Proponents of the efficient market hypothesis claim that markets are fully rational and able to incorporate new information correctly into asset prices. They accept that some abnormality may arise in the formation of asset prices but assume that competition amongst investors trying to take advantage of such abnormalities will drive prices back to their “correct values”. Therefore, they consider Bayesian rationality as a good description of investor behavior.
Our research shows that even if taken at only its bayesian definition, irrationality does not always explain the biases and paradoxes that are empirically observed. it is rather our assumptions about the world around us which are wrong than the way we derive statistical inferences from them.
Project: we seek to understand why people who know that a bubble is ongoing in a given market continue to participate in that market.
Johann Wolfgang von Goethe wrote that Man was not born to solve the problem of the universe, but to find out what he has to do; and to restrain himself within the limits of his comprehension.
It is most certainly true that the universe will remain at a level of complexity beyond our comprehension. Yet, we are beings of desire and we naturally wish to push the limits of what we understand.
The Santa Fe institute, a leading center on complexity science, sees complexity as arising in any system in which many agents interact and adapt to one another and their environments. These interactions and adaptations result in evolutionary processes and often surprising "emergent" behaviors at the macro level.
Our research shows that it is possible to find some order in an evolving complex system. A financial market, for instance, is a complex open system where prices derive from all the different expectations of market participants about other participants. Yet, we can identify periods of equilibrium (i.e., bubbles) where the market is less complex.
Project: we hypothesize that market participants cannot see the overall increase of complexity in a system but can detect local patterns of complexity.
The Random Walk Hypothesis states that the evolution of a stock price is random and thus unpredictable. Randomness is such a large concept that it could define almost any process. How can we be sure that an observed pattern is not due to randomness itself?
After all, a roulette ball can fall on black 20 times in a row and the person sitting next to you in a waiting room may have the same birthday as you. Those are random events with a very peculiar pattern: they are unlikely to be observed.
Since we derive statistical inferences from what we observe in the world, we deem unlikely events as being "not random". In the same way, we consider as random any event that displays features of randomness.
Our research shows that there exists a form of non-randomness that can be identified and that is not sustainable in the long-run. This from of non-randomness constitutes a bubble.
Project: we seek to better understanding randomness in human-driven events such as a financial market. A market is governed not by a fair coin toss but by decisions of people; accordingly, its apparent randomness should be generated by people who believe that this market should be random.
Project: we seek to detect non-random processes in the way a disease spreads and the way cells multiply.