Saturday, June 21, 2008

More means-testing from Mickey (Kaus)

Mickey Kaus returns to the idea of means-testing, arguing that only a means-test can save enough money to both balance Social Security and help fund universal health care.

But any amount of money you want to save through a means-test, I can match through a change in the traditional benefit formula. (Really. Name an amount and I'll put it together.) Let's say you have a means-test that cuts off the top 25% of people from benefits altogether. I can similarly construct a traditional benefit cut that does the same thing. There's no magic here.

The difference is that if you tell higher income folks ahead of time that they won't get any benefits, they'll save more on their own in order to fund their retirements. But if you implement a means-test on total retirement income, this means that the more they save the more their benefits get cut -- and so it becomes a tax on saving in a way that a traditional benefit cut isn't.

I don't have any particular beef with Mickey's idea to reduce benefits for high earners, but I'm puzzled with the fixation on a means-test.


Jim Glass said...

"... but I'm puzzled with the fixation on a means-test. "

I'd think political forces will drive SS towards means testing rather than reducing the benefit formula for high-earners because:

1) The "social insurance" rationale of SS is better served with means testing.

I.e., someone with a long record of high earnings can wind up poor and needy due to little fault of one's own (bad financial luck, some unhappy reports from the doctor, etc.) while someone else with a low earnings record winds up rich.

So a high earner who's wealth has been wiped out by family medical bills has benefits reduced due to his high earnings record, while Paris Hilton with her low earnings record keeps her benefits in full. That seems contrary to the "social insurance" ideal.

2) We can expect Congress to again put off enacting benefit or formula cuts of any kind until the very last moment, when the cash flow crunch is really bearing down.

At that point, changing the benefit formula to reduce benefits 20 years in the future for those with high earnings records will produce cash flow savings 20 years too late. (I don't see how they could change the benefit formula for the past 20 years retroactively.) Means testing will be needed to save cash right away.

Of course, the two options are not mutually exclusive, both could be enacted, to save costs right away and even more in the future.

As to the idea that SS should be designed to best increase savings and investment, that's laudable enough -- but my cynical self says that's never been a concern during the enactment of the many changes made to SS in the past, so why would it become a consideration in the future? What incentives will change for the polticians?

I'd wager the politics of SS will remain all social insurance arguments versus the cost of coming tax hikes.

That said, Edmund Phelps had a piece in the WSJ a while back in which he argued that if people had actually saved for their coming retirement-years living costs and medical expenses -- instead of being told the government would pay for them via SS and Medicare -- the national savings rate would have been much higher for a generation, so the trade account would be better, the dollar stronger (oil might still be around $80), the future economic growth picture much brighter, etc.

But Medicare I'd think would have a far bigger effect on that than any plausible change in SS from here on. So if this is the concern maybe we should be arguing to get the rich out of the future of Medicare one way or another, and to let them know it now.

Anonymous said...

Scrapping the 25% of benefits that go to the highest career earners would mean no benefits to individuals making today's equivalent of about $60,000.

Now I realize Mickey's means-testing would be based on income in retirement, not career earnings, but this figure gives an approximate sense of who might be hit with a 100% benefit cut.

Consider that $60,000 career-earners get roughly $20,000 a year from social security -- if they retire at the Normal Retirement Age. Therefore, a means test would impose a $20,000 annual cost on saving for such individuals.

To overcome this 100% benefit cut, a $60,000-earner would have to save about 7.5% of their income each year.

So under Mickey's plan, such an individual would have to save 20% of their income (including the 12.4% SS tax) in order to receive retirement benefits equal to about 33% of their annual income -- or less than half of what financial planners suggest people should be saving.

This is far too severe to be realistic.

Andrew G. Biggs said...

Good points by both. To Jim, I'd simply raise the example of Medicaid for long-term care: in theory it's well-targeted toward the poor, but we know there have been problems over time of older folks spending down their assets to acquire eligibility. That's not good policy (which is one reason they've had to crack down on it).

To anon, I'd basically just agree. But my argument against Mickey is that a lot of people won't do that extra saving, precisely because it would result in a direct reduction in their Social Security benefits. So both personal saving and improvements to system financing would be smaller than a static analysis would imply.

Marc said...


I'll preface that I am not a fan of means-testing. However, I take issue with your claim that anything done through means-testing could be done through traditional benefit cuts.

Of course, from a macro perspective, such cuts could have the same fiscal result -- but the burden will not be born by the same people. With traditional benefit cuts, for example, there is no way to differential between someone who has made around 100,000 (wage-adjusted) dollars a year and someone who has been making $1 million a year.

Means-testing has the advantage of being able to reduce benefits more for the latter person than the former.

To mitigate savings disincentives, means-testing could be based (or based in part) on lifetime earnings, rather the income or assets at the time of retirement.

Let me know your thoughts,

Andrew G. Biggs said...


You're right, although that could be fixed by eliminating the wage cap ($102,000) for the purposes of calculating benefits. That could allow benefits to be paid to people making $100k or so, but not to people making $1 million, etc. Although Mickey is looking for such big savings that I'm not sure it would make much difference in practice -- they'd all have to be cut off.

One issue I'll try to touch on a future post is the UK's experience with means-testing through the Pension Credit. This created a situation where most lower/middle class workers had very poor incentives to save, such that a financial adviser would almost be negligent in telling them to do so. Obviously, this worked at a different end of the income spectrum, but the problems are roughly the same.