The Dismal Science

Kill the 401(k)?

The accounts cost the government billions—and they might not be helping us save.

 A man deposits his tax return into a mailbox. Americans get tax breaks for saving for retirement, but do those breaks encourage us to save?

Photograph by Erik S. Lesser/Newsmakers.

As America hurtles toward the fiscal cliff, there’s an increasingly frantic search for ways to shore up the country’s deteriorating balance sheet. Republicans want to cut spending; Democrats would prefer to raise taxes on the wealthy. But a paper released today by Harvard and Danish researchers highlights just how much room there could be to generate more government revenue without sacrificing economic efficiency—in other words, the type of policies that both parties could conceivably learn to love. The study analyzes the responses of Danish taxpayers to savings incentives—much like those that exist for American 401(k) and IRA accounts—and also behavioral “nudges” that automatically deduct retirement savings from workers’ paychecks. It turns out that savings incentives had scarcely any impact on the rate at which Danes accumulated nest eggs, while the nudges were very effective in making people save. These findings suggest that 401(k) plans and their brethren—which cost the U.S. government as much as $100 billion a year in lost revenue—don’t do much to further their stated objective of boosting retirement savings. Even if $100 billion wouldn’t go all that far toward solving America’s debt problems, it suggests that smart approaches to eliminating or improving government programs could quickly add up to fiscal solvency—and might help the two sides find common ground.

The reason we have tax shelters like the 401(k) is to change the relative cost of spending money today versus saving for tomorrow. Exempting retirement investments from taxation increases the saver’s return on his investment, so a rational cost-benefit calculation should lead most people to put something away for the future. In theory, such tax shelters should go some way toward correcting Americans’ problem of undersaving.

But these policies assume that, despite being impulsive spenders, we will respond like textbook economic agents by saving more when tax exemptions make saving cheaper—an unlikely proposition. Even if savers do put more money in a 401(k) as a result of tax incentives, they might simply do so by reducing their investments in other non-exempt accounts, rather than saving more overall. Whether tax shelters actually encourage higher savings is a matter for the data to decide.

The lack of necessary data on Americans’ incomes and savings is what motivated Harvard economists Raj Chetty and John Friedman to partner with researchers in Copenhagen, who had access to the complete tax records of every Danish citizen over the past couple of decades. What’s more, some recent changes in the Danish tax system allowed them to examine how taxpayers respond to shifts in tax-based incentives to save.

The study focuses on a Danish tax reform in 1999 that, for taxpayers in the highest tax bracket, reduced the subsidy for pension contributions by 13 cents on the dollar (or more precisely, 13 ore on the Danish krone). As a result, saving got more expensive for earners above the top bracket cutoff of 268,000 DKr (about $41,000), while remaining unchanged for those earning under the threshold. To assess the impact of savings incentives on pension contributions, the researchers examined the change in savings for high earners who lost their subsidy, using those below the cutoff as a control group.

It turns out that a little less than 85 percent of Danes affected by the change did nothing to respond to the shift in savings incentives—they put money in their pensions at about the same rate in 1999 as they did the year before. Even those among the minority that did cut their pension contributions in response to the new rule didn’t save much less overall—they simply put more away in other investments. Based on these findings, the authors estimate that every dollar that the Danish government spends to encourage pension contributions generates only about a penny in extra savings.

The fact that most people are too distracted from their finances to respond to any shift in incentives opens the door for a different approach to promoting savings—by simply taking advantage of the incredible inertia that seems inherent to the human condition. By putting in place high default contribution rates to pension plans, we might get people to save more simply because they never get around to changing the amount of their monthly payments.

This approach has already proved effective in raising 401(k) contributions in this country, and it’s a hypothesis that the authors of the current study also apply to the Danish data. They benefit from access to the annual tax returns of every Danish citizen, which allows them to follow individuals when they move jobs, sometimes to companies that have higher default levels for pension contributions, sometimes to companies with lower ones. Most people—again, around 85 percent—never change their default option. If they move to a higher-default employer, they save more, and if they move to a lower-default one, their savings go down. Further, this group of inattentive savers doesn’t shift money in or out of other savings, so a $1 increase in the default contribution translates into an extra dollar put away for retirement. No government subsidy required. (The authors also study a government-mandated 1 percent contribution to retirement savings and similarly find that most people didn’t react by reducing other savings, even if they were already saving more than 1 percent of their income.) The remaining 15 percent did actively manage their contribution rates, but when you put the two groups together it still adds up to a $0.90 increase in savings from a $1 increase in the default amount. If the U.S. government’s objective is to increase savings rates, it seems like it could accomplish this by simply putting in place a high standard default level for pension contributions, an approach that has the benefit of not resulting in any lost tax revenues.

Harvard economists Chetty and Friedman embarked on this project primarily because they wanted to inform American rather than Danish tax policy, and they have gotten some pushback on whether it’s appropriate to apply their Danish findings to U.S. taxpayers. This skepticism strikes me as misguided. It’s surely the case that Danes and Americans have different preferences—Danes love herring, rye bread, and socialized medicine in a way that many Americans find puzzling. But it’s harder to make the case that the average American is more attentive to his pensions and taxes than the average Dane.

If the study’s findings do indeed have universal application, it suggests that the world is divided into two types of people—what the authors call “passive” and “active” savers. Most of us—about 85 percent—fall into the passive camp, where behavioral “nudges”’ make us save more without even thinking about it. By contrast, as a means of boosting savings, tax incentives seem twice damned—most of us don’t pay enough attention to respond to them, and those who do simply undo their effects by shifting funds in and out of other savings accounts; there’s not necessarily a net gain in savings.

It’s unlikely that Republicans would agree to a national “nudge” like a defaultpension contribution that could be changed only through deliberate action. It would strike libertarians as another case of heavy-handed government treading on individual liberty. And they may have a point—as Glenn Hubbard, the dean of Columbia Business School (where I teach) and economic adviser to the Romney campaign points out, who is to say what the “right” level of default savings should be? For example, it’s important to have savings on hand for a rainy day, and locking people into savings instruments that can’t be touched until retirement could have many negative consequences. This problem could be exacerbated by the fact that it’s the passive, unsophisticated savers who are most likely to blindly stick with the default—at least some of these unsophisticated savers are also likely to find themselves with pressing money problems.

But getting rid of the tax shelters that burn a $100 billion hole in the government balance sheet might be one reform both sides can get behind. In a recent conversation with me, Hubbard echoed the study’s sentiments that 401(k)s and IRAs are essentially subsidies to relatively wealthy households that are well-informed enough to take advantage of them. Or, in Hubbard’s words, these types of programs act as “a tax on the unaware.”

If we decide it is a national priority to promote savings, maybe there’s a middle ground that would at least partially satisfy both sides. For example, the government could provide tax incentives to companies—who presumably are sophisticated about their tax bills—to provide savings ”nudges” to their workers, perhaps choosing the default that best suits their employees. That way, at least the decision is made by a local manager rather than a Washington bureaucrat.

But the biggest takeaway from this study may be that we don’t need across-the-board cuts in deductions or exemptions—they’re not all created equally. Some—like savings tax shelters—starve the government of tax revenue without doing much to encourage retirement savings or furthering other policy objectives. Distinguishing which well-meaning programs accomplish their objectives and which ones misfire is a delicate task. And it’s where research like the Danish tax study comes in, to guide policymakers toward more efficient policies and a more effective government.