Party Membership and Participation

Knut Heidar. Handbook of Party Politics. Editor: Richard S Katz & William Crotty. 2006. Sage Publishing.

What is Party Membership?

Party membership is defined in this chapter as ‘an organizational affiliation by an individual to a political party assigning obligations and privileges to that individual.’ The role of party member is a formal role, to be distinguished from the behaviorally defined role of party activist. How parties organize themselves and administer membership varies widely. Parties usually keep a register of their members, and issue membership cards. Members are generally obliged to pay annual membership fees (dues) and to pledge not to be members of other parties simultaneously. Socialist parties may expect their members to also be members of trade unions; Christian parties may expect party officers to be practicing Christians. The trend is to lower the threshold for party membership by limiting obligations and reducing dues. Membership privileges include participating in party activities: electing party officials, nominating candidates for public office, debating policies, and participating in decision-making and in social events. Sometimes the affiliation is collective, as when a non-party organization, such as a trade union, signs up parts of its membership as a bloc. But in this case identifying the individual party member would be difficult.

Both parties with formal membership and those with informal membership have party activists. Having formal membership, however, is not a criterion for an organization to meet the definition of ‘party.’ US political parties do not have formal memberships, but do have party activists (see Eldersveld, 1986; Stone et al., 2004). In Africa there are parties without formal membership (Carbone, 2003) and parties whose members belong to several parties simultaneously (Erdmann, 2004: 65).

Party membership means different things depending on the situation. In states with one-party systems it is often difficult to distinguish between party members and public officials (Giliomee and Simkins, 1999). The definition of ‘party member’ is more varied and culturally contingent than what is signified by the terms ‘citizen’ and ‘voter’—which are defined by public law and election behavior, respectively. Party membership is usually a more demanding form of participation than voting. Duverger (1964: 61) mentions a series of ‘concentric circles … of … ever-increasing party solidarity’ -suggesting terms such as ‘supporters,’ ‘adherents,’ ‘militants,’ and ‘propagandists’ as useful descriptions of party attachment. These circles are described behaviorally, and are, according to Duverger, closer to the ‘real nature of participation’ than is formal membership. Accordingly, the levels of party attachment ‘define the content of the sociological bond which unites the members of the community to which we give the name “party”.’ Party membership bonds may also be described organizationally, according to the nature of obligations imposed and privileges bestowed on members (Scarrow, 1996: 16-18).

Varieties of Membership

The significance of party membership springs from the character of these bonds, and to some extent also defines the type of party organization. In top-heavy ‘catch-all’ parties, party members are assigned less importance than in ‘mass’ parties. In ‘mass’ parties the members generally have strong links to the party organization. Likewise, in ‘caucus’ parties, where the parliamentary party wing has a tradition of autonomy, the extra-parliamentary party is left to fill the role of fan club, helping to mobilize voters.

In the early 20th century, debates in US political parties regarding their nomination rules for primaries (local, state, and national) illustrate the point that member linkage affects the way the party works. The custom was that the party had the discretion to decide which party members qualified to participate in the nomination of the party’s electoral candidates. Abuses led to public regulations (Merriam, 1907). In some states a voter who wanted to influence the party’s nomination was required to declare his or her intention to support the party’s candidates in the election. In other states there were ‘tests,’ including a declaration that one sympathized with the party’s goals, or a declaration that one believed in most of the principles of the party. Or one was simply asked: ‘Are you a Republican or a Democrat?’ This public regulation of the nomination process left the US parties without party members in the formal sense defined above. When the Libertarian Party of California in 2004 advertised for ‘members’ on the web, the ‘sociological bond’ reflected the US ‘no-member’ tradition ( One can either register to vote as a Libertarian or sign up with the party as a ‘dues-paying member.’ As a dues-paying member one is eligible to vote on central committee business, and to be a delegate to the annual state convention. However, it is also necessary to declare that one does not ‘believe in or advocate the initiation of force as a means of achieving social or political goals.’

In the 20th century socialist parties had tough membership requirements, reflecting a strong member-party bond. To join Argentina’s Socialist Party during the 1930s one needed to declare in writing acceptance of party statutes, principles, methods, and programs. The applicant also needed sponsorship by two people who had been party members for at least six months. Moreover, one had to wait one year to be eligible to vote for party officials and six months to vote on all other questions (Wellhofer, 1972). This high threshold was more ‘socialist’ than ‘Argentine’ in its origin. The Peronist populist party of the 1950s had no formal membership policy. It accepted ‘opportunists from all sectors’ and exercised ‘overt and tacit coercion’ to make public employees join the party (Little, 1973: 658).

Party membership in one-party dictatorships illustrates another membership bond. According to Leninist party theory only one party can be allowed to control the state. Under the Soviet regime, prospective party members were carefully screened. For a time they were ‘candidate members’ before being trusted to practice ‘democratic centralism’ (Ware, 1987). Party membership in one-party states gave access to privileges: jobs, information and education. For these reasons the ‘party’ label and consequently the institution of party membership may be suspect in new democracies recently evolved from one-party to multi-party systems (Bratton, 1999).

