Income Inequality and State Parties: Who Gets Represented?

Abstract

Recent studies of representation at the national and state levels have provided evidence that elected officials’ votes, political parties’ platforms, and enacted policy choices are more responsive to the preferences of the affluent, while those with average incomes and the poor have little or no impact on the political process. Yet, this research on the dominance of the affluent has overlooked key partisan differences in the electorate. In this era of hyperpartisanship, we argue that representation occurs through the party system, and we test whether taking this reality into account changes the story of policy dominance by the rich. We combine data on public preferences and state party positions to test for income bias in parties’ representation of their own co-partisans. The results show an interesting pattern in which underrepresentation of the poor is driven by Democratic parties pushing the more liberal social policy stances of rich Democrats and Republican parties reflecting the particularly conservative economic policy preferences of rich Republicans. Thus, we have ample evidence that the wealthy, more often than not, do call the shots, but that the degree to which this disproportionate party responsiveness produces less representative policies depends on the party in power and the policy dimension being considered. We conclude by linking this pattern of influence and “coincidental representation” to familiar changes which define the transformation of the New Deal party system.

Keywords

parties representation economic inequality public policy party organizations and interest groups

Introduction

Until recently, empirical studies of representation in the United States have told a rather positive story about the responsiveness of elected officials (and of the policy process generally) to the broad preferences of the public. Following in the tradition of Miller and Stokes’ (1963) classic “Constituency in Congress,” and after some methodological corrections (Achen 1977, 1978; Erikson 1978, 1981), a good deal of research concluded that despite the public’s lack of interest, information, and real attitudes, there exists considerable correspondence between districts’ ideological orientations and the roll call voting of members of Congress (Ansolabehere and Jones 2010; Ansolabehere, Snyder, and Stewart 2001; Canes-Wrone 2015; Canes-Wrone, Brady, and Cogan 2002; Erikson and Wright 1980, 2000, 2009; Hall 2015; Hollibaugh, Rothenberg, and Rulison 2013; Miler 2016; Sulkin, Testa, and Usry 2015; Wright and Berkman 1986) and even state legislatures (Birkhead 2015; Hogan 2004, 2008). Beyond legislative voting, studies of state policy (Erikson, Wright, and McIver 1993) and national policy making over time (Erikson, MacKuen, and Stimson 2002) add up to a large body of work which indicates that indeed, even without attentive, well-informed citizens, elected officials are reasonably responsive to public preferences. Interestingly, as the methods and data sources have developed in quantity and sophistication, the evidence for responsiveness appears to be even stronger (Ansolabehere and Jones 2010). In general, the thrust of this, and the larger body of work of which these studies are a part, was that representative democracy in the United States was in pretty good shape. A lot of the time, the public seem to get the policies they want from their elected officials.

Virtually, all of this work treated the pubic as a single entity as though it were a single force in pulling and pushing politicians. However, scholars then began to look not at just “public opinion” but to ask whose public opinion matters. Gilens (2005, 2011, 2012) mapped the preferences of citizens on over 1,800 policy questions and matched these to policies passed by the national government. He reports an overwhelming pattern in which national policies are much more likely to change consistent with the preferences of the affluent (90th percentile in income) than those of the median wage earners or the poor (10th percentile). Larry Bartels (2008, 2016) comes to the same conclusion in examining the responsiveness of roll call voting of Senators and members of the House of Representatives to different income groups: they are quite responsive to their better-off constituents (upper third), much less so to middle-income constituents, and virtually not at all to the least well-off of those they supposedly represent. Rigby and Wright (2013) identified a pattern that largely mirrored those of Gilens and Bartels—but across states, in particular focusing on the policy positions adopted by state parties compared to the preferences held by their state’s citizens, whereas Flavin (2012) found the same pattern looking at the policies states enacted. In studies that systematically contrast the influence of elites, experts, organized interests, and public opinion, the public comes out way behind in terms of its relative influence (Gilens and Page 2014; Jacobs and Page 2005). From the perspective of democratic theory, the message of these studies is quite pessimistic: rather than “government by the people,” it seems to “be government by and for the rich.” Before accepting this troubling conclusion about the character of American politics, we feel that we should focus on the 800-pound gorilla of contemporary American politics that all of these studies pretty much ignore: political parties.

Bringing Parties into the Picture

A problem with the existing treatments of income group differences in the unequal representation studies is that they treat these groups as though they are more or less unified actors. But we know that the primary dividing line in American politics today is partisanship (Layman, Carsey, and Horowitz 2006; Poole and Rosenthal 1997; Shor and McCarty 2011). Enns (2015), for example, in arguing that the apparent dominance of the rich in the existing studies is much less severe than it seems, makes the point that the differences in preference between income groups are much less than the huge gulf in issue stances between Democratic and Republican partisans. Furthermore, we see significant differences in income between party members—with more affluent Americans more likely to identify as Republican and poorer Americans more often Democrats (Stonecash 2000). This raises the possibility that patterns in parties’ representation of different income groups may simply be a consequence of the differences in the income profile of their core supporters.

This leads to a conjecture, which is consistent with our traditional (i.e., New Deal) view of the parties that the affluent are represented by the Republican Party (with its reputation for catering to the interest of the wealthy and business interests), whereas the Democratic Party represents the interest of working people and the poor. If this is, in fact, the case, then party representation could go a long way to mitigate the apparent problem of powerless middle and working classes. This is a conclusion by Brunner, Ross, and Washington (2013) in which they conclude that electing Republicans results in overrepresentation of the affluent, and conversely, electing Democrats results in overrepresentation of those living in poorer precincts in their study of representation in the California Senate. In an analysis of roll call votes, Rhodes and Schaffner (2017) also find the preferences of lower income Americans reflected in Democrats’ votes, which they characterize as a populist model of representation.

