David R. Henderson  

Means Testing Isn't "Awesome"

Means-Testing and Political Ec... Matt Zwolinski on Sweatshops...

Over two years ago, Bryan Caplan posted on why means testing is "awesome". He didn't actually make that case, though. The case he actually made is that means testing is a good idea. "Awesome" is a step above "a good idea."

I wrote a post critical of his. In it, I made three criticisms of means testing. The first is that the implicit marginal tax rates that means testing would require could be quite high. I wrote:

Just eye-balling the data, I estimate that the bottom decile (about 8 million families) in 2008 had income between $0 and $17,500 and the second from the bottom decile had income from $17,500 to about $28,500. So that's $11,000 of income over which to phase out a $5,000 benefit. The implicit marginal tax rate from that phase-out alone is, therefore, $5,000/$11,000 or 45%. That's on top of other tax rates.

I also pointed out that there are two measures of means: wealth and income. Wealth is a better measure of means but I noted that income is the measure likely to be used. I wrote:
Therefore means testing would discriminate in favor of wealthy people with low income. Because the most expensive programs for which means testing is advocated tend to be for the elderly, this is an even bigger problem. Among the retired population, the correlation between income and wealth is even weaker, I believe, than among the population in general.

Finally, I pointed out that there's a fairness issue here. I wrote:
Again, because the programs at issue tend to be for the elderly, there can be huge differences in income because one family saved a large percent over the years, and is earning interest and dividends on that income, and the other family saved 0 and relies on Social Security. This could be so even though the two families had a similar age-earnings profile.

Bryan argued with me:
Right now we already means-test a lot of programs, such as Medicaid, food stamps, and housing vouchers. Question: Should we make the entire population eligible for these programs, regardless of income and wealth? If not, why not? If you don't want to transform existing means-tested programs into universal programs, why don't you want to transform existing universal programs into means-tested programs?

That persuaded me. Here's where I cried uncle. Specifically, I wrote:
I don't want to see means-testing removed for those programs because I think my arguments I gave earlier are outweighed by the huge cost of expanding those programs greatly. I don't have a good argument and I think you have correctly identified something that I didn't even know was there: status quo bias.

My point with this post is not to take back what Bryan persuaded me of. Rather, it's to try to persuade Bryan and others who share his view to admit the criticisms that many others, including Scott Sumner and I, have made. Bryan re-advocated means testing in his latest post without even hinting that these three problems I raised are indeed problems.

Comments and Sharing

COMMENTS (14 to date)
Mike W writes:

Social Security and Medicare Parts B & D are already means tested.

A couple receiving SS benefits with greater than $44,000 of Modified Adjusted Gross Income would have 85% of their benefits taxed on their 1040 at their marginal tax rate. At a 25% marginal tax rate they would pay back about 21% of their benefits.

So...what's the point of this discussion?

David R. Henderson writes:

@Mike W,
You really don’t understand, in light of what Bryan and I wrote, what the point is of discussing means testing?

MikeP writes:

At a 25% marginal tax rate they would pay back about 21% of their benefits.

That's small solace to the waitress in the 0% marginal tax bracket who gives up 12% of her income to support them.

So...what's the point of this discussion?

That there should be much more means testing.

In fact, I think that both Social Security and Medicare should be phased out in favor of welfare and Medicaid paid out of the general fund. Those programs are welfare that pretends to be insurance, and it is criminal to tax 18-year-olds and send the proceeds to millionaires.

Mike Rulle writes:

There should be no means testing. This would be an ex-poste tax increase.

Having said that, there can still be reform. While the government confiscated payroll taxes to spend on other things (the amount diverted is exactly equal to the amount in the literally Orwellian named SS Trust Fund). Individuals did pay into the system. Perhaps, like the Picard suit in the Madoff Case, we should seek to take back from who received the $5 Trillion in the "Trust Fund". That is sarcasm.

All SS income should be taxed like 401k withdrawals. I assume they are. After all, federal income tax is not paid on payments into the system. So why means test it, unless all that is meant by that is simply paying a normal income tax rate at your bracket? Why then don't we means test 401k payments too? We can do other things like lower the growth rate of SS payments to CPI; or push out the years to 68-70; or even max out the number of years it is received. But the fair future value that was paid into the system should be paid out. That at least is the starting point before reform.To the extent we need to lower overall payments, then it should be done so without imposing an ex-post tax increase on some select few.

Just because our legislators effectively misallocated the funds, it does not make SS an entitlement. It is savings from the individual's perspective.

