About us    Campaigns    Research    Support us    Publications    Media Room    Join Us    Contact us

Parental Valuation of School Attributes in Developing Countries: Evidence from Pakistan


Pedro Carneiro
University College London, Institute for Fiscal Studies and
Centre for Microdata Methods and Practice
Jishnu Das
World Bank
Hugo Reisy
University College London

June 2013

A surprising explosion of private for pro t schools in very poor settings has been recently documented for areas of India, Ghana, Nigeria and Kenya by (among others) Tooley & Dixon (2005)1 and Dixon (2012), and for Pakistan by Alderman, Orazem & Paterno (2001) and Andrabi, Das & Khwaja (2010). These are areas where the majority of children attends a large number of small private schools. Much of this growth in private schooling appears to be driven by rising discontent with the quality and availability of public education, even when it is freely provided, and even in environments where parents are very poor. Even though we think of education policy as mostly geared towards the public sector, this research shows that private schools are increasingly playing a central role in the education of poor children. Unfortunately, there is little research on what drives school choices among the poor, when there exists an abundance of choice possibilities (one exception is Alderman, Orazem & Paterno (2001)). Without knowledge of what drives parental choices it is impossible to understand the role of the private sector in this market, and to inform education policy. This paper studies the demand for primary school attributes in Pakistan. We use a rich dataset with information on parental characteristics and their school choices, and a detailed set of attributes for all schools in various education markets (villages), as well as all their costs. We estimate a model for the demand of di erentiated products (Berry, Levinsohn & Pakes (1995) and Berry, Levinsohn & Pakes (2004)). Based on the resulting estimates we compute the marginal willingness to pay for di erent school attributes, and we simulate the welfare implications of abolishing or regulating private schools. Finally, we assess whether the school attributes most valued by parents are strongly correlated with student success. We nd that the distance between a student’s residence and each school is a verystrong determinant of choice. The average distance between home and school (for those enrolled) is 500 meters for girls and 670 meters for boys. A 500 meter increase in this distance decreases the likelihood that a school is chosen by 5.1 percentage points for girls, and 4.2 percentage points for boys.

In contrast, an increase in annual tuition by USD$1 (from an average of USD$13 for private school) leads to declines of only 0.64 and 0.37 percentage points in the likelihood that a school is chosen, for girls and boys respectively. A 10 percentage point increase in the proportion of teachers with 3 years of experience (from an average of 60%), or a 10 percentage point increase in the proportion of teachers with a university degree (from an average of 25% for girls and 31% for boys), have less than a 1 percentage point impact on the probability that a school is chosen. Increases in the average test scores of students or the average test scores of teachers in each school have almost no impact on school choice (except for the poorest girls), perhaps because they are difficult to observe (when compared to the other attributes we consider). More generally, there is some but imperfect correspondence between the school attributes that are more strongly associated with student performance within each village, and the school attributes that are most valued by parents.

To read more http://ipl.econ.duke.edu/bread/papers/1013conf/carneiro.pdf

Comments Off on Parental Valuation of School Attributes in Developing Countries: Evidence from Pakistan

  Disclaimer: The copyright of the contents of this blog remains with the original author / publisher.