Elite Framing of Inequality in the Press: Brazil and Uruguay Compared*

Current elite studies argue that inequality produces negative externalities to elites, who may either promote democracy or adopt authoritarian measures in order to shield their interests from the actions of the rebellious poor. This article argues that elite framing of poverty and inequality in the press is a good thermometer of elite public response to such externalities. The press represents a communication tool shared by elites in the state, market, civil society, and, most evidently, the media itself. If inequality threatens elite rule, elites should share their concerns in order to move towards a solution. Since the literature links inequality and elite response, I propose undertaking a comparison of elite public responses to poverty and inequality in two South American cases with opposite records of inequality: Brazil and Uruguay. The article approaches elite framing of poverty and inequality in the press by analyzing opinion pieces and editorials in the main newspapers of both countries. Results invert the expected link between inequality and elite response. Elite framing of inequality in the Brazilian press did not suggest elite concern with externalities, neither an elite turn towards more democracy or authoritarianism. Contrastingly, a few Uruguayan elites did frame the poor as menacing.

[…] The intellectual pressure that people make through newspaper articles […] is valid because one starts redirecting state action.Quite often they [state bureaucrats] are [just] sitting there in their glory, and they think that they are trying to solve problems in the best way possible, but […] they don't know the communities, they don't even know everybody inside the state.Therefore, the most articulated layers of society, the ones that are able to make diverse pressures in an efficient way, are the ones that will be first and foremost beneficiated.(Extract from an in-depth interview with a male corporate leader from Brazil 1 ).
he most popular models of elite behavior state that elites feel threatened by the poor in contexts of high income inequality and that significant political change is likely to follow-up threat perceptions (ACEMOGLU and ROBINSON, 2005;BOIX 2003).With that in mind, several elite studies use survey data and in-depth interviews as metrics of elite attitudes toward the poor (e.g., CLARKE and SISON, 2003;HOFFMANN-LANGE, 2010;HOSSAIN and MOORE, 2005;LÓPEZ, 2013aLÓPEZ, , 2014;;REIS, 2011;REIS and MOORE, 2005).This article proposes a different approach by examining insertions of elites in the editorial and op-ed pages of the press.It tests whether the basic premises of the mentioned models apply to Brazil and Uruguay in terms of the manner in which they frame poverty and inequality.I therefore take debates in the press as a proxy of elite public response to externalities of inequality.
The Brazil-Uruguay comparison contrasts two opposite cases in South America.
Brazil holds a record of high inequalities, while Uruguay poses a more homogeneous society and historically lower levels of income inequality.Moreover, Brazil is characterized by extreme levels of urban violence (BRINKS, 2008;CANO and SANTOS, 2007), which elites tend to attribute to poverty and inequality (REIS, 2000b(REIS, , 2011)), while Uruguay poses low levels of violence and criminality (LÓPEZ, 2013a).In that sense, Brazil and Uruguay fit the two opposing poles described by recent literature.
All things considered, how much elite public mobilization can be expected in Brazil and Uruguay?Contradicting the expectations of the state-of-the-art literature, the results indicate that elites can promote public discussions regarding the poor without any sense of urgency, even in the context of high inequality in Brazil.Concordantly, a greater sense of urgency can emerge out of a context of lower inequality, as seen in 1 The interview is part of a round of in-depth interviews with political, corporate and civil society elites in five countries, conducted by NIED.See Reis and Moore (2005).

T Matias López
(2016) 10 (1) e0002 -3/31 Uruguay.For the most part, elites in these countries use the printed media to urge the economic action on the part of the state.In the discussion section, I will suggest that rather than worrying about externalities, elites ought to concentrate on economic policy as a means to demand the reproduction of a longstanding corporatist and patrimonial relationships with the state.
However, first it is important to understand why the press is a suitable and promising data source for elite studies, and especially to those concerned with elite reactions to the poor.To begin with, elites are aware of the press as a powerful tool for inducing state action.The opening quote in this article illustrates elites' perceptions of their influence through news media.Furthermore, in the quote, a Brazilian corporate leader argues that, if left alone, the state would be incapable of diagnosing actual needs and that elites are likely to benefit from channeling state action toward the "right direction".
Elites are aware of their use of news media; therefore, one can assume it as a channel of inter-elite communication.As such, elite public statements in the press should echo externalities of inequality.The more inequality disturbs elite rule, the more elites should debate the matter in order to find solutions.Furthermore, I argue that it is important to understand how elites frame poverty and inequality.Even if elites perceive the poor as a threat, different frames may lead to different courses of action.
Thus, I will argue that debates in the printed media are a good measure of elites' propensity to react to social constrains.First, I introduce Acemoglu and Robinson (2005) and Boix's (2003) models of elite behavior, to which I later contrast their findings with my case studies.Afterwards, I discuss the intersection between elites and the press and why opinion pieces are a good measure of elite public response to externalities.The "data and methods" and "results" sections follow, in which I present the content analysis of opinion pieces that address poverty and inequality.I end the paper with a discussion of results and a conclusion.

