WHAT IS CLIENTELISM? MAPPING THE CONCEPTUAL TERRAIN OF DISTRIBUTIVE POLITICS

The precise definition of clientelism and related terms has sparked considerable controversy. Most basically, some defend a narrow usage, where the term properly describes only the most literal quid-pro-quo exchanges—including paradigmatic patronage and vote-buying. On such accounts, clientelism describes exchanges where patrons can plausibly monitor behavior well enough to make distribution strictly conditional on individual-level support (Hicken Reference Hicken2011, 291; Stokes et al. Reference Stokes, Dunning, Nazareno and Brusco2014, 7). Taking distributive politics as the most general category of interest, they define programmatic politics as a system where resources are distributed according to public, universal rules, and non-programmatic politics as one where they are targeted to particular people. Clientelism is thus a subset of non-programmatic distributive politics, in which targeting is contingent on individual support.

Other scholars, however, use the term more expansively. In practice, they observe, the strict individual-level conditionality required by narrow accounts is quite rare (Auerbach et al. Reference Auerbach, Bussell, Chauchard, Jensenius, Nellis, Schneider and Sircar2022, 280). Rather, contingency comes in degrees and is often subject to uncertainty—as when clients are unsure about how accurately patrons can monitor their activity (Ferree and Long Reference Ferree and Long2016). Also, many real-world relationships that appear paradigmatically clientelist—with certain actors clearly inhabiting roles of “patron,” “client,” and “broker”—persist without any pretense of contingency (Hicken and Nathan Reference Hicken and Nathan2020), often because clients feel morally obliged to reciprocate (Lawson and Greene Reference Lawson and Greene2014). As ethnographic accounts across diverse contexts have shown, systems of particularistic exchange and distribution are often governed by complex norms of reciprocity, gratefulness, and honor (Finan and Schechter Reference Finan and Schechter2012; Piliavsky Reference Piliavsky2014). Many scholars thus identify clientelism with expectations of reciprocal exchange, rather than strict contingency (Aspinall and Berenschot Reference Aspinall and Berenschot2019, 2; Gonzalez-Ocantos and Oliveros Reference Gonzalez-Ocantos, Oliveros and Hannah2019, 3; Pellicer et al. Reference Pellicer, Wegner, Bayer and Tischmeyer2022, 933).

Our account follows expansive conceptions in highlighting continuities across a wide variety of particularistic practices. To be clear, though, this does not reflect a verdict on the analytical or empirical virtues of these rival conceptions, only our distinct normative aims in this article (see Kincaid, Pellicer, and Wegner Reference Kincaid, Pellicer, Wegner, Kincaid and Van Bouwel2023). While contingency does affect the democratic merits of distributive practices, it is hardly the only relevant factor. As we explore below, their scale, structure, and publicity all seem relevant for democratic values too. An expansive account enables us to track all of these variables, rather than privileging one over the others.

Still, expansive accounts of clientelism may risk stretching the concept too far. Consider the phenomenon of “pork-barrel politics,” where representatives channel resources toward localized “club goods” that disproportionately benefit their voters. Despite exhibiting expectations of reciprocal exchange, pork is widely tolerated in broadly “programmatic” contexts as a crude way to satisfy diverse interests. Indeed, even the most universal policies help some groups more than others, while parties generally expect greater support from those who benefit disproportionately from their policies. As such, a reciprocal exchange standard may seem to count even classically programmatic appeals as forms of clientelism.

This challenge reflects the continuous nature of the underlying variable here. In practice, that is, politicians’ choices about how to distribute resources are motivated to varying degrees by the expectation that recipients will support them; similarly, voters’ choices are motivated to varying degrees by expected material rewards. Many empirical researchers set an arbitrary cutoff to facilitate case selection: for instance, Pellicer et al. (Reference Pellicer, Wegner, Bayer and Tischmeyer2022) classify relationships as clientelist when expectations of reciprocity are “the main rationale” for the actors (see also Kincaid, Pellicer, and Wegner Reference Kincaid, Pellicer, Wegner, Kincaid and Van Bouwel2023, 402–3). Given our normative aims, however, we prefer terminology that preserves the underlying continuity and enables us to include scale, structure, and publicity (alongside reciprocal expectations) as additional dimensions of “particularism.” We thus interpret “clientelism” and “programmatic politics” as ideal types, situated at opposite poles of a continuum, rather than discrete categories defined by necessary and sufficient conditions.

