Random vs. Certain Death

Here is a moral puzzle in the spirit of Eliezer’s Torture vs. Dust Specks post.  It is directed to free market admirers.

A company has 100 employees.  It has the opportunity to make $1 billion but only if a task gets completed.  There are two ways of completing the task.

(1)  A specific employee, named John, must die.
(2)  Three of the 100 employees will be randomly chosen and killed.

The company can’t force its employees to take any actions, but it can bribe them.  John will not accept any amount of money to give up his life with certainty.  But all 100 employees would gladly risk a 3% chance of death in return for $5 million.  Consequently, the company intends to pay each employee $5 million and complete the task using option (2).

Now imagine that you are a government regulator who has the power to change what will happen.  You can:

(A) Forbid the company from completing the task.
(B)  Not interfere.
(C)  Force the company, and John, to complete the task by using method (1).  You could then force the company to give $5 million to each employee.

Assume that all 100 employees are exactly alike except that only John can complete the task by himself.  What should you do?  Does it matter if John was randomlly chosen right before the game started?

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • http://profile.typekey.com/robinhanson/ Robin Hanson

    Option B is one that John likes ex ante, but if it turns out that he is chosen he will not like it ex post. Similarly, if the public is understood as having made an ex ante agreement to give the government the power to make certain choices according to certain principles, then the government could reasonably follow through with that ex post, even if some people are then made worse off. I can certainly imagine that it could have been reasonable for John to ex ante support such a government policy, which then ex post made him worse off.

  • http://www.writerandeditor.co.uk Richard

    The problem with this puzzle is that the outcome of the job is significant. If John must die because his liver contains a unique enzyme, only extractable after death, that may lead to a patentable (i.e. $1bn-value) drug to save millions of lives, it’s a bit different to John being executed for the pleasure of Warren Buffett who just happens to be a snuff fan willing to pay $1bn for the experience.

    In fact, option two is pretty commonly observed in the financial services industry, where stress-related mortality is acknowledged and the rewards are of the magnitude you suggest. The employees voluntarily take on the stress, knowing it’s harmful to their health, and the government does nothing to intervene.

    This kind of question makes much more sense in a military context, of course, where one explosives expert, for example, may be required to conduct a suicide mission as an alternative to a massed attack with a 3% (or probably higher) fatality rate for the broad spread of troops. But in that case, of course, it’s the government taking the decisions; and there’s no $5m on the table for the troops.

  • http://www.jmrozendaal.com John Mark Rozendaal

    The answer to this one seems very obvious to me – as a government regulator, you forbid the company from completeing the task. (The problem here is that we don’t know what the “task” is; it might have some value, some end that would justify such immoral means in some people’s minds.)

    “Free markets?” Who is actually in favor of “free markets?’ I have never heard any of those who claim to be in favor of free markets say that they would like to do business without the support and regulation that government provides for banking, currency, civil law, and civil courts to enforce contracts. In fact the wealthiest markets that we have could not exist without a great deal of government support and oversight. The people who say that they favor free markets actually favor market regulations that advantage them and not others.

  • fraac

    If we can make $1 billion for killing 3 out of 100, how much for the lot?

  • Constant

    “Free markets?” Who is actually in favor of “free markets?’

    Libertarians, for example minarchists but particularly anarchists (anarcho-capitalists).

    I have never heard any of those who claim to be in favor of free markets say that they would like to do business without the support and regulation that government provides for banking, currency, civil law, and civil courts to enforce contracts.

    There is a substantial libertarian literature critiquing all of these, especially the government control of banking and currency, but also (among the anarchists) government monopoly over the courts. The history of private law comes up frequently, in particular the law merchant. Bryan Caplan has some stuff up on private law.

    In fact the wealthiest markets that we have could not exist without a great deal of government support and oversight. The people who say that they favor free markets actually favor market regulations that advantage them and not others.

    So that’s what it comes down to, a smear on the character of libertarians, based on the writer’s own abysmal ignorance.

    Why are you attacking market admirers? I doubt Robin Hanson wants his blog to be used for an argument like this. But the blog entry invited your sort of comment. So maybe the blog entry is where the mistake started. For my part, I don’t see why I should be expected to keep mum in the face of such an attack.

  • Constant

    James,

    Your hypothetical employs a key assumption which cannot be made in real life when the government coerces its subjects. In your hypothetical, ex hypothesi every worker would willingly risk his life by 3% in exchange for $5 million. Given this, then the government solution (C) is arguably equal to or better than the voluntary solution in every respect, since (properly implemented) each employee still gets $5 million and now has (in some sense) only a 1% chance as opposed to a 3% chance. Given your assumptions, the government-coerced solution is superior to the voluntary solution. (That is not exactly true because John is known beforehand, but I am bending over backward to affirm the superiority of the government-coerced solution in this hypothetical, for the sake of argument.)

