Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 17 February 2020

Political Clientelism

Abstract and Keywords

This article studies political clientelism. In the first section, the term clientelism can be defined as giving material goods in return for electoral support, where the criterion of distribution that the patron uses is simply: did you/will you support me? This section includes definitions of vote buying and patronage. The two waves of studies of clientelism and the link between clientelism and commitment are discussed in the subsequent sections. In the latter half of the article, the discussion tries to determine if clients are swing voters or core supporters, and looks at the causes and consequences of clientelism.

If most scholars of the topic are right, political clientelism slows economic development, vitiates democracy, and allows dictators to hold onto power longer than they otherwise would. It slows economic development by discouraging governments from providing public goods and by creating an interest in the ongoing poverty and dependency of constituents. Its vitiates democracy by undermining the equality of the ballot, allowing some voters to use their votes to communicate policy preferences while others use their votes only as an exchange for minor side payments. And it keeps dictators in power by allowing them to stage elections in which competition is stifled in which voters who would prefer to vote against the regime are kept from doing so by fear of retaliation. Given these critical effects, we need to understand clientelism’s internal dynamics, its causes, and its consequences.

1 Definitions

1.1 Clientelism

The concept of clientelism suffers more than most from a lack of consensus about its meaning. Focusing on clientelism as a method of electoral mobilization, I define it (p. 649) as the proffering of material goods in return for electoral support, where the criterion of distribution that the patron uses is simply: did you (will you) support me?1

It is worth noting that “proffering of material goods” in reality sometimes takes the form of threats rather than inducements. We have the government of Singapore threatening to withhold improvements of housing in districts that elect opposition legislators (Tam 2005), Christian Democratic operatives in Naples and Palermo threatening to cite opposition-supporting grocers for health violations (Chubb 1982), and the local magnate threatening to fire citizens who vote against his favored candidates in Misiones, Argentina (Urquiza 2006), to cite just a few examples.

It is the distributive criterion of electoral support that distinguishes clientelism from other materially oriented political strategies. Consider, by contrast, what is known in the USA as pork barrel politics, in which benefits are paid to one or a few districts while costs are shared across all districts (Aldrich 1995, 30).2 The implicit criterion for the distribution of pork is: do you live in my district? Or consider programmatic redistributive politics, in which parties in government emit public policies that withdraw resources from some groups and distribute them to others, almost always with electoral considerations in mind. The criterion for who will benefit from redistributive programs is: do you occupy a given class of beneficiaries (those who are unemployed, or have retired, or fall into a given tax bracket, etc.)?3 Programmatic benefits therefore have a public good quality: they redistribute resources from classes of non-beneficiaries to classes of beneficiaries, but within a class of beneficiaries, particular people who qualify cannot be excluded. By contrast there is a quid pro quo aspect to clientelist redistribution: it is only available on condition that the client complies by providing political support.

My definition is not worlds apart from Kitschelt and Wilkinson’s, who note that citizen–politician linkages are often “based on direct material inducements targeted to individuals and small groups of citizens whom politicians know to be highly responsive to such side-payments and willing to surrender their vote for the right price.” This they call a “patronage-based, voter-party linkage” (2007, 10). But alternative (or at least different) definitions abound. One defines “patron–client relationships” more generically as a “vertical dyadic alliance … between two persons of unequal status, power or resources each of whom finds it useful to have as an ally someone superior or inferior to himself (Landé 1977, p. xx). The “dyadic” part of the definition underscores the face-to-face quality of clientelism; the “alliance” part emphasizes the repeated character of the relationship.

(p. 650) Other students of clientelism define it narrowly as an exchange of a public sector job for political support (see e.g. Robinson and Verdier 2003, 2)—what many call patronage. Still others define it in terms of what it is that patrons and clients exchange. According to James Scott, the relation is an “instrumental friendship in which an individual of higher socioeconomic status (patron) uses his own influence and resources to provide protection or benefits, or both, for a person of lower status (client) who, for his part, reciprocates by offering general support and assistance, including personal services, to the patron” (1972, 92, emphasis mine). Whitaker makes a similar point in his discussion of politics in emirates, writing that in clientelist relations, “patronage, economic security, and protection can be exchanged for personal loyalty and obedience” (cited in Lemarchand 1977, 102).

Scott’s definition raises the question: under what conditions would a client not simply purchase protections and benefits in the market, rather than eliciting them from someone whom he knows personally and who is of a higher status than he? Markets may not exist or be well developed for the kinds of protections or benefits sought. Or these protections and benefits may be available on the market but their potential consumer (the client) has insufficiently plentiful resources (income) to secure them from an impersonal seller. The low-income, limited-assets client has other resources in greater abundance: time, a vote, insertion into networks of other potential supporters whom he can influence, and the like. We do not have to get very far into definitions of clientelism before we are reminded of the material poverty of the client.

Scott’s definition also focuses our attention on the clients’ interest in securing security and protection. In many polities security and protection are provided by the state as a public good. Hence, taking Scott’s two points together, all else equal we would expect patron–client ties to be prevalent in societies with widespread poverty and with a relatively weak and ineffective state apparatus.

Having explored definitions of clientelism (and offered my own), I now do the same for the related concepts of patronage and vote buying. In my usage, patronage and vote buying are subclasses of clientelism. Whereas clientelism involves the dyad’s inferior member giving electoral support broadly construed, including her own vote and efforts to secure for the patron the votes of others, vote buying is a more narrow exchange of goods (benefits, protections) for one’s own vote. In contrast, again, to pork and programmatic redistribution, the criterion for selecting vote sellers is: did you (will you) vote for me?

Patronage, in turn, is the proffering of public resources (most typically, public employment) by office holders in return for electoral support, where the criterion of distribution is again the clientelist one: did you—will you—vote for me? Hence patronage is distinct from the broader category of clientelism. In clientelism, the more powerful political actor may or may not hold public office, and therefore may or may not (p. 651) be able to credibly promise to secure public resources (as opposed to, say, party resources) for the client. In patronage, the patron holds public office and distributes state resources. This definition concurs with those of others, such as Mainwaring, who defines patronage as “the use or distribution of state resources on a nonmeritocratic basis for political gain” (1999, 177). The clientelism–patronage distinction corresponds to Medina and Stokes’s (2007) one between economic monopoly over goods which the patron controls independent of the outcome of an election, and political monopoly over goods that he controls only if he retains office. An example of an economic monopoly is a grain elevator in a rural community, access to which its owner can limit to those who voted or will vote for him, whether or not he wins the election. An example of a political monopoly is public employment, which a patron can use to reward or punish voters only in the case that he wins.

Whether a relationship is of more general clientelism or of patronage—whether it is based on an economic or a political monopoly—is consequential. Under a political monopoly, voters who wish to throw a patron out of office may face a collective action problem: his exit represents a public good, yet the voter who votes against him when a majority of others does not risks suffering the patron’s retaliation. Each voter minimizes her risk and maximizes her payoff when she votes for the unpopular patron but all other voters (or at least a majority) vote against him. Yet because all voters face this same incentive, the unpopular patron remains in power.

An implication is that, in polities in which patronage or political monopoly is widespread, one cannot infer a party’s (or its program’s) popularity from its electoral successes. Mexico’s PRI offers an example of a ruling party that remained in power and continued to win elections, probably long after its underlying popularity had been severely eroded (see Magaloni 2006).