Paradigmatically “clientelist” practices lie near the maximally particularistic pole of this continuum, exhibiting the greatest reciprocal motivation as well as the most individualized, personalistic, privatized modes of distribution. Meanwhile, the ideological competition practiced by paradigmatically programmatic parties lies closer to opposite pole, pairing low reciprocal motivation with highly collective, institutionalized, public modes of distribution. That said, few real-world practices are pure realizations of either ideal type: most clientelist parties proclaim certain programmatic goals, while traces of reciprocal motivation persist in most “ideological” appeals made by broadly programmatic parties.Footnote ² Rather than insisting on purity, we use “clientelism” and “programmatic politics” to refer to practices situated close to one ideal-typical pole or the other. This tracks standard usage and avoids stretching either concept, while also enabling us to make better sense of the many intermediate practices that include a mix of particularistic and programmatic features—separating terminological controversies from genuine normative disputes, and focusing our attention on the latter.

WHAT IS DEMOCRACY? CLARIFYING THE NORMATIVE STAKES OF DISTRIBUTIVE POLITICS

Of course, the second key term in our title is no less controversial than the first. Again, however, settling complex debates about the definition and value of democracy is beyond our scope. Instead, we identify three mid-level theses that must be accepted in order to vindicate the specific suggestion we set out to examine: namely, that some particularistic practices serve democratic ends by fostering collective power among disempowered groups. In this section, we establish that each thesis is broadly plausible—and compatible with diverse foundational assumptions—before examining their implications for political ethics in subsequent sections.

The first two theses draw primarily from normative democratic theory, holding that (a) leaders have significant but defeasible reasons to avoid particularistic exchanges; and (b) leaders also have strong reasons to foster organized forms of collective power, especially among disempowered groups. The third thesis draws out the practical implications of the first two, turning to empirical research on what facilitates such collective power, concluding that (c) social–organizational density and identities grounded in material interests are uniquely conducive to the urgent democratic goal of building power among disempowered groups.

To be sure, none of these theses is entirely uncontroversial, and none can be sufficiently defended from all worries in such a short space. What we demonstrate in this section is rather that a wide range of people accept (or have good reason to accept) all three. And if we accept them—the rest of the article contends—we should also accept that some particularistic tactics may be justified as the best way to promote democratic ends in certain non-ideal conditions.

CAN “COLLECTIVE CLIENTELISM” FOSTER COUNTERVAILING POWER AMONG DISEMPOWERED GROUPS?

Countless real-world distributive practices exist somewhere between the particularistic and programmatic poles on the continuum outlined above—and the differences among them matter for the question at hand. For instance, some patrons recruit clients on an individual basis, bribing supporters with cash, cell credit, or household goods, in transactions that most closely resemble the ideal-typical “vote-buying” considered by philosophical critics. Even when they do not become coercive, such individualized transactions atomize voters, pitting them against one another and precluding broader, more collective forms of identification and organization. Other parties, however, recruit larger voting blocs as their “clients,” for instance, by providing collective goods like sports fields or health centers preferentially to supportive districts or neighborhoods. In doing so, they interact with voters as members of groups—often through established organizations like churches and schools—and their distributive practices thus create multi-level organizational structures, sustain social networks, and highlight the shared material interests of group members. Where small-scale distribution typically threatens democratic ends, in other words, large-scale particularistic distribution may foster just the sort of social–organizational density and materially grounded collective identity that helps counter-hegemonic groups overcome obstacles to collective action.

Relatedly, some brokers operate independently, competing for access and constituencies from election to election, whereas others are integrated into dense, stable networks, often via institutions such as unions, churches, and parties. Here, the democratic stakes might seem more ambiguous: when brokers must compete, after all, individual clients have greater leverage, and may be able to extract more resources for themselves. Yet because it entrenches individualized, small-scale patterns of distribution, such personalistic competition also hinders the formation of collective power. By creating enduring social and organizational infrastructure that supports genuine countervailing power, by contrast, more institutionalized networks of distribution will enable disempowered groups to extract much larger and more enduring concessions from political elites than they could as individuals (see Bika Reference Bika2011).