    But this is not at all easy to apply to real life, because it makes an assumption about the workers’ preferences. In real life the required omniscience does not exist. People’s preferences are revealed in their action, and their action reveals their preferences only if the action is uncoerced (coerced action does reveal certain preferences, but different ones from uncoerced action). So there’s a real difficulty in knowing what individuals would have freely chosen while at the same time coercing them to act a certain way.

    You can of course design your hypothetical so that, to begin with, people freely choose, but then before they are able to act on their choice the government swoops in and coerces everyone to dance to its tune, which in turn is carefully constructed to respect everyone’s revealed preference. But in real life this trick can be played on people only a few times before they wise up to it and expect the government to swoop in and deprive them of liberty. Knowledge that the government will swoop in will inevitably affect their choices. Their choices become a conscious pantomime, rendering it difficult or impossible to infer their preferences from their choices. The realistic result of this approach, in which the government attempts to anticipate the workers’ wants and needs and then coerces the workers to act accordingly, is the Soviet worker’s paradise. The Soviet Union was, as we will recall, a totalitarian dictatorship. Both deep and probably intractable problems of knowledge and deep and probably intractable problems of motivation lie in the way of the dreamers who imagine a wise and benevolent government which flawlessly anticipates the wants and needs of the proletariat.

    I am commenting on the relationship between your hypothetical scenario and the real world because you did, after all, say that you were directing your question at market admirers. You yourself established that connection, which I am commenting on.

  • Roger

    The correct answer is (A). Profit is not an end capable of justifying the mean of murder. You see this sort of calculus in extremist religious movements. The business becomes a church or mosque, the employees are adherents to the cause, the profit is the disruption, confusion and shame caused and John is a prospective suicide bomber.

    The extremists even one-upped you on this example – it is common practice for suicide bombers’ families to be reimbursed for their child’s sacrifice, making it even more ‘fair’. John, in the original example, isn’t stated as gaining anything from giving his life for the company.

    You see an interesting interplay of social dynamics here. I suspect that since the cost to the employer is much higher for a random lottery in the case of the extremists (and certainly much higher than 3%), you’ll find a much harder social push for the Johns of the community to ‘step up’ for ‘the greater good’. They’ll be motivated by more than monetary remuneration, they’ll become martyrs, receive the adulation of their close-knit community (well, for a time…), and are ‘guaranteed’ rewards in the afterlife.

    In the end, however, the churc^H^H employer gets what it wants. Pretty sad if you ask me.

  • Constant

    The correct answer is (A). Profit is not an end capable of justifying the mean of murder.

    However, different jobs have different risks of death. The risks are part of the job. Mining carries a certain risk, movie stunts carry a certain risk, policing carries a certain risk. We try to minimize the risk, but accepting the risk is part of performing the job. It is always possible to reduce the risk by (say, in the case of police) never patrolling and never answering 911 calls. The risk of death is balanced against the need to do the job. And when we’re talking large numbers, the risk translates into the near certainty that some approximate number of police (etc.) will die in a given period of time. Those lives are in effect traded away.

  • Silas

    James_Miller: I think your question would be more interesting if:

    1) Companies didn’t already do exactly this (with different numbers) in terms of paying employees to accept higher risks of death. (cf. compensating differential; salaries of smokestack painters vs. ground floor painters)

    2) The government didn’t already ban some instances of this while permitting others. (Paying employees to accept risk of falling off ladder: okay. Paying employees to accept risk of asbestos: not okay.)

    For my “market fanatic” part, the company should be allowed to do this and the government shouldn’t intervene. It doesn’t matter if John was randomly selected from the beginning, for the reasons Robin_Hanson gave.

    Btw, I’m not sure even utilitarian logic supports C, except for the bizarre tunnel-visioned utilitarianism that no one supports but everyone pretends people support. You have to consider what it’s like for people to live in a world in which the government can arbitrarily order executions, in substitution of voluntarily-made arrangements, because of its estimates of what improves social utility. (Compare to the moral dilemma about whether doctors should randomly kill some patients to harvest organs.)

  • Caledonian

    The correct answer is of course B.

    We consider acceptable military operations with have a certain (or high certainty) of death for a few to save many, but people joining the military know what they’re getting into. If John knew when being hired that there was a possibility he might be the employee who needed to die – that there was a 1% chance that his accepting the position would be lethal to him – and he did it anyway, then the first option becomes acceptable.

