Democratic Stability and Backsliding

A. Two Types of Stability

Stability can be provided by a modus vivendi, which exists when the various sides to a conflict understand that further pursuing the conflict will be more costly than any expected gains—perhaps because the sides have similar levels of power. Under such circumstances, it is “wise and prudent” for the sides to avoid open conflict, but they remain “ready to pursue their goals at the expense of the other.”⁸ Rawls identifies two shortcomings of modus vivendi stability. The first is practical: it is not robust to shifts in the underlying distribution of power. The second is normative: in a society stabilized by a modus vivendi, citizens are effectively forced to live under institutions that reflect the balance of power in society, rather than institutions and policies that reflect their values.

Rawls hopes to show that a democratic society could be stable in a more robust and normatively attractive sense, which he refers to as “stability for the right reasons.” Stability for the right reasons exists when it is common knowledge that (a) citizens share a conception of justice and (b) view that conception as a regulative constraint on their behavior. Stability for the right reasons avoids the two weaknesses of modus vivendi stability. From a practical perspective, it is superior because citizens “will not withdraw their support of it should the relative strength of their view in society increase and eventually become dominant.”⁹ From a normative perspective, decisions are not merely “imposed on citizens by brute force”;¹⁰ instead, “the reasons from which citizens act include those given by the account of justice they affirm.”¹¹ In this regard, stability for the right reasons makes possible a form of political autonomy that is absent under a modus vivendi.

B. Reciprocity and Stability

Having distinguished between these two types of stability, we can now see the sense in which Rawls’s account of stability turns on the importance of reciprocity. The difference principle represents a commitment not to “take advantage of contingencies of native endowment, or of initial social position, or of good or bad luck over the course of life, except in ways that benefit everyone, including the least favored.”¹² Indeed, Rawls argues that the difference principle is the “only” arrangement, “that meets the following reciprocity condition: those who are better off at any point are not better off to the detriment of those who are worse off at that point.”¹³ He refers to this as the “deeper idea of reciprocity implicit in” the difference principle.¹⁴ How, though, does the difference principle’s reciprocity-based appeal underpin Rawls’s account of stability?

To address this question, one needs to consider the sources of instability under the difference principle, as compared to the sources of instability under alternatives, such as restricted utility. Rawls notes that, unlike the difference principle, restricted utility “is a maximizing aggregative principle with no inherent tendency toward either equality or reciprocity.”¹⁵ He then argues that because of this lack of a tendency toward reciprocity, there are reasons to worry about whether, under restricted utility, the least advantaged will accept the basic structure. After all, they are asked to have less in order to benefit those who are already better-off for reasons that are arbitrary from a moral point of view (e.g., having been lucky to have been born with socially valued natural aptitude). Rawls says that this can reasonably be expected to lead the least-advantaged to, “grow distant from political society” and to become “withdrawn and cynical.”¹⁶ In short, restricted utility’s failure with respect to reciprocity poses a threat to stability.¹⁷

By contrast, in a society organized around the difference principle, it is the more advantaged who are “most likely to be discontent.”¹⁸ However, Rawls argues that the threat thereby posed to stability is less significant. Such citizens recognize that while they are worse off than they could be, their being so is a natural consequence of a commitment to reciprocity:

The more advantaged see themselves as already benefited by their fortunate place in the distribution of native endowments, say, and benefited further by a basic structure (affirmed by the less advantaged) that offers them the opportunity to better their situation, provided that they do so in ways that improve the situation of others.¹⁹

Thus, it is because the difference principle better satisfies the demands of reciprocity than salient alternatives that Rawls takes it to have a stability-based advantage over them.²⁰

Importantly, it is not sufficient for this stability argument that the basic structure is, in fact, consistent with a commitment to reciprocity. If the basic structure is organized in accordance with the difference principle, but the well-off think that that principle imposes an unfair burden on them (perhaps because they disregard the contingent nature of their advantages), then it will not be conducive to stability. Thus, although it is not often recognized, it is critically important from the perspective of the Rawlsian stability argument that citizens perceive the basic structure to be consistent with a commitment to reciprocity and respond to it accordingly.