The idea that elected representatives stand for the preferences of their partisans rather than the average constituent is not new (Bafumi and Herron 2010; Brady, Han, and Pope 2007; Clinton 2006; Hopkins and Stoker 2011; Hussey and Zaller 2011; Wright 1989b). The important point here is that while economic class differences are one significant characteristic of American politics, they pale in comparison with the ideological and partisan difference between the parties. Furthermore, if each party simply represents the median member of their own party constituency, they may represent different income groups as a consequence—not a cause—of the party-centric representation process.

This conceptualization follows from the long-standing view of the role of parties: they provide coherence in a system that constitutionally disperses power and accountability; for many citizens they are the primary heuristic for making sense of politics. Parties do this by aggregating policy preferences into recognizable and familiar brands with which citizens can identify (Huddy, Mason, and Aarøe 2015; Snyder and Ting 2002). And the distinctiveness, and therefore importance, of party differences has greatly increased at both the elite and mass levels (Hetherington 2001; Iyengar and Westwood 2015; Poole and Rosenthal 1997).

This set of findings leads us to appreciate the error in the popular vision of democratic elections, moving policy to the preferences of the median voter (Downs 1957). Rather, particularly in the current polarized period, elected officials are especially attuned to their “base” rather than to the preferences of overall constituencies. This suggests two things. First, representation occurs via partisanship—enacted policy is a function of party dominance in government rather than of overall public opinion (Grumbach 2018). And second, preferences of the affluent are achieved with Republican victories, with lower income groups perhaps receiving policies they prefer under Democratic control of government. If so, and given the potential (not to mention tendency) for party control to thermostatically change with policy mood (Erikson, MacKuen, and Stimson 2002; Wlezien 1995), we might believe that, on balance, the whole electorate is not that badly represented.

Does this view allay concerns about the dominance of the affluent and the impotence of the poor in the policy process? It might, but only if certain conditions are met. Keeping in mind that there are still large income differences within the parties, representative party platforms could occur if (1) the influence of income groups within the party is equal (e.g., Republican policy positions respond to the preferences of both poor Republicans and wealthier Republicans), or (2) “coincidental representation” (Enns 2015; Wlezien and Soroka 2008) is achieved because there are no significant differences in the policy preferences of different income group (e.g., poor Democrats and rich democrats agree on a policy such as the minimum wage). In the latter case, even if one group has a more significant political voice, all groups get policies closer to what they prefer.

As we look toward the within-party analysis, there are three theoretically possible outcomes. One is equal influence in which there are income differences in preferences among co-partisans but each has proportionate influence on state party ideology. In the case of the Democratic Party, which is expected to have a larger proportion of lower income adherents, that would mean those people would realize better representation by electing Democrats. A second possible outcome is coincidental representation where although there are differences in policy preferences between income groups in the aggregate (Gilens 2015; Rigby and Wright 2013), it may be that there are quite minimal differences within the parties with Democrats, rich and poor, being consistently liberal and Republicans, rich and poor, being conservative. As stated above, in terms of practical policy outcomes, it would not matter that much if one income group has more sway in determining party policy if the income groups agree.

The third possibility is continued unequal representation. It may well be that within the parties, income group preferences diverge, and one group’s preferences are much more important than others. An example of this might be that affluent Republicans dominate within their party, while lower income groups prevail in fashioning Democratic Party policies. That would be consistent with New Deal images of the parties with Republicans favoring those with higher incomes and business interests, while Democrats attend to the concerns of lower income working people. Alternatively, the preferences of the affluent might be reflected in both parties, only with the expectation that the preferences of affluent Democrats would not equal that of affluent Republicans. Such an outcome would be consistent with the logic Bartels (2016, 257–68) offers for the disproportionate influence of the high-income groups of his analysis: the affluent are reported to participate more, to be much more likely to contact elected officials, to contribute to candidates and parties, and finally to share the class-based worldview of those serving in office who seldom share the working-class origins of those they represent.

As we will see, evidence shows that state party policies are a combination of coincidental representation and dominance of the well-to-do. This combination varies by party and policy dimension, and the pattern we find helps fill out a story of the larger patterns of partisan change of the last several decades.

Data and Measures

Assessing inequality in party responsiveness to affluent versus poorer co-partisans requires generating measures, at the state level, of mass preferences by income group, as well as ideological positioning of the parties on the same issue dimensions. We describe our method for generating these data in the following sections.

Preferences of the state electorates

For the mass public, we are using data from the National Annenberg Election Surveys of 2000 (N = 58,373) and 2004 (N = 81,422). These data include a large number of survey respondents from the 48 continental states, with all but five states having at least 500 respondents across the two survey years.¹ The limitation of these Annenberg survey data is that respondents were interviewed across a series of rolling, cross-sectional surveys in which different sets of (somewhat overlapping) questions were asked of each cross section. To generate measures of policy preferences across these surveys, we pooled the surveys for 2000 into one data set and the surveys from 2004 into a second merged data set. For each pooled data set, we identified a set of commonly asked policy items in each survey. These included 30 items from the 2000 surveys and 28 items from the 2004 surveys—capturing respondents’ positions on a range of issues: tax and spending levels, health and social welfare, abortion, gay rights, energy and environment, campaign finance, and legal/ethical concerns. As all respondents were not asked all questions, we imputed the missing data using imputation by multiple chained equations (see Royston 2005), which predicts survey participants’ responses based on their preferences among items they were asked.

We used factor analysis to extract two dimensions from the Annenberg surveys: one set of issues taps the dimension of economic policy preferences and the other taps the social issues. The former includes respondents’ preferences on tax policy, health care, social welfare, and K-12 school funding. The latter captures responses to questions on abortion, death penalty, gay rights, school prayer, and gun control. The specific items used are listed in Table 1 of the supplemental appendix. These two dimensions were estimated separately for 2000 and 2004, using the first dimension in each case (with eigenvalues from 1.58 to 3.22). The second dimension eigenvalues are all below 0.50.² The factor scores were then standardized (mean = 0, SD = 1.0), allowing us to combine the 2000 and 2004 survey files for an overall working N of 136,056.

