Voting with intensity

Stanford Law Review

Volumn 53, Issue 1, 2000, Pages 111-160

  Levmore, Saul ^a  

^a UNIVERSITY OF CHICAGO   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0346025393     PISSN: 00389765     EISSN: None     Source Type: Journal    
DOI: 10.2307/1229419     Document Type: Article

References (85)

1. See Saul Levmore, Precommitment Politics, 82 VA. L. REV. 567, 581-82 (1996) (suggesting that the Speech and Debate Clause may present one problem with enforcing political precommitments asserted by a member of Congress). I say "probably" unenforceable because a failure to make the payment might result in a claim for fraud. As for the criminality question, note that this is not a payment to a "person" in return for a vote, 18 U.S.C. § 597 (West, WESTLAW through P.L. 106-170), but rather presents a question of corrupt payment or bribe, unless the argument is that it is an indirect payment - but one that is close enough to a direct payment to a person - that offends the statute.
2. Richard L. Hasen, Vote Buying, 88 CAL. L. REV. 1323, 1329-31 (2000). I will use the expression "vote buying" instead of "vote selling" when it seems appropriate to emphasize something about the purchasers. "Selling" draws attention to the fact that sellers, though they are poorer than buyers, might be made better off through sales. The equality argument is perhaps the most common one found in the literature.
3. See, e.g., Daniel R. Ortiz, The Democratic Paradox of Campaign Finance Reform, 50 STAN. L. REV. 893, 897-901 (1998) (analyzing the position that campaign finance reform is necessary to prevent the wealthy from exerting a disproportionate influence on politics).
4. This "efficiency" argument might of course be used to limit campaign contributions as well, but existing law in this regard is taken as given. As far as I know, the argument appears first in Richard A. Epstein, Why Restrain Alienation?, 85 COLUM. L. REV. 970, 987-88 (1985). For a discussion questioning the "efficiency" argument, see note 24 infra.
5. An idea to be credited to Margaret Jane Radin, Market-Inalienability, 100 HARV. L. REV. 1849, 1849-50 (1987).
6. A mixture of the anti-commodification and equality arguments appears in Pamela S. Karlan, Politics By Other Means, 85 VA. L. REV. 1697, 1709-14 (1999).
7. Legalized vote selling is found in some special districts created to manage resources, in corporate law as discussed in the text accompanying notes 41-45 infra, and perhaps in closed residential communities and the like. See Hasen, supra note 2, at 1348, 1364. It may, of course, be found in slightly camouflaged form in legislative (or informal) logrolling and in political campaign expenditures and fundraising. See id. at 1338, 1359; text accompanying notes 71-80 infra.
8. For a discussion on indirect vote buying, see text accompanying notes 33-37 infra.
9. Higher compensation can attract better candidates and encourage agents to perform well in order not to risk losing these highly compensated positions. On the other hand, increased compensation might encourage corrupt practices by those who aspire to these highly compensated posts. I would prefer to label the "efficiency" argument an agency argument because it is about the behavior of the political agent whose honest activities may be inefficient. But I will abide by the prevailing nomenclature.
10. This seems to be the case regarding other possible "markets." For example, if school vouchers turn out to improve public schools and benefit poor students, commodification will surely become more acceptable. For instance, commodification became more attractive in the case of staffing the armed services as it appeared that conscription was itself unequal and otherwise flawed. See generally, William A. Fischel, The Political Economy of Just Compensation: Lessons From the Military Draft for the Takings Issue, 20 HARV. J.L. & PUB. POL'Y 23 (1996);
11. Susan Rose-Ackerman, Inalienability and the Theory of Property Rights, 85 COLUM. L. REV. 931, 964 nn.92-93 (1985). The case against the Civil War era system of drafting people but allowing them to buy replacement labor is surely reinforced by arguments that this sort of commodification might feed back and affect the country's willingness to go to war and the method of engaging in war and so forth. In all these cases, the inclination against (or for) commodification seems driven in large part by expected consequences.
12. Thus, the anti-commodification argument seems neither weaker nor stronger in the presence of rules permitting (or restricting) campaign contributions, despite the fact that monetary contributions commodify political participation.
13. There are, to be sure, many things that are more transferable than alienable. Body organs and babies, for example, are somewhat transferable in the sense that donors' and biological parents' wishes are normally given substantial consideration. See generally Susan Rose-Ackerman, supra note 8 (arguing that efficiency, specialized distributive goals, and responsible democracy justify inalienability rules in certain contexts).
