Full text of Review (Federal Reserve Bank of Dallas) : Third Quarter 2000

View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

•••••••••••••••

• •

Alan D. Viard

Gasoline and Crude Oil Prices:
Why the Asymmetry?
Stephen P. A. Brown and Mine K. Yucel

Financial Statements and Reality:
Do Troubled Banks Tell All?
Jeffery W. Gunther and Robert R. Moore

This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library (FedHistory@dal.frb.org)

[conomic ~nd
rin~nci~1 Review
Federal Reserve Bank of Dallas

Robert D. McTeer, Jr.
President and Chief Executive Officer

Helen E. Holcomb
Fi,st Vice President and
Chief Operating Office!'

Robert D. Hankins
Se11ior Vice President, Banking Supe,visio11

Harvey Rosenblum
Senior Vice President and Director ofResearch

W. Michael Cox
senior VICe President and Chief EC0110mist

Editors
Stephen P. A. Brown
Senior Economist and Assistant Vice President

Jeffery W. Gunther
Research Officer

Mark A. Wynne
Research Officer

Director of Publications
Kay Champagne
Associate Editors
Jennifer Afflerbach
Monica Reeves
Design and Production
Gene Autry
Laura J. Bell

Economic and Financial Review (ISSN 1526-3940),
published quarterly by the Federal Reserve Bank of
Dallas, presents in-depth information and analysis
on monetary, financial, banking, and other economic
policy topics Articles are developed by economists in
the Bank's Economic Research and Financial Industry
Studies departments, The views expressed are those of
the authors and do not necessarily reflect the positions
of the Federal Reserve Bank of Dallas or the Federal
Reserve System,
Articles may be reprinted on the condition that the
source is credited and the Public Affairs Department is
provided with a copy of the publication containing the
reprinted material.
Subscriptions are available free of charge. Please
direct requests for subscriptions, back issues, and
address changes to the Public Affairs Department,
Federal Reserve Bank of Dallas, P.O. Box 655906,
Dallas, TX 75265-5906; call 214-922-5254; or subscribe
via the Internet at www dallasfed.org Economic and
Financial Review and other Bank publications are
available on the Bank's web site, www dallasfed.org

Contents
The Trdn~ition to (on~umption
Tdxdtion, Pdrt 1: The ImpdCt
on [xiHin~ Cdpitdl
Alan D. Viard
Page 2

Gd~oline dnd Crude Oil Pri(e~:
Why the H~ymmetry?
Stephen P. A. Brown and Mine K. YOcel
Page 23

findncidl hdtemenB dnd Redlity:
Do Trou~led Bdnb Tell HII?
Jeffery W. Gunther and Robert R. Moore
Page 30

Alan Viard reviews the transitional impact on existing capital
from replacing the income tax with a consumption tax. This
replacement generally reduces the real value of existing capital
because it does not receive the tax relief given to new investment.
If the income and consumption taxes had stylized forms and capital were produced without adjustment costs, the proportional
decline would equal the consumption tax rate-a 25 percent tax
would uniformly reduce the value of existing capital by 25 percent.
Under more realistic assumptions, however, the actual decline
is likely to be smaller and less uniform and some types of capital
may even increase in value. The burden on owners of existing
capital is also mitigated because the tax reform increases the rate of
return they earn from reinvestment.

Many consumers complain that gasoline and crude oil prices
have an asymmetric relationship in which gasoline prices rise more
quickly when crude oil prices are rising than they fall when crude
oil prices are falling. Many also regard the asymmetry they observe
as evidence of market power in the petroleum industry. Most previous research provides econometric evidence of the asymmetry,
confirming at least part of what consumers suspect. In this article
Stephen Brown and Mine Yucel extend the inquiry by examining the market conditions underlying the asymmetric relationship
between gasoline and crude oil prices. They find the observed
asymmetry is unlikely to be the result of monopoly power. The
remaining explanations for the asymmetry suggest that policies to
prevent an asymmetric relationship between gasoline and crude oil
prices are likely to reduce economic efficiency.

Each quarter, banks file a call report, or Report of Condition
and Income, containing hundreds of accounting items pertaining to
their financial condition. This article analyzes call report revisions
to assess the extent to which regulatory exams promote accurate
data. The findings indicate banks with new or emerging difficulties
often significantly underreport these problems, intentionally or not.
In addition, the findings point to a significant role for exams in
uncovering financial problems and ensuring bank accounting statements reflect them. To the extent the loan-loss accounting in call
reports is widely used to assess loan quality, these results support
the view that exams are important in the public dissemination of
accurate information on banks' financial condition.

Many tax reform proposals call for replacing the individual and corporate income taxes
with a consumption tax. Supporters contend
that such a reform would increase national saving, boosting future income and consumption.
They also argue that consumption taxation is
fairer and simpler than income taxation. An
extensive literature discusses the wide-ranging
transitional effects of such a reform, particularly
the potential devaluation of the capital stock in
existence on the reform date. In this article, I
review and synthesize this literature.
The effects on the capital stock’s real value
arise from the differences in how the income
and consumption taxes treat investment. A stylized income tax applies to gross output minus
depreciation, while a stylized consumption tax
applies to gross output minus gross investment.
In other words, the income tax allows a deduction only as the investment depreciates, while
the consumption tax allows new investment to
be deducted immediately. As explained below,
the consumption tax eliminates the net tax burden on (marginal) investments because the
savings from the initial deduction offset the
taxes on the subsequent output. Since the
switch to a consumption tax removes the tax
penalty on new investment, it is likely to
expand the capital stock.
However, the timing of the consumption
tax—an initial deduction offset by subsequent
taxes—has unfavorable implications for the
existing capital stock on the date of the reform.
This capital is subject to the future taxes but
does not receive the tax deduction granted to
new investment because the deduction was unavailable when the capital was produced. This
capital also loses the depreciation deductions
the income tax system provides. The introduction
of the consumption tax therefore tends to reduce the real value of the existing capital stock.
Assuming the income and consumption
taxes have stylized forms and capital is produced without adjustment costs, the real value
of the initial capital stock is reduced uniformly
by a proportion equal to the consumption tax
rate. In this simplified analysis, a 25 percent
consumption tax reduces the real value of all
capital by 25 percent. However, the tax reform
is likely to increase after-tax rates of return,
which benefits the owners of existing wealth.
The decline in value can be mitigated by transition relief (such as maintaining depreciation
deductions), at the cost of a higher tax rate on
current and future workers.
I show that under more realistic assumptions, the decline in the real value of the initial

The Transition to Consumption
Taxation, Part 1: The Impact
on Existing Capital
Alan D. Viard

his article discusses only the

impact of tax reform on the real
value of the capital stock.
Part 2, which will appear
in a future issue of Economic
and Financial Review,
will examine the effects on
the valuation of outstanding
debt and the resulting
distributional implications.

Alan D. Viard is a senior economist and
policy advisor in the Research Department
at the Federal Reserve Bank of Dallas.

FEDERAL RESERVE BANK OF DALLAS

ably. Capital is the result of any current production that allows an increase in future output
and includes intangible investments, such as
research and development.
In my initial simplified analysis, I assume
that capital can be produced at a constant marginal cost (in terms of consumption goods) and
that there are no adjustment costs associated
with investment. An unlimited number of units
of new capital can be produced, with each costing one unit of consumption. Conversely, any
unit of existing capital can be converted back
into one unit of consumption. Because capital
does not become more expensive as more is
produced, the supply of capital goods is infinitely elastic. I also assume that new capital is
economically identical to, and therefore a perfect substitute for, existing capital. I modify
these assumptions below.
I assume that capital depreciates at an
annual geometric rate δ. One unit of current
investment increases the capital stock n years
later by exp(–δn) units. The equation of motion
for the capital stock is
(2)
K· = I – δK ,

capital stock is likely to be smaller in aggregate,
but less uniform, than the simplified analysis
suggests. Consumer-owned capital and government-owned capital escape the decline in value
because they receive special treatment under consumption tax proposals. Also, tax reform forgives
the deferred income tax liabilities that many
types of capital currently face, which mitigates
or potentially reverses the decline in their real
value. Furthermore, tax reform is likely to increase investment in most types of capital,
which in the presence of adjustment costs
drives up the price of new capital and mitigates
the decline in the value of existing capital. However, tax reform is likely to reduce investment in
types of capital that are fully or partly exempt
from federal income tax, driving down their value.
The combined effect of these factors is
that some types of capital experience little or no
decline in value or even rise in value, while
other types experience significant declines. In
many cases, the magnitudes are uncertain.
I begin by comparing income and consumption taxes, with particular attention to their
treatment of investment. I explain how the timing of the consumption tax tends to cause a
decline in the real value of the initial capital
stock and present the simplified analysis, concluding that the proportional decline equals
the tax rate. I then describe the implications
of special treatment for consumer and government capital, deferred income tax liabilities,
and adjustment costs and summarize the overall
impact.

I use a simple economic model to compare income and consumption taxes. I assume
that there is no risk or uncertainty and that
the economy is closed to international trade
and investment. Gross output is produced
from labor and capital in accordance with a
production function that may vary over time,
F (Kt ,Lt ,t ). Since output and income are identical
in this closed economy, I use the terms interchangeably.
Output can be used as the economy’s
single consumption good or as capital. At each
date t, gross output is divided between consumption and gross investment (production of
new capital),
F (Kt ,Lt ,t ) = Yt = Ct + It .

Since saving is equal to investment in this
closed economy, I use the terms interchange-

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

where the dot denotes rate of change.
The annual after-tax real rate of return
demanded by the savers who supply funds to
the firm is r. To simplify notation, I assume
that r is constant over time. The wage rate
demanded at date t by households supplying
labor to the firm is wt .
In this closed economy, aggregate wealth
equals the value of the aggregate capital stock.
The combined value of the debt and equity
each firm issues must equal the value of its capital. Some households may issue debt to each
other, but this does not change aggregate
wealth since the lending household’s asset is
offset by the borrowing household’s liability.
Therefore, the impact of tax reform on the real
value of the capital stock also determines its
impact on real aggregate wealth.
However, the distribution of the wealth
changes among households depends on how
tax reform affects the real value of firms’ and
households’ outstanding debt. If the real value
of debt is unchanged, changes in the value of
a firm’s capital are borne solely by its equity
holders as residual claimants and there are no
wealth effects for households borrowing and
lending to each other. But if the real value of
outstanding debt changes, part of the change in
the value of capital is shifted to firms’ debt
holders and wealth is also transferred between
borrowing and lending households. This article

COMPARISON OF INCOME AND
CONSUMPTION TAXES

(1)

discusses only the impact of tax reform on the
real value of the capital stock. Part 2, which will
appear in a future issue of Economic and
Financial Review, will examine the effects on
the valuation of outstanding debt and the resulting distributional implications.
I now consider the treatment of labor
supply and investment under the different tax
structures. It is simplest to assume that tax revenues are rebated back to households.

The first-order condition for investment
can be derived from the following analysis.
Consider a small deviation that increases the
capital stock by one unit at date t, followed at
time t + dt by an increase in consumption that
returns the capital stock to its original path. The
investment at date t reduces consumption by
one unit. Between t and t + dt, the additional
unit of capital depreciates to 1 – (δdt ) units but
produces FK,t dt units of output, on which taxes
of τy,t (FK,t – δ)dt are paid. Consumption at date
t + dt is 1 + (FK,t – δ)(1 – τy,t )dt units. Since the
firm cannot have an incentive to deviate from
the optimal path, the consumption gained at
t + dt must be 1 + (rdt) times greater than that
sacrificed at date t, which requires

Labor Supply and Investment Under
Stylized Income Tax
In my initial simplified analysis, I consider
a pure, or stylized, income tax that accurately
measures and taxes net income. The base of this
stylized income tax is Y – δK, gross output
minus the depreciation of capital. The tax
allows depreciation deductions that match true
economic depreciation, δK, and provides no tax
credits. Corporate dividends, corporate retained
earnings, and the capital income of noncorporate firms are taxed uniformly. A single tax rate
applies to all households, although there may
be a refundable exemption to provide relief for
poorer households. I later consider a somewhat
more realistic description of the federal income
tax that reflects the deferred liabilities it imposes
on many types of capital.
A well-established principle of public
finance states that in a stylized model of this
type, it makes no difference whether taxes are
collected from buyers or from sellers. It is simplest to assume that a single firm carries out all
production and that the income tax is collected
from this firm. If τy,t is the income tax rate at
time t, the firm’s tax liability is
(3)

(6 )

Tt = τy,t [F (Kt ,Lt ,t ) – δKt ].

Real Value of Capital Under Stylized Income Tax
I next examine the real value of capital,
which I define as the number of units of consumption its owner(s) can obtain by selling
one unit of capital and consuming the after-tax
proceeds. I show that under the maintained
assumptions, this value is always unity in the
no-tax economy and under the stylized income
tax, regardless of the age of the capital and the
income tax rate.1
This result follows from a simple arbitrage
relationship. With no adjustment costs and constant costs of capital production, one unit of
new capital can be obtained at an opportunity
cost of one unit of consumption and is therefore
worth one unit of consumption. One unit of
existing capital —for example, the surviving
remnant of exp(δn) units of investment made n

(1 − τ y ,t )F (K t , Lt , t ) 
∞
(4 ) Max ∫ t = 0 exp(−rt )
 dt ,
+ τ y ,t δK t − w t Lt − I t 
subject to the constraints of Equations 1 and 2.
The first-order condition for labor is
FL ,t =

r
.
1 − τ y ,t

It is easy to show that if savers demand a timevarying rate of return, a condition of this form
holds at each instant using the contemporaneous rate of return. Also, with several types of
capital, each decaying at a different (geometric)
rate, a condition of this form holds separately
for each type.
With no taxes, the firm invests until the
pretax rate of return, FK – δ (marginal product
minus depreciation), equals savers’ required
after-tax rate of return. With income taxation,
however, this pretax return exceeds savers’
required after-tax rate of return. This wedge
between pretax and after-tax returns reflects
an economic distortion between consumption
and saving, which is widely viewed as a major
disadvantage of the income tax. As shown
below, a constant-rate consumption tax avoids
this distortion.

The firm chooses the quantity of labor L
and gross investment I to maximize the present
discounted value of Y – wL – I – T, which is its
payment to the savers who provide its funds,

(5)

FK ,t = δ +

wt
.
1 − τ y ,t

With no taxes, the marginal product of labor
equals the wage rate. With income taxation, the
marginal product is higher than the wage rate,
reflecting a distortion in the trade-off between
work and leisure.

FEDERAL RESERVE BANK OF DALLAS

The difference between this base and the
income tax base Y – δK is net-of-depreciation
investment I – δK, which, from Equation 2,
equals the change in the capital stock. In the
United States and other growing economies, the
capital stock increases over time, so net investment is positive and consumption is lower than
income. A stylized consumption tax requires a
higher tax rate than a stylized income tax to
meet a given revenue target.
In an economy with multiple firms, a consumption tax can be collected in different ways.
A retail sales tax is collected solely from the firm
that sells to the consumer, while a value-added
tax is collected from firms at different stages of
the production process. A flat tax is similar to
the value-added tax but is collected partly from
firms’ workers. A personal consumption tax
(sometimes called a personal expenditures tax
or a consumed-income tax) is collected from
consumers. Koenig and Huffman (1998, 25–26),
Congressional Budget Office (1997, 7– 22),
Gillis, Mieszkowski, and Zodrow (1996), McLure
and Zodrow (1996), Slemrod (1996), Auerbach
(1996, 43 –46), Gravelle (1996a, 1423 –28), and
Joint Committee on Taxation (1995, 51– 52,
57–58) describe and compare these different
methods of imposing a consumption tax. Part 2
of this series will examine the differing implications of these taxes for the real value of outstanding debt and for the distribution of wealth
changes across households.
However, since these taxes have economically similar effects on the real value of capital
and aggregate wealth, I do not distinguish them
here. I again assume the tax is collected from
the representative firm. Letting τc denote the
consumption tax rate,3 which is assumed to be
constant over time, the firm’s tax liability is

years ago—must have the same value as the
unit of new capital because both units have the
same marginal product and, under the stylized
income tax, are subject to the same taxes. Their
marginal products are the same because old and
new capital are perfect substitutes in production. Each unit of capital, new or old, bears the
same tax, τy,t (FK,t – δ), at each date t.
A different calculation confirms that each
unit of capital is worth one unit of consumption.
The value of each unit of capital must equal the
present discounted value of its after-tax cash
flows. From Equation 6, the after-tax cash flow
(marginal product minus tax liability) is
(7)

FK,t – τy,t (FK,t – δ) = r + δ.

For each unit of existing capital, this cash flow
declines at rate δ as the unit depreciates. So the
present value (discounted at rate r ) of future
cash flow is
(8 )

∞

V = ∫ t = 0 exp(−rt )(r + δ ) exp(−δt )dt
r +δ
=
= 1.
r +δ

The real value of capital is always unity
under the stylized income tax, regardless of
the tax rate or fluctuations in the production
function. The after-tax cash flow remains equal
to r + δ, due to changes in the quantity of
capital (which alter its marginal product) or
changes in the after-tax rate of return. Due to
the infinite supply elasticity, fluctuations in the
production function or changes in the income
tax rate alter the quantity of capital or after-tax
returns but not the real value of each unit. Since
this result also holds for a zero tax rate, adoption or repeal of the income tax does not
change the real value of capital.

