Bryan Caplan  

The Welfare State as Extended Warranty

Piketty, Piketty, Piketty, Pik... Spot the Problem...
"Extended warranty?  How can I lose?"

            -- Homer Simpson

Value is subjective, and taste for risk varies widely.  But every economist I've asked - and virtually every savvy consumer - concludes that extended warranties are a lousy deal.  The short case against extended warranties:

1. You normally get a very good warranty for free.  When you buy an extended warranty, you only receive a marginal increase in security. 

2. This marginal increase is usually grossly overpriced.  You often pay 20% of the purchase price to protect yourself against a 1% risk. 

3. In any case, by the time your extended warranty kicks in, your product has already lost most of its value.  Most of the time, the vendor just repairs and returns your original semi-obsolete device. 

The lesson: Even if you're a deeply anxious person, self-insurance is the prudent course: Just take your chances, and use all the money you save to buy replacements when the need arises.

If extended warranties are such a bad deal, why are they so popular?  Psychological bias.  Vendors are preying on your absurd yearning for certainty - your desire to know that you'll never lose what you have.  Why absurd?  Because true certainty is unavailable at any price.  Can you be certain that the vendor will remain in business?  No.  Can you be certain that the vendor will honor the contract?  No.  Can you even be certain that you'll still be alive tomorrow?  The answer, I'm afraid, is no.  Extended warranties sell because they provide the illusion of certainty - and many human beings are too soft-headed to disbelieve the illusion.

Readers often misinterpret my The Myth of the Rational Voter as arguing that people are rational in markets, but irrational in politics.  My actual view is simply that people are markedly more irrational in politics than in markets.  Given the bustling market in extended warranties, then, my natural reaction is to ask: Does this absurd yearning for certainty play an even larger role in politics?  My answer is a resounding yes.  The absurd yearning for certainty lies at the heart of support for the welfare state.

Think about the populist case for fully socialized medicine.  What's supposed to be so great about it?  The certainty.  You know that no matter what happens, the government will provide your health care.  Of course, this certainty is merely rhetorical.  Can you be certain that the government will cover the treatments you actually need?  No.  Can you be certain that the government will competently deliver whatever treatments it covers?  No.  Can you even be certain that your medical problems are curable?  The answer, I'm afraid, is no.  What is the point of this rhetoric of certainty?  To confuse voters so they don't ask the questions that matter to savvy consumers.  Namely:

1. What is the marginal health benefit of socialized medicine?

2. How much extra does the government charge for this marginal health benefit?

If voters did ask these questions, they'd find considerable evidence that government health spending, like extended warranties, is a crummy deal.  You pay an arm and a leg, and the marginal improvement in health is slight.

The same goes for government retirement programs.  What's so great about Social Security?  You know that no matter what happens, the government will provide your retirement income.  But once again, this certainty is merely rhetorical.  Can you be certain that Social Security will give you a decent standard of living?  No.  Can you be certain that Social Security will suffice for your basic needs?  No.  Can you even be certain that you'll still be eligible for Social Security by the time you retire?  The answer, I'm afraid, is no.  But the rhetoric of certainty silences the awkward questions every savvy consumer should ask:

1. What marginal income security benefit does Social Security provide?

2. How much extra does the government charge for this marginal security benefit?

If voters did ask these questions, they'd find that projected Social Security returns are pretty crummy for most people - even according to the Social Security Administration itself.  The SSA's own estimates look worse yet when they factor in probable future changes in the benefit formula.  There's "certainty" for you!

Are there any important differences between the market's extended warranties and the government's welfare state?  Sure.  Most economists will highlight a difference that makes the welfare state look better: Extended warranties usually insure against relatively small losses; the welfare state, in contrast, often insures against relatively large losses.  These economists have a point, but exaggerate: The welfare state's biggest programs "insure" against the rather foreseeable problems of old age and unprotected sex.   

Relatively few economists, though, will highlight a difference that makes the welfare state look worse: If you don't want an extended warranty, all you have to do is tell the vendor, "No thanks, I'll take my chances."  Once you realize that extended warranties are a bad deal, you never have to get ripped off again.  The welfare state, in stark contrast, won't take no for an answer.  Once you realize that the welfare state is a bad deal, the rip off continues until the day you die.  Or flee the country.

Comments and Sharing

COMMENTS (26 to date)
Steve Roth writes:

I think you're double-dealing here, straw-manning the opposition.

re: MOTR, you disclaim an absolutist argument, saying you're making an argument of degree:

"My actual view is simply that people are markedly more irrational in politics than in markets."

