| Description: |
We develop an innovative two-step model in which contact with a public service is a necessary but not sufficient condition for paying a bribe. This identifies three groups: those without contact; those having contact but not paying a bribe; and those paying a bribe for a service they have contacted. We test hypotheses about why contact and bribery vary with individual characteristics, public services and national context using survey data from the 24-country 2012 Latin American Public Opinion Project. Heckman’s multi-level probit statistical method controls for the effect of selection bias in contacting education, health, courts and local government services, refining and altering influences on the payment of bribes found in a single-step model. Variations in individual social and political capital and national context are more often significant than individual socio-economic resources.
|