2 Two Waves of Studies of Clientelism

The post-war literature on clientelism comes basically in two waves, the first one inspired by the emergence of new nations, the second by the democratization of large swaths of the developing world. The papers gathered and reissued in Schmidt, Scott, Landé, and Gausti’s influential 1977 reader, Friends, Followers, and Factions, had first appeared in print between 1950 and 1974, the bulk of them during the ten years after 1964. Important monographs appeared in the 1980s (such as Judith Chubb’s studies of clientelism and patronage in southern Italy), and the theoretical ground for studies of clientelism began to shift in the late 1980s and 1990s.

After a hiatus of several decades, studies of political clientelism are again legion. In addition to reappearance in journals and monographs, after 2007 we will have not (p. 652) one but three major new collections on related topics.4 The two waves of writings differ in many ways: in the political regimes studied (the early wave was indifferent to regimes, the second focused mainly on clientelism under democracy); in the basic conceptual categories employed; in the modes of analysis used; and in their disciplinary influences. The early wave was inspired mainly by anthropology and secondarily by sociology, the later one by economics.

Relations of patron and client present us with a paradox. They entail unequal actors—slave and master, serf and lord, sharecropper and landowner, worker and manager, voter and party boss—who enter into a relationship that is both voluntary and, from the less-powerful member’s vantage point, exploitative. In Kitschelt’s words, clientelism “involves reciprocity and voluntarism but also exploitation and domination” (2000, 849). By extension, we would expect clientelist relations to be full of opportunities for defection and betrayal. Why does the relationship persist, even though the client might be better off severing the link? We look, then, for some social cement to keep the client and patron together.

In many early (and even some more recent) studies, the cement is a norm. A norm is a consequential and broadly held idea that takes the form: “do x” (Elster 1989). Gouldner claimed as universal a “moral norm of reciprocity” that makes two demands: “(1) people should help those who have helped them, and (2) people should not injure those who have helped them” (1960, 171). Under clientelism, superior members of dyads reinforce the norm of reciprocity by giving their inferiors ceremonial gifts, which, like spontaneous and useful gifts, (presumably) create a sense of obligation that the gift must be reciprocated. Scott reflects on the normative and psychological tenor of relations that are personalized and ongoing. The patron–client dyad is distinguished by

the face-to-face, personal quality of the relationship. The continuing pattern of reciprocity that establishes and solidifies a patron–client bond often creates trust and affection between the partners. When a client needs a small loan or someone to intercede for him with the authorities, he knows he can rely on his patron; the patron knows, in turn, that “his men” will assist him in his designs when he needs them. Furthermore, the mutual expectations of the partners are backed by community values and rituals.

(1972, 94; emphasis in the original)

According to many norms-oriented students of clientelism, norms of reciprocity have the effect of pushing obligations from one sphere of a relationship into others. “Reciprocities that were once restricted to a specific type of exchange have thus led to cumulative or alternative exchanges among parties” (Lemarchand 1977, 106). Lemarc-hand cites as examples the overflow of patron–client ties in Senegal from spheres of feudal and religious obligations into clan politics; the generalizing of patron–client (p. 653) ties between Hutu and Tutsi in Rwanda from land ownership into cattle ownership; and the spillover of mercantile clientelism into local politics in Ibadan, Nigeria. This generalized set of obligations also spills over into relations between patron-politicians and voter-clients. Citing Wurfel, Scott notes that a Filipino politician “does favors individually rather than collectively because he wishes to create a personal obligation of clientship” (1972). In sum, if the emphasis on social norms as a cement of clientelism is not ill placed, their effect is not only to keep the subordinate member of the dyad from rebelling but also to generalize his subordination.

A very different way of thinking about clientelist exchanges is that they tie the client to the patron not by encouraging a norm of reciprocity but by encouraging a fear that the flow of benefits will be cut off. Such a perspective is more consistent with turn-of-the-century (twentieth to twenty-first) sensibilities than with the sensibilities of students of clientelism in the 1960s and 1970s. But it was by no means absent from the first-wave literature. In his classic monograph Political Leadership among the Swat Pathans, Frederick Barth insisted that “gifts can be cancelled out by an equivalent return, and do not imply any authority of the giver over the receiver … Unilateral gift-giving … does not effectively put the recipient under an obligation to respond to the command of the giver, as does the payment of bribes or salaries.” Instead, he explained, “gift-giving and hospitality are potent means of controlling others, not because of the debts they create, but because of the recipient’s dependence on their continuation. A continuous flow of gifts creates needs and fosters dependence and the threat of its being cut off becomes a powerful disciplinary device” (1959, 77).

Barth’s interest-oriented explanation is nearly identical to the one put forth forty years later by Brusco and her co-authors, that voters (in Argentina) comply with an implicit clientelist contract “because they anticipate that, should they not comply, they would be cut off from the flow of minor payoffs in the future” (2004, 76). (The only difference is that “satisfied men” among the Swat Pathan have incomes hundreds of times greater than those of the “hungry men” to whom they give gifts, which are therefore, from the latter’s perspective, by no means “minor.”)

Just as some first-wave studies posited that the cement binding clients to patrons was the client’s fear of the patron’s cutting off of rewards, so some second-wave studies continue to emphasize the client’s normative sense of obligation to the patron. A Filipino observer explains the power of campaign gifts thus: “Once a person has granted us something, a favor, we would do everything to pay that favor back to him or her, sometimes even at the expense of ourselves” (cited in Schaffer and Schedler 2006, 32).

In a similar vein, a client of the Argentine Peronist party responded in the following way to a question about whether a local party broker asked her to attend rallies in exchange for free medicines:

No … I know that I have to go with her instead of with someone else. Because she gave me medicine, or some milk, or a packet of yerba or sugar, I know that I have to go to her rally in order to fulfill my obligation to her, to show my gratitude.

(p. 654) The brokers use these feelings of friendship and gratitude to harvest votes. Following a long day of handing out goods and favors at Children’s Day celebrations, a Peronist broker remarked: “After what you just saw … votes will come. I don’t have to go and look for them … votes will come anyway” (Auyero 2001, 82).

Yet Auyero’s research shows that, in the minds of clients, the instrumentalism underlying the friendship is never far below the surface. The same client who attended rallies out of gratitude for medicine, milk, and sugar added that “if I do not go to [the Peronist broker’s] rally, then, when I need something, she won’t give it to me. [She would say,] ‘go ask the person who went to the rally with you’ ” (cited in Auyero 2001, 160).

Generally, scholars have fallen into the norms or self-interest camp without subjecting their inclinations to anything like an empirical test. An exception is the study by Brusco, Nazareno, and Stokes (2004). We asked our Argentine samples whether people who received targeted goods during election campaigns felt obliged to vote for the party who had proffered the goods. Not many more answered that recipients do feel an obligation than that they do not (51 percent to 43 percent). But those who said, as a factual matter, that people do feel obliged tended to be the kind who would never receive such a handout: compared to the means for our samples they were wealthy, non-Peronist, and came from big cities. They seemed to be interpreting other people’s experiences, not their own. We also asked whether people who receive goods should feel an obligation to reciprocate with their vote. If there is a norm of political reciprocity in Argentina saying that “gifts” demand a response, this norm has far-from-widespread acceptance: nine out of ten people whom we sampled said that recipients of campaign handouts should not feel obliged to return the favor of a handout with a vote (2004, 81). We cannot be sure that these results are not country and time specific, but they do suggest that norms of political reciprocity are not universal and are perhaps vulnerable to political mobilization against them.

