National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
THE AMERICAN
RETAIL SALES TAX:
CONSIDERATIONS ON
THEIR STRUCTURE,
OPERATIONS, AND
POTENTIAL AS A
FOUNDATION FOR A
FEDERAL SALES TAX
JOHN L. MIKESELL
*
Abstract - Americans are familiar with
the retail sales tax. Therefore, it is not
surprising that Congress would consider
such a tax as a way to tax consumption
expenditure, should it choose to shift
from the present federal structure that
emphasizes income taxation. While the
sales taxes are impressive revenue
producers for state and local government, the taxes are poorly designed as
consumption taxes: they tax too few
household services, they exempt too
many household purchases of goods,
and they tax too many business inputs,
especially capital asset purchases. State
and local sales tax rates are relatively
low, so compliance appears not to be a
major problem, and economic distortions, while real, have not been a great
difficulty. The much higher rates
needed to replace the federal income
tax would create many more problems.
Most national governments choose the
credit-invoice value-added tax if they
seek substantial revenue from an
indirect consumption tax. That is almost
certainly a better option than the retail
sales tax for a national indirect consumption tax in the United States as
well.
INTRODUCTION
Retail sales taxes have been an American fiscal success.1 The tax that Mississippi initiated in 1932 by converting its
fractional rate general business tax into
a two percent tax on retail sales gave
the state a tax that produced considerable revenue at low statutory rates and
could be easily collected in relatively
painless amounts on each transaction.
These taxes gave states an alternative to
help them weather the collapse of
property tax revenue during the Great
Depression. The spread of the tax was
dramatic: by 1938, 26 states (plus
Hawaii) had adopted the tax. After that
*
School of Public and Environmental Affairs, Indiana University,
Bloomington, IN 47405.
149
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
fiscal catastrophe, retail sales taxes
provided important support for the
considerable increase in state government in the post–World War II era.
the percentage of sales tax voluntarily
reported against total tax liability at 95.9
percent in that state, and a Washington
state study (Washington Department of
Revenue, 1990) estimated its rate to be
98.3 percent.2 Few believe sales tax
compliance rates to be a severe problem, although that belief stems more
from faith rather than from research.
Retail sales taxes have now been the
largest single source of tax revenue for
states for 50 years, currently yielding
more than $130 billion annually. In fiscal
1994, these taxes produced more
revenue than any other tax for 23 of the
50 state governments. Revenues
amount to a bit more than one-third of
all state tax revenue; 12 states collect 40
percent or more of their tax revenue
from their general sales tax. These sales
taxes also yield ten percent of local
government revenue, a share so low
only because of the near-total domination of the property tax in the finances
of local school districts. The retail sales
tax is a vital contributor to the finances
of state and local government as the
only broad-based tax on consumption in
the United States revenue system.
Because the retail sales tax has been so
productive for state governments, it is
not surprising that some propose a
national retail sales tax as a replacement
for federal income taxes in the pursuit
of fundamental tax reform. It does offer
one alternative for moving the federal
revenue system from heavy reliance on
income taxes toward taxation based on
consumption. However, such a change
requires a clear understanding of the
structure of the state retail sales taxes
and what problems a heavier use of
these taxes might create. The following
sections will examine the logic of the
retail sales tax, the extent to which state
sales taxes are structured according to
that logic, and the potential for a
national sales tax that follows the
design of these state taxes.
The sales tax seems the model of
simplicity to the customer making a
purchase: the tax is added easily and
conveniently to the price at the time of
purchase. That appearance is misleading. As Simons (1950, p. 9) wrote many
years ago, the sales tax “is a simple tax
only in the sense that most people have
no part in its technical operation.”
Because the tax levied by the states
excludes many business and household
purchases and requires vendors to
distinguish which sales are exempt and
which are taxed, business compliance
with and government administration of
the tax is far more complex than its
appearance at the cash register.
