Market Structure and Media Diversity - Research Program on Digital ...
Market Structure and Media Diversity - Research Program on Digital ...
Market Structure and Media Diversity - Research Program on Digital ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
REPORT SERIES<br />
FALL 2012<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g><br />
Scott J. Savage<br />
D<strong>on</strong>ald M. Waldman<br />
Scott Hiller<br />
University of Colorado at Boulder<br />
www.twcresearchprogram.com
For more informati<strong>on</strong>:<br />
Fern<str<strong>on</strong>g>and</str<strong>on</strong>g>o Laguarda<br />
901 F Street, NW<br />
Suite 800<br />
Washingt<strong>on</strong>, DC 20004<br />
Ph<strong>on</strong>e: (202) 370-4245<br />
www.twcresearchprogram.com<br />
twitter.com/TWC_RP
Table of C<strong>on</strong>tents<br />
Foreword .................................................................................................................................................. 3<br />
By Fern<str<strong>on</strong>g>and</str<strong>on</strong>g>o R. Laguarda, Time Warner Cable<br />
Abstract .................................................................................................................................................... 4<br />
1. Introducti<strong>on</strong> ..................................................................................................................................... 5<br />
2. Empirical Approach ....................................................................................................................... 8<br />
2.1 Step One: The Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for News Services .................................................................................8<br />
2.2 Step Two: The Supply of News Services ....................................................................................10<br />
2.3 Step Three: Estimating C<strong>on</strong>sumer Benefits from a Change in <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> ...............11<br />
3. Data ...................................................................................................................................................12<br />
3.1 Experimental Design ....................................................................................................................12<br />
3.2 Survey Administrati<strong>on</strong> .................................................................................................................12<br />
3.3 <str<strong>on</strong>g>Media</str<strong>on</strong>g> Sources <str<strong>on</strong>g>and</str<strong>on</strong>g> News Service ...............................................................................................13<br />
3.4 <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> ...........................................................................................................................13<br />
4. Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> Estimates ....................................................................................................................... 18<br />
5. Impact <strong>on</strong> Welfare from a Change in <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> ....................................................22<br />
5.1 The Supply of News Services ......................................................................................................22<br />
5.2 <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> C<strong>on</strong>sumer Welfare .................................................................................24<br />
5.2.1 Algorithm .............................................................................................................................24<br />
5.2.2 Estimates of C<strong>on</strong>sumer Welfare ........................................................................................24<br />
6. C<strong>on</strong>clusi<strong>on</strong>s ................................................................................................................................... 27<br />
Endnotes ................................................................................................................................................28<br />
References ............................................................................................................................................ 30<br />
Acknowledgements ........................................................................................................................... 33<br />
About the Authors ...............................................................................................................................34<br />
Key C<strong>on</strong>tact Informati<strong>on</strong> ...................................................................................................................36<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 1
2<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Foreword<br />
By Fern<str<strong>on</strong>g>and</str<strong>on</strong>g>o R. Laguarda, Time Warner Cable<br />
We have just lived through another presidential electi<strong>on</strong> cycle <str<strong>on</strong>g>and</str<strong>on</strong>g> <strong>on</strong>e thing<br />
is clear: media plays a critical role in the democratic process. Even in the<br />
Internet age, strengthening local broadcast televisi<strong>on</strong>, particularly the provisi<strong>on</strong><br />
of local news, remains a priority for communicati<strong>on</strong>s policy-makers.<br />
Yet, by allocating large amounts of spectrum <strong>on</strong> a geographic basis that<br />
ensured few licenses in each local market, lawmakers created a structure<br />
where local competiti<strong>on</strong> was purposefully limited. Today, c<strong>on</strong>sumers have<br />
a much wider range of choices for local news <str<strong>on</strong>g>and</str<strong>on</strong>g> informati<strong>on</strong> than they<br />
did back when broadcast licenses were first assigned. But media ownership<br />
remains the subject of regulati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the extent to which these rules<br />
promote the availability of local c<strong>on</strong>tent c<strong>on</strong>tinues to be an important policy questi<strong>on</strong>.<br />
This report, by Scott Savage, D<strong>on</strong>ald Waldman, <str<strong>on</strong>g>and</str<strong>on</strong>g> Scott Hiller, all from the University of<br />
Colorado at Boulder, helps to begin to answer that questi<strong>on</strong> by examining how market structure<br />
affects the availability of local news <str<strong>on</strong>g>and</str<strong>on</strong>g> informati<strong>on</strong> c<strong>on</strong>tent to c<strong>on</strong>sumers.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> structure itself is the result of government acti<strong>on</strong> to allocate broadcast spectrum. That<br />
structure changes as stati<strong>on</strong>s c<strong>on</strong>solidate, merge, or partially merge their news operati<strong>on</strong>s—as<br />
increasingly happens today. In this report, the authors develop an innovative approach to measure<br />
c<strong>on</strong>sumer preferences for local news. Their approach includes using data obtained by a nati<strong>on</strong>wide<br />
survey of U.S. households that showed diversity of opini<strong>on</strong>, coverage of local events, <str<strong>on</strong>g>and</str<strong>on</strong>g> multicultural<br />
issues are important characteristics of local news services.<br />
By measuring the “value” of these characteristics, the authors propose to study what a reducti<strong>on</strong> in<br />
the number of independent televisi<strong>on</strong> voices in the marketplace means for c<strong>on</strong>sumers. Of course, the<br />
numerical value the authors propose is intended to represent a dollar amount, or “willingness to pay”<br />
for certain types of c<strong>on</strong>tent. They d<strong>on</strong>’t expect to determine the value of “diversity” as a cultural or<br />
political objective. Instead, they are attempting to indicate the relative magnitude of the impact of a<br />
reducti<strong>on</strong> in this type of program c<strong>on</strong>tent <strong>on</strong> c<strong>on</strong>sumers when market structure changes. The results<br />
are informative for policymakers c<strong>on</strong>sidering the impact of rules designed for a media marketplace<br />
that has evolved from analog to digital <str<strong>on</strong>g>and</str<strong>on</strong>g> is moving toward a fully Internet-enriched envir<strong>on</strong>ment.<br />
When we launched the Time Warner Cable <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong> Communicati<strong>on</strong>s, we<br />
hoped to fund scholarship <strong>on</strong> topics that assessed <str<strong>on</strong>g>and</str<strong>on</strong>g> challenged status quo assumpti<strong>on</strong>s. This<br />
report provides insight into evaluating regulati<strong>on</strong>s for media ownership <str<strong>on</strong>g>and</str<strong>on</strong>g> televisi<strong>on</strong> programming<br />
intended to promote localism. In particular, there is a significant impact <strong>on</strong> c<strong>on</strong>sumer welfare<br />
from the reducti<strong>on</strong> of diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism following a decrease in the number of independent<br />
televisi<strong>on</strong> stati<strong>on</strong>s in a market. Moreover, in a “virtual merger,” where stati<strong>on</strong>s share news operati<strong>on</strong>s,<br />
there are no benefits to local c<strong>on</strong>sumers offsetting the loss of program c<strong>on</strong>tent. Policymakers<br />
struggling with legacy rules designed to protect traditi<strong>on</strong>al broadcast media in an increasingly<br />
diverse <str<strong>on</strong>g>and</str<strong>on</strong>g> competitive digital envir<strong>on</strong>ment should c<strong>on</strong>sider these findings as they weigh how<br />
c<strong>on</strong>sumers are best served.<br />
We hope this report stimulates debate <str<strong>on</strong>g>and</str<strong>on</strong>g> encourages a more thoughtful policy discussi<strong>on</strong>. As<br />
always, we look forward to your comments <str<strong>on</strong>g>and</str<strong>on</strong>g> feedback.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 3
4<br />
Abstract<br />
We estimate dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for local news service, described as the offerings from newspapers, radio,<br />
televisi<strong>on</strong>, the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es. Results show that the representative c<strong>on</strong>sumer values<br />
diversity in news reporting, more coverage of multicultural issues, <str<strong>on</strong>g>and</str<strong>on</strong>g> more informati<strong>on</strong> <strong>on</strong> community<br />
news. About two-thirds of c<strong>on</strong>sumers have distaste for advertising, which likely reflects<br />
their c<strong>on</strong>sumpti<strong>on</strong> of general, all-purpose advertising delivered by traditi<strong>on</strong>al media. Dem<str<strong>on</strong>g>and</str<strong>on</strong>g><br />
estimates are used to calculate the impact <strong>on</strong> c<strong>on</strong>sumer welfare from a marginal decrease in the<br />
number of independent televisi<strong>on</strong> stati<strong>on</strong>s that lowers the amount of diversity, multiculturalism,<br />
community news <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising in the market. Welfare decreases, but the losses are smaller in<br />
large markets. For example, small-market c<strong>on</strong>sumers lose $54 milli<strong>on</strong> annually while large-market<br />
c<strong>on</strong>sumers lose $16 milli<strong>on</strong>. If the change in market structure occurs in all markets, total losses<br />
nati<strong>on</strong>wide would be about $830 milli<strong>on</strong>.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
1. Introducti<strong>on</strong><br />
Informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs can raise political awareness <str<strong>on</strong>g>and</str<strong>on</strong>g> promote a range of ideas.<br />
On the assumpti<strong>on</strong> that unregulated media markets supply too little variety, many societies have<br />
charged regulators with ensuring sufficient opportunities for different, new <str<strong>on</strong>g>and</str<strong>on</strong>g> independent viewpoints<br />
(which we shall refer to as diversity), <str<strong>on</strong>g>and</str<strong>on</strong>g> with requiring that media resp<strong>on</strong>d to the interests of<br />
their local communities (localism).<br />
In the United States, the Federal Communicati<strong>on</strong>s Commissi<strong>on</strong> (FCC) traditi<strong>on</strong>ally has limited<br />
comm<strong>on</strong>- <str<strong>on</strong>g>and</str<strong>on</strong>g> cross-ownership of newspapers, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> televisi<strong>on</strong> (TV) stati<strong>on</strong>s. Recently, the<br />
FCC relaxed ownership rules <str<strong>on</strong>g>and</str<strong>on</strong>g> refocused its attenti<strong>on</strong> <strong>on</strong> market forces—for example, c<strong>on</strong>sumer<br />
preferences <str<strong>on</strong>g>and</str<strong>on</strong>g> new media, such as satellite, the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es—to meet its diversity<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> localism goals. Given the increased choices available through new media, supporters of greater<br />
ownership c<strong>on</strong>centrati<strong>on</strong> argue that traditi<strong>on</strong>al media should be free to c<strong>on</strong>solidate <str<strong>on</strong>g>and</str<strong>on</strong>g> use the<br />
efficiencies to provide more diverse <str<strong>on</strong>g>and</str<strong>on</strong>g> local news programming. Opp<strong>on</strong>ents questi<strong>on</strong> whether such<br />
efficiencies are achievable, <str<strong>on</strong>g>and</str<strong>on</strong>g> argue that large, c<strong>on</strong>solidated media corporati<strong>on</strong>s are not flexible<br />
enough to serve the interests of local <str<strong>on</strong>g>and</str<strong>on</strong>g> minority communities.<br />
Evaluati<strong>on</strong> of these arguments requires, am<strong>on</strong>g other things, measurement of the expected societal<br />
benefits that would arise from increased media diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism, <str<strong>on</strong>g>and</str<strong>on</strong>g> how these benefits<br />
change with regulatory interventi<strong>on</strong>s that shape media market structure. In this paper, we estimate<br />
c<strong>on</strong>sumer preferences, with regard to selected characteristics, for their local news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs<br />
service (“news service”) described as the offerings from newspapers, radio, TV, the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
smartph<strong>on</strong>es.<br />
The selected characteristics of news service are:<br />
• diversity of opini<strong>on</strong> in the reporting of informati<strong>on</strong>,<br />
• the amount of informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events,<br />
• coverage of multicultural issues, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
• the amount of space or time devoted to advertising.<br />
We use our dem<str<strong>on</strong>g>and</str<strong>on</strong>g> estimates to calculate the impact <strong>on</strong> c<strong>on</strong>sumer welfare from a change in<br />
media market structure that reduces the number of independent TV stati<strong>on</strong>s in the market. 1<br />
Specifically, we employ the willingness-to-pay (WTP) c<strong>on</strong>struct to measure the difference in<br />
c<strong>on</strong>sumer welfare effects between the news service supplied before the change in market structure<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> the service supplied after the change. We focus <strong>on</strong> broadcast TV stati<strong>on</strong>s because they are<br />
the main source of news for most households <str<strong>on</strong>g>and</str<strong>on</strong>g> because the FCC has direct oversight of their<br />
ownership structure. 2 By relating c<strong>on</strong>sumer valuati<strong>on</strong>s of news service to a measure of TV market<br />
structure, it is possible to indirectly assess the extent to which ownership rules address the policy<br />
goals of diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism.<br />
Note: The views expressed are those of the authors <str<strong>on</strong>g>and</str<strong>on</strong>g> not necessarily those of Time Warner Cable, the Time Warner Cable <str<strong>on</strong>g>Research</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong> Communicati<strong>on</strong>s, or the Federal Communicati<strong>on</strong>s Commissi<strong>on</strong> (FCC).<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 5
6<br />
We estimate our dem<str<strong>on</strong>g>and</str<strong>on</strong>g> model with discrete choice data obtained from a nati<strong>on</strong>wide survey of<br />
U.S. households during March 2011. Results show that diversity of opini<strong>on</strong>, community news,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> advertising are important characteristics of local news services. The representative c<strong>on</strong>sumer<br />
is willing to pay from $21 to $25 per m<strong>on</strong>th for an increase in diversity of opini<strong>on</strong> (<str<strong>on</strong>g>and</str<strong>on</strong>g> approximately<br />
the same for community news) from a low to a medium level (defined in Table 1), but <strong>on</strong>ly<br />
an additi<strong>on</strong>al $6 to $7 to move to a high level of diversity of opini<strong>on</strong> (or community news). The<br />
representative c<strong>on</strong>sumer also values a low-to-medium improvement in informati<strong>on</strong> that reflects<br />
the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities ($7) more than an improvement from low to high ($4).<br />
Many c<strong>on</strong>sumers have distaste for advertising, which likely reflects their c<strong>on</strong>sumpti<strong>on</strong> of general,<br />
all-purpose advertising delivered by traditi<strong>on</strong>al media such as radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV. The representative<br />
c<strong>on</strong>sumer is willing to pay about $5 to avoid a movement from a low to a medium level of advertising,<br />
but a much higher amount of $16 to avoid a movement from a medium to a high level.<br />
Using FCC (2011) data <strong>on</strong> media market structure, we present evidence that indicates the amount<br />
of diversity, localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising in the news services supplied to c<strong>on</strong>sumers is lower following<br />
a marginal decrease in the number of independent TV stati<strong>on</strong>s. As a result, the average “small<br />
market” (i.e., five or fewer TV stati<strong>on</strong>s) c<strong>on</strong>sumer loses $0.99 per m<strong>on</strong>th, whereas the average<br />
“large market” (i.e., 20 or more TV stati<strong>on</strong>s) c<strong>on</strong>sumer loses $0.44 per m<strong>on</strong>th. These losses are<br />
equivalent to $54 milli<strong>on</strong> annually for all small-market households in the U.S. <str<strong>on</strong>g>and</str<strong>on</strong>g> $16 milli<strong>on</strong> for<br />
