Political Parties as Mechanisms of Social Choice

Marjorie Randon Hershey. Handbook of Party Politics. Editor: Richard S Katz & William Crotty. Sage Publications, 2006.

Introduction: Rationality Approaches and the Study of Political Parties

If the use of rational choice theory in the study of political parties were a political candidate, it would be considered a front-runner with high negatives. Approaches that emphasize rationality have produced a large theoretical and empirical literature and a number of fruitful and provocative findings and have also generated intense criticism. The aim of this chapter is to set out the central tendencies of these approaches to the study of political parties, summarize some of their main applications in the field, discuss ways in which they have added to our understanding of party politics, and evaluate some of the main critiques of their use.

Rationality approaches, the general class to which social choice analysis belongs, begin with the premise that people behave purposefully Individuals are goal-directed actors. These approaches do not attempt to explain why people hold particular goals; rather, taking the goals as given, they assume that individuals will try to achieve their goals through instrumental, efficient means. In practice, the use of efficient means can be understood as the effort by individuals to increase the benefits they expect to receive from their actions relative to the costs they expect to pay.

This purposeful behavior takes place in an environment whose characteristics affect the individual’s ability to achieve his or her goals. Any environment will have a given level of information, a set of institutional rules, and a historical context. Political environments have some unique qualities that affect individuals’ choices. For one, governments are responsible for public goods those shared by all members of the society, whose provision cannot be limited to some members and withheld from others. Although individuals may vary in their ability to take advantage of such goods as improved air quality or stronger national defense, for example, they cannot be delivered to only those people who pay taxes to fund them while excluding those who do not.

One of the most intriguing insights of the rationality approaches is that to create these public goods—to clean the air and defend the nation—requires collective action, but collective action is vulnerable to specific and serious problems. In particular, the actions that would be required to achieve a public good may not be in the individual interest of any of the actors. Why should a member of Congress choose to spend her valuable time negotiating the content of bills and making appropriations to cut air pollution—a policy that will benefit all her congressional colleagues and all their constituents—while her colleagues are then freed to spend their own time pursuing pork barrel projects, casework, and other activities that will help ensure their own re-election? And if it is in nobody’s individual interest to devote their personal resources to pursuing collective goals, then how can those goals, which are in everyone’s interest, be achieved?

Analysts have shown that when a number of legislators or other individuals make choices on policies based on their own preferences, acting independently of one another, they can produce an outcome that all of them would consider worse than outcomes they could gain by organizing, even though each is behaving in ways that are entirely rational. Consider, for example, the chief executives of two cellulose plants located on the same lake. Because they are in competition, each plant chooses independently to reduce its costs by dumping chemical wastes into the lake. To recycle wastes would increase the expenses of the plant that did it and weaken its competitive position in the market. But dumping the wastes causes water quality to deteriorate. As a result, both plants, which need a relatively unpolluted source of water in order to produce their product, suffer long-term costs. In the short run it is rational for each plant to find ways to cut its costs. But the longer-range result is to increase the costs paid by each plant and by the larger community, which is clearly a suboptimal outcome.

These problems of collective action can be solved in a variety of ways. Collections of individuals or legislators can agree to act cooperatively in making their decisions. They can create sets of rules and expectations to structure their joint activity: these decision rules are termed ‘processes of social choice’. Stable sets of rules and expectations become organizations and institutions. These institutions can deal with collective action problems by creating inducements for individuals to act so as to further the interests of the collectivity or community as a whole.

A political party is one of these institutions: a means by which people can agree to behave cooperatively over the long term so as to secure benefits that they would not have been able to gain as individuals. A party could provide the mechanism for solving several kinds of social choice and collective action problems. Within a legislature, parties can make it possible to create lasting majorities for policies on important issues. In campaigns, parties can help candidates coordinate their behavior (their fund-raising, for example) in order to achieve mutual gain. They can make organized efforts to register and turn out voters so as to reduce the individual campaigns’ costs.

In elections, affiliating with a party lets a candidate offer potential supporters a means of reducing their information costs; knowing that a candidate is a Republican or a Democrat permits the citizen to infer a series of conclusions about the candidate’s policy stands and general approach to public life. Lower information costs increase the likelihood that the citizen will vote. For these and other reasons, practitioners of rationality approaches argue, it is instructive to conceptualize parties as institutions designed to solve problems of collective choice.

Rationality Approaches to the Study of Parties

Every functioning democracy has political parties. Nascent parties were created in the United States within the first decade after the signing of the Constitution. Interestingly, they were created by political leaders who had expressed strong anti-party sentiments, who had denounced the idea of parties as self-serving, dissension-provoking, and likely to be harmful to the new republic. Why did some political leaders choose to create parties at this time? What is the evidence that parties were created as a vehicle to solve problems of collective action? And why did these politicians form parties rather than some other kind of political organization?

