The Myth of the Rational Voter by Bryan Caplan

Rating - 3: worth reading once (borrow it from a library)

This is the new book from my favorite young economist, so it was almost a foregone conclusion that I would recommend it. Take that as your disclaimer. The other side is that this fairly academic book may not appeal to a general audience. The whole point of this book is that the general audience is ignorant, irrational, and does not want to hear it. I can recommend this to you if you are rational and politically interested. I cannot recommend this to you if you are stupid or willfully ignorant.

Voters are systematically irrational. They vote on the basis of what feels good rather than what actually works. This leads to democracies' performing worse than they should, not because special interests distort the system, but because politicians mostly do what the voters tell them to. Everyone would be better off if voters were less biased (willfully ignorant) or if more decisions were left in individual hands (markets) rather than subject to the whims of democracy.

The Myth of the Rational Voter offers a lot to different audiences. While a general audience may not read the book, an article-length treatment will cover most of it. Indeed, just reading the Coase quote before the table of contents will give the gist: people make more careful decisions about what to do with their own money than with the public treasury, or else they run out of money and stop making those bad decisions.

For the academic, about one-fifth of the book is citations. Hear that, reader, the actual text is about two hundred pages, so this is a quicker read than you think. Ahem. If you want further reading or to check Prof. Caplan's claims, he has given you everything you need, down to page numbers and regressions. This is an advantage of professors over the more popular writers: they cite things properly. You can see Prof. Caplan arguing with other theories and researchers. Follow the citations if you want a sense of where this fits into the literature.

Chapter two gives you the best simple chunk of the argument: more informed voters are more skeptical of the government and friendly to markets. This holds across preferences and demographics, rather than being a case of educated people voting their self-interest. The less you know about policies and politicians, the more likely you are to favor economic mistakes like protectionist tariffs and farm subsidies.

You may have paused there to wonder if Prof. Caplan has inserted his policy preferences as "right" and labeled all opposition as "irrational" or "economic mistakes." While you have reached the most popular critique, it is not a valid one: he has quite a bit of empirical evidence showing that people vote differently when they know more about economic, government, the effects of policies, etc. The informed view falls somewhere between his views and the median voter. He quotes well-informed sources from across the political spectrum. (Your proper next step is to ask why he is not then irrational for being more pro-market than the median informed voter, to which I will let him respond.)

Or perhaps you did not so much pause as shout in outrage that your trade and farm policies are not mistakes. Sadly, you are wrong. This is not the place to argue it, but all trade barriers and agricultural subsidies make society poorer, with the largest costs usually borne by the poorest people. If we really think farmers need more money or steel mills should have higher profits, everyone would be better off if the government simply wrote checks, rather than having economy-distorting tariffs and ethanol subsidies. If you do not understand why, you probably need to read this book. Seriously, you have no economic leg to stand on.

Not everything is economics? Yes, that would be a major reason why the median informed voter has different policy preferences than Prof. Caplan: different values. He understands this, but it does not change the fact that economically ignorant voters will support policies that work against their values because they sound like nice things to do. Becoming better informed will shift your policy preferences towards his even if you disagree with him on where the country should be headed, because you will cause less collateral damage along the way. Also, he takes economics as his example because it is what he knows; the general population is probably similarly ignorant and misguided about what you know well.

How bad is it out there? Let's say a pollster called and asked you to name your senators; if you said, "I couldn't name either," you would be ahead of much of the population because you knew there were two. Most voters could not pass a test of basic economic concepts, but they still get to vote on who sets our trade policies. The theory of our republic is that the voters can pick people of good judgement who will then become informed and select good policies. As it turns out, most elected politicians really will respect the median voter's biases and legislate accordingly, or else he will get voted out of office and the new guy will.

In case you do not follow links, let me explain what the most common biases are. Most people are insufficiently supportive of markets, because they think that profit comes at someone else's expense rather than being the result of a mutually beneficial exchange; worry about trade with foreigners, while such trade is almost always good for everyone involved; favor make-work under the heading of "creating jobs" and oppose creative destruction that frees labor for more valuable employment; and think that we are in dire times, rather than the best moment in history ever to have been alive. In case that was insufficiently clear, the correct answers are: markets are good, foreign trade is good, eliminating non-productive jobs is good, and the world has never been wealthier. If those are not intuitively obvious, read the book.

There are not new biases or temporary mistakes. Quotes from the 1700s are still germane, and I defy you to guess which Quarterly Journal of Economics citations are from 1893 and 1993. Adam Smith wrote about these biases before the Industrial Revolution, and we still have not learned.