large-market households. If the change in market structure occurs in all markets, aggregate losses<br />
nati<strong>on</strong>wide would be about $830 milli<strong>on</strong>.<br />
Other studies have measured the relati<strong>on</strong>ship between informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> market structure. However, these studies measure supply from just <strong>on</strong>e of the media sources<br />
that make up the c<strong>on</strong>sumer’s news service; for example, Milyo (2007), Gentzkow (2007) <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
Gentzkow <str<strong>on</strong>g>and</str<strong>on</strong>g> Shapiro (2010) for newspapers, <str<strong>on</strong>g>and</str<strong>on</strong>g> Siegelman <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldfogel (2001) <str<strong>on</strong>g>and</str<strong>on</strong>g> Crawford<br />
(2007) for radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV. Our research also is related to studies that quantify the relati<strong>on</strong>ship<br />
between quality <str<strong>on</strong>g>and</str<strong>on</strong>g> market structure for different industries. For example, Mazzeo (2003) shows<br />
that average flight delays are l<strong>on</strong>ger in more c<strong>on</strong>centrated airline markets. Goolsbee <str<strong>on</strong>g>and</str<strong>on</strong>g> Petrin<br />
(2004) estimate that cable TV channel capacity, number of over-the-air channels <str<strong>on</strong>g>and</str<strong>on</strong>g> number of<br />
premium movie channels increased in resp<strong>on</strong>se to satellite entry. Matsa (2011) finds that supermarkets<br />
facing more intense competiti<strong>on</strong> have more products available <strong>on</strong> their shelves, while<br />
Olivares <str<strong>on</strong>g>and</str<strong>on</strong>g> Cach<strong>on</strong> (2009) show that the inventories of General Motors dealerships increase with<br />
the number of competitors. In c<strong>on</strong>trast, Domberger <str<strong>on</strong>g>and</str<strong>on</strong>g> Sherr (1989) find no correlati<strong>on</strong> between<br />
the threat of new entry <str<strong>on</strong>g>and</str<strong>on</strong>g> a customer’s satisfacti<strong>on</strong> with his/her attorney used for home purchases.<br />
3 Since we measure the change in market structure by reducing the number of independent<br />
TV stati<strong>on</strong>s, our paper also is related to structural models of differentiated oligopoly that predict<br />
the price effects from a simulated merger; for example, Nevo (2000) for breakfast cereals, Pinske<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Slade (2004) for U.K. brewing, <str<strong>on</strong>g>and</str<strong>on</strong>g> Ivaldi <str<strong>on</strong>g>and</str<strong>on</strong>g> Verboven (2005) for car manufacturing.<br />
Relative to these literatures, our study makes several c<strong>on</strong>tributi<strong>on</strong>s.<br />
• First, we offer new evidence from media markets by examining the c<strong>on</strong>sumer welfare effects<br />
from a news service bundled from newspapers, radio, TV, the Internet, <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es.<br />
• Sec<strong>on</strong>d, the predicti<strong>on</strong> of n<strong>on</strong>-price effects appears to be novel in the simulated merger<br />
literature.<br />
• Finally, by looking at a vector of n<strong>on</strong>-price effects we are able to document a new <str<strong>on</strong>g>and</str<strong>on</strong>g> notable<br />
tradeoff between the diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism characteristics of news service <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
That is, the amount of diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism declines following a decrease in the number of<br />
independent TV stati<strong>on</strong>s. This decline is a cost to the typical c<strong>on</strong>sumer—but the amount of advertising<br />
declines at the same time, which is a benefit to the typical c<strong>on</strong>sumer. This finding should<br />
be significant to antitrust officials <str<strong>on</strong>g>and</str<strong>on</strong>g> policy-makers because it highlights an additi<strong>on</strong>al potential<br />
benefit for c<strong>on</strong>siderati<strong>on</strong> during the analysis of a media market merger. It also provides a new<br />
angle from which to assess the efficacy of media ownership rules.<br />
The remainder of this paper is organized as follows. Secti<strong>on</strong> 2 outlines the empirical model. Data<br />
are described in Secti<strong>on</strong> 3. Secti<strong>on</strong> 4 presents dem<str<strong>on</strong>g>and</str<strong>on</strong>g> estimates <str<strong>on</strong>g>and</str<strong>on</strong>g> calculates c<strong>on</strong>sumer valuati<strong>on</strong>s<br />
for the diversity, localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising characteristics of a news service. Secti<strong>on</strong> 5 presents<br />
estimates of the impact <strong>on</strong> c<strong>on</strong>sumer welfare from a change in market structure, <str<strong>on</strong>g>and</str<strong>on</strong>g> Secti<strong>on</strong> 6<br />
summarizes our c<strong>on</strong>clusi<strong>on</strong>s. A detailed technical descripti<strong>on</strong> of our empirical approach is provided<br />
in the appendices (available <strong>on</strong> the TWC <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong> Communicati<strong>on</strong>s<br />
website: http://twcresearchprogram.com/pdf/savageappendix.pdf).<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 7
8<br />
2. Empirical Approach<br />
We examine the relati<strong>on</strong>ship between market structure <str<strong>on</strong>g>and</str<strong>on</strong>g> media diversity by asking two<br />
questi<strong>on</strong>s:<br />
• What are the expected societal benefits that arise from increased media diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism?<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g><br />
• How do these benefits change with regulatory interventi<strong>on</strong>s that shape media market structure?<br />
We employ a three-step empirical approach to answer these questi<strong>on</strong>s. In Step One we estimate<br />
a mixed logit model of the dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for local news service with discrete choice data. The estimated<br />
preferences from the representative household’s utility (or satisfacti<strong>on</strong>) functi<strong>on</strong> are used to<br />
calculate the c<strong>on</strong>sumer’s WTP for each of the n<strong>on</strong>-price characteristics of their news service. In<br />
Step Two, using 2011 data from the FCC, we estimate the relati<strong>on</strong>ship between the number of TV<br />
stati<strong>on</strong>s in the market <str<strong>on</strong>g>and</str<strong>on</strong>g> the amount of diversity, localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising supplied within each<br />
household’s news service (2011). In Step Three, we use the estimated dem<str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> supply resp<strong>on</strong>se<br />
parameters from Steps One <str<strong>on</strong>g>and</str<strong>on</strong>g> Two, respectively, to calculate the impact <strong>on</strong> c<strong>on</strong>sumer welfare<br />
from a change in media market structure that reduces the number of independent TV voices by <strong>on</strong>e.<br />
2.1 Step One: The Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for News Services<br />
There are several problems when estimating dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for news service with market data. First,<br />
households c<strong>on</strong>sume a bundle of entertainment <str<strong>on</strong>g>and</str<strong>on</strong>g> news services. Typically there are offerings<br />
from newspapers, radio, TV, the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es. Yet data <strong>on</strong> these bundles, their n<strong>on</strong>price<br />
characteristics <str<strong>on</strong>g>and</str<strong>on</strong>g> prices are not readily available. Sec<strong>on</strong>d, even when available, these data<br />
are unlikely to exhibit sufficient variati<strong>on</strong> for the precise estimati<strong>on</strong> of dem<str<strong>on</strong>g>and</str<strong>on</strong>g> parameters. For<br />
example, the levels for the diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism characteristics are often highly positively correlated.<br />
Third, news services are a mixture of private <str<strong>on</strong>g>and</str<strong>on</strong>g> public goods <str<strong>on</strong>g>and</str<strong>on</strong>g> many households, e.g.,<br />
those who bundle broadcast radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV stati<strong>on</strong>s, make no direct payment for c<strong>on</strong>sumpti<strong>on</strong>. Since<br />
detailed data <strong>on</strong> the amount of advertising within household bundles are not available, it is not possible<br />
to accurately measure the full cost of news services.<br />
We overcome these problems by using an indirect valuati<strong>on</strong> method, similar to that used in the<br />
envir<strong>on</strong>mental <str<strong>on</strong>g>and</str<strong>on</strong>g> transportati<strong>on</strong> choice literature, which employs market <str<strong>on</strong>g>and</str<strong>on</strong>g> experimental data.<br />
The market data is the news service households currently c<strong>on</strong>sume. The experimental data is a set of<br />
c<strong>on</strong>structed news services. We design a choice set that manipulates the characteristics of the c<strong>on</strong>structed<br />
news services to obtain the optimal variati<strong>on</strong> in the data needed to estimate the dem<str<strong>on</strong>g>and</str<strong>on</strong>g><br />
parameters precisely. Resp<strong>on</strong>dents choose between a pair of c<strong>on</strong>structed news services, <str<strong>on</strong>g>and</str<strong>on</strong>g> then<br />
between that choice <str<strong>on</strong>g>and</str<strong>on</strong>g> their actual news service at home. As our design exogenously determines<br />
the levels of the characteristics of each news service, <str<strong>on</strong>g>and</str<strong>on</strong>g> r<str<strong>on</strong>g>and</str<strong>on</strong>g>omly assigns the levels across<br />
resp<strong>on</strong>dents, we limit measurement <str<strong>on</strong>g>and</str<strong>on</strong>g> collinearity problems. Further, by asking resp<strong>on</strong>dents<br />
to complete eight such choice occasi<strong>on</strong>s, we increase parameter estimati<strong>on</strong> precisi<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g> reduce<br />
sampling costs by obtaining more informati<strong>on</strong> <strong>on</strong> preferences for each resp<strong>on</strong>dent.<br />
Appendix A describes our methodology for estimating the dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for news services in which<br />
c<strong>on</strong>sumer satisfacti<strong>on</strong> (or utility) from their news service depends <strong>on</strong> five characteristics: COST,<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Table 1. Summary of News Service Characteristics<br />
COST<br />
Characteristic Levels<br />
DIVERSITY OF OPINION<br />
COMMUNITY NEWS<br />
MULTICULTURALISM<br />
ADVERTISING<br />
The total of m<strong>on</strong>thly subscripti<strong>on</strong>s to all of the household’s media sources,<br />
plus any c<strong>on</strong>tributi<strong>on</strong>s to public radio or public TV stati<strong>on</strong>s (ranging from<br />
$0 to $250 per m<strong>on</strong>th).<br />
The extent to which the informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs in the<br />
household’s overall media envir<strong>on</strong>ment reflects different viewpoints.<br />
Low: <strong>on</strong>ly <strong>on</strong>e viewpoint.<br />
Medium: a few different viewpoints.<br />
High: many different viewpoints.<br />
The amount of informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events in the household’s<br />
overall media envir<strong>on</strong>ment.<br />
Low: very little or no informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events.<br />
Medium: some informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events.<br />
High: much informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events.<br />
The amount of informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs in the household’s<br />
overall media envir<strong>on</strong>ment that reflects the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
minorities.<br />
Low: very little or no informati<strong>on</strong> reflecting the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
minorities.<br />
Medium: some informati<strong>on</strong> reflecting the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities.<br />
High: much informati<strong>on</strong> reflecting the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities.<br />
The amount of space <str<strong>on</strong>g>and</str<strong>on</strong>g>/or time devoted to advertising in the household’s<br />
overall media envir<strong>on</strong>ment.<br />
Low: barely noticeable.<br />
Medium: noticeable but not annoying.<br />
High: annoying.<br />
NOTES: The upper limit of $250 per m<strong>on</strong>th for COST is the total cost for a media envir<strong>on</strong>ment with a seven-day subscripti<strong>on</strong> to a<br />
premium newspaper, such as The New York Times or San Francisco Chr<strong>on</strong>icle ($25), an “All of XM” subscripti<strong>on</strong> to satellite radio ($20), a<br />
premier subscripti<strong>on</strong> to cable or satellite televisi<strong>on</strong> ($110), a subscripti<strong>on</strong> to very-fast Internet service ($45), an unlimited data subscripti<strong>on</strong><br />
for a smartph<strong>on</strong>e ($30), <str<strong>on</strong>g>and</str<strong>on</strong>g> $10 m<strong>on</strong>thly memberships to both NPR <str<strong>on</strong>g>and</str<strong>on</strong>g> PBS. Detailed descripti<strong>on</strong>s of the characteristics as they<br />
appeared in the survey questi<strong>on</strong>naire are available in Savage <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldman (2011).<br />
DIVERSITY OF OPINION, MULTICULTURALISM, COMMUNITY NEWS, <str<strong>on</strong>g>and</str<strong>on</strong>g> ADVERTISING.<br />
Table 1 describes the levels of these n<strong>on</strong>-price <str<strong>on</strong>g>and</str<strong>on</strong>g> price characteristics. COST is the dollar amount<br />
the household pays per m<strong>on</strong>th for its news service—that is, the total of m<strong>on</strong>thly subscripti<strong>on</strong>s to<br />
all media sources, plus any c<strong>on</strong>tributi<strong>on</strong>s to public radio or public TV stati<strong>on</strong>s. DIVERSITY OF<br />
OPINION is the extent to which the informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs in the household’s<br />
news service reflects different viewpoints. COMMUNITY NEWS is the amount of informati<strong>on</strong><br />
<strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events in the household’s news service. MULTICULTURALISM is the<br />
amount of news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs in the household’s news service that reflects the interests of<br />
women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities. ADVERTISING is the amount of space <str<strong>on</strong>g>and</str<strong>on</strong>g>/or time devoted to advertising<br />
in the household’s news service.<br />
The weights <strong>on</strong> the individual characteristics of a news service (hereafter, marginal utilities) have<br />
the usual interpretati<strong>on</strong>: the change in utility from a <strong>on</strong>e-unit increase in the level of the characteristic.<br />
We expect positive marginal utilities <strong>on</strong> DIVERSITY OF OPINION, COMMUNITY NEWS<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 9
10<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> MULTICULTURALISM for the representative household. For example, an estimated marginal<br />
utility of 0.4 for DIVERSITY OF OPINION indicates that a <strong>on</strong>e-unit improvement in DIVERSITY<br />
OF OPINION, measured by a discrete improvement from “Low = 1” to “Medium = 2”, increases<br />
utility by 0.4 for the representative household. A higher cost <str<strong>on</strong>g>and</str<strong>on</strong>g> a higher amount of advertising<br />
provide less satisfacti<strong>on</strong>, so negative marginal utilities <strong>on</strong> COST <str<strong>on</strong>g>and</str<strong>on</strong>g> ADVERTISING are expected.<br />
For example, an estimated marginal utility of -0.02 for COST indicates that a <strong>on</strong>e-dollar decrease<br />
in COST increases utility by 0.02 for the representative household.<br />
Since the estimated marginal utilities, such as an increase in utility of 0.4 described above, do<br />
not have an underst<str<strong>on</strong>g>and</str<strong>on</strong>g>able metric, it is c<strong>on</strong>venient to c<strong>on</strong>vert these changes into dollars. This<br />
is d<strong>on</strong>e by employing the ec<strong>on</strong>omic c<strong>on</strong>struct of willingness-to-pay. For example, the WTP for a<br />
<strong>on</strong>e-unit increase in diversity of opini<strong>on</strong> (WTP d) is defined as how much more the news service<br />
would have to be priced to make the c<strong>on</strong>sumer just indifferent to whether they received the old<br />
(cheaper but with <strong>on</strong>ly <strong>on</strong>e viewpoint) service <str<strong>on</strong>g>and</str<strong>on</strong>g> the new (more expensive but with a few different<br />
viewpoints) service. Given estimated marginal utilities of 0.4 for DIVERSITY OF OPINION<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> -0.02 for COST, the WTP for an improvement in diversity of opini<strong>on</strong> from “low” to “medium”<br />
is $20 (= -0.4/-0.02). Note that this linear interpretati<strong>on</strong> of the marginal utilities implies that the<br />
representative household would also be willing to pay the same amount ($20) for an improvement<br />
in diversity of opini<strong>on</strong> from “low” to “medium” as it would to move from “medium” to “high.”<br />
This c<strong>on</strong>straint can be relaxed during ec<strong>on</strong>ometric estimati<strong>on</strong> so that the marginal utility for an<br />
improvement in diversity of opini<strong>on</strong> from “medium” to “high” can be different from the marginal<br />
utility for an improvement in diversity from “low” to “medium.”<br />
This approach to estimating c<strong>on</strong>sumer valuati<strong>on</strong>s is used for all other n<strong>on</strong>-price characteristics<br />
of local news service. The WTP for COMMUNITY NEWS, MULTICULTURALISM, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
ADVERTISING is the negative of the ratio of its marginal utility to the marginal utility of COST.<br />
In summary, the WTP c<strong>on</strong>struct provides a theory-driven, intuitive (dollar) measure of the value<br />
c<strong>on</strong>sumers place <strong>on</strong> the characteristics of their local news service.<br />
2.2 Step Two: The Supply of News Services<br />
Previous studies of media markets typically use academic <str<strong>on</strong>g>and</str<strong>on</strong>g> industry databases from BIA<br />