One of the most insightful accounts of the rise of parties in the United States that explicitly adopts the rational choice perspective is that of John Aldrich in Why Parties? For Aldrich, the provocation for the formation of parties was a dilemma of social choice. Members of the infant federal government faced a large number of decisions on vital issues, ranging from the nature of government involvement in the economy to the location of the capital. The choice situation at this time, however, was fluid and unstable. Members of Congress and other federal notables were pulled and tugged on several dimensions simultaneously: their ideological beliefs about the scope and power of the federal government, their attitudes toward other issues, sectional concerns, and inter-state and intra-state rivalries. Majorities were fleeting and unstable; vote-trading deals were made and then unraveled. In short, the Congress was in the grip of a social choice problem.

One of those most concerned with the fluidity of this situation was Alexander Hamilton. Hamilton had initially enjoyed majority support in Congress for his vision of a strong central government. But he was having trouble holding that support; Hamilton’s chief antagonist, James Madison, won some key debates by prospecting for votes among Hamilton supporters on an issue-by-issue basis. If Hamilton hoped to win consistent majorities for the principle of a federal government with expanded power and to establish a clear precedent for future action consistent with this principle, then he needed a means of keeping his supporters together over a range of votes.

As Aldrich shows, Hamilton succeeded by laying the groundwork for a legislative party. As Secretary of the Treasury, he was able to set the congressional agenda by submitting a plan for dealing with the public debt. He worked with allies in the House to coordinate which bills were offered and to enlist his supporters to vote for them. In these ways, Hamilton was able to ensure that those in Congress who shared his views about the role of the federal government could overcome the problem of fleeting and unstable majorities and translate their preference into government action. In social choice terms, a form of institutional structure was created to reduce the number of competing bases for choice, to bring equilibrium instead of disequilibrium, and thus to produce an answer to the fundamental question that faced the new government.

Once Hamilton’s legislative ‘party’ had come into existence, as the majority, it was able to win on any votes where it organized. The opposing group, headed by Madison and Thomas Jefferson, then had no alternative but to try to alter the status quo by changing the membership of Congress in the next congressional elections. They set up a number of ‘committees of correspondence’ to encourage the election of other Jeffersonians and succeeded in winning control of the House in 1792. By that time, voting behavior in the House and Senate had moved from ‘widely dispersed scatters’ to ‘two clear and quite separate blocs … not just polarized but also very strongly related to partisan affiliations,’ consistent with Aldrich’s expectation.

A second major dilemma of collective action, Aldrich argues, can be seen in relation to the formation of the first true mass-based American party, the Democrats, in the late 1820s. Recall that a central issue in collective action is that even though something might be in the collective interest of a large number of people, no individual may see it in his or her personal interest to contribute to that goal. The collective interest in this case was the election of a president. As will be shown later, individuals will be tempted to abstain from voting because even the small costs of gathering information and going to the polls are likely to outweigh the minimal benefits the individual expects to get from the outcome. But mass abstention can be a big problem for politicians, because they need votes in order to fulfill their ambitions.

In the relatively fluid politics of the 1820s, a number of ambitious politicians, notably Martin Van Buren and Andrew Jackson, came together to create an institutional structure—a mass party—that would enable them to attract more votes and thus benefit from control of government. This form of political party used mass rallies, parades, and other means to provide citizens with information about the election. By doing so, the mass party reduced citizens’ information costs and, by working to mobilize citizens, helped bring more prospective voters into the polling booths.

But to help solve the collective action problem for prospective voters, those who wanted to organize a mass party faced a collective action problem of their own: organizing a mass party is a costly enterprise for individuals, given that the benefits will be shared. Aldrich posits that the party’s midwives would have tried to reduce these costs by building on the organizations in states where there was already the start of a party infrastructure. Once these new forms of party were developed, they would focus their efforts at voter mobilization in those states and in elections that were expected to be close, because those are the races in which a party’s effort would be most likely to make a difference.

Consistent with his hypotheses, Aldrich finds major increases in turnout in the New England and ‘middle’ states (New Jersey and Delaware), where there was high competitiveness and where the Democratic Party had successfully organized. Because these existing local organizations were more likely to join the mass party if they were able to maintain a substantial amount of independence (a selective incentive), this choice by the founders of the mass party laid the groundwork for a degree of local party autonomy in American politics that has long been unusual among the world’s democratic party systems.

Aldrich’s third main contribution is to use rationality assumptions to explain the major recent change in the American parties: the movement from the mass party to the ‘party in service’, which developed around the 1960s. This development was caused by several long-term changes in the parties’ environment, which altered the structure of rewards and opportunities. Changes in the electorate, such as increasing educational levels and the assimilation of immigrant populations, reduced the number of people who needed the help provided by the older mass parties at the local level. Changes in national policy, especially expansion of the New Deal entitlement programs, further reduced the need for the local party’s assistance. The steady inroads of Progressive and Populist reforms and of ticket splitting restricted the party organization’s ability to control access to office and to control office-holders once they were elected.

