[bzbarsky]

This presupposes that all money, and in general all resources, including labor, belongs to the state (or society) except that which is explicitly allowed to people.

While that is a coherent point of view, it has some really unpalatable implications. The first one is that the assumption that society owns everyone's labor until proven otherwise is a great justification for various forms of effective slavery. The second one is that taking your "useful to society" statement at face value leads to things like the eugenics movements of the early 20th century, withholding (or confiscating) resources from those deemed "useless" to society, etc.

But even within this framework, making more money than you spend corresponds to producing more resources for society's benefit than you consume, and thereby building up a sort of "social credit". You then draw on this credit later in life (e.g. in retirement). It seems like the main claim you are making regarding inheritance is that this sort of social credit should not be transferable, right?

But that raises the question of whether people be able to give gifts to someone else at all. In this framework the answer is basically "no", because that would represent a transfer of the non-transferable social credit. But if they _should_, then should there be a substantive difference between a gift given 10 minutes before someone has a heart attack and receiving the same gift 20 minutes later? From an ethical point of view, I have a hard time with there being a difference between those two cases.

[Matumio]

You can also earn money just by inheriting a large pile and hiring someone who manages it. Or work for someone who did that. Contributing to society at large is entirely optional for some people.

It's not a problem when someone inherits a sum that can be earned within an average person's lifetime. It's the proportions. A single person should not inherit the right over several thousand lifetimes of work output.

No, it's not better when "the state" or "society" decides as a group where this work output should be directed instead. But that's not the only alternative. When someone argues against inheritance using the "meritocracy" argument, I don't think this implies that people who would have worked for the heir's money should instead work on a project that all members of society have sanctioned.

[bzbarsky]

> You can also earn money just by inheriting a large pile and hiring someone who manages it.

Just to be clear, in this situation what is going on is that you are holding the right to call on some resources (i.e. money) but are not exercising that right immediately. What you are doing instead is letting others use your right to get the resources they want now, with the understanding that in the future they will give you more resources. The "let others make use of resources that you can call on but don't need right now" part is in fact a _very_ useful social function; someone doing that is in fact contributing to society. Whether and how much they should be compensated for that is a good question, and I am open to arguments that the typical compensation for it is too high compared to the social utility of the activity.

And of course the exponential growth aspect makes things unsustainable over the long run here: people can't promise you more resources in the future indefinitely in the real world, though they have been able to do that for the last few centuries for various reasons. In conditions prior to that, here were various attempts at mechanisms for avoiding this exponential growth problem (e.g. the jubilee system described in the Bible is an interesting example of attempting to address the issue).

> A single person should not inherit the right over several thousand lifetimes of work output

Should a single person earn such a right to start with? In some cases, I'd think maybe: consider someone who saved society thousands of lifetimes of work effort in some way, e.g. via an invention. But I feel that this is an interesting question to ask as a baseline. Once we posit that someone _can_ own that much right-to-output, we're back to whether people can gift it and under what conditions...

> When someone argues against inheritance using the "meritocracy" argument, I don't think this implies that people who would have worked for the heir's money should instead work on a project that all members of society have sanctioned.

It seems like the basic options for dealing with inheritances are:

1) Some individual or group gets the money. This is inheritance as normally understood. This is the thing being argued against on "meritocracy" grounds above, afaict.

2) Society at large (the government) gets the money. This is estate taxes as normally understood, and corresponds to the "work on a project that all members of society have sanctioned" option.

3) The money just disappears. That is, there is effectively debt forgiveness for whoever owed money to the person who died. If they held cash, that would be the Federal Reserve. If they had bank accounts, that would be the relevant banks. If they held bonds, it's the bond issuer. For stocks maybe you can effectively "un-dilute" the other stockholders, sort of like a 0-price buyback. I have no idea what the real estate equivalent of this would be, or the equivalent for goods someone owns (books, clothes, furniture, piles of grain, etc).

I'm not thinking of other options so far, but open to ideas on what other options would be.

It seems to me that one glaring flaw with option (3) is that it can be converted to option (1) with some planning: if the main assets you hold at death are IOUs from people and such IOUs get canceled at death, then you can just carefully choose who owes you money and they end up effectively inheriting it. It's a lot less flexible from just holding whatever assets you want and then having a will, but for the really large fortunes it would not be difficult to structure it to be equivalent, I suspect, modulo the extent to which the "heirs" have control over the money before they come into their inheritance.

The basic issue with (3) is that it fundamentally doesn't _cancel_ the right-to-output; it just transfers it to someone else, still. I haven't been able to think of a good way so far to actually _destroy_ such a right, but I am not an expert in this area and haven't spent _that_ much time thinking about it.