One of the central assumptions of the efficient market theory is the – “rational actor”. The market consists of rational individuals who factor in all available information to arrive at the price of a stock. This should mean that the price always reflects the current prospects of that stock/company.
However, things get a little murky when we accept “bounded rationality”. Bounded rationality accepts the limits of our minds. This can be an outcome of heuristics (short-cuts that the mind takes), biases, prejudices or just plain laziness (vast amount of data to be processed)
Thinking fast and slow – by Daniel Kahneman is an absorbing book that describes these biases and heuristics. The book has been written in a very layman –friendly style, allowing me to get some insights into this complex topic.
The book talks about two systems in the mind – system 1 is the intuitive machine that makes snap judgments and is prone to biases. System 1 is always switched on, it acts like the default machinery. System 2 is the more deliberate and slow. It is the more analytical of the two. Using system 2 takes effort and energy (literally). Usually system 2 is called into action only if system 1 thinks there is something unusual and it cannot process it.
Some common features of system 1 are –
- It is uses associative memory to retrieve relevant information while taking a decision,
- it is prone to priming (use of anchors, “liking”, “cognitive ease”, familiarity reduces suspicion)
- It has difficulty with probabilistic thinking
- It is looking for confirmation of beliefs (does not invert)
- It is searching for a coherent story – even with partial information
A story in probabilistic blindness –
Let us assume that we are looking at a company that sells a product X. You are evaluating the option of investing in this company. You have read in the news about a scandal in one of the overseas unit of this company. Later you talked to two of your neighbors who have used the product and they are extremely unhappy.
The news of the scandal evokes strong emotions about the company’s prospects – system 1 scanned its associative memory to pick out this item. Now in a case of intensity matching it tries to calibrate the severity of the scandal to what the future prospects of the company are.
The future prospects of the company are correlated to many factors – Eg: regulations, competition, market size etc. However it is difficult to do a comprehensive analysis of all these factors and then arrive at a weighted average probability. Instead we substituted this difficult question with a much simpler one.
Further in a case of confirmation bias we have found confirming evidence in the neighbors. The law of small numbers dictates that extreme outcomes are more likely in a small sample size. A sample size of 2 is definitely not representative. The use of the scandal and the neighborly opinions as a decision framework is a severe case of availability bias.
Typically after a scandal like this, companies in this sector shut down under-performing units and make the right noises. This creates a temporary effect of the price moving back up. This phenomenon is called regression to the mean. Any implied correlation between the scandal and improved performance would be a mistake. Almost all things tend to display a tendency to regress towards the mean over extended periods.
The complete disregard for the probability of companies in that sector to emerge from an overseas scandal unscathed is called base rate blindness.
A more rational approach would be to start with an understanding of the “base rate” prospects for companies in this sector and your “intuition” for this company. Next would be to understand the factors that correlate to the success of the company like Business economics, regulatory environment, corporate governance and management capability. Give a weightage/ranking to each of these correlated factors.
Now adjust from the base rate for the rankings on all these factors. If the company is an “average” on all factors but is 30% below the average on corporate governance (which gave rise to the corporate scandal). Then you want to adjust the base rate 30% downwards (towards your more intuitive guess).
Prospect theory
Another big idea in this book is called the prospect theory – I want to add a post script because It is so powerful. Expected utility (probability weighted utility) is a function of the “change” you experience. And we are more averse to loss than we desire gains – loss aversion.
So in probabilistic thinking the 2 key variables are the odds of something happening (calculated from base rate and intuition) and the payoff (change from a reference point). Loss aversion dictates that we should look for bets where the odds of a positive outcome are large (stocks with Business/People & Price better than the average) and has a large payoff in the case of a positive outcome (mis-pricing which provides a margin of safety and boosts returns)