Ethics of Strategic Voting in Popular Elections

1. Role of elections in a democracy

Popular elections are the condition for every representative democracy, hence acknow ledging the practices which govern such elections is essential to understand the modern democratic state.²There are some who would argue that a fair process of voting is not in any way the essence of democracy.³ According to their approach, democracy should be associated with numerous institutions that counterbalance any discretionary powers of public authorities out of consideration for the fundamental individual rights and liberties. Fair elections are then not the only condition of democracy. It is what we would call the liberal democracy approach. Yet, the advocates of this approach do not underestimate the act of voting. Increase of voters’ turnout is supported by “get-out-the-vote” actions or enforcement of compulsory voting, for instance in Australia. Moreover, polling days are generally considered a public holiday by national authorities around the world to encourage citizens to participate in the elections. Therefore, regardless of the particular variant of democracy, be it ancient Athenian, liberal, populist, popular elections stand as an important measure of legitimization, which supports the moral significance of a polling day.

This paper offers a comprehensive discussion of the issue of ethics of strategic voting in popular election with overview through the typical mathematical deliberations in quantitative reasoning of strategical voting. However, the following chapters require no background in mathematics and so might be particularly interesting for liberal art students or those who seek an instant research insight into this area. Although some experience with global political system is assumed, the plain language helps to follow the argumentation with even a little knowledge in the field. This paper is meant to scrutinize the moral underpinning of strategic voting in the very particular case of popular elections defined as any means of indirect democracy, i.e., voting for a representative or for a body of representatives. It should be considered

a narrowing down of the subject since it appears that strategic voting may be attrib uted to any process of combining individual inputs with a collective output, where the decisive body requires no less than two agents, who may choose between no less than three options.⁴ For example, let us suppose that a board of a successful company is made of five members who jointly choose between three drafts of an annual budget for the upcoming year. For the sake of clarity of this example, let me illustrate the set of three proposals: X = {x, y, z} to be chosen by a body of i, j, k, l and m and arranged in the Table from the most to the least preferable choice. Note that the model does not consider any intensiveness of preference. The board member j may be excited to cast a vote for a well-calculated budget x over y, yet honestly hates the z variant. Despite that, the degree of discrimination between these items stays the same. For the sake of simpli city and clarity, let us assume that a simple majority rule decides which budget will be implemented and each board member marks only the top‑ranked option in a ballot. This is what we call the plurality rule. Any of the variants with the highest number of votes wins, no matter of the percentage it would score.

If each of the board members claimed their actual preferences, the result would be a draw between z and y. The outcome z is, however, j’s last option. Then, keeping in mind the preferences of other board members, j forecasts a plausible outcome as not optimal for his or her goals and decides to misreport his or her true preferences out of a desire for a more favourable end-result. Assuming that j submits y instead of x as his or her highest-ranked option, then Rj’s order would be: y>x>z. Assuming that Ri, Rk, Rl and Rm remain the same, the y budget draft wins. Clearly, j exercised strategic voting. This is the most popular strategy that can be noticed across a plurality voting sys tem, and it is often called “compromising” or “voting for the lesser evil”. However, this strategy does not cover a wide array of tactical voting methods. Such strategic voting, alternatively known as tactical voting, is commonly found to be controversial on account of the deceptive behaviour employed. Driven by the intention to enhance plausibility of certain alternative to be elected, a strategic voter takes advantage of the group decision-making process, which any election certainly is, casting a vote that does not represent their top preferences, which raises the issues of accountability, transparency and manipulation. The idea behind the thesis is to scru tinize the moral underpinning of strategic voting, which refers to the voter who casts a vote that does not represent one’s top electoral preference out of a desire to obtain

a more favourable outcome. The intention of the author is to argue that tactical voting need not be considered a moral wrongdoing, if implemented in popular elections.