3 Clientelism and Commitment

The second-wave shift away from norms and toward fear of retaliation draws on economics and political economy. A series of papers in the 1980s and 1990s explained formally how parties could use individualized or targeted inducements to mobilize electoral support. The basic idea was that, rather than using public policies to effect transfers from some classes of voters to others, parties could deliver inducements to individual voters and thus bolster the parties’ electoral prospects. Dixit and Lon-dregan (1996) call this “tactical” as opposed to “programmatic” redistribution.5 A (p. 655) central finding of this formal literature was that, when parties knew their constituents well and could efficiently deliver goods to those who would be most likely to return the favor with their vote, then doing so could be cost effective. Parties that practiced this strategy would target core constituents: voters whose needs (Dixit and Londregan 1996) or electoral predispositions (Cox and McCubbins 1986) were well known to the party and hence who presented little risk that the party would waste resources on them. (See also Lindbeck and Weibull 1987.)

Following these theoretical leads, Magaloni, Diaz-Cayeros, and Estévez (2006) posited that risk aversion is a variable quality, even of a single party over time; varying tolerance for risk could explain the mix of strategies—programmatic in some constituencies and in some periods, clientelistic in others—that parties deployed.

A difficulty with this line of theorizing about clientelism as an electoral strategy is that it deals inadequately with problems of commitment.6 A voter who receives a bag of food with the understanding (implicit or explicit) that she will return the favor with a vote can easily renege on the deal on election day, especially when she is protected by the secret ballot. Indeed, the secret ballot was introduced to free voters of the kind of tacit coercion that vote buying entails. And the commitment problem runs in both directions: a party that before an election promises patronage in exchange for votes may well forget its promise afterwards.

To illustrate, consider a patronage game: members of a favored group enjoy the tactically redistributed goods, following Dixit and Londregan, only when the patronage party wins. First the voter decides whether to comply and vote for this party or defect and vote against it. Then nature makes the party either win or lose the election. If the party wins it then chooses whether to reward the voter or withhold a reward. If the party loses it cannot disburse rewards and the game ends. Figure 31.1 is the extensive form of the game. The sequence laid out there implies that the voter chooses whether to comply or defect before the outcome of the election is known (the party, if it gets to act, obviously knows that it has won).

The party gains v when the voter votes for it and assumes a loss of −rp whenever it pays a reward. By assumption, v > rp: the best outcome for the party is to gain a vote without paying for it but it prefers to pay and get the vote than not to pay and forgo the vote. In this game the voter is mildly ideologically predisposed against the patronage party: its most preferred outcome is to win the reward and vote against the party but it would vote for it if necessary to win the reward. Thus it prefers the outcomes in the following order: (rv + d) > (rv + c) > d > c.

Fig. 31.1. A one-shot game between a voter and a would-be patron (rewards dependent on incumbency)

If the party wins, its dominant strategy is to withhold a reward from the voter. Whether the voter has complied or defected, the party does better denying him a reward. Knowing that the party, even if it wins, will not give a reward, the voter is always better off defecting and voting for the party which, on ideological grounds, (p. 656) he prefers. Hence the party never offers a reward and the voter’s decision is entirely driven by its pre-redistributive preference for or against the party.

Even if the voter were ideologically inclined to vote for the patronage party, this preference, rather than the promise of patronage, would motivate the vote. The party still would do better reneging on its promise. Therefore the voter would ignore the promise and vote purely on ideology. Again, no patronage is meted out and its promise does not motivate electoral choices.

(p. 657) But we know that patronage and vote buying do occur and hence we need to rethink the model. Rather than the one-shot games developed in the literature, we do well to return to the insights of many students of clientelism that the relationships involved are face to face and ongoing. The “gift giving” of clientelism and patronage does not only motivate people to vote for the patron’s party directly but also reinforces a social network in which patron and client are embedded. That clientelist relationships are ongoing—that the dyad is embedded in a social network—is theoretically important for several reasons. Networks provide information about their members to other members: we know whether our neighbor or co-worker votes or abstains, voices support for one party or another, and comes from a family of communists or Christian Democrats (Democrats or Republicans, Peronists or Radicals, etc.), none of which we know about strangers. Clientelist parties use operatives who are embedded in these networks and, like the Taiwanese campaign managers described by Wang and Kurzman, are “walking encyclopedia[s] of local knowledge” (2007, 94). This local knowledge allows them to make informed guesses about whether a voter to whom the party gave goods or employment actually followed through and supported the party or defected to another. Networks allow clientelist parties to sidestep the secret ballot.

The party can then use this information to reward the voter who has cooperated and punish the voter who has defected—it can hold the voter accountable for his or her vote. Yet in contrast to the kind of accountability celebrated in democratic theory, this is “perverse accountability,” in which voters are held accountable for their actions by parties (Stokes 2005).

Given that patrons and clients are embedded in networks, we can model clientelism as a repeated game. The voters’ preferences are given by

$Display mathematics$
where vi = {x1, x2} is a vote for either the clientelist party or the opposition, xi is voter i’s position on the ideological spectrum, and rvi = {0, r} is the value to the voter of the reward offered by the machine in exchange for votes, relative to the value of voting according to the voter’s preferences. Thus $− 1 2 ( v i − x i ) 2$ is the expressive value of voting for one of the two parties.

Table 31.1 presents the normal form of the game, and Table 31.2 simplifies the payoffs. Consider the case of a voter who is mildly opposed, on ideological or programmatic (p. 658) grounds, to the clientelist party. Without any inducements she will vote for its opponent, but it can offer her a reward (r) that would improve her payoff over voting against the clientelist party and not receiving the reward. That is, her preference order is:

$Display mathematics$

Table 31.1. Normal form of a game between the clientelist party and a voter

Voter

Party

Reward

No reward

Comply

$1 2 ( x i − x l ) 2 + r , v − r$

$1 2 ( x i − x l ) 2 , v$

Defect

$1 2 ( x i − x 2 ) 2 + r , − r$

$1 2 ( x i − x 2 ) 2 , 0$

Table 31.2. Normal form of the game between the clientelist party and a weakly opposed voter with simplified payoffs

Voter

Party

Reward

No reward

Comply

3,3

1, 4

Defect

4, 1

2, 2

The clientelist party, in turn, would most like to receive her vote without having to pay a reward but is willing to pay for the vote if necessary. Hence its preference order is:

$Display mathematics$

Table 31.2 makes clear that the clientelist party and the mildly opposed voter are in a prisoner’s dilemma. Both would like to “cooperate” in a vote-buying arrangement, with the party paying a reward and the voter supporting the party. But if the voter supports the party then the party does better by withholding the reward, and if the party withholds the reward then the voter does better voting against it.

In the one-shot game, vote buying fails. But in the repeated game, they can cooperate to traffic in votes. They can do so by playing a grim-trigger strategy, with either side responding to a defection from the pair of strategies vote for the clientelist, pay reward by punishing the other side in every subsequent election in the future. The voter’s minmax value, the lowest payoff that the clientelist party can hold her to, is $− 1 2 ( x i − x 2 ) 2$: even if she is forced to forgo the reward, she can always vote for the opposition party. The party’s minmax value is 0: in response to the voter who defects, the party can at least always withhold a reward.

These minmax values allow us to define the feasible and individually rational payoffs of the two players, as illustrated in Figure 31.2. Among these payoffs is the pair that results when vote buying occurs: $− 1 2 ( x i − x 2 ) 2 + r , v − r$. By the folk theorems, in a repeated-play setting, if the players are patient then all feasible and individually rational payoffs enforce an equilibrium. The reason is that “when players are patient, any finite one-period gain from deviation is outweighed by even a small loss in utility in every future period” (Fudenberg and Tirole 2002, 153). Yet note also in Figure 31.2 that the vote-buying equilibrium is only one of several possible; the entire shaded area consists of feasible and individually rational payoffs.