In addition, the rise of national media enabled individual candidates to reach voters, contributors, and issue- and candidate-oriented campaign activists without needing to work through a party organization. The size of the new voting population and the changes in communications media meant that campaigns would have to be directed by experts in these new media and other technologies: professionals rather than volunteers. In the mass party of earlier years, candidates were generally subordinate to the party organization in campaigns. But by the 1960s, ambitious candidates could (and usually had to) win office without the party’s direction. Candidates, not parties, became dominant. The parties, adapting as they have throughout their long lives in American politics, thus had to gear up to become ‘parties in service’ to provide services and expertise in order to retain at least an active role, if no longer a dominant one, in candidates’ campaigns.

Do we need the logic of social choice to explain why parties formed and why they changed during the last half of the 20th century? Probably not. It might be possible, alternatively, to argue that people’s affiliative needs lead them to identify with and become active in parties and that the rise and fall of partisan behavior over time reflects changes in the available alternatives for affiliation. Some might respond that the need for affiliation itself could be interpreted as a rational process; but if so, then it is fair to ask whether rational choice is a falsifiable system of analysis.

Whether or not it is the only defensible explanation, does the rationality approach give us new insight into these vital periods of party change? Clearly it does. In the case of the parties’ founding, for instance, the approach calls attention to the fact that it was not enough for most members of Congress to agree that the new national government should assume the debts incurred by the states before the Republic was formed. Most members did agree, yet they were unable to act initially on that agreement. Institutional arrangements had to be created to produce incentives for the members of that majority to vote on the basis of their felt need for a strong national government rather than on other possible bases. As an approach, then, it has considerable power to stimulate insights about the behavior of political parties.

Mobilizing for Collective Action

If we accept Aldrich’s premise that a party is a means to an end, created and shaped by officeholders because they find it to be in their personal interest, then we are led to ask a variety of questions about the relationship between institutional structure and the ambition of candidates and public officials. Analysts in the rational choice school have examined this relationship in three major areas. The first is voter choice; as in the case of Van Buren and Jackson, candidates need to get voters to support them in elections and to mobilize activists on their behalf. Second, elected officials need to resolve conflicts over leadership recruitment and succession: who will have the opportunity to run for which office at a given time. Finally, officeholders continue to deal with the challenges that the early Congresses faced: what alternatives to public policy problems will be considered, in what order, and who will get what, when, and how from government?

In a democracy, citizens must have some role in choosing important government personnel. Writers and activists may differ on the thresholds they would prefer, but it is fair to assume that at least some minority of officials with significant decision-making authority must be elected by citizens. Yet with even a minimal definition of ‘democratic’, we quickly run into a major problem of collective action. Although governments in modern industrial democracies make decisions about virtually every aspect of their citizens’ lives, citizens’ awareness of that impact may be slight. The effects of many government decisions are complex and long-range. People’s immediate needs for food, shelter, and entertainment dominate their thinking and their time; developing an interest and a desire to become informed about an entity that seems to be faceless and remote is not common. Yet without at least some minimal citizen participation, a democratic government lacks legitimacy. So how is it possible to get citizens to go to the polls to choose elected officials if they would rather watch football, if their interest in politics and information levels are low, and if they feel confident that government will continue functioning without their input?

When a democracy begins to expand, nascent party organizations have a straightforward incentive to selectively stimulate citizen participation. The more voters they are able to mobilize, the greater their likelihood of electing their candidates. The early Democratic Republicans (Jeffersonians) clearly followed that path in reaching out to citizens in the localities; party leaders further had an incentive to expand the right to vote in order to gain power using the votes of these newly enfranchised groups. Although the Federalists may also have seen the logic of enlisting new voters, they were less well suited to do so because the policies that animated the Federalist Party had much greater appeal to merchants, bankers, and other elites—who already had the right to vote and who were limited in numbers—than to the immigrants and frontiersmen who initially lacked that right.

As a democracy matures and the franchise expands to approach its natural limits, the challenge then becomes the need to motivate qualified voters to go to the polls. Parties can serve this need in several ways. For Anthony Downs, the use of a party ideology can be a ‘rational habit’ for voters. When an individual has carefully informed herself about the parties and policy stands in preceding elections and has always come to the same decision, she could reasonably conclude that the probability of changing that decision is too small to warrant investing time and energy in gathering information about the current election. As long as conditions remain the same, this habitual decision helps to reduce the citizen’s costs and thus to increase the probability that she will vote. If conditions change beyond a certain threshold, she should see the need to gather information again.

Others also view the existence of parties in the electorate, or party identification, as a means of simplifying voters’ choices and therefore making it easier for them to choose to vote. For Morris Fiorina, for example, citizens can use the device of retrospective analysis to reduce their costs. Although politics is confusing, Fiorina writes, voters ‘typically have one comparatively hard bit of data: they know what life has been like during the incumbent’s administration’. Presuming that the parties are reasonably consistent in their policies over time, this retrospective judgment asks less of citizens than does an approach that expects considerable data-gathering; information about the past is cheaper to obtain than information about the future. Retrospective voting could refer only to a particular administration but it could also be extended to a standing assessment of the administration’s party.