Financial Networks, Nielsen <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> ProQuest Newsst<str<strong>on</strong>g>and</str<strong>on</strong>g> to measure the quantity<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> quality of news provided by newspapers, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV stati<strong>on</strong>s. For example, Yan <str<strong>on</strong>g>and</str<strong>on</strong>g> Napoli<br />
(2006) <str<strong>on</strong>g>and</str<strong>on</strong>g> Crawford (2007) count the minutes of local programming provided by TV stati<strong>on</strong>s,<br />
while Gentzkow <str<strong>on</strong>g>and</str<strong>on</strong>g> Shapiro (2010) measure diversity with an index that examines the similarity<br />
of a newspaper’s language to that of a c<strong>on</strong>gressi<strong>on</strong>al Republican or Democrat. Since we are investigating<br />
a household’s news services from all of its media sources, similar measures are not practical<br />
here. Instead, we use informati<strong>on</strong> <strong>on</strong> the c<strong>on</strong>sumer’s news service at home to measure the characteristics<br />
supplied by news service alternatives in different TV markets. 4<br />
C<strong>on</strong>sider a reducti<strong>on</strong> in the number of independent TV voices in a market as it impacts the<br />
single news service’s characteristic diversity of opini<strong>on</strong> (d). Equati<strong>on</strong> 4 in Appendix A shows that<br />
diversity of opini<strong>on</strong> is modeled as a linear functi<strong>on</strong> of the number of independent TV voices in<br />
the market (VOICES), the number of TV stati<strong>on</strong>s in the market (STATIONS), the interacti<strong>on</strong> of<br />
these two variables, <str<strong>on</strong>g>and</str<strong>on</strong>g> several c<strong>on</strong>trol variables. The interacti<strong>on</strong> term is included to measure the<br />
different impacts from a change in market structure in small versus large TV markets. We estimate<br />
Equati<strong>on</strong> 4 to obtain the relati<strong>on</strong>ships between the number of TV stati<strong>on</strong>s in the market <str<strong>on</strong>g>and</str<strong>on</strong>g> the<br />
four n<strong>on</strong>-price characteristics of news service (hereafter, marginal effects), <str<strong>on</strong>g>and</str<strong>on</strong>g> use these estimates<br />
to approximate the supply-side resp<strong>on</strong>ses from media outlets.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
2.3 Step Three: Estimating C<strong>on</strong>sumer Benefits from a Change in<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
It is tempting to apply a chain rule argument <str<strong>on</strong>g>and</str<strong>on</strong>g> multiply the estimated marginal effect of a<br />
decrease in the number of independent TV voices by the appropriate marginal utility (described<br />
in Secti<strong>on</strong> 2.1) to calculate the value to society from a change in the number of independent TV<br />
voices that influences the market’s provisi<strong>on</strong> of diversity of opini<strong>on</strong>. This would not estimate<br />
the effect that we are interested in, however. The problem is that we cannot observe the scale of<br />
diversity of opini<strong>on</strong>. Instead, we apply a new technique to our estimates, explained below, which<br />
takes advantage of the fact that we need to estimate the scale of diversity of opini<strong>on</strong>. This alternative<br />
approach uses our sample estimates from Equati<strong>on</strong>s 2 <str<strong>on</strong>g>and</str<strong>on</strong>g> 4 in Appendix A to predict how<br />
changes in the number of independent TV voices affect the c<strong>on</strong>sumer’s expected benefit from the<br />
amount of diversity of opini<strong>on</strong> supplied in his/her local news service.<br />
The representative c<strong>on</strong>sumer’s expected benefit from the diversity of opini<strong>on</strong> in his/her local news<br />
service is a weighted average of the benefit of being supplied with either the low, medium, or high<br />
diversity of opini<strong>on</strong> news service multiplied by the probability of being supplied that particular<br />
level of the diversity characteristic. We are not able to estimate the benefits from being in any<br />
particular news service envir<strong>on</strong>ment, but we are able to estimate (from Step One) the c<strong>on</strong>sumer’s<br />
WTP for a change from a low to a medium diversity of opini<strong>on</strong> service, <str<strong>on</strong>g>and</str<strong>on</strong>g> similarly the WTP for<br />
a change from low to high diversity. The effect of a change in the number of voices <strong>on</strong> the expected<br />
benefit from diversity of opini<strong>on</strong> can be shown to be a functi<strong>on</strong> of <strong>on</strong>ly the WTP for a change out<br />
of the low level of a characteristic, <str<strong>on</strong>g>and</str<strong>on</strong>g> the changes in probability for the supply of medium <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
high levels of the characteristic. See Equati<strong>on</strong> 7 in Appendix A for a detailed derivati<strong>on</strong> of the<br />
c<strong>on</strong>sumer’s expected benefits equati<strong>on</strong>.<br />
Equati<strong>on</strong> 7 provides the basis for calculating the value to society from a change in market structure<br />
that affects the provisi<strong>on</strong> of diversity of opini<strong>on</strong> in local media markets. Estimates of the<br />
valuati<strong>on</strong>s for n<strong>on</strong>-price characteristics for the typical c<strong>on</strong>sumer are obtained from the dem<str<strong>on</strong>g>and</str<strong>on</strong>g><br />
estimates <str<strong>on</strong>g>and</str<strong>on</strong>g> marginal WTP calculati<strong>on</strong>s in Step One. Estimates of the probability of each<br />
individual c<strong>on</strong>sumer’s being supplied a medium <str<strong>on</strong>g>and</str<strong>on</strong>g> high level of the characteristics are obtained<br />
from the ordered probit model of media supply resp<strong>on</strong>ses in Step Two. In Step Three, we use<br />
our estimated coefficients from the ordered probit model <str<strong>on</strong>g>and</str<strong>on</strong>g> the sample data to calculate the<br />
predicted probability distributi<strong>on</strong>s for low, medium <str<strong>on</strong>g>and</str<strong>on</strong>g> high diversity of opini<strong>on</strong> in the “before”<br />
envir<strong>on</strong>ment. Holding all other things c<strong>on</strong>stant, we then reduce the number of independent TV<br />
voices by <strong>on</strong>e in the sample data to approximate the change in market structure, <str<strong>on</strong>g>and</str<strong>on</strong>g> recalculate<br />
the predicted probability distributi<strong>on</strong>s for low, medium <str<strong>on</strong>g>and</str<strong>on</strong>g> high diversity of opini<strong>on</strong> in the “after”<br />
envir<strong>on</strong>ment. The difference in before-<str<strong>on</strong>g>and</str<strong>on</strong>g>-after predicted probabilities are used to form the<br />
change in probabilities, ∆PdM ∆X <str<strong>on</strong>g>and</str<strong>on</strong>g> ∆PdH . These calculati<strong>on</strong>s are repeated for the multi culturalism,<br />
∆X<br />
community news <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising characteristics of news service, <str<strong>on</strong>g>and</str<strong>on</strong>g> then aggregated to reflect the<br />
general populati<strong>on</strong>.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 11
12<br />
3. Data<br />
3.1 Experimental Design<br />
The willingness-to-pay (WTP) for local media envir<strong>on</strong>ment features is estimated with data from<br />
an <strong>on</strong>line survey questi<strong>on</strong>naire employing repeated discrete-choice experiments. The questi<strong>on</strong>naire<br />
begins with the cognitive buildup secti<strong>on</strong> that describes the resp<strong>on</strong>dent’s local news service<br />
in terms of the offerings from newspapers, radio, TV, the Internet, <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es. Resp<strong>on</strong>dents<br />
are asked questi<strong>on</strong>s about their media sources, how much informati<strong>on</strong> they c<strong>on</strong>sume from each<br />
source, the cost of their media sources, <str<strong>on</strong>g>and</str<strong>on</strong>g> the levels of the four different characteristics of their<br />
news service described in Table 1. 5<br />
Cognitive buildup is followed by the choice scenario. Informati<strong>on</strong> from the cognitive buildup<br />
questi<strong>on</strong>s is used to summarize each resp<strong>on</strong>dent’s actual entertainment <str<strong>on</strong>g>and</str<strong>on</strong>g> news service at home<br />
with respect to their media sources, the levels of the n<strong>on</strong>-price characteristics of their service,<br />
DIVERSITY OF OPINION, COMMUNITY NEWS, MULTICULTURALISM <str<strong>on</strong>g>and</str<strong>on</strong>g> ADVERTISING,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> their COST. A table summarizing the sources <str<strong>on</strong>g>and</str<strong>on</strong>g> characteristics of the resp<strong>on</strong>dent’s<br />
actual media envir<strong>on</strong>ment at home is presented before the choice scenario. The resp<strong>on</strong>dent is<br />
then instructed to answer questi<strong>on</strong>s presenting eight choice occasi<strong>on</strong>s. In each occasi<strong>on</strong>, they<br />
choose between their actual news service at home <str<strong>on</strong>g>and</str<strong>on</strong>g> two c<strong>on</strong>structed new service alternatives,<br />
A <str<strong>on</strong>g>and</str<strong>on</strong>g> B, that differ by their levels of DIVERSITY OF OPINION, COMMUNITY NEWS,<br />
MULTICULTURALISM, ADVERTISING <str<strong>on</strong>g>and</str<strong>on</strong>g> COST.<br />
We used market data from newspapers, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV stati<strong>on</strong>s, Internet <str<strong>on</strong>g>and</str<strong>on</strong>g> mobile teleph<strong>on</strong>e<br />
service providers, a pilot study <str<strong>on</strong>g>and</str<strong>on</strong>g> three focus groups to test <str<strong>on</strong>g>and</str<strong>on</strong>g> refine our descripti<strong>on</strong>s of the<br />
characteristics for news service alternatives (See Savage <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldman 2011 for more discussi<strong>on</strong>).<br />
Measures developed by Huber <str<strong>on</strong>g>and</str<strong>on</strong>g> Zwerina (1996) <str<strong>on</strong>g>and</str<strong>on</strong>g> Zwerina et. al. (1996) were used to<br />
generate an efficient n<strong>on</strong>-linear optimal design for the levels of the c<strong>on</strong>structed news characteristics.<br />
A fracti<strong>on</strong>al factorial design created 72 paired descripti<strong>on</strong>s of A <str<strong>on</strong>g>and</str<strong>on</strong>g> B news services that<br />
were grouped into nine sets of eight choice questi<strong>on</strong>s. The nine choice sets were rebalanced to<br />
ensure that each household faced a range of costs that realistically portrayed the prices for media<br />
sources in their local market. For example, a resp<strong>on</strong>dent who indicated that they pay nothing for<br />
their news source was exposed to a range of costs that included zero dollars per m<strong>on</strong>th. 6 The nine<br />
choice sets, al<strong>on</strong>g with the order of the eight A-B pair choice alternatives within each choice set,<br />
were r<str<strong>on</strong>g>and</str<strong>on</strong>g>omly distributed across all resp<strong>on</strong>dents.<br />
3.2 Survey Administrati<strong>on</strong><br />
Knowledge Networks Inc. (KN) administered the <strong>on</strong>line survey. Panel members were recruited<br />
through nati<strong>on</strong>al r<str<strong>on</strong>g>and</str<strong>on</strong>g>om samples, almost entirely by postal mail. For incentive, they were<br />
rewarded with points for participating in surveys, which can be c<strong>on</strong>verted to cash or other<br />
rewards. 7 During the week of March 7, 2011, KN r<str<strong>on</strong>g>and</str<strong>on</strong>g>omly c<strong>on</strong>tacted a gross sample of 8,621<br />
panel members to inform them about the survey. The survey was fielded from March 11 to<br />
March 21. A total of 5,548 resp<strong>on</strong>dents from all 50 states <str<strong>on</strong>g>and</str<strong>on</strong>g> the District of Columbia completed<br />
survey questi<strong>on</strong>naires. Owing to incomplete survey resp<strong>on</strong>ses, the sample was trimmed by 417<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
esp<strong>on</strong>dents. The median completi<strong>on</strong> time for the sample of 5,131 resp<strong>on</strong>dents with complete<br />
informati<strong>on</strong> was about 16 3/4 minutes.<br />
Table 2 presents selected demographic characteristics of the U.S. populati<strong>on</strong>, all KN’s panel<br />
members, <str<strong>on</strong>g>and</str<strong>on</strong>g> those panel members who were invited to participate in this survey (United States<br />
Census Bureau, 2009; Knowledge Networks, Inc., 2010). The demographics for all KN panel<br />
members are similar to those reported by the Census Bureau for the U.S. populati<strong>on</strong>. Columns<br />
Four <str<strong>on</strong>g>and</str<strong>on</strong>g> Five of Table 2 also show that, apart from race <str<strong>on</strong>g>and</str<strong>on</strong>g> employment status, the demographics<br />
for the gross sample of panel members invited to participate in this study <str<strong>on</strong>g>and</str<strong>on</strong>g> the final sample<br />
of resp<strong>on</strong>dents who completed questi<strong>on</strong>naires also are similar to those reported by the Census<br />
Bureau. However, estimates from the probit model that compares resp<strong>on</strong>dents’ characteristics<br />
between the gross sample <str<strong>on</strong>g>and</str<strong>on</strong>g> the final sample also indicate potential differences in age, gender,<br />
educati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> Internet access between our final sample <str<strong>on</strong>g>and</str<strong>on</strong>g> the populati<strong>on</strong>. We remedy this<br />
possible source of bias in our results from Step One <str<strong>on</strong>g>and</str<strong>on</strong>g> Step Two by estimating with weighted<br />
maximum likelihood. See Savage <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldman (2011) for the probit model estimates <str<strong>on</strong>g>and</str<strong>on</strong>g> the<br />
procedures used to develop the post-stratificati<strong>on</strong> weights.<br />
3.3 <str<strong>on</strong>g>Media</str<strong>on</strong>g> Sources <str<strong>on</strong>g>and</str<strong>on</strong>g> News Service<br />
Table 3 presents summary statistics for resp<strong>on</strong>dents’ media sources. Columns Two <str<strong>on</strong>g>and</str<strong>on</strong>g> Three show<br />
that almost 95 percent of sample resp<strong>on</strong>dents watch TV, 81 percent listen to the radio, <str<strong>on</strong>g>and</str<strong>on</strong>g> 80 percent<br />
use the Internet. About 46 percent of resp<strong>on</strong>dents read a paper or <strong>on</strong>line newspaper regularly,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> virtually <strong>on</strong>e-quarter (24.8 percent) of sample resp<strong>on</strong>dents own a smartph<strong>on</strong>e. 8 On average,<br />
TV viewers spend about 1.9 hours <strong>on</strong> a typical day watching TV to get informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
current affairs, radio listeners spend 1.4 hours listening to the radio to get informati<strong>on</strong> <strong>on</strong> news<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs, <str<strong>on</strong>g>and</str<strong>on</strong>g> Internet users spend <strong>on</strong>e hour <strong>on</strong>line (e.g., MSN, Yahoo, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV<br />
stati<strong>on</strong> Web sites, journalists’ blogs) to get informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs. Newspaper<br />
readers also spend about <strong>on</strong>e hour <strong>on</strong> a typical day reading the newspaper, while smartph<strong>on</strong>e<br />
owners use their ph<strong>on</strong>e to go <strong>on</strong>line for 0.6 hours a day to get informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current<br />
affairs. The most popular media source combinati<strong>on</strong>s are radio, TV <str<strong>on</strong>g>and</str<strong>on</strong>g> the Internet—about 30<br />
percent of sample resp<strong>on</strong>dents—<str<strong>on</strong>g>and</str<strong>on</strong>g> newspaper, radio, TV <str<strong>on</strong>g>and</str<strong>on</strong>g> the Internet, for about 26 percent<br />
of sample resp<strong>on</strong>dents.<br />
Summary statistics for news service characteristics are presented in Table 4. These data indicate<br />
that, <strong>on</strong> average, the levels of the DIVERSITY OF OPINION, COMMUNITY NEWS,<br />
MULTICULTURALISM <str<strong>on</strong>g>and</str<strong>on</strong>g> ADVERTISING characteristics are about “medium.” About 58 percent<br />
of resp<strong>on</strong>dents indicate that they bundle their subscripti<strong>on</strong> TV service with the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g>/<br />
or teleph<strong>on</strong>e service. The price (or COST) for the typical media envir<strong>on</strong>ment ranges from zero<br />
to $447 per m<strong>on</strong>th, with an average of $111.20 per m<strong>on</strong>th. Significantly, about 10 percent of the<br />
sample indicate that they have c<strong>on</strong>tributed $117, <strong>on</strong> average, to public radio <str<strong>on</strong>g>and</str<strong>on</strong>g>/or TV stati<strong>on</strong>s<br />
during the past 12 m<strong>on</strong>ths. 9<br />
3.4 <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
We use data from the FCC (2011) to measure media market structure. The important variables are<br />
the number of full-power independent TV stati<strong>on</strong>s in the market (VOICES) <str<strong>on</strong>g>and</str<strong>on</strong>g> the total number<br />
of full-power independent <str<strong>on</strong>g>and</str<strong>on</strong>g> n<strong>on</strong>-independent TV stati<strong>on</strong>s in the market (STATIONS). VOICES<br />
is measured by first combining all the TV outlets within each market. The listing of the unique<br />
parent company identifiers of all attributable owners of an outlet (“voiceprint”) is then created,<br />
sorted alphabetically, <str<strong>on</strong>g>and</str<strong>on</strong>g> duplicate voiceprints are eliminated. The parent identifier is then used<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 13
14<br />
Table 2. Demographic Distributi<strong>on</strong>s (%)<br />
Regi<strong>on</strong><br />
Census KN panel KN sample<br />
Gross<br />
sample<br />
(Invited)<br />
Full sample<br />
(Completed)<br />
Final sample<br />
(Completed)<br />
Omitted<br />
(Completed)<br />
Northeast 18.4 18.5 18.0 18.5 18.3 21.8<br />
Midwest 21.8 22.1 23.4 24.7 24.9 21.3<br />
South 36.5 35.9 35.6 34.2 34.4 31.4<br />
West 23.2 23.5 22.9 22.6 22.4 25.4<br />
Age<br />
18–24 years 12.6 10.7 10.7 8.6 7.6 21.8<br />
25–34 years 17.8 17.4 15.0 12.7 11.3 29.7<br />
35–44 years 17.8 18.9 16.6 16.1 15.8 20.1<br />
45–54 years 19.5 18.5 20.1 20.3 20.9 13.2<br />
55–64 years 15.5 18.5 20.3 22.2 23.3 9.1<br />
