|
Commentaries
Tax Reform
Savings,
Retirement and Social Security Reform
Contributing
Members Commentaries
|
Integrating Corporate and Shareholder Taxes
Tax Notes,
September 17, 1990
By Ernest S. Christian, Jr.
Ernest S. Christian, Jr., is a partner in the law firm
of Patton, Boggs & Blow, Washington, D.C. Mr. Christian
is a former Deputy Assistant Secretary of the Treasury (Tax
Policy). This article was derived from a letter to Kenneth
W. Gideon, Assistant Secretary of the Treasury.
Summary
In this article, Mr. Christian sets forth five criteria for
evaluating methods for fully integrating the corporate and
individual income taxes, and then examines five models of
integration in light of these criteria. The article urges
that integration be achieved in a manner that is least likely
to introduce new distortions and most likely to enhance investment
per dollar of revenue loss. To achieve these goals, an integrated
system should operate to increase the amount of after-tax
earnings available for retention and reinvestment in the business
that generated the income.
The article observes that a corporate tax rate cut is implicit
in the basic concept of integration at today's tax rates.
Each of the integration models considered in the article either
implicitly or explicitly reduces the corporate tax rate to
28 percent and excludes dividends from income. The article
concludes that the essence of integration can be achieved
by simply reducing the corporate tax rate to 28 percent and
excluding dividends from income. The 1986 Act's elimination
of the principal structural barrier to integration -- highly
progressive individual income tax rates -- makes this type
of integration possible now.
Table of Contents
- Introduction
- New Ability to Achieve Integration
- The Legacy of the 1986 Act
- Organization and Assumptions
- Criteria for an Integrated System
- The Importance of the Cash Flow Test
- The Partnership Model of Full Integration
- The Credit Model of Full Integration
- The Basis Model of Full Integration
- The Average Rate (AR) Model
- The Dividend Exclusion Model of Full Integration
- The Issue of Pension Funds That Own Stock
- Issues Not Discussed
- Conclusion
Introduction
Nearly all methods of integration will, one way or another,
result in reducing the "corporate" rate to the top
shareholder rate and excluding dividends from tax. This paper
urges that we do it the easy way -- in the manner that is
least likely to introduce new distortions and most likely
to enhance investment per dollar of revenue loss.
New Ability to Achieve Integration
The present rate structure should permit one or more simple
methods of integration that satisfactorily accomplish the
purpose of reducing -- to the normal level paid on other income
-- the present high rate of tax on income from equity capital
invested in corporate form. Past attempts foundered on complexity
primarily because the corporate rate was lower than the top
shareholder rate and higher than the low and intermediate
shareholder rates. Simple approaches either fell far short
or violated the then-prevailing canon of progressivity.
The Legacy of the 1986 Act
The present ability to achieve tax integration may be the
most important legacy of the 1986 Act. Because the rates for
taxable shareholders are now so near to being flat and because
the corporate rate is now higher instead of lower than the
top individual rate, various flat-rate methods of integration
are the most plausible options. Most of the traditional approaches
to integration are of little practical relevance today. Indeed,
most traditional methods would introduce far more complexities
and distortions than are warranted by today's rate structure.
Because gains on stock are now fully taxed, the dimension
of the problem presented by our two-tier system has also been
expanded. Today, both retained and distributed earnings are
taxed twice. Even so, integrating the corporate and shareholder
taxes should be a far more manageable process than before.
The key is not to seek perfection.
Organization and Assumptions
First, the article states the tests for a proper form of tax
integration under any rate schedule and briefly discusses
the importance of taking into account the effect on corporate
cash flow. Then the tests are applied to several methods of
integration. To illustrate how much the situation has changed
as a result of the 1986 Act, the text begins with the traditional
Partnership Model (which is the most complex) and ends with
the Dividend Exclusion Model (which is the simplest). In between
lie several related methods, some familiar and some not so
familiar -- such as the Basis Model and the Average Rate Model.(1)
Also discussed are some issues which seem fundamental to all
methods of integration -- including gains on stock attributable
to retained earnings, the treatment of foreign shareholders,
the treatment of tax-exempt shareholders in general, and the
treatment of tax-exempt pension funds where there is a deferred
tax. For the most part, transitional issues are ignored because
the transition to a properly integrated system would be relatively
uncomplicated.
