Democracy is about including those who are potentially affected by collective decisions in making those decisions. It follows that contemporary democratic theory is mostly concerned with institutions and practices through which the voices of citizens find a place in collective decision making. And it also follows that contemporary democratic theory should focus on the means for voice-based inclusions: distributions of rights, including rights to vote, speak, and associate; individuals’ standing as citizens and their capacities of citizenship; and the practices, organizations, and institutions that translate voice into collective decisions. For these reasons, contemporary democratic theory is modeled primarily on membership combined with empowered voice.

An alternative to voice is exit. As conceived in Hirschman's (Reference Hirschman1970) classic work Exit, Voice, and Loyalty, given the opportunity, dissatisfied members may choose to exit a collectivity rather than to seek reform or response by voicing their dissatisfaction. If voice is associated with democracy, exit is associated with markets: Individuals choose organizations that best serve their needs and exit those that fail them. They vote with their feet.

The trade-offs and interactions between voice and exit have been extensively and brilliantly examined by Hirschman and a wave of literature generated by his work (Dowding and John Reference Dowding and John2008; Dowding et al. Reference Dowding, John, Mergoupis and Van Vugt2000; Hirschman Reference Hirschman1970, Reference Hirschman1974, Reference Hirschman1978). Hirschman's goal in Exit, Voice, and Loyalty was to introduce the concept of voice to economists, habituated to conceiving exit as the only mechanism through which firms are held accountable by their customers. Voice, Hirschman argued, is essential to renewing organizations because it conveys information from the most quality-conscious and active members/customers to organizational leaders, enabling the kinds of responsiveness necessary for organizations to recuperate from mistakes and to hold at bay the more destructive option of exit.

If Hirschman sought to introduce the concept of voice into economics, my aim here is to reintroduce the concept of exit into democratic theory. I use the term “reintroduce” because the concept has played a major, though undertheorized, role in the institutions and practices of democracy. Exit mechanisms have always been recognized as essential to electoral democracy, as citizens leave one party and join another. Rights of exit are logically implied by all democratic theories with roots in liberal contract theory (Green Reference Green1998; Kukathas Reference Kukathas2003). Theorists who see a civil society composed of voluntary associations as a key element of democratic government assume that “voluntariness” brings with it rights as well as capacities for exit from associations (Rosenblum Reference Rosenblum1998; M.E. Warren Reference Warren2001). More recently, theorists of multiculturalism and feminism have advocated rights of exit from groups as a hedge against the coercive capabilities of ethnic and religious groups within liberal-democratic societies (Benhabib Reference Benhabib2002, 131–32; Deveaux Reference Deveaux2006, 47–53; Kukathas Reference Kukathas2003; Okin Reference Okin2002; Philips Reference Philips2007, chap. 6).

But with a few important exceptions (Gastil Reference Gastil2000, chap. 2; Shapiro Reference Shapiro2003, 43–64; Sørensen Reference Sørensen1997), exit has not been theorized as a generic feature of democratic theory and practice. Tellingly, “exit” is not to be found in the indexes of works of contemporary democratic theory. When the concept does make an appearance, it is often viewed as an undesirable alternative to democracy understood as common choice, one that atomizes collectivities and damages collective loyalties, erodes collective choice and undermines provision of common goods, encourages excessive individualism, drains collectivities of their best and brightest, and substitutes silence for voice (see, e.g., Barry Reference Barry1974; Elster Reference Elster, Bohman and Rehg1997, 33–34; Healy Reference Healy2007; Held Reference Held2006, 94–98, 235; Newman Reference Newman2007; Olson Reference Olson2006, 48–52, 92).

I argue here that exit-based empowerments—individual rights and capacities for exit—are as central to the design and integrity of democracy as the distributions of votes and voice long considered its key structural features. My theoretical aim is to conceptualize exit-based empowerments as generic features of democratic institutions. My normative aim is to identify the contributions of exit-based empowerments to democratic goods. The primary scope of my analysis is the strong, liberal-democratic constitutional states that have effective monopoly powers over public goods provision, and have some capacity to design and reform institutions in ways that increase their responsiveness to citizens. Many of my examples of exit-based empowerments are drawn from these contexts, including, for example, competitive party systems and civil societies composed of voluntary associations, within which exit-based empowerments are central. Although the primary context of my analysis is that of strong liberal-democratic states, because I conceptualize exit-based empowerment as a generic feature of democracy, it will also have some applications in weak state contexts, as well as within non-democratic regimes and contexts. But I shall comment upon these contexts only in passing.

