Revisiting the relationship between ability and sheepskin effects of schooling on individual earnings: The case of Pakistan

Significant amount of recent research continues to produce evidence in support of the presence of sheepskin effects in returns to schooling both for developed and developing countries. However, researchers have not made many attempts to identify or empirically test the possible mechanisms that may generate such effects. A few noteworthy exceptions are Flores-Lagunes and Light (2010) for the U. S., Riddle (2008) for Canada and Shabbir & Ashraf (2011) and Shabbir (2013) for Pakistan. Shabbir and Ashraf (2011) summarily reports that the sheepskin effects for rural Pakistan persist in the face of controls for measures of innate and cognitive ability. The present paper revisits this issue and adds value by presenting and discussing all of the relevant empirical estimates in full detail. Further, the present analysis fully updates the review of the literature as well as the various aspects of the pertinent debate surrounding the nature of the sheepskin effects. This study reconfirms that significant sheepskin effects exist for rural Pakistan for diplomas obtained by completing primary, high school and perhaps also FA and BA levels. Further, according to the detailed empirical regression results presented and discussed in this paper, the sheepskin effects prove to be robust both to an inclusion of measure of innate ability (Raven Progressive Matrices) and "cognitive" ability (specially administrated tests of literacy and numeracy). This implies that sheepskin effects "signal" individual characteristics unrelated to these measures of ability. The findings have significant policy implications about the nature of the private vs. social returns to schooling.