WHYPARTIESFORM 45

equilibrium the ordinary circumstance, there would be no incentive to form a party based on the social choice problem. But even when an equilibrium exists, the political party need not make its members worse off than without the parry; they can always choose to take no “partisan” actions. With PIEs generally considered impossible, there is a strong incentive for parties to form, precisely because of the likeli-

1altves hood of disequilibrium. Riker’s dismal conclusion turns out to provide a strong case for the formation of political parties.

The new institutionalism (e.g., Shepsle 1979) emerged in response to the ordinary absence of (pure) voting equilibria. Two points discussed below and in later chapters are also relevant here. First,

‘:ss transactions many different institutional arrangements can be sufficient to yield Wus making C (structure-induced) equilibria, such as committee systems, agenda

designs, and even separated powers. None of these are necessary— l)ority coalition like parties, all yield possibility results. Second, partisan institutions cs” of C. Sup- are one of those sets of sufficient institutions.

• id B find more sa. As Axelrod nal conflict of .i

•nnected, hay- COLLECTIVE ACTION AND ELECTORAL

coalition must MOBILIZATION

A-B-D, in- The Problem of Collective Action in Elections 2-C or B-C-D, cong C’s ideal The Federalist and Jeffersonian Republican parties began with the

government as a means of solving a social choice problem (see chap. ihere is an equi- 3). Such parties-in-government may also become electoral parties. ijine we should The most obvious motivation lies with the minority. The examples ci preferences, above demonstrated incentives for some majority to form a party. If

Incentives to this happens, some or all of those excluded might form a parry in reac- he disequilib- tion, seeking to become the legislative majority. Failing to reach major- ) has formed ity size, the minority would naturally turn to the public, seeking to 2.1 sometime elect more of its members. That is essentially what the Jeffersonians agree on C’s did when facing a Hamiltonian majority. Later parties, notably the

n on this pol- Jacksonian Democratic party, formed more directly for electoral pur-

e

would have poses (see chap. 4). The question for this section and the next, then, t Won at least is what set of incentives candidates for elective office might have that

Presumably would lead them to form or join a political party. In this section we at of prefer- examine incentives that arise from attempting to mobilize the elector-

ate. Mobilizing the electorate by definition is getting the public to turn that PIEs are out to vote for, or otherwise support, a candidate. Examining the logic Indeed, were of voting among citizens introduces the second form in which prob-

46 Political Parties and Democracy

lems of collective action are studied, and in this case turnout is the quintessential example.

The Nature of Problem of Collective Action and Mobilization

Turnout is ordinarily seen as a problem in individual decision making, unlike the prisoners’ dilemma. Both can be put in game theoretic terms, but in the latter case, the strategic interaction between the play- ers is central. Both players have an immediate and direct impact on the outcome, and each player would be wise to at least consider the strategic possibilities of the other player. In large electorates the out- come depends on the actions taken by all, but strategic interaction is so remote that it can be effectively ignored: how one citizen decides to act has very little effect on the decisions of any others. Sheer size all but eliminates strategic interaction, reducing the problem to one of individual decision making.

The standard theory, called the “calculus of voting,” employs ex- pected utility maximization (see Downs 1957; Riker and Ordeshook, 1968, 1973). If there are two candidates, the calculus, like all rational choice models, predicts voting for the more preferred one. The ques- tion is whether to vote at all. The calculus for choosing whether one votes or abstains is

(2.1) R=PB+D—C.

R denotes the reward (expected utility) for casting a vote, and one votes if R is positive and abstains if not. P represents the probability that the vote will affect the outcome, roughly the probability of casting the vote that makes or breaks a tie)9 B represents the differential bene- fit the citizen receives from the election of the more preferred candi- date. The D. for duty, term measures any positive rewards received from the act of voting itself, which may include the satisfaction of hav- ing done one’s duty as a citizen, the value of expressing support for the preferred candidate or party, and so on. Finally, C stands for the costs of voting, including the time and effort needed to register and go to the polls and the costs of decision making.2° C and D, therefore, come with the act of voting itself and do not depend on the outcome. Only B depends on the outcome, and it is discounted by the impact of P, the effect this one vote would have on determining that outcome.