A key issue is the average shareholder rate, i.e., what would
be the aggregate amount of tax payable by shareholders as
a group if total corporate taxable income were imputed and
taxed to them in the year earned? The answer is critical to
the Average Rate Model. The average shareholder rate on fully
imputed taxable income is also important in comparing the
results of the Partnership Model to the results of various
flat-rate approaches such as the Basis Model and the Dividend
Exclusion Model.
For purposes of this analysis, the weighted average federal
income tax rate on fully imputed corporate taxable income
is guesstimated to be around 20 percent.(2)
This average rate was estimated under the assumption that
corporate taxable income imputed to tax-exempt and foreign
shareholders would be treated the same as dividends under
present law. In fact, taxable income imputed to such shareholders
may be more akin to unrelated business taxable income (or
effectively connected income in the case of foreigners) than
to dividends.
This does not suggest that tax-exempt and foreigner shareholders
actually would be required to pay a net income tax on the
amount of corporate taxable income imputed to them under the
Partnership Model. Indeed, this analysis assumes that they
would not. Therefore, the Partnership Model (and the comparable
Credit Model with refundability) would produce a larger revenue
loss than the flat-rate models (with the exception of the
Average Rate Model where the additional revenue loss goes
to enhance corporate cash flow instead of to tax-exempt and
foreign shareholders).
The analogy to unrelated business taxable income and effectively
connected income should be taken into account in evaluating
flat-rate approaches such as the Basis Model and the Dividend
Exclusion Model. For example, if a flat tax at a 28-percent
rate were collected from corporations and dividends were by
one means or another excluded from tax, that 28-percent tax
should not necessarily be compared to the suggested 20-percent
average shareholder rate on fully imputed taxable income.
Instead, the flat 28-percent tax might more appropriately
be compared to an average shareholder rate computed on the
assumption that amounts imputed to tax-exempt and foreign
shareholders were fully taxable. Such a recomputed average
rate would be close to 28 percent. Viewed in that way, the
flat-rate approaches would produce essentially the same tax
result as if the shareholders had earned the income directly.
For purposes of this analysis, the top shareholder rate is
assumed to be 28 percent. The five-percent surcharge, or bubble,
has been ignored. The surcharge probably should be eliminated
-- as indeed it would be if a flat tax at the top shareholder
rate were collected from corporations and dividends were excluded.
Even if the "bubble" still existed -- as it might
under the Partnership Model or the Credit Model -- the five-percent
surcharge serves only to recapture the benefit of the 15-percent
bracket. The maximum amount of tax payable by a shareholder
can never exceed 28 percent of his taxable income.
The analysis also assumes that the amount of income earned
in corporate form that is subject to tax under an integrated
system would not be greater than under present law. Had the
shareholders actually earned the income directly, they would
have had the benefit of tax-exempt bond interest, the excess
of accelerated over straight-line depreciation, and other
similar provisions. Therefore, corporate taxable income, not
earnings and profits, should be taxed to shareholders under
imputation methods such as the Partnership Model.
Similarly, corporate taxable income is the tax base under
the Average Rate, Basis, and Dividend Exclusion Models where
the flat tax collected from the corporation is a proxy for
the tax that would be payable by the shareholders had they
earned the income directly.
The concept of earnings and profits can, however, be used
to further assure that the shareholders are in the same position
as if they had earned the income directly. Earnings available
for distribution or retention typically will include both
taxable income (minus tax) and amounts properly not included
in taxable income. The clearest example is tax-exempt bond
interest.
The Partnership Model would, at the least, increase shareholder
stock basis by the amount of corporate taxable income. Thus,
corporate taxable income distributed as a dividend would not
be taxed to shareholders. Gains on stock attributable to retained
taxable income would be offset by basis and would also not
be taxed a second time.
The Basis Model illustrates one way to go the next step and
assure that earnings and profits are not rendered taxable.(3)
Shareholder stock basis would, in general, be further increased
by positive adjustments to earnings and profits. For example,
tax-exempt interest received by a corporation and distributed
to shareholders would be a return of basis and be nontaxable
to the shareholders just as if they had earned the interest
directly. Gains on stock attributable to retained tax-exempt
interest also would be offset by basis.
