Our analytical framework has served to identify a number of channels through which mass media can influence policymaking. Information provided via the media can be used in voting decisions. This can both increase the salience of particular issues and the probability of selecting politicians who act in the public interest. A free press can also serve as a direct check on politicians’ excesses. Therefore we might expect some effects of the media on corruption. In general, finding reliable evidence to match the
Mass Media and Political Accountability 53
richness of the theoretical possibilities is not easy. This section reviews the limited available evidence. What we know comes mainly from fairly reduced-form cross-country evidence. This is a notoriously difficult context to study, with directions of causality being virtually impossible to discern and highly imperfect measures of most variables. A more promising approach is to exploit data from countries that for one reason or another exhibit variation in media activity. A case in point is India, where state-level media institutions vary significantly. Given its greater reliability we dis-cuss the within-country evidence first, and follow this with an assessment of cross-country studies.
Evidence from India
The tradition of a free and independent press has permeated somewhat into the developing world. A prominent example is India, which has a newspaper industry that is distinguished from that in the bulk of other low-income countries by being both free and independent (Ram 1991). Sen (1984) attributed a major role to this freedom and independence in explaining why India has not experienced any major famines since achieving independence. He observed that:
India has not had a famine since independence, and given the nature of Indian politics and society, it is not likely that India can have a famine even in years of great food problems. The government cannot afford to fail to take prompt action when large-scale starvation threatens. Newspapers play an important part in this, in making the facts known and forcing the challenge to be faced (Sen 1984, p. 84).
In contrast, investigators have pointed to China’s lack of democracy and of free-dom of information as reasons why it experienced a major famine between 1958 and 1961, with excess mortality figures ranging between 16.5 and 29.5 million. They have also identified representative democracy and the media as factors in African coun-tries that have succeeded in preventing famines (see Dreze and Sen 1989). As the quote from Sen makes clear, the media increase the salience of government perfor-mance in famine situations by providing information on politicians’ actions that citi-zens can use in their voting decisions.
Though suggestive, Sen’s analysis does not establish a robust link between the development of mass media and government responsiveness. Besley and Burgess (forthcoming) extend the analysis of the role of media in influencing government policy. Using panel data for Indian states for 1958–92 they look at two policy re-sponse systems: first, public distribution of food as a rere-sponse to falls in food pro-duction associated with droughts, and second, spending on calamity relief as a response to crop damage caused by floods. They then examine how newspapers and
politics affect the responsiveness of Indian state governments to these shocks. They find that higher newspaper circulation is associated with increased government re-sponsiveness in both cases. A 10 percent drop in food production is associated with a 1 percent increase in public food distribution in states at the median in terms of newspaper circulation per capita, whereas for states that are in the 75th percentile, a 10 percent drop in food production is associated with a 2.28 percent increase in pub-lic food distribution. An interesting finding is that newspapers published in regional languages are driving the results.3 This makes sense, because regional language news-papers report on localized shocks and local vulnerable groups and politicians are more likely to read them than national newspapers. These results hold up in the face of an array of robustness checks. Thus we have strong evidence that even within India, variation in newspaper circulation can explain how responsive the govern-ment is to the needs of its citizenry.
In line with political agency theory, the interplay between mass media and politi-cal institutions is what determines government responsiveness. Besley and Burgess (forthcoming) examine how various political factors influence government respon-siveness. They find that political turnout increases state governments’ responsive-ness to drought and flood shocks. Greater political competition is also associated with greater responsiveness. This makes sense, as higher turnout and more intense political competition will increase politicians’ incentives to build reputations for be-ing responsive to citizens’ needs.
Table 3.1 ranks 16 Indian states on their responsiveness to the need for public food distribution and their per capita incomes and newspaper circulation. The re-sponsiveness measure indicates how much food, on average, the state government distributes as a response to food production changes in that state. According to this measure, Kerala is the most responsive state and Bihar is the least responsive. The striking feature of this responsiveness ranking is how weakly it correlates with the income per capita ranking; however, the newspaper circulation ranking follows the responsiveness ranking quite closely.