This is important because if there is disagreement about justice, then it is virtually assured that there will also be disagreement over whether or not the basic structure is consistent with a commitment to reciprocity. (After all, many of those who deny that the difference principle is called for by justice are also likely to see it as inconsistent with the demands of reciprocity.) A shared conception of justice is, therefore, central to the Rawlsian account of stability because it allows for a shared perception that a society’s basic structure satisfies the demands of reciprocity. And, this shared perception is, on Rawls’s account, critical for stability.

C. A Dilemma for the Rawlsian Approach to Stability

This section will argue that the Rawlsian account faces an important dilemma that undermines its ability to explain democratic stability. If we interpret the overlapping consensus as requiring convergence on a particular conception of justice, it is – in light of the burdens of judgment – infeasible. Yet, if we interpret the consensus broadly enough to satisfy worries related to feasibility, then the reciprocity-based mechanism that the Rawlsian account of stability relies upon should no longer be expected to operate. Simply put: the mechanism underlying the Rawlsian account of stability is not compatible with due regard for the burdens of judgment.

Begin, then, with the narrower conception of the overlapping consensus, which requires convergence on a particular conception of justice. If everybody accepts justice as fairness, then considerations related to reciprocity can stabilize the society by explaining to both the least- and most-advantaged why they ought to comply. This possibility coheres with Rawls’s description of the advantage that justice as fairness has over alternatives with respect to stability.

However, such an account faces an important feasibility-based challenge. Over time, Rawls came to view the expectation – endorsed in Part III of A Theory of Justice – that all citizens may come to accept the same comprehensive doctrine as “unrealistic” and “utopian” because it conflicts with the burdens of judgment.²¹ These burdens include the difficulty of assessing and weighing conflicting and complex considerations, the inevitable vagueness of moral and political concepts, the way in which our personal experiences shape our perspective, and the wide variety of relevant considerations. Yet, these burdens also affect people’s reasoning about questions of justice. Thus, as has often been noted, just as it is unrealistic (even under favorable conditions) to expect convergence on a given comprehensive doctrine, so too it is unrealistic to expect convergence on a given conception of justice.²² The implication is that even if a shared conception of justice among citizens could help underwrite stability for the right reasons by leading all citizens to view the basic structure as satisfying the requirements of reciprocity, this would require a uniformity of views that is unrealizable in democratic political life.

In light of the challenges associated with achieving consensus on a particular conception of justice, it is tempting to interpret the consensus as merely requiring agreement on a family of conceptions of justice that share certain features. This could include the enumeration of certain rights, the assignment of some priority to those rights, and a commitment to ensuring that all citizens have adequate means to make use of such rights. Rawls embraces such a modification in his later work, asserting that it is “more realistic” to suppose that the focus of an overlapping consensus would be “a class of liberal conceptions.”²³ This approach may be buttressed by the observation that although citizens did not share a commitment to a particular conception of justice even during democracy’s most stable periods, those periods may seem to have been underwritten by a narrowing of the range of disagreement about justice.

This approach plausibly allows a society that is divided, for instance, between followers of Rawls and followers of Milton Friedman to count as having an overlapping consensus. Friedman can reasonably be interpreted as arguing for the strong protection of certain basic liberties and for a negative income tax that provides all citizens with a floor level of resources.²⁴ If, as an example, (nearly) everyone in the society accepted one of these conceptions of justice, then there would be agreement on some priority for the basic liberties and the need for citizens to have adequate means to make use of them, even while there remained considerable disagreement about the best interpretation of these commitments. Because positions that insist on protecting a floor level of resources for all, while embracing utilitarianism above that floor, should be expected to attract support in a free democratic society, a plausible account of democratic stability needs to be compatible with their presence. By recognizing this, the broader conception of the overlapping consensus may overcome the feasibility-based concerns faced by the narrower conception.

Unfortunately, the broader conception purchases this realism at the cost of undermining the power of the reciprocity-based mechanism on which the Rawlsian account relies. To see this, notice that Friedman’s position – which may win out in the democratic politics of such a society – suffers from the stability-based liabilities that Rawls associates with restricted utility and its failure to satisfy the demands of reciprocity. Under such a system, Rawls says, the least advantaged “cannot affirm the principles of justice” in their “thought and conduct over a complete life.”²⁵ This is because, as noted above, they see their fellow citizens supporting, and benefiting from, a basic structure that gives the less advantaged lower life expectations than they would have had under available alternatives, and does so for reasons that ought to be considered arbitrary from a moral point of view (e.g., their being born to a less advantaged position). As we have seen, Rawls argues that a society that organizes itself around restricted utility therefore risks the least advantaged growing “distant from political society” and becoming “withdrawn and cynical.”²⁶ The concern is that restricted utility attempts to maximize social welfare without taking seriously enough the cost that some are asked to bear for the sake of others and, as a result, risks alienating those citizens. The critical point is that interpreting the consensus broadly enough to escape concerns of feasibility comes at the cost of undermining the reciprocity-based mechanism for stability that Rawls’s account depends upon.