Table 1.

Income and State as Predictors of Citizen Ideology (R²).

Regression model	Economic issues index	Social issues index
Income terciles (dummies) alone	.016	.004
Place (46 state dummies) alone	.008	.031
Income and place	.024	.035

Source. 2000 and 2004 National Annenberg Election Surveys. Individual level N = 131,069. All coefficients and full models are significant at <.001.

Our approach here is to analyze responsiveness in terms of two of the large dimensions of conflict in American politics: economic and social issues. While a lot of work examines the American policy space as one large liberal-conservative continuum at both the elite (Poole and Rosenthal 1997) and mass levels (Tausanovitch and Warshaw 2013), there remain good reasons to examine the separate dimensions of economic and social policy (Carmines and D’Amico 2015; Feldman and Johnston 2014; Lupton, Myers, and Thornton 2015; Treier and Hillygus 2009). This is particularly important when considering differences among income groups since Americans with higher incomes tend to hold more conservative preferences on economic policy, but more liberal stances on social policy. These differences are evident in Figure 1 which shows the results of fitting a simple linear model for income (grouped into low, medium, and high terciles). In trying to sort out how income groups are represented, this cross-cutting income-issue gradient could obviously confound the analysis if we used a one-dimensional measure of ideology.

Figure 1.

Predicted ideology from simple income dummies with 95% CIs.

The second reason for looking at the economic and social dimensions separately is they have different relative relationships to individual income and geographic location (Rigby and Wright 2011). More of the variance in economic preferences is within states—with smaller differences across the states. In contrast, social policy preferences are more strongly correlated with location—with state of residence telling you much more about someone’s likely policy preferences than knowing their income would. These differences are summarized in Table 1 which compares the predictive power of income and place (state of residence) for the two dimensions. This is important because differences in the variances of explanatory variables at different levels of analysis can alter the relative importance of those explanatory factors (Wright 1989a). Although not directly relevant to the analysis here, these differences imply that social issues may play a larger role in accounting for the behavior of state electorates than is the case at the individual level where analysts report that economic issues are clearly dominant in voter decision-making (Ansolabehere, Rodden, and Snyder 2006; Bartels 2006).

Estimating income-based differences in preferences

Research in this area typically uses one of three ways to estimate the policy preferences of different income groups across the states. One is to separate respondents by reported income. If we want to compare say high-, medium-, and low-income groups’ preferences, respondents can be divided by their answers to the survey income questions. The problem here is that the distributions of responses to income questions in surveys do not conveniently break into nice thirds, necessitating the use of weights in regression equations to account for the different sizes of the income groups (Bartels 2016, chap. 8; Bhatti and Erikson 2011). A second approach accounts for this unequal distribution by allocating respondents proportionally into equal thirds through use of a weighted average of the mean preferences (by state) across the income groups.³ Developed by Bhatti and Erikson (2011) and used by Rigby and Wright (2013) in earlier analysis of responsiveness of state parties to their constituents as a whole, this approach offers a more straightforward interpretation of income group effects.

A third method, developed by Gilens (2005), adopts a different approach by using a regression model to predict preferences by income and then uses those predicted values in later models testing responsiveness. This approach works better in this analysis where we want to estimate not just the average preferences of each state’s income terciles but the distinct preferences of Democratic and Republican identifiers within terciles. Gilens’ innovation was to estimate issue preferences by first regressing preferences on income (and income squared) and then for his analyses to estimate from that equation the preferences for the 90th, 50th, and 10th percentiles on individual issues. In our case, we calculated respondents’ incomes in terms of percentiles by state, assigning the mean percentile (p) to respondents according to their answers to the nine income categories of the Annenberg surveys. Then for each state, we regressed preferences (economic or social) on p and p². Using these equations, we estimate (impute) the preferences for the 17th, 50th, and 83rd percentiles to represent the median preference of the three terciles. These are highly correlated with the alternative measures of income group preferences discussed above: r >.95 in each case. In analyses described later, we repeat this process for Democratic and Republican partisans separately.

Measures of state party positions

We conceptualize party ideology as the ideological positioning that stems from the collective issue stances of parties’ candidates in each state. This approach captures the policy profiles parties present to voters, which we believe is better than relying on platform statements which are seldom seen and which have little impact on what the candidates actually stand for or what elected officials do in office. We used actual campaign positions rather than inferences of positions based on roll call votes. This bases the measures on the positions of losers as well as winners, and has the advantage over roll calls of not being affected by distortions due to differences in partisan control, agenda setting, or other legislative procedures that vary across the states.

We pooled data collected by Project Vote Smart (PVS) that administers the candidate questionnaires—formerly called the National Political Awareness Test (NPAT) and currently termed Political Courage Test (PCT). We used responses collected over the period 1995–2005 from candidates for Congress, governor, and state legislatures (N = 18,467). Aggregated over the years, these data provide healthy sized candidate samples on which to estimate party ideologies, ranging from 115 in Louisiana to 968 in New Hampshire with a mean N of 373. Like the Annenberg surveys, some of the questions asked in the PVS questionnaires vary from state to state and over time. However, there was a good deal of overlap across the 255 surveys which we draw on. We based our measures on the 87 questions that were asked of at least half of the respondents. Missing responses were imputed using the same chained imputation strategy employed for the mass surveys.

