[EM] decision process design: wealth tax

[James Green-Armytage]

Dear election methods fans:

	I'm interested in enacting a progressive wealth tax in place of the
income tax, property tax, capital gains tax, etc. 
	Although this is a bit beside the point, I'll give a quick argument for
it here. Imagine that person A has $1 billion in the bank, and makes
$200,000 in 2006, whereas person B is absolutely dirt poor and then by
some wonderful stroke of luck or effort earns $200,000 one time and one
time only. My opinion is that person A should pay more tax.
	Anyway, since this is a voting systems list, I'll not focus on that, but
rather focus on how a public vote could shape a wealth tax, i.e. determine
how progressive it is, or whether it is progressive at all. This fits into
a broader interest of mine, which is reducing real social issues to
manageable public votes, preferably ones where Condorcet cycles are not
possible, e.g. a series of yes or no choices or one dimensional spectra.
	Here is the wealth tax decision process that I have in mind: The basic
idea is to use a series of brackets, as we do with the income tax, to have
people vote on the rates for the different brackets, and to base the tax
rates on the median vote for each bracket. If the revenue generated by the
tax is either too great or too small to meet the target revenue, we can
multiply all rates by a scalar (that is, if you multiply the top bracket
rate by 1.3, you have to multiply all the other rates by 1.3), but this
can be decided by a separate process. In other words, this process
determines the "shape" rather than the "size" of the tax.
	To be a bit more specific about the brackets, let me give a simple
example: The lowest bracket is from $0-$50,000, and the next lowest
bracket is from $50,000-$100,000. The tax rate for the lowest bracket is
0%, and the rate for the next lowest bracket is 1%. If my wealth totals
$70,000, then I will owe ($50,000)*0 + ($70,000-$50,000)*0.01 = $200.
	To be a bit more specific about the definition of a wealth tax, assume
that a person's "wealth" is measured as the total of their money holdings
and the approximate value of their non-money property (stocks, house, car,
etc.). It is understood that no one will be able to produce an exact
figure for their total wealth, but as long as what you report is "close
enough", you pass.
	Note that the boundaries of the different brackets should be adjusted for
inflation over time.
	Here's one example of how the brackets might be constructed, and a series
of rates that might be reasonable. I personally like the general
progressivity of these rates, but to be honest I don't know whether it
would generate too much or too little revenue; this is intended as a very
rough sketch rather than a serious proposal.

wealth	tax per year
0-100k	0%
100k-300k	0.25%
300k-600k	0.5%
600k-1m	0.75%
1m-3m	1%
3m-6m	2%
6m-10m	3%
10m-15m	4%
15m-30m	5%
30m-50m	6%
50m-100m	7%
100m-500m	8%
500m-1b	9%
1b+		10%
(k=1,000; m=1,000,000; b=1,000,000,000)

	I have no idea whether this level of progressivity is anywhere close to
the kind of response that you would get if a representative subset of the
population participated in the vote, but I would like to find out, if
possible. In particular, I'm interested in the interplay that would occur
between different motives such as self-interest, altruism, and ideas of
fairness. If we inhabited a simplified world where the tax structure did
not affect the amount of wealth that people chose to earn, an
self-interest was the only motive, then you might expect all brackets
above the median wealth level to be voted a tax rate of 100%. However, I
would not necessarily expect to see this in practice, in part because of
people's idea of fairness, in part so that the very wealthy wouldn't be
driven out of the country, or something like that.

my best,
James Green-Armytage