Bryan Caplan  

Means-Testing and Status Quo Bias

Bad Predictions, Durant Editio... Response on Means Testing...
I'm a big fan of means-testing the welfare state, but many of my favorite people disagree.  I've got a challenge for these hold-outs.  Here goes:

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? 

Comments and Sharing

COMMENTS (17 to date)
Richard writes:

A universal hand-out discourages work through the income effect. But a means-tested hand-out discourages work through both the income effect and the substitution effect. So the inefficiency -- the loss of economic production -- is much greater.

Of course, any hand-out has to be paid for with taxes, and a universal hand-out requires higher taxes than a means-tested one. But, at least with Social Security, there is a relationship between how much a person puts into the system when he's working and how much he can take out after he retires. So the work penalty created by the Social Security tax is weaker than that created by, say, a general income tax. But you'd destroy this relatively beneficial quality of the Social Security tax if you severely means-tested the benefits.

Dave writes:

I would be for means testing Social Security and Medicare if the means test were based on lifetime income--not retirement income. We don't need more policies that penalize saving.

Nikonman writes:

One of the objections to means-testing of Medicare is the idea that it would lead to a two-tier health system, where those with means would take advantage of better doctors, treatments, etc., than those allowed under Medicare.

My understanding of the 1993 Clinton-era policy of tying Medicare to Social Security benefits was to prevent seniors from withdrawing from Medicare, despite the runaway costs at the time. The major objection was that seniors who pursued their health care options privately would create a two tier system. Tying it to Social Security benefits prevented many seniors from withdrawing from Medicare.

It seems to me that means-testing would do the same thing. Not that I am against it, mind you, but this objection has been out there for 17 years now.

Tom West writes:

Three considerations. One, a lot of people already complain about the marginal rate in certain tax brackets, and making more programs means-tested adds to that burden.

Second, means-tested vs. universal is often a political question, not an economic one. Certain programs would only garner enough support if they are made universal.

Third, there is a psychological effect of "getting something for your money". Despite the difficulties and wait-lists, government provided health-care is enormously popular with Canadians ("Tommy Douglas, the father of medicare was voted "The Greatest Canadian" in a widely watched CBC program).

However, more than that, it's something that every Canadian can point to and say "at least I'm getting *something* for my taxes". In fact, more than once I've heard Americans say that about our taxes. Logically you can pick this apart, but it doesn't change the base emotion. There are a lot of Canadians, including myself, who feel grateful for the protection that medicare provides, even though they probably pay rather more in taxes than an equally good or better insurance policy would provide.

I suspect that universal programs make *all* taxes easier to bear, which is a would make them very attractive to progressives.

JPIrving writes:

@Tom West

I think I agree with you, and thus think that means testing is if nothing else, a good way to wean societies off the welfare state.

If there is a two tier system with one group living off the other we are sure to create a political situation where the suppliers of the transfers resent the "bums" living off their hard work. This happened with the old pre clinton welfare system in the U.S., and in Sweden during the 1990s to an extent.

This actually encourages me. It is widely accepted that some type of means testing is needed in the world's welfare states around the 2012-2015 period. The bigger the non recipient group grows the better the odds for reforming and shrinking the whole monstrosity! Hooray demographic apocalypse....

Devin Finbarr writes:

Yes, absolutely. Charles Murray proposes exactly this, and I agree with him.

The only test I might add is a children test for social security - a person should get more social security based on how many children they should have. Social security is essentially a socialization of the old method of children paying for the retirement of parents. But the socialization destroys the incentive to have children, thus leading to the great demographic problem.

A government is the property owner of a large piece of land. All governments will try and raise their rent (taxes) to the optimal (laffer curve maximizing) level. The government will the distribute the proceeds to the people who control it. The U.S. is a one person, one vote democracy. The most obvious solution is to distribute the money completely equally to voters. Again, the government sets taxes at a rate that will maximize revenue-purchasing power. It expends money first on goods and services that will generate more revenue-purchasing power (infrastructure, basic defense, etc). The surplus is then returned as a "national dividend" to all voters. If implemented efficiently, this would end up being about $10,000-$15,000 per voting American.