Fig. 31.2. Feasible and individually rational payoffs in repeated playbetween the clientelist party and a weakly opposed voter

(p. 659)

The game outlined above generates several hypotheses:

• the greater a clientelist party’s ability to directly observe an individual’s vote, the more likely vote buying is to occur;

• the closer the vote comes to being fully opaque and anonymous, the less likely vote buying is to occur;

• the closer parties are to one another ideologically, the more likely vote buying is to occur;

• the more valuable the reward or private inducement is to the voter, the more likely vote buying is to occur (Stokes 2005).

4 Are Clients Core Supporters or Swing Voters?

Clientelism as a repeated game between voters and parties embedded in social networks also tells us something about the kind of voter whom clientelist parties (p. 660) target. Its blandishments are wasted on loyal voters, who would support it anyway, independent of rewards, and hence cannot credibly commit to withdrawing their support should the flow of rewards be cut. Blandishments are also wasted on die-hard opposition voters. In contrast to the mild opponents discussed earlier, these die-hards’ ideological distaste for the party outweighs the value of the reward; they therefore cannot credibly commit to vote for it.7 Indifferent voters, and those who are slightly opposed to the party—whose distaste for voting for it is outweighed by the value of the reward—are the ideal targets of vote-buying efforts. The model thus predicts a marginal or swing voter strategy: parties will shower targeted rewards on people who are indifferent or slightly against them on ideological or programmatic grounds, not those who strongly support them.

A handful of studies have tested core versus swing voter hypotheses. These studies face methodological difficulties. In those using ecological data, the investigators typically consider bias in the distribution of resources toward districts that have voted heavily for the governing party in the past as evidence of a core voter strategy. Yet the voters who support the party and are then showered with public resources may have supported the party in the past because they were also, in the past, showered with resources. Hence the term “core” or “loyal” should be interpreted with caution. “Loyal” voters are those with a proven receptiveness to targeted goodies rather than those with an ideological predisposition in favor of the clientelist party. Ideally, ecological studies would control for the effect of targeted gift giving in earlier elections. Studies that rely on survey data face their own kind of endogeneity problem: does a person’s self-reported friendliness toward a party cause him or her to receive handouts, or do handouts make him or her friendly? For that matter, are self-reported indifferent voters actually opponents whose opposition has been mitigated by rewards?

With these caveats in mind, empirical studies paint a mixed picture. Ansolabehere and Snyder (2002) find that US counties that traditionally provide high levels of support for the governor’s party receive larger transfers from the state to local governments. They also find that a change in which party controls state government is followed by a shift in the distribution of these transfers in favor of counties that vote heavily for the new governing party.

Elsewhere in the hemisphere, several studies test core versus swing voter hypotheses with ecological data. Studying Mexico, Magaloni (2006) finds that the PRI spent lightly on social programs in regions controlled by the opposition, and spent heavily in regions in which voters would be most likely to defect by abstaining, again suggesting a marginal voter strategy. By contrast, Pérez Yarahuan (2006) finds that the ruling party favored its own electoral strongholds in the distribution of social programs and discriminated against opposition strongholds, suggesting a core strategy, at least in the 1994 election (see also Hiskey 1999). In a study of political manipulation of expenditures on an anti-poverty program in Peru, Schady (2000) finds that the (p. 661) Fujimori government favored “marginal” districts, ones in which the president had come close to winning or losing in his first election and in a referendum.

In Argentina, the evidence is also mixed. Weitz-Shapiro (2006) finds that toward the end of Carlos Menem’s time in power and at the beginning of the De la Rúa administration (1999–2001), the distribution of unemployment compensation funds was biased in favor of swing districts—those where margins of victory of one party or another were thin. Nazareno, Brusco, and Stokes (2006), also using ecological data, show that the smaller a party’s margin of victory in past elections (or, sometimes, soon-to-occur future elections), the greater the proportion of its budget that went to expenditures on personnel. Employing survey data on campaign handouts, Stokes (2005) finds the evidence to be less decisive. Although she finds that the Argentine Peronists withheld campaign handouts from strong opponents (handouts which, according to her model, would have been a wasted on these kinds of voters), the Peronists gave inducements indifferent or marginal voters but also voters who favored the party. She speculates that clientelist parties often combine clientelist gift giving with ideological appeals: the ward-heeler who appears at a person’s door with a bag of food or the offer of employment will also explain the party’s programmatic advantages. Alongside their face-to-face material mobilization the parties also proselytize, and the once-indifferent voter begins to look more like a loyal constituent.

In sum, the bulk of the empirical data points toward clientelist parties targeting marginal or swing voters, but a few important exceptions find that they target core constituents. No one has accounted for this variation. Following Dixit and Lon-dregan, Cox and McCubbins, and Magaloni and her co-authors, it may be that the key variable is risk: when politicians face ideologically diverse groups and wish to minimize their risk of wasting goodies, or when the politicians are themselves risk averse, they deploy a core voter strategy. Recent work that focuses not (following Cox’s 2006 terminology) on “persuasion” (changing voters’ minds about whom to vote for) but on “mobilization” (including them to vote, rather than to abstain) may shed further light on this matter.

5 Causes and Consequences

5.1 Why does Poverty Encourage Clientelism?

We saw that the very definition of clientelism points toward the poverty of the client. Indeed, it is impossible to survey the qualitative literature on political clientelism without concluding that it is a feature disproportionately of poor countries. It gets studied in the contexts of eighteenth-century Holland (Randeraad and Wolffram 2001), eighteenth- and nineteenth-century England (O’Gorman 2001), (p. 662) inter-war Greece (Mavrogordatos 1983), US cities in the late nineteenth century through the 1950s (Wilson and Banfield 1963; Reynolds 1988), Iceland in the 1940s and 1950s (Kristinsson 2001), southern Italy in the 1950s and 1960s (Graziano 1977; Chubb 1981, 1982), South Asia (Weiner 1967; Chandra 2004, 2007; Wilkinson 2006; Krishna 2007), South-East Asia (Landé 1965; Scott 1972), Africa (Lemarchand 1977; Wantchekon 2003; van de Walle 2007), Bulgaria (Kitschelt et al. 1999); and in Latin America from Mexico (Fox 1994; Magaloni 2006, Magaloni, Diaz-Cayeros, and Estévez 2007) to Argentina (Levitsky 2003; Calvo and Murillo 2004; Brusco, Nazareno, and Stokes 2004) and everywhere in between. We lack quantitative cross-national studies of the subject—it is hard to develop cross-nationally comparable quantitative measures of clientelism—but a mere glance at the qualitative literature shows that, while it is not an exclusive feature of the developing world (yesterday’s or today’s), one is much more likely to encounter it there than in the advanced democracies.8

Less obvious is why this is so. Without noticing that they are doing so, scholars have posited two distinct explanations for the link between clientelism and poverty. In the first and more common account, poor people value a handout more highly than do wealthy people; hence, if one is going to hand out goodies, one will target the poor (diminishing marginal utility of income—see e.g. Dixit and Londregan 1996; Calvo and Murillo 2004). In the second account, poor people are risk averse and hence value more highly a bag of goodies in hand today than the promise of a redistributive public policy tomorrow (see e.g. Desposato 2007; Wantchekon 2003; Kitschelt 2000; Scott 1976). As Kitschelt explains, “poor and uneducated citizens discount the future, rely on short causal chains, and prize instant advantages such that the appeal of direct, clientelist exchanges always trumps that of indirect, programmatic linkages promising uncertain and distant rewards to voters” (2000, 857).

