It is the one year anniversary of the Open Internet Advisory Committee (as noted earlier). Today the committee issued a report of its work over the last year. You can access it here. Today’s post discusses the report about data Caps, which was written by the Economic Impacts working group.
I am a member of the committee and the Economic Impacts Working Group, and I like the work we did. I chair the group. “Chair” is a misleading title for what I really do, which is take notes of the groups’ discussions and transcribe them. Every now and again, I do a little more. As one of the members without any stakes in the outcome, occasionally I offer a synthesis or compromise between distinct views.
The report aims to analyze data caps in the context of the Open Internet Report and Order. The Open Internet Report and Order discusses usage-based pricing (UBP), but does not expressly mention data caps except by implication in that data caps can be considered a form of UBP. The Order left open the possibility of many experiments in business models and pricing.
Moreover, the Internet had evolved over time, and the Order anticipated that the Internet would continue to evolve in unexpected ways. The Order set up the advisory group to consider whether aspects of the Order remain consistent in its effects on the Internet as the Internet evolves, and it is in that spirit that this conversation was undertaken.
The report’s first goal was seemingly simple: clarify relevant terminology (e.g., cap, UBP, thresholds). That was harder than it looked. Many of advocacy pieces on the net chose to talk past one another. The report’s second goal was to identify a common fact-basis for discussion. We took steps in that direction. That took a bit of sleuthing, but the Internet makes many things possible.
The third and fourth goals were the hardest. We sought to analyze different perspectives, and identify unaddressed open questions. That was very hard, since it required getting consensus on topics where policy advocates are not accustomed to reaching consensus.
The Report concludes that there is considerable variance and experimentation in the market by ISPs. The fact base did not permit us to go much beyond that conclusion. It is difficult to interpret even the highest data caps and highest thresholds in the situations in which they arise, as there is no definitive public source on household usage per month, nor any other data to use as a benchmark. In addition, more than one benchmark was required, as usage varies depending on ISP and technology.
All public measurements show great skew in usage, and suggest that caps do not yet impact users other than the highest users. It is obvious that the present situation will not last forever, but the committee could not reach anything like a consensus on how quickly many households would approach certain utilization rates. (Side comment: Indeed, I was struck by the near absence of consensus among industry estimates).
So, at best, the committee could reach only tentative conclusions. Although caps do not seem to be affecting a large number of US users now, the situation may change in the future, as user habits, supplier experimentation, vendor policy, and applications all change.
Looking at stakeholders
The report also elaborates on many of the key concerns of three stake-holders prominently identified in the Order, namely, users, broadband providers, and edge-providers.
I believe this is the aspect of the report that will most intrigue my academic colleagues at other universities. The focus on distinct perspectives departs from many economic models of access in communications markets. These standard models bring with them many hidden assumptions from telephony that may or may not carry over to Internet markets. Those differences should motivate a variety of new thinking about the key economic factors at work.
The discussion begins with an analysis about users. It focuses on user understanding about perceptions of caps and thresholds. The report concludes that this topic may require future monitoring, especially given the importance of consumer education to user perceptions of caps and thresholds.
It is not yet apparent whether many issues of importance to users are a transitory or permanent concern. The experience of ISPs with providing customers with tools to monitor or control data usage could also be valuable for insights about the perceptions of caps by consumers. There is not much experience to tell, and we will have to wait and see.
The discussion about broadband providers focuses on many topics and questions: whether data caps, tiers and related forms of UBP may encourage end users nearing that cap to act efficiently; whether data caps, tiers and related forms of UBP may spur efficiency and innovation on the delivery of services; whether data caps, tiers and related forms of UBP may help manage network growth; whether data caps, tiers and related forms of UBP might encourage heavy users to change their usage, and if so, in what way; whether data caps may shape the future and conduct of other service providers (i.e. application developers).
As an aside, I think many academics will be struck by the variety and range of open questions. No single theory will address all the open questions, nor would any sensible analyst rely on any single theory. I also expect many academic economists will identify open questions that require new thinking.
The discussion about edge-providers considers how data caps, tiers and UBP can shape other providers of services in broadband ecosystem, e.g., entrepreneurs who provide applications, build web pages, and operate other services in the cloud. This part of the report identifies areas where ISPs and edge providers have different perspectives on open questions.
This section also examines competition policy for specialized services, which is a very hot topic (and this sub-group was not the only one to address aspects of this topic this year). In general, competition policy is concerned about situations where one firm, such as a broadband provider, supplies a service and also controls aspects affiliated with the cost, performance, and user-experience in a competing service, provided by an edge-provider. The report identifies how the ISP’s perspective and the edge provider’s perspective diverge on this topic. The report concludes the situation yields no easy answers in general, and, at a minimum, merits further monitoring.
In general, the committee concludes that these debates cannot be easily summarized in a brief set of bullets or summary paragraphs. The report contains many perspectives, as well as many open questions, and it identifies many issues that the FCC could further monitor.
Takeaways for an economist
I have already hinted that an academic might take-away a few interesting insights from this report. Let me highlight a few of these:
First, for those interested in regulatory policy, these policy issues do not look like telephone regulatory policy. That departure is important and interesting, and requires dropping of old views, and adoption of new thinking.
