An open access publication of the American Academy of Arts & Sciences
Summer 2017

Collusion in Restraint of Democracy: Against Political Deliberation

Ian Shapiro
View PDF

Recent calls to inject substantial doses of deliberation into democratic politics rest on a misdiagnosis of its infirmities. Far from improving political outcomes, deliberation undermines competition over proposed political programs–the lifeblood of healthy democratic politics. Moreover, institutions that are intended to encourage deliberation are all too easily hijacked by people with intense preferences and abundant resources, who can deploy their leverage in deliberative settings to bargain for the outcomes they prefer. Arguments in support of deliberation are, at best, diversions from more serious threats to democracy, notably money's toxic role in politics. A better focus would be on restoring meaningful competition between representatives of two strong political parties over the policies that, if elected, they will implement. I sketch the main outlines of this kind of political competition, differentiating it from less healthy forms of multiparty and intraparty competition that undermine the accountability of governments.

Ian Shapiro, a Fellow of the American Academy since 2000, is the Sterling Professor of Political Science and Henry R. Luce Director of the MacMillan Center at Yale University. His recent books include Politics against Domination (2016) and The Real World of Democratic Theory (2011).

Advocates of political deliberation usually defend it as a collaborative activity motivated by the possibility of agreement. Even when agreement proves elusive, deliberation helps people come to grips with one another’s views, draw on their different experiences and expertise, and better understand the contours of their enduring disagreements. People’s views will be better informed, and the decisions they make will be of higher quality than if they had not deliberated. When study after study reveals most people to be appallingly ill-informed about much public policy, deliberation’s appeal seems obvious. Two minds are better than one, three better than two, and so on. Democracy will be improved if its decision-making can incorporate, and build on, the benefits of deliberation. Or so it is frequently claimed.1

Deliberation should not be confused with argument. When people argue, there is an expectation that one of them will, or at least should, win. Even when we speak of one person making an argument, we see this as something that stands until it is contradicted, or challenged and beaten by a better argument. Like the deliberationists, proponents of argument believe it will enhance understanding and improve the quality of decisions. This was the essence of John Stuart Mill’s defense of the robust clash of opinions in On Liberty: it would lead people to hold better-informed and more accurate views. Mill even went so far as to worry – needlessly, it turned out – that as advancing science expanded the realm of settled knowledge, people would be deprived of argument’s benefits. No longer forced to sharpen their wits by defending their views in the marketplace of ideas, they would become mediocre dullards; less able to think for themselves and more easily manipulated by others.2

My claim here is that the argumentative and deliberative ideals should be more clearly distinguished than they usually are. They support different and incompatible institutional arrangements. I also maintain that the argumentative ideal is superior because, when appropriately institutionalized, it helps hold governments accountable for their actions. By contrast, the deliberative ideal cannot easily be institutionalized – and perhaps cannot be institutionalized at all – because people who prefer to bargain can easily abuse rules designed to promote deliberation. But deliberation’s difficulties run deeper. Its defenders fail to appreciate that, in politics, deliberation and the search for agreement are – to borrow an antitrust analogy – unhealthy forms of collusion in restraint of democracy. They should worry less about voter ignorance, which, as Anthony Downs noted long ago, might well reflect sensible budgeting of scarce time, and worry more when office-seekers fail to engage in robust public debates over the policies that, if elected, they will enact.3

Joseph Schumpeter’s competitive model of democracy, in which governments acquire power by prevailing in a “competitive struggle for the people’s vote,” gives institutional expression to the argumentative ideal.4 This was perhaps best exemplified in the Westminster system as it existed from 1911, when the Parliament Act stripped the House of Lords of its real powers, until the late 1990s, when the Lords was reformed to enhance its legitimacy as a second chamber and the Commons began ceding authority to European and other courts, the Bank of England, and independent agencies. The twentieth century’s middle eight decades were the heyday of Parliament’s supremacy within the British political system and of the Commons’ supremacy within Parliament. Epitomized at Prime Minister’s Questions, the sometimes overwrought weekly gladiatorial clashes over the famous wooden despatch boxes, it thrives on the ongoing contest between opposing policies and ideologies.

Schumpeterian democracy depends on alternation between two strong parties in government. The party that wins the election exercises a temporary power monopoly, but the loyal opposition – a government-in-waiting whose leaders hope to take power at the next election – continually challenges its policies. This system depends on combining first-past-the-post single member plurality (SMP) electoral systems with parliamentary democracy. The SMP electoral system produces two large parties, so long as the political makeup of the constituencies more or less reflects the political makeup of the national population.5 Parliamentary systems ensure that the parties will be strong because the leader of the majority party is also the chief executive. Government and opposition clash across the aisle continually, and compete during elections by offering voters the different programs they plan to implement.