65 years or older 16.8 16.0 17.3 20.1 21.2 6.0<br />
Race<br />
N<strong>on</strong>-white 18.9 20.9 30.3 26.0 25.4 33.3<br />
White 81.1 79.1 69.7 74.0 74.6 66.7<br />
Gender<br />
Female 51.7 52.6 53.0 50.8 51.1 47.5<br />
Male 48.3 47.4 47.0 49.2 49.9 52.5<br />
Marital status<br />
Married 55.1 52.5 53.5 55.1 55.8 46.3<br />
Not married 44.9 47.5 46.5 44.9 44.2 53.7<br />
Educati<strong>on</strong><br />
< High school 13.8 12.9 10.8 9.6 9.5 10.8<br />
High school 30.7 29.6 29.0 30.2 30.0 32.6<br />
Some college 28.2 29.1 31.3 29.8 30.0 27.6<br />
Bachelor’s degree or<br />
higher<br />
Household income<br />
27.4 28.3 28.9 30.4 30.5 29.0<br />
< $10,000 6.6 7.0 7.1 6.3 6.3 6.7<br />
$10,000–$24,999 16.8 16.1 15.1 14.4 14.3 14.6<br />
$25,000–$49,999 26.2 26.1 24.3 24.8 24.6 27.1<br />
$50,000–$74,999 19.5 20.3 18.3 18.8 19.1 15.9<br />
> $75,000 30.8 30.4 35.2 35.7 35.7 35.7<br />
Employment<br />
In labor force 66.1 67.3 56.8 55.2 54.7 60.7<br />
Not in labor force 33.9 32.7 43.2 44.8 45.3 39.3<br />
Internet access 64.0 66.0 73.0 81.2 80.6 88.2<br />
Observati<strong>on</strong>s n.a. n.a. 8,621 5,548 5,131 417<br />
NOTES: Census data are from December 2009. KN panel data are from January 2010. Remaining data are from March 2011<br />
SOURCE: United States Census Bureau (2009); Knowledge Networks, Inc. (2010).<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Table 3. Summary Statistics for <str<strong>on</strong>g>Media</str<strong>on</strong>g> Sources<br />
<str<strong>on</strong>g>Media</str<strong>on</strong>g> source Obs.<br />
Sample<br />
share (%)<br />
Mean<br />
Time Spent<br />
Using (Hours)<br />
s.d. Min Max<br />
Newspaper 2,342 45.6 1.015 1.766 0 24<br />
Radio 4,154 81.2 1.423 1.873 0 24<br />
Satellite radio 558 10.9 1.522 2.221 0 24<br />
Televisi<strong>on</strong> 4,856 94.6 1.953 2.172 0 24<br />
Cable televisi<strong>on</strong> 2,736 53.4 1.976 2.210 0 24<br />
Satellite televisi<strong>on</strong> 1,381 27.0 2.071 2.197 0 24<br />
Own Internet 4,135 80.6 1.074 1.659 0 24<br />
Smartph<strong>on</strong>e 1,270 24.8 0.580 1.344 0 24<br />
NOTES: Obs. is the number of observati<strong>on</strong>s. Sample share is the percentage of the sample that uses the media source; s.d. is st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard<br />
deviati<strong>on</strong>. Min is minimum value. Max is maximum value. “Own Internet” is home Internet service not provided by KN.<br />
Table 4. Summary Statistics for News Service Characteristics<br />
Feature Obs. Mean s.d. Min Max<br />
DIVERSITY OF OPINION 5,131 2.09 0.655 1 3<br />
COMMUNITY NEWS 5,131 1.99 0.711 1 3<br />
MULTICULTURALISM 5,131 1.83 0.705 1 3<br />
ADVERTISING 5,131 2.29 0.682 1 3<br />
COST ($ per m<strong>on</strong>th) 5,131 111.2 76.03 0 447<br />
CONTRIBUTION ($ annual) 535 111.5 161.5 0.25 1,500<br />
BUNDLE 3,688 0.576 0.494 0 1<br />
NOTES: 1 = “low”, 2 = “medium” <str<strong>on</strong>g>and</str<strong>on</strong>g> 3 = “high” for DIVERSITY OF OPINION, COMMUNITY NEWS, MULTICULTURALISM, <str<strong>on</strong>g>and</str<strong>on</strong>g> ADVERTISING.<br />
CONTRIBUTION is value of c<strong>on</strong>tributi<strong>on</strong>s to public radio <str<strong>on</strong>g>and</str<strong>on</strong>g> public TV stati<strong>on</strong>s during the past 12 m<strong>on</strong>ths. BUNDLE = 1 when subscripti<strong>on</strong><br />
televisi<strong>on</strong> service is bundled with Internet service <str<strong>on</strong>g>and</str<strong>on</strong>g>/or other teleph<strong>on</strong>e services. Obs. is the number of observati<strong>on</strong>s. s.d. is<br />
st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard deviati<strong>on</strong>. Min is minimum value. Max is maximum value.<br />
Table 5. <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
Variable Descripti<strong>on</strong><br />
HOUSEHOLDS Number of households in the market.<br />
NEWSPAPERS<br />
Number of daily newspapers with a city of publicati<strong>on</strong> located in a county<br />
in the market.<br />
RADIO STATIONS Number of radio stati<strong>on</strong>s in the market.<br />
STATIONS Number of full-power TV stati<strong>on</strong>s in the market.<br />
MEDIA OUTLETS NEWSPAPERS plus RADIO STATIONS plus STATIONS.<br />
NEWSPAPER VOICES Number of parent entities owning a daily newspaper in the market.<br />
RADIO VOICES Number of independent radio voices in the market.<br />
VOICES Number of independent TV voices in the market.<br />
MEDIA VOICES NEWSPAPER VOICES plus RADIO VOICES plus VOICES.<br />
TV-NEWSPAPER VOICES Number of independent newspaper <str<strong>on</strong>g>and</str<strong>on</strong>g> TV voices in the market.<br />
TV-RADIO VOICES Number of independent radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV voices in the market.<br />
SOURCE: FCC (2011).<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 15
16<br />
Table 6. Summary Statistics for <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
All markets<br />
Variable <str<strong>on</strong>g>Market</str<strong>on</strong>g>s Mean s.d. Min 25th 75th Max<br />
HOUSEHOLDS 203 1,670,158 1,842,396 4,145 447,396 2,228,143 7,444,659<br />
SMALL MARKETS 203 0.084 0.278 0 n.a. n.a. 1<br />
MEDIA OUTLETS 203 138.7 71.25 4 80 182 291<br />
MEDIA VOICES 203 73.11 35.97 3 44 97 152<br />
NEWSPAPERS 203 12.76 8.206 0 6 19 32<br />
RADIO STATIONS 203 113.2 59.41 3 64 157 241<br />
STATIONS 203 12.74 5.879 1 8 17 27<br />
NEWSPAPER VOICES 203 7.634 4.076 0 4 10 19<br />
RADIO VOICES 203 55.12 28.75 2 31 73 119<br />
VOICES 203 10.36 4.626 1 7 13 22<br />
TV-NEWSPAPER VOICES 203 11.91 4.758 1 8 15 24<br />
TV-RADIO VOICES 203 63.06 30.95 2 38 85 129<br />
Small markets (five or fewer TV stati<strong>on</strong>s)<br />
HOUSEHOLDS 68 195,814 98,806 4,145 116,273 264,844 395,620<br />
MEDIA OUTLETS 68 46.97 15.90 4 37 57 86<br />
MEDIA VOICES 68 26.36 8.695 3 20 34 41<br />
NEWSPAPERS 68 4.160 2.347 0 2 6 11<br />
RADIO STATIONS 68 38.60 13.85 3 30 48 75<br />
STATIONS 68 4.211 1.060 1 4 5 5<br />
NEWSPAPER VOICES 68 3.308 1.900 0 2 4 8<br />
RADIO VOICES 68 19.00 6.608 2 14 25 31<br />
VOICES 68 4.046 1.097 1 3 5 5<br />
TV-NEWSPAPER VOICES 68 5.734 1.302 1 5 7 8<br />
TV-RADIO VOICES 68 22.54 7.316 2 17 28 35<br />
NOTES: <str<strong>on</strong>g>Market</str<strong>on</strong>g>s is the number of televisi<strong>on</strong> markets. s.d. is st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard deviati<strong>on</strong>. Min is minimum value. Max is maximum value. 25 th is<br />
25 th percentile. 75 th is 75 th percentile. n.a. is not applicable.<br />
to count the number of voices in the voiceprint for each outlet. Voiceprints composed of a single<br />
voice are added to the voice count of the market, while any voiceprint that includes <strong>on</strong>e of the<br />
voices counted at the previous stage of the calculati<strong>on</strong> is eliminated. These are voices that are not<br />
independent because they have been heard <strong>on</strong> another outlet. This process is sequentially repeated<br />
based <strong>on</strong> the number of voices in the voiceprint. Table 5 describes the remaining market structure<br />
variables c<strong>on</strong>sidered in this analysis.<br />
Table 6 presents summary statistics. Our sample covers 203 of the nati<strong>on</strong>’s 210 televisi<strong>on</strong> markets. 10<br />
As of December, 2009, the total number of newspaper, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV outlets in a market ranged<br />
from four to 291, with an average of 139 per market. On average, about 81 percent of media outlets<br />
are radio stati<strong>on</strong>s, which partially reflects that the geographical definiti<strong>on</strong> of a TV market can<br />
include several radio markets. When examining the market structure data at the 75 th percentile,<br />
we observe that most markets are served by about 182 or fewer media outlets. The bottom panel<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
in Table 6 shows a similar pattern for small TV markets with five or fewer stati<strong>on</strong>s. In December<br />
2009, the total of newspaper, radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV outlets in small markets ranged from four to 86, with<br />
an average of 47 per market. On average, about 82 percent of media outlets in small markets are<br />
radio stati<strong>on</strong>s, <str<strong>on</strong>g>and</str<strong>on</strong>g> as indicated by the 75 th percentile, most small markets are served by about 57<br />
or fewer media outlets.<br />
C<strong>on</strong>trol Variable Methodology<br />
The survey provided the household data used to c<strong>on</strong>struct the c<strong>on</strong>trol variable vectors Y n <str<strong>on</strong>g>and</str<strong>on</strong>g> Z j in<br />
Supply Equati<strong>on</strong> 4 in Appendix A. The Y vector measures the head of the household’s age (AGE = 1 if<br />
18–24 years; 2 if 25–34; 3 if 35–44; 4 if 45–54; 5 if 55–64; 6 if 65–74; 7 if 75 years or over), educati<strong>on</strong> (EDUC<br />
= 1 if less than high school; 2 if high school; 3 if some college; 4 if bachelor’s degree or more), gender<br />
(GENDER = 1 if female; 0 if male), household income (INCOME = 1 if less than $10,000; 2 if $10,000–<br />
$24,999; 3 if 25,000–$49,999; 4 if $50,000–$74,999; 5 if $75,000 or more), <str<strong>on</strong>g>and</str<strong>on</strong>g> race (RACE = 1 if white; 0<br />
otherwise.). The Z vector includes dummy variables to c<strong>on</strong>trol for the 16 different media source combinati<strong>on</strong>s<br />
in our sample that are comprised from newspapers, radio, TV, the Internet <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es. 11<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 17
18<br />
4. Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> Estimates<br />
The choice data described in Secti<strong>on</strong> 3.1 are used to estimate a discrete-choice model of household<br />
utility from their local news service. Since 29 resp<strong>on</strong>dents lacked geographical identifiers<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> could not be assigned to an appropriate TV market, they were dropped from the final sample<br />
of 5,131. Since each of the choice scenarios represents informati<strong>on</strong> <strong>on</strong> preferences from three<br />
alternatives, A, B, <str<strong>on</strong>g>and</str<strong>on</strong>g> actual news service at home, the sample size for ec<strong>on</strong>ometric estimati<strong>on</strong> is<br />
5,102×8×3 = 122,448. Table 2 displays some demographic differences between our final sample<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> the populati<strong>on</strong>. We remedy this possible source of bias by estimating the discrete-choice<br />
model by weighted maximum likelihood, where the c<strong>on</strong>tributi<strong>on</strong> to the log likelihood is the poststratificati<strong>on</strong><br />
weight times the log of the choice probability for the choice occasi<strong>on</strong>.<br />
Table 7 reports weighted maximum likelihood estimates of household utility. As c<strong>on</strong>sumers may<br />
have heterogeneous preferences for unmeasured aspects of news service, we estimate utility with<br />
an alternative-specific c<strong>on</strong>stant to capture differences in tastes between the actual <str<strong>on</strong>g>and</str<strong>on</strong>g> new (A <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
B) news services. For purpose of comparis<strong>on</strong>, in Model (i) we begin by reporting estimates from a<br />
st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard c<strong>on</strong>diti<strong>on</strong>al logit model with fixed marginal utility parameters. Model (ii) displays estimates<br />
from a mixed logit model specificati<strong>on</strong> where the four n<strong>on</strong>-price marginal utility parameters are<br />
assumed to be independently normally distributed. 12 Preferences may be correlated, for example,<br />
c<strong>on</strong>sumers who like more diversity of opini<strong>on</strong> may also like more informati<strong>on</strong> <strong>on</strong> women <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
minorities. Accordingly, the mixed logit model Model (iii) permits correlati<strong>on</strong> between the n<strong>on</strong>-price<br />
parameters. 13 Model (iv) reports estimates from a mixed logit model specificati<strong>on</strong> with correlated<br />
n<strong>on</strong>-price parameters plus COST×MED_INCOME <str<strong>on</strong>g>and</str<strong>on</strong>g> COST×HIGH_INCOME. The additi<strong>on</strong>al<br />
observed c<strong>on</strong>sumer characteristics are MED_INCOME, which equals <strong>on</strong>e when household income is<br />
greater than $25,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> less than $50,000; <str<strong>on</strong>g>and</str<strong>on</strong>g> zero otherwise, <str<strong>on</strong>g>and</str<strong>on</strong>g> HIGH_INCOME, which equals<br />
14 15<br />
<strong>on</strong>e when household income is greater than $50,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> zero otherwise.<br />
The data fit all model specificati<strong>on</strong>s reas<strong>on</strong>ably well as judged by the sign <str<strong>on</strong>g>and</str<strong>on</strong>g> statistical significance<br />
of most parameter estimates. We focus our discussi<strong>on</strong> <strong>on</strong> the results from Model (iv)<br />
because that model permits the marginal disutility of cost to vary by income. The means of each<br />
of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om marginal utility parameters for DIVERSITY OF OPINION, MULTICULTURALISM<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> COMMUNITY NEWS are positive <str<strong>on</strong>g>and</str<strong>on</strong>g> significant at the <strong>on</strong>e percent level, while the mean of<br />
the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om parameter for ADVERTISING is negative <str<strong>on</strong>g>and</str<strong>on</strong>g> significant. These estimates imply that<br />
utility to the representative c<strong>on</strong>sumer increases when there is more diversity in the reporting of<br />
news, more informati<strong>on</strong> <strong>on</strong> women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities, more informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
less space <str<strong>on</strong>g>and</str<strong>on</strong>g>/or time devoted to advertising. The fixed parameter for COST is negative <str<strong>on</strong>g>and</str<strong>on</strong>g> the<br />
corresp<strong>on</strong>ding parameters for COST×MED_INCOME <str<strong>on</strong>g>and</str<strong>on</strong>g> COST×HIGH_INCOME are positive.<br />
These estimates imply that utility to the c<strong>on</strong>sumer decreases when the dollar amount paid for their<br />
news service increases but that the effect diminishes as resp<strong>on</strong>dents’ household income increases.<br />
The st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard deviati<strong>on</strong>s of each of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om marginal utility parameters are significant at the<br />
<strong>on</strong>e-percent level, indicating that tastes vary in the populati<strong>on</strong>. Together, the estimated means <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard deviati<strong>on</strong>s of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om parameters provide useful policy informati<strong>on</strong> <strong>on</strong> the percentage<br />
of the populati<strong>on</strong> that places a positive value <strong>on</strong> the n<strong>on</strong>-price characteristics of news service.<br />
The mean <str<strong>on</strong>g>and</str<strong>on</strong>g> st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard deviati<strong>on</strong> of the parameter estimator for DIVERSITY OF OPINION are<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Table 7. Mixed Logit Estimates of the Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for Local News Service<br />
Model (i) Model (ii) Model (iii) Model (iv) Model (v)<br />
Mean Mean s.d. Mean s.d. Mean s.d. Mean s.d.<br />
-0.042 ***<br />
(0.001)<br />
-0.037 ***<br />
(0.001)<br />
-0.028 ***<br />
(0.0003)<br />
-0.028 ***<br />
(0.0003)<br />
-0.020 ***<br />
(0.0002)<br />
COST<br />
0.005 ***<br />
(0.001)<br />
0.004 ***<br />
(0.001)<br />
COST×MED_INCOME<br />
0.013 ***<br />
(0.001)<br />
0.011 ***<br />
(0.001)<br />
COST×HIGH_INCOME<br />
0.801 ***<br />
(0.019)<br />
0.443 ***<br />
(0.016)<br />
0.805 ***<br />
(0.019)<br />
0.448 ***<br />
(0.016)<br />
0.810 ***<br />
(0.019)<br />
0.433 ***<br />
(0.016)<br />
0.383 ***<br />
(0.009)<br />
DIVERSITY OF OPINION<br />
0.590 ***<br />
(0.019)<br />
0.449 ***<br />
(0.014)<br />
0.591 ***<br />
(0.019)<br />
0.450 ***<br />
(0.014)<br />
0.649 ***<br />
(0.018)<br />
0.433 ***<br />
(0.014)<br />
0.461 ***<br />
(0.009)<br />
COMMUNITY NEWS<br />
0.604 ***<br />
(0.021)<br />
0.041 ***<br />
(0.015)<br />
0.599 ***<br />
(0.021)<br />
0.041 ***<br />
(0.015)<br />
0.685 ***<br />
(0.019)<br />
0.015<br />
(0.015)<br />
0.012<br />
(0.009)<br />
MULTICULTURALISM-<br />
0.681 ***<br />
(0.020)<br />
-0.244 ***<br />
(0.016)<br />
0.692 ***<br />
(0.020)<br />
-0.229 ***<br />
(0.016)<br />
0.695 ***<br />
(0.019)<br />
-0.227 ***<br />
(0.016)<br />
-0.357 ***<br />
(0.010)<br />
ADVERTISING<br />
1.295 ***<br />
(0.043)<br />
0.748 ***<br />
(0.032)<br />
MEDIUM DIVERSITY OF OPINION<br />
1.076 ***<br />
(0.043)<br />
0.991 ***<br />
(0.037)<br />
HIGH DIVERSITY OF OPINION<br />
1.197 ***<br />
(0.045)<br />
0.894 ***<br />
(0.033)<br />
MEDIUM COMMUNITY NEWS<br />
0.510 ***<br />
(0.072)<br />
1.116 ***<br />
(0.034)<br />
HIGH COMMUNITY NEWS<br />
0.473 ***<br />
(0.075)<br />
0.253 ***<br />
(0.027)<br />
MEDIUM MULTICULTURALISM<br />
0.159 ***<br />
(0.062)<br />
0.147 ***<br />
(0.032)<br />
HIGH MULTICULTURALISM<br />
-0.029 ***<br />
(0.049)<br />
-0.169 ***<br />
(0.023)<br />
MEDIUM ADVERTISING<br />
0.291 ***<br />
(0.081)<br />
-0.739 ***<br />
(0.039)<br />
HIGH ADVERTISING<br />
0.816 ***<br />
(0.026)<br />
0.888 ***<br />
(0.022)<br />
0.888 ***<br />
(0.021)<br />
0.911 ***<br />
(0.021)<br />
0.769 ***<br />
(0.018)<br />
ALTERNATIVE SPECIFIC CONSTANT<br />
Likelihood -59,453 -32,714 -32,477 -32,303 -32,523<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 19<br />
NOTES: Estimated by simulated weighted maximum likelihood. (i) is estimated with the c<strong>on</strong>diti<strong>on</strong>al logit model. (ii) through (v) are estimated with the mixed logit model. Mean <str<strong>on</strong>g>and</str<strong>on</strong>g> s.d. are the estimated means <str<strong>on</strong>g>and</str<strong>on</strong>g> st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard<br />
deviati<strong>on</strong>s of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om marginal utility parameters. Covariances of correlated r<str<strong>on</strong>g>and</str<strong>on</strong>g>om parameters are not reported but are available from the authors up<strong>on</strong> request. ALTERNATIVE SPECIFIC CONSTANT equals <strong>on</strong>e for actual news<br />
service alternative at home <str<strong>on</strong>g>and</str<strong>on</strong>g> zero for news service alternatives A <str<strong>on</strong>g>and</str<strong>on</strong>g> B. Robust st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard errors in parentheses. *** denotes significant at the <strong>on</strong>e-percent level. ** denotes significant at the five-percent level. * denotes significant at<br />
the ten-percent level. Number of observati<strong>on</strong>s is 122,448.