(9)

Labor Supply and Investment Under
Stylized Consumption Tax
I now consider the effects of a stylized
consumption tax. The tax base is consumption,
which equals gross output minus gross investment, Y – I, in accordance with Equation 1. The
stylized consumption tax differs from the stylized income tax solely in deducting gross
investment rather than depreciation from gross
output. In other words, the firm may deduct
capital investment costs immediately rather than
as the capital depreciates. Many economists
have noted that an income tax can be transformed into a consumption tax simply by
replacing depreciation allowances with expensing, which is an immediate deduction for investment costs.2

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Tt = τc [F (Kt ,Lt ,t ) – It ].

The firm again chooses the quantity of labor L
and gross investment I to maximize the present
discounted value of Y – wL – I – T,
(1 − τc )[F (K t , Lt , t )
∞
(10 ) Max ∫ t = 0 exp(−rt )
 dt ,
− I t ] − wt L t

subject to Equations 1 and 2.
The first-order condition for labor is
unchanged from Equation 5, except that the
consumption tax rate replaces the income tax
rate,
(11)

FL ,t =

wt
.
1 − τc

rent revenues and provide a significant refundable exemption. I compare this income tax
with a 25 percent consumption tax, which
would raise approximately the same revenue,
with a similar refundable exemption.5 I use
these values—δ = 0.08, r = 0.04 (both before
and after tax reform), τy = 0.2, τc = 0.25—as a
standard example throughout this article. In
the no-tax economy, the equilibrium marginal
product is 0.12. With the 20 percent income tax,
the marginal product is 0.13, in accordance with
Equation 6; the pretax rate of return is 0.05; and
the tax payment is 0.01. With the 25 percent
consumption tax, the marginal product is 0.12,
in accordance with Equation 12. Since the consumption tax payment is 0.03, the after-tax cash
flow is 0.09.
Using these parameters, Figure 1 illustrates
the source and timing of the consumption tax’s
favorable treatment of a marginal new investment. The figure compares the time paths of the
income and consumption tax payments that
result from one additional unit of gross investment, with no later change in gross investment.
The difference in timing is dramatic. The income tax payment at each point in the life of
the investment is positive, declining as the capital depreciates. The initial consumption tax payment is negative, but subsequent payments are
positive and three times larger than under the
income tax (because the tax rate is higher and
depreciation is not deductible). As I explain below, this timing difference plays a crucial role
in the transition between the two tax systems.
One other effect of the consumption tax
should be mentioned. If the firm makes inframarginal investments that generate pure profits

However, the first-order condition for investment takes a different form. Again, consider
a small deviation that increases the capital stock
by one unit at date t, followed at date t + dt by
an increase in consumption that returns the
capital stock to its original path. The investment
of one unit at date t provides tax savings of τc ,
so after-tax consumption falls by only (1 – τc ).
Between t and t + dt, the initial unit of capital
depreciates to 1 – (δdt ) units but produces
FK,t dt units of output. Pretax consumption at
date t + dt is 1 + (FK,t – δ)dt units, and aftertax consumption is (1 – τc )[1 + (FK,t – δ)dt ].
Around the optimal path, the consumption
gained at date t + dt must be 1 + (rdt ) times
greater than that sacrificed at date t, which requires (1 – τc )[1 + (FK,t – δ)dt ] = (1 – τc )[1 + (rdt)]
or
(12)

FK,t = δ + r.

Unlike the income tax, the constant-rate
consumption tax does not impose a net tax
burden on the marginal new investment. The
tax savings from expensing exactly offset (in
present discounted value) the taxes on the subsequent cash flows. The reason is that the marginal unit of investment generates cash flows
with a present discounted value of exactly one
unit. Because there is no net tax burden, the
pretax rate of return equals savers’ required
after-tax rate of return, as in the no-tax economy. Unlike the income tax, the constant-rate
consumption tax does not distort the investment
decision.4
Because it removes the wedge between
pretax and after-tax returns, the replacement of
the income tax by a constant-rate consumption
tax has major economic implications. It either
reduces the marginal product of capital or increases the net return savers receive, or both.
In the long run, both effects are likely to occur,
as after-tax returns rise to prompt more saving
and the resulting expansion of the capital stock
drives down the marginal product. The breakdown depends on the elasticity of investment
with respect to rates of return (the rate at which
marginal product declines as the capital stock
expands) and the corresponding elasticity of
saving.
Consider an example in which the depreciation rate is 0.08. Assume that savers’ required
rate of return does not vary in response to tax
changes and always equals 0.04. If the actual
U.S. individual and corporate income taxes
were of the stylized form assumed in this simplified analysis, a tax rate of about 20 percent or
slightly higher would be sufficient to raise cur-

Figure 1

Tax Payments Resulting from
One Unit of Gross Investment
Tax payment
.05

–.05

–.10
20% Income tax

–.15

25% Consumption tax

–.20

–.25
Year 1

Year 2

Year 3

Year 4

NOTE: δ = .08; both before and after tax reform, r = .04.

FEDERAL RESERVE BANK OF DALLAS

Figure 2

(cash flows with present discounted value
greater than initial investment costs), the consumption tax takes a fraction τc of the profits.
However, since the firm still retains (1 – τc ) of
the pure profits and an investment with any
pure profits is worth making, the tax does not
deter any of these investments. (The income
tax, in addition to taking τy of the pretax return
to the marginal investment, also takes τy of any
pure profits generated by inframarginal investments.)6
The consumption and income taxes have
different effects on investment incentives. I now
show that they also have different effects on the
real value of capital.

Relationship of Tax Bases
I – δK

Business cash flow,
YK – I

C
wL

Real Value of Capital Under
Stylized Consumption Tax
The expensing provided by the consumption tax reduces the real value of each unit of
capital to (1 – τc ) units of consumption. The
opportunity cost of an additional unit of new
capital is now (1 – τc ) units of consumption
because the investment provides τc in tax savings. Conversely, since converting a unit of capital into one unit of consumption triggers a tax
payment of τc , the net yield of such a conversion is now only (1 – τc ).
The reduction in the real value of capital
is confirmed by a reduction in the present discounted value of its cash flows. From Equation 12, the marginal product of each unit of
capital is r + δ, so the after-tax cash flow is
(1 – τc )(r + δ). With depreciation rate δ and discount rate r, the present discounted value of
each unit’s cash flow is

Wage
tax
base

Consumption
tax
base

Income
tax
base

and the wage rates workers demand. However,
the income tax creates an additional distortion
by driving a wedge between the pretax and
after-tax rates of return on investment, while the
consumption tax does not. Also, the consumption tax depresses the value of capital below its
pretax replacement cost, while the income tax
does not.
A decomposition of the taxes helps clarify
their similarities and differences. Define gross
capital income as gross income minus wages,
(14)

YK,t = Yt – wt Lt .

The stylized income tax consists of a tax on wages,
wL, plus a tax on net-of-depreciation capital
income, YK – δK. The consumption tax can also
be decomposed. Combining Equations 1 and 14
reveals that consumption equals wages plus
capital income minus investment,

∞

(13) V = ∫ t = 0 exp(−rt )[(1 − τc )(r + δ)]exp(−δt ) dt
(1 − τc )(r + δ )
=
= 1 − τc .
r +δ

(15)

Ct = wt L + (YK,t – It ).

The excess of capital income over investment is
called business cash flow because it is the portion of capital (or business) income that is not
used to produce new capital and that is distributed (flows back) to savers. Business cash flow
measures capital’s net contribution to consumption (the output it produces minus the portion
reinvested to produce it). Equation 15 states that
consumption equals wages plus business cash
flow. If business cash flow is positive, as in the
United States, capital is productive, permitting
consumption to exceed wages.7 A consumption
tax is a wage tax plus a business-cash-flow tax.8
Figure 2 shows the relationship of these
quantities. Because cash flow is positive, wages
are lower than consumption. Because net investment is positive, as noted above, consumption
is lower than net income.

The timing of the consumption tax causes the
lower value of capital. Tax savings at the date of
investment offset τc units of the investment
costs. For the marginal investment, the subsequent after-tax cash flows (which must offset
the remainder of the costs) have a value of only
1 – τc units. Each unit of existing capital has
already received the initial tax savings and has
only 1 – τc units of remaining value.
Decomposing the Income and Consumption Taxes
The above analysis indicates that income
and consumption taxes are similar in one
respect but different in two others. Both taxes
distort the labor supply decision by driving a
wedge between the marginal product of labor

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

YK – δK

Since the income and consumption taxes
both include a wage tax, it is useful to consider
its properties. The wage tax drives a wedge
between the pretax product of labor and the
after-tax wage rate and distorts the labor supply
decision. This explains why the income and
consumption taxes both have this effect.
However, the wage tax has no effect on
the investment decision or on the value of capital. An addition to Figure 1 charting the wage
tax payments triggered by an additional unit of
gross investment would show zero at every
date. The initial investment does not change the
wage tax, since output produced by labor is
taxed whether it is used as consumption or as
capital. The subsequent output the investment
generates is untaxed, because the wage tax
applies only to the marginal product of labor,
not capital. Each unit of capital is still worth one
unit of consumption.
The effects of the income and consumption taxes on investment and the value of capital therefore arise from their net-capital-incometax and business-cash-flow-tax components rather
than their common wage-tax component. A netcapital-income tax creates an investment distortion by imposing a tax burden on the marginal new investment. In contrast, a constant-rate
business-cash-flow tax creates no distortion because it imposes zero present-discounted-value
burden on the marginal new investment. With a
constant rate, a cash-flow tax is a lump-sum tax
that does not distort economic decisions.
Since business cash flow is positive and
substantial, the cash-flow tax raises significant
revenue. The tax is lump sum because this revenue is not raised from labor or (in present discounted value, on the margin) from new investment, the two economic activities that can be
distorted in this model. Where then does the
revenue come from? As noted above, the tax
collects revenue from any pure profits generated by inframarginal new investments; a tax on
pure profits has long been recognized as lump
sum. But the bulk of the revenue is collected
from the capital stock in existence when the tax
is introduced. It is well known that a tax on
existing capital is also lump sum. I now describe
the effects on existing capital in more detail.

real value of the existing capital stock declines
by a proportion equal to the consumption tax
rate. This result has been widely noted and can
be regarded as canonical.9
Real Value of Capital Declines by
Proportion Equal to Tax Rate
Each unit of capital is worth one unit of
consumption under the stylized income tax and
(1 – τc ) units under the stylized consumption
tax. The tax reform therefore causes the real
value of capital to decline by proportion τc . For
example, the introduction of a 25 percent consumption tax reduces the value of existing capital by 25 percent.
Many believe the switch to a consumption
tax reduces the value of existing capital because
it increases the taxes on that capital. As explained below, this belief is largely incorrect.
However, it is useful to consider the change in
the tax burden on existing capital. These taxes
are likely to increase, largely because depreciation is no longer deductible. Also, if the tax
reform is revenue-neutral, the stylized consumption tax rate is higher than the stylized
income tax rate was. (The tax increase can be
readily seen in Figure 1.) For any capital already
in existence on the reform date, the tax payments increase due to the midstream change.
Nevertheless, the devaluation of existing
capital does not occur because it is taxed more
heavily under the consumption tax than it was
under the income tax. The decline in value
equals the consumption tax rate, regardless of
whether any income tax previously existed or
what was done with it. For example, repealing
a 90 percent income tax and adopting a 25
percent consumption tax (with spending cuts
financing the revenue shortfall) reduces the tax
burden on existing capital. Yet in this simplified
analysis, the real value of existing capital still
declines by 25 percent.
Instead, the existing capital stock is devalued because it is taxed more heavily than new
investment. Existing capital declines in value
either when it is subjected to a tax increase from
which new capital is spared or when it is excluded from a tax cut given to new capital. The
relative treatment of existing and new capital is
crucial because they are perfect substitutes for
each other in production and have the same
marginal product.
For new investment to remain viable
under an income tax, the tax payments on it
must be offset by a marginal product that
exceeds savers’ required returns, as stated in
Equation 6. Since existing capital is a perfect

SIMPLIFIED ANALYSIS OF THE TRANSITION
The above analysis permits a simple description of the transitional effects of repealing
a stylized income tax and introducing a consumption tax. The change is assumed to be
unexpected. Under the stated assumptions, the

FEDERAL RESERVE BANK OF DALLAS

Table 1

Effect of Tax Reform on Existing Capital
substitute for new capital, it also enjoys this
higher marginal product (or lower required rate
of return). Therefore, the introduction of an
income tax or an increase in its rate does not
affect the value of existing capital. The marginal
product or the required return or both change
so that cash flow remains equal to r + δ, which
ensures the value remains equal to unity.
Under the consumption tax, however,
Equation 12 makes clear that the future tax
payments from new investments are not offset
by a marginal product that exceeds savers’
required rate of return. Instead, they are offset
by the tax deduction granted on the date of
investment. Since this tax deduction is not
given to the existing capital stock, it receives
no offset for its future tax payments and its
value declines.
This point can be clarified by returning
to the standard example, in which δ = 0.08,
r = 0.04 (both before and after tax reform),
τy = 0.2, and τc = 0.25. Column A of Table 1
shows that with the income tax, the marginal
product is 0.13 and the tax payment is 0.01, so
the after-tax cash flow is 0.12. As shown in column B, when the consumption tax replaces
the income tax, the capital stock expands until
the marginal product falls from 0.13 to 0.12.
The consumption tax payment from each unit
of capital is 0.03, so the after-tax cash flow
is 0.09.
The 25 percent decline in the value of
capital reflects the 25 percent decline in the
after-tax cash flow, from 0.12 to 0.09. Of the
0.03 decline in after-tax cash flow, 0.02 is due to
higher tax payments. Tax payments rise from
0.01 to 0.03, due to the higher consumption tax
rate and the loss of depreciation allowances.
The other 0.01 of the decline in after-tax cash
flow is due to the decline in pretax marginal
product, which is an equilibrium response to
the tax cut for new investment. The annual tax
burden on new investment under the income
tax is 0.01, but the effective annual burden
under the consumption tax is zero because
expensing offsets the 0.03 annual tax payment.
With an unchanged after-tax rate of return, the
0.01 effective tax reduction causes an expansion
of the capital stock that reduces the pretax
marginal product by 0.01. In this case, twothirds of the reduction in value is due to a
tax increase on existing capital; because new
investment is spared this increase, there is no
offsetting increase in the equilibrium pretax
marginal product. The other one-third is due
to the exclusion of existing capital from a tax
cut given to new investment—a tax cut that

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

(A)
20 percent
income tax
(τy = .2)

(B)
25 percent
consumption tax
(τc = .25)

Marginal product, FK

.13

.12

.48

Pretax rate of return, FK – δ

.05

.04

.40

Tax payment T

.01

.03

.36

After-tax cash flow, FK – T

.12

.09

.12

1.00

.75

1.00

Present value, (FK – T )/(r + δ)

Income tax: FK = δ + [r /(1 – τy )]; T = τy (FK – δ)
Consumption tax: FK = δ + r ; T = τc FK
NOTE: δ = .08; both before and after tax reform, r = .04.

drives down the equilibrium pretax marginal
product and thereby reduces the value of existing capital.
Column C shows the outcome with a 90
percent income tax. Replacing this onerous
income tax with a 25 percent consumption tax
provides a large tax cut of 0.33 to existing capital, reducing annual payments from 0.36 to 0.03.
However, it grants an even larger tax cut of 0.36
to new investment. (Its annual tax burden is
0.36 under the income tax but is effectively zero
under the consumption tax, due to expensing.)
In equilibrium, this 0.36 tax reduction causes
the capital stock to expand until the marginal
product falls from 0.48 to 0.12. Due to the 0.36
decline in marginal product, the after-tax cash
flow from existing capital falls by 0.03 despite
the 0.33 tax cut.
Regardless of the rate of the income tax
being replaced, the tax disparity between existing capital and new investment is 0.03, which is
25 percent of pretax marginal product. The
value of the existing capital therefore always
declines by 25 percent.
For simplicity, this standard example
makes the extreme assumption that tax reform
does not change the after-tax rate of return,
which remains equal to 0.04, while the capital
stock expands to drive the pretax rate of return
down to that value. Realistically, savers may
demand a higher after-tax return to provide the
additional funds required for this expansion.
However, this does not alter the conclusion
that the real value of existing capital falls by
25 percent. Consider the opposite assumption,
in which the pretax return remains equal to
0.05, while the after-tax return rises to this
value. (This extreme is also unrealistic since the
higher after-tax return is likely to prompt additional saving, which expands the capital stock

and reduces the pretax rate of return.) The marginal product is 0.13, the consumption tax payment is 0.0325, and the after-tax cash flow is
0.0975. Although this cash flow is higher than in
the standard example, it must be discounted at
a higher after-tax rate of return. The value of
each unit of capital is

to 0.05, the household’s annual consumption
falls to 4.34 units. Consumption declines only 13
percent, not 25 percent. More dramatically, consider a household that consumes nothing for H
years after the tax system changes and then
splurges by consuming W exp(rH ). With the
above parameters, a household that plans to
wait forty years can consume 554 units under
the consumption tax, an increase of 12 percent
from the 495 units available under the income
tax. Some wealth holders, therefore, may support the adoption of a consumption tax, despite
the reduction in the value of their wealth. Of
course, the rare household that consumes its
entire wealth immediately after tax reform is
unaffected by future returns and suffers the full
25 percent reduction in consumption.10
The rise in after-tax returns is due to the
repeal of the income tax rather than the introduction of the consumption tax. If the consumption tax is introduced as a supplement to
the income tax (or as a replacement for a wage
tax), rates of return do not increase and all
wealth holders suffer a proportional decline in
consumption equal to the tax rate.

