How Do ISP Data Caps Affect Subscribers?

20 Pages Posted: 28 Mar 2013 Last revised: 14 Sep 2015

Wei Dai

University of California, Irvine

Scott Jordan

University of California, Irvine - Donald Bren School of Information and Computer Sciences

Date Written: March 27, 2013

Abstract

It has become common for ISPs to place caps on the monthly usage of cellular and broadband data plans. ISPs commonly claim that caps benefit most users. They cite statistics that show that a small percentage of users consume a high percentage of network capacity, and claim that flat-rate pricing is unfair to the majority of users. They claim that caps affect only heavy users, result in lower tier prices, and increase the incentive for ISPs to add capacity to the network.

In contrast, many public interest groups claim that caps hurt most users. They claim that caps discourage the use of certain applications and encourage a climate of scarcity. They claim that caps and their corresponding overage fees do not correspond to the cost for network capacity, and that their use may decrease an ISP’s incentive to add capacity.

There is a vigorous debate over the use of caps. A Senate bill would require the FCC to evaluate data caps to determine whether they reasonably limit network congestion without unnecessarily restricting Internet use.

However, there is little academic literature that addresses the impact of data caps. In this paper, we propose models to evaluate the impact of data caps upon subscribers and ISPs. The model includes the critical elements of both Internet architecture and economic motivations. We model user utility for the two dominant applications: web browsing and video streaming. Utility is represented as a function of a user’s relative willingness-to-pay, the time devoted to each application per month, and the performance of each application. User surplus is expressed as utility minus the opportunity cost of the time devoted. We model a monopolist ISP that sets tier prices, tier rates, network capacity, data caps, and overage charges in order to maximize subscription plus overage revenue minus the cost of capacity. Combining these two models, users choose a tier and decide how much time to devote, and correspondingly whether to violate the cap.

The analysis is based on optimization methods. We show how users fall into three categories: those unaffected by a cap, those who are capped but do not choose to exceed the cap, and those who exceed the cap and pay overage charges. We give closed form expressions for which users fall into each category, based on a user’s willingness-to-pay, opportunity cost, and the level of the cap. We then examine a monopolist’s use of caps. We compare the optimal tier rates, tier prices, and network capacity without caps to the same quantities when caps are added. We first consider the case in which an ISP institutes caps in order to ensure that heavy users pay an amount equal to the cost of their usage, and show how an ISP will set the cap and which users it will affect. We then consider the case in which an ISP sets caps and overage fees to maximize profit. We show that in this case, an ISP will increase the tier rate and decrease the tier price. We also give closed form expressions, under certain assumptions, for which users will be hurt and which will benefit from the cap and changes in tier rate and price.

Finally, we give numerical results when users have a constant elasticity of demand. We illustrate how the tier rate, tier price, cap, and overage fees vary with the standard deviation in Internet usage amongst subscribers. We also illustrate the increase in ISP profit when caps are used, the corresponding change in user surplus, and the change in total social welfare.

Keywords: Pricing, Utility

Suggested Citation

Dai, Wei and Jordan, Scott, How Do ISP Data Caps Affect Subscribers? (March 27, 2013). TPRC 41: The 41st Research Conference on Communication, Information and Internet Policy 2013. Available at SSRN: https://ssrn.com/abstract=2240424 or http://dx.doi.org/10.2139/ssrn.2240424

Wei Dai

University of California, Irvine ( email )

Campus Drive
Irvine, CA 62697-3125
United States

Scott Jordan (Contact Author)

University of California, Irvine - Donald Bren School of Information and Computer Sciences ( email )

Bren Hall
Irvine, CA 92697-3440
United States
(949) 824-2177 (Phone)

Paper statistics

Downloads
192
Rank
129,552
Abstract Views
1,393