    If it’s suddenly discovered that one employee needs to die, and it is known who it is, that person is being sacrified against their will for the good of the others, and no one wants to be in a society where they could be so sacrified.

    If everyone is willing to accept the risk of death, even if three people will die instead of one, that’s a society we’d prefer – one that permitted us which risks to assume and which risks to avoid.

    Besides, we let people risk their lives for far less than $5 million. Why should we prevent the free choice just because it’s made as a group?

  • Nick Tarleton

    But this is not at all easy to apply to real life, because it makes an assumption about the workers’ preferences. In real life the required omniscience does not exist.

    You have to consider what it’s like for people to live in a world in which the government can arbitrarily order executions, in substitution of voluntarily-made arrangements, because of its estimates of what improves social utility.

    This is why I think simplistic moral dilemmas of this type are not terribly useful. The problem implicitly asks you to assume perfect knowledge (of probabilities if not outcomes), and the possibility of choice (or coercion) without any externalities (even creating bad precedent). These assumptions are (almost) never realistic, and once they’re disposed of I think that many of our “irrational” moral intuitions end up being quite justifiable as checks on the abuse or mistaken use of power.

    Still, I don’t know what I would do in this case.

  • Stuart Armstrong

    In a vacuum, I’d go for (B). In the real world, I’d go for (A).

    Why? because one of my goals is to increase the real value of human life. Allowing such a procedure to go forwards would acheive the reverse. It pains me to admit it, but the delusion that life is infinitely precious may well be a positive bias – or at least an aspirational hypocrisy.

    The Soviet Union was, as we will recall, a totalitarian dictatorship.

    Not everyone who favours regulation is advocating ressurecting the USSR, just as not every libertarian is in admiration of Somalia.

    Constant’s substantive point is correct, in general. But it does fail on occasion. Sometimes, you can be pretty sure of people’s preferences. People prefer to eat than to starve; they prefer to be acknowledged than insulted. A government which operates along those principles (with exceptions for those who clearly demonstrate that they are in a different case) would not suffer much from this unknown preference issue.

  • Brandon Reinhart

    (B) Not interfere.

    The company’s employees and their family members are the ones who should determine if the risk is worth it. There are lots of ways to interpret this risk: maybe they are going to launch their own asteroid mining program or contract on a military operation or maybe something less epic but no less risky. Unless they are endangering those who don’t willfully participate, I don’t see how government interference is justified.

  • Silas

    Stuart_Armstrong: You’re saying that “not permitting anyone, ever, to pay or be paid for absorbing an increased risk of death” is “what you would do” and is “consistent with a respect for human life”, the latter being “infinitely precious”?

    In that case, the next steps would be to explain:

    -Why you paid or were paid for absorbing higher risk of death several times during the past week.
    -How you would get law enforcement to take on the increased risk of death in enforcing regulations such as these.

  • Constant

    Not everyone who favours regulation is advocating ressurecting the USSR, just as not every libertarian is in admiration of Somalia.

    But my point is not that people advocate resurrecting the USSR, but that the way it turned out is a real, recorded unintended consequence of taking a certain approach to the problems of society. Let us not ignore history for no better reason than that people today mean well. On the other side, Somalia is a legitimate thing to worry about. An anarcho-capitalist needs to grapple with the possibility that the fall of the state might lead to a Somalia, whether or not he wants it to. There are, of course, responses to this, but it is a legitimate concern.

    People prefer to eat than to starve; they prefer to be acknowledged than insulted. A government which operates along those principles (with exceptions for those who clearly demonstrate that they are in a different case) would not suffer much from this unknown preference issue.

    But what is the government competent to do, given this knowledge? Produce food? Recall the Soviet and Chinese famines. Oops, there I go again with the historical examples.

  • Dagon

    There may be causes that John will die for, and a $14M donation to one of them may be sufficient motivation for him to take the job. But fundamentally, specifics matter a lot – moral reasoning (for me, at least) gives weight to a very wide range of factors (including what the company will do with that $1b, what other effects the transaction will have, etc.).

    As a government regulator, it also depends on specifics. Specifically, what do my bosses (and their constituents) want in this case? There are cases where high risk and a large pension are socially acceptable, and cases where it’s frowned upon, and you don’t give enough data to know which it is.

    For a soldier, I go with B. $5m is more than most 3% risk takers have historically gotten. In many cases, I go with D – force the company to spend some amount of money on third party counseling about risk and reward, then let people make their choices.

    I don’t think the government should ever go with option A, but I cannot examine all possible scenarios, so I may just not have thought of one that fits.

  • Darin

    Constant is right on the money. This is the whole problem with regulation — not that its worse in theory (with an omnipotent bureaucrat-god), but that it doesn’t work in practice (with real bureaucrats).