After imputation, we factor-analyzed subsets of the items (25 economic policy items related to redistribution, social spending, and safety net and 20 items relating to abortion, gun control, civil rights, and school prayer) to achieve summary party positions on economic and social issues. We used the first factor score for each set of issues scales so that higher values indicated more liberal policy positions (as they do in the public opinion measures used in these analyses). This first dimension captured 78% of the variance in economic policy positions (eigenvalue = 7.03) and 65% of the variance in social policy positions (eigenvalue = 4.53) across candidates. These scores varied significantly across candidates, with economic policy positions ranging from −3.42 to 1.92 and social policy positions ranging from −2.01 to 1.89. As expected, Democrats scored higher on each scale than did Republicans—gaps of more than 1 standard deviation on each scale—capturing the well-established polarization in party politics. For additional information on these measures, see Rigby and Wright (2013).

Results: Unequal Representation

The first thing we do is replicate earlier results which showed that the poor have no direct impact on party preferences using the pooled Annenberg survey data from 2000 and 2004 (rather than only the 2000 data used in Rigby and Wright 2013).⁴ In the analysis of income group representation, we examine two aspects of the relationship between citizens and the state parties. One is alignment which asks simply whether the relative positions of the parties across the states coincide with that of the preferences of the different income groups. Alignment measures do not make any claim about influence, just relative congruence. The second measure, influence, assesses groups’ impact on party positions controlling for the preferences of other groups. That is, when the groups disagree, which seems to have the greatest impact on the positioning of the parties? The distinction between what we are calling here alignment and influence is central to our analysis below of income group representation within the parties. Where a group has alignment with policy outcomes, roll call voting patterns, or, in our case, state party ideologies, but without influence, we have what Enns (2015) calls “coincidental representation.”

Figure 2 shows the results of replicating the earlier analysis with the expanded data set. Plotted in the figure are the estimated regression coefficients (solid dots) and their 95% confidence intervals (solid lines) that result from regressing the scores of state party economic issues and social issues on the estimated preferences of the income group thirds in the states. The general pattern is the same as reported by Rigby and Wright (2013) who only relied on the 2000 Annenberg data. The lowest income groups never have any independent effect on either economic or social party positioning. The highest income group clearly sets the agenda for both parties in the states on economic issues. Similarly, the affluent are the only group with independent influence on the Democratic parties on social issues. The one exception to the rule by the rich that we find is the influence of Republican Party on social issues where middle-income groups’ preferences appear to have more influence than either the poor or the affluent. These analyses fully replicate the pattern of the earlier analysis, but they do not consider the respondents’ own party identifications nor the reality that elected officials are more likely to represent their supporters than those who oppose them that this article takes up.

Figure 2.

Income group effects on party ideologies: combined NAES 2000 and 2004 Surveys.

Now that we have determined that working with the larger data set does not change the patterns of unequal income group influence reported earlier, we can turn to the focus of this article: bringing in parties in the electorate. The importance of distinguishing between co-partisans and the rest of the state electorate is largely born out in Table 2 which shows the results when we regressed party economic and social ideologies on Democratic, Republican, and independent identifiers’ economic and social policy preferences. Both parties’ positions on economic policy are clearly responsive to their own partisans: the Democratic Party positions reflect the preferences of Democratic partisans in the states and the same for Republicans. This is further evidence of parties acting strategically—playing to the base, while the preferences of the other party and independents have no independent impact. The pattern we found for the parties’ positions on social issues is a bit different. On the Republican side, it is clear that only the Republican identifiers in the states have influence on Republican Party positions. For the Democratic parties in the states, in contrast, it appears that the social policy preferences of both parties matter, although the coefficient for Democratic identifiers is larger.

Table 2.

Effects of Partisans’ Preferences on State Party Economic and Social Ideologies.

Group preferences	Economic issues		Social issues
Group preferences	Democratic parties	Republican parties	Democratic parties	Republican parties
Democrats	1.104** (0.501)	0.662 (0.592)	1.317*** (0.265)	−0.011 (0.266)
Independents	0.480 (0.327)	−0.126 (0.385)	−0.232 (0.381)	−0.122 (0.381)
Republicans	−0.224 (0.271)	1.377*** (0.320)	0.950*** (0.290)	1.488*** (0.290)
Constant	0.196 (0.169)	−0.363* (0.200)	0.429*** (0.134)	−0.117 (0.134)
Observations	47	47	47	47
R ²	.231	.384	.789	.620

Note. Standard errors in parentheses.

p < .1. **p < .05. ***p < .01.

Income and Party

This brings us to our main research question: Does consideration of party affiliation in the electorate mitigate the disproportionate responsiveness of parties and elected officials to the affluent? The hopeful hypothesis is that the economic distortions in representation disappear. Instead, as argued by Brunner, Ross, and Washington (2013), electing Democrats results in an overrepresentation of the poor and electing Republicans in an overrepresentation of the affluent—with differences in the composition of party bases explaining away much of the differential responsiveness that challenges our notion of a democratic policymaking process.

We repeat the earlier analysis of responsiveness to income group preferences, but next we examine this income gradient within each party. We began by estimating the preferences for each income tercile among Democratic identifiers and then separately among Republican identifiers. We included independents who lean toward one party or the other within these estimates since these “leaners” are likely part of the primary constituency to which parties wish to appeal, and leaners have been shown to be more like the partisans they lean toward than pure independents (Keith et al. 2009). To calculate these estimates, we ran regressions with economic (and then for social preferences) as the dependent variable, with a dummy for party (independent leaners included), the respondent’s income percentile in her state, and her income percentile squared.⁵ We then used the coefficients from these regressions to predict values for the policy attitudes of the low-, middle-, and high-income groups within each party and state (values of .17, .50, and .83 for the income percentile variables).