Elites and distribution
The concept of "elites" in elite theory is based on the notion that every society holds a ruling minority, a group that controls and disputes the main sources of political, economic, and symbolic power (LÓPEZ, 2013b).According to Higley and Burton (2006, p. 07), elites are actors capable of affecting political outcomes, both regularly and substantially.Recent studies have modeled elite political action as a function of income distribution, assuming that extreme inequality generates unbearable externalities to elites.
For instance, Acemoglu andRobinson (2000, 2005) argue that the poor are more likely to rebel in contexts of high inequality.In such contexts, the poor pose a credible threat of rebellion and elites tend to either adopt democracy or to improve the scope of an existing democracy, as a means of nullifying such threats.Boix (2003) otherwise argues that elites are more likely to tolerate democracy (assumed as distribution) in contexts of low inequality, when revolutionary tendencies are improbable.
Despite differences in their models, these authors agree that elites are reactive to pressures brought about by the poor, but such pressures are dependent on income distribution.Inequality, they claim, triggers the poor's propensity to rebel against the status quo.Elite reaction differs depending on the scale of the threats and the costs of repression.In some contexts, elites are likely to, as a defense mechanism, endorse a more politically equal society.In others, they counterattack with authoritarian measures.In an equivalent reasoning, de Swaan (1988de Swaan ( , 2005) ) argues that elite attitudes towards distribution depend on the level of threats posed by the poor.His argument is that Western welfare states resulted from elites acknowledging that they could actually prevent the effects of negative externalities by improving the life conditions of the poor.
Therefore, the literature poses the distribution of both political and social rights as outcomes of elite reactions to threatening the poor.Moreover, the literature assumes the level of conflict between elites and the poor as a function of income inequality.
Specifically in Latin America, previous theories of authoritarianism (O'DONNEL, 1973) and current empirical studies (STEVENS et al., 2006) maintain that elites are more likely to endorse repression as a means to prevent distributional demands.
Indeed, Latin American history is filled with cases of authoritarianism, often associated with high inequality.Unequal countries such as Brazil, Peru, and Mexico systematically failed to sustain democracy during the twentieth century.Meanwhile, it is also true that these cases and many others did manage to democratize during the third wave of democratization (HUNTINGTON, 1991), however, not on account of reductions in inequality rates.Brazil, for instance, posed a Gini coefficient of 0.63 at the time of its first presidential election after the military regime.

Elites and the press
Hughes and Prado (2011) argue that Latin American media groups carry political power because they are linked to powerful corporations and families.I would add that the Latin American media is also often linked to the state, be it through state concession, official advertising, or directly through state ownership.As such, news media stands as a key actor in current Latin American democracies (see also BECERRA andMASTRINI, 2009 andSORJ, 2010 for accounts in local scholarship).
Modern-day democracy often functions through mechanisms of mediated deliberation, with regular citizens reacting to, following, and trusting opinion leaders (GASTIL, 2008;MAIA, 2012;PAGE, 1996).This also applies to the elite.State officials are quite often the source of regular news and quite active in opinion journalism, thus encouraging a strong synergy between the press and the state (BENNET, 1990).
Moreover, editors, highly ranked journalists, and commentators usually have a much higher income than the median voter (PAGE, 1996).Thus, they too belong to elite circles and play a role in the political game.In a way, the press is much closer to political and financial power than to its readership.Consequently, it makes sense to use opinion pieces from the mainstream press as a proxy of public elite debate.
However, this is not without problems.For instance, Bourdieu (1996) argues that the media uses experts and intellectuals' opinions for the sole purpose of legitimizing its own opinions.Previous case studies show that newspapers are often inclined to print outside views when such opinions are in accordance with their own (DAY and GOLAN, 2005;PAGE, 1996)-a phenomenon often understood as "gatekeeping" (SHOEMAKER et al., 2008).Thus, the press is not an open arena, but a restricted one.Nevertheless, the same literature points out that newspapers' bias favors governmental and corporate interests, meaning that the voices left out mostly belong to less powerful actors and organizations (PAGE, 1996).
The fact is that elites and the press are strongly interconnected.During military regimes in South America, civilian political elites used the news media as a way to remain influential in the public sphere.Throughout democratization, the media was an important resource for shaping debate and introducing new leadership.As Tomaselli and Teer-Tomaselli (2008) argue, the contribution of the media's critical perspective was often an important source of internal opposition in competitive authoritarian regimes.However, as democratization succeeded in South America, new and old political leaders, corporate leaders, and emergent civil society organizations still found themselves struggling with major social problems.Thus, public debates in the news media remained heated (CASTRO, 1990).
Generally speaking, why do elites choose to print their opinion?Different elites may look forward to printing their opinion for different reasons.For instance, political elites may be looking for support, bureaucratic elites may wish to publicize policy success, business elites may want to criticize "over-taxation", and so forth.They use the press because they realize that other elite members are likely to end up receiving their message.
As for the press, it could be seen as both a source of political power and merely part of the political environment (KUNELIUS and REUNANEN, 2012).Several measurements have pointed to both elite shaping of media coverage and media shaping of elite attitudes (VALENZUELA and ARRIGADA, 2011).Beyond the "who shapes who" debate, it is certain that elites and the press are strongly interconnected.
In addition to the connection between elites and the press, some communication studies also refer to "elite media", meaning those vehicles that carry elite status.A typical example is The New York Times.Although this study relies on data from what could be considered "elite media" (i.e., media that addressees educated audiences), it does not focus on the influence of elite media but elites' discourse in the

