2016 Economics of Media and Communications Conference

December 1–2, 2016

(All day)

University of Chicago
Organizers
Matthew Gentzkow, Stanford University
Emir Kamenica, University of Chicago Booth School of Business
Jesse M. Shapiro, Brown University

The nature of media production and consumption is changing rapidly. And the ways we consume and share information leaves in its wake a wealth of data that can help us better understand myriad political, economic, legal, and cultural aspects of our society.

At a recent conference hosted by the Becker Friedman Institute, the third on this topic organized by Matthew Gentzkow of Stanford University, Emir Kamenica of University of Chicago Booth School of Business, and Jesse M. Shapiro of Brown University, researchers presented their work on several of these areas.

Throughout all of the presentations, the researchers emphasized the ways in which technology and the Internet have drastically changed the consumption patterns of media consumers. While these changes are not particularly surprising, the nuanced ways in which these changes in the media market have affected other aspects of society suggest avenues for interesting lines of research.

One topical question explored at the conference was how media bias affects demand for news and competition in news markets. Much research has explored the role of political bias in mass media in democracies, but this research has not extended to authoritarian regimes. Two papers presented at the conference begin to fill that void, examining media production and consumption in China and Russia.

David Strömberg presented work with Bei Qin and Yanhui Wuo studied the interplay between the dual goals of media outlets—economic and political. After analyzing different classifications of newspapers with varying levels of bias, the researchers find that where the media is entirely government-run at varying administrative levels, both competition and lower-level government control limit the amount of political bias. Empirical research confirmed the predictions of their model; they found that “lower-level governments produce less-biased newspapers and launch commercial newspapers earlier.”

What makes this study novel is its incorporation of the vertical element of government control, as well as its focus on timing of product entry to the market. This approach is pivotal for understanding media bias particularly in autocratic environments, and also useful in other contexts. Strömberg pointed out that the bias measure developed in this paper predicts political control in other markets, whether the firms have dual politico-economic goals, or singular profit-driven goals.

On a similar line, Andrey Simonov and Justin Rao investigated the demand for biased and unbiased news in a context where both government-controlled, government-influenced, and independent news sources are readily available. Leveraging the availability of browser data, the researchers were able to analyze traffic at 48 news outlets to evaluate media consumer preferences and determine whether and when consumers prefer biased news.

They found that in general, consumers prefer news with less pro-government bias, but there was considerable variation in this finding. This is particularly critical as, as the authors note, 68 percent of countries have been press classified as “not free” or “partly free” by the 2015 Freedom of the Press report.

What makes consumers seek out biased news? And if they prefer unbiased, independent news, how effective is government control through censorship and propaganda? The researchers assessed that by identifying news topics that were sensitive for the government, and studying variations in the amount of sensitive news over time. They found that consumers do prefer independent news outlets for reporting on politically sensitive topics.

This exacts a cost in market share; the researchers calculate that audiences would be higher if government-controlled and –influenced outlets covered news the way independent outlets did. The profit loss in foregone advertising revenue, however, appears to be minimal.

So, what drives some consumers to prefer biased information? The researchers conclude that what consumers prefer isn’t necessarily the bias in and of itself, but rather other aspects of the bias, such as the emotional language used. The authors note that this finding has possible impacts for other areas, peripheral to media, such as advertising and product reviews.

Another paper addressed an issue of concern to many in the wake of November’s elections: the effect of media in democracies on the electorate and election results. This is an area of research that has been explored with particular scrutiny since the advent of television advertisement. In their paper Political Advertising and Election Outcomes, Jörg Spenkuch and David Toniatti observe  that a great deal of empirical evidence—including work done by University of Chicago’s Scott Ashwoth—suggests that “political advertising is ineffective at engaging the electorate.”

Using a nuanced approach, exploiting effectively random variation between neighboring counties, Spenkuch and Toniatti find three main conclusions: (1) the amount of political advertising has practically no effect on voter turnout; (2) advertising does exert an important impact on candidates’ vote shares; (3) partisan differences in election turnout is effected by partisan differences in advertising. As the authors point out, this last point is critical for political scientists and public policy practitioners alike. Not only would eliminating partisan differences in advertising in 2008 have substantially narrowed the margin between Barack Obama and John McCain, it would most likely have changed the outcome of the 2000 election between George W. Bush and Al Gore.

Aside from political concerns, other papers focused on the production and sharing of online content. In a digital online world where simple cut-and-paste sharing amplifies information flow, how do content producers protect their economic investment in generating news? Julia Cagé presented work with Nicolas Hervé and Marie-Luce Viaud that explores these issues.

With the rise of extremely rapid reporting and news breaking, it becomes possible to track the timeline of a story as it develops online across media outlets. Cagé reported that this phenomenon results in a huge amount of verbatim copying without attribution. The researchers find that only 36% of online content is original. With copying so prevalent, scooping the competition and being the first to break a story ultimately has little effect on the number of online views.

These results are disturbing from the point of view of copyright law, but, as the authors note, the implications are difficult to parse. Applying copyright laws to online news production is extremely difficult, if not impossible. Moreover, government control of media production has mixed effects, as discussed in the previously mentioned papers. Ultimately, future research is necessary to really understand the course of online news media.

In another take on copyright issues, Petra Moser looked at the impact of copyright on science. Other researchers also shared work on communications in non-digital realms—on juries and within firms. Kelly Shue showed how consistency reinforces good news and mitigates bad news within firms.

Melissa Kearney assessed the effectiveness of Sesame Street as an educational tool, exploiting differences in broadcast reception to assess whether children who could easily view the show did better in school. She reported a positive impact.

All of these areas exist in a precarious and shifting environment in which further research is critical. One of the most remarkable things about the research presented at this conference is its applicability in a multitude of diverse areas that are not immediately related to media production and consumption. The lessons we learn from the study of media and communications have far-reaching impacts on many critical areas of modern society.

—   Amelia Snoblin

December 1, 2016 (All day) December 2, 2016 (All day)