UMass Amherst Study Finds Affordability, Not Infrastructure, is Major Barrier to High-Speed Internet Connectivity

With a federal subsidy that has provided less expensive or free broadband internet to more than 23 million American households due to run out of money by the end of May, a new University of Massachusetts Amherst study reveals that direct-to-consumer subsidies can be far more effective at connecting households to high-speed internet than expanding broadband infrastructure.

The research, believed to be the most comprehensive accounting of U.S. broadband demand and usage to date, suggests that government policies aiming to expand internet access should be refocused on subsidies rather than infrastructure upgrades to help close the “digital divide” between connected and unconnected households, particularly among households with lower incomes.

Christian Rojas, professor of resource economics at UMass Amherst, found that affordability is the main driver of disparities in broadband connectivity—not lack of availability. In fact, more than 95% of the U.S. population could connect, but only seven in 10 households do. Rojas and postdoctoral researcher Augusto Espín show that, dollar-for-dollar, consumer subsidies can result in at least four times more households adopting high-speed internet compared with funding expansions of broadband networks.

“The infrastructure is already there for most people, they just need some help affording it,” Rojas says. “When you break down how the subsidies impact different people, we see that internet adoption among the lowest income bracket increases the most dramatically. If you think about the gap between rich and poor, that’s where the subsidies are able to bridge it almost completely.”

The $65 billion 2021 Bipartisan Infrastructure Law dedicated $42.25 billion to deploy new internet infrastructure and just $14.2 billion for direct subsidies. Under the legislation, eligible households have been able to receive up $30 per month toward internet service ($75 on certain Tribal lands). In February, the subsidy program stopped accepting new applications and April will be the last month of full subsidies, according to the Federal Communications Commission. 

The study indicates that only 56.4% of households in the bottom quartile of the income distribution (less than $53,150) had fixed broadband service in 2018, compared with 83.2% of households in the top income distribution ($92,700 or more). Rojas and Espín estimate that direct-to-consumer subsidies prescribed in the infrastructure law would increase internet adoption in the bottom income quartile to 79.4%.

“I think the policy of devoting some resources to expand the broadband network is important but recalibrating it and devoting more resources to subsidies would make it even more effective,” Rojas says.

The study recommends a variable subsidy based on income, where households with the lowest income receive a larger subsidy. It also suggests that the government subsidize satellite internet service to households in remote areas, rather than funding costly expansions of terrestrial broadband to connect relatively few homes. 

Rojas says money that the government does earmark for internet infrastructure upgrades would be better spent on improving internet speeds and network capacity in urban areas with high population density.

The study was based on a comprehensive dataset of household internet usage and prices used to estimate broadband demand for the years 2016 to 2018, the latest periods for which complete information was available. Rojas and Espín then used the estimates to model policies focusing on subsidies to low-income households versus network expansion to unserved areas.

The full study, “Bridging the digital divide in the US,” appears in the International Journal of Industrial Organization. It was recently highlighted by the European Association for Research in Industrial Economics.