In his (most recent) piece on class warfare, Paul Krugman writes:
Consider, as a prime example, the push to raise the retirement age, the age of eligibility for Medicare, or both. This is only reasonable, we’re told — after all, life expectancy has risen, so shouldn’t we all retire later? In reality, however, it would be a hugely regressive policy change, imposing severe burdens on lower- and middle-income Americans while barely affecting the wealthy. Why? First of all, the increase in life expectancy is concentrated among the affluent; why should janitors have to retire later because lawyers are living longer? Second, both Social Security and Medicare are much more important, relative to income, to less-affluent Americans, so delaying their availability would be a far more severe hit to ordinary families than to the top 1 percent.
But is that true? It’s first important to acknowledge that Social Security reforms don’t occur in a vacuum: Plans that increase the normal retirement age (NRA) almost always make other changes, such as progressive benefit cuts, an increase in the taxable maximum wage, the introduction of minimum benefits for low earners, etc. – that would offset any regressive effects of raising the retirement age. It’s the package that matters and Krugman should look at the package.
Second, you need to consider interactions with disability benefits. Truly low-income individuals are disproportionately on disability, which wouldn’t be cut by an increase in the retirement age. Moreover, the share of individuals’ claim disability would likely rise, since there would be more years prior to the NRA for a disabling event to occur. For the program as a whole, raising the retirement age might make it more progressive.
To check, I simulated an increase in the Normal Retirement Age for members of the 1990 birth cohort using the Policy Simulation Group’s models. I assumed an increase in the NRA to 70 by the year 2075, so for the 1990 cohort the NRA would be around 69. Since taxes wouldn’t change, I focused on benefits: For each quintile of the lifetime earnings distribution, I calculated the percent of total lifetime benefits received. For instance, under current law the top 20% of earners receive around 37% of total lifetime benefits. That’s both because they earn more and get higher monthly benefits and because high earners tend to live longer.
The results are a little hard to interpret. The share of benefits for the bottom earnings quintile drops by around 0.6 percentage points. This group has very, very low average benefits since it generally captures the people who fail to survive to retirement age. The benefit share for the second quintile rises, as it does for the third (middle) quintile and the fourth. Total benefits decline by around 3.3 percentage points for the top earnings quintile. These folks rarely claim disability and so would be hit hardest by the higher retirement age.
As with many distributional questions, it’s a little tough to say whether raising the retirement age makes Social Security more or less progressive. But it’s pretty clearly wrong to say, as Krugman does, that a higher retirement age is “hugely regressive.”