N.R. Miller
10/21/04
PARTY COMPETITION FOR VOTES: ELECTORAL
DIVERGENCE AND CONVERGENCE
In the handout on Strategic Effects in Voting Systems, we considered strategic entry or exit by political parties (or candidates) as well as the strategic question of how many candidates a party should run in a given multi-member district. Beyond this there is the question of how political parties (or candidates) should position themselves with respect to issues or on the ideological (left-right or liberal-conservative) spectrum. By changing their positions, political parties (or candidates — but henceforth I will refer only to parties) can change voters’ preference orderings over parties and perhaps change their voting choices. We assume that parties seek policy positions that will best advance their goals.
Recall that in our early preview of the origin and nature of political parties, we distinguished among three ideal types of goals political parties might have. (Real parties of course have a mix of these and other goals.) An ideology-expressing party aims to educate the electorate by “sticking to its principles” in the hope converting more voters to its principles in the distant future and therefore will not strategically change its position in order gain more votes in the present. However, an office-seeking party aims to win elections in order to gain the offices, power, patronage, etc., that result from electoral victory. Such a political party has no true policy or ideological preferences and is ready to adjust its positions in any way that will increase its support among voters in any particular election. (An office-seeking — or vote maximizing — party is analogous to a profit-maximizing firm in economic theory.) A policy-seeking party has policy preferences but its goal is to act strategically in way that gets its preferences enacted into public policy as closely as possible; therefore it must in accordance with the aphorism that “if you want to be a statesman (i.e., enact preferred policies), you must first win an election.”
We may expect that almost all “serious” parties in systems with free elections will be some mix of the office-seeking and policy-seeking types, since ideology-expressing parties almost by definition are not electorally successful. And it turns out that, office-seeking and policy-seeking parties behave in much the same way when competing for electoral support (because the latter must typically must compromise its policy goals in order to win the election needed to enact even a compromised policy). However, electoral competition works in fundamentally different ways in two-party systems and multi-party systems. Thus Duverger’s Law implies that the nature of the electoral system shapes the nature of electoral competition as well as the number of political parties.
Seventy five years ago, the economist Harold Hotelling published an article on “Stability in Competition” (Economic Journal, March 1929) that analyzed how strategic considerations often led competing stores like Woolworth’s and Kresge (think K-Mart and Target but before the era of suburban malls) to locate almost adjacent to each other in the middle of town, even though consumers would be better off if they located on the opposite sides of town so that most customers would be closer to the closest store. In general, Hotelling complained that “buyers are confronted with an excessive sameness.” And in a famous side comment, he though the same principle applied to political competition.
So general is this tendency that it appears in the most diverse fields of competitive activity, even quite apart from what is called economic life. In politics it is strikingly exemplified. The competition for votes between the Republican and Democratic parties does not lead to a clear drawing of issues, an adoption of two strongly contrasted positions between which the voter may choose. Instead, each party strives to make its platform as much like the other's as possible. Any radical departure would lose many votes, even though it might lead to stronger commendation of the party by some who would vote for it anyhow. Each candidate “pussyfoots,” replies ambiguously to questions, refuses to take a definite stand in any controversy for fear of losing votes. Real differences, if they ever exist, fade gradually with time though the issues may be as important as ever. The Democratic party, once opposed to protective tariffs, moves gradually to a position almost, but not quite, identical with that of the Republicans. It need have no fear of fanatical free traders, since they will still prefer it to the Republican party, and its advocacy of a continued high tariff will bring it the money and votes of some intermediate groups.
Thirty years later another economist Anthony Downs wrote a book on An Economic Theory of Democracy (1957) in which (among other innovative things) he developed Hotelling’s idea quoted above and turned it into what now generally referred to as the Downsian model of two-party competition, which can be sketched out as follows.