Firstly, some light is shed on the social choice theory and a theory of democracy with a view to offer some context for a better understanding the place of this thesis in a philo sophical discourse. Then, the consequentialist argument is introduced, followed by the express value argument and the sincere argument, which suggest manipulation since the agent does not reveal one’s profound choice, so to say, does not vote naively. I regard it as non-conclusive, whereas it presupposes a not commonly accepted view on the role of election in the democracy itself. Further on, I challenge five of Satterthwaite’s “transparency arguments”: 1) inequality of skills; 2) inefficiency; 3) non-transparency of voters’ preferences; 4) non-transparency of representatives’ preferences; and 5) ran domness, which I treat by and large as valid, with minor comments. However, I believe that few of the “transparency arguments” can be adopted as a virtue, rather than a vice, of democracy because they encourage cooperation and adjusting ongoing coalitions. Finally, I differentiate between a weak and a strong position against treating strategic voting as a wrongdoing. Fulfilling a methodological requirement, the evaluation of strategic voting is possible thanks to the social choice theory, the roots of which date back to 1950s. It has merged political philosophy, normative economics, mathematics and ethics, creating an interdiscip linary approach in the analysis of the phenomenon of collective decision-making process. Guided by the clarity of reasoning, wherever possible, I have skipped any unnecessary deliberations and referenced them in an extensive bibliography in the footnotes. Hence, this thesis is focused primarily on the conceptual rather than mathematical reasoning.

2. Voting studies

The rational choice theory encompasses three branches: the game theory, the decision theory and the social choice theory.⁵ Yet, aggregating the preferences of multiple individuals with respect to several available candidates, which is colloquially known as elections, is generally scrutinized under the social choice theory.⁶ Therefore, it will be the exclusive focus of our studies. Nevertheless, all the three branches follow a common assumption in modelling human behaviour, i.e., that jointly generated actions exercise the virtue of rationality if each individual has a complete set of preferences – that is, if they follow the order of superiority, inferiority, or indifference among all pairs of choices as you might have noticed in Table. Again, the intensity of preferences is not taken into account at all. Hence, we can use the tools of the social choice theory to follow a process of elections as a decision-making scenario oriented towards maximization of utility which, I argue, might be helpful in revealing the ethical reasoning of the voting mechanism. Yet, a disclaimer must be added.

From the standpoint of a voter, moral obligation may emerge at least twice during an election, that is the general duty of voting, regardless of the outcome – what I am not willing to argue, and voting for a certain candidate who guarantees the expected result. If the candidate’s program is good, then it would be moral to support the particular candidate so that they are elected to the office to act on the good programme. This is the principle of maximization of utility,⁷ which I believe should be limited by mathematical findings of the social choice theory. Note that this model is relevant for voters who enjoy the right to a free election. Sometimes the requirement of secret ballot is added. Why does the condition of a secret ballot is so relevant for some? On the basis of an open ballot system, where the voter’s choice is not confidential, the general argument of representativeness arises, known in literature as the Satterthwaite’s non-transparency of representatives’ preferences argument, which has been arguably established in his desideratum in 1973.⁸ As the open ballot is commonly used in parliaments, strategic voting would blur the voting records of politicians. Politicians may have voted tactically on some occasions. If in assessing the performance of a politician one relies on their voting records only, a distorted picture may emerge: the records do not reveal the reasons why politicians voted as they did. As a result, the voters may have difficulty in ascertaining whether their representative did indeed repres ent their interests.⁹ This argument refers to every case of voting empowered by another (e.g., voting by proxy, deputy, etc.), if the representative votes without taking into consideration the view of their principal. In other words, the representative needs to breach or at least neglect the authorization by lack of consultation with the one the representative acts for. However, according to the contemporary models of representation, free mandate is the most favourable one. It means that MPs represent an entity bigger than a nation, i.e., neither their electorate nor even the whole body of voters. They represent minors, disabled people, non‑citizens, animals, or even future generations. Therefore, free man date invalidates Satterthwaite’s non-transparency of representatives’ preferences argument. Moreover, there are other means of communications that MPs may employ in order to deliver a reason for having voted in a certain way. Keeping the foregoing in mind, I believe that the argument is negligible, if implemented in popular elections.

