Rockeymore: “Don't save Social Security by raising retirement age”

Maya Rockeymore, president and CEO of Global Policy Solutions and an adjunct professor at American University, writes in the Houston Chronicle that an increasing longevity gap between rich and poor means we shouldn't try to fix Social Security's funding problems by raising the retirement age. The normal retirement age is currently 66, and rising under current law to 67 by the early 2020s, and many analysts argue it should rise further as life spans continue to increase. Rockeymore writes:

Some policy experts argue that increasing the retirement age is a sure-fire way to extend the solvency of Social Security. But lawmakers overlook how this option would unfairly disadvantage people with shorter life expectancies. Declining life expectancies for vulnerable populations such as low-income and lesser educated individuals means that increasing the retirement age would have the unfair effect of using their lifetime payroll tax contributions to subsidize the retirements of people who live longer. Often, these people are wealthier, better educated and white.

Rockeymore is correct that, while there has always been a longevity gap between rich and poor, this difference appears to be widening. By itself, this would make Social Security less progressive, since high earners would tend to collect benefits longer than low earners. The CBO put out a nice paper on this subject back in April.

However, it's not clear that increasing the normal retirement age would exacerbate this effect. Remember that an increase in the normal retirement age is nothing other than a benefit cut. If the retirement age rises by a year, that means that everyone will receive around 7 percent lower benefits than they otherwise would have. Raising the retirement age is no different than an across the board 7 percent benefit cut. An across the board benefit cut wouldn't disproportionately hurt low earners or those with short life spans; everyone – rich and poor, short or long-lived – will receive 7 percent less than they otherwise would have.

On the other hand, raising the early retirement age of 62 would disproportionately hurt those with short life spans, since more of them would die before being able to claim any benefits. This simple spreadsheet shows that that raising the normal retirement age would reduce everyone's benefits equally, even if the rich live longer than the poor, while raising the early retirement age would tend to hit the poor/short-lived more than others. I did some more detailed modeling here.

So I think we've established that Rockeymore's basic underlying premise isn't really correct. Raising the retirement age doesn't have a hugely disproportionate effect on low earners or short-lived individuals. However, she goes on to make a policy recommendation that I think is also working looking at:

There are fairer policy options that would achieve the solvency goal. For example, Sen. Barack Obama proposes to increase Social Security's cap on taxable earnings. True, this taxes higher earners without giving them more monetary benefits. But that is neither a policy reversal nor a detour from the shared national values that make Social Security the most popular law in history. Rather, it returns Social Security to its intended progressive course, correcting the unfair and unintended regressive impact that widening life expectancies are creating for women in some areas of the country and for low-income, less-educated and minority workers.

First, it is true that a widening longevity gap would make Social Security less progressive and that applying a new payroll tax on people earning over $250,000 would make the system more progressive. I have no idea (and I suspect neither does Rockeymore) whether the increased progressivity from the new tax would be the same size as the reduced progressivity from the widening longevity gap. My guess is that the tax would be significantly more progressive than the longevity gap is regressive, but we'd have to run some numbers on that. The point here is that you want things to be at least roughly proportional.

Second, Rockeymore says that applying a new tax on workers earning $250,000 without paying them any extra benefits "is neither a policy reversal nor a detour from the shared national values that make Social Security the most popular law in history." Huh? Let's see: Social Security has never levied taxes above the payroll tax ceiling (currently $102,000), and it has never levied taxes without paying additional benefits in return. So to me it seems to be exactly a policy reversal from how the program was founded and how it has continued to run since the 1930s.

If we wish to control for the effects of a widening longevity gap there are some simple ways to do it. For instance, Social Security current replaces a progressive proportion of workers' pre-retirement earnings. In 2008, the benefit at the full retirement age equals:

(a) 90 percent of the first $744 of his/her average indexed monthly earnings, plus

(b) 32 percent of his/her average indexed monthly earnings over $744 and through $4,483, plus

(c) 15 percent of his/her average indexed monthly earnings over $4,483.

If the richer are living increasingly long lives relative to the poor, all we need to do is tweak these parameters to restore whatever level of progressivity we desire. For instance, we might increase the 90 percent replacement factor to slightly raise benefits for low earners, while slightly reducing the 15 percent replacement factor that applies mostly to high earners' benefits. This doesn't demand anything radical.