20<br />
0.443 <str<strong>on</strong>g>and</str<strong>on</strong>g> 0.801, respectively. Using the cumulative normal distributi<strong>on</strong>, this implies that about 70<br />
percent of the populati<strong>on</strong> prefer more differing viewpoints in the reporting of news <str<strong>on</strong>g>and</str<strong>on</strong>g> 30 percent<br />
prefer fewer viewpoints. Similar calculati<strong>on</strong>s show that about 80 percent of the populati<strong>on</strong> prefer<br />
more community news, <str<strong>on</strong>g>and</str<strong>on</strong>g> more news that reflects the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities is<br />
preferred by about <strong>on</strong>e-half of the populati<strong>on</strong>. Approximately two-thirds of the populati<strong>on</strong> prefer<br />
having fewer ads.<br />
Our descripti<strong>on</strong> of advertising measures the amount of space <strong>on</strong> a newspaper or Web page, or<br />
the amount of time devoted to commercial advertising <strong>on</strong> radio or TV. Using this definiti<strong>on</strong><br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> informati<strong>on</strong> <strong>on</strong> public news c<strong>on</strong>sumpti<strong>on</strong> from the Pew <str<strong>on</strong>g>Research</str<strong>on</strong>g> Center (2010), we use<br />
our dem<str<strong>on</strong>g>and</str<strong>on</strong>g> estimates to shed light <strong>on</strong> the value of informative vs. n<strong>on</strong>-informative advertising.<br />
Given that 58 percent of the U.S. public gets its news from TV, the estimated negative valuati<strong>on</strong>s<br />
for ADVERTISING likely reflect the c<strong>on</strong>sumpti<strong>on</strong> of general, all-purpose advertising delivered by<br />
traditi<strong>on</strong>al media such as radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV. In other words, most c<strong>on</strong>sumers will indicate distaste for<br />
n<strong>on</strong>-informative advertisements because they do not want to view them or listen to them. In c<strong>on</strong>trast,<br />
the estimated positive valuati<strong>on</strong>s likely reflect the c<strong>on</strong>sumpti<strong>on</strong> of more informative, targeted<br />
advertisements delivered by new media such as the Internet, smartph<strong>on</strong>e <str<strong>on</strong>g>and</str<strong>on</strong>g> video-<strong>on</strong>-dem<str<strong>on</strong>g>and</str<strong>on</strong>g>.<br />
Here, c<strong>on</strong>sumers indicate their preference for advertisements because they are positively informed<br />
about something specific to their needs <str<strong>on</strong>g>and</str<strong>on</strong>g>/or they have some choice in the advertisements they<br />
view. 16<br />
In this discussi<strong>on</strong> the coding of the four n<strong>on</strong>-price features in the household utility functi<strong>on</strong> is<br />
linear, which implies that the marginal utilities are the same when moving from low to medium<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> from medium to high. We now relax this restricti<strong>on</strong> by replacing each of the four n<strong>on</strong>-price<br />
characteristics with a pair of dichotomous variables. For example, MEDIUM DIVERSITY OF<br />
OPINION equals <strong>on</strong>e when DIVERSITY OF OPINION equals “medium” <str<strong>on</strong>g>and</str<strong>on</strong>g> zero otherwise, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
HIGH DIVERSITY OF OPINION equals <strong>on</strong>e when DIVERSITY OF OPINION equals “high” <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
zero otherwise. Here, the estimated parameter <strong>on</strong> MEDIUM DIVERSITY OF OPINION measures<br />
the change in utility from moving from informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs in the household’s<br />
overall news service reflecting <strong>on</strong>ly <strong>on</strong>e viewpoint (low diversity) to a few different viewpoints<br />
(medium diversity). The estimated parameter <strong>on</strong> HIGH DIVERSITY OF OPINION measures the<br />
change in utility from moving from informati<strong>on</strong> <strong>on</strong> news <str<strong>on</strong>g>and</str<strong>on</strong>g> current affairs reflecting <strong>on</strong>ly <strong>on</strong>e<br />
viewpoint to many different viewpoints (high diversity). This approach to estimating n<strong>on</strong>-linear<br />
c<strong>on</strong>sumer valuati<strong>on</strong>s is used for all other n<strong>on</strong>-price characteristics of the local news service.<br />
Mixed logit estimates of the utility model with n<strong>on</strong>-linear preferences are presented in Table 7.<br />
Model (v) reports estimates with correlated n<strong>on</strong>-price parameters plus COST×MED_INCOME<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> COST×HIGH_INCOME. Focusing <strong>on</strong> the means of each of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om marginal utility<br />
parameters, the results indicate declining marginal utility for the representative c<strong>on</strong>sumer with<br />
respect to diversity of opini<strong>on</strong>, multiculturalism <str<strong>on</strong>g>and</str<strong>on</strong>g> community news. Accompanying WTP<br />
calculati<strong>on</strong>s by household income are reported in the bottom panel of Table 8. For comparis<strong>on</strong>,<br />
WTP calculati<strong>on</strong>s from the linear estimates of utility are reported in the top panel. In Column<br />
Two we observe that the representative medium-income c<strong>on</strong>sumer is willing to pay $20.82 per<br />
m<strong>on</strong>th for an improvement in diversity of opini<strong>on</strong> from low to medium, but <strong>on</strong>ly another $6.76<br />
per m<strong>on</strong>th for an additi<strong>on</strong>al improvement to high diversity of opini<strong>on</strong>. Similarly, the representative<br />
medium-income household is willing to pay $24.88 per m<strong>on</strong>th for an initial improvement<br />
in informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events from low to medium, but <strong>on</strong>ly another $6.18 per<br />
m<strong>on</strong>th for an additi<strong>on</strong>al improvement to high. The marginal utility estimates for multiculturalism<br />
indicate that households value an improvement in informati<strong>on</strong> that reflects the interests of<br />
women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities from low to medium (i.e., WTP = $7.04) more than an improvement from<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
low to high (i.e., WTP = $4.09). In other words, the representative medium-income household<br />
wants more, but not a lot more informati<strong>on</strong> reflecting the interests of women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities.<br />
The marginal utility estimates for advertising indicate a pattern similar to diversity of opini<strong>on</strong><br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> community news, albeit in reverse. The representative household is willing to pay about<br />
$15.87 per m<strong>on</strong>th to move from high to medium advertising, but would pay <strong>on</strong>ly an additi<strong>on</strong>al<br />
$4.70 per m<strong>on</strong>th to move from medium to low advertising. 17<br />
Table 8. Willingness-to-Pay by Household Income<br />
Linear preferences<br />
Low income <<br />
$25,000<br />
Medium income<br />
> $25,000 ≤<br />
$50,000<br />
> $50,000 High<br />
income<br />
DIVERSITY OF OPINION $12.14 $13.61 $17.65<br />
COMMUNITY NEWS $12.30 $13.79 $17.89<br />
MULTICULTURALISM $1.12 $1.25 $1.62<br />
ADVERTISING $(6.68) $(7.49) $(9.72)<br />
N<strong>on</strong>-linear preferences<br />
MEDIUM DIVERSITY OF OPINION $18.16 $20.82 $26.34<br />
HIGH DIVERSITY OF OPINION $24.05 $27.58 $34.89<br />
MEDIUM COMMUNITY NEWS $21.70 $24.88 $31.48<br />
HIGH COMMUNITY NEWS $27.09 $31.06 $39.30<br />
MEDIUM MULTICULTURALISM $6.14 $7.04 $8.91<br />
HIGH MULTICULTURALISM $3.57 $4.09 $5.18<br />
MEDIUM ADVERTISING $(4.10) $(4.70) $(5.95)<br />
HIGH ADVERTISING $(17.94) $(20.57) $(26.02)<br />
NOTES: Willingness-to-pay is calculated using the mean of each of the r<str<strong>on</strong>g>and</str<strong>on</strong>g>om marginal utility parameters <str<strong>on</strong>g>and</str<strong>on</strong>g> the marginal disutility<br />
of COST. The marginal disutility of COST varies by household income <str<strong>on</strong>g>and</str<strong>on</strong>g> is β 1 + β MMED_INCOME + β HHIGH_INCOME, where MED_<br />
INCOME equals <strong>on</strong>e when household income is greater than $25,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> less than $50,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> zero otherwise, <str<strong>on</strong>g>and</str<strong>on</strong>g> HIGH_INCOME<br />
equals <strong>on</strong>e when household income is greater than $50,000 <str<strong>on</strong>g>and</str<strong>on</strong>g> zero otherwise. Linear calculati<strong>on</strong>s use utility estimates from model<br />
(iv) in Table 7. N<strong>on</strong>-linear calculati<strong>on</strong>s use utility estimates from model (v) in Table 7. The parentheses <strong>on</strong> MEDIUM ADVERTISING indicate<br />
WTP to move from a medium to a low level of advertising. The parentheses <strong>on</strong> HIGH ADVERTISING indicate WTP to move from a high to<br />
a low level of advertising.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 21
22<br />
5. Impact <strong>on</strong> Welfare from a Change in <str<strong>on</strong>g>Market</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
The dem<str<strong>on</strong>g>and</str<strong>on</strong>g> estimates provide informati<strong>on</strong> <strong>on</strong> the expected societal benefits from increased<br />
media diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism. The questi<strong>on</strong> of interest now is how do these benefits change with<br />
regulatory interventi<strong>on</strong>s that shape media market structure? We shed light <strong>on</strong> this questi<strong>on</strong> by<br />
estimating the relati<strong>on</strong>ships between the number of TV stati<strong>on</strong>s in the market <str<strong>on</strong>g>and</str<strong>on</strong>g> the amount of<br />
diversity, localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising supplied within each household’s news service (see Appendix A,<br />
Equati<strong>on</strong> 4). The resulting supply resp<strong>on</strong>se parameters then are combined with WTP calculati<strong>on</strong>s<br />
to measure the impact <strong>on</strong> c<strong>on</strong>sumer welfare from a change in media market structure that reduces<br />
the number of independent TV voices by <strong>on</strong>e.<br />
5.1 The Supply of News Services<br />
Because unobserved cost <str<strong>on</strong>g>and</str<strong>on</strong>g> dem<str<strong>on</strong>g>and</str<strong>on</strong>g> factors affect both media market structure <str<strong>on</strong>g>and</str<strong>on</strong>g> the supply<br />
of news service characteristics, the estimated coefficient <strong>on</strong> VOICES in Equati<strong>on</strong> 4 is likely to suffer<br />
from omitted variable bias. For example, a market with higher unobserved costs of producing<br />
advertising will be less profitable <str<strong>on</strong>g>and</str<strong>on</strong>g> will attract fewer TV stati<strong>on</strong>s. This market may also have<br />
more advertising because stati<strong>on</strong>s need additi<strong>on</strong>al revenue to cover their higher costs. A st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard<br />
ordered probit model would bias the estimated relati<strong>on</strong>ship between ADVERTISING <str<strong>on</strong>g>and</str<strong>on</strong>g> VOICES<br />
in a negative directi<strong>on</strong>. We account for this endogeneity with a two-stage selecti<strong>on</strong> model similar<br />
in approach to Mazzeo (2002), Singh <str<strong>on</strong>g>and</str<strong>on</strong>g> Zhu (2008), <str<strong>on</strong>g>and</str<strong>on</strong>g> Chen <str<strong>on</strong>g>and</str<strong>on</strong>g> Savage (2011). See Appendix<br />
C for a detailed descripti<strong>on</strong> of this two-stage model.<br />
Table 9 presents the estimates of the supply resp<strong>on</strong>se. Focusing <strong>on</strong> the important variable of<br />
interest, we observe that the estimated coefficients <strong>on</strong> VOICES are positive for all n<strong>on</strong>-price news<br />
characteristics, while the estimated coefficients <strong>on</strong> VOICES×STATIONS are negative. These results<br />
suggest that following a decrease in the number of independent TV stati<strong>on</strong>s in the market, c<strong>on</strong>sumers<br />
are more likely to have less diversity of opini<strong>on</strong>, multiculturalism, community news <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
advertising in their news service. For example, the sample means of the predicted probabilities of<br />
supply presented in Table 10 show that following the change in market structure, the percentage of<br />
households in a low diversity of opini<strong>on</strong> state will increase by 0.016, the percentage of households<br />
in a medium state will increase by 0.003, <str<strong>on</strong>g>and</str<strong>on</strong>g> the percentage of households in a high state will<br />
decrease by 0.019. The results with respect to diversity of opini<strong>on</strong>, multiculturalism <str<strong>on</strong>g>and</str<strong>on</strong>g> community<br />
news are reas<strong>on</strong>ably intuitive. C<strong>on</strong>solidati<strong>on</strong> of TV stati<strong>on</strong>s is associated with the softening<br />
of media competiti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the provisi<strong>on</strong> of less diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> less local news, which is costly to<br />
produce. 18 The result with respect to advertising is c<strong>on</strong>sistent with Crawford’s (2007) finding<br />
that independent TV stati<strong>on</strong>s provide more advertising per program but charge lower prices to<br />
advertisers. 19<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Table 9. Sec<strong>on</strong>d-Stage Ordered Probit Estimates of <str<strong>on</strong>g>Media</str<strong>on</strong>g> Supply Resp<strong>on</strong>ses<br />
VOICES<br />
VOICES×STATIONS<br />
STATIONS<br />
AGE<br />
EDUC<br />
GENDER<br />
INCOME<br />
RACE<br />
λ m<br />
DIVERSITY OF<br />
OPINION<br />
0.0767 ***<br />
(0.0263)<br />
-0.0011<br />
(0.0011)<br />
-0.0195<br />
(0.0161)<br />
0.0396 ***<br />
(0.0098)<br />
0.141 ***<br />
(0.0164)<br />
-0.0397<br />
(0.0295)<br />
0.0574 ***<br />
(0.0124)<br />
0.0623 *<br />
(0.0334)<br />
-0.0550 **<br />
(0.0260)<br />
COMMUNITY<br />
NEWS<br />
0.0491 *<br />
(0.0270)<br />
-0.0018 **<br />
(0.0008)<br />
0.0006<br />
(0.0167)<br />
0.0928 ***<br />
(0.0109)<br />
0.0824 ***<br />
(0.0195)<br />
0.1310 ***<br />
(0.0303)<br />
-0.0022<br />
(0.0122)<br />
-0.1210 ***<br />
(0.0366)<br />
-0.0250<br />
(0.0170)<br />
MULTI-<br />
CULTURALISM<br />
0.1120 ***<br />
(0.0230)<br />
-0.0015 *<br />
(0.0008)<br />
-0.0414 ***<br />
(0.0147)<br />
0.0252 **<br />
(0.0103)<br />
0.1240 ***<br />
(0.0157)<br />
0.0523 *<br />
(0.0301)<br />
-0.0300 **<br />
(0.0152)<br />
-0.0788 *<br />
(0.0406)<br />
-0.0540 ***<br />
(0.0210)<br />
ADVERTISING<br />
0.0445 ***<br />
(0.0257)<br />
-0.001<br />
(0.0008)<br />
-0.0007<br />
(0.0142)<br />
0.111 ***<br />
(0.0101)<br />
0.0999 ***<br />
(0.0191)<br />
0.0196<br />
(0.0250)<br />
0.0295 **<br />
(0.0117)<br />
0.2210 ***<br />
(0.0380)<br />
0.0074<br />
(0.0166)<br />
Likelihood -4,844.4 -5,218.4 -5,182.8 -4,910.8<br />
NOTES: Estimated by weighted maximum likelihood. Bootstrapped st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard errors in parentheses. ***denotes significant at the <strong>on</strong>e-<br />
percent level. **denotes significant at the five-percent level. *denotes significant at the ten-percent level. Estimated cutoff parameters<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> estimated parameters for the media alternative dummy variables are not reported. Number of observati<strong>on</strong>s is 5,102.<br />
Table 10. Mean Change in Predicted Probabilities<br />
DIVERSITY OF<br />
OPINION<br />
MULTI-<br />
CULTURALISM<br />
ADVERTISING<br />
COMMUNITY<br />
NEWS<br />
ΔPLΔX 0.0159 0.0342 0.0066 0.0084<br />
ΔPMΔX 0.0031 -0.0116 0.0053 -0.0003<br />
ΔPHΔX -0.0190 -0.0226 -0.0119 -0.0081<br />
NOTES: Sample means are calculated from each individual resp<strong>on</strong>dent’s predicted probabilities. ΔP L = P L1 - P L0, ΔP M = P M1 - P M0, ΔP H = P H1 -<br />
P H0, <str<strong>on</strong>g>and</str<strong>on</strong>g> ΔX = ΔVOICES = -1.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 23
24<br />
5.2 <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> C<strong>on</strong>sumer Welfare<br />
5.2.1 Algorithm<br />
We use our dem<str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> supply resp<strong>on</strong>se estimates from Steps One <str<strong>on</strong>g>and</str<strong>on</strong>g> Two to measure the<br />
impact <strong>on</strong> c<strong>on</strong>sumer welfare from the change in media market structure. The procedure to calculate<br />
the changes to c<strong>on</strong>sumer welfare is:<br />
(i) With the existing data, use the estimated coefficients from Table 9 to predict each resp<strong>on</strong>dent’s<br />
“before” probability distributi<strong>on</strong> of low, medium <str<strong>on</strong>g>and</str<strong>on</strong>g> high values for each of the four<br />
n<strong>on</strong>-price news service characteristics. Let PL0 be the before probability of a low level of the<br />
characteristics, PM0 is the before probability of a medium level, <str<strong>on</strong>g>and</str<strong>on</strong>g> PH0 is the before probability<br />
of a high level.<br />
(ii) Approximate the change in media market structure by reducing the number of independent<br />
TV stati<strong>on</strong>s in the sample by <strong>on</strong>e, all other things held c<strong>on</strong>stant. Use the estimated coefficients<br />
from Table 9 to predict each resp<strong>on</strong>dent’s “after” probability distributi<strong>on</strong> of low,<br />
medium <str<strong>on</strong>g>and</str<strong>on</strong>g> high values for each of the four news service characteristics. Let PL1 be the after<br />
probability of a low level of the characteristic, PM1 is the after probability of a medium level,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> of the feature; <str<strong>on</strong>g>and</str<strong>on</strong>g> PH1 is the after probability of a high level.<br />