∞

∫ t = 0 exp(−.05t )(.0975) exp(−.08t )dt
.0975
=
= .75.
.13
It can be seen that intermediate, more realistic
assumptions about the rate of return also yield
the same result.
Although these different assumptions
about after-tax returns result in the same value
of capital, they have different implications for
the well-being of wealth holders.
Wealth Holders Benefit from
Higher After-Tax Rates of Return
The real value of the capital stock (aggregate wealth) is not the only factor affecting the
well-being of its owners. This value merely
measures the consumption the owners can
obtain by immediately liquidating their wealth.
Since most wealth holders do not intend to consume their entire wealth immediately, their
well-being also depends on the future rates of
return they will earn.
As discussed above, tax reform is likely to
cause after-tax returns to rise and pretax returns
to fall. The former effect is likely to be larger in
the short run because it takes time for additional
new investment to expand the capital stock and
drive down pretax returns. Therefore, wealth
holders are likely to enjoy significantly higher
after-tax returns for some time. Koenig and
Huffman (1998), Congressional Budget Office
(1997), Auerbach (1996), and Auerbach and
Kotlikoff (1987, 66) provide quantitative estimates.
For most wealth holders, the increase in
net returns mitigates the effects of switching to
a consumption tax. Consider a household with
initial wealth W that intends to consume at a
level rate over the next H years. If the after-tax
return is r, it consumes at an annual rate of

Effects of Transition Relief
Some consumption tax proposals call for
transition relief to mitigate or offset the decline
in the real value of existing capital. A simple
approach is to remove the business-cash-flowtax component of the consumption tax and
impose only a wage tax. Since the wage tax has
an even narrower base than the consumption
tax (Figure 2 ), a higher rate is necessary to raise
the same revenue, which imposes a heavier
burden on workers. Workers finance transition
relief for the owners of existing capital (and any
investors who receive pure profits). The higher
tax rate also exacerbates the labor supply distortion; since the cash-flow tax is lump sum, its
removal offers no offsetting efficiency gains.
As described above, a wage tax imposes
zero tax payments on investment at every date.
It does not distort the investment decision, and
the real value of capital remains equal to unity.
Replacing the income tax with a wage tax therefore leaves the value of existing capital unchanged. Existing capital escapes any decline in
value because it receives the same tax reduction given to new investment. The effects of a
wage tax (for any rate) are shown in column C
of Table 2 for the standard example, in which
δ = 0.08 and r = 0.04, both before and after tax
reform. Columns A and B are repeated from
Table 1 to show the effects of the 20 percent
income tax and 25 percent consumption tax.

Wr
.
1 − exp(−rH )
If the annual after-tax real return is initially 0.04,
a household with initial wealth equal to 100
units of consumption and a forty-year horizon
consumes 5.01 units per year. If tax reform
reduces its real wealth from 100 to 75 units but
raises the annual after-tax real return from 0.04

FEDERAL RESERVE BANK OF DALLAS

Table 2

Effect of Transition Relief on Existing Capital
(A)
20 percent
income tax
(τy = .2)

(B)
25 percent
consumption tax
(τc = .25)

(D)
Depreciation
relief
(τc = .25)

Marginal product, FK

.13

.12

Pretax rate of return, FK – δ

.05

.04

Tax payment T

.01

.03

.00

.01

After-tax cash flow, FK – T
Present value, (FK – T )/(r + δ)

.12

.09

.12

.11

1.00

.75

1.00

.917

Income tax: FK = δ + [r /(1 – τy )]; T = τy (FK – δ)
Consumption tax: FK = δ + r ; T = τc FK
Wage tax: FK = δ + r ; T = 0
Depreciation relief: FK = δ + r ; T = τc (FK – δ)
NOTE: δ = .08; both before and after tax reform, r = .04.

cient to prevent its devaluation because this
treatment is less generous than the expensing
provided for new investment. In general, it can
be shown that existing capital declines in value
by τcr/(r + δ) under this policy rather than by
the τc that occurs without transition relief. This
form of transition relief is therefore more favorable to short-lived types of capital (those with
high δ).12

Since replacing the income tax with a wage tax
increases after-tax rates of return while leaving
the real value of wealth unchanged, it improves
the well-being of wealth holders.11
A more modest (and more common) proposal allows firms to deduct the depreciation
allowances on existing capital that they would
have deducted under the income tax. The revenue loss from this relief also raises the tax rate
required to meet a given revenue target, which
exacerbates the labor supply distortion and
increases the burden on workers and any investors with pure profits. However, these effects
are smaller than under the wage-tax option.
This transition relief is insufficient to avoid
a decline in real value. As shown in column D
of Table 2, the provision of depreciation allowances (with the consumption tax rate still equal
to 25 percent, for simplicity) results, for these
parameters, in a tax liability of 0.01 on existing
capital, the same as under the 20 percent income
tax. (With higher depreciation rates, it results in
a tax reduction; with lower depreciation, tax reform still increases tax liability but by less than
without relief.) However, the pretax marginal
product of capital still declines by 0.01, the equilibrium response to the 0.01 tax cut for new
investment, so the after-tax cash flow falls from
0.12 to 0.11. Existing capital still declines in value,
though only by 8 percent rather than 25 percent.
As emphasized above, existing capital
declines in value if its tax treatment deteriorates
relative to new investment. Maintaining depreciation deductions for existing capital is insuffi-

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Limitations of Simplified Analysis
Several aspects of this simplified analysis
are unrealistic. The analysis does not reflect the
special treatment capital owned by consumers
and state and local governments would receive
under consumption tax proposals. Also, the stylized income tax system does not accurately represent the U.S. income tax and the assumed
production technology is unrealistic. The remainder of this article extends the analysis to
address these limitations.
EXTENSION 1: SPECIAL TREATMENT OF
CONSUMER AND GOVERNMENT CAPITAL
The simplified analysis assumes that all
production is done by a single firm and is
treated uniformly within each tax system. However, consumers and state and local governments, rather than private business firms, own
significant portions of the U.S. capital stock.13
Consumer capital includes owner-occupied homes
and consumer durable goods; government capital includes public buildings and roads. These

ployees. The course’s gross capital income is not
taxed, and construction costs of new courses
are not deductible. For rental housing, both
wages and business cash flow are taxed. For
owner-occupied housing, only wages are taxed.
Imputed rental is not taxed, and new home construction costs are not deductible.
One way to describe this treatment is that
the production of consumer and government
capital is treated as if it were consumption and
the actual consumption subsequently produced
by this capital is ignored. Any “consumption”
tax that taxes investment as if it were consumption and exempts the output produced by capital is simply a wage tax; only wages remain
if business cash flow (capital income minus
investment) is removed from the consumption
tax base.
Because wages in these sectors are taxed
in the same way as in the rest of the economy,
the income and consumption taxes affect labor
supply in these sectors in the manner indicated
by the general first-order conditions, Equations
5 and 11. Under the income tax, these sectors’
exemption from the tax on net capital income
causes a misallocation of capital. Since capital
income is untaxed in these sectors, the marginal
product of capital is δ + r.15 However, the marginal product is δ + [r/(1 – τy )] elsewhere, in
accordance with Equation 6. Pretax rates of return are higher for rental housing and private
golf courses than for owner-occupied housing
and municipal golf courses, implying that total
output increases if capital is shifted from the
latter to the former.
The consequences of the disparate treatment are less problematic under the consumption tax. Since the business-cash-flow tax is
lump sum, its uneven application across sectors
does not distort investment decisions. The marginal product of capital is r + δ throughout the
economy, in accordance with Equation 12. To
be sure, the rental payments on a new rental
home are taxed, while the imputed rental of
a new owner-occupied home is not. However,
the construction costs of the new rental home
are deductible, while those of the new owneroccupied home are not. Since the present discounted value of the rentals equals the construction costs (for the marginal home), the
present discounted value of the tax burden is
zero in each case. Treating the original production as consumption misstates the timing of consumption but not its present discounted value.16
Nevertheless, the disparate treatment has
important implications for the value of consumer and government capital existing on the

types of capital are currently taxed differently
than they would be under a stylized income tax,
and they would be taxed differently under
major consumption tax proposals than they
would be under a stylized consumption tax.
Because of this different treatment, they escape
the decline in value other capital experiences
on the reform date.
Under the stylized income tax, the marginal product of this capital is taxed and its
depreciation is deducted. Under the stylized
consumption tax, the marginal product of this
capital is taxed and new production of such
capital is deducted. The marginal product of
consumer capital is its imputed rental value; the
marginal product of government capital is the
imputed value of the public services it produces. However, valuation difficulties arise
because these services are not sold in the marketplace; consumers do not pay rent to themselves, and governments generally provide
public services without charge. Consumers and
recipients of public services receive the capital
income in-kind rather than as cash payments.
Whether due to these valuation difficulties
or other reasons, neither the current income tax
system nor leading consumption tax proposals
follow the stylized treatments described above.
The current system and the proposals take the
same approach to consumer and government
production, treating these sectors as they would
be treated by a wage tax. The current system
exempts them from the net-capital-income tax,
and the proposals exempt them from the business-cash-flow tax.
At a private golf course, the income tax
applies to both the wages of course employees
and the net capital income the course generates.
At a municipal golf course, however, the federal
income tax applies only to the employees’
wages. The gross capital income of the course
is not taxed, and depreciation is not deducted.
For a rental housing unit, the income tax applies
both to the wages of the landlord’s employees
and the landlord’s net-of-depreciation capital
income. For an owner-occupied home, the
income tax applies only to the wages of workers who perform services (such as plumbers
and carpenters).14 The home’s imputed rental
income is not taxed, and depreciation is not
deducted.
Similarly, proposed consumption taxes
apply to both the wages of the private course’s
employees and the course’s business cash flow,
its gross capital income minus new investment.
At the municipal course, however, the proposed
taxes apply only to the wages of course em-

FEDERAL RESERVE BANK OF DALLAS

reform date. Since the value of capital is unity
under a wage tax, tax reform causes no decline
in value. On the reform date, owner-occupied
homes and municipal golf courses escape the
decline in value experienced by rental housing
and private golf courses.17 Any pure profits
generated by inframarginal investments of consumers and state and local governments also
escape taxation.
Of course, the exemption of consumer
and state and local government capital from the
cash-flow tax reduces consumption tax revenue.
A higher tax rate is necessary to meet any given
revenue requirement, which exacerbates the
labor supply distortion and increases the tax
burden on workers, owners of other types of
existing capital, and any investors who receive
pure profits from other types of capital.

tax return is 0.04. Suppose instead that this type
of capital is tax exempt for the first 5.776 years
after it is produced and thereafter is subject to a
40 percent tax on its subsequent income. Since
this system imposes the same tax burden (in
present discounted value) as the 20 percent
income tax, the effective tax rate is 20 percent.18
The equilibrium pretax return is still 0.05, and
the marginal product is still 0.13. Since all units
of capital, regardless of age, are perfect substitutes in production, they all have this marginal
product.
However, the after-tax cash flow varies
with age. It is 0.13 for each unit of capital that
is less than 5.776 years old. Each unit of older
capital faces a tax of 0.4 (0.05), or 0.02, and has
an after-tax cash flow of 0.11. Due to this agevarying treatment, the value of capital also
varies with age. Consider a unit of capital that is
more than 5.776 years old. Since its after-tax
marginal product is permanently 0.11, its real
value (with a depreciation rate of 0.08 and a discount rate of 0.04) is 0.11/0.12, or 0.917.
Newly produced capital has an opportunity cost of one unit of consumption and, by the
familiar arbitrage argument, is worth one unit of
consumption. But capital that is older than 5.776
years is worth only 0.917 units of consumption.19 The new capital is worth more because it
still has 5.776 tax-free years left, while the older
capital does not. It can also be shown that the
real value of capital gradually falls from 1 to
0.917 during its first 5.776 years as it uses up its
tax-free period.
Since the stylized income tax (which imposes the same tax on capital of all ages) does
not reduce value, the devaluation arises solely
from the fact that the tax burden is greater in the
later part of the investment’s life. It is helpful to
view the deferred taxes as a liability the owners
of capital owe the government. The economic
burden of the tax system is 0.01 at all ages, since
marginal product always exceeds r + δ by this
amount. During the tax-free period, the owners
effectively borrow from the government. Later,
when they are paying 0.02, they effectively service this loan. The outstanding liability reduces
the value of the capital. Indeed, the same effect
could be achieved by imposing the 20 percent
uniform tax and making an explicit loan. The
capital, encumbered by the debt, would be
worth less than unity.
This devaluation has implications for the
effects of tax reform. Replacing the income tax
with a 25 percent consumption tax reduces the
value of each unit of existing capital (that is
more than 5.776 years old) by only 18 percent,

EXTENSION 2: INCOME TAX DEFERRAL
The simplified analysis assumes that the income tax is a stylized tax that measures income
accurately and treats old and new capital neutrally. Under such a tax, the value of capital
equals its replacement cost. However, under more
realistic assumptions about the timing of the
income tax, many types of capital are already
worth less than replacement cost. Replacing the
income tax with a consumption tax therefore
causes a smaller reduction in (or may even
increase) the real value of existing capital.
Income Tax Deferral Reduces the
Real Value of Capital
As explained above, the differing impact
of the stylized income and consumption taxes
on the real value of capital result from differences in their timing. The stylized income tax
collects the same tax from each unit of capital
regardless of age, while the consumption tax
grants an initial tax deduction offset by subsequent taxes (recall Figure 1). Unlike the stylized
income tax, the consumption tax is a deferred tax.
However, as detailed below, the federal
income tax system frequently imposes heavier
burdens on old capital than on new investment.
I show that this tax deferral reduces the real
value of capital below unity. A tax reform that
combines the introduction of a consumption tax
with repeal of the income tax reduces the value
of existing capital by less than the simplified
analysis concludes. In some cases, the value of
existing capital may increase.
Consider the standard example of a 20
percent income tax on a type of capital with a
0.08 depreciation rate, when the required after-

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

law still provides a 20 percent credit for research and experimentation, a 10 percent credit
for business equipment that uses solar or geothermal energy, and a 10 percent to 20 percent
credit for rehabilitation of historic structures.
Under accelerated depreciation, the tax
law computes depreciation deductions as if
capital depreciates more rapidly than it does.
For example, capital with a 0.08 depreciation
rate may be depreciated for tax purposes at
a 0.12 rate. One unit of investment results in
a depreciation deduction t years later of 0.12
exp(–0.12t ), rather than the true depreciation of
0.08 exp(–0.08t ). The deductions are higher
than true depreciation for capital that is less
than 10.137 years old and lower for capital that
is older. This deferral of tax liability results in
capital having a value less than unity. The tax
law provides accelerated depreciation for nearly
all types of tangible capital.

from 0.917 to 0.75. Income tax repeal effectively
forgives the deferred tax liability and increases
the value of the capital, partly offsetting the 25
percent decline caused by the introduction of
the consumption tax.
Therefore, to the extent current income
taxes are imposed on a deferred basis, capital is
already valued at less than replacement cost and
the replacement of the income tax by a consumption tax has a smaller impact than the
previous analysis suggests. If the devaluation
under the income tax is sufficiently large, the
switch to consumption taxation may even raise
the real value of existing capital.
Note that the offsetting rise in value is due to
the repeal of the income tax (and its associated
deferred liabilities) rather than the introduction of
the consumption tax. If a consumption tax is introduced as a supplement to the income tax (or as a
replacement for a wage tax), the value of existing
capital still declines by a proportion equal to the
consumption tax rate, as in the simplified analysis.
Similarly, if the consumption tax rate is subsequently raised to meet an increase in revenue
needs, the real value of existing capital still declines by a proportion equal to the rate increase.
As described below, numerous instances
of deferred taxation exist in the current system.
Front-loaded investment and saving incentives
result in deferred taxation. Also, under one
theory of corporate financial policy, the taxation
of corporate dividends at a higher rate than retained corporate earnings results in tax deferral.

