by Aaron Swartz
The candidate, like most people, has never purchased a poll before and so has no idea what they actually cost. And the pollster never discloses the actual amount of their markup. If they’re ever questioned about a discrepancy in price, they point to all sorts of difficult-to-measure factors. “Oh, our polls cost more because they’re conducted by specially trained operators—because we work with you to develop the most scientific questions—because we put a lot of effort into properly interpreting the results.” These claims never stand up to even basic scrutiny (the operators are poorly paid temps, the question wording violates basic principles of professional practice, the results are incorrectly calculated through spreadsheets so bad the pollsters must be borderline innumerate), but in the rush of a campaign who has time for this kind of investigation? And who’s going to look a gift horse in the mouth—they’re doing this at cost, remember?
The mail and TV and other consultants play exactly the same game, each with slightly different lies and gimmicks, but the pollster has special influence because of their control over “the evidence.” Their supposed expertise is not in any particular aspect of campaign tactics, but in that most basic question: what it is the people actually want. And by controlling that, they can come to control a great deal of campaign strategy. As a result, the pollster is usually first among equals in these strategic councils.
Observers of the political scene often complain about the high-tech calculations and incredible brainpower that goes into properly packaging a candidate.* But in reality it’s difficult to overstate the general level of incompetence. Political consultants are largely shielded from market competition by the tribal instincts of politicos. If you were to buy mailers for a commercial company, you’d talk to different print shops and compare their rates and reviews. But if you’re a left-wing political candidate, you cannot go to a standard print shop—you have to go to a political print shop. And certainly not a Republican print shop or even the standard Democratic print shop, but one subdivided to cater to your specific political grouping (left-wing vs. centrist, moderate Republican vs. Tea Party). After all, who wants to support the enemy?
Chomsky, Noam. March 12/13, 2005. “The Toothpaste Election.” Counterpunch.
“Up, Simba!” in Consider the Lobster: And Other Essays (Little, Brown). ISBN 0316156116.
Of course the market for left-wing Democratic political candidates is pretty small, so there’re not many prospective competitors eager to home in on the business. And even if they do, these firms’ marketing departments largely consist of going to the right cocktail parties. Particular cliques (e.g., left-wing Democratic electoral activists) all tend to know and recommend each other, in that way that loose social circles do, where genuine good feeling towards acquaintances merges with good business sense.
But the biggest problem is that the scientific basis for their vaunted and expensive expertise is practically zero. Psychologists have long recognized that to become an expert at some skill, you need a great deal of practice with rapid feedback.† There are lots of expert basketball shooters, because when you miss a basket you know right away and can adjust your shot next time. There are very few expert long-term economic forecasters because your forecast comes true months or years after you make it, when you’ve long forgotten what it was you did right or wrong.
Ericsson, K. Anders, Neil Charness, Paul J. Feltovich, Robert R. Hoffman. June 26, 2006. The Cambridge Handbook of Expertise and Expert Performance (Cambridge University Press). ISBN 0521600812.
Expertise in politics is much more like prediction than basketball. At the end of an election, you get basically just one bit of information: you either won or lost. And it’s easy for everyone involved to (rightly!) point to circumstances outside of their control. The candidate didn’t take their advice, the strategy was derailed by a late-breaking scandal, the campaign didn’t have enough money to fully execute on the plan, etc.
When you suggest a candidate emphasize a particular issue or put a particular photo on their mailer, you’ll simply never know whether you were right or wrong. The candidate will succeed or fail months after you make a decision, there’s no way to measure how much an individual decision affected the results, and even if you were somehow the only one responsible for a candidate’s entire campaign, the results could always have been skewed by some surprise in the news or some fluke of your opponent’s.
But the end result is that nobody ever learns from their mistakes, and without learning there can be no real expertise in politics. So, in the absence of real knowledge, practitioners naturally tend to believe in themselves and their products. The TV consultant insists what’s needed is more and different TV commercials, the mail consultant argues late-campaign mail has a proven effect, the targeting consultant says we need to spend more money on targeting to make sure our other dollars aren’t being wasted, and so on. The result is that campaigns get very expensive very fast.
But most expensive of all is the fund-raising consultant. Because fund-raising directly involves money, fund-raising consultants are able to set incredible prices—rates like 1/3 of all money they raise. In part, this is because fund-raising consultants are in more demand than any of the other consultants—the thing you do on day one of a campaign is not buy TV ads or conduct polls, but raise money, so even campaigns that never get off the ground need fund-raising consultants. But also, these rates seem to be justified by improperly specified hypotheticals. The candidate thinks, “If I didn’t hire a fund-raising candidate, I’d have no money—so what’s wrong with giving up a third of money I don’t even have?” instead of “How much of that money could I raise without a consultant?”