In any of several ways, then, citizens can use party as a means of drawing inferences about the candidates’ characteristics and policy stands. It is a form of stereotyping that can be used to shortcut the process of gathering information about the choices in an election. It can summarize the likely impact of each candidate or party or simply reduce the voter’s choices to a manageable number. Through whatever means, the party symbol’s function as a stereotype or heuristic is a means of getting a lot of useful information cheaply, and that, in turn, increases the likelihood that a citizen will vote.

In addition to these two means by which party can solve this collective action problem -as a symbol by conveying a fairly cheap package of useful information and as an organization by reaching out to citizens who might not otherwise vote—party can also cut citizens’ costs of involvement in government by linking them to their representatives. Some evidence suggests that party identifiers are more likely to contact and request services from a US House representative if he or she shares the identifier’s party affiliation.

Recruiting Leaders and Regulating Access to Political Office

Democracies need to recruit leaders as well as voters. Political leadership is a public good so, as we have seen, problems of collective choice arise in its selection. Collective action problems should be mitigated in the recruitment of leaders by the fact that those who win office get selective benefits—power, importance, and other perquisites of holding office, at least at higher levels. These benefits can be offset, however, by the many challenges that face elected officials in the form of public cynicism, media criticism, and the larger salaries and bigger perks that are available in comparable positions in the private sector.

Why would parties serve as useful devices for leadership recruitment rather than, say, non-profit groups like the United Way or for-profit organizations? A new party would have the motivation to do so as a means of securing power for itself. Federalists in the new American government sought like-minded allies to establish Hamilton’s proposals as the blueprint for the new central government. Then the Federalists’ opponents saw the need to serve as an alternative channel of candidate recruitment. By encouraging both the spread of local party organizing among leaders who held the same general sentiment and the connections among these local parties in this recruitment process, the Democratic Republicans helped to further their own ambitions as well as to strengthen the existence of the federal system.

As this party competition becomes more established, a party hoping to maintain (or obtain) dominance in government should not only seek like-minded individuals to run for office under the party’s label but also offer help in the form of campaign expertise, money, and other resources to increase their chances of winning. This seems to be an almost universal activity of party organizations, though the extent of its provision varies over time. The locus of parties’ provision of resources varies as well; as many local party organizations struggled to deal with changes in society and in communications media during the 1960s and 1970s, organizations of the party in government such as the congressional campaign committees (and currently state legislative campaign committees as well) began to collect resources to aid party candidates in their campaigns.

Parties have never been able to monopolize the provision of resources to candidates in American society; the rise of new types of organized interests and methods of communication has made this point more obvious in recent decades. But the major parties have found a way to compensate for this weakness. They have used their dominance of the legislative and judicial branches to pass rafts of laws giving easy access to the ballot to candidates running as Democrats or Republicans and seriously disadvantaging those who want to run as independents or under a minor party label. These ballot access rules and a range of other laws and court decisions have protected the two major parties’ status as the main channel through which all but a very few resource-rich candidates must run for office.

The parties’ institutional rules for selecting candidates also offer a means of solving problems of collective choice. Larry Bartels has examined the ways in which the individual preferences of voters in presidential primaries are combined into a single collective choice. Bartels notes that the nominating process differs from a classic public choice situation in that it is dynamic; it consists of a series of primaries on successive dates over a five-month period. The set of competing candidates can change from primary to primary, the electorate clearly changes, and voters in each primary contest can learn from the primaries that took place before. So learning and momentum characterize the process. These characteristics help to produce ‘a genuine majority for a single alternative’, as opposed to a situation in which no candidate can muster a real majority.

Making Decisions and Allocating Resources in Government

Finally, governments must make decisions as to what will be done and in what order. How are these decisions to be made? If majority rule is the chosen mechanism, as is commonly the case in a democracy, then how will a majority be created? In particular, how can the possibility of continually cycling majorities, the social choice problem described by Arrow, be resolved?

Creating and sustaining parties within government is one such means. As we have seen, partisanship can provide a tool with which groups of decision-makers who have common goals can build relationships of trust and predictable patterns of agreement. It can also simplify deliberation and act to sanction trust-breakers in order to protect that predictability.

Much of the rational choice work on governmental decision-making, like most other analyses of national political institutions, focuses on Congress. The work of the national legislature is more public and produces more usable data than do bureaucratic agencies and federal courts. Within Congress, analysts usually focus on roll call voting, the final step in House and Senate action on a bill. Rationality approaches to committee and subcommittee markup and the work of conference committees, which are just as vital to the legislative process, are limited by the problems all congressional researchers face in getting access to these aspects of congressional behavior.