Nevertheless, state sales tax compliance
seems to reach levels that federal
income tax collectors can only dream
about, although good data on the
question are sparse: a Tennessee study
(Adams and Johnson, 1989) estimated
THE IDEA OF A RETAIL SALES TAX
A retail sales tax as a practical application of consumption taxation does have
considerable appeal for improvement of
national economic prospects and for
improvement of fairness in the distribution of the cost of government. More
than 40 years ago, Kaldor (1955) made
the case for consumption as an equitable basis for taxation, particularly in
regard to complications of defining a
comprehensive income base: “...each
individual [measures tax capacity] for
himself when, in the light of all his
present circumstances and future
prospects, he decides on the scale of his
personal living expenses. Thus a tax
150
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
based on actual spending rates each
individual’s spending capacity according
to the yardstick which he applies to
himself. Once actual spending is taken
as the criterion all the problems created
by the non-comparability of workincomes and property-incomes, of
temporary and permanent sources of
wealth, of genuine and fictitious capital
gains resolve themselves; they are all
brought into equivalence in the measure
in which they support the actual
standards of living” (1955, p. 47).
Assigning a higher tax burden to those
who spend more surely represents an
ability-to-pay standard consistent with
consumer sovereignty and market
choice. How much the consumer
believes he or she can afford to purchase from the private market becomes
the standard according to which a
portion of the cost of government will
be distributed. Furthermore, by being
collected on each transaction instead of
on some end-of-year accounting, the
retail sales tax variant of the consumption tax is particularly convenient for
taxpayers.
disincentive to save that results from
taxing the returns to saving under an
income tax, and thus possibly to
improve the prospective path of real
economic growth, provides a tempting
argument for consumption taxation.
One prominent advocate of the national
retail sales tax maintains that substituting it for the income tax would produce
“...a capital formation boom with
strongly increased productivity, higher
paying jobs, and new investment from
around the world” (Lugar, 1995, p. 3).
A retail sales tax following the ideal of a
tax on consumption expenditure would
require two structural principles (Due
and Mikesell, 1994, p. 16):
2
“a. It should apply to all consumption
expenditures, and thus to all sales for
2consumption purposes, at a uniform rate.
Failure to do so will distort relative outputs of various goods and services, discriminate among various families on the
basis of consumer preferences, and, frequently, complicate compliance and administration because of the need to distinguish between taxable and nontaxable
items and among sales at various rates.
The consumption base also has a strong
foundation on grounds of economic
efficiency, particularly as an alternative
to a tax on income. As Cnossen and
1
Sandford
(1988, p. 32) note, switching
from income to consumption tax bases
“will reduce the difference between the
pre-and-post tax return to saving that
1
encourages
taxpayers to consume rather
than save, so saving will be encouraged
by the change and the growth path of
the economy may subsequently move
upwards.” This effect has generally
defied precise measurement in the
United States, partly because of the
difficulty of understanding what
influences private saving and partly
because the country has no experience
with a tax solely on household consumption. Nevertheless, to end the
“b. It should apply only to consumption
expenditures, and thus not to savings or
to purchases for use in production. Taxation of savings or uses of savings would
contradict the consumption intent of the
tax. Taxation of production inputs has
several undesirable consequences, including that of producing a haphazard and
unknown final pattern of distribution of
burden among various families.”
The retail sales tax is limited to consumption expenditure by suspending
application of tax to business purchases.
Businesses apply tax to their sales,
except when the purchaser provides
suspension documentation—a certificate of exemption—ordering that the
retail sales tax not be applied. The
certificate identifies (1) the purchaser as
151
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
a registered business eligible to make
exempt purchases and (2) the business
use of the item that makes it nontaxable.3 When the suspension system
works, no business in the chain of
production and distribution bringing
product to the household consumer
would have paid tax on its purchases,
because each business would have
presented its tax suspension certificate
to the seller. The consuming household
is not a registered business, can offer no
suspension certificate, and must pay the
tax on its purchases. Through the
suspension process, the tax rests
uniformly and exclusively on household
consumption. If business purchases of
what they use to produce what they sell
are entirely excluded and all household
purchases of goods and services are
taxed, the retail sales tax will be a tax on
final sales to consumers—a general
consumption tax.
percent or higher. That is considerably
increased from the range in 1970 from
two percent (five states) to six percent,
with a mean rate of 3.54 percent.
