Bryan Caplan  

Insurance versus Charity

PRINT
The Veterans Administration Mo... Punk Songs for Classical Liber...

One of the most effective ways to take the sting out of charity - for recipients and involuntary donors alike - is to give it a new name: "insurance." If you buy fire insurance and your house burns down, you don't feel like a bum when you cash the check. Why? Because you paid market price for a service, and are just getting what's coming to you.

If you want government to redistribute money to the unemployed, or the poor, or the old, then, it's foolish to call it what it is. Don't call it "welfare." Call it "social insurance."

Economists strangely seem even more prone to this kind of obscurantism than the rest of the world. Whenever economists see anyone giving money to anyone else without hope of repayment, they call it "insurance." Maybe it is; but isn't it possible that it is what it seems to be - something for nothing, or at least something for a lot less than it costs?

A case in point: Economists who study religion are very interested in its "social insurance" function. But hold on! Is this really a case where people are pooling risks to mutual benefit? Or is the real story that predictably rich church members are helping predictably poor church members?

One simple test: If you don't have to pay any premiums until you're ready to file a claim, it's probably charity rather than insurance. When I've probed economists who work in this area, they've generally admitted what I suspected: When new members show up at church and ask for help, their new friends usually give them a hand, without worrying too much about whether they'll ever pay them back. Who makes the books balance? Not the temporarily lucky, as the insurance story would predict, but the predictably rich.

One reason why economists confuse insurance and charity is that they think people are even more selfish than they really are. People are highly selfish, but many are nevertheless willing to give away 1% of their income to causes they believe it.

The other reason, I suspect, is that economists feel like they have enough bad press already. When politicians call welfare "social insurance," the obvious sarcastic question for economists to ask is:

If it's really "insurance," then why don't the rich pay less than the poor? After all, the rich are a lot less likely to file a claim.

But even though this is right on target, who but a Non-Bleeding Heart Libertarian economist would want to publicly raise the question?


Comments and Sharing





COMMENTS (18 to date)
Lord writes:

If you don't have to pay any premiums until you're ready to file a claim, it's probably charity rather than insurance.

One can make it insurance though by charging enough to cover the next reported claim, a multi-year policy.

How about the ancient Roman practice of collecting to put out the fire? The premium did fall the further the fire advanced though. If you weren't home, maybe they would take a chance you would pay up.

Vorn writes:

The idea that economists have an overly pessimistic view of human motive seems to be correct. While of course economists know intellectually that the assumption that individuals are profit-maximizing, atomistic, and rational beings is just that, an assumption, I think quite naturally it is hard to not let an assumption that one works with day in and day out influence one's thinking.

I think this is illustrated by a game used to teach negotiation. Briefly stated, its a variation on the prisoners dilemma. Two sides pretend they are oil exporters and the only source of oil. They cannot communicate with each other; the only signals they can send are the prices they charge. They can charge $30, $20, or $10 for a unit of oil. The more they jointly charge, the more they make in profits. However, if one side undercuts the other side, they will make an even bigger profit and the other side loses big. The joint gains will be smaller over all, but the defecter will make more individually. Since this is a repeat game, with multiple rounds, defecting leads to a loss of trust. If both sides defect, the joint returns are not very good, though better than if you try to cooperate but are defected on.

Of course, as a classic prisoners dilemma, the dominant strategy for any one round is to defect. However, with a loss of trust and lost profits in subsequent rounds, this may not be the case with multiple rounds. However, given that there is a fixed number of rounds, on the last round it makes sense to defect. Trust is no longer relevant to the game after that. But knowing that the other side realizes this and will defect on the last round, you realize that trust is not really valuable in the round before last either. In effect, the round before last is just like the last round, because you know they are going to defect in the last round no matter what. So you plan to defect one round earlier instead. Then you realize that they also realize this, so the logic is that you should defect even sooner, and so on, until, even with a game lasting multiple rounds, the dominant strategy is to defect in each and every round. All based on the assumption, of course, that the other side is atomistic, profit-maximizing, and rational.