Good argument, IMO.

But re: socialized medicine, you claim that proponents are using an argument of absolute certainty.

"You know that no matter what happens, the government will provide your health care."

Give them the credit you give yourself:

"More people will be more certain that if something bad happens, they'll be covered."

That expanded certainty has utility. So if socialized medicine increases aggregate certainty, that effect (there are many others) increases aggregate utility.

Peter H writes:

I think the extended warranty example actually tilts a bit more in favor of the socialized medicine example.

Specifically, when it comes to extensive medical interventions, the following is the case in comparing socialized systems to private insurance schemes:

1.) I have little certainty in either case. Everything you said about providing insufficient or inadequate care applies to a private health insurance company as well. To the extent that the security I get from a government run system is rhetorical, the security I get from private insurance is comparably rhetorical. And the mechanism I would use to compel treatment in either case (regulatory complaints followed by a lawsuit) is substantially identical.

2.) The guarantee of ongoing coverage is actually incredibly valuable, because severe illness is likely to correspond to an income shock. If I am sick enough to need $100,000 worth of medical treatment, I am almost surely not working, and may have been fired by my employer. So I may be unable to pay my insurance premiums and may be dropped by my insurer. Many if not most people do not have the savings to pay their costs of living plus health insurance premia for the period of an extended illness.

To your questions:

What is the marginal health benefit of socialized medicine?

Little if any.

How much extra does the government charge for this marginal health benefit?

Probably a negative amount, at least as compared to current total expenditures. Government in the US spends more on healthcare than government in Canada (by far the most comparable country to the US as far as demographics, culture, and lifestyle). For that greater spending, the US gets much less than universal healthcare. I don't argue that the US could implement a Canadian style system at Canadian prices, but they probably could do so at less than is currently spent by the combined government and private insurance cost.

Steve Roth writes:

IOW, do what you do at your best: argue with their best argument.

Steve Roth writes:


If Peter H is right that private and public insurance are equally uncertain, then aggregate certainty is a purely function of the number of people covered.

Socialized covers more, ergo...

Peter, Canada doesn't have universal health care. They have 'universal' health care financing ('insurance'). A different thing entirely.

I have Canadian friends, one of whom told me, after having to shell out $50,000 for his father's surgery at the Mayo Clinic--he had stomach cancer, and was put on a waiting list for tests in Ontario--that, 'They tell you it's free, but, you know what, you get what you pay for.'

That same Canadian, a few years later, had to avail himself of Mayo's services to save his own life. The conventional wisdom among Canadians is, if you're really sick, go to the USA for treatment.

Also, the most amusing example of ridiculously over-valuing an extended warranty, came from none other than Barack Obama. During the so-called Health Care Summit, prior to the passage of Obamacare, he opened his presentation with his own story of not having been sold collision insurance on his beater automobile when he was a community organizer. He actually seemed to think his auto insurance company had done him a disservice by not taking his money for something that was essentially useless for his situation.

Jeff H writes:

I call straw man. Bryan asks:

Can you be certain that Social Security will give you a decent standard of living? No. Can you be certain that Social Security will suffice for your basic needs? No. Can you even be certain that you'll still be eligible for Social Security by the time you retire? The answer, I'm afraid, is no.

The first two questions ask things Social Security was never intended to do. It was always intended to part of a persons income in retirement, not the full amount. The last question may not be a certainty, but the probability is quite high. If past is precedent, then a good starting assumption is that, yes, there will be some minor changes in eligibility rules, but it won't disappear barring the demise of the United States. So let's just say the probability of you being eligible for Social Security at retirment is above 0.95.

I'm not going to get into the fact that SSA has quite low actual operating costs relative to what it takes from some people and doles out to other people. But if you actually care about the individual benefit, try looking at it. Here is a link to the estimated replacement rates.

And then ask what you think people would do without Social Security and then look back at what people actual did without Social Security.

Phil writes:

I think people know it's just an illusion of certainty. My view is that people are scared of making a bad decision in a free market of choices, and this lifts from them the worry that they may regret what they did. It's peace of mind they want, not certainty.

Especially when some options cost more than others. People with children ... they might really want to buy the cheaper healthcare plan, and that way they can afford that new car. But they'd feel worried and guilty. Even if government health care is crap, now, it's not their fault for not buying a better plan.

Peter H writes:

Patrick, I have lived in Quebec for several years and am currently seeing a guy in Ontario and travelling up there often, so I have a decent familiarity with the structure of the system. (Or systems, plural, if we want to be pedantic).