My personal quest is against pessimistic bias. Life is good and has never been better. You are reading this on a global communications network, and we have enough prosperity and leisure time to chat online about books. The computer in your car is more powerful than ones that took us to the Moon. A minimum wage job in the United States will give you a standard of living above the world's median, and it will give you luxuries that the wealthy could not afford a century ago. Things are only going to get better as technology accumulates, lifespans increase, and productivity grows.

By now, you should have questioned the premise that economists agree on things. You see them disagreeing all the time, and you can find someone to support any position. True, but economists do not disagree on core issues. You see them arguing at the margins. Economists argue about which forms of trade are most beneficial, not whether trade is beneficial. They argue about how much deadweight loss taxes create, not whether there are losses. You can see scientists arguing all the time too, but those are disagreements at the edges of inquiry, not questions about the existence of gravity or natural selection.

I'm tired of addressing counterarguments. The text anticipates them. Read.

A key argument is that the public is rationally irrational. Some beliefs are comforting but false, and you can believe them as long as they have very little impact on your life. Prof. Caplan's view is that a voter's chance of swinging an election is approximately zero, so the expected loss from voting on the basis of irrational beliefs is approximately zero. You will never be the tie-breaking vote in a national-level election. Given that, believe and vote as you will, and you suffer no harm. Smuggling in another counterargument, you could argue that a one in a million shot at altering a multi-trillion dollar budget is worthwhile. We are, however, in a system of divided government with checks and balances in which both parties do relatively similar things on a great many issues, due to the limit range of options that public support gives them. Discuss.

Prof. Caplan also argues that the self-interested voter hypothesis is clearly false. There is very little correlation between what is good for voter X and what party/policy s/he supports. Worse, the system would be better if the hypothesis were true. If you vote for a policy to help someone else and it actually hurts them, then you feel like you have done a good deed while the problem metastasizes. You have moved on to worsening the next problem while wondering why those people you helped are not grateful.

As a final substantive note, I particularly liked the distinction between the policies that voters want and the policy outcomes that voters want. For example, voters favor both low prices and shielding of domestic industries from foreign competition. Protectionism causes higher prices. Therefore, voters will always be unhappy with democratic outcomes, because they cannot have both the policies and the outcomes they want. I want to eat lots of chocolate and lose weight, and I will fire any nutritionist who will not give me both. A related note shows why the deluded idealist is more likely a successful politician in this environment than the cynical panderer: it is easier to be a passionate advocate for self-contradictory falsehoods if you really believe them.

On the structure of the book, there are at least a dozen endnotes that could productively be converted to footnotes. Not everyone will flip to the back to check each citation, so they will miss good parentheticals. Chapter I note 68 is a good example of this. In the endnotes, chapter I note 61 and chapter VI note 28 both reference a decades-old source to show that men are more pro-choice than women. I suppose "have been" would solve this nit-pick, but there must be a more recent source. Finally, twenty-one self-citations in the bibliography?

The SAEE charts in chapter III are notable for using the same scale. Many authors would shift those around to exaggerate the differences, but this is a good use of simple and honest descriptive statistics. It is easier to be honest when the facts support you, as these graphs do.

Amazon link
EconLog, Bryan Caplan's blog

Gödel, Escher, Bach by Douglas Hofstadter

Rating - 2.5: parts of it are worth reading once (borrow it from a library)

You will notice that it has been a long time since my last review. This is a long book that does not go quickly. Some parts are mind-numbingly tedious, so I was stuck on one chapter for more than a week. Other parts are very insightful, and you pause after the chapter to contemplate them. Some require re-reading to get their full meaning, and you are encouraged to pause and do some sample problems from number theory. This book is not a small project.

GEB explores how recursion and self-reference allow meaning and complexity to arise from the operation of very simple units. Number theory is a window on how formal statements can mirror reality and how we impart meaning to the manipulation of symbols. We see similar patterns in our neurons and DNA, suggesting this as a path to understanding how consciousness and identity arise.

The structure and content of the book mirror each other. An idea will be explained in a way that uses the idea. Each chapter is proceeded by a dialogue that demonstrates this. Bach is our musical structure, Escher is our visual representation, and Gödel is our guide to its formal logic; all three (and others) deal with recursion, so the ideas bend back onto themselves.

Yes, this book gets two paragraphs of summary. It is that hardcore.

I wanted to like this book more than I did. I found myself at the usual 20% point realizing that I was not enjoying it, but I stuck it out because this is one of the most highly recommended books that I know. I find that I would have few people to whom I could recommend it, and even then I might recommend using a summary for some chapters. In the spirit of the book: the chunked version is more valuable.