Of course, both sides would be better off if they could find a way to cooperate; but it is irrational to cooperate with the other side if you make these assumptions about them. The only rational strategy when you are an atomistic, profit-maximizing, and rational and know the other side to be the same is to always defect.

Interestingly, when this negotiation game is played with economists, they typically start off the game playing a 10. Silicon Valley entrepreneurs, in contrast, will typically start off trying to cooperate from the beginning by playing a 30. (Obviously, these are generalizations with many exceptions.) Typically, in terms of maximizing profits in this game, the Silicon Valley entrepreneurs do better than the economists. They do not automatically assume that the other side has the traits listed above and take more risks concerning the characteristics of the other side in an effort to create mutually beneficialy cooperation.

So, I think that Byran Caplan is right on about economists. Many are more pessimistic about human nature than is justified by empirical observation. It is of course completely natural. Intellectually, we know that the assumptions we use in our models are just that, assumptions. They are there to make the models manageable and, theoretically at least, once we have the basic model down, we can relax one or more of those assumptions and consider the results. But in practice, many economists seem to forget at times that some of their simplifying assumptions are not accurate descriptions of reality. Especially when those assumptions are not obviously wrong for so many people.

Nonetheless, I think that this bias is important. Whenever you take policy advice from an economist (perhaps this applies more generally) you should really inquire into what assumptions (if any) they are making and whether you really think those assumptions are a good fit with reality, and if not, the extent that those deviations from reality are important with respect to the policy proposal in question.

Paul N writes:

Once again, I have some problems with the way Bryan's defined the problem.

To me "social insurance" refers, at least in part, to the fact that by giving poor people money, you're making it less likely that they and their children will commit crimes (esp., but not limited to, theft). So there really is an "insurance" value to it.

This may be economically justified when (1) people are loss-averse, i.e. it's a total bummer when someone steals your hubcaps even though they're not really THAT expensive and (2) a large portion of poor people steal, as stealing is a much less efficient form of wealth transfer than cash (like how much can you sell those hubcaps for, anyway?).

spencer writes:

Given that social security and unemployment compensation are the biggest "insurance" progams and are both financed by payroll taxes the rich do pay less then the poor.

Cann't you ever get any facts right.

Bill Stepp writes:

Social Security and other tax-and-spend schemes are not insurance, at least not the way an insurance analyst or actuary would define that term. The idea of insurance is for insureds to pool their risk by paying premia determined by applying the law of large numbers to specific perils, and to receive benefits in accordance with the contractual provisions of the plan. The plan itself is voluntary and based on a contract under the rule of law.

All of this is alien to Social Security and other welfare plans, which rely on involuntary payments (taxation) and pay out benefits that don't have the same kind of relationship to the recipients that insurance claimants have.
Most welfare plan payments go to recipients who never paid a dime in income tax, or at least never paid anying like the premia forked over by insurance buyers.

As an insurance actuary told me, if life insurers ran their businesses like the Social Security trustees run SS, they'd be in jail and their shareholders would lose their investments (for stock-based insurers). Their policyholders (for both mutual and stock-based insurers would get smaller payouts than they thought they would when they bought their insurance. The class action plaintiffs' bar would be circling like the buzzards they are.

R.J. Lehmann writes:
BRYAN: If you don't have to pay any premiums until you're ready to file a claim, it's probably charity rather than insurance.

"Probably" is an important qualifier there. Not all insurance plans rely on upfront premium. Some, particularly when you get into the complex field of structured reinsurance, rely entirely on post-event financing. And even the common insurance plans like auto or life will generally pay a claim as soon as a policy contract is in place, even if the first premium payment has not been delivered. Hell, when the World Trade Center was attacked, Larry Silverstein DIDN'T even have a contract in place yet -- just binders with a host of insurers laying out the intent to sign a contract -- which is why it was so difficult to determine what the terms were on the question of whether the attacks counted as one event, or two.

BRYAN: If it's really 'insurance,' then why don't the rich pay less than the poor? After all, the rich are a lot less likely to file a claim.