But my analogy still stands. My US health insurance would not cover me travelling to the Mayo Clinic for a $50,000 procedure. If I were told my in-network providers were fully booked and I had to wait, I would have a similar problem to your friend's dad. If I flew out to MN and got the procedure done out of network and out of state, it would be from my own pocket.

Just because the Ontario system is not comprehensive (and it varies a bit by province, but you used ON as your example) does not imply that private insurance in the US doesn't have the same or similar problems for patients.

Jeff writes:

Here's the relevant question I would ask SS defenders: given the awful returns we can expect from the program, how much better off would the average person be if the program were means tested and the resulting savings either invested in broad-based index funds or simply returned to people paying into the system?

Peter H writes:


Not a particular fan of Social Security, but I wonder how a means test for it could really work? Based on lifetime income? What happens to people who were not thrifty or had a business failure or a large lawsuit or somesuch which wiped out their savings?

Further, this acts as a pretty hefty increase to marginal tax rates at the high end. If I know that a year working at over $100k reduces my SS benefit by so many dollars per year, that has the same disincentive effects as raising my marginal income tax rates.

Based on assets? Seems like it would result in a lot of gamesmanship, asset hiding, and fraud, like we see now with medicaid nursing home care...but ramped up to 11.

The NPV of a $1500/month social security check is in the neighborhood of $250,000. People have done a lot of fraud for a lot less money than that.

It also creates a huge disincentive to save. If I lose social security benefits based on my savings, that will function like a tax on savings.

Robert writes:

Jeff H,

I checked the link for the estimated replacement rates that you provided to me, and it is not clear how good the benefit is. What are they compared to?

Even if the administrative costs are low, what is the net present value of my expected social security benefit compared to the net present value of 12% of my lifetime income?

From what I understand, it provides inexpensive financing for the government. Shouldn't we factor in the probability of uncertainty due to the fact that the money saved in a bank or other retirement account is more liquid? Was it not determined that Congress can change the benefits at any time?

At the end of the day, I can understand the popularity of SS, since even if I make a bad loan and do not get what was promised, I would still like to get back as much of my principle as possible plus any interest I can get. Oh, how much of the flattening of incomes of the middle class from the 1980s was due to the increase in the payroll taxes from the reform of social security?

As far as socialize medicine, is it normal in economics to consider organizations, which have monopolistic power, have the power to influence the courts, law enforcement, and taxation, and have limited accountability to their individual customers, to provide better, more innovative services and at lower cost?

What are peoples' feelings about the VA?

Are there any concerns about the incentive effects of reporting cost figures by the organizations trying to increase revenues? If I pass my costs to suppliers to get it off my books, have the costs gone away?
If I do not prevent or report fraud, have the costs to the system gone down? Is this an issue?

Anyway, just some thoughts and questions, not that it really matters. I vote in Queens, and next year my vote will probably be diluted even more. I do really enjoy this blog, and I enjoy reading peoples’ comments. Thank you!

Jeff H writes:

Peter H,

The United States already has a means-tested benefit program for the elderly--it's called the Supplemental Security Income (SSI) program. They say there's nothing new under the sun, and if you think the US is going to change Social Security to a means-tested program, it will probably look a lot like SSI. And who do you think operates SSI? Why, the Social Security Administration, of course.


You can compare the replacement rates to whatever you want, that's the beauty of them. I have no idea what the rate of return on your personal investments would be. The point of my providing the link was to at least make sure an actual comparison is made before making broad statements--and you can add in whatever uncertainty you want. If you just look at the financial issue, you at least can back out what return you would need to get to "beat" Social Security. (Which you probably won't know with certainty anyway, contrary to what Jeff implies...)

Above all, though, let's not fool ourselves into thinking government would stop trying to "do something" for the elderly, the poor, etc. The questions should not be about what is ideal, they should be about what is realistic.

PaulS writes:

OK, by the time my car is ready for extended warranty, it isn't worth much (on the scale of cars.) At least in principle it could be replaced in kind without overwhelming trouble. For smaller consumer items, the "insurance" is worth so little that just the "paper"-work can add absurdly to the 'load' (ratio of pay-in to cost of provision.)

But Peter H is on to something.

Medical "care" is altogether different. It is at the very best an experience good, and more usually a credence good, both of which carry long illustrious histories of fraud. One cannot ever know whether the Coffee Study of the Week, or the Statin Study of the Month, or any of their multitudinous brethren and sistren, contain even the tiniest kernel of truth. Even when a genuine need for treatment is evident, it will often have arisen too quickly for "shopping" to be even a remote possibility. (And no lifetime could ever be long enough to pre-"shop" everything: even doctors specialize.)

