Scott Sumner  

The wrong way to fix Social Security

PRINT
The Libertarian Target... Did Edwin Cannan Think a 100 P...

Governor Chris Christie wants to reduce Social Security benefits for the rich:

As part of the plan, he'll propose phasing out Social Security payments for those making more than $80,000 in other income and eliminating them for those making $200,000 or more a year.
This is a good idea in general, but the specific proposal is awful. He should be trying to make the program more progressive, but ends up with a plan that is unfair to savers. Christie should be proposing higher payroll taxes on the rich, but is actually proposing higher income taxes on the rich.

Many people have trouble understanding this distinction, so consider the following thought experiment:

Imagine identical twins that each make $150,000/year from age 25 to 65, and then retire.

Can we agree that these two people are equally affluent? Now suppose one chooses to spend all that money when it's earned and relies on Social Security when he is old. The other invests half his income each year and becomes quite affluent by age 65. Neither brother is better off than the other; they both have identical lifetime consumption in present value terms. It's just that the thriftier brother chooses to defer his consumption to later years, and thus consumes more in nominal terms. The other brother could have done the same, but chose not to. Just as you can't say apples are better than oranges if their price is the same, you can't say four apples in 40 years is better than one apple today if their price is the same.

Economic theory suggests that both brothers should face the same tax rate. Their lifetime consumption should be reduced equally, relative to the no-tax case. A simple payroll tax would be neutral regarding the saving vs. consumption decision. So would a VAT. Indeed even a progressive payroll tax would be neutral. But an income tax would discriminate against the thriftier brother, imposing a higher tax rate on future consumption than current consumption. It would double tax money that is saved. That's perverse, especially given that saving provides the funds for investment, and hence future economic growth.

Now suppose there was a third brother, who made only $50,000/year. Is there an argument for taxing that person at a lower rate? Yes, you can easily make an argument for progressive taxation on utilitarian grounds. Not everyone will buy the argument, as it's hard to do interpersonal utility comparisons, but you can make a plausible case for progressive taxation.

But there is no plausible case to be made for Christie's proposal, which is to raise income tax rates on the affluent elderly. If he wants to fix Social Security in a way that makes the system more progressive, he should have proposed trimming retirement benefits for those whose lifetime wage income was above a certain threshold. Unlike the Christie plan, a phase out based on lifetime wage income would not distort the decision as to whether to consume now or in the future. It would not discourage saving and investment.

I usually visualize the Democrats as representing the "one marshmallow" Americans. Tax policy is one of the very few areas where the GOP appeals to voters favoring free markets and small government. If the GOP also begins to favor anti-saving tax policies then we are really in deep trouble. Let's hope Christie is an outlier, one of those RINOS who will soon switch to the Democratic Party.


Comments and Sharing


CATEGORIES: Taxation




COMMENTS (19 to date)
John Thacker writes:

Yes, I have stressed this before too. When we have access to lifetime earnings, it seems stupid to raise the taxes of poor and middle class savers rather than richer people who spent everything instead of saving.

Brad writes:

But the problem with fixing social security by taxing today's high earners is that it is a big intergenerational transfer.

Today's old rich folks got to pay low taxes during their earnings years and reap the benefits of full social security today on the backs of today's young rich.

Scott Sumner writes:

Brad, That may be correct, but Christie's plan is unjustified either way. Perhaps we should cut the benefits of the currently retired, but it should be based on their previous wage income, not their current total income.

Seth Green writes:

"I usually visualize the Democrats as representing the "one marshmallow" Americans."

This is a great line. But then, whom do Republicans represent: people who don't want to be confronted by unwelcome social and demographic changes?

David W writes:

It's unfair to penalize people for saving, true. But so what? We're out of money, and we've got to save somewhere. Better to cut outflows than increase taxes again.

Once someone has made it to 65 with zero in the bank, we're not going to let them starve to death. Not even if the reason they're broke was extravagance.

Further, this isn't a huge disincentive. The maximum benefit is 2452/mo, ~30K/yr. A linear phase-out over 120K is a marginal rate of 25%, an average rate that's low and never higher than 15%. You're still better off for having saved, better off for finding ways to boost the return on your assets, and even better off for continuing to work.

Floccina writes:

Everyone who retires at a given age should get the same payout.

ThomasH writes:

This is another iteration of the theme that progressive consumption taxes are better than progressive income taxes. Like most Liberals I agree. So ideally we should fix SS with revenues from a progressive consumption tax. Alternatively maybe a higher, uncapped, payroll tax with the first x thousand of income exempt could work.

RogC writes:

[Comment removed pending confirmation of email address. This is your final notice. We have attempted to contact you several times previously. Email the webmaster@econlib.org to request restoring this comment and your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Silas Barta writes:

I'm not sure this is a natural category you're drawing. If someone invest his savings into amplifying his salary (eg training), but does not otherwise save, does that count as a high saver (not meriting a higher tax rate) or a high income person (and thus meriting)?

[broken url fixed--Econlib Ed.]

Bostonian writes:

"Unlike the Christie plan, a phase out based on lifetime wage income would not distort the decision as to whether to consume now or in the future. It would not discourage saving and investment."

It is odd that Sumner does not even try to quantify the disincentive to work caused by his phase-out based on lifetime wage income.

