Emerging Trends in The Social and Behavioral Sciences

Partisan News: A Perspective from Economics

Item

Title

Partisan News: A Perspective from Economics

Author

Stone, Daniel F.

Research Area

Social Institutions

Topic

Mass Communication

Abstract

I briefly summarize the economics literature on ideologically slanted political media (which I call, for short, partisan news), and discuss directions for future research. In the literature review, I take a history of thought approach, describing how theory and empirical work have fed off one another and real‐world events. I also note ways in which the work of economists differs from comparable work from other disciplines. In the discussion of future research, I identify open questions and policy options, and assess the relationship between research from economics and other disciplines.

Related Essays

Understanding American Political Conservatism (Political Science), Joel D. Aberbach

Learning Across the Life Course (Sociology), Jutta Allmendinger and Marcel Helbig

Public Opinion and International Conflict (Political Science), Adam J. Berinsky

Empathy Gaps between Helpers and Help‐Seekers: Implications for Cooperation (Psychology), Vanessa K. Bohns and Francis J. Flynn

Mental Models (Psychology), Ruth M. J. Byrne

Heuristic Decision Making (Political Science), Edward G. Carmines and Nicholas J. D'Amico

Elites (Sociology), Johan S. G. Chu and Mark S. Mizruchi

Neoliberalism (Sociology), Miguel Angel Centeno and Joseph N. Cohen

Misinformation and How to Correct It (Psychology), John Cook et al.

Expertise (Sociology), Gil Eyal

Cognitive Processes Involved in Stereotyping (Psychology), Susan T. Fiske and Cydney H. Dupree

Political Advertising (Political Science), Erika Franklin Fowler

Controlling the Influence of Stereotypes on One's Thoughts (Psychology), Patrick S. Forscher and Patricia G. Devine

Media and the development of Identity (Psychology), Adriana M. Manago

Media Neuroscience (Communications & Media), J. Michael Mangus et al.

News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck and Alina Feinholdt

Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E. Tetlock

Identifier

etrds0419

extracted text

Partisan News: A Perspective from
Economics
DANIEL F. STONE

Abstract
I briefly summarize the economics literature on ideologically slanted political media
(which I call, for short, partisan news), and discuss directions for future research. In
the literature review, I take a history of thought approach, describing how theory and
empirical work have fed off one another and real-world events. I also note ways in
which the work of economists differs from comparable work from other disciplines.
In the discussion of future research, I identify open questions and policy options, and
assess the relationship between research from economics and other disciplines.

INTRODUCTION
The political news media has been studied extensively by social science and
journalism scholars for decades.1 Economists mostly neglected the Fourth
Estate until the early 2000s. Interest in the topic was motivated then by
two main factors: (i) growing public distrust in the media as an institution
and (ii) significant technology-driven changes occurring in the content and
competitive structure of media markets. Interest was strengthened as we
(perhaps belatedly) recognized that the media as an industry has greater
social and political importance than most, maybe all, others. The time was
also right for economics media research due to the recent development of
applicable research methods (computational text analysis, natural and field
experiments, and ideas from behavioral economics), allowing economists to
address questions that beforehand may have seemed outside of our domain.
A good portion of this new economics political media literature focuses
on the specific issue of ideologically slanted political media content, which
I call, for short, partisan news. For most industries, economists believe that
increasing the diversity of goods serves consumers and society well. However, the overall benefits of greater diversity in news viewpoints are not as
1. A classic example is McCombs and Shaw (1972).
Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2016 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

clear. A more competitive marketplace of ideas may make truth more likely
to eventually emerge. On the other hand, a more partisan news landscape
might skew political beliefs and actions. Which of these effects dominates?
How do the effects depend on context? What other more subtle issues should
be considered? How can partisan news be studied empirically in a scientific
manner?
In this essay, I summarize findings from the economics literature on
partisan news and present thoughts on where the literature stands and
directions for future work. In the brief literature review,2 I take a history of
thought approach, discussing how theory and empirics fed off one another
and real-world events, and write to a multidisciplinary audience, noting
specific ways in which economics work differs from that of other fields,
admittedly with a bias toward economists’ contributions. In the discussion
of potential future work, I take a more critical perspective. I suggest that
economics media research make a stronger attempt to contribute to policy
analysis (broadly defined), and that economics research can benefit from
better incorporating past findings, and future reactions, of media scholars
from other disciplines.
THE LITERATURE
THE “LIBERAL MAINSTREAM MEDIA” AND THE SUPPLY–DEMAND FRAMEWORK
Economics research on partisan news was first motivated in part by concerns
about a liberal bias in the “mainstream media” in the United States. Scholars from other disciplines had analyzed bias using “content analysis,” which
involves researcher judgment calls to code the ideological slant of media content. See, for example, D’Alessio and Allen (2000). While this tool is flexible
and continues to be useful, it has limitations: it is time- and labor-intensive
and not precisely replicable. In the 2000s, a number of economics papers
developed more formulaic text analysis methods less subject to these issues.3
Researchers from other fields quickly adopted similar methods; see Groeling
(2013) and Puglisi and Snyder (2016) for recent reviews.
The liberal media claim also raised the theoretical question of how such a
bias could persist in a competitive market. Sutter (2000) provided an early,
informal economic analysis, noting the distinction between forces from the
supply side (characteristics of owners, journalists, or advertisers) and the
2. Gentzkow, Shapiro, and Stone (2014) (GSS) and Puglisi and Snyder (2016) are other recent surveys
of theory and empirical economics work on partisan media, respectively. Shleifer (2015) also discusses the
history of thought on this topic, but has a more limited scope. Strömberg (2015) and other Handbook of
Media Economics chapters review related issues in media economics, and Andina-Díaz (2011), Prat and
Strömberg (2013), and Sobbrio (2013) are earlier surveys that include coverage of partisan news.
3. Groseclose and Milyo (2005) was perhaps the first and has since been particularly highly cited (and
has attracted criticism; see, e.g., Nyhan, 2012).

Partisan News: A Perspective from Economics

3

demand side (consumer preferences). This framework has been useful in
much subsequent literature. A famous example is Gentzkow and Shapiro
(2010), an empirical analysis of “what drives media slant.” They study US
newspapers in 2005 and find that consumer ideology explains much more of
the variation in slant than owner ideology.
CAUSES OF DEMAND-DRIVEN BIAS AND PARTISAN SELECTIVE EXPOSURE
Concerns about aggregate liberal bias seemed alleviated, at least somewhat,
in the later 2000s. This may have been partly due to research results like those
of Gentzkow and Shapiro. However, it was also apparent that media content
in reality was becoming more ideologically diverse. Consumers could essentially choose their media slant(s) among online and cable television news
outlets (in addition to more traditional network television and print news
options). Slant was no longer something imposed by the supply side on a consumer with few options. Researchers observed this trend and were motivated
to pursue deeper analysis of causes and effects of consumer demand for partisan news, and in particular, the preference for like-minded news. Scholars
from other fields call the resulting segregated news consumption “partisan
selective exposure.” Sunstein (2001) sounded an early warning about the
potential emergence and socially harmful effects of “echo chambers,” essentially an extreme version of partisan selective exposure in which consumers
limit themselves to narrowly like-minded viewpoints.
Scholars from other disciplines generally assumed that the preference for
like-minded news is motivated by psychological factors (Hart et al., 2009). In a
nutshell, most people feel good when their politics are confirmed, and find it
off-putting to be challenged. Economists have incorporated such psychological factors in a number of papers (Mullainathan & Shleifer, 2005). However,
perhaps the greater contribution of economists was to also explore the range
of more rational forces that could drive like-minded news demand.
In Gentzkow et al. (2014), we present a simple game theoretic model to capture key elements from this literature.4 The model also serves as the basis for
a formal definition of reporting bias, which we taxonomize into two types.
It is useful to discuss this model here for a few reasons: to give readers who
may be unfamiliar with such modeling a better sense of how it works, to
illustrate how important conclusions follow from model assumptions, and
to provide a vocabulary for later discussion in this essay. The model consists
of the following assumptions:

4. Fang (2014), Oliveros and Várdy (2015), and Perego and Yuksel (2015) are more recent papers that
analyze innovative, richer models of rationally demanded bias.

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

•
•

•

•

•

•

There are two types of players: news consumers and firms.
There is an uncertain “state of the world,” either L or R. If L is true, this
can be thought of as a situation that (truly) favors a leftist political view,
and if R is true, this favors rightists.
Media firms observe private information on the state (a “signal”) but
before doing so, each must announce a “reporting strategy” publicly: a
way of transforming the signal into a news report.
Firms observe their signal, then report news based on their signal and
reporting strategy.
Consumers choose which firm to consume news from, or to not consume
news at all, based on the reporting strategies and a private opportunity
cost, and, if news is consumed, Bayesian update their prior belief (probability) about each state.
Finally, each consumer chooses an action (political, such as a vote, or
private consumption, e.g., whether to buy organic or conventional
food) and receives “payoffs” given her action, the true state, and the
consumer’s utility function (preferences).

GSS’s definition of reporting bias is, in words, that strategy x is biased to the
right of strategy y if consumer beliefs that R is the true state likely increase
when the firm switches from y to x without the consumer being aware of the
switch. The definition of left bias is analogous.
The two types of (each) bias are “distortion bias” and “filtering bias.” The
former occurs when it is possible for a firm to directly report its signal as the
news. Call this the neutral strategy. In this case, no other strategy can increase
the (expected) payoff of a consumer who seeks news only for instrumental
value, that is, to inform her choice of action. Thus, any strategy biased right
or left of the neutral strategy cannot be rationally demanded for instrumental
value.
The second type of bias, filtering bias, occurs when signals are more complex than news reports, and so firms must filter private information into
simpler news. In this case, there is no strategy that maximizes the payoff of
all rational instrumental value-seeking consumers. There may still be a seemingly neutral strategy, but a strategy biased to the right or left of this may
actually be (instrumental) payoff-maximizing for some consumers, and thus
rationally demanded. Bias in this case is not necessarily “payoff-reducing”
for consumers.
To see these ideas more clearly consider a specific example in which there is
just one firm whose signal is more complex than the news it can report. Suppose this signal consists of three “subsignals,” each of which is binary and
supports L or R, but the news report is limited to a single binary statement
(supporting either L or R). A seemingly neutral reporting strategy would

Partisan News: A Perspective from Economics

5

be for the firm to report news supporting L if the majority (two or three)
of its subsignals support L, and report news supporting R if the majority of
subsignals support R. A strategy in which “R” was reported only if all three
subsignals support R would be “biased left” of such a neutral strategy. However, this left-biased strategy could increase the payoff of some consumers,
namely, if they have a strong left-leaning utility function. Suppose the action
is “vote L” or “vote R” (and suppose also that the vote is meaningful). If a
consumer would only rationally switch from (vote) L to R if the evidence
supporting this switch was as strong as possible (three R subsignals), then
this consumer would prefer the left-biased strategy to the neutral strategy.
Thus, rational choice can explain partisan selective exposure (leftists seeking left-biased news; the logic applies to rightists symmetrically). Figure 1
depicts the game tree for each of these two reporting strategies. Chiang and
Knight (2011) supports the rationality of consumer interpretation of partisan news.
Groeling (2013) and Puglisi and Snyder (2016) discuss other views on
defining and categorizing types of bias. Groeling splits bias into two types,
“selection” (bias in stories to report on) and “presentation” (bias in what
information to present for a particular story). Groeling refers to both of these
types of bias as involving distortion, which seems to imply that they cannot
be rationally demanded, although GSS’s filtering bias potentially applies to
both of Groeling’s types of bias.
The GSS model implies that rational demand for instrumental value cannot
explain distortion bias. GSS discuss how distortion bias can be caused by two
other demand-side factors: (i) noninstrumental preferences for like-minded
news (the “feels good” motive) and (ii) firm reputational incentives (firms
may “look good” by distorting news toward what consumers believe is likely
true). GSS also discuss how if bias is caused by such reputational motives,
bias is relatively likely to be reduced by competition, but demand-side biases
(distortion or filtering) driven by other causes tend to become more diverse
via competition.
Economists have also contributed to empirical research on consumer preferences for partisan news. Across the social sciences this research has moved
away from relying on surveys, and toward data that more directly captures
behavior (Prior, 2013). Economics research has been consistent with this trend
(Halberstam & Knight, 2014) and has proposed new ways of measuring and
interpreting the extent of partisan selective exposure in different contexts.
However, the economics literature on this specific topic is limited and the
large majority of researchers working in this area come from other fields.

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

The “neutral”
reporting strategy
Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Nature
draws
state

Firm reports
pro-L news
Firm sees 2
pro-L
subsignals

The “left-biased”
reporting strategy

State = L

State = R

Firm reports
pro-R news
Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals
Firm reports
pro-L news
Firm sees 3
pro-L
subsignals

Firm sees 2
pro-L
subsignals

State = L

Nature
draws
state

State = R

Firm sees 3
pro-R
subsignals
Firm reports
pro-R news
Firm sees 3
pro-R
subsignals

Firm sees 2
pro-R
subsignals

Figure 1 Partial game trees, for example, described in text in which firm
observes three binary subsignals and makes one binary report. All events within a
dashed box (an information set) are observationally equivalent to consumers (i.e.,
each event within a dashed box results in the same news report, which is the only
new information a consumer can observe before choosing her action). Note the
state and subsignals are random variables and occur with probabilities that the
game trees omit (probabilities for events that may arise from a given box (node)
must sum to one).