The nature of party membership depends on social context. At one extreme membership is an expression of belonging—a simple reflection of religious, class or ethnic identity. Entering the party is a natural occurrence within one’s social milieu—parties and party membership are expressions of ‘segmented’ or ‘encampment’ societies. One grows up in a Catholic family, a Catholic neighborhood, goes to Catholic schools, reads the Catholic press and belongs to Catholic organizations—including the Catholic Party (Beyme, 1985: 192). Most famous were the Austrian cradle-to-grave parties where one lived within the class ‘laager’ of society, where most individual needs had organizational outlets. The Austrian social democrats, to counter the omnipresence of the Catholic Church, developed in the early interwar period an extensive organizational flora, including the Workers’ Stamp Collecting Association. Party membership was only one of many expressions of living in a strong, closed class community.

Forms and Entitlements

There are basically three forms of party membership: individual, auxiliary, and collective. Individual membership is established when an individual signs up with the party—generally at the branch level, but increasingly at the national level by signing up through the mail or over the Internet. One is expected to agree with basic party goals and to be of a certain age. Parties often require that members have citizenship. The Irish Fianna Fáil used to ask only for ‘a connection to Ireland through birth, residence or Irish parentage,’ while its adversary Fine Gael was only open to Irish citizens (Beyme, 1985: 168). Most Canadian parties do not require members to be Canadian citizens, but in fact almost all members are (Cross and Young, 2004: 435). With implementation of transnational elections for the European Parliament and new (local) voting rights for immigrants, citizenship may not be as widely required for membership in European parties.

Often parties have auxiliary organizations for youths, women, and pensioners. These organizations are internal subdivisions of the party organization. Youths’ organizations generally have both minimum and maximum age requirements; however, being of voting age is not necessary. Dual memberships are often possible—in both a party branch and the youths’ or women’s organization—which makes an accurate count of party members difficult.

Finally, collective (‘corporate’) membership is found when an organization that is not formally of the party enlists all or some of its members in the party. Obviously these organizations are close to the party, for example trade unions or farmers’ associations. In the early European labor movement, the relationship between party and trade unions was described as ‘one body, two arms.’ A local, industrial trade union branch might be a subunit of both its national trade union and a particular party, dealing with the business of both at its meetings. As trade unions grew less partisan, particularly after the schism following the Russian Revolution in 1917, individual trade union members sometimes had the option of stating formally their reservations against being registered as party members. Collective membership was—and to some extent still is -particularly important to many European social democratic parties. Sometimes collective membership was arranged locally, as in Scandinavian countries. Sometimes it was arranged nationally, as in Britain. National trade unions generally enlisted their members in a party for financial and/or power reasons. The individual trade unionist might not realize he was also a party member. Such corporate arrangements make it hard to define ‘membership’ as an individual attachment, although the transaction is registered as membership in party statistics. This is sometimes labeled ‘affiliated membership’ in the literature (Katz and Mair, 1992). Generally, collective membership within social democratic parties has been replaced by other ways to maintain close relationships with friendly external organizations.

Counting Party Members

Accuracy in establishing membership totals is difficult; parties’ membership claims must be treated with caution. Because the character of party membership can vary according to time and place, establishing uniformity in counting (and eventually comparing) is difficult. In fascist militia groups, communist cadres, and ‘catch-all’ party branches—all of which have unique membership bonds—membership numbers would signify very different organizational capabilities in each case. Nevertheless, it would be useful to know when and where such parties gained or lost members.

Basically there are two ways to estimate party membership levels: probing party registers and studying party membership claims made in surveys. Party registers are often based on varied and changing operative principles, and are seldom up-to-date. Membership files, donor registers and mailing lists might be combined indiscriminately by the party to arrive at a membership total. Ambitious activists may register half-hearted individuals as members, taking a ‘perhaps’ for a ‘yes,’ and disregarding whether individuals pay dues. To look good or to qualify for more delegates at national conferences, or for increased public subventions, party branches might falsify membership totals. Keeping up-to-date files to reflect members who are deceased, are not paying dues, or have terminated membership is difficult. For example, should secretariats count, for a certain time, members who stop paying dues? The introduction of computerized files in the 1980s and 1990s, and greater professionalism among party administrators, have reduced the impact of such problems.

Using the survey method could be problematical. Apart from chance uncertainties, people may not know or remember that they are members, or may even falsely claim membership out of embarrassment at not being a member.

Although party membership claims should be handled with care, there is usually an interesting story behind the figures. Critical evaluations of membership figures may give important insights into a party’s evolution and party trends. Changes in party membership numbers have, for example, been discussed in the ‘decline of parties’ debate, although most authors stress that membership decline is only one factor among many in evaluating party strength. Empirical works interpreting changes in party membership are generally focused on Western European countries (Katz and Mair, 1992), and most studies register a decline in recent decades (Scarrow, 2000; Mair and van Biezen, 2001). Many new democracies in Southern Europe (since the 1970s) and in Central Europe (in the 1990s), however, experienced increases in party membership. Earlier, relatively high party membership levels in the 1950s and 1960s may reflect a boom in participation following World War II, and not a ‘golden age’ of party vitality as during the advent of democracy in the early 20th century. Due to a lack of reliable data, however, these long-term tendencies are not much studied (but see Bartolini, 2000). Despite the recent overall decline in membership—often interpreted as a decline in the segmented social structure of post-industrial societies (known as ‘individualization’)—there are large differences between levels of party membership across Europe. Austria still has a high level of member density (party members as share of party vote), despite having declining party membership. Likewise, parties within the same country show persistent, wide differences in their ability and/or willingness to recruit members.