I also wonder why someone who highly values progressivity is writing at a libertarian blog. I thought libertarians generally favored the protection of private property and opposed redistribution.

Anand writes:

The bigger problem is that means testing will not produce much money, unless one defines "rich" as something like $40,000.

Scott Sumner writes:

Seth, Yes.

David, You said:

"Better to cut outflows than increase taxes again."

Cutting outflows is exactly the same as higher taxes. They are identical. Suppose we kept outflows the same, and imposed a tax equal to the proposed cuts. Identical.

"Once someone has made it to 65 with zero in the bank, we're not going to let them starve to death. Not even if the reason they're broke was extravagance."

I agree, and never suggested cutting benefits to zero. Perhaps the benefits for the high earners could be cut to the Social Security minimum.

"Further, this isn't a huge disincentive. The maximum benefit is 2452/mo, ~30K/yr. A linear phase-out over 120K is a marginal rate of 25%, an average rate that's low and never higher than 15%. You're still better off for having saved, better off for finding ways to boost the return on your assets, and even better off for continuing to work."

Yes, but you forget that the tax code already has massive disincentives to save and work. This is being piled on top of the others. It's a bad policy, especially given that a superior policy option is easily available (and actually far easier to administer.)

Floccina, That's reasonable, but it's not even close to being politically feasible right now. Perhaps that sort of system could be very gradually phased in over time.

Thomas, Or adjust the formula for benefits, to make it more progressive.

Silas, Yes, that's a problem with my argument. I'd make two responses. We already subsidize education much more than physical capital acquisition. And let's not let the perfect be the enemy of the good.

Bostonian, Many libertarians like Milton Friedman and Friedrich Hayek favored a modest amount of redistribution. So do I. I call myself a libertarian because I favor a radical reduction in the size of government. If I had to dispense with either redistribution or open borders, I'd get rid of redistribution and keep the open borders. That's a strongly libertarian view.

Both my plan and Christie's plan creates a disincentive to work. His is also a disincentive to save, whereas mine is not.

Anand, Yes, this is mostly symbolic.

Bostonian writes:

"Both my plan and Christie's plan creates a disincentive to work. His is also a disincentive to save, whereas mine is not."

If someone wants to earn more with the intention of spending the extra income before retirement, your plan creates a disincentive to work that Christie's does not.

Joe The Economist writes:

Scott, Social Security is already massively progressive. The marginal dollar of contribution for a high wage worker is 15% vs 90% for a lower-wage worker.

Since inception the system has collected about $15 trillion - In Total. It has in taking that money created roughly $25 trillion in empty promises. Maybe the problem isn't in the revenue but in the promises.

Justin writes:

That's a huge tax on saving for people who plan to earn $80,000/yr-200,000/yr in retirement, which I'm imagining is the target range for most affluent workers.

The maximum retirement benefit is $33,672/yr, which means the government would take away up to 28% of the gap between $80,000 and $200,000 in retirement income, on top of already existing taxes.

I don't know about anyone else, but if this happened I personally would cut my 401k contributions from the IRS limit to the amount needed to earn the match and just spend the savings on overseas vacations and European cars. I'm already forecasting negative real returns on my Social Security contributions, but that beats -100%.

Justin writes:

David W,

It's a huge disincentive to save. 401(k) withdrawals are taxable as ordinary income and most of the $80k-$200k phase out range falls into the 28% tax bracket. Combine that with the 28% implicit tax of losing all social security benefits and several additional percentage points from state income taxes, and we're talking a 60% marginal tax rate on money I might die before I ever see. No thanks.

Scott Sumner writes:

Bostonian, Yes, but again my plan is more efficient on that basis, as it's better to have a lower tax rate on labor income spread out over a larger number of years.

Joe, The system was a mess from the very beginning, starting out with the fact that (as you point out) it was pay as you go. It should have been fully funded, with individual accounts.

Justin, Exactly.

AS writes:

Justin, I think you are still missing some taxes, such as estate, and corporate. Mankiw made a good post about taxes.

BC writes:

Scott, I agree with your point that taxes ideally should be neutral with respect to savings decisions, which Christie's plan is not.

However, savings effects may not be the only incentives to consider. People often characterize Social Security as intergenerational transfer. However, it doesn't seem to be any more of a transfer than the previous system of having children take care of their parents. What's different about Social Security is that seniors' children now must instead take care of Other Seniors. In turn, seniors without income-producing children can rely on Other Seniors' children. Thus, Social Security externalizes the previously internal retirement benefits of having children. People deciding to have fewer children has coincided with this 80-yr experiment, which in hindsight probably should not have been a surprise.

There are two possibilities:
(1) After 80 years of economic growth, enough people are wealthy enough to finance their own retirements, even after a lifetime of paying Social Security taxes, so that we can limit Social Security to only the truly poor; or
(2) It's unrealistic to expect one generation to escape dependency on the next through their own savings. In this case, Social Security causes its own demographic demise, and it would seem unwise to continue telling young workers to not bother having children because they can always rely on Someone Else's children in their retirement.

Either case implies limiting Social Security benefits (over time) to only the truly poor. Sure, there are negative work and savings incentives, but no more so than any other program for the poor. Then, Christie's proposal could be justified if it is just the first step in phasing out Social Security for all but the most poor.

Comments for this entry have been closed
Return to top