Here, then, is the dilemma—either Rawlsian accounts require too much uniformity of belief to plausibly arise in a large democratic society or they admit positions that are inconsistent with the reciprocity-based mechanism for stability upon which they rely. Might there be a “goldilocks solution” to this dilemma—that is, a consensus that allowed for the reasonable disagreement about justice that is sure to arise in a democratic society, even while excluding views that raise concerns of instability by failing to satisfy the demands of reciprocity? The discussion of restricted utility should make one skeptical. Bracketing its ultimate appeal, restricted utility seems like a relatively obvious candidate conception of justice—the kind of possibility that is likely to arise and attract support in any democratic society that allows for free expression and vibrant political disagreement. If, as Rawls argues, even that possibility is inconsistent with the reciprocity-based mechanism of stability, then it seems doubtful that one can rely on that mechanism to stabilize a democratic society while also being realistic about the extent of disagreement that is likely to arise.²⁷

The broad implication of the discussion in this section is that the Rawlsian explanation is not well-positioned to help us to understand what could facilitate stability in real democratic societies. Yet, if the Rawlsian account is not well-positioned to help us to understand what makes democratic stability possible, it is also unlikely to help us understand, or think productively about, how to respond to democratic breakdown.

A. Elections

I begin by explaining the crucial role that elections can play in rendering a political system stable. In a sense, it is puzzling that incumbents hold elections and that losing candidates comply with the results. While some may do so because they are intrinsically committed to the fairness of electoral procedures, it is inevitable that others will lack a sufficiently strong normative commitment of that kind. If electoral systems are to determine who controls government, there must be reasons to hold elections and comply with the results that even these individuals find compelling.

Such individuals will compete electorally when they view their expected return from doing so as superior to the return that they will get from attempting to subvert the system.³⁶ But, why – from this perspective – would competing factions view electoral competition favorably?

Begin with the opposition. They can organize to subvert the existing system so that they do not need to accept its results. However, doing so is likely to be dangerous and to carry with it remote prospects for victory. Thus, if the government is willing to hold reasonably fair elections, the opposition will often find competing in them attractive.³⁷ These elections must be reasonably fair to attract opposition participation—not because the opposition is assumed to be independently wedded to norms of electoral fairness, but because the opportunity to participate in rigged elections cannot substitute for the appeal of regime subversion.

Why, though, would the government offer its opposition the opportunity to compete on reasonably fair terms? The simplest case is that the government is very confident that it will win elections, but – by holding them – it induces the opposition to compete by the rules of the game rather than work to subvert the regime extra-institutionally. The government may prefer this because – from its perspective – attempts at regime subversion carry some (even if low) probability of a cataclysmic outcome—perhaps involving the execution or expulsion of its leading officials. Furthermore, even when they fail, attempts to subvert the regime may be costly to the government in other ways—for instance, they may destabilize the economy, leading capital (human and financial) to flee.

Why, though, might the governing party hold elections and comply with the results even when it believes it is likely to lose? It may do so when (A) the costs associated with the opposition’s attempts to subvert the regime outweigh the expected benefits of canceling elections and (B) holding elections can be expected to forestall efforts at regime subversion. (A) is more likely to obtain when the governing party anticipates that (1) it will have a realistic prospect of competing for, and gaining access to, office in the foreseeable future³⁸ and (2) its important interests will be protected even if it surrenders office. In sum, canceling elections – even when one expects to lose – may not be particularly appealing if doing so: carries with it a significant risk of revolt and severe accompanying personal costs, promises to destabilize the regime in ways that will undermine one’s economic position,³⁹ and when one’s rights – including the right to compete to regain office – will anyway be protected.