In addition to their scale and structure, finally, the publicity of distributive practices is relevant to their democratic value. Some patrons draw on private largesse to finance rewards, for instance, while others manipulate the distribution of public resources—giving government jobs, desirable school places, or municipal services like street repairs to supporters. Again, both models have serious democratic costs: practices that subvert bureaucratic neutrality, for instance, may intensify corruption and public distrust, ultimately weakening state capacity. Nevertheless, greater publicity is ultimately more conducive to the building blocks of countervailing power, and for that reason we take it to be preferable on balance. Hidden and privatized forms of distribution exacerbate clients’ dependence on the whims of patrons, as well as the difficulties facing disadvantaged groups in their efforts to understand and contest distributive practices. When patrons reward clients by manipulating the distribution of public resources, by contrast, everyone still has a nominal claim to those resources, and may have some official recourse if they are denied. Even when the public rules for distribution are manipulated, and the public reasons offered for policies are insincere, their existence creates structures of claim-making that can later be wielded strategically by democratic forces, and may constrain the most blatant partiality via the “civilizing force of hypocrisy” (Elster Reference Elster and Elster1998). Compared to openly partial, privatized practices, finally, those that wear the mask of public justification are more likely to facilitate a gradual expansion of the identity group demanding and receiving benefits—as discussed below—rather than “locking in” smaller-scale practices.

As noted, these dimensions help constitute the particularism scale outlined above, with highly individualized, personalistic, and privatized practices—such as vote-buying—sitting near the maximally particularistic pole; and highly collective, institutionalized, and public practices—like those of paradigmatically programmatic parties—sitting near the maximally programmatic pole. Meanwhile, the cluster of practices most central to the suggestion examined here—that is, those that might serve democratic ends by fostering collective power among disempowered groups—sit somewhere between these poles.

The term “collective clientelism” is often used to refer to such practices (Bénit-Gbaffou Reference Bénit-Gbaffou2011; Burgwal Reference Burgwal1995; Gay Reference Gay1998; Reference Gay1999; Montambeault Reference Montambeault and Hilgers2012; Pellicer et al. Reference Pellicer, Wegner, Bayer and Tischmeyer2022), and we follow that usage here. After all, their “collective” character is the key reason they are thought to play a more positive democratic role than more paradigmatic forms of clientelism. Whereas individualized, personalistic, and privatized practices of particularistic exchange clearly exacerbate the political demobilization and disempowerment of those poor voters who are most in need of better representation (Stokes Reference Stokes2005), practices of “collective clientelism” that engage such voters as members of a group appear to facilitate the sort of “collective organization and identity” (Gay Reference Gay1998, 14) they need to achieve better representation. Though they depart from the programmatic ideal in absolute terms, diverse scholars have suggested that the way they do so may create the conditions for a more egalitarian balance of power among social groups, and perhaps, over time, a more programmatic form of competition.

As an illustration of how “collective clientelism” can create organizational structures that disempowered groups subsequently use to wield collective power in more programmatic ways, consider the following example from Argentina, recounted by Javier Auyero—an early pioneer of the suggestion we are examining (e.g., Reference Auyero2001). In June 1996, the residents of several Patagonian oil towns blockaded access roads into their region to protest joblessness and the privatization of the state oil company, accusing the political class of dishonesty and demanding “genuine sources of employment” (Auyero Reference Auyero2004, 321). Federal troops were mobilized to clear the road but were intimidated by the number of protesters—20,000, according to some estimates—and within a week, the provincial Governor had conceded most of their demands. Crucially, Auyero, Lapegna, and Poma (Reference Auyero, Lapegna and Poma2009, 14) note that this was not merely a spontaneous eruption of popular energy: rather, its success depended on the distribution of resources and information via “well-established patronage networks.” For one, broadcasts by a former mayor on allied radio stations framed unemployment as an injustice to citizens of the region, which required government redress. More concretely, that same former mayor is widely believed responsible for the free taxi rides that brought picketers to the blockades; the food, gasoline, firewood, cigarettes, and diapers they received there; and the trucks and bulldozers that showed up in support (Auyero, Lapegna, and Poma Reference Auyero, Lapegna and Poma2009, 13–4).