Data emerging from one study cast doubt on the risk aversion explanation. Brusco, Nazareno, and Stokes (2006) find that poor Argentines were indeed more risk averse than were their wealthier compatriots. But risk aversion had no independent effect on one’s propensity to sell one’s vote. Comparing two equally poor voters, the more risk-averse one was no more likely to sell her vote than was the less risk-averse one.

A third account, slightly different from the others, is that not poverty per se but income inequality encourages clientelism (see e.g. Hicken 2007; Stokes 2007; Robinson and Verdier 2003). If the resources available for vote buying rise at the same rate as a country’s average income, then development will not make it too expensive. But if clientelism must be paid for by a growing (upper) middle class and if its targets are themselves increasingly from the (lower) middle class, then the transfer will increasingly be as painful for those on the giving side as they are profitable for those on the receiving side, and one should encounter more resistance from the givers. (p. 663) Furthermore, middle-class citizens have higher opportunity costs for monitoring the activities of other voters; parties will find it more difficult to retain a presence in social networks and hence to monitor individuals’ voting behavior when the distribution of voters by incomes is dense in the middle and thinner at the lower end.

Another perspective altogether is that, rather than poverty generating clientelism, clientelism generates poverty. As a strategy to stay in power, clientelist parties may develop an interest in holding back income growth. Chubb blamed underdevelopment in Italy’s Mezzogiorno on clientelism and patronage carried out by the Christian Democratic Party, which wished to keep its constituents poor and dependent. Other accounts point toward the underprovision of development-enhancing public goods in polities in which office holders focus on the provision of private goods (see e.g. Robinson and Verdier 2003), and on the declining relative productivity of, and dependence on, monopolized-goods sectors as countries undergo economic development (Medina and Stokes 2007; see also Lyne 2006). Of course it may be true both that poverty causes clientelism and that clientelism causes poverty.

5.2 Institutions

Electoral rules. Some analysts find it intuitive that electoral systems that encourage the personal vote also encourage clientelism (see e.g. Hicken 2007; Kitschelt 2000). Kitschelt reasons that

Personalized contests permit candidates and constituencies to organize, monitor, and enforce direct trades of support for favors flowing from office. In multimember electoral districts, personal preference votes for individual candidates rather than entire party lists make possible personalized trades. Politicians’ incentives to pursue clientelism further increase when the votes that different candidates for the same party receive individually are not pooled to calculate the seats won by the entire party and/or when the party leadership does not control the nomination of list candidates.

(2000, 859)

The flaw in this account is that it elides the personalist appeal of candidates with personalized, face-to-face voter mobilization. Although the two have in common a downplaying of issues and programs, they are in other ways quite distinct. A campaign that focuses on the personal qualities of the candidate may invest in mass media appeals and rely on a highly centralized party structure. By contrast, clientelist parties and campaigns rely on an army of brokers, intermediaries, and campaign workers to monitor the actions of voters at a fine-grained level. Clientelist parties in fact are decentralized parties, and decentralization is the price that the party leadership has to pay if it is to sustain an army of brokers to, in effect, spy on voters. Were the centralized leadership, or an individual charismatic leader, able to circumvent these decentralized brokers—were they able to replace decentralized vote buying with rousing oratory or compelling ideology—they would gladly do so. The personalization of electoral campaigns is hence at odds with clientelism, and we may be on the wrong track when we (p. 664) link personalizing electoral rules with the prevalence of clientelism. Again, however, this dispute remains theoretical in nature, and probably will until we discover a cross-nationally robust measure of clientelism that allows us to assess the impact of electoral rules on the phenomenon.

Legal restrictions on patronage and vote buying. Today, the vast majority of democracies place legal restrictions on patronage and vote buying. They enact laws and regulations regarding the recruitment of personnel into the bureaucracy, and they prohibit vote trafficking. According to Shefter’s (1977, 1994) influential analysis, patronage was strictly limited in polities in which the civil service was professionalized before the franchise was broadened, but was more widespread in polities in which mass enfranchisement pre-dated any serious efforts at civil service reform. In the latter, parties were freer to mobilize mass support by offering public employment.

Shefter’s account leaves some questions unanswered. Why do civil servants professionalize early in some places and later in others? In some settings where a professionalized civil service is in place, ambitious politicians mount an effective assault on it, reversing its autonomy and turning it into a source of patronage. This was the experience, for instance, of Pakistan under Zulfikar Ali Bhutto: the country inherited an autonomous and professionalized civil service from British India but Bhutto’s 1973 reform turned the Pakistani civil service into a patronage machine for his party, the PPP (see Baxter et al. 2002). In this and other instances like it, laws and regulations limiting patronage and vote buying appear malleable in the hands of politicians.

Ballots. The technology that governments and parties allow voters to use to express their electoral choices also influences clientelism. The main reason is that clientelism is greatly aided by less-than-opaque and -anonymous voting; ballots that facilitate the monitoring of voters’ choices also facilitate clientelism. By the early twentieth century most countries that were independent and had democratic intervals had eliminated voice voting and gone over to the secret ballot, greatly reducing the scope for vote trafficking. Nevertheless, some ballots and balloting systems still in use today help parties infer the vote choices of individual voters. Spain, for instance, combines ticket ballots—ones containing only the names of candidates from one party’s list—with the option of voting semi-publicly: Spanish voters have a choice of placing their preferred party list in an envelope in a curtained booth or doing so openly. (France and some former French colonies, as well as several other democracies, also use ticket ballots.) A handful of countries, among them Argentina, Panama, and Uruguay, use party ballots, ballots that political parties produce and in which individual candidates (for executive offices) or party lists appear on separate ballots.

Ticket and party ballots are both distinct from Australian ballots, in which all candidates or lists of candidates for a given office appear listed in the same format, and in which public agencies produce and distribute the ballots on the day before or day of the election, under controlled conditions (Converse 1972). Reynolds and Steenbergen (2005) calculate that 85 percent of countries today use Australian ballots, 15 percent what I am calling ticket and party ballots.

(p. 665) Party ballots may particularly facilitate vote buying not because of the format of the ballot but because parties produce them and usually can distribute them well before the election. The distribution of ballots becomes part of the process of mobilizing voters, and when parties distribute ballots along with bags of food, building materials, or other individualized inducements, the message comes across especially clearly that the favors are expected to be reciprocated with a vote. Brusco, Nazareno, and Stokes (2004) find that Argentine voters who received their ballots directly from party representatives are significantly more likely to have received campaign handouts and to report that these handouts influenced their vote.

Because it is politicians who write regulations discouraging patronage and design ballots that work against vote buying, one wonders whether these rules have an independent effect of reducing clientelism or whether they are brought to life only after other factors, such as economic development, have made clientelism less effective and hence less tempting for politicians. Still, in some settings parties compete against one another by shining a light on vote buying and electoral corruption, creating electoral incentives to carry out ballot and other institutional reforms. This dynamic has been observed in Mexico, in Taiwan, and in Argentina; in all three countries, opposition politicians admonish voters, as an Argentine politician put it, to “receive with one hand and vote with the other” (quoted in Szwarcberg 2001).

Accounts of transitions from clientelist to programmatic politics offer clues about the factors encouraging clientelism. Lehoucq and Molina (2002) attribute the decline of vote buying in Costa Rica in the 1940s to the introduction of the secret ballot and the increasing costliness of payments to voters, the latter suggestive of development as a cause of this decline.