    If the regulators really had the perfect information they are imagined to have, regulation would probably be a great thing. Reality is different.

  • Cassio

    It is quite obvious for me that, morally, option A is more adequate: the governor regulator should forbid the company from completing the task. No one gets 1 billion, no one gets killed.
    Evidently, this is not what governors/presidents/neocons think, vide Iraq & Oil War.

  • douglas

    Constant– It’s OK with me if you take any and every opportunity to defend libertarians.

  • Constant

    I got it off my chest. I’m probably good for a few days.

  • Roger


    -Why you paid or were paid for absorbing higher risk of death several times during the past week.
    -How you would get law enforcement to take on the increased risk of death in enforcing regulations such as these.

    The case as given is a guarantee of death, not a risk of it. Either 1 person or 3% are guaranteed to die. Don’t conflate statistics with risk. Just because an actuarial table will tell you that 5% of a police workforce die annually, this is not a guarantee that 5% WILL die this year, or next or the one after, that’s the Gambler’s Fallacy. The absolutism involved here and lack of important extenuating circumstances (The company is a company of firemen, and the ‘profit’ is saving lives) makes the company’s decision murder. As such, the correct answer is (A) to stop it completely.

    The lack of important extenuating circumstances is the clincher here. Motive isn’t everything, but it is enough to turn murder to sacrifice. Without information on motive, are you comfortable assuming that the company is working for some public good?

  • Andrew2

    John will not accept any amount of money to give up his life with certainty.

    John is wrong.

    Now, I suspect you meant to say John is not willing to accept any amount of money less than the $1 billion the company has to pay.

    In that case, and assuming that I have perfect knowledge and there are no unintended consequences to worry about (or at least know myself to be unbiased enough and in the posession of “enough” knowledge), then C: kill John AND distribute the profits if the $1 billion gain is worthwhile on the whole, otherwise A: forbid. But that is a silly assumption unless I happen to be a mature post-singularity sysop or something.

    In today’s world, I have neither the knowledge nor the computational resources to make that call, so I use heuristics to take such side effects into account. Particularly, and on this scale, then B: don’t interfere. In the long run, and with non-catastrophic risks, these things tend to sort themselves out by removing the decision-making ability of poor decision-makers.

  • http://jamesdmiller.blogspot.com/ James D. Miller

    I strongly believe that (B) is better than (A). Many people rationally risk their lives for money and the government should not try to stop this. If it did, then the government would have to outlaw dangerous jobs such as construction work and commercial fishing.

    I also strongly believe that (A) is better than (C). I don’t believe the government should force a worker to die so a firm can benefit.

    But I also think that (C) is better than (B). If all employees are equal it seems better for one of them to die than for three of them to die. Clearly, if everyone was ignorant of who would be the employee to die in method (1) then method (1) would be better than method (2). I don’t see why actually knowing the name of the employee who will die in method (1) should change anything if all employees are the same.

    I therefore believe that B>A>C>B. My beliefs, therefore, are intransitive meaning that they must be irrational. This is why I presented the problem. Does anyone else share my intransitive irrationality?

  • http://www.ProductivityShock.com/ Jason Briggeman

    D) Force the company to complete the project, turn its fruits over to me, and make it give me hundreds of millions of dollars as well. Why not? You’ve basically posited a government with unlimited powers of coercion.

  • Silas

    Roger: I’m familiar with the distinction between a “known” probability and a historical probability, but I don’t think it’s relevant here, or to the examples I gave. While people are typically never given the direct choice of a known probability of death, they are nonetheless offered risks, which they estimate as having some probability distribution. While a law enforcement agent might not know exactly how much his death risk increases by going to enforce some regulation, he believes it does increase, and considers the compensation sufficient. Choosing option (3) therefore necessarily involves making exactly the choice the employer just tried to, with a different utility metric and risk/compensation numbers.

    (And as an FYI, the employer does not consider the “guaranteed death” an option. It can’t force employees to do things without paying them; only the government here will cause them to do the “one guaranteed death” option.)

  • http://profile.typekey.com/sentience/ Eliezer Yudkowsky

    James Miller, you’ve certainly got a problem there. My guess: By saying it’s “better for one worker to die than three workers to die”, you are ignoring the consequences of having the government in particular enforce this tradeoff, yet you take the consequences of government into account in saying that (A) is better than (C).

    The position I take for human societies is (B). If I recall correctly, the revealed economic value of a First World human life (naturally, these are more expensive than Third World lives) is on the order of $3 million. If you don’t like that, ban automobiles.

    However, it would be proper for a sufficiently smart superhuman intelligence (AI or otherwise) to intervene with either (A) or (C).