The argument unfolds in six sections. In the first, I note that exit-based mechanisms are poorly theorized within democratic theory, largely because they are equated with market mechanisms. Democratic theorists have instead implicitly adopted a “voice–monopoly” model characterized by limited exit and focused on voice-based institutions and practices. Although the costs of exit are often emphasized in democratic theory, the costs of the voice–monopoly model to basic democratic norms are not. In the second section, I identify and justify two norms that are increasingly common in contemporary democratic theory: nondomination and the all-affected principle. I consider the challenges of instantiating these norms under several contemporary circumstances of politics, which include challenges of size, scale, high resource thresholds for collective action, and the migration of loci of decision-making across borders and into nondemocratic organizations. Within the domain of each challenge, I argue, the voice–monopoly model masks damage to democratic norms. In the third section, I begin the task of recovering the democratic possibilities of exit as an alternative by noting that the mechanism is, almost always, constructed into political systems, either by being enabled through rights and protections, or by being institutionalized by designing competition into states, as in the case of competitive party systems. Exit is, typically, connected to democratic goods in two ways: as a means of providing individuals with empowerments with low thresholds of effective use, and as ways of inducing responsiveness from organizations. Two further potentially democratic effects follow: Exit can often induce representatives or other elites to engage deliberatively those who can exit, and exit induces elites to cultivate reputations for trustworthiness. Each of these goods counts as a fundamental contribution to the basic norms of democracy. In section four, I suggest that these democratic potentials of exit have been masked because its political functions have not been well conceptualized. In particular, exit and voice do not trade off, because they refer to two different features of democracy: empowerment of individuals, and communication of values and preferences. Hirschman's exit–voice dichotomy confused the dimensions of empowerment and communication. So the alternative to rights-based inclusion is exit, whereas the alternative to voice is what I shall call “signal,” which is the communicative feature of exit. In the final section, I show how this conceptualization of exit—particularly, the concepts of enabled and institutionalized exit, and the distinction between exit as empowerment and exit as communication—pays off in institutional analysis. I use these distinctions to map the ways exit might be, should be, and often are designed into democratic institutions. Understanding exit-based empowerment, I conclude, will enhance our capacity to imagine institutions that deepen democracy under the contemporary circumstances of politics.

EXIT AND THE VOICE–MONOPOLY MODEL OF DEMOCRACY

The democratic potentials of exit-based mechanisms have gone largely untheorized. The key reason is that democratic theorists have adopted, usually implicitly, what I shall call here a “voice–monopoly” model of democracy. Exit, in this model, is a market mechanism based on individual choice. Democracy, in contrast, is about collective choice, enabled and empowered by strong, constitutional states. Exit, in this view, has little place in such contexts, not least because it would undermine collective decision and action. Rawls’ theory of a liberal-democratic society, for example, assumes a collectivity that members enter by birth and exit only through death (Rawls Reference Rawls1993, 41).Footnote ¹

The one explicit justification of the voice–monopoly model is found outside of democratic theory proper, in A. O. Hirschman's (Reference Hirschman1970) brilliant Exit, Voice, and Loyalty. Hirschman introduced the exit–voice dichotomy to explain how organizations repair and regenerate themselves in response to the dissatisfaction of their members. Voice and exit are two ways members can pressure organizations to change: They can speak out or they can leave. Voice presupposes that an organization is structured in such a way that it can respond to its members. Exit assumes that members have the opportunity and capacity to choose goods from competing organizations. In organizations that are responsive, voice and exit are complementary: The availability of exit induces organizations to provide members with voice, precisely so they can reduce the likelihood that members will leave with their dollars, their votes, their good will, or their capacity to contribute to the organization (Hirschman Reference Hirschman1970, 36). Loyalty, the third term in Hirschman's model, is a resource organizations can cultivate to counter exit and stabilize a core of their most committed and “quality sensitive” members, which in turn helps to ensure their long-term capacity to regenerate, reform, and thrive.

In the absence of loyalty and the presence of choices for members, however, exit and voice may trade off: Dissatisfied members will prefer exit to voice. In Hirschman's (Reference Hirschman1970) model, exit has a low threshold of difficulty relative to voice. It has the advantage of being “impersonal—any face-to-face confrontation between customer and firm with its imponderable and unpredictable elements is avoided and success and failure of the organization are communicated to it by a set of statistics.” By contrast, voice is “a far more ‘messy’ concept because it can be graduated, all the way from faint grumbling to violent protest; it implies articulation of one's critical opinions rather than a private, ‘secret’ vote in the anonymity of a supermarket; and finally, it is direct rather than roundabout. Voice is political action par excellence” (15–16).

The messiness of voice can be a barrier to its use, so that voice-based political action—the essential feature of democracy—is relatively disadvantaged whenever exit is available, at least when loyalty is not factored into the model. When individuals choose exit, they have control over the consequences of their acts, and they bear the costs and reap the rewards. When they exercise voice, however, they are subject to inherent qualities of politics: the “slow boring of hard boards”—as Max Weber ([1919] Reference Weber, Gerth and Mills1946, 128), so nicely put it; the inherent unpredictability of collective actions—in Hannah Arendt's (Reference Arendt1958, chap. 5, section 34) terms; the risks of speaking up; the possibility of large investments with little or no payback; and the psychological burdens of overt conflict. Thus, Hirschman writes,

Without members who are willing to take on the burdens of voice, organizations, including democratic states, are unlikely to be able to rejuvenate, reform, and perform. For these reasons, Hirschman (Reference Hirschman1970) suggests, a “no-exit” situation is sometimes preferable, assuming “(1) exit is ineffective as a recuperation mechanism, but does succeed in draining from the firm or organization its more quality-conscious, alert, and potentially activist customers or members; and (2) if voice could be made into an effective mechanism once these customers or members are securely locked in” (55).

