Article body

Hopeful Losers?

Existing liberal democracies encourage citizen participation in politics, and protect freedoms to pursue myriad (reasonable) ways of life. Yet many of these regimes also fill elected offices and decide important policy questions using electoral and legislative systems at various geographic and civic scales that are less inclusive than alternative arrangements, ceteris paribus. The arrangements in question are less inclusive in two distinct but related senses: first, some voters are more likely than others to influence outcomes; and, second, institutions are more responsive to some values and interests than others. We expect some citizens to remain persistently hopeful about democracy, even as they lose out time and again on matters deeply important to them and when no principled reasons are offered for this asymmetry, aside from pragmatic variations on the platitude “You can’t always get what you want.”

Suppose we can agree on a reasonable definition of inclusion (and correlative standards of influence and responsiveness). Is this a problem?

Much of political theory and philosophy would not be especially troubled here, worrying instead about whether the basic structure of institutions is just, or at least approximately so. Typically, the institutions under consideration are at the level of the sovereign territorial state, but sometimes, at a broader level. We should expect a range of political arrangements to be roughly consistent with justice, and thus political philosophy and theory (even allegedly nonideal theorizing) don’t have much to say about the vagaries of electoral mechanics and legislative structures. Fairness doesn’t mandate that all people always get what they each want; it may not even mandate strictly equal likelihood of influence over outcomes, and it certainly doesn’t require that institutions ought to be uniformly responsive to all values and interests—even supposing we could agree on what ‘equal responsiveness’ might mean.

Against this tendency, I contend that political theorists and philosophers should be arguing for particular institutional forms with particular geographies, consistent with justice. Some institutions for collective decision making and some spatial arrangements of those institutions are more inclusive than others, and this matters: there is an ideal-theoretic and practical moral case for mixed electoral systems in particular configurations over time and space.

The Thesis, Elaborated

You can’t always get what you want, but some people inevitably get more of what they want than others.

It would be a near-trivial exercise to cast influential debates in ethics and political philosophy in light of this truism. We have a variety of beautifully refined arguments about which principles ought to determine when, and the extent to which, people get how much of what they want. Some obvious candidates are the following: need, merit, preference, allocative efficiency, luck sensitivity, non-envy, and neutrality of aim. You cannot always get everything you want, but philosophers and theorists have abundant resources at their disposal for crafting plausible arguments about how to regulate inevitable inequalities in who gets how much of what.

That said, the platitude has, I believe, a more troubling resonance in the case of democratic theory, where what we want is, among other things, government that is fair, in both how citizens influence outcomes and how responsive institutions are to our sincere and reasonable preferences.[1] Here, the connection between careful moral reasoning about a fair degree of influence and sufficiently responsive institutions, on the one hand, and the play of actual politics, on the other, is often exceedingly unclear, even presuming we can agree on precisely what we mean by influence and responsiveness.

In a free society under the rule of law, people are at liberty to pursue what they want, as long as they do not unduly interfere with their fellow citizens doing the same. In democratic politics, however, those same free citizens come together as moral equals to rule themselves, inevitably through some kind of electoral procedure and legislative arrangements. In that complex institutional setting, some citizens inevitably have more influence over outcomes than others—at the very least, in the sense that their votes are more likely to correspond with the winning outcome, but also in the closely related sense that their votes are more likely to decide a given outcome than others. On a range of issues, then, these voters are more likely to be on the side of the majority. Some citizens, furthermore, but in a similar vein, enjoy more responsive institutions—in at least the intuitive sense that laws and policies emerging from and enforced by those institutions tend to converge with their desires and expectations. Even when they disagree with a particular outcome, they are more likely to have access to means of redress, and to have available ways of mitigating negative consequences of their minority status with respect to the effective worth of their liberties. These differences in influence and responsiveness, thus understood, are obviously not due to merit, need, or some principled and plausibly argued claim. Rather, the inequality at issue is largely a function of demographic contingencies, paired with institutional realities that are exceedingly difficult to modify once entrenched.

