In this paper I test the prediction that the reference standard for income comparisons increases in a person's productive ability, which is proxied for by education. The prediction is derived from a model by Falk and Knell (2004), who are among the first to endogenize the reference standard. In their model, individuals trade off a self-enhancement motive (choosing a low comparison income to make oneself feel good) against what Falk and Knell call a self-improvement motive: aiming high in one's comparison in order to be more motivated, and thus more productive, at work. Falk and Knell's main result is that, at least under their assumptions about functional form, the optimal comparison income increases in a person’s productive ability.The difficulty with any empirical test is that, to my knowledge, the reference standard for income comparisons is not observed in existing large data sets. As a way around, I propose to focus on immigrants (an application that Falk and Knell themselves suggest), and I assume that return visits are a way of adjusting one's standard for income comparisons. By extending Falk and Knell's model in this way, I am able to derive the two predictions that other things equal better educated immigrants (1) will return to their former home countries less often, and (2) will have greater difficulty feeling at home when they do return to their countries of origin. I test the two predictions using data from five waves of the German Socio-Economic Panel, which has a large number of observations on immigrants and a rich array of questions related to immigration. Both predictions are tentatively confirmed. In probit and ordered probit models controlling for a large number of other factors, I find that better educated immigrants are indeed less likely to have visited their countries of origin, and that when they did return for a visit, they report greater difficulty feeling at home. The effects are statistically significant, but rather small in size: a one standard deviation (i.e., about 2.5 years) difference in years of schooling reduces the probability of a return visit by approximately 0.7 percentage points, and is associated with a probability of feeling at home straight away that is lower by about 1.5 percentage points. Both marginal effects are small compared to the marginal effects of variables such as household income, distance to the country of origin, or dummies for having left the country of origin because of a war or because of a lack of political freedom.

Keywords

reference standard, social comparison, immigration, Germany