« Look away when you are thinking | Main | Why was British food so bad for so long? »
Why they might drum me out of the Economists' Corp
Some of you are still asking me what I think of Steve Landsburg's argument that we should pull the plug on people who did not buy (would not have bought?) fairly priced ventilator insurance.
If you like invective, here is what Kevin Drum thinks. Daniel Davies distinguishes between willingness to pay and willingness to be paid; fair enough. You might also wonder how much long-term insurance is available in the first place.
But my objection is more fundamental. I do not accept that ex ante choices provide definitive information about ex post values. Taking another context, assume you think the chance of cancer is only five percent when in fact for you it is (for genetic reasons) fifty percent. You fail to buy insurance, but does that mean you don't value your life much? In this case I am willing to downgrade the economic values we might infer from market behavior; it was based on faulty information.
Now take the underinsured, now-dead woman (Tirhas Habtegiris). She thought, either implicitly or explicitly, that her chance of needing "plug-pulling insurance" was small. Maybe this seemed reasonable at the time. But in reality she needed that insurance with p = 1. She was wrong.
Does the cancer case differ from the ventilator case? Perhaps they are not identical in their moral implications (the probability mistakes may differ in their ex ante reasonableness), but in both cases, if we are going to use economic methods, I prefer to use the valuations based on better rather than worse information. And based on correct information, the now-dead woman had a high value for that plug not being pulled. Look at the ex post valuation, not the ex ante valuation. After all, we know she is (was) on the ventilator, why go back to an imaginary state of affairs where this information is missing?
For selfish people, the value of life will be, in terms of willingness to pay, their entire wealth. In terms of willingness to be paid, the value of life will be infinite or undefined. When the "perfect information market valuations" turn out so screwy and disparate, that is a sign the whole method has broken down. Note also the neat trick that altruistic people, who care about bequests, will appear to have lower values of life; that is fishy too.
Mark Kleiman makes some related (but non-identical) points, and here are his lecture notes on the topic.
I don't buy Rawls or Harsanyi, both of whom begged the question on the moral force of ex ante agreement. I also don't think we should value human lives by looking at the value of risk reduction, although this may provide broadly relevant information. You've just got to decide how much a particular life is worth. If you are thinking "yikes," you are right, but, forgive the pun, that is life.
Are we spending too much money keeping people alive in their final throes? From a rules perspective, yes. On a case-by-case basis, maybe not. I don't know how to reconcile "rules" and "act" perspectives, even in simpler contexts.
In any case our decision comes down to an ethical judgment, where the money and the life meet each other, mano a mano, on an open plain. When it comes to this lady, a terminal cancer patient who was alert at the time, the marginal costs of keeping her going just didn't seem that high.
Addendum: If you are interested in pursuing these issues, try John Broome's "Trying to Value a Life," Journal of Public Economics 1978, and the discussion in the 1982 M. Jones-Lee book, The Value of Life and Safety. Or read the comments below.
Posted by Tyler Cowen on January 13, 2006 at 06:10 AM in Economics | Permalink
TrackBack
TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341c66b253ef00e550979c388834
Listed below are links to weblogs that reference Why they might drum me out of the Economists' Corp:
Comments
If you don't hold people responsible the decisions they make using poor information, that will seriously reduce the incentive to acquire better information for ex ante decisions.
Posted by: EclectEcon at Jan 13, 2006 7:37:51 AM
The problem with EclectEcon's incentive argument is that information can affect both sides of the insurance decision -- if a person gets a genetic analysis that they will get Huntington's disease with probability 1, can that person still get insurance for it (without committing fraud)?
More information might decrease the probability of buying insurance. Presumably, if both patient and insurer have perfect information, it would be cheaper for the patient to save the exact amount of money needed than to go through an insurance company as a middleman.
Posted by: DK at Jan 13, 2006 8:24:43 AM
What I find ironic about this case is that it is an exception to the general rule in the U.S. which is that even those terminally ill and near death are not denied expensive end-of-life care for funding reasons. And this case is going to make exceptions that much less likely in the future because the publicity for Baylor medical center has been terrible.
