The Evolution of Trust
Post: July 3, 2013 9:34 pm
Author: Michael L. Manapat
Source: Special Issue
Economics has long been interested in deviations from rationality in human decision making, but it generally provides proximate rather than ultimate explanations for irrationality. “Inequity aversion,” for example, might be invoked as the reason why people reject small offers in the Ultimatum Game, but this tells us little about how or why humans developed such preferences. Evolution, on the other hand, can provide ultimate explanations for seemingly irrational human behavior.
In our paper “Information, irrationality, and the evolution of trust
,” David Rand, Martin Nowak, and I show how evolutionary dynamics can lead to the emergence of trust and trustworthiness. The study of trust is formalized with the “Trust Game.” The “investor” starts with one monetary unit and can either keep it or transfer it to the “trustee.” If the transfer is made, the trustee can choose to keep any amount between 0 and 1. Whatever is not kept is returned to the investor after being multiplied by a factor b > 1. The probability that the investor makes the transfer is a measure of trust, the fraction returned by the trustee is a measure of trustworthiness, and the multiplicative factor b corresponds to the gains that arise from mutually beneficial interactions based on trust.
This model is applicable to a variety of real-world scenarios—for example, investors place money with investment management organizations, which can either honor their fiduciary responsibilities and return the principal together with investment gains after fees or abscond with the entirety of the investment.
In a non-repeated Trust Game involving anonymous individuals, the rational investor will not make the transfer and the rational trustee will return nothing after receiving a transfer. In behavioral experiments, however, humans transfer with high frequency and return substantial fractions of what they receive. How can this seemingly irrational behavior be explained?
We show that the existence of “information” can lead to the evolution of trust and trustworthiness. If there is some chance that the investor knows the fraction that the trustee will return before the interaction begins, then cheating trustees are caught sufficiently often that all trustees are incentivized to be trustworthy. Investors are thus, in turn, incentivized to be trusting even in interactions without information. Because humans evolved in a setting in which information is available, their inclinations are to be trusting and trustworthy even when exchanges are entirely anonymous. We verify our theoretical findings with large-scale, online behavioral experiments.
Our study also reveals some interesting properties about the evolutionary dynamics of human preferences. When the probability of having information is sufficiently high, the population consists entirely of trusting investors and trustworthy individuals, one of the formal equilibria of the Trust Game when information is added. On the other hand, when this probability is in a certain intermediate range, the population is constantly cycling between states of trust and distrust. Trustees become trustworthy, investors start to trust, trustees begin to exploit trusting investors and turn untrustworthy, and investors respond by trusting less. Thus, while behavior in the aggregate seems to be at an equilibrium, behavior at the individual level is constantly fluctuating in response to changing conditions.
Evolution can thus provide us with insights beyond those reached by classical economic theory, and these insights can have important social, economic, and policy implications. Online marketplaces need to design reputation systems to ensure trust and trustworthiness among buyers and sellers, and our results can shed some light on how much information is necessary. Government officials and NGOs often need to devise policy interventions to facilitate commerce, of which trust is a critical component, and the observation that trustworthiness leads trust (and that cycles are inevitable if information is not sufficiently prevalent) can guide these actions. Given that human behavior was shaped by evolution on a macro level and continually evolves via social forces on a micro level, the evolutionary perspective should be a central part of any analysis of human interaction.
Read the paper: Manapat, M., Nowak, M., & Rand, D. (2003). Information, Irrationality and the Evolution of Trust.