3. Theory of democracy and ethical consequentialism

Jean-Jacques Rousseau once wrote: “Politics and Morals cannot be separated, and who wants to study one without the other is bound to misunderstand both”.¹⁰ It is the closing passage of his Social Contract book. It was not an explicitly original view, though. The roots of this thinking go back to Aristotle who considered both politics and ethics to be practical studies, where morals collide with individual interests, thus creating the phenomenon of politics.¹¹ Therefore, ancient scholars reflected on the decision -making process, but until the 20th century this approach was linked with benevolent

government rather than fair conferral of power. The former political theory covered decision-making over the collective, while the latter – collective decision-making.¹² According to William H. Riker, “democracy is both an ideal and a method”.¹³ In this sense, democracy is always linked with core values, and it is not just a device to express the will of the majority.¹⁴ Democracy as a method is the process of participation in managing society, where the procedure of voting stands next to expressing free speech, public manifestation, founding political parties, to name a few. Thus, elections are not tantamount to democracy. “Only voting that facilities popular choice is democratic. This condition excludes voting both in oligarchic bodies and in plebiscites in communist and military tyrannies, where voting is no more than forced approbation”.¹⁵ Hence, voting resembles a political program rather than a conceptual condition of democracy, which is neither necessary nor sufficient. Although the discussion over the relationship between democracy and voting is complex; some believe that there is no interaction between them at all, but scoping the thesis to representative democracy, the alternatives are limited to voting and sortition. The latter is strategy-proof, free of any manipulation, but has a major flaw, namely lack of participation.

Riker states that democracy relies on three common cores. These are: participation, equality, and liberty.¹⁶ The theory of democracy then deals with a normative goal, while the social choice theory validates whether these goals are attainable. “By the use of the latter, it is possible to assess, at least in part, whether it is sensible to pursue demo cratic ends by democratic means”.¹⁷ But why is the theory of social choice even needed to evaluate voting? Isn’t counting votes trivial? Since Arrow’s paradox, political scholars have learned that there is no such voting mechanism that might simply reflect the “will of people”.¹⁸ Explicating it formally would take us too far afield, so let me just quote probably the best handbook in the field: Consider a (resolute) voting rule that is defined for some number m of alternatives with m ≥ 3, with no restrictions on the preference domain. Then, this rule must be at least one of the following: 1) dictatorial: there exists a single fixed voter whose most-preferred alternative is chosen for every profile; 2) imposing: there is at least one alternative that does not win under any profile; 3) manipulable (i.e., not strategy-proof). Properties 1 and 2 are not acceptable in most voting settings. Hence, under the conditions of the theorem, we are stuck with property 3: there will exist profiles such that at least one of the voters has an incentive to misreport her preferences.¹⁹ Keith Dowding and Martin Van Hees conclude that voting mechanism is a game, not a truth-tracking procedure for underlying the preferences of the electoral body.²⁰

The outcome depends as much on the features of a decision mechanism itself as on the preferences of its members. The above-mentioned vulnerability of the decision mech anism has generally been viewed as unfortunate, so there have been invented varieties of electoral systems to marginalize manipulation of strategic voting.²¹ Joseph Isidoro Morales wrote in 1797:

In the methods of election currently in use (…) an area lies open to private o personal injustice by the electors, as, depending on the situation, one, two, three or more of them can prevent the election of the most deserving candidate if they thus wish to contravene the course of justice. This system is so well known and occurs so often that an explanation of it is redundant.²² Hence, Morales comes up with solution – the Borda’s electoral system. In such an election, merit and justice are safeguarded by censorship of other electors in the case of a public election, and pangs of conscience if it is secret. Even if men’s passions cause them to lean toward injustice, their pride will lead them to conceal it.²³ Despite Morales’ wishes, none of the electoral systems is manipulation-free. Gibbard- -Satterthwaite’s findings in 1970s have shown that it is not logically possible to establish strategy-proof electoral system, so every each of them is unavoidably vulnerable to stra tegic voting.²⁴ Hence, we can never consider the outcome of any ballots as a straight forward reflection of the will of people. While the Gibbard-Satterthwaite’s theorem shows that for every rule there is a manipulation in a profile, perhaps there are rules for which those manipulations are too rare for us to care about. Even in the plurality rule, which is the most manipulable, the probability of manipulation is about .²⁵ When there are millions of n voters, it becomes negligible. However, things look less rosy when we consider groups of voters that can manip ulate the outcome collectively. If we allow for the groups to be sufficiently large, then we will almost always have some groups (or even many groups) who can attain a bet ter outcome for all the voters in the group by manipulating collectively. Even if the designing of the voting rules were manipulable, but so complex that it would be com putationally difficult for a voter with limited resources to find a strategy, collective effort extends such resources to even a quantitative mathematics scholar. Here comes the consequentialist argument against strategic voting, i.e., the desire that election ought to be like voting poll, which is intended to reveal the sincere pref erences of voters. Therefore, strategic voting inferences such order, which is sequen tially morally wrong. Here, the concept of God of lack crosses my mind. If we cannot otherwise establish a strategy-proof electoral system, ethics enters the game to enforce the desired rules. I believe that advocates of such statements should prove that their proposition has any moral burden in the first place.