14. See, e.g., CONSCRIPTION: A WORLD SURVEY: COMPULSORY MILITARY SERVICE AND RESISTANCE TO IT (Devi Prasad & Tony Smythe eds., 1968) (surveying more than 70 jurisdictions and finding not one that permits private sales of military service obligations).
15. Thus, there may be externalities when citizenship rights are sold from poor to wealthy persons, but perhaps less so when body organs, babies, and military service responsibilities are sold. Where A is wealthier, A might trade good w for B's good v. But if B were wealthier, we would not see this trade of v for w. B is still better off trading with A, unless a ban on exchanges will somehow lead to wealth redistribution or other relevant changes. Therefore, the argument for inalienability needs to include some external effect on C.
16. The question is whether wealth effects, and the likelihood of missing out on high-valuing users, can explain those things held to be inalienable. Even if we have some fabulous method of distributing commodities to the highest valuing users, we must of course recognize that as we remove commodities from circulation, in a manner of speaking, we reduce the incentive to work and to acquire wealth. If this is the best way to undertake redistribution, there is some optimal degree to it.
17. Wealth redistribution will, of course, generate new allocations and therefore new production levels of most goods. However, the point is that a market at least provides some signals to producers.
18. Note that I use the awkward term "fixed" rather than "inelastic" to describe the output level of such things as citizenship rights and military forces. The idea is not that there is price insensitivity but rather that the level is fixed by some external decisionmaker or constraint.
19. Other inalienable things, such as babies and some body parts, are obviously not fixed in volume, and so the positive and normative arguments must be different. We might focus on the disutility of sensitive observers who do not like seeing certain goods move to wealth-empowered buyers. But this is a slippery slope, and perhaps it is sufficient to say simply that in these areas it is plausible that a central planner could do better in finding a higher valuing user.
20. It is "safer" in the sense of reducing the fear that the university would sometimes behave strategically and create shortages in order to maximize its revenue.
21. Note that the transferability of queue positions might encourage socially wasteful steps on the part of potential sellers. Opportunistic new entrants might wait in queues in order to have spots to sell. But there are complexities to this argument, and I do not pursue them here.
22. Thus, rules against selling children can be explained with reasonable assumptions about the "external" effects on these children, or with various claims about social policy. Yet, there is little doubt that in the universe of prospective adoptive parents, determined, wealthy "buyers" are advantaged, if only because they can employ agents and travel to foreign countries in search of available children.
23. Direct transfers are barred in the case of voting (and military conscription) but not in the other examples. In some cases, the possibility of occasional direct transfers removes pressure on the general rule of inalienability. Thus, if transferability were an all-or-nothing question, there might be overwhelming sentiment for making kidneys completely alienable, rather than forbidding sales and donations altogether. And if the choice were between forbidding every proposed child adoption or allowing all transfers, it might even be that most observers would prefer a free market rather than a freezing of family units with no "transfers" at all.
24. The argument is similar to many other collective action problems. In the corporate context, it has been noted that the collective action problem of shareholders rushing to sell their votes (if permitted to do so) would exert a "strong influence over the market price of votes" that would not ordinarily be entirely offset by competition among bidders. Frank H. Easterbrook & Daniel R. Fischel, Voting in Corporate Law, 26 J.L. & ECON. 395, 411 (1983). This anticipates several of the steps taken here. It is interesting that Easterbrook and Fischel do not say that this collective action problem explains or justifies a ban on vote selling, perhaps because such an approach to collective action problems might then justify some of the very defensive tactics in the face of tender offers that these authors are disinclined to favor. In the corporate context, to be sure, there are market tests as to the wisdom of defensive tactics, but in the voting context it is not as if we can see how the price of a voting right changes with stricter or more relaxed inalienability rules.
25. It is interesting to imagine the legal hurdles that might be placed in front of an attempt to legalize vote selling in general elections (although I do not recommend vote selling in general elections here). Perhaps the Equal Protection Clause would be thought to require equal voting rather than equal ability to vote or sell. Certainly a centralized auction, with no popular election, would run afoul of our constitutional structure if the auction allocated constitutional posts or contained enough legislation to take away the decisionmaking authority of elected officials.
26. As such I do not dwell on such fundamental but familiar questions as how persons who cannot coordinate ever agreed on a precommitment strategy or what to make of positive theories that are unaccompanied by evolutionary tales.
