clock menu more-arrow no yes mobile

Filed under:

Two eminent political scientists: The problem with democracy is voters

Why almost everything you think about democracy is wrong.

Shutterstock

Consider the curious case of New Jersey in 1916: That summer, there was a string of deadly shark attacks along the Jersey Shore. As a result, Woodrow Wilson lost his home state in the presidential election.

Why, you ask? Because the beachfront towns (which rely on tourism) were negatively impacted by the attacks. Though Wilson wasn’t responsible for the hungry sharks, he was the incumbent, and people vote against incumbents when things are bad.

This is a story political scientists Christopher Achen and Larry Bartels tell in Democracy for Realists, in service of a sobering thesis: Voters don’t have anything like coherent preferences. Most people pay little attention to politics; when they vote, if they vote at all, they do so irrationally and for contradictory reasons.

The book lays waste to a reassuring theory about democracy that goes something like this: Ordinary citizens have preferences about what the government ought to do; they elect leaders who will carry out those preferences and vote against those who will not; in the end, we’re left with a government that more or less serves the majority.

Even voters who pay close attention to politics are prone — in fact, more prone — to biased or blinkered decision-making. The reason is simple: Most people make political decisions on the basis of social identities and partisan loyalties, not an honest examination of reality.

“Election outcomes,” Achen and Bartels conclude, “turn out to be largely random events from the viewpoint of democratic theory.”

If Achen and Bartels are right, democracy is a faulty form of politics, and direct democracy is far worse than that. It virtually guarantees that at some point, you’ll end up with a grossly unfit leader.

And that, of course, is what we now have.

Last month, I sat down with Achen and Bartels to talk about their critique of democracy, and how it might explain our current political predicament.

A warning: This is a long and sweeping discussion, touching on a number of thorny issues. I offer pushback when and where I can. But to be perfectly honest, I found many of their criticisms of democracy compelling. Which is not at all what I had hoped.

I want to believe in those comforting myths about democracy, mostly because the alternatives are worse. But even if democracy is the least bad form of government, we still ought to know how it works and, more importantly, how it doesn’t.


Voter behavior is erratic and incoherent

Sean Illing

What actually drives voter behavior?

Christopher Achen

I think people are doing the best they can. They just don't have a lot of information, and so they substitute guesses and views of the world that make them feel comfortable. I think people are looking for ways to make sense of what is a very complicated reality out there. It's hard for those of us who get paid to think about it all the time to make sense of it, and it's very hard for people with a lot of other demands in their lives.

So they're doing the best they can but, as we said in the book, we think that we need institutional structures that would get them some help and do what the Federalist Papers suggest should be done, which is to have a popular voice in government but to supplement it with the opinions of people with more expertise and more experience.

Larry Bartels

One of the more important building blocks for our work was [Austrian-born economist and political scientist who wrote mostly in the early 20th century] Joseph Schumpeter’s work on democracy. As an economist, he emphasized more clearly than people had previously the significance of the distinction between economic life, where people make choices that directly affect their own well-being (i.e., you stop buying products that you don’t like) and the political world, where the connection between individual behavior and the outcomes that I experience is so indirect that it almost makes no sense for me to try to perceive instrumentally.

It turns out, when it comes to political outcomes, most people are not making rational decisions based on the real-world impact they will have on their life, in part because they just don’t know.

So much of politics, not surprisingly, turns out to be about expressive behavior rather than instrumental behavior — in other words, people making decisions based on momentary feeling and not on some sound understanding of how those decisions will improve or hurt their life. And so if you think about people using the democratic levers that they have available to them to express themselves, rather than to make instrumental choices, you're probably more often than not going to be closer to the actual psychology of what they're up to.

Sean Illing

Does the average voter even have what we might call policy preferences?

Christopher Achen

Well, they do adopt some. They take in some information. So with the Trump phenomenon, for example, people clearly recognize that for certain identities, he was a vocal spokesman for those identities, and they did learn that. And the combination of that and his familiarity from reality TV and so on made him successful in their minds at being the kind of leader they were looking for. And I think people are pretty good at that, actually. They're pretty good at picking out who's on their side.

Sean Illing

To be clear, they’re good at picking people who appear to be on their side, who play the right rhetorical game.

Christopher Achen

Good point. Now, what they're much less good at is thinking about whether it makes any sense to build a wall across the southern border with Mexico. Is that going to solve the problem? How much is that going to cost? Is that how I would want to spend my money? Voters tend not to think about these sorts of questions very well, and their incoherent and shifting positions suggest as much.

