What is the market demand for broadband by U.S. households? It is well known that demand for broadband Internet has grown substantially in the last decade. That growth has driven the demand for computers, routers, fiber-optic cable and much more. Our research puts some statistical numbers behind this widely recognized growth.2 We estimated consumer willingness-to-pay (WTP) for broadband service in late 2009 and early 2010. Our WTP estimates show a high valuation for broadband for average consumers, and even for inexperienced users.
There are many proposals for increasing the deployment and use of broadband infrastructure, most importantly the FCC's National Broadband Plan (see http://www.broadband.gov/plan/). Formal cost-benefit evaluation of these proposals requires, among other things, some understanding of the potential benefits from more widespread access to broadband Internet service. Our study provides more understanding of those benefits.
Choice experiments were used to estimate household preferences for the different features that comprise Internet service, including the basic service features speed and reliability, in addition to recent and hypothetical uses of the Internet. A carefully designed choice experiment manipulates the features for a series of hypothetical Internet services to obtain the variation in the data needed to estimate the parameters precisely. Choice experiments also allow us to estimate the marginal utilities (satisfaction) for service features that are not currently available or are only available in limited geographical areas.
The WTP for Internet service is estimated with data from a nationally representative online survey with more than 6,200 respondents. Respondents face repeated discrete-choice experiments. Respondents choose from a pair of hypothetical Internet service alternatives, labeled A and B. A follow-up question is then presented that asks respondents to make an additional choice between their hypothetical choiceA or Band their "status quo" Internet service they now experience at their homes.
The hypothetical alternatives differ by the levels of the three Internet features: cost, speed, and reliability, and one of five Internet activities: the ability to designate some uploads and downloads as high priority (labeled "Priority" in the accompanying table); the ability to interact with their health professionals and view their records online ("Telehealth"); the ability to connect remotely regardless of WiFi availability ("Mobile Laptop"); a built-in Skype-like service ("Videophone"); and a video-on-demand movie service for viewing full-length movies ("Movie Rental").
The typical U.S. household is involved in Internet activities and applications at home that do not require very fast download and upload speeds.
We estimate the WTP for a subsample of experienced users, as well as for a subsample of inexperienced users (that is, those with less than 12 months of online experience). This difference provides some indication of valuations for households that have recently connected to the Internet.
Speed and reliability are important features of Internet service with consumers willing to pay approximately $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. These estimates indicate a very fast service is worth only about $3 more than fast service. The latter finding requires an explanation. As it turns out, the typical U.S. household is involved in Internet activities and applications at home that do not require very fast download and upload speeds, such as reading and writing email or light Web usage. Households are sensitive to other aspects of their service. They are also willing to pay an additional $6 per month for Priority, $4 for Telehealth, $5 for Videophone, and $3 for Movie Rental. Mobile Laptop is not valued by respondents.
For comparison with the findings of previous studies, we calculate the own-price elasticity of the demand for broadband Internet of 0.44, which is less elastic than estimates using older data. This means consumers are less willing now to give up broadband service if prices increase because they value it more highly than they had previously. In many households the cost of broadband is treated as a necessary expenditure.
Household valuations for Internet service may vary with the number of years the household has been connected to the Internet. Inexperienced households with slow-speed connections are willing to pay about $16$17 per month for an improvement from slow to fast speed, but they do not value an improvement from fast to very fast speed. Inexperienced households with a high-speed connection are willing to pay approximately $26$27 per month for an improvement from slow to fast speed.
To better illustrate our results, the accompanying table shows four levels of Internet service. Basic service has fast speed and less reliable service. Reliable Internet service has fast speed and very reliable service. Premium service has fast speed, very reliable service, and the ability to designate some downloads as high priority. Premium Plus service has fast speed, very reliable service, plus all other activities bundled into the service. We then assume that household valuation for a dialup-like service: less reliable, slow, and with no other special activities, is $14 per month.
The estimates shown in the table suggest the representative household would be willing to pay $59 per month for a Basic service, $79 for a Reliable service, $85 for a Premium service, and $98 for a Premium Plus service. The bottom half of the table shows that an inexperienced household with a slow connection would be willing to pay $31 per month for a Basic service, $41 for a Reliable service, $59 for a Premium service, and $71 for a Premium Plus service.
Experienced users are more aware of the full range of economic, entertainment, information, and social benefits the Web has to offer.
These estimates suggest the value of broadband has changed in the last decade. In 2003 the representative household was willing to pay approximately $46 per month for Reliable broadband service3 compared to about $79 in 2010. Given that the price of broadband has not changed much in this period, these estimates suggest experienced households get more for their money today than in the recent past. Using the language of economics, we would say that monthly consumer surplus per household has increased substantially between 2003 and 2010.
Choice experiments show that reliability and speed are important features of Internet service, but that very fast Internet service is not worth much more to households than fast service at this point in time. Using these results, we calculate that a representative household would be willing to pay about $59 per month for Basic Internet service. In contrast, an inexperienced household with a slow connection would be willing to pay about $31 per month for Basic Internet service.
One interpretation of these results is that experienced users are more aware of the full range of economic, entertainment, information, and social benefits that the Web has to offer. Inexperienced users may also have less technical ability when using high-technology goods and service. As such, they would be relatively less productive when using the Internet to produce household income and/or savings in time. The large increase in WTP indicates the Internet is not only becoming a much more important part of people's lives, possibly because they are learning more about its capabilities, but also that its capabilities have increased through new applications and increased reliability.
Our data does not allow us to differentiate two potential explanationshouseholds with a higher WTP for Internet subscribed sooner or experience increases valuation. Likely both contribute to the difference. The FCC's National Broadband Plan discusses a number of proposals for subsidizing broadband. Our findings stress the importance of a well-targeted subsidy program over a poorly targeted one.
If experience causes increased valuation, then correctly targeted private or public programs have the potential to increase overall broadband penetration in the U.S. (see Ackerberg et al.1). These programs could educate households about the benefits from broadband (for example, digital literacy training), expose households to the broadband experience (for example, public access), and/or directly support the initial take-up of broadband (for example, discounted service and/or hookup fees).
On the other hand, the high WTP for broadband indicates that large and poorly targeted subsidy programs could be quite wasteful. Subsidizing broadband, especially for non-poor households in higher-cost areas, may not be an efficient use of government spending because many such households would have been willing and able to pay market prices for broadband. As a result, the subsidy would have no effect on broadband adoption for these households.
1. Ackerberg, D., Riordan, M., Rosston, G., and Wimmer, B. Low-income demand for local telephone service: Effects of lifeline and linkup. SIEPR Discussion Paper, 0847. Stanford University, Palo Alto, 2009.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2011 ACM, Inc.