The deliberative model, by contrast, calls for institutions that create incentives to seek agreement rather than victory – or at least agreement as a condition for victory. Rules that require concurrent majorities in bicameral chambers force representatives to find common ground when they can, and compromise when they cannot. Executive vetoes and supermajority provisions to override them create similar incentives. Proponents of deliberation often find proportional representation (PR) congenial for comparable reasons. Instead of two catchall parties that must submerge their disagreements in order to win elections, PR leads to party proliferation, bringing a more diverse array of voices to the political table. In addition to the left-of-center and right-of-center parties characteristic of SMP systems, in PR systems, liberals, religious groups, Greens, separatists, and nationalists, among others, can all elect representatives to the legislature to be part of the conversation. Because one party seldom wins an absolute majority, coalition government, which forces parties to seek and perhaps even manufacture common ground, is the norm.

The U.S. system is a hybrid. The SMP electoral system produces two large parties, but the independently elected president weakens them, and the system of checks and balances forces consensus-seeking and compromise to the extent possible. The American founders intended the Senate, in particular, to be a constraining body made up of what Jefferson would later refer to as an “aristocracy of virtue and talent.” It has been heralded as such by commentators dating back at least to Alexis de Tocqueville.6 The idea that the Senate is the world’s greatest deliberative body, which first gained currency with Daniel Webster’s three-hour soliloquy in defense of the Union in 1850, has been repeated to the point of banality, no matter how scant its connection with reality.7 I will have more to say about the kind of competition the U.S. system fosters shortly. As a prelude to this, notice that, unlike the Westminster model, which gives temporary control of the government’s power monopoly to the majority party and relies on alternation over time as its main mechanism of accountability, the U.S. model divides up the control of power on an ongoing basis. Madison’s slogan was that “ambition must be made to counteract ambition.”8 The checks and balances force the players in the different branches to accommodate themselves to one another; hence its affinities with the deliberative ideal.

Up to a point. A major limitation of institutions that encourage deliberation is that they can produce bargaining instead. Juries, for example, are traditionally subject to unanimity requirements that put pressure on their members to talk out their differences until they reach agreement. When this works well, it produces thorough exploration of all the arguments and evidence provided by the contending parties: a poster child for the benefits of deliberation. But a jury can also be held hostage by a recalcitrant crank who has nothing better to do when everyone else wants to go home. His superior bargaining power and stubbornness might enable him to extract agreement from the others, but this will not be deliberative consensus on the merits of the case. What holds for juries also holds for other institutions that we might hope will induce deliberation. When they produce bargaining instead, those with the most leverage will prevail. So it is that small parties often exert disproportionate influence over coalition governments, U.S. Senators can use holds and filibuster rules to thwart the will of the majority, and various other super majority and concurrent majority rules can be deployed to similar effect.

In short, deliberation requires people to act in good faith, but it is not possible to design institutions to induce good faith. “If men were angels,” Madison wrote, “no government would be necessary.”9 Indeed, when power is at stake and representatives must answer to constituents, the impulse to bargain will likely overpower even genuine desires to reason collaboratively. In 2009, a number of centrist Republican Senators showed an interest in working with the Obama White House for “cap-and-trade” legislation on toxic emissions control. They soon bolted, however, when confronted with Tea Party – orchestrated threats of primary challenges in their constituencies, should they choose to persist.10 Since power is endemically at stake in politics, it seems unlikely that there will be much genuine deliberation or that politicians will resist the impulse to exploit rules that might maximize their leverage instead.

An exception that proves the rule is the British House of Lords. It functioned most effectively as a deliberative body after it lost most of its real powers in 1911. Peers who participated were mainly public-spirited individuals who specialized in particular areas and were often nonpartisan or crossbenchers. But the Lords has become more partisan and assertive since the 1999 reforms restored a measure of its legitimacy as a somewhat democratic institution, albeit one at a considerable distance from the ballot box.11 What the Lords has gained in legitimacy has come at the price of diminished effectiveness as a deliberative institution.12

The various deliberative institutions that have been tried out or proposed in recent years are exclusively consultative. Deliberative Polls and citizens’ juries have no authority to decide anything. They might affect how people vote, but it is the voting that will be decisive. Objects of theoretical conjecture like ideal speech situations are even more radically divorced from politics, since they depend on armchair speculation about what people would decide in settings that are devoid of power relationships. Questions can and have been raised about whether such speculations add up to anything we should believe, or whether the changes in people’s views produced by Deliberative Polls and other consultative mechanisms tried thus far are really improvements on their pre-deliberative views or simply changes.13 These issues need not detain us here, however, since my present point is that – whatever its merits – institutionalizing deliberation turns out to be an elusive endeavor. If it is purely consultative, it is not clear why anyone will or should pay attention to it. Yet if rules are created to institutionalize deliberation and give it real decision-making teeth, they can all too easily undermine political competition and empower people with leverage to appropriate them for their own purposes.