Riker argues that once the voting mechanism has been accepted by a community, whatever result it generates, it should be claimed to reflect the people’s will. Otherwise, we would have to rely on the Rousseau-like populist tradition of the collective general will not being derived from individual preference, but via a mysterious “common agent”.²⁶ Under the Rousseauian approach, Riker argues, the accuracy of general will is not subject to an any formal reasoning. In popular elections it is accurate, since each voter chooses a candidate guided only by the common interest rather and not by any personal or private desire. “Thus, by summing the common interest regarding wills of a real person, one can arrive at the will of the great artificial person, the Sovereign”.²⁷ Is it a popular misconception that such populist approach simply remains in the trust for the majority rule, as for example aggregative or electoral (minimal) democracy positions, which encourage expressing only a particular, individual interest in the ballot box.²⁸ Riker sophisticates Rousseauian (populist) approach by claiming that the opinion of the majority must be right and must be respected because the will of the people is the liberty of the people, since the Liberty, as Rousseau says, “is the obedience to a law we have prescripted for ourselves”.²⁹ As can be seen, Rousseau thus invokes morally-driven notions of Liberty or Sovereign written with capital letters. Under the liberal approach, Riker suggests, the outcome of voting is just a decision, without any moral interpretation. A function of voting is to control the officials and nothing more, whilst the Rousseauian (populist), aggregative or electoral (minimal) democracies emphasize electoral output. Liberals are aware that the government’s power can easily be directed against citizens. They believe that officials will be deterred from taking advantage of their power out of fear of losing the power by the next election. Since the future majority cannot be clearly specified, the officials remain self-controlled. Thus, not only the past election encourages officials to maintain discipline, nor the unexpected outcome of a future election, but both the last and the next polling, jointly aggregating limited tenure. In other words, in the liberal view it is not assumed that the electorate is right, therefore strategic voting in liberal approach is not even problematic for democracy, at least in scope of Riker, and has no moral burden at all.³⁰

5. Electoral ethics

Voting is a moral issue since it requires consideration for others as well. Majority of voters do not just choose for themselves, but for everyone, including dissenting minor ities, children, non-voters, resident aliens, and people in other countries affected by their decisions. For this reason, voting seems to be a morally charged activity.

Strategic voting encourages us to vote while being affected by concerns and expecta tions about the likely outcome of the elections, which leads to voting for an option other than the truly preferred one. Though scholars have begun to address the phenomenon in their first papers in the 1950s, strategic voting also known as manipulation has been known for a long time. The oldest description of elections in which manipulation took

place can be found in the Epistulae of Pliny the Younger, which describes a trial over the slaves of the Roman consul Gnaeus Afranius Dexter who were accused of murdering their master.³¹ The most famous examples of elections suspected of strategic voting are the presidential elections in France in 2002 and the parliamentary elections in Slovenia in 2011 – it is estimated that even as much as 30% of voters could have voted strategic ally in these elections.³² In fact, the so-called “voting for a lesser evil” is a very common phenomenon, exercised by nearly one in three voters in the UK.³³ For their manipulation to be successful, agents must be aware of: 1) their preferences; 2) the preferences of all other voters; and 3) the technical operation of the voting mechanism. Moreover, the preferences relation R is 1) complete; 2) reflexive; and 3) transitive. R is complete if, comparing any alternative, the alternatives can be ranked in a chain of preferences, although the comparing individual may be indifferent to some of them. R is reflexive: it is conventionally reflected as aRa for any a ∈ A. R is transitive: in every case in which an individual prefers some a to some b, and b to some c, they also prefer a to c.³⁴

