QUANTITATIVE TRADING STRATEGIES

BIRTH OF A NEW DISCIPLINE
The results seen in the above-mentioned studies have led to researchers to investigate
the reasons why some market inefficiencies can withstand over time. The
most popular theories study patterns in human behavior. The tendency of individuals
to move as a crowd and create market bubbles led to new thinking about how
markets operate. Focus began to turn to the behavior of individuals and whether
this behavior, predictable or not, leads to panics and manias in the markets. This
new branch of finance studies psychology and sociology as it applies to financial
markets and financial decisions.
CHAPTER 1 Introduction to Quantitative Trading 21
Behavioral Finance and the Flaw of Human Nature
Behavioral finance, as it is now known, has attracted some of the top minds in academic
finance. It is a combination of classical economics and the principles of behavioral
psychology. This new science has been used as a vehicle to study potential causes of market
anomalies and inefficiencies that inexplicably seem to repeat over time. By studying
how investors systematically make errors in their decision-making process, academics
can explain and traders can exploit the psychological aspect of investing.
Many ideas of behavioral finance were spawned by the work of Amos
Tversky and Daniel Kahneman, psychologists who studied how people made
choices regarding economic benefit. Among the most popular of Kahneman and
Tversky’s discoveries were prospect theory and framing. Prospect theory, as we
will see later in this chapter, deals with the fact that individuals are reluctant to
realize losses and quick to realize gains. Framing deals with how answers can be
influenced by the manner in which a question is posed. One example of framing
from a 1984 study by Kahneman and Tverksy illustrates this point. The pair asked
a representative sample of physicians the following two questions:
Imagine that the United States is preparing for the outbreak of an unusual
Asian disease, which is expected to kill 600 people. Two alternative programs
to combat the disease have been proposed. Assume that the exact
scientific estimates of the consequences of the program are as follows: If
program A is adopted, 200 people will be saved. If program B is adopted,
there is a one-third probability that 600 will be saved and a two-thirds
probability that no people will be saved.
Which of the two programs would you favor?
Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease,
which is expected to kill 600 people. Two alternative programs to
combat the disease have been proposed. Assume that the exact scientific
estimates of the consequences of the program are as follows: If program
C is adopted, 400 people will die. If program D is adopted, there is a onethird
probability that nobody will die and a two-thirds probability that 600
people will die.
Which of the two programs would you favor?
Both these questions present the exact same scenario. In programs A and C,
200 people would live and 400 people would die. In programs B and D, there is a
one-third probability that everyone would live and a two-thirds probability that
everyone would die. Programs A and C lead to exactly the same outcome, as do
programs B and D. The only difference between the first and second question is in
framing. The first question is positively framed, viewing the dilemma in terms of
lives saved. The second question is framed negatively, the results measured in lives
lost. This framing affects how the question is answered.
Kahneman and Tversky discovered that while 72 percent of the physicians
chose the safe and sure strategy A in the first question, 72 percent voted for the
22 PART 1 Structural Foundations for Improving Technical Trading
risky strategy D in the second question. This is illogical, as anyone who picks strategy
A should also pick strategy C, since the stated outcome in both cases are exactly
the same. The experiment shows that how we frame questions can influence the
responses we receive.
Much of Kahneman and Tversky’s work displays a tendency for people to
make inconsistent decisions when it comes to economic decisions. Other economists
have taken the pair’s work and applied its value to the question of market
efficiency. This brings up the logical question: If individuals make inconsistent
decisions, can this lead to inefficient financial markets due to irrationality?
Irrational Decision Makers
While most people believe that all investors must act “rationally” for a market to be
efficient, this is not accurate. Buyers will buy to the point of their perceived fair
value, and sellers will sell down to the point of their perceived fair value. The price
at which an equal number of buyers and sellers meet is the clearing market price.
When positive information is released, investors rationally bid a stock higher on the
revised fair value of business prospects. Even if a handful of investors and traders
act irrationally by buying and selling based on irrelevant information (such as moon
phases or what their pets bark), the market should still be priced efficiently.
Chances are that if one irrational investor is buying, then another is selling.
