Globalisation has changed job opportunities in much of the developing world. In India, outsourcing has created a new class of high-skill jobs which have increased overall returns to schooling. Existing evidence suggests education may broadly respond to this change. We use microdata to evaluate the impact of these jobs on local school enrollment in areas outside of major IT centers. We merge panel data on school enrollment from a comprehensive school-level administrative dataset with detailed data on Information Technology Enabled Services (ITES) center location and founding dates. Using school fixed effects, we found that introducing a new ITES center causes a 4% to 7% increase in the number of children enrolled in primary school; this effect is localised to within a few kilometers. We show the effect is driven by English-language schools, consistent with the claim that the impacts are due to changes in returns to schooling, and is not driven by changes in population or income resulting from the ITES center. Supplementary survey evidence suggests that the localisation of the effects is driven by limited information diffusion.

In the last thirty years, globalisation has dramatically changed job opportunities in the developing world. In many countries this change has increased the skill premium. In India, the focus of this paper, this change has been particularly striking. The number of individuals employed in outsourcing-related businesses has increased from roughly 50,000 in 1991 to over 2 million in 2010 (NASSCOM, 2010); these jobs demand employees with high levels of education and a good command of English, and pay high salaries by Indian standards. The availability of these new opportunities increases the return to education which may, in turn, increase school enrollment. Popular median suggests the availability of jobs of this type may have broad geographic impacts in India, including impacts outside of areas where these new jobs locate (Giridharadas, 2010). Understanding the magnitude of this change, and how widespread the impacts are, is important for understanding the consequences of globalisation.

Existing evidence on India suggests that cities and districts with a major IT presence have experienced changes in education patterns with the growth in these jobs. In this vein, Munshi and Rosenzweig (2006) demonstrate evidence of increased returns to English in Mumbai between 1980 and 2000, and simultaneous increases in English-language enrollment among groups best able to take advantage of new job opportunities. Shastry (2010) shows evidence that districts with greater IT growth over the 1990s have greater schooling growth following this period. This paper makes two signicant contributions to the evidence on globalisation’s impact on school enrollment.

First, we estimate the effect of the introduction of these businesses outside of the major IT areas. This allows us to evaluate the validity of the popular claim that these businesses will have broad geographic impacts in all of India, and consider the question of whether these changes will ameliorate or exacerbate inequality across areas. Our data is sufficient to allow us to distinguish the magnitude of impacts over quite small distances, and we argue we are able to make strong causal statements about the impact of BPOs. As we detail below, we find that the impacts of call centers on school enrollment are large but very localised. Second, with a more qualitative survey we are able to provide some preliminary evidence on the mechanisms behind these effects and their relatively narrow geographic range. We argue this effect is due to limited information dissemination across areas. This suggests that in the absence of any intervention impacts may not be geographically broad, although better information provision about job opportunities could have large impacts.

The paper proceeds in two parts. We first use panel data on school enrollment and Information Technology Enabled Services (ITES) to estimate the impact of new businesses on school enrollment, both overall and in English-language schools specially. This estimation is done within school, using the sharp timing of ITES center introduction, and we argue the impacts we observe can be interpreted causally. The second part of the paper uses data from a survey in one area which provides GPS data on the location of households and ITES centers and detailed data on the quality of information about these jobs. This allows us to estimate how information decays as people move further from the BPO locations. Although it is more difficult to make causal claims in this case, we will be able to focus our estimation on people within a very small area (within 1 kilometer of a BPO) which limits possible confounds.

We begin with panel data on enrollment at the school level from a comprehensive administrative dataset in three states in India (Karnataka, Andhra Pradesh and Tamil Nadu); each school is observed for a period of four to eight years between 2001 and 2008. We combine this with a newly collected dataset on ITES business locations and founding dates. Our ITES center data includes areas outside of Chennai, Hyderabad and Bangalore, which allows us to estimate the impact of jobs in areas which have not had an overwhelming IT presence. Our ITES center location data allows us to identify the PIN code (similar to a ZIP code) location of each center, which we can link to school location. We use a school fixed effects estimator to analyse how enrollment changes within an individual school upon the introduction of a new ITES center to the area.