The Evolutionary Politics of Minimal Winning Coalitions: On the Work of Bruce Bueno de Mesquita and Alastair Smith

The apparent closeness of the American presidential race between Obama and Romney is a reminder of the fundamental principle of human politics--the political survival of leaders depends on their having the support of a minimal winning coalition. 

No ruler can rule alone.  Even an absolute dictator needs a small coalition of powerful people who are loyal to him, and to win and maintain that loyalty, the dictator must buy them off with money and status.  Once a dictator is abandoned by his loyalists, he loses his power.  But as long as he has the support of that small winning coalition, he can rule successfully even when he oppresses the great majority of the people under his rule.  The private interest of his small coalition of supporters is advanced at the expense of the public interest of the people at large.

A democratic leader differs from a dictatorial leader in that the democratic leader depends on a larger coalition of supporters.  Because of the large size of a winning democratic coalition, democratic leaders must persuade a large number of supporters that he will advance public policies that serve the general welfare of this big coalition.  But still this large democratic coalition is less than the whole community, and it does not have to be a majority of the citizens.

In the American electoral college system for the presidency, the minimal winning coalition requires the support of those voters in various states who will give a candidate at least 270 of 538 electoral votes.  As we know, the winner in the Electoral College does not necessarily need to win the majority of the popular votes.  Moreover, if neither Romney nor Obama wins at least 270 electoral votes, perhaps because they have evenly split the electoral votes (269 for each), then the election will be decided by a vote of the House of Representatives, with each state delegation casting one vote.  In that case, the minimal winning coalition would be 26 state delegations, which could represent a minority of the voters.  To avoid this outcome, the tie in the Electoral College could be broken if one or more of the electors could be persuaded to betray their party commitment and vote for the popular vote winner.  There is a very slight possibility that this could happen in this election.

According to Frans de Waal, this principle of minimal winning coalitions holds true not only for human politics but also for chimpanzee politics.  The natural drive of male chimps for dominance leads them to compete for rule over each chimp community.  Success in this competition depends on the exercise of strategic intelligence in which chimps must form coalitions that will support them as the alpha chimp in the hierarchy (see chapter 5 of de Waal's Chimpanzee Politics). 

This suggests that the human political principle of minimal winning coalitions could be rooted in an evolutionary history of politics shared with our primate ancestors, and thus it would be grounded deep in our natural history as political animals.

That all of politics can be explained through the principle of minimal winning coalitions is the argument of Bruce Bueno de Mesquita and Alastair Smith in The Dictator's Handbook: Why Bad Behavior Is Almost Always Good Politics (PublicAffairs, 2011).  They have summarized the argument of their book in an article for Foreign Policy--"A Dictator's Handbook for the President"--in which they suggest that if Obama is to win reelection, he will have to act more like a dictator.

Bueno de Mesquita and Smith assume that "politics is about getting and keeping political power" (xviii).  To explain how that is done, one needs to understand one fundamental fact about political power:  "No one rules alone; no one has absolute authority.  All that varies is how many backs have to be scratched and how big the supply of backs is available for scratching" (4).

They explain the variability of political life by explaining how leaders handle three groups of people--the nominal selectorate, the real selectorate, and the winning coalition.  "Fundamentally, the nominal selectorate is the pool of potential support for a leader; the real selectorate includes those whose support is truly influential; and the winning coalition extends only to those essential supporters without whom the leader would be finished.  A simple way to think of these groups is: interchangeables, influentials, and essentials" (5).

In the United States, the nominal selectorate or interchangeables for the presidency are those legally qualified to vote; the real selectorate or influentials are those who actually vote in any particular election; and the winning coalition or essentials are the smallest group of voters scattered properly across the states to win a majority in the electoral college.  By contrast, in North Korea, the nominal selectorate includes all the voters, although the elections are rigged; the real selectorate includes a tiny group of people who actually select the leader; and the winning coalition is probably no more than a few hundred people whose support is essential for sustaining the power of North Korea's leader.

Consequently, Bueno de Mesquita and Smith argue, there is no difference in kind between types of governance throughout political history, because they differ only in the relative size of their nominal selectorates, real selectorates, and winning coalitions.  Nevertheless, it is convenient to distinguish between dictatorships as small-coalition regimes and democracies as large-coalition regimes. 

In every regime, politics is about the leader paying off the right people--those who constitute the minimal winning coalition.  A dictatorial leader must pay off a small coalition of people.  A democratic leader must pay off a large coalition of people.  Thus, all politics is about "pay to play"--"Paying supporters, not good governance or representing the general will, is the essence of ruling" (25).  Consequently, all governments are corrupt, and they differ only in the level of their corruption.  "If corruption empowers, then absolute corruption empowers absolutely" (127).