In our terms, the success of this disempowered group in exercising collective power to defend their interests depended on social connections, organizational structures, and collective identities originally fostered and sustained by practices of “collective clientelism.” Of course, we have neglected many details in this case, and a comprehensive normative assessment of it would require much further exploration. Our aim in discussing it is not to treat any of the actors involved as morally exemplary but simply to provide a “proof of concept” for the basic process at the center of the normative conjecture we examine: that is, to show that practices of collective clientelism can create key building blocks of countervailing power. And as the ethnographic literature demonstrates, this case is hardly unique: Auyero is only one of many scholars to observe something like this basic sequence, and to suggest on its basis that something like “collective clientelism” may have democratic value. While many empirical questions remain unsettled, therefore, we take these convergent results—helpfully summarized in a meta-analysis by Pellicer et al. (Reference Pellicer, Wegner, Bayer and Tischmeyer2022)—as sufficient for our purposes.

The key question for political ethics, more specifically, is not only whether some particularistic strategies could create building blocks for countervailing power, but whether adopting them is ever the best way of doing so. And given our pro tanto reasons to avoid particularism, this is plausible only when more programmatic alternatives are inaccessible or radically insufficient. The next two sections examine two such contexts in detail.

COLLECTIVE CLIENTELISM AS A TRANSITION TO PROGRAMMATIC DISTRIBUTION

One reason programmatic methods of building countervailing power might be unavailable is that habits of small-scale clientelist exchanges are too deeply ingrained among poor voters. In such contexts, a party that tries to attract these voters on purely programmatic grounds has little chance of success. Instead, we suggest, democracy-seeking leaders may be justified in temporarily resorting to strategies further toward the particularistic pole. More specifically, a party seeking to build a constituency among disempowered voters may permissibly engage in whatever sort of particularistic distribution those voters are accustomed to—including, if necessary, the most individualized, paradigmatic forms of clientelist exchange. As time goes on, however, they can (and must) shift linkages onto more collective, programmatic grounds.

This suggestion has analogs from other competitive contexts. Where certain strategies are clearly advantageous, yet cannot be prevented, abstaining from them altogether may have perverse effects, simply granting systematic advantages to the least scrupulous competitors. Insofar as it is undesirable for more scrupulous competitors to withdraw from the competition entirely, then, some adversarial ethicists argue they may employ such tactics—though always subject to certain limits (Bagg and Tranvik Reference Bagg and Tranvik2019; Schedler Reference Schedler2021). This section applies that logic to the present case, articulating the conditions in which certain particularistic practices may be permissible, and the limits democracy-seeking actors must respect in using them.

First, of course, parties eschewing particularistic methods must really be at a severe competitive disadvantage. And at times, it may be difficult to tell whether this condition is met: that is, whether programmatic methods of fostering collective power are sufficiently closed off to justify a resort to particularistic alternatives. Nevertheless, it is clearly met in some real-world cases. A common theme of recent empirical literature, for one, is that clientelist exchanges are frequently driven by bottom-up demand from voters, rather than imposed by politicians (Anciano Reference Anciano2018; Cooperman Reference Cooperman2024; Nichter and Peress Reference Nichter and Peress2017). Moreover, clientelist relationships are often governed by well-developed norms (Finan and Schechter Reference Finan and Schechter2012; Lawson and Greene Reference Lawson and Greene2014; Piliavsky Reference Piliavsky2014), which can amount to an alternative interpretation of the social meaning attached to democratic mechanisms like voting (Borges Reference Borges2019). And whether or not we accept this alternative interpretation as legitimate, the perceived legitimacy of many forms of particularistic exchange can make them quite difficult to dislodge. In such conditions, even parties with clear programmatic goals—such as the Brazilian Workers’ Party—feel pressured to adopt particularistic practices in order to survive (Johannessen Reference Johannessen2020).