Why Parties Want Members and why People Want to Join

Why do parties recruit members and why do people join parties? Parties may want members to help in campaigning, to provide electoral legitimacy, to run and finance the organization, to recruit new candidates for public office, to anchor the party in civil society, to sound out grassroots opinion and to develop new policies (Katz, 1990; Scarrow, 1994: 28; Ware, 1996: 63-84). No doubt party leadership will welcome differently the contributions of members in each of the above areas. If the intention originally was to recruit members to run successful electoral campaigns, leadership may find members wanting, even demanding, to influence party policies. New, alternative sources of income—such as public subvention—may change the calculus on the benefits of membership, reducing the incentive to generate party income from membership dues.

The 19th-century parliamentary party factions—the ‘caucus’ or ‘cadre’ parties—had mostly local supporters. These parties emerged from parliamentary politics without a formal organization linking supporters to the parliamentary party through a nation-wide extra-parliamentary structure. Many liberal and conservative parties in Europe long resisted the challenge of the socialist mass-membership parties, and fought elections with the aid of informal networks. In the early 20th century the US campaign-based parties showed that parties could operate without a membership organization, even in a mass democracy, if forced to do so by public regulations. Still, most parties today organize a membership, even though the distinction between members and supporters is often vague, as in the French Gaullist party in the 1960s or in Berlusconi’s Forza Italia since the mid-1990s.


Why do individuals want to become party members? The question applies only if membership requires an active choice. In the British labor movement a great difference has been found between members exercising an active choice and simply being signed up. When the rules for collective trade union membership of the Labour Party were changed from ‘opting in’ (declaring that membership is wanted) to ‘opting out’ (declaring membership unwanted) after 1945, the percentage of trade union members also affiliated to the Labour Party rose from 49% to 91% (Beyme, 1985: 175).

The calculus on the benefits of party membership depends on party type as well as on political and social setting. People join to gain influence, material favors, information, social benefits or mental satisfaction. Among the things members may lose are money, time and alternative opportunities. What is known about people’s motivations for joining a party is scattered; there is no generally acknowledged typology for it. Most studies build, however, on the Clark and Wilson (1961) distinction between material, social and purposive incentives for organizational commitment. Member surveys in the UK, Ireland, and Scandinavia show that people mostly express political (purposive) motives for joining—from ideological convictions to fighting for or against particular policies/politicians (Seyd and Whiteley 1992; Gallagher and Marsh, 2002; Heidar and Saglie, 2003). Paul Sabatier (1992) has developed a ‘commitment theory,’ according to which individuals join and become active in a political organization because of their strong ideological sympathies with the organization’s political goals. Purposive motives appear to be especially prominent in newly formed parties (Clarke et al., 2000). Some members also report social reasons, such as family tradition and social norms. A study of the Italian Socialist Party in the 1960s found that about one-third listed influence of family and friends as the reason for membership (Barnes, 1967). With movement networks declining, however, and with a weaker social element in party organization, social motivations for party membership may well have weakened (Ware, 1987). Studies of US party activism show that ideological motivation is more frequently present in affluent counties (Conway and Feigert, 1968). In patronage parties, such as early US local party organizations (‘party machines’), material incentives—for example, public employment or preferential treatment—were dominant. As late as 1990 the New York Times reported that to get a civil service job in Illinois, applicants had both to pass an examination and to get approval from a precinct captain and the county party chairman (quoted in Beck and Sorauf, 1992: 117). Today, European member surveys show that very few members express career benefits or material rewards as their motives for joining. However, these motives are less socially acceptable and therefore probably underreported. Motivational research is in any case difficult as motives may be vague, complex and volatile (McCulloch, 1990).

Who are the Members?

Who will become party members? The answer is—as for political participation in general -that the most resourceful individuals sign up for party membership. Within the specific social, cultural or geographical segments mobilized by the party, people with relatively high scores on education, income, and socio-economic status (SES) are disproportionately filling membership ranks (Widfeldt, 1995). But membership is not only pursued on an individual basis, creating the usual ‘high-SES’ biases. Parties are mobilizing agents, often organizing recruitment campaigns among special groups. European social democratic parties around 1900 often had significant numbers of leaders from the middle or upper classes, but still worked especially hard to enlist working-class members. They targeted the trade unions, the industrial plants, and the poorer neighborhoods. Ideology obviously was important, as the goal of the early socialists was to liberate the working class. Youths and women have always been groups targeted by many parties, as evidenced by the special party organizations often created for them. Some parties based their politics on promoting agrarian, religious, and ethnic interests, and sought to enlist members from these groups. Recruitment drives served two purposes besides increasing membership in general. First, parties sought to increase internal legitimacy by recruiting among the people on whose behalf they fought. Second, they sought to increase electoral legitimacy by giving voters a sense of social representation through party membership. When a party pursues a people’s party strategy its targeting of particular groups probably declines.