What is the nature of the legislative parties in the rational choice perspective? Let us begin, as many students of congressional behavior do, with David Mayhew’s analysis of Congress members as single-minded seekers of reelection. Mayhew shows how this single goal can explain a series of important institutional features of Congress. Members of Congress, in order to get re-elected, have designed an institution that enables them to meet their needs for three important activities: advertising themselves favorably, claiming credit for desirable behavior (such as casework), and taking positions on bills and other issues. One aspect of this institutional design, Mayhew writes, is that the congressional parties were organized to allow each member to take positions that benefit his or her re-election. Party leaders, then, were inclined to tell their members to vote their constituency’s interests rather than to risk losing an election by supporting a party position that is unpopular in their district.

In Mayhew’s Congress, in short, the congressional parties and their leaders have very limited power; the primacy of individual members’ re-election needs does not leave room for party leaders to demand obedience to a ‘party line’. Yet since Mayhew’s analysis was published in 1974, other researchers in the rational choice tradition have described congressional parties that are considerably stronger. Gary W. Cox and Mathew D. McCubbins, for example, begin with the same assumption as does Mayhew: that legislators are guided by a desire for re-election. But for Cox and McCubbins, members’ re-election depends not just on their own personal qualities and votes but also on public perceptions of their party’s record, as can be seen in such collective phenomena as electoral tides. Somebody, then, needs to marshal the forces for creating and protecting that record.

Who has the motivation and the ability to do so? The legislative party leaders, in Cox and McCubbins’s view, are a means of dealing with the collective action problem of getting someone to do what members of the party in Congress all want done but do not want to spend the time and resources doing. Some in the mainstream of their legislative party will find it attractive to take on these legislative tasks—to monitor the situation and use selective incentives to encourage other legislators to act on the party’s preferences—in return for the rewards of holding leadership positions, and particularly because the successful completion of these tasks could increase their party’s chance of becoming (or remaining) the majority.

Because they are in their party’s preference mainstream and usually from safe districts, these party leaders are not likely to suffer in their own re-elections from acting on their party colleagues’ preferences. So they internalize the party’s collective interests; they structure the legislative process so as to keep members of the majority from ‘cheating’ on deals and unraveling the majority coalition.

Using statistical evidence, Cox and McCubbins show that party leadership does structure the legislative process. They demonstrate that the legislative party finds a variety of ways of controlling the legislative agenda in the interests of its members. They show that members are more likely to receive desirable committee assignments if they demonstrate more loyalty to the party leadership. More generally, they confirm that members of Congress behave so as to support the partisan structures that can constrain their self-interested activities. The result, they argue, is that the legislative parties in the House, and especially the majority party, have become ‘legislative cartels’ in that they have taken the power to make the rules that guide the movement of legislation.

That leaves us with an interesting puzzle. Two sets of researchers, each working within the rationality tradition, have reached different conclusions as to the strength of party in the legislative process. How can these differences be explained? David Rohde has developed and John Aldrich has added to a theory they call conditional party government. In brief, they contend that the party’s members in Congress will grant their leaders greater power under certain conditions. In particular, the amount of power the party leaders will hold, and thus the degree to which the party is able to affect the voting behavior of its members, depends on two factors: the extent of internal agreement within each legislative party (preference homogeneity) and the distance between the two parties on important issues (preference conflict).

Rohde and Aldrich argue that these two factors have electoral roots. As the parties outside of Congress—voters and activists—have become more homogeneous internally and more polarized relative to one another, they have elected representatives who are closer to the center of their party’s caucus than was the case in earlier years. In particular, the shifting partisanship of conservative Southerners, from their traditional Democratic identification to Republican affiliation, has ‘purified’ the ideological composition of the two parties in the House; as a result, members of each House party caucus have become more inclined to delegate power to their party leaders. Others suggest that the members’ own beliefs are probably central to this change as well. It could be that the decisions of ambitious candidates have played an independent role in reorienting the party preferences of conservative Southerners and other partisans.

It is clearly true, as we will see, that House members have voted with the majority of their party colleagues to a much greater extent since the 1980s than was the case when Mayhew wrote Congress: The Electoral Connection. But not everyone agrees that this demonstrates the increasing power of the legislative parties as a means of rationalizing decision-making. Keith Krehbiel has argued that party power in Congress is an illusion; instead, what we term ‘party’ is simply a surrogate for individual preferences. When members’ preferences on issues (as measured by interest group ratings, which are based in turn on roll call votes) are controlled, Krehbiel writes, the effects of party and partisanship are rare. Rather, he argues, members of Congress vote with fellow members simply because they agree on the policy in question. To Krehbiel, then, what Rohde regards as the conditions for strong legislative parties are nothing more than a change in Congress members’ preferences, which Rohde has misla-beled as ‘party’.

Krehbiel’s argument is a major challenge to the view that parties structure the legislative process. But it has been critiqued from a number of perspectives. One is his reliance on a measure based on roll call voting; any such empirical test does not rule out the possibility that parties act at other points in the legislative process, for example, in helping to shape the drafting of bills, their markup in committee, and their placement on the legislative calendar. Another is the observation that the legislative parties do have incentives to offer their members in return for party support and that recent changes in the organization of the congressional parties have increased the ability and the willingness of party leaders to use these incentives to pressure their members.