I repeat my meme I use often. We are less wealthy than we think we are, in large part because of Government spending policies of past decades. But lets not change the meanings of what things are. All should have to experience some belt tightening---but not by redefining SS by pretending it is welfare.

Jim Glass writes:

Means testing is very much like democracy as famously described by Churchill, a terrible way to do things except in comparison to the alternatives.

Alas, the benefit of proposed program reforms must be determined relative to workable alternatives, not abstract ideals.

"Status quo bias" (or path dependence) in politics and political economics is *huge*. Milton Friedman notably used Social Security as an example of this: as a new program it could never possibly be enacted in today's world on its existing terms, either whole or in its major component parts. (Hey, let's make America more progressive by starting more new social spending programs that benefit millionaires along with everyone else -- indeed, giving the rich the largest dollar benefits -- funded by new payroll taxes applying from the first dollar of the very lowest-earner's wages!)

Yet given its existence, any attempt to reform it receives only virulent attacks, and so far remains the kiss of death in politics.

Jim Glass writes:
"A couple receiving SS benefits with greater than $44,000 of Modified Adjusted Gross Income would have 85% of their benefits taxed on their 1040 at their marginal tax rate. At a 25% marginal tax rate they would pay back about 21% of their benefits."
That is an effective means test, being that the "income tax" on SS does not become general revenue as all other income tax does, but instead is quietly remitted back to the SSA, which results in a net reduction of benefits based on level of income.

But note how carefully that means test has been disguised as ordinary income tax by the politicians. And successfully so. You won't find one person in 20 who calls it a means test. SSA doesn't call it a means test. It's just "income tax".

In no way has means testing of SS been politically acknowledged, much less embraced. See the Democrats' response to Pozen's proposed progressive indexing of benefits the last time the Repubs dared suggest SS reform.

"So...what's the point of this discussion?"
As to Social Security the point is that it is underfunded by $20 trillion at present value according to the SS Trustees -- or starting in about another dozen years by 1.5 to 2 points of GDP annually in terms of cash flow, which is the measure that really matters. In today's terms that's around $250 billion to $300 billion annually.

How is this cash flow gap going to be closed? When the era of denial finally ends and push finally comes to shove, by far the most likely political resolution is just what happened in 1983, a combination of tax increases and benefit cuts (calculated in 1983 at 50%-50%, the classic compromise!). Whose benefits are going to get cut?

Again just like 1983: (1) the young who "have time to adjust" (and aren't going to lynch the politicians for cutting their current benefit income) and (2) the well off who can afford to have their benefits cut. After all, as long as SS justified as "social insurance" against poverty, how can they protect the benefits of the rich by cutting the benefits of the poor? This time though, they won't be able to disguise the means test they use as anything but a means test.

But what other politically plausible option will there be to close the cash flow gap? Raise taxes on everybody rich and poor to preserve full unreduced benefits for the rich? Minimize tax increases by reducing the benefits of the poor as much as those of the rich?

As to Medicare ... multiply the above by about 3X.

Mike W writes:


I guess I was confused because Bryan's post dealt with folks at the low end of the income range and yours addressed the wealthy elderly. But I guess you're both saying, "those who can should take care of themselves and government assistance should be only for the truly needy". (While I enjoy the rhetoric of Libertarianism...the approval of the markets and personal responsibility and the empahsis on less government...I sometimes forget that the philosophy is as fanciful as that of the social justice crowd.)

Bryan seems to be OK with government assistance as long as it is just for orphans and cripples and you appear to be OK with it as long as it is just for the less than wealthy, so the only question seems to be where do we draw the line as to who qualifies as "needy"?

In the real world (i.e., not that of academic theory) folks in the lower half of the income scale need help with pension and senior health care. SS and Medicare fulfill this need...and the reality is that they are funded by progressive taxation.

The stuff about those programs being "insurance" or "savings" is nonsense. They are funded by taxes on income assessed through payroll. They are promises by our society that we will ensure that the elderly do not live in poverty after a lifetime of work. As they are essentially income taxes...to which everyone, including waitresses, should contribute...then means testing is just another way of arriving at a funding scheme that has those who are better off contributing to the welfare of those who are not so well off.

David R. Henderson writes:

@Mike W,
Actually, my post above dealt explicitly both with those who are wealthy and those who are at the bottom end.
And, no, neither Bryan nor I are “OK” with it forcibly taking from some to give to others: what we’re both wrestling with is what to do about the current system.