Hirschman's reasoning provides the structural logic that maps onto the normative logic of contemporary democratic theory: Democracy is, in the first instance, about empowering the voices of citizens. But Hirschman's analysis also reveals a feature of democracy—or, rather, thinking about democracy—that is less remarked: The democratic case for voice usually assumes monopoly organizations. It does so normatively—voice is most important within the context of monopolistic organizations. And it does so structurally—monopoly induces voice by restricting exit.

In these two respects, Hirschman's analysis tracks the fact that modern democracy was born of a specific kind of monopoly—that of states. In its eighteenth- and nineteenth-century origins, the democratic project focused on increasing inclusions within states that had effectively consolidated power by controlling territory, developing administrative capacities, and regularizing sovereignty through constitutional means (Poggi Reference Poggi1990, chaps. 1–2). The justifications for voice are closely related to these elements of monopoly in two ways. First, when a collectivity controls key features of livelihood, such as security, and solves collective action problems through coercion, then individuals subject to that coercion should have a say in how it is deployed. Second, the greater the costs of exit to individuals, the greater the need for voice. Though liberal-democratic states do not legally restrict exit from their territories, they recognize that exit is costly: It is disruptive of family, social support networks, careers, language, and culture, and can mean giving up the protections and welfare entitlements of citizenship. These monopoly-like effects are well recognized and justified by the existence of voice mechanisms—that is, democratic processes that legitimate the monopoly-like properties of the state. Thus it is appropriate that democratic theory has focused on equalities of political resources, secured by positive political rights (voting, speech, association) and related welfare rights (education and income security), as well as on the mechanisms such as electoral systems, judicial systems, public sphere discourse, and civil society activism through which citizens’ voice is translated into influence over law and policy (see, e.g., Habermas Reference Habermas and Rehg1996, chap. 8).

The depth of attachment to monopoly within democratic theory stems from the fact that it is structurally necessary for the provision of common goods. As Hirschman's analysis suggests, democracies are sensitive to problems of collective action: Defectors from collectivities undermine democracy by undermining the possibility of common choice (Barry Reference Barry1974). Union organizing is the archetypal case: The worker who breaks with the solidarity of the bargaining unit also undermines the capacity of the union to serve its members. More generally, as Olson (Reference Olson1971) famously detailed, when individuals are left to weigh the costs and benefits of collective action, larger groups tend to return fewer benefits, causing individuals to exit the collectivity, which in turn undermines the provision of public goods. As Hobbes understood, monopoly removes the threats to common security and provision posed by defection. Similarly, democratic theorists—particularly those focused on the important relationship between solidarity and collective choice—view exit opportunities as harmful, indeed, so much so that, as Hirschman (Reference Hirschman1970) observes more generally, exit is often branded criminal or treasonous (17).

Similarly, democratic theorists typically understand opportunities for exit from public and publicly provided goods as undesirable, as they drain collectivities of their resources, degrading public provision (Barry Reference Barry1974; Healy Reference Healy2007; Newman Reference Newman2007; Olson Reference Olson2006, 48–52, 92). Hirschman (Reference Hirschman1970) identifies a particularly “noxious combination” of exit and voice that transforms public services into expensive, low-performing public entities (44–46). The presence of some amount of exit—from government rail transport to private trucking in Nigeria, or from public to private schools, and from high tax to low tax municipalities in the United States—drains organizations or jurisdictions of their “quality conscious” members, who are, typically, those who have the resources that enable them to exit. In this way, organizations—rail in Nigeria, and some public schools and municipalities in the United States—develop into what Hirschman calls “lazy monopolies”—expensive, inefficient, unresponsive, and poorly performing entities that have a lock on those without the resources to exit. The “lazy monopoly” phenomenon probably extends to pressures for democratization as well. Nineteenth-century Britain sought colonies to relieve internal pressures on internal labor markets, which elites feared could result in disorder. Their strategies also reduced pressures for democratic reform. Many authoritarian states—China, for example—encourage their most vocal dissidents to exit, thus reducing internal pressure for voice.