Front-Loaded Personal Savings Incentives
Another form of deferred income taxation
is the provision of front-loaded incentives for
personal saving. The amount saved is deducted
and the proceeds are fully taxed when withdrawn from a designated account. These incentives apply to pensions, conventional individual retirement accounts (IRAs), 401(k)s,
403(b)s, medical savings accounts, education
IRAs, and Keogh accounts for self-employed
taxpayers. In some cases, a 10 percent penalty
may also apply to withdrawals. The taxation of
the proceeds reduces the value of the assets
to (1 – τy ) or, if the penalty is applicable, to
(0.9 – τy ). Repeal of the individual income tax
forgives the tax and penalty.21
The incentives provided to Roth IRAs do
not have the same effects. The household receives no deduction for the original saving, but
the return on the account is exempt. Although
these incentives are often referred to as backloaded, their timing is actually neutral. There is
no deferral because the same zero tax rate
applies at all times. Income tax repeal does not
increase the net value of Roth IRAs.
Another instance of deferred taxation is
the delay in taxing capital gains until the gains
are realized. Repeal of the income tax forgives
the tax on unrealized capital gains.22

Front-Loaded Investment Incentives
Front-loaded investment incentives include
expensing, investment tax credits, and accelerated depreciation.20 Current income tax law
allows the cost of some investments to be expensed (as all investment would be under a
consumption tax) rather than depreciated. Most
intangible investments —such as research and
development, worker training, and business
planning—can be expensed. Small businesses
are allowed to expense $20,000 (scheduled to
rise to $25,000 by 2003) of equipment investment per year. Each unit of expensed capital is
worth (1 – τy ) under the income tax, so the net
change resulting from tax reform is (τy – τc ).
In some instances, the income tax allows
a credit against tax liability equal to a fraction k
of investment costs when the investment is
made. Each unit of capital then has a value of
about (1 – k), depending on how depreciation
allowances are adjusted to account for the
credit. Although the general credit for equipment investment was abolished in 1986, U.S. tax

Taxation of Corporate Dividends
and Retained Earnings
Another potential source of deferred taxation has broad applicability. Shareholders in
many corporations must pay individual income
tax on dividends and on the capital gains that

FEDERAL RESERVE BANK OF DALLAS

equal to a fixed fraction x of their net capital
income, which implies

result from corporate retained earnings (in
addition to the corporate income tax imposed
at the firm level).23 The tax on dividends is
generally higher than the effective tax on retained earnings. Under one theory of corporate
financial policy, this differential taxation is a
deferral of tax liability and reduces the value of
existing capital.
Consider a corporation that uses equity to
finance all its investment. Let ISSUE denote the
funds raised by issuing and selling new shares,
DIV denote the dividends paid to existing shareholders (gross of dividend tax), and RETAIN
denote the earnings retained on behalf of existing shareholders. The business cash flow distributed to stockholders consists of dividend
payments to existing stockholders minus equity
issuance proceeds received from new stockholders. By definition, business cash flow equals
gross capital income minus gross investment, so
DIV – ISSUE = YK – I. Retained earnings equal
the net increase in the capital stock existing
stockholders own, consisting of the increase
in the firm’s capital stock (gross investment
minus depreciation) minus the portion sold to
new stockholders, so RETAIN = (I – δK ) – ISSUE.
Rewriting these equations yields
(16)

(18)

Since increases in I result in one-for-one increases in ISSUE with no changes in DIV or
RETAIN, new share issuance is the marginal
source of finance for new investment.
Under this theory, the tax imposed on
stockholders is
(19)

Tt = τd DIVt + τr RETAINt
= [x τd + (1 – x)τr ](YK,t – δKt ).

Under the old view, the system imposes a tax
on net capital income, with the effective rate
equal to a weighted average of the rates on dividends and retained earnings. The lower rate on
retained earnings merely reduces the overall
tax rate with no deferral. Since investment is
financed from new share issuance, with no
change in dividends or retained earnings, no tax
saving or payment occurs at the time of investment. The subsequent output from the investment generates a stable mixture of dividends
and retained earnings at each date, so its tax
treatment is the same at each date.
However, King (1974) and later writers
challenge the old view’s assumption that the
corporation simultaneously issues new equity
and pays dividends. Even in the no-tax economy, raising funds from new stockholders and
paying them to current stockholders generates
unnecessary transaction costs. More important,
this behavior raises shareholders’ taxes. Lowering both dividends and share issuance by one
dollar, which increases retained earnings by one
dollar, reduces taxes by (τd – τr ). Also, the small
amount of equity issuance by mature corporations casts doubt on the old view’s assertion that
such issuance is the marginal source of investment finance.
King and those who followed him advocate an alternative theory of corporate financial
behavior, known as the “new view.” 24 Under this
theory, a mature corporation, defined as one
with positive cash flow, pays dividends equal to
its business cash flow and issues no new equity:

DIVt + RETAINt = YK,t – δKt ;
RETAINt + ISSUEt = It – δKt .

For given values of real variables (capital income, investment, and depreciation), Equation
16 places two restrictions on the three financial
variables (dividends, equity issuance, and retained
earnings). The corporation has one degree of
freedom in choosing its financial policy.
If a common tax rate τ applies to both dividends and retained earnings, the liability under
any financial policy satisfying Equation 16 is
(17) Tt = τ (DIVt + RETAINt ) = τ(YK,t – δKt ).
Regardless of the corporation’s financial policy,
this is simply a tax on net capital income with
no tax deferral.
However, since the federal income tax system offers a preferential tax rate on long-term
capital gains, the dividend tax rate, τd , exceeds
the effective tax rate on retained earnings, τr .
Because this system treats different financial
flows differently, its effects depend on the corporation’s financial policy.
Economists have considered two major
theories of corporate financial policy. The earliest theory, generally called the “old” or “traditional” view, assumes that firms pay dividends

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

DIVt = x (YK,t – δKt );
RETAINt = (1 – x)(YK,t – δKt );
ISSUEt = It – (1 – x)YK,t – x δKt .

(20)

DIVt = YK,t – It ; RETAINt = It – δKt ;
ISSUEt = 0.

Since increases in I result in one-for-one increases in RETAIN and decreases in DIV, with
no change in ISSUE, retained earnings (foregone
dividends) are the marginal source of finance
for new investment.

firm level, causing the marginal installation costs
of capital at each firm to rise when the firm installs more capital. Since both forces have similar
implications when investment increases throughout the economy, I discuss them together and
refer to them as adjustment costs.26
As discussed previously, tax reform is
likely to cause the aggregate capital stock to
expand, so that the equilibrium pretax rate of
return is reduced. With adjustment costs, this
expansion is associated with a rise in the cost of
capital goods, which mitigates the decline in
value for existing capital. Adjustment costs also
restrain the expansion of the capital stock.
With adjustment costs, the pretax rate of
return equals (FK + q· )/q – δ rather than FK – δ,
where q is pretax replacement cost and the dot
denotes rate of change. Dividing by q is necessary because q units of consumption must be
sacrificed to obtain one unit of capital. Adding
the change in q is necessary to reflect the capital
gain (or loss) earned by holding capital. The firstorder conditions for investment under income
and consumption taxation are (respectively)

Under this theory, the tax imposed on
equity holders equals
(21)

Tt = τd DIVt + τr RETAINt
= τr (YK,t – δKt ) + (τd – τr )(YK,t – It ).

The tax combines a net-capital-income tax with
effective rate τr and a cash-flow tax with effective rate equal to (τd – τr ), the “extra” tax on
dividends. Under the new view, the extra dividend tax is a deferred tax. Since new investment
is financed by retained earnings (a reduction in
dividends), a tax savings of (τd – τr ) is received
at the time of investment. The subsequent output is distributed as dividends, in accordance
with Equation 20, on which taxes are imposed.
This tax timing—an initial tax savings offset by
subsequent tax payments— is similar to that of
the consumption tax and front-loaded investment incentives.
Under the new view, only the tax on
retained earnings (capital gains) is a net-income
tax that distorts investment. The extra tax on
dividends is a business-cash-flow tax that leaves
investment undistorted but reduces the value of
capital to 1 – τd + τr . The combined effect of
income tax repeal and introduction of a consumption tax at rate τc is to change the capital
stock’s value by (τd – τr – τc ). The real value of
existing capital may even increase, depending
on the tax rates.
Although the new view is a theoretically
appealing description of the behavior of mature
corporations with positive business cash flow,
its validity remains controversial. Zodrow (1991)
reports that some empirical evidence favors the
old view, while Auerbach and Hassett (2000)
report evidence favoring the new view. Zodrow
discusses the difficulty of conclusively testing
the two theories, particularly in light of additional modifications (not considered here) that
can be made to each, including models of the
interaction of equity and debt finance.

(22)

FK ,t + q˙t
qt

=δ+

r
,
1 − τ y ,t

and
(23)

FK ,t + q˙
qt

= δ + r,

rather than Equations 6 and 12. If the after-tax
required return r remains constant, the pretax
return still declines after tax reform, but a
decline in FK is now only part of the likely
response (and is smaller because the capital
stock does not expand as much). Another part
is a rise in q (and still another is a negative
value of q· since the increase in q following tax
reform is likely to decay over time). Ignoring
the income tax deferral considered above, each
unit of capital is still valued at q under the
income tax and (1 – τc )q under the consumption tax. However, this no longer implies a
proportional decline of τc , because q is higher
under the consumption tax.
Note that the increase in investment and
the resulting rise in q are due to the repeal of
the income tax rather than the introduction of
the consumption tax. If a consumption tax is
introduced as a supplement to the income tax
(or as a replacement for a wage tax), the value
of existing capital is still likely to decline by a
proportion equal to the consumption tax rate, as
in the simplified analysis. Similarly, if the consumption tax rate is subsequently raised to meet

EXTENSION 3: ADJUSTMENT COSTS
Incorporating more realistic assumptions
about production generally (but not always)
mitigates the decline in value implied by the
simplified analysis.25
The simplified analysis assumes the supply
of capital is infinitely elastic because unlimited
amounts of each type of capital can be produced and installed at constant cost (in terms of
consumption). It is more realistic to assume that
capital’s marginal production cost rises when
economy-wide production increases. Also, some
evidence suggests adjustment costs exist at the

FEDERAL RESERVE BANK OF DALLAS

increased revenue needs, the real value of existing capital is still likely to decline by a proportion equal to the rate increase.
In the simplified analysis, the decline in
value is always τc , regardless of the response of
r to tax reform; if r rises, it simply dampens the
decline in FK . With adjustment costs, however,
an increase in r dampens the rise in q as well as
the decline in FK . If after-tax rates of return rise,
the decline in value of existing capital is greater
than it would otherwise be.
The extreme form of adjustment costs
occurs when adjustment is impossible and the
quantity of capital is fixed; additional units cannot be produced at any cost and the existing
units cannot be converted back into consumption. If all types of capital are in fixed supply
and changes in labor supply are unimportant,
the marginal product of capital is unchanged by
tax reform. Consider the standard example, in
which δ = 0.08 and r = 0.04 both before and
after tax reform. With no adjustment costs, the
marginal product declines from 0.13 to 0.12 and
the value falls from 1 to 0.75. But if capital is in
fixed supply and the marginal product remains
unchanged at 0.13, the value of each unit of
capital is

sis is then reversed. Adjustment costs reduce the
replacement cost of these types of capital, reinforcing any decline in the value of existing
capital. As noted above, consumer and government capital is not taxed, while many types of
intangible capital face effective tax rates of zero
because they are expensed. This analysis is
likely to apply to these types of capital.27
POTENTIAL MAGNITUDES OF EFFECTS
This section combines the various extensions discussed above and incorporates them into
the simplified analysis. The aggregate decline in
value is smaller than the simplified analysis suggests, but its distribution is less uniform.
Table 3 summarizes the possible impacts
for five categories of capital. (Even so, the
analysis is somewhat aggregated; the actual
impact may vary significantly across different
types of capital within each category.) The consumption tax rate is assumed to be 25 percent.
Column A lists the percentage declines that
apply when there is no income tax deferral and
no adjustment costs. Column B lists the modification attributable to front-loaded investment
incentives. Column C lists the modification
attributable to adjustment costs; due to uncertainty about the magnitude, only the sign of
the effect is reported. Column D combines
columns A through C. Column E states the
appropriate modification if the new view is
valid, and column F combines that modification
with column D.
The first row refers to consumer and government capital. Column A shows zero because
this capital is exempt from the cash flow tax.
The entry in column B is also zero because this
capital does not receive any front-loaded investment incentives. Its preferential treatment under
the current income tax consists of facing a zero
tax rate at every stage of its life.
Column C reports a negative effect because consumer capital and government capital
are exempt from income tax and are likely to
contract after tax reform, as discussed above.
Congressional Budget Office (1997, 45–46 ) and
Gravelle (1996b) survey the literature on possible reductions in value of the largest category of
consumer capital, owner-occupied homes. Allowing for some increase in after-tax interest rates,
Gravelle estimates that homes would decline in
value by about 22 percent under the extreme
assumption that they are in fixed supply. With
more realistic assumptions about the magnitude
of adjustment costs, however, she concludes
that the decline may be as low as 9 percent.

∞

∫ t = 0 exp(−.04t )[(.75)(.13)] exp(−.08t )dt
.75(.13)
=
= .81
.12
and the decline in value is 19 percent rather
than 25 percent. If some adjustment costs are
present, but capital is not in fixed supply, an
intermediate outcome occurs. Also, if tax reform
raises the required after-tax return, the decline
in value is greater than 19 percent, even under
the fixed-supply assumption.
Adjustment costs have opposite implications for lightly taxed types of capital. Consider
a system in which some types of capital face
effective tax rates of 30 percent and others face
effective rates of zero. If the after-tax rate of
return is initially 0.04, pretax rates of return are
0.057 for the former and 0.04 for the latter. If tax
reform causes the after-tax rate of return to rise
to .045, the stock of the heavily taxed capital
expands until its pretax rate of return falls from
0.057 to 0.045, while the stock of the untaxed
capital contracts until its pretax rate of return
rises from 0.04 to 0.045. The untaxed capital
does not benefit from tax reform, and it is
crowded out by the rise in after-tax interest
rates.
Therefore, untaxed (or lightly taxed) types
of capital are likely to contract rather than
expand after tax reform. All of the above analy-

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Table 3

Summary of Transition Effects on Real Value of Capital
Percentage change in value from replacement of income tax with 25 percent consumption tax
(A)
Simplified
analysis*

(B)
Front-loaded
investment
incentives

Adjustment
costs

Tangible capital except equityfinanced mature corporation

– 25

> –17

Tangible capital, equity-financed
mature corporation

– 25

> –17

Intangible capital except equityfinanced mature corporation

– 25

< 10

Intangible capital, equity-financed
mature corporation

– 25

< 10

< 35

Consumer and government capital

(C)

(D)
Net effect
(if old view
is valid)

(E)
Additional
effect (if new
view is valid)

(F)
Net effect
(if new view
is valid)

* Modified to reflect special treatment of consumer and government capital.

The entry in column E is zero because the
new view is inapplicable to noncorporate capital, including consumer and government capital.
The net impact shown in columns D and F is a
value decline whose magnitude depends on
adjustment costs.
The next two rows refer to tangible capital,
such as plant and equipment. The first of these
rows refers to investment for which the new
view is inapplicable. This includes debt-financed
corporate investment, equity-financed investment
by immature corporations with negative cash flow,
and investment by noncorporate firms (proprietorships, partnerships, and limited liability companies) and S corporations that are taxed in the
same manner as noncorporate firms. The other
row refers to investment for which the new view
(if valid) is applicable, which is equity-financed
investment by corporations (other than S corporations) with positive cash flow.
Under the simplified analysis, the value of
tangible capital declines by 25 percent, as
shown in column A. However, this type of capital has deferred tax liabilities due to accelerated
depreciation and other front-loaded investment incentives. Following Auerbach (1996, 51),
I assume these incentives reduce the value of
this capital by an average of 8 percent and enter
this number in column B.
I report a positive effect in column C,
since this type of capital should expand after tax
reform and adjustment costs should increase
its value. The magnitude is highly uncertain.
Auerbach (1996, 62) estimates an increase of
about 10 percent under one assumption about
adjustment costs but notes that a smaller or

larger increase is possible.
The combined effect shown in column D is
a decline in value of less than 17 percent. An increase is possible if adjustment costs are quite large.
For equity-financed investment by mature
corporations, the net impact is more favorable,
if the new view is valid. This is the category
Koenig and Huffman (1998) consider. Following
their assumption that dividends are taxed at 25
percent and capital gains at zero, I enter 25 percent in column E. The net impact is an increase
in value of more than 8 percent. If the old view
is valid, this adjustment does not apply.
The last two rows refer to intangible capital, again distinguishing between capital for
which the new view (if valid) is applicable and
that for which it is inapplicable. The initial
impact is 25 percent. Because this capital is
expensed, it is currently devalued by a proportion equal to the income tax rate. I enter 35 percent, the top corporate income tax rate, in column B. The adjustment-costs effect is negative
since, as explained above, tax reform is likely to
reduce investment in this capital. The net impact is an increase in value of less than 10 percent. The value may decline if adjustment costs
are sufficiently large. For equity-financed investment by mature corporations, the net impact is
an increase in value of less than 35 percent, if
the new view is valid.
These estimates do not include frontloaded saving incentives. The removal of the
income tax (and penalty) on withdrawals from
employer pension plans and tax-deferred
accounts constitutes an increase in value that
should be added to the numbers in Table 3.