6. Moral dilemmas

I would like to present the three main attempts to challenge strategic voting. These are: firstly, the express value argument; secondly, the sincere argument, and thirdly, the transparency arguments.

6.1. The express value argument

The express value argument relating to strategic voting claims that, even if all the voting mechanisms are vulnerable to manipulation, we should always submit our top-pref erence alternative to fulfil the expressive role of elections in a society.

To shed some light, let us recall the theory of voting behaviour, which holds that most citizens vote not in order to influence the outcome of the elections or government policies, but to express themselves.³⁵ They vote to signal that they are loyal to certain ideas, ideals, or groups. They may have voted for Jarosław Kaczyński in the 2010 elec tions in Poland to signal that they are brave, strong and self-sufficient, or for Bronisław Komorowski to prove that they are solid, hard-working and effective.³⁶ Under this theory, although the act of voting is private, voters regard voting as an apt way to demonstrate and express their commitment to their political team. Sports fans who paint their faces with team colours do not generally believe that they, as indi viduals, will change the outcome of the game, but instead wish to demonstrate their commitment to their team. Even when they watch games alone, sports fans cheer and clap for their teams. Perhaps this is exactly what voting is.

However, even if social science clearly explains why people participate in those collective actions, it fails to answer the question “why people should vote and how to do it”. The gap between the positive and the normative scope does not allow deducing the existence of any ethical concussion in the ethics of voting. Perhaps voters indeed “perform X not to generate (or do) Y, but want to be an X-performer and a member of the collective of (abstractly defined) fellow X-performers”,³⁷ but I am still not convinced why I should follow this strategy and prioritize the expressive role of elections in eval uating my strategic voting.

6.2. The sincere argument

According to this view, manipulation of social choice is regarded dishonest and, as such, immoral from the point of view of deontological value of honesty, regardless of how accurately popular elections track the “peoples’ will”.³⁸ Dowding and Van Hees claim that not all forms of strategic voting deserve being called “dishonest voting” in a moral sense. An honest game is not equivalent to a straightforward game. The difference lies in sincere and insincere manipulation. The former occurs when an “agent 1) votes for a compromise alternative whose chances of winning are thereby increased and 2) genu inely prefers that compromise alternative to the alternative that would otherwise win”.³⁹ Being consistent, since strategic voting is biased on information how others vote, casting a vote can be called sincere manipulation if an agent signals truthfully (for example in a poll) that she will vote in particular manner.

Further on, if one manipulates sincerely, then one hopes that the alternative voted for will indeed be the outcome and, in that sense, one’s choice still expresses one’s will. However, insincere manipulation, i.e., hiding our preferences in order to improve the chances of the other alternative’s winning, may seem to be less expressive of one’s will.

6.3. The transparency arguments

Now I turn to scrutinize Satherwaithe’s transparency arguments.⁴⁰ Note how they over lap with the moral doubts of strategic voting occurring in the consequentialist argu ments introduced in chapter 3 in fine, however, Satherwaithe uses argumentation based on conflicts of utilities.

