Positional Option Trading (Wiley Trading)

  augments, not replaces, traditional economics.

  What have we learned from behavioral finance?

  First, behavioral finance has added to our understanding of

  market dynamics. Even in the presence of rational traders and

  arbitrageurs, irrational “noise” traders will prevent efficiency. And

  although it is possible to justify the existence of bubbles and

  crashes within a rational expectations framework (for example,

  40

  Diba and Grossman, 1988), a behavioral approach gives more

  reasonable explanations (for example, Abreu and Brunnermeier,

  2003, and De Grauwe and Grimaldi, 2004).

  Second, we are now aware of a number of biases, systematic

  misjudgments that investors make. Examples include the

  following:

  Overconfidence: Overconfidence is an unreasonable belief

  in one's abilities. This leads traders to assign too narrow a

  range of possibilities to the outcome of an event, to

  underestimate the chances of being wrong, to trade too large,

  and to be too slow to adapt.

  Overoptimism: Overconfidence compresses the range of

  predictions. Overoptimism biases the range, so traders

  consistently predict more and better opportunities than really

  exist.

  Availability heuristic: We base our decisions on the most

  memorable data even if it is atypical. This is one reason teeny

  options are overpriced. It is easy to remember the dramatic

  events that caused them to pay off, but hard to remember the

  times when nothing happened and they expired worthless.

  Short-term thinking: This thinking shows the irrational

  preference for short-term gains at the expense of long-term

  performance.

  Loss aversion: Investors dislike losses more than they like

  gains. This means they hold losing positions, hoping for a

  rebound even when their forecast has been proven wrong.

  Conservatism: Conservatism is being too slow to update

  forecasts to reflect new information.

  Self-attribution bias: This bias results from attributing

  success to skill and failure to luck. This makes Bayesian

  updating of knowledge impossible.

  Anchoring: Anchoring occurs when relying too much on an

  initial piece of information (the “anchor”) when making a

  forecast. This leads traders to update price forecasts too slowly

  because the current price is the anchor and seems more

  “correct” than it should.

  And there are at least 50 others.

  41

  It is these types of biases that traders have tried to use to find trades with edge. Results have been mixed. There are so many

  biases that practically anything can be explained by one of them.

  And sometimes there are biases that are in direct conflict. For

  example, investors underreact, but they also overreact. Between

  these two biases you should be able to explain almost any market

  phenomena. The psychologists and finance theorists working in

  the field are not stupid. They are aware of these types of

  difficulties and are working to disentangle the various effects. The

  field is a relatively new one and it is unfair and unrealistic to

  expect there to be no unresolved issues. The problem is not really

  with the field or the serious academic papers. The problem is with

  pop psychology interpretations and investors doing “bias mining”

  to justify ideas.

  It is common in science for a new idea to be overly hyped,

  particularly those that are interesting to lay people (traditional

  finance is not interesting). In the 1970s there were popular books

  about catastrophe theory, a branch of physics that was meant to

  explain all abrupt state changes and phase transitions. It didn't. In

  the 1990s, chaos theory was meant to explain practically

  everything, including market dynamics. It didn't. Behavioral

  finance is being overexposed because it is interesting. It provides

  plenty of counterintuitive stories and also a large amount of

  schadenfreude. We can either feel superior to others making

  stupid mistakes or at least feel glad that we aren't the only ones

  who make these errors.

  And people love intuitive explanations. We have a great need to

  understand things, and behavioral finance gives far neater

  answers than statistics of classical finance theory. Even though

  behavioral finance doesn't yet have a coherent theory of markets,

  the individual stories give some insight. They help to demystify.

  This is reassuring. It gives us a sense of control over our

  investments.

  A science becoming interesting to the general public doesn't

  necessarily mean it is flawed. For example, there have been

  hundreds of popular books on quantum mechanics. However,

  behavioral finance does have some fairly serious problems to

  address.

  Just as in conventional finance theory, behavioral finance studies

  individual decision-making despite the fact that people do not

  42

  make investing decisions independently of the rest of society.

  Everyone is influenced by outside factors. Most people choose

  investments based on the recommendations of friends (Katona,

  1975). And professionals are also influenced by social forces

  (Beunza and Stark, 2012). Over the last 30 years the sociology of

  markets has been an active research field (for example, Katona,

  1975, Fligstein and Dauter, 2007, and references therein), but this

  work hasn't yet been integrated into behavioral finance. Because

  behavioral finance largely ignores the social aspects of trading and

  investing, we don't have any idea of how the individual biases

  aggregate and their net effect on market dynamics. This is

  necessary because, even though we don't understand how

  aggregate behavior emerges, it is very clear that markets cater to

  irrational behavior rather than eradicate it. For example, the

  services of financial advisors, stock brokers, and other financial

  intermediaries made up 9% of the US GDP (Philippon, 2012)

  despite the fact that they are almost all outperformed by much

  cheaper index funds and ETFs.

