The Impact of COVID-19 on School Choice and Household Education Expenditures: Evidence from India

• In Event: Economics of Education

Proposal

In October 2021, the Government of India opened most government schools across the country, ending the world's second longest school closure as a result of COVID. The return to in-person schooling brought more than a return to physical classrooms. Given the economic impacts of one of the world's strictest lockdowns, commentators have speculated that many families that had previously been sending their children to private schools no longer had the resources to do so and moved their children to government schools (Central Square Foundation, 2020). As children return to schools, government and policymakers will be faced with the task of how to serve a student body that has been out of school for nearly two years and whose composition has changed in that period.

In this paper we ask what the effects of the COVID-19 lockdowns have been on school choice and household expenditures on education. In particular, we look to identify what impact the lockdowns have had on school choice and household-level educational expenditures and where households send their children after schools re-opened.

To identify the effects of lockdowns on school choice and household expenditures on education, we leverage a long-run high-frequency unbalanced panel available from the Centre for Monitoring the Indian Economy (CMIE). The Consumer Pyramids Survey from CMIE contains detailed consumption data on approximately 120,000 Indian households from 2014 to 2022. The data includes questions on how much households spend on major educational items, which allows us to track how individual household expenditure has changed over time. CMIE also collects data on time use for individual members of the household, which combined with expenditure data, allows us to estimate whether children in the household are going to private or government schools, or have dropped out of schooling altogether.

We combine this panel with two sources of data to identify variation in when schools were closed. First, we use Google Mobility Data that provides aggregated mobility from mobile phones using Google Maps to estimate the effects of increased or decreased mobility on household-level expenditure decisions. Similar data has been used to estimate changes in mobility during COVID in a number of contexts (Andersen, 2023). Second, we manually code school reopening dates across all states in India. While the Central Government could set a national-level date for school re-openings, states were free to re-open (and close again) schools after the national-level opening date. For example, the state of West Bengal, the fourth largest state by population, did not reopen primary schools until February 2022, while the neighboring state of Assam reopened schools on the same day as the national government, closed them again in May 2022, and then reopened them once again in October 2022.

We first provide descriptive results of the impact of COVID-19 lockdowns on school choice and household educational expenditures. We find a large decrease in private school attendance from 40 to 20 percent of households with children that has not rebounded since 2020. This was matched almost equally with an increase in public school attendance and school dropouts in 2020 and 2021, and a decrease in out of school children and increase in children in public school in 2022. We also leverage differences in COVID severity between districts using district-level Google Mobility data. Here, we find that greater COVID severity led to a decrease in most household level expenditure, but only expenditure on private tuition was significant. This is largely consistent with descriptive results that shows broad decreases in expenditures on education. Students also spent less time on learning if they experienced greater COVID severity.

Given India's federal structure and separate reopening dates for different grades within states, we leverage between- and within-household variation in when children were able to return to school in a staggered difference-in-differences framework to causally estimate changes in school choice and educational expenditures (Callaway, 2020; deChaisemartin, 2020). We find that the total household expenditure on education increased significantly as schools re-opened. However, household spending on private school tuition as well as individual student's learning time remained similar to that while schools were closed, suggesting the impacts of COVID persist after schools reopened.

Our paper makes three major contributions. First, we contribute to a rapidly emerging literature on the impact of COVID on human capital formation in India and globally (Andrew and Salisbury, 2022; Angrist, Bergman and Matsheng, 2020; Li et al., 2023; Moscoviz and Evans, 2022; Singh, Romero and Muralidharan, 2022). The more optimistic findings suggest that low-cost technological and remediational measures have been able to mitigate and reverse the impacts of time out of school (Angrist, Bergman and Matsheng, 2020; Singh, Romero and Muralidharan, 2022), but these have been for children that governments and NGOs have been able to reach. Our paper also estimates dropouts, providing policy-relevant evidence on how much governments will need to target out-of-school children. Given the long-run nature of the panel, we can also track the impacts over a long period of time.

Second, we contribute to a literature on household expenditures on education during economic uncertainty more broadly. Families often adjust their investment levels in education when experiencing income shocks, which may negatively impact children's educational performance and outcomes (Ananat et al., 2011; Oreopoulos, Page and Stevens, 2008; Lunn and Kornrich, 2018). In this paper, we examine how households adjust their expenditure during and after lockdowns and further analyze the impacts on an important education investment decision: school choice.

Finally, for policymakers, we provide estimates of household-level human capital responses to the COVID-19 pandemic. COVID has shifted children from private to government schools and vice-versa, and in and out of schooling (Central Square Foundation, 2020). Understanding which children and households are most likely to be affected are of critical importance to policymakers as they look to mitigate COVID-related learning losses (Moscoviz and Evans, 2022). The data we use can also be continuously updated, allowing policymakers to apply our methods to later rounds of the CMIE data to better understand the medium-to-long-run effects of COVID-19 on school attendance and household human capital decisions in India.

Authors

• Ashutosh Bhuradia

• Emmerich Davies, Harvard Graduate School of Education

• Fei Yuan, Harvard Graduate School of Education