You are currently browsing the category archive for the ‘elections’ category.

A few days ago, I linked to a paper on Voter Ignorance by Ilya Somin. There are a few points in it that I would like to talk about:

1) Here is Somin talking about retrospective voting (p. 15)

The retrospective-voting argument does, however, possess a kernel of truth. As Fiorina puts it, retrospective voting can impose a kind of “rough justice” on political leaders who have failed badly. If a policy failure is large, highly visible, and easily attributable to a particular set of leaders, it is certainly likely that they will be voted out of office, as the elections of 1932, 1952, 1968, and 1980 suggest. Moreover, the bigger the failure, the less likely it is that the opposing party’s performance will be worse. The ability of voters to punish large and obvious policy failures by incumbents is one of the major advantages of democracy over dictatorship.

This is how I think about it too. Democracy in general gives politicians a weak but significant incentive to work for good policy, because while voters are quite ignorant about politics, they are not completely ignorant; they do generally know when politicians have done an especially poor job. Unfortunately democracy also gives politicians a host of other incentives that aren’t so great.

2) Somin argues that even altruistically motivated voters are rationally ignorant because the chance that they affect the election is vanishingly small. However, as Gelman and others show, altruistically motivated voters do have an incentive to show up to vote, and extending their logic, also to be informed. What Somin forgets is that while the chance that a particular voter changes the outcome of the election is minuscule when the voting population is large, when the voting population at large, the social welfare effects are also large. Moreover, the chance that a voter decides the election and the size of the welfare effect are inversely proportional to eachother, so that when we calculate expected altruistic utility of voting, it is constant.

However, I don’t think that Gelman’s analysis is quite right either because voters aren’t interested in being altruistic per se, they are interested in the psychological rewards of feeling altruistic, and that feeling doesn’t heavily depend on how informed the voter is. I should say that while it seems intuitively obvious to me that the rewarding altruistic feeling from behaving altruistically doesn’t heavily depend on how informed a voter is, I don’t have a good explanation for why that would be the case. I suspect there’s some sort of cognitive bias at work here.


In a comment on my article on Professional Voting, Robin Hanson questions how well the Election Council in a Professional Voting system would work

The obvious question is how well can ordinary voters monitor whether the Electoral Council members are choosing neutral fair test questions.

I have thought about how low levels of monitoring by ordinary voters (due to rational ignorance) would affect the effectiveness of the Electoral Council, and I have concluded that low monitoring levels would not affect Electoral Council effectiveness very much at all. To a large extent, voters do not need to monitor the Electoral Council members in order to give them incentives to do their job well because those incentives are inherent to the job.

Essentially, the only non-monetary rewards more readily available to Electoral Council members than to other people would be small level fame and the ability to pursue sociotropic goals (ideological or otherwise) for example, getting  the minimum wage raised. Electoral Council members would have little opportunity to be corrupt or wield direct power, because the Electoral Council would not control very much public money nor control powerful government functions.  Because Electoral Council positions have a comparative compensation advantage in providing the opportunity to persue sociotropic goals and because sociotropic motivation is quite strong for some people, sociotropic motivation would be especially strong among Electoral Council members because they would self select for being highly sociotropically motivated (I discuss this logic here). Therefore, we should expect Electoral Council members to act mostly to maximize the monetary rewards of being reelected as well as maximize how much they advance some set of sociotropic goals.

The only way to pursue sociotropic goals as an Electoral Council member, however, would be to influence the arguments and facts that voters learn by influencing the questions that appear on the political knowledge test that the Electoral Council would administer. For a Council member to advance their set of sociotropic goals as much as possible, they must ensure that the questions on the test cover the most convincing arguments and facts supporting their sociotropic goals, and to ensure that those questions measure how much individual knows about such arguments as accurately as possible. Luckly, this is exactly the job of the Electoral Council.

Working to include biased questions on the test would not further the influence of individual Electoral Council members because biased questions are very easy to work around. For example, including the question “Are you a Democrat?”, on the political knowledge test would not help Democrats, because non-Democrats can easily claim to be Democrats. Furthermore, even if such biased questions did help the factions which included them, other factions would strongly resist such questions because they would be percieved as unfair. Attempts to include questions biased in favor of specific demographics would be almost totally muted by the weighting of the final voting result by the demographic survey.

