Does Trump read public opinion or form public opinion on the issue of free trade agreements?

(1)

Page 1 of 49

DOES TRUMP READ PUBLIC OPINION OR FORM

PUBLIC OPINION ON THE ISSUE OF FREE TRADE

AGREEMENTS?

Author: Alexander Jordaan

Supervisor: Armen Hakhverdian 24/06/2019

(2)

Page 2 of 49

ABSTRACT

***

1. INTRODUCTION

Even before announcing his candidacy for president from the golden escalators of his apartment building in lower Manhattan, Donald J. Trump was never far from public attention. Yet his ascendance to the leadership of the Republican party, and from there his meteoric rise to the presidency means that his polarizing opinions and positions now not only continue to fill headlines, but also form United States policy outcomes. Running on a platform that included policy attitudes that deviate from longstanding traditional Republican positions, the Trump’s republican campaign has been described by pundits as being uniquely attuned to the policy preferences of its voters.

The study of this described political attunement is of importance in evaluating the current quality of party democracy in the United States (US). Classic work on democratic representation values democracy in terms of the degree of causality between the opinion-policy link. Meaning that democracy is strongest when public preferences are most determinant of policy outputs (Katz 1997; Thomassen 1994). In a system of representative democracy like that which we see in the US, the mechanism that translates public preferences into policy outputs is that of political parties. This caveat adds an important distinction, it means that public preferences do not directly decide public policy, rather through electoral cycles the public vote for the political party that best match their individual preferences (Dahl 1971; May 1978; Burstein 2003; McDonald 2005). In theory the success of a party is based on its ability to align its attitudes to match that of the electorate. It is up to the party leadership to decide which policy

(4)

Page 4 of 49 attitudes will best resonate with voters. It is precisely this caveat in the opinion-policy link that this paper seeks to explore.

This research will look to understand the dynamic relationship between the policy preferences of political parties and the policy preferences of their electorate. It is undoubtedly true that public opinion shapes political leadership through electoral outcomes, yet a central question in the field of democratic representation is the degree to which party leadership shapes political opinion (Gabel and Scheve 2007; Gilens and Murakawa 2002; Ray 2003; Steenbergen et al. 2007). The inclusion of this facet is integral in understanding the broader cues that influence this opinion-policy link. In order to test this normative issue of directional causality we look for empirical evidence in the party competition of the most recent American presidential election. The question being to what degree did changes in policy attitudes by party elites represent an alignment with existing preferences and more importantly, to what degree does it help shape public preferences. This research looks to apply this question to the modern context in asking the question; Is the policy attitudes of the Republican party under Trump a reflection of already pre-existing policy demands by republican voters, or is the Trump Republican party forming the policy demands of their voters?

In order to investigate the causal connection between party policy attitudes and public preferences we will be using a time-series approach (Jacobs and Shapiro 1994). In choosing an issue domain where there has been an explicit change in party policy attitude we will look to see if the new policy position derives from existing voter support or rather changes voter support. Testing the directional causality between public opinion and party attitudes is the primary goal of this paper. The specific empirical case this paper will look to test is the explicit change in communicated policy attitudes towards the policy position on free trade by the Republican party during the run up to 2016 presidential campaign as primary candidate Donald Trump ascertained leadership of the Republican Party. To this end, the following main research has been formulated as. To what degree is the change in Republican Party’s position on Free trade during the 2016 Presidential election a response to existing public opinion on the issue, or has it led public opinion on the issue?

(5)

Page 5 of 49 A further distinction made will be to understand how communicated party attitudes work in a competitive party system whereby people self-identify as being a party supporter. In analyzing this we will not only to measure the degree to which a change in party attitude affects public support among its own supporters, but also the degree to which it changes the opinions of those that self-identify with other parties. This is of importance in understanding how communicated party attitudes affect public opinion differently along partisan lines. To further understand the influence communicated party policy attitudes has on public opinion, we will look to see if there are differences in how those from different age categories, education levels and ideological standpoints are affected by political party persuasion.

1.1 Motivation

My motivation for studying how party attitudes push public opinion is twofold. The first is to add to the understanding of how individuals form their political policy positions. In particularly if they evaluate their membership of political parties through appraising their own positions towards that of their party. Or do we instead see that voters cue their policy positions off their existing favored party. Evaluating the strength of parties in shaping mass opinion is imperative in understanding how different groups can perceive political questions so differently. How those from similar backgrounds can have such stringently different positions undermines the idea of policy positions being formed out of rational self-interests. In looking at the degree to which we take policy cues from political parties indirectly measures the publics commitment to previously expressed values. The degree to which individuals are led by the views of their party has an important reflection on the degree to which we are principled and committed to the value systems we have proclaimed to have in the past. This work looks at the role of ideology as a moderating effect on party policy influence.

My second motivation is to link the strength of political parties in forming public opinion to the issue of polarization. Given that the two-party majoritarian competitive system in the US requires parties to campaign and highlight on differences in policy positions, it is of debate whether the power of parties to form the opinions of their supporters leads to increased polarization and thus a less democratically representative government. If the

(6)

Page 6 of 49 juxtaposition of policy differences that occurs during campaigning cause greater differences of opinion between different party supporters, then the winner of the election will be less representative of the population than had they been elected without a campaign. The question I thus want to highlight is whether the amplification of party policy communication is indeed healthy for democratic outcomes.

1.2 Outline

The research report proceeds as follows. The following section will include a literature review on the normative theories and debates that are of relevance to studying the causal flow between political party attitudes and public opinion. The subsequent section will explain how the case, hypothesis and variables chosen are appropriate to the studying of policy-opinion linkage. Following this will be the research design which will offer explicit clarification for the methods and operationalization of the case study. The succeeding section is the presentation of the outcomes and subsequent findings. The main results of which will be explained in an evaluation that will also offer recommendations for future research.

2) Literature review

2.1 Bottom-Up Democratic Representation

Within the study of the structure of political competition scholars look particularly towards the relationship between public opinion and the actions of political agents (Deegan-Krause, 2010, p.1), again this opinion-policy nexus. The role of politicians as agents of public sentiment is a product of the mechanisms of representative democracy. How representative democracies function is through a principle-agency structure whereby public policy preferences are carried out on behalf of the public by elected representatives. What is significant of this structure is that political representation although a product of public support is independent from it. Though the demand comes from public preference the mandate of responsibility for policy outcomes lies in the hands of political agents. The consequences of this representational form of principle-agency politics is that there can exist differences between the policy preferences of political agents and those that vote them into position.