It’s true that the best fund-raising consultants have connections to networks of wealthy donors. In the same way that television consultants and pollsters advise on how to market the candidate to the public, these fund-raising consultants advise on how to market a candidate to the donor scene. They know what issues different wealthy people especially care about, how to talk to them, and they can often set up meetings to pitch a candidate to wealth.
The wealthiest don’t meet with candidates directly, of course, but have full-time professionals who advise them on their giving. These professionals typically advise an entire wealthy family or support a wealthy person who acts as a kingmaker themselves. Some wealthy people have more interest in politics than others: they like to vet candidates themselves and recommend them to their circle of less politically engaged wealthy friends.
And there are the classic fund-raisers, as mentioned above, where you persuade a circle of existing donors to invite their own social networks to a party at one of their houses so you have a chance to woo their friends.
But the vast majority of fund-raising is much simpler than any of this, almost ridiculous in its simplicity. It is: call time. The fundraising consultant uses public records about campaign contributions to pull the names of people who have donated to similar candidates (or, even better, if you have a good relationship with a similar candidate, you can get explicit permission to use their donor list). Phoning people straight off public records is illegal, but if you find their phone number some other way, it’s OK to research their donation history. The fund-raising consultant looks up their phone number in the phonebook or on Google, along with any other basic info they can find or glean about the donor, and prints it out on a piece of paper (a call sheet). This is called “prospecting.”
A stack of such sheets is always kept in a binder and whenever a candidate gets a free moment, they are dragged to a closet with a phone and forced to do their call time. This is the real substance of the fund-raising consultant’s job: forcing the candidate to do the most humiliating and degrading and torturous work of the campaign—to become a telemarketer.
/> The closet is typically kept far away from campaign headquarters and contains nothing besides the binder and the phone (step one: no distractions). Then the fund-raising consultant uses every psychological tactic in the book to sit there and force the candidate to make calls. And, eventually, they do—with all the results you’d expect. (“How did you get this number?” people demand. “I don’t know,” the candidate lies, “my fund-raising consultant gave it to me.”)
But because these are people with a history of donating, the calls are not always so overtly angry. Sometimes you reach practiced professionals who know just the questions to ask to determine if you’re the kind of candidate they’ll support. They’ll angrily quiz you on their pet issue, sound out your support for business in general, or even begin deal making right over the phone.
I’ve heard of countless candidates who abhor call time. Running for office seems like a glamorous and important profession; it’s difficult to lower yourself from that image to the reality that it consists mostly of begging strangers for money. Like children who hate doing homework, candidates devise all sorts of excuses and devices to avoid having to do it. But it is inescapable, and never-ending—even winning is no escape.
After the consultants comes the world of political staffers. These are the people you actually employ, as opposed to the consultants you just rent. Being a political staffer is a dark life. You spend one year out of every two working 90-hour weeks, rarely leaving the office, sleeping only with the other staffers who work on the campaign, and giving up any semblance of an outside life. Then you typically spend the other year unemployed.
Occasionally your candidate succeeds and you manage to get a job in their administration, but this is rare and limited. Most campaign skills don’t translate well into office and even when they do, an office employs far fewer people than the campaign. So you try to pick up work on “issue campaigns,” organizations that try to use campaign tactics to get particular legislation passed. But without a strong model and a clear deadline, issue campaigns are a pretty demoralizing experience on their own.
So the job attracts a specific kind of person, an odd type of person, who would give up a steady job to throw themselves into semiannual fits of obsession over a random person. (Every two years the staffer claims, “This guy is really special, he’s the real deal,” without even a hint of self-consciousness.)
But the campaign staffers play little causal role in the campaigns, other than through the difficulty of hiring reliable people for such an unreliable job. So let us dwell no longer on their plight.
We have our team and their squabbling semblance of a strategy. But what does the actual campaign look like?
First, obviously, you raise money. A healthy war chest scares off potential opponents, gets you taken seriously by the press (before any actual votes are cast or polls are taken, the press usually judges candidates by their success in the “money primary”), and supplies you with the resources to run the rest of your campaign. This means the ability to raise money is crucial in its own right—before even a single dollar gets spent, the race is biased toward those who can fund-raise.
Then there is the “inside game” of collecting endorsements from local public officials, unions, interest groups, and the rest. These endorsements sometimes come with practical benefits—like lists of donors and volunteers that can be tapped—but mostly, like early money, they help give the candidate the air of viability to the press and public.
“Viability” is especially crucial in a primary because in a race without the usual two-party labels, so much of voting is based on the bandwagon effect: people like to pick a winner. Furthermore, in a simple plurality system like America’s, voting for someone other than the top two candidates is essentially throwing your vote away. So it’s crucial to make sure everyone considers you one of the top two.
Once you’ve met the threshold to be viable, you need to actually start moving people toward your side: persuasion. In a typical campaign, there is a long-running argument held in the media—through debates, dueling quotes to the press, competing public events, and so on—which is then underscored through things like TV ads (in big campaigns) and mailers. The highly informed voters watch this argument like a prize fight, but they usually come into it rooting for a particular side and not waiting to be won over. The rest of the public glimpses it only through a dark glass. Key phrases and arguments glance their consciousness, perhaps enough to sway them one way or another, but rarely with any degree of significant thought.