The worry here is not the straightforward complaint that, over some range of important issues, majorities will persistently vote their shared preferences against losing minorities, although this obviously happens, and is sometimes morally troubling. Still, we have quite sophisticated theoretical arguments and institutional correctives to avoid the moral quagmire of brute majoritarianism. Indeed, some would argue that the very marriage of liberalism and democracy (that we now take for granted in most constitutional republics and federations) is precisely a history of reining in the majoritarian excesses of democracy. This line of argument seems to me to be, on the whole, rather convincing.

Yet suppose we have carefully crafted institutions, informed by liberal-egalitarian principles (of equal moral standing, extensive liberties, and neutrality of aim), within which exercises of power are justified by public reasons. We would have, in such happy circumstances, a well-ordered, political-liberal republic of reasons. Even under these ideal conditions, however, Arrow (1951), Gibbard (1973), Satterthwaite (1975), and many others have shown us that the very mathematical structure of collective decision making forces us to make difficult choices. Some electoral and legislative arrangements are more sensitive than others to a wider range of values and preferences, but that sensitivity may be purchased at the expense of greater vulnerability to forms of manipulation. Other arrangements may be robust against a range of likely vulnerabilities, yet tend toward longer debates and more complicated compromises, making resulting decisions unstable; in contrast, more stable arrangements may be less sensitive to diverse values and preferences.

To be sure, there has been fruitful debate about the practical relevance of these well-known properties of decision making arrangements—that is, their varied degrees of sensitivity and vulnerability to manipulation.[2] That debate, however, tends to confirm my motivating concern here: when political theorists and philosophers turn to the question of institutional choice, especially electoral and legislative procedures, they often seem to throw up their hands, morally speaking, and leave these concerns as, for the most part, mere technical issues of implementation.[3]

John Rawls is exemplary here, although hardly unique. After dismissing any thought that “what the majority wills is right” (Rawls, 1999, p. 313), he accepts “that a variant of majority rule suitably circumscribed is a practical necessity” (Rawls, 1999, p. 311). We accept the potential moral vagaries of majority rule, Rawls thinks, only on condition that basic liberties are preserved, and that “in the long run the burden of injustice should be more or less evenly distributed over different groups in society, and the hardship of unjust policies should not weigh too heavily in any particular case” (Rawls, 1999, p. 312).

The view of institutions here is either quietist or concessive: our political arrangements are inevitably flawed vehicles for realizing a shared political conception of justice, and so all that can really be said, as a matter of philosophy, is that democratic authority is invoked so as “to share equitably in the inevitable imperfections of a constitutional system,” in light of which we ought “not to invoke the faults of social arrangements as a too ready excuse for not complying with them, nor to exploit inevitable loopholes in the rules to advance our interests” (Rawls, 1999, p. 312). Justice can help us here, perhaps, if it can show us a way to fix these imperfections, close those loopholes; or, if this is not (yet) possible, then justice tells us how to distribute the burdens of injustice more fairly.

My aim here is to challenge the concessive view. Choice of institutional arrangements, and especially electoral procedures, is not a mere technicality to be muddled through once the (ideal) moral and constitutional architectures have been settled. Nor, however, is it a problem requiring a separate body of non-ideal theory to address some yawning chasm between our philosophical ideals, on the one hand, and our decidedly messy practices, on the other.

Rather, I mean to show that ideal theorizing about such things as justice and legitimacy has clear consequences for institutional design, in ways that reflect empirical properties of those institutions: how they are designed, how they tend to work, and how they can reasonably be expected to work under favourable and less-than-favourable conditions. Furthermore, if you sing your political liberalism in a roughly Rawlsian key, you should favour more inclusive mixed electoral and legislative arrangements over less inclusive options, other things being roughly equal. Just what those “other things” are I will now explain.

A Reasonable Expectation of Influence: Implications

This last claim may seem strange: why would a political liberal, especially a Rawlsian political liberal, strongly endorse particular institutions, rather than allowing the play of (reasonable) argument to determine the institutional specifics of a just basic structure?