Partisan News: A Perspective from Economics

7

MEDIA EFFECTS: THEORY AND EMPIRICS
The effects of various forms of reporting strategies on media consumer beliefs
and actions are what ultimately drive “social welfare effects”—the implications for societal well-being. Sometimes, models measure these effects just in
terms of consumer and/or producer surplus; often models include political
effects (given voter preferences, whether voters elect the optimal candidate
and/or politicians take optimal actions, allowing for formalization of “political failure”). In general, theory models imply that supply-driven bias is likely
harmful to consumers and society overall, while the effects of demand-side
bias are more ambiguous and depend on various contextual factors. Positive
effects may be due to rationally demanded filtering bias, cross-checking, and
increased engagement of more partisan consumers. Cross-checking and competition of advocates are likely more beneficial when advocates have stronger
incentives to search for information than neutral media. Advocate competition is less socially beneficial when “truth” is more ambiguous and difficult
to verify (e.g., forecasts for climate in 50 years vs tomorrow’s weather).
Negative effects (of demand-driven bias) may be due to consumers choosing socially suboptimal news due to psychological biases, or preferences
for noninstrumental value from news (entertainment or ego confirmation).
Choices resulting from such preferences may be individually rational but
socially harmful as they neglect positive externalities from more informative
news. The engagement and biased processing effects may even go hand in
hand, as more biased consumers may be more engaged by biased media.
Endogenizing actions by politicians likely exacerbates effects [good ones
from more informative media, e.g., Chan and Suen (2009), and bad effects
otherwise, as in Stone (2013)].
To go into more depth on how modeling assumptions affect welfare implications, I refer again to the results from GSS. We show that the private welfare
effects of demand-driven filtering bias are positive if the cause of bias is a partisan utility function. However, welfare effects are ambiguous if demand for
bias is caused by “partisan beliefs”—prior beliefs biased toward one state or
the other. Intuitively, ideal reporting is truly subjective in the first case. People with different “values” can be better off with different ways of filtering
facts into news. In the second case, as ultimately people share the same goals
but disagree on how to reach them, the ideal reporting strategy is not subjective. If preferences for reporting differ, at least some of those preferences
must be (objectively) off base. Noneconomics literature often refers to consumer
demand for biased news as being driven by ideological “attitudes.” It can
be unclear to what extent attitudes include inherently subjective preferences
(e.g., preferences for apples or oranges) versus beliefs about objective facts
(e.g., whether apples or oranges most improve one’s health).

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

A final contribution of economics is a focus on credible identification of
real-world media effects. Much of the prior research depended on lab or
survey experiments or considered empirical evidence that, while still valuable, has less well identified causality.5 Most economics empirical research
uses random, or quasi-random, variation in media exposure in the “field”
to identify causal outcomes. Many papers from other disciplines use these
methods as well (Arceneaux, Johnson, Lindstädt, & Wielen, 2016) and perhaps even more have used field experiments. DellaVigna and Kaplan (2007)
is a key paper from economics on media effects, which indeed reports substantial voting effects caused by the entry of a partisan outlet. Subsequent
papers have found results confirming these conclusions and developed them
further.
WHERE WE STAND
Summing up very briefly, economists have contributed new empirical methods for analysis of partisan news, and new theoretical frameworks for analyzing the causes and consequences of partisan bias. In particular, we have
fleshed out the range of rational forces that may drive demand for bias and
the mechanisms by which bias may influence social welfare, and contributed
to precision in measurement of media bias and its effects, helping confirm
that partisan news effects are real and substantial. Still, it seems fair to say
that in economics and beyond we remain unclear on the really big questions:
to what extent is partisan media good or bad for society? What can be done
about socially harmful aspects of partisan news?
PARTISAN MEDIA EFFECTS: POLITICS AND POLICY
Broadly speaking, if “more partisan” media offered more benefits than costs,
we would expect political systems to function better as media outlets grew
more partisan. Furthermore, broadly speaking, this does not appear to be the
case, at least in the United States: most political scientists would agree the
US political system has not fared well in the new era of more partisan media.
It is well established that political elites have become more polarized, and
political gridlock has increased, in the United States in recent decades (Barber
& McCarty, 2015). This suggests that the growth in partisanship of media
that has occurred in roughly the same period has, overall, done more harm
than good. Bandyopadhyay, Chatterjee, and Roy (2015) discuss examples of
similar associations between media and polarized politics from other parts
of the world.
5. See, for example, Socolow (2008) for a discussion of the challenges in measuring media effects (and
a defense of the strategic sophistication of media consumers).

Partisan News: A Perspective from Economics

9

The most straightforward way that partisan media may have caused the
observed polarization of elites is via partisan selective exposure: if voters
tended toward like-minded media, and media became more ideological, then
this may have ideologically polarized the electorate, causing strategic or sincere polarization of elected officials. Prior (2013) reviews research related
to this idea and finds that the evidence supporting it is thin. He concludes
that although partisan media may have caused already partisan citizens to
become more extreme, most Americans continue to be mostly moderate or
indifferent toward politics. Still, there are papers offering direct evidence
supporting the link between media and general polarization, including Campante and Hojman (2013), Levendusky (2013), Martin and Yurukoglu (2014),
and Stroud (2010).
There are several other mechanisms by which partisan media may have
contributed to the negative political trends. First, even if just the tails of the
citizen ideology distribution were polarized, this could still have had important political effects. Party activists could exist mainly in these tails and be
the ones who apply the key pressure on politicians causing them to take more
extreme positions. Arceneaux et al. (2016) and Clinton and Enamorado (2014)
show that partisan media has indeed caused elected officials to act in a more
polarized way.6 The economics literature has not analyzed the behavior of
party activists as a distinct and important group within the electorate; this
may be a fruitful area for future work.
Second, even if the masses’ fundamental values were not polarized, beliefs
and information may have become more skewed. The literature could
certainly dig further into information and knowledge effects of new media
(Schroeder & Stone, 2015). Ahler (2014) and Levendusky and Malhotra
(2015) are part of a related literature outside of economics presenting related
evidence, which claims that citizens are misinformed about (overestimate)
growth in ideological differences across the parties.
Third, research on growth in “affective polarization” and “partyism”—that
opposing partisans have grown to dislike each other more over time—has
found much stronger empirical support (Garrett et al., 2014; Iyengar, Sood,
& Lelkes, 2012; Westwood et al., 2015). Lelkes, Iyengar, and Sood (2015) use
quasi-random variation in broadband access to show that online media has
played a causal role in these trends.
The economics literature (on media or otherwise) typically does not consider the concept of “dislike.” While a good portion of economics work on
partisan news incorporates behavioral factors, any disagreement occurring
in these models seems, well, agreeable. In the variations covered in GSS,
news consumers may disagree before getting news due to heterogeneous
6. Bandyopadhyay et al. (2015) and Stone (2013) make related theoretical points, showing how media
can cause polarized political actions or gridlock without mass polarization.

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

priors, or disagree after getting news with different slants. However, in neither case would the model imply that voters harbor hard feelings against
one another. A natural new direction for the economics literature, empirical
and theory, would be to incorporate a broader range of “behavioral” factors,
such as affective polarization. The economics theory papers that avoid any
behavioral or noninstrumental assumptions often seem to take this modeling approach as a point of pride. As it is harder to explain the existence of
slanted news without consumers having a direct preference for slant, a model
that avoids such assumptions is more impressive, in a sense. It is important
to understand what phenomena can and cannot be explained with different
sorts of models. However, of course, we also want our models to be consistent
with basic aspects of reality.
To be clear, partisan new media has certainly also had positive welfare
effects. Some types of watchdog journalism have improved; examples
include the Drudge Report exposing the forgery of documents criticizing
George W. Bush’s National Guard service, and “citizen journalist” videos of
police violence. Partisan media may do a better job of avoiding distortion in
news to appear neutral (Burke, 2009) or uninformative “he-said-she-said”
reporting (Chan & Suen, 2009). Many observers point out that US journalism’s embrace of the “ideal of objectivity” in the late twentieth century was
an exception, not the rule (with respect to both media around the world, and
historical news media in the United States). A careful analysis comparing
costs and benefits of partisan news would certainly be a worthwhile, though
challenging avenue for future work.
Even if the net effects of a move away from the ideal of objectivity are
ambiguous, it is clear that unfettered media markets do not function perfectly
and there is a role for public policy. So far, we (in economics and beyond)
have not cracked the policy nut. A simple and flexible policy approach with
relatively broad appeal, like, for example, libertarian paternalism or Pigouvian taxation, might be ideal. This goal is likely too ambitious, but still worth
keeping in mind. Preservation of competition is the closest thing historically
to such a general policy principle. The literature supports the importance of
competition for avoiding negative impacts from supply-side bias, and this
risk is still relevant (Baum & Zhukov, 2015).7 Even the Newspaper Preservation Act, which allows two newspapers in the same metropolitan area to
collude on pricing to maintain both firms’ viability, requires that the news
operations remain separate and competitive. However, at the national level,
supporting competition is not what we need policy to do. The national media
7. Historically, a key concern was cross-ownership of outlets across platforms (print, television, and
radio), and this is still a concern even with the lower costs of entry online (see, e.g., Cagé, 2015, on the
trend of wealthy individuals buying newspapers), though it is also possible consolidated ownership can
result in more differentiated content (George, 2007).

Partisan News: A Perspective from Economics

11

market is plenty competitive, and the political problems seem to exist despite,
or perhaps because, of this.
A number of other policy options have been considered. The Fairness Doctrine is still regularly discussed. It is almost surely politically infeasible and
its value is questionable regardless (Hazlett & Sosa, 1997). Public funding
of media is another commonly discussed policy, with mixed real-world success. An obvious concern is government capture or the perception thereof.
Another is lack of incentives to maximize audience size and social benefits.
A more radical option sometimes mentioned is Canada’s “prohibition
on falsity” in news (On The Media, 2011). This might appear politically
infeasible in the United States given strong protection of political speech;
however, numerous states have laws against false political ads that have
been shown to reduce the incidence of misleading advertising (Zhang, 2015).
Regardless, this type of policy is worthy of study, along with other variations
on policy used across nations. For example, is there a correlation across
nations between freedom of speech protection and polarization in the new
media era?
An approach that is likely even more radical and less feasible would be
to apply Pigouvian taxation to externalities from political information: tax
actions with negative externalities (perhaps consumption of certain types of
media known to report misleadingly partisan news) and subsidize actions
with positive externalities (demonstration of political knowledge or even tolerance?). Perhaps some type of field experiment could be conducted to examine the effects of such policies?
Cagé (2015) makes the case for a creative but reasonable policy proposal:
a new type of organization specific to media, akin to universities with both
nonprofit and private sector elements. This proposal is meant to address the
issue of declining revenue and quality of the press in general, and not partisanship, in particular. However, perhaps there is a link—maybe biased news
is cheaper to produce than neutral news, or journalists prefer to be neutral,
and only resort to slant under competitive pressure? (Cagé argues something
similar is the case for entertainment and soft news.)
Academic research should not be bounded by practical constraints, but
should take them into consideration. Given political constraints, it might
be more productive for the literature to provide recommendations, or even
just potentially practical findings, for conscientious media consumers. For
example, can we tell consumers what effects partisan news consumption has
on stress and mental health? Or on social outcomes, such as diversity and
number of friends, and family relationships, or overall happiness? Can the
literature advise on technologies, or even heuristics, that consumers could
use to improve the neutrality or partisan diversity of news they obtain (for
those who aspire to do this)? How can we enlighten those of us who do not

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

(to help address affective polarization)? Are there techniques informed by
science that we can better use to de-bias others in casual conversation? We
all know the importance of social pressure, but its power can still surprise
us, especially in the absence of financial incentives [e.g., DellaVigna, List,
Malmendier, and Rao (2014), or the emphasis of Kahneman (2011), on the
importance of water cooler gossip]. Can we make consuming one-sided partisan news, or holding biased partisan views, or being intolerant or hostile to
opposing partisans, socially unacceptable—that is, uncool, or just politically
incorrect?
This literature may also be of service for conscientious journalists, or those
who train them. Can we provide more clear guidance on the fundamental
issue of the ideal of objectivity? What other research questions could we
address of practical value to journalists? Can feedback about ideology or
accuracy to journalists affect the quality of their output? Field experiments
and text analysis are likely the best tools for making progress on these and
similar questions.
METHODS AND DISCIPLINES
It will not shock anyone if I report that papers written by economists tend to
cite few (if any) papers from other fields, and those written by noneconomists
tend to neglect our work. This focus on papers from one’s home discipline
is at times reasonable. However, a slant in the way we cite work seems less
appropriate for papers intended to enhance understanding of a question that
crosses disciplinary boundaries.8
Are authors unaware of the range of relevant work from other disciplines?
Or do we cite minimal research from other fields because its value is underestimated? Or because the value is judged correctly to be not that high? I am
not foolish enough to try to answer these questions, but think they are worth
considering for scholars from all fields.
In fact, I think that citation by high-quality surveys from other fields is a
useful goal for economists to strive for in our work. Such citation is prima
facie evidence that the question addressed is important, and that the results
are compelling and clearly communicated. Most economics papers that have
been widely cited outside of economics are empirical and use text analysis,
8. I have already mentioned several examples of findings that seem underappreciated outside their
home discipline; another example is the third-person effect, referring to the tendency to overestimate media
influence. This is a widely used term and concept outside of economics (Davison, 1983), but not mentioned
in any economics paper to my knowledge. Furthermore, while reviewing the proofs for this essay I became
aware of Knobloch-Westerwick (2014), which seems to exemplify this issue. Her book cites numerous
studies of selective exposure—some of which focus on information—that have been neglected by economics, while also largely neglecting the more recent, closely related work by economists.