  • michael vassar

    Eliezer: Isn’t it better, given that you are hypothetically a regulator in a position where you *do* have the power to do (A) or (C) to be accused of hypocrisy and to do (A) or, more probably (C). Your decision to use that power is not the same as the decision to give you that power in the first place, and it is that decision that you think is a bad idea. Sometimes hypocrisy is even a good talking point. You can more credibly argue against regulation and be seen as a nuanced thinker rather than an idealogue if you say something like “yes regulation can make the world a better place, in fact, I can think of situations where as a regulator I have been able to save lives and/or bring about more efficient outcomes, but even with my awareness of this potential good I think that the risk of abuse of power outweighs the benefits that society derives by giving regulators such as myself the life and death power that we have been given. Here are some ways in which our powers could be curtailed to reduce tis risk of abuse. They are not without their costs relative to the current situation, but I judge those costs to be greater than those of allowing the current situation to continue. If you trust my judgment enough to give me the power you are trying to give me you should also trust it enough to seriously reconsider when I tell you that my considered opinion at this time is that people should not be trusted with power in this manner”

    Robin: Lots of new commentators here. Any idea where they are coming from?

  • http://entitledtoanopinion.wordpress.com TGGP

    John Mark Rozendaal, you really shouldn’t shoot your mouth off like that unless you don’t mind coming off looking silly. There are a considerable number of libertarians who are anarchists and advocate the elimination of absolutely all government regulation. I am not one of them, but I don’t accuse them of being disingenuous. Kevin Carson, in particular, focuses a lot on how regulations help big businesses at the expense of smaller ones and uses the term “vulgar libertarian” for those who ignore this issue, but he doesn’t claim all anarcho-capitalists (he is on the left and supports the free-market but not “capitalism”) don’t actually want the elimination of regulations favorable to business.

    Alex Tabbarok discussed a similar issue here.

    Some economists claim that Somalia actually improved when it lost its government.

    My answer is that if John had agreed to accept such a risk, the government should do nothing. If the rest of the company simply discovered that they could kill John to accomplish this task and John refuses, his wish should be upheld by the authorities. Employees willing to randomly risk their lives should be allowed to do so.

  • Cássio

    I strongly believe that (B) is better than (A). Many people rationally risk their lives for money and the government should not try to stop this.

    Yes, indeed. Many people risk their lives for money. Life is full of policemen, firemen, soldiers, bank robbers, drug dealers, lion hunters, and so on, all of them are solid examples that some risks are worth enough to be taken. They risk their lives for subsistence. Risk is probability, not a certainty. With actuarial mortality of 5% per year, most policemen (95%, right) will arrive home alive every night during this time span.

    Usually people don’t offer themselves in holocaust in exchange for any amount of money that they obviously will not use, otherwise you can tell me exactly how to take my millions to paradise or wherever after death.

    If the State(s) is not able to keep any single private institution from taking the life of an employee in exchange of money, then what does the State(s) stands for? Ask Hobbes the answer, maybe ask Washington – not the government, but George.

  • Roger

    Silas: In principle, I agree with you, but the information given isn’t enough to make the determination of the motives of the people involved. I’m perfectly OK with allowing people to risk, or even sacrifice, their lives for the greater good. It’s a selfless and altruistic choice. Personally, I’m not OK with allowing people to risk or end their lives for purposes that are blatantly evil.

    The corporation could be firefighters, but then again, it could be a charity set up by extremists who want to bomb subways. We don’t know and our default answer should be to prevent such an action until we know more.

  • Silas

    Roger: I assume any information not given in a moral dilemma can be taken to be morally neutral. Therefore I assume that the project in question is otherwise not morally objectionable. Otherwise, you’re barring yourself from ever being able to answer any such question because, “But I don’t know if such-and-such is on his way to murder newborn dolphins!” Your methodology justifies preventing all action.

    And nothing in my justification involved a “sacrifice for the greater good” as such; these workers are presumably taking the risk because they are sufficiently compensated for it.

  • http://www.luchko.ca/~aaron/wp/ Aaron Luchko

    At first my thought is to say as a government regulator I’d rule out (A), if they’re willing to risk a 3% chance of death for $5 million I’m fine with that, that just leaves (B) and (C). I’m tempted to let employees vote on the matter but is my decision over their welfare any less valid than their decision over John’s welfare, in each case a 3rd party is sacrificing John’s life? The main difference I can see in the regulator forcing 1) as opposed to the employees voting for 1) is that the weight of the decision rests on only 1 person, who never knew John. Thus I think I’d choose (C), assuming the final stipulation that the task is validly worth $1 billion (ie creating usable wealth, not just redistributing the 1 billion without any greater benefit). The $5 million is based on the fact that at some point in the past John was a random employee and the employees are willing to take the 3% for $5 million.