Supposing that at least some democracy-seeking leaders do in fact face such conditions, then, how can they pursue particularistic practices responsibly? To start with, the tactics we discuss are only plausibly justifiable among groups that are actually counter-hegemonic: particularism remains impermissible among groups that already enjoy major advantages in organizational capacity. Meanwhile, one hard constraint their would-be leaders must respect is a prohibition on the most obviously coercive strategies: regardless of the competitive advantage they may appear to bring, they are simply too cruel, demoralizing, and (thus) damaging to democracy (Mares and Young Reference Mares and Young2019). Still, that leaves many other particularistic tactics on the table. To render them permissible, we argue, democracy-seeking leaders must respect five additional constraints: namely, they must (a) continually increase the scale and publicity of their distributive tactics, (b) consciously shift voter loyalty and linkages onto ideological grounds, (c) publicly maintain a programmatic orientation, (d) gradually eliminate opportunities for particularistic distribution for everyone, and (e) avoid fomenting intractable inter-group conflicts. In the rest of this section, we explore each requirement in greater detail.

First, how might a party gradually expand the scale of distribution? In a context where building materials are the salient currency of clientelist exchange, a new party may need to distribute such materials to individual families, at first, to establish its viability. Over time, however, it could expand to the block and then the neighborhood as a whole, ultimately transitioning to public works projects serving the entire district. Voters who expect quid-pro-quo exchanges may be more tolerant of these gradual shifts, rather than an abrupt transition to a new mode of political engagement. Yet by the time it is targeting district-level public goods toward broadly supportive districts, the party is simply engaged in pork-barrel politics—a form of collective clientelism—and from there, it is well positioned to keep scaling up.

As voters shift their expectations about the scale of distribution over time, meanwhile, they may also change how they understand themselves as political agents in relation to the state, political parties, and one another. As noted, family-level distribution separates and individualizes voters, pitting them against one another in a competition for limited resources, and making collective identification and organization more difficult. By contrast, public works that serve an entire neighborhood or district will likely have the opposite effect—especially if they provide physical spaces where individuals can build community and coordinate political action. Such strategies are likely to have desirable “policy feedback” effects (Pierson Reference Pierson1993), in other words, and can facilitate “constituency-making” (Disch Reference Disch2021) in service of democratic ends. Insofar as the party also defends a consistent ideology, more broadly, voters initially attracted by particularistic benefits may begin to identify with that platform—making their support less contingent on particularistic distribution of any kind.

As this suggests, it is crucial for democracy-seeking parties to publicly articulate their long-term programmatic vision, even when that is not their primary means of generating support. In practice, many broadly clientelist parties attract at least some voters on a programmatic basis. Meanwhile, voters often mix particularistic, ideological, and identity-based motivations in ways that are obscure even to themselves. Indeed, the gradual transition we propose relies on this ambiguity, which allows linkages to shift incrementally over time in a more democratic direction. (As we explore below, public articulation of an egalitarian, universalist, counter-hegemonic ideology can also serve as a key commitment device.)

In this way, our account preserves two key normative roles envisioned for parties—as vehicles of public justification (Rosenblum Reference Rosenblum2010; White and Ypi Reference White and Ypi2016) and mechanisms for linking citizens to state institutions (Wolkenstein Reference Wolkenstein2016)—while allowing that under certain non-ideal circumstances, the presumptive hierarchy among these roles may be reversed. Ideally, linkage would follow justification, as voters choose to affiliate with parties whose justifications are most compelling, and only then gain valuable linkages to state institutions. But where that process fails, collective clientelism can still create valuable linkages, which may gain further value over time as their justificatory component gains salience.

To entrench these changes, of course, democracy-seeking parties must also ensure that clientelist exchange is inaccessible to everyone. After using particularistic tactics to gain power, more specifically, they must change the rules that make those tactics possible—engaging in “mutual disarmament” to ensure no one can employ them going forward. To stay competitive and deny the least scrupulous actors a monopoly on power, for instance, a democracy-seeking party may need to manipulate the distribution of state resources like street repairs and school places in the short term. Once solidly in power, however, it must eliminate discretionary control, for instance, by universalizing or randomizing distribution.