Following this logic, democratic theorists implicitly tie the very possibility of democracy to monopoly, whereas the concept of exit is left as an undertheorized residual loosely identified with market organization, and collective choice is modeled as an alternative to markets (see, e.g., Ackerman and Fishkin Reference Ackermann and Fishkin2002; Barber Reference Barber1984; Bowles and Gintis Reference Bowles and Gintis1987, chap. 5; Elster Reference Elster, Bohman and Rehg1997, 33–34; Held Reference Held2006, 94–98, 235; Macpherson Reference Macpherson1977; Olson Reference Olson2006, 48–52, 92; cf. Shapiro Reference Shapiro2003, 56). In this way, the concept of exit is identified with features of markets that undermine collective choice: Exit means that investors with mobile capital shop for the cheapest places to do business—places with weak labor and environmental standards, for example. Exit enables the wealthy to free ride on societies that are well ordered because they have strong public services and security (Hirschman Reference Hirschman1978, 96; Lindblom Reference Lindblom2001, chap. 17). Competition among local governments in federal systems, especially for capital, tends to corrode central state capacities to provide public goods (Cai and Treisman Reference Cai and Treisman2004). Libertarians and public choice theorists buy into the same dichotomy from the other side, preferring markets to government—the latter necessarily dangerous and almost always oppressive owing to its monopoly character. Democracy is dangerous on this view just because it builds on the monopoly powers of the state, which are no less onerous because political elites wield their powers on behalf of a majority or plurality (Buchanan Reference Buchanan1975; Hayek Reference Hayek1944). This same dichotomy was replayed in the resurgence of “neoliberalism” in the 1980s—the British-originated concept that identifies the privatization of government services with more market, less government—and, thus by implication, a shrinking of the domain of democracy in favor of the market. In these ways, the meanings of exit in most democratic theories have been tied to “private” zones of choice, particularly markets, protected but not controlled by the state, and hence beyond the proper domain of democracy.

It also follows from this model that when voice is effective and monopoly extensive, the bare fact of monopoly is, on average, good for democracy: Violence monopolized can also be tamed, turned into capacities for collective provision of common goods, and legitimated through various kinds of democratic devices (Warren Reference Warren, Dryzek, Honig and Phillips2006). Voice is not only normatively required by monopoly-based democracy, but also enabled: Dissatisfied citizens will pressure for change, and democratic institutions should both empower and channel this dissatisfaction. Under these circumstances, democracy is not only inclusive, but also directive: Through public opinion formation and electoral representation, peoples’ interests, preferences, and values are formed into voice and translated into policies.

For these kinds of reasons, most contemporary democratic theorists view exit as a threat to democracy, with the result that they have taken on, sometimes explicitly but more often implicitly, a theoretical commitment in favor of tight monopoly organization combined with voice—or what I am calling the voice–monopoly model of democracy.Footnote ²

COSTS OF THE VOICE–MONOPOLY MODEL

Although the costs of exit are often noted in contemporary democratic theory, the costs of the voice–monopoly model for democratic goods are rarely tallied. The costs are best appreciated by surveying the fit between the contemporary circumstances of politics and the voice–monopoly model in light of the basic norms of democracy: that is, the norms by which we should assess institutions and practices to be more or less democratic.

Contemporary democratic theorists are increasingly converging on two complementary norms. The first of these is nondomination. The norm that individuals should not be subject to domination has a long history traceable to early modern republican thought, but has recently been reemphasized by a number of thinkers who seek to relate the concept of democracy more directly to the good of self-determination (Bohman Reference Bohman2007, chap. 1; Pettit Reference Pettit1999, chap. 3; Richardson Reference Richardson2003, 28–36; Young Reference Young2000, chap. 1). Although there are relatively sophisticated accounts of domination as a form of power, for present purposes the intuitive meaning is sufficient: Individuals should not be subject to relations of power that deny them capacities and opportunities to self-determine. Stated positively, individuals should have empowerments—political, economic, and social—sufficient for them to avoid relations of domination and to exercise individual autonomy (see also Habermas Reference Habermas and Rehg1996, 84–87; Held Reference Held2006, chap. 10)

The second norm is the all-affected principle: Because many of the relationships that enable self-determination are collective in nature, self-determination requires that those potentially affected by collective decisions have the opportunity to influence those decisions (Goodin Reference Goodin2007; Habermas Reference Habermas and Rehg1996, 107, 365; Young Reference Young2000, 23). The all-affected principle is a principle of proportional inclusion: Individuals have a normative claim to influence collective decisions just to the extent that they are affected by collective decisions. Nondomination is a necessary but not sufficient condition for instantiating the all-affected principle insofar as it focuses normative attention on conditions of individual autonomy. The all-affected principle is more demanding and positive: Individuals should have effective opportunities to influence collective decisions that potentially affect them, including the voice to communicate their values, interests, and preferences and deliberate their relationship to collective decisions. The nature and scope of this norm remain under debate, not least because it is quite demanding in a world in which the effects of collective decisions on individuals are complex, extensive, and intertemporal (Goodin Reference Goodin2007; Näsström 2010; Young Reference Young2000, 23). For my purposes here, however, it is sufficient to note that the principle taps a long-standing normative intuition first expressed in the Code of Justinian from the sixth century: “What touches all ought to be considered and approved by all” (Cod. 5.59.5.2: quod omnes tangit, omnibus tractari et approbari debet). The updated version of the intuition is that, whatever else democracy might require, it is fundamentally about self-governance through collective decisions and actions.