In elaborating their theory, Bueno de Mesquita and Smith formulate 5 basic rules for all leaders (17-18):

Rule 1:  Keep your winning coalition as small as possible.
Rule 2:  Keep your nominal selectorate as large as possible.
Rule 3:  Control the flow of revenue.
Rule 4:  Pay your key supporters just enough to keep them loyal.
Rule 5:  Don't take money out of your supporter's pockets to make the people's lives better.

I find much of what they say persuasive.  I am particularly impressed by some of the insights that come from their theory.  One example is how they explain the remarkable stability of a dictator's rule when he has steady access to valuable natural resources (like oil, natural gas, gems, and forest products) that provide all the income he needs to pay off his small coalition of supporters.  If a dictator has this, he doesn't need his people to be economically productive, and so he doesn't need to limit his oppression in order to motivate them to work.  This explains the dictatorship of Senior General Than Shwe of Burma and others with control over vast oil resources.  By contrast, a dictator like Hosni Mubarak, who did not have access to steady income from natural resources, was exposed to revolt once the decline in foreign aid revenue and a faltering economy weakened his ability to pay off his small coalition of supporters.

Another example of the insights from Bueno de Mesquita and Smith is how they explain the failure of foreign aid to foster democracy and economic development:  foreign aid for countries ruled by dictators will inevitably be diverted to the benefit of the dictators, and by saving them from financial crisis, foreign aid takes away from dictators any need to expand their coalition of supporters.

But despite my general agreement with Bueno de Mesquita and Smith, I see five main problems with their argument:

(1) Following the traditional assumptions of game theory, they assume that all human beings are rational maximizers of their self-interest.  But they don't respond to recent research in behavioral game theory that shows human beings as moved by some moral concern for others.

(2) They denigrate political philosophy.  But then one notices that their ideas add little to the ideas of Aristotle, Machiavelli, and James Madison.

(3) They adopt the traditional stance of value-free social science, because they assert that they are concerned with studying what is and not with what ought to be the case in politics, and consequently they disparage any appeal to "good governance" or the "general welfare."  But then they endorse the values of liberal democracy as the regime that most closely approximates the ideal of good government.

(4) They assume that history is progressive in moving towards liberal democracy.  But they never explain or justify their theory of historical progress.

(5) They assume that there is a "natural order governing politics" (279).  But they never explain the ground of that "natural order."

(1) The axiom of self-interest
Bueno de Mesquita and Smith advise us to reject "fuzzy ideas like the national interest, the common good, and the general welfare."  Rather, we should see that "politics, like all of life, is about individuals, each motivated to do what is good for them, not what is good for others" (xix).  Thus, we should assume that all politicians are "self-interested louts" (xxv).  As one can see in his TED talk, Bueno de Mesquita is following the traditional assumptions of Rational Choice Theory about human beings as rational egoists. 

I agree that it is reasonable to assume that most human beings most of the time are moved by individual self-interest.  But it is unreasonable to assume that human beings never care for others, and that other-regarding behavior is never important for social and political life.  As products of natural and cultural evolution, we are both selfish and social beings.  We cannot explain our moral and political lives if we don't see this complex combination of selfishness and sociality.  Adam Smith was correct in the first sentence of The Theory of Moral Sentiments: "How selfish soever man may be supposed, there are evidently some principles in his nature, which interest him in the fortune of others, and render their happiness necessary to him, though he derives nothing from it except the pleasure of seeing it."

Moreover, as I have often noted on this blog, research in behavioral game theory (for example, in experiments with the Ultimatum Game) indicate that not all human beings act as rational maximizers of their selfish interest, because many people show sympathy or "fellow-feeling" for the experiences of others, just as Adam Smith said.  And this sympathy leads them to respect norms of fairness and social concern, even to the point of punishing violations of those norms at some cost to themselves.

Furthermore, scientific research in evolutionary psychology and neuroscience suggest that normal human beings are moved by moral sentiments or emotions such as guilt, shame, indignation, and love.  Perhaps the only people not moved by such emotions are pure psychopaths, who might be the only people who fully satisfy the assumptions of rational choice theory.  In fact, the most brutal and callous dictators are probably people with psychopathic personalities.  To explain all this, we need not just Rational Choice Theory, but also Emotional Choice Theory.

Bueno de Mesquita and Smith contradict themselves on this point.  On the one hand, they assume that human beings never care about others.  On the other hand, they concede that at least some human beings show benevolence and concern for others, because there are some "benevolent dictators" (like Singapore's Lee Kwan Yee) (104), and there are some "benevolent leaders" in democratic regimes (127).

