Platform Pricing and Foreclosure: Evidence from an Internet Service Provider (Job Market Paper)

Abstract: This paper studies the joint pricing decisions of internet service providers (ISPs), who sell broadband internet access and pay TV subscriptions. I estimate a model of consumer choice over ISP and third-party online video subscriptions (such as Netflix) using novel household-level data containing online video usage information at the hourly level. I find that the elasticity of demand for internet access is -0.99, and that TV elasticities are between -6.45 and -3.13, implying much higher margins for internet than TV. When access to online video is removed from the average household’s preferred bundle of subscriptions, willingness-to-pay falls by 20%, or $38. Next, I use a model of bundle pricing to study the implications of alternative ISP strategies for pricing internet content. I find that foreclosure of online video is not profitable due to (i): the large contribution of online video access to internet valuations and (ii): low ISP margins on TV relative to internet. When given the option to set add-on prices for access to online video, the ISP chooses positive prices, and new surplus is unlocked through substitution from online video to TV.

Steering Incentives on Platforms: Evidence from the Telecommunications Industry (with Brian McManus, Aviv Nevo, and and Jonathan W. Williams)

Abstract: Internet Service Providers (ISPs) offer both TV packages and access to the internet, which allows customers to view streaming video that competes with TV and can increase ISPs’ network costs. This provides ISPs with an incentive to steer its customers toward more profitable subscriptions and viewing choices. We study these incentives using a unique dataset that documents individual consumers’ internet usage choices and TV subscriptions, all in a setting where an ISP introduced a new policy of internet usage allowances and overage charges. We extend the textbook monopoly bundling model to describe the policy’s main effects, including how ISPs’ incentives to encourage or discourage streaming video varies with its ability to steer consumers. We then analyze empirically the price policy’s impact on consumers’ choices. Consistent with our theoretical model, the new policy steered internet-only consumers into bundled TV and internet subscriptions; this effect was greatest for heavy users of streaming services most similar to conventional TV. Internet usage growth was curtailed for consumers of all types, regardless of choices about subscriptions, and it reduced usage of and subscriptions to third-party streaming video services. Finally, we discuss the implications of these findings for antitrust and regulatory issues in the telecommunications industry, including net neutrality.

The Unbundling of the Telecommunications Industry: Evidence from Cord-cutting (with Jacob Malone, Aviv Nevo, and Jonathan W. Williams)

Abstract: We study the impact of the introduction of popular over-the-top video (OTTV) services on the telecommunications industry using household-level panel data. Poorer, younger, and smaller households that watch less TV and heavily engage broadband services are most likely to drop TV services during this period. Upon “cutting the cord”, which reduces an ISP’s revenue by an average of 50%, households increase total usage by 31% with OTTV responsible for 60% of the increase. Among common OTTV services, usage of Netflix, Hulu, and Sling TV increase 25%, 72%, and 662%, respectively. We discuss the implications of our findings for ongoing policy and regulatory debates in the telecommunications industry, and strategies firms are likely to employ to offset the loss of revenue and increase in cost associated with cord-cutting.