FEDERAL RESERVE BANK OF DALLAS

CONCLUSION

NOTES

Replacing income taxation with consumption taxation would have wide-ranging effects
on the value of the capital stock. A simple result
can be obtained by assuming that the income
and consumption taxes have stylized forms and
that capital goods can be produced at constant
cost (with no adjustment costs). In this simplified analysis, the real value of existing capital
would decline by a proportion equal to the consumption tax rate. This decline would occur
because existing capital would be treated less
favorably than new investment. The harm to
owners of existing capital would be mitigated
because income tax repeal would increase aftertax rates of return.
However, under a more realistic specification of the consumption tax, consumer and
government capital would receive preferential
treatment that would allow them to escape
this value decline. Also, under more realistic
assumptions about income tax design and the
production process, income tax repeal would
offset part of the negative impact from the introduction of the consumption tax. Income tax
repeal would enhance the value of business
capital by forgiving deferred tax liabilities.
Repeal would also increase investment in many
types of business capital, which, in the presence
of adjustment costs, would drive up the value of
existing capital. However, repeal would reduce
investment in consumer and government capital
and some lightly taxed forms of business capital, which would drive down their real value.
The net result is that the decline in the real
value of existing capital is smaller and less uniform than the simplified analysis suggests. Some
types of capital might even rise in value. The
reduced overall impact on the real value of
capital weakens the argument for broad transition relief. However, the uneven nature of the
effects might support an argument for targeted
relief for types of capital that are more adversely
affected. Uncertainty about the magnitude of
adjustment costs and the appropriate theory
of corporate financial behavior complicates decisions on transition policies.
The analysis in this article offers an incomplete description of the transition. The distribution of the wealth decline depends on how tax
reform affects the real value of outstanding
debt. What are the relative effects on stockholders and bondholders? When owner-occupied
homes decline in value, is the loss borne fully
by the owners, or do mortgage lenders bear part
of the loss? In Part 2, I address these issues.

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

I am grateful to Mark Wynne, Mine Yücel, Gregory
Huffman, and V. Brian Viard for extremely helpful
comments and to Monica Reeves for careful editing.
Many authors discuss this result, including Bradford
(2000, 319 – 25), Lyon (1990), Sandmo (1979),
Johansson (1969), and Samuelson (1964).
See Koenig and Huffman (1998, 31), Gentry and
Hubbard (1997, 8), Gravelle (1996a, 1427), Joint
Committee on Taxation (1995, 57), and Auerbach and
Kotlikoff (1987, 133).
I express the tax rate in a tax-inclusive manner rather
than the tax-exclusive manner in which retail sales tax
rates are usually expressed. For example, the sales
tax rate is usually said to be 33.33 percent if a $33.33
tax is imposed on a consumption good with a pretax
price of $100 and an after-tax price of $133.33. I refer
to this levy as a 25 percent tax (τc equals 25 percent)
because the tax is 25 percent of the after-tax price.
The tax-inclusive approach is consistent with the
manner in which income tax rates are stated. Koenig
(1999, 685), Gale (1999, 443 – 44), Gillis, Mieszkowski,
and Zodrow (1996, 730, note 17), and Joint Committee
on Taxation (1995, 54, note 83) discuss the relationship between tax-exclusive and tax-inclusive rates.
A time-varying consumption tax rate is distortionary
because it penalizes investment in years when the tax
rate is temporarily low. Time-varying rates generally
pose greater difficulties for consumption taxes than for
income taxes. See Bradford (2000, 311– 31) and
Auerbach and Kotlikoff (1987, 62, 83 – 87).
Estimates of the revenue-neutral consumption tax rate
include 21.8 percent (Koenig 1999, 697); 27.7 percent
(Gale 1999, 455, replacing excise as well as income
taxes and allowing for tax avoidance and evasion and
base erosion); 25.2 percent to 25.7 percent (Ventura 1999,
1445, imposing revenue neutrality only in steady state);
and 21.4 percent (Feenberg, Mitrusi, and Poterba 1997, 75).
The Treasury Department has also estimated a revenueneutral rate of about 25 percent. Gilles, Mieszkowski,
and Zodrow (1996, 763) make a similar estimate.
Bradford (2000, 71–72, 91– 92), Gentry and Hubbard
(1997, 6), Congressional Budget Office (1997, 29),
and Joint Committee on Taxation (1995, 58, note 94)
discuss the taxation of pure profits. Also, in an economy with uncertainty, the returns from various risky
investments may be higher or lower than the return on
a safe investment. Under the income and consumption
taxes, the government shares in these surpluses and
shortfalls in the proportions τy and τc , respectively.
The market value of the surpluses and shortfalls is
zero, as Bradford (2000, 93; 1996, 129) and Gentry
and Hubbard (1997, 7– 9) explain. Also, see Joint
Committee on Taxation (1995, 94 – 99).
Abel et al. (1989) demonstrate that business cash flow
has been consistently positive in the United States. If it

were consistently negative, more resources would be
devoted to producing capital than were produced by
it. Such an economy would be considered dynamically
inefficient because it could increase both current and
future consumption by reducing its capital stock.
Bradford (2000, 68), Koenig and Huffman (1998, 26),
Gentry and Hubbard (1997, 5), Congressional Budget
Office (1997, 66), Auerbach (1996, 30 – 31), and Joint
Committee on Taxation (1995, 55) note this equivalence.
See Bradford (2000, 80, 99), Diamond and Zodrow
(1999, 25), Lyon and Merrill (1999, 308), Hall (1997,
147), Congressional Budget Office (1997, 66),
Auerbach (1996, 47) and Gravelle (1996a, 1443).
These authors differ in the extent to which they
acknowledge the qualifications to this conclusion that
I address in the text when I modify the simplified
analysis. Lewis and Seidman (2000, 100), Gentry and
Hubbard (1997, 10 –11), Feenberg, Mitrusi, and
Poterba (1997, 85), Gillis, Meiszkowski, and Zodrow
(1996, 747), Joint Committee on Taxation (1995, 84),
and Auerbach and Kotlikoff (1987, 62, 79) also note in
more general terms that the adoption of a consumption tax reduces the real value of existing wealth.
Lewis and Seidman (2000), Diamond and Zodrow
(1999, 26), Congressional Budget Office (1997, 67),
Gentry and Hubbard (1997, 11), Feenberg, Mitrusi,
and Poterba (1997, 85), Gillis, Mieszkowski, and
Zodrow (1996, 748), Auerbach (1996, 60), Bradford
(1996, 139 – 40), and Joint Committee on Taxation
(1995, 87) note the effects of higher after-tax returns.
In the text, I consider households that experience a
wealth decline equal to the 25 percent reduction in the
value of the capital stock (aggregate wealth). Recall
that this article does not address how the wealth
decline is divided between debt and equity holders
or household lenders and borrowers. I will explain in
Part 2 that wealth changes may vary greatly across
households with different portfolios.

If homeowners perform their own services, their imputed wages are also exempt from income tax and
proposed consumption taxes. This exemption distorts
the allocation of labor but has no implications for the
valuation of capital.
This statement assumes that state and local governments make their investment decisions using the same
profit-maximization criteria as private firms. In reality,
their decisions may be affected by a variety of political
factors.
Some argue that the difficulty of measuring these
imputed service flows is a disadvantage of the consumption tax. The opposite is true, since the exemption of these flows causes capital misallocation under
the income tax but not under the consumption tax.
Capital income must be measured to be taxed at a
positive rate but need not be if it is to be taxed at a
zero effective rate, which is the objective of the
consumption tax. See Bradford (2000, 10 –12, 94 – 95).
Bradford (2000, 107; 1996, 140), Lewis and Seidman
(2000, 100), Diamond and Zodrow (1999, 25), Hall
(1997, 149), Congressional Budget Office (1997,
66–67), and Sullivan (1996, 342) note that consumer
capital escapes the decline in value that affects other
capital. In the absence of special rules, the preferential
treatment applies to capital consumers and governments own on the reform date, regardless of subsequent transactions. For example, housing that is
owner-occupied on that date still benefits, even if it
is converted to rental use the next day; although
subsequent rental payments are taxed, an immediate
deduction equal to the home’s value is granted
because the conversion is treated as new investment.
Conversely, housing used for rental on the reform date
does not benefit, even if it is converted to owner use
the next day; although the owner’s subsequent imputed
rental income is exempt, an immediate tax is imposed
on the home’s value because the conversion is treated
as disinvestment and consumption.

This equivalence follows because

A related approach maintains the cash-flow tax but
gives each unit of existing capital offsetting rebates
with a present value of τc . This may be done through
an immediate rebate of τc or, as Bradford (2000,
327– 28) discusses, a permanent stream of rebates
equal to τc (δ + r ) exp(–δt ) in year t. The rebate
approach differs from the wage-tax option only in
taxing any future pure profits and maintaining the
appearance of a tax on capital.
Bradford (2000, 110; 1996, 143) discusses the relative
treatment of long-lived and short-lived capital under
this transition policy.
The federal government also holds capital. It is
economically irrelevant, however, whether the federal
government taxes itself on this capital. Capital held by
nonprofit organizations would also receive preferential
treatment under the major consumption tax proposals,
and the analysis in the text generally applies to this
capital.

∞

.4∫ t=5.776 exp(–.12t )dt = .2∫ t=0 exp(–.12t )dt.

Note that exp[– (5.776)(.12)] equals 0.5.
If firms can convert a unit of older capital back into
one unit of consumption without any tax liability, the
value of this capital cannot fall below unity. Indeed, no
older capital remains in existence since all firms make
such conversions. To prevent this outcome, the tax
system must require the firm to pay a tax of 0.083 units
to recapture the benefits of the earlier tax-free period.
Under the federal income tax, most front-loaded
investment incentives are accompanied by recapture
taxes. Since the tax law often allows firms that purchase existing capital from other firms to claim the
same front-loaded benefits as if they had produced
new capital, it also imposes recapture taxes on selling
firms to prevent tax-motivated sales. Auerbach and

FEDERAL RESERVE BANK OF DALLAS

Kotlikoff (1983, 129 – 36) describe U.S. recapture
taxes then in place.
Under the tax law, each owner of a sole proprietorship,
partnership, limited liability company, or small (S)
corporation pays individual income tax on his or her
share of the firm’s income. The front-loaded investment
incentives discussed in the text are used in computing
each owner’s tax liability. Each corporation (other than
an S corporation) pays a firm-level corporate income
tax on its taxable income, and (as discussed in the
text below) each of its shareholders also pays individual income tax on his or her dividends and capital
gains. The front-loaded investment incentives are used
to compute the corporate income tax but not the
shareholders’ individual income taxes. Auerbach and
Kotlikoff (1987, 131– 35; 1983) discuss the effects of
front-loaded investment incentives on the value of
capital. Lyon and Merrill (1999, 307), Diamond and
Zodrow (1999, 26), Gillis, Mieszkowski, and Zodrow
(1996, 748), Bradford (1996, 137), Auerbach (1996,
36), and Joint Committee on Taxation (1995, 84 – 85)
discuss the resulting implications for the transitional
effects of tax reform.

(1997, 41), Congressional Budget Office (1997, 67),
and Auerbach (1996, 48, 61– 63) discuss the implications of adjustment costs for the transitional effects of
tax reform.
27

REFERENCES
Abel, Andrew B., N. Gregory Mankiw, Lawrence H.
Summers, and Richard J. Zeckhauser (1989), “Assessing
Dynamic Efficiency: Theory and Evidence,” Review of
Economic Studies 56 (January): 1–19.
Auerbach, Alan J. (1996), “Tax Reform, Capital Allocation,
Efficiency, and Growth,” in Economic Effects of Fundamental Tax Reform, ed. Henry J. Aaron and William G.
Gale (Washington, D.C.: Brookings Institution), 29 – 81.
Auerbach, Alan J., and Kevin A. Hassett (2000), “On the
Marginal Source of Investment Funds,” NBER Working
Paper Series, no. 7821 (Cambridge, Mass.: National
Bureau of Economic Research, August).

Nondeductible IRAs and tax-deferred annuities also
receive front-loaded incentives. Lewis and Seidman
(2000, 100), Gillis, Mieszkowski, and Zodrow (1996,
747), and Auerbach (1996, 38, 69) discuss the
transitional effects of front-loaded savings incentives.
Lewis and Seidman (2000, 100), Congressional
Budget Office (1997, 67), and Gillis, Mieszkowski, and
Zodrow (1996, 747) discuss the transitional implications of accrued capital gains.
The analysis in this section does not apply to investment by S corporations, noncorporate firms, or
consumers and governments.
Sinn (1991a; 1991b) thoroughly discusses the new
view. Congressional Budget Office (1997, 67), Gentry
and Hubbard (1997, 39), Gillis, Meiszkowski, and
Zodrow (1996, 748), and Auerbach (1996, 37, 69) note
its implications for the transitional effects of tax reform.
Koenig and Huffman (1998) assume the validity of the
new view in their analysis.
If new investments are imperfect substitutes for old
capital because they incorporate different technologies, the decline in value of existing capital is also
mitigated. In the standard example, when tax reform
reduces the marginal product of new capital from
0.13 to 0.12, I assume the marginal product of existing
capital falls by the same amount, which (along with
the increase in its tax payments) causes its value to
decline. However, if the two vintages of capital are not
perfect substitutes, the expansion of new investment
may not drive down the return on existing capital to
the same extent.

Auerbach, Alan J., and Laurence J. Kotlikoff (1983),
“Investment Versus Savings Incentives: The Size of the
Bang for the Buck and the Potential for Self-Financing
Business Tax Cuts,” in The Economic Consequences of
Government Deficits, ed. Laurence H. Meyer (Boston:
Kluwer-Nijhoff), 121– 49.
——— (1987), Dynamic Fiscal Policy (Cambridge:
Cambridge University Press).
Bradford, David F. (1996), “Consumption Taxes: Some
Fundamental Transition Issues,” in Frontiers of Tax
Reform, ed. Michael J. Boskin (Stanford: Hoover
Institution Press), 123 – 50.
——— (2000), Taxation, Wealth, and Saving (Cambridge:
MIT Press).
Congressional Budget Office (1997), The Economic
Effects of Comprehensive Tax Reform (Washington, D.C.:
Government Printing Office).
Diamond, John, and George R. Zodrow (1999), “Housing
and Intergenerational Redistributions Under a Consumption Tax Reform,” Proceedings, Ninety-First Annual
Conference on Taxation (Washington, D.C.: National Tax
Association), 25 – 31.

Bradford (2000, 104 – 5; 1996, 138), Huffman and
Koenig (1998, 27–29), Lyon and Merrill (1999, 307– 09),
Diamond and Zodrow (1999, 26), Gentry and Hubbard

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Bradford (2000, 105, 107; 1996, 138, 140), Hall (1997,
149 – 50), Gentry and Hubbard (1997, 12, 43), and
Joint Committee on Taxation (1995, 85) discuss the
effects of tax reform on lightly taxed capital.

Feenberg, Daniel R., Andrew W. Mitrusi, and James M.
Poterba (1997), “Distributional Effects of Adopting a
National Retail Sales Tax,” Tax Policy and the Economy,

vol. 11, ed. James M. Poterba (Cambridge: MIT Press),
49 – 89.

Lewis, Kenneth A., and Laurence S. Seidman (2000),
“Transitional Protection During Conversion to a Personal
Consumption Tax,” Public Finance Review 28 (March):
99 –119.

Gale, William G. (1999), “The Required Tax Rate in a
National Retail Sales Tax,” National Tax Journal 52
(September): 443 – 57.

Lyon, Andrew B. (1990), “Invariant Valuation When Tax
Rates Change Over Time: Confirmations and Contradictions,” Journal of Political Economy 98 (April): 433 – 37.

Gentry, William M., and R. Glenn Hubbard (1997),
“Distributional Implications of Introducing a Broad-Based
Consumption Tax,” Tax Policy and the Economy, vol. 11,
ed. James M. Poterba (Cambridge: MIT Press), 1– 47.

Lyon, Andrew B., and Peter R. Merrill (1999), “The Stock
Market and Consumption Tax Reform,” Proceedings,
Ninety-First Annual Conference on Taxation (Washington,
D.C.: National Tax Association), 307–13.

Gillis, Malcolm, Peter Mieszkowski, and George R.
Zodrow (1996), “Indirect Consumption Taxes: Common
Issues and Differences Among the Alternative
Approaches,” Tax Law Review 51 (Summer): 725 –74.

McLure, Charles E., Jr., and George R. Zodrow (1996),
“A Hybrid Approach to the Direct Taxation of Consumption,” in Frontiers of Tax Reform, ed. Michael J. Boskin
(Stanford: Hoover Institution Press), 70 – 90.

Gravelle, Jane G. (1996a), “The Distributional Effects of
Fundamental Tax Revisions,” San Diego Law Review
33 (Fall): 1419 – 57.

Samuelson, Paul A. (1964), “Tax Deductibility of Economic
Depreciation to Insure Invariant Valuations,” Journal of
Political Economy 72 (December): 604 – 06.

——— (1996b), “Effects of Flat Taxes and Other
Proposals on Housing,” Congressional Research
Service Report for Congress (April).

