Digital Divide In Economic Literacy: How Unequal Access To Technology Shapes Student Outcomes

Abstract

The digital divide  the systematic inequality in access to information and communication technology (ICT) across socioeconomic, geographic, and gender dimensions  is widely hypothesised to exacerbate educational inequality in developing countries. Yet rigorous causal evidence on the magnitude and mechanisms of this effect, particularly for economics learning outcomes in low- and middle-income country (LMIC) contexts, remains sparse. This paper addresses that gap using a uniquely constructed five-wave panel dataset of 18,640 secondary and higher-secondary students across 186 schools in four Pakistani provinces, collected between 2020 and 2025. Exploiting quasi-random variation in district-level ICT infrastructure roll-out generated by the Pakistan Tehreek-e-Insaf government's Digital Pakistan initiative as an instrumental variable, we estimate the causal effect of technology access on a newly developed and psychometrically validated Economic Literacy Index (ELI). Our two-stage least squares (2SLS) estimates indicate that moving from the bottom to the top quartile of the district-level Digital Access Index (DAI)  a composite measure of device ownership, broadband availability, electricity reliability, and digital skill  is associated with a 0.74 standard deviation (SD) improvement in the ELI (p < 0.001). The effect is substantially larger for female students (0.88 SD), students in the lowest household wealth quintile (0.97 SD), and students in rural government schools (0.82 SD), confirming that the digital divide amplifies pre-existing educational inequalities rather than operating uniformly across the student population. Structural equation modelling (SEM) reveals that approximately 41 percent of the total effect operates through the channel of access to digital learning resources, 28 percent through teacher digital pedagogy quality, and 31 percent through student self-directed study time. A counterfactual simulation suggests that eliminating the observed inter-district digital access gap in Pakistan would reduce the ELI variance attributable to socioeconomic background by 34 percent. These findings have direct and urgent implications for Pakistan's educational technology policy, ICT infrastructure investment strategy, and the design of digital inclusion programmes targeting the most educationally marginalised populations.

JEL Classification: I24, I26, O15, O33, C33, D63