(iii) Use the probabilities in (i) <str<strong>on</strong>g>and</str<strong>on</strong>g> (ii) to form, for each resp<strong>on</strong>dent, ∆PL ∆X , ∆PM ∆X <str<strong>on</strong>g>and</str<strong>on</strong>g> ∆PH ∆X , for<br />
each news service characteristics, where ΔPL = PL1 - PL0 , ΔPM = PM1 - PM0 ,<br />
ΔPH = PH1 - PH0 , <str<strong>on</strong>g>and</str<strong>on</strong>g> ΔX = ΔVOICES = -1.<br />
(iv) Use the estimates of marginal WTP in the bottom panel of Table 8 <str<strong>on</strong>g>and</str<strong>on</strong>g> the change in predicted<br />
probabilities in (iii) above to evaluate the change in expected benefit (see Appendix A,<br />
Equati<strong>on</strong> 7) for each resp<strong>on</strong>dent <str<strong>on</strong>g>and</str<strong>on</strong>g> for each n<strong>on</strong>-price news service characteristic.<br />
(v) Sort the expected welfare changes in (iv) for each resp<strong>on</strong>dent by the number of TV stati<strong>on</strong>s,<br />
ranging from five to 20. Calculate the mean expected welfare change per m<strong>on</strong>th for all<br />
resp<strong>on</strong>dents in a market with five stati<strong>on</strong>s, six stati<strong>on</strong>s, … , <str<strong>on</strong>g>and</str<strong>on</strong>g> 20 stati<strong>on</strong>s.<br />
(vi) Use the FCC (2011) data to count the number of populati<strong>on</strong> households in a televisi<strong>on</strong> market<br />
with five stati<strong>on</strong>s, six stati<strong>on</strong>s, … , <str<strong>on</strong>g>and</str<strong>on</strong>g> 20 stati<strong>on</strong>s.<br />
(vii) Calculate the aggregate annual change in c<strong>on</strong>sumer welfare for each market size by multiplying<br />
(v) by 12 by (vi) for each level of the number of televisi<strong>on</strong> stati<strong>on</strong>s, i.e., five stati<strong>on</strong>s, six<br />
stati<strong>on</strong>s, … , <str<strong>on</strong>g>and</str<strong>on</strong>g> 20 stati<strong>on</strong>s.<br />
5.2.2 Estimates of C<strong>on</strong>sumer Welfare<br />
Table 11 <str<strong>on</strong>g>and</str<strong>on</strong>g> Figure 1 present estimates of the impact <strong>on</strong> c<strong>on</strong>sumer welfare from a marginal decrease<br />
in the number of independent TV stati<strong>on</strong>s for all market sizes ranging from five to 20 TV stati<strong>on</strong>s.<br />
Columns 3 through 6 of Table 11 report average c<strong>on</strong>sumer welfare per m<strong>on</strong>th, <str<strong>on</strong>g>and</str<strong>on</strong>g> columns 7 through<br />
12 report annual aggregate welfare. 20 The first important observati<strong>on</strong> is that the average welfare<br />
effects per m<strong>on</strong>th depend <strong>on</strong> market size, with smaller markets experiencing larger effects in absolute<br />
terms. The intuiti<strong>on</strong> for this finding is clear. The impact from the loss of an independent voice in<br />
the market will be more acute when there are fewer competitors to fill the void. As a result, the<br />
average c<strong>on</strong>sumer in a small market loses $0.99 per m<strong>on</strong>th, whereas the average c<strong>on</strong>sumer in a<br />
large market loses $0.44 per m<strong>on</strong>th. These losses are equivalent to about $54 milli<strong>on</strong> annually for<br />
all small-market households in the U.S. <str<strong>on</strong>g>and</str<strong>on</strong>g> $16 milli<strong>on</strong> for all large-market households. 21 If the<br />
change in market structure occurs in all markets, for example if two of the big four networks,<br />
ABC, CBS, FOX <str<strong>on</strong>g>and</str<strong>on</strong>g> NBC c<strong>on</strong>solidated, annual aggregate losses nati<strong>on</strong>wide would be about $830<br />
milli<strong>on</strong>. For comparis<strong>on</strong>, this represents about seven percent of the total operating costs for CBS<br />
in 2010. 22<br />
Given the WTP estimates in Table 8, it is not surprising that the average welfare losses per<br />
m<strong>on</strong>th from DIVERSITY OF OPINION <str<strong>on</strong>g>and</str<strong>on</strong>g> COMMUNITY NEWS are greater than for<br />
MULTICULTURALISM in almost all markets. However, while DIVERSITY OF OPINION<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Table 11: Impact <strong>on</strong> C<strong>on</strong>sumer Welfare From a Change in <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
Size<br />
Pop.<br />
Share<br />
DIV MCULT ADV CNEWS DIV MCULT ADV CNEWS Total<br />
Average c<strong>on</strong>sumer welfare per m<strong>on</strong>th<br />
(Dollars per m<strong>on</strong>th)<br />
Annual aggregate welfare in market<br />
(Dollars in milli<strong>on</strong>s)<br />
Total less<br />
ADV<br />
5 0.050 -0.61 -0.23 0.30 -0.45 -32.87 -12.58 16.49 -24.40 -53.36 -69.85<br />
(0.05) (0.07) (0.04) (0.03) (2.87) (3.60) (2.13) (1.67)<br />
6 0.061 -0.59 -0.22 0.29 -0.42 -38.84 -14.72 19.19 -27.51 -61.87 -81.06<br />
(0.07) (0.09) (0.05) (0.04) (4.36) (5.69) (3.36) (2.44)<br />
7 0.091 -0.59 -0.23 0.29 -0.40 -57.83 -22.67 28.55 -39.29 -91.24 -119.8<br />
(0.06) (0.07) (0.04) (0.03) (5.57) (7.20) (3.91) (2.84)<br />
8 0.081 -0.57 -0.22 0.28 -0.37 -49.66 -19.47 24.49 -32.84 -77.47 -102.0<br />
(0.06) (0.07) (0.04) (0.03) (4.99) (6.11) (3.57) (2.36)<br />
9 0.095 -0.54 -0.21 0.27 -0.35 -55.97 -21.94 28.06 -36.21 -86.06 -114.1<br />
(0.05) (0.07) (0.04) (0.03) (5.46) (7.28) (4.30) (2.76)<br />
10 0.056 -0.55 -0.21 0.27 -0.34 -33.08 -12.90 16.20 -20.59 -50.38 -66.57<br />
(0.06) (0.09) (0.04) (0.03) (3.91) (5.30) (2.69) (1.75)<br />
11 0.099 -0.53 -0.21 0.26 -0.32 -57.29 -22.36 28.33 -34.51 -85.83 -114.2<br />
(0.05) (0.06) (0.03) (0.02) (5.49) (6.76) (3.70) (2.29)<br />
12 0.069 -0.52 -0.20 0.25 -0.30 -38.83 -15.00 18.40 -22.13 -57.57 -75.97<br />
(0.06) (0.08) (0.04) (0.03) (4.23) (5.63) (2.80) (1.89)<br />
13 0.024 -0.51 -0.20 0.24 -0.28 -13.31 -5.12 6.31 -7.31 -19.43 -25.74<br />
(0.09) (0.12) (0.06) (0.03) (2.34) (3.07) (1.44) (0.88)<br />
14 0.093 -0.47 -0.17 0.24 -0.26 -47.76 -17.06 23.77 -26.01 -67.05 -90.83<br />
(0.05) (0.06) (0.03) (0.02) (4.64) (6.43) (3.14) (1.80)<br />
15 0.03 -0.48 -0.18 0.22 -0.24 -15.57 -5.99 7.24 -7.72 -22.04 -29.28<br />
(0.08) (0.12) (0.05) (0.03) (2.74) (3.77) (1.70) (1.00)<br />
16 0.079 -0.45 -0.16 0.23 -0.23 -38.73 -14.06 19.58 -19.41 -52.62 -72.20<br />
(0.05) (0.07) (0.03) (0.02) (4.62) (6.21) (2.93) (1.55)<br />
17 0.072 -0.45 -0.16 0.21 -0.20 -34.85 -12.56 16.45 -15.83 -46.79 -63.24<br />
(0.06) (0.08) (0.03) (0.02) (4.29) (6.10) (2.72) (1.39)<br />
18 0.043 -0.45 -0.17 0.21 -0.19 -21.00 -8.03 9.67 -8.74 -28.10 -37.77<br />
(0.07) (0.10) (0.05) (0.02) (3.38) (4.71) (2.19) (1.05)<br />
19 0.026 -0.44 -0.19 0.21 -0.17 -12.44 -5.35 5.90 -4.87 -16.76 -22.66<br />
(0.08) (0.11) (0.05) (0.02) (2.23) (3.17) (1.43) (0.60)<br />
20 0.032 -0.38 -0.14 0.17 -0.09 -13.26 -4.98 5.93 -3.20 -15.51 -21.44<br />
(0.03) (0.04) (0.02) (0.01) (1.01) (1.50) (0.62) (0.19)<br />
Total 1 -0.52 -0.20 0.25 -0.31 -561.3 -214.8 274.6 -330.6 -832.1 -1,107<br />
(0.01) (0.01) (0.01) (0.01) (15.35) (15.35) (10.46) (6.82)<br />
NOTE: Bootstrapped st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard errors in parentheses. The change in market structure is a <strong>on</strong>e-unit reducti<strong>on</strong> in the number of independent TV voices in the market,<br />
all other things held c<strong>on</strong>stant. There are 90,193,905 populati<strong>on</strong> households in markets from five to 20 TV stati<strong>on</strong>s (FCC, 2011). Pop. share is the number of populati<strong>on</strong><br />
households in the market divided by populati<strong>on</strong> households. DIV is diversity of opini<strong>on</strong> in the reporting of informati<strong>on</strong>, MCULT is coverage of multiculturalism<br />
issues, ADV is amount of space or time devoted to advertising, <str<strong>on</strong>g>and</str<strong>on</strong>g> CNEWS is amount of informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events. Total losses of $832.1 milli<strong>on</strong><br />
are the sum of the individual market losses.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 25
26<br />
c<strong>on</strong>tinues to have significant negative impacts in both small (-$0.61) <str<strong>on</strong>g>and</str<strong>on</strong>g> large (-$0.38) markets,<br />
the effect for COMMUNITY NEWS quickly dissipates from -$0.45 to -$0.09 as the number of<br />
stati<strong>on</strong>s in the market increases. MULTICULTURALISM follows a similar trend to DIVERSITY<br />
OF OPINION, losing about 40 percent of its negative impact from small (-$0.23) to large<br />
(-$0.14) markets. ADVERTISING also follows a similar trend to DIVERSITY OF OPINION <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
MULTICULTURALISM, losing about 40 percent of its positive impact from small ($0.30) to large<br />
($0.17) markets. 23<br />
A final important observati<strong>on</strong> is the potential tradeoff between the amount of diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism<br />
in news, <str<strong>on</strong>g>and</str<strong>on</strong>g> the amount of space <str<strong>on</strong>g>and</str<strong>on</strong>g>/or time devoted to advertising. In the c<strong>on</strong>solidati<strong>on</strong> of<br />
two independent TV stati<strong>on</strong>s, c<strong>on</strong>sumers lose <strong>on</strong> certain characteristics because there is less diversity<br />
of opini<strong>on</strong>, less coverage of multiculturalism issues, <str<strong>on</strong>g>and</str<strong>on</strong>g> less community news, but they gain<br />
because there is less space <str<strong>on</strong>g>and</str<strong>on</strong>g> time devoted to advertising. 24 Specifically, Columns 3 through 6 of<br />
Table 10 show that, <strong>on</strong> average, about 24 percent of the annual m<strong>on</strong>thly losses to c<strong>on</strong>sumers from<br />
less diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism in each market are offset by less exposure to advertising. This illustrates<br />
an important feature of the news service experience in our data: the first-order effects from c<strong>on</strong>solidati<strong>on</strong><br />
are, potentially, not all bad for c<strong>on</strong>sumers. Nevertheless, c<strong>on</strong>sumers <str<strong>on</strong>g>and</str<strong>on</strong>g> policy-makers<br />
should be c<strong>on</strong>cerned about the impacts from a “virtual merger,” where TV stati<strong>on</strong>s in the same<br />
market combine their news operati<strong>on</strong>s under joint operating <str<strong>on</strong>g>and</str<strong>on</strong>g> marketing agreements without<br />
actually merging. Since a virtual merger is likely to result in less diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism but not<br />
less advertising, the welfare reducti<strong>on</strong>s in Table 11 would be even more pr<strong>on</strong>ounced. For example,<br />
Column 12 shows that if the virtual merger occurred in all markets, annual aggregate losses<br />
nati<strong>on</strong>wide would be about $1.1 billi<strong>on</strong>.<br />
Figure 1. Change in Average C<strong>on</strong>sumer Welfare Per M<strong>on</strong>th From a Change in <str<strong>on</strong>g>Market</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Structure</str<strong>on</strong>g><br />
0.40<br />
0.20<br />
0.00<br />
0.20<br />
0.40<br />
0.60<br />
0.80<br />
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Diversity</str<strong>on</strong>g> of Opini<strong>on</strong><br />
Multiculturalism<br />
Advertising<br />
Community News<br />
NOTES: Vertical axis is dollars per m<strong>on</strong>th <str<strong>on</strong>g>and</str<strong>on</strong>g> horiz<strong>on</strong>tal axis is number of TV stati<strong>on</strong>s. The change in market structure is a <strong>on</strong>e-unit<br />
reducti<strong>on</strong> in the number of independent TV voices in the market, all other things held c<strong>on</strong>stant.<br />
SOURCE: Table 11.
6. C<strong>on</strong>clusi<strong>on</strong>s<br />
This study examined market structure <str<strong>on</strong>g>and</str<strong>on</strong>g> media diversity. A differentiated-product model was<br />
used to estimate c<strong>on</strong>sumers’ dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for certain characteristics in their local news service, which<br />
includes the offerings from newspapers, radio, TV, the Internet, <str<strong>on</strong>g>and</str<strong>on</strong>g> smartph<strong>on</strong>es. The model<br />
captures both private- <str<strong>on</strong>g>and</str<strong>on</strong>g> public-good aspects of news services by including the amount of advertising<br />
in the household’s full cost of c<strong>on</strong>sumpti<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g> by characterizing service in terms of diversity<br />
of opini<strong>on</strong> in the reporting of informati<strong>on</strong>, coverage of multiculturalism issues, <str<strong>on</strong>g>and</str<strong>on</strong>g> the amount<br />
of informati<strong>on</strong> <strong>on</strong> community news <str<strong>on</strong>g>and</str<strong>on</strong>g> events. The empirical results show that the representative<br />
c<strong>on</strong>sumer values diversity in news reporting, more informati<strong>on</strong> <strong>on</strong> women <str<strong>on</strong>g>and</str<strong>on</strong>g> minorities, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
more informati<strong>on</strong> <strong>on</strong> community news. Many c<strong>on</strong>sumers, however, have distaste for advertising,<br />
which likely reflects their c<strong>on</strong>sumpti<strong>on</strong> of general, all-purpose advertising delivered by traditi<strong>on</strong>al<br />
media.<br />
The dem<str<strong>on</strong>g>and</str<strong>on</strong>g> estimates are used to calculate the impact <strong>on</strong><br />
c<strong>on</strong>sumer welfare from a marginal decrease in the number<br />
of independent TV stati<strong>on</strong>s that lowers the amount<br />
of diversity, localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising in the market. The<br />
predicti<strong>on</strong> of n<strong>on</strong>-price effects is appropriate for media<br />
markets, where some households make no direct payment<br />
for c<strong>on</strong>sumpti<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g> appears to be novel in the simulated<br />
merger literature. Our results show that c<strong>on</strong>sumer welfare<br />
decreases following the change in media market structure,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> that the losses are smaller in large markets. For example,<br />
small-market c<strong>on</strong>sumers lose $54 milli<strong>on</strong> annually while<br />
large-market c<strong>on</strong>sumers lose $16 milli<strong>on</strong>. Should the<br />
change in market structure occur in all markets, total<br />
losses would be about $830 milli<strong>on</strong>.<br />
Policy-makers struggling with legacy<br />
rules designed to protect traditi<strong>on</strong>al<br />
broadcast media in an increasingly<br />
diverse <str<strong>on</strong>g>and</str<strong>on</strong>g> competitive digital<br />
envir<strong>on</strong>ment should c<strong>on</strong>sider these<br />
findings as they weigh how c<strong>on</strong>sum-<br />
ers are best served.<br />
—Fern<str<strong>on</strong>g>and</str<strong>on</strong>g>o R. Laguarda<br />
We make no claims as to whether media ownership rules should be relaxed or tightened. We<br />
note that the estimated total losses of $830 milli<strong>on</strong> approximate the extreme case of c<strong>on</strong>solidati<strong>on</strong><br />
between two major nati<strong>on</strong>al media players <str<strong>on</strong>g>and</str<strong>on</strong>g>, as such, this is an upper-bound calculati<strong>on</strong>. The<br />
large c<strong>on</strong>sumer losses in small TV markets relative to large markets are potentially important. The<br />
tradeoff between diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism <str<strong>on</strong>g>and</str<strong>on</strong>g> advertising also is notable because it highlights an<br />
additi<strong>on</strong>al potential benefit for c<strong>on</strong>siderati<strong>on</strong> during the analysis of a media market merger. It also<br />
provides a new angle from which to assess the efficacy of media ownership rules.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 27
28<br />
Endnotes<br />
1. The number of independent TV stati<strong>on</strong>s is determined by counting all stati<strong>on</strong>s within a market. For every stati<strong>on</strong><br />
with a comm<strong>on</strong> parent, we then count <strong>on</strong>ly the first of those stati<strong>on</strong>s al<strong>on</strong>g with the remaining stati<strong>on</strong>s with no comm<strong>on</strong><br />
parent. See Secti<strong>on</strong> 3.4.<br />
2. In 2010, 58 percent of the public turned to TV for news, 44 percent used the Internet or cellular teleph<strong>on</strong>e, 34 percent<br />
relied <strong>on</strong> radio, <str<strong>on</strong>g>and</str<strong>on</strong>g> 31 percent read newspapers (Pew <str<strong>on</strong>g>Research</str<strong>on</strong>g> Center; 2010).<br />
3. Some papers exploit a law or regulatory change to document the effect of a change in market structure <strong>on</strong> the supply<br />
of media <str<strong>on</strong>g>and</str<strong>on</strong>g> telecommunicati<strong>on</strong>s services. Berry <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldfogel (2001) show that, following the Telecommunicati<strong>on</strong>s<br />
Act (“Act”) of 1996, c<strong>on</strong>solidati<strong>on</strong> reduced radio stati<strong>on</strong> entry <str<strong>on</strong>g>and</str<strong>on</strong>g> increased product variety. Ec<strong>on</strong>omides et al. (2008)<br />
show that following the Act, households benefited more from the new plan <str<strong>on</strong>g>and</str<strong>on</strong>g> quality differences offered by entrants<br />
into local teleph<strong>on</strong>e markets than from price decreases.<br />
4. Appendix B presents external evidence indicating that the informati<strong>on</strong> reported by individual survey resp<strong>on</strong>dents<br />
is a reas<strong>on</strong>ably good proxy for the diversity of news service alternatives in different markets. Specifically, Table B1<br />
reports the estimates from a simple ordered-probit model of DIVERSITY OF OPINION, MULTICULTURALISM or<br />
COMMUNITY NEWS <strong>on</strong> various measures of diversity <str<strong>on</strong>g>and</str<strong>on</strong>g> localism for radio <str<strong>on</strong>g>and</str<strong>on</strong>g> TV stati<strong>on</strong>s c<strong>on</strong>structed from FCC<br />
