Let me begin by saying that Democracy without Shortcuts is an excellent—and timely—book. It makes so many valuable contributions to the literature that I could spend this entire review discussing only those contributions. I would single out for special mention the participatory conception of deliberative democracy that Lafont develops (especially, in chapter 6)—one that does not take, as she quotes Oscar Wilde, “too many evenings”; Lafont’s discussion of the role of judicial review in a democracy (chapter 8), which makes it an essential part of a true democracy, not an anti-democratic add-on; and her discussion of the power of considered public opinion (especially, in chapter 2) to generate, never total, but partial agreement over time - for example, on a right to interracial marriage, women’s right to vote, or the prohibition of slavery and peonage (Lafont 2020, 61), rights to life and physical integrity and against rape (Lafont 2020, 64), and a right to same sex marriage (Lafont 2020, 64)- rather than the endless, irresolvable disagreement that some authors emphasize.

Lafont’s book has also narrowed a disagreement on which she and I have had a long-running conversation over many years. In an earlier book, I argued that a rights-respecting democracy is the best form of government for human beings, but that it is, is itself a contingent fact (Talbott 2010). To illustrate the contingency, I argued that legislation by deliberative poll could be a better form of government for human beings.

In a deliberative poll, a stratified random sample of the population is selected to be statistically representative of the entire population on traits that are regarded as politically significant—for example, age, sex or gender, race, religion, income, party affiliation, urban or rural, etc. They constitute a minipublic—a representative sample of the total population. They are tasked with addressing a particular issue or question; they are provided with briefings on all sides of the issue; they deliberate among themselves and, ultimately, they vote. Of course, in contemporary society, minipublics have no legislative power. An empowered minipublic would be a deliberative poll with legislative power. This kind of government is called a lottocracy.

In my book, I argued that lottocracy would be a better form of government than democracy, were it not for the potential for corruption. In the intervening ten years since I wrote that book, American democracy has become so much more corrupt that today the potential for corruption in a lottocracy seems to me less serious than the actual existing corruption in American democracy, so now I am prepared to argue that some kind of lottocracy probably could be a better form of government for the U.S. than its current form of democracy, if it were adopted in a step-by-step process that I will describe shortly. I feel empowered to make this argument because it seems to me that in her arguments against lottocracy Lafont actually provides me with the tools I need to complete my argument. In her argument against lottocracy, Lafont makes a strong case for the use of minipublics as sources of information for democratic deliberations. I am going to try to show that there is a plausible pathway from this informational use of deliberative polls to some form of legitimate lottocracy, and I am going to argue in favor of taking that path.

There are two main inter-related prongs of Lafont’s argument against lottocracy. The first is that a lottocracy would require democratic citizens to blindly defer to the judgment of their “representatives” in the minipublic (Lafont 2020, 10-14 and 109). The second is that because such a system of legislation would bypass considered public opinion, “over time the actual public opinion of the citizenry would become more and more disconnected from the laws and policies to which they are subject” (Lafont 2020, 121). This disconnect would not only produce alienation, but it would also significantly impact enforcement of the laws themselves, since in a democracy, which laws get enforced and whom they are enforced against depends, in substantial part, on considered public opinion.^{^[1]} I summarize these two concerns as the concern about blind deference and the concern about alienation.

I think that both of these concerns can be largely neutralized, if we recognize that the changeover from a democracy to a lottocracy can be a gradual, step-by-step process, which need not terminate with a complete lottocracy, but rather with some kind of hybrid of democracy and lottocracy. I, myself, believe that the process would culminate in a near-complete lottocracy, but I am willing to wait and see where the process would lead.

There would be no way to motivate a move to lottocracy if the citizenry were happy with the results of their democracy. I am going to focus on the U.S., because in the U.S. there is general agreement on the right and on the left that the political system is broken.^{^[2]} Senators representing 16.2 percent of the U.S. population control half the votes in the Senate. Of course, it is unlikely that all of those Senators would agree on any single piece of legislation, but because rural states are overwhelmingly represented by Republicans, it is often the case that legislation fails even though it is supported by Senators representing a large majority of the population, or succeeds even though it is supported by Senators representing a minority of the population. Here is one example: The Brett Cavanaugh Supreme Court nomination was approved by a 50-48 vote in the Senate. The 48 who voted against the nomination represented 34 million more people than the 50 who voted for it (Schoen 2018). So the U.S. is already far from being a majoritarian democracy.