Differential premia are not an essential feature of insurance, and in most lines, they are a relatively recent development made possible by more sophisticated risk modeling that allows the insurer to stave off adverse selection. Historically, a standard mutual fire policy, like the ones Ben Franklin developed, were much more likely to charge everyone an equal amount.

But if the argument is that the rich are paying MORE than the poor even though their risk is lower, then certainly, I'd agree that's an unusual arrangement for an insurance pool, and one that is clearly incorporating elements of charity. But I don't know that the one necessarily has to preclude the other. If every participant is transferring risks they all face to a common pool, the plan is an insurance plan, even if the premium collected versus the benefit structure paid out appears to have charity, rather than profit, as a goal.

PAUL N: To me 'social insurance' refers, at least in part, to the fact that by giving poor people money, you're making it less likely that they and their children will commit crimes (esp., but not limited to, theft). So there really is an 'insurance' value to it.

That may count as a form of risk management, mitigating potential risks and ameliorating the chance of loss, but it would not count as insurance, which is the transfer of existing risk to a common pool.

BILL STEPP: The idea of insurance is for insureds to pool their risk by paying premia determined by applying the law of large numbers to specific perils, and to receive benefits in accordance with the contractual provisions of the plan. The plan itself is voluntary and based on a contract under the rule of law.

The law of large numbers only operates in a relatively few forms of insurance. They happen to the ones people are most familiar with -- auto, home, life -- but as you get into more complex commercial risks, the "numbers" become much smaller and the potential variance much greater. When you come to, say, terrorism risks, or the risks underwritten in the Lloyds of London market, they're almost all completely unique punts, with no prior "like" risks with which to compare them.

Private insurance is typically voluntary, but I wouldn't call that a defining feature of insurance. In some areas, the state compels insurance for those who choose to engage in certain activites (if you wish to drive, you must purchase third party liability; if you wish to practice medicine, you must have malpractice insurance.) The involuntary nature of those requirements doesn't make the underlying product something other than insurance.

BILL STEPP: All of this is alien to Social Security and other welfare plans, which rely on involuntary payments (taxation) and pay out benefits that don't have the same kind of relationship to the recipients that insurance claimants have. Most welfare plan payments go to recipients who never paid a dime in income tax, or at least never paid anying like the premia forked over by insurance buyers.

That's certainly not true in Social Security, which is an indexed annuity. The value of your benefit is specifically tied to how much you've paid in, so there is a direct relationship between what has been paid in and what is paid out. That it is a poorly capitalized, poorly reserved and bound-to-be insolvent annuity plan is a separate question altogether. It's not as if there's never been poorly capitalized, poorly reserved and bound-to-be insolvent insurers operating in the private market.

BILL STEPP: Most welfare plan payments go to recipients who never paid a dime in income tax, or at least never paid anying like the premia forked over by insurance buyers.

Many insurance benefits are paid out to claimants whose premiums were quite small in comparison to their claims. Many others contribute loads in premium over an extended period of time, and never make a claim. In and of itself, that doesn't have any bearing on whether something IS insurance, although it certainly has a bearing on how long the insurer can keep operating under that structure.

BILL STEPP: As an insurance actuary told me, if life insurers ran their businesses like the Social Security trustees run SS, they'd be in jail and their shareholders would lose their investments (for stock-based insurers). Their policyholders (for both mutual and stock-based insurers would get smaller payouts than they thought they would when they bought their insurance. The class action plaintiffs' bar would be circling like the buzzards they are.

All of this is absolutely true, but again, it's not as if private insurers have never operated in that way. I'd acquaint you with the history of companies like Legion Insurance, Mission Insurance Group, Integrity Insurance Co., and Transit Casualty -- all companies that vastly understated their liabilities, attempted to use investment returns to cover their underpriced premiums (which allowed them to capture huge market share) and ultimately, went belly up.