  Next, behavioral finance has largely limited itself to the study of

  cognitive errors. There are many other types of nonrational

  behavioral inputs into decision-making, including emotion,

  testosterone levels, substance abuse, and the quest for status.

  And behavioral finance gives no coherent alternative theory to the

  EMH. A catalog of biases and heuristics—the mistakes people

  make—is not a theory. A list of facts does not make a theory. Of

  course, sometimes observations are necessary before a theory can

  be formulated. Mendeleev drew the periodic table well before the

  atomic structure of matter was understood. We knew species

  existed well before we understood the process of speciation by

  natural selection. Still, to be scientific, behavioral finance

  eventually needs to lead to a unifying theory that gives

  explanations of the current observations and makes testable

  predictions.

  Behavioral finance can still help. Whenever we find somethi
ng

  that looks like a good trading idea we need to ask, “Why is this

  trade available to me?” Sometimes the answer is obvious. Market-

  makers get a first look in exchange for providing liquidity. Latency

  arbitrage is available to those who make the necessary

  investments in technology. ETF arbitrage is available to those with

  the capital and legal status to become authorized participants. But

  43

  often a trade with positive edge is available to anyone who is

  interested. Remembering the joke about the economists, “Why is

  this money sitting on the ground?” Risk premia can often be

  identified by looking at historical data, but behavioral finance can

  help to identify real inefficiencies. For example, post-earnings

  announcement drift can be explained in terms of investor

  underreaction. Together with historical data, this gives me enough

  confidence to believe that the edge is real. The data suggest the

  trade, but the psychological reason gives a theoretical justification.

  High-Level Approaches: Technical Analysis

  and Fundamental Analysis

  Technical analysis is the study of price and volume to predict

  returns.

  Technical Analysis

  Aronson (2007) categorized technical analysis as either subjective

  or objective. It is a useful distinction.

  Subjective technical analysis incorporates the trader's discretion

  and interpretation of the data. For example, “If the price is over

  the EWMA, I might get long. It depends on a lot of other things.”

  These methods aren't wrong. They aren't even methods.

  Subjectivity isn't necessarily a problem in science. A researcher

  subjectively chooses what to study and then subjectively chooses

  the methods that make sense. But if subjectivity is applied as part

  of the trading approach, rather than the research, then there is no

  way to test what works and what doesn't. Do some traders succeed

  with subjective methods? Obviously. But until we also know how

  many fail, we can't tell if the approach works. Further, the

  decisions different traders who use ostensibly the same method

  make won't be the same or even based on the same inputs. There

  is literally no way to test subjective analysis.

  Some things that are intrinsically subjective are Japanese

  candlesticks, Elliot waves, Gann angles, trend lines, and patterns

  (flags, pennant, head, and shoulders, etc.). These aren't methods.

  In the most charitable interpretation, they are a framework for

  (literally) looking at the market. It is possible that using these

  methods can help the trader implicitly learn to predict the market.

  But more realistically, subjective technical analysis is almost

  44

  certainly garbage. I can't prove the ideas don't work. No-one can.

  They are unfalsifiable because they aren't clearly defined. But

  plenty of circumstantial evidence exists that this analysis is

  worthless. None of the large trading firms or banks has desks

  devoted to this stuff. They have operations based on stat arb, risk

  arb, market-making, spreading, yield curve trading, and volatility.

  No reputable, large firm has a Japanese candlestick group.

  As an ex-boss of mine once said, “That isn't analysis. That is

  guessing.”

  Any method can be applied subjectively, but only some can be

  applied objectively. Aronson (2007) defines objective technical

  analysis as “well-defined repeatable procedures that issue

  unambiguous signals.” These signals can then be tested against

  historical data and have their efficacy measured. This is essentially

  quantitative analysis.

  It seems likely that some of these approaches can be used to make

  money in stocks and futures. But each individual signal will be

  very weak and to make any consistent money various signals will

  need to be combined. This is the basis of statistical arbitrage. This

  is not within the scope of this book.