Framing
Elite theory often relies on elites' interest as a key variable in order to explain their actions (LÓPEZ, 2013b).What it often misses is that sharing interests implies a complex set of inter-elite communication.Elites present their cases to each other using arguments, ideas, values, and ideologies.As several authors in the frame literature argue, ideas carry political power (CAMPBELL, 2002;HALL, 1989;McNAMARA, 1998;SCHÖN, 1994;SMITH, 1991;SNOW et al., 1986).In this regard, framing is a key dimension of elite public debate, and consequently of elite action.
A major point in the present study is to delimit the main frames of poverty and inequality used by Brazilian and Uruguayan elites.My argument is that such frames are key variables for understanding current distributional patterns as well as the potential of change.Elites react to menaces from below, and mainstream newspapers are an important part of that reaction.
It should be noted that since Goffman (1974) introduced the term "framing", it has been often criticized for its conceptual vagueness (ENTMAN, 1993;SCHEUFELE, 1999).Indeed "framing" is not a very precise concept and this is probably why it is widely used.Frames can be defined as principles that structure the interpretation of reality (REESE, 2007); they are also capable of shaping reality in a certain way (MATTHES, 2012).
With such a definition, frame researchers often advocate for causal effects of normative ideas (i.e., frames) in social policy (CAMPBELL, 2002;SCHÖN, 1994;SNOW et al., 1986) and economic policy (HALL, 1989;McNAMARA, 1998;SMITH, 1991).The connection between frames and policy is an important premise in the present work, as it contends that elite framing of poverty and inequality may affect the patterns of distribution and, ultimately, the scope of democracy.

Why editorials and opinion pieces?
According to Yin (1999), elite leadership of opinion has a causal effect on political behavior.Opinion leadership is routinely performed by elites whose prestige and perceived professional background allows them to generate public debates and to Elite Framing of Inequality in the Press: Brazil and Uruguay Compared (2016) 10 (1)  e0002 -8/31 influence decision making.In the press, the editorial and opinion sections incorporate the role of opinion leadership.
The editorial and opinion sections are meant to present a broad view on public issues, including the newspaper's official positioning, and that of influential opinion leaders.These sections are an important channel of inter-elite communication, where eminent figures in the state, market, and civil society present their views (BROWN et al., 2001).
Nevertheless, it is also true that editorial policies and newspapers' own interests bias and reduce this plurality (CIAFALO and TRAVERSO, 1994;DAY and GOLAN, 2005;PAGE, 1996).At the end of the day, opinion sections do not present a balanced view.However, such bias tends to favor economic and political interests of powerful actors; accordingly, they actually work in favor of elite studies such as this one.
It is in opinion sections and editorials that the framing power of the press finds its ultimate expression (WAHL-JORGENSEN, 2008).Opinion pieces help frame topics within the elite world; thus, they are part of the mechanism of decision making.Yet, opinion sections remain largely understudied in both elite research and political communication research.
One exception is Page's study on the 1992 Los Angeles riots.In this study, the author argues that elite framing of rebellion in editorials and op-ed sections helped weaken the Bush administration.Page (1996) argues that this outcome resulted from high dissatisfaction among elites, who saw the administration as incapable of delivering both profits and social order.The case of the 1992 riots in the US is a good example of externalities of inequality and segregation, and of how powerful actors associate in order to pressure the state to develop solutions.In the present study, I intend to investigate this type of mechanism.