The simple Dividend Exclusion Model (and the Average Rate
Model) automatically would eliminate all tax on dividends
whether paid out of corporate taxable income or earnings and
profits. The question presented by the Basis Model is whether
it is necessary or practically feasible to add on the earnings
and profits basis adjustment so as to achieve a symmetrical
result with capital gains and dividends.
Criteria for an Integrated System
A proper form of tax integration would: (i) result in the
tax on income from equity capital invested in corporate form
being no greater than if the shareholders had earned it directly;
(ii) result in the tax on retained and distributed earnings
being the same; (iii) be designed to maximize the amount of
corporate cash flow available for retention and reinvestment;
(iv) be evenhanded between corporations that do and do not
pay dividends; and (v) not artificially influence dividend
policies or otherwise interfere with the normal relationships
between corporations and their shareholders.
Any integrated system which reduced the tax to the normal
level paid on other income -- as if the shareholders had earned
it directly -- would result in a large revenue loss. Even
so, revenue considerations did not constrain the range of
choices discussed in this paper. Instead, the analysis seeks
to identify the integration method that is most likely to
enhance economic growth per dollar of revenue loss and least
likely to disrupt capital markets.
The Importance of the Cash Flow Test
The integrated system should operate -- much in the manner
of a corporate rate cut -- to increase the amount of after-tax
earnings available for retention and reinvestment in the business
which produced the earnings to start with. Whether all or
part of the corporation's after-tax earnings actually are
retained for reinvestment in the business or are distributed
to be consumed or reinvested elsewhere by the shareholders
should be left to the corporation and its shareholders. Increases
or decreases in the amount distributed should not be dictated
by the mechanical operation of the particular method of integration.
From a capital formation standpoint, many observers would
argue for enhancing retained earnings. Any integrated system
that reduces the present high rate of tax on income earned
in corporate form will leave additional dollars in the private
sector that are available for savings, consumption, or some
combination of both. To the extent that these additional dollars
are saved instead of being consumed and result in greater
capital investment, economic growth will ensue. Retained earnings
are by definition savings that are concentrated directly and
immediately at the point where most business capital investment
is made, where nearly all research and development occurs,
and where essentially all competition in international markets
is undertaken. Therefore, the combination of greater retained
earnings and a lower cost of capital -- which is presently
extraordinarily high in part because of the high rate of tax
on income earned in corporate form -- should result in the
greatest degree of economic growth per dollar of Treasury
revenue loss.(4)
From a structural standpoint, the present rate schedule leads
naturally -- perhaps inevitably -- to an integration method
which operates much in the manner of a corporate rate cut.
The function of an "integrated" tax is, after all,
to collect tax at the shareholders' rates in the year income
is earned in corporate form.
When, as now, the corporate rate (34 percent) is higher than
the top shareholder rate (28 percent), the first result of
any properly integrated system is automatically to reduce
the maximum tax on income earned in corporate form to a level
which is less than the corporate rate that applied in the
nonintegrated system. The administratively feasible way to
assure current collection is to require the corporation to
pay the tax under one of the flat-rate methods -- just as
it previously paid the separate corporate tax. The difference
is the rate of tax (28 percent instead of 34 percent). Provided
the income is not taxed again when dividends are paid, no
further adjustment would be necessary to assure that the corporation
has not underpaid any shareholder's tax.
Overpayments of tax might be said to occur with respect to
shareholders who would not be taxed at a 28-percent rate if
they had earned the income directly. Under the present rate
schedule, any such overpayments are an imperfection that probably
can be ignored. The average rate of tax payable by fully taxable
shareholders on fully imputed income probably is pretty close
to 28 percent anyway. It is primarily tax-exempt and foreign
shareholders who might be said to be overpaid. While it would
hardly seem necessary, any overpayments by 15- percent bracket
shareholders could be handled under a flat-rate system. For
example, they could be allowed a refundable credit when dividends
are paid.(5)
Even though it is correct to say that a flat-rate 28-percent
tax collected at the corporate level would result in little
overpayment for fully taxable shareholders, there would be,
in the aggregate, a substantial "overpayment" of
tax if it is assumed that income earned in corporate form
by foreigners and tax-exempt entities should not be taxed
at all. Elimination of that overpayment obviously would result
in a substantial additional revenue loss. The magnitude of
that additional loss is indicated by comparing the 28-percent
flat tax to the suggested 20-percent average tax rate on fully
imputed corporate taxable income. But for the large amount
of stock owned by tax- exempt pension plans, etc., the average
shareholder rate would be close to 28 percent.