In any case, there's an important difference to mention. Private insurers must purchase reinsurance or otherwise tap the capital markets to cover the potential of insolvency. Social Security and other government-run plans need not due that because they have a backstop -- the American taxpayer -- who serves as ultimate guarantor of the benefits they promise. Now this moral hazard is not a desirable state of affairs, obviously, but if we're focusing solely on the semantics of the word "insurance," I think the government programs still qualify, even if they happen to be especially poorly structured insurance plans.

Bill Stepp writes:

I wouldn't call government-sponsored "insurance"
insurance any more than I would equate taxation with paying your electric bill. I pointed out that private insurance is voluntary to demarcate it from SS, which is a fraud.
Social Security discriminates against the poor and blacks (as Milton Friedman and others pointed out), who have lower life expectancies than wealthier people and whites.
Social Security is not an annuity because the latter is an insurance product with a death benefit. SS has no death benefit. There is a tenuous link between taxes paid in and benes paid out, but this by itself doesn't qualify it as an annuity. Something based on a crime (taxation) can't qualify as a contract.
Re: reinsurance, moral hazard and dodgy private insurers, yes there are bad insurance companies. We can add Conseco and Equity Funding to the list (the latter pumped up earnings with fake insurance policies--sold amazingly to other insurers; the former go into deep do do with shaky real estate and mobile home loans). Reinsurance and tapping capital markets doesn't cause a moral hazard problem, which can only come about by the government's subsidizing and underpricing risk. Even if an insurer underprices risk, which usually doesn't happen over a long enough stretch, its shareholders (who presumably are holding voluntarily) take a hit. Welcome to capitalism.

David J. Balan writes:

I think most people who call it insurance mean insurance from the standpoint of being behind a Rawlsian veil of ignorance, not from the standpoint of where people are after their endowment has been revealed. At least that's what I mean.

Bob Knaus writes:

Having grown up amongst the Old German Baptist Brethren (a sect often confused with the Amish) I can say a thing or two about religiously-based insurance versus charity.

I think my family would want to call it "charity" because that's a Biblical precept which insurance isn't. In fact many of the believers are dead set against insurance, unless the government makes it mandatory. Church members will support each other through pretty much any financial hardship. That said, there are some pretty steep barriers to entry. You have to wear odd clothes, and subscribe to odd beliefs. Oh yeah, and there is the prohibition against bankruptcy, even if it is involuntary it will get you kicked out of the fellowship.

As to Brian's observation of giving 1% of one's income to charity, that is off by an order of magnitude for the devout, many of whom are not at all wealthy. A long time ago, I bought a tractor from a fervently Pentacostal Cuban-American farmer. It was a Saturday, and as was the rural custom, I just offered him a personal check for $3500. He wanted to know if he could have some of it in cash so he could offer up a 10% "tithe" at his church service the next morning. Not 10% of the profit, 10% of the gross! Fortunately my dad had enough cash on him to let the guy make his offering.

No data, just anecdotes :-)

Jason writes:

What? People pay tithes on their net?

Dog of Justice writes:

Of course, as a classic prisoners dilemma, the dominant strategy for any one round is to defect. However, with a loss of trust and lost profits in subsequent rounds, this may not be the case with multiple rounds. However, given that there is a fixed number of rounds, on the last round it makes sense to defect. Trust is no longer relevant to the game after that. But knowing that the other side realizes this and will defect on the last round, you realize that trust is not really valuable in the round before last either. In effect, the round before last is just like the last round, because you know they are going to defect in the last round no matter what. So you plan to defect one round earlier instead. Then you realize that they also realize this, so the logic is that you should defect even sooner, and so on, until, even with a game lasting multiple rounds, the dominant strategy is to defect in each and every round. All based on the assumption, of course, that the other side is atomistic, profit-maximizing, and rational.

This standard analysis is brittle.

In practice, by deciding to always defect on the second-to-last round, you swing a significant percentage of opposing last round plays from cooperate to defect. Not everyone thinks like an economist -- and economists need to take that into account to properly analyze the problem!