Furthermore, only one state levied a six
percent rate and only five levied a five
3
percent
rate. The pace of rate increases
has declined during the 1990s, as strong
state finances have reduced the need to
increase rates.
Although states tend to copy law from
their neighbors, no state sales tax
exactly matches any other and the
different structures take dramatically
different shares of their state economies. Nevertheless, the sales taxes share
some elements. (1) All are levies
“imposed upon the sales, or elements
incidental to the sales, such as receipts
from them, of all or a wide range of
commodities” (Due, 1957, p. 3). (2) All
have a system for suspending tax on
items purchased for resale so that the
cost of inventory to a retailer will not
include sales tax paid on its acquisition.
(3) All levy rates of one percent or
higher on retail transactions. (4) All
encourage, if not require, separate
quotation of the tax in each transaction
except in special circumstances and all
but Arizona prohibit retailers from
advertising that they will absorb or
refund the tax (Due and Mikesell, 1994).
Beyond those elements, the coverage
choices dramatically influence the
nature of the tax base.4
Because state sales taxes miss the
standard, they do not have all the
advantages claimed for a tax on general
consumption expenditure. The taxes
now are partial taxes on both household
consumption and business input
purchases. Surprisingly enough, many
states continue to have sales taxes that
discourage capital formation by taxing
the acquisition of many business inputs,
including real capital assets.
AMERICAN RETAIL SALES TAXES AND
THE IDEAL
Table 1 shows the retail sales tax yield
for each state in fiscal 1994. Highest per
capita yields, after adjusting for statutory rates, are in Hawaii, New Mexico,
South Dakota, Nevada, and Arizona; the
yield per percentage point in those
states averages $159 per capita. The
lowest yields are in Rhode Island, West
Virginia, Pennsylvania, Oklahoma, and
Vermont: the yields averaged only $70.
The high yield states have particularly
Forty-five states plus the District of
Columbia (all states except Alaska,
Delaware, Montana, Oregon, and New
Hampshire) and more than 6,000 local
governments levy retail sales taxes.
Among the sales tax states, statutory
rates ranged from three percent
(Colorado) to seven percent in mid1996, with a mean of 5.18 percent.
Seventeen states levied rates of six
152
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
TABLE 1
AN INTERSTATE COMPARISON OF SALES TAX BASES, 1994
State
Adjusted Sales Tax
Revenue
($ Thousand)
Alabama
1,340,142
Arizona
2,463,337
Arkansas
1,211,806
California
16,871,660
Colorado
1,125,265
Connecticut
2,184,089
Florida
10,042,360
Georgia
3,370,936
Hawaii
1,268,686
Idaho
544,145
Illinois
4,794,082
Indiana
2,601,382
Iowa
1,388,742
Kansas
1,297,170
Kentucky
1,855,338
Louisiana
1,888,695
Maine
617,008
Maryland
2,220,825
Massachusetts
2,303,139
Michigan
4,538,124
Minnesota
2,848,810
Mississippi
1,586,879
Missouri
2,195,890
Nebraska
743,240
Nevada
1,184,850
New Jersey
3,778,427
New Mexico
1,267,035
New York
6,117,517
North Carolina
2,574,512
North Dakota
296,557
Ohio
4,479,907
Oklahoma
1,096,600
Pennsylvania
5,134,300
Rhode Island
412,820
South Carolina
1,685,727
South Dakota
371,251
Tennessee
3,081,250
Texas
11,709,340
Utah
984,287
Vermont
220,777
Virginia
2,178,352
Washington
4,169,570
West Virginia
756,622
Wisconsin
2,427,900
Wyoming
199,428
District of Columbia
509,857
Adjusted Sales Tax
Revenue as
Percent of Tax
Revenue
28.1%
43.5%
38.2%
34.0%
27.1%
32.2%
56.4%
38.4%
42.4%
33.7%
31.0%
35.7%
33.6%
35.3%
32.6%
43.1%
35.0%
29.3%
20.9%
29.4%
32.9%
47.7%
37.2%
34.7%
49.8%
28.0%
41.9%
18.6%
24.5%
33.5%
31.6%
25.7%
30.0%
28.7%
37.4%
56.3%
53.8%
60.2%
40.7%
26.5%
27.1%
43.0%
29.6%
28.8%
27.0%
20.4%
Per Capita Sales
Tax Revenue per