One potential problem would occur if the primary cause of disagreement between electoral council members is differences in values and not about what arguments and facts are most true. In this case, Electoral Council members, may devote significant energy towards preventing questions focusing on strong arguments supporting policies which they oppose from appearing on political knowledge test, instead of towards including questions focusing on strong arguments supporting policies which they support. This would obviously be a bad outcome.

This is an interesting potential problem because it means that we should actually prefer Electoral Council members to be be relatively biased towards believing arguments and facts that support what they support and against believing arguments and facts which do not support what they support. However, monitoring of Electoral Council members by ordinary voters (however little of it there is) and Electoral Council members valuation of honesty in general should reduce Electoral Council member’s efforts to keep legitimate arguments that opposing factions support from being covered by the political knowledge test. Additionally, this would not be a potential problem for Professional Voting as a corporate governance institution, because Electoral Council members would value profit almost exclusively and homogeneously, so disagreements would only exist about which arguments are persuasive.

Illya Somin argues that the rational ignorance of voters gives relatives of former successful politicians an advantage stemming from better name recognition.

Because voters know very little about the details of candidates’ ideology and issue positions, they use a candidate’s family affiliation with a popular political leader as an information shortcut. Voters could instead analyze each candidates’ qualifications and ideology in detail (though, as Bhutto noted, that may be impossible for those who are illiterate or poorly educated). However, rational ignorance ensures that most of them have neither the time nor the incentive to do so. Bhutto herself, of course, rose to power in Pakistan in large part because voters associated her with her father, a popular politician who had been executed by a military dictator in 1979.

He also argues that this advantage also means they will be less competent on average then other politicians, presumably because there is a smaller population of politicians with familial name recognition. This also suggests that in elections where the electorate is more uninformed, familial name recognition should be more of an advantage. I would be interested to see a study try to support this empirically.

Update: Discussion in the comments.

The concept of rational voter ignorance has been around since the 50’s. The reality of rational voter ignorance seems obvious to me, but I have been searching for a good source of strong evidence for broad voter ignorance, a paper or survey that I can point people who doubt voter ignorance to. Several papers that I have run across mention that voter ignorance is well documented, but they do not cite anything, and I have been unable to find any survey paper that specifically focuses on the extent of voter ignorance. There are also several papers I have run across that note significant policy preference shifts when voters become more informed (like this and this), which this implies voter ignorance, but I would like a paper specifically focused on the extent of voter ignorance.

Does anyone know of such papers?

In the past, I have advocated both Direct Representation, a form of proportional representation, and the institutional protection of federalism by having state governments at least partially elect the national government, as was the case in the U.S. before the 17th Amendment. How these two institutional features might fit together is not immediately obvious, but the solution is relatively simple:

At the state level, each population elects it’s representatives (parties) in the normal Directly Representative way. At the national level, Direct Representation elections are also held normally except the votes are cast by all the individual state parties using the votes they have been delegated instead of by voters in each of the states. In this way, both the state and national legislatures remain Directly Representative, and the national legislature remains representative of the state governments.

I have considered discussing vote selling for a while, and now Greg Mankiw has brought up the subject

Professor Sandel asked a fascinating question (and I am paraphrasing), “If you economists are so in favor of voluntary exchange, would you extend that conclusion to letting a person sell his right to vote to another?”

I said No. It is true that both parties in the transaction must be better off if they agreed to the deal. Nonetheless, the standard argument for unfettered voluntary exchange does not apply because there are externalities. That is, when one person sells his vote to another, that transaction may affect unrelated third parties through the electoral process.

I think Mankiw’s answer is correct. Economists should not favor vote markets, at least vote markets which can include transactions other than pure vote for vote trading. I also think the reason for this is that voting decisions have large externalities, but I’d like to elaborate a little bit on Mankiw’s reasoning.

As I have discussed here many times before, people generally do not vote in their material self interest, and thus voting is a reasonably good way of providing for public goods. However, if vote selling is allowed, the people buying votes are likely to use the votes in their material self interest, so there would be a shift in the type of people who make voting decisions from those who are likely to seek public goods to those who are likely to seek private goods. This is the essence of the problem.

