AltInst: Tax Inferred Ability to Pay

Taxes have the problem that they can distort incentives to save, work, and
buy things. If you tax particular goods, people buy those goods less.
Taxing income is better, but that still distorts incentives to save. Taxing
consumption works better still, but still distorts incentives to work (vs.
"leisure", i.e. non-"work" activity.)

If rat-racing signaling problems make people work too much, then moderate
consumption taxes might actually improve things. But if tax disincentives
to work are still a problem I have a suggestion: tax better inferred ability
to pay or consume (or earn), rather than current consumption (or earnings).

The tax that least distorts choices is a poll tax, a fixed amount per
person. The poorest are unable to pay such a tax, however, and electorates
seem to prefer to tax "the rich" more. Presumably what we really mean to do
is tax more those who are *able* to earn (or pay or consume) more. But what
we actually do is tax more those who actually *do* earn more. And this is
what creates the possibly problematic incentives to not work.

What we are doing is inferring ability to earn/consume from actual current
earnings/consumption. This creates an incentive for people who *can* earn
more to hide this ability by working less. By mimicing those who earn
less, they can pay less in taxes.

We can reduce this problem by using more information when inferring ability
to earn. For example, we can make taxes this year depend not just on income
or consumption this year, but also on income or consumption in previous
years. Taxes could depend on one's education level, IQ score, parent's
education level, or zip code of residence. If one expected that such
methods would better reveal one's ability to earn, one should be less
inclined to work less in order to try to mimic those less able to earn.

This approach doesn't really solve the problem of taxes discouraging people
early in life from doing things, like getting an education, which can
increase their ability to earn later in life. But it doesn't make these
problems any worse, and it *does* help reduce distortions in incentives to
work/earn/consume later in life.

Of course one might naturally be hesitant to trust congress to tax based on
these various characteristics, fearing they would mainly look to transfer
wealth from some groups to others, rather than inferring ability to pay. I
have some further thoughts on this topic.

In all fairness, I should point out some disturbing implications of this
proposal. First, consider the fate of those who graduate in English
literature from a good school. Even if this person chooses to live poor
while writing the great American novel, we can infer that this person
probably could have done well in some more marketable degree such as
business, and could do well in some other more boring but better paid job.
Should we tax this person based on what they could earn if they got such a
better paying job or degree?

A similar scenario is someone like me, who was once in the well-paid
computer industry, but chose to switch careers into academia, which not only
pays less but required a long break as a student where I was paid much less.
Should I have been taxed all along based on what I could have earned in the
computer industry?

How far do we want to go in taxing ability to earn, rather than earnings?

Robin Hanson rhanson@gmu.edu http://hanson.gmu.edu
Asst. Prof. Economics, George Mason University
MSN 1D3, Carow Hall, Fairfax VA 22030
703-993-2326 FAX: 703-993-2323

[To drop AltInst, tell: majordomo@cco.caltech.edu to: unsubscribe altinst]
Received on Fri Aug 27 13:55:32 1999