Moderated by : Madhu Veeraraghavan
Contributed by: Rajni Dhameja, CFA
The 10th India Investment conference hosted one session on Taming the Uncertainty by Ralph Hertwig. He is a director of the Centre of Adaptive Rationality at the Max Planck Institute for Human Development at Berlin.
The contours of the session were designed around how the uncertainty is different from risk and how in real life we deal with uncertainty more often than not. The hypothesis was that considering the fact that people are not fully rational as assumed by financial models, can we invest more time and effort to understand how people take decisions and train the models accordingly. Hence it is an attempt to consider the approach of “moving from people to models as against the existing approach of models to people“.
Uncertainty is a human condition. It is a space where outcomes of a particular event are either not known or incompletely known hence the surprises can happen. For instance climate change in an uncertain situation rather than risky situation because full impact of climate change is unknown to a large extent. Following are few real life examples depicting uncertainty and the way people deal with it:
- Miracle of Hudson river in 2009, wherein plane was stuck above the water due to some issue in the engine. Between the choices of going back to the starting point and landing over the water, pilot made the less popular choice of landing over the water. In hindsight, pilot explained that for this decision pilot relied only on one piece of information i.e. direction of movement of windshield which indicated that if pilot attempts to land he will be able to do so.
- Decision making by cab drivers in the city of New York (approximately 6000 times in year) as to who they can let inside the cab so that they do not let any unwanted person inside the cab. Their decision making is based on simple heuristics of preference towards women over men, individuals over group, older people over younger ones etc.
- 80% of decision making by firefighters is in less than 1 minute
The above three completely different examples of uncertain situations point out to one common thing i.e. way of decision making in these situations. Decision maker’s reliance on simple heuristics rather than complex algorithms.
Let us understand the difference between risk and uncertainty:
Risk is perfect knowledge of possible outcomes of a space along with the associated probabilities of occurrence of those outcomes whereas, uncertainty is where the all outcomes are either not known or incompletely known hence the surprises can happen.
In real life lot of situations are such which are uncertain in nature rather than being risky. Mervyn King, former governor of Bank of England once mentioned that : “If only banks were playing in a casino, then we probably could calculate approximate risk weights”.
Further, the outcomes can be known in the problems which are small world problems hence they are risky nature and can be solved with Bayesian principles. eg: outcome of a lottery ticket. But if you consider the uncertain situation like planning a picnic, it cannot be solved through Bayesian principle as the outcomes are unknown.
In real world people deal with uncertainty in a manner which is completely different than the rational way as defined in neo classical economic theory. The neo classical economic theory specifies the problem solving approach as defining the objective, understand the attributes of decision making, define the alternatives and associated probabilities and choose the outcomes. But, real life decision making is far from this in most of the situations. The real life problem solving is through bounded rationality as against the idealistic Olympian rationality of neo classical theory.
The moot point over here is, in certain situations which are uncertain in nature heuristics leads to better decision making than the decisions taken through fully rational methods.
A research program was conducted in simple heuristics to develop a framework which helps to improve the quality of decisions. The framework is as follows:
- Adaptive toolbox: What are the heuristics available which people use
- Ecological Rationality: It is important to understand, in which environment do heuristics succeed. The argument is not that heuristics are always better than the models or other way round. The argument is that there are certain situations where heuristics lead to better decision making than the models and figuring out those situations is the key. Heuristics can be applied where uncertainty is high, alternatives are many, data available is in smaller quantities and there are time pressures to take decision. For instance comparing the 14 models of asset allocation including Markowitz mean variance model and other variations of it with 1/ N diversification, 1/N was not necessarily the top most in terms of results but one among the top ones. For complex model to generate superior results, 500 years of stock data is required. This clearly states the trade off. Hence it is important to understand the ecological rationality.
There are alternative views which suggest that heuristics cannot be effective in complex situations because complex situation need complex solutions for instance catching the liar or game theory. One research was conducted to apply the heuristics to a complex problem. Complexity was introduced in the game theory by increasing the proportion of fat tails i.e. uncertain outcome. The final result of this experiment was that the simple heuristics lead a better result than the complex algorithms. Hence, this indicates that heuristics can be useful to solve complex problems as well.
- Boosting: How can the people’s competencies be fostered to improve decision making as against the popular idea that human beings are flawed decision makers and nothing can be done about it. The popular belief is that human beings are biased and lazy thinkers hence rely on heuristics. The boosting believes that by training people about the abstract concepts in simple heuristics manner can improve the decision making. eg: training the micro entrepreneurs on abstract skill of separating the household and business accounting. Instead of taking complex route, if u explain the home accounts and business accounts as two buckets and tag the expenditure in relevant bucket will help people understand the accounting in easier manner. A real life experiment in this area has shown significantly improved results.
This brings us back to the idea which was introduced at the beginning of this write up as to “approach of moving from people to models as against the existing popular approach of models to people”.
Two important takeaways from Q&A are application of this principle to the financial world in terms of valuation and risk communication. In valuation it is important to consider that the landscape of projections is more of uncertain nature due to which we may not be aware about all the outcomes and hence factor in the black swan events.
In risk communication, if we adopt the communication style which is more heuristic in nature as against the current approach of using technical terms the investors might be in a better position to understand them.
List of books which were suggested to understand the decision making process of human mind :
- Predictably Irrational
- Irrational Exuberance
- Blind Spot
- The Biased Mind
- Human Error
- Why we make mistakes
- Thinking Fast and Slow
- Taming the uncertainty