What makes trusters trust?
The vast majority of studies investigating the antecedents of interpersonal trust are limited by their reliance on cross-sectional data. In this paper, we make use of highquality repeated measures data from the British Household Panel Survey (BHPS) to examine the effect of a range of theoretically related explanatory variables on subsequent levels of interpersonal trust over a six year period. We employ a fixed effects specification with lagged covariates to control for the time-invariant characteristics of individuals that might spuriously link our explanatory variables to subsequent changes in trust. We contrast the results of our fixed effects model with cross-sectional and a repeated measures random effects specifications, which do not control for unobserved individual heterogeneity. While the cross-sectional and random effects models show substantial effects of life events on subsequent levels of trust, the fixed effects specification shows only two significant explanatory variables. The only events found to predict a future increase in trust are obtaining a post-compulsory educational qualification, and improving one's perception of the financial situation of the household. These results lead us to conclude that existing correlational studies of the causes of interpersonal trust are likely to be affected by endogeneity bias. Additionally, the small number of significant predictors of trust found in the fixed effects model lends support to the idea that interpersonal trust is a relatively stable social value, developed early in the life-course and relatively resistant to 'life events' in the short to medium term.