1% Tax Rate ($)
79.42
120.90
109.79
89.47
102.60
111.14
119.96
119.45
269.12
96.05
79.60
90.45
98.17
103.65
80.81
109.43
82.92
88.72
76.25
113.78
103.96
84.94
98.48
91.59
125.11
79.68
153.25
84.17
91.04
92.97
80.70
74.79
71.00
59.17
92.01
128.70
99.23
101.94
103.18
76.10
95.00
120.06
69.21
95.55
107.44
144.23
Sales Tax Revenue
as Percentage of
Personal Income
per 1% Tax Rate
0.44%
0.63%
0.65%
0.40%
0.46%
0.38%
0.55%
0.59%
1.12%
0.52%
0.34%
0.45%
0.49%
0.50%
0.46%
0.62%
0.43%
0.36%
0.30%
0.51%
0.47%
0.54%
0.48%
0.44%
0.53%
0.29%
0.90%
0.33%
0.47%
0.50%
0.39%
0.42%
0.32%
0.27%
0.52%
0.66%
0.51%
0.52%
0.60%
0.38%
0.42%
0.53%
0.40%
0.46%
0.53%
0.47%
Source: U.S. Bureau of Census: State Tax Collection: 1994. Sales tax collections are adjusted according to the Due
and Mikesell technique (1994) to allow for inconsistencies caused by census reporting peculiarities, nonstandard
rates, and special excises levied in lieu of expanded sales tax coverage in some states.
153
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
broad-based coverage and, except for
South Dakota, distinct tourist economies; the low yield states have generous
exemptions. However, even more
information about the structural
differences can be garnered by comparing yield per percentage point of tax as
a percentage of state personal income.
The broadest by this comparison
(Hawaii, New Mexico, South Dakota,
Arkansas, and Arizona) average 0.79
percent of personal income, compared
with an average of 0.30 percent for the
narrowest (Rhode Island, New Jersey,
Massachusetts, Pennsylvania, and New
York). That means, holding size of
economy constant, each percentage
point of tax from the broadest coverage
taxes will yield more than twice the
revenue of the narrowest coverage tax.
While these states all levy a sales tax,
the legislative choices made in structuring each tax have caused dramatically
different tax bases. Major differences
include the taxation of household
purchases of goods, taxation of services,
and exclusion of business purchases.5
exempt commodity, reduces revenue
from the given statutory rate, makes
collection more difficult for vendors and
tax authorities, and may distort use of
productive resources. The presumption
is against exemption unless there are
extraordinary reasons otherwise, and
any exemption goes against the
principle of general consumption
taxation.
Food exemptions
States exempt food purchased for
preparation at home to reduce
regressivity and thereby remove roughly
one-fifth of all household consumption
on goods from the tax base.7 The
exemption makes the sales tax more
vulnerable to business recessions, causes
tax payments by families to vary
according to preferences for food, and
complicates compliance and administration (Mikesell, 1996b). Furthermore,
Congress requires those states choosing
to otherwise tax food to exempt
purchases made with food stamps, as a
price for participation in the program.
Only 19 states, all west of the Mississippi River, now tax food fully, 25
exempt food fully, and three provide a
reduced rate.8 Twenty-five years ago,
the numbers were reversed: 28 states
taxed food, 16 exempted food, and 1
state provided a reduced rate.
The Taxation of Household Purchases
of Goods
Retail sales taxes are typically general
taxes on the purchase (or sale) of
tangible personal property and selective
taxes on certain purchases (or sale) of
service. States have chosen, however, to
exempt certain household purchases of
goods from the tax. State legislatures
understand that households with low
annual income spend a higher portion
of that income than do higher income
households and seek to avoid the
appearance of a regressive tax under
which tax liability as a fraction of
income declines with household
income.6 They respond by removing
certain expenditure categories from tax.