Recall that, as illustrated in Figure 31.2, the clientelist equilibrium is only one of many possible equilibria. Schaffer and Schedler point out that, whereas markets in consumer goods are generally considered morally legitimate, “the explicit purchase of votes runs counter to present norms of democratic liberty and equality” (2006, 6). It is therefore vulnerable to ideological attacks. In Peru, a progressive military regime, followed by leftist organizers and radical clergy, encouraged people to think of themselves as citizens who should receive public services in exchange for their taxes, rather than as clients who needed to plead for special favors (Stokes 1995).

Several authors note a decline in clientelism and vote buying in Mexico with that country’s gradual democratization, which culminated in 2000. A Mitofsky poll found that 5 percent of voters received a handout before the 2000 elections; the Mexico 2000 panel study put the number at just under 15 percent (see Cornelius 2004).9 He interprets both numbers as a decline over past practice. Lehoucq (2006) and Cornelius give credit to Mexico’s Federal Electoral Institute (the Instituto Federal Electoral, IFE), overhauled in 1994, for reducing clientelism and vote buying. Magaloni concurs, but goes a step further and analyzes the reasons why the PRI, Mexico’s ruling party, was (p. 666) willing to grant the IFE independence. It granted this independence, in her analysis, as a way of inducing the opposition to endorse the legitimacy of elections. Fair elections with the possibility of losing were a better outcome for the PRI than illegitimate elections and the social instability that followed.

Institutional consequences of clientelism. Little research has delved into the institutional causes of clientelism, but even less into its institutional consequences. But in a highly original paper, Desposato (2007) studies clientelism’s effects on political parties in legislatures. He reasons that parties that use clientelist strategies will behave differently in the legislature than parties that mobilize electoral support by providing public goods. Clientelist parties will work determinedly to secure public resources for distribution throughout their personal networks; when they are in opposition, they will display little legislative cohesion. Public-goods oriented parties will work determinedly to provide these goods and claim credit for their provision, and will display more legislative cohesion, whether in government or in opposition. Desposato compares two state legislatures in Brazil, one in a state (Piauí) where clientelism is widespread, the other in a state (Brasília) in which it is nearly absent. He finds differences in the behaviors of the legislative parties—greater frequency of roll-calls in Brasília, less cohesion of opposition parties in Piauí—that accord with his theory.

5.3 The Effectiveness of Clientelism and Patronage

Most students and casual observers of clientelism assume that it works as an electoral strategy—that, all else equal, a party that disburses clientelist benefits will win more votes than it would have had it not pursued this strategy. In general we do not expect parties to pursue strategies that are ineffective. And yet we have some theoretical reasons for believing that conditions are not always ripe for clientelism. In the repeated game outlined earlier, the vote-buying equilibrium is just one of many possible equilibria.

Voters who are offered patronage by an unpopular incumbent may find ways to overcome the collective action problem of voting against him or her, as they eventually did in Mexico (Magaloni 2006). And if we assume that parties sometimes lack information about the consequences of their strategies, particularly in new democracies, then we should not be surprised that they sometimes undermine themselves.10

