AN ILLUSTRATION OF THE USE OF INSTRUMENTAL VARIABLES - DO NEIGHBORHOOD CONDITIONS AFFECT A YOUNG PERSONS CHANCE OF FINISHING HIGH-SCHOOL

Instrumental variables estimation (IVE) is useful in addressing a vari ety of estimation problems. In particular, IVE is useful in situations where explanatory variables are measured with error, are correlated w ith unobserved determinants of the dependent variable, or are simultan eously determined along with the outcome. IVE estimation depends on th e analyst's ability to identify an instrument, a variable that influen ces the troublesome regressor but does not affect the outcome of inter est. An illustration of the use of IVE is presented that examines the relationship between the neighborhood dropout rate and an individual's chance of finishing high school. In this illustration, city-level cha racteristics serve as instruments. The empirical results suggest that the neighborhood dropout rate affects a girl's chance of finishing hig h school. Tests described in the text suggest that the instruments cho sen are invalid for boys.