Tax Title of
Testimony of William W. Beach, Heritage Foundation Director
of the Center for Data Analysis, Before the
and Means Committee: Measuring the Fairness of a Tax System
June 9, 2005
the Fairness of a Tax System
Committee on Ways and Means
House of Representatives
of William W. Beach
Center for Data Analysis
of William W. Beach
The President's call for fundamental tax reform combined with this
committee's continued interest in repairing and improving our tax code
provides an enormous opportunity for expanding the social and economic
well-being of all Americans. Attaining a simple, fair, and pro-growth
tax system, however, involves disciplined thinking by policy makers
about a number of important changes to current law. I would like to draw
your attention to some of the considerations you should make when
thinking about fairness.
Let me ask you to hold a mental construction in mind for the next few
minutes. It is this: in a perfect tax world, every taxpayer at each
income level would be treated equally and the more people made in
taxable income the more tax they would pay. In this world, as well, the
taxes levied to raise the necessary revenues for needed government would
not interfere with the equal right of all taxpayers to use their labor
and capital in such a way as to achieve their economic and social goals.
That simple mental construction is crucial to the work you do day in and
day out and especially to the product we all hope will flow from this
committee once the President's Advisory Panel on Federal Tax Reform
completes its work. You need to have a model against which you can
evaluate the horizontal, vertical, and forward equity of changes to our
current tax code.
If you lived in this simple tax world, then every change to the nation's
tax law would have to pass the test: does the change treat equals
equally, does it re-enforce vertical proportionality of our tax system,
and does the change disturb the peaceful and lawful work of taxpayers
toward their economic and social goals.
Unfortunately, we do not live in this perfect world, even though this
model is a key to the survival of good policy in a political environment
awash with conflicting interests. Also unfortunately, the analytical
tools you have at your disposal for evaluating the equity elements of
proposed changes are rather crude, easily abused, and not well suited
for answering these key equity questions.
As I observed, nearly every major tax bill is challenged to prove that
it is fair. Fairness, however, can (and probably does) mean something
different to each person who thinks about it. I imagine there are
differences on this subject even on this committee. Since you cannot
entertain an infinite number of different definitions of fairness but
must instead be governed by a definition that enjoys wide support and
also allows you to make decisions on fairness, it is appropriate to
start with the question: What is tax fairness?
I think we all can agree that "tax fairness" at least means
that everyone pays their fair share. That is, the total amount of taxes
a person pays is proportional to their economic ability to pay taxes.
Thus, taxes paid are proportional to income or to consumption or to some
other measure of our use of government.
"Tax fairness" also should mean (and I think generally does
mean) that tax policy enacted today will act on each person's taxable
income so as to disadvantage no type of taxpayer over another in
achieving their economic ends. This forward equity of the tax code is a
crucial but seldom-noted fairness consideration. Lawmakers should
consider whether policy change facilitates individual economic, social,
and personal choices that set in motion a sequence of activities that
lead to goals a person sets for him or herself. For example, do tax
policy changes made today raise barriers to women re-entering the
workforce years from now after raising a family, or to immigrants
starting micro-businesses, or to retiree pursuing part-time work? Do
policy changes make it more or less difficult for young people to
achieve their goals?
Economists have developed techniques for analyzing how tax policy
changes affect taxpayers and non-taxpayers. This family of techniques,
known as distribution analysis, provides policy makers with crude but
sometimes effective tools for determining whether their policy changes
meet the tests of vertical, horizontal, and forward equity.
Distribution analysis, however, often flounders on two, central
problems: 1) what should we use to measure tax incidence against and 2)
how does the passage of time affect the distribution of taxes.
What policy makers frequently want to know is simply enough stated (how
will tax policy change affect the economic well-being of taxpayers), but
just as frequently is hard to answer. How do you measure the
relationship between tax policy and economic well-being? Because we
cannot measure all of the things that affect a taxpayer's well-being,
economists often settle on proxies for those data we cannot obtain or
activities we cannot observe. Certainly, the most common of such proxies
However, what is income? Most people think of income as the total amount
of money they make each year. But, does that amount count the income
from previously taxed income, like interest on a savings account or
dividends from an investment? Is "income" the total amount
that is spent on all goods and services and leisure? Does it include net
worth? Do we count non-cash compensation when distributing the effects
of a tax policy change?
Even if we could settle on an income concept that most analysts would
accept, how good are the income data that we would use to create
distribution tables. For example, the U.S. Census Bureau obtains a
pretty good idea about household and individual income at each decennial
census. During the intervening decade, Census regularly surveys the
population and produces updates to its decennial estimate of income
(most notably the March supplement each year) that form the basis for so
much of our economic work on taxes.
This important dataset, however, is composed of only 60,000 households
out of total population of over 110,000,000 households. While that
survey size assures statistical significance on most demographic
concepts, it produces at best a crude representation of the types and
ranges of income, particularly among high-income households.
What about distributing tax policy changes by consumption? Consumption
generally is a public act, and the very fact that consumption leaves
highly visible footprints means that using it for distributional
purposes avoids many of the definitional problems surrounding
"income." If we were to use consumption as the metric against
which to measure the fairness of a tax system, we would assume that
levels of tax payments would follow levels of consumption.
Simple enough, but what do you do with young taxpayers? They are
consuming very expensive education that they pay off over time, buying
homes to start a family that are paid through mortgages, buying their
first car, their furniture, and raising children (by itself an expensive
proposition). Short-term and long-term consumption get mixed together in
real life, which raises problems for distributional analysts.
Anyway, consumption patterns tend to follow the cycle of life: high
consumption and debt early on, followed by increases in net worth and
less consumption in middle life, which ends with low consumption and
depletion of savings over retirement. If a tax system followed that
pattern of consumption, would it be fair? Probably not.
Finally, some analysts argue that we can learn a great deal about the
fairness of a tax system by studying the actual marginal tax rates faced
by taxpayers across income. If a tax system meets the vertical and
horizontal tests for fairness, then marginal tax rates will be roughly
the same for all taxpayers in each income class.
However, our current tax policy is, if anything, one of targets, not of
equal treatment. That is, Congress has decided to use the tax system to
achieve specific social and economic goals, which has resulted in a
significant decay in vertical equity. To illustrate this point, I have
provided in my full testimony a wonderful graph prepared by Kevin
Hassett of the American Enterprise Institute, a tax economist well known
to this committee. Dr. Hassett compares the current tax code to tax law
in 1986 and 1988 and how tax policy has affected the marginal income tax
rates faced by a family of four. While this graphic shows many things,
its single most important message is how targeting tax relief has
produced significant equity distortions in the code.
As Dr. Hassett's chart shows, drifting away from a tax system governed
by principles has led to tax law that is less just. Achieving a
significantly better tax code obviously involves major legislative
efforts. Having guiding principles before the members of the House and
the Senate should help tem extract our tax code from the dramatic
difficulties into which it has fallen..