Let’s do a little math to show that buying a presidential election would be possible in a vote market. I suspect that in any given election, most people would currently be willing to sell their vote for about $15, and there are about 220 million potential voters in the U.S (source). That means that to buy 1/2 of all the votes in the U.S. would cost approximately $1.7 billion. It is definitely within the power of the President to divert much more than $1.7 billion to a private company over the course of 4 years, so a presidential candidate could buy the election through a vote broker by promising to pay them over $1.7 billion if he is elected.

There are two major bad outcomes that would result from a vote market. First, in such a political system, government policy will be designed to maximize rent extraction, through high taxes, tariffs etc.. Consider the simple world where politicians buy votes to be elected dictator and make all policy decisions except how to conduct elections. Votes would have a rather large market price because they would be worth a great deal to politicians because the government manages a great deal of money. I envision that votes would be sold contingently; people would sell their vote to a politician on a contract which pays out several thousand dollars if the politician wins. In order to continue offering more and more money for votes, politicians would have to extract more and more money via rent extraction. Politicians would have little incentive to provide for public goods because, by definition, public goods are non-excludable and would not help them get votes. The resulting wealth distribution would be very egalitarian because the winning politician would switch quite often, so everyone would get paid for their vote fairly often, so this might sound good, but this would be achieved by maximizing rent extraction, which is clearly a very harmful and inefficient practice.

The second bad outcome is that the country would likely eventually collapse into an actual authoritarian system. In a political system with a vote market, there would be no way to limit control of a ruler over elections and law enforcement because politicians would compete solely on vote prices and not at all on policies. Politicians who are completely free to establish their own policies would inevitably do away with elections and establish some type of oligarchy. Politicians could even state so openly because vote holders would have no incentive to favor non-authoritarian politicians over authoritarian politician.

Addendum: Bryan Caplan also weighs in

The lesson: If you’re afraid of vote-selling, you should be afraid of voting as well.

I think he is too pessimistic.

Students at New York University claim their voting rights are worth a lot to them,

Only 20 percent [of NYU students surveyed] said they’d exchange their vote for an iPod touch. But 66 percent said they’d forfeit their vote for a free ride to NYU. And half said they’d give up the right to vote forever for $1 million.

The most obvious explanation for this is that you can’t always trust people to do what they say they will do. Talk is cheap. I find it extremely unlikely that more than 1% of students would not give up their voting rights for $1 million. If some people value their voting rights in excess of $1 million how voting come voter turnouts are around 30%? These students are not giving realistic answers.

Now let me draw a parallel between surveys about voting and elections. In both cases the costs of answering “incorrectly” are very low. In surveys there is no individual benefit for answering truthfully. In elections, the material benefits of an individually preferable outcome to any individual voter are significant, but the individual benefit of voting intelligently is very small because of the tiny impact that an individual vote has on the outcome of the election. However, In both cases there are significant psychological benefits to answering or voting in ways that feel good. In the case of survey about voting, it feels good to fool yourself into thinking that you value your voting rights very highly because thinking that you are a civically minded person is pleasurable. In the case of elections, it feels good to vote for proposals which sound good without investing many resources into determining if proposals are actually a good idea. has an interesting post about campaign ads which are “99% Fact-Free”:

In this article we examine two examples of what we call “fact-free” advertising, which we see in abundance. These ads seek to associate the candidate with a string of positive words and images but are void of specifics. Voters should beware.

I think this generally supports the ‘Vision’ theory of campaign advertising. As the article explains, however, these ads do serve a useful purpose:

Even hot air has its uses. “These ads do have a lot of meaningless rhetoric but are not completely empty,” says Kathleen Hall Jamieson, a professor who teaches courses in political communication at the University of Pennsylvania. “Actually these two ads signal two different sets of priorities. Ask how you would react if Edwards spoke of a ‘strong military’ or Romney said he’d ‘lift families out of poverty.’ Romney uses traditional Republican language to signal that he would spend more on defense. Edwards speaks of ‘the middle class’ to signal that although his policies will address poverty he will focus on middle class needs as well.”