A few studies have explored the electoral consequences of clientelism, most of them in the Americas. In the USA, Levitt and Snyder (1997) study the effect of pork-barrel spending (by the definition offered above) on subsequent elections for the House of Representatives. They find that an additional $100 in federal spending can increase (p. 667) the popular vote of the incumbent by as much as 2 percent.11 In Peru, Roberts and Arce (1998) find a positive correlation between the Fujimori government’s per capita expenditures by department on anti-poverty programs in 1994 and early 1995 and Fujimori’s vote share in his 1995 (re)election. But clientelism and patronage are not always a plus. In Argentina, Calvo and Murillo (2004) find that, in provinces governed by the Peronist party, the larger the number of public employees per thousand, the greater the Peronist vote share in subsequent elections. But in provinces governed by Argentina’s other main political party, the Radicals, public employment has no significant effect on Radical vote shares. Using more disaggregated (municipal) data, Nazareno, Brusco, and Stokes (2006) find similar—and even starker—results: patronage spending by Radical mayors depresses their party’s vote share. The same is true of spending by the federal government (controlled by the Peronists) on targeted unemployment relief programs in the late 1990s. Controlling for other factors (such as poverty), Peronist mayors could expect to receive about twice as much funding per capita as Radicals, and spending in Radical-controlled municipalities significantly reduced the Radical vote share in the next election. In contrast, when these same targeted funds were distributed in Peronist municipalities, local Peronists increased their vote share in subsequent elections. Like Levitt and Snyder, Nazareno and his associates reason that mayors who knew they were in trouble heading into an election might spend more; even if the extra spending enhanced their vote share over the level it would have stayed at had they not intensified their patronage efforts, the patronage would appear, misleadingly, to depress the incumbent’s support. We thus face a possible endogeneity problem: (anticipated) poor electoral performance might cause more spending, rather than spending causing poor electoral performance. To deal with this possibility, Nazareno and his co-authors employ an instrument: spending in an earlier election. Still, even correcting for endogeneity, the results hold: Radical patronage depresses the Radical vote. Why would clientelist spending ever be bad for the party that does the spending? For some kinds of constituents—especially wealthier, more autonomous constituents—such spending may indicate inefficient, pandering governments. It is highly suggestive that the best case we have of such negative effects is spending by the Argentine Radical Party, a party of relatively middle-class constituents. 6 Conclusions Political clientelism, the giving of material resources as a quid pro quo for political support, is best understood as part of an ongoing exchange between patron and (p. 668) client, with threats of defection instead of, or perhaps in addition to, norms of reciprocity sustaining it. Incorporating the old observation that patron–client linkages are face to face and ongoing allows us to model the exchange as a repeated game, and hence one that can overcome problems of commitment and defection on either side. Theoretical and empirical studies identify conditions under which both core and marginal voters will be the targets (or beneficiaries) of clientelist parties. Clientelism is intimately linked to poverty and inequality, of which it is probably both a cause and a consequence. Institutions such as personalized campaigns, ballot design, and legal restrictions may also influence whether parties deploy clientelist or programmatic strategies. Much theoretical work and empirical research remain to be done. The affinity between poverty (inequality) and clientelism is settled fact, but the mechanisms linking the two, and the direction of causality, are not. We tend to treat clientelism as involving a dyadic link between patron and client, in an electoral context, between party and voter. But really the strategic interactions of at least three actors should be considered: party leaders, party brokers, and voters. We have some detailed empirical information on brokers, such as Wang and Kurzman’s (2006) fascinating study of Kuomintang brokers in Taiwan, but the implications of their presence for theory have not been sufficiently worked out. Furthermore, we would like to know more about the interactions between parties as they strategize about which methods to pursue. Finally, our understanding of the relationship between clientelism and institutions—from macro institutions such as electoral systems to micro institutions such as ballot design—is in its infancy. Not until we achieve fuller theoretical accounts and test them with more systematically comparative data will we have the tools to tame political clientelism. References Aldrich, J. H. 1995 Why Parties? The Origin and Transformation of Political Parties in America. Chicago: University of Chicago Press.Find this resource: Ansolabehere, S. and Snyder, J. M., Jr. 2002. Party control of state government and the distribution of party expenditures. Typescript. Massachusetts Institute of Technology.Find this resource: Auyero, J. 2001. Poor People’s Politics: Peronist Survival Networks and the Legacy of Evita. Durham, NC: Duke University Press.Find this resource: Barth, F. 1959. Political Leadership among the Swat Pathans. London: Athlone.Find this resource: Baxter, C., Malik, Y. K., Kennedy, C. H., and Oberst, R. C. 2002. Government and Politics in South Asia. Boulder, Colo.: Westview Press.Find this resource: Brusco, V., Nazareno, M., and Stokes, S. 2004. Vote buying in Argentina. Latin American Research Review, 39 (2): 66–88.Find this resource: ———2006. Clientelism and risk. Typescript. Yale University.Find this resource: Calvo, E. and Murillo, M. V. 2004. Who delivers? Partisan clients in the Argentine electoral market. American Journal of Political Science, 48 (4): 742–57.Find this resource: (p. 669) Chandra, K. 2004. Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in India. Cambridge: Cambridge University Press.Find this resource: ——2007. Counting heads: a theory of voter and elite behavior in patronage democracies. Pp. 183–239 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Chubb, J. 1981. The social bases of an urban political machine: the case of Palermo. Political Science Quarterly, 96 (1): 107–25.Find this resource: —1982. Patronage, Power, and Poverty in Southern Italy. Cambridge: Cambridge University Press.Find this resource: Converse, P. E. 1972. Change in the American electorate. Pp. 263–337 in The Human Meaning of Social Change, ed. A. Campbell. New York: Russell Sage Foundation.Find this resource: Cornelius, W. A. 2004. Mobilized voting in the 2000 elections: the changing efficacy of vote buying and coercion in Mexican electoral politics. Pp. 47–65 in Mexico’s Pivotal Democratic Election: Candidates, Voters, and the Presidential Campaign of 2000, ed. J. I. Domínguez and C. Lawson. Stanford, Calif.: Stanford University Press.Find this resource: Cox, G. 2006. Voters, core voters, and distributive politics. Typescript, University of California, San Diego.Find this resource: Cox, G. W. and McCubbins, M. D. 1986. Electoral politics as a redistributive game. Journal of Politics, 48 (2): 370–89.Find this resource: Desposato, S. W. 2006. How does vote buying shape the legislative arena? Pp. 144–79 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Schaffer. Boulder, Colo.: Lynne Rienner.Find this resource: Dixit, A. and Londregan, J. 1996. The determinants of success of special interests in redistributive politics. Journal of Politics, 58: 1132–55.Find this resource: Elster, J. 1989. Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press.Find this resource: Fox, J. 1994. The difficult transition from clientelism to citizenship: lessons from Mexico. World Politics, 46: 151–84.Find this resource: Fudenberg, D. and Tirole, J. 2002. Game Theory, 8th edn. Cambridge, Mass.: MIT Press.Find this resource: Gouldner, A. 1960. The norm of reciprocity: a preliminary statement. American Sociological Review, 25 (2): 161–78.Find this resource: Graziano, L. 1977. Patron–client relationships in southern Italy. Pp. 360–78 in Friends, Followers, and Factions: A Reader in Political Clientelism, ed. S. W. Schmidt, J. C. Scott, C. Landé, and L. Guasti. Berkeley and Los Angeles: University of California Press.Find this resource: Hicken, A. 2007. How do rules and institutions encourage vote buying? Pp. 68–89 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Schaffer. Boulder, Colo.: Lynne Rienner.Find this resource: Hiskey, J. 1999. Does democracy matter? Electoral competition and local development in Mexico. Ph.D. dissertation. University of Pittsburgh.Find this resource: Kitschelt, H. 2000. Linkages between citizens and politicians in democratic polities. Comparative Political Studies, 33 (6–7): 845–79.Find this resource: —and Wilkinson, S. (eds.) 2007. Citizen–politician linkages: an introduction. In Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition. Cambridge: Cambridge University Press.Find this resource: —Mansfeldova, Z., Markowski, R., and Toka, G. 1999. Post-Communist Party Systems: Competition, Representation, and Inter-party Cooperation. Cambridge: Cambridge University Press.Find this resource: Krishna, A. 2007. Politics in the middle: mediating relationships between citizens and the state in rural north India. Pp. 298–383 in Patrons, Clients, and Policies: Patterns of Democratic (p. 670) Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Kristinsson, G. H. 2001. Clientelism in a cold climate: the case of Iceland. Pp. 172–92 in Clientelism, Interests, and Representation: The European Experience in Comparative Perspective, ed. S. Piattoni. Cambridge: Cambridge University Press.Find this resource: Landé, C. H. 1965. Leaders, Factions, and Parties: The Structure of Philippine Politics. Yale Southeast Asia Monograph Series 6. New Haven: Yale University Press.Find this resource: ——1977. Introduction: the dyadic basis of clientelism. Pp. xiii–xxxvii in Friends, Followers, and Factions: A Reader in Political Clientelism, ed. S. W. Schmidt, J. C. Scott, C. Landé, and L. Guasti. Berkeley and Los Angeles: University of California Press.Find this resource: Lehoucq, F. E. 2007. When does a market for votes emerge? Theoretical and empirical perspectives. Pp. 48–67 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Schaffer. Boulder, Colo.: Lynne Rienner.Find this resource: ——and Molina, I. 2002. Stuffing the Ballot Box: Fraud, Electoral Reform, and Democratization in Costa Rica. Cambridge: Cambridge University Press.Find this resource: Lemarchand, R. 1977. Political clientelism and ethnicity in tropical Africa: competing solidarities in nation-building. Pp. 100–23 in Friends, Followers, and Factions: A Reader in Political Clientelism, ed. S. W. Schmidt, J. C. Scott, C. Landé, and L. Guasti. Berkeley and Los Angeles: University of California Press.Find this resource: Levitsky, S. 2003. Transforming Labor-Based Parties in Latin America: Argentine Peronism in Comparative Perspective. Cambridge: Cambridge University Press.Find this resource: Levitt, S. D. and Snyder, J. M., Jr. 1997. The impact of federal spending on House election outcomes. Journal of Political Economy, 105 (1): 30–53.Find this resource: Lindbeck, A. and Weibull, J. 1987. Balanced budget redistribution as the outcome of political competition. Public Choice, 52 (3): 273–97.Find this resource: Lyne, M. 2007. Rethinking economics and institutions: the voter’s dilemma and democratic accountability. Pp. 335–80 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Magaloni, B. 2006. Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico. Cambridge: Cambridge University Press.Find this resource: ——Diaz-Cayeros, A., and Estévez, F. 2007. Clientelism and portfolio diversification: a model of electoral investment with applications to Mexico. Pp. 381–429 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Mainwaring, S. P. 1999. Rethinking Party Systems in the Third Wave of Democratization: The Case of Brazil. Stanford, Calif.: Stanford University Press.Find this resource: Mavrogordatos, G. T. 1983. Still-Born Republic: Social Coalitions and Party Strategies, 1922– 1936. Berkeley: University of California Press.Find this resource: Medina, L. F. and Stokes, S. 2007. Monopoly and monitoring: an approach to political clientelism. Pp. 150–82 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Nazareno, M., Brusco, V., and Stokes, S. C. 2006. Réditos y peligros electorales del gasto público en Argentina. Desarrollo económico, 46 (181): 63–86.Find this resource: O’Gorman, F. 2001. Patronage and the reform of the state in England, 1700–1860. Pp. 54–76 in Clientelism, Interests, and Representation: The European Experience in Comparative Perspective, ed. S. Piattoni. Cambridge: Cambridge University Press.Find this resource: (p. 671) Pérez Yarahuan, G. 2006. Policy making and electoral politics: three essays on the political determinants of social welfare spending in Mexico, 1988–2003. Ph.D. dissertation. University of Chicago.Find this resource: Piattoni, S. (ed.) 2001. Clientelism, Interests, and Representation: The European Experience in Comparative Perspective. Cambridge: Cambridge University Press.Find this resource: Randeraad, N. and Wolffram, D. J. 2001. Constraints on clientelism: the Dutch path to modern politics, 1848–1917. Pp. 101–21 in Clientelism, Interests, and Representation: The European Experience in Comparative Perspective, ed. S. Piattoni. Cambridge: Cambridge University Press.Find this resource: Reynolds, A. and Steenbergen, M. 2005. How the world votes: the political consequences of ballot design, innovation, and manipulation. Electoral Studies, 25 (4).Find this resource: Reynolds, J. 1988. Testing Democracy: Electoral Behavior and Progressive Reform in New Jersey, 1880–1920. Chapel Hill: University of North Carolina Press.Find this resource: Roberts, K. and Arce, M. 1998. Neoliberalism and lower-class voting behavior in Peru. Comparative Political Studies, 31 (2): 217–46.Find this resource: Robinson, J. and Verdier, T. 2003. The political economy of clientelism. Typescript. University of California, Berkeley.Find this resource: Safire, W. 1993. Safire’s New Political Dictionary. London: Random House.Find this resource: Schady, N. 2000. The political economy of expenditures by the Peruvian social fund (FONCODES), 1991–1995. American Political Science Review, 94 (2): 289–304.Find this resource: Schaffer, F. C. (ed.) 2007. Elections for Sale: The Causes and Consequences of Vote Buying. Boulder, Colo.: Lynne Rienner.Find this resource: ——and Schedler, A. 2007. What is vote buying? Pp. 24–47 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Shaffer. Boulder, Colo.: Lynne Rienner.Find this resource: Schmidt, S. W., Scott, J. C., Landé, C., and Guasti, L. (eds.) 1977. Friends, Followers, and Factions: A Reader in Political Clientelism. Berkeley and Los Angeles: University of California Press.Find this resource: Scott, J. C. 1969. Corruption, machine politics, and political change. American Political Science Review, 63: 1142–58.Find this resource: ——1972. Patron–client politics and political change in Southeast Asia. American Political Science Review, 66: 91–113.Find this resource: ——1976. The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia. New Haven: Yale University Press.Find this resource: Shefter, M. 1977. Party and patronage: Germany, England, and Italy. Politics and Society, 7: 403–51.Find this resource: ——1994. Political Parties and the State: The American Historical Experience. Princeton: Princeton University Press.Find this resource: Stokes, S. C. 1995. Cultures in Conflict: Social Movements and the State in Peru. Berkeley and Los Angeles: University of California Press.Find this resource: ——2005. Perverse accountability: a formal model of machine politics with evidence from Argentina. American Political Science Review, 99 (3): 315–25.Find this resource: ——2007. Is vote buying undemocratic? Pp. 117–43 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Schaffer. Boulder, Colo.: Lynne Rienner.Find this resource: Szwarcberg, M. L. 2001. Feeding loyalties: an analysis of clientelism, the case of the Manzaneras. Typescript. Universidad Torcuato di Tella.Find this resource: Tam, W. 2005. Political insecurity and clientelist politics: the case of Singapore. Typescript. University of Chicago.Find this resource: (p. 672) Urquiza, E. Y. 2006. Las eternas internas: política y faccionalismo en un municipio radical, 1983–1999. Pp. 57–80 in Democracia local: clientelismo, capital social, e innovación política en Argentina, ed. S. Amaral and S. Stokes. Buenos Aires: Eduntref.Find this resource: Van de Walle, N. 2007. Meet the new boss, same as the old boss? The evolution of political clientelism in Africa. Pp. 112–49 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Wang, C.-S. and Kurzman, C. 2007. Logistics: how to buy votes. Pp. 90–116 in Elections for Sale: The Causes and Consequences of Vote Buying, ed. F. C. Schaffer. Boulder, Colo.: Lynne Rienner.Find this resource: Wantchekon, L. 2003. Clientelism and voting behavior: evidence from a field experiment in Benin. World Politics, 55: 399–422.Find this resource: Washington Post Foreign Service 2006. Dirty politics “ingrained” in Mexico. June 26, p. A16.Find this resource: Weiner, M. 1967. Party Building in a New Nation. Chicago: University of Chicago Press.Find this resource: Weitz-Shapiro, R. 2006. Partisanship and protest: the politics of workfare distribution in Argentina. Typescript. Columbia University.Find this resource: Wilkinson, S. I. 2007. Explaining changing patterns of party–voter linkages in India. Pp. 238–97 in Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. H. Kitschelt and S. Wilkinson. Cambridge: Cambridge University Press.Find this resource: Wilson, J. Q. and Banfield, E. 1963. City Politics. Cambridge, Mass.: Harvard University Press.Find this resource: Notes: (*) I thank Carles Boix, Valeria Brusco, Noam Lupu, and Marcelo Nazareno for their comments. (1) A different phenomenon, which would be labeled campaign finance or corruption (depending on a country’s laws), is when private actors give money to politicians and parties in exchange for legislative concessions and other favors. In this relation, the flow of money is the reverse of the flow in clientelism: it goes not from politician to private actor but from private actor to politician. (2) Safire notes that the phase “probably is derived from the pre-Civil War practice of periodically distributing salt pork to the slaves from huge barrels” (1993). (3) These distinctions are conceptual, not empirical: a politician who deploys clientelist strategies may simultaneously provide public and programmatic-redistributive goods (see Magaloni, Diaz-Cayeros, and Estérez 2006). (5) Their distinction does not exactly follow the one I offer above. By programmatic redistribution they mean redistribution grounded in ideological visions of redistributive justice and implemented through income (and, less commonly, wealth) taxes. Tactical redistribution (which they also refer to as “pork”) is in a sense between classes of people: the examples they offer are subsidies to some industries and the location of military bases in some districts. Hence they shy away from truly individualized benefits, although their “machine case” would seem to involve individualized distribution and monitoring. (6) For an exception, see Robinson and Verdier (2003). (7) Hence the preference order of the clientelist party’s mild opponent is: –1/2(xix2)2 + b > −1/2(xix1)2 + b > −1/2(xix2)2 > −1/2(xix1)2. The preference order of the die-hard opponent is: −1/2(xix2)2 + b > −1/2(xix2)2 > −1/2(xix1)2 + b > −1/2(xix1)2. (8) Although the quantitative research has not been cross-national, we have learned much from systematic data generated by field studies (see Wantchekon 2003) and from survey research (see Brusco, Nazareno, and Stokes 2004). The latter study shows that parties were substantially more likely to try to buy the votes of poor than of wealthy Argentines. (9) Yet during the 2006 presidential campaign, “millions of poor Mexicans have been threatened with exclusion from health care and social assistance programs if they do not vote for various candidates [and] others, mostly in rural areas, have been given cash payoffs of$40 to \$60 for their votes” (Alianza Civica, cited in Washington Post 2006).

(10) Strategic debates within parties, in the present context over whether, for instance, to pursue “air” (advertising, propaganda) or “ground” (vote buying, face-to-face mobilization) strategies, are an indication of some uncertainty about what will work and what will not.

(11) An important feature of their paper is that it deals with the fact that effort to attract federal dollars by House members is a potential omitted variable. The authors deal with this problem by introducing federal expenditures in other districts in the same state as an instrumental variable.