(7)

Page 7 of 49 Electoral turnover is the mechanism by which the electorate can ensure that public preferences remain the main causal influence of the actions of political agents (Stimson, 1995). The outcomes of elections are a direct way to check that the mandate of those in political power is representative of the preferences of the public. The electorate chooses in theory those that best reflect their own preferences. If those preferences have changed or have not been carried out, then those political agents are voted out of office at the next election. Not only does electoral turnover act as a control mechanism to choose those that govern, but given that agents generally want to stay in their role the ability to remove them means that there is a vested interest for the agent to carry out policy that is in the preference of the public. The democratic mechanism of electoral turnover is the solution to the principle-agent problem that comes from the detachment of political representation (Stimson, 1995). The responsiveness of this causal link between public opinion and policy outcomes is the alignment between the actions of those representing in the same preferred direction as those that vote for them (Dahl, 1971; Ramney, 1954). This one directional understanding of representative democracy inherently requires a structured time sequence (Stimson et al, 1995, p.1). The public has policy preferences, it elects those that best reflect those attitudes, and then these representatives carry out the policies that got them elected or else face removal. The causal flow between public preference and policy outcomes must thus in representational theory be structured in such a hierarchy, meaning there will always be a time difference between public opinion change and the response of political representation (Stimson et al, 1995).

While this bottom up design of representative government means that there is a strong structural link between public opinion and representative policy attitudes, so much so that on a macro level it seems that public opinion will always direct policy (Stimson et al, 1995). Yet the process of representation means that political representatives are not solely a reflection of public opinion. The reality of the matter is that the competitive challenge of an election requires voters to choose between representatives that offer different attitudes on a wide distribution of positions (Stimson, 1995, p2). The choice of who one wants representing oneself is made out of available options. In practical politics there is space for policy dealignment between representatives and their supporters, so long as they remain within the margins of safety of what is supported by the public. A

(8)

Page 8 of 49 result of this is that we can support candidates despite differences on certain policy positions, so long as he or she feels to be the most representative out of the available candidates. Representatives are not static in their policy attitudes and weigh up positions based on their own policy opinions versus those they feel will be appreciated by their supporters. Election turnover occurs for those that stray too far from the second. In sum it is the existence of electoral turnover that ensures policy-oriented behavior emerges that is generally representational of public opinion, differences however are probable.

2.2 Dynamic Representation

Though the causal arrow may generally flow from masses to elites existing literature demonstrates that public opinion should not be viewed as isolated from the influences of political elites (Gabel and Scheve 2007; Gilens and Murakawa 2002; Ray 2003; Steenbergen et al. 2007). The vote-seeking behavior of elected representatives means that they are structurally incentivized to align their own policy preferences with public opinion. The means to this alignment can take two forms, a politician can change his personal policy preferences to reflect existing public opinion, or he or she can attempt to shift public opinion (Stimson et al, 1995). Developments in political science has shown us that the structural relationship between electorate and representatives has a reciprocal process. One in which political actors are not only translators of public preferences but act also as persuaders of public opinion (Deegan-Krause, 2010, p.2; Mair 1997). The evidence for this is seen in studies that highlight the degree to which political actors forego policy work and instead focus a considerable amount of time and effort to trying to shaping political preferences. Political actors put great effort into shaping political social identities and linking themselves to interests, values, cultural milieus and social networks, framing arenas for group-interaction and pushing emotional distances between their own voters and rival parties (Deegan-Krause, 2010, p.2; Colomer and Puglisi, 2005). These attempts by politicians to change the broader social and cultural environment is understandable due to the constructivist nature by which public opinion is formed.

(9)

Page 9 of 49 Traditions of scholarly work have shown that the formation of public preferences are constructed from numerous cues. One of which that is of importance to this paper is the fact that citizens take cues from political elites (Steenbergen et al, p.17). Adjusting their views to the information they receive from party leaders. The fact that voters look to representatives in order to form their policy opinions adds a two-directional element to the opinion-policy causal flow outlined above. Dependent on the persuasion strength of elites on public opinion the causal flow of policy outcomes can become elite driven. Just as how a bottom-up opinion-policy causal flow requires a time for political agents to respond, so does a top-down elite causal flow. If political elites communicate a change in policy direction and seek to persuade the public to follow suit, then we should expect to see a time discrepancy as public opinion aligns to the attitudes of political leaders. The directional flow of the opinion-policy nexus thus can always be identified based on the sequential hierarchy by which either public opinion forms the opinions of political leaders, or political leaders form public opinion.

2.3 Policy Opinion as a top-down process

In order to understand the features of what allows for successful elite driven persuasion it is important to understand how voters form their opinion. An overarching assumption by scholars is that opinion formation is primarily inspired by the motivation to ‘’get it right,’’. Meaning to arrive at an opinion that is most consistent with the individual’s present values, interests and surroundings. In understanding new information citizens process it through motivated reasoning, comparing the new information with pre-existing preferences (Gilens, 2002, p.18). The following section will look to show how elite cues allow citizens to arrive at different preferences dependent on their existing values, interests and surroundings (Deegan-Kraus, 2010; Gabel, 2005, Boninger, 1995). Not only do these factors define the direction of the attitude formation, but they also form the strength of the attitude.

2.4 Social Identification

The first cue that is of importance is an individuals social identification with groups. The groups citizens self-identify with act as important reference groups that individuals look to to formulate how and to what degree they should feel towards an issue. An issue might

(10)

Page 10 of 49 not be of direct relevance to an individual yet if the issue can be perceived as important to their in-group then the attitude can become significant to that person (Boninger, 1995, p63). It seems we interpret politics differently depending on the groups we identify with. For example, evangelical Christians have been shown to interpret candidate’s policy positions through a more ethical lens as opposed to a materialistic frame (Shah et al. 1996; Cramer and Walsh, p.21). Existing social identities act as tools for making sense of the world. For political actors to best persuade public policy they are encouraged to frame themselves as being representative of their interests. Successful persuasion can also be achieved by communication new positions as relating to other issues opinions groups hold. For example, elites have often linked social welfare discourse to immigration in order to convey existing group familiar sentiment on one issue onto interpretation of a current issue (Cramer and Walsh, p.23).