27. This is not the place to discuss the Williams Act and whether it is really a good thing, but it might be useful to observe an important difference between voting in the context of tender offers and in general political elections. In the tender offer context, those who do not sell their shares remain on board to enjoy future corporate performance, to sell in a later transaction, or to litigate if they are somehow excluded. But in elections, those who do not sell votes in an initial round enjoy no special or comparable position later on.
28. The "efficiency" argument is based partly on a variety of collective action problems. The argument is that buyers might work harder to extract benefits sufficient to justify their investment in votes. With no collective action problem, the sellers might contract to guard against this "misbehavior" and then perfectly monitor the elected buyer and enforce their contract. In order to avoid juggling all these collective action problems at once, we might simply think of the efficiency problem as arising out of the sellers' inability to specify all that the buyers may not do. This is a familiar problem in politics where there are serious and even insurmountable legal obstacles to political commitments. See Levmore, supra note 1. One weakness of the efficiency argument is that it imagines the elected official as doing more harm if the official has spent more money to gain power. If the model assumes a selfish actor, it is not obvious why explicit vote buying should make anything worse. A profit-maximizing politician will get away with whatever he can even if it is cheap to be elected. As we will see, the efficiency story is also a bit just-so in character. Where vote selling is in fact permitted, we take for granted that outright theft and other agency problems by those who are elected must be handled by straightforward rules and their enforcement. In the corporate context, for example, a buyer might spend money in order to gain a position from which insider trading or other non-pro-rata extractions are possible. If vote selling is not permitted, then we might say that the law hopes to discourage such buyers, and that more direct enforcement is costly. But if vote selling is permitted, then the claim is that there are advantages to transactions involving voting rights, and that antisocial behavior can and must be dealt with through more straightforward prohibitions.
29. I avoid the question of whether there are more likely to be multiple buyers or multiple, competing groups of organized sellers.
30. The most obvious possibility is to ask for a price that reflects the per-voter expenditures in lively campaigns. It goes almost without saying that we have minimum wage legislation even though there is a comparable valuation problem. Interest groups simply look at market prices, however infected by various collective action problems and other imperfections, and "search" in the direction of prices that are slightly higher than the inherited baseline. This is not a bad strategy. If employment rates suddenly dropped, these same interests could easily obtain a rollback to lower minimum wages. Similarly, minimum vote prices could be set through trial and error.
31. "Low value" is, of course, partly a function of wealth, so that low-valuing shareholders might simply sell first because they attach greater value to the money that they will receive.
32. Another simple possibility is that non-voters are lethargic or, in rationality terms, inclined to place a high value on their own leisure activities. They may therefore attach a high price, or cost, to the steps necessary to vote - but also to those necessary to hear from buyers and sell one's vote.
33. These assumptions or simplifications are fairly reasonable, I suspect, but mostly they serve analytic convenience rather than some particular conclusion.
34. Imagine that A is representative of a majority of the likely voters, L, and that A prefers candidate X over Y by $20. If A would sell his vote, he would want at least $20 (more) before selling the vote to Y rather than to X. A might also be willing to buy votes, but he knows that X's election will benefit many voters, and so A has no intention of spending his own money in the hope that he is the marginal, or swing, vote purchaser. The same can be said of B, C, and D, the other dispersed likely voters who prefer X by $20, so to speak. Meanwhile, E, F, and G form the well-organized minority of L, and they each prefer Y over X exactly as A prefers X over Y. Finally, H and I are unlikely voters; each would vote for X if voting were costless (but it is not). H (and I) would sell her vote to X (or to A) for $3 (or even for $0), but H would only sell to Y (or to the EFG group) for $23. In this example the EFG group can overcome its collective action problem and raise up to $60 to secure a victory for Y. H and I will accept less than that amount for their votes. A, B, C, and D thus prefer a world without vote buying because vote buying allows the organized minority of likely voters to buy votes from unlikely voters. Some unlikely voters, like H and I, and the organized minority of likely voters, including E, are made better off by these sales, but those who are in A's (unorganized) group are worse off. Their votes will lose value if vote buying takes place. The example supports the intuition that many citizens are worse off if organized interests have more power - and that vote buying is something organized interests are likely to do and not something most dispersed citizens will do. Slightly less obvious is that sellers of votes (drawn from U) can be made better off as these interest groups are more powerful. But the point in the text is that although the majority of L may not be well enough organized to raise money for vote buying, this majority of L may be sufficiently organized and perceptive to pass rules against vote buying. An alternative argument might focus on politicians as a group. These actors can make promises but they might be better off if they cannot buy votes because their line of work might then be more expensive. The argument essentially turns the agency explanation discussed above on its head. See text accompanying note 6 supra. One problem with this explanation is that these same politicians should prefer to be able to make binding promises to the electorate and, as already noted, they cannot.
