Unfolding the Mechanisms of Compensatory Advantage: An Instrumental Variable Approach

In this paper we study how previous school results affect parental responses and expectations concerning the age at which their children will leave the educational system. We are particulartly interested in how these responses and expectations vary depending on parental socio-economic status, in the case of poor academic results. Our analysis draws upon the longitudinal survey “Panel d'élèves du second degré, recrutement 2007-2013”. We start with naife OLS regressions with measures of parental responses and expectations as dependent variables, and social origins and previous school performance as independent variables. Previous school performance is measured with the results in a national test at the beginning of lower secondary education and with a dummy that distinguishes those have repeated a year during primary education. In the next step, we address the issue of the possible endogeneity of previous school performance using an IV framework. The IV analysis is our specific contribution to the session. As instrument we use month of birth. Previous studies have shown that in countries like France with a strick rule for admission to primary education based on the date of birth, those born just before the cut-off date for admission are at higher risk of retaking one year and have lower school results. Previous studies have also shown that month of birth is not related to social class or to other factors that might affect parental responses or expectations, outside the causal path through school performance. Preliminary results suggest that the effect of previous school performance and also its heterogeneity by social origin is larger in the IV estimation. Our study is important because it sheds light on the mechanisms underlying compensatory advantage ie. the disproportionally large likelihood to move on to higher education for initially low performing students from socio-economically advantaged social background.