Then, within parties, we regressed the state party ideologies on income group preferences—estimating both alignment (bivariate association) and influence (marginal association controlling for the opinion of the other two income groups). The key results are presented in Figure 3, and Table 2A in the supplemental appendix provides the full regression models. The results are clear for economic policy: the positions adopted by state parties reflect those of the wealthier income groups. We first consider responsiveness conceptualized as alignment; the preferences of the middle-income partisans are positive, but the coefficients are smaller than the alignment coefficient for the top income group, and the results are not statistically significant for those with the lowest income. In terms of influence, the impact of the affluent is most evident, in terms of both stronger slopes and significance. For economic policy positioning, it appears to be a case of the elites listening to the elites within their own party rather than the median voter in their primary constituency. The extent to which other income groups get what they want from their party on economic policy appears to be instances of “representation by coincidence” (Enns 2015; Gilens 2015; Gilens and Page 2014; Wlezien and Soroka 2011), although we have more to say about that below.

Figure 3.

Within-party income group effects on state party positions.

The pattern for social policy is the same for Democrats. In the figure (lower left plot), we see that only the preferences of the top group are statistically significant when we consider independent influence when the income groups disagree. So for social policies of the Democratic parties in the states, the conclusion is the same, elites listening to elites within their own party—disregarding the more conservative social policy positions of lower income Democrats. These findings challenge the hopeful notion that bringing parties into the study of representation will explain away much of the differential responsiveness by income, as well as hopes that at least one party represents the poor. Instead, Democratic parties align more closely with the wealthier members of their own party. Democratic parties pursue policies supported by less affluent Democrats, which occurs primarily through a process of coincidental representation in which the preferences of poor and middle-income Democrats are shared by the wealthier members of their party.

The exception to our pattern of parties aligning with their wealthiest co-partisans is found for Republican parties’ social policy platforms. Earlier analysis, which omitted the role of mass partisanship (Rigby and Wright 2013 and replications with these data presented in Figure 2 of this paper), found that Republican social policies were most responsive to the middle class in the state electorates. Here, we see that this carries through; it is the middle-income Republicans who have the most influence on Republican parties’ social policy positions. This seems to support Thomas Frank’s (2004) argument that the Republican Party has grown in adherents by appealing to the relatively conservative social values of the middle class. In the multidimensional Republican world, the state parties reflect the preferences of rich Republicans on economic policy, but middle-income Republicans on social policies.

This simple pattern is entirely consistent with the long-term growth of Republicanism in mass public. As long as the New Deal coalition held and the chief differences between the parties centered on economic issues, the Republican Party, with its reputation of favoring the affluent, remained a minority party. With the extension of the party differences to race and then social issues from the 1960s and beyond, more and more of the electorate found that, on balance, they favored the Republican Party (Layman and Carsey 2002). It is hard to see how the story of partisan change of recent decades can be told without acknowledging the role of social issues in the Republican gains among the mass electorate.

Our analysis of responsiveness within the parties does not change the fundamental story of the positioning of parties and legislative candidates: most of the time the high resource folks are attended to, with the important exception being that the Republican parties bend to their middle-income partisans’ conservative preferences on social issues. We argue that this is central to their increased numbers in the electorate and that this increase is reflected in the election of more Republicans; furthermore, that has served the goals of affluent Republicans on economic policy.

Party Control and Misrepresentation

We finish with a look at the potential implications of these findings on the conservative or liberal policies parties may pursue when in control of their state government. Grumbach (2018) shows that state party control is strongly correlated with policy liberalism (Democratic control) or conservatism (Republican control) in recent years. We are also in a period of relatively high levels of unified control (National Conference of State Legislatures 2018; New York Times 2012) making the assumption that representation primarily occurs when one’s preferred party is in control of the state government.

Yet, our findings suggest that even when one’s preferred party is in control, it is the preferences of the wealthier members of that party that are reflected in the policies parties pursue. This is most problematic in the cases where affluent and poor Democrats (or affluent and poor Republicans) hold distinctly different preferences. If partisans of different income groups are in agreement, then greater attention to wealthy co-partisans would have little impact (Enns 2015; Wlezien and Soroka 2011). On the contrary, if there are significant differences between the preferences of income groups within the parties, then an imbalance in responsiveness means that the poor have parties—that once in control of state government are not offering up the kinds of social or economic policy they want—not even for their most loyal co-partisans.

This means that understanding the consequences of each party’s prioritization of their wealthiest co-partisans requires assessing the degree of consensus or disagreement between income groups within each party. We did this by estimating reduced-form quadratic regression models in which social and economic policy preferences for Democrats (and then for Republicans) are regressed on individual-level measures of income percentiles. These results are presented graphically in Figure 4, which plots the estimated relationship of economic and social policy preferences of Democrats and Republicans by income (along with their 95% confidence intervals). We find a distinct asymmetry between the parties—one that yields interesting predictions for who is likely to get represented when their party takes control of the state government.

Figure 4.

Income and partisans’ policy preferences.

The income gradient is most steep—indicating differences in preferences across income groups—for Democrats on social policy issues and for Republicans on economic policy issues. This suggests that when Democratic parties gain control of state government, Democrats of all income levels should be pretty much satisfied with economic policies pursued. However, there is likely to be a bigger gap between the social policy preferences of low-income Democrats and the social policies their party pursues when elected. This is in sharp contrast to poor and affluent Republicans who are more likely to agree on social policy issues. Instead, the income gradient among Republicans is on economic policy. Therefore, when Republicans take control of state government and pursue policy preferences held by their wealthy co-partisans, their lower income partisans may be in general agreement with the social policies they pursue. Yet, the economic policies they pursue are likely to be quite conservative—reflecting the more extreme economic policy preferences held by affluent Republicans, despite the more moderate views held by other Republicans, not to mention Democrats in the state.

This analysis suggests that the consequence of political inequality operates through the party system and also differs for the two parties. For Democrats, political inequality is most likely to take the form of Democratic parties pursuing the more liberal social policies favored by wealthy Democrats. For Republicans, we are more likely to see the consequences of political inequality in the parties’ efforts to enact the more conservative economic policy proposals and policies favored by wealthy Republicans.