The overpayment could be eliminated by a corporate-oriented
approach such as the Average Rate Model, where the corporation
would pay tax at the 20-percent (or whatever) average shareholder
rate. The additional revenue loss would be channeled into
corporate cash flow available for reinvestment or for proportionate
distribution to shareholders of all categories.
The alternative is to resort to one of the traditional shareholder-oriented
methods such as the Partnership Model or the Credit Model,
where the additional revenue loss would go directly to the
tax-exempt and foreign shareholders. Indeed, the only discrete
purpose served by the Partnership Model and the Credit Model
is to funnel large amounts of cash to tax-exempt and foreign
shareholders at the expense of both Treasury and corporate
cash flow. All other purposes served by the Partnership Model
and the Credit Model could better and more simply be accomplished
by other methods.
As a general proposition, the Partnership Model and the Credit
Model would influence dividend payments, provide the cash
flow equivalent of a greater corporate rate cut to one corporation
than to another, and result in the total amount of tax on
one incorporated business being greatly different from the
tax on another otherwise identical incorporated business.
The Partnership Model of Full Integration
A corporation's taxable income would be imputed and taxed
to the shareholders in the year earned. Shareholder stock
basis would be increased by the amount of the corporation's
taxable and tax-exempt income. Dividends paid out of taxable
and tax-exempt income would be a return of basis and excluded
from further tax. Gains on stock attributable to retained
taxable or tax-exempt income would be offset by basis and
excluded from further tax. To put shareholders entirely in
the same position as if they had earned the income directly,
stock basis also should be adjusted for the tax deferrals
included in earnings and profits. See discussion of the Basis
Model. The idea of adjusting stock basis for corporate debt,
and the alternative of providing for a negative basis, both
seem to go beyond reason -- particularly in the case of a
method of integration which is only of academic interest anyway.
Although corporations would have to pay no tax as such --
and in that sense the result is the same as repeal of the
corporate tax -- they would have to pay an additional "tax-compensation"
dividend sufficient to permit the top bracket shareholders
to pay their tax on imputed retained income. Because corporations
cannot pay disproportionate per share dividends to shareholders
of the same class, they would have to pay a combined regular
and tax-compensation dividend equal to 28 percent of imputed
taxable income -- not ony to 28-percent bracket shareholders,
but to tax-exempt and foreign shareholders as well.(6)
Because a corporation which now pays no dividend would have
to pay a tax-compensation dividend equal to 28 percent of
its taxable income, the effect of the Partnership Model on
its pre-dividend cash flow would be exactly the same as if
the corporate tax rate had been reduced from 34 to 28 percent.
A corporation which now is paying a dividend would have to
pay a proportionately smaller tax-compensation dividend, would
experience a proportionately greater increase in its pre-dividend
cash flow and, therefore, would receive the equivalent of
a larger corporate rate cut.(7) A closely held corporation
entirely owned by tax-exempt, foreign, and low rate domestic
shareholders would have to pay little or no tax-compensation
dividend. The effect on it would be essentially the same as
outright repeal of the present 34-percent corporate tax.
A more typical situation might be where 50 percent of the
corporation's stock is owned by a combination of foreign and
tax- exempt shareholders. Wholly apart from whatever effect
the Partnership Model might finally have on that corporations'
retained cash flow, it is clear that the income from its business
would be taxed at a much lower rate than the income from some
otherwise identical business where there were very few tax-exempt
or foreign shareholders.
Therefore, not only is the Partnership Model the most complex
method; it also appears to fail nearly all the tests for a
proper form of integration.
The Credit Model of Full Integration
The traditional structure, often referred to as the partnership-
withholding method, is as follows: (i) the corporation pays
tax at the top shareholder rate (28 percent) in the year taxable
income is earned; and (ii) the corporation's taxable income
is imputed and taxed to the shareholders who are allowed refundable
credits for the tax paid by the corporation on their behalf.
As in the case of the Partnership Model, shareholder stock
basis should at least be adjusted for imputed taxable income
(in this case 72 percent of taxable income or taxable income
minus credits) and, more correctly, should be adjusted for
earnings and profits.