Market efficiency runs into trouble when the actions of irrational investors
do not cancel out. Consider the situation where irrational investors all pile on and
buy the market at the same time or they all run for the exits at the same time. If all
the irrational investors buy or sell together, they can overwhelm the rational
investors and cause market inefficiencies.
Let’s take, as an example, XYZ Inc., and say it’s trading at $100, with 100
investors following it. The 80 rational investors have decided $100 is the fair value
of the company based on future business prospects. The other 20 buy and sell based
on irrelevant information. As XYZ’s revenues increase, rational investors bid the
stock up to $120—buyers are willing to pay the higher price based on improved
business prospects. Now the 20 irrational investors, all momentum players, begin to
buy the stock due to its performance, driving the stock up to $135. They buy from
rational people who are willing to sell their stock above their perceived fair value.
Ten of the rational investors, who either believe they’re misinterpreting the information
or who feeling pain because they are not long in XYZ, become irrational and
also buy the stock, driving it up to $145. This process can spiral out of control
and create a positive feedback loop, causing unbelievable valuations. All this started
with 20 irrational investors and a small amount of positive news in XYZ Inc.
When irrational investors move together, market irrationality can exist and
take hold for quite some time, eventually leading to bubbles, panics, and crashes.
This theory might explain the technology boom and bust of the late 1990s and
early 2000s. As public investors and day traders craved technology stock exposure,
their thought process shifted from rational methods of valuation to the irrational
CHAPTER 1 Introduction to Quantitative Trading 23
belief that valuation was not important. Like many, I saw my peers making a lot of
money in the market and felt compelled to get on board and not be left behind. My
buying the market had nothing to do with fair valuation or expected business
prospects of the companies I bought. Instead, it was motivated by fear of being the
only one not getting rich in the market. Similar herd mentality was also present
during panics in 1987, 1989, and 1997. Afraid that they would be left holding the
bag when the market made a low, otherwise rational investors can become irrational.
As a result, they sell investments to raise cash, either hoping to miss some
of the decline or, at the very least, to outperform their peers.
An article originally published in the Journal of Finance is credited with
beginning the behavioral finance revolution. In 1986, Werner DeBondt and
Richard Thaler studied the return differences of the best and worst performing
stocks from 1926 through 1982. Stocks with the best three year returns were
assembled into a portfolio of winners, and those with the worst three year
returns were gathered into a portfolio of losers. DeBondt and Thaler noticed
that over one to five years after the portfolios were created, those that contained
previously underperforming stocks significantly outperformed portfolios of
previously outperforming stocks by between 4 and 6 percent per year. This outperformance
occurred whether a narrow (35) or a broader number of stocks
(80+) were chosen for each portfolio. They concluded that investors overreact
to unexpected news events, placing too much emphasis on recent news and
earnings.
Investors begin to expect companies that have consistently beat earnings estimates
to continue to do so in the future. Then, at some point business prospects
slow, the company only meets or even misses estimates, and investors run for the
exit in the stock. Similarly, companies that continually perform worse than expected
are labeled as “terrible” and with no turnaround potential. Eventually, their
business prospects also recover, and investors run to buy the stock. This short-term
thinking among investors can create market inefficiencies.
Robert Shiller of Yale University might be the most widely known behavioral
economist. Shiller’s ground-breaking work on market volatility in the 1980s
redefined how economists look at the stock market. In 1981 he suggested that
market prices were as much as 5 to 13 times too volatile based on their drivers of
value: cashflow. While stock prices should move proportionally to changes in a
company’s expected cashflow (cashflows will eventually be passed on to
investors via dividends), Shiller found that stock prices were more volatile than
what would be predicted by the volatility in underlying dividends. He hypothesized
that the excess volatility could be attributed to investors’ psychological
behavior and the fact that investors overreact to both positive and negative news.
Although some have criticized Shiller’s methods (Schwert, 1991), his arguments
have set off a new wave of thinking about the effect of investor psychology on the
movement of prices.
24 PART 1 Structural Foundations for Improving Technical Trading
Despite the academic research suggesting that inefficiencies do exist in
financial markets, many people question the profitability of quantitative trading.