These results, along with those in Besley and Burgess (forthcoming), strongly sup-port the notion that the mass media help solve political agency problems and make governments more accountable. By making politicians’ actions more transparent, the media are informing citizens about the likelihood that they will be protected in the future. In turn, citizens are using this information in making their voting deci-sions. Politicians realize this, which creates an incentive for them to be responsive to shocks. Note that this incentive is present even when politicians have no inherent
3. Data on newspaper circulation were broken down into 19 different languages. Hindi and English tend to be national in scope, while other languages tend to be state specific. Poor, vul-nerable populations will tend to be conversant in the state-specific language. In our regressions we therefore broke out newspaper circulation into Hindi, English, and “other.”
Mass Media and Political Accountability 55
interest in protecting citizens, and are only doing so in the interests of garnering the votes of vulnerable citizens. The mass media thus affect responsiveness both by in-creasing the salience of the social protection issue and by affecting the selection of politicians via voting.
The Indian evidence is consistent with other recent country studies suggesting that the media can affect policymaking. For example, Yates and Stroup (2000) looked at pesticide decisions by the U.S. Environmental Protection Agency and found that it sets more draconian standards when newspapers have published more articles about safety.
This fits with the idea that the media can change issue salience. Using data from the United Kingdom, Larcinese (2001) found that the mass media both determine the po-litical knowledge of citizens and drive voter turnout. Strömberg (2001) related New Deal spending in county-level data for the United States to radio ownership, finding a positive association between the two, which suggests that areas with a higher penetra-tion of radios were more successful at attracting New Deal spending.
Evidence from Cross-Country Data
A number of corruption measures have recently become available for different coun-tries. In accordance with the foregoing theoretical discussion, we would expect that Table 3.1. Ranking of 16 Indian States, Selected Variables, 1958–92
(ranking of 1 is the highest)
Per capita Per capita newspaper
State Responsiveness income circulation
Kerala 1 13 1
Maharashtra 2 3 2
West Bengal 3 5 4
Tamil Nadu 4 8 3
Gujarat 5 4 6
Assam 6 10 15
Uttar Pradesh 7 11 8
Andhra Pradesh 8 9 10
Karnataka 9 6 7
Rajasthan 10 15 9
Punjab 11 1 5
Orissa 12 12 16
Haryana 13 2 13
Jammu and Kashmir 14 7 11
Madhya Pradesh 15 14 12
Bihar 16 16 14
Source: Besley and Burgess (2001).
greater press scrutiny would be associated with lower corruption. Ahrend (2001) and Brunetti and Weder (1999) carried out exercises along these lines and demonstrated the existence of a negative correlation between press freedom measures and corrup-tion in cross-country data; however, drawing causal influences from such findings is difficult, because if corrupt governments can capture the media and then get away with even greater corruption, then the two would be co-determined without one causing the other. To make further progress, we need to measure features of the media market that may make media capture more or less likely, that is, we need to measure proxies for the transactions costs discussed in Besley and Prat (2001).
One promising route takes advantage of the data on media ownership collected by Djankov and others (forthcoming). A plausible notion is that state ownership of the media will lower the cost of capturing the media. Suggestive of this idea, Djankov and others (forthcoming) found that corruption is lower in countries with fewer state-owned newspapers. They found no effect for television.
Using the same data Besley and Prat (2001) also looked at corruption as an out-come. Using three different sources of data on corruption, they found that corrup-tion is negatively correlated with foreign ownership of the media, a finding that is robust to including a wide variety of different control variables. Besley and Prat interpreted this finding as evidence that foreign ownership may be correlated with factors that make the media more effective at generating information.
Djankov and others’ (forthcoming) and Besley and Prat’s (2001) results together point to the need for a better understanding of what determines media capture. A rough way of measuring media capture empirically is to look at whether a country’s press freedom is given a score of less than or equal to 2 on Freedom House’s 6-point scale. We can then ask which characteristics of a country’s media market are signifi-cantly correlated with media capture so measured. To do this, we exploit the data of Djankov and others (forthcoming). Specifically, we include three variables: the ex-tent of foreign ownership, the exex-tent of state ownership, and a measure of ownership concentration. In line with Besley and Prat’s (2001) theoretical predictions, capture is more likely if state ownership of newspapers is more prevalent and newspaper own-ership is more concentrated. The latter suggests that media capture is affected by media plurality. Unlike the case of corruption as an outcome, foreign ownership has no significant effect on the probability of capture.