Case selection
This study is based on a two-case comparison: Brazil and Uruguay.Case selection is inspired in the classical most different systems design (PRZEWORSKI and TEUNE, 1970), therefore privileging the opposite dimensions of inequalities in both cases.As mentioned, current models would predict high elite concern in Brazil and low Brazil and Uruguay can easily be considered as opposite cases: the first holds a record of extreme inequality (current Gini of 0.51), and the second is a small country often considered egalitarian by regional standards (current Gini of 0.43 features shared by most Latin American countries (CRUZ-COKE, 2008).Therefore, there are also reasons to believe that frames of poverty and inequality will be similar.
Predictions based on political culture have become extremely unpopular in political science and to a lesser degree in political sociology.As results will demonstrate, similarities in frames in Brazil and Uruguay can be accounted to basic parallelisms in their political repertoire.In the discussion section, I will argue that the data reflects the existence of similar patrimonial strategies in addition to similar political cultures.

Data
Content analysis in this study uses data from the project "Public and Private Strategies Toward Poverty, Inequality and Difference", coordinated by NIED at the Federal University of Rio de Janeiro.Quite often, communication and social science studies analyze how media framing influences public opinion.This is not the goal of this study.I deliberately did not include news or letters to the editor in order to restrict the analysis to intentional, authorial, and direct communications from elites.
The data consists of editorials and opinion pieces written by frontrunners from diverse fields and published in the opinion section (not the reader's section) of the main news publications in Brazil and Uruguay, totalizing 124 opinion pieces in Brazil and 62 opinion pieces in Uruguay.The analysis relies on qualitative content analysis.
Two clipping services were hired in mid 2011 (in Uruguay) and early 2012 (in Brazil); they were asked to survey all opinion pieces and editorials that included the words "poverty" and/or "inequality" anywhere in the text.also generate bias.In this case, adding observations may inflate the relevance of poverty and inequality in elite discourse, but it also allows me to saturate the qualitative analysis.Since I mainly use qualitative discourse analysis, increasing observations was appropriate.Also, Uruguayan news publications usually have strong ideological and political affiliations (see Table 01).Adding Brecha, which is a mainstream leftwing publication, helped me to counterbalance the weight of rightwing publications.
In both cases, the publications chosen were those with the highest circulation and who addressed educated audiences.In this regard, they can be considered as agenda setting newspapers (see Table 01).They also have similar editorial styles, as they include daily editorials and op-ed pages in a single opinion section.In the Brazilian newspapers, this section is located in the first pages and in Uruguay it is usually located later in subsequent sections in the publication.Other than that, they follow a similar pattern, with a couple of editorials and editorial notes and three or four opinion articles.It is noteworthy that the number of opinion pieces is not the same as the number of authors, given that individuals often published more than one piece in that period.In Brazil, apart from the newspaper editors who naturally signed all the editorials, the most frequent author was a journalist who signed nine opinion pieces.
The second most frequent author was a senator from Democratic Labour Party (PDT) (a leftwing labor party) who signed seven opinion pieces.The third most frequent author signed four opinion pieces and is a senator from the Workers Party (the governing leftwing party).In Uruguay, with the natural exception of newspaper editors, most authors published only once.The most frequent author was a politician from Partido Independiente (a small leftwing opposition party) who signed three opinion pieces.Two politicians from Frente Amplio (the governing leftwing party) published two opinion pieces each and one union leader from FUCVAM (a union of housing cooperatives) also published two opinion pieces.