Sean Illing

In graduate school, I read a book called The Macro Polity, which was published in 2002 by political scientists Robert Erikson, Michael Mackuen, and James Stimson. The thesis was something like: voters as a collective aggregate tend to act with purpose and predictability, even though most individual voters do not.

So the idea was that the vast majority of voters are capricious and uninformed, but in the end they tend to cancel each other out. What pushes elections in an intelligible direction is the minority of educated and engaged voters. But your book claims that elections are basically a “coin toss.”

Larry Bartels

Well, that argument, which goes back to the 18th century, works pretty well so long as the errors and political choices are distributed equally on both sides of whatever the option is. But in many cases, that's unlikely to be true. But we don't want to say that politics is essentially random. There are lots of elements, certainly of presidential elections, that are quite predictable. We have this little picture in Chapter 6 of the book about the relationship between economic conditions and how long the incumbent party has been in office and the election outcome. That suggests a good deal of predictability:

Here’s a graphical summary of the historical relationship between short-term income growth, incumbent party tenure, and presidential election outcomes. Election years are arrayed along the horizontal axis on the basis of growth rates in real disposable personal income per capita in the six months leading up to each election, adjusted for how long the incumbent party has held the White House. The vertical axis represents the popular vote margin (in percentage points) for the incumbent party’s presidential candidate.
Javier Zarracina/Vox

But the point is that if the underlying conditions suggest that the election is likely to be close, as has been the case for most recent presidential elections, then the small factors that determine the outcome at the margin are likely to be random and in any case aren't the kinds of things that democratic theory focuses on. Like, for example, people's preferences with respect to policy issues.

Christopher Achen

The points that Larry just made about the correlated errors are the root of the problem. Those theorems don't go through if people all tend to make the same kind of mistakes at a given time. And there's plenty of that, right? If you look at shares of the two-party vote in the 20th century, the presidential candidate who got the biggest two-party share was Warren Harding, now widely regarded as the worst president in American history. But people thought he was charming. And they all made the same mistake at the same time. And that's the kind of thing that's just fatal to that particular argument you find in The Macro Polity.

Sean Illing

I take your broader points about voter behavior, but let me push back just a little bit. Plenty of people will say: Okay, the average voter is misguided in all the ways you note, but there are still heuristics or cognitive shortcuts that voters use to make sense of the political world. For example, they choose the candidate who most aligns with their ideological worldview or they rely on signals from party leaders or trusted authorities. It’s not perfect, but it does allow voters to make more or less reasonable decisions.

Larry Bartels

When that kind of argument has been operationalized, it usually takes as given the basis on which people are choosing what heuristics to rely on. And that itself is quite problematic. If people knew where to go to get the informative cues about questions they need to answer, they probably would do better. But most of the empirical evidence we have suggests that it's only the more interested, better-informed people who have enough context to be able to interpret those cues appropriately and to use them sensibly.

Voters don’t understand cause and effect

Sean Illing

There’s a model in contemporary political science called the “retrospective theory of voting,” which basically says that elections aren’t really about ideas or policy preferences, but voters are nonetheless able to accurately judge the performance of incumbent administrations. What’s wrong with this model?

Christopher Achen

Well, as an empirical matter, we think it works pretty well, and that's why political scientists have increasingly fastened on it as an alternative to the more idealistic notion of an issue-voting populace decision-making. The downside is that in their enthusiasm to fasten on it, they've attributed a kind of normative sheen to it that we don't think is warranted by the quality of the decisions that result.

Sean Illing

Can you clarify what you mean there?

Christopher Achen

Well, people aren't very good at attributing the implications of the decisions that are made by policymakers to those policymakers. In order to do this efficiently, they'd have to have a pretty canny understanding of how the behavior of any particular elected official or party contributes to the good or bad outcomes they experience. And they're not very good at that. They'd have to take a long view of the effects of policy, and they're not very good at that. So they make these retrospective judgments, but in a kind of haphazard way that doesn't seem to promote accountability in the way that political scientists would like to think it does.

Sean Illing

So you found no evidence to suggest that voters understand cause and effect in any coherent way?

Christopher Achen

Again, we're saying, yeah, the world is immensely complicated, so to say that one has a good understanding of cause and effect in this domain would really be asking quite a lot of people. Certainly, economists don't agree about the impact of economic policies on the well-being of individuals or groups. And so what we do in the book in order to try and get around those difficulties is to focus on some cases where it seems pretty clear that the incumbent politicians ought not to be paying for people's bad fortunes and find that even those cases, there seems to be a pretty systematic pattern of punishment.

