Rosston’s Nightmare: Broadband Prices Go Up, Up, Up

Prices in the U.S. have about doubled since 2010, when the Stanford economist’s paper predicted people want broadband so much doubling was possible.

Morgan Stanley found average prices of $66 in 2017, up 12%. Hidden fees often raise that. When Rosston wrote in 2010, most prices were $30-$45. An economist reading Greg’s paper could predict prices likely would double if the companies could just signal each other to raise prices. The paper was written for the Broadband Plan where the staff could see large increases coming. Since costs of delivering broadband are down, regular price increases imply we only have weak competition.

Rosston found people were willing to pay ~$45 more for better broadband, which would increase if broadband became even more essential. (It has.) The only way to stop that was strong competition or regulation. Plan leader Blair Levin could see that, but he told me action for either was blocked at a higher level. I guessed that was Larry Summers in the White House, but never could confirm that.

The disease of monopoly-like price increases is spreading to Germany and Britain. Mike Fries and Liberty Global are doing regular price increases and Deutsche Telekom is following. In Britain, BT has just raised prices and I’m watching the others to see if they match. The strategy is like the Americans. Increases that aren’t so high the regulator would step in but enough to lead to much higher prices in a few years.

Half the U.S. has only one choice at 25 megabits or better. Few of the rest have more than two.

Few of the rest have more than two, which almost always signal each other to raise prices. The resulting are among the highest in the developed world. Companies like Comcast have been raising prices 3%-8% per year. That’s enough to make a huge difference in margins over time, but not enough for organized outrage or FCC action.

The Broadband Plan looked closely at whether wireless was likely to change this supply/demand situation. Even then, we could see wireless was a partial substitute. The FCC’s dream was that the new spectrum would bring in enough new providers to make a difference. Unfortunately, that dream looks to be going the way of weapons of mass distruction in Iraq.

Some projected wireless would take 10-25% of the market within five years, enough to have a (modest) influence on prices but not enough to bring them down. I tended towards that opinion.

I was wrong. Seven years later, it looks like well under 10% have gone wireless only in broadband. The speeds are now fine, averaging ~ 20 megabits down in much of the country. Until this year, the caps were too low, at 2-10 gigabytes a month. That’s less than an hour of TV a day, much less than the typical U.S. home watches.

In 2017, the four U.S. wireless companies are advertising “unlimited,” with a soft cap of 23-29 gigabytes. That’s more like it. Things are better in France; Jennie paid $23 for a month with a 100 gigabyte cap this spring. That’s enough for most people who only modestly watch video and do little downloading, although France still is adding landline customers.

Verizon estimates wireless cost per bit is coming down at 40% per year, meaning wireless offerings will get much more robust where competition is working.

Household Demand for Broadband Internet Service

TPRC 2010

47 Pages Posted: 19 Jan 2012

Gregory L. Rosston

Stanford Institute for Economic Policy Research

Scott Savage

University of Colorado at Boulder – Department of Economics

Donald M. Waldman

University of Colorado at Boulder – Department of Economics

Date Written: August 15, 2010


This paper uses data from a nationwide survey administered during late 2009 and early 2010 to estimate a random utility model of household preferences for broadband Internet service. Reliability and speed are important service characteristics: the representative household is willing to pay $20 per month for more reliable service; $45 for an improvement in speed from slow to fast; and $48 for an improvement in speed from slow to very fast. A representative household would be willing to pay $79 per month for a fast, reliable Internet service. Internet valuations increase with experience, and there has been an estimated two- to three-fold increase in consumer surplus per household between 2003 and 2010. If experience causes increased valuation, targeted programs that educate households about the benefits from broadband, expose households to the broadband experience and/or directly support the initial take-up of broadband have potential to increase overall penetration in the United States.

Full paper is free to download.



Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to top