Great attention to detail is both its strength and its failing. The book does an excellent job of showing, not telling. Rather than just saying that number theory proves x, Mr. Hofstadter walks through the steps. Better yet, he goes back one step earlier to show how you derive number theory. The chapter on DNA takes you down to the base pairs. If you want all that, it is there. If you do not want all that, it is still there. This is healthy. To really get the ideas, you should work through them. Most audiences, however, probably do not need to get the ideas that deeply, and it makes it seem suspect when there is any hand-waving about what is assumed but not demonstrated.

The best example is Typographical Number Theory. Mr. Hofstadter demonstrates a notation for number theory by which all derivations are simply manipulations of symbols according to rules. No thought is necessary, just the repeated application of a set of tools to a set of theorems. If number theory is too wild for you, this could be a way to calm things. You might want to skip TNT, but it is a central part of the book, so you need to have at least some understanding of what he is talking about. A later dialogue inadvertently reflects TNT in asking why we do not read books one letter at a time: "Well, it sounds like that would turn the experience of 'reading' The Pickwick Papers into an indescribably boring nightmare. It would be an exercise in meaninglessness, no matter what concept I associated with each letter." (I presume that Mr. Hofstadter thinks otherwise, since the next dialogue cites number theory as a great entertainment for an evening, explicitly not something that will help you get to sleep.)

So that is why we have a *.5 rating: punishing detail. It is good intellectual vigor to read through it, but it is going to be far more detail than you need in your life, and it is not clear what is skimable.

The book is clever, too clever by half. There are anagrams and puns everywhere. The structure mirrors the content, which helps and leads to some very interesting bits. In one dialogue, the lines of the second half mirror the first half. Another discusses music and arranges the characters' lines to act out the fugue elements as they come up. There are nested dialogues, nested pictures, and theorems that represent themselves. Typographical Number Theory sets up a bit that reaches an unexpected punchline a few hundred pages later. One bit uses six letters to make a kind of joke about DNA, the periodic table, music, and the book itself.

That is a lot to pack into six letters. The book is very densely packed, like DNA. This is another way that form follows function, because the book so often shows how many (and self-reflecting) layers of meaning are packed into simple-seeming structures.

At the extreme, we have an unreliable narrator. That can be great in fiction, but it is troubling in non-fiction. One never knows if we can fully trust the author. Is this next bit a joke that we are meant to catch? An intentional mistake or trap, which may or may not be pointed out? Mr. Hofstadter discusses how a message may or may not successfully signal that there is a message to be decoded there. Have you ever sent signals that no one realized were signals? It seemed so obvious to the author... The joke bibliography and index entries are some of the most visible, and some puzzles are pointed out along the way. Add thirty years of progress since the book was written to whatever gaps in knowledge the author had at the time, and we really cannot trust the book to be fully accurate. (Think of computers in 1979, although the important fundamentals are the same.)

One famous idea from GEB is how an author can hide the true ending of a book. A reader can always see how many pages are left, so the author needs to pad the end of the book with other stuff after the true ending. The lengthy bibliography and index can serve that function here, but we have an explicit statement about including extraneous ideas and characters. Of course this is demonstrated in the dialogue that proposes it, but would you care to guess where the real ending of GEB is?

Or, given the structure of the book, the real beginning? Note that odd first word, and the last dialogue.

You should have the feeling that you are missing something. You probably are. I surely did. The next question would be if that is an important something, or just another instance of Mr. Hofstadter subtly hiding his name in the book (the index gives some pointers there). There could be a variety of esoteric messages hidden within. You could spend a long while sifting through and playing. Any demand for an obvious payoff defeats the purpose of it all.

That is how clever things are: wonderful to get, but you can usually do without them. You do not get the full meaning from a book if you do not catch its allusions, but I would be hard-pressed to say what you lose other than some ill-defined "flavor." While I may appreciate Buffy the Vampire Slayer on a much deeper level than you, you will somehow get by without extensive knowledge of pop culture, classical mythology, and in-jokes. I may get slightly more utility from the show, but it takes me much longer to explain the full impact of a scene with many references. If you have to explain the joke, it isn't funny.

See, that was an example. I just spent too long explaining why it is a bad idea to spend too long explaining. And there I go again. The book rubs off.

I have mostly addressed structure and the way of telling, rather than the content. Given the book, that might mean that I have adequately covered the content. I think the clever tricks are probably too distracting, so most people talk about GEB as an exercise in self-reference, rather than really addressing self-reference and the emergence of meaning from complexity. Mr. Hofstadter may have thought so, since he has a recent book that goes back to the core topic in a purer way. Gödel's Incompleteness Theorem is meant to be deeply meaningful, not a clever trick.

There book yields much, but you must put much work into it. I think that's called learning.

Amazon link
Gödel
Escher
Bach