Coding
The primary coding targeted authors' institutional belonging, gender and party affiliation (if suitable).The editorials were excluded from this coding given they were officially signed by the newspaper, not the editor.Among the authors, in both Brazil and Uruguay, men were predominant.In Uruguay, only one female author was sampled, a senator from Frente Amplio.In Brazil, 10 female authors were identified.Most of the Brazilian female leaders belong to the top bureaucracy (ministers, state secretaries, and so on) followed by civil society (mostly journalists).Only one female leader belongs to the political sphere, a senator from the Workers Party (PT).
I divided the authorship into five elite sectors (see Table 02).Congressmen and party leaders were clustered in the category "politicians".The category "bureaucrats" includes non-elected public officials that belong to governmental agencies or international organizations such as the World Bank and the United Nations.The category "businessmen" clusters leaders that belong to the corporate world.In Brazil it included, for instance, the CEO of a chain of language schools (who signed two opinion pieces) and the president of the National Federation of Industries (CNI).In Uruguay, it included the CEO of a public opinion research company, the CEO of an investment company, and the president of a business association.The category "civil society leaders and intellectuals" clusters journalists, professors, economists, free thinkers and in Uruguay, one union leader.This category is not necessarily related to the idea of organized civil society (e.g., NGOs), but to actors outside of the state and market who are active in political life.Finally, the category "editorial" clusters opinion pieces that were signed by the newspaper itself.Content analysis of the opinion pieces was carried out on the following three code categories: a) mentioning of externalities of poverty and/or inequality; b) mentioning of responsible/causal agents and spheres; c) mentioning of means to solve social problems.
Category "a" aimed for statements about externalities of poverty and/or inequality such as criminality, immorality and so on.Category "b" aimed for statements about the social actors or institutions that should do something about poverty and/or inequality.Category "c" aimed for statements about how to properly address poverty, which policies to follow, and so on.A single opinion piece or editorial could be coded using all code categories simultaneously.The code categories follow de Swaan's model of "elite social awareness" (de SWAAN, 1988;de SWAAN et al., 2000).According to de Swaan, elite acknowledgement of externalities, responsibility, and means are accountable for elite general attitudes towards distribution.Table 03   Only a few coding procedures differed among various cases in order to capture context specific data, for example, the mentioning of "blacks" in Brazil and an "egalitarian past" in Uruguay.Although I read and checked coding in all articles in both samples, I counted the Brazilian material with a second coder3 .I used Cohen's Kappa to test for inter-coder reliability in a subsample of 15 (about 10%) randomly selected articles, which were re-coded by the two coders.The level of agreement found was .79,with p < .0005.Since this is a small n study, I could read all opinion pieces in the data set to assure coding accuracy.
Code frequencies were used to build a co-occurrence matrix of subjects.
Naturally, the subjects "poverty" and "inequality" were excluded once articles were sampled through these two keywords.The matrixes were used to build two subject networks, one for each country.The networks reveal how different subjects relate to each other.After building the two networks, I engaged in qualitative speech analysis on those subjects that clustered many other relevant topics.

Results
Recent political history in Brazil and Uruguay accounts for intense elite mobilization in news media.During democratization, for instance, Brazilian elites (including media elites) often debated social issues in the press and television.
According to Castro (1990), media elites had political reasons, rather than market reasons, to invest in such debates.On the one hand they wanted to have a role in the political transition in order to remain powerful after the regime change.Conversely, political leaders benefited from having air time and space in newspapers.
In Uruguay, politicians also used the press as a tool to mobilize political support during the democratic transition (ARNDT, 1994).Overall, it is clear that the press was an important tool in elite-to-elite communication during the democratic transition, and it remained so afterwards.
Currently in Brazil and Uruguay, elites continue to use the press as an arena to convey their messages.Political leaders, public officials, clerics, corporate leaders and social activists are among the key political actors that routinely use the press in order to publish statements.Thus, elite framing of poverty and inequality should have important effects in political life.Moreover, if we assume that externalities play a key role in elite strategy, we should expect the public debate to mirror elite concern.

