Remarks by Dr. N. Gregory Mankiw Chairman Council of Economic Advisers at the National Bureau of Economic Research Tax Policy and the Economy Meeting National Press Club November 4, 2003.
My remarks today will focus on the estate tax, with particular attention to its
incidence and its revenue effects. The estate tax is a timely issue. As you know,
President Bush won repeal of the estate tax as part of the 2001 tax cut. Repeal does not
take full effect until 2010 and is then scheduled to sunset at the end of that year, along
with the rest of the 2001 tax cut. The President has consistently advocated making repeal
permanent. This proposal won majority support in both houses of Congress last year, but
failed to win the necessary 60 votes in the Senate. The future of the estate tax is likely to
be a major topic of debate during the next few years.
The debate about the estate tax illustrates some general economic principles that
are relevant to many areas of tax policy. I will focus on two in particular. The first is the
distributional impact of taxes—who wins and who loses. The second concerns the effects
of tax changes on government revenue. I believe that, along both dimensions, public
discussion and official analysis of the estate tax are often fundamentally flawed.
Many of the problems arise from a fact about which all economists agree—taxes
affect how people behave. These behavioral responses have implications for how the
burden of the tax is distributed and for the revenue effects of a tax change. These
implications, however, are often ignored.
Let me begin with the issue of incidence. Defenders of the estate tax often claim
that it is a highly progressive tax. It is certainly the case that the tax is levied only on the
largest 2 percent of estates. From this fact, defenders of the tax claim that the burden of
the tax falls only on the richest 2 percent of Americans.
If you look more closely at this argument, you will see that it rests on a particular,
and I believe untenable, theory of tax incidence. This argument is coherent only under
the assumption that the burden of the estate tax falls entirely on the decedent. In other
words, it makes sense if the rich dead guy takes the tax hit. When estate taxes are
included in official distributional analyses, this is precisely what is assumed. This
assumption is easy and natural for tax analysts because, by law, the decedent’s estate is
responsible for paying the tax.
This approach, however, reflects a theory of incidence that the economics
profession has repudiated for at least a century. Official distributional analyses reject it
in many other settings. We know that taxes do not stay where Congress puts them.
In technical terms, the economic incidence of a tax does not match the statutory
incidence. Who bears the burden of a tax depends on the underlying economic
fundamentals, not on who writes the check to the government. When the government
taxes car companies, for instance, the burden falls not only on the company shareholders.
It also falls on car buyers and car workers, and most likely on consumers and workers in
other industries as well.
Incidence studies recognize this principle in many ways. The economists who
prepare distributional tables wisely pay no heed to Congress’ statements about who pays
excise taxes. Instead, they always assume that the burden of excise taxes falls on
consumers. Similarly, these economists ignore Congress’ declaration that the Social
Security-Medicare payroll tax burden is split equally between employers and employees.
Instead, they generally assume that the burden is borne entirely by the employees.
Although these conclusions about where the tax burden falls may not be exactly right,
they are reasonable conjectures based on solid economics.
Unfortunately, the same insights have not been applied to the estate tax. Under
what circumstances would the estate tax actually fall only on the decedent? That would
happen if the tax prompted the decedent to reduce his consumption during his lifetime, so
that he could satisfy the tax obligation without diminishing the after-tax bequests left to
his loved ones. In other words, the estate tax would have to reduce lifetime consumption
and promote estate accumulation.
Simply stating this assumption casts doubt upon it. A good rule of thumb is that
when you tax an activity, you get less of it. The estate tax makes estate building less
attractive and probably reduces the size of bequests. Empirical research confirms that, in
fact, the estate tax reduces the amount that decedents accumulate and pass on to their
heirs. As a first approximation, it would make more sense to distribute the burden of the
tax to the estate’s beneficiaries rather than to the decedent.
What would happen if we allocated the estate tax burden to heirs rather than
decedents? At first blush, one might think that it would not make much difference. After
all, are not the children of rich people rich?
It turns out that the answer is “not always.” A number of economists have taken a
careful look at this difficult question, using a variety of data sets and methodological
approaches. Their results are roughly similar. The correlation between the lifetime
earnings of successive generations is around 0.4 or 0.5. Even adding in inheritances, the
figure increases to only about 0.7. This is nowhere near a perfect correlation. And the
correlation is far smaller when we look at the link between grandparents and
grandchildren, and probably smaller still if we consider nephews, nieces, and other
possible heirs. The bottom line is, once we move away from the standard assumption
that the entire burden falls on the decedents, the tax appears much less progressive than
one might have guessed.
But this departure from the standard assumption, as dramatic as its implications
may be, is only the beginning of the story. The estate tax is a tax on capital. As such,
one would naturally expect it to discourage capital accumulation. Now, put this together
with the fact that a smaller capital stock reduces productivity and labor income
throughout the economy and the implication is clear: the repeal of the estate tax would
stimulate growth and raise incomes for everyone, even those who never receive a
The average worker has little reason to know that his weekly paycheck is smaller
because of the existence of the estate tax. He may never realize that he bears part of the
burden of the estate tax. But these subtle, indirect effects are at the heart of how
economies work. Giving this worker (or, at least, his elected representatives) that
information is an important part of an economist’s professional responsibility.