Besley and Prat (2001) also asked whether the observed correlation between me-dia ownership and political outcomes occurs because private or foreign ownership makes media more efficient or because it makes them less susceptible to political capture. Using perceived media independence as an instrumental variable, one can run a test to check whether, conditional on being noncaptured, media ownership influences political outcomes. In the case of private ownership, no efficiency gain is apparent. The beneficial effect of having private media arises only because they are less likely to be captured. In the case of foreign media, the overidentification test
Mass Media and Political Accountability 57
fails, and one cannot exclude the possibility that both the efficiency and the noncapture channels are active.
Figures 3.1 and 3.2 present these relationships graphically. Figure 3.1 plots press freedom against the degree of state ownership of newspapers using the data from Djankov and others (forthcoming) and Besley and Prat (2001). A high press freedom score denotes more freedom. If we define press capture as having a press freedom score of less than or equal to 3, we can predict the probability of press capture. We do so on the basis of three variables: state ownership of newspapers, foreign ownership of newspapers, and concentration of newspaper ownership. We then graph the Inter-national Country Risk Guide measure of corruption against our predicted probabil-ity of capture (figure 3.2). The upward sloping pattern in the data should now be clear: countries with a higher probability of being captured are more corrupt. While the results are crude, they illustrate how the cross-country data can be used to in-form the theoretical discussion.
Djankov and others (forthcoming) also cast light on the political salience argu-ment, because they found that a broad array of welfare and policy indicators re-sponds positively to a lower fraction of state ownership of newspapers. If one considers
State ownership of newspapers Linear prediction
0
Figure 3.1. Press Freedom and State Ownership of Newspapers
Source: Authors.
Press freedom 6
1
1
ownership to be correlated with the quality of the information generation process, as would be true, for example, if media motivated by profit invest more in researching important news stories, then the changing policy priorities found in the data are consis-tent with the idea that issue salience changes with the development of the media.
Concluding Remarks
A drive to make governments more accountable to citizens’ needs is apparent world-wide. Actors ranging from domestic and international nongovernmental organiza-tions to international financial organizaorganiza-tions have pushed the governance agenda.
While consensus on the need to improve accountability is widespread, what mecha-nisms might be used to achieve this is much less clear. This chapter marshals evi-dence that suggests that a free and independent press working in conjunction with democratic institutions can make governments more responsive to citizens’ needs.
Political agency models have proved useful in examining the role of the media, because they focus on the importance of information in the political process. We have discussed
Predicted probability of capture .044003
Figure 3.2. Corruption and Predicted Probability of Press Capture
Source: Authors.
International Country Risk Guide measure of corruption 6
1
.9213
Mass Media and Political Accountability 59
the various chains of influence by which the media can influence the policy process according to such models. We then argue that the models provide some evidence that ties the electoral process, the media, and government actions together. While the empiri-cal literature is in its infancy, the data certainly suggest that the media have a role to play in political agency. Improved understanding of what makes governments better ser-vants of the people and how strengthening institutions supports this role defines a rich agenda for future work in political economy.
Our central conclusion is that free and independent media should not be viewed as a luxury that only rich countries can afford. Instead our analysis suggests that they should be viewed as a requisite and integral part of representative democracy.
Thus a key question is, what kinds of reforms can strengthen the role of the media in promoting accountability? The question of media regulation is central in this context.
In the past, most countries had extensive industry-specific rules for the press and for broadcasting. By the 1990s many industrial countries, including the United States and the countries of the European Union, saw a dramatic shift in the philosophy underlying media regulation. The new model is that the media should be governed by the same basic competition policy principles that are applied to other industries, which are mostly based on laissez faire except when consumer welfare is demonstra-bly hurt. This has led to the elimination of media-specific rules, such as restrictions on ownership, concentration, and pricing. However, because of their role as political watchdogs, the media differ from other industries. Consumer welfare, defined in terms of the interests of customers (viewers and advertisers), is a limiting notion in that it does not take into account the industry’s effect on voters’ welfare. Future research in the area should re-evaluate existing regulatory regimes in this new light.
References
The word “processed” describes informally reproduced works that may not be commonly avail-able through libraries.
Ahrend, Rudiger. 2001. “Press Freedom, Human Capital, and Corruption.” London School of Economics, London. Processed.
Barro, Robert. 1973. “The Control of Politicians: An Economic Model.” Public Choice 14(Spring):
19–42.