Indeed, for Rawls, the principles we would agree to under the constraints of an original position are informed by “the strains of commitment” and “the burdens of judgement.” Consideration of the former (the “strains”) gives us good reason to think that rational parties “will not enter into agreements they know they cannot keep or can do so only with great difficulty” (Rawls, 1999, p. 126). The burdens we find ourselves bearing in a just regime will be consistent with our basic needs and fundamental interests, including our equal moral standing as citizens, secure in our self-respect; otherwise, we would not accept them. Consideration of the latter (the “burdens”) gives us reason to think that citizens who affirm just institutions will tolerate fellow citizens whom they think mistaken on moral and philosophical matters, because we accept that, on these deep issues, evidence is complex and uncertain, and how we interpret and weigh the significance of evidence is shaped by our interests, values, and experiences (Rawls, 2005, pp. 54-57). We would accept these “burdens of judgement” as the cause of persistent and legitimate disagreement whether or not we find ourselves in electoral majorities vis-à-vis the moral and philosophical questions that generate those disagreements. A just regime cannot and should not satisfy everyone, all the time. Such is life within sufficiently just institutions consistent with the strains of commitment and the burdens of judgement.

On the Rawlsian conception of stability, then, what matters is an overlapping consensus among reasonable comprehensive views on a political conception of justice, the resulting institutions of which are likely, over time, to foster reasonable citizens committed to that political conception, and to just social arrangements that flow from it. There is no legitimate expectation, on this view, that our comprehensive doctrines will ever be decisive in public decisions. What matters to legitimacy and stability is the possibility of finding the resources for affirming shared political values and associated institutions within our respective comprehensive doctrines.

Nonetheless, there is an asymmetry here that may warrant concern. Some comprehensive doctrines—most likely a cluster of related doctrines closely convergent around liberal conceptions of justice—will indeed find that significant elements of their comprehensive doctrines do prevail in many political decisions. Outside of this privileged cluster, citizens will enjoy no such expectation. So, while an overlapping consensus on a political conception of justice may hold, and sustain the legitimacy of associated institutions, it is not the case that the resulting authoritative institutions are uniformly sensitive to all of the comprehensive doctrines party to that overlapping consensus. Some comprehensive doctrines will do better than others, politically speaking.

It seems a strange sort of loyalty that political liberals ask of reasonable minority groups—over repeated decisions, and perhaps over several generations—to affirm just but significantly unresponsive electoral and legislative institutions. Certainly these institutions, by virtue of being just (and justifiable within the public reason of a plural society) do satisfy the strains of commitment as Rawls presents them: they secure the conditions necessary for us to live meaningful, satisfying lives together, whatever our ultimate (but ultimately reasonable) values and aspirations may be. Yet, for some, this is all that is on offer: they will never further enjoy the privileges of power that come from being a stable electoral majority.

The (not unreasonable) political-liberal response is to note that, if the burdens of judgement are widely accepted by citizens of a just regime, then, while some reasonable minority views may not ever be decisive in politics, the majority will certainly tolerate those views. They will, furthermore, take care to respect and even accommodate them, so far as possible in public affairs.

In general, reasonable citizens will treat one another as free moral equals who share a political conception of justice, but the strains of commitment ensure that citizens will not accept any such conception that does not ensure extensive personal liberties, the social bases of self-respect, and sufficient means to pursue reasonable conceptions of the good. Rawls’s principles were designed with these expectations in mind, so if they can be approximated in our basic structure of fundamental institutions, then that is sufficient.

So, understanding responsiveness as a specific kind of influence—that is, in terms of the likelihood that our reasonable values and preferences will be decisive in outcomes—is too strong a condition, and ignores the ways in which liberal values constrain democratic politics just because some reasonable views will never be decisive in this way. Still, if “responsiveness as likelihood of decisiveness” is too strong, why wouldn’t some expectation of more diffuse but still significant influence in the apparatus of political life be part of the strains of commitment?

Consider this possibility in terms of Rawls’s famous argument from the original position: parties who anticipate a persistent lack of influence may argue that they will not accept principles without some guarantees that public institutions will allow them some fair expectation of influence over matters important to them in public life. If there is no such implied expectation, then it seems odd for Rawls to follow Mill in valuing the “education in public spirit” and “affirmative sense of political duty and obligation” that follow from participation in public life as free and equal citizens (Rawls, 1999, p. 206). After all, if we cannot reasonably expect our sincere and cogent judgements to be influential on at least some important matters, some of the time, then why would self-respect correlate with the “education” and “affirmative sense” to which Rawls refers?