The greater the strength of identification with a group the greater the degree to which an individual will align their attitudes to reflect the same views. Evidence shows that the degree to which an individual self-identifies as being Catholic the greater the likelihood that they share the Catholic Church’s strong stand on the issue of abortion. The greater the allegiance to a self-identified group the stronger the influence the groups leadership has on the formation of the attitudes of its followers (Boninger et al, p.63).

Political parties are suspect to positioning themselves in the frames of social groups. The US Democratic party has consistently communicated a narrative of being the party of civil rights champions and has seen great success as a result (Miller and Schofield, 2008). Performing well among those from minority backgrounds, winning consistently around 80% of the African American vote in recent competitive elections (Stonecash, 2018). These framing of political parties as being representational of certain groups can change over time. Democrats have historically been viewed as the party of the poor, the working class, unionists, urbanites, Catholics, and Jews while Republicans have generally been viewed the party of the wealthy, business executives, small town, rural, protestants (Abramowitz, 2006, p.183; Green et al, 2002). Yet these links to social identities have changed over time, today things like gender, marital status, sexual orientation and religion have risen as salient correlators of party affiliation. Those from traditional family backgrounds and of strong religious conviction now tend to be hold

(11)

Page 11 of 49 Republican policy attitudes while singles, gays, and less religious voters tend to lean towards the Democrat policy attitudes (Abramowitz, 2006, p.183).

Yet while political parties associate with social identities they can in and of themselves perform independently the function of being a social identity. Association with a political party does not need to be an outcome of other group identities but can become an identity itself. Just as one can identify with a religion, class, profession, or ethnicity one can similarly identify as being a Republic or Democrat. The exceptionality of those labels of self-identity is that they are understandably incredibly strong indicators of political policy preferences. Just like other self-identities the strength of the loyalty is salient to the degree to which members conform to the policy preferences of its leaders. When party loyalty is high it expresses a strong influence on citizens policy attitudes. In being independently a social constructed identity means that regardless of different social class, ethnicity and religious backgrounds those that self-describe themselves as ‘’being’’ of a party are incredibly likely to align their own policy positions to those communicated by the leaders of the group they follow.

2.5 Party Competition and Policy Emphasis

We have now acknowledged that like other self-identities such as occupation, sexual orientation, ethnicity, political parties similarly play an independent role as a group that is influential in guiding how its members form their political belief systems. What is however different about forming an identity and loyalty to a political party is that political parties perpetually operate within a background of competition. Because of the structural contestable nature of political elections, the success of a party depends on its ability to outperform the other. This dichotomous facet to politics means that part of absorbing it as an identity promotes a rivalry in opposition to the other identity (Ray, 2003; Walsh and Cramer, 2004; Shapiro, 2004). Much like supporting Barcelona football club leads to an antagonistic relationship with Real Madrid, identifying as Republican implicitly demands an identity that exists in contrariety with the Democrats. The actions and positions of the one group is thus quickly perceived as being a threat to the other (Abramowitz, 2006, p.183). As a result, the greater the attachment to one party the more negative one evaluates the political actions of the other. This is of consequence

(12)

Page 12 of 49 as it means that the cues communicated by political leaders not only influence the political opinions of those that identify with their party, but that they also influence the policy opinions of those identifying with the competing party. The influence effect on the opposition supporters will run counter directional to the persuasive narrative promoted by the other, with degree of loyalty and partisan animosity strengthen the effect.

2.6 Issue differences and consensus

So far literature has highlighted that the cues of political leaders have a causal effect on the policy opinions of the public, but that the degree to which it is believed is dependent on the contextual conditions of the receivers. However, we have yet to consider the fact that the strength of the persuasive narrative espoused is also of importance. Political leaders advocate opinions on a diverse range of policy positions, yet they are not equally emphasized towards the public. The realities of political communication and campaigning mean that parties are selective on the issue areas they choose to differentiate themselves on. Literature on issue importance shows us that political parties emphasize issues for strategic advantage, choosing to highlight and associate with certain policies in the hope of gaining votes (Ian Budge, 2015). This emphasis on certain policy issues can be campaign-specific to gain advantage of specific circumstance, or it can be consistent over multiple campaigns. When dominance of a particular issue carries over from one campaign to the other it can establish a longstanding party-issue association known as issue ownership (Budge & Farlie, 1983, p.25). Whether it is new or long-standing the chosen policy emphasis by political leaders is a result of parties attempting to compete with other parties. Issue salience thus nearly always occurs in order to offer voters a choice between two sides of a debate, emphasizing the differences between parties (Rabinowitz and Macdonald 1989; Budge, 2015).

The degree to which a specific policy position is accentuated by political leaders has shown to enhance the persuasive influence they have on that single issue. If voters play up the importance of an issue to their party those that identify with the party will cue off those messages and align themselves to the reflect the position, they will also similarly value the position as being important (Ray, 2003). The directional influence of the

(13)

Page 13 of 49 leadership effect again occurs along party lines. Those that don’t support the party align their opinions on the issue counter directionally to the discourse of the other. The result of this is that issue salience in political contestation increases polarization on the issue. The success of issue salience as a tool for influencing public policy positions however is conditional on it occurring on issues where parties disagree. Studies show that when there is a consensus in political policy messaging the persuasive influence of the attitude is lower as voters fail to adopt the same level of issue salience. It is the existence of competition that seems to particularly rile voters policy standpoints, without it a voters become permissive (Ray, 2003, p.16)

2.7 Age

In evaluating opinions on policy changes voters not only look to those they identify with, but also look inwardly and reference their current positions with positions they have held in the past. The values and behavioral habits acquired at early stages in voters’ lives act as cues as to how they should behave in the present. Studies have shown that as citizens age, they pick up habitual tendencies, this applies as well to political behavior (Van der Brug, 2010). The party’s voters favor when they are young and first orientating themselves in party politics result from a rationalization of the present political context. Yet as we advance in age, we carry our previous rationalizations of policy and party positions along with us. This can lead to long-term commitments to policy positions even when the rational for them has changed (Van der Brug, 2010; Tilley, 2003.). The longer one routinely supports a party or a specific policy the harder it is for it to be changed. Those that are young have fewer historic reference points by which to distinguish parties and policies are thus more likely to rationalize them on their present-day merits. Those that have become habitually committed to certain positions can continue to give the policy lip-service even when the present context is radically different to the context in which they first formed their policy/party positions (Van der Brug, 2010). The result of this is that younger and older voters seem to react differently to change, with younger voters seemingly more likely to alter their opinions to new information than older voters.

