Behavioral Economics Foils an Obama Tax Cut?

New research finds that a trendy economic theory backfired on the Obama Administration. Or did it?

The Making Work Pay tax credit, part of President Barack Obama’s 2009 stimulus bill, was one of the least noticed tax cuts of all time. Rather than coming as a check from the government with the sum written on the dotted line, Making Work Pay—a credit of up to $400, or $800 for couples filing jointly—was disbursed in a steady dribble. The middle-class workers who were its target had their paychecks grow slightly for part of 2009 and 2010, as the government withheld less. When asked, many didn’t realize they’d gotten anything.

Why do it this way? It’s rarely good politics to spend more than $100 billion in public money on something few people notice. But policymakers believed that parceling out the money piecemeal, rather than sending it to taxpayers in a lump sum, had two advantages: It could start sooner, since people didn’t have to wait for the check; and it was more likely to work. By giving people the sense that their incomes had grown, doling out the money paycheck by paycheck was supposed to make recipients more likely to spend it, thereby lifting the economy. That was the theory, anyway, according to a school of thought favored in the Obama White House: behavioral economics.

In the past decade, this new set of ideas about economic behavior has gone from the margins of academia to the intellectual mainstream. In 2002 one of its godfathers, the psychologist Daniel Kahneman, won the Nobel Prize in Economics, and the years since have seen a growing list of best-sellers (Malcolm Gladwell’s Blink, Richard Thaler and Cass Sunstein’s Nudge, Dan Ariely’s Predictably Irrational) describing and drawing on its findings. Unlike neoclassical economists, behavioral economists don’t see people as rational actors coolly weighing costs and benefits. Behaviorists argue instead that people rely on a set of instincts, biases, and cognitive shortcuts to make decisions, which often lead them to choices they come to regret. We save too little and spend too much, we stick with the status quo even when it costs us money, we avoid smart risks and take dumb ones.

In 2009 this theory held obvious appeal to the incoming Administration. If the country’s ills were in part the result of poor financial decisions people made unconsciously, perhaps those problems could be fixed through behaviorally informed public policy. The Administration’s first big legislative push, its stimulus bill, presented an opportunity to test some exciting new ideas on a national scale.

It didn’t work. That, at least, is the finding of the first study to look specifically at the behavioral economics element of the stimulus. In a forthcoming paper, three economists—Claudia Sahm of the Federal Reserve, and Joel Slemrod and Matthew Shapiro of the University of Michigan—found that Making Work Pay didn’t get people to spend more money. In fact, it got them to spend less. The study is a prism through which to view both the efficacy of Obama’s stimulus and whether a set of discoveries about the foibles of human decision-making can be translated into effective government policies.

The foundations for behavioral economic theory were laid by two Israeli psychologists: Kahneman and the late Amos Tversky. One of their best-known studies was a simple survey. They presented people with a scenario: As you enter a theater, you realize you have lost the ticket you bought. Would you pay the $10 for a new one? Fewer than half said they would. Kahneman and Tversky asked another set of people the same question, but with one tweak: Rather than lose the ticket, they lost a $10 bill on the way to the theater. Would they go ahead and buy a ticket? Nearly 90 percent said they would.

Financially the two situations are identical, so what explains the dramatic difference in responses? Richard Thaler, a University of Chicago economist and one of the founders of behavioral economics, argues that people keep “mental accounts” in their heads for different sorts of expenses. Certain sums are earmarked, even if unconsciously, for entertainment, rent, or groceries. Buying a second ticket to replace the lost one felt like spending $20 to go the play, and for most people that was too big a piece of their entertainment account. (The study was carried out three decades ago.) A $10 bill that falls out of one’s pocket, however, hasn’t been assigned to a particular account, so its loss isn’t felt as much.

The design of Making Work Pay plays off of mental accounting. One of Thaler’s findings is that people are more likely to spend money that they have filed in their “current income” mental account rather than their “assets” mental account—in other words, they measure their spending against the size of their paycheck instead of the size of their bank account. A lump-sum tax rebate feels like an increase in wealth and is more likely to be saved. A series of slightly bigger paychecks feels like an increase in income and is more likely to be spent.