More specifically, and as noted on an earlier post, the working group did not make policy for data caps. Rather, the report mapped the policy landscape. It pointed out where the FCC or another consumer-protection policy-making body, such as FTC, might want to monitor events and where issues remained unsettled. It identified some choices too, and tried to identify bridges between general arguments and specific observable behavior and facts.
I believe we were the first organization to ever do something this comprehensive on this topic – namely, map the landscape by bringing all the arguments into one place. We did not settle anything, nor was it within our purview to settle anything. Rather, we attempted to move the conversation to a more productive place. I would hope that our report would motivate others to look at the issues with a new perspective.
Second, we began using a model for this setting that differs from the standard model that still dominates regulatory economics textbooks. The standard regulatory model examines when government has to intervene to prevent monopoly providers from exploiting consumers. In my experience with this committee, that model had very limited value in a few areas (e.g., consumer protection), but it did not provide a useful conceptual foundation for most of the discussion.
Instead, the report primarily used a model of three types of market participants – access providers, consumers, and edge-providers. Occasionally the report is concerned with distortions arising from monopoly power, but often that was not the only concern, or even the most central one. A big part of our analysis focused on the divergent interests of different participants. Many of those concerns would arise without or with monopoly power.
In other words, it is time to rethink the standard model for this context (and this is an important enough context to motivate it). That observation should interest many academic economists.
Third, it was striking how much uncertainty became embedded into the discussions. Don’t get me wrong. The presence of uncertainty, per se’, was not a surprise. Rather, the pervasiveness and its causes surprised me. Even though broadband markets are (so-called) mature markets in the US, growth in demand/utilization was hard to forecast, which is rather unusual for a so-called mature market. So too was changes in costs. Those inadequacies were exacerbated by the lack of clear forecasts about even the most basic elements of demand. All those were further made worse by the absence of data about the key issues and behaviors of users.
Fourth, there is plenty of room for new research efforts in academic economics. Let me take a moment to outline a few issues:
• It was striking how inadequate classic price discrimination theory sounds when more than just a monopolist and user care about prices. Can we have a theory of price discrimination with multiple players, and see what insight that generates?
• It was striking how few economic theories identified the key divergence in incentives between access providers and edge providers, and even if they did, few theories provided clear policy guidance. This seems like a very ripe area for basic applied micro economic theory.
• While plenty of classic price theory discusses price discrimination for supplying an input into the production of another firm, that is not precisely the situation in the Internet. Rather, one firm provides one complement to users, and a wide variety of firms provide a heterogeneous set of the other complements. What happens in that setting when one supplier of inputs price discriminates? That is a pretty ripe question.
• Most of all, despite the pervasive respect this industry has for data, it was surprising how little data made it into the public discussions.
I could go on, but let me finish with a few smaller open questions. Here are some questions that arose and require new economic research:
• What is the substitution between broadband and television, and other forms of mass market entertainment?
• What is the elasticity of substitution of data traffic between broadband and other traffic access points, such as between work and home, and between broadband access and other wireless devices?
• There has been little experimentation with usage based pricing with time-of-day elements. What would be user response in this setting? Since many of the heaviest data users are also computer-literate, one might expect some responsiveness from those users to the introduction of time-of-day options, but there is no way to know without trials.
• What is the elasticity of substitution of popular applications (such as email, YouTube, etc) across access points, and what does that tell us about the long run trends in demand for data?
• What is the economic shadow value of capacity utilization during different times of the day? Basic economics says that the cost of data to an access provider during times of low capacity utilization ought to be quite different from cost during peak times? Let’s get an estimate.
• Related to the last point, what costs do edge-providers with data-intensive applications really place on the system operators?
Looking forward to more!
3 Replies to “The Economic Policy of Data Caps”
I’m surprised you used the word “market” (for the US at least). If there was indeed a market for consumers then this stuff wouldn’t matter and it would sort itself out. In practise there is a very limited market, with many residential areas having a choice of two ISPs (cable and telco) if they are lucky.
Areas with more competition seem to have done best by separating the piping (cable, phone lines etc) from the provision of service over that piping.
As with net neutrality there are many flavors of data caps or usage based pricing and how they are perceived by different stake holders (not just 3).
To posit that telco and internet models are somehow different is a mistake.
In fact they all share the same immutable laws of supply and demand; governed by moore’s, metcalfe’s and zipf’s laws, among others, at every layer and boundary point.
As the author indicates, what’s needed is a framework which distills the issues objectively. That framework needs to account for a core and edge perspective, an upper and lower layer perspective and a market segment agnostic segment. By the latter I mean a method of understanding pricing and consumption on a marginal basis from the highest to lowest volume/density and not getting caught up in arbitrary and artificial market segment or geographic or content or technical boundaries.
The blueprints for such a framework exist from the competitive WAN, competitive data (aka Internet) and competitive wireless markets of the 1980s-90s; all of which sprang from open access approaches. It’s just that with consolidation and remonopolization of the sectors we’ve forgotten those lessons and unlearned actual history.