Sandmo, Agnar (1979), “A Note on the Neutrality of the
Cash Flow Corporate Tax,” Economics Letters 4 (2):
173 – 76.

Hall, Robert E. (1997), “Potential Disruption from the
Move to a Consumption Tax,” American Economic
Review Papers and Proceedings 87 (May): 147– 50.

Sinn, Hans-Werner (1991a), “Taxation and the Cost of
Capital: The ‘Old’ View, the ‘New’ View, and Another
View,” in Tax Policy and the Economy, vol. 5, ed. David F.
Bradford (Cambridge: MIT Press), 25 – 54.

Huffman, Gregory W., and Evan F. Koenig (1998), “The
Dynamic Impact of Fundamental Tax Reform, Part 2:
Extensions,” Federal Reserve Bank of Dallas Economic
Review, Second Quarter, 19 – 31.

——— (1991b), “The Vanishing Harberger Triangle,”
Journal of Public Economics 45 (August): 271– 300.

Johansson, Sven-Erik (1969), “Income Taxes and Investment Decisions,” Swedish Journal of Economics 71
(June): 104 –10.

Slemrod, Joel (1996), “Which Is the Simplest Tax System
of Them All?” in Economic Effects of Fundamental Tax
Reform, ed. Henry J. Aaron and William G. Gale
(Washington, D.C.: Brookings Institution), 355 – 84.

Joint Committee on Taxation (1995), Description and
Analysis of Proposals to Replace the Federal Income
Tax, JCS-18-95 (Washington, D.C., June).

Sullivan, Martin A. (1996), “Housing and the Flat Tax:
Visible Pain, Sudden Benefits,” Tax Notes 74 (January 22):
340 – 45.

King, Mervyn A. (1974), “Taxation and the Cost of
Capital,” Review of Economic Studies 41 (January):
21– 35.

Ventura, Gustavo (1999), “Flat Tax Reform: A Quantitative
Exploration,” Journal of Economic Dynamics and Control
23 (September): 1425 – 58.

Koenig, Evan F. (1999), “Achieving ‘Program Neutrality’
Under a National Retail Sales Tax,” National Tax Journal
52 (December): 683 – 97.

Zodrow, George R. (1991), “On the ‘Traditional’ and ‘New’
Views of Dividend Taxation,” National Tax Journal 44
(December): 497– 509.

Koenig, Evan F., and Gregory W. Huffman (1998), “The
Dynamic Impact of Fundamental Tax Reform, Part 1: The
Basic Model,” Federal Reserve Bank of Dallas Economic
Review, First Quarter, 24 – 37.

FEDERAL RESERVE BANK OF DALLAS

The United States consumes 8.5 million
barrels of gasoline daily—nearly half its daily
consumption of all petroleum products. The
average automobile tank is filled weekly, and
gasoline prices are posted at every street corner
where there is a gasoline station. Consequently,
most U.S. consumers are very aware of movements in gasoline prices and closely observe
the asymmetry when crude oil and gasoline
prices fluctuate. Many consumers complain that
gasoline prices rise more quickly when crude
oil prices are rising than they fall when crude
oil prices are falling, exhibiting an asymmetric
relationship.1 To the naked eye, movements
in spot crude oil and retail gasoline prices
may lend some credence to consumers’ complaints (Figure 1 ).
Furthermore, in some instances when
gasoline prices have risen sharply and swiftly
following a rise in crude oil prices —such as
occurred in 1999 and 2000 and during the Gulf
War in 1990—consumers and politicians have
called for policies to put a stop to what is seen as
unfair pricing practices for petroleum products.2
Such reactions seem to stem from a popular suspicion that large, integrated companies have
monopolized the oil industry. The public seems
to take the asymmetric relationship between
gasoline and crude oil prices as evidence that
the petroleum industry is monopolistic.
Most of the previous research on the subject confirms at least part of what consumers
suspect: it provides econometric evidence of an
asymmetric relationship between gasoline and
crude oil prices. This article extends inquiry into
the issue by considering competing explanations for the asymmetry. The available evidence

Gasoline and Crude Oil Prices:
Why the Asymmetry?
Stephen P. A. Brown and Mine K. Yücel

he available evidence

suggests that asymmetry
is unlikely to be the result
of monopoly power exercised
by large, integrated
oil companies.

Figure 1

Detrended Crude Oil and
Retail Gasoline Prices
Cents per gallon

Dollars per barrel

25
Crude oil,
spot price, WTI

Stephen P. A. Brown is director of energy economics
and microeconomic policy analysis in the Research
Department at the Federal Reserve Bank of Dallas.
Mine K. Yücel is a senior economist and
research officer in the Research Department
at the Federal Reserve Bank of Dallas.

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

–10

–5
Unleaded regular,
self-serve gasoline price

–20

–10

–30

–15
’85

’87

’89

’91

’93

’95

’97

SOURCES: Department of Energy; Haver Analytics.

’99

suggests that asymmetry is unlikely to be the
result of monopoly power exercised by large,
integrated oil companies. An examination of
the possible explanations for the asymmetry
also suggests that government intervention to
prevent the asymmetry between gasoline and
crude oil prices is likely to reduce economic
efficiency.

shown in Panel 1 of Figure 2. In contrast, retail
gasoline prices respond only gradually to a
falling crude oil price (Panel 2). The net effect
is an asymmetric response in gasoline prices
(Panel 3). Retail gasoline prices respond more
quickly when crude oil prices are rising than
when they are falling.
EXPLANATIONS OF ASYMMETRY

THE EVIDENCE FOR ASYMMETRY

With a number of studies showing that
gasoline prices respond more quickly when
crude oil prices rise than when they fall, economists have offered numerous explanations for
the phenomenon.4 Explanations include market power, search costs, consumer response to
changing prices, inventory management, accounting practices, refinery adjustment costs, and the
behavior of markups over the business cycle.
For the gasoline markets, however, no one has
posited a formal econometric test that would
allow the testing of the various explanations—
including market power —for price asymmetry
against the available data. In the absence of
such tests, judgment and economic theory
must be used to sort through the explanations
and determine whether the asymmetric response
of gasoline prices to movements in crude oil
prices is the result of market power or more
benign forces.

Most of the previous research provides
econometric support for public claims that gasoline prices rise more quickly when crude oil
prices are rising than they fall when crude
prices are falling. Bacon (1991) finds asymmetry
for the UK gasoline market. Karrenbock (1991);
French (1991); Borenstein, Cameron, and Gilbert
(1997); Balke, Brown, and Yücel (1998); and a
GAO report (1993) all find some evidence for an
asymmetric response in U.S. gasoline markets.3
In contrast with the other studies, Norman and
Shin (1991) find a symmetric response in U.S.
gasoline markets.
Of these studies, one of the most visible
and comprehensive is that of Borenstein,
Cameron, and Gilbert (1997), hereafter identified
as BCG. They use weekly and biweekly data
from 1986 to 1992 in a series of bivariate errorcorrection models to test for asymmetry in price
movements between gasoline’s various stages
of production and distribution—from crude oil
through the refinery to the retail pump. They
find strong and pervasive evidence of asymmetry in all segments of the market.
Shin (1992) argues, however, that the
periodicity of the data, the sample period of
estimation, and the model specification may
affect the results obtained in various studies. To
examine the issues that Shin raises, Balke,
Brown, and Yücel (1998), hereafter identified as
BBY, extend the work of BCG by using several
different model specifications and various subsamples of weekly data from 1987 through early
1996. BBY confirm BCG’s supposition that most
of the price volatility originates upstream (in or
closer to markets for crude oil) rather than
downstream (in or closer to final consumer markets). They also find that asymmetry is sensitive
to model specification but not to sample period.
With their most preferred specification, however, BBY find evidence that asymmetry is pervasive across the stages of gasoline production
and distribution.
For example, BBY find retail gasoline
prices initially rise sharply after the crude oil
price rises and then increase more gradually, as

Market Power

Market power is probably the greatest
concern to those who observe that gasoline
prices respond more quickly when crude oil
prices rise than when they fall. For the banking
industry, Neumark and Sharpe (1992) show that
market concentration is an explanatory variable
for the asymmetry found in interest rate movements. In a comprehensive study of U.S. industry, however, Peltzman (2000) finds no evidence
that market power is related to price asymmetry.
In addition, neither we nor Peltzman could find
a theoretical model that relates market power to
an asymmetric response of downstream prices
to changes in upstream prices.5 Were such a
model to exist, it might involve consumer search
costs or firms concerned with maintaining a tacit
collusion or both.
Consider an industry with a few dominant
firms that are engaged in an unspoken collusion
to maintain higher profit margins. Reputation
can be important to maintaining such a tacit
agreement (Tirole 1990). If the firms value the
agreement and have imperfect knowledge of
the upstream prices their competitors are paying, each firm would face an asymmetric loss

FEDERAL RESERVE BANK OF DALLAS

gasoline outlets in 1996. Of these, 114,452 were
branded outlets (that is, they sold brand-name
gasoline) belonging to 21 companies with at
least 1,000 outlets each. Citgo, a subsidiary of
the Venezuelan PDVSA, had the most retail outlets, with 14,529 in 48 states; Texaco came in
second with 13,785 outlets in 25 states. The top
six companies had 55 percent of the branded
market and 33 percent of the total retail market,
none of which provides strong evidence of

function where it would be more reluctant to
lower its selling price than to raise it. When
upstream prices rise, each firm is quick to raise
its selling price because it wants to signal its
competitors that it is adhering to the tacit agreement by not cutting its margin. When the
upstream price falls, each firm is slow to lower
its selling price because doing so runs the risk
of sending a signal to its competitors that it is
cutting its margin and no longer adhering to the
tacit agreement. In the gasoline markets, such
an explanation could be applied to each upstream price and its adjacent downstream price.
Despite popular wisdom and an explanation linking concentration to the asymmetry
between movements in crude oil and gasoline
prices, there does not appear to be much evidence of monopolization in any segment of the
gasoline market. The United States consumed
123 billion gallons of gasoline in 1996. The market share claimed by the four largest gasoline
refiner/marketers (37.7 percent), as well as a relatively low Herfindahl–Hirschman Index of 650,
suggests that U.S. gasoline production is competitive when viewed at the national level.6
Because refined products are harder and
more expensive to ship than crude oil, however,
gasoline markets tend to be regionalized. In
addition, regional variation in the environmental
regulation of gasoline formulation may be increasing the regionalization of gasoline markets.
Furthermore, changes in technology and environmental regulation have caused some smaller
refiners to go out of business and increased the
market share of the remaining refiners —most
notably in California, where the clean air rules
are more stringent than the national average
and the number of refiners has decreased (from
31 in 1990 to 23 in 1996).
If gasoline markets were strictly regional,
the number of refiners serving a region would
be limited by the size of the regional market
and economies of scale. In those regions with a
few refiners, market power would be a possibility. Nonetheless, gasoline shipments between
regions seem sufficient to establish workable
competition in most areas, and in most regions
of the country one can find a number of competing brands of gasoline.
The case for market power also seems
difficult to make for the retail sector. In rural
areas and small towns, regional monopolies
could exist, and gasoline stations have often
been cited as examples of monopolistic competition. But, the sheer number of retail gasoline
stations makes complete monopolization unlikely. The United States had 190,246 retail

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Figure 2

Asymmetric Response of Retail Gasoline
Prices to Movements in Crude Oil Prices
Percent

Panel 1: Response to Rising Prices
(with confidence bands)
.40
.35
.30
.25
.20
.15
.10
.05
0

10 11 12 13 14 15 16

Weeks

Panel 2: Response to Falling Prices
(with confidence bands)
.10
.05
0
–.05
–.10
–.15
–.20
–.25
–.30
–.35

10 11 12 13 14 15 16

Weeks

Panel 3: Difference in Response
(with confidence bands)
.25

.20

.15

.10

.05

10 11 12 13 14 15 16

Weeks
SOURCE: Balke, Brown, and Yücel (1998).

Components of the Retail Gasoline Price
The cost to produce and deliver
gasoline to consumers includes the cost
of crude oil to refiners, refinery processing
costs, marketing and distribution costs,
and retail station costs and taxes. In July
2000 crude oil costs made up 44 percent
of the total cost of gasoline. Refining costs
and profits were 13 percent, and distribution,
marketing, and retail costs were 16 percent.
Federal and state taxes (not including
county and local taxes) were 27 percent of
the total price, on average. Both federal
and state taxes have been increasing in the
past two decades. After staying constant at
4 cents per gallon until 1983, federal taxes
rose gradually to 18.4 cents per gallon by
1994, where they have remained. State
taxes have increased steadily from the
1920s to the current rate of 19.96 cents.

What Do We Pay For in a Gallon
of Regular Grade Gas?

car consumes 504 gallons of gasoline per year.
For a person filling up the tank every week, that
comes to 9.7 gallons per week. The price differential between gasoline stations is usually not
more than a couple of cents. If the difference were
10 cents (which is much higher than average), it
would amount to 97 cents per week, about the
price of a cup of coffee, which is likely to be
less than the value of the time used in an
aggressive search for lower-priced gasoline.8

(July 2000)
Retail price: $1.551
Refining costs
and profits

13%

Distribution and
marketing costs
and profits

16%

Taxes

27%

More Benign Explanations

Crude oil

Beyond market power and search costs,
economists have offered a number of explanations for the asymmetric response of gasoline
prices to movements in crude oil prices. Alternative explanations include markups that vary
over the business cycle, consumer response to
changing prices, inventory management,
accounting practices, and refinery adjustment
costs. Other than the variation in markups over
the business cycle, none of the explanations can
be ruled out on either theoretical or empirical
grounds.
If markups vary over the business cycle,
the difference between the crude oil and retail
gasoline price could increase as overall prices
rise. Reagan (1982) and Reagan and Weitzman
(1982) offer a theoretical explanation for such
a relationship based upon the variation in demand over the business cycle. Haltiwanger and
Harrington (1991) further suggest that the fluctuations in margins may result from variations in
the degree of collusive behavior. However, BBY
find that the shocks to crude oil and gasoline
prices originate with supply rather than demand,
which renders the explanation inapplicable.
The consumer response to changing
gasoline prices may contribute to the asymmetry
between movements in crude oil and gasoline
prices at the retail level. If consumers accelerate
their gasoline purchases to beat further increases
when its price is rising, they will increase inventories held in automobiles and quicken the pace
at which the price rises. If drivers fear running
out of gasoline and do not slow their purchases
when its price is falling by as much as they
accelerated their purchases when prices rose,
the price of gasoline will fall more slowly than
it rose.
Similarly, firms in the oil industry may
view the short-run costs of unexpected changes
in their inventories as asymmetric (see BCG). If
operation costs rise sharply when inventories
are reduced below normal operating levels, a reduction of upstream supply could lead a firm to
raise its output prices aggressively to prevent a

44%

SOURCE: Energy Information Administration,
Office of Oil and Gas, online publication
“A Primer on Gasoline Prices.”

market concentration or market power. Nonetheless, Borenstein and Shepard (1993) find some
evidence of coordinated pricing in a study using
data from 1986–91 for 59 U.S. cities.
Limited Market Power and Search Costs

In the retail gasoline market, consumer
search costs could lead to temporary market
power for gasoline stations and an asymmetric
response to changes in the wholesale price of
gasoline. (See BCG, Norman and Shin 1991,
Borenstein 1991, Deltas 1997, and Peltzman
2000.) Each gasoline station has a locational
monopoly that is limited by consumer search.
After consumers have searched, the profit margins at each gasoline station are pushed down
to a roughly competitive level. When wholesale
prices rise, the owner of each station acts to
maintain profit margins and quickly passes the
increase on to customers.7 When wholesale
prices fall, however, each station temporarily
boosts its profit margins by slowly passing the
decrease on to customers. Only after the customers engage in a costly and time-consuming
search to find the lowest prices are the stations
forced to lower prices to a competitive level.
A factor slowing the search process is that
the costs of an intensive search are likely to be
much higher for most consumers than the corresponding gains from finding a cheaper price
for gasoline. The money saved is a very small
part of the consumer’s budget, so that consumers will not search unless the price differential is very high. How large is this differential for
the average consumer? The average passenger

FEDERAL RESERVE BANK OF DALLAS

loss of inventories. If an increase in inventories
above normal operating levels has a relatively
small effect on costs, the firm could be less
aggressive in reducing its selling prices when it
experiences an increase in upstream supply.
Hence, inventories would buffer downstream
price movements less when prices are rising
than when they are falling.
If oil supply shocks cause asymmetric
movements in inventories—with higher inventories when oil supply is plentiful and lower
inventories when oil supply is reduced —the
asymmetry of price movements could be
enhanced by FIFO (first in, first out) accounting.
If inventories are lower when upstream supply
is reduced, the firm will sell the products incorporating the higher upstream price sooner. If
inventories are higher when upstream supply is
increased, the firm will sell the products incorporating the lower upstream price later. These
actions help foster asymmetric pricing.
Refiners also face adjustment costs to
changing their output or their product mix and,
consequently, adjust their output slowly when
possible. When crude oil supplies are reduced,
refiners as a group have little choice but to
reduce output quickly, which would lead to
fairly quick increases in gasoline prices. When
crude oil supplies are increased, however, refiners don’t necessarily have to increase output
quickly. They can increase output slowly and
delay the decreases in gasoline prices.