Importantly, these principles abstract from institutions and practices, so that “democracy” is defined normatively rather than identified with specific institutions and practices such as competitive elections, majority rule, political rights and liberties, representative government, or deliberation. By abstracting democratic norms, we can ask how each kind of institution or practice advances (or damages) these norms.

The voice–monopoly model of democracy is one model for organizing these two norms into society. The norm of nondomination is closely linked to two long-standing and widely appreciated features of democratic institutions. On one hand, the familiar catalogs of rights protect against political sources of domination: rights such as those of speech, conscience, protections and self-incrimination, due process, and so on. On the other hand, many of these rights combine with others to produce positive powers, or choices. These include rights to vote, petition, and associate, as well as welfare rights aimed at material vulnerabilities. The voice–monopoly model of democracy theorizes these protections and enablements as working within the monopoly powers of the state. The protections function as limitations of state powers. But the model also assumes, importantly, that state powers can be directed to underwriting several dimensions of citizenship. Thus political citizenship is enabled by electoral systems and representative institutions. Legal citizenship is enabled by judicial institutions that provide standing, protection, and due process. And social citizenship is enabled by a variety of welfare entitlements as well as some regulation of economic powers. These enablements presume (rightly) states with monopoly powers of regulation, taxation, and redistribution. With respect to the all-affected principle, the voice–monopoly model of democracy assumes that the primary domain of collective affectedness is circumscribed by the state. Thus, the locus of collective self-governance is the state, and the primary means of self-governance are the rights, standings, and representative structures that empower voice within state or statelike entities.

The voice–monopoly model of democracy is not “wrong”: It instantiates these norms of democracy in ways that respond to important and legitimate problems of collective action—problems that remain with us. But the voice–monopoly model is increasingly insufficient. As I suggest below, it is insensitive to some long-standing problems of democracy, as well as being outgrown by new problems, in such a way that is not only costly to democratic norms, but also makes it difficult to conceptualize one key mechanism for instantiating democratic norms: exit-based empowerment. I detail the problems below and then theorize exit-based empowerment in the remainder of the article.

There are at least five problems areas, two of which are well recognized, and three of which are increasingly part of the contemporary landscape of politics. The first three problems generate normative deficits primarily with respect to the all-affected principle, whereas the remaining two problems generate normative deficits primarily with respect to nondomination.

RETHINKING EXIT: ENABLED AND INSTITUTIONALIZED EXIT

Democratic theory is not without resources to respond to the deficits of the voice–monopoly model. One of these resources, I am arguing, can be found in rethinking the democratic potentials of exit. It is because much contemporary democratic theory has slipped with little notice into the voice–monopoly model that such rethinking has not occurred. The reason for the slippage, I suggested above, is that the voice–monopoly model gets its traction primarily from problems of collective action, combined with its normative emphasis on voice, and (often implicitly) the attractions of a solidaristic conception of citizenship. The concept of exit is left as a residual category associated with the market. So when this residual concept drifts back into the political realm, the implications (on this implicit logic) are negative: Exit is a strategy associated with those who are excessively individualistic, who fail to recognize common dependencies, and who free ride on the common contributions of others.

These features of exit are, of course, empirically possible and even common. This article is neither an unqualified argument against the voice-monopoly model nor is it an unqualified argument for exit. What I am arguing is that, as a democratic theory, the voice–monopoly model fails to identify the conditions under which exit might function as a key mechanism for connecting democratic norms to contemporary circumstances of politics, disarming a key democratic device where it is needed the most.

That the voice–monopoly model undertheorizes exit is evident in the range of positive and negative normative valances that attach to its common meanings. Exit may mean shirking collective responsibilities. But it might also mean breaking a relationship of domination, as in quitting a bad marriage, a poor job, or a domineering ethnic or identity group. Exit might mean leaving an underperforming service provider, such as a clinic or school, in favor of another. Exit could mean leaving a political party by voting for the opposition. Or it might mean, simply, quitting, as does a drop-out who is not engaged by a neighborhood school. Exit might carry penalties for the organization that is exited, as when a voter sends his or her vote to one candidate rather than another. In such cases, exit generates attentiveness from the organization. But exit might also indicate lack of choice, as when citizens become discouraged with their limited choices for representatives, and choose not to vote at all. Or exit may, sometimes, simply indicate apathy, as when a social relationship or membership falls into disuse.