Schumpeter’s competitive model of democracy trades on analogies between the political marketplace of ideas and the economy. Political parties are the analogues of firms; voters mirror consumers. Schumpeter treats the policies that parties propose to enact if they become governments as the political analogues of the goods and services that firms sell, and the votes that politicians seek as analogues of the revenues that firms try to earn. Democratic accountability is the political equivalent of consumer sovereignty: the party that does best at satisfying voters wins their support.

Schumpeter’s illuminating analogy is nonetheless strained in several ways, two of which matter here. One is that political parties are vying to control a monopoly, a fact that constrains competitive possibilities. As I argue below, the best option is competition between two large, centrally controlled parties. The Schumpeterian analogy also falters because there is no unproblematic equivalent of a firm’s shareholders for political parties. Some will single out party members or activists as the appropriate political shareholders, but parties that empower them run into trouble. Membership in political parties is typically free or very cheap, rendering them susceptible to hostile and anomalous takeovers, like that perpetrated by Donald Trump in the 2016 Republican primaries, or that which occurred in the British Labour Party in the summer of 2016. Party leader Jeremy Corbyn lost a confidence vote in the Parliamentary Labour Party by 172 to 40 in June, triggering a leadership challenge, but an easily augmented membership nonetheless reelected him as leader with 61.8 percent of the vote three months later.14 As this example underscores, grass roots activists tend to be unrepresentative of a party’s supporters in the electorate. This imbalance can be especially pronounced in two-party systems, which, as I argue below, are nonetheless best from the standpoint of robust public debate.

Representation should be geared to maximizing the chances that public debate will center on the policies that parties, if elected, will implement as governments.15 This is why SMP beats PR, and why strong, centralized parties are better than weak, decentralized ones. Supporting a party in a multiparty system can help voters feel better represented because their representatives’ views are likely closer to their ideals than would be the case in a two-party system. But this is an illusion. What really matters is the policies that governments will implement. That cannot be known until after the coalition is formed, post-election. Coalition governments decrease accountability, since different coalition members can blame one another for unpopular policies.16 Americans got a taste of this when unusual conditions produced a cross-party coalition to enact the Budget Sequestration Act in August 2011, putting in place $1.1 trillion of automatic spending cuts over eight years split evenly between defense and domestic programs, unless Congress passed an alternative by January 2013. The Sword-of-Damocles proposal was widely said to be sufficiently draconian that the representatives would be forced to find a compromise. In the event, they did not and the sword fell, with each side blaming the other for intransigence. Perhaps it was a cynical way for both parties to achieve cuts without being savaged by their electoral bases. Whether due to blundering or collusive cynicism, the result was that everyone had an alibi and no one was undeniably responsible for the outcome. Coalition governments live perpetually on such ambiguous terrain, undermining accountability for what governments actually do.

Competition enhances political accountability, but some kinds of competition are better than others. As we have seen, competition between representatives of two parties, one of which will become the government, enhances accountability because they run on the platform they will be judged on as governments. Moreover, the need to sustain broad bases of voter support gives them strong incentives to advocate policies that will be good for the country as a whole, or at least for large swaths of the population. Smaller parties represent more narrowly drawn interests: business, organized labor, and ethnic and religious groups. This loads the dice in favor of clientelism, because politicians know that they will be held accountable for how effectively they advocate or bargain for their group’s interests in a governing coalition. It is better for parties to compete over what is best for the country as a whole than to bargain over the rents they can extract for their clients. This contrast can be overdrawn, to be sure, because large catchall parties consist of different interests among whom implicit bargains must be struck to keep them in the party. But that bargaining is constrained by the need to propound and defend platforms that can win support from other groups as well, otherwise they cannot hope to become the government.