Another factor impacting membership profiles is competition from alternative organizational networks for a party’s recruitment base. In countries with strong organizations mobilizing low-SES citizens—for example, like an established network of organizations mobilizing broad segments among farmers, religions denominations, or ethnic groups—party recruitment (and party voting) may be enhanced among the lower-SES groups (Rokkan and Campbell, 1960). On the other hand, alternative organizations may be detrimental to general party recruitment by being more attractive, as when single-issue political action committees appear more attractive to people who ‘want things done,’ or when environmentalist groups are more attractive to young activists than the ‘generalist’ parties (Lawson and Merkl, 1988).

Women in the Party

In the latter decades of the 20th century, integration of women was a major goal of many parties. Some parties, such as the Icelandic Women’s Alliance established in the 1980s, were open only to women. But women’s parties are rare and usually small. Most parties have had predominantly male memberships. However, in many countries male membership is declining, particularly in Scandinavia. In Denmark and Norway, female membership rose from roughly one-third to roughly one-half the total membership from the early 1970s to 1990 (Sundberg, 1995). Several parties today have a majority female membership. At the other extreme, women in Tanzania (Tenga and Peter, 1996) and Malaysia (Rogers, 1986) struggle to be included in political parties at all. However, reports show rising levels of female involvement at lower party levels in those countries.

Causes of Change in Party Membership

An obvious starting point in explaining membership trends is the party’s general political support. Strong support among the people creates the potential for a large membership. Still, organizational factors enter the process: parties may increasingly/decreasingly be willing or able to recruit supporters. An example is the change by center-right parties from caucus into membership parties in order to contain the influence of socialist mass parties—what Duverger called the ‘contagion’ from the left. Changing organizational structures and the status of members may be a deliberate strategy of established party elites to pursue their interests (Panebianco, 1988: 191). Also contextual political, social and economic factors enter the process. Public regulations changing rules for collective membership—requiring individuals to opt in or opt out—have (as discussed above) huge effects on membership size. Public party finance may also have consequences for the number of members when subvention is based on membership. Indeed, membership numbers were falsified in Danish and Norwegian youths’ parties in the 1980s and 1990s in order to obtain public funds. Parties winning office or taking over the state, as in communist countries after a revolution, may offer security and career prospects that make them attractive to prospective members. Social forces may enhance membership, as when parties reach a ‘critical mass’ in a community, making it appear easy (or even necessary) to join the party—cf. Tingsten’s (1937) ‘law of social gravity.’ Finally, economic fluctuations may influence membership trends, although not always in clear-cut ways (Beyme, 1985: 175-88). Indirectly, the economic slump between the First and Second World Wars led to increased membership in fascist parties. Recessions after World War II did not. Two factors explaining declines in party membership in many Western European parties since the 1970s are the rise of affluence and the political consumerist attitudes prevalent in post-industrial societies.

Party Member Activities

What do party members do? Members’ activities vary substantially according to the nature of the party and also to how much time members have, their interests, and the opportunities available to them. The following concerns studies of parties in advanced industrial societies, primarily Western Europe and Canada. First, these studies find that many members join their party primarily as an expression of support, and that after joining they are inactive and have no intention of becoming active. In the German CDU in the 1970s roughly one-third of members attended at most one party function annually (Falke, 1982: 73). In Britain, up to 50% of Labour Party members, and 75% of Conservative Party members, reported being inactive during an average month. Most, however, took part in at least one activity during the five years preceding surveys (Whiteley and Seyd, 2002: 95-7). Surveys (1991 and 2000) of Norwegian party members showed that more than 50% did not take part in any party activities whatsoever during the preceding year, and that about 20% stated they had no intention of being active (Heidar and Saglie, 2003: 770).

Second, studies indicate that the proportion of members participating in party activities on a regular basis varies from 10% to 45% (Scarrow, 2000: 95). What constitutes ‘on a regular basis’ can be debated. For example, at the turn of the last century 7-8% of members in Danish liberal and Christian parties were classified as ‘active’ although they reported spending only slightly more than 5 hours monthly on party activities. Using the same criterion, 25% of Socialist People’s Party and Red-Green Alliance members could be considered ‘active’ (Pedersen et al., 2004: 375). In the 1990s, studies of British parties showed that, by the same criterion, 10-20% of party members could be considered ‘active’ (Whiteley and Seyd, 2002: 95). The criteria for being an ‘active’ party member, however, will vary according to both the particular party cultures and the methods of measurement used in the research. The way activism is operationalized will naturally reflect the analytical focus of the particular research, making cross-country comparative summaries difficult.

Third, members’ activities occur both inside and outside the party. Inside the party they attend meetings, engage in debates, and organize party affairs. Outside the party they take part in electoral campaigns, argue the party’s case at work, write articles and run for public office. Bringing out the vote was the main task of members in the early caucus parties and remains important—despite centralized media campaigns (Scarrow, 1996; Carty and Eagles, 2003). Offers of extra financial support to the party are often counted as activity, although one could argue that these are merely expressions of party support, much like when one joins the party without intending to be an active member. There are no clear ‘activity thresholds’ which allow for a simple, unambiguous definition of ‘party activity.’ Membership activity profiles come in all shapes and sizes. On the one hand, about 80% of Canadian party members attended branch meetings during the last year and more than 70% volunteered in election campaigns (Cross and Young, 2004: 440). On the other, about 25% attended a leadership convention and 6% sought a federal nomination.