To top it off, entry to the field is obstructed by barriers far higher than actually necessary for reasonable safety. Self-treatment of trivial conditions is often likewise obstructed. In our litigious culture, even the most contrived scenario of a one-in-a-gazillion chance of something going wrong "justifies" quasi-infinite rent-seeking.

Put it all together, and it should be no surprise that medical "care" is becoming a wholly corrupt racket replete with both massive overtreatment, and utterly unpredictable charges ranging over factors of 10 or even 100 for the identical course of treatment.

The bottom line: over a human lifetime, the fairly likely penalty for lacking "insurance" is not merely having to buy a new TV unexpectedly; it's utter and permanent penury and financial ruin unless one is a multimillionaire. It simply doesn't matter that both the insurance and the "care" itself are often raw deals.

Demanding the socialized form becomes essentially an act of desperation. One knows full well that the "service" will often resemble that cluster-thingie of shiftlesness, sloth, corruption, and ineptitude known as public road "construction" ... but the visible alternatives look even worse.

One would have to invent some alternative that doesn't leave people bankrupt, and doesn't find an excuse to drop them the instant they need anything. Cash on the barrelhead will not do. It's more suited (or used to be) to toasters and TVs, or trivia like flu shots. There's no guarantee that a serious ailment or accident will hold off politely until enough money can be accumulated in a savings account (which everlasting ZIRP does nothing to help.)

So in the end, I haven't the foggiest notion what to do when what we have now finally collapses of its own ridiculous Byzantine complexity - beyond setting up some sort of NHS-like thing, holding my nose to ward off the stench emanating from same, and praying loudly and often for good "health".

Peter H writes:

Jeff H,

SSI is only available if one has not worked enough to qualify for the more generous Social Security program. If you work for 10 years or more in your lifetime, you're not going to get SSI.

And it's rife with fraud, especially for the payments SSI makes for the non-elderly disabled.

If you abolished Social Security and just had SSI or something like it for elderly people, you would suddenly have a lot more fraud and asset hiding, since people who had worked and been able to save would have incentive to hide their assets. Likewise, the under/over threshold (which is a firm cutoff for SSI) would be the difference for a present value of $100,000+ That's a big implicit marginal rate.

My personal preference would be to just further progressivize Social Security payments, which would have some of the means testing effect, without as much distorting of incentives.

Kevin Erdmann writes:

Jeff H. Says
"The first two questions ask things Social Security was never intended to do. It was always intended to part of a persons income in retirement, not the full amount."

FICA started out at 2%. And, slowly, slowly, it has crept up over the years so that SS now takes over 12% of wages. But it was never intended to provide your entire retirement....

Try to sell private annuities with these tactics while keeping yourself out of prison...

Floccina writes:

I go even further than that on Medical care. Why it is better to have health insurance than to pay for medical care after the fact? I think that we only need health insurance only because most people have health insurance. In fact if we paid after the fact for care our interests and the interests of he providers would in some ways be more aligned. The providers would need to consider that if the treatment kills us or fails to help us they are unlikely to be paid.

LD Bottorff writes:

Perhaps Bryan should have written a book called The Myth of the Rational Consumer. Voters are consumers in the marketplace of political power. People who are skeptical of the ability of consumers to make wise choices as they purchase goods and services should be more skeptical of the ability of voters to make wise choices as they choose a government. Personally, I agree with Bryan; people are less irrational when buying goods and services compared to when they are buying the lines of politicians.

Floccina writes:

@Peter H
I see no excuse for SS not paying out the same amount to everyone, maybe about $700.month. We should admit that SS is a welfare program. Also roll the whole payroll tax into the income tax or better yet a progressive consumption tax.

Tom West writes:

Phil nails it. Bryan thinks extended warranties are a ripoff because they provide the illusion of certainty.

But the illusion of certainty, especially for rarely occurring events, is often far more valuable than the actual service itself.

Yet another example where economists fail to understand the true value of a service by looking only at the measured value.

Personally, I don't buy into the extended warranty illusion, but I certainly have gotten decades of benefit from the Canadian healthcare system, including the illusion that other Canadians care enough about me that they're willing to pay my medical expenses even if I can't.