This calculus is a typical example of expected utility maximizing. It thus serves as a template for a large number of other expected utility maximization problems. One example is the “calculus of candidacy”

WHYPARTIES FORM 47

that will be examined in the next section. It also serves as a calculus for political participation more generally. Olson (1965) analyzed the problem of collective action for participating in interest groups, for example, and his logic is effectively equivalent to this calculus.

The calculus is a model of individual decision making, but the out- come sought is a public good. The winning candidate is “jointly sup- plied,” no one can be excluded from “consuming” the good, and in- deed no one can avoid consuming it, no matter whether they voted for or against the winner or did not vote at all. The question, then, is under what conditions it is rational for the individual to contribute to (or “cooperate in”) the provision of this public good.

The collective action problem follows immediately from the calcu- lus and the observation that, in any large group, the P term is almost invariably very small. A near zero P makes the PB term tiny unless B is immense. Thus all those who share an interest in seeing a candidate elected nonetheless are motivated to act primarily on the D and C terms, that is, the intrinsic costs and benefits to voting, and very little in terms of their collectively shared interest in the candidate. If we set aside the D term for the moment (as Barry 1970 and others argue should be done), then one votes if PB> C. If P is effectively zero, then no one should vote. As in the prisoners’ dilemma, the rational citizen should “defect” by abstaining.2′

The calculus of voting includes a second, prior “collective action” problem: becoming informed. A citizen concerned about the electoral outcome needs to determine what outcome is desired. Which candi- date, in other words, does the citizen want to see elected, and how important is the outcome—that is, how large is the B term? The citi- zen must expend decision-making costs to gather and process infor- mation to determine this, but if a vote has a negligible impact on the outcome, why should anyone pay these costs? Downs (1957) ex- plained why it is rational for citizens to be ill informed except as they “accidentally” acquire information or obtain it for other reasons.

Incentives for Candidates in Electoral Mobilization

Candidates want to win elections. To do so, they need to convince more citizens to prefer them than prefer their opponent(s), and they need to convince these supporters to vote in greater numbers than their opposition. Citizens may not have incentives to turn out or even to ascertain their preferences over candidates. Candidates, however,

do have strong personal incentives to solve these collective action prob- lems for citizens, if only for their supporters. Campaigns therefore can

48 Political Parties and Democracy

be understood as attempts to create supporters and get them to turn out in the face of these two collective action problems.

There are a number of ways candidates can generate supporters and get them to vote, and these can be seen as attempts to manipulate terms in the calculus of voting. Most important are the common efforts to lower the costs of voting, such as exhortations to register and vote and formally organized mobilization drives. Candidates also can lower decision-making costs for voters by providing as much infor- mation as possible in a readily available form, seeking to “instruct” voters that the candidate values what they do, thus also seeking to generate a favorable B term as well as lowering C. At the same time, “allocating emphasis,” to use Page’s term (1976), or even outright ex- aggeration, may make the B differential appear large.

Exhortations that all citizens should do their duty by voting seek to increase the intrinsic rewards of voting, while claims that “everyone’s vote counts” seek to make the P term seem high. These claims strike everyone, however, opponents and supporters alike, so candidates typ- ically leave them to editorial writers and the League of Women Voters. But candidates do manipulate the P, B, C, and D terms more selec- tively. Thus candidates and parties focus their campaign appeals and mobilization drives on those they believe already are, or are most likely to become, their supporters.

Although candidates employ many particular tactics to make it seem in their supporters’ personal interests to turn out and vote, the general points are that candidates have private incentives to seek to overcome these collective action problems, and that these tactics, to be successful, must be chosen in light of the collective action problems facing the electorate. Implementing these tactics takes resources. It is probably not very expensive to generate the largely private benefits sufficient for overcoming the free riding incentives an individual citi- zen faces, but these small per capita costs become substantial in a large electorate. Yet as a great deal of empirical work has demonstrated (e.g., Patterson and Calidera 1983), wise expenditures of resources pays off in increased turnout.

Incentives for Party Affiliation for Candidates

That candidates have private incentives to reduce collective action problems among their supporters does not necessarily mean they have incentives to form a party. Today’s elections are typically described as “candidate centered,” and a large part of that claim is that it has be- come feasible for individual candidates to raise and expend resources

WHYPARTIES FORM 49

on their own (see chaps. 6 and 8). So part of the answer must be his- torically contingent, but part must continue to apply, since candidates with any serious hopes are almost invariably partisan.