35. These collective action stories point to important differences between voting rights and most other goods. For voters who wish to be part of winning coalitions (and surely for those who suspend narrowly-defined rationality and work to influence voting outcomes) voting rights are different because their value drops as other players exercise their own voting rights. These voters might oppose a market in voting rights much as any producer might try to enlist the government as an ally in reducing competition. Even voters who are accustomed to losing in politics might oppose a market in votes, because they can see that other interests will enjoy yet greater advantages in organizing and buying votes, and then in exercising political power. And for voters who gain some sort of consumption value from the act of voting, it is likely that little additional value would be garnered from multiple voting opportunities. These voters appreciate the presence of markets for standard commodities, where there is reason to increase consumption from time to time, but they see no comparable benefit in a market for voting rights.
36. This is of course a speculative proposition both because we do not know what free market prices would look like and because the number of truly swing voters is contestable. The important point is that the amounts are not trivial. Candidates for a U.S. Senate seat can easily spend $10 per voter in an election and obviously more than that if we divide only by the number of voters a candidate actually attracts or needs to attract, or if we add in expenditures related to party primaries. Moreover, if we think of no more than one-third of the likely voters as actually open to suggestion, the price (or cost) per vote seems quite significant. Prices might be higher if enough sellers were organized or enough buyers present as to solve the collective action problem. It goes almost without saying that there are many differences between what I have called "indirect" and "direct" vote buying, and these distinctions must mean a good deal to those who are comfortable with the status quo. But this is generally the case with arguments about near-substitutes. In any event, I do not mean to imply that it is inconsistent to favor a ban on direct trades but to allow expenditures and promises that might produce equivalent results. To the contrary, my argument, straddling the already thin line between the positive and the normative, leads to the idea that it is direct vote buying that runs the greater risk of producing perverse results and that this distinction might "explain" the conventional legal rules.
37. See Samuel Issacharoff, Pamela S. Karlan, & Richard H. Pildes, THE LAW OF DEMOCRACY: LEGAL STRUCTURE OF THE POLITICAL PROCESS 618-64 (1998) (presenting cases and comments on the Supreme Court's view that sometimes "money is speech" but that some regulation of contributions, as well as spending caps in return for public financing, is tolerable).
38. It would not be terribly surprising if cross-sales were permitted because endowments are somewhat evened out by giving everyone points to sell or use. A student who might not buy any of the contested resources in a pure auction system (in order to make attendance at the university as inexpensive as possible) might, by virtue of the initial distribution of course enrollment rights (or parking lot spaces or dormitory rooms), now keep some of this currency in order to compete for a scarce resource. It is also possible that a university would allow transferability in order to encourage some utility-enhancing exchanges, but discourage accompanying side payments in order to minimize the effect of wealth. If such in-kind trades are prohibited, it is likely to be because the university fears its inability to enforce the ban on outside currency.
39. Some of the analysis changes if voting takes place over a period of days or weeks rather than on a single election day.
40. The "best price rule" is found in 15 U.S.C. § 78n (d)(7) (West, WESTLAW through P.L. 106-170).
41. From the equality and efficiency perspectives, if wealthier people are more likely to vote than are their poorer counterparts, then it should be politically correct (as it is) and perhaps economically correct to welcome payments or other policies that encourage voter turnout. And so long as Candidate X cannot target expenditures to attract those members of U who are more likely to vote for X, X's efforts to turn out the vote should be welcome - though a self-interested X might then not invest in voter turnout. From the efficiency perspective, if the fear is that, once elected, X will do more harm the more X needed to spend to acquire office, then we should have no objection to X's expenditures where X is subject to a binding cap on expenditures. Moreover, the greater the turnout the harder it is for X to know to whom debts are owed. But as already suggested, it is very hard to imagine a practical vote-encouraging plan that would not target some voters rather than others, and perhaps even harder to imagine why a candidate would fund a neutral vote-encouraging campaign. Candidates will advertise their encouragements in those places where they think it more likely that they will attract supporters (or not at all). Note that under the view advanced in the text we should not expect members of L to work hard to forbid encouragements to U. Members of L may prefer not to share the attention they presently enjoy, but the marginal impact of higher turnout on each member of L is negligible. Of course, the entire matter may be academic because, apart from compulsory voting, it may be thought impossible to avoid or disallow the targeting of encouragements. In this case, everyone's favorite argument against vote selling becomes a suitable argument against vote encouraging.
