The paradox of voting and the ethics of political representation

Philosophy and Public Affairs

Volumn 38, Issue 3, 2010, Pages 272-306

  Guerrero, Alexander A ^a  

^a NEW YORK UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 77954812860     PISSN: 00483915     EISSN: 10884963     Source Type: Journal    
DOI: 10.1111/j.1088-4963.2010.01188.x     Document Type: Article

References (25)

1. The probability of being decisive was initially suggested as being 10-8 but has been put at 10-90 by some more recent work for U.S. presidential elections. See, respectively, William Riker and Peter Ordeshook, "A Theory of the Calculus of Voting," American Political Science Review 62 (1968): 25-42; Alan Carling, "The Paradox of Voting and the Theory of Social Evolution," in Preferences, Institutions and Rational Choice, ed. Keith Dowding and Desmond King (Oxford: Clarendon Press, 1995), pp. 20-43.
2. See, e.g., Anthony Downs, An Economic Theory of Democracy (New York: Harper & Row, 1957), p. 274.
3. See, e.g., Stanley Benn, "The Problematic Rationality of Political Participation," in Philosophy, Politics, and Society: Fifth Series, ed. Peter Laslett and James Fishkin (New Haven, Conn.: Yale University Press, 1979).
4. See Alvin Goldman, "A Causal Responsibility Approach to Voting," Social Philosophy and Policy 16 (1999): 201-17, at p. 217; Richard Tuck, Free Riding (Cambridge, Mass.: Harvard University Press, 2008), pp. 50-62. These accounts highlight that one may cause an effect even though the effect does not counterfactually depend on what one does.
5. 5.See John Ferejohn and Morris Fiorina, "The Paradox of Not Voting: A Decision Theoretic Analysis," American Political Science Review68 (1974): 526-35.
6. 6.See Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), pp. 73-75.
7. I find these responses unsatisfactory, but it is worth noting that many of them are compatible with the response I offer - there might be a number of different but compatible reasons why voting is rational. Although there is not room to discuss these other responses fully, let me say something about where I think they run into trouble. Those responses that reject P1 but highlight the importance of other kinds of desires make voting come out as rational, but at the cost of making the reasons one might have for voting seem normatively insignificant. Additionally, these accounts require that one have certain beliefs or desires in order for voting to be rational, but these are not beliefs and desires that we endorse as grounding the decision to vote: desires to 'cheer' for a particular candidate, to avoid social stigma, et cetera. Other views that reject P1 require that we take an interest in being able to claim partial causal credit for bringing about an outcome, even though that outcome would have obtained whether we voted or not. Those responses that reject P1 but argue that voting is rational because voting is morally required run into a different kind of problem. Namely, the first premise of the argument - that voting is morally required - is arguably false. Those responses that reject P2 require that we adopt an implausible conception of rationality, or that we make arguably implausible empirical assumptions about the expected utility that flows from the chance that our vote might be pivotal. There is more to be said about these accounts, certainly, but we might want an account of the rationality of voting that did not require taking on these various commitments.
8. Why just adults? Why not the mentally incompetent? What about requirements of citizenship and other conditions of voter eligibility? In general, the view I advocate casts a suspicious eye on all non-enfranchisement for the simple reason that all who will be subject to the laws have an interest in, if not a right to, participation in the authorization of the lawmakers. That said, there will be some legitimate limits imposed on who counts as a member of the political community. Here, I will leave aside discussion of these important boundary-setting questions.
9. Hanna Fenichel Pitkin, The Concept of Representation (Berkeley: University of California Press, 1967), p. 145.
10. 1Ibid., pp. 145-67.
11. Could we have a non-electoral measure of a candidate's normative mandate, say, by the use of sophisticated polling? Possibly, but there are concerns worth noticing. First, the procedural safeguards necessary to make a poll an adequate measure will typically make polls come to resemble elections, different only in that they use sampling techniques to achieve their results. Second, whatever the sampling technique used, there would likely be objections both to the particular technique and to the idea of a 'democracy' in which every person interested in being counted would not actually be counted.
12. The world contains a remarkable diversity of systems of electoral representation: systems in which multiple people represent a single district; where voters can vote for more than one candidate or can rank all the candidates; where voters vote not for candidates but for political parties or lists of individuals; and so on. I focus on SMDP systems because they remain dominant in the United States and in use in many places. There are significant questions about whether and how the account I offer might extend to other electoral systems. This should hardly be surprising. After all, it is plausible that the answer to the questions of whether it is rational to vote, and, if so, why, might well depend on the details of the system in which one is voting. Because I think that a primary purpose of elections is to help measure the extent to which candidates are actually supported and authorized to govern, and because all systems of representation involve the transfer of political power from the many to some particular few (a few who will then have the ability to make decisions on behalf of the many), I do think that some extension of my account will be possible for most systems that employ elected representatives. But I will not pursue that project here.
13. The word 'mandate' already has a use in elections, a use that I intend to piggyback on, but which should be kept distinct. 'Mandates' in this other sense have several features which make them different from normative mandates: (1) they concern claims about specific instructions from voters to candidates regarding particular issues; (2) they are a function of the margin of victory measured in terms of the percentage of actual voters who voted for them, not measured as a function of support from all mentally competent adults; and (3) they are often described in threshold terms, so that a candidate either does or does not have a mandate. These three features are not features of normative mandates.
14. The precise nature of these norms is complicated; in particular, the way in which constituents' preferences and beliefs should be aggregated is controversial. One might defer to a simple majority, to the majority of those affected, to those with the most intense preferences, and so on. I will ignore these complications here.
