In his 1998 State of the Union Address, facing the prospect of budget surpluses for the first time in decades, President Bill Clinton had a simple plan: "Save Social Security first." He wanted to dedicate future budget surpluses to the Social Security trust funds, pushing back for decades the day when it would run out of money. In 2000, presidential nominee Al Gore took up the refrain with a memorable, and occasionally mocked metaphor: the "lockbox."
Well, that didn't happen. The latest projection by the Social Security Board of Trustees is that the trust funds will be exhausted in 2033. At that point, payouts will be limited to the amount brought in by payroll taxes--which means that everyone's benefits will fall by about 25 percent.
This looming threat is often cited as a reason why we need to "reform" Social Security, which is code for cutting benefits. Most recently, President Barack Obama proposed changing the calculation of cost-of-living adjustments, which amounts to a progressively larger benefit cut as you get older. The problem, however, is that Social Security is the only significant source of income for many older Americans, and its net benefits are falling anyway due to increasing taxes and higher Medicare premiums (which are deducted from Social Security).
Against this backdrop, it's a mistake to look at Social Security in isolation. We have to look at our overall retirement security "policy." And when we do that, it's obvious that we're throwing billions of dollars away.
That money is being wasted on tax subsidies for "private" retirement accounts such as 401(k)s and IRAs. In 2012, these subsidies added up to $199 billion, or about 1.3 percent of GDP. (See Toder, Harris, and Lim, Table 5.) The motivation to protect these accounts isn't totally crazy. Since we want people to save for retirement, we offer tax breaks for doing so.
More Tax Breaks, But Not More Savings
There are two major problems with these subsidies.
The first is that they go overwhelmingly to people who don't need them -- like my wife and me. As two university professors living in Western Massachusetts, where the cost of living is low, we make more than we need to support our lifestyle. We max out our defined contribution plans every year, and because we're in a relatively high tax bracket (28%, I think), we save thousands of dollars a year on our taxes. This is the problem with most subsidies that are delivered as tax deductions. Their cash value depends on the amount you can deduct and on your marginal tax rate. In this case, fully 80 percent of retirement savings tax subsidies goes to households in the top income quintile. (See Toder, Harris, and Lim, Table 5.)
The second problem is that these tax incentives don't work. They don't cause people to save more. In my case, the amount we save is just our income minus our consumption, and our consumption isn't affected by the tax code. If there were no tax subsidy, we would save the same amount and just pay more in taxes. And it's not just us. A recent and widely discussed paper by Raj Chetty, John Friedman, Soren Leth-Petersen, Torben Heien Nielsen, and Tore Olsen looked at what happened when the Danish government reduced tax subsidies for retirement savings by rich people. The short answer is that decreases in retirement savings were almost perfectly matched by increases in non-retirement savings. The overall effect, they estimate, is that for every dollar in tax subsidies, total savings go up by one cent. The other ninety-nine cents is just a handout to people who would have saved anyway.
In White House Burning, we did not propose eliminating these tax subsidies because we did not want to remove incentives for people to save more. Since it turns out that those incentives don't work, I now say, get rid of them.
The Right Retirement Policy
The single most important step we can take toward improving retirement security for all Americans is eliminating tax preferences for private retirement accounts and earmarking the resulting tax revenue for the Social Security trust funds. There would have to be some transition period for existing defined benefit pension plans, but once phased in, this would increase tax revenues by about 1.3 percent of GDP.
That is almost double the 75-year gap in Social Security's finances, which the CBO estimates at 0.7 percent of GDP.* This means that we could ensure the program's solvency for generations to come and still have money left over. Those additional revenues could be used to expand Social Security benefits through a guaranteed, fixed pension, as recommended by Michael Lind, Joshua Freeman, and Steven Hill of the New America Foundation. Or we could grandfather in employer-sponsored defined benefit plans, if we think those are preferable to defined-contribution plans. Or we could maintain savings incentives for lower- and middle-income workers, for example by providing a 15 percent refundable tax credit for contributions to retirement accounts up to $5,000 per year and phasing it out around the 60th income percentile.** Or we could pay down the national debt.
All of these options are preferable to what we're doing today: handing out free money to people who would save plenty for retirement anyway. Furthermore, bolstering the Social Security trust funds would eliminate--at least for most of our lifetimes--the argument that we have to cut benefits because we can't afford them.
We do have the money. It's just going to the wrong people.
* Last year, the CBO estimated the 75-year gap in Social Security's finances at 0.7 percent of GDP in its extended baseline scenario, which conforms to current tax law. (The alternative scenario adds the arbitrary assumption that payroll taxes will be capped as a percentage of GDP, resulting in a 75-year gap of 0.9 percent of GDP.)
** Currently, the three bottom income quintiles only enjoy about 7 percent of the retirement savings tax subsidies, or about 0.1 percent of GDP. So if we want to preserve 401(k) plans for people who really need them, we can certainly afford it.
We want to hear what you think about this article. Submit a letter to the editor or write to firstname.lastname@example.org.