Over the space of one, or even two decades, in a reasonably industrialised nation, the economy does not grow perceptibly.

This statement is not in agreement with reality.

Even if we take the growth numbers reported at face value, and ignore the fact that part of the growth in most Western countries stems from funny-money being moved around between different accounts

a decade of growth - say, the five-year periods immediately before and after a policy is adopted - would grow the economy by less than forty percent.

And that is not "perceptibly". Sir, you must be jesting.

This makes median income relevant, because significant changes in the income distribution quickly become a zero-sum game.

Income distribution is per definition a zero-sum game, since it's a matter of how the total income is distributed. And the total income is always, guess what, 100% of the total income. Duh.

INCOME is never a zero-sum game, though.

Thus, if you wish to measure how a policy affects poverty, some fraction of the median income offers a quite reasonable proxy. In the sense of evaluating concrete policies - which was the original topic of this discussion - that makes median income a perfectly relevant tool.

To measure income distribution yes. To meaure poverty, no.

Your example of North Korea having low relative poverty, while correct, is a red herring

No, because the point is that with your argumentation, North Korea is a less poor nation than France, because the income distribution is more even. And that is absurd. Which is my point.

Furthermore, even though I defended - and will continue to defend - median income as a valuable proxy for short-term calculations, you completely sidestepped the fact that I proposed an semi-absolute (absolute in space, but relative in time)

Space?

measure of poverty: Shelter, heat, food water, education, access to information, access to standard of care-level medical care, access to medicine ('access' in this context means reliable access). I would ask you to evaluate this poverty metric.

In fact, that's pretty much what HDI does. Which you didn't like.

I also believe that you overlook an important fact in your discussion of the arbitraryness of boundries used in the calculation of relative poverty: The natural boundary to use is the area of jurisdiction in which the policy is being contemplated, since the value of fraction-of-median-income seems to me to be in short-term-evaluation of policies.

You see, I don't agree with that, because in my opinon, poor people continue to be poor even if policies change in neighbouring countries.

Furthermore, I will happily acknowledge both that fraction-of-median income is meaningless outside the evaluation of reasonably industrialized economies (a criterion that neither North Korea nor Beverley Hills fulfills) and that it is not directly comparable to most other metrics of poverty.

This is not a problem, however

No, it's not a problem, but that fact that it doesn't relate to other metrics of poverty means either it or the other metrics doesn't really measure poverty at all. And we already know which one that doesn't.

It's time to stop this stupid charade. Percentage of mean income is not and will never be a valid measurement of poverty. No matter how much you twist and turn and start using fancy words that make you feel like you understand things, it is a measurement of income distribution, not poverty. And that's that.