Recent changes in information technology have made new activities available to party members (Römmele, 2003). The Internet has opened new communication channels and debate arenas. Members, however, appear slow to adapt to these new opportunities. In 2000 only about 10% of party members in Norway -which early on had, as a nation, a relatively high level of access to the Internet—visited their party’s home page at least once a month, and even fewer used e-mail in party affairs (Heidar and Saglie, 2003). However, among office-holders, and particularly among younger office-holders, the Internet was more widely used. Forty percent of office-holders aged under 40 used e-mail at least once monthly to keep in contact with fellow party members.

Fourth, there are different types of activists. Party activists may differ both in their type (internal-external) and in their level of activity (high-low). The ‘party builder’ would be high in intensity and focused on internal activities, while the ‘party supporter’ would occasionally argue the party’s case among friends, neighbors or workmates. Very little is known about different types of activists. In the old days, when people belonged to communist ‘cadre’ parties, the fascist militia or liberal ‘caucus’ parties, one assumed the differences in activist types to follow the party type. In contemporary advanced industrial societies, differences in levels of activism between parties are not that marked. Levels of activism vary just as much within a party as between parties. An extensive study on ‘high-intensity’ party members in Britain in the early 1990s found that, in the Labour Party, about 10% of members reported working for the party more than 10 hours monthly, while the figure was 5% in the Conservative Party (Whiteley and Seyd, 2002). In the research literature on specific parties or countries, levels of activism have been found to be in decline during the 1990s (Zielonka-Goei, 1992: 102). Whether this is a general trend, however, is hard to know. Systematic, comparative data are not readily available (Selle and Svåsand, 1991). Due to declining membership, parties may require more activity from remaining members to keep up their organization (Scarrow, 1993). In Norway, however, levels of party member activism were fairly stable from 1991 to 2000—although party membership declined significantly, suggesting that other contextual variables had an impact (Heidar and Saglie, 2003).

Why Differences in Party Activism?

One’s decision to join a party as a member (see above) and one’s decision to engage in various party activities may involve parts of the same calculus, but the two decisions need not be taken for the same reason(s). Both decisions are dependent on both supply and demand factors. Duverger (1964: 116) distinguished between totalitarian and restricted parties, that is, parties that demanded virtually total commitment and parties that demanded very little from their members. Party ideology plays a central part when party members are expected to act as the vanguard of historical necessities. The Leninist (totalitarian) party model demanded total involvement by members. This requirement made for an absurd situation when a new ‘party activity’ was introduced by the Soviet Communist Party in 1926: members were asked to repent their views. Stalin’s ‘organic theory’ of the party was put into practice when the Central Committee decided that a party task should be ‘to try to make the opposition bloc admit that its views are wrong’ (van Ree, 1993: 43).

A party generally wants its members to be active, although the reasons will vary both from party to party and over time according to the party’s required/needed level of activism. The ‘contagion from the right’ argument of Leon Epstein (1967: 260) asserted, with reference to Duverger, that modern media would cause the displacement of existing membership functions by new campaign techniques. This development, it was asserted by Epstein, would make European mass parties more similar to American campaign parties. By the same token, one could also argue that the advent of the Internet will change the character of future party activism (Party Politics, 2003). Finally, as discussed above, an organization’s size may affect member activism, as when a smaller membership must perform tasks previously performed by a larger membership.

Party members give different reasons for engaging in party activities. There are efforts to map their motivations (see above) and also to explain them. A standard explanatory approach is to analyze sociodemographic and socioeconomic motivations. In a study of British Labour Party members, Seyd and Whiteley (1992) compared active and inactive party members, and found a tendency for socioeconomic variables to be correlated with party activism.

A member’s activism may also be explained by his or her preferences for particular policies. In the terminology of rational choice theory, a party’s favored policies take on the character of a ‘collective good’ when implemented (Whiteley and Seyd, 1996: 218). Policy preferences interacting with an actor’s objective ability to make a decisive contribution towards a goal may explain activism. Rational choice theory predicts a positive correlation between party members’ ideological convictions and their levels of participation. A study of Danish party members, however, found only weak correlations between ideological radicalism and party member activism (Hansen, 2002: 191). Also, central to research on party activism is the notion that social norms induce party activism. The ‘expectations-values-norms’ theory sees actors as ‘embedded in networks of social norms and beliefs, which provide internal and external motivations to behave in certain ways’ (Whiteley and Seyd, 2002: 45). Hence, party members surrounded by family and peers are likely to be more active than party members who are not part of such networks. Party identification is another factor to consider, because it has proved to be a strong predictor of members’ activism levels. The basic notion is that party activism is not the result of party members’ cognitive (cost-benefit) evaluations, but rather of members’ loyalty and affection for a group or party (Whiteley and Seyd, 2002: 55). Finally, it is likely that some party activists are driven by political ambition, either to improve their own welfare or to improve the welfare of others. ‘The relevance of political ambition for party organization should be obvious’ (Schlesinger, 1991: 33). The argument assumes that the individual’s behavior is explicitly goal-directed. As a theoretical explanation for party activism, political ambitions can be seen as the antithesis of a theoretical explanation stressing ‘expressive incentives.’