If self-respect is in part implicated with the public activities fostered by a just basic structure, as Rawls hopes, then those activities must surely be for something: we argue to persuade, and through persuasion we hope to influence outcomes.[4] It seems perverse to encourage argument that is unlikely to ever have such influence. The best we might hope for—and this is not much, morally speaking—is that self-respecting citizens at least be reasonably civil in treating one another “as rivals, or else as obstacles to one another’s ends” (Rawls, 1999, p. 206).

Certainly reasonable citizens cannot expect to be routinely influential in the sense of their votes being decisive, or to wield strictly equal influence on all, or even many, issues. But we might reasonably expect more than some kind of diffuse, indirect influence in the sense of having had our say in the cacophonous public sphere. Somewhere between decisive and diffuse influence is an expectation of our reasonable views having a serious hearing in the formal public sphere of legislatures, assemblies, and political campaigns. The appeal to public argument of a certain character—which is an abiding theme throughout Rawls’s later formulations of his theory (2001), and is central to so much subsequent work by theorists of political liberalism—entails an expectation of possible influence of just this sort. Persistent losses in elections, legislatures, and referenda in spite of reasonable values and plausible arguments will challenge any such expectation.

If an expectation of influence in the formal public sphere is persistently denied, then the connection between legitimacy and participation in public argument will likely come to be seen as a merely formal guarantee of voice; real influence, however, must be sought elsewhere, and by other means. Suppose that one or a few issues matter deeply to those who find themselves effectively excluded from political influence: under such conditions, and in particular for profoundly important moral issues, reasonable citizens, thwarted over decades and generations, may be pushed toward increasingly antagonistic, even unreasonable challenges to the institutional status quo.

Furthermore, some reasonable citizens—perhaps those with more diffuse but longstanding complaints of insufficient influence, such as minority linguistic or religious communities—may have plausible grounds for thinking that they could achieve comparable benefits and gain effective influence under their own sovereign institutions. Exit may be perceived as a Pareto-improving strategy for some citizens without sufficient influence, if an alternative sovereign jurisdiction promises to secure the same basic rights, and provide the same (or comparable) public goods at similar costs. This problem cannot be resolved simply by rejecting such strategic reasoning as unreasonable. After all, some demands for recognition may be valid public claims in a liberal democracy, and such recognition may well require the institutions for effective self-government according to shared reasonable political values (cf. Patten, 2002). If so, then the basis of loyalty may be further undercut insofar as persistent minority status at some institutional scale seems inimical to such recognition.

Still, why give much weight at all to this expectation, however reasonable? Influence over legislative debates and party platforms is all well and good, and it is a reasonable thing to expect, other things being equal; but other things are often not equal, and if a just regime provides sufficient space and resources to pursue our projects together with those who believe and desire as we do, then where is the pressing problem? True, persistent minority status in electorates might motivate some of these reasonable citizens to retreat from mainstream politics, but why suppose that their retreat will generate perverse motivations and eventual failure of reasonableness? Why assume that the expectation of influence, however reasonable, is so pressing a concern for these reasonable minority groups?

There is a clear prudential reason for worrying about the retreat from public life: we have ample and growing evidence from political science and social psychology to show that insular groups can easily be driven to more extreme positions than their members may have originally held (e.g., Sunstein, 2002). The more we remove ourselves from public life, the more likely we are to find ourselves in smaller groups of those who think and believe and act as we do. Furthermore, some emerging evidence suggests that the retreat from public life correlates with a range of attitudes and behaviours that do not seem especially favourable to reasonableness if they are allowed to persist. In their analysis of community-level data from across the United States, Robert Putnam and his colleagues find tentative but troubling evidence that living in more diverse communities is associated with citizens participating less in civic associations and traditional political activities, but also having fewer friends, reading less, watching more television, and caring less about politics (Putnam, 2007). Although other studies and especially other countries reveal interesting complexity, similar results obtain (e.g., Stolle, Soroka, and Johnson, 2008; Stolle and Harrel, 2012).

