> The problem is that eating food IS consumption. For very poor people, food is the main consumption.

Then tax consumption enough to pay for the food accounting for the consumption tax that will apply to the food. That the amount of money you need to pay for the food increases as the consumption tax used to pay for the food increases is not a problem as long as anything at all other than food is being consumed in the economy. (If it happened to be the case that nothing other than food was being consumed, that would just mean it isn't possible to supply enough food—an income or wealth tax wouldn't be able to change anything about this.) The consumption tax needed for the government to pay for the food will be lower than the income or wealth tax that would be needed for the same thing.

Taxing investment to pay for food is only going to work insofar as it reallocates resources from the production of whatever the investments were in to the production of food. A consumption tax can already do this reallocation more efficiently. If your goal is to make the economy produce more of food and less of other things, this goal can be achieved more efficiently with a consumption tax than with an income or wealth tax.

> That's the bad of taxing consumption - that it naturally taxes more the poor, making them even poorer, than the rich.

It only taxes the poor more in proportion to their wealth, to the extent a higher proportion of the wealth of the poor is consumed instead of invested, but not in absolute terms. In absolute terms, it is the reverse, since the rich consume on average more than the poor even if this consumption is less as a proportion of their wealth. The tax on consumption taxes your absolute consumption, not your consumption as a proportion of your wealth, so it does make the poor better off in relation to the rich.

> But if you admit that some public services are necessary (eg police, defense from external threats) then it's obvious that a 0 tax rate is bad - 0 taxes mean 0 public services.

Even if these public services could not be provided without taxes, the reason having no taxes would be bad would not be that it is an extreme. But public goods can be provided without taxes. For the police, courts and laws, chapter 29[1] of The Machinery of Freedom describes one way this could be done. National defense can be funded with dominant assurance contracts, as can many other public goods (even when they can't be made excludable), including lighthouses, scientific research, information goods such as books, software, or art, clean air, and free-to-air television. So even if some public services are necessary or even just desirable, it is not obvious that having no taxes is bad as they can also be provided without taxes.

[1]: http://www.daviddfriedman.com/Libertarian/Machinery_of_Freed...

You can't design rationally a new human society and expect it to work as designed. It's the same error that Marxists did.

Is that sort of meta reasoning an efficient way to get at the truth? There are much better arguments that can be found against Marxism. Saying "But what if the Marxists said the same thing?" is a low-effort excuse that can be used against the proponents of any belief you don't like. It's better to take the reasoning at face value and address the arguments directly.

I'm not saying "But what if the Marxists said the same thing?". I'm saying that trying to revolutionize how society works based on a theory never worked as expected in history, and cited the Marxists as the biggest example.

My first comment said that sensible economic policies are not politically feasible. I certainly don't predict that sensible policies are going to be adopted or sensible systems created through our current political institutions. Meanwhile, Marxism failed because the theory was wrong, not because it was politically unfeasible. It is possible for a system to never happen even if the theory behind it isn't wrong—Nash equilibria can remain and not spontaneously turn into Pareto optima even when there exists a Nash equilibrium that is a Pareto improvement over the current equilibrium.

There haven't been that many large scale attempts to revolutionize society in history and some have worked just the way some people expected without hindsight. So it is in fact possible to predict the outcomes of a large scale system through causal reasoning and to have a reasonable idea of their desirability.

> There haven't been that many large scale attempts to revolutionize society in history and some have worked just the way some people expected without hindsight.

Which ones worked "just the way some people expected without hindsight"?