42. See Levmore, supra note 1, at 585-91 (noting that vote selling that permits third parties to become power brokers is similar to the role played by irrevocable proxies).
43. The effects will depend on how low the threshold is set for gaining a seat.
44. See, e.g., DEL. CODE ANN. tit. 8, § 212(e) (1991) ("A . . . proxy shall be irrevocable if . . . it is coupled with an interest sufficient in law to support an irrevocable power.").
45. See Easterbrook & Fischel, supra note 21.
46. See Schreiber v. Camey, 447 A.2d 17, 26 (Del. Ch. 1982) (holding that loan to controlling shareholder arguably in return for shareholder's vote on a merger found inoffensive because the agreement was meant to benefit all shareholders and the deal had been worked out by disinterested directors with full disclosure). Schreiber has fared reasonably well in subsequent opinions. See IXC Communications, Inc. v. Cincinnati Bell, Inc., No. C.A. 17324, C.A. 17334, 1999 Del. Ch. LEXIS 210, at *22 (Del. Ch. Oct. 27, 1999) (citing Schreiber for proposition that vote-buying agreements are not illegal per se, but rather voidable only upon a showing that the agreement defrauds or disenfranchises the shareholders); Haft v. Hart, 671 A.2d 413, 420-21 (Del. Ch. 1995) (upholding contract granting irrevocable voting proxy); Weinberger v. Bankston, Fed. Sec. L. Rep. (CCH) ¶ 93,539 (Del. Ch. Nov. 19, 1987) (citing Schreiber for rule that a vote-buying transaction will be upheld if it is intrinsically fair and its object or purpose is not to defraud or disenfranchise stockholders); Kass v. Eastern Air Lines, Inc., 1986 WL 13008, at *3 (Del. Ch. Nov. 14, 1986) (quoting Schreiber in holding that transfer by corporate shareholder of stock voting rights without ownership is "not necessarily illegal" and "must be examined in light of its object and purpose"); Wincorp Realty Inv., Inc. v. Goodtab, Inc., No. 7314, 1983 WL 8948, at *4-5 (Del. Ch. Oct. 13, 1983) (citing Schreiber in stating that a vote-buying agreement is not per se illegal).
47. See note 41 supra and accompanying text.
48. The idea is that minority shareholders await conflicts among major shareholders in order to become part of ruling coalitions. See generally Zidell v. Zidell, Inc., 560 P.2d 1091 (Or. 1977) (holding that specific minority shareholders not entitled to special protection when stock sale shifted controlling majority stock ownership to another stockholder).
49. The single-subject rule, found in many state constitutions, is often liberally construed to favor legislative deals. See, e.g., Steadman v. Hindman, 992 P.2d 27 (Colo. 2000) (holding that the initiative did not contravene the state constitutional single-subject requirement); Masters v. Commissioner, 604 N.W.2d 134 (Minn. Ct. App. 2000) (holding that the statute did not violate the single-subject requirement); Bernstein v. Commissioner of Pub. Safety, 351 N.W.2d 24 (Minn. Ct. App. 1984) (determining that statute does not violate prohibition against laws embracing more than one subject). But the idea is that construction is most liberal, or that various provisions are found most germane to one another, in the case of a budget bill where vote trading is most explicit. Put differently, the common requirement that initiatives be single-subject in scope can be explained as anti-logrolling precisely because it is here that the drafters (of the initiative) have a substantial, competition-reducing, agenda-setting advantage, if only because it is difficult to get a vote on a competing proposal in short order. E.g., CAL. CONST. art. II, § 8(d) ("An initiative measure embracing more than one subject may not be submitted to the electors or have any effect.").
50. On the other hand, we do not readily permit legislators to use "outside" money to trade for votes in the legislature. See generally Levmore, supra note 1, at 586-87 & n.41 ("At the very least, payments made directly to promise-keeping representatives would quickly run afoul of the legislated limits on supplemental income."). Similar constraints, ranging from anti-bribery statutes to campaign finance laws to supplemental income restrictions, apply to state legislators. Id. This suggests either that there is more to the ban than collective action problems or that it is understood that outside money would come from dispersed citizens or interests so that the collective action problem would reappear.