Conclusion

We extend the line of research which asks whether the quality of representation depends on the incomes of voters. We use the states as research laboratories, examining the responsiveness of the state political parties on the dimensions of economic and social policies. We take the analysis beyond previous work which suggested that indeed the state parties are more responsive to the affluent, with the exception of the Republican parties which seemed to heed middle-income group preferences on social policy. Most importantly, we extend this work by situating the representation process in the very partisan—and increasingly polarized—political context in which policy making (and therefore representation) occurs. Some scholars have suggested that when partisanship is taken into account, the inequalities in responsiveness to income groups are much less important; partisan differences swamp the much smaller differences in preferences between income groups. And others argue that lower income Americans are actually better represented by Democratic parties—making the question one less of who elected officials represent and more tied to inherent competition between two class-based political parties. Together this research supports the optimistic notion that much of what looks like income-based political inequality may simply represent partisan differences in the electorate.

By differentiating between Democratic and Republican members of state constituencies, we did find that state parties are much more responsive to the preferences of their own partisans than those of independents or the other party’s supporters, and those differences are substantial just as we would expect in the era of polarized parties. However, when we examined the impact of income groups within the parties, we see a continued pattern of inequality—in which parties are better aligned with the more affluent members of their own party—on both parties’ stances on economic issues and the Democratic parties’ stances on social issues. The well-off in the parties call the shots. The one exception is that the state Republican parties appear to be influenced by the social preferences of their middle-income supporters. Thus, while there is no denying the huge impact of partisanship, indeed leaving it out of the analysis is like ignoring the proverbial 800-pound elephant in the room—even with partisanship squarely in the analysis, the parties reflect the issue stances of their better-off partisans, who represent an even smaller segment of the state electorate than suggested in earlier research.

Does this matter for the types of policies parties are likely to pursue when in office? That depends on the issue and party. Consider a state in which Democrats win control of state governments, and the parties are heeding the preferences of their most affluent supporters. On economic policy, we would expect the party to pursue policies more liberal than the median voter in the state, but generally reflective of low-income, middle-income, and affluent Democrats. This deviation from the preferences of the median voter is still important, but less troubling from a normative perspective than one based primarily on income. However, on the social dimension, the party’s policies are likely to reflect the much more liberal preferences of affluent Democrats—leaving less well-off Democrats, not to mention Republicans, underrepresented and likely dissatisfied with direction of social policies under Democratic regimes.

Of course, as we explained above, Republican control would result in more conservative social policies, but these would not be drastically out of line with Republicans across the income spectrum. However, on economic policies, which reflect high-income earners’ preferences, we expect the policies to be more conservative than those desired by middle- or low-income Republicans, not to mention far out of alignment with what Democrats want.

Thus, we have ample evidence that the wealthy, more often than not, do call the shots, but that the degree to which this disproportionate party responsiveness produces less representative policies depends on the party in power and the policy dimension being considered.

It strikes us that the Democratic parties, as the traditional party of the working class, do a pretty good job of providing coincidental representation to their lower income partisans on the central long-term defining dimension of economic policy. It is on the newer social issue dimension where responsiveness to the affluent Democrats’ relative liberalism departs from the more conservative attitudes of lower income Democrats. In contrast, Republicans seem to have gained adherents with appeals to social conservatives, and the party seems to be in relative consensus on conservative positions on social issues. But on economic issues, the party continues to reflect the more conservative economic preferences of affluent Republicans. In short, we can see how the greater salience of social issues has placed strains within the Democrats, and at the same time while providing a basis for Republican appeals to low- and middle-income citizens whose support in turn makes possible the party’s adoption of the conservative economic policies favored by affluent Republicans.

Supplemental Material

Appendix-_Wright_and_Rigby – Supplemental material for Income Inequality and State Parties: Who Gets Represented?

Supplemental material, Appendix-_Wright_and_Rigby for Income Inequality and State Parties: Who Gets Represented? by Gerald C. Wright and Elizabeth Rigby in State Politics & Policy Quarterly

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: The authors are grateful to the Russell Sage Foundation and the National Science Foundation for their financial support as well as to Project Vote Smart for providing us with data from their NPAT survey of candidates. We profited from comments from presentations of earlier versions of this work at the Midwest Political Science Association (2016) and the University of Houston Conference on Representation in the State Legislatures (2016).

ORCID iD

Gerald C. Wright

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Gerald C. Wright is professor of Political Science at Indiana University, Bloomington. His research focuses on representation, public opinion, political parties, and state politics.

Elizabeth Rigby is an associate professor of Public Policy and Public Administration, and of Political Science at George Washington University. Her research examines the interplay of politics, policy, and social inequality in the contemporary United States.

References

Achen

Christopher H.

1977. “Measuring Representation: Perils of the Correlation Coefficient.” American Journal of Political Science 21 (4): 805–15.

Achen

Christopher H.

1978. “Measuring Representation.” American Journal of Political Science 22 (3): 475–510.

Ansolabehere

Stephen

Jones

Philip Edward

. 2010. “Constituents’ Responses to Congressional Roll-Call Voting.” American Journal of Political Science 54 (3): 583–97.

Ansolabehere

Stephen

Rodden

Jonathan

Snyder

James M.

Jr.

2006. “Purple America.” Journal of Economic Perspectives 20 (2): 97–118.

Ansolabehere

Stephen

Snyder

James M.

Jr. Stewart

Charles

III . 2001. “Candidate Positioning in U.S. House Elections.” American Journal of Political Science 45 (1): 136–59.

Bafumi

Joseph

Herron

Michael C.

2010. “Leapfrog Representation and Extremism: A Study of American Voters and Their Members in Congress.” American Political Science Review 104 (3): 519–42.

Bartels

Larry M.

2006. “What’s the Matter with.” What’s the Matter with Kansas? Quarterly Journal of Political Science 1:201–26.

Bartels

Larry M.

2008. Unequal Democracy: the Political Economy of the New Gilded Age. New York: Russell Sage Foundation.

Bartels

Larry M.