Suppose that voters are distributed (with respect to their first preferences or “ideal points”) over an ideological spectrum that ranges from the far left to the far right. (Typically most voters may near the middle of the ideological road, but the Downsian model does not assume this.) With respect to their lower preferences, voters prefer positions on the ideological spectrum that are closer to their ideal points to those that are farther away (cf. the description of the preferences of Congressional committee members in Problem #4 of Exercises Pertaining to Voting Rules). Two parties are also situated on this ideological spectrum, one (party L) initially at a distinctly leftwing position and the other (party R) initially at a distinctly rightwing position. The simple Downsian model focuses on a two-party system, so additional parties cannot enter the electoral arena and each voter votes for whichever party is closer to his or her ideal point. Which party get the most votes depends on (i) exactly where on the spectrum L and R are located and (ii) exactly how voters are distributed across the spectrum. Suppose that more voters find themselves ideologically closer to R than find themselves closer to L, with the result that R would win the election if neither party changes its position. If L and R were ideological-expressing parties, the story would end at this point — neither party could better achieve its goals by changing its position. But suppose L and R are both office-seeking parties, shamelessly willing to change their positions to win more votes. This being the case, the initial positioning of the parties fails to be an equilibrium: the prospective losing party finds it advantageous to change its position. In this case, the L party finds it expedient to shift its position somewhat in the direction of the middle of the ideological road since, as a result, some middle-of-the-road voters who previously were closer to R than L are now closer to L than R and will switch their votes from R to L. If L shifts far enough towards the center, L will win the election instead of R. But equilibrium still has not been reached, because it is now party R that has an incentive to shift its position closer to the center of the spectrum. If it shifts enough, R can not only cancel out the advantage that L has gained by its shift but can also regain the electoral advantage. You can see where this story is heading. No equilibrium is achieved until both L and R are positioned adjacent to one another right in the middle of the ideological road. That is, the Downsian model of two-party system competition has the following basic implication.
(a) A competitive two-party system entails party convergence, i.e., both parties offer similar middle-of-the- road policies to the electorate (both as election promises and as actual policies carried out when the party holds office).
From this, several corollaries follow.
(b) In a competitive two-party system, voters do not greatly care which party wins the election (since they offer convergent, not divergent, policies).
(c) In a competitive two-party system, elections are very close, with neither party getting much more or less than 50% of the vote.
(d) In a competitive two-party system, there is frequent alternation in party control of government (that is not, however, accompanied by big changes in government policy).
We have said that the two parties tend to converge in the middle of the road, but this it to speak a bit loosely. Where exactly is “the middle of the road”? There is a theoretically precise answer. This ideological “middle of the road” corresponds to the most preferred position of the median voter (relative to whom half the voters are more leftwing and half the voters are more rightwing). You will recall from the discussion of Problem #4 in the Exercises Pertaining to Voting Rules, this position is the Condorcet winner among all positions along a spectrum of spending levels and they same logic applies to an ideological spectrum. In a two-party system, all potential elections are straight fights (i.e., between L and R). A party located at the position of the median voter cannot be beaten in an election no matter where the other party locates along the ideological spectrum. Therefore both parties seek to occupy this unbeatable position.
Further Questions to Think About
1. We assumed above that the L and R two parties are both office-seeking. Suppose both parties are policy-seeking (as their L and R names may suggest). Does party convergence still result?
2. An implicit assumption above is that all voters vote (for the ideologically closest party). Is this assumption compatible with the conclusion? What would change if turnout was less than 100%. How might the level of turnout be a function of party positions?
3. The Downsian model takes it as a given that there are only two parties? How would the logic of party competition change if additional parties could readily enter the competition? Suppose that four vote maximizing parties are initially located at more or less equal intervals along the ideological spectrum. What are the incentives for convergence or divergence?
4. The Downsian model is (perhaps) persuasive in explaining why parties in a two- party system tend to stay pretty close to the middle of the ideological road — even when their historical origins and most loyal supporters are very distinctive, e.g., the British Conservative Party with its roots in the aristocracy and the British Labour Party with its roots in the working class and trade unions. Nevertheless, in practice such parties never converge completely (i.e., to the point of being literally indistinguishable) and sometimes they actually do diverge quite a bit (e.g., U.S. parties today are said to be quite “polarized”). What are some factors (not taken account of in the simple Downsian model) that may keep the parties in a two party system from converging as much as the Downsian logic suggests? Some answers are suggested by the following excerpt from recent review article. (You can find the full article at a link on the course web page.)
