Does personality matter?

Does personality matter in economics? Two new papers suggest not, or at least not much.

First, Julia Muller and Christiane Schwieren got people to play a trust game, in which one player (the trustor) is given some money and invited to choose how much to share with player 2, the trustee. Any gift is tripled, and the trustee then chooses how much to return.

They found that among trustors, there were correlations between the amounts they chose to give and some of the “big five” personality factors. An inclination to trust - give money to player 2 - was positively correlated (0.28) with agreeableness and negatively correlated with conscientiousness (-0.26), as well as with anxiety. Correlations with other big five factors were insignificant. And personality factors taken together explain around a quarter of the variation in trustors’ behaviour. I don’t know whether this is a lot or a little.

However, among trustees, there were no correlations between measured personality characteristics and the amounts returned. The only (very) significant predictor of the amount returned was the amount received; if people got a lot, they gave a lot.

A paper by Armin Falk and colleagues largely corroborate this. They used lab experiments to test for links between personality factors and time preference, risk aversion, trust, reciprocity and altruism. And they found weak correlations. For example, there’s a negative correlation between neuroticism and risk appetite, but only -0.12, and a positive correlation between agreeableness and altruism, but only 0.2.

They also used a different data set to look for links between wages and personality factors amongst Germans. They found that openness and, to a lesser extent, conscientiousness are positively correlated with higher wages, as are having an internal locus of control and being more inclined to trust people; see also this paper by James Heckman and colleagues. Altruism, extraversion and agreeableness are associated with lower wages; the latter corroborates a finding from the UK. However, only around 10% of the variation in wages is explained by observable personality factors.

I reckon there are (at least) two interpretations of all this. One is that personality factors are dispositions which are only weakly correlated with actual behaviour. A generous man on a bad day will be mean, a neurotic one on a good day will take a risk, and so on.

The other possibility is that Marxists (and  neoclassical economists!) have a point. Behaviour is determined more by the situation we are in than by our character. When we impute character as an explanation of behaviour - as when we infer that successful men are highly motivated or that ones who give to charity are generous - we are committing the fundamental attribution error, of overweighting personality and underweighting the environment.

Comments

[Rob]

The first set of results does not necessarily suggest that personality isn't important, rather that reciprocity is a near-universal trait. This is fairly close to what most game theorists would predict. In fact, it suggests that personality might matter a lot, because the person who makes the first move sets the tone for all later interactions, since the later interactions are all governed by reciprocity.

[chris]

@ Rob - but personality only accounts for a quarter of the variation in 1st moves.

[Moving storage]

I am totally agree with Rob,that "personality only accounts for a quarter of the variation in 1st moves". This theme made me think about a lot of things around us.