Affiliation with a party provides a candidate with, among other things, a “brand name.” In advertising, successful brand names con- vey a great deal of information cheaply: they cue an established repu- tation (see Downs 1957). Travelers, for example, know little about a local hamburger stand but know that McDonald’s provides a certain type of product with standards for cleanliness, service, and so on. A party label can convey a great deal of information as well. The Ameri- can Voter popularized the view that political parties provide cues and partisan images (Campbell et al. 1960). Key (1966) referred to party identification as a “standing decision”: partisans vote for their pre- ferred party’s candidates until and unless given good reasons not to. The candidate’s party affiliation therefore provides a very inexpensive way to infer a great deal: what a typical Democrat or Republican is like. To be sure, other sources of reputation could serve much the same as party affiliation. A reputation as a liberal or conservative, for example, is a similar cost-saving device for voters. The empirical dom- inance of party cues (and their not coincidental relation to what is popularly understood by “liberal” and “conservative”) in the public suggests, of course, that the affiliation of a candidate with a party has proved useful. Thus the collective action problem for voters of becom- ing sufficiently informed to make a (possibly preliminary) determina- tion of whom they favor is greatly attenuated, given party affiliation and perhaps other reputational cues. This effect is exaggerated to the extent that voters’ choices are correlated among candidates of the same party. The correlation is, of course, partially endogenous to the actions and the stances of a party’s candidates (as well as to institu- tional features such as ballot forms that ease or hinder split-ticket vo- ting). Even today, however, many vote straight tickets, or close to it, and as Cox and McCubbins (1993) have shown, there is a substantial impact of party identification on even the highly candidate (especially incumbent) centered voting for Congress.

Affiliation with a party not only brings the candidate a “natural” reputation, it also provides economies of scale. This is especially im- portant for turnout. Campaigns may reduce free riding incentives in the public, but they are costly for the candidates. The campaign bud- get imposes real constraints, especially at lower levels of office and for nonincumbents. A turnout drive by the party’s presidential nominee reduces or eliminates the costs of getting partisans to the polls for

50 Political Parties and Democracy

other candidates of that party, for example. Once the voter is there, the additional costs of voting for remaining offices are very small, es- pecially if party-line votes are possible. Thus the tide of partisans turn- ing out to vote for president lifts the boats for all of the rest of that party’s nominees.

The combination of office-seeking ambition and the very nature of electoral institutions generates incentives for candidates to solve the two collective action problems affecting voters: becoming informed and turning out to vote. Candidates have two kinds of incentives to affiliate with a political party, ameliorating both of the public’s collec- tive action problems. Party affiliation provides an initial reputation that reduces decision-making costs and provides a core of likely sup- porters. Party campaign efforts, whether conducted by the party or- ganization itself or by its various candidates, provide economies of scale for all of the party’s candidates as they seek to reduce the costs and increase the benefits for supporters to come to the polls. As be- fore, these incentives create the possibility that candidates might want to affiliate with a party. There are other means of reaching the same ends. Moreover, affiliation is not costless. The reputational effects of being a Democrat or Republican need not be entirely positive and at times can be quite negative. Until recently being a Republican in the South provided a reputation, but one that made winning all but im- possible. Any partisan image undoubtedly mixes positives and nega- tives for any candidate. Yet the ambitious politician seeking a long and successful career almost invariably affiliates with one or the other ma- jor parry, in part owing to reputational effects and economies of scale.

One of the tensions facing partisan candidates is the need to solve another collective action problem, that of generating the many activ- ists needed to secure the labor and financial (and other) resources needed to achieve mobilization. This may yield tension, because the best appeal to activists may differ from what would best mobilize vot- ers. Resolution of such competing pressures depends in part on the activists’ incentives. For example, political machines generated selec- tive incentives for securing activists’ support that were largely inde- pendent of policy appeals to the public. The reduction of such private incentives is a substantial part of the forces that reduced or eliminated partisan machines. In place of the private benefit seekers of the machine era are today’s more policy-motivated activists. The conse- quences of such activists for inducing more divergence between candi- dates of opposing parties is developed and tested in chapter 6.

  • Page 1
  • Page 2
  • Page 3
  • Page 4