2016. Unequal Democracy, 2nd edition. Princeton: Princeton University Press.

10.

Bhatti

Yosef

Erikson

Robert S.

2011. “How Poorly Are the Poor Represented in the Us Senate?” In Who Gets Represented?, eds. Enns

Peter

Wlezien

Christopher

. New York: Russell Sage Foundation, 223–246.

11.

Birkhead

Nathaniel A.

2015. “The Role of Ideology in State Legislative Elections.” Legislative Studies Quarterly 40 (1): 55–82.

12.

Brady

David W.

Han

Hahrie

Pope

Jeremy C.

2007. “Primary Elections and Candidate Ideology: Out of Step with the Primary Electorate?” Legislative Studies Quarterly 32 (1): 79–106.

13.

Brunner

Eric

Ross

Stephen L.

Washington

Ebonya

. 2013. “Does Less Income Mean Less Representation?” American Economic Journal-Economic Policy 5 (2): 53–76.

14.

Canes-Wrone

Brandice

. 2015. “From Mass Preferences to Policy.” Annual Review of Political Science 18 (1): 147–65.

15.

Canes-Wrone

Brandice

Brady

David W.

Cogan

John F.

2002. “Out of Step, out of Office: Electoral Accountability and House Members’ Voting.” American Political Science Review 96 (1): 127–40.

16.

Carmines

Edward G.

D’Amico

Nicholas J.

2015. “The New Look in Political Ideology Research.” Annual Review of Political Science 18 (1): 205–16.

17.

Clinton

Joshua D.

2006. “Representation in Congress: Constituents and Roll Calls in the 106th House.” Journal of Politics 68 (2): 397–409.

18.

Downs

Anthony

. 1957. An Economic Theory of Democracy. New York: Harper & Row.

19.

Enns

Peter K.

2015. “Relative Policy Support and Coincidental Representation.” Perspectives on Politics 13 (4): 1053–64.

20.

Erikson

Robert S.

1978. “Constituency Opinion and Congressional Behavior: A Reexamination of the Miller-Stokes Representation Data.” American Journal of Political Science 22 (3): 511–35.

21.

Erikson

Robert S.

1981. “Measuring Constituency Opinion: The 1978 U. S. Congressional Election Survey.” Legislative Studies Quarterly 6 (2): 235–45.

22.

Erikson

Robert S.

MacKuen

Michael

Stimson

James A.

2002. The Macro Polity of Cambridge Studies in Political Psychology and Public Opinion. New York: Cambridge University Press.

23.

Erikson

Robert S.

Wright

Gerald C.

1980. “Policy Representation of Constituency Interests.” Political Behavior 2 (1): 91–106.

24.

Erikson

Robert S.

Wright

Gerald C.

2000. “Representation of Constituency Ideology in Congress.” In Continuity and Change in U.S. House Elections, eds. Brady

David

Cogan

John F.

Fiorina

Morris P.

Stanford: Stanford University Press, 148–77.

25.

Erikson

Robert S.

Wright

Gerald C.

2009. “Voters, Candidates, and Issues in Congressional Elections.” In Congress Reconsidered, eds. Dodd

Lawrence C.

Oppenheimer

Bruce I.

Washington, DC: CQ Press, 71–95.

26.

Erikson

Robert S.

Wright

Gerald C.

McIver

John P.

1993. Statehouse Democracy: Public Opinion and Policy in the American States. New York: Cambridge University Press.

27.

Feldman

Stanley

Johnston

Christopher

. 2014. “Understanding the Determinants of Political Ideology: Implications of Structural Complexity.” Political Psychology 35 (3): 337–58.

28.

Flavin

Patrick

. 2012. “Income Inequality and Policy Representation in the American States.” American Politics Research 40 (1): 29–59.

29.

Frank

Thomas

. 2004. What’s the Matter with Kansas? How Conservatives Won the Heart of America. New York: Metropolitan Books.

30.

Gilens

Martin

. 2005. “Inequality and Democratic Responsiveness.” Public Opinion Quarterly 69 (5): 778–96.

31.

Gilens

Martin

. 2011. “Policy Consequences of Representational Inequality.” In Who Gets Represented?, ed. Gilens

Martin

. New York: Russell Sage Foundation, 247–84.

32.

Gilens

Martin

. 2012. Affluence and Influence: Economic Inequality and Political Power in America. Princeton: Princeton University Press.

33.

Gilens

Martin

. 2015. “The Insufficiency of “Democracy by Coincidence”: A Response to Peter K. Enns.” Perspectives on Politics 13 (4): 1065–71.

34.

Gilens

Martin

Page

Benjamin I.

2014. “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens.” Perspectives on Politics 12:564–81.

35.

Grumbach

Jacob M.

2018. “From Backwaters to Major Policymakers: Policy Polarization in the States, 1970–2014.” Perspectives on Politics 16 (2): 416–35.

36.

Hall

Andrew B.

2015. “What Happens When Extremists Win Primaries?” American Political Science Review 109 (1): 18–42.

37.

Hetherington

Marc J.

2001. “Resurgent Mass Partisanship: The Role of Elite Polarization.” The American Political Science Review 95 (3): 619–31.

38.

Hogan

Robert E.

2004. “Challenger Emergence, Incumbent Success, and Electoral Accountability in State Legislative Elections.” Journal of Politics 66 (4): 1283–1303.

39.

Hogan

Robert E.

2008. “Policy Responsiveness and Incumbent Reelection in State Legislatures.” American Journal of Political Science 52 (4): 858–73.

40.

Hollibaugh

Gary E.

Rothenberg

Lawrence S.

Rulison

Kristin K.

2013. “Does It Really Hurt to Be out of Step?” Political Research Quarterly 66 (4): 856–67.

41.

Hopkins

David A.

Laura

Stoker

. 2011. “The Political Geography of Party Resurgence.” In Who Get Represented? eds. Enns

Peter K.

