns in the data may be due to pure chance. Enough
analysis applied to enough data is sure to produce apparent patterns that do not reflect real
relationships that can be counted on to persist in the future.
Classify each of the following strategies as directional or nondirectional.
a. The fund buys shares in the India Investment Fund, a closed-end fund that is selling at a discount to
net asset value, and sells the MSCI India Index Swap.
b. The fund buys shares in Petrie Stores and sells Toys “R” Us, which is a major component of Petrie’s
balance sheet.
c. The fund buys shares in Generic Pharmaceuticals betting that it will be acquired at a premium by Pfizer.
CONCEPT CHECK
26.1
2
Rules for deciding relative “aggressiveness” of pricing may vary. In one approach, a computer scans for stocks
whose prices historically have tracked very closely but have recently diverged. If the differential in cumulative
return typically dissipates, the fund will buy the recently underperforming stock and sell the outperforming one.
In other variants, pricing aggressiveness may be determined by evaluating the stocks based on some measure of
price to intrinsic value.
bod61671_ch26_926-950.indd 930
bod61671_ch26_926-950.indd 930
7/25/13 2:04 AM
7/25/13 2:04 AM
Final PDF to printer
C H A P T E R
2 6
Hedge
Funds
931
An important implication of the market-neutral pure play is the notion of portable alpha.
Suppose that you wish to speculate on a stock that you think is underpriced, but you think
that the market is about to fall. Even if you are right about the stock being relatively under-
priced, it still might decline in response to declines in the broad market. You would like
to separate the stock-specific bet from the implicit asset allocation bet on market perfor-
mance that arises because the stock’s beta is positive. The solution is to buy the stock
and eliminate the resultant market exposure by selling enough index futures to drive beta
to zero. This long stock–short futures strategy gives you a pure play or, equivalently, a
market-neutral position on the stock.
More generally, you might wish to separate asset allocation from security selection. The
idea is to invest wherever you can “find alpha.” You would then hedge the systematic risk
of that investment to isolate its alpha from the asset market where it was found. Finally,
you establish exposure to desired market sectors by using passive products such as indexed
mutual funds, ETFs, or index futures. In other words, you have created portable alpha that
can be mixed with an exposure to whatever sector of the market you choose. This proce-
dure is also called alpha transfer, because you transfer alpha from the sector where you
find it to the asset class in which you ultimately establish exposure. Finding alpha requires
skill. By contrast, beta, or market exposure, is a “commodity” that can be supplied cheaply
through index products and offers little value added.
Do'stlaringiz bilan baham: 
