A Beautiful Science

The Neuroeconomics of Trust, Paul J. Zak

The Genetic Basics of Political Cooperation, James H. Fowler, Laura A. Baker, and Christopher T. Dawes

Two more articles from the Hendricks Symposium on Genetic Factors in Politics. I have recently blogged several others, as well as noting another in the context of Enterprise Resilience farther back.

The notes are, as always, below the fold. Above the fold is a nifty reference to the film A Beautiful Mind, which is used to argue that economics’ “rational man” appears to suffer from an atypical mental disorder. Secondly, more findings on genetic and environmental individuality that repeats Hibbing’s and Alford’s finding of a genetic basis for belief.

A Rational Mind?

There is clearly a problem with the SGP Nash equilibrium in this game since those who play out of equilibrium earn more money. Though John Nash did not directly analyze the trust game (which is a sequential-play prisoner’s dilemma), his well-publicized illness reveals why the SGP Nash equilibrium concept does not apply here. As most people know, John Nash suffers from the neuropsychiatric disorder schizophrenia. Schizophrenics are typically socially withdrawn, and analogously the SGP Nash equilibrium for the trust game does not recognize that the game is embedded in a social interaction. (Zak 9)

Turnout = Genetics (A Lot) + Unique Experiences (Somewhat) + Shared Experienced (A Little)

Figure 1. Ternary diagram of 10,000 draws from posterior Bayesian distribution of estimated components of variance in voter turnout (left panel) and party affiliation (right panel). The probability that the true coefficients lie outside the region of these draws is p=0.0001. The mean of the distribution is noted by the blue solid circle. These results suggest that heritability plays a significant role in political behaviour. Variation in political cooperation (turnout) is more than half heritable, while variation in political affiliation (partisanship) is only about one fifth heritable. (Fowler, Baker, and Dawes 5)

Partisanship = Shared Experiences (A Lot) + Genetics (A Little) (But No Shared Influence?)

You know what’s below the fold…

Matching public voter turnout records to an adult twin registry, we compare concordance in political behaviour between monozygotic and dizygotic twins. The results show that the decision to cooperate by choosing to vote is primarily determined by genetic factors. (Fowler, Baker, and Dawes 1)

Moreover, contrary to the expectation that the environment would decrease concordance over time, MZ twins living apart tend to become more similar with age (Fowler, Baker, and Dawes 3)

To assess the impact of genetics on turnout, we obtained electronic voter registration records for 3.8 million voters from Los Angeles County with complete vote histories for 8 elections (three primary, two statewide, and three general) from 2000- 2005 and matched them to the Southern California Twin Registry,13 a volunteer adult registry of MZ and DZ twins who live in the Los Angeles area. (Fowler, Baker, and Dawes 3-4)

Fig.1 shows the posterior distribution of the estimated factors. About 53% (95% C.I. 23%,73%) of the variance in turnout behaviour can be explained by genetic factors. The environment also plays a role, but the unshared environment accounts for much more variance (30%, C.I. 22%,39%) than the shared environment (17%, C.I. 1%, 43%). These results are based on data that pool same-sex male and female pairs. When we conduct the same analysis on each gender separately we do not find significant differences in the estimates. The genetic component for males is 51% (95% C.I. 6%,79%) and for females is 59% (95% C.I. 19%,81%). (Fowler, Baker, and Dawes 6)

To a certain degree, these results should not be surprising. Scholars have already documented a genetic basis for altruism and prosocial behavior17,18 and these behaviours in turn have been linked to voter turnout.19-21 (Fowler, Baker, and Dawes 7)

More importantly, our results have general implications for the vast literature on cooperation in humans. Although previous laboratory studies examined cooperation between twins, 27,28 none has focussed on the question of how twins cooperate with unrelated members of the general population. Thus, this article represents the first attempt to estimate the genetic basis of cooperative behaviour in large, well-mixed populations. (Fowler, Baker, and Dawes 7)

The traditional view in economics is that individuals respond to incentives, but absent strong incentives to the contrary selfishness prevails. Moreover, this “greed is good” approach is deemed “rational” behavior; without extreme self-interest, the standard models predict that money will be left on the table during a transaction and therefore an equilibrium cannot have been reached. (Zak 1)

Similarly, a substantial body of research has examined variations in efficiency by individuals within a firm, called X-inefficiency (XE) by Leibenstein (1966) (reviewed in Franz, 1997; and Franz, this book). The consensus view is that XE arises from bounded rationality and psychological motives that militate against efficiency. One way to reduce XE is to provide incentives for individuals to behave “more rationally.” (Zak 1)

A third example of a failure of the fully rational agent model is the degree of cheating during intertemporal transactions with asymmetric information (Zak & Knack, 2001)… demonstrate that low rates of investment occur primarily because of weak formal and informal institutions that inadequately enforce contracts. Nevertheless, even in institutional environments that do not enforce contracts well, a substantial number of investments still occur, presumably without undo duress, suggesting that some or even many money managers are reasonably trustworthy (or, alternatively, that investors are poor monitors of advisors, but since investors have a substantial incentive to monitor, this explanation is unlikely). (Zak 1-2)

