Reversing losses in learning after Covid
The pandemic wreaked havoc on education. But a new study shows severe learning losses can be reversed, even if partially, by keeping schools open and pushing after-school remediation schemes, which can be helpful for marginalised students
The Covid-19 pandemic and resulting school closures delivered a massive shock to education systems worldwide. The shock was highly unequal both across and within countries. Education was disrupted in low-middle income countries such as India more severely than in richer countries which were better equipped to pivot to remote instruction. Within countries, poorer households were more exposed to both the health and income shocks of the pandemic. Further, children from poorer households were less likely to receive any instruction, whether remote or in-person. The World Bank, Unesco and Unicef estimated the potential long-term education costs of the pandemic to be up to $17 trillion.

Thus, understanding the extent of Covid-19 learning loss in India, and the impact of possible actions to mitigate it, should be a priority for education policy. Yet, we know remarkably little about fundamental questions related to pandemic-related learning losses and recovery, including the severity of the loss, the extent of recovery after reopening, the inequality in losses and recovery and the possible policy interventions to accelerate recovery and mitigate the increase in learning inequality.
In a new paper, we aim to answer these questions using an original dataset of over 19,000 primary school age children in rural Tamil Nadu whom we tested in July-August 2019 (before the pandemic), and then again in December 2021 (after 18 months of school closure), and April-May 20221 (after six months of school reopening), using comparable assessments.
We found large learning losses in December 2021. Children aged five to eight scored much lower on comparable tests than children of the same age and in the same villages in July 2019. This “learning loss” was large in both maths and Tamil and larger for older children (who had missed more schooling). Children who were eight in December 2021 scored roughly the same as those aged six in 2019, a developmental lag of two years. For children aged six, this developmental lag was smaller and equivalent to around eight months in Tamil and 10 months in maths. We also found that learning losses were larger among children with less educated mothers, and that school closures led to an increase in learning inequality.
How persistent are these losses? The big worry is that children who fall behind on foundational skills may never catch up, especially in an education system that prioritises finishing the syllabus, even when student learning levels are considerably behind curricular standards. There is also evidence that even a few months of school closure due to natural disasters can lead to growing learning gaps over time.
The good news is that we find evidence of considerable learning recovery after schools reopened. In our data, nearly two-thirds of the learning loss measured in December 2021 was recovered by May 2022. This recovery seems to be similar across ages, and faster for more disadvantaged children. Thus, learning inequality fell after schools reopened in Tamil Nadu.
One potential contributor to both the recovery and the reduction in inequality is a large-scale remediation programme called Illam Thedi Kalvi (ITK) or “education at doorstep” run by the Tamil Nadu government. The programme used community volunteers — local women who completed Class 12 or college — to deliver 60-90 minutes of remedial instruction in the evenings, using an engaging activity-based curriculum not directly tied to the school curriculum. ITK was among the largest Covid-19 learning recovery initiatives globally: It featured intensive community mobiliasation, and reached over three million children using 200,000 community volunteers between January and June 2022.
Programme awareness and participation were both high. Over 90% of surveyed households had heard about ITK, 57% reported their children attended, and attendance rates among participating children were also high. Attending the programme improved learning in both maths and Tamil, and we estimate that ITK accounts for about one-fourth of the total cohort-level recovery. Thus, about half of the learning loss would have been remedied six months after schools reopened even without ITK, and the ITK programme increased this to two-thirds. Importantly, participation in ITK was higher for more disadvantaged students. Thus, the programme effects from ITK accrue more to students from less-advantaged backgrounds, which likely contributed to the progressive learning recovery.
ITK was also very cost-effective, even in comparison to some of the most promising primary school interventions on which we have rigorous evidence. Field reports suggest participating in ITK provided substantial recognition and respect for ITK volunteers in their community. This may partly explain the large number of applicants (nearly 750,000 for 200,000 positions) even though volunteers received only a modest stipend of ₹1,000 a month.
Our findings are relevant for other Indian states, as well as other countries at similar income levels to India. First, the severity of learning losses that we document — where children are up to two years behind expected levels (from a low base) — highlights the seriousness of the Covid-19 shock for education. Second, we show that rapid recovery, even if partial, is possible and that opening schools and keeping them open can, by themselves, contribute substantially to learning recovery. Third, after-school remediation programmes such as ITK can be a template for other states. Such approaches may make sense even beyond the pandemic. In particular, the disproportionate benefits of ITK for disadvantaged communities suggests that such programmes may help accelerate India’s ability to deliver universal foundational literacy and numeracy.
Karthik Muralidharan is Tata Chancellor’s Professor of Economics at the University of California, San Diego. Abhijeet Singh is associate professor of economics at the Stockholm School of Economics. Mauricio Romero is assistant professor of economics at ITAM, Mexico CityThe views expressed are personal