  • Caledonian

    Robin: Lots of new commentators here. Any idea where they are coming from?

    I can’t speak for the others, but razib specifically praised this blog – and linked to it.

  • Julian Morrison

    A principled libertarian / free marketer would say (1) you MAY NOT kill coercively, neither one nor three, and the task must go undone if this were the only option. The actual worth of the task is completely immaterial. (2) In the worst successful case, you might be able to find suicidal volunteers willing to accept payment in the form of favors for third parties. (3) In the best successful case, you can and should break the arbitrary and imposed rules of the game by discovering a means to complete the task without killing anyone. All reasonable effort should be bent toward this option.

    Moral dilemmas are creatures, themselves, of the regulating mind, because they say “by my command, the world is thus, now solve an impossible problem”. The free marketer sees the world as existing without fiat, and will attempt to find a way to bisect the dilemma.

  • Julian Morrison

    I’ll give you an example of what I mean: Global Warming. That’s a classic “moral dilemma”. Either we permit the planet’s climate to warm, with various semi-predictable and probably on balance unpleasant consequences, or we drag industrial civilization to a screeching halt, killing many thousands in a worldwide super-recession, and stranding the third world in perpetual low-carbon poverty.

    The typical entrepreneurial thinker would refuse to tolerate the dilemma, and instead cut right down the middle with questions like “how about nuclear power instead” and “why not engineer the climate, there are probably hundreds of ways to do that, we already cause global dimming with particulate aerosol emissions”.

  • Anna

    Random vs. Certain Death

    I would take certain death as I wouldn’t like to be tortured.
    Anna:)

  • michael vassar

    TGGP: I agree that John Mark Rozendaal came across as silly, but if he had said “The people with actual power who say that they favor free markets actually favor market regulations that advantage them and not others.” he would be more or less correct, would he not?

  • Anonymous

    I think the key to solving this puzzle is the phrase “you are a government regulator”. Your goals in order of priority are:
    1. You need to be perceived as increasing safety.
    2. You need to increase the number of employees working in your agency.
    3. You want to reduce the number of people who die.
    Goal 3 is important, but it is significantly lower in priority than goals 1 and 2.
    One fairly good strategy would be to do nothing beforehand, and then slap a huge fine on the company afterwards. You should be able to collect at least $1.5 billion because the company will be getting $1 billion profit, plus the company will have reserved $5 million for each of its 100 employees.
    After you have the $1.5 billion, give a large amount to the families of the people who died. Then give a large amount to lawyers who are working on contingency. With the remainder, hire a PR firm. Don’t give any of the money to the employees who didn’t die (explain that this is blood money and they shouldn’t profit from other peoples’ tragedies). Also, put the CEO in jail.

  • http://www.luchko.ca/~aaron/wp/ Aaron Luchko

    There’s one interesting aspect to the idea of consent.

    Clearly at some point in the past, before this scenario arose, John was essentially random, therefore the chance that he would be the employee to die in the scenario is a 1% chance. Also note that he is one of the employees willing to accept the 3% for $5 million, so clearly he would be willing to accept a chance of 1% for $5 million (option C).

    The main thing here is he is never asked if he would accept this wager, we know he would have accepted if he was asked but the fact remains that he was not asked.

    Therefore how beholden is John to a deal he would certainly have made, but wasn’t given the opportunity to make? Note that while John is receiving the poor end of the deal the other 99, who also never agreed to the deal, reap the benefits.

  • Cassio

    One fairly good strategy would be to do nothing beforehand, and then slap a huge fine on the company afterwards. You should be able to collect at least $1.5 billion because the company will be getting $1 billion profit, plus the company will have reserved $5 million for each of its 100 employees.

    It seems a clever resolution of the problem, but it is not consistent with democracy. In a democratic and constitutional State one cannot apply fines and punishment to whoever, if there is not a legal support for this. So, to create this system of regulatory acts and laws, legislators must act beforehand and. For example, it is forbidden to cross when the red light is on. If you do it, you can be fined or even arrested, depending on the consequences. However, there is no specific prohibition on hanging a coat on a chair. No one can fine or arrest you for that. If someone does, without legal support, you are closer to suppression of individual guarantees than you’d like to be.

  • http://www.jmrozendaal.com John Mark Rozendaal

    Comment writes:
    “So that’s what it comes down to, a smear on the character of libertarians, based on the writer’s own abysmal ignorance.”

    My posting above was indeed ignorant, and certainly contained a straw man. I do not believe that it was ad hominem.