Partisan News: A Perspective from Economics

13

large data sets, or a clever identification strategy. Theory papers from economics (e.g., the work on rational demand for bias), on the other hand, have
been more ignored outside of our field. What could make our theoretical
results more compelling to noneconomists? Exploring dynamically and/or
socially richer models, which do not necessarily have analytical solutions
(e.g., agent-based models)? Or perhaps the Bayesian framework and theory
results will be appreciated across fields in good time? Or perhaps questions
addressed by economic theory papers are not of as broad interest as those of
empirical papers? Again, I do not offer answers here, just questions.
CONCLUDING REMARKS
Recent developments in the media landscape have been great fodder for
research but perhaps not great for social welfare. Media failure may directly
cause political failure with enormous implications. Economists and scholars
from other disciplines have explored a wide range of theoretical issues and
made progress with empirical analysis in understanding reality. Going forward, I suggest a focus on work that will be of interest across disciplines
and/or be policy relevant, and that we make a better effort to be aware of,
and learn from, work across disciplines. We have yet to “save the media” (as
called for by Cagé). However, we should not give up hope that we can help
to do so.
ACKNOWLEDGMENTS
I thank (especially) the Emerging Trends editors, Bob Scott and Marlis
Buchmann, and also Kevin Arceneaux, Siddhartha Bandyopadhyay, David
D’Alessio, Stefano DellaVigna, Brendan Nyhan, Gaurav Sood, and Michael
Socolow for very helpful comments and suggestions. This essay was written
while I was a visitor at the University of Virginia’s department of economics;
I am grateful for their hospitality.
REFERENCES
Ahler, D. J. (2014). Self-fulfilling misperceptions of public polarization. The Journal of
Politics, 76(03), 607–620.
Andina-Díaz, A. (2011). Mass media in economics: Origins and subsequent contributions. Cuadernos de Ciencias Económicas y Empresariales, 61, 89–101.
Arceneaux, K., Johnson, M., Lindstädt, R., & Wielen, R. J. (2016). The influence of
news media on political elites: Investigating strategic responsiveness in congress.
American Journal of Political Science, 60(1), 5–29.
Bandyopadhyay, S., Chatterjee, K., & Roy, J. (2015). Manufacturing extremism: Political consequences of profit-seeking media (No. 15-14).

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Barber, M., & McCarty, N. (2015). Causes and consequences of polarization. Solutions
to Political Polarization in America, 15, 15–58.
Baum, M. A., & Zhukov, Y.M. (2015). Media ownership and news coverage of international conflict. Working paper.
Burke, J. (2009). Unfairly balanced: Unbiased news coverage and information loss.
Available at SSRN 1020627.
Cagé, J. (2015). Sauver les médias. Capitalisme, financement participatif et démocratie.
Paris, France: Le Seuil (La République des Idées).
Campante, F. R., & Hojman, D. A. (2013). Media and polarization: Evidence from the
introduction of broadcast TV in the United States. Journal of Public Economics, 100,
79–92.
Chan, J., & Suen, W. (2009). Media as watchdogs: The role of news media in electoral
competition. European Economic Review, 53(7), 799–814.
Chiang, C. F., & Knight, B. (2011). Media bias and influence: Evidence from newspaper endorsements. The Review of Economic Studies, 78(3), rdq037.
Clinton, J. D., & Enamorado, T. (2014). The National News Media’s effect on congress:
How Fox News affected elites in congress. The Journal of Politics, 76(04), 928–943.
D’Alessio, D., & Allen, M. (2000). Media bias in presidential elections: A
meta-analysis. Journal of Communication, 50(4), 133–156.
Davison, W. P. (1983). The third-person effect in communication. Public Opinion Quarterly, 47(1), 1–15.
DellaVigna, S., & Kaplan, E. D. (2007). The Fox News effect: Media bias and voting.
The Quarterly Journal of Economics, 122(3), 1187–1234.
DellaVigna, S., List, J. A., Malmendier, U., & Rao, G. (2014). Voting to tell others (No.
w19832). National Bureau of Economic Research.
Fang, R. Y. (2014). Media bias, political polarization, and the merits of fairness.
Garrett, K. R., Gvirsman, S. D., Johnson, B. K., Tsfati, Y., Neo, R., & Dal, A. (2014).
Implications of pro- and counterattitudinal information exposure for affective
polarization. Human Communication Research, 40(3), 309–332.
Gentzkow, M., & Shapiro, J. M. (2010). What drives media slant? Evidence from US
daily newspapers. Econometrica, 78(1), 35–71.
Gentzkow, M., Shapiro, J. M., & Stone, D. F. (2014). Media bias in the marketplace:
Theory (No. w19880). National Bureau of Economic Research.
George, L. (2007). What’s fit to print: The effect of ownership concentration on product variety in daily newspaper markets. Information Economics and Policy, 19(3),
285–303.
Groeling, T. (2013). Media bias by the numbers: Challenges and opportunities in the
empirical study of partisan news. Political Science, 16(1), 129.
Groseclose, T., & Milyo, J. (2005). A measure of media bias. The Quarterly Journal of
Economics, 120(4), 1191–1237.
Halberstam, Y., & Knight, B. (2014). Homophily, group size, and the diffusion of political information in social networks: Evidence from twitter (No. w20681). National
Bureau of Economic Research.

Partisan News: A Perspective from Economics

15

Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009).
Feeling validated versus being correct: A meta-analysis of selective exposure to
information. Psychological Bulletin, 135(4), 555.
Hazlett, T. W., & Sosa, D. W. (1997). Was the fairness doctrine a “chilling effect”? Evidence from the postderegulation radio market. The Journal of Legal Studies, 26(1),
279–301.
Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology a social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431.
Kahneman, D. (2011). Thinking, fast and slow. New York, NY: Macmillan.
Knobloch-Westerwick, S. (2014). Choice and preference in media use: Advances in selective
exposure theory and research. New York, NY: Routledge.
Lelkes, Y., Iyengar, S., & Sood, G. (2015). The hostile audience: Selective exposure
to partisan sources and affective polarization. American Journal of Political Science,
1–16.
Levendusky, M. S. (2013). Why do partisan media polarize viewers? American Journal
of Political Science, 57(3), 611–623.
Levendusky, M., & Malhotra, N. (2015). Does media coverage of partisan polarization affect political attitudes? Political Communication, 1–19 [ahead-of-print].
Martin, G. J., & Yurukoglu, A. (2014). Bias in cable news: Real effects and polarization
(No. w20798). National Bureau of Economic Research.
McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media.
Public Opinion Quarterly, 36(2), 176–187.
Mullainathan, S., & Shleifer, A. (2005). The market for news. American Economic
Review, 1031–1053.
Nyhan, B. (2012). Does the US media have a liberal bias? Perspectives on Politics,
10(03), 767–771.
Oliveros, S., & Várdy, F. (2015). Demand for slant: How abstention shapes voters’
choice of news media. The Economic Journal, 125(587), 1327–1368.
On The Media. (2011). Lying is illegal in Canada(’s news broadcasts). Retrieved from
http://www.onthemedia.org/story/133101-lying-is-illegal-in-canadas-newsbroadcasts/transcript/
Perego, J., & Yuksel, S. (2015). Media competition and the source of disagreement.
Working paper.
Prat, A., & Strömberg, D. (2013). The political economy of mass media. In Advances
in economics and econometrics: Volume 2, applied economics: Tenth world congress (Vol.
50, pp. 135). Cambridge, England: Cambridge University Press.
Prior, M. (2013). Media and political polarization. Annual Review of Political Science,
16, 101–127.
Puglisi, R., & Snyder, J. M. (2016). Empirical studies of media bias. In S. Anderson, J.
Waldfogel & D. Stromberg (Eds.), Handbook of media economics. Elsevier Press.
Schroeder, E., & Stone, D. F. (2015). Fox News and political knowledge. Journal of
Public Economics, 126, 52–63.
Shleifer, A. (2015). Matthew Gentzkow, Winner of the 2014 Clark Medal. The Journal
of Economic Perspectives, 29(1), 181–192.

16

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Sobbrio, F. (2013). The political economy of news media: Theory, evidence and open issues.
Handbook of alternative theories of public economics. Cheltenham, England: Edward
Elgar Press.
Socolow, M. J. (2008). The hyped panic over ’War of the Worlds’. The Chronicle of
Higher Education.
Stone, D. F. (2013). Media and gridlock. Journal of Public Economics, 101, 94–104.
Strömberg, D. (2015). Media and politics. Annual Review of Economics, 7, 173–205.
Stroud, N. J. (2010). Polarization and partisan selective exposure. Journal of Communication, 60(3), 556–576.
Sunstein, C. R. (2001). Republic.com. Princeton, NJ: Princeton University Press.
Sutter, D. (2000). Can the media be so liberal—the economics of media bias. Cato
Journal, 20, 431.
Westwood, S. J., Iyengar, S., Walgrave, S., Leonisio, R., Miller, L., & Strijbis, O. (2015).
The tie that divides: Cross-national evidence of the primacy of partyism. Working
paper.
Zhang, Z. (2015). Swiftboating: Misleading Advertising in Presidential Campaigns.
Unpublished manuscript.

DANIEL F. STONE SHORT BIOGRAPHY
Daniel F. Stone is an assistant professor of economics at Bowdoin College. He
specializes in the study of belief formation and choice under uncertainty, and
political beliefs and choices, in particular. He received his PhD in economics
from Johns Hopkins University in 2008.

RELATED ESSAYS
Understanding American Political Conservatism (Political Science), Joel D.
Aberbach
Learning Across the Life Course (Sociology), Jutta Allmendinger and Marcel
Helbig
Public Opinion and International Conflict (Political Science), Adam J.
Berinsky
Empathy Gaps between Helpers and Help-Seekers: Implications for Cooperation (Psychology), Vanessa K. Bohns and Francis J. Flynn
Mental Models (Psychology), Ruth M. J. Byrne
Heuristic Decision Making (Political Science), Edward G. Carmines and
Nicholas J. D’Amico
Elites (Sociology), Johan S. G. Chu and Mark S. Mizruchi
Neoliberalism (Sociology), Miguel Angel Centeno and Joseph N. Cohen
Misinformation and How to Correct It (Psychology), John Cook et al.
Expertise (Sociology), Gil Eyal

Partisan News: A Perspective from Economics

17

Cognitive Processes Involved in Stereotyping (Psychology), Susan T. Fiske
and Cydney H. Dupree
Political Advertising (Political Science), Erika Franklin Fowler
Controlling the Influence of Stereotypes on One’s Thoughts (Psychology),
Patrick S. Forscher and Patricia G. Devine
Media and the development of Identity (Psychology), Adriana M. Manago
Media Neuroscience (Communications & Media), J. Michael Mangus et al.
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E.
Tetlock



Partisan News: A Perspective from
Economics
DANIEL F. STONE

Abstract
I briefly summarize the economics literature on ideologically slanted political media
(which I call, for short, partisan news), and discuss directions for future research. In
the literature review, I take a history of thought approach, describing how theory and
empirical work have fed off one another and real-world events. I also note ways in
which the work of economists differs from comparable work from other disciplines.
In the discussion of future research, I identify open questions and policy options, and
assess the relationship between research from economics and other disciplines.

INTRODUCTION
The political news media has been studied extensively by social science and
journalism scholars for decades.1 Economists mostly neglected the Fourth
Estate until the early 2000s. Interest in the topic was motivated then by
two main factors: (i) growing public distrust in the media as an institution
and (ii) significant technology-driven changes occurring in the content and
competitive structure of media markets. Interest was strengthened as we
(perhaps belatedly) recognized that the media as an industry has greater
social and political importance than most, maybe all, others. The time was
also right for economics media research due to the recent development of
applicable research methods (computational text analysis, natural and field
experiments, and ideas from behavioral economics), allowing economists to
address questions that beforehand may have seemed outside of our domain.
A good portion of this new economics political media literature focuses
on the specific issue of ideologically slanted political media content, which
I call, for short, partisan news. For most industries, economists believe that
increasing the diversity of goods serves consumers and society well. However, the overall benefits of greater diversity in news viewpoints are not as
1. A classic example is McCombs and Shaw (1972).
Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2016 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

clear. A more competitive marketplace of ideas may make truth more likely
to eventually emerge. On the other hand, a more partisan news landscape
might skew political beliefs and actions. Which of these effects dominates?
How do the effects depend on context? What other more subtle issues should
be considered? How can partisan news be studied empirically in a scientific
manner?
In this essay, I summarize findings from the economics literature on
partisan news and present thoughts on where the literature stands and
directions for future work. In the brief literature review,2 I take a history of
thought approach, discussing how theory and empirics fed off one another
and real-world events, and write to a multidisciplinary audience, noting
specific ways in which economics work differs from that of other fields,
admittedly with a bias toward economists’ contributions. In the discussion
of potential future work, I take a more critical perspective. I suggest that
economics media research make a stronger attempt to contribute to policy
analysis (broadly defined), and that economics research can benefit from
better incorporating past findings, and future reactions, of media scholars
from other disciplines.
THE LITERATURE
THE “LIBERAL MAINSTREAM MEDIA” AND THE SUPPLY–DEMAND FRAMEWORK
Economics research on partisan news was first motivated in part by concerns
about a liberal bias in the “mainstream media” in the United States. Scholars from other disciplines had analyzed bias using “content analysis,” which
involves researcher judgment calls to code the ideological slant of media content. See, for example, D’Alessio and Allen (2000). While this tool is flexible
and continues to be useful, it has limitations: it is time- and labor-intensive
and not precisely replicable. In the 2000s, a number of economics papers
developed more formulaic text analysis methods less subject to these issues.3
Researchers from other fields quickly adopted similar methods; see Groeling
(2013) and Puglisi and Snyder (2016) for recent reviews.
The liberal media claim also raised the theoretical question of how such a
bias could persist in a competitive market. Sutter (2000) provided an early,
informal economic analysis, noting the distinction between forces from the
supply side (characteristics of owners, journalists, or advertisers) and the
2. Gentzkow, Shapiro, and Stone (2014) (GSS) and Puglisi and Snyder (2016) are other recent surveys
of theory and empirical economics work on partisan media, respectively. Shleifer (2015) also discusses the
history of thought on this topic, but has a more limited scope. Strömberg (2015) and other Handbook of
Media Economics chapters review related issues in media economics, and Andina-Díaz (2011), Prat and
Strömberg (2013), and Sobbrio (2013) are earlier surveys that include coverage of partisan news.
3. Groseclose and Milyo (2005) was perhaps the first and has since been particularly highly cited (and
has attracted criticism; see, e.g., Nyhan, 2012).

