Working Paper: CEPR ID: DP18354
Authors: Julian Hidalgo; Michelle Sovinsky
Abstract: Due to global lockdowns driven by the COVID-19 pandemic, the pervasiveness of inequalities in digital access is more salient. As a result, finding strategies to narrow the digital divide has moved to the forefront of public policy. We examine the impact of a pricing subsidy in Colombia to learn about what hinders low-income populations from adopting internet services. Our model allows take-up to depend on the types of plans offered as well as the rate of diffusion in the neighborhood. We find that increasing the diffusion rate (via internet literacy programs) is more beneficial to takeup among households in the lowest socioeconomic markets and in less technically savvy markets relative to a pricing subsidy. Pricing subsidies reduce the digital divide by 13% points but over half of that is due to the impact on the diffusion rate. Our results suggest that non-price policies are equally important to bridge the digital divide.
Keywords: digital divide; network effects; internet diffusion; limited choice sets; data restrictions; discrete choice model; price subsidies
JEL Codes: L15; L51; L86; D12; D31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Increasing diffusion rate through internet literacy programs (L86) | More beneficial for households in lower socioeconomic strata (D19) |
| Pricing subsidies (D49) | Overestimated impact (F69) |
| Non-price policies (L49) | Important for less technically savvy consumers (L96) |
| Pricing subsidy (D49) | Reduction in digital divide (I24) |
| Pricing subsidy (D49) | Increase in adoption rates (J13) |
| Diffusion rate (C69) | Reduction in digital divide (I24) |