Effects of Market Structure on Broadband Quality in Local U.S. Residential Service Markets

Abstract

This paper investigates the effects of changes in market structure on residential service quality in local U.S. broadband markets. Our analysis focuses on identifying causal effects on quality from entry (or exit) by firms using both legacy and non-legacy broadband network technologies. We analyze a very large and highly disaggregated census block-level data set we have constructed using Federal Communication Commission (FCC) data on residential broadband services deployment in populated U.S. census blocks between December 2014 and December 2018. We employ a difference-in-differences framework combined with both spatial fixed effects and spatial instrumental variables. The latter econometric tools are used to address unobserved market heterogeneity and concerns about significant error in measured ISP counts for fixed wireless ISPs. Our results suggest that entry by legacy technology (cable and DSL) service providers has large impacts on available service quality in local broadband markets, effects which diminish rapidly with continued entry. Contrary to expectations, entry by a single new fiber ISP seems to have no impact on available quality; entry by two fiber ISPs, by contrast, raises maximum available quality by a modest but statistically significant 90 Mbps. Entry by a fixed wireless ISP appears to have had a large, negative impact on maximum available download speeds offered by incumbent legacy ISPs. Economic explanations for these results are discussed and evaluated. We establish that (i) both fixed wireless and fiber ISPs compete in sets of speed tiers that are distinct from those offered by legacy ISPs, and speculate that (ii) price competition (as opposed to quality competition) may be more intense in markets with fixed wireless entry.