Media Ownership Study 2-Peer Review
July 15, 2011
Jonathan Levy, Deputy Chief Economist, Federal Communications
David F. Layton, Professor, Daniel J. Evans School of Public Affairs,
University of Washington.
Review of Consumer Valuation of Media As A Function of Local Market
Structure, by Professors Scott Savage and Donald Waldman of the
University of Colorado
Per your request I have reviewed “Consumer Valuation of Media As A Function of Local
Market Structure”, by Professors Scott Savage and Donald Waldman of the University of
Colorado. As requested, in conducting my review I have focused on “whether: (1) the
methodology and assumptions are reasonable and technically correct; (2) whether the
methodology and assumptions are consistent with accepted theory and empirical (e.g.,
econometric) practices; (3) whether the data used are reasonable and of sufficient quality
for purposes of analysis; and (4) whether the conclusion, if any, follow from the
analysis.” Further as requested I do not offer any advice on policy or evaluate the policy
implications of the study. Before proceeding, I state that I do not believe that I have any
conflict of interest in reviewing this study.
“Consumer Valuation of Media As A Function of Local Market Structure” collects and
analyzes original data to estimate models that provide willingness to pay (WTP)
estimates of the value that consumers place on a number of features of their local media
environment. WTP is a theoretically driven and justified approach that economists use to
measure value. The study focuses on the “Diversity of Opinion”, the local “Community
News” component, amount or quality of the media content focused on “Multiculturalism”
(related to ethnic, gender, or minority issues), and the amount of “Advertising”.
The data collection approach is survey based. It uses an established methodology that
has come to be called a “Stated Preference” or “Choice Experiment” approach. This
approach has been frequently used (and published in peer-review literature) in a number
of fields of economics including (but not limited to) environmental and health economics
as well as in market research in addition to the media/telecommunications context here.
In reviewing the study, I find that they used good and accepted professional practices in
designing their survey. Alternatives are constructed based on the local media factors
above plus costs. They take appropriate care in eliciting and designing the variables.
They used an established choice experiment regime – choose an alternative media
package “A” versus alternative media package “B” , and then based on that choose their
current real media package or their preferred of “A” or “B”. In designing their survey,
they employed good and established pre-testing procedures. Their experimental design is
well considered and is based on a well known and peer-reviewed published approach.
The survey administration and sampling is conducted by a well known and established
firm – Knowledge Networks. The sample size is quite large compared to most studies
that use these techniques. In summary, I believe that the data used are reasonable and of
sufficient quality for purposes of analysis.
The econometrics has two basic pieces. The first is a discrete choice model (a form of
bivariate probit model) which ultimately provides the WTP or value associated with
altering a consumer’s local media environment. This model is well explained and
follows from standard theory and practice in the area. The second part is an ordered
probit model (also a form of discrete choice model) which is used to understand and later
to predict under potential mergers how consumers might perceive the new levels of
“Diversity of Opinion”, the local “Community News” component, amount or quality of
the media content focused on “Multiculturalism” (related to ethnic, gender, or minority
issues), and the amount of “Advertising”. In conducting their analysis, the authors
provide alternative models as robustness checks. In summary, I believe that
methodological approach and assumptions are consistent with accepted theory and
In order to draw conclusions regarding how mergers might affect the overall economic
welfare of consumers as measured by WTP, the researchers combine their discrete choice
model and order-probit model to conduct a prediction of what WTP would be if the
number of television stations reduced by one in each market. This component is well-
explained and flows from their econometric analysis.
There are a few place where I believe that further discussion, modeling, or information
might be useful to the FCC. These do not change my overall assessment of the study.
1) It may be helpful to some users of the study to see in its own individual table the
final post-stratification weights as used in the econometric modeling.
2) It may be helpful to note that some users of these types of models may interpret
equation (3) in expectations form and consider all the WTP’s to be “expected or
3) A number of different discrete choice WTP models are estimated and discussed –
a baseline, demographic based models, and the non-linear model in Table 19 used
to provide the estimates in table 22. First, I wonder whether it is useful to the
FCC to have a model like the non-linear one in table 19 reported with
demographics? Second, would it be helpful to the FCC to have the authors
identify which WTP estimates from the various models should be used for which
In summary, I conclude that overall, the methodology and assumptions are appropriate
and consistent with economic and econometric theory and practice. Their estimate of the
economic welfare changes associated with potential mergers follows from their economic
and econometric modeling.
David F. Layton
Note: We are currently transitioning our documents into web compatible formats for easier reading. We have done our best to supply this content to you in a presentable form, but there may be some formatting issues while we improve the technology. The original version of the document is available as a PDF, Word Document, or as plain text.