They recognize that dictators are exposed to assassination and revolutionary violence, but they say nothing about the willingness of assassins and revolutionaries to sacrifice their lives in the attempt to overthrow a dictator.  For example, they mention the assassination of Anwar Sadat, but they don't reflect on the fact that the assassins were killed.  Rational self-interest cannot explain the willingness of people to sacrifice their own lives for what they regard as a just cause.  Machiavelli was astute enough to notice that any ruler can be killed by anyone willing to give up his life in assassinating him.

Bueno de Mesquita and Smith are also silent about how dictators can be overthrown by the "people power" of social movements practicing nonviolent resistance.  Erica Chenoweth and Maria Stephan (Why Civil Resistance Works [2011]), using data from the past 100 years, have shown that every nonviolent resistance campaign that achieved the active and sustained participation of just 3.5 percent of the population was successful in overturning dictators.  This provides quantitative evidence for Locke's account of government by consent of the governed.  It seems that the small coalition of elites supporting the dictator can be persuaded by a small nonviolent resistance movement to turn against the dictator.  This shows that dictators cannot rule without at least the passive acquiescence of the great body of the people, and so dictators can be overthrown by the active resistance of 3.5 percent or less of the people, even when that resistance is nonviolent.

(2) Political Philosophy
In a section of their book entitled "Great Thinker Confusion," Bueno de Mesquita and Smith assert that while "politics is not terribly complicated," "history's most revered political philosophers haven't explained it very well" (xix).  They indicate that people like Plato, Aristotle, Machiavelli, Hobbes, Montesquieu, and Madison all failed to explain politics because they "lacked modern tools of analysis (which we, luckily, have at our disposal)" (xxi).

But anyone who has read those political philosophers might notice that there is almost nothing in the explanations of Bueno de Mesquita and Smith that were not laid out by those philosophers.  In fact, Bueno de Mesquita and Smith occasionally admit that what they say is very close to what Madison said about democracy and what Machiavelli said about dictatorship (xx-xxi, 11).

But since Bueno de Mesquita and Smith believe that "politics is not terribly complicated," they generalize about "leaders" without distinguishing the different motivations of politically ambitious people.  By contrast, Aristotle, Machiavelli, and Madison recognize that the greatest statesmen are moved by a love of fame, honor, or glory that sets them apart from ordinary politicians.  People like George Washington, Abraham Lincoln, or Winston Churchill are moved by "the love of fame, the ruling passion of the noblest minds."  One might say that this is still self-interest, and yet it's an expansive self-interest that moves magnanimous statesmen to serve the public good on a grand scale for the sake of a kind of immortality through glory.  (Consider, for example, what Lincoln told Joshua Speed about how issuing the Emancipation Proclamation would satisfy the longing for glory that he had sought his whole life.)  But the crudely simplified political psychology of Bueno de Mesquita and Smith cannot recognize magnanimous statesmanship, because this would make political science too complicated for them, and they are confident that "politics is not terribly complicated."

What Bueno de Mesquita and Smith say about dictatorship adds nothing to what Aristotle and Machiavelli said about what tyrannts have to do to preserve their rule.  For example, what Bueno de Mesquita and Smith say about the two ways that dictators can respond to revolutionary threats corresponds closely to what Aristotle says about the two ways of preserving tyranny (198).  What they say about democracy adds nothing to what Aristotle, Machiavelli, and Madison said about republican government as a mixed regime that balances the power of one, few, and many.

This distinction in political philosophy between one, few, and many corresponds to what Bueno de Mesquita and Smith identify as the three groups of people in political life--"leaders," "essential supporters," and "masses."  As opposed to a dictatorship, a modern liberal democracy is a mixed regime in which some large portion of the minimal winning coalition includes the "many" or the "masses."

(3) Value-Free Social Scientists Supporting Democratic Values
Unlike the political philosophers, who were concerned with "'the big questions'--what the highest nature of man ought to be, or what the 'right' state of government really is, or what 'justice' truly means," Bueno de Mesquita and Smith indicate that their concern is with what really is not what ought to be (xxii). 

But then, throughout their book, they assume that liberal democracy is the ideal regime, or at least the best approximation to the best regime, because while "autocratic politics is a battle for private rewards," "democratic politics is a battle for good policy ideas" (44).  While a small-coalition regime promotes the interests of the few, a large coalition regime promotes the interests of the many.

This creates contradictions in their writing.  They say that "paying supporters, not good governance or representing the general will, is the essence of ruling" (25).  But they also say that "paying off the right people is the essence of good government" (31).  So it seems that paying supporters both is and is not good government.

If "good governance" is identified with "representing the general will" or promoting "the general welfare," then, Bueno de Mesquita and Smith conclude, "democrats are closer to this good governance ideal than autocrats" (75), because in a democracy, politicians compete in offering good policy ideas about how best to secure the public good (43-44).  Once again, they contradict themselves--both denying that politics is concerned with "good governance" and affirming that democracy is concerned with "good governance."