Partisan News: A Perspective from Economics

3

demand side (consumer preferences). This framework has been useful in
much subsequent literature. A famous example is Gentzkow and Shapiro
(2010), an empirical analysis of “what drives media slant.” They study US
newspapers in 2005 and find that consumer ideology explains much more of
the variation in slant than owner ideology.
CAUSES OF DEMAND-DRIVEN BIAS AND PARTISAN SELECTIVE EXPOSURE
Concerns about aggregate liberal bias seemed alleviated, at least somewhat,
in the later 2000s. This may have been partly due to research results like those
of Gentzkow and Shapiro. However, it was also apparent that media content
in reality was becoming more ideologically diverse. Consumers could essentially choose their media slant(s) among online and cable television news
outlets (in addition to more traditional network television and print news
options). Slant was no longer something imposed by the supply side on a consumer with few options. Researchers observed this trend and were motivated
to pursue deeper analysis of causes and effects of consumer demand for partisan news, and in particular, the preference for like-minded news. Scholars
from other fields call the resulting segregated news consumption “partisan
selective exposure.” Sunstein (2001) sounded an early warning about the
potential emergence and socially harmful effects of “echo chambers,” essentially an extreme version of partisan selective exposure in which consumers
limit themselves to narrowly like-minded viewpoints.
Scholars from other disciplines generally assumed that the preference for
like-minded news is motivated by psychological factors (Hart et al., 2009). In a
nutshell, most people feel good when their politics are confirmed, and find it
off-putting to be challenged. Economists have incorporated such psychological factors in a number of papers (Mullainathan & Shleifer, 2005). However,
perhaps the greater contribution of economists was to also explore the range
of more rational forces that could drive like-minded news demand.
In Gentzkow et al. (2014), we present a simple game theoretic model to capture key elements from this literature.4 The model also serves as the basis for
a formal definition of reporting bias, which we taxonomize into two types.
It is useful to discuss this model here for a few reasons: to give readers who
may be unfamiliar with such modeling a better sense of how it works, to
illustrate how important conclusions follow from model assumptions, and
to provide a vocabulary for later discussion in this essay. The model consists
of the following assumptions:

4. Fang (2014), Oliveros and Várdy (2015), and Perego and Yuksel (2015) are more recent papers that
analyze innovative, richer models of rationally demanded bias.

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

•
•

•

•

•

•

There are two types of players: news consumers and firms.
There is an uncertain “state of the world,” either L or R. If L is true, this
can be thought of as a situation that (truly) favors a leftist political view,
and if R is true, this favors rightists.
Media firms observe private information on the state (a “signal”) but
before doing so, each must announce a “reporting strategy” publicly: a
way of transforming the signal into a news report.
Firms observe their signal, then report news based on their signal and
reporting strategy.
Consumers choose which firm to consume news from, or to not consume
news at all, based on the reporting strategies and a private opportunity
cost, and, if news is consumed, Bayesian update their prior belief (probability) about each state.
Finally, each consumer chooses an action (political, such as a vote, or
private consumption, e.g., whether to buy organic or conventional
food) and receives “payoffs” given her action, the true state, and the
consumer’s utility function (preferences).

GSS’s definition of reporting bias is, in words, that strategy x is biased to the
right of strategy y if consumer beliefs that R is the true state likely increase
when the firm switches from y to x without the consumer being aware of the
switch. The definition of left bias is analogous.
The two types of (each) bias are “distortion bias” and “filtering bias.” The
former occurs when it is possible for a firm to directly report its signal as the
news. Call this the neutral strategy. In this case, no other strategy can increase
the (expected) payoff of a consumer who seeks news only for instrumental
value, that is, to inform her choice of action. Thus, any strategy biased right
or left of the neutral strategy cannot be rationally demanded for instrumental
value.
The second type of bias, filtering bias, occurs when signals are more complex than news reports, and so firms must filter private information into
simpler news. In this case, there is no strategy that maximizes the payoff of
all rational instrumental value-seeking consumers. There may still be a seemingly neutral strategy, but a strategy biased to the right or left of this may
actually be (instrumental) payoff-maximizing for some consumers, and thus
rationally demanded. Bias in this case is not necessarily “payoff-reducing”
for consumers.
To see these ideas more clearly consider a specific example in which there is
just one firm whose signal is more complex than the news it can report. Suppose this signal consists of three “subsignals,” each of which is binary and
supports L or R, but the news report is limited to a single binary statement
(supporting either L or R). A seemingly neutral reporting strategy would

Partisan News: A Perspective from Economics

5

be for the firm to report news supporting L if the majority (two or three)
of its subsignals support L, and report news supporting R if the majority of
subsignals support R. A strategy in which “R” was reported only if all three
subsignals support R would be “biased left” of such a neutral strategy. However, this left-biased strategy could increase the payoff of some consumers,
namely, if they have a strong left-leaning utility function. Suppose the action
is “vote L” or “vote R” (and suppose also that the vote is meaningful). If a
consumer would only rationally switch from (vote) L to R if the evidence
supporting this switch was as strong as possible (three R subsignals), then
this consumer would prefer the left-biased strategy to the neutral strategy.
Thus, rational choice can explain partisan selective exposure (leftists seeking left-biased news; the logic applies to rightists symmetrically). Figure 1
depicts the game tree for each of these two reporting strategies. Chiang and
Knight (2011) supports the rationality of consumer interpretation of partisan news.
Groeling (2013) and Puglisi and Snyder (2016) discuss other views on
defining and categorizing types of bias. Groeling splits bias into two types,
“selection” (bias in stories to report on) and “presentation” (bias in what
information to present for a particular story). Groeling refers to both of these
types of bias as involving distortion, which seems to imply that they cannot
be rationally demanded, although GSS’s filtering bias potentially applies to
both of Groeling’s types of bias.
The GSS model implies that rational demand for instrumental value cannot
explain distortion bias. GSS discuss how distortion bias can be caused by two
other demand-side factors: (i) noninstrumental preferences for like-minded
news (the “feels good” motive) and (ii) firm reputational incentives (firms
may “look good” by distorting news toward what consumers believe is likely
true). GSS also discuss how if bias is caused by such reputational motives,
bias is relatively likely to be reduced by competition, but demand-side biases
(distortion or filtering) driven by other causes tend to become more diverse
via competition.
Economists have also contributed to empirical research on consumer preferences for partisan news. Across the social sciences this research has moved
away from relying on surveys, and toward data that more directly captures
behavior (Prior, 2013). Economics research has been consistent with this trend
(Halberstam & Knight, 2014) and has proposed new ways of measuring and
interpreting the extent of partisan selective exposure in different contexts.
However, the economics literature on this specific topic is limited and the
large majority of researchers working in this area come from other fields.

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

The “neutral”
reporting strategy
Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Nature
draws
state

Firm reports
pro-L news
Firm sees 2
pro-L
subsignals

The “left-biased”
reporting strategy

State = L

State = R

Firm reports
pro-R news
Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals
Firm reports
pro-L news
Firm sees 3
pro-L
subsignals

Firm sees 2
pro-L
subsignals

State = L

Nature
draws
state

State = R

Firm sees 3
pro-R
subsignals
Firm reports
pro-R news
Firm sees 3
pro-R
subsignals

Firm sees 2
pro-R
subsignals

Figure 1 Partial game trees, for example, described in text in which firm
observes three binary subsignals and makes one binary report. All events within a
dashed box (an information set) are observationally equivalent to consumers (i.e.,
each event within a dashed box results in the same news report, which is the only
new information a consumer can observe before choosing her action). Note the
state and subsignals are random variables and occur with probabilities that the
game trees omit (probabilities for events that may arise from a given box (node)
must sum to one).

Partisan News: A Perspective from Economics

7

MEDIA EFFECTS: THEORY AND EMPIRICS
The effects of various forms of reporting strategies on media consumer beliefs
and actions are what ultimately drive “social welfare effects”—the implications for societal well-being. Sometimes, models measure these effects just in
terms of consumer and/or producer surplus; often models include political
effects (given voter preferences, whether voters elect the optimal candidate
and/or politicians take optimal actions, allowing for formalization of “political failure”). In general, theory models imply that supply-driven bias is likely
harmful to consumers and society overall, while the effects of demand-side
bias are more ambiguous and depend on various contextual factors. Positive
effects may be due to rationally demanded filtering bias, cross-checking, and
increased engagement of more partisan consumers. Cross-checking and competition of advocates are likely more beneficial when advocates have stronger
incentives to search for information than neutral media. Advocate competition is less socially beneficial when “truth” is more ambiguous and difficult
to verify (e.g., forecasts for climate in 50 years vs tomorrow’s weather).
Negative effects (of demand-driven bias) may be due to consumers choosing socially suboptimal news due to psychological biases, or preferences
for noninstrumental value from news (entertainment or ego confirmation).
Choices resulting from such preferences may be individually rational but
socially harmful as they neglect positive externalities from more informative
news. The engagement and biased processing effects may even go hand in
hand, as more biased consumers may be more engaged by biased media.
Endogenizing actions by politicians likely exacerbates effects [good ones
from more informative media, e.g., Chan and Suen (2009), and bad effects
otherwise, as in Stone (2013)].
To go into more depth on how modeling assumptions affect welfare implications, I refer again to the results from GSS. We show that the private welfare
effects of demand-driven filtering bias are positive if the cause of bias is a partisan utility function. However, welfare effects are ambiguous if demand for
bias is caused by “partisan beliefs”—prior beliefs biased toward one state or
the other. Intuitively, ideal reporting is truly subjective in the first case. People with different “values” can be better off with different ways of filtering
facts into news. In the second case, as ultimately people share the same goals
but disagree on how to reach them, the ideal reporting strategy is not subjective. If preferences for reporting differ, at least some of those preferences
must be (objectively) off base. Noneconomics literature often refers to consumer
demand for biased news as being driven by ideological “attitudes.” It can
be unclear to what extent attitudes include inherently subjective preferences
(e.g., preferences for apples or oranges) versus beliefs about objective facts
(e.g., whether apples or oranges most improve one’s health).

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

A final contribution of economics is a focus on credible identification of
real-world media effects. Much of the prior research depended on lab or
survey experiments or considered empirical evidence that, while still valuable, has less well identified causality.5 Most economics empirical research
uses random, or quasi-random, variation in media exposure in the “field”
to identify causal outcomes. Many papers from other disciplines use these
methods as well (Arceneaux, Johnson, Lindstädt, & Wielen, 2016) and perhaps even more have used field experiments. DellaVigna and Kaplan (2007)
is a key paper from economics on media effects, which indeed reports substantial voting effects caused by the entry of a partisan outlet. Subsequent
papers have found results confirming these conclusions and developed them
further.
WHERE WE STAND
Summing up very briefly, economists have contributed new empirical methods for analysis of partisan news, and new theoretical frameworks for analyzing the causes and consequences of partisan bias. In particular, we have
fleshed out the range of rational forces that may drive demand for bias and
the mechanisms by which bias may influence social welfare, and contributed
to precision in measurement of media bias and its effects, helping confirm
that partisan news effects are real and substantial. Still, it seems fair to say
that in economics and beyond we remain unclear on the really big questions:
to what extent is partisan media good or bad for society? What can be done
about socially harmful aspects of partisan news?
PARTISAN MEDIA EFFECTS: POLITICS AND POLICY
Broadly speaking, if “more partisan” media offered more benefits than costs,
we would expect political systems to function better as media outlets grew
more partisan. Furthermore, broadly speaking, this does not appear to be the
case, at least in the United States: most political scientists would agree the
US political system has not fared well in the new era of more partisan media.
It is well established that political elites have become more polarized, and
political gridlock has increased, in the United States in recent decades (Barber
& McCarty, 2015). This suggests that the growth in partisanship of media
that has occurred in roughly the same period has, overall, done more harm
than good. Bandyopadhyay, Chatterjee, and Roy (2015) discuss examples of
similar associations between media and polarized politics from other parts
of the world.
5. See, for example, Socolow (2008) for a discussion of the challenges in measuring media effects (and
a defense of the strategic sophistication of media consumers).

Partisan News: A Perspective from Economics

9

The most straightforward way that partisan media may have caused the
observed polarization of elites is via partisan selective exposure: if voters
tended toward like-minded media, and media became more ideological, then
this may have ideologically polarized the electorate, causing strategic or sincere polarization of elected officials. Prior (2013) reviews research related
to this idea and finds that the evidence supporting it is thin. He concludes
that although partisan media may have caused already partisan citizens to
become more extreme, most Americans continue to be mostly moderate or
indifferent toward politics. Still, there are papers offering direct evidence
supporting the link between media and general polarization, including Campante and Hojman (2013), Levendusky (2013), Martin and Yurukoglu (2014),
and Stroud (2010).
There are several other mechanisms by which partisan media may have
contributed to the negative political trends. First, even if just the tails of the
citizen ideology distribution were polarized, this could still have had important political effects. Party activists could exist mainly in these tails and be
the ones who apply the key pressure on politicians causing them to take more
extreme positions. Arceneaux et al. (2016) and Clinton and Enamorado (2014)
show that partisan media has indeed caused elected officials to act in a more
polarized way.6 The economics literature has not analyzed the behavior of
party activists as a distinct and important group within the electorate; this
may be a fruitful area for future work.
Second, even if the masses’ fundamental values were not polarized, beliefs
and information may have become more skewed. The literature could
certainly dig further into information and knowledge effects of new media
(Schroeder & Stone, 2015). Ahler (2014) and Levendusky and Malhotra
(2015) are part of a related literature outside of economics presenting related
evidence, which claims that citizens are misinformed about (overestimate)
growth in ideological differences across the parties.
Third, research on growth in “affective polarization” and “partyism”—that
opposing partisans have grown to dislike each other more over time—has
found much stronger empirical support (Garrett et al., 2014; Iyengar, Sood,
& Lelkes, 2012; Westwood et al., 2015). Lelkes, Iyengar, and Sood (2015) use
quasi-random variation in broadband access to show that online media has
played a causal role in these trends.
The economics literature (on media or otherwise) typically does not consider the concept of “dislike.” While a good portion of economics work on
partisan news incorporates behavioral factors, any disagreement occurring
in these models seems, well, agreeable. In the variations covered in GSS,
news consumers may disagree before getting news due to heterogeneous
6. Bandyopadhyay et al. (2015) and Stone (2013) make related theoretical points, showing how media
can cause polarized political actions or gridlock without mass polarization.