Putnam himself is optimistic that this is a short-term tension, and that American democracy in particular has both the resources and the historical tendency to foster engagement and trust across diversity. But how? Aggressive assimilation, or the imposition of a strong unifying identity, are not in the cards for political liberals, who agree with Rawls that “the extent to which we make engaging in political life part of our complete good is up to us as individuals to decide, and reasonably varies from person to person” (Rawls, 2001, p. 144). Assimilation tactics are simply not part of the calculus of social planning for political liberals, nor for multicultural liberals. Furthermore, the political liberal also agrees with Rawls that justice as fairness is consistent with a classical republican concern that continued liberty requires an engaged citizenry: “If we are to remain free and equal citizens, we cannot afford a general retreat into private life” (Rawls, 2001, p. 144).

Institutional Design and Dominated Options

In addition to serious prudential concerns about the dynamics of insular groups and disengaged citizens, there is a moral-philosophical case for taking seriously these patterns of exclusion. Whether or not you think that eventual instability is a serious problem, and whatever weight you are inclined to give to a reasonable expectation of political influence, so long as the weight assigned is positive, it should lead you to favour institutional arrangements that are more rather than less responsive to reasonable claims, other things being equal.

Quite specific constraints on institutions are arguably woven into the structure of Rawls’s theory—indeed Rawls thought as much: “Certain institutional forms are embedded within the conception of justice” (1999, p. 231). This follows from the structure of the choice situation central to Rawls’s justificatory logic: behind a veil of ignorance, the possibility of being in a sensitive linguistic minority, say (e.g., Kymlicka, 2001), or in a persistent electoral minority of the kind at issue here, may be serious enough to warrant parties behind the veil to favour institutional realizations of their favoured principles of justice that are more, rather than less, responsive to a diversity of reasonable moral claims and ways of life. I want to draw out two features of this line of argument.

First, this moral decision rule—rejecting dominated institutional options—is itself affirmed by Rawls early on in Theory, where he insists that “other things being equal, one conception of justice is preferable to another when its broader consequences are more desirable” (Rawls, 1999, p. 6). We can simply apply the same rule to collections of institutions consistent with a conception of justice: some configurations dominate others in this sense, other things being roughly equal with respect to basic justice.

Second, the institutions that are most likely to determine whether or not particular basic (and supervening) structures are dominated are electoral; the salient feature of those systems is their relative informational richness with respect to the range of values, preferences, and interests in the electorate. It is well known in the literatures on voting and elections that some voting systems tend to be more informationally rich than others. Specifically, ‘first past the post’ plurality systems tend not to be especially rich in this way, encouraging voters to vote strategically and candidates to compete for the center, making it difficult for third parties to find purchase in electoral campaigns.

To be sure, informational richness may come at the cost of stability in legislatures: pure proportional and mixed voting rules tend to make coalitions necessary to form governments, meaning that strong stable majority governments may be less common. I offer the following reason to think this isn’t a major concern: the kind of instability associated with unstable voting systems may in fact foster a kind of loyalty among diverse and often disgruntled citizens, although perhaps not the kind of loyalty we should praise with too much enthusiasm unless it is tempered by other considerations—such as the independent moral-philosophical attractiveness of more responsive institutional arrangements.

Students of elections since Borda, Condorcet, Dodgson, and Black have worried that, when several voters attempt to choose among more than two options by voting according to their preferences, a cycle may result such that no clear outcome is favored by the majority; more precisely, the collective ordering is intransitive (e.g., Black, 1958; Riker, 1982). This problem cannot be avoided without violating at least one of several intuitively plausible assumptions about fairness and rational consistency (Arrow, 1951). This makes outcomes sensitive to the order of pairwise contests among options, leaving majoritarian procedures vulnerable to manipulation (Riker, 1980) avoidance of which involves either constraints on admissible preference orderings, or on the agenda-setting process, or both (Dodgson, 1887; Black, 1948; Plott, 1967; Brams and Fishburn, 1978; Shepsle, 1979; McKelvey and Schofield, 1987; Saari and Tataru, 1999). But in two intriguing essays, Nicholas Miller (1983; 1996) argues that we should favour majoritarian procedures just because they are unstable in the way social choice theory suggests. For my purposes here, the relevant theme in Miller’s work is the idea that, when majoritarian procedures fail to identify a clear social preference, there is sufficient uncertainty to motivate participants to continue playing subsequent rounds of the political game: future wins and losses are not sufficiently predictable to inspire unreasonable resistance or rational exit. Loyalty is ensured by instability, because that is evidence that the playing field is fair, that teams are evenly matched, and that continued play is thus worthwhile.

