In 2008, the political scientists Marty Cohen, David Karol, Hans Noel, and John Zaller released a book called The Party Decides: Presidential Nominations Before and After Reform. The basic argument is that, despite the ostensible democratic machinery of state primaries, party elites were the ones who really made the decisions. The authors showed that even after the presidential primary reforms in the late ’60s and early ’70s made the nomination process more reliant on voters, the candidate who collected the most endorsements from party grandees still ultimately locked up the nomination.
A semi-bastardized version of this argument quickly became conventional wisdom among the pundit class. Although The Party Decides itself contained extensive caveats and qualifications (especially concerning its minuscule sample size), the book came to be treated as all but ironclad proof that the voters had no say whatsoever in the selection of their party’s nominee. When it comes to presidential primaries at least, The Voters Are Stupid. They might think they have some say in choosing their party’s nominee—said the wonks, nodding sagely to one another—but in reality, they were merely validating the pre-existing choices of the elite class.
But by 2015, this consensus was melting like snow before a stream of hot urine, as Donald Trump contemptuously bulldozed the Republican establishment and locked up that party’s nomination. Indeed, not only did he casually brush aside unified opposition of nearly the entire Republican elite, but he did it despite having no formal political experience of any kind. It seemed the voters had some kind of a voice after all.
But this changed political context did not spell the end of the Stupid Voter narrative: It merely changed form. Whereas voters were previously deemed stupid because they had no influence on political outcomes, they were now deemed stupid because they had too much influence, influence that thwarted the wise and sensible aims of political elites who otherwise would have governed in the public interest. In 2016, political scientists Christopher Achen and Larry Bartels published a much more ambitious book called Democracy for Realists: Why Elections Do Not Produce Responsive Government. It is, basically, The General Theory of Stupid Voters. It has become the latest conventional wisdom about democracy, garnering near-universal praise in the elite press, from the London School of Economics to Foreign Affairs to the New York Review of Books. The Economist deems it “the most influential recent book on voting.”
At wonk central HQ Vox, it has been cited as definitively proving that “everything you know about democracy is wrong,” namely that “the problem with democracy is voters.” It was cited as “the best book to help you understand the wild 2016 campaign,” because “people in general have cast their votes for no particularly good reason at all, so there’s no reason to expect Trump supporters to be any different.”
There is much to recommend about the book, and many of its assertions are certainly at least partly right. But its general conclusion that voters are basically incapable of reason, and never vote based on ideological considerations — one which has been similarly stripped of caveats in popular discussion — is far too strong, and even dangerous. Voters are often ignorant, but they are not completely insensible to reality. Blinding oneself to that fact could easily knock the legs out from any attempt to confront Trump, and even undermine our very democratic institutions.
First, let’s take a look at the central argument of Democracy for Realists. Achen and Bartels assemble a huge body of evidence to demonstrate that the voting public is vastly ignorant about policy, tends to rationalize pre-existing biases, and blames the incumbent party for things they could not possibly control, like shark attacks. Even when voters can be shown to be making a sort of judgment about political success on the merits—namely, voting the bums out during times of economic crisis—their decision tends to be severely myopic. Voters generally judge economic performance only on the last few months before election day, not based on how the whole last electoral term has gone.
The authors take aim at something they call the “folk theory” of democracy, which they define as the idea that democracy is about simply representing the will of the voters. The “folk theory” of democracy is the animating force behind initiatives to increase popular participation in decision-making by adding ballot initiative and referendum procedures to various constitutions, so that “the people” can have a direct voice on policy matters. The authors demonstrate many severe problems with this sort of direct democracy — most notably, that it is at least as vulnerable to elite influence as any other sort of democracy, if not more so. Ordinary citizens really are not equipped to make decisions on complicated policy questions, and have often shot themselves in the foot by voting down water fluoridation measures and so forth (often egged on by well-funded right-wing extremists).
Illustration by Matt Lubchansky.
The trouble starts with their formal model of the folk theory, which they represent with an elaborate mathematical system descended from neoclassical economics called the “median voter theorem.” By this view, voters select candidates closest to their own ideology, and assuming voter preferences are represented by a single left-right spectrum with two equal-sized peaks, parties will rationally appeal to the median voter directly in the political middle. This predicts that each party will have the exact same centrist platform. The “rationally ignorant” median voter doesn’t have to do anything to see his preferences validated by the political system.