The redundancy in allowing both credits to eliminate shareholder
tax on imputed income and a basis adjustment to eliminate
shareholder tax on dividends (and gains attributable to retained
earnings) emphasizes that the Credit Model serves primarily
to provide large refunds to tax-exempt and foreign shareholders.
An amount equal to the disproportionately large tax-compensation
dividends paid by corporations to tax-exempt and foreign shareholders
under the Partnership Model would be paid as tax to Treasury,
which would then refund that amount to tax-exempt and foreign
shareholders. (Refunds also would be made to 15-percent bracket
shareholders although, under today's rate schedule, that seems
more coincidental than anything else.)
Compared to the Partnership Model, the only purpose served
by the Credit Model is to permit Treasury to audit and collect
a single tax at the corporate level./8/ Otherwise, the Credit
Model produces the same overall results as the Partnership
Model -- including the same lack of evenhandedness among corporations
that do and do not pay dividends.(9)
Even though for the purpose of illustration, the Credit Model
allows full refunds to all shareholders, it seems improbable
that refunds actually would be allowed to tax-exempt and foreign
shareholders. It is one thing to say that such shareholders
would not themselves be required to pay a regular tax on income
attributed to them under the Partnership Model. Political
and treaty considerations probably would prevent it. Giving
cash refunds to tax- exempt and foreign shareholders for taxes
paid at the corporate level is quite another matter.(10)
If the credits were not refundable, the Credit Model would
serve no purpose that could not better be accomplished by
other means. The Credit Model is essentially an anachronism
that should be disregarded. If adopted, it would fail nearly
all the tests for a proper form of integration.
The Basis Model of Full Integration
Although there are several variations on the same theme, the
Basis Model would be structured as follows. Corporations would
pay a flat 28-percent tax on taxable
income in the year earned. No amount of corporate earnings
would be imputed and taxed to shareholders. Shareholder stock
basis, however, would be increased by the amount of the corporation's
earnings and profits (i.e., taxable income plus positive adjustments
to earnings and profits in excess of taxable income minus
tax paid by the corporation on taxable income).
Dividends would be a return of basis and effectively excluded
from tax./11/ Gains on stock attributable to retained earnings
also would be a return of basis. Thus, there would be no second
tax at the shareholder level under either of the two alternative
means by which equity holders can realize a return on their
investment.
Although the basis adjustment effectively excludes dividends
from tax, the basis adjustment does not produce the same result
as exempting from tax all gains on stock -- and is not intended
to. The proposition is that a dollar of earnings retained
and reinvested by a corporation is equal to the discounted
present value of the future income stream that the corporation
expects to be produced by that dollar. Therefore, only gains
on stock attributable to retained earnings would be excluded
from tax as each increment of that income stream is, when
realized by the corporation, added to the basis of stock.
The taxable gains realized by shareholders would consist of
(i) built-in gains on stock purchased prior to the effective
date; (ii) inflation gains; and (iii) gains attributable to
real appreciation in the value of assets which has not been
realized at the corporate level.
The effect of the basis adjustment with respect to both dividends
and capital gains generally is to produce the correct results
under an integrated tax system: (i) the amount of income earned
in corporate form that is subject to tax is no greater than
if the shareholders had earned it directly; and (ii) the rate
of tax is no higher than if the shareholders had earned the
income directly. On the assumption that corporate earnings
allocable to shares owned by tax-exempt or foreign shareholders
is closely akin to unrelated business taxable income or effectively
connected income, the only real departure from the norm is
in the case of 15-percent bracket shareholders. If necessary,
that defect could be handled easily within a system that already
undertakes to allocate earnings and profits on a per-share
basis.
But for the complexity of allocating basis among the constantly
changing group of shareholders in a publicly held corporation
and concerns about the potential effects on markets of the
new "changing basis" factor, the Basis Model has
great merit.
While it would never be simple, the process of allocating
basis to shareholders may be manageable. Drawing upon experience
with master limited partnerships, it may be that the Treasury
study will show that large corporations could handle a system
similar to the Basis Model -- much in the same way they now
handle information reporting on dividends. Although the scope
of the task is larger and more complex in the case of big
publicly held corporations, they are probably more able to
handle it than smaller companies. If imputation of income
is feasible, so is imputation of basis.