The incentives available to House party leaders were expanded by the Democratic caucus beginning in the early 1970s. Reforms that reduced the autonomy of committee chairs increased the power of both the caucus and the party’s legislative leadership. Then, after the 1994 Republican takeover of the House and Senate, the Republican House party restricted the committee chairs’ independence further, term-limiting the chairs and freeing the party leadership from seniority rules in selecting the chairs. More recently, after the Republican victories in 2002, which left the Republican House delegation even more uniformly conservative than before, the party’s leaders in the House gained greater power. The new reforms eliminated the term limit adopted in 1995 for the Speaker, weakened the seniority system again so as to permit the House Republican leadership to award subcommittee chairs on the powerful Appropriations committee to their loyal allies (and thus increase leadership control over bills dealing with federal spending), and cracked down on Republican members who opposed the leadership. In addition, the Republican leadership restricted Democratic participation on conference committees and resolved more of the details of legislation at the leadership level rather than in committee and on the floor.

The stated aim of the 2002 reforms was to help the leadership push the Bush economic and social agenda through the contentious House and thus move the inevitable compromise with the Senate more toward the House’s stance: in other words, to create rules that would offer an incentive for party legislators to accept the discipline necessary to pass the president’s program. A key House Republican pointed out: ‘By abandoning the seniority system, the leadership was able to ensure that it had its team on the committees. It reinforced the idea that chairmen are not autonomous. They owe their allegiance to the leadership.’ This certainly seems to suggest that the legislative parties have provided inducements to members to vote with their party, which goes beyond a story of the simple expression of shared policy goals.

Several researchers also demonstrate that the increased voting cohesion of the congressional parties goes beyond what we would expect from similarity of preferences alone. Cox and Poole, for example, measure the difference between votes on policy and votes on procedure. The latter, they argue, would be more likely to elicit pressure from congressional party leaders. In fact, they find that members are more likely to vote with their party on procedure, suggesting that the difference is caused by pressure from the leaders. Snyder and Groseclose write that when we measure Congress members’ votes on what are assumed to be ‘free votes’—issues on which one party is overwhelmingly likely to prevail, and thus members are probably free to vote their own and/or their constituency’s preference—and then take into account the party’s stance, there is consistent evidence of party influence affecting the individual legislator’s ‘true’ preference.

Some ingenious quasi-experimental studies provide further evidence that party makes a difference independent of members’ preferences on issues. Wright and Schaffner, for example, compare the organization and the coalition formation in the non-partisan Nebraska state legislature with that of the partisan Kansas legislature ‘next door’. They find that although candidates for the Nebraska legislature do identify as Democrats and Republicans and show evidence of party polarization in their views on issues, the roll call voting of the elected representatives in Nebraska, where the legislature is not organized by party, does not show any clear pattern. Roll call voting in the Kansas Senate, on the other hand, shows clear partisan divisions. ‘Party’, they conclude, ‘lends order to conflict’, just as a rationality approach would predict.

There is persuasive support, then, for the idea that the American legislative parties developed not just as an expression of members’ shared preferences but also as means of stabilizing and controlling the congressional agenda and of regularizing and satisfying members’ institutional ambitions. Governing parties formulate policies in order to further the interests -policy, re-election, or other—of the individual office-holders and office-seekers who comprise them. These sets of positions on issues have the effect of structuring conflict on policies, and thus structuring the approach of a legislature to its environment.

The Value and Limits of Rationality Approaches

How much value do we gain from the use of rationality approaches to study political parties? The author of a recent article in the American Political Science Review described rational choice theory as ‘arguably the most popular and fastest-growing theoretical orientation in contemporary political science.’ It has substantially changed the way political scientists study political parties, he wrote. In fact, because rationality approaches could be considered a universal theory of political and social behavior, these approaches can be the means of helping social scientists integrate the study of parties with that of other institutions.

Many would agree with this assessment. The use of social choice analysis has been central to the institutional approaches that have undergone a renaissance in political science. One of the most valuable contributions of these approaches is the reminder they provide that democratic institutions are not self-generating. The simple and powerful fact that the collective interest is not sufficient in itself as a motivator for any one individual’s behavior, without some specific benefit to the individual making the decision, demands the effort of political scientists to explain why democratic rules develop, how they are maintained, and what undermines them.

These approaches have generated a great deal of criticism, however, some of it intense. In one of the best-known critiques, Donald Green and Ian Shapiro charge that rationality approaches, including the social choice approach to the study of political parties, practice ‘post hoc theory development’: theorists, they argue, first look at empirical evidence and then design a model that fits it, rather than making falsifiable predictions, testing them, and then adapting the model. In a later defense of their critique, they suggest that ‘Rather than ask “What causes X?” [a rational choice model] begins with the question “How might my preferred theoretical or methodological approach account for X?”’ If this is so, then the approach might produce an interpretation of a series of events, and perhaps a most interesting interpretation, but not a prediction that can be falsified. To prove useful, a theoretical approach has to go beyond simply relabeling existing findings or showing that results are consistent with expectations that may, of course, have been influenced by prior knowledge of the results.