    TGGP writes:
    “John Mark Rozendaal, you really shouldn’t shoot your mouth off like that unless you don’t mind coming off looking silly.”

    Actually I don’t mind coming off looking silly. I don’t mind exposing my abysmal ignorance either. I learned several things here, and I hope that my learning didn’t cost the others who participate in this blog too much. A good debate on ethics has value far higher than my concealing my abysmal ignorance or appearing clever.

  • ehj2

    A consequence of our technical/medical capacity to harvest human organs to extend the lives of the privileged wealthy class means we have people in the world being offered big money for their organs, or simply being killed for those organs (not my idea of a “good” unregulated market) right now.

    History teaches us that wealth always conspires to maintain wealth, and we should know that corporations are required by law to pursue profit over social and moral considerations — so there will always be a class of people susceptible to this offer of “cash for death.” Will we offer a family in poverty large sums to harvest the organs of one healthy child? Will we decide to harvest organs from those with mental disease? Given the cost of parenting, should we encourage a family to have an “extra” child (they pick the weakling of the litter) to sacrifice for the good (college education) of the others? (I suppose a libertarian might argue this is quite similar to the tradeoff made in nature.)

    How about criminals? Larry Niven has written at length and quite persuasively on this subject. When people want something, they will gradually eliminate the barriers to getting it. We might start with only those criminals on death row. After all, why waste their organs? But that won’t be enough. We’ll need more hearts and livers and cataract-free eyes. So we’ll make more crimes deserving of the death penalty.

    The price of life has always been the price of some other life. Now, medical technology offers the possibility that the “other life” we’re living on can be human.

    ~~~

    A very real problem for libertarianism is that people can’t live as islands (individually), and no group of people has ever persisted for more than a few moments without a member wanting to impose their will (and regulate) the collective. Even the libertarian impulse is the imposition of a certain kind of rule — a “no-rule” rule. We’re a social species with extremely aggressive tendencies — and that should end the argument. To have a “no-rule” state would require a strong central authority to impose and maintain a “no-rule” rule. The trouble is, a state strong enough to maintain a “no-rule” rule always works to benefit a “leader-class” (hence, the unrealizable utopian goal of libertarians to be rid of regulation).

    The best we can hope for is regulation (rule-of-law) that supports the individual as an equal (as difficult as that is to define) to all others in the collective (yes, worldwide), with powers vested in the people to manage and “own” the leadership. Our experience with even the approximations of “no-rule” rule in unbridled free-market capitalism should inform us of its winner-take-all rapaciousness — and the danger of that short-sighted rapaciousness to the future of all life on this planet.

    /ehj2

  • http://www.satisfice.com James Bach

    This is more bad science fiction. There’s no real moral dilemma here worth analyzing, here. You might as well talk about the ethics of PacMan eating up the little blue monsters in the eponymous video game. Whatever you guys think you’re doing, wouldn’t it be at least a little more interesting to try doing it with a situation that has some remote possibility of being encountered in your actual lives?

    History magazine did an article, this month, called “What Would You Have Done?” based on the book The Darkest Hour, by Laurence Rees. I must get this book. The article describes moral dilemmas faced by specific people during WWII and how they dealt with them.

    Examples:

    – As a teenager, Estera Frenkiel, a secretary in a Lodz ghetto, was given ten certificates excusing Jews from the death camps. How did she decide who to excuse?

    – Having joined SMERSH, the notorious Soviet counter-intelligence unit, Zinaida Pytkina was expected to kill a German prisoner. How did she justify it?

    – British intelligence officer Nigel Nicolson was told to lie about the likely fate of Croats handed over to Tito’s partisans. Could he obey orders and still find a way of telling the truth?

    — James

  • Thaddeus McMonster

    Each employee has a right to his own life, which he can voluntarily waive, but not transfer (if he could transfer this right, then slavery would be justifiable). So, while any employee would trade a 3% chance of death for $5 million, he has a right to exit the lottery at any time, even after it takes place.

  • Stuart Armstrong

    Dear Silias:

    Stuart_Armstrong: You’re saying that “not permitting anyone, ever, to pay or be paid for absorbing an increased risk of death” is “what you would do” and is “consistent with a respect for human life”, the latter being “infinitely precious”?

    In that case, the next steps would be to explain:

    -Why you paid or were paid for absorbing higher risk of death several times during the past week.
    -How you would get law enforcement to take on the increased risk of death in enforcing regulations such as these.