If the monopolization of gasoline markets is a
concern, policies will be more effective directed
at monopolization than at market phenomena
that can be the result of either competitive or
monopolized markets.
Refining and Wholesale Markets

Because there is little evidence of monopolization in the refinery and wholesale markets
for gasoline, the observed asymmetry between
wholesale gasoline and crude prices is most
likely the result of competitive market forces.
Calculations based on the BBY estimates also
suggest the degree of asymmetry of response in
wholesale gasoline prices to changes in crude
oil prices is quite small and of short duration.
Given a 1 percent increase and a 1 percent decrease in the crude oil price, the difference in
response of wholesale gasoline to these changes
is only 0.35 percent and persists only for two
weeks. The asymmetry of response in wholesale gasoline prices starts around the third week
and becomes insignificant around the fifth week.
If competitive market forces account for the
asymmetry between wholesale gasoline and
crude oil prices, any policies to eliminate it are
quite likely to involve higher costs than living
with the asymmetry.
Even if it is the result of market power,
the asymmetry is so fleeting that the likely costs
of the unintended consequences of a policy to
prevent price asymmetries probably would outweigh the benefits. If policymakers are concerned about the monopolization of refinery or
wholesale markets for gasoline, the most prudent policy is to watch for mergers that increase
market concentration without providing gains in
the economies of scale, rather than to take
direct steps to suppress asymmetry.

THE POLICY RESPONSE

If we adhere to the traditional view that
economic policy should be directed only at market failures or imperfections, policy probably
should not be directed at eliminating the asymmetry between crude oil and retail gasoline
prices. The evidence of monopolization in refining and wholesale markets for gasoline is weak
at best. Peltzman (2000) finds that asymmetry
itself is not indicative of a monopolized market.
Any market power that might exist at the retail
level appears to be related to the costs of product differentiation — most likely in the form of
locational differences.
Furthermore, Peltzman finds that an asymmetric relationship between an upstream and
a downstream price is as likely in competitive
markets as in markets thought to be monopolized. If competitive market forces and asymmetry coexist, steps to suppress or eliminate the
asymmetry are likely to prove costly because
government interference in natural market processes typically reduces economic efficiency.9

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Retail Markets

Compared with the upstream markets,
price asymmetries in the retail market are longer
in duration and smaller in magnitude. Locational differentiation and consumer search costs
could contribute to market power, and Borenstein and Shepard (1993) find evidence of coordinated pricing in the retail gasoline market. But,
asymmetric pricing can arise whether or not
there is market power. Consequently, the benefits of policies to eliminate asymmetry in the
retail gasoline market are likely to be small,
while the costs could be high.
Calculations made with the BBY estimates
suggest that a 1 percent increase and a 1 percent decrease in the price of oil lead to a peak
differential of only 0.2 percent in the response

of the retail gasoline price. To illustrate, suppose the current prices for oil and gasoline are
$30 per barrel and $1.50 per gallon, respectively. The peak difference in the response of
the retail gasoline price to a $6 increase and
decrease in the per barrel price of crude oil
would be only 6 cents per gallon.10 For the average driver, this differential would amount to
about 60 cents in the peak week. Because the
differential is so small and search costs are high,
it is not surprising that the price asymmetry persists longer than 16 weeks.
Since there is no evidence or theory suggesting that asymmetry necessarily arises from
market power in the retail market, policies
aimed at eradicating asymmetry are likely to
reduce efficiency. Even a simple policy of requiring retail margins to remain constant over
time could have unintended consequences for
inventories and lead to shortages when prices
are rising. More complicated policies would
be more difficult to administer. Again, the best
policy seems to be to watch for mergers that
increase market concentration, rather than to
take direct steps to suppress the asymmetry.

NOTES
The authors thank Nathan Balke, Mark French,
Bernard Gelb, Roger Hemminghaus, Donald Norman,
Pia Orrenius, Jason Saving, and Mark Wynne for
helpful comments and suggestions while retaining
responsibility for any errors or omissions.
1

Previous research does not find this type of asymmetry. Increased environmental regulation of refinery
operations and increased taxation of gasoline appear
to have been offset by productivity gains. See Borenstein, Cameron, and Gilbert (1997) and Balke, Brown,
and Yücel (1998).

For examples, see Ferguson (2000) and Ivanovich
(2000).

Peltzman (2000) finds that the fuel component of the
consumer price index responds asymmetrically to the
fuel component of the producer price index.

Pricing asymmetries have been observed in many
industries, including banking (Neumark and Sharpe
1992) and agriculture (Mohanty et al. 1995). Peltzman
(2000) finds pricing asymmetry exists in about twothirds of U.S. industry.

Variations of the kinked-demand model of oligopoly
do not suggest an asymmetrical movement in the
output price of an industry in the response to common
shocks to the input prices of the firms in that industry.
The model explains why prices are less likely to
change in either direction. See Scherer (1980) and
Neumark and Sharpe (1992).

The Herfindahl – Hirschman Index (HHI) is a summary
measure of market concentration.

CONCLUSIONS

A number of econometric studies confirm
casual observations that gasoline prices respond
asymmetrically to crude oil price movements
by rising more quickly when crude oil prices
are rising than falling when crude oil prices
are falling. Although popular opinion seems to
attribute the asymmetry to market power,
Peltzman (2000) shows that price asymmetries
arise independently of market structure. In addition, no formal theory relating market power to
asymmetry has been tested (to our knowledge),
nor is there much evidence of concentration in
U.S. markets for gasoline. Consumer search costs
and locational advantages may provide market
power to some retailers, but such market power
might be viewed as the costs of product differentiation under monopolistic competition.
With the evidence pointing away from
market power as an explanation, asymmetry is
likely to be the consequence of other market
factors. As such, policies to suppress asymmetric price movements are likely to lead to undesirable outcomes. If one is concerned about
market power in the production, distribution,
and marketing of gasoline, the best policy
seems to be watching for mergers that increase
market concentration without increasing economies of scale, rather than taking direct steps
to suppress asymmetry.

HHI = ∑ S 2i ,
i =1

where Si is the market share of the i th firm. A monopolistic industry with one firm would have an HHI of
10,000 (where market shares are measured in percentage terms).
7

Given that retail margins are very small (see Deltas
1997 and BCG), a large increase in input prices could
quickly turn margins negative. Hence, retailers hasten
to pass on input price increases.
The low individual costs associated with limited market
power should not be taken as an argument that
asymmetry has little aggregate cost. We are simply
pointing out that the individual benefit – cost calculations made by rational individuals are likely to result in
a relatively slow search. When multiplied by the tens
of millions of people who drive on a daily basis, the
aggregate costs of asymmetry are significant, but
these costs are presumably lower than the aggregate
search costs that would be necessary to eliminate
asymmetry.
Such policies might include government manipulation
of inventories or a requirement that oil companies,
distributors, and retailers use LIFO (last in, first out)

FEDERAL RESERVE BANK OF DALLAS

pricing for gasoline with constant markups over time.
To the extent that either policy interfered with free market outcomes, implementation would reduce economic
efficiency. A policy of varying taxes inversely with oil
prices would be ineffective in eliminating asymmetry
because it would not produce additional gasoline
when prices are rising.
Six dollars per barrel is equal to the variance of oil
prices in the past 10 years.

Ivanovich, David (2000), “Inquiry into Midwest GasolinePrice Surge Is Still Pending, Commission Says,” Houston
Chronicle online, July 30.
Karrenbock, Jeffrey D. (1991), “The Behavior of Retail
Gasoline Prices: Symmetric or Not?” Federal Reserve
Bank of St. Louis Review, July/August, 19 – 29.
Mohanty, Samarendu, E. Peterson, F. Wesley, and Nancy
Cottrell Kruse (1995), “Price Asymmetry in the International Wheat Market,” Canadian Journal of Agricultural
Economics 43 (November): 355 – 66.

REFERENCES
Bacon, Robert W. (1991), “Rockets and Feathers: The
Asymmetric Speed of Adjustment of UK Retail Gasoline
Prices to Cost Changes,” Energy Economics 13 (July):
211–18.

Neumark, David, and Steven A. Sharpe (1992), “Market
Structure and the Nature of Price Rigidity: Evidence from
the Market for Consumer Deposits,” Quarterly Journal
of Economics 107 (May): 657– 80.

Balke, Nathan S., Stephen P. A. Brown, and Mine K. Yücel
(1998), “Crude Oil and Gasoline Prices: An Asymmetric
Relationship?” Federal Reserve Bank of Dallas Economic
Review, First Quarter, 2 –11.

Norman, Donald A., and David Shin (1991), “Price
Adjustment in Gasoline and Heating Oil Markets,”
American Petroleum Institute Research Study no. 060
(Washington, D.C., August).

Borenstein, Severin (1991), “Selling Costs and Switching
Costs: Explaining Retail Gasoline Margins,” Rand Journal
of Economics 22 (Autumn): 354 – 69.

Peltzman, Sam (2000), “Prices Rise Faster Than They
Fall,” Journal of Political Economy 108 (June): 466 – 502.

Borenstein, Severin, A. Colin Cameron, and Richard
Gilbert (1997), “Do Gasoline Prices Respond Asymmetrically to Crude Oil Prices?” Quarterly Journal of Economics
112 (February): 305 – 39.

Reagan, Patricia (1982), “Inventory and Price Behavior,”
Review of Economic Studies 49 (January): 137– 42.
Reagan, Patricia, and Martin Weitzman (1982), “Asymmetries in Price and Quantity Adjustments by the
Competitive Firm,” Journal of Economic Theory 27
(August): 410 – 20.

Borenstein, Severin, and Andrea Shepard (1993),
“Dynamic Pricing in Retail Gasoline Markets,” NBER
Working Paper Series no. 4489 (Cambridge, Mass.:
National Bureau of Economic Research, October).

Scherer, F. M. (1980), Industrial Market Structure and
Economic Performance, 2nd ed. (Chicago: Rand McNally).

Deltas, George (1997), “Retail Gasoline Price Response
Asymmetries to Wholesale Price Shocks” (Paper presented at Western Economic Association Meeting,
Seattle, July).

Shin, David (1992), “Do Product Prices Respond
Symmetrically to Changes in Crude Oil Prices,” American
Petroleum Institute Research Study no. 068 (Washington,
D.C., December).

Ferguson, Ellyn (2000), “House Seeking Answers on High
Gas Prices,” USA Today online, June 22.

Tirole, Jean (1990), The Theory of Industrial Organization
(Cambridge, Mass.: MIT Press).

French, Mark (1991), “Asymmetry in Gasoline Price
Changes” (Washington, D.C.: Board of Governors of the
Federal Reserve System, August, Draft paper).

U.S. General Accounting Office (1993), “Analysis of
the Pricing of Crude Oil and Petroleum Products,”
GAO/RCED-91-17 (Washington, D.C.: Government
Printing Office, March).

Haltiwanger, John, and Joseph E. Harrington Jr. (1991),
“Impact of Cyclical Demand Movements on Collusive
Behavior,” Rand Journal of Economics 22 (Spring):
89 –106.

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Efforts requiring banks to divulge information on their financial condition have a long history. In the early 1800s, some states required
banks to file reports of condition with the governor or legislature, arguing the state was a
shareholder in the banks and therefore entitled
to the information (Robertson 1995). However,
the reports contained only broad breakdowns
of assets and liabilities and no information bank
directors did not wish to disclose. In 1869,
Congress empowered the comptroller of the
currency to “call” for a full statement of condition from national banks several times a year.
Regulators have since made many changes to
the resulting call report, but its purpose remains the same—to provide timely information
regarding the condition of banks.
The modern call report, or Report of
Condition and Income, is filed quarterly by all
banks and contains hundreds of accounting
items that regulators and private analysts use to
characterize the financial condition of both individual banks and the industry. Call reports now
include detailed measures of assets, liabilities,
revenues, expenses, and off-balance-sheet activity. The level of detail is somewhat greater for
large banks than for small ones, but even small
banks file an extensive report.
This article analyzes call report revisions
to assess the extent to which regulatory exams
promote accurate data. If the loan-loss accounting in call reports is widely used to measure
loan quality, the findings support the view
that exams are important in the public dissemination of accurate information on banks’ financial condition.

Financial Statements and Reality:
Do Troubled Banks Tell All?
Jeffery W. Gunther and Robert R. Moore

his analysis provides

direct evidence of exams’
significant role in uncovering
financial problems and
ensuring bank accounting
statements reflect them.

DISCLOSURE OF LOAN LOSSES
An old saw in banking is that making loans
is easy, but getting paid back is hard. While
banks have developed a substantial tool kit for
identifying creditworthy borrowers, all loans
entail some risk and inevitably, some will not be
repaid. A loan that appears sound at origination
may deteriorate in quality and eventually become
a loan for which repayment is highly unlikely.
Because loans are a primary banking product, a
true picture of a bank’s overall financial condition often depends on the accuracy with which
loan portfolio problems have been identified and
measured. As a result, the accuracy of the line
items pertaining to loan quality and performance
has the potential for determining a call report’s
usefulness in tracking financial developments.
The banking industry uses a specialized
system to account for loan quality problems, at

Jeffery W. Gunther is a senior economist and research
officer and Robert R. Moore is a senior economist
and policy advisor in the Financial Industry Studies
Department at the Federal Reserve Bank of Dallas.

FEDERAL RESERVE BANK OF DALLAS

the heart of which is the allowance for loan and
lease losses (ALLL). Through provision for loan
and lease losses, banks add funds to ALLL. These
provisions are an expense item and reduce a
bank’s net income.1 The ALLL balance is subtracted from total loans, so that loans on the
balance sheet are reported net of ALLL. When
loans are charged off, total loans are reduced by
the amount of the losses, but the losses are
charged against ALLL, leaving net loans unaffected.2 If a bank recovers some of the losses
on loans previously charged off, the recoveries
are added back to ALLL.
When a bank charges off a loan, the
resulting loss does not affect reported profitability, since the charge-off is against ALLL.
Credit quality problems affect reported profitability when a bank incurs the provision expense, since the expense directly reduces net
income. As a result, timely disclosure of information on credit quality and its impact on overall operating results depends on the degree to
which provisions are made in anticipation of, or
concurrent with, actual impairment in the loan
portfolio. If adequate provisions are made only
after the impairment occurs, profitability prior to
the provisions is overstated.
Regulatory guidance directs banks to
make provisions if ALLL is insufficient to absorb
estimated credit losses. However, the definition
of estimated credit losses highlights the difficulty
in pinpointing an appropriate level for provisions: “Estimated credit losses are anticipated
losses that are reasonably expected to occur but
whose amounts or obligors cannot be specifically identified” (Federal Reserve Board of Governors 1999). Because assessing the adequacy
of ALLL and the need for provisions is based
on an estimate of losses, in many cases it may
only be possible to determine a range of suitable levels for provision expense, rather than
the single most appropriate level. In addition,
exam findings can lead banks to charge off
some existing loans, thereby reducing ALLL and
potentially requiring additional provisions.

were to deteriorate. This form of income smoothing might lead outside observers and investors
to regard banks as more stable and less risky
than they are.3 An undesirable aspect of such
income smoothing is that it could make a bank’s
financial condition less transparent to the users
of financial statements.4
Potential tax benefits are another incentive
for manipulating loan-loss provisions. Before the
1986 Tax Reform Act, provisions were treated as
a deductible expense, and setting higher provisions often lowered taxable income. However,
for banks with assets over $500 million, the act
linked the amount of a bank’s deduction to its
actual charge-off experience (Walter 1991).
Another factor that might prompt banks to
set an inappropriate level of ALLL and provision
expense involves regulatory or market-based
penalties for a deterioration in financial condition. Risk-based capital requirements allow
banks to count ALLL only in Tier 2 capital and
only up to 1.25 percent of risk-weighted assets
(Kwan and O’Toole 1997). By not making the
necessary provisions, banks with asset-quality
problems can raise reported net income and retained earnings, thereby boosting Tier 1 capital
and potentially avoiding the numerous restrictions regulators typically place on troubled
banks.
Given the current institutional framework,
which assigns regulators a large role in the
monitoring and disciplining of banks, this latter
incentive provides a particularly strong reason
for regulatory exams. The Commercial Bank
Examination Manual states that “the examiner’s
responsibility to determine the adequacy of a
bank’s ALLL is one of the most important functions of any examination” (Federal Reserve
Board of Governors 1999). In verifying the
adequacy of ALLL, examiners consider information obtained during the current and prior
exams, loan quality trends and peer group data,
processes for internal credit review, past-due
and restructured loans, and economic conditions.
If after considering these factors an examiner
finds that a bank’s ALLL is too low, the institution normally is required to increase its provision expense and raise ALLL to the desired level.5
Several studies support the view that troubled banks often have insufficient ALLL and that
exams are important in helping correct the
problem. The General Accounting Office (1990,
1991) finds troubled or failing banks often have
insufficient ALLL. Similarly, Berger, King, and
O’Brien (1991) discuss the potential for insufficient ALLL, particularly when a bank has not
been examined recently. Gilbert (1993) provides

EXAMS AND TRANSPARENCY
Various incentives may induce banks to set
provisions outside the range commensurate with
credit quality. In response, policymakers have
sought to blunt or counteract these incentives.
Banks can reduce the variability of
reported income by making higher provisions
than necessary when credit quality and net
income are high. In this case, provisions would
not have to increase as much if credit quality

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Table 1

Provision Expense as a Percentage of Average Assets,
Year-End 1996–98
No revision
Reported
Percentile
25
50 (median)
75
Average
Percentage
of banks

Downward revisions
Reported

.03
.12
.24

.09
.26
.42

.18

.37

98.83

Revised

.17

benefits of exams in assessing loan quality and
the sufficiency of ALLL may be limited mostly to
the supervisory process itself, as opposed to the
promotion of financial transparency in general.
These earlier studies address important
issues. Do banks sometimes set provisions below what is needed to cover their loan losses? If
so, how often does this occur and by what magnitude do banks underreport? Are exams effective in promoting adequate levels of ALLL? And
finally, do provisions and ALLL convey useful
information to outsiders about banks’ financial
condition? The following analysis addresses
some of these issues further by providing evidence based on the incidence and size of revisions to call report data.