This "good governance ideal" corresponds to what Aristotle, Machiavelli, and Madison recognized as the common good or the "general and aggregate interests of the whole" (Federalist number 10).  They would agree with Bueno de Mesquita and Smith that "working out what makes people do what they do in the realm of politics is fundamental to working out how to make it in their interest to do better things" (xxii).

So while Bueno de Mesquita and Smith set out to be value-free social scientists, they recognize that political life is value-laden, and therefore a purely value-free social science cannot explain the reality of politics.

(4) The End of History
Implicitly, Bueno de Mesquita and Smith have a philosophy of history as a progressive movement towards liberal democracy as the best regime.  But they never fully explain or justify this philosophy of history.

They assert: "Until recently, and with very few exceptions, small-coalition systems have been the dominant form of government" (226).  But there is "hope for the future," they say.  "Every government and every organization that relies on a small coalition eventually erodes its own productivity and entrepreneurial spirit so much that it faces the risk of collapsing under the weight of its own corruption and inefficiency.  When those crucial moments of opportunity arise, when the weight of bad governance catches up with despots, then a few changes can make all the difference" (281).  Consequently, "sooner or later every society will cross the divide between small-coalition, large-selectorate misery to a large coalition that is a large proportion of the selectorate--and peace and plenty will ensue" (282).

Notice, again, the contradiction: they denigrate the idea of "good governance," but then they say that every despotic government faces collapse when "the weight of bad governance catches up with despots."

They never explain why there has been such a sudden and unprecedented shift in recent history from small-coalition regimes to the large-coalition regimes.  After all, if Rule 1 of politics is "Keep your winning coalition as small as possible," why would politicians suddenly, for the first time in history, turn to large-coalition arrangements in a way that would make this the wave of the future?

To begin to explain this, we need to see that Bueno de Mesquita and Smith are wrong when they say that small-coalition systems have dominated human history until recently.  In fact, through most of human history, human beings lived in foraging bands of hunter-gatherers that were large-coalition regimes, because they were largely egalitarian societies in which leaders were severely limited in their power by popular resistance to exploitative dominance.  It was not until about 5,000 years ago that the emergence of agrarian states based on agricultural production made it possible for leaders to rule in small-coalition regimes in which the majority of people were exploited by the ruling elite.  And even in those ancient small-coalition states (like Mesopotamia), there were frequent rebellions against despotic rule.  The emergence of modern liberal democracies over the past two or three centuries is in some ways a revival of the popular constraints on power that prevailed in ancient foraging societies.

This points to a scheme of historical progress that moves through three epochs.  Recently, Douglass North, John Wallis, and Barry Weingast (in Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History [Cambridge University Press, 2009) have explained this as a history of three social orders--"the foraging order," "the limited access order," and the "open access order."  In the open access order, prohibitions on the use of violence or coercion allow for open entry and free competition in the formation of social, economic, and political organizations.  This allows for a vibrant growth in social, economic, and political life beyond any previous social order.

To fully understand this movement of history, we would need to understand it through an evolutionary science of natural selection and cultural selection.

5. The Natural Order
If there is a "natural order governing politics" (279), as Bueno de Mesquita and Smith say, how do we explain that "natural order"?

As I have just suggested, the only comprehensive science of natural order in political life is a Darwinian science of evolution.  This Darwinian science could explain the evolution of the complex interaction of selfishness and sociality in human nature.  It could explain the evolved human nature studied by political philosophers.  It could explain the value-laden character of politics as a moral order.  And it could explain the history of political order as passing through an evolutionary process of trial and error-- from the foraging order of hunter-gatherers to the limited access small-coalition regimes of agrarian states to the open access orders of large-coalition liberal democracies.

As Bueno de Mesquita and Smith indicate, freedom is the greatest public good.  Freedom is the natural condition for human flourishing, materially and spiritually (120-21, 124-25, 180, 214, 273-74, 278-82).  Although the evolutionary path to free societies is not predetermined, because evolutionary history is an undesigned, contingent process of trial and error, the greater wealth, inventiveness, and happiness of the free society make it likely that it will eventually prevail.

Here, again, they contradict themselves: having denied the reality of the public good, they affirm freedom as the greatest public good!

The evolutionary historical logic of this trend towards a free society has been elaborated in recent books by Robert Wright (Nonzero), Matt Ridley (The Rational Optimist), and Steven Pinker (The Better Angels of Our Nature).  It's an evolutionary logic that leads to expanded cooperation, greater prosperity, and less violence.

Here we see the argument for a new political science of Darwinian liberalism, according to which the minimal winning coalition in a mixed regime is broad enough to promote the general welfare of the one, the few, and the many.