Because implementing empowered minipublics in the U.S. would require amending the U.S. Constitution, there is no real prospect of its happening soon. However, because the failures of democracy are so stark in the U.S. Congress, I use it as an example of how a step-by-step process could, in theory, lead to some kind of lottocracy, without concerning myself with the practical details.

The step-by-step process to some form of lottocracy could begin with issues of two kinds:

(1) Issues on which a large majority of public opinion favors one alternative. How large a majority? Well, we could start with 2/3 (66.6%). Here are some examples: universal background checks on gun sales (³ 85%);^{^[3]} enacting limits on campaign contributions (77%) (Pew Research Center 2018); providing a path to citizenship for currently undocumented immigrants in the U.S. (81%) (Norman 2019); and raising the federal minimum wage from $7.25 to $15 per hour (67%) (Leslie and Hartig 2019).

These proposals have repeatedly failed to garner majority support in the U.S. Congress. No one thinks that a minipublic would overrule these large majorities. If minipublics were empowered to enact legislation favored by large majorities, all of these measures would be adopted easily. But the process by which they would be adopted would also educate the citizenry about deliberative polls.

Imagine that an important issue is being addressed by a minipublic with the power to make law on that issue. Of course, the sessions would be televised, because, for many people, they would be must-see reality TV. Many voters who tune into Congressional debates in the U.S. are surprised when they see so many white men doing the arguing. A minipublic would be very different. It would be a representative sample of the people. Of course, the sample size (e.g., 500) would rule out representatives for every sub-group. There would almost surely be no professional philosophers and no equity fund or hedge fund managers. But more than 50% of the representatives would be female, and there would be a large number of different races and ethnicities among the representatives.

It is hard to overestimate the shock citizens would feel when they saw a genuinely representative sample of the American population.^{^[4]} Senators and Representatives are not required to report all of their wealth (e.g., the value of their residence), but given only what they do report, the median net worth of a Senator or Representative is around $511,000 (Hawkings 2018). In a representative sample of the U.S. population, the median net worth would be approximately $97,000, which for many, if not most of them, would largely reflect the value of their residence.

There has never been anything like an empowered minipublic in the history of democracy. As I said, it would be must-see TV. Many more people, especially people who do not see themselves represented by the current members of Congress, would tune in to view some of the process, or would tune in to view shows that presented highlights of the process, than follow Congressional deliberations on TV today. Imagine the audience appeal of a forum in which the “other 99%” would have 99% of the political power. Of course, very few people would have time to watch all day every day as the representatives listened to presentations from the various sides and deliberated about them. But they would not have to. Very quickly, social media would identify representatives for salient demographic groups. For example, many Catholics would be particularly interested in the Catholic representatives whose views most closely resembled theirs; similarly for other groups. I would expect the representatives to have Twitter accounts and perhaps also to write blogs for their followers, in which they summarized the highlights of the proceedings and explained how their own views had evolved, if they had. Imagine you are following your favorite representatives, those whose views align closely with yours. You would be most interested in information that your favorite representatives found surprising, or in arguments that your favorite representatives found persuasive, especially information or arguments that led your favorite representatives to reconsider their position on the issue. Just by following your favorite representatives, you would get the information most relevant to you, and thus you could get the educational benefit of the proceedings most relevant to you without having to go through the entire process. Here I am describing some of the educational benefits of minipublics that Lafont discusses at much greater length. All that I would add is that the educational benefits would be much more extensive if the minipublics were empowered, because when the general population realized that their deliberations would actually make law, they would pay a lot more attention to them.