15. This is one strain of the norms of guardianship - that which focuses on the interests of one's constituents. Another strain focuses on acting as a guardian for the interests of the nation as a whole. See, e.g., Edmund Burke, Reflections on the Revolution in France (London: Penguin Books, 1968), p. 115. For simplicity's sake, I will consider norms of guardianship focusing on the interests of one's constituents, though much of what I say applies as well to norms that direct representatives to act as a guardian of the nation's interests.
16. 16.Something like this approach is apparent in Pitkin, The Concept of Representation, pp. 150-67.
17. For alternative presentations of the representative's normative predicament (including the complexities mentioned in footnotes 14 and 15, above), see Andrew Rehfeld, "Representation Rethought: On Trustees, Delegates, and Gyroscopes in the Study of Political Representation and Democracy," American Political Science Review 103 (2009): 214-30; Jane Mansbridge, "Rethinking Representation," American Political Science Review 97 (2003): 515-28.
18. For relevant discussion, see Dennis F. Thompson, Political Ethics and Public Office (Cambridge, Mass.: Harvard University Press, 1987), pp. 96-104.
19. There is an interesting kind of case in which a candidate runs on a platform of pure deference, making explicit that, if elected, she will act as a simple delegate. People might support such a candidate without thereby intending to defer to that person's decision making, and it might seem inappropriate for such a candidate to act as a trustee even if the candidate has a strong MNM. Though somewhat contrived, such a case is certainly possible. I think the correct thing to say in response is that this is an instance where omitting discussion of norms of fidelity may distort the picture. Even representatives 'licensed' to act as trustees should take into account norms of deference when deciding what to do. But representatives who have pledged to act as delegates also have norms of fidelity inclining them in this direction, and that does put more pressure on them to defer to their constituents in cases of disagreement. These cases are complicated, however, and worth further discussion.
20. One might argue, based on the Condorcet jury theorem (CJT), that this is backwards: it is better, on epistemic grounds, for a larger group to make decisions and, failing that, for the members of the smaller group to defer to the views of the larger group (to act as delegates) when making decisions. For relevant discussion, see Adrian Vermeule, Law and the Limits of Reason (Oxford: Oxford University Press, 2009), pp. 50-53. This might be correct, if the CJT conditions of competence (members of the group each have a probability p > 0.5 of picking the correct option) and independence were satisfied by the larger group. Here, I only suggest that in practice, for the actual complex policy decisions that have to be made, these conditions are not satisfied.
21. Someone might suggest that interest groups of various sorts might appropriately prefer that their representative not act like a trustee, even if they support their representative. Instead, they might prefer that their representative act like a delegate, but a delegate who defers just to their interests and preferences, rather than to the interests and preferences of all the people in the jurisdiction. A variant of this suggestion is that elections are 'winner take all', and so what one wants from a candidate one supports is not for her to act as a trustee, but for her to act as a delegate who only takes the preferences of her supporters into account. Both suggestions run afoul of plausible moral requirements on the conduct of political representatives. The reason that the preferences and interests of those who supported the winning candidate often and permissibly 'win out' in determining policy is not that the preferences and interests of those who backed the loser can be permissibly ignored - it is because there will typically be more people who have those preferences and interests. Some of the details here will depend on the particular way in which preferences and conflicting preferences are to be aggregated - a question I leave aside. What I want to stress, however, is that it is not morally permissible for a representative to simply ignore the preferences and interests of the losers (or those who are not members of the politically successful interest groups).
22. It is rational to vote, now, because by voting one can alter the moral situation in a direction that one believes to be good (by making it morally permissible for the candidate to act as a trustee). If candidates respond to their MNMs as I argue they should, then voting is rational both because it alters the moral situation in a good direction and because of the benefits that attend having a candidate one supports act as a trustee rather than a delegate.
23. there have been a few technical discussions of how concern about the margin of victory, and the information it communicates to candidates, might be related to the instrumental rationality of voting. see micael castanheira, "victory margins and the paradox of voting," european journal of political economy 19 (2003): 817-41; ronny razin, "signaling and election motivations in a voting model with common values and responsive candidates," econometrica 71 (2003): 1083-1119; george stigler, "economic competition and political competition," public choice 13 (1972): 91-106, esp. pp. 103-4.
24. My focus is limited to systems of electoral representation in which representatives, once elected, are able to exercise independent judgment in governing. One reason for this focus is that having an MNM is a property that can be had only by those entities that it might make sense to support, in the sense of being content to authorize them to govern. A piece of legislation put to a referendum vote can be more or less popular, but cannot have an MNM, because it does not make sense for individuals to support a piece of legislation in the sense of being content to authorize that piece of legislation to govern. In certain electoral systems, particularly those referred to as 'strong' party systems, representatives may have little or no ability to exercise their independent judgment in governing. I will not here discuss whether or how the account might extend to 'representative' systems of this kind, though there is no question that the ethics of representation are considerably altered if representatives are dramatically constrained - through formal or informal mechanisms - in what they can actually do. One possibility is that a version of the account might apply, but at the level of the behavior of political parties (as group agents), rather than at the level of the behavior of individual representatives.
25. Throughout, I will be less concerned with the empirical question of how voters actually behave, and more concerned with the normative question of how voters might sensibly behave in various situations. If voters actually behave in a way that is not sensible, this is a reason for education of a certain kind, not for criticizing the account I've offered (unless there is some further reason for thinking that education will necessarily be ineffectual).