High-speed broadband, school closures and educational achievements

In this study, I shed new light on the short-run effects of access to high-speed internet on educational disparities, before and after the pandemic shock. By following 3 million students belonging to 6 different cohorts over the period 2012-2022, I estimate the effect of the broadband infrastructure on student performance. While most previous contributions use discontinuous jumps in the available broadband connection speed across space at a given moment in time, this study exploits the actual roll-out of an infrastructural policy associated with an increase in 30 Mbit/s household broadband coverage from 40% to 80% over a 5-year period. The estimation strategy relies on a unique dataset, combining panel data on student performance with a rich set of school- and student-level information and broadband data measured at a very fine spatial scale. Results show an average null effect of high-speed broadband on 8th grade student performance in both numeracy and maths. However, this results masks substantial heterogeneity: lower performers in grade 5 and students with better backgrounds gain from internet speed, whereas the opposite is true for other students. Interestingly, the stronger effect on low-performers tends to disappear during the lockdown, suggesting a negligible mitigating role for high-speed internet during the period of school closure. On the other hand, the broadband infrastructure might have further amplified the gap between students with different socioeconomic background.