Access and Excess - The Effect of Internet Access on the Comsumption Decisions of the Poor

This paper tests for the existence of a causal relationship between internet access and the consumption decisions of the poor. In particular, we develop theoretical models that provide two testable hypotheses. Firstly, that, by better aligning beliefs with reality regarding the value of future income, information attained via the internet will either increase or decrease one’s propensity to consume, depending on the nature of their occupation. Second, that, by heightening one’s concern for their relative status, access to social media via the internet will cause users to increase their conspicuous consumption. The analysis exploits a natural experiment, the 2015 roll out of a free internet provision program (Free Basics) in Colombia, as a source of exogenous variation in potential internet access. By taking advantage of a unique geo-coded dataset of telecommunications pylons in Colombia this paper is able to compare households covered by the free internet provision program to those who are not. We estimate a number of 2SLS instrumental variable and difference-in-differences specifications on rich data from the Encuesta Longitudinal Colombiana (ELCA) panel survey. Our estimates provide tentative evidence in support of our second hypothesis as we estimate a local average treatment effect (LATE) of a 12.2 percentage point increase in the share of conspicuous consumption in the overall consumption bundle as a result of internet access. We find no evidence of a statistically significant LATE on overall consumption. Our difference-in-differences, intention to-treat (ITT), estimates for both outcomes are small and insignificant, however this could be a result of a lack of power given the low take-up of the internet provision program.