Even in the aggregate, voters aren’t rational

Sean Illing

One of the many assumptions your book undercuts is this idea that large groups of voters can deliberate reasonably — that if you give people the appropriate information and allow them to exercise judgment, they will do so more or less intelligently. Or at the very least, they won’t steer the country into an abyss.

Larry Bartels

I think it's hard to see how the public as a whole would steer the country in any particular direction. Usually when we think about public input, we think about public input in response to particular kinds of choices that have been framed by political elites of one kind or the other, whether they're party leaders or elected officials. And whether people come to the right conclusions about the choices that are offered to them, I think this is most of what is interesting and consequential, which is how the choices get framed in the circumstances under which people are allowed to have input into deciding what path to take.

Sean Illing

How do choices get framed? How do opinions get formed?

Larry Bartels

A lot of it is people simply taking cues from political figures, from public figures, that they've identified themselves with one way or the other, whether they're party leaders or the leaders of social groups or interest groups that they feel some attachment to. If, for example, you look at the change in views about Russia that we've observed after Trump made admiring comments about Vladimir Putin, you might think, given the history of the US and Russia over the past century, that people would have pretty ingrained views about what they think of the Russian system. But that turned out not to be the case.

People's views shifted pretty quickly and pretty dramatically in the wake of fairly casual elite cues they were receiving. But from somebody who they trusted and whose cue they were happy to take about something that they hadn't really thought much about. The relationship between Russia and the US is a pretty important thing, but the ordinary American hasn't spent a lot of time thinking about how they should think about that. Indeed, they don’t spend much time really thinking about political issues of consequence. This idea that people have fixed or informed views about central issues doesn’t square with most of the data we have.

Sean Illing

Can you give me some examples of the data you used or case studies you analyzed? What are you basing these claims about voting behavior on?

Larry Bartels

A substantial scholarly literature on electoral politics and public opinion has accumulated over the past several decades, and we relied on it heavily. But we also added original statistical analyses designed to test our arguments in particularly illuminating times and places.

For example, the notion that voters blindly reward or punish incumbent presidents for good or bad times led us to the presidential election of 1936; political scientists have portrayed that election as a historic ideological mandate for Franklin Roosevelt’s New Deal, but we found that Roosevelt’s support hinged crucially on how much incomes grew in each state in the year leading up to the election.

The same logic led us to the Jersey Shore in 1916, where a dramatic series of shark attacks — hardly something a president can control — turned out to produce a significant dent in the vote for Woodrow Wilson.

To disentangle the effects of policy preferences and social identities, we examined support for John Kennedy in 1960 — an unusually pure case of religious affinity (and animosity) with little or no real policy content. Our analysis in that case was bolstered by the fact that repeated interviews with the same people allowed us to relate voting behavior in 1960 to measures of the strength of voters’ Catholic identities from 1958, before Kennedy emerged as a candidate.

In the same spirit, our analysis of the impact of partisanship on views about abortion employed a decades-long study of changing political attitudes to show that more than half of Democratic men who expressed pro-life views in 1982 were pro-choice by 1997; the corresponding rate of change among Republican men was less than 30 percent. Political attitudes and behavior are enormously complex, and so we are shameless opportunists, delighted to exploit clear glimpses of underlying patterns and processes wherever we can find them.

Shutterstock

Strong parties won’t solve the problem of democratic control

Sean Illing

The parties — in the United States, at least — seem to have lost much of their control over the process. At the same time, we’ve seen a spike in partisanship. Do you think a stronger or different party system can correct some of the fundamental problems you’ve identified?

Larry Bartels

I wouldn’t say a “spike” in partisanship — we’ve had a long, gradual increase in the intensity of partisan loyalties over the past 35 or 40 years. As with many aspects of the contemporary political system, that’s probably a sort of “return to normalcy” following the unusual period in the middle of the 20th century which saw the breakdown of the New Deal party system, a loosening of partisan attachments, and electoral landslides in both directions.

I wouldn’t say that we’ve had “a decline in the power of the parties,” either. The parties in Congress are immensely powerful. All of President Obama’s big legislative accomplishments — the stimulus bill, the Affordable Care Act, Dodd-Frank — were passed with few or no Republican votes. Conversely, President Trump has so far relied entirely on the Republican Party for support, despite his railing during the campaign against the political establishment on both sides of the aisle. When it comes to fundraising and candidate recruitment, the party committees and their allied networks of interest groups have a national reach and ideological cohesion that 19th-century party boss Mark Hanna would have envied.

