To what extent should government seek to achieve equality? In order to keep it simple, I will assume a single value judgment: that preferences are the best guide to judging the rightness or wrongness of any particular situation. If you don’t like this value judgment (or think we should have others), that’s fine, but I don’t want to get this sidetracked into a debate about whether coercion is intrinsically wrong or equality is intrinsically good.

What are the factors involved in creating economic disparities and more broadly, welfare disparities? One that I believe is not relevant is “preferences over work and leisure”. I accept that policy might not be able to practically separate them, but from a philosophical standpoint a sole focus on material equality seems arbitrary. Why make a distinction between someone who prefers to work longer to make their leisure more enjoyable and someone who prefers a greater quantity of lower quality leisure? The only way to do so is to judge people’s preferences as intrinsically good or bad. Therefore, any measure of inequality should be based on the best possible measure of total welfare, rather than select aspects.

There are two genuine, broad factors involved in creating inequality. They are:

- Disparity in endowments. These include natural differences (such as intelligence, beauty and inherited health problems) as well as environmental differences such as quality of parenting and early educational opportunities. Finally, another factor is straight transfers of wealth by means of inheritance.
- Disparity in results of gambles. Or in slang, “running hot” vs. “running bad”. Quite a lot of actions have a considerable degree of luck in them. There are the obvious ones like “winning the lottery”, but getting spotted by the right talent agent or accidentally tapping a large undiscovered market niche are less apparent.

These are roughly correlated with ensuring “equality of opportunity” and “equality of outcome”, respectively. The important question though, is how much should we ensure, based on preferences. Looking at utilitarian arguments for greater equality:

- Economic equality is correlated with social cohesion and a number of other factors (e.g. homicide rates and inequality). Such research does attempt to correct for other variables, making causation reasonable likely.
- Diminishing marginal utility of money means that poor people would tend to benefit more from their last dollar than a rich person would. Hence, the theory suggests total welfare may be able to be increased by redistributing income.

These may seem like compelling cases for redistribution of income, but you have to consider why the free market could not already solve them. For a tool already exists to capture the benefits of greater equality: insurance. If the rich and the poor person could go back to a point where they were both equal, they could strike a deal that would have the richer of the two compensate the poorer one in the future. This already happens with current types of insurance: everyone agrees in advance to compensate the person whose house burns down, with the idea that the now comparatively poorer person would value money more than the others. Sure, there is moral hazard, but evidentially the expected costs of this are outweighed by the expected benefits of spreading the risk for the contract to have gone ahead. The question of whether the government should be involved in redistribution, therefore, should be to the degree that a market failure exists in insurance.

For disparities in endowments, this certainly seems to be the case. The “genetics lottery” as well as the “quality of family lottery” are both involuntary gambles. Given observations of insurance in real life and the general tendency for humans to be risk-averse, we can conclude that if they were given the option, potential people under the veil of ignorance would want to purchase at least some "endowment insurance." (For a good non-academic article on this issue, see "How big a safety net?" by Steven Landsburg). The exact amount would depend on the costs – if it were completely costless to redistribute endowments (you can equivalently do this by taxing their purchasing power), it is likely they would go for completely equality, but this is not the case: endowments are hard to observe. Still, they probably would purchase some insurance. The problem is, as Steven Landsburg puts it, “there are no telephones in the world-before-birth, so insurance salesman couldn’t contact you.” You would like to have purchased insurance – but you can’t rely on collecting premiums from people after the event. It would be like if you could decide on whether to purchase travel insurance after you came back from your holiday.

This implies a reasonable case for having government engage in some redistribution of the benefits of endowments. Note, however, that there is less of a case for government intervention when parents can intervene: parents could purchase some “skill insurance” if they wanted to, even if it was limited to payouts for severe genetic defects and nothing else. Factors that are already observable, such as the genetic characteristics of parents (giving the potential child more ‘genetic lottery tickets’) and the wealth of parents cannot be insured against at that point, since they are already known.

Now we look at matters of luck after birth. Take Bill Gates – are his endowments really worth tens of billions of dollars? His genetic gifts are probably fairly substantial, but it is unlikely. A great deal of his wealth is likely a result of simply running way above expectation. If you wanted to tax the value of his endowments, he’d still be extremely wealthy. Diminishing marginal utility of money should still hold, so could redistribution be welfare enhancing here?