Wlezien

Christopher

. New York: Russell Sage Foundation, 93–128.

42.

Huddy

Leonie

Mason

Lilliana

Aarøe

Lene

. 2015. “Expressive Partisanship: Campaign Involvement, Political Emotion, and Partisan Identity.” American Political Science Review 109 (1): 1–17.

43.

Hussey

Wesley

Zaller

John

. 2011. “Who Do Parties Represent?” In Who Gets Represented?, eds. Enns

Peter

Wlezien

Christopher

. New York: Russell Sage Foundation, 311–44.

44.

Iyengar

Shanto

Westwood

Sean J.

2015. “Fear and Loathing across Party Lines: New Evidence on Group Polarization.” American Journal of Political Science 59 (3): 690–707.

45.

Jacobs

Lawrence

Page

Benjamin I.

2005. “Who Influences U.S. Foreign Policy?” American Political Science Review 99 (1): 107–23.

46.

Keith

Bruce E.

Magleby

David B.

Nelson

Candice J.

Orr

Elizabeth

Westlye

Mark C.

Wolfinger

Raymond E.

2009. “The Partisan Affinities of Independent ‘Leaners.’” British Journal of Political Science 16 (2): 155–85.

47.

Layman

Geoffrey C.

Carsey

Thomas M.

2002. “Party Polarization and “Conflict Extension” in the American Electorate.” American Journal of Political Science 46:786–802.

48.

Layman

Geoffrey C.

Carsey

Thomas M.

Horowitz

Juliana Menasce

. 2006. “Party Polarization in American Politics: Characteristics, Causes, and Consequences.” Annual Review of Political Science 9:83–110.

49.

Lupton

Robert N.

Myers

William M.

Thornton

Judd R.

2015. “Political Sophistication and the Dimensionality of Elite and Mass Attitudes, 1980−2004.” The Journal of Politics 77 (2): 368–80.

50.

Miler

Kristina

. 2016. “Legislative Responsiveness to Constituency Change.” American Politics Research 44(5):816–43.

51.

Miller

Warren E.

Stokes

Donald E.

1963. “Constituency Influence in Congress.” The American Political Science Review 57 (1): 45–56.

52.

National Conference of State Legislatures. 2018. “State Partisan Composition.” https://www.ncsl.org/research/about-state-legislatures/partisan-composition.aspx (accessed March 2, 2020).

53.

New York Times . 2012, November 22. “State Government Control since 1938.” https://archive.nytimes.com/www.nytimes.com/interactive/2012/11/23/us/state-government-control-since-1938.html?ref=politics

54.

Poole

Keith T.

Rosenthal

Howard

. 1997. Congress: A Political-economic History of Roll Call Voting. Oxford: Oxford University Press.

55.

Rhodes

Jesse H.

Schaffner

Brian F.

2017. “Testing Models of Unequal Representation: Democratic Populists and Republican Oligarchs?” Quarterly Journal of Political Science 12: 185– 204.

56.

Rigby

Elizabeth

Wright

Gerald C.

2011. “Whose Statehouse Democracy? Policy Responsiveness to Poor Versus Rich Constituents in Poor Versus Rich States.” In Who Gets Represented?, eds. Enns

Peter

Christopher

Wlezian

. New York: Russell Sage Foundation. 189–222.

57.

Rigby

Elizabeth

Wright

Gerald C.

2013. “Political Parties and Representation of the Poor in the American States.” American Journal of Political Science 57 (3): 552–65.

58.

Royston

Patrick

. 2005. “Multiple Imputation of Missing Values: Update.” The Stata Journal 5: 188–201.

59.

Shor

Boris

McCarty

Nolan

. 2011. “The Ideological Mapping of American Legislatures.” American Political Science Review 105 (3): 530–51.

60.

Snyder

James M.

Jr. Ting

Michael M.

2002. “An Informational Rationale for Political Parties.” American Journal of Political Science 46 (1): 90–110.

61.

Stonecash

Jeffrey M.

2000. Class and Party in American Politics. Boulder, CO: Westview Press.

62.

Sulkin

Tracy

Testa

Paul

Usry

Kaye

. 2015. “What Gets Rewarded? Legislative Activity and Constituency Approval.” Political Research Quarterly 68 (4): 690–702.

63.

Tausanovitch

Chris

Warshaw

Christopher

. 2013. “Measuring Constituent Policy Preferences in Congress, State Legislatures, and Cities.” The Journal of Politics 75 (2): 330–42.

64.

Treier

Shawn

Hillygus

D. Sunshine

. 2009. “The Nature of Political Ideology in the Contemporary Electorate.” Public Opinion Quarterly 73 (4): 679–703.

65.

Wlezien

Christopher

. 1995. “The Public as Thermostat: Dynamics of Preferences for Spending.” American Journal of Political Science 39 (4): 981–1000.

66.

Wlezien

Christopher

Soroka

Stuart N.

2011. “Inequality in Policy Responsiveness?” In Who Gets Represented?, eds. Wlezien

Christopher

Soroka

Stuart N.

New York: Russell Sage Foundation, 285–310.

67.

Wlezien

Chrostopher

Soroka

Stuart

. 2008. “On the Limits to Inequality in Representation.” PS: Political Science and Politics 41:319–27.

68.

Wright

Gerald C.

1989a. “Level-of-Analysis Effects on Explanations of Voting: The Case of the 1982 Us Senate Elections.” British Journal of Political Science 19 (3): 381–98.

69.

Wright

Gerald C.

1989b. “Policy Voting in the U. S. Senate: Who Is Represented?” Legislative Studies Quarterly 14 (4): 465–86.

70.

Wright

Gerald C.

Berkman

Michael B.

1986. “Candidates and Policy in United States Senate Elections.” The American Political Science Review 80 (2): 567–88.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.28 MB