51. I have suggested but hardly proved that poor people would actually be better off with legalized vote selling. Moreover, there is always the response that wealth redistribution can be done more directly. In the present context this is something of an ironic reaction because it is an argument that might have been raised in the first place against the equality argument. Yet, as every public finance student knows, we would be better off redistributing wealth through an income tax and permitting vote selling than regulating bargains in the name of wealth redistribution. Louis Kaplow & Steven Shavell, Why the Legal System Is Less Efficient Than the Income Tax in Redistributing Income, 23 J. LEGAL STUD. 667 (1994).
52. Put differently, the explanations or justifications for the ban on vote selling are a bit just-so in character. If vote selling were legalized, while other things remained inalienable, it would be possible to "explain" the legalization with reference to intensities, a new view of wealth effects, and so forth.
53. I stress the likelihood of unequal and intense preferences because intense preferences alone are unproblematic. A may intensely wish to cause a redistribution from B to A, while B hopes to do the opposite. We do not wish to empower such preferences (and they are zero sum) but we might want to give weight to A's intense interest in policy toward Cuba or animal rights, while B has opposite but much less intense preferences on these matters. It is unequal, intense preferences that are most interesting.
54. I use this weak language because private voters who favor X over Y are free to vote for Y if they sense that other citizens intensely prefer X.
55. See, e.g., DENNIS C. MUELLER, CONSTITUTIONAL DEMOCRACY 158-59 (1996) (discussing Condorcet's 1785 work).
56. Id. at 152 (quoting E.B. White).
57. One way to think about deliberation among jurors, legislators, and voters quite generally is as an opportunity for participants to evaluate what they hear in order to decide who is "expert" and how much to change one's vote in light of this expert opinion. But it is noteworthy that although jurors and legislators are virtually required to attend deliberations, voters in large-scale elections are not.
58. See MUELLER, supra note 52, at 159-71 (discussing the incorporation of the equal intensity idea in Condorcet's Theorem and elsewhere).
59. Social practice with regard to these sorts of tastes is interesting. Few people become vegetarians in order to satisfy others' intense preferences. On the other hand, we would not be surprised to find an informed Q and R altering their eating habits in vegetarian S's presence. Perhaps it is easier for them to believe that the sight of meat-eating will upset S more than the act pleases themselves, but harder to assess S's claim about the mere knowledge of a few more or a few less meat-eaters. This is also a good setting for a kind of altruism. If S tried to engineer a vote for a legal intervention forbidding Q and R to eat as they pleased, Q and R might well react with a strong libertarian preference of their own. Another explanation for our (or Q and R's) failure to bend to the will of the passionate vegetarian minority is that while we (or the passionate minority) attach great value to passionate vegetarianism, S might actually prefer for Q and R to have the chance to come around to the vegetarian perspective on their own. If Q and R simply behave like S because of S's intense preferences, they remove this possibility of genuine conversion. If, for example, T is a fervently religious person and P is not, it would be easy to imagine that while T hoped for P's devotion, T would not want P to act as does T simply because T intensely wishes it so.
60. And we will value it more, the more likely it is that non-experts are wrong and the more expensive it is to bring in (mostly right) non-experts. The theorem does not quite hold where some identifiable voters are better informed than others. If there is one expert with a 9/10 chance of being right and two non-experts each with a 2/3 chance of being right, we are more likely to get the matter "right" asking the one expert than we are by relying on a majority vote of the group of three. One in ten times the expert will be wrong, but only 4/9 of the time will the non-experts save the day (because we need them both to get it right (2/3 × 2/3) in order to outvote the incorrect expert. Meanwhile 9/10 of the time the expert had it right, but there is a 1/9 chance that the two non-experts will outvote the expert and ruin a good thing. Relying on the known expert gets it right 90% of the time; relying on a majority of the three gets it right about 85% of the time. Had we abided by a majority vote of a pool consisting of our expert, with a 90% chance of correctness, and one million non-experts, each of whom had a 2/3 chance of being right, we could indeed improve things from 90% to more than 99%.
61. The bargain will be required by Q if Q is selfish and not interested in maximizing social utility alone or if Q despairs of comparing his preferences with S's. Note that the discussion here assumes that there are not many voters with mild preferences to offset S's intensity.
62. There is the familiar idea of being impressed by an intense opinion emanating from someone who rarely has strong views.