Tom Myers writes:

I'd like to end means-testing as such, because of what I see as several kinds of disincentive effects, but I suspect that any effort to simply remove it (e.g. from Medicare, as you say) without making other changes would make things worse, not better. It's quite possible, though, that you'd think of my preferences as just means-testing by another name. Still, the disincentives are different.
In my tax system, the waitress at the diner and Bill Gates would each pay the same proportion of income (or consumption, by preference) as tax, but each would also get the same size of weekly "redistribution" check, which Bill might not notice. (A net linear tax of y=mx-b, essentially Friedmanite.) I used to think that would be all the redistribution I would do, and that the redistribution check would cover healthcare insurance vouchers, but the recent debates have changed my mind about that. My universalist Medicare would be a tax-funded universal "insurance" program with a deductible set to 16% of your income (in some sense, that's our average nationally) so that prices are set in a market where BillG pays his own bills, I pay almost all of mine, and the waitress...well, she pays for anything small. I will skip some of the necessary details in the hopes of improving the chance that you'll read this, and respond -- do you think this is "means-testing" as you use the phrase?

Huxley writes:

I think if the law requires that people be poor to receive benefits, many people will take the government up on it. If the grasshoppers get treated the same as the ants, we will find more ants transforming into grasshoppers.

Doc Merlin writes:


In response to the "two tier" problem, we already have that, because medicare payouts are so low and dropping.

Doc Merlin writes:

@Tom Meyers
Your idea has a few flaws: it separates the payers from the payees even more than the current system does. It also has the problems of normal welfare systems in that it reduces the incentives to work harder.

Martin writes:

Which currently universal to be means-tested programs are we talking about? Examples?

Martin writes:

Which currently universal to be means-tested programs are we talking about? Examples?

PrometheeFeu writes:

I oppose means testing because I think it discourages innovation and entrepreneurship. Having a universal welfare program means you can take the risk to become an entrepreneur even if you don't have much of a cushion to fall back on. I think that if people knew they would always be able to rely on having some healthcare and a survival income, they would be much more likely to quit their job and try out something new which might work and be of economic benefit.

Means testing creates in-between periods when those who qualify haven't had their application approved yet. Also, means-testing requires big bureaucracies to validate applications, investigate fraud, etc... A waste of good resources which could be used for something better.

Tom Myers writes:

Doc Merlin, I'm not sure of the "separation" when everybody is both payer (in proportion to income, or perhaps consumption) and payee. Could you clarify?

I do agree that any social support, means-tested or not, does produce some disincentive, and that this is bad: the acrobat with a safety net will never try quite as hard to avoid falling as the one without. But I also agree with Prometheefeu that the safety net encourages entrepreneurs (and labor mobility between companies, and willingness to move to different regions and try new things). Also I admit to being basically just plain soft. So I like having some sort of safety net, and I agree again with Prometheefeu that conventional means-testing requires a bureaucracy (and motivates gaming that bureaucracy, which is an extremely unfortunate misincentive). My proposal doesn't have the same sort of administrative requirements; I believe that its misincentives and disincentives (and bureaucracy) would be less than what we've got now. (But I could be wrong; I usually am.)

Vangel writes:

I am against means testing because if we make it easier for the state to keep redistributing money from the productive to the unproductive it will keep doing so and the broken system that we are living with will last longer than it should. The sooner we wean ourselves from the cancer that is the welfare state the safer, healthier and better off we will be.

Dan Weber writes:

Like others have said, I dislike means-testing because of the negative disincentives. Means testing is a marginal tax.

People should be able to choose when they want to leave the public program and enter the one they have to pay for. So give everyone either health care or a small tax credit to buy their own. People of means will buy their own because they don't like the limits of the public system, but they aren't forced out of it if they work overtime shifts.

Simon Kinahan writes:

No. I would prefer to abolish all of the those programs and replace them with a non-means-tested per-capita refundable income tax credit (aka basic income guarantee) indexed to a basket of basic necessities with some minimal level of forced saving.

Why? Because means testing is just part of the general problem with all of these programs - they involve pre-judging the worthy versus unworthy poor, and worthwhile versus non-worthwhile expenditures. The rules for what food spending is and isn't eligible for food stamps, for example, are Byzantine. Except thats unfair to the Byzantine emperors, who just gave away free bread. As such they limit the options of the poor and effectively trap them in poverty. The problem of massive effective marginal tax rates is only one aspect of that problem.

People - even (possibly especially) poor people - are better judges of their circumstances than bureaucrats and politicians. So if you're going to help them, as far as possible you should help them with what they believe they need help with.

Comments for this entry have been closed
Return to top