Besley, Timothy, and Robin Burgess. 2001. “Political Agency, Government Responsiveness, and the Role of the Media.” European Economic Review 45(4–6): 629–40.
———. Forthcoming. “The Political Economy of Government Responsiveness: Theory and Evi-dence from India.” Quarterly Journal of Economics.
Besley, Timothy J., and Anne C. Case. 1995. “Does Political Accountability Affect Economic Policy Choices? Evidence from Gubernatorial Term Limits.” Quarterly Journal of Economics 110(3):
769–98.
Besley, Timothy, and Andrea Prat. 2001. “`Handcuffs for the Grabbing Hand? Media Capture
and Political Accountability.”’ London School of Economics, London. Processed.
Brunetti, Aymo, and Beatrice Weder. 1999. “A Free Press Is Bad News for Corruption.” Univer-sity of Basel, Basel, Switzerland. Processed.
Djankov, Simeon, Caralee McLeish, Tatiana Nenova, and Andrei Shleifer. Forthcoming. “Who Owns the Media?” Journal of Law and Economics.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: HarperCollins.
Dreze, Jean, and Amartya Sen. 1989. Hunger and Public Action. Oxford, U.K.: Clarendon Press.
Ferejohn, John. 1986. “Incumbent Performance and Electoral Control.” Public Choice 50(1–3): 5–25.
Larcinese, Valentino. 2001. “Information Acquisition, Ideology, and Turnout: Theory and Evi-dence from Britain.” London School of Economics, London. Processed.
Persson, Torsten, and Guido Tabellini. 2000. Political Economics: Explaining Economic Policy. Cam-bridge, Massachusetts: MIT Press.
Przeworski, Adam, Susan C. Stokes, and Bernard Manin, eds. 1999. Democracy, Accountability, and Representation. Cambridge, U.K.: Cambridge University Press.
Ram, N. 1991. “An Independent Press and Anti-Hunger Strategies: The Indian Experience.” In J. Dreze and A. Sen, eds., The Political Economy of Hunger, vol. 1. Oxford, U.K.: Oxford Univer-sity Press.
Sen, Amartya. 1984. “Food Battles: Conflicts in the Access to Food.” Food and Nutrition 10(1): 81–89.
Strömberg, David. 2001. “Radio’s Impact on the New Deal.” Department of Economics, Institute of International Economic Studies, Stockholm. Processed.
World Bank. 1997. World Development Indicators. Washington, D.C.
Yates, Andrew J., and Richard L. Stroup. 2000. “Media Coverage and EPA Pesticide Decisions.”
Public Choice 102: 297–312.
4
The Media and Markets in the United States
Edward S. Herman
Do the media support markets? There are, of course, a variety of media in virtually every country, as well as major differences in media structures—and media relation-ships with government and markets—between countries. This suggests that media attitudes toward markets might well vary accordingly, both within and between coun-tries. At the same time recent decades have seen an accelerated trend toward dis-placing government-sponsored, and sometimes government-owned and government-controlled, noncommercial media with commercial, advertiser-funded media. A steady trend toward concentration and conglomeration has also been ap-parent, along with more extensive cross-border operations and control of this pri-vate sector of the media (see Bagdikian 2000; Herman and McChesney 1997;
McChesney 1999).
These trends have made media structures and media outputs across the globe more alike than in the past, partly because of the cross-border flows from common sources like CNN, the BBC, and News Corporation and its affiliates; partly because of increased cross-border ownership and alliances; and partly because commercial-ization and competition have tended to homogenize media outputs as media man-agers have sought to reach affluent audiences and please their owners and advertisers.
These trends are making the media within the United States more centralized and the media elsewhere more like the media of the United States. The U.S. media sys-tem is the most powerful in the world, and much of the spread across borders is associated with the parallel movement abroad of its television channels like CNN, its motion pictures, its syndicated television shows, its advertising agencies, and its overseas alliances and ownership. Even where cross-border growth occurs from out-side the United States, as with the Brazilian Globo’s and Mexican Televisa’s effective production and global sale of Spanish soap operas and other programs, this has closely followed the U.S. model (Herman and McChesney 1997, chapter 7; Straubhar 1996, p. 225).
61
Because of the centrality of the U.S. media in the global system, their trend-setting character, and their representation of the most mature dominantly commercial sys-tem in a world in which commercialization is becoming ever more important, this chapter focuses on the relationship between media and markets in the United States.