With the exception of socioeconomic variables, the explanatory factors described above have been merged into one overarching ‘general incentives model,’ and Whiteley and Seyd have on several occasions applied this model to explain party activism in British parties (Seyd and Whiteley, 1992; Whiteley et al., 1994; Whiteley and Seyd, 2002). Based on member surveys, their conclusion is that the general incentives model tends to outperform its rivals. They also found support for their general incentives model in their study of Irish Fine Gael members (Gallagher et al., 2002). They concluded that the ‘SES model adds nothing to what can be explained by the general incentives model’ (p. 111). A test of the general incentives model based on survey data of Danish party members found that it fits the facts better than did other models (Hansen, 2002: 251-59).

Other models fare no better or worse than the general incentives model. Empirical studies suggest that a number of factors—even theoretically antithetical ones—are relevant in explaining party activism. A supplementary explanatory approach that, so far, has received surprisingly little attention in empirical research focuses on the party variable. Since the work of Duverger, it has been generally accepted that different party types, or ideological party ‘families,’ have distinctive participatory cultures, for example totalitarian versus restrictive parties. Bringing back ‘party’ or ‘party family’ as an explanatory variable will introduce a version of the ‘expectations-values-norms’ theory. Different parties are expected to attract different kinds of people, and to shape them through different party cultures. The German Greens attracted high-intensity members in the 1980s (Poguntke, 1992). Research on Norwegian party members has shown that party is strongly correlated with levels of party member activism, even when controls are made for SES variables (Heidar, 1994: 76).

Are Party Members Special? May’s Law of Curvilinear Disparity

May (1973) argued that ‘sub-leaders’ in parties hold more extreme views than both party voters (non-leaders) and the party leadership. His position differed from the standard one, which assumed that party leaders held more ‘ideological’ or ‘party correct’ views than their followers (McClosky et al., 1960). May predicted, and claimed to find, hierarchical contrasts between the sub-leaders and others in opinions on policy alternatives. Party sub-leaders would, according to his theory, be devoted activists. Their recruitment and socialization would make them more ‘ideological’ than rank-and-file supporters. On the other hand, top leaders must moderate (or appear to moderate) their views for two reasons: compromises are necessary in public office; and competition for moderate voters. One could reasonably expect to find, according to May’s law, empirical differences in political views at different hierarchical levels in parties. The expectation that one would find ‘militants,’ ‘ideologues’ or ‘true believers’ among party activists is commonplace among political commentators. But does empirical research confirm such expectations?

Studies within both the formal context of European party organizations and the informal context of US party organizations suggest that the answer is ‘no’ (Norris, 1995: 33; Herrera and Taylor, 1994). Norris studied the British Conservative and Labour parties, focusing on party candidates for the 1992 election, members attending selection meetings for candidates, and party voters. She found that members of both parties tended to hold views located between the moderate voters and the more radical leaders, and that the relationship between the different layers in the party was more complex than suggested by May First, the motives of the party sub-leaders and senior leaders were more varied (than suggested by May’s law)—as were the forces shaping political opinions at different levels. Norris found that high ideological commitment was among the major factors inducing party leaders to stand for election. On the other hand, one might expect that the sub-leadership faced moderating forces when fighting local elections. Observers also easily overrate how representative the extremist sub-leaders and their factions are within parties: they are usually very vocal in order to put pressure on the leadership, but do not necessarily speak for the average party member. Norris explains the negative test of May’s law with the argument that mixed ideological and electoral incentives shape the opinions of both party leaders and their members. A study of Irish Fine Gael shows that the members were far from being extremists (Gallagher and Marsh, 2004). Instead they held ‘impeccably middle-of-the-road opinions’ (p. 418). A study of several Norwegian parties of political views at different levels on a number of issues found a curvilinear pattern in one-third of the cases, but failed to find a clear pattern—which left the authors wondering why curvilinearity occurred in some cases but not in others (Narud and Skare, 1999).

There is no final verdict on how special the political opinions of party members (or sub-leaders) are within parties. Even if the literature has failed to produce a consensus on May’s law (Scarrow et al., 2000: 131), the law of curvilinear disparity, with its ‘grain of truth’ and clear predictions, continues to generate empirical research on opinion formation processes within political parties. Perhaps one need to be reminded that May presented more than one hypothesis on how opinion formation takes place within parties.

Michels’ Law of Oligarchy

The research inspired by Robert Michels’ ‘law of oligarchy’ is both older and more voluminous than the research inspired by May’s law. While May basically held an optimistic view on the impact of members—it matters what sub-leaders/members think and do—Michels was essentially pessimistic, arguing that neither leadership’s rules nor members’ opinions matter. In its original version, Michels’ law stated: ‘to say organization is to say a tendency to oligarchy’ (Michels, 1925: 25; quoted in Beyme, 1985: 232). In the English book based on a translation of the Italian edition of Michels’ work, the ‘fundamental sociological law of political parties’ is formulated in the following terms: ‘It is organization which gives birth to the domination of the elected over the electors, of the mandataries over the mandators, of the delegates over the delegators. Who say organization, say oligarchy’ (Michels, 1962: 365). In this spiced-up version, the moderating ‘tendency to’ has been left out, which makes a substantial difference. But as with May, it is the forceful hypothesis embedded in Michels’ law, the argument that democratic parties do not and cannot exist, that has made his law of oligarchy so widely researched and hotly contested.