Upward revisions
Reported

Revised

.00
.16
.29

.11
.25
.49

.30
.52
1.00

.19

.42

.87
1.00

NOTES: The categories are based on revisions to the level of provisions. The data show the distribution
of the ratio of provision expense to average assets.
SOURCE: Original and revised files of the Report of Condition and Income, Federal Financial Institutions
Examination Council.

Sample
The analysis examines call report revisions
to gain insight into exams’ role in promoting
accurate financial information. Given the importance of information on loan quality in assessing banks’ overall financial condition, the analysis is limited to provision expense revisions. To
focus on safety and soundness concerns, most
of the analysis is directed at upward revisions. If
examiners determine the provision expense a
bank reports is inadequate, they may require
the bank to make additional provisions and
refile one or more recent call reports to reflect
the change.6 It is important to note, however,
that not all exam findings on provisions necessarily require call report revisions. If additional
provisions are necessary, the expense may simply be reflected on a bank’s subsequent call
report. Nevertheless, the revisions provide a
unique window through which to view the
results of exam activity.
The data this study uses are obtained from
files at the Federal Reserve Bank of Dallas and
are limited to commercial banks, year-end
1996–98. The originally reported data are from
files transmitted from the Federal Reserve
Board, seventy to eighty days following the
report dates. The revised data are for the same
report dates but were transmitted from the
Board in May 2000. Any differences between
the original data and the data obtained in May
reflect revisions made sometime after the data
were published as “final,” which typically occurs
about sixty-five days after the report date.7
Several additional restrictions frame the
analysis. First, the sample is limited to banks that
received a satisfactory rating on the last exam
prior to the report date.8 Focusing on these banks
facilitates an assessment of whether new or
emerging problems are freely divulged by banks

evidence of atypical movements in call report
data for the quarters in which banks are downgraded by examiners.
Setting provisions requires detailed knowledge about a bank’s loan portfolio. Regulators
and, especially, bank managers are more likely
than outsiders to have such detailed knowledge.
If an exam aligns provision expense and ALLL
with credit quality, it may facilitate the public
communication of important bank-specific
information and thereby enhance banking system transparency.
Consistent with this view, Docking, Hirschey, and Jones (1997) find a bank’s announcement of loan-loss provisions adversely affects
that bank’s stock price and sometimes the stock
price of other banks as well. Berger and Davies
(1998) provide evidence that quarterly financial
statements are a conduit for transmitting exam
findings to financial markets. And Flannery and
Houston (1999) find exams affect the relationship between a bank holding company’s market
and book value, possibly reflecting the improved accuracy of financial statements following an exam or a certification effect whereby
exams serve as a stamp of approval on published financial statements.
Other researchers have reached a different
conclusion, however, arguing essentially that
outsiders can see through a bank’s loan-loss
accounting and discern the true quality of its
loans, even if provisions and ALLL are lower than
necessary. Wall and Koch (2000) cite several
studies that indicate investors often do not react
to announcements of loan-loss provisions, presumably having already effectively estimated
the extent of the deterioration in bank loan
portfolios. If, without substantial cost, outsiders
can accurately estimate losses in a bank’s loan
portfolio on the basis of other information, the

FEDERAL RESERVE BANK OF DALLAS

Table 2

Percentage of Banks with Upward Revisions to
Provision Expense, Year-End 1996 – 98

or reported only at the behest of examiners. In
addition, banks less than four years old are
excluded, since young banks typically exhibit
unique financial characteristics and are not
directly comparable to more mature banks. The
resulting sample contains 24,519 year-end call
reports for the period from 1996 through 1998.

Quarter of first exam in subsequent year
Second

Third

Fourth

None

.76
1.32
.57
.40
1.26
1.37
.99
1.76
3.26
9.60

.00
.46
.43
.23
.21
.41
.69
1.83
3.02
3.37

.23
.95
.89
.23
.92
.00
.46
1.09
.93
1.84

.00
.28
.00
.00
.00
.00
.59
.59
1.56
1.12

.00
.70
.29
.28
.14
.56
.54
.71
.84
1.46

Asset-quality category
1
2
3
4
5
6
7
8
9
10

Results
The analysis reveals an interesting relationship between revisions to call report data
and exam results.
Frequency and Magnitude of Revisions. While
banks in general seldom revise their financial
reports, when they do, the revision is often substantial. As Table 1 shows, reported provision
expense was unrevised nearly 99 percent of the
time.9 Downward revisions were made to only
0.17 percent of the reports. However, banks
revised their provisions upward in 1 percent of
the cases examined, and these revisions tended
to be large. The median ratio of provision expense to average assets originally reported by
the banks that revised upward is 0.25 percent,
compared with a median ratio of 0.52 percent
based on the revised reports. The same comparison holds for the average of the revised provision expense ratio, which is more than twice
as high as originally reported. These revisions
are sufficient to lower reported profitability
appreciably. For the banks that revised their
provisions upward, the return on assets originally reported is 1.02 percent, on average, compared with 0.71 percent for the revised reports.
The number of upward revisions in our
sample is fairly small. However, these were
good times for the banking industry. Because
financial problems were few, the need for
increases in provisions could be expected to
have been low. The analysis below controls for
this factor by examining sound and troubled
banks separately.
Financial Problems, Exams, and Revisions. To
investigate whether revisions to provision
expense are driven by examiners’ findings, the
sample is divided into five groups. The first
group contains banks for which an exam began
in the first quarter of the year immediately following the fourth quarter report date. The second group is banks for which the first exam in
the subsequent year occurred in the second
quarter, and so on for the third and fourth
groups. The fifth group contains banks that
were not examined in the year following the
call report date.
The banks are also divided into ten assetquality categories. These categories are based

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

First

NOTES: The percentages are based on the number of upward revisions to the level of provisions.
The asset-quality categories are based on the ratio to assets of loans past-due thirty days or
more and still accruing and nonaccrual loans, where the first group is banks with the lowest
problem-asset ratios and the tenth group those with the highest ratios. The ratios are calculated
using revised data. Each asset-quality group contains 10 percent of the sample.
SOURCE: Original and revised files of the Report of Condition and Income, Federal Financial Institutions
Examination Council.

on the ratio to assets of loans past-due thirty
days or more and still accruing and nonaccrual
loans. The first group is banks with the lowest
problem-asset ratios and the tenth group those
with the highest ratios. The ratios are calculated
using revised data.10 Each group contains 10 percent of the sample.
This exercise indicates that banks with
severe asset-quality problems are more likely
than other banks to revise their loan-loss provision upward. In addition, banks that are
examined—particularly in the first quarter of
the year after the call report date—also are
more likely to raise their provision. As Table 2
shows, almost 10 percent of the banks in the
worst asset-quality group that were examined in
the first quarter revised their loan-loss provision
upward. In contrast, only 1.46 percent of the
banks in the worst asset-quality group that were
not examined in the year following the call
report date revised their provision upward.
Moreover, only 0.76 percent of the banks in the
best asset-quality group that were examined in
the first quarter revised their provision upward.
None of the unexamined banks in the best assetquality group raised their provision expense.
Further evidence that exams are a significant impetus for call report revisions is
obtained by dividing the banks into two categories based on whether they were downgraded by examiners to problem status. Of the
banks examined and downgraded in the first
quarter, 36 percent revised their loan-loss provision upward. In contrast, of the banks examined
in the first quarter that were not downgraded,

only 1.4 percent raised their provision expense. For banks examined in the fourth quarter, about 2 percent of those that were downgraded restated their provision at a higher level,
while under 1 percent of those that remained
in nonproblem status revised it upward. These
figures indicate banks that are downgraded by
examiners, particularly early in the year, often
revise the preceding year’s fourth quarter call
report to reflect a greater degree of financial difficulty than originally reported.11
Summary of Findings. These data provide a
look at the frequency and magnitude of call
report revisions and their relationship to exams.
Upward revisions of provisions are large
enough to reduce profitability appreciably. For
banks in general, the revisions are infrequent.
However, banks with new or emerging problems often significantly underreport provision
expense. There is a strong relationship between
examiner downgrades of banks and upward
revisions to the provision expense reported for
the previous year, especially when the downgrades occur early in the current year.

report data.) However, if call report information
does not accurately reflect financial conditions
when published, the report’s usefulness in
tracking financial developments between onsite exams could be reduced. Additionally, if
inaccuracies in the call report data are ultimately
corrected, the revisions might overstate the
report’s usefulness in tracking financial developments in real time, as Cole and Gunther
(1998) point out. These considerations suggest
the need to analyze early-warning models based
on originally published data to assess whether
these models’ ability to identify financial problems is appreciably lower than that of models
based on revised data.
NOTES
1

CONCLUSION
This analysis provides direct evidence of
exams’ significant role in uncovering financial
problems and ensuring bank accounting statements reflect them. The auditing role of exams
directly manifests itself in the difference
between original and revised call reports. For
the report dates used in the analysis, more than
one-third of the banks that fell into problem
status had to revise their most recent call report
to reflect a greater degree of financial difficulty than originally reported. To the extent outsiders use provisions and ALLL in assessing loan
quality, these results support the view that exams
are important in the public dissemination of accurate information on banks’ financial condition.
The findings also point to the need for
further research. Because call reports are filed
quarterly, whereas banks are typically examined
about once every twelve to eighteen months,
call report data potentially provide a more upto-date picture of a bank’s condition than onsite exams alone. For this reason, regulators use
call report data extensively in a variety of efforts
to monitor banks’ condition. One such effort
involves the construction and implementation of
statistical early-warning models to identify
emerging financial problems. These statistical
systems typically rely heavily on call report data
for input variables. (Cole, Cornyn, and Gunther
1995 provide an example of this use of call

Banks can use a reverse provision to remove funds
from ALLL.
For simplicity, the text refers exclusively to loan
performance. Losses on leases are treated similarly.
Greenawalt and Sinkey (1988) find evidence of income
smoothing and further discuss banks’ motivation for
the practice. See also Wall and Koch (2000).
An interagency statement on March 10, 1999, directs
banks to maintain “prudent, conservative, but not
excessive, loan-loss allowances that fall within an
acceptable range of estimated losses” (Securities and
Exchange Commission et al. 1999). The statement
also discusses plans for interagency cooperation in
issuing guidance on appropriate methodologies,
documentation, and disclosure.
In addition to examiner review, a bank’s loan-loss
accounting may be reviewed by independent auditors.
While all commercial banks are subject to exams, not
all are subject to external audits. The Federal Reserve
requires bank holding companies with consolidated
assets of $500 million or more to have an annual
external audit. New banks are also required to have
external audits. The Securities and Exchange Commission requires audits for publicly traded companies,
including bank holding companies. Finally, the Federal
Deposit Insurance Corporation Improvement Act of
1991 requires annual external audits for any bank
insured by the Federal Deposit Insurance Corp. with
assets greater than $500 million (Federal Reserve
Board of Governors 1994).
The analysis assumes the judgment of examiners is
correct — that is, an upward revision to provision
expense is taken to mean the initial level of ALLL was,
in fact, too low.
As regulators process call report data, substantial
effort is devoted to validating the reported information.
The primary goal is to ensure the data are accurate
before they are published as “final.” While the data
are typically published about sixty-five days after the

FEDERAL RESERVE BANK OF DALLAS

report date, revisions can be made for up to five
years.
Satisfactory status corresponds to a safety and
soundness rating of 1 or 2. Safety and soundness
ratings of 3, 4, and 5 are considered unsatisfactory.
In first quarter 1998, banks began reporting provision
for credit losses and no longer reported provision for
loan and lease losses. The new provision covers loan
and lease losses but also includes provisions for
losses on certain types of off-balance-sheet activity.
For simplicity, we refer to provision expense in all
years as provision for loan and lease losses. Banks
continue to report ALLL and now also report an allowance for credit losses, which includes the allowance
for losses on off-balance-sheet activity. Comparing the
two quantities makes it possible to estimate the size
of the provision for losses on off-balance-sheet activity.
The provision for losses on this activity is very small
overall in comparison with the provision for loan and
lease losses, and for the vast majority of banks the
provision for off-balance-sheet losses is zero.
Similar but not identical results are obtained when the
originally reported data are used to calculate the ratios.
Upward revisions to provision expense tend to occur
mostly at small and midsize banks. The incidence is
fairly equal for banks with assets under $100 million,
from $100 million to $500 million, and above $500
million but below $1 billion. Banks with assets of
$1 billion or more have a substantially lower incidence
of upward revisions. However, in our sample few
downgrades occur in the largest size category.
The relative lack of financial problems among the
large banks may help explain their low incidence of
upward revisions to provision expense.

Docking, Diane Scott, Mark Hirschey, and Elaine Jones
(1997), “Information and Contagion Effects of Bank
Loan-Loss Reserve Announcements,” Journal of
Financial Economics 43 (February): 219 – 39.
Federal Reserve Board of Governors (1994), “Internal
Control,” Commercial Bank Examination Manual, November.
——— (1999), “Allowance for Loan and Lease Losses,”
Commercial Bank Examination Manual, November.
Flannery, Mark J., and Joel F. Houston (1999), “The Value
of a Government Monitor for U.S. Banking Firms,” Journal
of Money, Credit, and Banking 31 (February): 14 – 34.
General Accounting Office (1990), “Bank Insurance
Fund: Additional Reserves and Reforms Needed to
Strengthen the Fund,” GAO/AFMD-90-100, September.
——— (1991), “Failed Banks: Accounting and Auditing
Reforms Urgently Needed,” GAO/AFMD-91-43, April.
Gilbert, R. Alton (1993), “Implications of Annual Examinations for the Bank Insurance Fund,” Federal Reserve
Bank of St. Louis Review, January/February, 35 – 52.
Greenawalt, Mary Brady, and Joseph F. Sinkey, Jr. (1988),
“Bank Loan-Loss Provisions and the Income-Smoothing
Hypothesis: An Empirical Analysis, 1976 – 84,” Journal of
Financial Services Research 1 (December): 301–18.
Kwan, Simon, and Randy O’Toole (1997), “Recent
Developments in Loan Loss Provisioning at U.S. Commercial Banks,” Federal Reserve Bank of San Francisco
Economic Letter, July 25.

REFERENCES

Robertson, Ross M. (1995), The Comptroller and Bank
Supervision: A Historical Appraisal (Washington, D.C.:
Office of the Comptroller of the Currency).

Berger, Allen N., and Sally M. Davies (1998), “The
Information Content of Bank Exams,” Journal of Financial
Services Research 14 (October): 117– 44.

Securities and Exchange Commission, Federal Deposit
Insurance Corporation, Federal Reserve Board, Office
of the Comptroller of the Currency, and Office of Thrift
Supervision (1999), Joint Interagency Letter to Financial
Institutions, March 10.

Berger, Allen N., Kathleen Kuester King, and James M.
O’Brien (1991), “The Limitations of Market Value
Accounting and a More Realistic Alternative,” Journal of
Banking and Finance 15 (September): 753 – 83.
Cole, Rebel A., Barbara G. Cornyn, and Jeffery W. Gunther
(1995), “FIMS: A New Monitoring System for Banking
Institutions,” Federal Reserve Bulletin, January, 1–15.

Wall, Larry D., and Timothy W. Koch (2000), “Bank LoanLoss Accounting: A Review of Theoretical and Empirical
Evidence,” Federal Reserve Bank of Atlanta Economic
Review, Second Quarter, 1–19.

Cole, Rebel A., and Jeffery W. Gunther (1998), “Predicting Bank Failures: A Comparison of On- and Off-Site
Monitoring Systems,” Journal of Financial Services
Research 13 (April): 103 –17.

ECONOMIC AND FINANCIAL REVIEW THIRD QUARTER 2000

Walter, John R. (1991), “Loan Loss Reserves,”
Federal Reserve Bank of Richmond Economic Review,
July/August, 20 – 30.

Full text of Review (Federal Reserve Bank of Dallas) : Third Quarter 2000

FRASER