data (2011). The results show that in general the FCC measures correlate well to the amount of diversity reported by<br />
individual households in our survey.<br />
5. Resp<strong>on</strong>dents were asked to c<strong>on</strong>sider what is available in their local media envir<strong>on</strong>ment, rather than what they usually<br />
listen to or view. This represents a statement about the amount <str<strong>on</strong>g>and</str<strong>on</strong>g> quality of informati<strong>on</strong> programming being produced<br />
by media sources for their c<strong>on</strong>sumpti<strong>on</strong>.<br />
6. Up<strong>on</strong> completi<strong>on</strong> of their cognitive buildup questi<strong>on</strong>s, an <strong>on</strong>line algorithm calculated each individual’s total cost of<br />
their local entertainment <str<strong>on</strong>g>and</str<strong>on</strong>g> news service <str<strong>on</strong>g>and</str<strong>on</strong>g> assigned the appropriate cost range for their choice occasi<strong>on</strong>s.<br />
7. KN recruitment uses dual sampling frames that include listed <str<strong>on</strong>g>and</str<strong>on</strong>g> unlisted teleph<strong>on</strong>e numbers, teleph<strong>on</strong>e <str<strong>on</strong>g>and</str<strong>on</strong>g> n<strong>on</strong>teleph<strong>on</strong>e<br />
households, <str<strong>on</strong>g>and</str<strong>on</strong>g> cellph<strong>on</strong>e-<strong>on</strong>ly households, as well as households with <str<strong>on</strong>g>and</str<strong>on</strong>g> without Internet access. If<br />
required, households are provided with a laptop <str<strong>on</strong>g>and</str<strong>on</strong>g> free Internet access to complete surveys, but they do not participate<br />
in the incentive program.<br />
8. Complementary data from the Nielsen Company (2010) <str<strong>on</strong>g>and</str<strong>on</strong>g> the U.S. Census Bureau (2011) indicate that about 93<br />
percent of people watch TV, 82 percent listen to radio, <str<strong>on</strong>g>and</str<strong>on</strong>g> 77 percent use the Internet. About 67 percent of resp<strong>on</strong>dents<br />
read a newspaper regularly, <str<strong>on</strong>g>and</str<strong>on</strong>g> 25 percent own a smartph<strong>on</strong>e.<br />
9. This is reas<strong>on</strong>ably close to the combined annual costs of membership in 2011. For example, Rocky Mountain PBS<br />
offers an annual membership for $40 <str<strong>on</strong>g>and</str<strong>on</strong>g> Colorado Public Radio for $120. These membership costs vary between<br />
states.<br />
10. Televisi<strong>on</strong> <str<strong>on</strong>g>Market</str<strong>on</strong>g> Area, or “market,” is a geographical regi<strong>on</strong> where all households receive the same offerings from<br />
TV stati<strong>on</strong>s. The seven markets outside our sample are: Bend, OR; Fairbanks, AK; Gr<str<strong>on</strong>g>and</str<strong>on</strong>g> Juncti<strong>on</strong>, CO; Missoula,<br />
MT; North Platte, NE; Ottumwa, IA-Kirksville, MO; <str<strong>on</strong>g>and</str<strong>on</strong>g> Presque, ME. All seven are small markets with five or fewer<br />
TV stati<strong>on</strong>s. As shown in Table 6, the remaining small markets in our sample cover 8.43 percent of households. FCC<br />
(2011) data show that 8.37 percent of populati<strong>on</strong> households were in small markets as of December 2009.<br />
11. For a robustness check, we specified an alternative set of dummy variables that also c<strong>on</strong>trolled for subsets of radio<br />
(i.e., satellite <str<strong>on</strong>g>and</str<strong>on</strong>g> broadcast radio) <str<strong>on</strong>g>and</str<strong>on</strong>g> TV (i.e., cable, satellite <str<strong>on</strong>g>and</str<strong>on</strong>g> broadcast TV). Ordered probit estimates of media<br />
supply resp<strong>on</strong>ses <str<strong>on</strong>g>and</str<strong>on</strong>g> estimates of the impacts <strong>on</strong> c<strong>on</strong>sumer welfare from a change in market structure, not reported<br />
here, are similar to those presented in Tables 9 <str<strong>on</strong>g>and</str<strong>on</strong>g> 11.<br />
12. All mixed logit models were estimated by simulati<strong>on</strong> using 100 Halt<strong>on</strong> draws. For robustness, we estimated several<br />
model specificati<strong>on</strong>s using 500 draws <str<strong>on</strong>g>and</str<strong>on</strong>g> the results were very similar.<br />
13. Model (iii) permits correlati<strong>on</strong> between the n<strong>on</strong>-price parameters.The correlati<strong>on</strong> matrices are available <strong>on</strong> request<br />
from the authors.<br />
14. By our definiti<strong>on</strong>s, low-income households (INCOME < $25,000) comprise 20.5 percent of the sample, mediumincome<br />
households ($25,000 ≤ INCOME < $50,000) comprise 24.6 percent, <str<strong>on</strong>g>and</str<strong>on</strong>g> high-income households ($50,000 ≤<br />
INCOME) comprise 54.9 percent.<br />
15. We also estimated a variant of model (iv) that included an additi<strong>on</strong>al interacti<strong>on</strong> between an indicator of college educati<strong>on</strong><br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> COST. This additi<strong>on</strong>al interacti<strong>on</strong> was not statistically significant at c<strong>on</strong>venti<strong>on</strong>al levels <str<strong>on</strong>g>and</str<strong>on</strong>g> the results,<br />
not reported, are similar to those reported in Table 7.<br />
16. For example, Comcast targets specific customer types through its Video-<strong>on</strong>-Dem<str<strong>on</strong>g>and</str<strong>on</strong>g> service <str<strong>on</strong>g>and</str<strong>on</strong>g> then permits the<br />
customer to select advertisements she/ he wants to view with via remote c<strong>on</strong>trol. See http://www.comcastspotlight.<br />
com/advertising-soluti<strong>on</strong>s/<strong>on</strong>-dem<str<strong>on</strong>g>and</str<strong>on</strong>g>.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
17. There are no systematic biases toward a specific alternative within the choice scenario. C<strong>on</strong>sumers chose their actual<br />
news service at home 29.1 percent of the time, news service alternative A 34.4 percent of the time, <str<strong>on</strong>g>and</str<strong>on</strong>g> news service<br />
alternative B 36.5 percent of the time. The number of sec<strong>on</strong>ds it took resp<strong>on</strong>dents to answer each choice occasi<strong>on</strong><br />
remained essentially c<strong>on</strong>stant over the eight choice occasi<strong>on</strong>s. Because some of our data are from repeated choices,<br />
we also need to be c<strong>on</strong>cerned with survey fatigue. For a robustness check, we estimated all model specificati<strong>on</strong>s’ (i)<br />
through (v) <strong>on</strong> the data for the first four choice questi<strong>on</strong>s versus the sec<strong>on</strong>d four questi<strong>on</strong>s. The results, not reported,<br />
show reas<strong>on</strong>ably similar estimates for the two subsamples of data. There is no systematic pattern that could be taken<br />
as evidence of survey fatigue.<br />
18. By definiti<strong>on</strong>, a reducti<strong>on</strong> in the number of independent TV stati<strong>on</strong>s means there are fewer viewpoints in the market,<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> as a result, less diversity of opini<strong>on</strong>.<br />
19. Brown <str<strong>on</strong>g>and</str<strong>on</strong>g> Alex<str<strong>on</strong>g>and</str<strong>on</strong>g>er (2004) find a positive correlati<strong>on</strong> between TV market c<strong>on</strong>centrati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the price of advertising<br />
per viewer. They argue that when c<strong>on</strong>sumers’ elasticity of viewing with respect to advertising is weak, a decrease<br />
in the fracti<strong>on</strong> of broadcast time devoted to advertising can lead to a decrease in the overall amount of advertising<br />
supplied <str<strong>on</strong>g>and</str<strong>on</strong>g> an increase in the price to advertisers (See also Cunningham <str<strong>on</strong>g>and</str<strong>on</strong>g> Alex<str<strong>on</strong>g>and</str<strong>on</strong>g>er, 2004).<br />
20. The reported st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard errors are calculated using a bootstrapping method. For example, for Row 1 we c<strong>on</strong>struct<br />
the benefit (Appendix A, Equati<strong>on</strong> 7) for each resp<strong>on</strong>dent in markets with five TV stati<strong>on</strong>s. We then draw marginal<br />
utility values from the multivariate normal distributi<strong>on</strong> implied by the mean <str<strong>on</strong>g>and</str<strong>on</strong>g> covariance parameters reported in<br />
Columns 8 <str<strong>on</strong>g>and</str<strong>on</strong>g> 9 of Table 7. These values are used to evaluate the benefit for each resp<strong>on</strong>dent <str<strong>on</strong>g>and</str<strong>on</strong>g> to obtain an estimate<br />
of the mean c<strong>on</strong>sumer welfare per m<strong>on</strong>th. We run this simulati<strong>on</strong> 500 times <str<strong>on</strong>g>and</str<strong>on</strong>g> report the mean <str<strong>on</strong>g>and</str<strong>on</strong>g> st<str<strong>on</strong>g>and</str<strong>on</strong>g>ard<br />
error of the sampling distributi<strong>on</strong> for c<strong>on</strong>sumer welfare per m<strong>on</strong>th.<br />
21. There are 90,193,905 populati<strong>on</strong> households in markets with from five to 20 TV stati<strong>on</strong>s. Total annual aggregate<br />
welfare for small markets is -$53.36 milli<strong>on</strong> = (0.050×90,193,905)×12×-0.99. Total annual aggregate welfare for large<br />
markets is -$15.51 milli<strong>on</strong> = (0.032×90,193,905)×12×-0.44.<br />
22. See CBS Corporati<strong>on</strong> income statements at: http://ycharts.com/financials/CBS/income_statement/annual.<br />
23. The FCC uses several measures of market structure when discussing ownership rules. For robustness, we examined<br />
how sensitive our results are to an alternative specificati<strong>on</strong> of the media supply resp<strong>on</strong>se Equati<strong>on</strong> 4 (in Appendix A)<br />
that c<strong>on</strong>trols for the number of daily newspapers in the market (NEWSPAPERS) <str<strong>on</strong>g>and</str<strong>on</strong>g> the number of radio stati<strong>on</strong>s<br />
(RADIO STATIONS). The estimates of the two-stage ordered probit model of supply, <str<strong>on</strong>g>and</str<strong>on</strong>g> the estimates of the impacts<br />
<strong>on</strong> c<strong>on</strong>sumer welfare from a change in market structure, not reported here, are similar to those presented in Tables 9<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> 11.<br />
24. The reducti<strong>on</strong> in advertising does not mean that the two merged will be worth less. Profits should increase as a result<br />
of higher advertising rates <str<strong>on</strong>g>and</str<strong>on</strong>g>/or cost efficiencies in the producti<strong>on</strong> of news programming.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 29
30<br />
References<br />
Berry, S., <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldfogel, J. 2001. “Do Mergers Increase Product Variety? Evidence from Radio<br />
Broadcasting.” Quarterly Journal of Ec<strong>on</strong>omics, 116(3), 1009–1025.<br />
Brown, K., <str<strong>on</strong>g>and</str<strong>on</strong>g> Alex<str<strong>on</strong>g>and</str<strong>on</strong>g>er, P. 2004. “<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g>, Viewer Welfare, <str<strong>on</strong>g>and</str<strong>on</strong>g> Advertising Rates in<br />
Local Broadcast <str<strong>on</strong>g>Market</str<strong>on</strong>g>s.” Ec<strong>on</strong>omic Letters, 86(3), 331–337.<br />
Brownst<strong>on</strong>e, D., <str<strong>on</strong>g>and</str<strong>on</strong>g> Train, K. 1999. “Forecasting New Product Penetrati<strong>on</strong> with Flexible<br />
Substituti<strong>on</strong> Patterns.” Journal of Ec<strong>on</strong>ometrics, 89(1), 109–129.<br />
Chen, Y., <str<strong>on</strong>g>and</str<strong>on</strong>g> Savage, S. 2011. “The Effects of Competiti<strong>on</strong> <strong>on</strong> the Price for Cable Modem Internet<br />
Access.” The Review of Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Statistics, 93(1), 201–217.<br />
Crawford, G. 2007. “Televisi<strong>on</strong> Stati<strong>on</strong> Ownership <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> the Quantity <str<strong>on</strong>g>and</str<strong>on</strong>g> Quality of TV<br />
<str<strong>on</strong>g>Program</str<strong>on</strong>g>ming,” FCC <str<strong>on</strong>g>Media</str<strong>on</strong>g> Ownership Study #3, Washingt<strong>on</strong>, D.C., FCC.<br />
Cunningham, B., <str<strong>on</strong>g>and</str<strong>on</strong>g> Alex<str<strong>on</strong>g>and</str<strong>on</strong>g>er, P. 2004. “A Theory of Broadcast <str<strong>on</strong>g>Media</str<strong>on</strong>g> C<strong>on</strong>centrati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
Commercial Advertising.” Journal of Public Ec<strong>on</strong>omic Theory, 6(4), 557–575.<br />
Domberger, S., <str<strong>on</strong>g>and</str<strong>on</strong>g> Sherr, A. 1989. “The Impact of Competiti<strong>on</strong> <strong>on</strong> Pricing <str<strong>on</strong>g>and</str<strong>on</strong>g> Quality of Legal<br />
Services.” Internati<strong>on</strong>al Review of Law <str<strong>on</strong>g>and</str<strong>on</strong>g> Ec<strong>on</strong>omics, 9, 41–56.<br />
Ec<strong>on</strong>omides, N., Seim, K., <str<strong>on</strong>g>and</str<strong>on</strong>g> Viard, B. 2008. “Quantifying the Benefits of Entry into Local Ph<strong>on</strong>e<br />
Service.” RAND Journal of Ec<strong>on</strong>omics, 39(3), 699–730.<br />
FCC. 2011. Government Furnished Informati<strong>on</strong> <strong>on</strong> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g>. February 17, <str<strong>on</strong>g>Media</str<strong>on</strong>g><br />
Bureau, FCC, Washingt<strong>on</strong>, D.C.<br />
Gentzkow, M. 2007. “Valuing New Goods in a Model with Complementary Online Newspapers.”<br />
American Ec<strong>on</strong>omic Review, 97(3), 713–744.<br />
Gentzkow, M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Shapiro, J. 2010. “What Drives <str<strong>on</strong>g>Media</str<strong>on</strong>g> Slant? Evidence from U.S. Daily<br />
Newspapers.” Ec<strong>on</strong>ometrica, 78(1), 35–71.<br />
Goolsbee, A., <str<strong>on</strong>g>and</str<strong>on</strong>g> Petrin A. 2004. “The C<strong>on</strong>sumer Gains from Direct Broadcast Satellites <str<strong>on</strong>g>and</str<strong>on</strong>g> the<br />
Competiti<strong>on</strong> with Cable TV.” Ec<strong>on</strong>ometrica, 72(2), 351–81.<br />
Heckman, J. 1979. “Sample Selecti<strong>on</strong> Bias as a Specificati<strong>on</strong> Error.” Ec<strong>on</strong>ometrica, 47(1), 153–61.<br />
Huber, J, <str<strong>on</strong>g>and</str<strong>on</strong>g> Zwerina, K. 1996. “The Importance of Utility Balance in Efficient Choice Designs.”<br />
Journal of <str<strong>on</strong>g>Market</str<strong>on</strong>g>ing <str<strong>on</strong>g>Research</str<strong>on</strong>g>, 33, 307–317.<br />
Imbens, G., <str<strong>on</strong>g>and</str<strong>on</strong>g> Wooldridge, J. 2007. “C<strong>on</strong>trol Functi<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> Related Methods.” Lecture Notes 6<br />
for What’s New in Ec<strong>on</strong>ometrics? NBER.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Ivaldi, M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Verboven, F. 2005. “Quantifying the Effects from Horiz<strong>on</strong>tal Mergers in European<br />
Competiti<strong>on</strong> Policy.” Internati<strong>on</strong>al Journal of Industrial Organizati<strong>on</strong>, 23(9/10), 669–691.<br />
Knowledge Networks Inc. 2010. KnowledgePanel® Demographic Profile January 2010. Government<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Academic <str<strong>on</strong>g>Research</str<strong>on</strong>g>, Knowledge Networks.<br />
Matsa, D. 2011. “Competiti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> Product Quality in the Supermarket Industry.” Quarterly<br />
Journal of Ec<strong>on</strong>omics, 126(3), 1539–1591.<br />
Mazzeo, M. 2002. “Competitive Outcomes in Product-Differentiated Oligopoly.” The Review of<br />
Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Statistics, 84(4), 716–728.<br />
Mazzeo, M. 2003. “Competiti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> Service Quality in the U.S. Airline Industry.” Review of<br />
Industrial Organizati<strong>on</strong>, 22(4), 275–296.<br />
Milyo, J. 2007. “The Effects of Cross-Ownership <strong>on</strong> the Local C<strong>on</strong>tent <str<strong>on</strong>g>and</str<strong>on</strong>g> Political Slant of Local<br />
Televisi<strong>on</strong> News.” FCC <str<strong>on</strong>g>Media</str<strong>on</strong>g> Ownership Study #6, Washingt<strong>on</strong>, D.C., FCC.<br />
Nati<strong>on</strong>al Climatic Data Center. 2011. Climate Data Online. Washingt<strong>on</strong>, D.C.: U.S. Department of<br />
Commerce, http://www.ncdc.noaa.gov/cdo-web/.<br />
Nevo, A. 2000. “Mergers with Differentiated Products: The Case of the Ready to Eat Cereal<br />
Industry.” R<str<strong>on</strong>g>and</str<strong>on</strong>g> Journal of Ec<strong>on</strong>omics, 31(3), 395–421.<br />
Nielsen Company. 2010. “iPh<strong>on</strong>e, Android Show Growth,” http://www.marketingcharts.com/wp/<br />
wp-c<strong>on</strong>tent/uploads/2010/08/nielsen-smartph<strong>on</strong>e-penetrati<strong>on</strong>-q2-10-august-2010.JPG.<br />
Olivares, M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Cach<strong>on</strong>, G. 2009. “Competing Retailers <str<strong>on</strong>g>and</str<strong>on</strong>g> Inventory: An Empirical<br />
Investigati<strong>on</strong> of General Motors’ Dealerships in Isolated U.S. <str<strong>on</strong>g>Market</str<strong>on</strong>g>s.” Management Science,<br />
55(9), 1586–1604.<br />
Pew <str<strong>on</strong>g>Research</str<strong>on</strong>g> Center. 2010. “Ideological News Sources: Who Watches <str<strong>on</strong>g>and</str<strong>on</strong>g> Why?” Pew <str<strong>on</strong>g>Research</str<strong>on</strong>g><br />
Center Publicati<strong>on</strong>, Accessed <strong>on</strong> August 14, 2012 at http://www.people-press.org/files/legacypdf/652.pdf.<br />
Pinkse, J., <str<strong>on</strong>g>and</str<strong>on</strong>g> Slade, M. 2004. “Mergers, Br<str<strong>on</strong>g>and</str<strong>on</strong>g> Competiti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the Price of a Pint.” European<br />