Again, you have to compare this to insurance. Why couldn’t Gates and a bunch of the other people you’re planning to redistribute his wealth to have organised an insurance contract that would have the wealthiest in the future pay money to the less wealthy? It would achieve all the benefits of equality but more importantly, do so based on the specific preferences of each individual rather than rely on a standardised estimation. The concept of the diminishing marginal utility of money can lead to erroneous conclusions, since everyone has a different utility function. If a group of people refuse to take up insurance despite the risky nature of their endeavour, that is a revealed preference that they make be risk-seeking and that forcing them to accept less risk will make them worse off. (For that reason, taxing lottery winners on distribution grounds is dubious: lottery players have shown by revealed preference that they are happier with a massive redistribution to a rich person than equality for the value of the ticket). If another group full of highly risk averse people is willing to give a lot of $EV to increase their expected utility, they can do as well. The key point is that unlike for initial endowments, opportunities do exist for people to purchase insurance. To begin a claim of justification for government intervention, you need to show what stops them from properly creating their ideal contracts.

Now, it is possible to make a case for market failure existing in insurance. Adverse selection due to asymmetric information is probably the biggest one. While moral hazard is a problem, government-enforced insurance creates pretty much the exact same problem in the distortionary nature of taxation. Additionally, even with some legitimate market failures, they may be sufficiently outweighed by government failures to make attempting to correct them even worse. But even if you can make a case for government to redistribute at this point, it is a much weaker one than for redistribution of endowments.

I'm confused as to whether you are referring to "insurance" in a metaphorical context, talking about government social programs, or proposing some new sort of program based on genetic screening that would compensate the least well-off...

"The key point is that unlike for initial endowments, opportunities do exist for people to purchase insurance."

???

How does one purchase insurance against adverse life outcomes? Is this a product offered by AIG or Berkshire?

The idea is that government efforts to redistribute wealth should be thought of as insurance, and compared with free-market insurance to see if there is a valid case for it. You could try genetic screening, but that isn't necessary - a less precise program can still attempt to alleviate the same problem, although with more distortion.

You certainly can purchase insurance against a number of bad events, like loss of property/life/job. Although many contracts aren't offered commercially, they can still be done informally. The idea at least is that people can potentially create a contract before the event with life outcomes (they can't before birth). If people don't purchase insurance against a certain risk, you can either conclude (a) they don't want any or (b) there is a market failure preventing people from creating an effective contract. If (b), you can consider whether "socialised insurance" will create a better outcome.

You are assuming that it constitutes a market failure if people don't get what they want. Of course, you could define it as such, but then there would be virtually unlimited market failures.

And how could you "consider whether 'socialsed insurance' will create a better outcome". "Socialised" implies a coercive action, where somebody is worse off as a result (otherwise no coercion would be needed). How can you say then whether the outcome is "better"?

It's not so much that they don't get what they want, but that mutually beneficial contracts cannot take place. A group of unborn children would all benefit ex ante from purchasing "endowment insurance" but since they can only begin to make decisions ex post, the contract can never go ahead.

My value judgment is that "an outcome is better if individuals under a veil of ignorance would prefer it." Or even without using that example, "if all individuals are better off ex ante".

That endowment insurance would benefit the babies is a huge assertion. Definitely depends on the insurance (the price for existing "endowment insurance" or social security is WAY too high).

Yes, I am aware the current level of insurance in the United States may not be at the optimal level. But that's not what I'm trying to determine. It's merely a method of evaluating different aspects of redistribution.

So because scientist say this is benefitial to these unborn, they sign the insurance contract in their name before they are born, is this correct? OP, do you assume that all unborn babies would "want" this type of contract?

All "potential beings" would either unanimously support or oppose such a contract because they are all identical - i.e. they know nothing about who they'll become, so they have no separate preferences yet, just the average preferences of the population to work on.

Might the estimate of insurance be wrong? Sure. But uncertainty in an area doesn't mean that nothing should be done. There is nothing particularly special about the status quo - it could easily be wrong as well.

What do you mean with the bolded part? Don’t you agree that - since they can’t think - they have no preferences. Only later when they are capable of thought on this subject will they have preferences.

I look forward to the discussion on this issue. I agree that the status quo is far from perfect. Also, I’m only half brainwashed by the AC crowd and still see some merits in equality VS aggregate wealth.

That a "potential being" has an equal probability of being born as anything, so the decisions it would make are based on the preferences of everything.

I think "what an agent would think if it could" is a valid ground for preferences. There's nothing unusual about this on an individual basis - people certainly take into account the welfare of their unborn children in their decisions despite the fact they can't form preferences yet. It would also create some fuzzy issues with say, comatose people: is it okay to do what you want to a comatose person who hasn't left explicit instructions because they can't (yet) form a preference on what they want?

ACists: Do you think potential parents should be able to sign a contract that would involve their children paying premiums (in the same method as taxation) which would be used to compensate children who got the short-end of the genetics stick? What if those parents would not have had children otherwise?

What about grandparents? If all four grandparents had signed the same contract when they were younger, can they insure the endowments of their grandchildren as well?

What if every single human signed this contract? Could we commit every single future generation to this (unless they signed a different contract)?