Examples such as these suggest that there is nothing inherently democratic (or undemocratic) about exit in itself. That is because we know little about what exit might signify or accomplish without knowing the social relationships it entails. For exit to count as an empowerment rather than silence, neglect, or shirking, an individual must (a) have the capacity to replace one kind of membership with another that is voluntarily chosen. To count as a democratic empowerment, (b) that capacity must be linked to consequences for the entity subject to exit in such a way that the entity has the incentive to respond to the individual's choice. From a normative perspective, conditions that meet (a), the capacity to choose, support nondomination, whereas conditions that meet both (a) and (b), linking choice to organizational responsiveness, support the all-affected principle.

The forms of exit that meet these conditions are, typically, constructed in one way or another by social and political institutions. Here I distinguish two forms:

• • Enabled exit is primarily about the ways and means though which individuals are empowered to make choices throughout a range of social, economic, and political relationships, including rights and protections, and welfare supports.

• • Institutionalized exit is primarily about organizing the monopoly powers of the state in such a way that individual decisions to enter or exit an organization, such as a political party within a competitive party system, serve to evoke responsiveness from that organization.

TWO ELEMENTS OF DEMOCRACY: EMPOWERMENT AND COMMUNICATION

Given the centrality of exit-based empowerment to democracy, why has it not played a key role in democratic theory, particularly in theories that focus on institutionalizing democracy within strong state contexts—that is, contexts in which states have effective (territorial) monopolies of coercive power? Why have recognitions been ad hoc rather than included among the generic mechanisms of democratic empowerment?

A key theoretical reason is that most democratic theories conflate two discrete problems of democratic organization in ways that obscure the potentially democratic functions of exit: (1) how power is distributed, and (2) how communications enter into collective decision-making. Exit, I have been arguing, is primarily about empowerments. But because contemporary democratic theory centers on voice—a modality of communication—it views exit not as an empowerment, but as silence. But this kind of judgment conflates distributions of power with communication, and leads to a mistaken polarization of exit-based and voice-based institutions, as if they trade off on the same register. They do not.

Ironically, it was a theorist of the thinnest version of democracy we have—Schumpeter—who understood this distinction, which he deployed to refocus “democracy” away from judgment and decision-making, and toward power. He develops the distinction in the course of introducing his famous (or infamous) definition of democracy:

This passage is usually noted within democratic theory for its view of the people as passive selectors of elites who make decisions (Held Reference Held2006, chap. 5). This (valid) critique notwithstanding, Schumpeter here suggests two important innovations.

First, he notes that democracy has two “elements”—one empowering the electorate and the other making collective decisions. Whether or not Schumpeter is accurate in his characterization of “classical” democratic theory (he was not: see Macpherson Reference Macpherson1977, 85; Pateman Reference Pateman1975, 3–5), he was right in charactering that theory as a theory of decision, conceived in terms of a directive voice of the people. Schumpeter argues that democracy should instead ensure that power resides with the people. He thus distinguishes empowered inclusion from decision, and thus mechanisms of empowerment from the substance of political judgment. (Schumpeter understood democracy as a theory of power, not judgment—though this point is not essential to the analysis here.)

Second, as Ian Shapiro (Reference Shapiro2003, 55) insightfully notes, Schumpeter identifies competition as the primary means of empowerment. “The underlying logic of his argument,” writes Shapiro, “is disarmingly simple. It reduces to a double claim: (1) That structured competition for power is preferable both to Hobbesian anarchy and to the power monopoly that Hobbes saw as the logical response to it, and (2) that the choices among anarchy, monopoly, and competition are the only meaningful possibilities.” Schumpeter's distinction highlights the fact that political systems are “democratic” only insofar as they are based on democratically distributed mechanisms of empowerment of the kinds that generates competition among political elites for followers who hold these empowerments. We do not need to follow Schumpeter in concluding that decision making is a matter only for elites. We simply need to note that empowerment and decision making are two distinct elements in a theory of democracy. Thus, a (democratic) theory of the distribution of power should focus on the means through which those affected by collective decisions gain standing to have influence over those decisions, effected through distributions of political, civil, and social rights, as well as distributions of substantive enablements, such as education. The theory of collective decision making should be about the ways and means though which participants form and communicate their preferences, and transform their preferences into collective decisions—through bargaining and deliberation, for example. One “element” has to do with how people get into a political conversation; the other with what happens in the conversation.

Once we make this distinction, we can see that exit, with its implied system requiring elites to compete for the power to decide, is primarily a means of empowerment of individuals, not a means of collective communication and judgment. For its part, voice now appears not primarily as means of empowerment, but rather as a feature of the communication necessary to collective decision making. We thus have two dimensions, empowerment and communication, in which democratic institutions can vary. Exit is mostly (but not only) about empowerment, and voice is mostly about communication. Thus the conceptual opposite of exit is monopoly, not voice. The conceptual opposite of voice is what I shall term signal, not silence. I specify these dimensions of exit more closely in the remainder of this section, and then put them to work to map the democratic configurations of exit in the final section.

