Clientelism and Party Politics

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


Clientelism is a term which describes the distribution of selective benefits to individuals or clearly defined groups in exchange for political support (although as we shall see below, there are a range of more precise definitions). Clientelism occupies a curious position in the scholarship on party politics, and there is a degree of confusion over its relationship to the electoral process. On the one hand, it appears to be associated with pre-modern social contexts and is therefore connotated with cultural and economic ‘backwardness.’ On the other hand, clientelistic dynamics can frequently be found in the most apparently advanced socio-economic contexts, as attested by the abundant scholarly literature on ‘pork-barrel’ political exchange in the contemporary United States. Moreover, the confusion is exacerbated by the wide and diverse range of political exchanges which can be accommodated by the concept of clientelism.

The first objective of this analysis is therefore to present a more precise conceptualization of clientelism. It will then proceed to examine the origins and modalities of political clientelism, and assess its implications for party democracy.

Conceptualizing Clientelism

Stripped down to the essentials, clientelism is a form of personal, dyadic exchange usually characterized by a sense of obligation, and often also by an unequal balance of power between those involved (see Eisenstadt and Roniger, 1984: 48-9; Piattoni, 2004). This definition reflects the origins of the concept as a descriptor of hierarchical patron-client relationships in traditional rural societies (Piattoni, 2001a: 9). These relationships involve ‘the patron providing clients with access to the basic means of subsistence and the clients reciprocating with a combination of economic goods and services (such as rent, labor, portions of their crops) and social acts of deference and loyalty’ (Mason, 1986: 489). In other words, clientelism is a way of describing the pattern of unequal, hierarchical exchange characteristic of feudal society, in which patrons and clients were tied to durable relationships by a powerful sense of obligation and duty.

This kind of ‘notables’ clientelism (Tarrow, 1967)—which is also referred to as ‘old clientelism’ (Weingrod, 1968)—survived into the modern democratic age in many parts of the world, and therefore became enmeshed in the dynamics of electoral politics. In clientelistic contexts patrons, or their agents, stand for election and their clients vote for them, sometimes out of a general sense of obligation and attachment, sometimes as part of a specific exchange for services rendered or promised. Either way, voting behaviour is a function of a personalized and instrumental view of political participation: voters simply use their vote to sustain their patrons, thus earning the patrons’ protection and help. Voters neglect the broader political consequences of their electoral choices, and representatives elected through clientelistic mechanisms cannot credibly claim a mandate to pursue a broad programme of public policies. Moreover, the rigid and unchanging nature of traditional clientelism undermines the ‘feedback’ function of electoral politics, making alternation in political power an artefact of elite decisions (as, for example, in the turno pacífico of late nineteenth- and early twentieth-century Spain, or the trasformismoof Italy in the same period).

In some cases, clientelism has evolved into something quite different from this kind of traditional social exchange. Studies of postwar Italy in particular have suggested the emergence of a ‘new clientelism’ (Tarrow, 1967; Weingrod, 1968; Caciagli and Belloni, 1981), in which political behaviour is still characterized by patterns of exchange, but of a new kind. Socioeconomic modernization brought greater geographical mobility and urbanization, higher levels of education, the replacement of agrarian by industrial employment, and the decline of traditional rural elites. These developments weakened traditional patron-client ties, which made way for new forms of exchange. Organized political parties, with relatively bureaucratized structures, replaced landlords and local notables as patrons. Clients, enjoying higher living standards and less instinctively deferential, demanded more immediate material benefits in exchange for their votes. In this new, ‘mass party’ clientelism, patrons have to ‘buy’ votes by distributing concrete excludable benefits and favours to individual voters or groups of voters. In the Italian context this is referred to as the ‘vote of exchange’ (Parisi and Pasquino, 1979; Katz, 1986).

This new clientelism shares some of the features of the old. The relationship is still instrumental, and the benefits provided to clients are still largely private and excludable. But there are also important differences. First, the relationship is less hierarchical, more ‘democratic.’ There remains an imbalance of power, in that the patron has control over resources that the client needs, but there is less of a sense of deference and dependency on the part of the client, who feels increasingly free to use her vote as a commodity to be exchanged for whatever maximizes her utility. Second, as a result of this less hierarchical and personalized context, the new clientelism is more conducive to fluidity and change in electoral behaviour, opening up possibilities of greater competition and elite turnover.