This model was directly based on similar economic models, which take a lot of assumed background conditions, run them through some intimidating math, and produce a result demonstrating that free market institutions automatically produce the best of all possible worlds. Voting, it’s just like buying peanut butter! It’s sort of an appealing notion, so long as it doesn’t make any close contact with reality.
Achen and Bartels blow this theory out of the water, thus defeating their conception of the folk theory of democracy. Most obviously, the parties do not have the same platform and never have, not even during the mid-20th-century period of relative political consensus when this kind of model was somewhat plausible. But since 1980 especially, the idea that the parties don’t have strong and increasingly stark disagreements is prima facie ridiculous.
The authors have a lot of smart things to say about the negative influence economics-style reasoning has had on political science. But they don’t consider the idea that using the median voter theorem to represent the folk theory may itself be misleading.
This can best be seen in their implicit theory of reasoning, which is based on the same neoclassical bullshit. They define it in exclusively individual terms—a fundamental premise of this style of economics. By their lights, political reasoning happens when someone has pre-existing, fully worked-out ideology, and perfect knowledge of how the political system has affected their personal well-being , who then calculates the most rational political decision in terms of their own pocketbook and principles.
It is true that virtually nobody behaves in this way. Many people don’t have a clue what each party stands for, while others are egregiously mistaken about who believes what. But more importantly, Achen and Bartels argue that even very well-informed people tend to rationalize their group identities by adopting whatever the consensus view is—and then argue that, by definition, adopting a consensus view cannot be a “reasoned” decision: “[T]he political preferences and judgments that look and feel like the bases of partisanship and voting behavior are, in reality, often the consequences of party and group loyalties … the more information a voter has, often the better able she is to bolster her identities with rational-sounding reasons.”
There are a lot of problems with the premise of this argument.
First of all, if the most informed people simply adopt the views of their most important identity groups, then where do those groups come by their notions? Presumably, they aren’t just distilled from the celestial ether. It could be a leader simply lays down a party line, which is adopted by rank-and-file partisans regardless of content or hypocrisy. That is perhaps a plausible picture for Republicans, who now apparently hate the FBI and love Vladimir Putin, but is it universally true of all voting blocs?
Take African Americans, for instance. Such people vote almost in lockstep for Democrats (routinely at over 90 percent), a fact which is repeatedly mentioned by Achen and Bartels. Blacks have tended to support Democrats since the 1930s, but not by such huge margins. In 1960, for instance, John Kennedy racked up only 68 percent of the black vote.
But in 1964, Lyndon Johnson racked up 94 percent. The reason, obviously, is that Johnson used his spectacular legislative legerdemain to pass the Civil Rights Act in July 1964, which his opponent Barry Goldwater opposed. Black voters made a collective decision that Johnson was genuinely committed to their interests while Goldwater was pushing disingenuous Dixiecrat politics, and shifted their votes accordingly.
It is frankly ridiculous to argue that there wasn’t at least some genuine reasoning about policy going on here, no matter where the change in opinion is coming from. Even if rank-and-file blacks were simply blithely taking marching orders from civil rights leaders (and if you read some histories of the civil rights movement, they clearly were not), it was still rational to follow that instruction. The parties really had made a profound ideological shift, and it was correctly identified by virtually every black voter.
That may be why the only time Achen and Bartels discuss black re-alignment into the Democratic column is during the New Deal, when Democrats were still the party of Southern white supremacists:
Was that sudden political mobilization inspired by support for New Deal policies, or by the Roosevelt administration’s cautious outreach to African-Americans, signaling that they, too, would be a welcome part of the Democratic coalition?
The authors are so committed to their stupid voter shtick that they generally assume that all political group loyalties must be the dumbest caricature of identity politics—policy-free group signaling, affective demonstrations, and so on. As if outreach to black people can’t be a policy. In reality, in 1941 President Roosevelt issued an executive order prohibiting racial discrimination in defense industries as part of the war effort—the start of a pro-black Democratic policy trend that culminated in the Civil Rights Act and the Voting Rights Act.
What’s most ridiculous about this idea that group behavior can’t be rational is that all thinking is collective in some way—and this reality affects elites as much as it affects voters! Every person learns a group language as a child, develops views heavily influenced by their parents and community, and generally attends some sort of church or government school. Of course all political thinking is mediated and influenced by various groups. That is what human reasoning is.