Exemption favors households with
relatively greater preference for the
Items subject to an excise
Motor fuel, tobacco product, and
alcohol beverage taxes are the most
important examples of excised items
that states have exempted from their
retail sales taxes. An excise implies
economic or social reasons for placing
an extra tax burden on the consumption
of the excised item. To then exempt the
taxed items from the general tax makes
no sense. Only ten states apply their
retail sales tax to motor fuel: Arizona,
California, Georgia (taxed at one
154
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
percentage point less than the basic
rate, plus a one percent special sales
tax), Hawaii, Illinois, Indiana, Michigan,
New York, South Carolina, and West
Virginia. Most states now tax liquor
and tobacco products, although
Mississippi exempts alcoholic beverages,
Virginia exempts state liquor store sales,
and Colorado and Texas exempt
cigarettes.
Utilities
Many states exclude from tax major
utilities—electric, gas, water, and
intrastate telephone—purchased by
residential customers. Even fewer states
tax interstate telephone charges. Until
1989, states believed that such taxes
would violate the commerce clause of
the constitution and thus avoided taxing
these charges. The U.S. Supreme Court
(Goldberg v. Sweet, 488 US 452, 109 S.
Ct. 582, 585 (1989)) held that they do
not, but states have been slow to react.
Prescription medicine and health related
products
States regularly exempt consumer
purchases of medicines and other goods
related to sickness or health. Only New
Mexico offers no exemption for prescription medicines. A number of states also
exempt over-the-counter medicines.
These states include Florida, Illinois
(taxed at a lower rate), Maryland,
Minnesota, New Jersey, New York,
Pennsylvania, and Rhode Island. When
the exemption is limited to prescriptions,
compliance and administration remains
relatively easy. Adding other health
products requires difficult interpretations
about what should be exempt and
complicates the collection process, as
well as erodes the notion of a tax on
consumption.
This brief review shows that the retail
sales taxes are not general taxes on
household consumption of goods, but
rather taxes that exempt broad categories of consumption. Each exemption (1)
violates the logic of a general consumption expenditure tax, (2) reduces revenue
yield from any particular advertised tax
rate, and (3) complicates compliance
and administration. Most have been
adopted to improve perceived fairness
of the system and to provide tax relief.
However, other means —notably annual
rebates or credits offered through other
portions of the overall tax system—can
provide assistance at lower revenue cost
and with less economic distortion. While
retail sales tax coverage of the purchases of goods is broad, these taxes
certainly are not uniform taxes on
household consumption of goods.
Clothing
A half-dozen states—Connecticut,
Massachusetts, Minnesota, New Jersey,
Pennsylvania, and Rhode Island—
exempt some clothing purchases,
presumably under the assumption that
exempting these necessities improves
fairness of the tax. Evidence suggests
that spending patterns across income
classes are such that this exemption
gives greater relief to high-income as
opposed to low-income families
(Schaefer, 1969). It therefore reduces
revenue yield and makes collection of
the tax somewhat more difficult,
without the desired equity effect.
Taxing Services
A tax on consumption should not treat
household expenditure on goods, e.g.,
the purchases of a television, differently
than household expenditure on a
service, e.g., the repair of that television.
Both are consumption expenditures. In
practice, the retail sales tax more often
than not reflects the pattern for such
taxes established in the 1930s: the base
is simply defined to be tax purchases of
155
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
tangible personal property. Taxation of
services usually amounts to a legislative
afterthought, often limited to transient
lodgings, rental of tangible personal
properties, property repairs and installation, and admissions, and not all taxes
go even that far. For instance, many
taxes do not tax the servicing of items
that are taxable when sold. States have
been especially reluctant to tax professional services provided by physicians,
dentists, lawyers, accountants, and
casualty insurers.
