Democracy, Heterogeneity and Representation: Explaining Representational Differences Across States

Benjamin G. Bishin, University of Miami

Introduction

Perfect substantive representation is impossible when constituents disagree.¹ In such cases, the elected must choose whom to represent. Simplified, this is the problem introduced by electoral heterogeneity. As an electorate becomes diverse in their preferences, politicians are forced to choose among constituents.² As a result heterogeneity bears critically on studies of congruence—the degree to which legislator behavior reflects the preferences of the citizenry. Recent research suggests that the process legislators use to make decisions varies according to state diversity (Bailey and Brady 1998). This essay proposes an alternative explanation for Bailey and Brady’s (1998) important, and perhaps underappreciated, research on the impact of citizen diversity on the representation process.

Heterogeneity and Representation

The degree to which heterogeneity conditions legislator congruence with citizens’ preferences has received scant attention from representation scholars. In one of the few studies to examine how representation varies according to state heterogeneity, Bailey and Brady (1998) find that the influences on legislators’ roll call votes on GATT and NAFTA differ significantly between homogeneous and heterogeneous constituencies.³ Perhaps the most important finding of their work is that the behavior of legislators in homogeneous states—those ranked among the 25 least diverse—is driven by constituent interests, while the behavior of those in the most heterogeneous is driven by political party and personal ideology. The authors take their results as evidence for dyadic representation, “that the representative process is different in homogeneous and heterogeneous states” (1998, 536).

This finding is groundbreaking. The results suggest that an important aspect of representation has been overlooked. Moreover, the results raise an entirely new set of research questions that are central to understanding how representation works. Theories of representation must account both for the statistical differences that seem to accrue as a result of heterogeneity and explain how legislators decide which cues to use in different contexts.⁴

Subconstituency Politics: A Theory of Representation

Dyadic representation does not uniquely explain these interstate differences. A group politics explanation also comports well with Bailey and Brady’s (1998) results. At its core, preference heterogeneity deals with group interests; groups of individuals support, oppose or ignore issues on the agenda. Politicians appeal to, represent, and build coalitions of individuals to gain support. Groups form when several individuals share an intense preference on one or more issues. Their intensity gives them political power disproportionate to their size (Olson 1965). Heterogeneity describes the diversity of group preferences in the district.

Groups, or subconstituencies, are fundamental to the way legislators view and service their districts (Fenno 1978). Goff and Grier (1993) point out that even same state senators’ behavior varies depending on how they view their districts and who they view as constituents. Schiller (2000) argues that these senators develop distinctive reputations to differentiate themselves, implicitly suggesting that they appeal to different groups. Bishin (2000) shows that legislators are responsive to the preferences of potential supporters rather than those of the average voter. Differences in the confluence of district preferences, and in the way legislators see their districts, bear directly on our ability to detect representation.

The subconstituency politics theory provides an alternative to dyadic representation. Instead of resulting from different representational processes, Bailey and Brady’s (1998) observed interstate differences are a product of the distribution of group preferences. In both cases, legislators appeal to group preferences, but because of preference diversity in heterogeneous districts, this influence is minimized by using measures of constituency preference that are based on district averages. Such measures better reflect constituent influence in homogeneous districts, where the group is large relative to the size of the district.

By definition, voters’ preferences in homogeneous states and districts are likely to vary less than they do in heterogeneous states and districts. In both cases, legislators may respond by reflecting the preference of the largest group of citizens with a position on the issue.⁵ If we estimate constituent preferences by accounting for the preferences of all citizens, rather than only the group to whom legislators actually appeal, the representative in the heterogeneous state mistakenly appears much less faithful to the electorate than does the legislator representing the homogeneous state. It is harder to substantively represent constituents that disagree.

Measures of constituent preferences will introduce more bias as the degree of homogeneity decreases. However, it is important to note that the crucial concept is not descriptive diversity, but preference diversity. Issues may arise on which even the most heterogeneous of constituencies entirely agree. Such issues might deal with public goods provided to a state or district. So while descriptively diverse states or districts may tend to have more substantive disagreement on issues, this need not be the case on particular issues.⁶

Indeed, inconsistent findings in the representation literature have previously been attributed to legislators varying their method of representation. For instance, Miller and Stokes’ (1963) seminal study suggests that their findings on the varying paths of influence on legislator behavior are a product of legislators changing their representational style across issues. The subconstituency politics theory implies that the manner in which legislators respond to groups of constituents does not vary across issues. Rather, our accuracy in assessing group preferences varies with diversity of constituent preferences. As constituent preferences become homogeneous within districts, legislators will appear more responsive to constituent preferences. This appears to be the case as Miller and Stokes find congruence highest on the salient civil rights issues—where intra-district opinion was likely homogeneous, and a lack of congruence on foreign policy issues—precisely those issues on which citizens’ preferences are likely to be most diverse.