Ec<strong>on</strong>omic Review, 48(3), 617–643.<br />
Revelt, D., <str<strong>on</strong>g>and</str<strong>on</strong>g> Train, K. 1998. “Mixed Logit with Repeated Choices: Households’ Choices of<br />
Appliance Efficiency Level.” The Review of Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Statistics, 80(4), 647–657.<br />
Savage, S., <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldman, D. 2011. “C<strong>on</strong>sumer Valuati<strong>on</strong> of <str<strong>on</strong>g>Media</str<strong>on</strong>g> as a Functi<strong>on</strong> of Local <str<strong>on</strong>g>Market</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Structure</str<strong>on</strong>g>.” FCC <str<strong>on</strong>g>Media</str<strong>on</strong>g> Ownership Study #2, Washingt<strong>on</strong>, D.C.: FCC.<br />
Siegelman, P., <str<strong>on</strong>g>and</str<strong>on</strong>g> Waldfogel, J. 2001. “Race <str<strong>on</strong>g>and</str<strong>on</strong>g> Radio: Preference Externalities, Minority<br />
Ownership, <str<strong>on</strong>g>and</str<strong>on</strong>g> the Provisi<strong>on</strong> of <str<strong>on</strong>g>Program</str<strong>on</strong>g>ming to Minorities.” In Baye, M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Nels<strong>on</strong>, J., eds.,<br />
Advances in Applied Microec<strong>on</strong>omics, Vol. 10: Advertising <str<strong>on</strong>g>and</str<strong>on</strong>g> Differentiated Products, Elsevier,<br />
Amsterdam.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 31
32<br />
Singh, V., <str<strong>on</strong>g>and</str<strong>on</strong>g> Zhu, T. 2008. “Pricing <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Market</str<strong>on</strong>g> C<strong>on</strong>centrati<strong>on</strong> in Oligopoly <str<strong>on</strong>g>Market</str<strong>on</strong>g>s.” <str<strong>on</strong>g>Market</str<strong>on</strong>g>ing<br />
Science, 27(6), 1020–1035.<br />
United States Census Bureau. 2009. American Factfinder. Washingt<strong>on</strong>, D.C.: United States Census<br />
Bureau.<br />
United States Census Bureau. 2011. Statistical Abstract of the United States: 2012 (131st Editi<strong>on</strong>).<br />
Washingt<strong>on</strong>, D.C.: 2011; United States Census Bureau.<br />
Yan, M., <str<strong>on</strong>g>and</str<strong>on</strong>g> Napoli, P. 2006. “<str<strong>on</strong>g>Market</str<strong>on</strong>g> Competiti<strong>on</strong>, Stati<strong>on</strong> Ownership, <str<strong>on</strong>g>and</str<strong>on</strong>g> Local Public Affairs<br />
<str<strong>on</strong>g>Program</str<strong>on</strong>g>ming <strong>on</strong> Broadcast Televisi<strong>on</strong>,” Journal of Communicati<strong>on</strong>s, 56(1), 795–812.<br />
Zavoina, R. <str<strong>on</strong>g>and</str<strong>on</strong>g> McKelvey, W. . 1975. “A Statistical Model for the Analysis of Ordinal Level<br />
Dependent Variables.” Journal of Mathematical Sociology, 103–120.<br />
Zwerina, K., Huber, J., <str<strong>on</strong>g>and</str<strong>on</strong>g> Kuhfeld, W. 1996. “A General Method for C<strong>on</strong>structing Efficient<br />
Choice Designs.” In <str<strong>on</strong>g>Market</str<strong>on</strong>g>ing <str<strong>on</strong>g>Research</str<strong>on</strong>g> Methods in the SAS System. Versi<strong>on</strong> 8 editi<strong>on</strong>, SAS<br />
Institute, Cary, North Carolina.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
Acknowledgements<br />
The Federal Communicati<strong>on</strong>s Commissi<strong>on</strong> (FCC) <str<strong>on</strong>g>and</str<strong>on</strong>g> the Time Warner Cable (TWC) <str<strong>on</strong>g>Research</str<strong>on</strong>g><br />
<str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong> Communicati<strong>on</strong>s provided funding for this research. We are grateful to<br />
Jessica Alm<strong>on</strong>d, Fern<str<strong>on</strong>g>and</str<strong>on</strong>g>o Laguarda, J<strong>on</strong>athan Levy <str<strong>on</strong>g>and</str<strong>on</strong>g> Tracy Wald<strong>on</strong> for their assistance during<br />
the completi<strong>on</strong> of this project. Y<strong>on</strong>gmin Chen, Nicholas Flores, Jin-Hyuk Kim, David Layt<strong>on</strong>,<br />
Edward Morey, Gregory Rosst<strong>on</strong>, Bradley Wimmer <str<strong>on</strong>g>and</str<strong>on</strong>g> seminar participants at the University<br />
of California at Merced, University of Colorado at Boulder, SUNY St<strong>on</strong>y Brook, <str<strong>on</strong>g>and</str<strong>on</strong>g> attendees at<br />
the C<strong>on</strong>ference in H<strong>on</strong>or of Professor Emeritus Lester D. Taylor at Jacks<strong>on</strong> Hole provided helpful<br />
comments <str<strong>on</strong>g>and</str<strong>on</strong>g> c<strong>on</strong>tributi<strong>on</strong>s.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 33
34<br />
About the Authors<br />
Scott Savage is an Associate Professor of Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Interdisciplinary<br />
Telecommunicati<strong>on</strong>s at the University of Colorado at Boulder. He teaches<br />
industrial organizati<strong>on</strong>, microec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> telecom ec<strong>on</strong>omics. His<br />
research interests include empirical industrial organizati<strong>on</strong>, telecom<br />
ec<strong>on</strong>omics, <str<strong>on</strong>g>and</str<strong>on</strong>g> applied ec<strong>on</strong>ometrics. Currently he is studying wholesale<br />
pricing in partially deregulated markets, market entry <str<strong>on</strong>g>and</str<strong>on</strong>g> competiti<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
c<strong>on</strong>sumer preferences for broadb<str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> media services.<br />
Educated in Australia, Savage earned a PhD from Curtin University of<br />
Technology, where he also did his undergraduate work. He matriculated at<br />
the University of Western Australia, where he received his Master’s degree.<br />
Professor Savage is a member of the American Ec<strong>on</strong>omic Associati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the Ec<strong>on</strong>omic Society<br />
of Australia, a <str<strong>on</strong>g>Research</str<strong>on</strong>g> Affiliate for Silic<strong>on</strong> Flatir<strong>on</strong>s, <str<strong>on</strong>g>and</str<strong>on</strong>g> is <strong>on</strong> the Editorial Board of Informati<strong>on</strong><br />
Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Policy. He has extensive experience in microec<strong>on</strong>omic modeling of behavioral<br />
choice, empirical analysis of the market outcomes from competiti<strong>on</strong>, <str<strong>on</strong>g>and</str<strong>on</strong>g> knowledge of communicati<strong>on</strong>s<br />
issues <str<strong>on</strong>g>and</str<strong>on</strong>g> instituti<strong>on</strong>s.<br />
Savage has used choice experiments to estimate dem<str<strong>on</strong>g>and</str<strong>on</strong>g> for various media <str<strong>on</strong>g>and</str<strong>on</strong>g> telecommunicati<strong>on</strong>s<br />
services for Optus, the Western Australian Department of Commerce <str<strong>on</strong>g>and</str<strong>on</strong>g> Trade, Cisco<br />
Systems, Knowledge Networks, the Nati<strong>on</strong>al Broadb<str<strong>on</strong>g>and</str<strong>on</strong>g> Plan <str<strong>on</strong>g>and</str<strong>on</strong>g> the Federal Communicati<strong>on</strong>s<br />
Commissi<strong>on</strong> (FCC). He was employed as a special c<strong>on</strong>sultant to the <str<strong>on</strong>g>Media</str<strong>on</strong>g> Bureau of the FCC<br />
during 2012.<br />
Professor Savage has published in numerous peer-reviewed scholarly journals such as The Review<br />
of Ec<strong>on</strong>omics <str<strong>on</strong>g>and</str<strong>on</strong>g> Statistics, Journal of Applied Ec<strong>on</strong>ometrics, Internati<strong>on</strong>al Journal of Industrial<br />
Organizati<strong>on</strong>, Journal of Law <str<strong>on</strong>g>and</str<strong>on</strong>g> Ec<strong>on</strong>omics, The Ec<strong>on</strong>omic Record, <str<strong>on</strong>g>and</str<strong>on</strong>g> Telecommunicati<strong>on</strong>s Policy.<br />
D<strong>on</strong> Waldman holds a BA from Cornell University, Ithaca, New York<br />
(1969), <str<strong>on</strong>g>and</str<strong>on</strong>g> a Masters <str<strong>on</strong>g>and</str<strong>on</strong>g> Ph.D. in ec<strong>on</strong>omics from the University of<br />
Wisc<strong>on</strong>sin, Madis<strong>on</strong>, Wisc<strong>on</strong>sin (1979). He is Professor of Ec<strong>on</strong>omics<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Associate Chair, Department of Ec<strong>on</strong>omics, University of Colorado,<br />
Boulder, CO, since 1984. He has held faculty positi<strong>on</strong>s at the University<br />
of North Carolina-Chapel Hill, <str<strong>on</strong>g>and</str<strong>on</strong>g> the University of Sydney, Sydney,<br />
Australia, <str<strong>on</strong>g>and</str<strong>on</strong>g> has been a research ec<strong>on</strong>omist for the Department of Health<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Human Services in Washingt<strong>on</strong>, D.C. <str<strong>on</strong>g>and</str<strong>on</strong>g> the RAND Corporati<strong>on</strong> in<br />
Santa M<strong>on</strong>ica, California. His research interests are in all areas of applied<br />
ec<strong>on</strong>ometrics, <str<strong>on</strong>g>and</str<strong>on</strong>g> he teaches statistics <str<strong>on</strong>g>and</str<strong>on</strong>g> ec<strong>on</strong>ometrics at the undergraduate<br />
<str<strong>on</strong>g>and</str<strong>on</strong>g> Ph.D. levels.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
He has presented his research at the American Ec<strong>on</strong>omic Associati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> the Ec<strong>on</strong>ometric Society<br />
annual meetings, <str<strong>on</strong>g>and</str<strong>on</strong>g> at the World C<strong>on</strong>gress of the Ec<strong>on</strong>ometric Society; participated in the Sixth<br />
Invited Choice Symposium, Estes Park, CO., <str<strong>on</strong>g>and</str<strong>on</strong>g> taught the ec<strong>on</strong>ometrics of discrete choice analysis<br />
at the Workshop for <str<strong>on</strong>g>Research</str<strong>on</strong>g> Training in Ec<strong>on</strong>omic Educati<strong>on</strong>, Princet<strong>on</strong> University Visiting<br />
<str<strong>on</strong>g>Research</str<strong>on</strong>g> Fellows <str<strong>on</strong>g>Program</str<strong>on</strong>g>, Princet<strong>on</strong>, N. J.<br />
He has published more than 50 refereed journal articles <str<strong>on</strong>g>and</str<strong>on</strong>g> c<strong>on</strong>tributed papers <strong>on</strong> a variety of<br />
ec<strong>on</strong>ometric topics including discrete choice <str<strong>on</strong>g>and</str<strong>on</strong>g> panel data, <str<strong>on</strong>g>and</str<strong>on</strong>g> <strong>on</strong> ec<strong>on</strong>omic topics including<br />
n<strong>on</strong>market valuati<strong>on</strong>, energy dem<str<strong>on</strong>g>and</str<strong>on</strong>g>, <str<strong>on</strong>g>and</str<strong>on</strong>g> producti<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> cost functi<strong>on</strong>s.<br />
Scott Hiller is a PhD c<str<strong>on</strong>g>and</str<strong>on</strong>g>idate at the University of Colorado at Boulder,<br />
where he is completing his dissertati<strong>on</strong>, Essays <strong>on</strong> the Organizati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g><br />
Operati<strong>on</strong> of <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> Entertainment Industries. His dissertati<strong>on</strong> advisors<br />
include Professors Scott J. Savage <str<strong>on</strong>g>and</str<strong>on</strong>g> D<strong>on</strong>ald M. Waldman, who<br />
also are co-authors with Hiller of <str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>,<br />
published as part of the Time Warner Cable <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong><br />
Communicati<strong>on</strong>s.<br />
Hiller’s research interests include empirical industrial organizati<strong>on</strong>, applied<br />
ec<strong>on</strong>ometrics, <str<strong>on</strong>g>and</str<strong>on</strong>g> music <str<strong>on</strong>g>and</str<strong>on</strong>g> media producti<strong>on</strong>. Currently he is studying<br />
the decisi<strong>on</strong> making of music festivals, offerings of instant video services, potential anticompetitive<br />
effects of exclusive dealing, <str<strong>on</strong>g>and</str<strong>on</strong>g> c<strong>on</strong>sumer valuati<strong>on</strong> for media services.<br />
In March 2012, Hiller presented a paper, Exclusive Dealing <str<strong>on</strong>g>and</str<strong>on</strong>g> its Effects: The Impact of Large<br />
Music Festivals <strong>on</strong> Local Music Venues, at the 10 th Annual Internati<strong>on</strong>al Industrial Organizati<strong>on</strong><br />
C<strong>on</strong>ference in Washingt<strong>on</strong>, D.C.<br />
He completed his M.A. in Ec<strong>on</strong>omics at the University of Colorado at Boulder <str<strong>on</strong>g>and</str<strong>on</strong>g> received a B.S.<br />
in Business Administrati<strong>on</strong>, cum laude, from the University of South Carolina H<strong>on</strong>ors College.<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g> 35
36<br />
Key C<strong>on</strong>tact Informati<strong>on</strong><br />
Scott J. Savage<br />
Associate Professor of Ec<strong>on</strong>omics & Interdisciplinary Telecommunicati<strong>on</strong>s <str<strong>on</strong>g>Program</str<strong>on</strong>g><br />
Department of Ec<strong>on</strong>omics<br />
256 UCB<br />
University of Colorado-Boulder<br />
Boulder, CO, 80309-0256<br />
(303) 735-1165<br />
(303) 492-8960 (fax)<br />
e-mail: scott.savage@colorado.edu<br />
D<strong>on</strong>ald M. Waldman<br />
Professor<br />
Department of Ec<strong>on</strong>omics<br />
256 UCB<br />
University of Colorado–Boulder<br />
Boulder, CO 80309-0256<br />
e-mail: d<strong>on</strong>ald.waldman@colorado.edu<br />
R. Scott Hiller<br />
Department of Ec<strong>on</strong>omics<br />
256 UCB<br />
University of Colorado-Boulder<br />
Boulder, CO 80309-0256<br />
(303) 735-5500<br />
e-mail: r<str<strong>on</strong>g>and</str<strong>on</strong>g>all.hiller@colorado.edu<br />
<str<strong>on</strong>g>Market</str<strong>on</strong>g> <str<strong>on</strong>g>Structure</str<strong>on</strong>g> <str<strong>on</strong>g>and</str<strong>on</strong>g> <str<strong>on</strong>g>Media</str<strong>on</strong>g> <str<strong>on</strong>g>Diversity</str<strong>on</strong>g>
About the Time Warner Cable <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong><br />
<strong>Digital</strong> Communicati<strong>on</strong>s<br />
The Time Warner Cable <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> <strong>on</strong> <strong>Digital</strong> Communicati<strong>on</strong>s will be<br />
dedicated to increasing public underst<str<strong>on</strong>g>and</str<strong>on</strong>g>ing of the benefits <str<strong>on</strong>g>and</str<strong>on</strong>g> challenges facing<br />
the future of digital technologies in the home, office, classroom, <str<strong>on</strong>g>and</str<strong>on</strong>g> community.<br />
The <str<strong>on</strong>g>Research</str<strong>on</strong>g> <str<strong>on</strong>g>Program</str<strong>on</strong>g> will focus <strong>on</strong> the following areas:<br />
• Increasing knowledge about the marketplace <str<strong>on</strong>g>and</str<strong>on</strong>g> the c<strong>on</strong>sumer<br />
• Increasing knowledge about digital technologies<br />
• Increasing knowledge about communicati<strong>on</strong>s policy<br />
• Increasing knowledge about innovati<strong>on</strong> in digital communicati<strong>on</strong>s<br />
About the <str<strong>on</strong>g>Research</str<strong>on</strong>g> Stipends<br />
Individuals receiving a stipend should produce a 25– to 35–page report. The report<br />
should be submitted no later than six m<strong>on</strong>ths after the start of the project.<br />
Proposals from any discipline with research interest in digital communicati<strong>on</strong>s will be<br />
c<strong>on</strong>sidered. Multidisciplinary research teams, c<strong>on</strong>sisting of two or more authors from<br />
different fields, are encouraged.<br />
Size of Stipend: $20,000<br />
Applicati<strong>on</strong> Deadlines for 2012/2013 Awards: November 1, 2012 <str<strong>on</strong>g>and</str<strong>on</strong>g> April 1, 2013<br />
Submitting Applicati<strong>on</strong>s: Applicati<strong>on</strong>s should be submitted <strong>on</strong>line at<br />
www.twcresearchprogram.com. Applicants should submit:<br />
• A three-page descripti<strong>on</strong> of the proposed project<br />
• A resumé (no more than three pages per author)<br />
Applicants will be notified when their applicati<strong>on</strong> is received <str<strong>on</strong>g>and</str<strong>on</strong>g> when the proposal<br />
review process is completed.<br />
About Time Warner Cable<br />
Time Warner Cable Inc. (NYSE: TWC) is am<strong>on</strong>g the largest providers of video, high-speed<br />
data <str<strong>on</strong>g>and</str<strong>on</strong>g> voice services in the United States, c<strong>on</strong>necting more than 15 milli<strong>on</strong> customers<br />
to entertainment, informati<strong>on</strong> <str<strong>on</strong>g>and</str<strong>on</strong>g> each other. Time Warner Cable Business Class offers<br />
data, video <str<strong>on</strong>g>and</str<strong>on</strong>g> voice services to businesses of all sizes, cell tower backhaul services to<br />
wireless carriers <str<strong>on</strong>g>and</str<strong>on</strong>g>, through its NaviSite subsidiary, managed <str<strong>on</strong>g>and</str<strong>on</strong>g> outsourced informati<strong>on</strong><br />
technology soluti<strong>on</strong>s <str<strong>on</strong>g>and</str<strong>on</strong>g> cloud services. Time Warner Cable <str<strong>on</strong>g>Media</str<strong>on</strong>g>, the advertising<br />
arm of Time Warner Cable, offers nati<strong>on</strong>al, regi<strong>on</strong>al <str<strong>on</strong>g>and</str<strong>on</strong>g> local companies innovative advertising<br />
soluti<strong>on</strong>s. More informati<strong>on</strong> about the services of Time Warner Cable is available at<br />
www.timewarnercable.com, www.twcbc.com <str<strong>on</strong>g>and</str<strong>on</strong>g> www.twcmediasales.com.
4<br />
901 F Street, NW<br />
Suite 800<br />
Washingt<strong>on</strong>, DC 20004<br />
www.twcresearchprogram.com