That’s as true for academics as it is for anyone else. Scholars build on findings and models developed by those coming before them, building on their successes, filling in the gaps, or overturning mistaken theories. Achen and Bartels themselves, naturally, cite a vast quantity of previous research. Voters may not fit the homo economicus model of rationality, but neither does literally anyone else who has ever lived. The authors rule out voter reasoning by foreclosing the only process by which political reasoning has ever actually happened. If all humans were held to the same intellectual standards that the authors hold “voters” to, we would all be gibbering idiots.
Achen and Bartels’ revisionist history of the New Deal period exemplifies how much the “stupid voter” framework can distort the lessons we draw from the elections and policy initiatives of the past. The traditional history of the New Deal era sees it as an unusually ideological period, in which voters were convinced to adopt radically more left-wing views. With Herbert Hoover-style conservatism profoundly discredited by the Great Depression, New Deal liberalism eventually swept in to take its place. (Incidentally, this is one period where an election unquestionably did produce responsive government, but I digress.) But Archen and Bartels argue that actually, these sweeping Democratic majorities were largely due to fortuitous election timing.
To be fair to the authors, they do present fairly compelling evidence that voters don’t make fair judgments of a whole presidential term, but instead short-sightedly consider only the quite recent past. FDR’s sweeping election victories were heavily—though not entirely—correlated with local income growth during just that period.
Republicans got swept out due to the Depression in 1932, while things were turning up fairly well right before election day in 1934 and 1936, leading to sustained huge Democratic majorities. Then, in 1938, the election took place during another bad recession. Democrats lost badly, and returned to their spending policies. Eventually, after many years of big Democratic victories, the voting public then adopted the ideology of its new masters: “retrospective judgments were … incorporated into durable partisan attachments[.]”
But stripped of the quasi-neoclassical presumption that reason only occurs individually, this mass shift in party loyalty looks less like the rube public being jerked around by circumstance, and more like a reasonably successful example of social political reasoning. Voters made a judgment in November 1932 that the economy was in severe crisis—a bleary, instinctive one, but a real and correct one nonetheless—and handed control of the government to the Democrats. A vast galaxy of interest groups, intellectuals, and politicians within the party then conducted a sharp debate about what should be done. Perhaps most importantly, FDR agreed that whatever it was, it needed to be big and aggressive.
The government conducted various policy experiments to fix the economic crisis, which managed to restore reasonable prosperity in time for the next election. Some efforts like the National Industrial Recovery Act were abandoned as quasi-failures, while others like the Works Progress Administration and Public Works Administration were celebrated as performing as advertised. Voters then ratified the leftward policy turn by returning the Democrats to power. Many reasoned correctly—a sort of thought shortcut, but a correct one nonetheless—that these New Deal Democrats must be on to something, and adopted the new ideology.
A particularly howling absence from Achen and Bartels’ account is that of unions, which were one of the key props of the New Deal coalition. Democrats passed a big expansion of labor organizing rights in the National Industrial Recovery Act in 1933, and a comprehensive organizing framework with the National Labor Relations Act in 1935. Union membership then exploded, and those new union votes went (with perfect collective rationality, just like that of African-Americans in 1964) overwhelmingly to Democrats.
On the other hand, the recession in 1937 was directly caused by a return to conservative austerity policy, namely trying to balance the budget long before full employment was reached. Voters lost trust in the Democrats, and their votes swung accordingly. The big 1938 loss harmed Democrats in Congress, but it also reinforced New Deal factions within the Democratic coalition by directly demonstrating that conservative policy causes economic disaster, and hence political losses.
Voters punishing parties at the ballot box for the adverse results of their policies is a critical part of the democratic reasoning process, because elites can be every bit as stupid as the most pig-ignorant voter. Achen and Bartels do note on a couple occasions that their analysis applies to everyone up into the elite, including “the authors of this book.” But nearly the entire book is dedicated to the irrationality of voters, and they repeatedly suggest that elites can be trusted to reason more effectively: “We need to learn to let political parties and political leaders do their jobs, too.” In their chapter on the downsides of voters’ economic myopia, they construct a neoclassical-style model of a competent legislator, showing that the greater the random economic fluctuations, the greater the chance that a better-than-average legislator will be turned out by accident.
Similarly, they reference some tentative evidence demonstrating the possibility of a “political business cycle,” or the possibility that the party in power will cynically—but rationally—juice the economy just before election day, instead of a more consistent approach. They note that since the 1936 publication of John Maynard Keynes’ book The General Theory of Employment, Interest, and Money, politicians have had a workable framework for carrying that out.