The sequester episode underscores the fact that the weakness of U.S. political parties is only partly due to republican institutional arrangements. Another source of party weakness is decentralized competition, an artifact of the wrongheaded idea that local selection of candidates somehow makes the process more democratic. In reality, because of their comparatively high rates of participation, activists, whose beliefs and preferences tend to be both more extreme and more intensely held than the median voter in their constituencies, dominate primaries and caucuses. This enables them to force representatives to pursue agendas that the median voter in their district abjures, or to serve the median voter only with the kind of subterfuge that might have been at work behind the Budget Sequester Act. The same is true of referenda, which sound democratic – “hooray for direct democracy!” – but which also enfranchise intense single-issue voters who turn out at disproportionately high rates. Thus it was with the Brexit referendum in June of 2016, when a majority of those who voted produced the result to leave, even though polling indicated that the median British voter favored the UK’s remaining in the European Union, as did substantial majorities of both major parties in the House of Commons.17

Some will say that making the system responsive to voters with intense preferences is a good thing. There is, indeed, a strand of democratic theory dating back to James Buchannan and Gordon Tullock’s Calculus of Consent in 1962 whose proponents defend vote trading and vote buying on the utilitarian ground that it improves the overall social utility.18 But democracy’s purpose is to manage power relations, not to maximize social utility. The contrary view would suggest that it was right for the U.S. government to abandon Reconstruction when Southern whites opposed it with greater intensity than most voters favored it, and that it was right for the intense preferences of neoconservatives who wanted the United States to invade Iraq in 2003 to override those of morenumerous but less-fervent skeptics.19 This is to say nothing of the fact that in politics, preferences are always expressed subject to budget constraints. The intense antiregulation preferences of the multibillionaires Charles and David Koch are massively amplified because their budget constraints differ vastly from those of the typical voter.20 In short, there are good reasons for the rules of democratic decision-making to reflect how many people want something, rather than how intensely they want it.

People have theorized about democracy for millennia, yet it is only in the past few decades that the idea has gained currency that democracy depends on, or at any rate can be substantially enhanced by, deliberation. I have sought to show here that this is a dubious proposition. It is hard, if not impossible, to create institutions that will foster deliberation in politics, and institutions designed to do so are all-too-easily hijacked for other purposes. But deliberation is in any case the wrong goal. Competition is the lifeblood of democratic politics, and not just because it is the mechanism by which governments that lose elections give up power. Institutions that foster competition also structure politics around argument, which Mill was right to identify as vital to the advancement of knowledge and good public policy.

But not any competition. The contestation over governing ideas that Mill prized is best served when two large parties are constrained to compete over potential governing programs. It is compromised by multiparty competition that encourages clientelism, as we have seen. And it is damaged even more by competition within parties, which empowers people with local agendas and intense preferences who participate disproportionately in primaries and caucuses. This can render parties vulnerable to the ideological capture of candidates by well-funded groups, as has happened with the Tea Party in Southern and Midwestern Republican primaries since 2009. But a more general problem is associated with local control of selection processes, in which candidates find themselves compelled to compete by promising to secure local goods. Once elected, they face powerful incentives to engage in pork barrel politics with other similarly situated politicians, protecting public funding for sinecures and bridges to nowhere in their districts. This problem is worse in districts – the vast majority in the United States – that have been gerrymandered to be safe seats, so that the primary is the only meaningful election. It is better for party leaders to seek candidates who can both win in their districts and support a program that can win nationally. The leaders, in turn, are held accountable by the backbenchers who remove them when they fail to deliver winning platforms. In sum, two large, centrally controlled parties are most likely to foster the programmatic competition that is best for democratic politics. By contrast, multi-party competition encourages wholesale clientelism, and intraparty competition encourages retail clientelism.

Deliberation can be rendered harmless and perhaps, occasionally, beneficial for democratic politics by relegating it to a purely consultative role; but in that case, it is hard to see what the hype surrounding deliberation amounts to. Regardless, the most pressing political challenges in the United States do not result from lack of deliberation. Rather, they stem from the increasing subversion of democracy by powerful private interests since the Supreme Court’s disastrous equation of money with speech in Buckley v. Valeo four decades ago, and the subsequent playing out of that logic in Citizens United and subsequent decisions.21 As politicians have become increasingly dependent on countless millions of dollars to gain and retain political office, those with the resources they need undermine the process by manufacturing – and then manning – huge barriers to entry, by contributing to both political parties in ways that stifle competition, by capturing regulators and whole regulatory agencies, by giving multimillionaires and billionaires the preposterous advantage of running self-funded campaigns, and by doing other end-runs around democratic politics. Unless and until that challenge can be addressed, debating what deliberation can add to politics is little more than a waste of time.