What we don’t (and never did) have is a way to manage the parties “democratically.” We’ve tried to give as much power as possible to ordinary citizens voting in primaries, but the task is too complex — too many candidates and issues, too little information, too much strategic uncertainty — for that to work reliably. We saw that in 2016. But we also saw a breakdown of the alternative theory of democratic control, in which general election voters keep the parties on a tight leash by refusing to support them if they happen to nominate someone who is too “extreme” or uninformed or unstable.

History clearly demonstrates that democracies need parties to organize and simplify the political world. But parties don’t make the fundamental problems of democratic control disappear; they just submerge them more or less successfully. When professional politicians are reasonably enlightened and skillful and the rules and political culture let them do their job, democracy will usually work pretty well. When not, not.

Christopher Achen

I would add only that I do favor stronger parties in one sense — that they should have more control over their nominations, both presidential and other. New Jersey does this well, actually, with the "party endorsement" being quite crucial in winning primaries here. We also make it hard to put an initiative on the ballot, and we have traditionally had a strong, competent state supreme court.

The result is good schools, no measles and mumps outbreaks because of low vaccination rates, and much else by way of good government because it's harder for the voters to harm themselves here than in other states where interest groups and nutty ideas have more control through unrestricted primaries and initiatives. Of course, New Jersey has lots of problems, but a bad state governmental structure isn't among them.

All politics is identity politics

Sean Illing

I want to ask you both about identity politics, which is a big part of your book. As you know, there’s a debate right now about the role and utility of identity politics. Your book argues, somewhat controversially, that in democracy all politics is identity politics, and that it’s foolish to pretend otherwise. Is that a fair reading of your thesis?

Christopher Achen

We do think that identities are fundamental. Now, in politics, no one framework is everything, and of course there are some other things going on. But we do think that identity is fundamental. The old left argument is that it's about class and that race and gender are side effects primarily of class issues. But class identification, working-class consciousness, and all of that framework, those are identities as well. So from our point of view, the proposal is to substitute one set of identities for another. That's a plausible argument, but the idea that you can propose policy, economic policy proscriptions, in a social vacuum with no attention to the other identities that are at work, that is something we just don't believe.

Sean Illing

So democratic elections, on your view, are essentially just a competition to see who can activate the most identities among the voters?

Christopher Achen

I would say there's a variety of identities people have that are more or less salient and can be made more or less salient politically. For many people, the principles become part of the identity and are important moving parts of the way they think about politics. But our claim is that the identities are more fundamental, the principles come later rather than the other way around.

Too much democracy or too little?

Sean Illing

There have been two broad reactions to the Brexit vote and to Donald Trump’s election. On the one hand, some believe we have too much democracy, too few barriers between popular will and the application of power; on the other hand, some argue that we have too little democracy, that we’re witnessing a righteous backlash against an anti-democratic and rigged system. Where do you come down?

Larry Bartels

The notion that the last election cycle somehow brought out a different kind of person or a different aspect of people's political character is misleading from our point of view. We didn't have Trump in our sights at all as we were working on this book over the course of 15 years, but I think the spirit of the book very much suggests that these kinds of things are likely to happen in democratic systems from time to time because of the way they work and the limitations they face all the time.

But what really triggers the kind of problems that people were concerned about in 2016 are mostly of elite-level actions, how the parties behave and what kind of messages they present to people and what kinds of alternatives they present to people. And so the idea that the American people are somehow different than they were five years ago or nine years ago I think is kind of mistaken. But the interaction between the elites and the masses is where the issues are, and I think much of political writing is not very well-suited to dealing with those kinds of interactions because the role of elites isn't very integrated into the overall way of thinking about what's going on. They are almost by necessity a kind of illegitimate piece of the system in a lot of popular thinking about the way politics works.

Sean Illing

You mention the problem of elites, and that really is a key dilemma in your analysis. It’s not so much about greater mass participation, which doesn’t necessarily make things better, as it is about getting elites to not rig the system in their favor.

Larry Bartels

Absolutely. If you think about democracy in the terms we prefer, you might say the biggest limitation at the moment is that we don't know how to incorporate the role of political elites in a constructive way into the governing process or to somehow make it possible to ensure that they're working on behalf of the interests of ordinary people.