That’s not what happened in practice, according to Sahm, Slemrod, and Shapiro. In a study of the 2009 stimulus, based on 500 telephone interviews, the authors found that only 13 percent of Making Work Pay recipients reported that the tax credit would lead them to increase spending. This was just half of the 25 percent spend rate the researchers found for the traditional lump-sum tax rebate in President Bush’s 2008 stimulus. Of course, 2009 was a worse economic climate than 2008, and that might have played a role in the change. To control for this, the researchers looked at one-time stimulus payments that went to retirees at the same time that Making Work Pay was going to working households. The retirees, too, reported much higher spending rates than the Making Work Pay households, who got their money in a steady drip.

The authors can only guess at what’s behind their results. Perhaps, they suggest, the relative invisibility of Making Work Pay worked against it. Recipients, not noticing that they were getting a tax rebate, didn’t spend it. A lump sum, on the other hand, might have been saved, but it might also have been spent by more households on a big-ticket item that they otherwise wouldn’t have bought.

Does that make behavioral economics bunk? On its own, no. Sahm, Slemrod, and Shapiro relied on survey data: People were asked in 2009 what effect the previous year’s lump-sum tax rebate had had on their spending and saving, and what effect Making Work Pay would have. To a behavioral economist there are three big problems with this approach: People are not very reliable sources of information about their own past spending; they’re even less reliable about future spending; and they’re often blind to the things that actually shape their financial decisions. “The work is certainly worth doing,” Thaler says, “but you can’t do it by asking people to remember what they did with the money.” Slemrod concedes that surveys have their limits, but he points to other research by economists Jonathan A. Parker and Nicholas S. Souleles that bolstered some of his paper’s findings without relying solely on self-reports.

The debate over Making Work Pay underscores a broader argument in Washington about whether Keynesian stimulus is effective in any form—Republicans have managed to turn “stimulus” into a dirty word. To its critics, the Administration’s interest in behavioral economics reflects a paternalistic, top-down approach to governance in which Washington elites stack the deck to achieve the outcomes they favor. Yet this wasn’t the first occasion that behavioral theory shaped a government program: In 2006, with bipartisan support, Congress passed the Pension Protection Act, which included a change to the rules around 401(k) plans. That change grew out of behavioral studies that showed people’s savings rates could be improved simply by flipping the default option so workers had to opt out of rather than into employer retirement plans. Research by the Employee Benefit Research Institute has found that the law worked, increasing both the number of people in the plans and the amount most people saved.

The biggest criticism of behavioral economics is that it remains little more than an interesting grab bag of exceptions to traditional economic theory—TED-conference fodder gleaned from lab studies that are pale approximations of real life. However, as both Thaler and Slemrod point out, there are important questions that neoclassical economics simply doesn’t deign to answer. Traditional economic theory predicts that the design of a tax credit like Making Work Pay should have no effect on its efficacy: A tax cut is a tax cut, whether it comes in a check, reduced paycheck withholdings, or, for that matter, a briefcase full of cash. How money is delivered should have no effect on whether people spend or save.

Yet if, as Sahm, Slemrod, and Shapiro argue, lump-sum payments are more effective than Making Work Pay, then behavioralists might have something to cheer about, too. That is, after all, exactly the type of effect that these economists would go looking for, even if it’s not the one they predicted. Should it hold up, the finding might best be seen as a refinement rather than a refutation. “I disagree with the idea that this disproves behavioral economics or contradicts it,” says Michael Barr, a University of Michigan law professor and former Assistant Treasury Secretary in the Obama Administration. “The basic insight is that how you give the money matters, and this supports that.”

So the behavioralists may be right after all: People don’t act the way economists predict they should. The trouble is that they don’t act the way policymakers want them to, either.

    Before it's here, it's on the Bloomberg Terminal.