After schools all over the world closed their doors in the first half of 2020 to curb the COVID-19 pandemic, an estimated 1.58 billion affected students are now at risk of suffering from learning loss, with those in less developed countries at a disproportionately higher risk (United Nations, 2020). This includes 69 million students in Indonesia, 60 percent of whom, as of October 2021, had yet to return to classrooms since their schools closed nineteen months prior in March 2020 (Pusdatin, 2021). With limited internet coverage, little to no live interaction between teachers and students, and a lack of teachers’ skills in adjusting the materials to each of their students’ ability levels, most Indonesian students are at a high risk of experiencing learning loss (Arsendy et al., 2020).

A recent study on the effect of fourteen-week school closures following an earthquake in Pakistan in 2005 showed that students who experienced learning loss due to the disaster continued to learn less than those who were not impacted even after they returned to school. The gap between the two groups continued to grow four years after the disaster (Andrabi et al., 2020). Recognising the potential long-term impact of the situation, we realise that a lack of monitoring student learning progress and effective recovery efforts could lead to permanent effects.

To add to the concern, Indonesian students have lagged behind their global peers since before the COVID-19 pandemic. Even in an actual classroom setting, Indonesian teachers still struggle to teach basic curriculum materials effectively and have a narrow focus on preparing students for high-stakes exams (Afkar et. al, 2018; Berliner, 2011). Student performance in both the Programme for International Student Assessment (PISA) and the Trends in International Mathematics and Science Study (TIMSS) have indicated that the majority failed to achieve the minimum proficiency level in all subjects and have consistently ranked amongst the lowest of the participating countries (OECD, 2019; Mullis et al., 2016). Furthermore, using nearly-nationally representative data, Beatty et al. (2021) revealed that numeracy abilities across the country have deteriorated between 2000 and 2014, despite rising average years of schooling and encouraging policies.

Given the risk of significant loss and permanence of the phenomenon in low- and middle-income countries, along with the particularly lengthy period of school closure in Indonesia, this paper aims to give an insight into the discussion on student learning progress during school closures. We will present the impact of the closures on primary school students’ achievement in Bukittinggi, the third-largest city on the island of Sumatra and a highly urbanised area. The city has consistently performed well in most education-related measures due to a strong cultural emphasis on education and a supportive government (Nihayah et al., 2020), but has been significantly affected during the pandemic as most students are confined to their homes with very limited teacher-student interaction.

We use a panel sample of students from RISE Indonesia’s other study in Bukittinggi. We reassessed the sample in 2020, one year after its initial assessment. We then compare the performance of each of the 2020 grade cohorts against the 2019 cohorts to measure the impact of school closures on learning outcomes. To our surprise, our findings suggest that Bukittinggi students’ abilities have actually improved during school closures. This improvement is mainly driven by students with low initial achievement with no differences between genders, while initially high-performing students did in fact suffer from learning loss. The increase is large and significant in both numeracy and literacy—with the gain in numeracy amounting to approximately half of a typical Indonesian student’s gain over their whole schooling years.