Aldrich avoids this trap in his analysis of parties; he posits that actors will seek solutions to the collective and social choice challenges that he spells out, but does not assume that these solutions will necessarily take the form of political parties. Not all rationality approaches are as careful. In the effort to explain why citizens vote, for instance, it has been generally accepted that the main determinants are the costs of voting and registration and the benefits to the individual: the likelihood that his or her vote will make a difference in the election outcome and the extent to which the citizen expects the election of one party to bring greater benefits to him or her than the election of the other party. The problem is that in many if not most cases, even the relatively small costs of voting are likely to overwhelm the benefits an individual can expect to derive from going to the polls. Thus one empirical study suggests that voters’ perception of a narrowing gap between the policy stands of the major British parties helps to explain the recent decline in British voter turnout. In the same sense, most people will not find much reason to try to inform themselves about the issues and the candidates. Why then, does anybody vote?

This problem can be handled. Some posit, for example, that parties or candidates use resources to mobilize voters in close races and thus overcome the collective action problem without having to claim that voters gain much benefit from going to the polls. In many cases, however, analysts have resorted instead to putting a figurative thumb on the ‘benefits’ side of the scale by adding an unspecified ‘sense of citizen duty’ or support for democracy, even in the absence of careful empirical data about these costs and benefits. Others simply posit that individuals can have a ‘taste for participation’ in social acts and for improving the lot of other people because they have group interests as well as individual goals. In the worst cases, this can amount to saying that people vote because they value voting, which does not carry us very far toward an understanding of voting behavior.

Another criticism raised by Green and Shapiro is that the body of work using rationality approaches in political science is empirically challenged. There has been an emphasis on developing theory and formal analysis but the effort to engage in systematic empirical testing has been given short shrift. This criticism, while widespread, is not entirely fair. There are notable examples of extensive empirical research in the rational choice tradition, such as Elinor Ostrom’s work on institutional development.

Nevertheless, the use of formal analysis can limit the approach’s ability to guide empirical research. The benefits of formal analysis are greatest when the model is elegant, when the number of variables can be reduced to the bare minimum and represented in mathematical form. But elegance of presentation—a great strength—is also a great weakness in that it prevents us from exploring the complexity of situations, which is, for many analysts, what makes politics interesting. Some practitioners suggest that this is not a reasonable criticism, in that when we consider the various models as pieces of one puzzle, then we might represent the complexity that is necessary to an understanding of political life. But it is fair to say that the practitioners have the burden of pulling these models together if they are to argue that the enterprise has value, rather than putting the onus on the reader.

The primary question in evaluating these approaches is whether the social choice framework gives us greater purchase on such questions as why parties form and why they delegate greater or lesser power to party leaders than would an explanation that did not make use of social choice tools. What are the alternatives to rationality approaches? Many political psychologists would contend that at least some behaviors are expressive, in the sense that individuals choose these behaviors for their own sake, rather than for the goals to which they lead. Others would offer cultural explanations for behavior: that people behave according to social norms or attachments—because the group indicates that the behavior is appropriate -rather than because the individual calculates that it will bring her closer to a particular goal. Dennis Chong writes: ‘In the political realm, it also appears that social norms, principled commitments, and expressive or symbolic benefits are often needed to motivate participation in collective action. Whether these so-called extra-rational incentives stretch the rational choice model or break it is the source of much controversy in the field.’

Still others have shown that outcomes are not the only important motivating forces. Psychologist Tom Tyler has demonstrated that individuals are more concerned with the fairness of a decision process, as they perceive it, than with the benefit that it produces for themselves. Political scientists John Hibbing and John Alford show that the value people place on outcomes depends more on what they see others as receiving than with the simple value of the outcome to themselves. Individuals are more inclined to value payoffs from decision-makers who do not seem to be taking too big a share for themselves and who seem reluctant to hold the power to make the decisions, presumably indicating that they are not greedy for more. Hibbing and Alford interpret their findings as evidence for the existence, even the prevalence, of non-maximizing behavior.

Proponents of rationality approaches would argue that many of these explanations can be subsumed under rational choice—in the case of expressive behaviors, for instance, by positing that the expressiveness itself is a goal, for which the particular behavior is chosen. But this must be done with care. It is straightforward to interpret material rewards for behavior. Classifying ‘civic duty’ and ‘affiliative needs’ as benefits can make it difficult to explain why anyone would not vote or participate.

The vehemence of this debate, which can be seen in mentions of the ‘pathologies’ of rationality approaches, or, on the other side, references to the ‘Luddite’ version of an interpretation, makes its resolution unlikely. As Michael Taylor points out, the gulf between practitioners of rationality approaches and others is wide enough to include disputes ranging from philosophy to methods. One survey found that a plurality of members of the American Political Science Association rejected the fundamental rational choice assumption that ‘individuals are rational utility-maximizers’, and only a third agreed with it, but that 78 percent of public choice scholars (largely economists) accepted the assumption. And a number of political scientists find the methodological tools of many social choice studies (formal modeling in particular) impenetrable.