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

priors, or disagree after getting news with different slants. However, in neither case would the model imply that voters harbor hard feelings against
one another. A natural new direction for the economics literature, empirical
and theory, would be to incorporate a broader range of “behavioral” factors,
such as affective polarization. The economics theory papers that avoid any
behavioral or noninstrumental assumptions often seem to take this modeling approach as a point of pride. As it is harder to explain the existence of
slanted news without consumers having a direct preference for slant, a model
that avoids such assumptions is more impressive, in a sense. It is important
to understand what phenomena can and cannot be explained with different
sorts of models. However, of course, we also want our models to be consistent
with basic aspects of reality.
To be clear, partisan new media has certainly also had positive welfare
effects. Some types of watchdog journalism have improved; examples
include the Drudge Report exposing the forgery of documents criticizing
George W. Bush’s National Guard service, and “citizen journalist” videos of
police violence. Partisan media may do a better job of avoiding distortion in
news to appear neutral (Burke, 2009) or uninformative “he-said-she-said”
reporting (Chan & Suen, 2009). Many observers point out that US journalism’s embrace of the “ideal of objectivity” in the late twentieth century was
an exception, not the rule (with respect to both media around the world, and
historical news media in the United States). A careful analysis comparing
costs and benefits of partisan news would certainly be a worthwhile, though
challenging avenue for future work.
Even if the net effects of a move away from the ideal of objectivity are
ambiguous, it is clear that unfettered media markets do not function perfectly
and there is a role for public policy. So far, we (in economics and beyond)
have not cracked the policy nut. A simple and flexible policy approach with
relatively broad appeal, like, for example, libertarian paternalism or Pigouvian taxation, might be ideal. This goal is likely too ambitious, but still worth
keeping in mind. Preservation of competition is the closest thing historically
to such a general policy principle. The literature supports the importance of
competition for avoiding negative impacts from supply-side bias, and this
risk is still relevant (Baum & Zhukov, 2015).7 Even the Newspaper Preservation Act, which allows two newspapers in the same metropolitan area to
collude on pricing to maintain both firms’ viability, requires that the news
operations remain separate and competitive. However, at the national level,
supporting competition is not what we need policy to do. The national media
7. Historically, a key concern was cross-ownership of outlets across platforms (print, television, and
radio), and this is still a concern even with the lower costs of entry online (see, e.g., Cagé, 2015, on the
trend of wealthy individuals buying newspapers), though it is also possible consolidated ownership can
result in more differentiated content (George, 2007).

Partisan News: A Perspective from Economics

11

market is plenty competitive, and the political problems seem to exist despite,
or perhaps because, of this.
A number of other policy options have been considered. The Fairness Doctrine is still regularly discussed. It is almost surely politically infeasible and
its value is questionable regardless (Hazlett & Sosa, 1997). Public funding
of media is another commonly discussed policy, with mixed real-world success. An obvious concern is government capture or the perception thereof.
Another is lack of incentives to maximize audience size and social benefits.
A more radical option sometimes mentioned is Canada’s “prohibition
on falsity” in news (On The Media, 2011). This might appear politically
infeasible in the United States given strong protection of political speech;
however, numerous states have laws against false political ads that have
been shown to reduce the incidence of misleading advertising (Zhang, 2015).
Regardless, this type of policy is worthy of study, along with other variations
on policy used across nations. For example, is there a correlation across
nations between freedom of speech protection and polarization in the new
media era?
An approach that is likely even more radical and less feasible would be
to apply Pigouvian taxation to externalities from political information: tax
actions with negative externalities (perhaps consumption of certain types of
media known to report misleadingly partisan news) and subsidize actions
with positive externalities (demonstration of political knowledge or even tolerance?). Perhaps some type of field experiment could be conducted to examine the effects of such policies?
Cagé (2015) makes the case for a creative but reasonable policy proposal:
a new type of organization specific to media, akin to universities with both
nonprofit and private sector elements. This proposal is meant to address the
issue of declining revenue and quality of the press in general, and not partisanship, in particular. However, perhaps there is a link—maybe biased news
is cheaper to produce than neutral news, or journalists prefer to be neutral,
and only resort to slant under competitive pressure? (Cagé argues something
similar is the case for entertainment and soft news.)
Academic research should not be bounded by practical constraints, but
should take them into consideration. Given political constraints, it might
be more productive for the literature to provide recommendations, or even
just potentially practical findings, for conscientious media consumers. For
example, can we tell consumers what effects partisan news consumption has
on stress and mental health? Or on social outcomes, such as diversity and
number of friends, and family relationships, or overall happiness? Can the
literature advise on technologies, or even heuristics, that consumers could
use to improve the neutrality or partisan diversity of news they obtain (for
those who aspire to do this)? How can we enlighten those of us who do not

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

(to help address affective polarization)? Are there techniques informed by
science that we can better use to de-bias others in casual conversation? We
all know the importance of social pressure, but its power can still surprise
us, especially in the absence of financial incentives [e.g., DellaVigna, List,
Malmendier, and Rao (2014), or the emphasis of Kahneman (2011), on the
importance of water cooler gossip]. Can we make consuming one-sided partisan news, or holding biased partisan views, or being intolerant or hostile to
opposing partisans, socially unacceptable—that is, uncool, or just politically
incorrect?
This literature may also be of service for conscientious journalists, or those
who train them. Can we provide more clear guidance on the fundamental
issue of the ideal of objectivity? What other research questions could we
address of practical value to journalists? Can feedback about ideology or
accuracy to journalists affect the quality of their output? Field experiments
and text analysis are likely the best tools for making progress on these and
similar questions.
METHODS AND DISCIPLINES
It will not shock anyone if I report that papers written by economists tend to
cite few (if any) papers from other fields, and those written by noneconomists
tend to neglect our work. This focus on papers from one’s home discipline
is at times reasonable. However, a slant in the way we cite work seems less
appropriate for papers intended to enhance understanding of a question that
crosses disciplinary boundaries.8
Are authors unaware of the range of relevant work from other disciplines?
Or do we cite minimal research from other fields because its value is underestimated? Or because the value is judged correctly to be not that high? I am
not foolish enough to try to answer these questions, but think they are worth
considering for scholars from all fields.
In fact, I think that citation by high-quality surveys from other fields is a
useful goal for economists to strive for in our work. Such citation is prima
facie evidence that the question addressed is important, and that the results
are compelling and clearly communicated. Most economics papers that have
been widely cited outside of economics are empirical and use text analysis,
8. I have already mentioned several examples of findings that seem underappreciated outside their
home discipline; another example is the third-person effect, referring to the tendency to overestimate media
influence. This is a widely used term and concept outside of economics (Davison, 1983), but not mentioned
in any economics paper to my knowledge. Furthermore, while reviewing the proofs for this essay I became
aware of Knobloch-Westerwick (2014), which seems to exemplify this issue. Her book cites numerous
studies of selective exposure—some of which focus on information—that have been neglected by economics, while also largely neglecting the more recent, closely related work by economists.

Partisan News: A Perspective from Economics

13

large data sets, or a clever identification strategy. Theory papers from economics (e.g., the work on rational demand for bias), on the other hand, have
been more ignored outside of our field. What could make our theoretical
results more compelling to noneconomists? Exploring dynamically and/or
socially richer models, which do not necessarily have analytical solutions
(e.g., agent-based models)? Or perhaps the Bayesian framework and theory
results will be appreciated across fields in good time? Or perhaps questions
addressed by economic theory papers are not of as broad interest as those of
empirical papers? Again, I do not offer answers here, just questions.
CONCLUDING REMARKS
Recent developments in the media landscape have been great fodder for
research but perhaps not great for social welfare. Media failure may directly
cause political failure with enormous implications. Economists and scholars
from other disciplines have explored a wide range of theoretical issues and
made progress with empirical analysis in understanding reality. Going forward, I suggest a focus on work that will be of interest across disciplines
and/or be policy relevant, and that we make a better effort to be aware of,
and learn from, work across disciplines. We have yet to “save the media” (as
called for by Cagé). However, we should not give up hope that we can help
to do so.
ACKNOWLEDGMENTS
I thank (especially) the Emerging Trends editors, Bob Scott and Marlis
Buchmann, and also Kevin Arceneaux, Siddhartha Bandyopadhyay, David
D’Alessio, Stefano DellaVigna, Brendan Nyhan, Gaurav Sood, and Michael
Socolow for very helpful comments and suggestions. This essay was written
while I was a visitor at the University of Virginia’s department of economics;
I am grateful for their hospitality.
REFERENCES
Ahler, D. J. (2014). Self-fulfilling misperceptions of public polarization. The Journal of
Politics, 76(03), 607–620.
Andina-Díaz, A. (2011). Mass media in economics: Origins and subsequent contributions. Cuadernos de Ciencias Económicas y Empresariales, 61, 89–101.
Arceneaux, K., Johnson, M., Lindstädt, R., & Wielen, R. J. (2016). The influence of
news media on political elites: Investigating strategic responsiveness in congress.
American Journal of Political Science, 60(1), 5–29.
Bandyopadhyay, S., Chatterjee, K., & Roy, J. (2015). Manufacturing extremism: Political consequences of profit-seeking media (No. 15-14).

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Barber, M., & McCarty, N. (2015). Causes and consequences of polarization. Solutions
to Political Polarization in America, 15, 15–58.
Baum, M. A., & Zhukov, Y.M. (2015). Media ownership and news coverage of international conflict. Working paper.
Burke, J. (2009). Unfairly balanced: Unbiased news coverage and information loss.
Available at SSRN 1020627.
Cagé, J. (2015). Sauver les médias. Capitalisme, financement participatif et démocratie.
Paris, France: Le Seuil (La République des Idées).
Campante, F. R., & Hojman, D. A. (2013). Media and polarization: Evidence from the
introduction of broadcast TV in the United States. Journal of Public Economics, 100,
79–92.
Chan, J., & Suen, W. (2009). Media as watchdogs: The role of news media in electoral
competition. European Economic Review, 53(7), 799–814.
Chiang, C. F., & Knight, B. (2011). Media bias and influence: Evidence from newspaper endorsements. The Review of Economic Studies, 78(3), rdq037.
Clinton, J. D., & Enamorado, T. (2014). The National News Media’s effect on congress:
How Fox News affected elites in congress. The Journal of Politics, 76(04), 928–943.
D’Alessio, D., & Allen, M. (2000). Media bias in presidential elections: A
meta-analysis. Journal of Communication, 50(4), 133–156.
Davison, W. P. (1983). The third-person effect in communication. Public Opinion Quarterly, 47(1), 1–15.
DellaVigna, S., & Kaplan, E. D. (2007). The Fox News effect: Media bias and voting.
The Quarterly Journal of Economics, 122(3), 1187–1234.
DellaVigna, S., List, J. A., Malmendier, U., & Rao, G. (2014). Voting to tell others (No.
w19832). National Bureau of Economic Research.
Fang, R. Y. (2014). Media bias, political polarization, and the merits of fairness.
Garrett, K. R., Gvirsman, S. D., Johnson, B. K., Tsfati, Y., Neo, R., & Dal, A. (2014).
Implications of pro- and counterattitudinal information exposure for affective
polarization. Human Communication Research, 40(3), 309–332.
Gentzkow, M., & Shapiro, J. M. (2010). What drives media slant? Evidence from US
daily newspapers. Econometrica, 78(1), 35–71.
Gentzkow, M., Shapiro, J. M., & Stone, D. F. (2014). Media bias in the marketplace:
Theory (No. w19880). National Bureau of Economic Research.
George, L. (2007). What’s fit to print: The effect of ownership concentration on product variety in daily newspaper markets. Information Economics and Policy, 19(3),
285–303.
Groeling, T. (2013). Media bias by the numbers: Challenges and opportunities in the
empirical study of partisan news. Political Science, 16(1), 129.
Groseclose, T., & Milyo, J. (2005). A measure of media bias. The Quarterly Journal of
Economics, 120(4), 1191–1237.
Halberstam, Y., & Knight, B. (2014). Homophily, group size, and the diffusion of political information in social networks: Evidence from twitter (No. w20681). National
Bureau of Economic Research.