(14)

Page 14 of 49 When it comes to forming political opinions, citizens take cues from a variety of sources. As mentioned above voters listen to that what they hear from the groups they affiliate with, they respond to the personalities and messaging of political elites, they reference their political habits of the past, further they pick up information from media sources, social media, internet, and any other online/offline channel possible. Yet regardless of where information comes from the degree to which is obtained, consumed and cognitively processed leads to differing outcomes.

When it comes to the rationalization different sources of information it is seen that there are variations in citizens performance (Luskin, 1990, p.332). Some voters put considerably more thought into their politics than others and have a more extensive domain knowledge on the issues. Voters use their own cognitive capabilities to value whether policies best server their own interest. Their level of cognitive complexity, ability to consume and evaluate different sources of information, and ability to evaluate across different levels of conceptualization, all affect the way in which they respond to different political cues. Those that are politically sophisticated have been shown to have highly organized and constrained political positions that are more resistant to persuasive appeals. They are also better at spotting and ascertaining policies that are in their own rational interests. Further it has been shown that they are less vulnerable to agenda setting and priming from the media. Further it has been found that they are more focused on specific policy issues and reasoned argumentation than that they are to candidates’ personalities or symbolic discursive narratives (Luskin, 1990; Wyckoff, 1980; Knight, 1985). Variation in the degree of political sophistication a voter has seems to also vary the hierarchy of preferences from which voters receive their political cues from. When political sophistication is low voters become less attentive towards policy and more susceptible to mediums that require low effort. When political sophistication is high policy implications are valued higher and voters are thus less likely to mistakenly support policies that are against their interests (Luskin, 1990 p.332).

In order to be politically sophisticated, one must absorb a certain quantity of political information, have the cognition to retain and comprehend the information, and be motivated enough to make the effort (Luskin, p.334). Given this model of political

(15)

Page 15 of 49 sophistication the variables that have shown to have a causal effect are; interests in politics, education, intelligence, and occupation (Luskin, p,333-336).

2.9 Moderating effect of Ideology

In self-identifying with a group citizen tend to align their views to correspond with the belief systems of the group. Yet as individuals can identify along a wide range of groups, this ability raises the possibility that different group memberships can hold conflicting opinions on the same subject matter. When different identities conflict on the same attitude the directional influence is based on the hierarchy of the citizens commitment to his group identities. If one group identity is valued as being more important than the other, then it will take precedence in the ladder of cue taking (Shapiro, 1994). What does however occur is that the conflicting group messaging does have a moderating effect on the persuasion influence of the other. For example, Christian Democrats are less likely to be supportive abortion than non-Christian democrats. The issue conflict between religious and party identities means that they moderate each other’s influence.

Given we are looking at public opinion on public policy positions the identity that comes first to mind is that of party loyalty. Self-identifying with a political party intrinsically means that the beliefs most influenced will be political ones. Yet there is another political identity that has a strong presence in how Americans identify themselves, that identity is political ideology (Barber & Pope, 2019, p.38; Wyckoff, 1980; Knight, 1985). Just as citizens gravitate towards identifying as Democrats or Republicans, we also see a commitment to the ideological labels Liberal and Conservative. Though ideology and partisanship are deeply connected and highly correlated they both represent different social identities. Partisans loyalty is an expression of attachment to an organization, the policy implications of this is that loyalists are committed to the party and not to specific policies (Barber & Pope, 2019, p.38; Wyckoff, 1980; Knight, 1985). Meaning that voters issue positions are amenably to the cues they receive from the party leaders. It is their support for policies that is flexible, not their support for the party. When it comes to ideology, we see instead a commitment to ideas and principles, in the US these being the ideological camps of Conservatism and Liberalism. Ideological citizens react to new party policy positions in reference to how the correspond with their ideological positions.

(16)

Page 16 of 49 Commitment to an ideological identity means that it is policy opinions that are stable, and party support that is flexible (Barber & Pope, 2019, p.38; Wyckoff, 1980; Knight, 1985).

In the United States there is a strong link between ideology and party, with the Democratic party identifying with Liberalism and the Republican party with Conservatism. On policy positions they are mostly seen in accordance with one another, so much so that the labels are often used interchangeable. This collaborating makes it difficult to separate the influence effect of partisan loyalty from ideological commitments (Barber & Poper, 2019, p.38). Due to ideological policy positions being relatively stable a conflict must arrive from the party side, meaning that we can observe the hierarchy of political identities when a party changes its position on a policy in a direction that is inconsistent with the ideology they associate most with. In such a situation we can see whether voters remain principled towards their ideological commitments or instead choose to realign their attitudes based on partisan loyalties. The degree of commitment to each side will act as a moderating factor on the directional attitude influence political parties have on the public. Disentangling and measuring the role of ideology and party loyalty in forming policy opinions with help further our understanding of their importance in the minds of the American populace.

2.10 Summary of Literature Review

The key finding from the literature is firstly that the influence of public opinion on political representation is dynamic. Direct intervention through elections causes political actors to be responsive to the policy attitudes of the public, vote seeking incentivizes political actors to represent popular policy (Dahl, 1971; Ramney, 1954). Yet political actors also communicate with the public and have a persuasion influence on them. The understanding of who influences who when it comes to forming policy outcomes can be understood through looking at time measures (Stimson et al, 1994). Policy changes that result from bottom-up support must occur in response to changes in public opinion. Policy changes that arise from situations where there is no change in public opinion, but that later directs public opinion can be observed to be elite driven. With the policy change occurring before the change in public opinion. Literature has shown that testing

(17)

Page 17 of 49 this sequential hierarchy is imperative to understanding the directional nature of this influence (Stimson et al 1994). For the next phase of the paper we will look to apply what we have learned from the literature to the current political context.

Secondly the literature outlines the conditional factors that influence the persuasive influence of political leaders. The conditions found from the literature are; 1) self-identification with a political party, 2) degree of party emphasis on a specific policy, 3) age, 4) education, and 5) ideological identification. The purpose of the following part of this paper will to be to apply these theories to a current day policy change. We will look to see if the conditional variables described can be observed empirically.