The flaws in the distributional analysis of the estate tax also apply to analyses of
capital income taxation in general, including the corporate income tax and the taxation of
capital gains and dividends under the individual income tax. The burden of these taxes is
almost always assumed to fall on the owners of capital. The burden shifted to labor is
Data limitations and resource constraints make applied incidence analysis very
difficult, and I cast no aspersions on those doing this important task. But I want to
emphasize that the distribution tables produced by tax analysts and reproduced in
newspapers are often based on faulty economics. In particular, the standard short-cuts
systematically bias the analysis by making capital taxation more progressive in
appearance than it is in reality.
Now, let me turn to the revenue effects of permanently repealing the estate tax.
This is the second aspect of discussions of the estate tax that is often at odds with sound
One argument for keeping the estate tax is that, if all else fails, it raises revenue.
This is not much of an argument for keeping an economically harmful tax when there are
better ways to raise revenue. Besides, estate and gift tax collections were only 1.4
percent of federal revenue in fiscal 2001.
But even that small number is likely to overestimate the tax’s net contribution to
total federal revenues. A more accurate number might be zero or negative.
Let me be perfectly clear here. Tax reductions are generally not self-financing.
In the vast majority of cases, the behavioral responses to changes in tax rates just are not
high enough to yield that result. But estate tax repeal could be an exception to that
general rule. It is conceivable that repeal could actually increase total federal revenue.
Or, if not, the revenue loss could be small.
The estate tax encourages people to take avoidance actions, such as making gifts
to their children. Since their children are almost always in lower tax brackets, these gifts
reduce income tax collections. Repealing the estate tax would remove the incentive for
such gifts and would thereby boost income tax revenues.
More important, however, is the effect I emphasized above – the depressing effect
of the estate tax on capital accumulation. As a matter of policy, revenue estimates
exclude these macroeconomic effects. I understand why such effects would be difficult
to routinely incorporate in revenue estimates. Nevertheless, their omission can seriously
skew the analysis of changes such as estate tax repeal.
It is hard to estimate how much capital we do not have thanks to the estate tax.
But the issue is of paramount importance. Research has established that intergenerational
transfers accounts for a large fraction of capital accumulation in the U.S. economy. It is
very possible that the depressing effect of the estate tax on economic growth offsets a
significant part of the revenue attributed to it.
This point also applies more broadly, to capital taxes in general. Any increase in
the tax burden on capital income is likely to reduce the amount of capital available to be
taxed, particularly in the long run. It will also reduce the amount of labor income to be
taxed, since capital accumulation is one determinant of labor productivity and wages.
Taking this feedback into account could greatly diminish the revenue attributed to capital
To provide accurate information to policy makers about the revenue effects, we
must find some way to incorporate these effects. The Joint Tax Committee and the
Congressional Budget Office have recently taken steps in this direction. The exercise is
called “dynamic scoring,” although “realistic scoring” would be a more accurate term.
Fairness: Back to Basics
Let us go back to where this discussion started—the notion that the estate tax is a
good tax because it is a fair tax. I’ve argued that when we make realistic assumptions
about behavioral responses, this assertion does not hold water. But let me go beyond
these economic arguments. I believe that, based on the principle of horizontal equity, the
tax is not at all a fair tax.
Let me explain. Consider the story of twin brothers – Spendthrift Sam and Frugal
Frank. Each starts a dot-com after college and sells the business a few years later,
accumulating a $10 million nest egg. Sam then lives the high life, enjoying expensive
vacations and throwing lavish parties. Frank, meanwhile, lives more modestly. He keeps
his fortune invested in the economy, where it finances capital accumulation, new
technologies, and economic growth. He wants to leave most of his money to his children,
grandchildren, nephews, and nieces.
Now ask yourself: Which millionaire should pay higher taxes? It seems natural
that they should face the same tax burden. They both started life with the same resources.
What notion of fairness suggests that they should face different tax burdens? What
principle of social justice says that Frank should be penalized for his frugality? None
that I know of.
Several years ago the book The Millionaire Next Door made bestseller lists with
the message that getting rich is more often the result of patience than of good luck.
Recent research suggests that this is right on the mark. Whether a person reaches old age
wealthy or penniless mostly depends on the percentage of his earnings he saved - not on
the total amount he made in his lifetime. This means that most of the burden of the estate
tax falls not on those who have been lucky throughout life but rather on those who have
been frugal. In other words, when the government taxes your estate, it is, literally, taxing
The estate tax unfairly punishes frugality, undermines economic growth, reduces
real wages, and raises little, if any, federal revenue. There are no principles of good tax
policy that support this tax, and I support the President’s call for its permanent repeal. I
hope that we can count on your support in achieving this objective.