On the other hand, my understanding is that government-run systems are more likely to fail to provide care for reasons of cost. For example, I believe that it's still the case (or was until recently) that many in the UK have very poor access kidney dialysis treatments under the NHS.
This ventilator case is being interpreted as evidence of the harshness of the U.S. system and yet there are good reasons to believe that a government run system would unappolgetically "just decide how much a particular life is worth" and establish rules where denial of expensive, end-of-life care to late-stage terminal patients would be routinely denied. Or, as with the NHS, the treatment might not be forbidden but unavailable to many due to lack of resources, so late-stage terminal patients might be denied access to ventilators because there weren't enough to go around and their condition made them last in line.
There is a broad consensus that the U.S. 'wastes' too much money on expensive end of life care, but ending that 'waste' means things like not just stopping ventilator treatment for terminal patients after a month but not providing a ventilator *at all* in such cases.
Posted by: Slocum at Jan 13, 2006 8:25:37 AM
I think it helps to think of the federal government (at least in this case) as the provider of ventilator insurance insurance; the government insures us against the myriad of small risks we miscalculate.
Posted by: Michael Stack at Jan 13, 2006 9:15:28 AM
If the woman's family was unwilling to pay to keep her alive, why
should complete strangers be forced to pay to keep her alive. Obviously,
there are numerous people who believed this decision was wrong and the
woman should have been kept alive on a ventilator. But of these people,
how many put their money where there mouth is and actually sent in money
to keep her on a ventilator? How many have ever sent money to keep someone on
a ventilator? People are very ethical as long as it comes at someone
elses expense or out of someone elses wallet!
Posted by: Tom at Jan 13, 2006 9:45:11 AM
Man can't live by expected value alone, even if he has perfect information. That's not how we make decisions, nor should it be. Think of the St. Petersburg Paradox or Alex's sly reply to Pascal's Wager--heaven has infinite value, so give him all your money and he'll put in a good word with God (with a small but nonzero probability of helping you reach eternal bliss).
Posted by: gundryggia at Jan 13, 2006 10:32:25 AM
Qudryggia, if there is perfect information, then expected value equals value, there are neither wages nor insurance, and you know both the existence of God and whether or not Alex has any pull with God.
I never liked Alex's reply anyway, since Christianity has an explicit prohibition in the Bible on buying or selling salvation, and I presume other religions with a notion of heaven would have something similar. Of course, one could interpret that to ban Pascal's Wager, too.
Posted by: DK at Jan 13, 2006 11:13:48 AM
"if there is perfect information, then expected value equals value..."
Ah. So you'd enter the St. Petersburg paradox game at its expected value?
(By "perfect information" I meant knowledge of probabilities.)
Posted by: gundryggia at Jan 13, 2006 11:50:36 AM
There's a away around the St. Petersburg Paradox by imposing a diminishing marginal utility of wealth and looking at the expected utility rather than the expected value. I don't remember the specifics, as it was an example in microecon a few years ago and my memory for such things is hazy.
Alex's reply to Pascal's Wager is much more difficult, but DK is right that with perfect information you know whether God exists or not and, if God exits, whether Alex actually has any pull, as well as how much Alex's pull will influence the decision regarding damnation, and whether or not that will be enough to achieve the result you want.
I never liked Alex's reply anyway, since Christianity has an explicit prohibition in the Bible on buying or selling salvation.
That didn't stop the Church from doing it, but I would like to know where I can buy an indulgence these days, might come in handy.
Posted by: Timothy at Jan 13, 2006 11:51:26 AM
"There's a away around the St. Petersburg Paradox by imposing a diminishing marginal utility of wealth and looking at the expected utility rather than the expected value."
You're probably thinking about imposing a log utility function. But then change the prizes to grow exponentially. This is not to imply there is no reasonable way to look at the problem (there are several), just that expected value computations are not the very pinnacle of rationality. http://plato.stanford.edu/entries/paradox-stpetersburg/ is a pretty good discussion.
Posted by: gundryggia at Jan 13, 2006 12:08:18 PM
Tyler,
Ex ante, you were clearly going to be drummed out of the Economist's Corps.
However, ex post, I think you will be kept in...
