1. Conception of Democracy

In some respects, Bai's conception of democracy remains undertheorized. He says that the principle of “one person, one vote” is “essential” to democracy (53), but clearly this is not the only defining commitment of democracy. In addition to a principle giving each voter one and only one vote, democrats also believe in universal suffrage (who gets to vote), majoritarianism (how the votes aggregate to determine a winner), and equal opportunity for informal political influence (the context in which voting occurs). These might seem like fussy reminders, but they matter to the empirical comparison of democracy and hierarchy. Like many current critics of democracy, Bai points to the United States as Exhibit A of the bad results associated with democracy. But in important respects the United States does not fit the profile of a pure democracy. Countermajoritarian institutions—including the Senate, Electoral College, and the Supreme Court—play very powerful roles in American politics. The equal suffrage of poor and minority voters is effectively limited by many obstacles that are deliberately created by public policy. And the richest American individuals and corporations are able to dominate public discourse. US failures around climate policy—one of Bai's main examples (56–58)—may be less an indictment of democracy than a consequence of some of the egregiously undemocratic features of the American political system.

A second weakness in Bai's discussion of democracy concerns its relationship with meritocracy. Bai rightly sees democracy and hierarchy as conceptually opposed to one another. But democracy need not be opposed to meritocracy. Designers of democratic institutions have always looked for ways to improve decision making within such institutions. Representation is one device intended to have this consequence, as are various electoral and legislative rules and procedures designed to promote better deliberation. Democracies have executives, and, in modern states, executives have access to professional civil services. Where officials have some autonomy, and are accountable to democratic representatives, meritocracy coexists with democracy.

2. The Service Conception of Government Legitimacy

A foundational commitment of Bai's Confucian approach is to the idea that “the legitimacy of the state lies in service to the people” (47). While this service conception of legitimacy seems straightforward enough, it can be interpreted in two ways. According to a “generic” interpretation, government is legitimate if and only if it promotes or realizes the legitimate interests of the people. According to an “instrumental” interpretation, government is legitimate if and only if its decisions advance the legitimate interests that people have outside of the sphere of governance. The language of “service” suggests the instrumental reading, but this reading makes the service conception quite controversial. By limiting attention to the results of government decision making, it rules out the possibility that the people have a legitimate interest in how decisions are made. A generic formulation, by contrast, is consistent with recognizing both result-based interests and process-based interests. While less controversial, this view does not clearly support meritocratic hierarchy. To do so, Bai would need to show that hierarchy outperforms democracy in producing good results and that people do not have compelling process-related interests that swing the overall comparative argument in favor of democracy. As I read Bai's book, it either simply assumes without argument the instrumental conception, or it adopts the generic conception but fails to explore possible process interests.

A further problem with a results-based theory is that it is unclear which results to count as good ones. Despite his apparent reliance on superior results to justify the hybrid system, Bai has surprisingly little to say about which results he has in mind. In several places, he mentions economic growth as relevant (52, 105), but this highlights a problem. Given that it conflicts with other possible social goals, how much to value economic growth is precisely the sort of question that a system of governance should help to resolve. It is question-begging to assume the superior value of growth when comparing alternative systems. Because the same could be said about other possible social goals, many political theorists have been drawn away from results and towards process values as the fundamental metric for making comparisons.

Bai might respond by insisting that the inferior intellectual and moral capacities of many voters imply that results achieved under democracy will be generally worse than under his preferred hybrid. Maybe so, but we should not discount the possibility that even voters with modest intellectual and moral capacities might contribute to some better outcomes, perhaps simply by being able to experience relevant forms of satisfaction and dissatisfaction. A theory based on results would need to examine the empirical evidence much more thoroughly than Bai does.

3. Asymmetric Idealization

Throughout Bai's book, democracy is portrayed very much as the realists see it. Bai notes that Rawls articulated a concept of “reasonable” citizens, and favored a form of “deliberative” democracy (59–62). Liberal thinkers like Rawls favor well-informed and actively participatory citizens who are guided by a sense of justice. “Unfortunately,” Bai informs us, “it seems that we cannot realistically expect the reasonable people to form a majority” (60).

With limited exceptions (e.g., 89–90), a corresponding realist perspective on meritocratic hierarchy is absent from the book. One might expect that concentrating significant power in a group of already successful people, and removing mechanisms of democratic accountability, would imply a serious risk that government will ignore the interests of the people. Such a system creates many opportunities for corruption and self-dealing on the part of the elite, and gives tremendous power to whichever agent is responsible for selecting people into the elite and judging their performance. It is predictable that loyalty to the particular vision and goals of that agent will quickly become the dominant factor. There is also significant risk that members of the elite, given their meritocratic achievements, will struggle to see things from the standpoint of the less advantaged, and for this reason fail to serve the people's interests.

Rather than compare a realist view of one regime type with an idealized view of another, a more balanced approach should consider how likely it is that each ideal can be realized, and, if it fails, what nonideal fallback regime will likely emerge. It seems possible that this sort of analysis might end up vindicating Churchill's dictum about democracy after all—that it is the worst form of government, except for all the others that have been tried (52). The ideal types in both the models explored by Bai are very hard to realize. But the corruption of hierarchy is really bad, while the corruption of democracy is something that is at least tolerable. Bai himself seems to hint at such a conclusion late in the book (283).

4. Liberalism, Hierarchy, Democracy

One striking feature of Bai's theory is its insistence on preserving certain liberal elements, even while compromising democracy with meritocratic hierarchy. Unlike some critics of liberal democracy, who favor nonliberal democracy, Bai argues for liberal nondemocracy. “The rule of law and human rights are endorsed and firmly established in this Confucian regime” (68; xv, 53, 244).

I am reluctant to criticize this feature of Bai's view, since it contributes to the humane and moderate character of his proposal. However, I worry that liberal nondemocracy does not form a stable equilibrium in today's world. Given the global pre-eminence of democracy, and given the ways in which democracy “flatters” ordinary people (as critics would put it), it is predictable that people will use their liberal rights to demand democracy. They will speak out, assemble, link up with others, march, protest, and so on, in favor of democratic transition.

If this is correct, then liberal nondemocracy may eventually face a dilemma. Either it will have to abandon the liberal guarantees in order to protect nondemocratic institutions, or it must accept democracy as the best form of governance that is consistent with respect for individual freedoms and the rule of law. The choice is authoritarianism or liberal democracy: there is no stable middle ground.