As with the rest of campaign strategy, it’s very unclear what affect any of this has. A massive experiment by Gerber et al. (2007) tested the effect of television and radio ad purchases during Rick Perry’s (R-TX) 2006 campaign for governor. It was a four-way race pitting then Lieutenant Governor Perry against a Democratic House member, humorist Kinky Friedman, and the Republican comptroller, who was running as an independent. By randomizing when ads were purchased in different areas, and measuring their effects with follow-up polls, the study concluded that the television ads boosted Perry’s vote share by about five points, irrespective of whether his opponents ran ads. However, the effects were short-lived, fading after a couple weeks.
All other measured effects have been on this scale. But such results are very hard to interpret because they come in the context of an already hard-fought race. We know the Republican is going to get a large percentage of the vote, and the Democrat will as well. But if Kinky Friedman bought a passel of television ads, would it boost his name recognition scores from zero percent to five percent (because TV ads are worth five points)? From zero percent to 25% (because they’d indicate he was a serious contender)? From zero percent to zero percent (because nobody will ever take him seriously)? It’s hard enough to measure the effect of a simple action on the wider social world. It’s practically impossible in a hard-fought zero-sum game like a political race.
But while time taken out to record ads is crucial, most of the candidate’s days are spent on what’s now called “earned media” (the old term, “free media,” was deemed misleading because it required too much work to really be “free”). These are the endless series of bogus campaign “events” held in the hopes of persuading some reporter to write about it or, even better, some television station to cover it. The candidate goes to a steel mill and shakes hands, the candidate declares his opponent a tax cheat, the candidate attends a debate. Ideally, it gets another mention for the candidate in the papers and maybe a quick chance to include a sound bite or two.
Finally, there is the issue of turnout. Campaign progress is always measured in percentages, as if there is a static population of voters and the goal is simply to win more and more of them over to your side. But in most elections, most eligible voters don’t vote—and this is especially true in the crucial primaries.
Most voters are pretty firmly tied to a party identity—political scientists have found that even folks who identify as “independent” voters usually vote straight party tickets. (For example, Tea Party activists may consider themselves independent of the Republican Party, but they’re never going to vote for the Democrat.) As a result, there’s very little room for persuasion in a general election (though that hasn’t ever stopped a major campaign from trying) and turnout becomes the crucial factor.
Despite that, turnout is treated with far less importance than persuasion in the average campaign. Part of this seems to be because of a widespread misconception among politicos about how persuadable people are. After all, the entire campaign seems to be a debate between competing ideologies—it’d be hard to understand why you’d go through all that effort if nobody ever changed their mind. Part of it seems to be a result of always seeing things in percentages. Part of it may be the result of the Median Voter Theorem.
The Median Voter Theorem, a key result in rational choice politics, says that candidates both move to adopt the policy views of the median voter so as to get the maximum number of votes. If both candidates are righ
t in the middle, with one just slightly to the left and one just slightly to the right, then they pick up the most votes—any move to an extreme would cede moderate voters to the remaining centrist.
Even a cursory look at any recent campaign will make clear the Median Voter Theorem doesn’t hold in real life. It’s hard to even think of a federal election where the candidates were barely distinguishable. This could be because of persistent “irrationality” on the basis of voters and candidates, but it could also be because of turnout effects. If moving to the center causes people at the extremes not to show up, then it’s not as costless as the MVT would suggest.
But there’s also a rational reason for not focusing on turnout: convincing people to turn out is hard. Under some analyses turning out an average voter is actually more expensive than persuading one, when it’d have to be half the price to be cost-effective. That’s because turning out a new voter increases your lead by just one vote, while persuading an existing voter increases it by two (one new vote for you and one vote less for your opponent). It takes a lot to get disaffected voters to the polls. By contrast, people who really like voting tend to do it every time no matter what’s going on.
Those tend to be the people who vote in primaries, when turnout is especially low. As a result, even though there’s more room to increase your vote through turnout, turnout is even less of a factor. Even if the entire primary electorate considers themselves to be hard-left Democrats, you can fight a vicious campaign about who’s the real hard-left Democrat in the race and who’s the corporate shill. Because turnout is so low these battles are usually fought among very high-information voters, who follow the twists and turns of a complicated campaign.
All the tactics of persuasion have been tested on turnout as well—and much more rigorously, since public voting records allow you to costlessly measure their effect. (There’s no need to poll the potential voters; you just look up their voting history.) Tactics like knocking on voters’ doors have been found to be surprisingly effective, and calling people up and even mailing them letters can be cost-effective under the right circumstances.* Measuring these effects is a burgeoning field of study, especially since IRBs will let you work to turn voters out but not to persuade them who to vote for.