A possible explanation for the substantial amount of “irrational” behavior observed in markets (and elsewhere) is that humans are a highly social species and to an extent value what other humans think of them… Loosely, it is possible that it “feels good” to fulfill others’ expectations in us. If such a cooperative instinct exists, it must be conditioned on the particular environment of exchange, including the history of interactions (if any) with a potential exchange partner. If conditional cooperation where not the case, individuals would be gullible, and the genes that code for gullibility would not have survived over evolutionary time (Boyd et al. 2003). (Zak 2)

The model shows that the degree of generalized trust in a country is inversely related to the transactions costs associated with enforcing an investment contract. In particular, trust depends on the social environment (how similar or dissimilar are those in a transaction; for example, think of the high degree of ethnic homogeneity in Norway, and how strongly social norms are enforced); the legal environment (how effectively contracts are enforced by formal institutions; for example, how readily redress can be obtained if one party of to the transaction believes that he or she obtained an unfair outcome); and the economic environment (as incomes rise, people will behave as if they trust others more because their time cost to investigate their trading partner rises; as income inequality rises, it is more likely that one’s trading partner will be untrustworthy because differences between parties to exchange, and therefore incentives to cheat, are greater). (Zak 4)

It also shows that societies that are less heterogeneous (in income, language, ethnicity, etc.) have higher trust because social ties between parties who are similar informally enforce contracts. For similar reasons, societies that are fair (have less economic discrimination) have higher trust. Alternatively, sufficiently strong formal institutions that enforce contracts can promote high levels of trust even in highly heterogeneous societies like the U.S. (Zak 4)

Empirically, trust is among the powerful factors economists have discovered that promote growth. The analysis in Zak & Knack (2001) shows that a 15 percentage point increase in the proportion of people in a country who think others are trustworthy raises income per person by 1% per year for every year thereafter. (Zak 4-5)

The results of Zak & Knack (2001), Knack & Zak (2003), and Zak & Fakhar (2005) demonstrate that the likelihood of two individuals who do not know each other exhibiting trust depends crucially on the social, legal, biological, and economic environments. (Zak 7)

This section surveys a variety of experimental studies that support the thesis that human beings are “wired” to be conditionally cooperative. President Abraham Lincoln said “… people, when rightly and fully trusted, will return the trust,” A substantial number of behavioral experiments by economists and psychologists have characterized the high degree of trust and trustworthiness in the laboratory consistent with Lincoln’s view that humans tend to reciprocate trust. (Zak 7)

As reported in Zak, Kurzban & Matzner (in press; and 2005) and as shown in Figure 2, OT levels in DM2s who receive an intentional trust signal are almost double that in DM2s in the Random Draw condition. (Zak 12)

Randomly, some women in our experiment were ovulating (progesterone > 3ng/ml) but none were pregnant (which is another time progesterone is high) by testing their levels of human chorionic gonadotropin (β-hCG). Progesterone has been shown to inhibit the uptake of OT by its receptor. This natural experiment where some female participants were ovulating allowed progesterone to disrupt the effect of OT on DM2 behavior: these women got the same OT surge when receiving a signal of trust but were less trustworthy (one-tailed t-test, p<.04). (Zak 13) OT appears to motivate DM2s to behave in a prosocial manner rather than play the SGP Nash strategy. (Zak 13) We also gave our experimental subjects an extensive survey inquiring about demographics, social behaviors, sexual behaviors (since OT is a reproductive hormone), and psychological profiles. Of 200 questions, almost none were related to OT levels or behavior in the trust game. Trust was related to three questions on whether DM1s thought others were mostly trustworthy or honest, but none of the survey questions were related to DM2 behavior. (Zak 13) Oliver Williamson (1993) coined the term “calculative trust” to denote the ability to use one’s experience to estimate the likelihood someone will be trustworthy. (Zak 15) The interpretation of these findings is that greater prefrontal activity is needed to forecast what another person will do and trust them compared to taking the sure payoff when playing the SGP Nash strategy or when interacting with a computer that plays using known probabilities. This theory of mind activation is a neural substrate associated with calculative trust. (Zak 15-16) These authors conclude that among the women studied, cooperation itself is rewarding, but requires the mediation of the conflicting concerns of making more money but behaving in socially less acceptable ways. (Zak 16) At the national level, trust can be raised by emphasizing the importance of education, reducing inequalities, and promoting freedom and democracy. National institutions that allow and encourage individuals to achieve their goals directly promote trust and therefore the creation of wealth. This is reflected in the higher rates of return on national stock markets for countries that have higher levels of generalized trust (Zak, 2003). English philosopher Bertrand Russell (1872-1970) wrote that “The most valuable things in life are not measured in monetary terms. The really important things are not houses and lands, stocks and bonds, automobiles and real state, but friendships, trust, confidence, empathy, mercy, love and faith.” The research reviewed here extends Russell’s statement. Friendships, confidence, empathy, mercy, love and faith all follow from trust and are likely mediated by oxytocin. As social scientists apply these findings to institutional design, not only will productivity be raised, but so will happiness. (Zak 18)