Handling negative data using Data Envelopment Analysis: a directional distance approach applied to higher education

This paper applies a data envelopment analysis (DEA) method to assess technical efficiency of both private and public universities in Italy. Moving from the traditional context where inputs and outputs are assumed to be non-negative, a directional distance function approach has been applied in order to handle both desirable (i.e. number of graduates) and undesirable (i.e. number of dropouts) outputs. The findings based on a panel from academic year 2003/2004 to 2007/2008 reveal the presence of interesting geographical (both by macro areas and regions) and ownership (private, public) effects. Several quality and quantity proxies have also been used in order to check whether the estimates depend on the output specification. Finally, the possible evidence of variation in the universities’ performances by subject of study has been taken into account in order to check whether the results are still consistent comparing universities within subject rather than across subjects.