But they do not discuss the many instances in which rube politicians have legislated themselves out of office by making gigantic, obvious policy mistakes. The most common and devastating is pushing austerity in a depression, or failing to juice the economy before election day. As discussed, Democrats did this in 1937. Perhaps that is understandable the year after Keynes. But a large swath of the party did the exact same thing in 2010, stymieing additional Keynesian stimulus after March 2009 over fears about the budget deficit. Those conservative Democrats were mainly situated in more vulnerable districts, and most of them were turfed out in the wave election that year.
The election of Republicans did not, of course, lead to a more sensible Keynesian approach. One of the many, many anti-democratic characteristics of America’s political system is that it doesn’t provide a broader menu of political options. Absent the two-party duopoly, we could theoretically, after a crisis, choose between a number of possible policies to ameliorate the disaster; but since there are only two choices, the rebellious instinct during times of economic duress redounds to the benefit of the out-of-power party, almost regardless of what they advocate.
But still, all that doesn’t change the fact that voters made a correct collective judgment that Democrats had screwed the pooch on the economy.
This question of elite stupidity gets almost no attention from Achen and Bartels. It’s a glaring absence, as it is at least as important as voter stupidity, if not more so. Ordinary citizens, it’s true, don’t have time or training to develop super-detailed or consistent policy views. But the whole supposed point of having a party system is that people will vote for the party most in line with their political instincts, and that this party will then train up an intellectual elite who can translate those instincts into workable policy. And yet very often, parties colossally fail to do anything of the kind. For example, Democratic politicians pushing for austerity when unemployment is 10 percent and their party has the majority — as most centrist Blue Dogs did in 2010, not to mention President Obama himself — is like a fire department mistakenly filling its tankers with gasoline. How to get party elites to actually follow through on the mandates they were elected on is at least as important to democratic functioning—and maybe even more important—as figuring out how to get voters to choose the right mandate.
Legitimizing the myth that voters have no actual beliefs or preferences is not only intellectually sloppy—it’s politically dangerous. One clear implication of the book’s argument is that neither ideology nor policy matters politically, because no voter actually understands anything. During the 2016 primary, Achen and Bartels argued exactly this about Bernie Sanders supporters. They cited (far from comprehensive) polling data showing his supporters somewhat to the right of Hillary Clinton supporters, concluding that his surprisingly strong challenge was merely the “reflections of social identities, symbolic commitments and partisan loyalties.”
As we have seen, this extreme view is clearly wrong. But worse, it leads directly to an enervating political helplessness. For instance, there is a strong case that Donald Trump got considerable political mileage out of railing against free trade deals in parts of the Rust Belt that were devastated by de-industrialization—suggesting that a more class-based campaign could peel off some of his voters and win the 2020 election, following the victorious Obama campaign strategy of 2012. But if we adhere to the view that policy appeals are always useless, then there is no point in even trying to win over any marginal Trump voter—or the vastly larger population of nonvoters—through any sort of rational appeal. The result of this belief is strategic inertia: It means fatalistically surrendering whole swathes of voters to the Republicans, without even attempting to appeal to their legitimate interests and anxieties.
Of course, virtually no one is going to be won over by brandishing a fully worked-out 2,000-page trade policy reform. But to foreclose all policy-based appeals—even those properly situated and framed with group- and identity-based signals, as would be indicated by an intelligent consideration of the research cited by Achen and Bartels—is to foreclose a potentially decisive political weapon, and possibly cede another term to Trump.
To their credit, the authors come back from the edge by the end of their book. They do not stoop to the odious libertarian arrogance of Jason Brennan in Against Democracy or Bryan Caplan in The Myth of the Rational Voter to say that democracy is basically bad in itself. They say that opportunities for voter choice and education should be expanded, not abandoned. Their call for reducing social and economic inequality is particularly welcome.
But their hyperbolic, elitist thesis—which wouldn’t have gotten half so much attention if it was more realistic and less contrarian—is still wrong. If the Democratic Party wants to get rid of Trump, it should remember that voters are not gormless sheep. Just because voters are not policy wonks or cool-headed logicians does not mean that they are categorically incapable of perceiving which party is (or isn’t) looking out for their interests. Party elites who blame voters for our country’s political ills would do better to look to some of their own egregious failings, and figure out what it would actually take to build a party worth voting for.
This article was originally published in the January – February issue of Current Affairs.
If you appreciate our work, please consider making a donation, purchasing a subscription, or supporting our podcast on Patreon. Current Affairs is not for profit and carries no outside advertising. We are an independent media institution funded entirely by subscribers and small donors, and we depend on you in order to continue to produce high-quality work.