3) Applied Theory and Case Selection

3.1 Case Selection and Structure of Time Measures

The policy change that we will be specifically studying is the Republicans party’s stance on the issue of free trade agreements. This case study looks to understand if the recent change in party position on the issue under the leadership of Donald Trump was representative of the opinions of his voters at the time or that the change in party position is elite driven and as a result has shaped public attitudes on the topic. In order to empirically test the causal flow, it requires a clear understanding on when exactly the party position on free trade agreements changed, what is also of importance is the relative emphasis placed on the issue of free trade agreements over time and the competing policy position of the Democratic challenger. In order to do so we will look periodically at how the Republican leadership have positioned themselves on the issue over the presidential elections that occurred in the 15 year period between 2003 and 2018. Emphasis will be placed on the period around the 2016 presidential election as this is where we see a radical change on the issue.

3.2 2004 United States Presidential Campaign

The 55th_{presidential election pitted Republican presidential incumbent against} Democratic nominee John Kerry. Both candidates had a strong history of supporting free trade deals. Republican leader Bush was coming out of a first term that involved the conclusion of numerous free trade agreements. With the United States having only three

(18)

Page 18 of 49 free trade agreements when Bush came into office, he added a further eight before his first term ended (Smith, 2016). During the 2004 presidential campaign and after its Bush argued ardently in favor of free markets and vehemently opposed government economic intervention. Bush is quoted as saying “One of the enduring lessons of the Great Depression is that global protectionism is a path to global economic ruin.” (Smith, 2016, p.586). Senator Kerry similarly had a background that was in favor of free trade agreements, having voted for NAFTA and legislation that led to the formation of the World Trade Organization (Debate USC, 2003). During Bush’s first term he had also supported prescribing China a Most Favored Nation status as well as giving more authority to the administration to Fast Track new negotiations on new trade agreements. During the campaign there seemed to be little emphasis on the specific topic of free trade. Among both the Democrats and Republicans there seemed to be an elite consensus in favor of free trade agreements. An issue related to free trade that did gain some attention was the topic of outsourcing jobs. When asked a specific question on this during a debate Bush answered not by criticizing free trade but instead by arguing that he would improve the public education system ‘’to gain the skills necessary to fill the jobs of the 21st century’’ (CBS Debate, 2004). Senator Kerry responded to the same question by saying ‘’Outsourcing is going to happen. I've acknowledged that in union halls across the country. I've had shop stewards stand up and say, will you promise me you're going to stop all this outsourcing, and I've looked them in the eye, and I've said, no, I can't do that. What I can promise you is that I will make the playing field as fair as possible; that I will, for instance, make certain that -- with respect to the tax system, that you as a worker in America are not subsidizing the loss of your job.’’ (CBS Debate, 2004). These quotes are demonstrative for the role the policy issue of Free Trade Agreements played in the 2004 presidential elections. Neither candidate emphasized the issue as they both candidates shared similar views on the policy, each preferring to talk about other issues in which they differed more substantially. Despite both parties being normally in favor of Free Trade Senator Kerry was slightly more critical than President Bush, criticizing NAFTA several times during his campaign (Fletcher, 2011).

(19)

Page 19 of 49 The 56th_{US Presidential Campaign occurred in the throes of the financial crisis meaning}

that much of the campaign rhetoric focused on the US economy and thus the topic of free trade agreements came more to the forefront than it had done in previous elections (Lewis-Beck, 2009). While Democratic candidate Senator Kerry had been reluctant to talk on the issue of trade in the 2004 elections the new Democratic candidate for the presidency Senator Obama was more responsive and critical on the issue. During the Democratic primary contest Senator Obama vowed to renegotiate NAFTA unilaterally and threatened to leave it if additional environmental and labor protections where not emplaced (Lewis-Beck, 2009). Once Senator Obama’s won the primary however he backed down from his opposition of NAFTA and in response to Republican claims that he was protectionist he amped up his enthusiasm for free trade (Chukwumerije, 2009). The communicated message from the Obama camp was one of balance, ’’It is absolutely critical that we engaged in trade, but it has to be viewed not just through the lens of Wall Street, but also Main Street, which means we’ve got strong labor standards and strong environmental standards and safety standards’’ (CNN debate, 2008). The Democratic message was one that was overall supportive of free trade agreements promoting it in combination with greater protections for workers rights. This exert from one of Obama’s speeches summarizes the democratic message; ‘’Trade has been a cornerstone of our growth and global development. But we will not be able to sustain this growth if it favors the few’’ (Berlin Speech, 2008; Jacobs, 2010).

Republican candidate Senator Johm McCain offered little different to the views of his predecessor in his staunch support of trade and free trade agreements. With a history of being a steadfast supporter of free-trade Senator John McCain entered the presidential race with a 100% pro-free trade voting record and made sure to communicate this during his campaign (CATO institute, 2008). Senator McCain highlighted free trade as a key part of his solution to the economic crisis, in his campaign plan named ‘’Bold Solutions for Economic Prosperity’’ he outlined the need to lower barriers to trade (Amedee, 2008). During a campaign debate he disavowed protectionism; ‘’Every time the United States has become protectionist and listened to the siren song that you’re hearing partially on this stage tonight, we’ve paid a very heavy price… Free trade should be the continuing principle that guides this nation’s economy’’ (NBC Debate, Oct

(20)

Page 20 of 49 2007). The message from the Republican camp could not be clearer, you could not get more pro-free trade than John McCain.

3.4 2012 Presidential Election

President Obama’s second crack at the presidency in 2012 represented no radical departure from his previous platform. Under his first term as president he pursued the pro-free trade plan he had been elected on, working with congress to pass free trade agreements with South Korea, Colombia and Panama. Further he spent his time in office continuing to negotiate on the Trans Pacific Partnership, expanding it to include Canada and Mexico (Meltzer, 2016). In his campaign rhetoric for the 2012 elections he continued with his measured support of free trade. Balancing imploring the positive effects of free trade agreements yet also stressing the importance of playing by the rules set by the trade agreements. In particular China was frequently mentioned as being culprit to unfair trade practices due to entering the World Trade Organization (WTO) under terms that Obama felt where not tight enough (OnTheIssues, 2012). The communicated campaign message of Obama could again be described as moderately positive on free trade.