How to make democracy work better

Sean Illing

The book calls for a return to something like a group theory of democracy, which amounts to a reluctant embrace of identity politics. What, exactly, are you advocating here, and why is it a smarter alternative to the folk or populist theory of democracy?

Christopher Achen

Partly what we want to do is to think about a variety of reform proposals that are out there and floating around from this framework. So, for example, making the presidential nomination process more "democratic" by getting rid of superdelegates is exactly the kind of blunder that we're criticizing.

Having the conservative majority on the Supreme Court say that there's a marketplace of ideas here, that there can't possibly be anything wrong with people publicizing their own ideas about policy proposals and candidates and so forth, and so therefore limitations on campaign spending are a violation of free speech. That, too, is the kind of blunder that we're criticizing.

Sean Illing

Okay, but where does that leave us?

Larry Bartels

We have to get out of this overly simplistic framework and create more sensible policy proposals. It seems clear to us that a lot of the actual ways in which people of ordinary education or ordinary means or just not much power, the ways in which they are disadvantaged are often occurring at the level of policymaking rather than at the level of elections themselves. The financial sector, for instance, is having a lot of policy success in Washington, in ways that ordinary people, if they really understood what was happening, would not approve. But they don’t follow it closely enough, they don’t understand, and the policy process is tilted toward moneyed interests that ordinary people have no chance.

So focusing reform on the places where the real problem is occurring as opposed to making fanciful proposals that ask us to do what none of us is really able to do. That's the kind of emphasis that we want to direct people's attention to.

Sean Illing

Yes, but how do we get there? Again, the book latches onto this group theory of democracy as a more intelligent means of channeling popular passions, yet it’s not clear to me what this vision of democracy looks like or how it’s very different from what we now have.

Christopher Achen

Well, in some ways it doesn't look so different from how politics works now. The policymaking president is heavily influenced by groups of all kinds. And some of those are identity groups, like African-American groups or Hispanic groups or LGBT groups. Others are occupational groups and religious groups and vocational groups, like the National Rifle Association. And so on. And so we already have a policy process driven by groups. And the idea is more than 100 years old. It's certainly not original with us. So I think our view is to not pretend that there's a magic wand that can make people uninterested in the groups to which they belong, but instead to direct reform proposals to instead making the balance among these groups fairer than it is now.

Sean Illing

So it’s your view that we ought to accept the fact that we think, act, and judge in terms of group affiliations of one sort or another, that we’re tribal creatures, and that our political system has to account for this structurally?

Larry Bartels

I think that’s very well put except for the term “tribal,” which implies that people are members of only one tribe at any given point. I think people's identities are complex and the way they bring them to bear in politics is complex, and that's part of what needs to be understood by us as analysts and also built into any idea of how one might use these groups attachments constructively in improving the political process.

Christopher Achen

I reacted exactly the same way Larry did. One would not say to African Americans, for example, that the experience of their lives in a heavily white society and the way in which that shapes the things that they want from politics, that that's a tribal attachment and that they should get over that and just think about economics, right? It kind of misses the point.

Sean Illing

Oh, I agree, I’m just trying to nail down your thesis.

Christopher Achen

Of course. To be clear, these group attachments are not some bad thing we do instead of being rational, well-informed creatures. They constitute who we are. You know, evangelical Christians don't regard themselves as a tribe. They have a way of thinking about their lives that makes sense. And secular people have a parallel set of views that makes sense of their lives. And so we all do this. We construct an interpretation of our lives, and we're loyal to that and we find other people with similar views. That's what human beings are like, and recognizing that seems to us a big step forward from the way we tend to think about politics now.

Sean Illing

So someone shouldn’t walk away from this book thinking democracy is too idealistic for high primates like us?

Larry Bartels

Well, democracy is happening; it's just not the kind of democracy that we hear about in Fourth of July speeches. So our complaint is not so much about democracy as it is about our misleading understanding of democracy and the bad implications it has for how we proceed democratically.

Christopher Achen

I think our reply to someone who walks away with that impression would be: What is it about the ideas of the American founders that you disagree with? Because that is the position that's being taken here. And as we say in the book, the response is often, well, a lot of them were slaveholders, and we don't have to take their ideas seriously. OK, yes, they were. But to not read them at all is a recipe for ignorance and not the one that we believe.

Sign up for the newsletter Today, Explained

Understand the world with a daily explainer plus the most compelling stories of the day.