The larger difficulty with the Basis Model may be the uncertainty
and potential for disruption, at least during a transition
period. A shareholder's basis would be constantly fluctuating
in amounts that could not be determined exactly until months
after a sale of stock. Even if corporations published quarterly
"capital changes" basis adjustment reports, basis
would be subject to substantial redetermination and revision
after year-end. Whether speculation about large swings in
basis adjustments would be a major new factor in evaluating
stocks is a question left to experts. Perhaps there would
not be a problem. To the extent that earnings and profits
may correspond to financial earnings, the amount of any tentative
basis adjustment as of any quarterly date (and any predicted
changes therein) may be merely the flip side of tentative
earnings as of that quarterly date (and any predicted changes
therein).
The Average Rate (AR) Model
The AR Model imposes a flat-rate tax at the corporate level
and excludes dividends from tax at the shareholder level.
Corporations would pay tax at the average shareholder rate
determined by Treasury for shareholders in the aggregate,
assuming full imputation of corporate taxable income (i.e.,
total shareholder tax payable expressed as a percentage of
total corporate taxable income).
Primarily, the AR Model illustrates that the large benefits
to tax-exempt and foreign shareholders under traditional methods
can be redirected to corporate cash flow. The AR Model also
illustrates the paradigm for integration -- a system which
produces the same overall tax and corporate cash flow result
as if the corporation's taxable income had been distributed
and taxed to the shareholders and, net of tax and any actual
dividends, reinvested in the corporation.
On an economy-wide basis, taking into account tax-exempt and
all other shareholders, the average shareholder rate on fully
imputed corporate taxable income probably is about the same
as the average rate on interest income and noncorporate business
income, and higher than the average rate on labor income.
The average shareholder rate, however, would be less than
the top rate on interest income, noncorporate business income,
and labor income. Although the top rate on retained corporate
earnings was for many years prior to the 1986 Act lower than
the top rate on other income, the combined corporate/shareholder
tax on distributed earnings was much higher than the top rate
on other income.
Given the potential distortions involved (and the obvious
political problem in taxing corporate earnings at an average
rate which is, by definition, lower than the top rate), the
AR Model, as such, probably is not a practical approach. Taxing
income from corporate equity capital at a rate that is substantially
lower than the highest marginal rate of tax on interest could
distort capital markets.
Setting aside politics and the large revenue loss, the AR
Model might be a more reasonable option if an "equalizing"
tax were imposed on dividends paid to top bracket shareholders
-- in order that the top bracket tax on income earned in corporate
form would be the same as the top bracket tax on interest
income, labor income, etc.
The Dividend Exclusion Model of Full Integration
All the foregoing discussion, including the similarities,
differences, and imperfections of the various models, leads
to the conclusion that the essence of integration today can
be accomplished by reducing the corporate rate to 28 percent
and excluding dividends from tax.
Indeed, a lower "corporate rate" probably is the
inevitable result if an integrated system is going to confront
and resolve the problem that has always been at the heart
of the matter -- how to collect a tax at the shareholders'
rates in the year income is earned despite the facts that
(i) all corporations can and must accumulate substantial portions
of their earnings; and (ii) shareholders hardly can be expected
to pay tax on income they have not received.
All the models implicitly or explicitly reduce the "corporate
rate" to 28 percent and exclude dividends. Whether one
goes further and engrafts onto dividend exclusion the earnings
and profits basis adjustment under the Basis Model is an option.
Whether one undertakes to compensate 15-percent bracket shareholders
through a refund when dividends are paid also is an option.
The only thing wrong with outright dividend exclusion is the
absence of any handy, precisely accurate way to prevent a
second tax on retained earnings when stock is sold. Perhaps
some way of approximating the results of a basis adjustment
could be devised.
The Issue of Pension Funds That Own Stock
Up to this point, pension funds have been lumped in with tax-
exempt charities, etc. The tax benefit accorded pension funds
is only a deferral of tax insofar as concerns the annuitants
and other beneficiaries of such funds. Ultimately, dividends
and gains on stock received by such funds will be taxed, although
at some real rate which is -- after taking into account the
deferral -- less than the present statutory rates of 15 and
28 percent.