Republican candidate Senator Mitt Romney’s views on free trade agreements can be seen to be rather similar to that of Barack Obama. Less enthusiastic on trade than McCain Romney pointed out the importance of cracking down on China infringing the WTO rules and also argued that free trade process often requires unions and suppliers to make painful adjustments. ‘’The case for trade makes good economic sense--trade improves the wages and standard of living for the average citizen. But trade can disrupt and devastate those individuals directly affected.’’ (OnTheIssues, 2012). Although Romney measured his support of free trade he continued to advocate it as integral for American economic growth, during the debates he criticized Obama for not doing enough ‘’European nations and China over the last three years have opened up 44 different trade relationships with various nations in the world. This president has opened up none.’’ (OnTheIssues, 2012).

(21)

Page 21 of 49 The Presidential Election of 2016 represented a change in narrative by the Republican party when it came to the issue of free trade agreements. The issue also enjoyed greater salience in the political debate and was a topic of greater discussion than in previous elections (Noland, 2016). During the Republican Primaries campaign (July 2015 – July 2016) nominee Donald Trump repudiated the traditional Republican support for free trade, this policy position allowed him to distinguish himself from other Republican politicians (Liu & Jacobsen, 2018). In particular he focused on the role of free trade agreements in outsourcing jobs, ‘’Our country doesn’t win anymore. We used to win, we don’t win anymore. We lose on trade’’, ‘’We’re going to bring jobs back from Japan, we’re going to bring jobs back from China, we’re going to bring, frankly, jobs back from Mexico’’ (CBS Debate, 2016).

Once winning the Republican Primary in July 2016 he carried his same disfavor of free trade and free trade agreements into the campaign to beat Hilary. While Romney in his 2012 presidential bid argued for a change in trade practices Romney still held a fundamentally pro-trade view of the world (Oreskes, 2016). Trump however campaigned on radically changing the approach to free trade and free trade agreements, his are views best summed up in his inaugural address on January 2017 when he stated that ‘’ From this day forward, a new vision will govern our land. From this moment on, it's going to be America First. Every decision on trade, on taxes, on immigration, on foreign affairs, will be made to benefit American workers and American families. We must protect our borders from the ravages of other countries making our products, stealing our companies, and destroying our jobs. Protection will lead to great prosperity and strength.’’ (CBS Debate, 2016). Trumps campaign platform railed against the North American Free Trade Agreement and called the creation of the World Trade Organization ‘’another Clinton disaster’’ (OnTheIssues, 2016). The Trump campaign communicated a narrative that all previous free trade agreements had been bad for the US economy, and that the only he could negotiate new and fairer ones. Candidate Trump’s strong critique of past free trade agreements represented a strong and evident change in communicated narrative by the Republican party.

(22)

Page 22 of 49 In a role reversal the Republican party became the party that was most skeptical on free trade agreements. While the Republican narrative changed dramatically the communicated position of Hilary Clinton remained relatively similar to her predecessors. The Democratic primary elections however pitted her against Senator Sanders who held more negative views on the issue, not believing in ‘’unfettered free trade’’ (OnTheIssues, 2016). The close primary contest between Senator Sanders and Secretary Clinton meant that the Democratic narrative on trade agreements was more split on its benefits. When Hilary won the candidacy in July 2016 the Democratic campaign narrative absorbed many of the positions of Senator Sanders but also somewhat returned to the customary stance seen in previous elections (Noland et al, 2016). That being that the benefits of trade are complex and that some free trade agreements need to be renegotiated, but that the benefits of free trade agreements are largely positive. ‘’I have a very clear view. We have to trade with the rest of the world. We are 5 percent of the world's population. We have to trade with the other 95 percent. And trade has to be reciprocal. That's the way the global economy works. But we have failed to provide the basic safety net support that American workers need in order to be able to compete and win in the global economy.’’ (MSNBC Debate, 2016).

Table 1: Summary of campaign positions of free trade agreements per party Presidential

Election Significant dates Political Party Party nominees General attitude towards free trade agreements

2004 Primaries: Jul 2004- Jul

2004

Presidential: Jul 2004- Dec

2004

Republican George Bush Good Democrat John Kerry Good

2008

Republican John McCain Good Democrat Barack Obama Good

2012

Republican Mitt Romney Good Democrat Barack Obama Good

(23)

Page 23 of 49 2016 Primaries: Jul 2016- Jul

2016

Republican Donald Trump Bad Democrat Hilary Clinton Good

3.2 Operationalization of Variables and Hypotheses

Now that we have an outlined how American party leaders have communicated on the issue of free trade agreements it is time to see if these positions are reflective of the views of the American public. What is key is that we will look at how certain conditions within the public have influenced public support on the issue, these conditions this paper will test derives from the existing literature explained above. We will now look to operationalize these factors into variables.

Dependent Variable (DV): Attitude towards free trade agreements [Good/Bad]

The dependent variable for this study is the public opinion on free trade. With the question posed to respondents being: In general, do you think that free trade agreements between the U.S. and other countries have been a good thing or a bad thing for the United States? [Good/Bad/Don’t Know-Refused]. With this question being posed to the public across several time periods the implications of this variable are twofold. Firstly, this paper seeks to understand if the policy attitudes communicated by party leaders during elections was indeed representative of the public opinion at the time. The second application of the variable is in studying changes in elite policy opinions. In examining the directional relationship between public attitude and elite attitude we can understand if the change is a response to existing public opinion or has a causal effect on public opinion. There are thus two hypothetical outcomes possible and this empirical study will thus highlight the causal ordering of the two.

H1: Mass driven Hypothesis: A change in public opinion on the issue of trade has led to a change of party policy attitude in the same direction

H2: Elite driven Hypothesis: A change in party communicated attitudes towards free trade has led to a change in public opinion in the same direction

(24)

Page 24 of 49 The independent variable that we will be looking at is how party identification effects public opinion on the issue of free trade. Do those that affiliate with the Republican party have different views than those that affiliate with the Democratic party. In order to study this we have respondents have been asked: In politics TODAY, do you consider yourself a Republican, Democrat, or independent?. Given the possible difference party identification may have on views on free trade we will again look to see if each party is representative of the views of their own supporters over time. Again, we will look to see the causal nature of public opinion on changes in party free trade opinion, only now the independent variable will look specifically at the responses of those that self-identify with the party.