1 For example, see Jane J. Mansbridge, Beyond Adversary Democracy, 2nd ed. (Chicago: University of Chicago Press, 1983); Amy Gutmann and Dennis Thompson, Democracy and Disagreement (Cambridge, Mass.: Harvard University Press, 1996); Amy Gutmann and Dennis Thompson, Why Deliberative Democracy (Princeton, N.J.: Princeton University Press, 2009); James S. Fishkin, The Voice of the People (New Haven, Conn.: Yale University Press, 1995); James S. Fishkin, When the People Speak: Deliberative Democracy and Public Consultation (Oxford: Oxford University Press, 2011); Hélène Landemore, Democratic Reason: Politics, Collective Intelligence, and the Rule of the Many (Princeton, N.J.: Princeton University Press, 2012); Jürgen Habermas, Communication and the Evolution of Society (Boston: Beacon Press, 1975); and Jürgen Habermas, The Theory of Communicative Action (Boston: Beacon Press, 1984). There are, of course, major differences among these and other theorists of deliberative democracy that do not concern me here.

2 John Stuart Mill, On Liberty, ed. David Bromwich (New Haven, Conn.: Yale University Press, 2003 [1859]), 86–120.

3 See Anthony Downs, An Economic Theory of Democracy (New York: Harper and Row, 1957), 244–246, 266–271.

4 Joseph Schumpeter, Capitalism, Socialism, and Democracy (London: George Allen & Unwin, 1942), 269.

5 Where there is substantial regional variation, by contrast, as in India, SMP systems can produce party proliferation.

6 Tocqueville described the Senate as peopled by America’s “ablest citizens”; men moved by “lofty thoughts and generous instincts.” By contrast, the House of Representatives consisted of “village lawyers, tradesmen, or even men of the lowest class” who were of “vulgar demeanor,” animated by “vices” and “petty passions.” Alexis de Tocqueville, Democracy in America, ed. J. P. Mayer (New York: Anchor Books, 1969 [1835, 1840]), 200–201.

7 Terence Samuel, The Upper House: A Journey behind Closed Doors (New York: Palgrave MacMillan, 2010), 68.

8 James Madison, “Federalist No. 51,” in The Federalist Papers, ed. Ian Shapiro (New Haven, Conn.: Yale University Press, 2009 [1787–1788]), 264.

9 Ibid.

10 Jane Mayer, Dark Money: The Hidden History of the Billionaires behind the Rise of the Radical Right (New York: Doubleday, 2016), 198–225.

11 The House of Lords Act of 1999 reduced the membership from 1,330 to 699 and got rid of all but ninety-two of the hereditary peers, who were allowed to remain on an interim basis, and an additional ten who were made life peers. On the recent evolution, see Meg Russell, The Contemporary House of Lords: Westminster Bicameralism Revived (Oxford: Oxford University Press, 2013), 13–35, 258–284.

12 For elaboration, see Ian Shapiro, Politics against Domination (Cambridge, Mass.: Harvard University Press, 2016), 73–78.

13 See Ian Shapiro, The Real World of Democratic Theory (Princeton, N.J.: Princeton University Press, 2010), 266–271.

14 James Lyons and Sanya Burgess, “Corbyn Reelected as Labour Leader with Increased Mandate,” The Times, September 24, 2016,

15 The argument in this and the next two paragraphs will be developed more fully in Frances Rosenbluth and Ian Shapiro, Democratic Competition: The Good, the Bad, and the Ugly (New Haven, Conn.: Yale University Press, forthcoming 2018).

16 Kathleen Bawn and Frances Rosenbluth, “Short versus Long Coalitions: Electoral Accountability and the Size of the Public Sector,”American Journal of Political Science 50 (2) (2006): 251–265.

17 Charlie Cooper, “EU Referendum: Final Polls Show Remain with Edge over Brexit,” The Independent, June 23, 2016,; and Anushka Asthana, “Parliamentary Fightback against Brexit on Cards,” The Guardian, June 26, 2016,

18 James Buchanan and Gordon Tullock, The Calculus of Consent: Logical Foundations of Constitutional Democracy (Ann Arbor: University of Michigan Press, 1962). Robert Dahl flirted with the notion that attending to intensity might be desirable from the standpoint of political stability, though he was skeptical that it could be measured. Robert Dahl, A Preface to Democratic Theory (Chicago: University of Chicago Press, 1956), 90–123.

19 For discussion of the dangers inherent in catering to intense preferences, see Shapiro, Politics against Domination, 46–61.

20 See Mayer, Dark Money, 120–158, 226–270, 354–387.

21 See Buckley v. Valeo 424 U.S. 1, 59 (1976); Citizens United v. Federal Election Commission 558 U.S. 310 (2010); and v. Federal Election Commission 599 F.3d 686, 689 (D.C. Cir. 2010).