Opinion

Want to Fix Social Security? Use the Right Wrench

Robert J. Shiller — June 2013

In a New York Times op-ed, Professor Robert Shiller writes that President Obama’s proposal to change how inflation is measured in Social Security benefit calculations “…solves the wrong problem, and, in doing so, undermines the integrity of the Social Security system.” One alternative, suggests Shiller, is to link retirees’ benefits to GDP per capita, in current dollars, which would align the interests of the retired with those of society as a whole.

Originally published in the New York Times on June 9, 2013.

Social Security is expected to run out of reserves by 2033—a mere 20 years from now. With the public apparently opposed both to tax increases and to benefit cuts, the main politically feasible way to avoid such a fate seems to involve some monkeying with obscure aspects of the definition of benefits.

Congress seems to want a ruse to disguise a cut in benefits as something else—like the discovery of a technical error that, once corrected, would let the government write smaller checks without taking the blame for cutting benefits.

In a spirit of compromise, President Obama has proposed changing how inflation is measured in benefit calculations. In his 2014 budget proposal released in April, he proposed that retirees' Social Security benefits be indexed to something called the Chained Consumer Price Index for All Urban Consumers or C-CPI-U, rather than the current benchmark, the Consumer Price Index for Urban Wage Earners and Clerical Workers, or CPI-W.

This seems to correct a real technical error. Economists have argued that the current index overstates the actual inflation rate, and that a switch to the C-CPI-U would make the inflation indexing more accurate, seemingly justifying the resulting gradual reduction in benefits. (And it would be a reduction, to the point that someone retiring today would be receiving about 5 percent less in 20 years.)

But here is the rub: the proposal solves the wrong problem and, in doing so, undermines the integrity of the Social Security system.

The purpose of Social Security is to help families. It reinforces the intergenerational sharing that families already—though imperfectly—provide. It helps retirees by stabilizing their income, and it helps their grown children, who are relieved of any excessive burden of supporting them. This purpose strongly suggests that the Social Security benefits should be indexed to some measure of the available, aggregate economic pie. That means a formula that looks completely different from the ones being discussed today.

Clearly, something needs to be done: if nothing changes, and the trust fund runs out in 2033, the system would be able to pay only about 75 percent of promised benefits.

The issues are complex, as economic theorists like Henning Bohn at the University of California, Santa Barbara, have shown. But now that an index change is on the table, we should take this opportunity to get it right.

One alternative that we should consider is a different kind of index switch, linking retirees' benefits to gross domestic product per capita, in current dollars. This measure responds to inflation just as the C-CPI-U does, but, in contrast, it also responds to changes in the nation's resources, as measured by real G.D.P. There could also be corrections for other factors, like the dependency ratio, which compares the number of "dependents" (retired people and children) to the number of working adults.

Such an index switch, however, has received hardly any public discussion, and that fact alone will make some people think that something is wrong with it. But political talk is limited; it is too focused on the identified problem of somehow fixing the projected Social Security insolvency. While this new idea won't solve insolvency—we may need to raise Social Security contributions to do that - it would support the principle that one generation shouldn't be more burdened than another.

Before 1972, Social Security wasn't indexed at all, and there was little public talk about tying it to inflation despite periodic, inflation-induced disasters for retirees. Congress simply made occasional, ad hoc, upward adjustments in the benefit formula. Proponents of these adjustments often justified the increases as offsetting the rising cost of living. But at the same time, they also described retirees as needy, suffering hardships and deserving dignity. While there were no explicit references to G.D.P., it's plausible that their sense of "needy" was influenced by comparison with the rising American standard of living for working adults. As a matter of fact, total Social Security benefits more than kept up with the rapid G.D.P. growth in the couple of decades before they were indexed to inflation.

Now that inflation indexing has been in place, some people see the formula as always providing the scientifically "right" amount of benefits. They don't think much about other factors that might enter into the determination of benefits. And, in fact, since 1982, Social Security benefits as a fraction of G.D.P. have generally been falling, at least until the latest recession.

The fact is that per capita G.D.P., in current dollars, has grown 1.2 percent faster a year, on average, than the CPI-W in the last 20 years, despite the Great Recession. If we indexed Social Security benefits to G.D.P. per capita, and not to the C-CPI-U, people who retire today and live 20 additional years would get a third more in real, inflation-corrected benefits—provided, of course, that the next 20 years are like the last.

If the economy grew unusually rapidly, however, they might get 50 or 60 percent more, in real terms. But even that wouldn't break the budget of the Social Security system, because contributions would be higher, too, in response to the greater economic growth. And if the economy stagnated in a deep depression in the next 20 years, retirees might see no growth at all, or conceivably even a decline, in real terms. That's O.K., too, because everyone else would be suffering as well. We have other programs, like Medicare, that deal with any extreme hardships that some older Americans with special conditions might encounter.

The point of G.D.P. indexing is to align the interests of the retired with society as a whole. Older Americans should share both the windfalls and the losses with other generations—with working adults, and with children. It shouldn't be otherwise. If the system's rules force us to maintain the real benefits of retirees at all costs, we may find ourselves, in difficult times, taking excessively from our children via cuts in education, or from our working adults. Why should retired people feel no effect at all from the recession while younger people are left suffering? They should be sharing the hardships, if we have familial feelings for one another.

The issue is especially salient today because the demographics have shifted: the aging of the baby boomers will create many retirees relative to adults who are still working. This huge shift could not have been foreseen by Social Security's designers.

Achieving the right benefit formula while fixing Social Security's solvency problem might require increasing the contribution rate—now in the form of 6.2 percent taxes on employees and employers. But so be it. We need to say what is right, keeping our eyes on the integrity of Social Security, which is crucial to our identity as a civil society.