Public debate on externalities of inequality
Bearing in mind that the theory predicts elite action to follow the effects of externalities, Brazilian elites should be more engaged in debating the effects of poverty and inequality in the press in comparison with Uruguayan elites.
As noted, the Uruguayan data is overrepresented in both time and scope.
Nonetheless, the average mention of poverty/inequality is much higher in Brazil than in Uruguay, as seen in figure 01.Even though the regularity of opinion pieces varies from case to case, the disparity is noteworthy.This discrepancy ultimately shows that Brazilian elites (including media elites) are keen to debate poverty and inequality through the press.
Elite This was, of course, expected since inequality is much higher in Brazil than in Uruguay.Nevertheless, there is no significant difference in the proportion of opinion pieces that framed poverty and inequality as sources of externalities.In Brazil, 10% of opinion pieces and editorials mentioned violence or criminality as a consequence of poverty, 5% mentioned poverty as a threat to development, and 7% mentioned other negative consequences of poverty.In total, 22% of the Brazilian opinion pieces mentioned or implied a threat, none of which were directly related to rebellion.It is important to note that this refers to the proportion of opinion pieces that mentioned externalities, meaning that the articles were not necessarily about externalities.In fact none of them primarily framed inequality or poverty as a threat.
Concordantly, opinion pieces in Brazil did not imply a sense of urgency or crisis.
Brazilian elites mostly addressed poverty and inequality using statistical jargon and their concerns orbited around economic policy.Thus, it is fair to conclude that poverty and inequality were not primarily framed as threats to elite rule in Brazil.This contradicts what we would expect if we assume current models of elite behavior as true.In Uruguay, 8% of all opinion pieces (including editorials) mentioned violence or criminality as a consequence of poverty, 3% mentioned it as a threat to development, and 8% made reference of other consequences such as filthiness, street children, and moral degradation.In total, 24% mentioned externalities, a proportion similar to the one found in Brazil, although relative to a much smaller amount of opinion pieces.
Nevertheless, the few opinion pieces that mentioned threats in Uruguay did imply some sense of urgency or crisis.
Overall, few articles mentioned threats in both countries, as seen in Table 04a and Table04b.In both cases, civil society leaders seemed to be more active in communicating the effect of negative externalities.Civil society leaders were also the ones more inclined to relate poverty and violence, which is a threat that certainly could affect elites' wellbeing.In absolute numbers, politicians were shy in relating poverty or inequality to externalities and businessmen did not emphasize this linkage at all, with Brazilian business elites did not mention poverty and inequality in their opinion pieces, and they generally did not relate them to externalities.Also, Brazilian elites in the private sector did not publically express concern about distribution demands, despite the fact that they were often their main target.Instead, Brazilian elites often framed poverty and inequality as something that is embarrassing for the country.
The following quotes illustrate the framing of poverty as a matter of national pride and as a source of threats.
We are the 7 th power in the world, but we have diseases like dengue, malaria, Chagas disease and leishmaniasis (Editorial, O Globo, March 09, 2011).
The country continues to show off African social rates […] in addition to a huge social inequality it carries in its shoulders (Editorial, Valor Econômico, January 06, 2009).
As we know, Brazil still has very high rates of poverty and inequality, which are incompatible with the county's average income.This is bad for our society, and moreover, [poverty and inequality] also have important consequences in violence and criminality (MENEZES FILHO, Naercio.In: Valor Econômico, August 21, 2011).
Poverty kills.It limits the present and destroys the future.It exterminates entire generations.The dramatic consequences of poverty should [be enough to] justify the implementation of a solid national agenda capable of fixing the problem in definitive terms, but this has not being possible.Fighting poverty is the state's responsibility and a demand from society (GARCIA, Marcelo.In: O Globo, September 08, 2009).
The first two quotes are from newspaper editorials, while the last two are comments by a prestigious economist and a former public official, respectively.As the quotes illustrate, Brazilian elites do debate inequality, often communicating some distress.Yet, elites do not tend to communicate concern about potential consequences of inequality, nor do they relate inequality with any threat to elite rule.The Brazilian elites that mentioned threats (e.g., violence), did not debate the suitability of distributional measures-such as cash transfer programs-or of political reforms that would lead towards improving democracy.They also did not demand more repression, or anything in the direction of limiting democracy.Instead, they addressed vague needs of better education, health and, most of all, of economic growth.In Uruguay, the dominant frame indicated poverty as a problem of social justice, without citing externalities to the non-poor.This dominant frame was present in articles signed by members of the ruling left-wing party, among leaders in civil society, and in editorials.Nevertheless, there was also a counter frame, which related the poor with the loss of values and moral degradation in Uruguay.This counter frame was present in articles signed by leaders identified with traditional parties.Among them, five articles framed the poor as victims of their own culture.For instance, a prestigious lawyer stated: Among us [Uruguayans] there were groups living in poverty.There were always cantegriles [slums] in the suburbs and pueblos de ratas [rat villages].What has changed is not the magnitude of poverty, or its extension, but its culture.[…] [work ethics] was the culture of traditional Uruguay, and what made us a great small country.But that culture is changing.Today we still have poverty, but in addition we also have marginalization.[…] Poverty is a transitory state, marginalization tends to stay.Marginalization is poverty plus cultural difference.When a culture is generated […] then national unity is lost" (FERRERE, Daniel. In: El Observador, June 27, 2009).
Concordantly, an editorial from Últimas Noticias in 2010 stated that "marginality is much more of a cultural problem than it is an economic problem", and that "it [marginality] was not present among the integrated society that characterized Uruguay in the first half of the 20 th century".Therefore, a few Uruguayan opinion pieces mentioned externalities and often had the "quality" of the poor to blame.
Those articles presented a strong sense of urgency or normativity, rather than the technical approach seen in Brazil.In other words, Uruguayan leaders were more vocal about those issues.This should not occur according to current models of elite behavior, i.e., we should not see any relevant indication of elite concern over the poor in Uruguay.These results indicate that lower levels of inequality do not necessarily account for the absence of tensions between the elite and the poor.(2000a, p. 171), is the belief that the state should pilot development by acting as an economic actor.Associated with statism is developmentalism, the belief that state economic planning assures social prosperity (REIS, 2000a, p. 171).This means that elites, and more so Brazilian elites, frame poverty and inequality as something related to the state's role in economic development.
In effect, previous perception research using closed questionnaires has shown high elite support for state-induced growth in both Brazil and Uruguay (see LÓPEZ, 2013a).Content analysis confirms that both Brazilian and Uruguayan elites tend to adopt state-centric frames when addressing poverty and inequality in the press.In such frames, growth plays a major role and distribution plays a minor role, as seen in figure 02.
Figure 02 shows how subjects relate to each other in the opinion pieces, thus providing an image for frames in each case.The subjects of "state", "growth", and "education" are the most powerful ones in both cases, indicating similar statist frames of poverty and inequality.The following quotes illustrate such frames.
In Brazil: The state is essential for regulating the financial system and services of infrastructure (such as energy and telecommunications) […] The state is supposed to execute social policy with the purpose of reducing inequality and poverty and offer quality education to the less fortunate.(NÓBREGA, Nelson da.In: Estado de São Paulo, April 27, 2008).
The country needs to move towards new methods of regulation which elevate production and its egalitarian transfer to the entire population.(YOUNG, Ricardo.In: Folha de São Paulo, March 28, 2011).In Uruguay: This country has no other path but that of a state-controlled capitalism, which should be as equal as it can be; call it social-democracy or not.The name doesn't matter (GATTO, Hebert.In: El País, January 19, 2011).
The participation of the state is crucial for achieving equity and equality in order to lead the economic process […] Reality has shown that the market is not in a condition to fulfill this task.
The Brazilian comments were made by two leaders from the corporate world.
The Uruguayan comments were made by a prestigious jurist and from a leftwing senator, respectively.In both cases, the main frame was typically statist.It presented economic growth as a solution to social problems, and the state as the actor that should promote it.
When opinion pieces made mention of causes or courses of action regarding poverty and inequality, they were almost unanimous in pointing to the state as the main responsible actor, as seen in Table 05a and Table05b.The opinion pieces gave a minor role to civil society, even those authored by civil society leaders.The initiatives of NGOs, for instance, were practically ignored.
Instead, elites demanded the state to induce economic growth and offer education to the poor, interconnecting the roles of growth and education through a statist frame.As one As noted, most elites did not communicate the need for distribution or altering (for better or worst) the scope of democracy.For instance, in both cases, growth was a much more powerful topic than cash transfer programs, minimum wage policy, and affirmative action.Yet, it should be noted that only a few Uruguayan elites advocated for repression as a solution to externalities of poverty.Once again, it is meaningful that arguments on repression appeared in Uruguay and not in Brazil, contradicting current models of elite behavior.
Overall, elite framing of poverty and inequality in Brazil and Uruguay mostly indicated the recovery of longstanding statist values, which minimize the need for distribution or democratization in favor of a stronger state capable of guiding the economy.In the following section, I discuss why this might be the case.