As regards the 1) inequality of skills; 2) inefficiency; 3) non-transparency of voters’ preferences; 4) non-transparency of representatives’ preferences; and 5) randomness, I do not question those arguments, yet I would like to argue that it is not as problematic as suggested at a first glance. The argument 4) was resolved in chapter 2. 1) Inequality of skills. The possibility of manipulation offers some voters an unfair advantage over others. Voting is supposed to be equal – one person one vote – but if some voters understand the strategic nature of the voting game and others do not, then those with the knowledge gain an advantage. They are more

likely to vote strategically. The ignorant will not understand the strategic possib ilities nor be disadvantaged.⁴¹ This argument fully relies upon information asymmetry. Armed with knowledge, one may take advantage of the ignorant. This argument is valid, but refers to every case of information asymmetry. Knowledge is power – but in many elections, it is not considered desirable. 2) Inefficiency. Strategic voting is wasteful. Finding out the preferences of others is a waste of resources. If there are strategic possibilities inherent in the process, then in order to discover them, voters need to know not only their own prefer ences, but the preferences of everyone else. Discovering these preferences is taken to be a wasteful and inefficient activity.⁴² This can be seen as a type of strategy of the commons; everyone would be better off if nobody spent effort on manipulation, but individually voters are still bet ter off when manipulating. This argument is particularly valid in fields such as sports tournaments, I believe, where fair-play is a beacon for judges. However, in popular elections or, more generally, in politics, I would assert that the cost is worth the result, because any democracy theory treats public debate as being beneficial. Knowledge of the preferences of others may challenge our own and induce our understanding other viewpoints. Getting familiar with the preferences of others stands as a part-and-parcel of a working democracy. 3) Non-Transparency of Voters’ Preferences. The possibility of manipulation gives people incentives to hide their preferences. If someone can hide their pref erences, they make it harder for others to manipulate and may make their own strategic possibilities easier.⁴³ As presented in section 5.2, the hiding of preferences is certainly a wrongdoing, but I doubt that transparency of voters’ preferences has any moral significance as outlined in chapter 3. It may be structured as follows: a weak position advocated by Dowding and Van Hees states that sincere manipulation indeed requires trans parency of voters’ preferences. Strong position represented by Riker, however, in order to invalidate non-transparency of voters’ preferences arguments needs to go as far as to admit that voting procedure has no moral burden at all, outweighed by other virtues of democracy like participation, equality, and liberty. But, by some reason these are not incorporated to elections procedure itself. For Riker, limited tenure exclusively manages to keep officials in the hands of an electoral body, which guarantees those three virtues of democracy. There is no need to extend them to the process of election itself. Again, non-transparency of voters’ preferences arguments is valid, but refers to every voting game, not only popular elections. For example, imagine only one round of two-candidates majority elections which are unmanipulable under Gibbard-Satterthwaite, because no transfer of preferences is possible. Next, let us suppose that two candidates race for a presidency and that Bush is expected to beat Gore by a whisker. Gore supporters may pretend Bush is a shoo-in to be counted beforehand as Bush supporters, which will result in a lower turn-in of genuine Bush supporters on the election day.

4) Randomness. Manipulation introduces an element of randomness into the voting process. If there is no dominant strategy, then there may be different ways of trying to secure one’s interests as well as possible. In those cases, it may be very difficult to predict the outcome of the decision process.⁴⁴ The fact that outcome is unpredictable does not mean it is random. Other altern atives are either to appoint officials by lottery or adopt a voting system with elements of dictatorship (according to Arrow’s theorem).

7. Conclusions

This paper answers a particular question whether it is possible to maintain with a sound reason that strategic voting is not a wrongdoing. The answer is positive and I deliver two positions, a weak and a strong one. I have addressed all major doubts of tactical voting without going into mathematical elaborations about equilibriums and econo mic assumptions, thus making this work accessible to everyone, especially freshmen of liberal art degrees. Unravelling the reasons for moral dilemmas induced me to take into consideration the theory of democracy, the social choice theory and political philosophy.

Being consistent, after Riker, I deny the Rousseauian approach which fails to hold its assumptions since the findings of social choice theory have gone public. The sincere argument remains valid, albeit with some added conditions. However, strong Riker’s position of gatekeeping the moral burden to some democratic institutions and limiting them to others seems arbitrary.

strategic voting may be “sincere”, and therefore morally acceptable under the argument 3). However, the latter rejects the claim that elections bear any moral burden and claims that voting itself should be perceived as an ethics-free decision-making tool. Keywords: ethics, moral, elections, social choice theory, strategic voting, Satterthwaite, Arrow’s paradox