You might ask: Why should fixed rules generating buy and sell signals ever be
superior to human discretion and the ability to evaluate problems on a situationby-
situation basis? (Indeed, they appear to, as seen in the results cited earlier in the
chapter showing that systematic money managers have outperformed their discretionary
counterparts.) The answer is: Human discretion has a habit of sabotaging
performance. Over the past decade, psychological and financial research have
come together to explain the nature of such emotional tendencies and to shed light
on why some market patterns continue to exist over the years.
Selling Winners and Holding Losers
Studies of human bias in economic situations shed light on how the mind affects
trading decisions. In one such study conducted in 1998, Terrance Odean, professor
of finance at the University of California, examined 10,000 accounts at a large
discount brokerage firm to determine if individuals’ trading styles differed
between winning trades and losing trades made from 1987 and 1993. He found a
significant tendency for investors to sell winning stocks too early and hold losing
stocks too long. Over his test period, investors sold approximately 50 percent more
of paper profits on winning trades than they sold of paper losses in losing trades.
Based on the data, Odean concluded that winning stocks were sold quicker
and more frequently than losing stocks. Although the results were a bit surprising,
this behavior by investors could make sense. When we buy stocks, we’re placing a
bet that a company is undervalued. Stocks that increase in value are logically
becoming less undervalued as they rise, while stocks that decrease in value are
logically becoming more undervalued. Winning stocks that have increased in value
could be considered not as cheap as when they were purchased. Losing stocks that
have declined in value could be considered cheaper than when purchased. In this
case, it makes sense to sell the winning stocks that have become less cheap and
hold losing stocks that have become cheaper.
Although the logic is sound, Odean’s results show that the opposite actually
occurs. Winning stocks that were sold continued to rise, while losing stocks that
were held continued to fall in value. In the year following sales, stocks sold with
gains by individual investors outperformed the market by an average 2.35 percent.
At the same time, losing stocks that were held underperformed the market by an
average of 1.06 percent. Odean discovered, on average, that investors underperform
the market by selling their winners too early and holding on to their losers too long.
Based on purely economic terms, it’s unclear why they behaved in this manner; psychology
may provide the missing link. Clearly, the more profitable course of action
suggested by the study is to buy winning stocks and sell losing ones.
Two theoretical underpinnings dominate the tendency to sell winners and
ride losers: prospect theory and mean reversion theory.
CHAPTER 1 Introduction to Quantitative Trading 25
Prospect Theory. Prospect theory adapts psychologists’ Daniel Kahneman and
Amos Tversky’s theories to financial markets. It suggests that investors are more
risk averse when dealing with profitable investments and more risk seeking in
investments with losses. We all enjoy winning and take pain in losing. As a result,
investors and traders take winners very quickly (to placate our psyche) and hold
on to losers (to hold on to hope that the losers may eventually become winners).
To demonstrate, give people the following choice:
Game 1
75 percent chance of making $1000
25 percent chance of making $0
or
100 percent chance of making $750
We can calculate the expected value of each game by summing the product
of each outcome’s probability by its payout:
Expected payout of risky choice = 75%
$1000 + 25%
$0 = $750
While the expected value of both options is the same in Game 1, individuals
tend to be very risk averse with gains. Most people will take the certain $750 rather
than take the risk for a higher payout. Now consider Game 2, which presents the
exact same choice, only among losses:
Game 2
75 percent chance of losing $1000
25 percent chance of losing $0
or
100 percent chance of losing $750
Expected payout of risky choice = 75%
–$1000 + 25%
$0 = –$750
In Game 2, most people will choose to risk the chance to come out even and
take the first option. While both options have the same expected value, the possibility
of coming out without a loss is often too much to pass up. Basically, the tendencies
revealed in both games show that individuals are risk averse with their
winnings and risk seeking with their losses.
This, of course, jibes with Terrence Odean’s research. From practical
experience, I can add that I’ve often caught myself holding on to losing trades
while thinking, “If I can only get out even on this trade,” or taking it so far as
to calculate breakeven points in hopes of avoiding the disappointment of closing
out a losing trade. If prospect theory is alive and well in the financial markets,
we may be able to take advantage of human nature by designing trading
strategies that are not susceptible to the inconsistent thinking embedded in
human nature.
26 PART 1 Structural Foundations for Improving Technical Trading