Notice, furthermore, that when the government openly works to subvert an existing electoral system (e.g., by censoring journalists, stripping opposition supporters of the right to vote, engaging in electoral fraud, etc.), this provides evidence that it lacks the public support to win election outright. After all, the incumbent who expects to handily win a fair election has reason to avoid fraud and the questions that it will raise about the extent of their actual public support. Thus, engaging in such behavior provides a public signal that the incumbent lacks broad support, which can simultaneously raise doubts about their competence and help coordinate opposition on non-electoral efforts at removal.⁴⁰

Finally, an electoral system makes it more difficult for a leader who wishes to subvert the existing regime to get support from other elites within their coalition for doing so. In a typical non-electoral system, such elites will depend on the leader for their power. It will, as a result, be difficult for these elites to resist the leader. By contrast, in electoral systems, many other elites will depend for their power, at least directly, on the support of their own constituents.⁴¹ This gives them a freer hand to resist their own party leaders than they would have in a typical non-electoral system. Moreover, they will often have reason to resist such attempts because they have valuable positions, and in many cases, reasonable hope for competing for higher office, in the existing system. Were the existing regime effectively subverted, not only is there is no reason for confidence that they would be as well-placed within its replacement, but they would also need to worry that the leader would regard them – given their popularity – as potentially dangerous competitors, the kind who new autocrats often work to eliminate. Thus, the existence of elections makes it more difficult for party leaders to get elite support, even from members of their own coalition, for attempts at regime subversion.⁴²

This subsection explains why opposition and governing elites will often prefer to compete within an electoral system, as well as why – once such a system is in place – it becomes particularly difficult to subvert. For our purposes, the crucial point is that political elites will often have very strong self-interested reasons to comply with electoral institutions. These reasons may often be dispositive even in the absence of a moral consensus on: a particular comprehensive doctrine, a conception of justice, or even the intrinsic fairness of the electoral system itself. While an influential stream of political science literature⁴³ has emphasized the potentially self-enforcing nature of electoral systems, this remains – in the philosophical literature – an oft-overlooked virtue of democratic systems, one that competing regime types (e.g., lottocracy, epistocracy, etc.) cannot be presumed to share.

To be sure, however, while we can identify mechanisms associated with elections that facilitate regime stability, the presence of elections – as is evident from any number of nation-building exercises – is not sufficient for such stability. The section’s more modest claim is that elections can stabilize a regime by incentivizing key players to focus on competition within the existing rules of the game. The next section describes supplementary institutions that can, when properly designed, enhance the stabilizing virtues of electoral systems.

B. Enumerated Individual Rights and an Independent Judiciary

Elections will have difficulty securing participation if contestants view the consequences of defeat as unacceptable. Individual rights can help protect against this possibility by assuring those who lose elections that their most fundamental interests are nevertheless secure. However, since it cannot be assumed that everybody who gains office will willingly respect such rights, the commitment to protecting them must itself be made credible. Is there an explanation, then, of why a government would avoid violating such rights that does not depend on officeholders being intrinsically committed to protecting them?

Enumerated rights and an independent judiciary can make it costly for the government to violate individual rights. To see why, following Weingast,⁴⁴ imagine a setting in which the government benefits from retaining office and from violating citizens’ rights (think of this as successful rent-seeking). The citizens, who are divided into two factions, do best when their rights are not violated. They can, if they jointly resist the government, overthrow it—thereby avoiding the costs associated with rights violations and leaving the government with its worst outcome. Resistance, however, is costly (albeit less costly than suffering rights violations). Under these specifications, citizens face a coordination problem with two pure strategy equilibria. If one faction knows that the other faction will acquiesce, then their best response is to also acquiesce (unilateral resistance would be costly, but unsuccessful). In such a setting, the government’s best response is to violate rights. However, if either faction will reliably challenge a government that violates rights, then the other’s best response is to do so as well, since their resistance will lead to the government being overthrown. In such circumstances, the government’s best response is to respect rights.

From this perspective, an obstacle to realizing the equilibrium in which the government respects rights is disagreement between the factions over what constitutes a rights violation. To avoid this, citizens need a shared understanding of what constitutes a rights violation and whether such a violation has occurred. In the absence of a robust shared moral view, enumerated rights can serve as a focal point, helping citizens to coordinate on what they will “count” as a rights violation for the purpose of coordinating resistance.⁴⁵ Meanwhile, the judiciary can provide an independent signal, indicating whether or not the government has violated enumerated rights. Together, then, enumerated rights and an independent judiciary can help to create a shared understanding among citizens about whether the government has violated rights, which can help citizens overcome their coordination problem. Since the government can anticipate this effect and wishes to retain office, it is incentivized to avoid behavior that would bring forth such a ruling. This contributes to democratic stability directly because some such rights are necessary preconditions of democratic government and, indirectly, by lowering the stakes of electoral loss.