63. It is therefore easy to see why we do not approve of vote buying among members of a jury. First, jurors do not internalize the social interest in adjudication or accurate factfinding. The social interest in deciding these things has little to do with how much a juror would require to vote one way or the other. If a party could pay all citizens to drop criminal charges, then there might be reason to value such a bargain, but the problem with paying a juror is that we exacerbate the agency costs inherent in the situation. There is perhaps a slightly more plausible case to be made for the proposition that we allow jurors to buy one another's votes, but the Condorcet Jury Theorem tells us why that is unwise if we seek to maximize the chance of reaching the "right" decisions. It is also easy to see why we do not want judges on a panel to sell votes to one another. The judges are our agents, and it is not their preferences, or even their intense ones, the system wishes to satisfy. If the judges were somehow representative of constituents with diverse intensities, as are elected politicians perhaps, then the case for vote trading or even buying (with constituents' funds) first becomes plausible. A judge might, of course, defer to another because of the latter's perceived expertise. In this case, the Condorcet Jury Theorem (and its assumptions) might help the judge decide when to defer. In any event, there is no precommitment to defer, except in the case of specialized courts, and we can think of each judge as weighing arguments including those in favor of deference to expertise.
64. See MUELLER, supra note 52, at 158-59 (discussing the idea that proportional representation is not about finding the "right" answer and that the choice of a "voting rule for a multiparty parliament must be made on the basis of a different set of assumptions from those Condorcet used to justify . . . [it] for jury decisions").
65. See Hasen, supra note 2, at 1338-48 (discussing examples of beneficial logrolling, inefficient logrolling, paradoxical logrolling, and equality-affecting logrolling); Saul Levmore, Voting Paradoxes and Interest Groups, 28 J. LEGAL STUD. 259, 263-68 & n.6 (1999) (discussing aggregation paradoxes and the possibility of resorting to single-subject restrictions).
66. Hasen cites a number of sources that make the familiar point that logrolling can be efficiency enhancing by allowing decisions to reflect relatively intense preferences, but there is no way to know whether, on the whole, logrolling does more good than harm. Hasen, supra note 2, at 1343. I have already suggested that legal rules about logrolling in legislatures and elsewhere may reflect intuitions about the presence of competitive markets, which would further reduce any apprehension about collective action problems.
67. Note that the organizer of these shadow markets has no particular interest in encouraging work effort, so he loses nothing by removing the scarce good in question from the universe of those that can be acquired through work. A central planner might be more cautious about handling wealth effects so bluntly. Note also that while these schemes can hardly guarantee allocations to the highest valuing users because participants may be unable to guess at market clearing prices (and there is a ban on post-auction trades in order to eliminate wealth effects), one suspects that they are used more in universities in the United States than in countries such as Cuba, where market mechanisms are less familiar and less admired (and yet where wealth effects are viewed with even more hostility). Queues may also work as market substitutes and in some circumstances they may even do a better job at allocating goods to those with the most intense preferences. Queues may eliminate wealth effects but they are also costly in that they take people away from more productive endeavors. When the social norms associated with queues allow those who acquire goods through queue-waiting to turn around and sell these goods, there is an obvious compromise between concerns about disparate wealth and about eliminating desirable trades. But this is obviously not the place to explore queues.
68. STV, as used in Australia and elsewhere to record ordered preferences and to credit "excess" votes to voters' subsequent choices so as not to "waste" votes, is discussed in Saul Levmore, Parliamentary Law, Majority Decisionmaking, and the Voting Paradox, 75 VA. L. REV. 971, 1037 n.171 (1989) (hereinafter Parliamentary Law). Another tempting solution to the problem of wasted votes is to deploy an intermediary who could gather points and information about preferences. But there is the obvious question of how this intermediary will discern intense preferences held by constituents and how strategic communication can be avoided.
69. An STV-style procedure could avoid wasted votes in one time period, but this does not solve the individual voter's problem of how many votes to put on the table in the early election period.
70. See note 78 infra and accompanying text.
71. See Parliamentary Law, supra note 65, at 991-96 (suggesting that rules evolve in response to dissatisfied majorities recognizing that prevailing rules had produced results that this majority could defeat).
72. It is of course true that markets can produce similarly perverse results because those who save for the future might have valued present consumption more than those who consume today. The result can be perverse because participants do not know future prices, incomes, and so forth. But, at least in most markets, participants do not know about this result. They have no reason to think that those who choose to consume today are not in fact higher-valuing users. A simpler difference may be that we somehow expect political elections to be decided by majorities in every single time period. Note that some voting innovations, such as cumulative voting in corporate law, do not go so far as to run the risk of "dissatisfaction," or perverse results, unless the majority simply (and grossly) miscalculates.