This is what democracy does well. Democracy is good at distilling voters’ altruistic values, but democracy is not good at selecting representatives who would choose intelligent ways of promoting those values. It is possible that it is actually lucky that these ads don’t contain policy details, because voters have poor incentives to evaluate policy well. If voters were really interested in deciding on public policy relatively directly, it could actually make policy worse if voters have bad biases.

This brings up a side point. I have heard several people attribute declining voting rates to opportunity cost; meaning that you can now do a lot of cool things with your time that you couldn’t before, but because voting has not become more attractive, people have substituted away from voting. Substitution takes two forms; either fewer people can vote (voter turnout has declined significantly in the last 40 years) or people can spend less time on voting activities, such as becoming informed about proposed policies, which I suspect is the case. If the second effect actually improves governance, as a country becomes richer, government should actually become more effective.

Both these points rely on the assumption that “a little knowledge is a dangerous thing;” is this a realistic assumption? I’m not sure.

Money is clearly important in politics, but I haven’t seen a good theory for why it is important. One of the biggest campaign expenditures is political advertising, and that is what I want to discuss right now. It is important to realize why and how people vote. Economists have long noted that it is not materially rational to vote; therefore, people vote because there are psychological benefits to doing so; doing your ‘civic duty’ feels good. Additionally, as I have discussed here before, people vote mostly altruistically, again, because voting altruistically feels good. This feeling does not, however, depend on whether a vote really did (or would) improve the world, so voters have essentially no incentive to make good political decisions.

This logic clearly restricts theories of political advertising to those which do not require voters to expend significant time or mental resources. For example, advertising cannot be meant as an argumentative device because voters are rationally uninterested in using time and mental resources to evaluate complex arguments.

I briefly discussed this subject with John M, and his major theory was that political advertising is primarily done to increase candidate-name recognition. Candidate names that are heard by voters often enough stick in their minds and voters are more likely to vote for candidates whom they have heard of before. The problem I see with this explanation for major races is that people who are indifferent enough to the election to vote based on which name they have heard most often, would be unlikely to be voting in the first place.

I will suggest two other explanations. First, political advertising could simply serve to signal popularity. People know that advertising is costly, so advertising signals that other people donated money or other resources to a candidate that advertises. The major problem with this theory is voters are certainly aware that politicians can get money from non-benign sources.

Another theory is that politicians use advertising to create a product which is “sold” to voters for their vote. Politicians use advertising to create an appealing ‘vision’ for public policy. This ‘vision’ would include ideology, emotional appeals, and candidate character and history traits. Voters vote for the candidate with the most appealing ‘vision’ because it gives voters the satisfying sense of leading the country in the direction of the candidate’s ‘vision’ or at least not towards the less appealing ‘visions’ of other candidates.

I think the ‘vision’ theory probably explains the bulk of political advertising in major races where voters are well aware of candidates. I think name recognition theory and/or the popularity signaling theory probably explain a lot of minor race political advertising. Consider campaign yard signs; yard signs do not have any content besides a candidate or issue name and sometimes a party identification. Both the name recognition theory and the popularity signaling theory can explain political yard signs, but the ‘vision’ theory cannot. It is my impression that yard signs play a much larger role in minor race campaigns than in major race campaigns.

It may be that this issue has already been well explored, but I have had trouble finding literature on the issue. What’s your theory for political advertising?

From The Free For All I learn that Washington state had campaign speech censor called the “Public Disclosure Commission” which could impose financial sanctions.

The official truth squad was something called the “Public Disclosure Commission.” Who would get to serve on this powerful commission? The majority stressed that PDC members “are appointed by the governor, a political officer. This group of unelected officials is empowered not only to review alleged false statements made in political campaigns but also to impose sanctions.” The possibility (or perhaps probability) that the commission’s notion of truth would reflect its political biases as well as ordinary human fallibility, seem clear, yet the statute did not require that the PDC’s decision be subjected to independent judicial review.

I am pretty shocked that I had never heard of this, and I am appalled that the state Supreme Court struck down the ruling by only 5-4. I have pretty a low opinion of voter’s ability to pick up on campaign lies, but it is impossible to see the commission working in any legitimate way. Here’s a Seattle PI article on the ruling.