Finally, democracy-seeking leaders must avoid strategies that place strict limits on the expansive trajectory we have outlined. For instance, many stubbornly persistent clientelist systems are stabilized in part by ethnic cleavages that are difficult to dislodge. This is one reason to emphasize identities grounded in shared material interests, where possible, since these tend to be more fluid and open-ended than ethnicity and other ascriptive identities, making them more amenable to coalition-building across cleavages (Shapiro Reference Shapiro2016). That said, the latter cannot be ruled out entirely: the shared structural position of many disempowered groups is closely associated with membership of ethnic, racial, or religious groups, and identification with those groups can be powerful sources for counter-hegemonic political solidarity. Importantly, marginalized groups can develop politically potent collective identities through shared experiences of oppression like racism or caste-based exclusion (Shelby Reference Shelby2002)—that is, by invoking shared structural positions and material interests—rather than “thick” identities grounded in race, ethnicity, culture, and other exclusionary categories.

Within the bounds set by these general constraints, much is left to contextual judgment. For instance, how can democracy-seeking leaders pick the right moment to scale up distribution or eliminate discretionary control—that is, when doing so will no longer jeopardize their competitiveness? Although it is often quite easy to identify which groups are hegemonic and counter-hegemonic, meanwhile, there may also be some genuinely difficult cases. Even if a group is disempowered relative to hegemonic elites, finally, enhancing its collective power may come at the cost of groups that are even more disadvantaged: how should we evaluate the overall democratic impact of doing so? Such tradeoffs are inevitable in political ethics, and many of the resulting dilemmas cannot be resolved in general terms.

Nevertheless, the constraints we have articulated offer some protection against the most troubling possibilities implied by the uncertainty inherent to this domain. By promoting a long-term programmatic vision among its members that includes gradual expansion in the scale of distribution, for instance, a party can ensure that its future leaders will face costs from the rank-and-file for unduly postponing the next steps in its transitional sequence. By publicly embracing a counter-hegemonic orientation and avoiding a reliance on fixed, inherently exclusive identities, meanwhile, it enables those left behind by an initial coalition to credibly demand inclusion—thereby incentivizing wider and wider coalitions over time, and limiting the danger that the initial phases of a transitional strategy might do lasting harm to other disempowered groups. While such risks cannot be eliminated entirely, we contend that a party following these guidelines can plausibly expect to foster collective power among an ever-expanding coalition of counter-hegemonic groups, ultimately sowing the seeds for a more fully programmatic and egalitarian form of competition.

Is this trajectory feasible in practice? Historically, shifts toward programmatic politics have had more to do with economic development, intra-elite competition, fiscal constraints, and other (largely) exogenous factors, rather than strategic uses of particularistic politics by democracy-seeking party elites (Buquet and Piñeiro Reference Buquet and Piñeiro2016; Kuo Reference Kuo2018; Mares Reference Mares2022). From the perspective of political ethics, however, the relevant question is not how frequently this strategy has been employed in the past, but how likely it is to be successful, if and when it is employed. If programmatic strategies fail to gain traction with disempowered voters because of deeply entrenched clientelist habits, in other words, should democracy-seeking leaders embark on this alternative mode of generating collective power, or is it simply doomed to fail? And framed in that way, the historical evidence is actually quite promising: while we know of no case following the precise trajectory outlined here, elements of it are present across many cases, which can collectively establish the strategy’s basic viability.

As Zografia Bika (Reference Bika2011) documents, for instance, party leaders in 1980s Greece gradually shifted rural voters from traditional, small-scale clientelist practices—anchored in “village brokers and local patrons operating on an undisputed personalised basis” (356)—toward a more “bureaucratic” form of clientelism “whose grassroots connections were mediated through the party machine, community council, co-operative, farmers’ union, coffee-shop and various village associations” (363). As Bika notes, “the brokering of deals and calling of shots never stopped”; it simply “changed face” as rural citizens “became clients and ideological subjects of the party” (363) rather than personally dependent on specific brokers.