    I’ll do this, as it’ll help me make my ideas coherent. I see no objection to paying people to accept increased risk (I personnally feel they should be paid more than they are, but that’s another issue for another day). I feel though, that there is a difference between the set-up “accepting a bribe to increase your risk of death by 3%” and “accepting a bribe to be one of a hundred people, three of which will be killed (by a corporation or an individual)”. It is the prohibition against murder that is being diluted in the second case (and it is being strongly diluted: someone is being killed, and that person can’t object, because “he accepted the risks”). The first case just dilutes the prohibition against being exposed to increase danger.

    Now there are cases where the two are indestinguishable (war, for example), but I feel it has been a great source of progress that we have maintained a prohibition against murder in times of peace – even if, from the point of view of the corpse, the result is the same.

    I feel the same about slavery, incidentally, and somewhat similarly about polygammy – the arguments for accepting them, among consenting adults, are very strong. But whenever they have been implemented, they have generally been disasterous, and the key word “consenting” seems to get sucked out of the system 🙂

    Please point out the weak points in this argument – I want to explore this area further, and counter-arguments help a lot.

  • Stuart Armstrong

    Dear Constant,

    But my point is not that people advocate resurrecting the USSR, but that the way it turned out is a real, recorded unintended consequence of taking a certain approach to the problems of society.

    Yes, and the point is good. But the comparaison is somewhat limited – the USSR, for instance, was a dictatorship. A closer example would be Western Europe after the second world war, up to the 70’s. There, a certain approach to problems in society, resulted in:
    1) Spectacular, record-breaking growth, relatively well spread across society,
    2) Fullish employement,
    3) Some public policy disasters,
    4) Long-term structural problems,
    5) An eventual reduction in the government role in the economy.

    A mixed picture, to be sure, but not an entirely negative one (and not exclusively due to starting from a low base, or to the Marshall plan, either). And, as for the particular example being discussed, as far as I know, every country would outlaw such behaviour. So I don’t think we can argue from history that any intervention by the state in private companies is a disaster (in certain specific instances, I’d argue they were very beneficial).

    People prefer to eat than to starve; they prefer to be acknowledged than insulted. A government which operates along those principles (with exceptions for those who clearly demonstrate that they are in a different case) would not suffer much from this unknown preference issue.

    But what is the government competent to do, given this knowledge? Produce food? Recall the Soviet and Chinese famines. Oops, there I go again with the historical examples.

    Historical examples as good! I’d personally prefer the government did nothing with the knowledge that people prefer praise to insults – but the point is, they know it, and if they acted on it, lack of information would not be the problem.

    As for the food, there are a whole host of things a government can do. The simplest and the most effective is susidise it – when it does so, more food is produced. Historical example: England, which went from an under-producer to a bloated over-producer through subsidies.

    Well-run governments also show skill in the distribution of food, in times of shortage (again, England, WWII, rationning – hell, even the Nazi puppet regimes of western europe could manage that). There are costs to both these interventions (distorted markets, black markets, long term unexpected consequences), but there are also costs to not doing anything.

    I feel there is a saliency issue here – if the government does something, and it has costs, then the government is to blame. If it does nothing, and there are costs, then the universe is to blame, or it’s part of the natural order of things, so people accept it better, even if the result is worse.

    Please don’t believe I am advocating unrestrained governmental interventions, by the way 🙂 Lacking information on people’s choices, fostering a culture of dependency or encouraging rent-seeking, excessively using coersive force, bloating laws and regulations, etc… – these are the costs of interventions. But I don’t feel they are high enough to rule out a priori all governmental interventions. In some cases (old style elite-only basic education) I feel these costs are insignificant in comparaison to doing nothing.

  • Kevin

    The correct answer, of course, is obviously (insert letter here)

    Isn’t it amazing how many people start like this? Anyway, I believe the Utilitarian thing to do here would be to allow the company to go ahead with the plan. UNLESS – It can be “kept secret” that the government ordered John to be killed after all. If it is almost positively certain that the government’s decision to have “only John” killed can be kept silent, and made it seem like the company ended up with option B, then there is no corruption of implicit social contract, there is no loss of belief in the system, and thus you can maintain stability while maximizing your consequentialist calculation.

  • Caledonian

    But then you have a corrupt government that manipulates the people for its own end, Kevin, and that never ends well.

  • Douglas Knight

    But whenever they have been implemented, they have generally been disasterous, and the key word “consenting” seems to get sucked out of the system

    What are you talking about?
    I doubt that polygyny leads to women-as-property, rather than the other way around (but I don’t know its history). Slavery is associated with raiding your neighbors for slaves; but (1) that’s the least of the problems and (2) there wasn’t much consenting going on in the first place.

    The classic example of coerced employment is at sea, but the problem there is more the physical constraints than the institution of absolute power of the captain.