The differences between these two types of clientelism are significant enough to undermine the precision of the concept. The client who votes automatically for her patron out of a sense of deference, and the implicit and imprecise promise of protection and aid, is a very different social actor from the client who shops around for the patron who offers the best deal, and may even switch patrons if the flow of benefits dries up (see Allum, 1997). The old clientelism is very much a form of social and political exchange, in that it ‘involves the principle that one person does another a favor, and while there is a general expectation of some future return, its exact nature is definitely not stipulated in advance’ (Blau, 1964: 93). The new clientelism instead resembles ‘economic’ or ‘market’ exchange, in which the client seeks to maximize utility irrespective of any sense of obligation towards or identification with another actor. Gellner (1977: 5-6) draws the distinction with striking clarity:

Economic benefits are, at least ideally, calculable, noncommital and single-shot: hence an economic operation is isolable, and does not need to give rise to any permanent relationship. … By contrast, the long-term imponderables which are being ‘exchanged’ in a political relationship, ipso facto give a much deeper colouring to the links between the parties to the transaction.

Although clientelism will rarely be exclusively of one kind or another, the extent to which ‘economic’ dynamics prevail has major implications for party democracy. If the only reason for supporting the party is a direct economic exchange which excludes feelings of loyalty or ideological affinity, there is little to anchor the clientele to the party if benefits are not forthcoming. This opens up the possibility of political instability and upheaval, rather than the continuity or even stagnation more often associated with clientelistic political systems.

Why Clientelism?

It is difficult to measure the role of clientelism in party politics in a given political system accurately (Wantchekon, 2003). Some statistical analyses, particularly for the United States, have studied budgetary allocations across territory in order to estimate the flow of selective, excludable benefits to particular localities (for example, Stein and Bickers, 1995; Frisch, 1998). However, such analyses focus on benefits to groups of varying size, and can rarely detect the kind of personalized, one-to-one exchange that is characteristic of much clientelistic politics. Most of our knowledge and understanding of clientelism in its most personalized form therefore comes from case studies which have used an ethnographic approach to identify the presence of clientelistic dynamics (for collections of such work, see Gellner, 1977; Eisenstadt and Lemarchand, 1981; Eisenstadt and Roniger, 1984; Piattoni, 2001b). Such case studies, naturally enough, select on the dependent variable, leaving us with little basis on which to assert that some political systems are more clientelistic than others. However, it is broadly held that advanced Western democracies are less clientelistic than developing countries, and that amongst the advanced democracies, those in Northern Europe and in the English-speaking world are the least clientelistic, although there is little in the way of hard comparative evidence to back up these claims.

A variety of explanations have been advanced to account for this apparent variation in clientelism across nations (for extensive reviews see Piattoni, 2001a, 2004). Structural explanations identify long-standing institutions and patterns of behaviour which prevent political systems from breaking out of practices of clientelistic exchange. Most famously, Putnam’s (1993) account of the varying performance of regional governments in Italy explained corrupt and clientelistic party politics in terms of the presence or absence of reserves of social capital. In some regions (in the Centre and North), traditions of ‘civicness,’ originating in early experiences of representative political institutions in the medieval period, established a pattern of associational-ism and collective action which formed the basis for public interest-oriented behaviour by voters once electoral democracy took root. In these areas, citizen participation in the political process is motivated by a sense of responsibility towards the community, and the electoral process revolves around competing packages of collective goods. In the ill-governed South of Italy, in contrast, no such civic tradition has been established, and citizens find it hard to overcome collective action problems. As a result, voters do not trust political parties to provide collective goods, and tend instead to seek selective, private benefits in exchange for political support (see Banfield, 1958, for a classic description). For Putnam, the absence of a tradition of ‘civicness’ is the prime cause of clientelism: without such a normative compass, citizens will tend to see the democratic process as an opportunity to win excludable benefits for themselves, and political parties will see it as their job to allocate such benefits in exchange for votes. Similar normative assumptions underpin the classic interpretations of ‘machine politics’ in American cities, where clientelism was directed at ‘working class people, especially immigrants unfamiliar with American ways and institutions’ (Banfield and Wilson, 1963: 118).