H3: Party Total Effect Hypothesis: Party identity will have a significant effect on attitudes towards free trade agreements of respondents.

H4 Partisan mass driven Hypothesis: Changes in public opinion on free trade agreements within those that self-identify with a political party will cause a response by their Party in the same policy direction.

H5: Partisan Elite driven Hypothesis: Changes in a communicated party attitudes on free trade agreements will affect public opinion differently based on which party citizens identify with.

Within-subjects condition 1 (C1): Age

Within those that self-identify with a specific party we will look to see if those from different age groups hold different opinions on the matter. The inclusion of this will help us to understanding if age is a factor in how party supporters perceive the issue of trade. In particular it is of value if the causal relationship of a party attitude change is elite driven, if this is the case then studying this variable will shed light on how the influenced of elite communication could affect different age groups differently. Information on voters age is garnered through directly asking respondents on the dependent variable their exact age. For the simplicity of analysis two categories have been created, supporters below the age of 30 (Young), and those above the age of 30 (Old).

(25)

Page 25 of 49 H6: Changes in a communicated party attitudes on free trade agreements will affect public opinion of party supporters differently depending on their age.

Within-subjects condition 2 (C2): Education

Just as with age we want to look if supporters within different parties hold differing views dependent on their political sophistication. Theory shows us that political sophistication is a causal effect of several variables that include; interests in politics, education, intelligence, and occupation (Luskin, p,333-336). For this paper we will be using educational attainment as an indicator for political sophistication. Just as with age we are particularly interested in the differences between how those from different educational backgrounds are influenced by the changing of policy opinions by political leaders. We thus hope to specifically study how the top down causal effect of party persuasion affect party supporters differently depending on their educational attainment. Respondents asked on their opinion on free trade agreements also answered the question: What is the highest level of school you have completed or the highest degree you have received? For the simplicity of the research this paper looks compares those that are college educated or higher with those without a college degree.

H7: Changes in a communicated party attitudes on free trade agreements will affect public opinion of party supporters differently depending on their education.

Within-subjects condition 3 (C3): Ideology

The purpose of this variable is to test the hierarchy of cues from which voters choose. Here we pit self-identification to a party against self-identification with an ideology. Looking to see if there is a difference in how citizens that affiliate with both identities respond when their political party changes opinion on trade. What will particularly insightful is the response of those who hold a ideological position that is inconsistent with the direction of the party policy change. In such a case we can value the hierarchical value of the two identities in the degree to which they are important to how citizens form policy opinions. The way in which we know a respondent’s ideological adherence is through the question: In general, would you describe your political views as... [Extremely liberal, liberal, moderate, conservative, extremely conservative, don’t know/refused]. For the statistical analysis extreme liberal and liberals are bucketed

(26)

Page 26 of 49 together as are extreme conservatives and conservatives. The analysis looked at the opinion on free trade of those with different ideologies within those that support the Democratic party and those that support the Republican party.

H6: Changes in a communicated party attitudes on free trade agreements will affect public opinion of party supporters differently depending on their self-described political ideology.

(27)

Page 27 of 49 In order to visualize the hypotheses the causal flow structures are outlined below.

Bottom-Up causal flow

H4 Partisan mass driven Hypothesis: Changes in public opinion on free trade agreements within those that self-identify with a political party will cause a response by their Party in the same policy direction.

Top-Down causal flow

H5: Partisan Elite driven Hypothesis: Changes in a communicated party attitudes on free trade agreements will affect public opinion differently based on which party citizens identify with.

Top-Down causal flow with condition variables

Measuring the effect of party influence on public opinion among varying conditions: Age (H6), Education (H7), and Ideology (H8).

(28)

Page 28 of 49

4) Research design

Now the variables and hypotheses have been outlined, the following section will describe the specifics of the research design. The purpose of this study is to observe the direction of the causal flow between public attitudes and party attitudes on the policy issue of free trade agreements in the United States, and to further examine the conditions that affect this influence relationship.

4.1 Methodology

The method this paper uses is a longitudinal measurement of the main variable over the 15-year period between 2003 and 2018. Repeated surveys with the question In general, do you think that free trade agreements between the U.S. and other countries have been a good thing or a bad thing for the United States? [Good/Bad/Don’t Know-Refused] was asked from different significantly sizeable sample groups. The poll was carried out by Pew Research center a total of 17 times during the examined period, being asked on the following dates: [December 2003], [March 2004], [July 2004], [December 2006], [November 2007], [April 2008], [April 2009], [November 2009], [November 2010], [March 2011], [March 2014], [May 2015], [March 2016], [August 2016], [October 2016], [April 2017], and [May 2018].

When the independent variable of party self-identification is added to the longitudinal study of attitudes towards free trade agreements, we split the measurement across two parties. We observe a measurement of attitudes towards free trade over time among those that self-identify as a Republican, and we observe a measurement of attitudes towards free trade over time among those that self-identify as a Democrat.

These measurements of attitudes on free trade among the supporters of the two parties will then be compared to the communicated policy positions of the corresponding political parties. This information we have as a result of the analysis done above, where we have defined the general policy position on free trade agreements [Good/Bad] for each of the two parties during the time periods surrounding the 2004, 2008, 2012, and 2016 United States presidential elections. In comparing public sentiment to party sentiment across these election periods we can observe whether Party positions on the issue is representative of the views of their supporters or not.

(29)

Page 29 of 49 To understand the directional causal relationship of who is cuing who it is necessary to look at a moment of party policy change on the issue. Such a moment was observed to have occurred during the 2016 presidential election as the Republican party under Trump changed the communicated attitude on free trade agreements from that of being generally good to generally bad. Such a change presents a methodological opportunity to study the causal relationship between public opinion and party policy. In order to disentangle the directional nature of the link we use apply the results of the longitudinal measurements on republican opinion on free trade agreements. Through looking at the sequential hierarchy in which either Republican supporters became more negative on trade and in delayed response the Republican party aligned itself to this view, or that it was the Republican Party that first changed its view and then through communicating their new view changed the opinions of Republican supporters. The hierarchy in which what occurs is consequential to defining the directional cause.