Partisan News: A Perspective from Economics

15

Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009).
Feeling validated versus being correct: A meta-analysis of selective exposure to
information. Psychological Bulletin, 135(4), 555.
Hazlett, T. W., & Sosa, D. W. (1997). Was the fairness doctrine a “chilling effect”? Evidence from the postderegulation radio market. The Journal of Legal Studies, 26(1),
279–301.
Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology a social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431.
Kahneman, D. (2011). Thinking, fast and slow. New York, NY: Macmillan.
Knobloch-Westerwick, S. (2014). Choice and preference in media use: Advances in selective
exposure theory and research. New York, NY: Routledge.
Lelkes, Y., Iyengar, S., & Sood, G. (2015). The hostile audience: Selective exposure
to partisan sources and affective polarization. American Journal of Political Science,
1–16.
Levendusky, M. S. (2013). Why do partisan media polarize viewers? American Journal
of Political Science, 57(3), 611–623.
Levendusky, M., & Malhotra, N. (2015). Does media coverage of partisan polarization affect political attitudes? Political Communication, 1–19 [ahead-of-print].
Martin, G. J., & Yurukoglu, A. (2014). Bias in cable news: Real effects and polarization
(No. w20798). National Bureau of Economic Research.
McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media.
Public Opinion Quarterly, 36(2), 176–187.
Mullainathan, S., & Shleifer, A. (2005). The market for news. American Economic
Review, 1031–1053.
Nyhan, B. (2012). Does the US media have a liberal bias? Perspectives on Politics,
10(03), 767–771.
Oliveros, S., & Várdy, F. (2015). Demand for slant: How abstention shapes voters’
choice of news media. The Economic Journal, 125(587), 1327–1368.
On The Media. (2011). Lying is illegal in Canada(’s news broadcasts). Retrieved from
http://www.onthemedia.org/story/133101-lying-is-illegal-in-canadas-newsbroadcasts/transcript/
Perego, J., & Yuksel, S. (2015). Media competition and the source of disagreement.
Working paper.
Prat, A., & Strömberg, D. (2013). The political economy of mass media. In Advances
in economics and econometrics: Volume 2, applied economics: Tenth world congress (Vol.
50, pp. 135). Cambridge, England: Cambridge University Press.
Prior, M. (2013). Media and political polarization. Annual Review of Political Science,
16, 101–127.
Puglisi, R., & Snyder, J. M. (2016). Empirical studies of media bias. In S. Anderson, J.
Waldfogel & D. Stromberg (Eds.), Handbook of media economics. Elsevier Press.
Schroeder, E., & Stone, D. F. (2015). Fox News and political knowledge. Journal of
Public Economics, 126, 52–63.
Shleifer, A. (2015). Matthew Gentzkow, Winner of the 2014 Clark Medal. The Journal
of Economic Perspectives, 29(1), 181–192.

16

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Sobbrio, F. (2013). The political economy of news media: Theory, evidence and open issues.
Handbook of alternative theories of public economics. Cheltenham, England: Edward
Elgar Press.
Socolow, M. J. (2008). The hyped panic over ’War of the Worlds’. The Chronicle of
Higher Education.
Stone, D. F. (2013). Media and gridlock. Journal of Public Economics, 101, 94–104.
Strömberg, D. (2015). Media and politics. Annual Review of Economics, 7, 173–205.
Stroud, N. J. (2010). Polarization and partisan selective exposure. Journal of Communication, 60(3), 556–576.
Sunstein, C. R. (2001). Republic.com. Princeton, NJ: Princeton University Press.
Sutter, D. (2000). Can the media be so liberal—the economics of media bias. Cato
Journal, 20, 431.
Westwood, S. J., Iyengar, S., Walgrave, S., Leonisio, R., Miller, L., & Strijbis, O. (2015).
The tie that divides: Cross-national evidence of the primacy of partyism. Working
paper.
Zhang, Z. (2015). Swiftboating: Misleading Advertising in Presidential Campaigns.
Unpublished manuscript.

DANIEL F. STONE SHORT BIOGRAPHY
Daniel F. Stone is an assistant professor of economics at Bowdoin College. He
specializes in the study of belief formation and choice under uncertainty, and
political beliefs and choices, in particular. He received his PhD in economics
from Johns Hopkins University in 2008.

RELATED ESSAYS
Understanding American Political Conservatism (Political Science), Joel D.
Aberbach
Learning Across the Life Course (Sociology), Jutta Allmendinger and Marcel
Helbig
Public Opinion and International Conflict (Political Science), Adam J.
Berinsky
Empathy Gaps between Helpers and Help-Seekers: Implications for Cooperation (Psychology), Vanessa K. Bohns and Francis J. Flynn
Mental Models (Psychology), Ruth M. J. Byrne
Heuristic Decision Making (Political Science), Edward G. Carmines and
Nicholas J. D’Amico
Elites (Sociology), Johan S. G. Chu and Mark S. Mizruchi
Neoliberalism (Sociology), Miguel Angel Centeno and Joseph N. Cohen
Misinformation and How to Correct It (Psychology), John Cook et al.
Expertise (Sociology), Gil Eyal

Partisan News: A Perspective from Economics

17

Cognitive Processes Involved in Stereotyping (Psychology), Susan T. Fiske
and Cydney H. Dupree
Political Advertising (Political Science), Erika Franklin Fowler
Controlling the Influence of Stereotypes on One’s Thoughts (Psychology),
Patrick S. Forscher and Patricia G. Devine
Media and the development of Identity (Psychology), Adriana M. Manago
Media Neuroscience (Communications & Media), J. Michael Mangus et al.
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E.
Tetlock


Partisan News: A Perspective from
Economics
DANIEL F. STONE

Abstract
I briefly summarize the economics literature on ideologically slanted political media
(which I call, for short, partisan news), and discuss directions for future research. In
the literature review, I take a history of thought approach, describing how theory and
empirical work have fed off one another and real-world events. I also note ways in
which the work of economists differs from comparable work from other disciplines.
In the discussion of future research, I identify open questions and policy options, and
assess the relationship between research from economics and other disciplines.

INTRODUCTION
The political news media has been studied extensively by social science and
journalism scholars for decades.1 Economists mostly neglected the Fourth
Estate until the early 2000s. Interest in the topic was motivated then by
two main factors: (i) growing public distrust in the media as an institution
and (ii) significant technology-driven changes occurring in the content and
competitive structure of media markets. Interest was strengthened as we
(perhaps belatedly) recognized that the media as an industry has greater
social and political importance than most, maybe all, others. The time was
also right for economics media research due to the recent development of
applicable research methods (computational text analysis, natural and field
experiments, and ideas from behavioral economics), allowing economists to
address questions that beforehand may have seemed outside of our domain.
A good portion of this new economics political media literature focuses
on the specific issue of ideologically slanted political media content, which
I call, for short, partisan news. For most industries, economists believe that
increasing the diversity of goods serves consumers and society well. However, the overall benefits of greater diversity in news viewpoints are not as
1. A classic example is McCombs and Shaw (1972).
Emerging Trends in the Social and Behavioral Sciences.
Robert Scott and Marlis Buchmann (General Editors) with Stephen Kosslyn (Consulting Editor).
© 2016 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.

1

2

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

clear. A more competitive marketplace of ideas may make truth more likely
to eventually emerge. On the other hand, a more partisan news landscape
might skew political beliefs and actions. Which of these effects dominates?
How do the effects depend on context? What other more subtle issues should
be considered? How can partisan news be studied empirically in a scientific
manner?
In this essay, I summarize findings from the economics literature on
partisan news and present thoughts on where the literature stands and
directions for future work. In the brief literature review,2 I take a history of
thought approach, discussing how theory and empirics fed off one another
and real-world events, and write to a multidisciplinary audience, noting
specific ways in which economics work differs from that of other fields,
admittedly with a bias toward economists’ contributions. In the discussion
of potential future work, I take a more critical perspective. I suggest that
economics media research make a stronger attempt to contribute to policy
analysis (broadly defined), and that economics research can benefit from
better incorporating past findings, and future reactions, of media scholars
from other disciplines.
THE LITERATURE
THE “LIBERAL MAINSTREAM MEDIA” AND THE SUPPLY–DEMAND FRAMEWORK
Economics research on partisan news was first motivated in part by concerns
about a liberal bias in the “mainstream media” in the United States. Scholars from other disciplines had analyzed bias using “content analysis,” which
involves researcher judgment calls to code the ideological slant of media content. See, for example, D’Alessio and Allen (2000). While this tool is flexible
and continues to be useful, it has limitations: it is time- and labor-intensive
and not precisely replicable. In the 2000s, a number of economics papers
developed more formulaic text analysis methods less subject to these issues.3
Researchers from other fields quickly adopted similar methods; see Groeling
(2013) and Puglisi and Snyder (2016) for recent reviews.
The liberal media claim also raised the theoretical question of how such a
bias could persist in a competitive market. Sutter (2000) provided an early,
informal economic analysis, noting the distinction between forces from the
supply side (characteristics of owners, journalists, or advertisers) and the
2. Gentzkow, Shapiro, and Stone (2014) (GSS) and Puglisi and Snyder (2016) are other recent surveys
of theory and empirical economics work on partisan media, respectively. Shleifer (2015) also discusses the
history of thought on this topic, but has a more limited scope. Strömberg (2015) and other Handbook of
Media Economics chapters review related issues in media economics, and Andina-Díaz (2011), Prat and
Strömberg (2013), and Sobbrio (2013) are earlier surveys that include coverage of partisan news.
3. Groseclose and Milyo (2005) was perhaps the first and has since been particularly highly cited (and
has attracted criticism; see, e.g., Nyhan, 2012).

Partisan News: A Perspective from Economics

3

demand side (consumer preferences). This framework has been useful in
much subsequent literature. A famous example is Gentzkow and Shapiro
(2010), an empirical analysis of “what drives media slant.” They study US
newspapers in 2005 and find that consumer ideology explains much more of
the variation in slant than owner ideology.
CAUSES OF DEMAND-DRIVEN BIAS AND PARTISAN SELECTIVE EXPOSURE
Concerns about aggregate liberal bias seemed alleviated, at least somewhat,
in the later 2000s. This may have been partly due to research results like those
of Gentzkow and Shapiro. However, it was also apparent that media content
in reality was becoming more ideologically diverse. Consumers could essentially choose their media slant(s) among online and cable television news
outlets (in addition to more traditional network television and print news
options). Slant was no longer something imposed by the supply side on a consumer with few options. Researchers observed this trend and were motivated
to pursue deeper analysis of causes and effects of consumer demand for partisan news, and in particular, the preference for like-minded news. Scholars
from other fields call the resulting segregated news consumption “partisan
selective exposure.” Sunstein (2001) sounded an early warning about the
potential emergence and socially harmful effects of “echo chambers,” essentially an extreme version of partisan selective exposure in which consumers
limit themselves to narrowly like-minded viewpoints.
Scholars from other disciplines generally assumed that the preference for
like-minded news is motivated by psychological factors (Hart et al., 2009). In a
nutshell, most people feel good when their politics are confirmed, and find it
off-putting to be challenged. Economists have incorporated such psychological factors in a number of papers (Mullainathan & Shleifer, 2005). However,
perhaps the greater contribution of economists was to also explore the range
of more rational forces that could drive like-minded news demand.
In Gentzkow et al. (2014), we present a simple game theoretic model to capture key elements from this literature.4 The model also serves as the basis for
a formal definition of reporting bias, which we taxonomize into two types.
It is useful to discuss this model here for a few reasons: to give readers who
may be unfamiliar with such modeling a better sense of how it works, to
illustrate how important conclusions follow from model assumptions, and
to provide a vocabulary for later discussion in this essay. The model consists
of the following assumptions:

4. Fang (2014), Oliveros and Várdy (2015), and Perego and Yuksel (2015) are more recent papers that
analyze innovative, richer models of rationally demanded bias.

4

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

•
•

•

•

•

•

There are two types of players: news consumers and firms.
There is an uncertain “state of the world,” either L or R. If L is true, this
can be thought of as a situation that (truly) favors a leftist political view,
and if R is true, this favors rightists.
Media firms observe private information on the state (a “signal”) but
before doing so, each must announce a “reporting strategy” publicly: a
way of transforming the signal into a news report.
Firms observe their signal, then report news based on their signal and
reporting strategy.
Consumers choose which firm to consume news from, or to not consume
news at all, based on the reporting strategies and a private opportunity
cost, and, if news is consumed, Bayesian update their prior belief (probability) about each state.
Finally, each consumer chooses an action (political, such as a vote, or
private consumption, e.g., whether to buy organic or conventional
food) and receives “payoffs” given her action, the true state, and the
consumer’s utility function (preferences).

GSS’s definition of reporting bias is, in words, that strategy x is biased to the
right of strategy y if consumer beliefs that R is the true state likely increase
when the firm switches from y to x without the consumer being aware of the
switch. The definition of left bias is analogous.
The two types of (each) bias are “distortion bias” and “filtering bias.” The
former occurs when it is possible for a firm to directly report its signal as the
news. Call this the neutral strategy. In this case, no other strategy can increase
the (expected) payoff of a consumer who seeks news only for instrumental
value, that is, to inform her choice of action. Thus, any strategy biased right
or left of the neutral strategy cannot be rationally demanded for instrumental
value.
The second type of bias, filtering bias, occurs when signals are more complex than news reports, and so firms must filter private information into
simpler news. In this case, there is no strategy that maximizes the payoff of
all rational instrumental value-seeking consumers. There may still be a seemingly neutral strategy, but a strategy biased to the right or left of this may
actually be (instrumental) payoff-maximizing for some consumers, and thus
rationally demanded. Bias in this case is not necessarily “payoff-reducing”
for consumers.
To see these ideas more clearly consider a specific example in which there is
just one firm whose signal is more complex than the news it can report. Suppose this signal consists of three “subsignals,” each of which is binary and
supports L or R, but the news report is limited to a single binary statement
(supporting either L or R). A seemingly neutral reporting strategy would

Partisan News: A Perspective from Economics

5

be for the firm to report news supporting L if the majority (two or three)
of its subsignals support L, and report news supporting R if the majority of
subsignals support R. A strategy in which “R” was reported only if all three
subsignals support R would be “biased left” of such a neutral strategy. However, this left-biased strategy could increase the payoff of some consumers,
namely, if they have a strong left-leaning utility function. Suppose the action
is “vote L” or “vote R” (and suppose also that the vote is meaningful). If a
consumer would only rationally switch from (vote) L to R if the evidence
supporting this switch was as strong as possible (three R subsignals), then
this consumer would prefer the left-biased strategy to the neutral strategy.
Thus, rational choice can explain partisan selective exposure (leftists seeking left-biased news; the logic applies to rightists symmetrically). Figure 1
depicts the game tree for each of these two reporting strategies. Chiang and
Knight (2011) supports the rationality of consumer interpretation of partisan news.
Groeling (2013) and Puglisi and Snyder (2016) discuss other views on
defining and categorizing types of bias. Groeling splits bias into two types,
“selection” (bias in stories to report on) and “presentation” (bias in what
information to present for a particular story). Groeling refers to both of these
types of bias as involving distortion, which seems to imply that they cannot
be rationally demanded, although GSS’s filtering bias potentially applies to
both of Groeling’s types of bias.
The GSS model implies that rational demand for instrumental value cannot
explain distortion bias. GSS discuss how distortion bias can be caused by two
other demand-side factors: (i) noninstrumental preferences for like-minded
news (the “feels good” motive) and (ii) firm reputational incentives (firms
may “look good” by distorting news toward what consumers believe is likely
true). GSS also discuss how if bias is caused by such reputational motives,
bias is relatively likely to be reduced by competition, but demand-side biases
(distortion or filtering) driven by other causes tend to become more diverse
via competition.
Economists have also contributed to empirical research on consumer preferences for partisan news. Across the social sciences this research has moved
away from relying on surveys, and toward data that more directly captures
behavior (Prior, 2013). Economics research has been consistent with this trend
(Halberstam & Knight, 2014) and has proposed new ways of measuring and
interpreting the extent of partisan selective exposure in different contexts.
However, the economics literature on this specific topic is limited and the
large majority of researchers working in this area come from other fields.