Discussion
Previous studies have demonstrated the effects of inequality on political and social conflict (see MULLER and SELIGSON, 1987;SCHOCK, 1996;SIGELMAN and SIMPSON, 1977) have historically benefited from big states, capable of absorbing and financing elites, offering them power, high income, and stability.Thus, elites may prefer to demand state economic action, rather than to sponsor distribution or democratization as a means to guarantee a safer social environment.
As demonstrated by the results, Brazilian and Uruguayan elites frame poverty and inequality through statist lenses.They communicate the need of a stronger state, rather than a stronger democracy, as a solution to social problems.Meanwhile, statism has been a key dimension of South American authoritarian regimes, as exhaustibly explored in the literature (e.g., CARDOSO, 1979;MALLOY, 1976;O'DONNELL, 1973;PRZEWORSKI, 1991;REIS, 2000b).Does this means that elite public debate is more likely to end up rescuing authoritarianism than it is of generating more democracy?Not necessarily.Though we are witnessing democratic erosion in some countries such as in Argentina and Venezuela (see FREEDOM HOUSE, 2014b) it is safe to assume that the elites in question benefit from democracy in its current shape.In Brazil and Uruguay, elites have not communicated the need to change the system in any direction, nor have they accused other elites of doing so.Nevertheless, it is noteworthy that a few Uruguayan rightwing elites have advocated the use of repression as an answer to externalities of poverty.
Assuming that Brazil and Uruguay are safe from authoritarian drawbacks, the opposite question comes to mind: Could elites' demands migrate towards demands for democratic improvement in the future?It is hard to know.In Brazil, democracy has proven stable, yet flawed.It is hard to picture an effective liberal democracy taking place in the current scenario of extreme inequality.Nevertheless, Brazilian elites have demonstrated their loyalty to the system over the decades.Despite showing little distress, at least elites agree that Brazil would be better with less inequality.Conversely, they do not frame distribution as part of their own interest, indicating weak incentives to pursue equality.