  However, we do need to be aware of a bad classic mistake when

  doing quantitative analysis of price or return data: data mining.

  This is where we sift through data using many methods,

  parameters, and timescales. This is almost certain to lead to some

  strategy that has in-sample profitability. When this issue is

  confined to choosing the parameters of a single, given strategy it is

  usually called overfitting. If you add enough variables, you can get

  a polynomial to fit data arbitrarily well. Even if you choose a

  function or strategy in advance, by “optimizing” the variables you

  will the get the best in-sample fit. It is unlikely to be the best out of

  sample. Enrico Fermini shared that the mathematician and

  economist John von Neumann said, “With four parameters I can

  fit an elephant , and with five I can make him wiggle his trunk”

  (Dyson, 2004).

  This mistake isn't only made by traders. Academics also fall into

  the trap. The first published report of this was Ioannidis (2005).

  Subsequently, Harvey et al. (2016) and Hou et al. (2017) discussed

  the impact of data mining on the study of financial anomalies.

  There are a few ways to avoid this trap:

  45

  The best performer out of a sample of back-tested rules will be

  positively biased. Even if the underlying premise is correct, the

  future performance of the rule will be worse than the in-

  sample results.

  The size of this bias decreases with larger in-sample data sets.

  The larger the number of rules (including parameters), the

  higher the bias.

  Test the best rule on out-of-sample data. This gives a better

  idea of its true performance.

  The ideal situation is when there is a large data set and few

  tested rules.

  Even after applying these rules, it is prudent to apply a bias

  correcting method.

  The simplest is Bonferroni's correction. This scales any statistical

  significance number by dividing by the number of rules tested. So,

  if your test for significance at the 95% confidence level (5%

  rejection) shows the best rule is significant, but the rule is the best

  performer of 100 rules, the adjusted rejection level would be

  5%/100 or 0.005%. So, in this case, a t-score of 2 for the best rule

  doesn't indicate a 95% confidence level. We would need a t-score

  of 2.916, corresponding to a 99.5% level for the single rule. This

  test is simple but not powerful. It will be overly conservative and

  skeptical of good rules. When used for developing trading

  strategies this is a strength.

  A more advanced test is White's reality check (WRC). This is a

  bootstrapping method that produces the appropriate sampling

  distribution for testing the significance of the best strategy. The

  test has been patented and commercial software packages that

  implement the test can be bought. However, the basic algorithm

  can be illustrated with a simple example.

  We have two strategies, A and B, which produce daily returns of

  2% and 1% respectively. Each was developed by looking at 100

  historical returns. We can use WRC to determine
if the apparent

  outperformance of strategy A is due to data mining:

  Using sampling with replacement, generate a series of 100

  returns from the historical data.

  46

  Apply the strategies (A and B) to this ahistorical data to get the

  pseudo-strategies A' and B'.

  Subtract the mean return of A from A' and B from B'.

  Calculate the average return of the return-adjusted strategies,

  A” and B”.

  The larger of the returns of A” and B” is the first data point of

  our sample distribution.

  Repeat the process N times to generate a complete

  distribution. This is the sampling distribution of the statistic,

  maximum average return of the two rules with an expected

  return of zero.

  The p-value (probability of our best rule being truly the better

  of the two) is the proportion of the sampling distribution

  whose values exceed the returns of A, that is, 2%.

  A realistic situation would involve comparing many rules. It is

  probably worth paying for the software.

  There is also a totally different and complementary way to avoid

  overfitting. Forget about the time series of the data and study the

  underlying phenomenon. A hunter doesn't much care about the

  biochemistry of a duck, but she will know a lot about their actual

  behavior. In this regard a trader is a hunter, rather than a

  scientist. Forget about whether volatility follows a GARCH(1,1) or

  a T-GARCH(1,2) process; the important observation is that it

  clusters in the short term and mean reverts in the long term. If the

  phenomenon is strong enough to trade, it shouldn't be crucial

  what exact model is used. Some will always be better in a sample,

  but that is no guarantee that they will work best out of a sample.

  As an example, this is the correct way to find a trading strategy.

  There is overwhelming evidence that stocks have momentum.

  Stocks that have outperformed tend to continue outperforming.

  This has been observed for as long as we have data (see Geczy and

  Samnov [2016], Lempérière et al. [2014], and Chabot et al.

  [2009]) and in many countries (for example, Fama and French,

  2010). The observation is robust with respect to how momentum

  is defined and the time scales over which it is measured. In the

  trading world, the evidence for stock momentum is overwhelming.