Michels analyzed the ‘new politics’ of the early 20th century that emerged with the advent of mass suffrage and extra-parliamentary parties. While Ostrogorsky (1902) had discussed how permanent party organizations outside parliament would pervert the reasoned debate among elites, Michels questioned whether the new mass parties really mattered politically, since the oligarchs ruled anyway. He argued -using the institutional approach of his mentor Max Weber—that the party leadership governed the party organization by necessity. Creating an organization would in itself create the basis for an oligarchy. The leaders would control the decision-making process and the channels of information, and they could manipulate the support of the uninformed and unprofessional membership, making empty rituals of formal democratic policy-making processes.

To prove his case, Michels selected for his empirical research the German Social Democratic Party, a party with a strong claim to being internally democratic. If oligarchy still prevailed in that party he argued, then his law would also apply to parties with less or even no intention of letting their members influence policies. To some, his law has appeared self-evident and applicable to all parties. E.E. Schattschneider found it hard to imagine what a democratization of the US parties actually would entail, let alone ‘whether democratization, if it were possible, would be appropriate to the legitimate functions of the parties in a modern political system’ (1942: 58). If true, if internal party democracy were impossible, there would be no point in studying party members in order to understand party decision-making. What happens inside the party would not impact the policies pursued by the party leaders. The study of members would, of course, still be interesting for other reasons, such as seeking to understand the basis for and dynamics of grassroots activism, the recruitment processes, the potential for the mobilization of voters, etc.

Michels’ law has been the starting point for numerous studies. As with May’s law, however, this is not because the studies generally confirm that members are unimportant, but because the ‘grain of truth’—or the ‘tendency to’—gives a useful analytical reference point for empirical studies. Most empirical research, in fact, concludes that even in parties dominated from the top, the party leadership cannot afford to overlook completely the political opinions of its members, regardless of the formal structures of the party organization (McKenzie, 1955). Michels did, however, also provide an extensive list of (researchable) factors that may limit the members’ opportunities to influence party decisions, such as members’ background and resources (Barnes, 1967), organization size (Tan, 1998), level of institutionalization (Panebianco, 1988), leadership types (Weber, 1964), etc. A study of the party organizations in a number of mature democracies summed it all up: ‘There are now many instances around the democratic world where party leaders operate a coalition of power in which grass-roots members are significant junior partners’ (Scarrow et al., 2000: 149). The impact of members varies, of course. Research indicates that, in the new Eastern European democracies, member input is fairly limited (van Biezen, 2000). In parties with weak or no membership, the discussions focus on relationships between leaders and activists or followers (Eldersveld, 1964). Changes in the party environment also impact the degree of member influence. State financing of parties may reduce the need to accommodate members, but little evidence for this is offered (Pierre et al., 2000). The Internet may be used both to enhance member influence and to strengthen leadership control (Party Politics, 2003). The Internet could make direct democracy within parties more workable, but any evidence of this is far from conclusive (Party Politics, 1999). It would also be difficult for members to influence, for example, candidate selection in the media-driven electoral campaigns of early 21st-century politics (Party Politics, 2001).

There are several ways to do empirical research on members’ participation in decision-making processes within political parties. The traditional approach is to study particular political issues to determine how the process evolved and who influenced the final outcome. Michels’ book is full of such cases. Another approach (also adopted by Michels) is to study organizational rules (Katz and Mair, 1992). Researchers can also interview or survey party leadership and/or members to get their evaluation of how the organization works (Party Politics, 2004). Finally, one may rely on ‘expert opinions’ to compare degrees of centralization of power in parties (Janda, 1980). One should note, however, that members’ influence may differ in different aspects of party work. In a study of party rules in about 18 democracies, members’ influence patterns turned out to be different for candidate selection, leadership selection, and policy-making (Scarrow et al., 2000).

Consequences of Party Membership

Party membership has an impact on party processes, the leadership, and the members themselves. The German Greens were in the 1980s very much occupied with building an organization that sustained members’ influence and hampered the ‘oligarchic tendencies’ of a party organization. Consequently they instituted collective leadership, rotation, and direct democracy, placing severe restrictions on their leadership, which they suspected of being unreliable (Poguntke, 1994). The effects of party members on internal party political decision-making have been discussed above; in this section I will look at the effects of party membership on the individual member and on the political system/society

Effects on the Individual Member

As noted, some party members are unaware of their membership, and consequently it does not affect them as individuals. At the other extreme (if one is aware of one’s party membership) such awareness can have dramatic effects. In illiberal regimes it may lead to persecution, imprisonment, even execution/assassination. Tsarist Russia jailed and exiled communist agitators in the early 20th century, just as the Soviet Communists jailed, hospitalized, and exiled their dissenters. In liberal regimes the consequences of membership are generally less drastic, although Communist party members in Western countries could lose their (non-party) jobs during the most intense periods of the Cold War.