73. Another possibility is to get at intensities in the financing arena. Wealth effects could be mitigated by giving each potential voter a (shadow) fund to contribute to political candidates. Voting would still be one-person-one-vote-per-election, but government funds would be allocated by individual decisionmaking. This system has promise, but needs to overcome various corruption and strategic behavior problems.
74. Note that if many L members switch from voting to selling in order to gain income, we will not do well in terms of revealing intense preferences.
75. In the faculty vote context, I suggested that disclosure was wise. See text accompanying note 59 supra. The situation here is a bit different.
76. Thus, D might feign opposition in order to extract a payment from B. There is literature on "demand-revealing" processes that suggests the value of a centralized process which asks everyone to reveal preferences, such as how much a new project is worth to a responding member, and then to abide by the aggregated results. See generally, T. Nicolaus Tideman & Gordon Tullock, A New and Superior Process for Making Social Choices, 84 J. POL. ECON. 1145 (1976) (explaining a process for selecting public goods in which each individual may change the outcome which occurs without her vote by paying a special charge equal to the net cost to others of including her vote). A problem with this approach in many real settings is that some voters have superior information about others' preferences.
77. I continue to imagine that preferences are arrayed from A to K, with A most in favor, K most opposed, and F in the middle, leaning toward K unless paid to do otherwise. The actors themselves do not have full information. Thus, B bids for a swing voter, searching for F and unaware that A might contribute (even more than B himself) to this cause or, indeed, that B might be able to free ride on intervention by A. It is thus easy to see that a rule permitting vote sales could eventually do the job; the more information the participants had about one another's preferences, the more we could expect of the market. But it is also quite plausible that a market will not produce the result that maximizes utility.
78. For example, if F requires her share of the project to be $100 less expensive in order to favor it, then F will require a $100 transfer or an $1100 payment to the central account. If B is only willing to pay between $100 and $1100, and if (an important "if") the other opponents, G through K, could be bought off by A and C through E, then we might miss some bargains that majority voting in this context seems inclined to welcome.
79. I have found no literature on this aspect of the rules of residential associations, and casual empirical inquiries have not produced anything but uncertain reactions. Perhaps this is true of many things that are fixed in output and that can be influenced by wealth disparities. Thus, students who win a lottery and secure parking privileges are often uncertain as to whether they would violate a rule or norm by selling the privilege to another student.
80. The problems might be mitigated by a system that allowed vote sales or some sort of auction only when there is a close election of the normal sort. The forty-eight percent can defeat the fifty-two percent by first showing that a standard election is fairly close, say within five points, and then by paying enough to induce swing votes in a second round. We might even handicap the minority in the second round. A disadvantage of this scheme is that resources will be wasted on indirect vote buying in the first round - in order to avoid the threat posed by the five-point rule. Advantages of the five-point rule include the likelihood of limiting wealth effects, diminishing the efficiency or agency problem (if only because the system will produce someone that might well have won a standard election), and promoting some sense of legitimacy. The reference to fundraising in the text suggests the question of why we encourage charitable organizations in some arenas.
81. I promised to defer the perverse results theme to another day, but it should be noted here that if voting law were entirely about avoiding perverse results, then we might expect compulsory voting because uneven voter turnout might lead to odd results. Similarly, uneven fundraising can lead to anti-majoritarian results.
82. I ignore different tax circumstances and other factors that might lead to clientele effects but also to periodic adjustments in corporate decisionmaking.
83. One way to think about the matter is that the advantages suggested by the Condorcet Jury Theorem are retained if we give each share a vote; we have no reason to think that those who now have more voting rights are less likely to be right than those with fewer shares and voting rights. At the same time, there may be settings in which those with more shares find it worthwhile to gather information. As such, when the straightforward theorem does not hold, because some (well-informed) identifiable voters are more likely to be right than others, we are well-positioned to make the right decision because we have assigned more votes to those shareholders.
84. There is the possibility that the voter thinks that G will win but behave better in office if G perceives a smaller margin of victory. Inasmuch as this voter will think the same about signals in most market systems, I will ignore this motive. Similarly, I ignore strategic voting in three-way elections.
85. Of course, our current political system offers this sort of danger when there are three or more candidates, because it is easy for a Condorcet winner to lose in our typical plurality voting systems. I suspect that if it turned out that C would defeat A and also B in pairwise competition, but that A emerged as president, there would be a significant move to change our voting rules. Such realities have arguably caused changes in our parties' primary election rules. The point in the text is that if strategic behavior appears to decide an election, there would be significant backlash against the limited market rules that spawned this behavior.