If, as I have been claiming, the primary contribution of exit to democracy has to do with empowerment, then we can identify the contribution more exactly by relating exit to a basic account of power. We already have some tools: I argued earlier that exit contributes to nondomination when it is constructed in ways that enable individuals to break a relationship with some provider of representation, services, or other goods. So a baseline meaning of exit is the capacity to break a social relationship: to leave a service provider, to fail to buy a good, to quit an association, to refuse to be counted as a constituent of a political party or advocacy group, to leave a marriage, or to leave a conversation. Breaking a relationship implies a choice; and a choice implies the possibility that the individual can reestablish a relationship that will serve to provide alternative goods: another source of income, a more responsive representative, higher-quality services, a better marriage, a more relevant conversation, and so on. Thus, exit is empowered just to the extent that there are alternative providers of goods, so that no provider holds a monopoly over goods in a way that might become the basis for an unjustifiable dependency (see, e.g., Alston and Ferrie Reference Alston and Ferrie1993; Domar Reference Domar1970).

We can redescribe such relationships of choice in terms of power, and thus specify more closely the idea that exit might function as an “empowerment.” Every power relationship is an asymmetrical relationship of choice structured by control over resources: A has power over B, just insofar as A controls a resource wanted or needed by B, and B must do A's bidding to gain access to that resource. In every power relationship, choice plays a key role: B's actions are, typically, valuable to A just to the extent that B wills the performance. Thus, even in a slave relationship, A's threats are effective just to the extent that B does A's bidding. A slave who simply refuses, even to the point of death, has effectively broken the relationship of domination, though at the cost of replacing it with a relationship of pure violence.

Within this simple model of power, exit functions to change asymmetrical relationships of power (with slavery a limiting kind) into symmetrical relationships of choice. A's power over B is diminished just to the extent that A's monopoly over resources B needs or wants is dissipated, which means that B can choose to exit a relationship with A, in favor of relationships with C, D, E, etc., who represent alternative providers of resources B needs or wants. As B's choices expand, A's asymmetrical power over B may be transformed into a relationship of choice. Note that this formulation does not count as an argument against the monopoly powers of the state as such: Monopoly is often justified and necessary for providing common goods. But it is an argument for designing exit-based empowerments into such institutions (think competitive party systems) as suggested earlier.

This simple model of power makes it easy to identify the conditions under which exit can function as a kind of empowerment that supports democratic goods. Exit is empowerment just to the extent that B can replace a relationship with A with alternative relationships. If A is not a monopoly supplier for goods B needs or wants, A cannot use control over these goods as a power relationship, nor can B merely suffer neglect owing to the lack of alternatives. It is also easy to identify the work that constructions of exit, discussed earlier, accomplish: Rights, protections, status, development of alternative suppliers (political parties, schools, health clinics, etc.) alter relationships of power at the individual level, away from asymmetrical relations of dependence or neglect and toward symmetrical relations of choice.

Framed in these terms, it is likewise easy to identify the contributions of exit to basic democratic norms. With respect to the good of nondomination, the primary meaning of exit-based empowerment here references individual choices, the effective existence of which blunts potential relationships of domination, and contributes to self-determination. Exit reduces vulnerability by increasing choice, and the possibility of exit helps to certify that relationships of mutual dependency—which are constitutive of social life as such—are not those of domination or neglect.

The work accomplished by exit with respect to the all-affected principle is also clear. In the limiting case of monopoly-based domination, the resource holder dictates the terms of dependency. In contrast, if exit is constructed in such a way that it induces a collectivity (or its elites, representatives, or decision-makers) to respond to its members, exit-based empowerment expands self-determination into collectivities, supporting the all-affected principle.

Now consider exit from the perspective of communication. If, as I have been arguing, exit contributes to democracy primarily through empowerment, then we should not be measuring exit on the register of voice, as do democratic theorists when they view exit as a silent act of private choice in contrast to the public act of voice. Exit carries communicative content just because every act, including an act of exit, leaves communicative traces. Or, stated otherwise, the capacity to exit is a form of power; but it is also a form of communication that indicates satisfaction or dissatisfaction, as Hirschman (Reference Hirschman1970, chap. 9) recognized.

Although communication can assume a wide variety of forms, here I identify voice-based communications with acts with linguistic content (expressions, assertions, arguments, etc.) or acts that can translate into linguistic content (protests, demonstrations, etc.), while using the term signal to identify the nonlinguistic communications that are consequences of exit from organizations or other kinds of social or political relationships. Although the act of exit may combine with voice, the act in itself constitutes a mode of communication—a signal—even when not combined with voice. Signals register in prices, membership numbers, and followings counted in votes.