In a very different context, Nam and Nethipo (Reference Nam and Nethipo2022) attribute the success of the Thai Rak Thai party in 2000s Thailand—in the face of serious repression—to its systematization of informal clientelist networks into more durable organizations, and its use of the resulting organizations to reinforce those particularist ties with programmatic content. As they explain, the party’s hybrid strategy initiated “palpable shifts in Thai political culture and the linkages that bound party to voters” (444), habituating voters to programmatic modes of competition. On Andrew Wyatt’s (Reference Wyatt2013) account, meanwhile, hybrid linkages cultivated by rival parties have had a similar effect in the Indian state of Tamil Nadu.

As a final illustration, consider the urban political machines central to US politics in the late 19th and early 20th centuries. Clearly, the eventual decline of particularistic exchange in this context had nothing to do with any long-term democratic commitments on the part of machine bosses. Nevertheless, the machines they created played a critical role in the political integration of millions of immigrants, at a time when the latter faced active hostility from nearly all major civic groups, trade unions, and other potential vehicles of collective organization (Golway Reference Golway2014). When programmatic tools were not available, in other words, particularist exchange served as an alternative method for creating key building blocks of collective power among disempowered groups—helping to equalize the relative power of different social groups. Meanwhile, the particularist origins of the resulting organizations and networks did not prevent an eventual programmatic transition: the danger of locking in perverse dynamics is real, but it is also far from inevitable. Although the strategy outlined in this section is hardly guaranteed to succeed in establishing durable collective power among disempowered groups, we conclude that its prospects are good enough to render it justifiable, in contexts where otherwise preferable programmatic methods face steep obstacles.

COLLECTIVE CLIENTELISM AS DEMOCRATIC SELF-DEFENSE

At the same time, it is also important not to valorize programmatic competition as such: after all, many serious threats to democracy emerge from voter engagement strategies that are perfectly programmatic in form. Within the realm of distributive politics, for instance, many formally universal policies governing resource distribution have the practical effect of enriching wealthy elites at the expense of disempowered groups (Bagg Reference Bagg2024, 91–107), which is widely understood to undermine democratic fairness. Yet the most dangerous programmatic appeals are often those that shift our focus away from material distribution altogether: after all, disputes about divisible material goods allow for compromise, whereas disputes about non-divisible goods (such as status) tend to foster more intractable forms of conflict (Shapiro Reference Shapiro2016). Think of strategies that rely on scapegoating minority groups, for instance, which clearly threaten democratic values despite appealing to voters on the basis of ideology rather than particularistic distribution.Footnote ⁵ At the limit, certain programmatic promises may even amount to directly limiting pluralism or eroding democratic protections—both catering to and fostering avowedly anti-democratic (yet ideologically defined) constituencies.

For obvious reasons, democracy-seeking leaders must fight these anti-democratic goals—ideally using entirely programmatic means. The trouble is that once anti-democratic parties gain power, they often seek to entrench themselves in part by capturing the channels of public communication and social organization within which programmatic challenges must be articulated (Levitsky and Ziblatt Reference Levitsky and Ziblatt2018). If they succeed, this constitutes a second context in which particularistic strategies might become permissible—subject to all constraints outlined in the previous section—as a way of breaking the incumbents’ stranglehold on the political system. Rather than constituting a tit-for-tat response that nullifies the advantages of an opponent’s norm-breaking in an adversarial competition, here the transgression of democratic principles is justified by the exigencies of “democratic self-defense” (e.g., Kirshner Reference Kirshner2014).

Why think practices of collective clientelism could play this role? Consider first that programmatic appeals must rely on episodic messaging through channels of mass media and public discourse that may be controlled by regime allies. By contrast, clientelist linkages rely instead on ongoing social relationships, cultivated via local brokers and intermediaries embedded within their target communities. As such, they offer democratic actors a channel for political interaction that allows them to bypass discursive institutions captured by elites. Meanwhile, programmatic parties must render their ideological appeals legible to members of disempowered groups: a task that is difficult under normal conditions, and especially so in degraded or captured information environments. By contrast, the benefits of particularistic exchanges are more immediate and tangible, meaning that clientelist strategies place fewer cognitive demands on voters. Where hegemonic elites may use their outsized control of mass communication to undermine the perceived legitimacy of programmatic challengers, finally, challengers relying at least partly on clientelist strategies will be less vulnerable to such attacks, because their claim to represent their constituents relies less on a public ideology that can be tarred as unpatriotic or illegitimate, and more on personal relationships and provision of tangible goods. In contexts where parties struggle to solidify enduring linkages through ideology alone, therefore, they can plausibly employ particularistic strategies to circumvent hegemonic capture of mediating institutions and the broader public culture, and thereby create the necessary building blocks for genuine countervailing power—building political loyalty and laying the groundwork for more ideologically driven mobilization over time.