The goals motivating one to become a party member (rectifying injustice, working with others on important matters, personal gain) may actually be realized by one’s membership in the party. Studies of Chicago machine politics during its heyday in the 1920s and 1930s showed that over half the precinct captains held public sector jobs (Gosnell, 1937; Epstein, 1967; Crotty 1986). Keeping those jobs depended on their ability to bring out the Democrat vote. The spoils from winning power generally go to a small number of top party politicians (locally and nationally) who enter public office, but electoral success may also bring spoils to a (varying) number of political-administrative personnel. Mostly these are party leaders, but at the lower levels party membership may be the qualification that in the end decides who gets the job.

The mechanisms through which party membership may benefit the rank-and-file members are as manifold as human imagination and corrupt practices allow. Access and friendships within the inner circle of top financial elites under the socialist governments in France under Mitterrand (1988-95) depended on social prestige, residence, and party membership—although most importantly on graduation from the school the École Nationale d’Administration for top bureaucrats, (Kadushin, 1995). In communist regimes, one’s job and career opportunities depended on membership in the party. After the Bolsheviks won power in Russia, they founded the Institute of Red Professors in 1921 to educate the new socialist intelligentsia of the Soviet Union (Fox, 1993). In Communist China, party membership influences recruitment into administrative and managerial positions (Bian, 1995). Evidence points, however, to a dual career pattern in China, whereby membership is always a prerequisite for administrative positions, but does not necessarily enhance professional careers (Walder et al., 2000).

On Polity and Society

The sum of micro membership experiences has macro consequences for the political system. Party membership may educate members or make them more cynical. Either development will affect the nature of political debate. Membership may give members a stake in the system—creating positive feedback on participation and legitimacy. Engaged members will take part in local-level electoral campaigns, bringing out the vote beyond what can be achieved by sophisticated national campaigns focusing on branding and personalities (Carty and Eagles, 2003). At the aggregate level, however, it is difficult to find clear relationships between party membership size and electoral strength (Scarrow, 2000). Party experiences might also frustrate members, causing them to go to the sidelines of politics for a while (Hirschman, 1982), or causing them to switch parties. If such switches occur in large enough numbers, it would ultimately result in the demise of old and the rise of new parties.

According to the participatory democracy school of thought, party member participation and debate in internal party affairs will provide a link between civil society and politics that supplements the link provided by competitive debate at the party system level (Teorell, 1999). The organizational encapsulation, by labor organizations, of the newly enfranchised, largely apolitical and underprivileged masses, helped to stabilize the new mass politics. These organizations contribute to making cross-national allegiances less important; shift the locus of conflict from the economy to the political arena; increase participation; and consequently strengthen the legitimacy of liberal democracy (Rokkan, 1962; Wellhofer, 1981). The argument that parties induce system stability and legitimacy has also been made, citing cases as diverse as the Rural African Party in Tanzania during the 1960s (Miller, 1970) and clientelist politics in the Philippines (Nowak and Snyder, 1974). The party organization in these cases and others provided services to its members and supporters while at the same time creating mobilization, control, and stability at the level of the political system.

Party membership may have more general implications for the individual and the society at large. To maintain its control over the Soviet Army, the Russian Communist Party filled its officer corps with loyal party members, and then constantly educated them in communism (Brzezinski, 1952). More commonplace is the ‘partyfication’ of public administration. In the USA, urban political machines were once heavily staffed with party members; incoming presidential administrations gave government and other jobs to loyal party supporters. Similar practices occurred in several European countries, including Austria and Belgium, where national ministries and public service apparatus also were once filled with party supporters (Daalder, 1987). These practices, however, are in decline due to civil service reforms.

Finally, one may ask whether democracy requires or needs party members. How ‘democracy’ is defined will play a big role in determining the answer to this question, but the standard answer would be ‘no.’ For example, for elections to be free and fair under political systems with universal suffrage, it is not necessary that competing parties have party members (or that competing candidates be party representatives, for that matter). However, the quality of democratic processes might be improved were more citizens to be party members, but that would depend, naturally, on the quality of party democracy case by case.

Party Member Research in the Future

Research on party membership is part of a broad effort to understand parties, citizen participation, and democratic processes. Sustained empirical research is required to improve our understanding of the individual party member and the institution of party membership and the effects it has on political processes. Most research on political parties has a European bias (Diamond and Gunther, 2001). The same bias applies to most research on party members. Therefore, more studies of party members and party membership in countries besides the advanced industrial countries of Europe, with their established liberal democracies, are called for. Such studies will call into question the accepted understanding of what the phenomenon of ‘party membership’ is, and will demand a more empirically applicable typology of membership than the variable geometry employed by Duverger.

Beyond extending the studies of membership in space and—to the extent possible—in time, the research would probably be most fruitful for general political science if it were pursued along two main lines: studying party membership as ‘political participation’ and party members as ‘political agents.’ Studying membership as participation would mean approaching this in the same way as other researchers approach voting. The relevant questions would be: Who are the members? How much do they participate? Why do they take part in varying degrees? One recent example of this line of research is the study of ‘high-intensity’ members’ participation by Whiteley and Seyd (2002). Studying members as political agents, on the other hand, would raise questions about the impact of member activity. The original questions were posed by Robert Michels: How is decision-making conducted in political parties? And to what extent do members influence decision-making? Answers to these questions will be difficult to find because searching for them places the same demands on researchers of political parties as those which have always been placed on researchers of power in general. They must answer the question: Who governs? In this field there is an obvious need for more empirical studies of the decision-making processes in political parties.