This view has been roundly criticized for its determinism, as well as its alleged empirical inaccuracy (Levi, 1996; Sabetti, 1996). Putnam’s interpretation of Italian history fails to account for a number of inconsistencies, and also fails to consider the role of socioeconomic marginal-ity (Pizzorno, 1971) in depressing civic endeavour in the geographically isolated South. Putnam’s broad-brush, structuralist approach is countered by other scholars who have placed greater emphasis on historical contingency and strategic interaction. Martin Shefter (1994) has argued that the extent of clientelistic mobilization by political parties is a function of the sequencing of the process of democratization: in countries where political parties and competitive elections emerged before the creation of an autonomous bureaucracy, politicians could exploit the resources available in the state administration for partisan ends, allocating public jobs on political grounds (to party activists and loyal voters). Filippo Sabetti’s (2000) analysis of the emergence of the Italian state shows that the process of institutional design and development is complex and multidimensional, and that the failure to decisively implant ‘good government’ in Italy was a contingent outcome rather than the inevitable result of its historical legacy. Simona Piattoni (2004) has countered deterministic views of clientelism and its origins by employing game theory to explain how the process of clientelistic political exchange can lead to a variety of outcomes, including the (more or less) optimal provision of public goods by government.

Given the difficulties involved in identifying and measuring clientelism, it is unlikely this debate can easily be resolved. Both the structuralist and the strategic interpretations of clientelism run into serious theoretical and empirical problems. Reductive determinism, especially analyses such as Putnam’s which see political systems as ‘doomed’ (Sabetti, 2000) to clientelism by distant events such as medieval conquests, lacks an adequate account of the mechanisms through which structures are reproduced and maintained. Strategic explanations, in contrast, neglect the role of structure in defining the ‘games’ within which actors develop their strategies. Empirically, structuralist theories need to account for the persistence of clientelism in supposedly virtuous England until well into the nineteenth century (O’Gorman, 2001), and for the extensive evidence of apparently vibrant ‘civic’ behaviour in the Italian South at various times since the medieval period (Sabetti, 2000). Strategic analyses need to show that the abundant evidence of clientelistic dynamics in postwar Italian parties, and the lack of such evidence for the Swedish parties, for example (Papakostas, 2001), can indeed be explained in terms of the peculiarities of strategic choice, rather than the very different socioeconomic conditions in which democracy developed in the two countries. Up to now there has not been enough systematic comparative-historical research to resolve these uncertainties.

Varieties of Clientelism: Patterns of Exchange in Electoral Politics

On a practical level, clientelism manifests itself in a variety of ways, depending on the kinds of resources available to patrons, and the kinds of citizen demands that have to be met. In the most traditional contexts, clientelism could draw on age-old reserves of loyalty and deference, so that patrons could obtain political support from their clients without providing too many concrete benefits. Banfield’s famous study of the Southern Italian village of ‘Montegrano’ in the 1950s found that ‘just before elections the Christian Democratic party distributes small packages of pasta, sugar, and clothing to the voters’ (1958: 26). In situations of dire poverty, such gifts may be enough to buy votes, particularly if there is a pre-existing foundation of deference towards the patron. The patron-client relationship in the rural context is not strictly reliant on the distribution of specific material benefits, however: for Tarrow (1967: 68), the patron is ‘a support in time of famine, his advice will be formally sought before marriages and land purchases, and he is asked for recommendation in the peasant’s frequent encounters with the bureaucracy.’ Where this type of ‘old’ clientelism is well entrenched, the establishment of democratic politics will be conditioned by the local notables’ mediating role between voters and the state. Patrons become proprietors of ‘packages’ of votes, and are able to trade these votes with the leaderships of the political parties, linking themselves with a larger network of clientelistic relationships. This kind of party organization compromises the cohesiveness and durability of the political party: factionalism and weak leadership authority are likely to result.