DOWNS AND TWO-PARTY CONVERGENCE by Bernard Grofman
Abstract. We take as our starting point the insights of Downs into two-party competition. A careful reading of Downs offers a much more sophisticated and nuanced portrait of the factors affecting party differentiation than the simplistic notion that, in plurality elections, we ought to expect party convergence to the views of the median voter. Later scholars have built on Downsian ideas to see what happens vis-a-vis party differentiation when we modify key assumptions found in the basic Downsian spatial model. Recent work allows us to turn what is taken to be the Downsian view on its head: Although there are pressures in two-party competition for the two parties to converge, in general we should expect nonconvergence. Moreover, contra the negative portrait offered by Green & Shapiro of the limited or nonexistent value of research on party competition models in the Downsian tradition, we argue that, when viewed as a whole, neo-Downsian models — especially those of the past decade — do allow us to reconcile theory and data in terms of a multi-componented theory of party competition with testable implications for comparative statics.
INTRODUCTION
The simple Downsian model of two-party competition under plurality is generally characterized as predicting party convergence to the policy position espoused by the median voter, and thus "Tweedledum-Tweedledee" political competition; yet this prediction violates empirical reality even for the United States, the country whose electoral politics provided the empirical inspiration for Downs' work. For example, when a given constituency elects members of opposite parties (e.g., when a congressional seat changes hands) to a member of the opposite party, or in states that are simultaneously represented by senators of opposite parties), the difference in voting records (as judged, say, by ADA scores) between the office-holders of different parties can be huge. This discrepancy between model and reality has led some scholars to argue that rational choice modeling of party competition is empirically vacuous.
Yet the standard Downsian convergence result rests on more than a dozen specific assumptions, such as the assumptions that parties/candidates are motivated solely by office seeking and that voters choose solely on the basis of their policy proximity to candidates' positions. As this essay will show, when one or more of these assumptions is violated, the usual convergence result can often be expected to disappear. Moreover, we no longer get other unrealistic predictions from the Downsian approach, such as the expectation that elections would be decided by relatively narrow margins because candidates of the two parties would be identical in their only relevant attributes (i.e., their policy platforms or anticipated policy choices).
There have been a number of different approaches to accounting for candidate/party divergence in plurality-based elections. Often scholars seize upon one single explanatory factor. Our purpose here is not to provide new results but rather to provide a synthesis that reveals how a whole series of disparate results about divergence, especially recent results, can be integrated into a common framework to instruct us about which of the basic assumptions of the standard Downsian model should be replaced. Almost any violation of the basic assumptions used by Downs to generate the two-party convergence result is likely to replace convergence with some degree of divergence in party positions.
The basic assumptions of the standard Downsian model are as follows:
1. There are only two political parties.
2. There is a single-round election for any office.
3. The election chooses a single candidate.
4. Elections take place within a single constituency.
5. The election is decided by a plurality vote.
6. Policies can be located along a single (left-right) dimension.
7. Candidate policy positions are well defined.
8. Candidate policy positions are accurately estimated by each voter.
9. Voters look no further than the next election.
10. Eligible voters go to the polls if the expected benefits of their vote 's contribution to the election of the candidate for whom they would vote exceed the "costs" of voting.
11a.. Voters care only about which candidate/party will enact policies closest to their preferences. They vote for the candidate closest to their own policy location.
11b. If there are no policy differences among the candidates/parties, then voters will be equally likely to support each of the candidates/parties.
12. Parties/candidates care only about winning.
13. Parties/candidates look no further than the next election.
14. Candidates/parties accurately estimate the policy preferences of voters, or at minimum, they can identify the location of the median voter overall and the median voter in each party.
15. Candidates are part of a unified party team.
By seeing what can happen to convergence when we replace one or more of the above assumptions with more realistic ones, we can assess how much convergence we might expect. Moreover, we can begin to identify the factors with the greatest impact on the degree of divergence. When we examine recent work modeling the forces that affect party convergence, we see that the standard view of the Downsian model as predicting convergence in two-party political competition gets it almost completely wrong. Yes, there are centripetal pressures; but in general they are partly or largely outweighed by the centrifugal ones, producing what my coauthors and I have labeled “spaced out politics.”
To simplify the exposition, this essay discusses the implications of violating individual assumptions while leaving other key assumptions intact. Although our focus is on two-party competition under plurality, and we take almost all of our illustrations from the United States, most of the ideas below are in no way restricted in their applicability to the United States, and may be extended to the multiparty-competition setting under various electoral systems.
From Annual Review of Political Science (2004) 7:25-46.