There is another category of legislation that seems tailor-made for resolution by minipublic: important issues that fly under the public’s radar. There are many issues on which the general public, by and large, has no opinion, either because they are too abstruse or too technical or because the general public can only attend to a relatively small number of issues. If all such issues were trivial, then it really would not matter what process was used to decide them. But many of them are far from trivial. Here is one example: the carried interest loophole in U.S. tax law. Private equity and hedge fund managers manage 2.8 trillion dollars in the U.S. alone. These managers typically earn 20% of the total profits. In a good year, individual managers earn billions. Under U.S. tax law, these earnings are taxed at the capital gains rate (20% on long-term capital gains) rather than the ordinary income rate (37%), even though the managers have not invested any of the capital on which the earnings are capital returns. In any reasonable world, the managers’ income would be taxed as ordinary income, just like the income of any other financial adviser, which would generate billions of dollars in additional tax revenue.

Under the current system, whenever the tax code is revised, there are backroom deals by politicians who receive large donations (and other benefits) from these managers. Somehow, whenever the tax law is revised, the carried interest loophole survives, as it did in the 2017 Tax Law, which only limited the reduced carried interest rate to assets held more than three years.

Large majorities of Americans believe that the wealthy are taxed too little, not too much.^{^[5]} But very few Americans know what “carried interest” means and very few could even explain what a hedge fund is. However, if the issue were decided by a deliberative poll of a statistically representative sample of the American population after having the opportunity to hear arguments from all sides, the vote to tax carried interest as ordinary income would probably be as close to unanimous as any vote on any issue.

In theory, it would be possible to have an educational campaign to educate the public on this issue. However, there are far too many issues on which the American public is ignorant for it to be feasible to have educational campaigns and referenda. On these issues, there is no way to avoid deferring to someone, either to legislators who make backroom deals out of the public eye, or to a statistically representative minipublic who have heard from all sides of the issue.

I have described here the first two steps of a process by which minipublics might be gradually empowered to legislate on more and more issues. My description of these first two steps is intended to neutralize Lafont’s concerns about blind deference and alienation. The step-by-step process would enable citizens to evaluate the results so that, for most citizens, their trust of minipublics would be based on evidence, not blind deference. Just taking the first two steps of the process that I describe here would go a long way toward eliminating the alienation generated by the current system, when even proposals supported by a large super-majority of citizens get buried in the legislative process.

One other objection that Lafont raises to empowered minipublics is that there would be a status quo bias (Lafont 2020, 125). This is an objection not to the process, but to its substantive results. It is hard to know what to make of this objection. The current system has a strong bias toward a status quo that favors the wealthy. It is true that a lottocracy would favor legislation that would be supported by a majority or super-majority, if they were well-informed. Whatever kind of substantive bias such a system would have, it would be expected that any well-functioning democracy would have it. The solution would not be to change the system, but to engage in public deliberation aimed at changing the laws that it produces.

It is instructive to compare Lafont’s discussion of lottocracy with her discussion of judicial review. The fact that she is willing to delegate an important function of democracy—determining the constitutionality of laws—to the majority vote of a small, unelected body, shows that she does not have a problem in principle with such delegations of power. In the case of the Supreme Court, she believes that she can avoid blind deference and alienation because of the public nature of the proceedings and the reasoning upon which the Court’s decisions are based (chapter 8). I am suggesting that there is a potential for the same kind of defense of empowered minipublics.

In this discussion I have only seriously considered two alternatives, the current system in the U.S. and a system of empowered minipublics. Perhaps we should just be looking at changes to the current system to make it more responsive. However, at least in the U.S., I think the situation is pretty hopeless. In Citizens United v. FEC^{^[6]} the Supreme Court constitutionalized the spending of unlimited amounts of money to influence elections. In two recent cases, Rucho v. Common Cause^{^[7]} and Lamone v. Benisek^{^[8]} the Court constitutionalized partisan gerrymandering. These issues may not be fixable in the U.S. electoral system, where the fact that they have been elected shows that Congressional representatives have themselves benefited from the existing system of campaign contributions and perhaps, also, gerrymandered districts. These issues disappear in a system of empowered minipublics, because the representatives in a deliberative poll are not elected.

These comments are much too brief to settle the issue of empowering minipublics, but I hope they help to keep the conversation going, because I am hoping that someday some democracy somewhere will undertake a limited experiment with lottocracy. I think it would be eye-opening to most people who live in what are today called “democracies” to see an empowered minipublic at work. As I say, it would be must-see TV. I think it would change most people’s conception of what a truly representative form of government might be.