Given the fact that pension funds own such large amounts of
stock, their presence will affect tax integration in numerous
ways. Some results are fairly obvious. If dividends are by
one means or another excluded from tax, dividends allocated
to an employee's account should be treated as a previously
taxed employee contribution and not taxed again when distributed
by the fund. If corporate taxable income were imputed and
taxed to pension plans (without refunds) or dividends and
other investment income of pension plans were taxed, a similar
"basis" adjustment would occur.
Issues Not Discussed
Several issues have not been addressed: debt versus equity,
foreign income and taxes, or the conundrum of the corporate
alternative minimum tax.
The purpose of integration is not to make debt the same as
equity -- they are different.(12) Eliminating the present
large penalty on the income from equity capital will eliminate
the artificial bias in favor of corporate debt that arises
from the two- tier tax. Nothing else can or should be done.
Even if all businesses were conducted in proprietorship form,
the deduction for interest paid still would be significant
and necessary. Allowing corporations to deduct interest as
well as dividends has a certain superficial symmetry about
it, but fails nearly all tests for a proper system of integration.
Excluding dividends from tax seems a far better approach.
Whether dividend exclusion would result in major shifts between
debt holders and equity holders and/or cause various discontinuities
is another question left to experts.
Insofar as foreign-source income and tax credits are concerned,
the dividend exclusion or some version thereof presents the
minimum amount of difficulty. Credit for foreign taxes paid
would be allowed at the entity level, just as it is today.
The Treasury study of integration, of course, would provide
the opportunity for a proper sorting out of how we are going
to tax foreign-source income in a global economy. Many of
the concepts reflected in the foreign income rules today appear
to be as much of an anachronism as a nonintegrated set of
corporate and shareholder taxes.
The AMT is a puzzle. It seems to be a double tax that would
apply to some income earned in corporate form and not to other
such income. Its primary function would continue to be to
add on tax when profits decline relative to capital investment
or capital investment increases relative to profits. Given
the present corporate taxable income base after the 1986 Act,
the AMT serves no positive purpose anyway and should be discarded.
If political thinking ever catches up sufficiently with good
economics to permit an integrated tax to be enacted, the AMT
probably also could be dispensed with.
Conclusion
It is certainly no newly discovered secret that the equivalent
of a corporate rate cut is implicit in applying the basic
concept of integration to today's rate schedule. It is also
obvious that the 1986 Act eliminated the principal structural
barrier to integration - - the need to deal with highly progressive
individual rates that exceed the corporate rate. The ability
to achieve an integrated tax now is a reality.
Because elimination of the double tax on distributed and retained
earnings need no longer be such a complicated task, those
who earlier worked on integration in the context of the classical
corporate tax have to unlearn a great deal. On the other hand,
many observers in the business community and elsewhere continue
to view integration primarily in terms of past experience
-- overly complex systems that fail nearly all the tests for
integration mentioned above. Others probably assume that the
normal consequences of applying integration concepts to the
present rate structure never will be permitted to occur, i.e.,
that integration would be accompanied by a combination of
rate increases and base broadening that would tend to reintroduce
many of the complexities, distortions, and controversies associated
with integration in the past. An emasculated version that
perpetuated the present high level of tax on income earned
in corporate form probably would not be worth the trouble
and turmoil involved.
Still others may foresee attempts to achieve total symmetry
between debt and equity financing, to impose some concept
of economic income for tax purposes, or to address other perceived
imperfections that seem naturally to be brought into focus
by the fundamental kind of thinking involved in an integration
study.
Very likely the Treasury study will contain a sophisticated
analysis of many of the collateral issues raised by integration
-- including the deficiencies of the present foreign-source
income rules.
We can hope that the Treasury study also will cut through
to the bottom line and fully focus everyone's attention on
the real point -- the primary purpose of integration now can
be readily and simply achieved. Whether we can afford the
revenue loss or can find some substitute revenue source that
does not distort and substantially defeat the process of integration
is a separate question. Curing any additional imperfections
in present law also is a separate matter that can come after
integration. The important thing is the first step. At this
particular point in time, we have the unique opportunity.
FOOTNOTES
1. Only "full integration" or what seems to be an
appropriately expanded definition of that term is discussed.