The ‘new’ clientelism, characteristic of more economically advanced settings, is subtly different in its impact on party politics. Whilst the local notables of the ‘old’ clientelism are sources of political legitimacy in and of themselves, the local party bosses in the ‘new’ clientelism have far less autonomy. Notables can conceivably change their political affiliation without compromising their relationship with their clientele, since what matters for voters is their continued willingness to provide protection and assistance. Local party leaders in the ‘new’ clientelism are far more dependent on their party affiliation to maintain their clienteles, and are therefore less inclined to change parties. The ‘machine politics’ of the American cities in the nineteenth and twentieth centuries illustrates this point. Banfield and Wilson’s (1963: 118-19) description of the big city machines in the 1950s shows how the party representatives are embedded in a formal organization with its own hierarchy and career structure:

The job of the precinct captain is to get out the vote for his party’s slate … [he] is chosen by and works under the direction of a ward leader, usually an alderman or elected party official …. It is up to him to … dispense the larger items of patronage, favors, and protection to those who have earned them. … Captains are often ‘payrollers,’ that is, they have appointive public jobs that they could not get or keep if it were not for the party …. The hope that the party will in due course run them for alderman keeps these captains at work.

In the new clientelism, the patron is the party organization, rather than any individual within it. Clientelistic favours are distributed by members of the party organization, who in turn receive authorization for this activity from the upper tiers of the party hierarchy. Clientelism therefore becomes bureaucratized, and less personalized, although the personal contact between party representatives and individual voters remains important for maintaining the relationship. In this respect, mass party clientelism is a significant departure from notables’ clientelism.

The new clientelism is consistent with internal party cohesion and formalized chains of command. Indeed, the case of the Italian Christian Democrats in the 1950s and 1960s shows that clientelism can be a key tool of party institutionalization. DC leader Fanfani used clientelism to overcome the party’s dependence on external interest groups and bureaucratize its organization, establishing systematic patterns of resource distribution through the party structures. Since the party controlled the key spending ministries in Rome, local party bosses on the ground needed to cement close ties with national leaders in order to gain access to the clientelistic benefits they distributed to their voters. This system, over time, became quite sophisticated, and similar techniques were also adopted by the Italian Socialists after they became a mainstay of government coalitions in the early 1960s.

The new clientelism is closely associated with the expansion of the economic and social role of the state. In traditional contexts, the state often has a more limited role, particularly in regard to its expenditure. Notables are often deployed by their clients to help with the bureaucratic requirements of the state, such as conscription, rather than to access material benefits. As the state’s role has expanded in much of the world to involve a detailed regulation of economic activity and the provision of a wide range of financial benefits (welfare and pensions, industrial and agricultural subsidies, public housing) and public services (education, health), the parties governing the state have had a greater ability to manipulate and channel these resources in exchange for political support. Often, the parties seek to make the criteria for the access to resources deliberately opaque, in order to enhance the discretionary nature of the distribution and extract greater political returns (Tanzi, 2000; Golden, 2003). Once again, the Italian case is one of the best-documented: standard practices ranged from strictly partisan allocation of jobs in the state-run postal service or railways, to the selective distribution of bogus sickness pensions and a variety of subsidies and development projects of questionable utility. Similarly, in Spain, the Socialist party established a rural employment subsidy (the PER—Plan for Rural Employment) which was directed at its own electoral strongholds in the South, and which gave local mayors a large degree of discretion over the allocation of the money (Hopkin, 2001: 128).

The growth in the role of the state has also led to a vast expansion in state personnel, which in many cases has been exploited by political parties to give jobs to their activists and supporters—what Lyrintzis (1984), in his analysis of the Greek case, calls ‘bureaucratic clientelism.’ A high profile example of this is the case of Jacques Chirac’s tenure as mayor of Paris, during which his party allegedly gave ‘no show’ council jobs to a number of activists who actually continued to work full-time on party business. In Austria, jobs in the state bureaucracy have been routinely allocated on the basis of party affiliation (Mueller, 1989). The Spanish Socialist Party, which won power in 1982 at a very early stage in its organizational development, by 1987 had appointed around 25,000 new state functionaries, bypassing the public administration’s normal recruitment procedures; many of these jobs went to party supporters (Hopkin, 2001: 126). Although patronage in the allocation of state jobs can be extensive, it cannot alone underpin a clientelistic electoral strategy, and, in the European case at least, it seems to have been deployed most often to shore up party organizations by providing salaries for committed party workers and facilitating party control of policy implementation. To this extent, clientelism may play an important role in bolstering the institutions of party democracy.

