Broadband Mergers and Dynamic Bargaining: An Application to Netflix

I measure how mergers in the market for broadband internet service affect short-run welfare. Mergers between internet service providers (ISPs) with non-overlapping markets may decrease welfare by increasing ISP bargaining leverage against content providers. However, study of this welfare channel has been stymied by a lack of data on interconnection fees between content and internet service providers. I estimate an industry model of demand, plan choice, pricing and interconnection bargaining using data on plan prices, consumer choice sets and bargaining delays between major U.S ISPs and the leading purveyor of streaming video content, Netflix. Intuitively, if delaying agreement over interconnection degrades quality of service to subscribers, then the opportunity cost of lost subscriptions identifies the fee. To map disagreement times and ISP competition into interconnection fees, I develop a multilateral dynamic bargaining model with asymmetric information. ISPs make take-it-or-leave it offers to learn about Netflix's benefit from interconnection, while simultaneously competing for subscribers who value Netflix quality of service. I structurally estimate the model and recover fixed interconnection fees ranging from 44 to 69 million USD. I find that a proposed merger between TimeWarner and Comcast that was challenged by the Federal Communications Commission would have slightly raised interconnection fees and bargaining length, reducing consumer welfare by 1.9 percent.