Having now sketched this institutional account of democratic stability, it is important to emphasize that there is no reason that a Rawlsian cannot accept it. After all, the Rawlsian is anyway committed to the importance of elections, enumerated rights, and an independent judiciary—albeit on different, and presumably independently sufficient, grounds. The point, then, is not that the institutional account requires one to take steps that the Rawlsian would reject, but – rather – that it draws attention to a set of mechanisms that are ignored by the Rawlsian explanation of stability, and that can be expected to be operative even in an environment marked by the diversity of conceptions of justice that one should expect in a democratic society. Unlike the Rawlsian account, then, the institutional alternative is well-positioned to help us to understand democratic stability and, because of that, to see how backsliding works and how its likelihood might be reduced.

A. The Emergence and Role of a Constitutional Consensus

In Section II, I argued that accounts of democratic stability rooted in the idea of a constitutional consensus are insufficient because they (1) lack an explanation of the mechanisms by which such a consensus could emerge and persist, and (2) fail to recognize that important mechanisms that drive backsliding can operate even in the presence of a constitutional consensus, which raises questions about the precise contribution that such a consensus can make to democratic stability. However, the institutional account allows us to address these lacunae and to identify the important role that a constitutional consensus can play.

What might cause a constitutional consensus to emerge and persist? It is plausible that, as many citizens experience the salutary effects of elections, enumerated rights, and an independent judiciary, they will come to value such institutions because they (1) provide a reasonably fair way of settling inevitable first-order disagreements, (2) provide some assurance that citizens’ most basic interests will be protected, and (3) facilitate peaceful and mutually beneficial social cooperation. For these reasons, citizens may value such institutions even when they do not get the first-order policy outcomes that they most prefer.⁵⁷ Thus, the institutional account may be able to help explain how a constitutional consensus can emerge and persist.⁵⁸

Is there an explanation of how a constitutional consensus can discourage backsliding that is consistent with recognizing its limited capacity to prevent stealth? When a constitutional consensus exists, particularly if citizens place significant weight on the importance of compliance with democratic institutions, leaders face a greater cost for eroding democratic institutions because the consensus forces them to worry that even those who otherwise support them on first-order policy grounds will join with the opposition if they attempt to erode democratic institutions. Thus, while it cannot guarantee that attempts at backsliding fail (citizens may still place greater weight on their first-order policy preferences), a constitutional consensus makes it more difficult for leaders to pursue packaging and uncertainty-based strategies. The two accounts are, therefore, complementary: the incentive-based stability on which the institutional account turns encourages, and can be importantly reinforced by, a constitutional consensus.

B. Institutions for Mitigating Backsliding

As I have noted, understanding what facilitates democratic stability, as well as what allows leaders to interfere with it, should help us think about how to mitigate the threat of backsliding.

The three types of backsliding strategies that I described above – stealth, packaging, and uncertainty – may seem, on first glance, relatively disparate. However, the institutional account reveals their shared underlying similarity: they all make it more difficult for citizens to respond in a coordinated fashion to government overreach. In other words, these strategies allow governments to bypass the mechanisms that, under favorable conditions, facilitate democratic stability. While evaluation of specific reform proposals will be context dependent, this account encourages reformers to identify measures that will help citizens to (1) recognize attempts at government overreach and then (2) respond to such attempts in a coordinated fashion.

Begin with helping citizens to recognize attempts at government overreach. Since a fundamental role for the judiciary is to make stealth unavailable to the government by reliably publicizing its actions, stealth strategies will be most effective when paired with attempts to undermine citizens’ faith in the judiciary. The government can do this in two related ways. First, the executive can politicize the judiciary, such that – although courts do provide an independent evaluation of the government’s behavior – citizens do not view them as doing so. Second, the executive can capture the judiciary by filling it with reliable supporters, such that its rulings are not, in fact, reliable indicators of whether rights have been violated. The account therefore emphasizes the importance of preventing the politicization and capture of the judiciary. It may be that the traditional means of doing so (e.g., long terms for judges) need to be supplemented by more inventive arrangements (e.g., allowing the judiciary itself to determine which judges will sit on high courts). Also important from this vantage point are other institutions (such as independent election monitors and non-partisan media sources) that can reliably monitor and convincingly publicize the government’s actions.

