OLS estimates of the rate of return to education are subject to a number of potential biases. Recent developments in the literature have focused particularly on exploiting alternative instruments, arising naturally in the data, to counteract these problems. A number of such instrumens are available in an Irish data set, including parental background variables (social class and educational attainment), and varaibles measuring changes in the schooling system (the introduction of free secondary schooling in the mid 1960s, accompanied by a rapid rise in educational participation rates, and the raising of the school-leaving age in 1972). The results suggest that OLS estimates of rates of return are not downward biased ( as would be suggested by recent arguments), and offer some support to the idea that discount rate bias may be of importance.
