The long term effects of early selection – a quasi natural policy experiment from Hungary

Using a unique institutional change this paper provides causal estimates on the long-term effects of early educational selection. During the postsocialist transition the structure of the Hungarian education system changed gradually as new earlyselective tracks were established. These elite academic tracks have cream-skimmed the best students at ages 10 and 12 - as opposed to the typical age of first selection at age 14 - in order to provide better education and thus better employment chances for the selected. Utilizing the spatial and time variance in the establishment of these early-selective tracks we provide difference in-difference estimates on the effect of these tracks on probability of employment and other outcomes. Using the 2011 Census data we estimate time and settlement fixed-effect and eventstudy models. The paper contributes to the literature in two ways. Firstly, contrary to previous policy evaluations on the effect of comprehensive, de-tracking school reforms we identify the effect from a re-tracking policy, which made the school system more selective. Secondly, the establishment of the early selective tracks did not overlap with other large-scale educational policy changes, which allows for an accurate identification of the effects of early educational tracking. Our results show that contrary to expectations early selection has no effect on the employment and unemployment chances or on tertiary participation probabilities of the young adults. The effects are precisely estimated zeros.