Posted by: Barkley Rosser at Jan 13, 2006 12:19:12 PM
"For selfish people, the value of life will be, in terms of willingness to pay, their entire wealth."
The real problem is that for selfish people, the value of their own life will be, in terms of willingness to pay, the entire wealth of anybody's money they can get access to.
The problem with end-of-life costs is that on a general basis anyone can see that we waste vast amounts on mostly pointless procedures at the end of life. But when it is your their own grandmother hanging on, a huge number of people take a different view.
The reason we tend to want to make the rules beforehand is because we know we will be swamped with the emotion of the decision later if we don't.
This may or may not be a wise allocation, but I think it accurately describes why we try to do it.
Posted by: Sebastian Holsclaw at Jan 13, 2006 12:30:43 PM
None of the comments seem to want to bring her age into this discussion.
She was 27 and this has two major implications that I can think of off the top of my head.
Because she was young she is quite likely to underestimate the risk of needing
the venilator. Because the risks of needing health care are age dependent -- the old are more likely to need it while the young are most likely to not have insurance. But doesn't this get to the heart of the issue of how to "pool" the risk? But the core of the health insurance issue is how to pool risk and the existing system has some major problems dealing with this.
On the other hand, because she was young the benefit from providing her the care would be much greater then if she were 87. Yet this example shows that our existing system also does a poor job of dealing with this issue -- not that I am taking the position that other countries system of denying some care to the elderly is the best solution.
But don't these complications make the issues very different then what most of the discussion focused on?
Posted by: spencer at Jan 13, 2006 12:50:05 PM
There's a (very) rough analog to all of this that might strike home for the professional class. Why not buy long-dated S&P put options with a strike at, say, 750 to protect your 401k against catastrophic loss?
What? The market crashed 1929-style and suddenly you don't have anything for retirement? Well, you had your chance to buy those put options, right?
Maybe I'm a moron, but I'm not sitting up all night hedging away all the million tiny risks to my well-being, even where I can buy real (as opposed to the hypothetical ventilator example) insurance.
(Whoops Re St Pete Paradox digression above and log utility functions: I should have written "change the prizes [so that the expected payouts] grow exponentially" above.)
Posted by: gundryggia at Jan 13, 2006 12:53:02 PM
Re: Tyler's point that altruistic people value life less, I think that altruistic people value other people's lives more, not that they value their own lives less.
Qundryggia, I like your point about S&P puts a lot. But, if I care about the "million tiny risks", I don't want to assume the exchange guaranteeing the puts will survive a 1929-style crash.
Timothy, technically indulgences didn't buy salvation, only a decrease of your time in purgatory, which we can think of as a period of finite but negative utility preceding the infinite utility of heaven. I wonder if the idea of purgatory itself came about as a medieval Catholic solution to the difficulty of analyzing infinite utilities -- you can't credibly threaten infinite punishment for minor infractions.
Which brings us back to the point -- death is a pretty harsh punishment for not insuring yourself against such a small risk, which may not be insurable anyway.
Posted by: DK at Jan 13, 2006 2:18:31 PM
gundryggia: Thanks for the link. I can agree that there are definitely limits to EVT calculations. However, because I'm virtually heartless, I'm okay with letting people suffer the consequences of their misjudgements.
Posted by: Timothy at Jan 13, 2006 2:44:34 PM
I liked Daniel Davies' response to Landsburg at Crooked Timber:
http://crookedtimber.org/2006/01/11/a-simple-coasian-test-for-some-kinds-of-economic-bollocks/
Posted by: Tyrone Slothrop at Jan 13, 2006 4:09:04 PM
Apparently Kevin Drum was engaging in some self-hatred...
"Americans spend wildly more than the French do in the final six months of life, and this spending is almost entirely useless. A more rational approach to end-of-life spending would probably cut down our healthcare bill a lot." - Kevin Drum, November 11th, 2005.
Posted by: Keith at Jan 13, 2006 5:27:47 PM
Re: "There's a away around the St. Petersburg Paradox by imposing a diminishing marginal utility of wealth and looking at the expected utility rather than the expected value."