6

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

The “neutral”
reporting strategy
Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Nature
draws
state

Firm reports
pro-L news
Firm sees 2
pro-L
subsignals

The “left-biased”
reporting strategy

State = L

State = R

Firm reports
pro-R news
Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals

Firm sees 3
pro-R
subsignals

Firm sees 2
pro-L
subsignals

Firm sees 2
pro-R
subsignals

Firm sees 3
pro-L
subsignals
Firm reports
pro-L news
Firm sees 3
pro-L
subsignals

Firm sees 2
pro-L
subsignals

State = L

Nature
draws
state

State = R

Firm sees 3
pro-R
subsignals
Firm reports
pro-R news
Firm sees 3
pro-R
subsignals

Firm sees 2
pro-R
subsignals

Figure 1 Partial game trees, for example, described in text in which firm
observes three binary subsignals and makes one binary report. All events within a
dashed box (an information set) are observationally equivalent to consumers (i.e.,
each event within a dashed box results in the same news report, which is the only
new information a consumer can observe before choosing her action). Note the
state and subsignals are random variables and occur with probabilities that the
game trees omit (probabilities for events that may arise from a given box (node)
must sum to one).

Partisan News: A Perspective from Economics

7

MEDIA EFFECTS: THEORY AND EMPIRICS
The effects of various forms of reporting strategies on media consumer beliefs
and actions are what ultimately drive “social welfare effects”—the implications for societal well-being. Sometimes, models measure these effects just in
terms of consumer and/or producer surplus; often models include political
effects (given voter preferences, whether voters elect the optimal candidate
and/or politicians take optimal actions, allowing for formalization of “political failure”). In general, theory models imply that supply-driven bias is likely
harmful to consumers and society overall, while the effects of demand-side
bias are more ambiguous and depend on various contextual factors. Positive
effects may be due to rationally demanded filtering bias, cross-checking, and
increased engagement of more partisan consumers. Cross-checking and competition of advocates are likely more beneficial when advocates have stronger
incentives to search for information than neutral media. Advocate competition is less socially beneficial when “truth” is more ambiguous and difficult
to verify (e.g., forecasts for climate in 50 years vs tomorrow’s weather).
Negative effects (of demand-driven bias) may be due to consumers choosing socially suboptimal news due to psychological biases, or preferences
for noninstrumental value from news (entertainment or ego confirmation).
Choices resulting from such preferences may be individually rational but
socially harmful as they neglect positive externalities from more informative
news. The engagement and biased processing effects may even go hand in
hand, as more biased consumers may be more engaged by biased media.
Endogenizing actions by politicians likely exacerbates effects [good ones
from more informative media, e.g., Chan and Suen (2009), and bad effects
otherwise, as in Stone (2013)].
To go into more depth on how modeling assumptions affect welfare implications, I refer again to the results from GSS. We show that the private welfare
effects of demand-driven filtering bias are positive if the cause of bias is a partisan utility function. However, welfare effects are ambiguous if demand for
bias is caused by “partisan beliefs”—prior beliefs biased toward one state or
the other. Intuitively, ideal reporting is truly subjective in the first case. People with different “values” can be better off with different ways of filtering
facts into news. In the second case, as ultimately people share the same goals
but disagree on how to reach them, the ideal reporting strategy is not subjective. If preferences for reporting differ, at least some of those preferences
must be (objectively) off base. Noneconomics literature often refers to consumer
demand for biased news as being driven by ideological “attitudes.” It can
be unclear to what extent attitudes include inherently subjective preferences
(e.g., preferences for apples or oranges) versus beliefs about objective facts
(e.g., whether apples or oranges most improve one’s health).

8

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

A final contribution of economics is a focus on credible identification of
real-world media effects. Much of the prior research depended on lab or
survey experiments or considered empirical evidence that, while still valuable, has less well identified causality.5 Most economics empirical research
uses random, or quasi-random, variation in media exposure in the “field”
to identify causal outcomes. Many papers from other disciplines use these
methods as well (Arceneaux, Johnson, Lindstädt, & Wielen, 2016) and perhaps even more have used field experiments. DellaVigna and Kaplan (2007)
is a key paper from economics on media effects, which indeed reports substantial voting effects caused by the entry of a partisan outlet. Subsequent
papers have found results confirming these conclusions and developed them
further.
WHERE WE STAND
Summing up very briefly, economists have contributed new empirical methods for analysis of partisan news, and new theoretical frameworks for analyzing the causes and consequences of partisan bias. In particular, we have
fleshed out the range of rational forces that may drive demand for bias and
the mechanisms by which bias may influence social welfare, and contributed
to precision in measurement of media bias and its effects, helping confirm
that partisan news effects are real and substantial. Still, it seems fair to say
that in economics and beyond we remain unclear on the really big questions:
to what extent is partisan media good or bad for society? What can be done
about socially harmful aspects of partisan news?
PARTISAN MEDIA EFFECTS: POLITICS AND POLICY
Broadly speaking, if “more partisan” media offered more benefits than costs,
we would expect political systems to function better as media outlets grew
more partisan. Furthermore, broadly speaking, this does not appear to be the
case, at least in the United States: most political scientists would agree the
US political system has not fared well in the new era of more partisan media.
It is well established that political elites have become more polarized, and
political gridlock has increased, in the United States in recent decades (Barber
& McCarty, 2015). This suggests that the growth in partisanship of media
that has occurred in roughly the same period has, overall, done more harm
than good. Bandyopadhyay, Chatterjee, and Roy (2015) discuss examples of
similar associations between media and polarized politics from other parts
of the world.
5. See, for example, Socolow (2008) for a discussion of the challenges in measuring media effects (and
a defense of the strategic sophistication of media consumers).

Partisan News: A Perspective from Economics

9

The most straightforward way that partisan media may have caused the
observed polarization of elites is via partisan selective exposure: if voters
tended toward like-minded media, and media became more ideological, then
this may have ideologically polarized the electorate, causing strategic or sincere polarization of elected officials. Prior (2013) reviews research related
to this idea and finds that the evidence supporting it is thin. He concludes
that although partisan media may have caused already partisan citizens to
become more extreme, most Americans continue to be mostly moderate or
indifferent toward politics. Still, there are papers offering direct evidence
supporting the link between media and general polarization, including Campante and Hojman (2013), Levendusky (2013), Martin and Yurukoglu (2014),
and Stroud (2010).
There are several other mechanisms by which partisan media may have
contributed to the negative political trends. First, even if just the tails of the
citizen ideology distribution were polarized, this could still have had important political effects. Party activists could exist mainly in these tails and be
the ones who apply the key pressure on politicians causing them to take more
extreme positions. Arceneaux et al. (2016) and Clinton and Enamorado (2014)
show that partisan media has indeed caused elected officials to act in a more
polarized way.6 The economics literature has not analyzed the behavior of
party activists as a distinct and important group within the electorate; this
may be a fruitful area for future work.
Second, even if the masses’ fundamental values were not polarized, beliefs
and information may have become more skewed. The literature could
certainly dig further into information and knowledge effects of new media
(Schroeder & Stone, 2015). Ahler (2014) and Levendusky and Malhotra
(2015) are part of a related literature outside of economics presenting related
evidence, which claims that citizens are misinformed about (overestimate)
growth in ideological differences across the parties.
Third, research on growth in “affective polarization” and “partyism”—that
opposing partisans have grown to dislike each other more over time—has
found much stronger empirical support (Garrett et al., 2014; Iyengar, Sood,
& Lelkes, 2012; Westwood et al., 2015). Lelkes, Iyengar, and Sood (2015) use
quasi-random variation in broadband access to show that online media has
played a causal role in these trends.
The economics literature (on media or otherwise) typically does not consider the concept of “dislike.” While a good portion of economics work on
partisan news incorporates behavioral factors, any disagreement occurring
in these models seems, well, agreeable. In the variations covered in GSS,
news consumers may disagree before getting news due to heterogeneous
6. Bandyopadhyay et al. (2015) and Stone (2013) make related theoretical points, showing how media
can cause polarized political actions or gridlock without mass polarization.

10

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

priors, or disagree after getting news with different slants. However, in neither case would the model imply that voters harbor hard feelings against
one another. A natural new direction for the economics literature, empirical
and theory, would be to incorporate a broader range of “behavioral” factors,
such as affective polarization. The economics theory papers that avoid any
behavioral or noninstrumental assumptions often seem to take this modeling approach as a point of pride. As it is harder to explain the existence of
slanted news without consumers having a direct preference for slant, a model
that avoids such assumptions is more impressive, in a sense. It is important
to understand what phenomena can and cannot be explained with different
sorts of models. However, of course, we also want our models to be consistent
with basic aspects of reality.
To be clear, partisan new media has certainly also had positive welfare
effects. Some types of watchdog journalism have improved; examples
include the Drudge Report exposing the forgery of documents criticizing
George W. Bush’s National Guard service, and “citizen journalist” videos of
police violence. Partisan media may do a better job of avoiding distortion in
news to appear neutral (Burke, 2009) or uninformative “he-said-she-said”
reporting (Chan & Suen, 2009). Many observers point out that US journalism’s embrace of the “ideal of objectivity” in the late twentieth century was
an exception, not the rule (with respect to both media around the world, and
historical news media in the United States). A careful analysis comparing
costs and benefits of partisan news would certainly be a worthwhile, though
challenging avenue for future work.
Even if the net effects of a move away from the ideal of objectivity are
ambiguous, it is clear that unfettered media markets do not function perfectly
and there is a role for public policy. So far, we (in economics and beyond)
have not cracked the policy nut. A simple and flexible policy approach with
relatively broad appeal, like, for example, libertarian paternalism or Pigouvian taxation, might be ideal. This goal is likely too ambitious, but still worth
keeping in mind. Preservation of competition is the closest thing historically
to such a general policy principle. The literature supports the importance of
competition for avoiding negative impacts from supply-side bias, and this
risk is still relevant (Baum & Zhukov, 2015).7 Even the Newspaper Preservation Act, which allows two newspapers in the same metropolitan area to
collude on pricing to maintain both firms’ viability, requires that the news
operations remain separate and competitive. However, at the national level,
supporting competition is not what we need policy to do. The national media
7. Historically, a key concern was cross-ownership of outlets across platforms (print, television, and
radio), and this is still a concern even with the lower costs of entry online (see, e.g., Cagé, 2015, on the
trend of wealthy individuals buying newspapers), though it is also possible consolidated ownership can
result in more differentiated content (George, 2007).

