on November 28, 2015
This book is an absolute classic, and I would heartily recommend it to anyone interested in social choice. Given the prominent role of large-scale political issues across society, small-scale social dynamics among friends, and everything in between, that means just about everyone should read it!
The fundamental point of the book is that group choices are inherently incoherent. This was first discovered by Condorcet in 1785, who showed that simple majority voting among three people, among three options, could result in circular preferences. That is, the group would choose, by majority vote, A over B, B over C, and C over A. Thus, A>B>C>A, making a circle. Thus, it is impossible to say that one option (Xb) is socially best, because there is always some Yb for which the group would choose over Yb over Xb. Similarly, it is impossible to say that one option, Xw, is socially worst, because there is always some Yw for which the group would choose Xw over Yw. Most of the book is taken up with a thoughtful, thorough, precise and yet accessible presentation of the many ways in which later workers have extended and reconfirmed this fundamental result. Crucially, the only formal system of voting that meets even minimal common-sense notions of fairness is one person, one vote, majority wins between exactly two options. The technical exposition is well-mixed with plain English discussion of the real-world meaning and moral implications of the math.
Given that social preferences are generally circular (and thus incoherent), much research has turned to the idea that institutions shape outcomes at least as much as do preferences. For example, a group choice between a dozen candidates for office would generally be subject to circular preferences. However, the usual institution is for a two stage process: everyone votes for all candidates, then there is a run-off between the top two (i.e. 1P1V, majority wins, between two options). While this procedure guarantees convergence to one option in two steps, rather than endless cycling among all options, Riker demonstrates in detail how this procedure, though seemingly reasonable, fails to meet even the minimal requirements for fairness. This may be astonishing but it is true. Similarly, the chairperson of a committee can manipulate the order in which amendments are presented so as to shape the outcome. A thorough discussion of both theoretical analysis and experimental results is included. For example, NASA's choice of trajectories for the two Mariner spacecraft was done by a deeply considered and thoroughly documented group choice procedure; details are presented in the book.
Of course, the institutionalist approach does not solve the problem but simply moves it to social choice over a different set of "options": how are institutions chosen? Riker barely addresses this question, though it has obvious real-world applications. After the institutions have been shattered by civil war, foreign invasion, or social revolution, how are new ones to be chosen?
Riker becomes successively less technical after that section. The second major discussion of "agenda manipulation" concerns the search for new issues that out-of-power politicians can use both to consolidate a new coalition and to split the ruling coalition in order to get elected. Of course, the search is also done by in-power politicans both to consolidate their existing coalition and to fracture potential challengers, though Riker does not mention this ancient "divide and rule" strategy. As this involves the constant search for new options, rather than just choice between existing options (or re-ordering the agenda to choose between existing options), this section is necessarily less technical and precise than the preceding sections. It closes with an excellent discussion of "the natural selection of political issues" based on constant experimentation with new issues, some of which succeed and most of which fail. The points are illustrated with a long discussion of how and why the East/West debate over economic policy in America was turned into a North/South debate over slavery. The slavery issue obviously existed before the Civil War, and the economic issues persisted after it, but somehow different issues became salient at different times, and the motives for manipulating their salience were a thorough mixture of morality and mendacity.
All that being said, the argument falls apart when it addresses the issue of its title: liberalism vs. populism. Fortunately, this issue is largely irrelevant to the main point of the book. The book paints a dismal picture of *all* systems for social choice, so the detailed comparison of these particular two is somewhat of a diversion, probably to justify a catchy, sales-enhancing title.
In the very front of the book Riker defined populism as a two-part belief system. First, the choice of the people is automatically good, i.e. the Roussauian General Will not only exists and is discoverable but also is always correct and moral. Second, the People as a whole realize Freedom by exercising the General Will. There is no reason for any constitutional limits to the power of The People, and no need to guard the liberty of individual people, because whatever The People choose will be right and good. Of course, the whole thrust of Riker's explication of social choice (i.e. the middle 3/4 of the book) is that because social preferences are inevitably circular, there is no coherent way to define a "general will" or "social preference". Hence, populism cannot find (and chose) the best options, because no such things exist. With all-encompassing circular preferences, almost every possible option will be both chosen and then rejected, no matter how good or bad any group of individuals considers that option to be. Therefore, both the first and second part of populism are nonsensical and impossible: they assume that a general will exists, but it does not. It's like starting a math proof by saying "First, let's assume 3=1/0": anything following from that is just nonsense, "not even wrong".
Similarly, Riker defines liberalism as the belief that regular elections will enable The People to reject less-desirable rulers, and thereby give the rulers (or those competing to replace them) an incentive to do better. Thus, it does not guarantee choosing good options, but over time it discourages choosing bad ones. Riker concludes that, because liberalism aims merely to avoid the worst rather than achieve the best, liberalism is practical and feasible.
However, both arguments can be turned on their heads, to demonstrate that populism is practical and feasible, while liberalism is nonsensical and impossible. The crux of Riker's argument is to demand impossible optimality from populism, while accepting mere sufficiency from liberalism. It can be inverted by demanding optimality from liberalism, while accepting mere sufficiency from populism. Because social preferences are generally circular, liberalism cannot find (and reject) the worst options, because no such things exist. With all-encompassing circular preferences, almost every possible option will be both chosen and then rejected, no matter how good or bad any group of individuals considers that option to be. On the other hand, one could define populism as the belief that regular elections will enable The People to select more-desirable rulers, and thereby give the rulers (or those competing to replace them) an incentive to do better. Thus, it does not guarantee avoiding bad options, but over time it encourages choosing good ones.
Again, the fundamental consequence of all-encompassing circular social preference is that "better" and "worse" cannot be meaningfully defined, and this invalidates the unspoken premise of almost all debates on the subject. This means not only that "the best" simply does not exist, but also that "the worst" simply does not exist. After all, if society has the circular preference A>B>C>D>A, we can neither pick out an element as the maximum, nor pick out an element as the minimum. They are all just points in a circle. As Riker wrote, "Truly, it is politics, not economics, that is the dismal science."