Can you get a bit more specific here? I think I sort of understand what you're saying but I'm not exactly sure.

The OP is making several good points here, but I don't quite understand how his idea of endowment insurance is different than just asking the question "What society would be people choose behind the veil of ignorance?"

As far a parents signing binding insurance contracts for their children, I don't see how this is any less coercive upon the children than forcing them to participate in state social safety net programs by virtue of their citizenship. Now, I have no general problem with what Acists call "state coercion". But the OP's solution seems strictly inferior to state insurance because it is similarly coercive but does not actually solve the problem of inequality that the OP wants to address (because some children will be involved in endowment insurance contracts and others won't by accident of birth).

First possibility: A group of parents planning to have a child within the next couple of years sign a contract. The contract states that for the rest of their lives, all of their children will be required to contribute thirty percent of their income to a fund. The fund will distribute assistance to needy members of the group. The parents have an incentive to make the fund as effective as possible - if they didn't think it was in their child's best interest ex ante, they wouldn't sign the contract.

That doesn't seem to me like something many ACists would support. It seems like it gives too much power to parents. However, there's somewhat of a grey area: how much choice should parents have?

Second possibility: The same as the first, except the children are liable to pay a lump sum payment according to the results of a high-quality genetic test. (So the child would be slapped with a large debt to pay in their adult life if they tested as a prodigy, whereas children with disabilities would receive a small trust fund).

If you were opposed to the first option, you'd likely oppose this one too, on the basis that it allowed parents too much control over their children. However, it's functionally similar, replacing lifetime payments with a lump sum - they might both have the same present value.

Third possibility: Parents pay premiums and receive payouts inversely correlated to their child's genetic profile.

I can't see how an ACist would oppose this - to be consistent, you'd have to oppose parents doing any form of gambling or insurance. However, from the point of the view of the child, the incidence on them would be very similar to the second option. That is, regardless of whether the parents or the children pay, there's a good chance that (at least with responsible parents) the child would be just as well off. If the parents are required to compensate the losers rather than the child, the child receives less inheritance/less childhood assistance, so is probably about as well off as under the second option.

Fourth possibility: Parents have no rights over the children.

Where do you draw the line, if you disagree with some or all of the above? Surely we couldn't give even potential children complete rights? That would imply that parents have a moral duty to have as many children as possible, which might get the state involved in forcing them to do so. Even only dealing with already-born children, it would still involve drastically regulating how parents can raise their children

It isn't really. Indeed the Landsburg article does exactly that: taking Rawls' argument and framing it as being exactly like an insurance contract. I find it to be a good rhetorical device however, especially when compared to "life outcome" insurance.

The question it raises is "to what degree to ACists think parents should be allowed to coerce their children?" Clearly allowing for absolutely nothing would be problematic, as would be selling them into perpetual slavery.

I'm an ACist and I have no idea what my answer to this question would be. I'm very interested in hearing logical arguments from all comers.

Oh, OK. This contract is invalid.

Edit: as is the second. Third seems fine to me. 4 and 3 do not seem mutually exclusive.

Children don't consent to other decisions their parents make for them either. A poor decision (or even a good decision with bad results) made the parent in early childhood could impose lifelong costs on the child. Should a third party (let's say, the state) monitor the behavior of parents to see if they're gambling with their children's future? They almost certainly will be. Many decisions a parent makes - say, what school their child goes to (school choice in AC-land presumably) result in making a gamble with that child's future. Maybe they could send them to an experimental school that has disproportionally high numbers of its students go on to be highly wealthy, but also with disproportionately high numbers of them becoming homeless.

Decisions are not contracts, though.

Does the child care?

Perhaps not right now. That doesn't matter, though. I see your point, both have long-term impact on the person, but one creates obligations for the child and the other doesn't. That's the difference.

Am I seriously the first to question this part of the OP? What about inequality created out of choice? As in, some people choose to work harder than others?

I believe this is addressed by the OP in the 2nd paragraph where he says "One that I believe is not relevant is “preferences over work and leisure”."

Essentially, gains in material wealth due to hard work are not actually inequalities. They just reflect different preferences over material goods versus leisure time. Some people "buy" extra leisure time by choosing not to work very hard, and we should not consider them unequal because they have chosen to buy time rather than goods.

However, if two people who work equally hard earn different amounts, this is either because one has greater endowed abilities, or because they ended up on different sides of variance. This is more morally problematic.

This is correct, although note that different preferences can create inequalities in welfare due to luck over prices. For instance, someone who enjoys driving may fluctuate from being better or worse off than someone identical except for some preferences depending on the price of oil. Now some of this could be easily alleviated by free-market insurance - there are certainly ways to hedge against price changes in your lifetime. However, you cannot insure against prices before you are born. So in that respect, preferences can be like endowments in their effect on welfare.