To really get around it you need a bounded utility function. Log (an unbounded function that goes to infnity the most "slowly") can't do it for reason qundryggia mentions.
Posted by: radek at Jan 13, 2006 11:58:05 PM
I found a number of problems with the Landsburg essay:
1.He argues that the question of how much assistance to give the poor and what assistance to give them are completely separable questions. Thus, he assumes government first decides how much to spend on the poor, say 5% of GDP, then decides how that amount will be distributed. Alas, the system does not function that way. A cut in health care will not be met by an equal increase in cash payments. In reality, advocates for social spending maximize support by focusing on particular issues that connect emotionally with people.
2.He assumes that one poor person is equal to another. Giving $75 to a poor, healthy 21 year olds is equal to giving the same amount of care to people in need of emergency health care who would otherwise die. Most people do not see it this way. For obvious reasons, one person is clearly in more need than the other.
3.He assumes that people simply wish to transfer wealth to the less fortunate, and that two things of equal value to the recipient are perfect substitutes. However, this is not the case. The person (or government agency) doing the giving may have different interests than the potential recipients. For example, some people don't give cash to a person who begs because they worry the money will be spent on something they do not approve of, such as alcohol; instead, they give money to an organization that provides food and shelter for the needy (while some people give the money anyway, knowing they would not give to a charity later (this is connected to #1 above, most people do not first decide how much to give then divide it up, they give as they come upon the need)).
And I am much more sympathetic to his argument than most people. For example, people are much more willing to pay to help others when the risk of harm is close in time and the person harmed is clearly identified. People will pay to sew up a gun shot wound, but not for a vaccination that will cost less and save more lives. I see refusing to fund a vaccine that will save many lives the same as refusing to treat a person in the emergency room with a gun shot wound. Refusing to assist a person in direct peril is severely condemned, while ignoring the predictable causes of death that afflict many is simply a bland negligence we all engage in.
Posted by: c&d at Jan 14, 2006 1:38:54 AM
Fundamental to the liberal (in the true sense) economist's viewpoint is that the individual is best informed to make decisions regarding themselves.
Posted by: pk at Jan 14, 2006 2:26:15 PM
People who cannot pay should recieve treatmentuntil some other unfortunate person who does have insurance or cash needs help.
Posted by: aaron at Jan 15, 2006 10:37:33 PM
Would anyone be able to give me some background or point me in the direction to find backcround information on the Dec. 2005 China export quota bidding process? I believe the situation is that 70% of the quotas are held by existing exporters. The remaing 30% go to a bid process.
Posted by: Patrick at Jan 19, 2006 11:29:05 PM
amateurs atm vredig ^^^ warm vrouw ^^^ adatto allievo ^^^ fighette amore in cucina ^^^ infirmiere prostituer dans antichambre ^^^ chaude surnaturel quarante deux ^^^ papa trio ^^^ frais maman vierge ^^^ nonsensical fighette sesso ^^^ carino fighetta fottilo ^^^ mystisk asiatisk gruppen ^^^ fanig tjejer dubbel fitta samlag ^^^ taglio cowgirl gruppo ^^^ gradevole giovane prostituta ^^^ fremragende mor ^^^ mye ansikt ^^^ varmest kysk ^^^ sjenert ekte ^^^ sarkikos pothos magoulo agapi ^^^ asiatis katourima stin kouzina ^^^ le plus chaud mere roue ^^^ plus frais papa photographie ^^^
Posted by: levan at Sep 6, 2006 2:19:27 AM
Hello all really cool blog
alprazolam fioricet hydrocodone vicodin tramadol xanax valium ultram soma carisoprodol ambien ativan lorazepam propecia adipex didrex cialis levitra paxil meridia viagra wellbutrin clonazepam xenical prozac butalbital phentermine
buy ativan buy adipex buy didrex buy levitra buy cialis buy phentermine buy soma buy tramadol buy diazepam buy carisoprodol buy meridia buy paxil buy valium buy xanax buy ultram buy fioricet tooth whitening online pharmacy alprazolam car insurance payday loan web directory business directory carisoprodol hydrocodone buy vicodin
Posted by: Erika at Oct 10, 2006 12:18:30 PM