Clientelism and Party Democracy

Most scholars have stressed the negative implications of clientelism for party democracy. Positive interpretations, noting clientelism’s ability to link political representatives to citizens, and to provide a mechanism for ensuring ‘constituency service’ (Cain et al., 1987) and party response to immediate citizen needs, have been far less common. The main reason is that the use of the vote as a currency to buy material benefits subverts the obstensible purpose of the electoral process in a representative democracy. Many populist conceptions of democracy envisage that citizens cast their votes in terms of their own understanding of the ‘public interest,’ and that the most widely shared view of this public interest will inform political decision-making through the workings of the representative institutions. In fact, this idealized view is conceptually problematic, as the social choice literature has shown: according to Kenneth Arrow’s ‘impossibility theorem,’ there is no adequate decision rule for aggregating individual preferences, and therefore no way of establishing what the ‘public interest’ actually is (Arrow, 1951; see also Schumpeter, 1994). This problem strengthens the case of liberal theorists who dismiss notions of collective interest and instead stress the sovereignty of individual choice in electoral politics (for the liberal-populist distinction in democratic theory, see Riker, 1982).

From a populist position, clientelism is criticized because it gives primacy to the distribution of individual, selective benefits to citizens, to the detriment of the provision of collective goods. Most populist theories emphasize equality of access to the political process, and the very unequal distribution of benefits in clientelism violates this principle. To take one example, the expansion of the welfare state in clientelistic polities has led to a wide range of inefficiencies and injustices, with some individuals and groups benefiting disproportionately at the expense of others, for the sole reason that their votes were deployed in clientelistic exchange. Moreover, if votes are cast purely in terms of the benefits received, then this leaves governing parties free to disregard popular opinion in all policy decisions which do not relate to the direct allocation of resources in exchange for votes. For this reason clientelism has often been associated with authoritarianism. Liberal theorists are less concerned with the inequality of outcome characteristic of clientelistic resource distribution. However, liberal theory does tend to stress equal citizen rights, and therefore liberals can object to clientelism on the grounds of the differential access to the political process that results from clientelistic exchanges. A further liberal objection is that clientelism has often involved the extensive deployment of government resources to satisfy clienteles, which implies heavy government intervention in private property rights in order to raise the necessary revenues. In short, clientelism is generally an unwelcome phenomenon from the point of view of mainstream normative democratic theory.

The picture is more ambiguous from the point of view of the practical implications of clientelism for democratic politics. Clientelism has been associated with both excessive continuity and violent change in party systems. The examples of Italy and Austria provide a neat illustration of this ambiguity: until the 1990s both countries were widely perceived as being locked into an immovable party system cartel by the mechanisms of patronage, clientelism and interparty collusion (Mair, 1997). But in the 1990s, both party systems underwent turbulent changes (spectacularly so in the case of Italy), changes which were widely interpreted as the result of voter protest against the clientelistic party cartels. However, periods of stifling continuity followed by abrupt change have also been noted in apparently far less clientelistic party systems (for instance, the UK). There is little strong comparative evidence to blame clientelism for the difficulties facing party democracy in advanced industrial democracies.

Clientelism is essentially a variant of ‘special interest politics’—a mechanism through which political parties and their representatives can obtain political support in exchange for selectively allocating benefits through state institutions. In many ways, it gives less cause for concern than the opaque money-raising practices of many contemporary parties which are willing to tailor public policies to corporate interests and various other lobbies in exchange for money. Corrupt party financing subverts citizen equality by allowing the wealthy to buy political favours which redistribute further advantage to them. Clientelism instead often allocates benefits to the least privileged, and since these clients often have little more than their vote to trade, the redistributive consequences of any specific clientelistic exchange will tend to be less significant. However, mass party clientelism on a large scale is ultimately both inegalitarian (because it does not respond to universalistic criteria) and economically unsustainable (because it feeds a continuing demand for redistribution). The case against clientelism as a form of linkage in party democracy therefore remains strong.