Conclusion
Several elite studies suggest that inequality is important in explaining elite response to poverty (ACEMOGLU and ROBINSON, 2005;BOIX, 2003;REIS and MOORE, 2005;de SWAAN, 1988;de SWAAN et al., 2000).Those authors argue that inequality generates negative externalities to elites, who may turn towards promoting distribution and further democratization in order to shield their interests from the action of rebellious poor as well as from other externalities of inequality.Acemoglu and Robinson (2005), Boix (2003) and several authors before them also realize that extreme inequality may conduct elites toward the other direction, i.e., authoritarianism.Latin American history is certainly a better example of the last trend than it is of the first.
In this article, my argument was that the elite debate in the press is a good thermometer of elite response to social pressures, since it represents a communication tool shared by leaders in the state, market, civil society and most obviously, the media.If inequality poses a threat to elite rule, elites ought to share their concerns with their peers in order to find a solution, be it a democratic solution or an authoritarian solution.
Since the literature claims that inequality triggers elite response, I proposed the comparison of two South American cases with opposite records of inequality: Brazil and Uruguay.In sync with the literature, greater inequality in Brazil should result in the framing of poverty as a source of externalities, the opposite being true in Uruguay.
However, this was not the case.Instead, Brazilian elites mostly framed inequality and poverty using statistical jargon, with no sense of emergency or threat.Uruguayan elites mostly followed the same pattern, but some were very vocal in framing the poor as menacing.This inversion of expectations regarding the Brazil-Uruguay comparison suggests that current modeling of elite behavior may be misleading and that new models need to be developed.
Revised by Cabo Verde Submitted in June 2014 Accepted in April 2015 specifically, I address how elites frame the subjects of poverty and inequality in the Brazilian and Uruguayan press.
Uruguay.Another important feature of the Brazil-Uruguay comparison is that both countries possess a free and independent press (see FREEDOM HOUSE, 2014a).
Brazilian newspapers do not have an official political or ideological affiliation, preferring to be presented as neutral.Only the Estado de São Paulo officially sided with the center-right, Brazilian Social Democracy Party 2 , here assumed as conservative.In Uruguay, elite newspapers are traditionally identified with political parties, but there is no institutional connection between them.El País and El Observador are considered to be rightwing or conservative newspapers, identified with the two Uruguayan traditional Elite Framing of Inequality in the Press: Brazil and Uruguay Compared (2016) 10 (1) e0002 -12/31 parties (Partido Colorado and Partido Nacional), while La República and the weekly publication Brecha are identified with the main leftwing party currently in office (Frente Amplio).Últimas Notícias stands as a neutral publication.

Figure 01 .
Figure 01.Average mention of poverty/inequality in opinion pieces of one Uruguayan corporate leader (out of three) who stated that poverty simply "bothered […] like a pebble in a shoe".
Contradicting current models of elite behavior, Brazilian elites convey poverty and inequality in a blasé and detached manner.They debate income distribution quite often in the press, but do so in a technical fashion using statistical jargon.Therefore, elites in Brazil demonstrate little public concern about potential political consequences of poverty and inequality.Uruguayan elites, on the contrary, should have few reasons to worry about externalities, yet some of them do frame the poor as a menace or as an Elite Framing of Inequality in the Press: Brazil and Uruguay Compared (2016) 10 (1) e0002 -20/31 important source of externalities.This is the opposite of what current theory would predict.What do elites want?As we have seen, current models of elite behavior fail in accounting elite framing of inequality in the press in Brazil and Uruguay.If not externalities, what then drives elites to mention inequality or poverty in opinion pieces?In this subsection, I will demonstrate that Brazilian and Uruguayan elites pose a strong statist drive when framing poverty and inequality.Statism, according to Reis Figure 02.Subject networks wrote: "The quality of basic education is crucial for the Brazilian industrial strategy" 4 .

Table 01 .
Information on news media in Brazil and Uruguay Project Private and Public Strategies in face of Inequality, Poverty and Difference.Note: Brazilian Population is about 200 million inhabitants.The Uruguayan population is about 3.4 million inhabitants.

Table 02 .
Authors Distribution in Brazil and Uruguay Project Private and Public Strategies in face of Inequality, Poverty and Difference.
exhibits three examples of how the material was coded.

Table 03 .
Examples of coding

Table 04a .
Mention of threats related to poverty and/or inequality in Brazil Project Private and Public Strategies in face of Inequality, Poverty and Difference

Table 04b .
Mention of threats related to poverty and/or inequality in Uruguay

Table 05a .
Mention of responsible actors or social sphere in Brazil Project Private and Public Strategies in face of Inequality, Poverty and Difference.

Table 05b .
Mention of responsible actors or social sphere in Uruguay Project Private and Public Strategies in face of Inequality, Poverty and Difference.
One possible explanation is that elites prefer to respond privately to such externalities.Elites have individual resources to protect themselves from violence, which is the main externality identified in this and other studies.Gated communities, private schooling and private security are just some of the resources elites can mobilize in order to insulate themselves from the poor.Meanwhile, they lack the capability that the state has for guiding economic life in their favor.Powerful groups in South America Elite Framing of Inequality in the Press: Brazil and Uruguay Compared (2016) 10 (1) e0002 -24/31