Jim Glass writes:

In the real world (i.e., not that of academic theory) folks in the lower half of the income scale need help with pension and senior health care. SS and Medicare fulfill this need...

While also making transfers to everyone in the top half from Warren Buffett on down. One might think that a program intended to benefit the lower half would be better targeted.

That it's not might reflect the fact that SS was *not* created to benefit the "lower half". As Paul Samuelson famously explained in his Newsweek article back during Social Security's heydays of the 1960s...

The beauty of social insurance is that it is actuarially unsound. Everyone who reaches retirement age is given benefit privileges that far exceed anything he has paid in -- exceed his payments by more than ten times (or five times counting employer payments)!
How is it possible? ... Always there are more youths than old folks in a growing population. More important, with real income going up at 3% per year, the taxable base on which benefits rest is always much greater than the taxes paid historically by the generation now retired...
A growing nation is the greatest Ponzi game ever contrived.
Not a word about targeting the needs of the "lower half" in that.

Which is why today Warren Buffett is collecting max benefits accumulated at an above-market rate on his contributions over his lifetime -- paid for by his Dairy Queen employees and their peers, who will pay more in tax than he did, for benefits reduced from his, earning only a below-market rate on their contributions.

And even then their benefits are underfunded by $20 trillion present value. Great way to help the lower half!

and the reality is that they are funded by progressive taxation ... essentially income taxes

A 15.3% flat tax on wages - and only on wages - from the first dollar up to $110k, became a "progressive income tax" ... when?

perfectlyGoodInk writes:

Why not employ more and more means-testing so that fewer and fewer people qualify for Social Security benefits, thus addressing its impending shortfall and then eventually phasing it out altogether?

Steve Miller writes:

The most common response to David Henderson's objection #3 (and it indirectly addresses some of what's going on behind objections #1 and #2) is: Don't means-test seniors based on *current* income or *current* wealth. Instead, means-test Social Security and Medicare based on *lifetime* earnings.

David R. Henderson writes:

@Steve Miller,
Correct. That’s the most common response. Alas, I think the probability that the government would give 0 Social Security to someone with close to 0 income and 0 wealth is close to 0. There would be a strong tendency to look at current income or current wealth.

Telnar writes:

In @Steve Miller's defense, the benefit could be split into two components: A welfare-style minimum income benefit which is means tested based on current income and which might develop a stigma over time (since going through the intrusive paperwork to claim it signals a failure to save and invest well), and a more traditional retirement benefit which is means tested by lifetime income. That model doesn't require letting those with $0 in income and wealth starve, but it also provides an incentive to save.

John R. Graham writes:

I may misunderstand Mike W.'s first comment. Means-testing for Medicare Part B kicks in at $85K MAGI for a couple, not $44K (www.socialsecurity.gov/pubs/10536.html#a0=5). SSA very carefully points out that this comprises only 5% of the SS population (http://www.socialsecurity.gov/pubs/10536.html#a0=0).

Medicare Part B is nominally voluntary but effectively universal. If we talk about capping the deductibility of out-of-pocket medical expenses, I think we have another question. It would be means-testing but not on general income. I think that makes more sense but I won't delve into it on this comment (except to note that it would be limiting a deduction and therefore reducing perverse incentives in the tax code).

In Canada, the federal government means-tested Old Age Security (at a high level of income)starting in the 1990s and I recall that personal-finance columnists immediately started writing articles about how to keep income below the cut-off (which is a slope, not a cliff). So, it does effect behavior.

So, there is a dynamic of means-testing at low levels of income for general programs (Medicaid, food stamps, etc.) and means-testing at high levels of income for age-tested programs (Medicare).

I think it's pretty obvious that this arrangement is attractive to politicians and will become even more attractive as the population continues to age. Perhaps there's no use fighting it.

What I have struggled with is the following: Is an increase in marginal tax rates more harmful when applied to low-income earners (poverty trap) than when applied to high-income earners?

Republican politicians have bought into the latter but they do not address the disincentive for high earners to keep earning. Because their earnings are more likely to be from investments rather than labor, we might have a real problem here. (There are already enough disincetives to saving in the system.)

I also appreciate that the means-testing should be a slope,not a cliff, but I can also see the political appeal for it to be a cliff for means-testing low-income earners and a slope for means-testing high-income earners, which is what we are seeing in the U.S. and Canada, at least, and probably other developed democracies.

[I suppose all readers are familiar with an idea started out by Milton Friedman, I believe, and lately proposed by Charles Murray "In Our Hands" (2006) of a guaranteed annual income.]

Comments for this entry have been closed
Return to top