Partisan News: A Perspective from Economics

11

market is plenty competitive, and the political problems seem to exist despite,
or perhaps because, of this.
A number of other policy options have been considered. The Fairness Doctrine is still regularly discussed. It is almost surely politically infeasible and
its value is questionable regardless (Hazlett & Sosa, 1997). Public funding
of media is another commonly discussed policy, with mixed real-world success. An obvious concern is government capture or the perception thereof.
Another is lack of incentives to maximize audience size and social benefits.
A more radical option sometimes mentioned is Canada’s “prohibition
on falsity” in news (On The Media, 2011). This might appear politically
infeasible in the United States given strong protection of political speech;
however, numerous states have laws against false political ads that have
been shown to reduce the incidence of misleading advertising (Zhang, 2015).
Regardless, this type of policy is worthy of study, along with other variations
on policy used across nations. For example, is there a correlation across
nations between freedom of speech protection and polarization in the new
media era?
An approach that is likely even more radical and less feasible would be
to apply Pigouvian taxation to externalities from political information: tax
actions with negative externalities (perhaps consumption of certain types of
media known to report misleadingly partisan news) and subsidize actions
with positive externalities (demonstration of political knowledge or even tolerance?). Perhaps some type of field experiment could be conducted to examine the effects of such policies?
Cagé (2015) makes the case for a creative but reasonable policy proposal:
a new type of organization specific to media, akin to universities with both
nonprofit and private sector elements. This proposal is meant to address the
issue of declining revenue and quality of the press in general, and not partisanship, in particular. However, perhaps there is a link—maybe biased news
is cheaper to produce than neutral news, or journalists prefer to be neutral,
and only resort to slant under competitive pressure? (Cagé argues something
similar is the case for entertainment and soft news.)
Academic research should not be bounded by practical constraints, but
should take them into consideration. Given political constraints, it might
be more productive for the literature to provide recommendations, or even
just potentially practical findings, for conscientious media consumers. For
example, can we tell consumers what effects partisan news consumption has
on stress and mental health? Or on social outcomes, such as diversity and
number of friends, and family relationships, or overall happiness? Can the
literature advise on technologies, or even heuristics, that consumers could
use to improve the neutrality or partisan diversity of news they obtain (for
those who aspire to do this)? How can we enlighten those of us who do not

12

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

(to help address affective polarization)? Are there techniques informed by
science that we can better use to de-bias others in casual conversation? We
all know the importance of social pressure, but its power can still surprise
us, especially in the absence of financial incentives [e.g., DellaVigna, List,
Malmendier, and Rao (2014), or the emphasis of Kahneman (2011), on the
importance of water cooler gossip]. Can we make consuming one-sided partisan news, or holding biased partisan views, or being intolerant or hostile to
opposing partisans, socially unacceptable—that is, uncool, or just politically
incorrect?
This literature may also be of service for conscientious journalists, or those
who train them. Can we provide more clear guidance on the fundamental
issue of the ideal of objectivity? What other research questions could we
address of practical value to journalists? Can feedback about ideology or
accuracy to journalists affect the quality of their output? Field experiments
and text analysis are likely the best tools for making progress on these and
similar questions.
METHODS AND DISCIPLINES
It will not shock anyone if I report that papers written by economists tend to
cite few (if any) papers from other fields, and those written by noneconomists
tend to neglect our work. This focus on papers from one’s home discipline
is at times reasonable. However, a slant in the way we cite work seems less
appropriate for papers intended to enhance understanding of a question that
crosses disciplinary boundaries.8
Are authors unaware of the range of relevant work from other disciplines?
Or do we cite minimal research from other fields because its value is underestimated? Or because the value is judged correctly to be not that high? I am
not foolish enough to try to answer these questions, but think they are worth
considering for scholars from all fields.
In fact, I think that citation by high-quality surveys from other fields is a
useful goal for economists to strive for in our work. Such citation is prima
facie evidence that the question addressed is important, and that the results
are compelling and clearly communicated. Most economics papers that have
been widely cited outside of economics are empirical and use text analysis,
8. I have already mentioned several examples of findings that seem underappreciated outside their
home discipline; another example is the third-person effect, referring to the tendency to overestimate media
influence. This is a widely used term and concept outside of economics (Davison, 1983), but not mentioned
in any economics paper to my knowledge. Furthermore, while reviewing the proofs for this essay I became
aware of Knobloch-Westerwick (2014), which seems to exemplify this issue. Her book cites numerous
studies of selective exposure—some of which focus on information—that have been neglected by economics, while also largely neglecting the more recent, closely related work by economists.

Partisan News: A Perspective from Economics

13

large data sets, or a clever identification strategy. Theory papers from economics (e.g., the work on rational demand for bias), on the other hand, have
been more ignored outside of our field. What could make our theoretical
results more compelling to noneconomists? Exploring dynamically and/or
socially richer models, which do not necessarily have analytical solutions
(e.g., agent-based models)? Or perhaps the Bayesian framework and theory
results will be appreciated across fields in good time? Or perhaps questions
addressed by economic theory papers are not of as broad interest as those of
empirical papers? Again, I do not offer answers here, just questions.
CONCLUDING REMARKS
Recent developments in the media landscape have been great fodder for
research but perhaps not great for social welfare. Media failure may directly
cause political failure with enormous implications. Economists and scholars
from other disciplines have explored a wide range of theoretical issues and
made progress with empirical analysis in understanding reality. Going forward, I suggest a focus on work that will be of interest across disciplines
and/or be policy relevant, and that we make a better effort to be aware of,
and learn from, work across disciplines. We have yet to “save the media” (as
called for by Cagé). However, we should not give up hope that we can help
to do so.
ACKNOWLEDGMENTS
I thank (especially) the Emerging Trends editors, Bob Scott and Marlis
Buchmann, and also Kevin Arceneaux, Siddhartha Bandyopadhyay, David
D’Alessio, Stefano DellaVigna, Brendan Nyhan, Gaurav Sood, and Michael
Socolow for very helpful comments and suggestions. This essay was written
while I was a visitor at the University of Virginia’s department of economics;
I am grateful for their hospitality.
REFERENCES
Ahler, D. J. (2014). Self-fulfilling misperceptions of public polarization. The Journal of
Politics, 76(03), 607–620.
Andina-Díaz, A. (2011). Mass media in economics: Origins and subsequent contributions. Cuadernos de Ciencias Económicas y Empresariales, 61, 89–101.
Arceneaux, K., Johnson, M., Lindstädt, R., & Wielen, R. J. (2016). The influence of
news media on political elites: Investigating strategic responsiveness in congress.
American Journal of Political Science, 60(1), 5–29.
Bandyopadhyay, S., Chatterjee, K., & Roy, J. (2015). Manufacturing extremism: Political consequences of profit-seeking media (No. 15-14).

14

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Barber, M., & McCarty, N. (2015). Causes and consequences of polarization. Solutions
to Political Polarization in America, 15, 15–58.
Baum, M. A., & Zhukov, Y.M. (2015). Media ownership and news coverage of international conflict. Working paper.
Burke, J. (2009). Unfairly balanced: Unbiased news coverage and information loss.
Available at SSRN 1020627.
Cagé, J. (2015). Sauver les médias. Capitalisme, financement participatif et démocratie.
Paris, France: Le Seuil (La République des Idées).
Campante, F. R., & Hojman, D. A. (2013). Media and polarization: Evidence from the
introduction of broadcast TV in the United States. Journal of Public Economics, 100,
79–92.
Chan, J., & Suen, W. (2009). Media as watchdogs: The role of news media in electoral
competition. European Economic Review, 53(7), 799–814.
Chiang, C. F., & Knight, B. (2011). Media bias and influence: Evidence from newspaper endorsements. The Review of Economic Studies, 78(3), rdq037.
Clinton, J. D., & Enamorado, T. (2014). The National News Media’s effect on congress:
How Fox News affected elites in congress. The Journal of Politics, 76(04), 928–943.
D’Alessio, D., & Allen, M. (2000). Media bias in presidential elections: A
meta-analysis. Journal of Communication, 50(4), 133–156.
Davison, W. P. (1983). The third-person effect in communication. Public Opinion Quarterly, 47(1), 1–15.
DellaVigna, S., & Kaplan, E. D. (2007). The Fox News effect: Media bias and voting.
The Quarterly Journal of Economics, 122(3), 1187–1234.
DellaVigna, S., List, J. A., Malmendier, U., & Rao, G. (2014). Voting to tell others (No.
w19832). National Bureau of Economic Research.
Fang, R. Y. (2014). Media bias, political polarization, and the merits of fairness.
Garrett, K. R., Gvirsman, S. D., Johnson, B. K., Tsfati, Y., Neo, R., & Dal, A. (2014).
Implications of pro- and counterattitudinal information exposure for affective
polarization. Human Communication Research, 40(3), 309–332.
Gentzkow, M., & Shapiro, J. M. (2010). What drives media slant? Evidence from US
daily newspapers. Econometrica, 78(1), 35–71.
Gentzkow, M., Shapiro, J. M., & Stone, D. F. (2014). Media bias in the marketplace:
Theory (No. w19880). National Bureau of Economic Research.
George, L. (2007). What’s fit to print: The effect of ownership concentration on product variety in daily newspaper markets. Information Economics and Policy, 19(3),
285–303.
Groeling, T. (2013). Media bias by the numbers: Challenges and opportunities in the
empirical study of partisan news. Political Science, 16(1), 129.
Groseclose, T., & Milyo, J. (2005). A measure of media bias. The Quarterly Journal of
Economics, 120(4), 1191–1237.
Halberstam, Y., & Knight, B. (2014). Homophily, group size, and the diffusion of political information in social networks: Evidence from twitter (No. w20681). National
Bureau of Economic Research.

Partisan News: A Perspective from Economics

15

Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009).
Feeling validated versus being correct: A meta-analysis of selective exposure to
information. Psychological Bulletin, 135(4), 555.
Hazlett, T. W., & Sosa, D. W. (1997). Was the fairness doctrine a “chilling effect”? Evidence from the postderegulation radio market. The Journal of Legal Studies, 26(1),
279–301.
Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology a social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431.
Kahneman, D. (2011). Thinking, fast and slow. New York, NY: Macmillan.
Knobloch-Westerwick, S. (2014). Choice and preference in media use: Advances in selective
exposure theory and research. New York, NY: Routledge.
Lelkes, Y., Iyengar, S., & Sood, G. (2015). The hostile audience: Selective exposure
to partisan sources and affective polarization. American Journal of Political Science,
1–16.
Levendusky, M. S. (2013). Why do partisan media polarize viewers? American Journal
of Political Science, 57(3), 611–623.
Levendusky, M., & Malhotra, N. (2015). Does media coverage of partisan polarization affect political attitudes? Political Communication, 1–19 [ahead-of-print].
Martin, G. J., & Yurukoglu, A. (2014). Bias in cable news: Real effects and polarization
(No. w20798). National Bureau of Economic Research.
McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media.
Public Opinion Quarterly, 36(2), 176–187.
Mullainathan, S., & Shleifer, A. (2005). The market for news. American Economic
Review, 1031–1053.
Nyhan, B. (2012). Does the US media have a liberal bias? Perspectives on Politics,
10(03), 767–771.
Oliveros, S., & Várdy, F. (2015). Demand for slant: How abstention shapes voters’
choice of news media. The Economic Journal, 125(587), 1327–1368.
On The Media. (2011). Lying is illegal in Canada(’s news broadcasts). Retrieved from
http://www.onthemedia.org/story/133101-lying-is-illegal-in-canadas-newsbroadcasts/transcript/
Perego, J., & Yuksel, S. (2015). Media competition and the source of disagreement.
Working paper.
Prat, A., & Strömberg, D. (2013). The political economy of mass media. In Advances
in economics and econometrics: Volume 2, applied economics: Tenth world congress (Vol.
50, pp. 135). Cambridge, England: Cambridge University Press.
Prior, M. (2013). Media and political polarization. Annual Review of Political Science,
16, 101–127.
Puglisi, R., & Snyder, J. M. (2016). Empirical studies of media bias. In S. Anderson, J.
Waldfogel & D. Stromberg (Eds.), Handbook of media economics. Elsevier Press.
Schroeder, E., & Stone, D. F. (2015). Fox News and political knowledge. Journal of
Public Economics, 126, 52–63.
Shleifer, A. (2015). Matthew Gentzkow, Winner of the 2014 Clark Medal. The Journal
of Economic Perspectives, 29(1), 181–192.

16

EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES

Sobbrio, F. (2013). The political economy of news media: Theory, evidence and open issues.
Handbook of alternative theories of public economics. Cheltenham, England: Edward
Elgar Press.
Socolow, M. J. (2008). The hyped panic over ’War of the Worlds’. The Chronicle of
Higher Education.
Stone, D. F. (2013). Media and gridlock. Journal of Public Economics, 101, 94–104.
Strömberg, D. (2015). Media and politics. Annual Review of Economics, 7, 173–205.
Stroud, N. J. (2010). Polarization and partisan selective exposure. Journal of Communication, 60(3), 556–576.
Sunstein, C. R. (2001). Republic.com. Princeton, NJ: Princeton University Press.
Sutter, D. (2000). Can the media be so liberal—the economics of media bias. Cato
Journal, 20, 431.
Westwood, S. J., Iyengar, S., Walgrave, S., Leonisio, R., Miller, L., & Strijbis, O. (2015).
The tie that divides: Cross-national evidence of the primacy of partyism. Working
paper.
Zhang, Z. (2015). Swiftboating: Misleading Advertising in Presidential Campaigns.
Unpublished manuscript.

DANIEL F. STONE SHORT BIOGRAPHY
Daniel F. Stone is an assistant professor of economics at Bowdoin College. He
specializes in the study of belief formation and choice under uncertainty, and
political beliefs and choices, in particular. He received his PhD in economics
from Johns Hopkins University in 2008.

RELATED ESSAYS
Understanding American Political Conservatism (Political Science), Joel D.
Aberbach
Learning Across the Life Course (Sociology), Jutta Allmendinger and Marcel
Helbig
Public Opinion and International Conflict (Political Science), Adam J.
Berinsky
Empathy Gaps between Helpers and Help-Seekers: Implications for Cooperation (Psychology), Vanessa K. Bohns and Francis J. Flynn
Mental Models (Psychology), Ruth M. J. Byrne
Heuristic Decision Making (Political Science), Edward G. Carmines and
Nicholas J. D’Amico
Elites (Sociology), Johan S. G. Chu and Mark S. Mizruchi
Neoliberalism (Sociology), Miguel Angel Centeno and Joseph N. Cohen
Misinformation and How to Correct It (Psychology), John Cook et al.
Expertise (Sociology), Gil Eyal

Partisan News: A Perspective from Economics

17

Cognitive Processes Involved in Stereotyping (Psychology), Susan T. Fiske
and Cydney H. Dupree
Political Advertising (Political Science), Erika Franklin Fowler
Controlling the Influence of Stereotypes on One’s Thoughts (Psychology),
Patrick S. Forscher and Patricia G. Devine
Media and the development of Identity (Psychology), Adriana M. Manago
Media Neuroscience (Communications & Media), J. Michael Mangus et al.
News Framing Effects and Emotions (Political Science), Andreas R. T. Schuck
and Alina Feinholdt
Implicit Attitude Measures (Psychology), Gregory Mitchell and Philip E.
Tetlock



Media

Partisan News: A Perspective from Economics