This longitudinal measurement of attitudes towards free trade agreements between Democrats and Republicans is then further broadened to include the conditional variables of age, education, and ideology to observe how these factors may alter how party supporters feel on the issue of trade. If the Republican party policy change is seen to be elite driven then measurements of these within-measures variables will observe which groups altered their opinions the most to align with their party.

Sources and data collection

The nature of the empirical material used comes from Political Survey data carried out by Pew Research Center. The sample groups for each survey derived from a combination of landline and cellular random digit dial. The samples thus included respondents from around the country allowing for a representational understanding of the entire population, limited only to those that own a phone. Sampling was performed in accordance with PSRAI specifications (Pew Methodology). As many as 7 attempts were made to contact every sampled telephone number and calls were staggered over times of day and days of the week to maximize the chance of reaching potential respondents (Pew Research center). All of the surveys conducted included questioning on all of the variables listen above.

(30)

Page 30 of 49 Here is an overview of the preliminary results

(DV) In general, do you think that free trade agreements between the U.S. and other countries have been a good thing or a bad thing for the United States?

Date Good thing Bad thing Don’t Care/Refused Total number of respondents

Dec-03 275 277 262 814 Mar-04 719 664 320 1703 Jul-04 463 354 186 1003 Dec-06 638 557 307 1502 Nov-07 503 618 278 1399 Apr-08 489 778 235 1501 Apr-09 809 797 425 2031 Nov-09 777 714 509 2000 Nov-10 410 577 268 1255 Mar-11 692 625 187 1504 Mar-14 1935 1018 382 3335 May-15 1158 670 174 2002 Mar-16 1105 931 218 2254 Aug-16 497 419 84 1000 Oct-16 581 583 148 1312 Apr-17 783 600 118 1501 May-18 830 471 202 1503

As shown the survey data offered significantly large sample groups, with only the first survey measuring below a thousand respondents.

Validity and Limitations of Method

The choice for a longitudinal study is that it allows for an observational study of the same variable over time. In doing so we can measure policy opinion at the individual level and compare it with the opinions of parties at a national level. The data we have on policy opinions at an individual level derived from Pew Research Center can be seen to be of strong validity. With large sample sizes the non-profit organization is widely acknowledged as a strong source for evidence-based research (Pew Research Center, 2018). Where the validity research is weaker is in the comparison to measurement of

(31)

Page 31 of 49 party attitudes. The analysis of communicated policies performed in the previous section observed the communicated messaging of party leaders through looking at their responses in speeches, debates and interviews during their campaigning. This form of analysis is limited in that it can only provide a general view of the communicated party positions on the issue. The nuances of the different party messages are not considered in this paper, the outcomes of the analysis are limited to describing party position on the issue of free trade agreements as either finding it good or bad for the US economy. As a result the findings of this paper will only give a general indication of if the party’s position is representative of the views of its supporters or not.

Further this paper is limited in that it only looks at the relationship between the public and national politics. While it is to be assumed that some citizens receive political cues from sub-national and supranational sources, this research looks at what it believes to be the primary venue for political interaction which is that with national political parties. Our assumption for this is based on the fact that policy towards free trade agreements is a national one and is of less relevance in local or regional politics.

5) Results

(32)

Page 32 of 49 The first results of the longitudinal study observe the general public opinion on the question: Do you think free trade agreements between the U.S. and other countries have been a good thing or a bad thing for the United States? [Good/Bad]. The measures of the dependent variable are then placed in the political context of campaign periods.

Findings 1: Public attitudes towards free trade agreements over time

Looking at these results we can observe that public opinion on free trade agreements has generally been positive among the American public in the 15-year period between 2003 and 2018. The cumulative opinions on free trade agreements for the total period showed a positive 55,2% approval for free trade agreements. Though this is a majority of the population, it does show that there is by no means a consensus on the benefits of free

(33)

Page 33 of 49 trade agreements for the United States. What we do observe however is that not only has the average support for free trade agreements been generally positive over the years but we also observe a trend that suggest that approval for free trade agreements is increasing. The least positive period measured was in March 2008 when only 39% of respondents saying they thought free trade agreements where good, the highest approval rating for free trade agreements was observed in the last survey conducted, with 64% of respondents approving of free trade agreements in April 2018.

These findings are important in support for the bottom-up understanding of policy. The generally positive support for free trade agreements is reflected by the generally positive positions the Democratic and Republican parties have taken during the past four election campaigns. Now we will look to see how public opinions on the issue are possibly different depending on affiliation with a particular party.

Findings 2: Public attitudes towards free trade agreements over time

(34)

Page 34 of 49 Table 2.2: Support for Free Trade Agreements among Self-identifying Republicans Table

2.2:

Table 2,3 Support for Free Trade Agreements among Republicans versus Democrats Significance testing between the Parties

(35)

Page 35 of 49 (DV) In general, do you think that free trade agreements between the U.S. and other countries have been a good thing or a bad thing for the United States?

Date Democrats Republicans Partisan Difference

Pearson Chi-Square Significance

Dec-03 54% 49% 5% 0,385 Mar-04 48% 58% 10% 0,003*** Jul-04 53% 63% 11% 0,014*** Dec-06 51% 58% 7% 0,073 Nov-07 47% 48% 1% 0,811 Apr-08 35% 48% 13% 0,000*** Apr-09 57% 49% 8% 0,021*** Nov-09 59% 51% 8% 0,015*** Nov-10 51% 32% 19% 0,000*** Mar-11 57% 53% 4% 0,296 Mar-14 69% 65% 4% 0,082 May-15 65% 61% 5% 0,139 Mar-16 64% 43% 21% 0,000*** Aug-16 67% 41% 26% 0,000*** Oct-16 72% 29% 43% 0,000*** Apr-17 72% 37% 35% 0,000*** May-18 76% 47% 30% 0,000***

Independent Variable Effect

The results of our independent variable (Party self-identification) on public opinion on free trade agreements show us that opinions on the issue are often split along partisan lines. While there have been periods where supporters of different parties have held significantly different views on the issue, they have never done so more than in the past few years. Of the 17 surveys conducted 11 surveys observed a statistically significant difference between how Democrats and Republicans viewed the issue. Of those surveys five of them where conducted between 2016 and 2018, it is during this period that we observe the greatest difference between along partisan lines, with differences in views

Does Trump read public opinion or form public opinion on the issue of free trade agreements?