India’s recent democratic history offers a helpful illustration of both key claims here: that is, that programmatic politics does not, in itself, serve democratic ends, and that particularist strategies could serve as a plausible form of democratic self-defense. Consider first the rise of the Bharatiya Janata Party (BJP) under Prime Minister Narendra Modi. Although the BJP occasionally employs certain quasi-clientelist practices, it is distinctive among Indian parties in part for its heavily programmatic orientation (Ammassari, Fossati, and McDonnell Reference Ammassari, Fossati and McDonnell2023)—grounded in relentless promotion of Hindu nationalism. Judged only in terms of the prevalence of particularistic distribution, therefore, the BJP’s rise looks like a democratic success story.

Of course, most scholars draw precisely the opposite conclusion, because despite their programmatic form, the BJP’s goals are substantively anti-democratic. Most obviously, its emphasis on Hindu nationalism is anti-pluralist, furthering the exclusion and oppression of Muslims. Perhaps more subtly, the axis of ideological contestation and identity cleavages it promotes through its hegemony on media and public culture have the effect of undermining solidarity among other genuinely disempowered caste and class groups, thereby discouraging collective organization against wealthy and upper-caste elites: that is, those who truly occupy a hegemonic position in society. The BJP’s rise thus presents a clear case of “divide-and-rule” tactics being used to undermine counter-hegemonic collective action and illustrates why a transition to programmatic politics is not merely insufficient for democratic goals but at times can be actively harmful. In such a context, we have suggested, democracy-seeking actors could permissibly turn to collective clientelism as a mode of democratic self-defense, so long as they respect the constraints outlined in the previous section.

Here, consider the career of another prominent Indian party: the Bahujan Samaj Party (BSP), which emerged as a powerful electoral force in the 1990s and 2000s with the explicit aim of “captur[ing] state power for the benefit of the oppressed majority” (Kumar Reference Kumar2017, 21). Where the BJP’s preferred terrain of ideological competition was marked by non-distributive issues of national or religious identities (positioning Hindus against Muslims), the BSPs centered economic and caste-based issues, positioning poor lower-caste groups (the bahujan) against wealthy upper-caste elites. These are precisely the sort of shared structural positions highlighted above as the most promising sources of counter-hegemonic collective identity. And crucially, it supplemented this egalitarian ideological message with practices of “collective clientelism” benefitting Dalits (formerly “untouchables”), such as caste quotas in government jobs. This strategy generated electoral dividends, enabling it to govern Uttar Pradesh—India’s most populous state, with around 240 million inhabitants—for many years.

The BSP’s early successes did not last, in large part due to the BJP’s success in exploiting sub-caste differences and splitting its core vote (Gautam Reference Gautam2023; Jaffrelot Reference Jaffrelot2019). Indeed, its fate reflects the danger of relying on ethnic identities of any kind, even when framed in structural terms. Still, its failure was not inevitable. An alternative version of the BSP—one that more faithfully respected the constraints outlined in the previous section, perhaps—could have built on its initial successes with collective clientelism in a more strategic and democracy-promoting way. After attracting supporters through particularistic rewards and integrating them into socially dense organizations, it could have fostered more enduring ideological loyalties and more open-ended, materially grounded forms of collective identity. If so, perhaps it could also have made a stronger stand against the BJP’s ethnonationalist authoritarianism—thereby using particularist distribution to protect democracy from a pathology of programmatic politics.

Thus far, the possibility that collective clientelism could be used as a tool of democratic self-defense remains hypothetical. Given that India is hardly the only democracy where traditional programmatic methods are struggling to contain growing anti-democratic ideologies, however, we believe that possibility is well worth our attention.