In fact, the illusions is strong enough that I consider myself fortunate that I'm in a position to pay the medical bills for several other Canadians rather than grousing about high taxes. Indeed, I've had decades of benefit from the illusion that government cares about me, and that benefit is not insubstantial.

Economists like Bryan are in danger of substantially reducing overall human happiness by undervaluing the illusions that keep the majority happier (if their voting is any indication).

Daublin writes:

It's a good metaphor.

As similar examples, I would toss out home-owner's insurance and comprehensive car insurance. In all of these cases, the insurance often falls short when you need it the most.

Your car insurance will send you to the most rinky-dink mechanic in town, and it won't cover damage that they say is due to negligence. Your home-owner's insurance will keep an ancient AC puttering along way after it should have been replaced by a newer model.

Jeff H writes:

Peter H,

Yes, I am well aware of SSI's requirements. And, contrary to your statement, you can actually receive both SSI and Social Security. SSI was actually intended to be a supplement to the Social Security program. For various reasons it has taken on a life of its own and is now dominated by the disability part of the program.

Additionally, the asset limit is really not that firm--there are various exclusions, including certain retirement accounts, a house, a car, personal effects, trust funds, etc. That said, I fully agree that there will be some behavioral effects like you mention.

Two points on your linked story. First, it is solely about the non-elderly, so rather tangential to an arguement abou the elderly. Second, the story is not exactly without flaws. If you poke around a bit more on the Internet you'll find various critiques.

Further, I think you missed the point of my comment, viz., Any means-tested Social Security program is going to look like the elderly SSI program. No judgement on that, just that I think that is the likely realistic alternative if Social Security went means-tested--It would probably be run by SSA using very similar policies as are currently in place for SSI. So the real question is whether that is preferred to what currently exists.

Kevin Erdmann,

I don't follow your point. How much of your income it takes up is irrelevant to whether is was "intended to provide your entire retirement". I'll grant that it sure makes it harder to sufficiently save on your own if Uncle Sam takes away 12% off the top, but that in no way means that's the program's purpose. Whether people agree with that purpose is another issue altogether and based on personal information and preferences. No judgment on that. But this is the system we have.

I think we can agree that intentions are pretty useless when it comes to judging a policy and its outcomes. But, if your frame of reference is that it will, and is intended to, provide for your full retirement, then you will behave differently than if your frame of reference is that it is intended to provide social insurance during retirment (i.e., be insurance against outliving your own savings). (Regardless of the tax rate.)

Unfortunately, I will not be able to participate in these particular chats further. Cheers!

david condon writes:

The timeframe for retirement is much larger, and people make very poor decisions over long timeframes. A large chunk of people simply don't save for retirement, and I find it hard to believe there's more than a small loss from the government managing these accounts. A lot of people would be much worse off without social security handling their retirement for them.

john hare writes:

As I have posted before about Social Security. If you had a thirty year mortgage on your house that used up a weekly paycheck per month, and if you could apply that 12.4% lost in Social Security deductions to the mortgage, then your house could have been paid off in 8-11 years depending on interest rate.

Twenty years of an extra paycheck and a half a month is substantial. Not to mention the quality of life from not worrying about a foreclosure. Should we all be held permanently financially responsible for the irresponsible?

Hazel Meade writes:

To be fair, reducing the level of uncertainty makes it easier for people to plan for the future.

I mean, seriously, we are pointing this out all the time with respect to regulation.

If a person knows that they are going to have a job in 5 years, they are much more likely to buy a house. If they know they aren't going to go broke if they get into a car accident, same thing. You can plan for the future if you know your economic situation is going to be fairly stable.

But then, that is what insurance is for, right?

I think the better argument is not that it's irrational to fear uncertainty. It's that instead of providing *insurance* against uncertainty the government often tries to regulate the uncertainty away (i.e. making it hard to fire people vs. unemployment insurance). Or in other cases, the government is administering the insurance in all sorts of unfair arbitrary ways (i.e. community rating in health care, mandatory coverages of politically expedient treatments, etc.)

david condon writes:

John Hare, that's an example of irresponsible planning. The interest on retirement is usually higher than the interest on a mortgage, so it's usually better to save for retirement. Foreclosure is only a risk if someone is irresponsible with their savings or happened to be unlucky enough to buy a house between 2004-2006 in the northeast or California and lost their job between 2008 and 2010.

Todd Kreider writes:

I've not read the entire post, and I've read no comments.

Of course extended warranties are B.S.

Bryan is supposedly telling us something we didn't know when we bought the Atari 2600 in 1982?

Comments for this entry have been closed
Return to top