Discussion and Conclusion

A group-based explanation provides a reasonable theoretical explanation for differences in representation by state heterogeneity. Instead of fundamental differences in the representation process, the theory suggests the same process occurs in both heterogeneous and homogeneous states. In both cases, legislators appeal to the preferences of groups. In homogeneous districts these groups are large relative to the size of the district. In such cases, scholars are more likely to observe the ‘statistical signal’ constituent groups send their legislators. Our measures of district preferences are influenced by district heterogeneity. In sum, while our ability to detect it varies, substantive representation occurs in both homogeneous and heterogeneous constituencies.

The subconstituency politics theory of representation is an alternative to Bailey and Brady’s dyadic representation hypothesis which asserts that the representational process varies between homogeneous and heterogeneous states. Clearly, empirical testing is needed but the theory seems widely applicable to inconsistencies in the policy congruence literature as well.

Notes

The author thanks Juliet Gainsborough for her comments.

1. Perfect substantive representation refers to the idea that legislators can act on behalf of all of the citizens in their Geographic Constituency.

2. In any system with single member electoral districts, heterogeneity induces representational bias. Politicians can represent groups of people, but cannot represent everyone when citizens disagree. I argue below that, while imperfect, substantive representation occurs when legislators act on behalf of groups.

3. The concept of heterogeneity has not been ignored. Fiorina’s (1974) inconclusive results have led to the investigation of the effect heterogeneity has on legislators’ reelection prospects (e.g. Bond 1983, Bond, Campbell and Cottrill 2001) and on the various ways in which heterogeneity should be conceptualized (Koetzle 1998). Heterogeneity is also crucial to Fenno’s (1978) discussion of how members see and service their districts.

4. This question seems deceptively simple. After all, when groups are of unequal size, which is usually the case, larger groups of potential supporters are clearly of more value to vote seeking legislators. However, such logic ignores the problem of intensity and the disproportionate influence small but intense groups can wield when opposed by large but apathetic ones.

5. For purposes of illustration we need only assume that which Fiorina (1974) has already demonstrated—that when groups with strongly held preferences disagree, legislators do not stake out a median position between the two (which is likely to anger both groups) but side firmly with the group that benefits them most.

6. It is precisely the conflation of these concepts that leads to the paradox of representation Lublin (1997) describes concerning majority-minority districts.

References

Bailey, Michael and David Brady. 1998. “Heterogeneity and Representation: The Senate and Free Trade.” American Journal of Political Science 42: 524-544.

Bishin, Benjamin G. 2000. “Constituency Influence in Congress: Does Subconstituency Matter?” Legislative Studies Quarterly 25: 389-415.

Bond, Jon R. 1983. “The Influence of Constituency Diversity on Electoral Competition in Voting for Congress, 1974-1978.” Legislative Studies Quarterly 8: 201-217.

Bond, Jon R., Kristin L. Campbell and James B. Cottrill. 2001. “The Puzzle of Constituency Diversity Revisited: Conditional Effects of District Diversity on Competition in Congressional Elections.” Typescript. Texas A&M University.

Fenno, Richard F., Jr. 1978. Home Style: House Members and Their Districts. Boston: Little, Brown.

Fiorina, Morris P. 1974. Representatives, Roll Calls and Constituencies. Lexington, MA: DC Heath.

Goff, Brian L. and Kevin B. Grier. 1993. "On the (Mis)measurement of Legislator Ideology and Shirking" Public Choice. 76: 5-20.

Hill, Kim Quaile and Patricia Hurley. 1999. “Dyadic Representation Reappraised.” American Journal of Political Science 43: 109-137.

Koetzle, William. 1998. “The Impact of Constituency Diversity upon the Competitiveness of U.S. House of Representatives Elections, 1962-1996.” Legislative Studies Quarterly 23: 561-574.

Lublin, David. 1997. The Paradox of Representation: Racial Gerrymandering and Minority Interests in Congress. Princeton, NJ: Princeton University Press.

Olson, Mancur Jr. 1965. The Logic of Collective Action. Cambridge, MA: Harvard University Press.

Schiller, Wendy. 2000. Partners and Rivals: Representation in U.S. Senate Delegations. Princeton, NJ: Princeton University Press.

Wright, Gerald R. 1989. "Policy Voting in the U.S. Senate: Who is Represented?" Legislative Studies Quarterly 24 (November): 465-486.

Ben Bishin’s email address is bishin@miami.edu.