I have argued that we have independent philosophical and prudential reasons to reject simple majority rule systems because they are dominated by more informationally rich options. Still, if the resulting instability of coalition-dominated governments in proportional and mixed electoral systems is offered up as a mitigating concern, then we should note that the same instability might well attend to majoritarian systems, but over longer time scales.

To explain: the apparent stability of informationally sparse majoritarian arrangements (a two-party Westminster system, with single-member districts and first-past-the-post plurality voting, is the obvious example) is rooted in the greater likelihood of clear majority governments being able to pursue their mandates without opposition support. Yet consider the matter across the electoral cycle, as the mandates of previous governments are painstakingly (and at considerable expense) reversed by subsequent governments.

Now, if Miller’s instability argument persuades, then why should it matter if the instability involves alternating majority mandates or unstable coalition governments? The instability argument doesn’t distinguish (at least not obviously) between informationally sparse or rich options: if the argument convinces you for Westminster-style two-party systems, then it ought to convince you for coalition governments under pure proportional or mixed-member proportional systems.

To be sure, the resulting loyalty to either sort of unstable system may seem less like that of reasonable citizens affirming a shared conception of justice within public reason, and rather more like the embittered loyalty of a chronic gambler to the roulette wheel. Indeed, as Miller himself notes, the structure of preferences is important to even the modest ‘loyalty through instability’ hope: given a deep cleavage along a single issue, majoritarian procedures are prone to generate a persistent minority rather than desirable instability. The volatility of majoritarian procedures depends on the absence of durable alliances that result in some citizens being persistently in the minority on a wide range of issues. The problem is that reasonable disagreements may pit a minority of citizens against the majority on a range of important moral issues over time.

This worry, however, dissolves when we turn to informationally rich proportional and mixed systems, which, however potentially unstable within legislatures, are unstable in the morally attractive sense that they allow more reasonable views to influence the political process.

Conclusion: Taking Responsiveness Seriously

I have presented a problem internal to multicultural and especially political liberalism, and in particular to Rawls’s definitive statement of the latter position: the political conception of justice as fairness. The problem is internal to Rawls’s project in the sense that his own account of reasonableness and the workings of just social arrangements together seem to permit a regime that encourages “propensities and aspirations that it is bound to repress and disappoint” (1999, p. 474), something Rawls thinks a just regime ought not to do.

I have offered reasons to think that this standard might be an apt criterion of legitimacy for a just regime. Other things being equal, we ought to favour that just regime that is more inclusive of, and more responsive to, the full range of reasonable values and claims in a free and fair society. Less responsive regimes are relatively unresponsive by virtue of electoral procedures and rules governing political campaigns. These regimes are dominated by those that are more responsive.

Furthermore, if you concede that, while not decisive, the first two concerns together are at least plausible—and I have suggested a prudential reason, rooted in extant social-scientific evidence concerning citizen engagement in plural societies, for thinking they might be—then the conclusion that we should avoid dominated just regimes is that much more compelling.

Finally, a lesson that follows from these arguments is that the often seemingly mundane institutional trappings of a constitutional liberal democracy—from electoral mechanics to campaign finance rules, zoning ordinances, local school funding issues, and curriculum design—are properly considerations of justice. It may be the case that there are a great many basic and supervening institutional structures consistent with a given political conception of justice that meet the test of an overlapping consensus. But a reasonable expectation of influence, and worries about the sort of ethos that will follow from generations lived under less responsive institutions, seem to constrain the choice among those institutional structures, favouring more responsive just regimes over less sensitive alternative arrangements.