Randall Holcombe’s Following Their Leaders: Political Preferences and Public Policy makes the case that contrary to the usual description of democracy, where voters call the shots with their votes and political leaders craft policies as the voters direct, political leaders control and craft policy, and voters follow their leaders, adopting their political preferences according to the platforms created by the elites. So how well does his case hold up? I’ll give my thoughts on what I see as the main claims of his case—the idea that votes reflect expressive preferences rather than instrumental preferences, the idea of anchor and derivative preferences, and the idea that policies are made by elites with voters following their leaders, rather than elites making policies according to voter preference.

The claim that votes are expressive rather than instrumental is where I see the most grounds for skepticism. The idea is theoretically sound—what people express when nothing is at stake often differs from what they actually choose when they directly create an outcome. And it’s factually correct to say that in all but the smallest elections, the act of casting a vote has essentially zero chance of creating an outcome. But I think the case may be overstated here. It takes for granted that voters are aware that their votes have essentially zero instrumental value—an assumption that was forcefully criticized by Jeffrey Friedman in another book I’ve covered in depth, Power Without Knowledge: A Critique of Technocracy.

Friedman argues this claim is simply asserted far more than analyzed, and Friedman also contends that the mathematical work showing votes have essentially zero instrumental value is actually an esoteric bit of knowledge that can’t simply be assumed as universally known. Additionally, when one openly declares “I don’t vote, because it’s not worth the time and effort since my one vote won’t make a difference,” the typical response from most people is bewilderment, because they consider that claim to be obviously wrong. Many—perhaps most—people will insist that of course your vote can make a difference, because most people genuinely have no idea of the mathematics implied by that claim. Successfully convincing people their vote has zero instrumental value takes a great deal of time and effort.

Another reason I see for skepticism is something that should be familiar to anyone with inclinations towards so-called “third parties” in the United States. I wasn’t quite old enough to have voted in the 2000 presidential election, but I was at least politically aware at that time. And I well remember a major point of contention was the candidacy of Ralph Nader and the concern that he would be a spoiler candidate for Al Gore. There was a fierce debate going on at the time among Nader supporters about whether they should cast their vote for Nader, which would effectively tilt the odds of what was being projected as a very close election towards George W. Bush, or if they should withhold their vote from Nader and vote in favor of Gore, whom the typical Nader supporter considered the lesser evil. Many voters went to the polls in that election preferring Nader as a candidate, but still cast their vote for Al Gore, because they knew that the outcome would realistically be either Bush or Gore, and Gore was the outcome they preferred between those two. This kind of behavior seems far more like voters who see themselves as using their vote to choose an outcome rather than express a preference.

This is not to say I think the distinction between expressive and instrumental preferences is without value, or that it never applies in voting. As I’ve commented before on this blog, I tend to interpret ideas like rational irrationality, or expressive versus instrumental preferences, as being more of a sliding scale than a binary switch. And if we take Holcombe’s strong claim (voters cast their votes expressively rather than instrumentally) and modify it to a weaker claim (many voters cast their votes more expressively than instrumentally), his point that voting aggregation methods can’t be used to validly infer instrumental social choices still holds.

The concept of anchor and derivative preferences seems solid to me. When looking at how people form their political preferences, the statement “I like the red tribe’s policies, so I’ll be on their team” seems to be much less reflective of reality than “I’m a member of the red tribe, so I’ll support their policies.” There is no official account of how many political issues there are, but when we begin to list off issues impacted by political policies, the list quickly becomes extensive. Gun control, abortion, trade and tariff policy, police policy, taxes and spending, military spending and foreign policy are all obvious examples, and each of them breaks off into multiple lines of inquiry. For example, “taxes” as a category contains all sorts of separate issues, such as what should be taxed (income, wealth, imports, externalities, etc.), how those taxes should be structured (flat rate, progressive rate, regressive rate, fixed payment), how those taxes interact with other taxes (deductions or credits), among other questions. Nobody has enough information or knowledge to have a well-formed opinion about all of these topics and subtopics simultaneously. And yet, the vast majority of voters do hold strong opinions on all of these topics, with high levels of certainty, and these beliefs are highly correlated with each other even when they have no direct connection. Holcombe’s contention that most voter beliefs are adopted derivatively—based on the elites, parties, or movements to which they anchor—both fits the facts and provides a highly plausible account for those facts.

I also find Holcombe’s contention that policies are formed by elites and voters follow the lead of the elites, rather than elites forming policies based upon voter input, to be sound and persuasive. Indeed, I have a hard time understanding how anyone who observes how politics actually works could possibly believe that elites base policy on voter input, or that policies are formed based on voters compromising among themselves as equals. If anything, I think Holcombe may be too generous in his description of how elites interact with voters. For example, Holcombe makes the following observation:

Recognizing that the demand for accurate and detailed information on the part of citizens is low, parties and candidates provide very little information of this type. Platforms are deliberately vague to broaden their appeal. Citizens will find little to disagree with in a vague platform.

But politicians don’t merely keep their policy intentions vague. They frequently engage in false advertising, knowing that most voters are inattentive enough that it will go unnoticed. To use just one example, in the 2008 presidential campaign, then Senator Barack Obama advertised himself as fiercely opposed to NAFTA and loudly proclaimed his intentions to undo this policy. Meanwhile, his main economic advisor, Austin Goolsbee, was quietly reassuring the Canadian government that this was all just political blustering and there were no real implications for policy. And upon winning the election, President Obama took none of the actions on NAFTA he had campaigned on.

Overall, I found Following Their Leaders to be a solid and important work. And as I mentioned in the first post of this series, my summary is no substitute for reading the book itself. However, I suspect the validity of the Holcombe’s argument also suggests why the argument will not find much traction. A key point Holcombe makes throughout the book is that, to a huge degree, people do not adopt parties based on policy, but instead adopt policy based on parties. Democracy is treated as sacrosanct, and its justness is taken for granted. I suspect that most people don’t come to support democracy because they are persuaded that democratic governments are accountable to the people—instead, they accept uncritically the idea that democratic governments are accountable to the people because it supports their pre-existing belief in the justness of democracy. Refuting the idea that democratic governments are accountable to the people will therefore have little effect. I wish I could end on a less dour note, and I sincerely hope to be proven wrong! But regardless of the impact it will have, Holcombe has written a well-reasoned and important book that deserves to be widely read, and one I can easily recommend.