The debate has even developed partisan overtones. Some scholars have claimed that the public choice perspective reflects an individualist philosophy with a normative slant toward tax-cutting, privatizing, and, in a word, conservatism, or even Republicanism. In fact, a survey of public choice scholars showed that most agreed with the statements that government does more to create and protect monopoly power than it does to prevent it (64 percent) and that when transaction costs are low, the market achieves the efficient allocation of resources regardless of how property rights are assigned (70 percent). But as Bernard Grofman observes, whether or not many proponents of rational choice analysis are themselves politically conservative, the form of analysis itself has no necessary partisan or ideological slant.

Where Can We Go from Here?

It would be a fatal weakness to find that a body of propositions that purports to be a general theory does not have universal application. But it is nevertheless plausible to suggest that a particular theoretical approach does a better job of explaining some aspects of politics than others. In the case of social choice approaches, their application to the behavior of political activists and leaders would seem to be more productive than their application to those who are closer to the periphery of political life.

As Joseph Schlesinger suggests, rationality approaches work best in well-defined institutional settings where the benefits are clear and observable and the costs are easy to calculate. A legislature is one such setting. The choices repeat more frequently in legislative voting and committee action than they do in a voter’s occasional forays into the polling place. That facilitates the calculation of costs and benefits. More generally, people who are involved in some active political role are likely to be in choice situations that are salient to them and characterized by fairly clear expectations and observable consequences. Activists, candidates, and public officials probably hold better-defined goals for their political activity than do those who play less active political roles as well as greater ability to determine whether or not these goals have been met. As Fiorina writes, ‘rational choice models are most useful where stakes are high and numbers low, in recognition that it is not rational to go to the trouble to maximize if the consequences are trivial and/or your actions make no difference’.

If we were to array individuals along a continuum from higher-level political actors to those who are relatively uninvolved with politics, then, the balance between instrumental and purely expressive behaviors probably shifts toward the instrumental the closer we come to the higher-level political actors. The less engaging the behavior, the more room there will be for expressive behaviors, for symbolic concerns rather than self-interest to prompt decisions. How else can we explain the importance of such issues as gay marriage in guiding the behavior of prospective voters, given that the level of self-interest for most Americans is nil? (Candidates and other political leaders, of course, can derive individual benefits from persuading voters that the issue affects them.)

We have seen that rationality approaches have been fruitfully applied to elements of the party in government and to Congress in particular. Applications to the study of party organization could be just as useful. There has been a resurgence of interest in party organization in the last two decades and in party activists more recently Joseph Schlesinger has made several important contributions to this enterprise. He notes, for instance, the important implications of the fact that parties compensate their participants indirectly rather than directly, and that the party organization maintains itself through market exchange, offering candidates and policies in return for votes with which to win office. Because winning office is so vital to parties, those who occupy the offices—the candidates and office-holders—inevitably hold positions of strength within the party.

Further work in this tradition could benefit us even more. It might be especially fruitful to draw on the insights generated by rational choice analysis of interest groups, which bear some interesting similarities to and differences from parties. For instance, consider Mancur Olson’s formulation of the ‘free rider’ problem with regard to interest groups, and especially large interest groups. Several analysts have focused on the role of selective incentives—individual benefits that they do not need to share—in explaining why individuals do in fact join large groups that are designed to achieve a collective goal, and on the possibility that some individuals overestimate their impact on such a group to the point where they believe that they are capable of making a difference in the group’s outcome even if, objectively, they ought to free-ride.

Jack Walker has offered a different explanation: that patrons and sponsors such as private foundations helped to reduce the costs of collective action for many organized interests. Does this happen in the case of parties? Surely it does in some third parties, as in the case of Ross Perot and the Reform Party. The Supreme Court’s recent ruling on the McCain-Feingold campaign finance reforms indicated that patrons can continue to donate unlimited amounts to newly-organizing parties. Major parties have sponsors too, including the corporate patrons who underwrite their national nominating conventions. Clearly, there are important differences between parties and organized interests. Yet the similarities are significant enough to raise questions about whether some of the insights into interest groups as mechanisms of social choice could be applied more fully to parties.

As the front-runner with high negatives, the social choice approach to studying political parties faces as big a challenge as do political candidates in the same position. If political scientists are to gain greater understanding of parties and their impact on politics and policy, it would benefit us to take insights from wherever we can get them. If the social choice proponents accept that challenge, then, like Aldrich, they need to lay aside their equations on occasion and speak directly to those who see parties in cultural, expressive, or other non-maximizing terms. And scholars on all sides need to recognize that a gentling of the rhetoric in continuing this dialog can go a long way toward reducing the high negatives and increasing our ability to learn from one another.