A second, less well-recognized, way to help facilitate citizens’ awareness of attempts at government outreach is to remove legal power over the design of electoral competition from the hands of contestants and their partisan allies.⁵⁹ Elected officials often have the power to design the terms of political competition. They can pass legislation that dictates: the rules of campaign finance, the rules for casting a countable ballot (e.g., should mail-in ballots be permitted? Is voter registration required?), the design of electoral districts, and so forth. When elected officials have a legal claim to design the rules of political competition, the question of whether their specific proposals count as eroding democratic institutions becomes a matter of partisan debate and legitimate disagreement. However, the design of political competition can be taken out of competitors hands, as when citizens vote directly on such matters or they are delegated to the judiciary or an independent body (such as a commission charged with designing electoral districts). Taking legal control over the relevant decisions out of the government’s hands makes efforts to manipulate such processes more transparent and less subject to partisan disagreement. It also reduces the anxiety associated with uncertainty by offering citizens some reassurance that their political adversaries will not, even upon taking office, be able to undermine democratic institutions—thus reducing their own willingness to, pre-emptively, support such attempts.⁶⁰

As important as it is to make attempts at government overreach publicly known, such efforts are insufficient—we have seen, after all, that efforts to erode democratic institutions can succeed even when citizens are aware that they are occurring (as in the packaging and uncertainty examples). In the face of these strategies, it is also important to find ways to increase the relative weight that citizens attach to democratic norms and institutions. This makes packaging strategies more difficult to execute and, if publicly known, undermines the prospects of the kind of preemptive sacrifice of democratic institutions associated with the uncertainty strategy. An obvious approach, which may increase the relative weight put on procedural considerations, is to limit polarization, so that there is less, in terms of substantive policy, at stake in elections. Unfortunately, given widespread empirical disagreement about the causes of polarization,⁶¹ it is difficult to say, with conviction, what may be useful here, though familiar institutional proposals – such as ranked choice voting, open primaries, and proportional electoral systems – may help, in large part, by advantaging more moderate candidates.

While the preceding paragraphs provide some direction in thinking about how to protect democracies against backsliding, I do not mean to suggest that there is a set of institutional recommendations that can, with certainty, forestall backsliding. To the contrary, given what is at stake from the perspective of ambitious officeholders, it will be an ongoing challenge to effectively publicize, and then motivate coordinated opposition in response to, actions that undermine democratic institutions. The more modest hope is that, because the account is rooted in a compelling explanation of what makes democratic stability possible, it can help us to better understand democratic backsliding and, then, to clearly identify the kinds of steps that can mitigate the risk of it.

A. Does the Institutional Account Implicitly Rely on an Overlapping Consensus?

In light of the article’s discussion, it is tempting to view the Rawlsian perspective as a kind of red herring—it focuses attention on a mechanism that cannot be expected to yield democratic stability under the relevant conditions and, in so doing, it diverts attention from more relevant considerations.

This way of glossing the point may, however, seem too strong, for even on the institutional account, one can identify (at least) two important ways in which a narrowing of disagreement about questions of justice can contribute to democratic stability. First, as we have seen, polarization is an important enabling condition for backsliding—it makes it easier to pursue each of the strategies described above. It could be, though, that having an overlapping consensus reduces polarization and, thereby, undercuts the capacity of executives to pursue such strategies.⁶² Second, if there is a complete absence of agreement on what constitutes important rights, this will make it difficult to enumerate rights in a way that will enlist the shared response to government overreach that the institutional account requires.⁶³ Here, then, are two ways that a kind of overlapping consensus could, even on the institutional account, contribute to democratic stability.

