Comcast and Netflix have come to a deal on ‘peering.’ Basically, what this means is that Netflix will pay Comcast to ensure that Netflix’s customers get good internet service for Netflix. While apparently, Netflix is not getting special treatment under the deal, as we economists all know, it is what not having a deal would have done that has driven this. So while the deal itself may be ‘neutral,’ the counterfactual (without the deal) may well not have. Of course, that counterfactual could also have involved Netflix complaining with regard to the proposed Comcast-Time Warner merger and so ‘who is paying who’ is perhaps a little less clear here. (Note also that my statements here are no different given that Netflix may just be substituting Comcast for another provider in accessing Comcast customers; the issue is not so much that but whether there is something ‘special’ in the Comcast-Netflix arrangement that may not have happened without the arrangement).
Tyler Cowen argues that this arrangement is not something we should be worried about per se — even if we are worried about Comcast’s market power. His argument invokes Coase — loosely, if negotiations between internet service providers, content providers and customers are efficient, then regardless of whether Netflix pays Comcast or not, the same outcome will arise. What he means is that you might be worried that Netflix is facing higher costs and that these will be passed on to its consumers in the form of a higher charge. That’s true but at the same time, Comcast has either lower costs or another revenue stream related to signing up those customers and so it will reduce charges to them. Since consumers only care about the sum of Comcast and Netflix charges, viva la difference!
Cowen is essentially correct in that argument although it could be a little more precise. Specifically, there are two relationships where we worry about special treatment or price discrimination. First, there is the relationship between ISPs and content providers whereby neutrality is defined as content providers — regardless of the ‘type’ of traffic (and not necessarily the quantity of it) — pay the same price to ISPs. Second, there is the relationship between ISPs and their customers that do not have the customers paying different prices to access different types of content. Weak net neutrality involves preventing just one of these types of discrimination. Strong net neutrality involves preventing both.
In this (really preliminary note) I work out the Coasian implications of this. I wrote it a few years ago thinking I might do a survey but didn’t get around to it. (Hmm, maybe a Kickstarter is in order?) The note shows that weak net neutrality does nothing compared to no regulation. In other words, so long as Comcast doesn’t discriminate in terms of what customers can choose, then this peering deal does not matter at all for outcomes or profits of content providers and ISPs. Strong net neutrality does change things. First, it may involve a distortion to allocative efficiency if there are different costs to the ISP of servicing different types of content providers because it prevents cost signals from being seen by consumers. It is not at all clear that is the case. Second, it results in a redistribution of profits from ISPs to content providers. The extent of that redistribution depends on the market power of ISPs. What this means is that strong net neutrality would stimulate incentives to invest in content.
The point is that Coase does not hold fully but is informative. The dynamic investments cannot easily be part of these negotiations and so static redistribution of profits matters.
Of course, Matthew Yglesias has argued that we can restore this potential inefficiency another way: by having content providers buy or become ISPs. He calls for Apple to start getting into the broadband game. Of course, Google are experimenting with this but it is far from clear that they are doing that to actually invest, threaten to invest or, most plausibly, just understand how the entire ‘pipe’ works to improve consumer experience. I can see how Yglesias might come to this conclusion from the textbook application of Coase but I think we can all agree that content providers should stick to their knitting and that really Apple has no capabilities to enter into broadband no matter how much we love the other stuff they do.