Equilibrium, linear regression, rationality… Most of textbook economics relies on strict assumptions and presents us with simplified models of reality. Microeconomics, especially, is facing a tough challenge: predicting behaviour on a smaller scale. For example, perfect market equilibria – as well as Cournot-oligopolistic – have proven to be prone to optimistic bias (Hu, 2008). The search for more accurate models in a time when quality information is precious gave rise to new branches of the traditional disciplines. Behavioural approaches are connecting psychology and modern experimental neuroscience with the other social sciences and are crafting tools never used before. The aim of this article, and ultimately my research as a whole, is to present a cognitive illusion called the optimism bias, discuss its significance for behavioural economics and game theory, as well as to explore possibilities of testing models with empirical data adjusted for the bias.
The Optimism Bias
The optimism (or optimistic) bias has been a topic of research from as early as the 50s (Armor & Taylor, 2002). In recent years, however, psychology has benefited from the modern tools of neuroscience and helped explain this intriguing characteristic of the human mind. The experiments of Dr Tali Sharot from UCL not only confirmed the presence of the optimism bias – the fact that “we overestimate the likelihood of positive events and underestimate the likelihood of negative events” – but also precisely identified its origin in a specific area of the brain (Sharot, 2012a). Defining the optimism bias as “the difference between a person’s expectation and the outcome that follows” (Sharot, 2011a) also allows for further experiments and gives a way to measure its effects.
Figure 1: Optimism Bias in Healthy Individuals. Participants in a study by Sharot and Dudai (cited in Sharot 2011a, p. R942, cited in Sharot 2011b) were asked to estimate the likelihood of different events happening to them in the next month. Later they reported back which of the events had actually occurred. The estimated probability of the positive events was higher than that of the neutral or the negative, whereas the realised likelihood of all the events was roughly the same.
The optimism bias has been observed in 80% of people independently of their personal characteristics such as nationality, gender, or age, as well as in certain animals. The notable group missing the optimism bias are those who suffer from depression. It turns out, patients with mild depressions see the world rather realistically and those suffering from more serious forms of depression are pessimistically biased. (Sharot, 2012b).
News in the Nature of Game Theory
In game theory, simple one-shot games such as the Prisoner’s Dilemma, as well as very complex extensive forms with imperfect information, can be solved using Nash equilibria as the underlying principle to describe how the players choose their strategies. Strictly logically evaluating all the payoffs in a selfish, rational way, a Nash equilibrium indeed seems like the desired outcome. (Un)fortunately, in most cases, the ‘rationality of agents’ assumption is violated. Human minds are complex; we are driven by irrational desires and survival needs, as well as swiftly shifting sentiments. Behavioural game theory evidenced that the traditional equilibrium approach to solving games often fails to predict the decisions of real individuals (Camerer, 2003). The hypothesis behind my research is that, to some extent, the optimism bias can be held accountable for the divergence between the results of models and experimental findings.
The Persistence of the Optimism Bias
Sharot further argues that a reason for the development of the optimism bias during our evolution is to prevent stagnation. When the human race gained the ability to look into the future, each one of us was faced with the realisation that, inevitably, we are going to die. Even by strict theoretical backwards induction, if the ultimate payoff is death, then players should be indifferent between their choices – no one will evolve. However, through our irrational optimism, a constant picture of our bright future allows us to live on. A whole school of economists has been built on the famous quote “In the long-run, we are all dead” by John Maynard Keynes (1923). It did not leave them hopeless; instead, they crafted theories to make a difference in the world. According to Sharot’s research, this “coding” in our brain is very difficult to break. She even provided evidence that being aware of the optimism bias does not make it go away, neither will people change their expectations significantly in the face of the reality.
A famous example of the stickiness of the bias, presented by Sharot (2011a), goes as follows: the participants in a study estimated their likelihood of having cancer as (on average) 10%, then they were told that the actual probability is as high as 30%. Being asked the second time, in the light of the new information, the responds would only estimate their probability as 11%. They did not believe that they are the ones to whom the 30% probability applies. Furthermore, even though people may be pessimistic about the government, the economy or global peace, they are still optimistic about their own future and the future of their family (Sharot, 2011a).
Armor and Taylor (2002) obtained similar results. They conducted an experiment, in which the respondents were asked questions about their future outlook. These questions included future salaries, number of children or the likelihood of negative events such as illnesses. Their results showed that “between 85% and 90% of respondents claim that their future will be better - more pleasant and less painful – than the future of an average peer.” Again, the respondents were optimistic about their own future but not necessarily about the future of others. Camerer and Lovallo (1999, cited in Armor & Taylor 2002) studied how this difference in relative expectations of success affects market entry decisions in their paper. I want to test whether the players in strategic situations expect better payoffs for themselves and worse for their opponents, effectively influencing their decision making.
Implications for the Real World
The optimism bias is now a well-established concept. Adjustment for optimism bias is possible and, in the field of business and project managements, the concept is applied on regular basis. By the Guidance of HM Treasury on appraisal and valuation (HM Treasury, 2013), projects have to be adjusted for excess optimism on key parameters. Bands and criteria for choosing the correct adjustments are laid out in their Green Book. The budget for the 2012 London Olympics, for example, had to be revised as well. The revised budget after performing all adjustments, including the optimism bias, was almost 300% higher than the original one (Jennings, 2011). Similarly, in his dissertation, Hu (2008) examined a model of the northwest European electricity market. For example, he found that uncertainty about the implementation of new technology resulted in an optimistic bias of the total surplus, quantified as 14 000 to 18 000 €/hr.
The new fields of behavioural economics and behavioural game theory are arising to devise new methods of explaining human actions. This is increasingly more important in an age of computers, widely available knowledge and the promotion of diversity. Each individual has a unique personality and are not one of thousands of uniform agents in the models anymore. The results of game theory and the effects of behavioural approaches can be carefully tested in experiments (Gintis, 2011). Games such as the Ultimatum game, Prisoner’s Dilemma (simple or repeated) and the Public Goods game have been played by many sample groups and the results have never quite matched mathematical predictions. In my research, I want to have my own go at these traditional economic experiments. I do not expect the results of the experiments to be in accordance with the predictions; my aim is to devise an individual measure of optimism bias and then apply it to the results as an adjustment mechanism. Consequently, I hope to clarify some of the effects of optimism bias on private decision making – to check and re-evaluate the predictions of game theory if I remove the effect of the optimism bias.