However, identifying ways in which an overlapping consensus might contribute to democratic stability does not vindicate the Rawlsian account, for the underlying mechanisms at stake here are different. Shared conceptions of justice matter here, not because they help sustain a shared sense of reciprocity (as in the Rawlsian account), but rather because they indirectly facilitate a unified response to an executive bent on disabling democratic institutions. This reliance on different mechanisms is important for two reasons. First, as we saw in Section I, the mechanisms underlying the Rawlsian account are problematic—the reciprocity-based explanation of democratic stability does not seem like it could be operative given a realistic account of the level of disagreement about justice that is likely to emerge in a democratic society. Second, as we saw in Section V, with different underlying mechanisms, comes a different understanding of what leads to democratic backsliding and, so, a different way of thinking about the steps that can be taken to mitigate the risk of it. Thus, even if there is some role for shared views on the article’s account, that does not imply that it is somehow parasitic on the Rawlsian account—instead, the underlying mechanisms enabling stability are dramatically different.

B. Is Institutional Stability Tantamount to a Modus Vivendi?

At this juncture, some readers are likely to think that even if the account that I have described provides a more satisfactory perspective for thinking about backsliding and democratic stability, it is nevertheless tantamount to a mere modus vivendi. This may seem concerning because, as we have seen, political stability rooted in modus vivendi has worrisome pragmatic and normative implications.

Rawls worries, pragmatically, that a modus vivendi will collapse with a significant change in the underlying distribution of power. However, the point of the article’s account is that – under the right conditions – there is no incentive for parties to abandon democratic institutions, even given considerable shifts in the underlying distribution of power. The discussion of elections’ self-enforcing characteristics showed, for instance, that even officials who expect to be voted out of office often have very strong incentives to hold those elections and to comply with the results. Thus, partly by focusing on the longer run incentives facing key players, the article’s position largely sidesteps the key pragmatic concern that Rawls associates with modus vivendi stability.⁶⁴

One should be careful not to claim too much here. As the preceding discussion of backsliding makes clear, even well-designed democratic institutions cannot be regarded as impervious to breakdown. In addition to the vulnerabilities described above (i.e., stealth, packaging, and uncertainty), exogenous events could destabilize an institutional equilibrium. Yet, this risk of destabilization also exists under stability for the right reasons. Even a consensus among ordinary citizens on a particular conception of justice would not prevent the backsliding associated with “stealth” strategies. Less obviously, it would not necessarily prevent “packaging” strategies since people may agree on a conception of justice while disagreeing substantially about the policies that best satisfy that conception. (Note, again, the wide disagreement among Rawlsians about the policies that best satisfy Rawls’s two principles.) Furthermore, some exogenous event could highlight the importance of a set of issues or concerns that had not previously been salient, and this could cause some people to doubt, and eventually abandon, their previous conception of justice, thereby upending the existing consensus.⁶⁵ From a pragmatic perspective, then, the article’s account should not be regarded as fragile in the way that Rawls often portrays a modus vivendi and it may, perhaps, not even be importantly inferior – from that perspective – to stability for the right reasons.

Turning now to normative considerations, it is important to recognize that the account described above is compatible with more than compliance via the tacit threat of “brute force.”⁶⁶ As described above, appropriate institutions may give rise to a constitutional consensus, which emerges because citizens recognize the value of society’s central political institutions. Insofar as citizens embrace a society’s procedures for making substantive decisions, they may view themselves as having powerful (procedural) reasons to accept policy decisions, even when they view those decisions as substantively mistaken.⁶⁷

Stability for the right reasons requires that “the reasons from which citizens act include those given by the account of justice they affirm.”⁶⁸ Consider two ways to read this requirement. One possibility is that a system characterized by institutional stability and the resulting constitutional consensus satisfies the requirement because, even though some citizens disagree with the winning tax policy (for example), they nevertheless see it as emerging from a set of procedures that is endorsed by their conception of justice. On this reading, the account satisfies a necessary condition for stability for the right reasons and, so, may count as providing a form of such stability.⁶⁹

A second possibility, though, is that stability is not normatively attractive unless all citizens live under policies whose value they accept. This reading would prevent us from seeing the article’s account as facilitating a form of stability for the right reasons. Yet, even the Rawlsian account seems unable to satisfy this standard, since even if citizens manage to converge on a consensus on a particular conception of justice, there will still be disagreement about the policies that best satisfy that conception. The normative value that this interpretation associates with stability for the right reasons does not, then, appear consistent with a proper recognition of the burdens of judgment. The upshot is that it is surprisingly difficult to identify a tenable normative advantage associated with Rawlsian stability.