No doubt these features of exit-based signal explain why exit is so often viewed as the (silent) opposite of voice. It is true that, in contrast to voice, exit fails to convey propositional content—and, indeed, does not (immediately) register in the medium of language. Strictly speaking, exit-based signals are residuals of acts of choosing. Whereas joining is positive and specific—a choice made to enter in this relationship rather than unspecified others—exit primarily signals dissatisfaction without the additional information that positive acts of choice convey through their specificity. But these features of exit should not lead us to overlook its residual (but potentially effective) communicative content. Contemporary philosophy of language emphasizes that in saying something we are also always doing something—typically, establishing a social relationship with an interlocutor, setting expectations for others and self, or committing oneself to the implications of what one says (Brandom Reference Brandom2000). The reverse is also true: In doing something, we are often saying something. But what is “said” through the act of exit is underspecified with respect to meaning unless (a) exit is matched with a positive choice, though which acting signals a specific intent (economists view such choices as communicative insofar as they “reveal preferences” (Tiebout Reference Tiebout1956, 410–17)), or (b) exit is matched with a content designed to specify the communicative content—focus groups, questionnaires, anonymous complaint boxes, exit interviews, and other mechanisms that seek to transform the implicit communicative content of a signal into explicit propositional content. For this reason, entities that are subject to exit—losing candidates and parties, failing businesses, languishing associations, undersubscribed service providers, jilted lovers—can respond only if they supplement signals with voice of the kinds that conveys specific, propositional content. Indeed, this was Hirschman's (Reference Hirschman1970) key recommendation: Healthy, innovative, responsive organizations anticipate exit by providing for voice (chaps. 3, 9).

These considerations further specify a key feature of exit relevant to democracy and mentioned above: its lower threshold for use relative to voice. If exit is constructed into a system or institution such that it can be exercised for relatively low cost, it also follows that communication through signal will have low effective thresholds of use relative to voice. Two potentially democratic implications follow: First, a lower threshold of use means that exit-based signals are less endowment-sensitive—and thus, in principle, may be available to individuals even when they lack the resources for voice. Second, to the extent that exit is low-cost, dissatisfaction may be signaled long before it is expressed, enabling organizations to respond to dissatisfactions and seek to cultivate voice as a proactive response. Because the communicative content of exit is not symmetrical with positive choices—to join, to buy, to enter a conversation, and so on—if it is institutionalized, it shifts the communicative burden to the entities that suffer loss. Both features of exit as signal underwrite the all-affected principle.

MAPPING THE DEMOCRATIC POSSIBILITIES OF EXIT

My analysis so far has been conceptual, aimed at developing the concept of exit in order to identify its place within democratic theory, namely, as an empowerment with a residual communicative content. One part of the analysis has been normative: I have been concerned with identifying the democratic norms—nondomination and the all-affected principle—that exit-based empowerments might advance. Another part has aimed at bridging normative effects and institutional design. As I have suggested, democratic institutions carry off two important feats: They empower people who are affected by collective decisions, and they communicate the preferences, values, and ideas that figure into decisions. I have been concerned with clarifying the circumstances under which exit might empower and communicate in ways that advance democratic goods. For these kinds of effects to follow, I have argued, exit must be constructed, either by being enabled or by being institutionalized. Although I have provided some examples, we can now put these distinctions to work more systematically to locate the places in which exit-based empowerments might be designed into the political systems of strong, constitutional states.Footnote ⁷

If we treat these elements of democracy as dimensions on which the design of institutions can vary, we can map the domain of institutional design, as indicated in Table 1. In the vertical dimension, representing forms of empowerment, we can distinguish three kinds of design, two of which are exit-based: enabled exit and institutionalized exit. Collective inclusion identifies positive empowerments to direct organizations with low exit (monopolies): typically actionable rights of organization and expression.

TABLE 1.

Mapping Exit-based Empowerments within Democratic Systems

I stylize the (horizontal) dimension of communication as varying from signal to voice and deliberation. Signals, as suggested above, are the nonlinguistic communicative consequences of individual decisions to join or leave organizations or other kinds of social or political relationships, indicated, for example, in votes, prices, membership numbers, and the like. The category labeled voice and deliberation includes communications with specifiable content, often (although not necessarily) in linguistic form. Although for other purposes we might make finer-grained distinctions, the key point here is to highlight the communicative residuals of exit.

Combining these dimensions in Table 1 represents the key theoretical argument spatially: Voice and exit are far from exhaustive of the possible workings of democracy (cells 2, 3, and 4, shaded in dark gray). Notwithstanding the brilliance of Hirschman's analysis, the exit–voice dichotomy (cells 1 and 5, shaded in light gray) hides as much as it reveals about the workings of democracy: It maps too easily back onto the distinction between exit-based markets (cell: 1 enabled exit–signal) and inclusion-based politics (cell 5: monopoly–voice), thus obscuring the roles of exit-based empowerments, actual and potential, within democratic systems. The broad point of the table is that it identifies democratic possibilities beyond the voice–monopoly model, which—as the example of electoral democracy suggests—was never adequate in any case. Three of these possibilities—represented by cells 2, 3, and 4—bear closer examination. (The empty cell represents an impossible combination: Where there is monopoly and no exit, there can be no residual communicative signal.)