Under today's rate schedule, various versions of the dividend
exclusion approach to integration would come very close to
producing the same overall tax result as if the shareholders
had earned the income directly. Traditional methods of "partial"
integration such as dividend deductibility and the gross-up/credit
method are inherently deficient under the tests applied here,
and are mentioned only in passing. The same is true of the
OECD variations discussed by Sijbren Cnossen in 1984 (International
Bureau of Fiscal Documentation Bulletin, Vol. 38, No. 11).
2. Extrapolations from various data sources seem to indicate
that somewhere around 70 percent of dividends are received,
directly or indirectly, by fully taxable persons or corporations.
Calculations for this paper assume that 75 percent of imputed
corporate taxable income ultimately would go to fully taxable
persons and on the average be taxed at pretty close to a 28-percent
rate.
3. If applied under the Partnership Model, section 705 also
would increase partners' (shareholders') basis by the amount
of tax- exempt interest received by the corporation. The Basis
Model would further increase shareholder stock basis by other
positive adjustments to earnings and profits, including tax
deferrals.
4. The rate of tax as well as the rate of depreciation and
other factors are taken into account in measuring the impact
of taxes on the cost of capital. This is discussed by Don
Fullerton and others in a paper at p. 173 of Compendium of
Tax Research 1987, Office of Tax Analysis, Department of the
Treasury. The cost of capital in the U.S. is higher than desirable
and, at least in the case of most capital equipment, is higher
than before the 1986 Act. See, for example, Christian and
Raboy, Increase in the Cost of Capital: Quantitative and Other
Results of the Tax Reform Act of 1986, The Tax Foundation,
1987.
5. Even though actually excluded from income under the Dividend
Exclusion Model, dividends could be grossed-up and presumed
to be taxable for the purpose of determining the amount of
the refund.
6. Essentially the same effect as a disproportionate dividend
could be achieved if the corporation knew each shareholder's
top rate. Under the hypothetical Variable Rate Method, corporate
taxable income would be allocated to shareholders. The corporation
would pay tax on each particular shareholder's portion at
a variable rate determined as follows: (The Shareholder's
Tax Rate) x (Undistributed Taxable Income/Total Taxable Income).
The entire amount of taxable income allocated to a shareholder
would be imputed and taxed to the shareholder who would be
allowed a credit for the tax paid by the corporation. If the
entire amount of taxable income allocated and imputed to the
shareholder had been retained by the corporation, the credit
always would equal the shareholder's tax. If a portion had
been distributed, the credit would equal the shareholder's
tax on the undistributed amount. The shareholder would pay
the tax on the distributed amount.
7. A dividend-paying corporation in effect gets credit against
its tax-compensation dividend for the tax already being paid
by its top bracket shareholders on regular dividends.
8. Such a system still would be complex. As in the case of
the Partnership Model, the complexities arise primarily in
the case of large publicly held corporations where there is
a constantly changing group of shareholders, where there often
are several classes of stock, and where there may be several
tiers of domestic and foreign corporations involved.
9. Under the Credit Model, a corporation which is presently
paying a dividend could reduce the amount of its dividend
to reflect the value of the shareholder credit to the top
bracket shareholders, and thereby obtain a larger cash flow
benefit and the equivalent of a larger corporate rate cut.
Such interference with dividend policies is common to all
methods of integration which use refundable credits.
10. Full refunds could be allowed to 15-percent bracket shareholders.
Tax-exempt and foreign shareholders could be allowed partial
refunds, e.g., a refund based on the assumption that corporate
taxable income imputed to them is taxable at the suggested
average shareholder rate of 20 percent. Alternately, the revenue
loss associated with those partial refunds could be channeled
into corporate cash flow by a version of the Variable Rate
Method. With little or no change in present law, corporations
could determine which of their shareholders are tax-exempt
or foreign. Corporations could pay tax at a 20-percent rate
on the portion of taxable income allocable to shares owned
by tax-exempts and foreigners. Such shareholders, of course,
would become valuable commodities. If corporations are to
any extent allowed the benefit of having such shareholders,
the benefit should be spread among all corporations, as under
the Average Rate Model.
11. Dividends in excess of cost basis and earnings and profits
basis would be taxable. As a practical matter, the result
is the same as excluding all dividends.
12. There is a suggestion to impose an entity tax on a base
that includes interest paid and to exclude both interest and
dividends from tax at the bondholder and shareholder level.
While interesting, any such approach to integration seems
outside the scope of a discussion focused primarily on enhancing
corporate cash flow.
|