The EV of defecting on the second-to-last round could very well be lower than that of cooperating. For the third-to-last round, this is almost certainly true. Thus, when actually playing this game with another unknown human, the optimal strategy is to cooperate on the first round (because of the effect this has on the other player's second round play), then tit-for-tat until a sufficiently large percentage of opponents can be expected to ignore your play -- which generally won't be the case until one of the last two rounds.

Vorn writes:

Dog of Justice,

You are absolutely right about how you SHOULD play given how humans really are. It should be kept in mind that tit-for-tat is probably the optimal strategy assuming a diversity of behaviors. It would not be optimal, however, if the other side consistently played in a manner following the assumptions of an economist. Such a player would defect in every round. If they are going to defect in every round, your optimal strategy is to defect in every round as well. Indeed, tit-for-tat practically does this; it does get screwed in the first round, but not afterwards. The economists are RIGHT if you make the assumptions they make.

The interesting thing, is that economists are much more likely to start this game with a 10; showing that they think like, well, economists. That is simply an empirical fact based on having many different groups play this game in different settings.

Dog of Justice writes:

The interesting thing, is that economists are much more likely to start this game with a 10; showing that they think like, well, economists.

Starting the game with a 10 is not thinking like an economist (or at least, not a good one). It is thinking like a pure mathematician. (Though unconditionally ending with a 10 is sound economic thinking.)

As you pointed out, any user of mathematical models needs to judge what model is appropriate for the given situation. In this situation, a pragmatic EV estimation informed by knowledge of likely opponent behavior is clearly more appropriate than a dominant strategy computation.

Vorn writes:

Dog of Justice,

Well, maybe this turns on what you think a "good" economist is. All I know is that, empirically speaking, economists are more likely to play a 10 in the first round. That doesn't mean ALL economists do this; maybe only the ones that aren't "good." I will say that the economists who play this way are not unintelligent.

I am sure that the vast majority of economists would adjust their playing after playing this game repeatedly. That is, they would adjust their assumptions about other players to more closely fit reality. But what is interesting is that economists not experienced in playing this game against humans start playing a 10. I think that shows something about the strength of their initial assumptions about others.

What is interesting about this is not what a "good" economist would do, but rather what many "actual" economists do.

Steven writes:
Given that social security and unemployment compensation are the biggest "insurance" progams and are both financed by payroll taxes the rich do pay less then the poor.

Do the rich pay less? My understanding is that as income increases, payment for payroll taxes increases, up to a maximum. Surely you aren't claimng that ordinary insurance premiums are based on percentage of salary...

liberty writes:

>> Given that social security and unemployment compensation are the biggest "insurance" progams and are both financed by payroll taxes the rich do pay less then the poor.

>Do the rich pay less? My understanding is that as income increases, payment for payroll taxes increases, up to a maximum. Surely you aren't claimng that ordinary insurance premiums are based on percentage of salary...

The way I understand it, the rich pay a percentage and then in the unlikely circumsance that they collect, there is a maximum on what they can recieve that is much less than they paid. That is the main way that these programs are funded - unlike regular insurance where everyone pays in, those with more risk pay more, and what you receive back is based on injury. The primary reason that people are in favor of these welfare programs is that the poor are subsidized by the rich.

Those who do best with unemployment insurance are those who earn just over the maximum paid out (so that 75% of earnings equals the maximum) and work just barely the minimum number of months required to recive the maximum number of months that it pays out.

I used to know many young adults in the US and UK that took advantage of this; they would work for 6 months, get dole for six months, in the UK while here they would work for 12 months I think then get unemployment for 12 months, something like that.

rvman writes:

>Maybe it is; but isn't it possible that it is what
>it seems to be - something for nothing, or at least
>something for a lot less than it costs?

Boy, is this materialist. The giver isn't buying insurance, and isn't getting nothing for something. He is purchasing the esteem of his peer group. Carnegie wasn't going to benefit by all of those libraries, he was buying a good reputation, and the kind of immortality which comes with that. The church giver may be buying the esteem of his fellow churchgoers, or of God himself. Charity is a consumption good.

rvman writes:

Or more precisely, the price of a consumption good.

Comments for this entry have been closed
Return to top