(although Minnesota misses repair and
installation services). Other states with
similarly extensive taxation of services
include the following: Arkansas,
Connecticut, Florida, Kansas, Louisiana,
Mississippi, New Jersey, New York, Ohio,
Pennsylvania, Tennessee, Utah, Washington, West Virginia, Wisconsin, and
Wyoming.
Limited or no coverage of service purchases
The tax applies to purchases of tangible
personal property and makes no
concerted attempt to tax services. A few
selected services may be specifically
taxed, but some of the states involved
tax no services at all. The 23 states not
previously mentioned in the other
categories fall into this one.
State sales taxes can be generally
divided into three broad groups for
taxation of service purchases (Due and
Mikesell, 1994; Research Institute of
America, 1996):
General coverage
A consumption expenditure tax that
includes household purchases of services
would have significant fiscal advantages. First, taxing services will increase
the yield from any given statutory tax
rate. The impact depends on what
services are added and what purchases
are already in the tax base. Adding
household purchases of repair, installation, and maintenance of tangible
personal property and services rendered
by commercial establishments—
including those for personal care—can
increase revenue by 10 to 15 percent,
much of that from services related to
automobiles. Taxation of privately
provided services—babysitting, housecleaning, etc.—is not administratively
feasible and taxation of medical and
dental services may create unacceptable
social problems. Furthermore, evidence
suggests that the broader base will
grow somewhat more rapidly, because a
growing percentage of household
spending is on services rather than
commodities (Duncombe, 1992).9
National income data show that, from
1985 through 1994, the service
This coverage excludes services rendered
by employees to employers and a few
categories such as financial intermediation. All service purchases are taxed
unless the statute specifically exempts
them. No retail sales tax fully integrates
taxation of services with taxation of
goods, but Hawaii, New Mexico, and
South Dakota provide the broadest
coverage, the first two from their
adoption and the latter by later expansion.
Extensive taxation of services
Repair, installation, maintenance, and
other services associated with tangible
personal properties are taxed, along
with a long list of specifically identified
services (parking, landscaping, pest
control, laundry and dry cleaning, and
cable television are common examples).
Medical, optical, and dental care, legal
services, and other professional services
seldom are taxed, however. Iowa,
Minnesota, and Texas provide the
broadest list in terms of economic
significance of the taxed services
156
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
component of personal consumption
spending increased at an annual rate of
7.4 percent, compared with a rate of
5.1 percent for the goods component.
Not all that service spending would
reasonably be taxed, but the difference
between the rates for services and for
goods illustrates the pattern.
product is considerable. As a result,
administration would be more difficult
for the authorities and compliance
would involve small firms with less skill
and sophistication with the tax system.
The authorities can deal with the
problem by taxing the purchases of such
service firms, thus letting their customers pay the tax indirectly. That is
generally the approach many valueadded taxes (VATs) use when they
exempt small businesses from collecting
tax on their sales, but not from paying
tax on their purchases. This is implicitly
the approach used by American states:
businesses selling services do not have
tax suspended on their purchases.
Where the labor content of the final
sales to the consumer is high, however,
a considerable portion of consumption
expenditure will go untaxed.
Second, taxing services can reduce
discrimination according to household
preferences. Without taxing services,
persons with high preferences for
services pay less tax than equivalently
situated consumers with lower preferences for services. The impact on
regressiveness is not so clear. Evidence
shows existing taxes with broad-based
extension to services leave the tax
regressive to income levels of around
$30,000, above which the tax is roughly
proportional (Fox and Murray, 1988).
Siegfried and Smith (1991) find that the
short-lived extension of the Florida sales
tax to selected services in 1987 reduced
regressivity. However, whether adding
services to any existing goods-oriented
sales tax would reduce regressivity
depends crucially on what the existing
tax has exempted (and taxed) and what
services are added.
Second, many services have mixed use,
being purchased by both households
and businesses. For example, attorneys
work for both businesses and households. The retail sales tax requires the
seller to segregate business use from
private use at purchase, apply tax to the
proper transactions and be prepared to
justify the remainder, and separately
account for the two flows. The burden
can be a problem for the seller. And the
tax authorities face a similar set of tasks
on unraveling flows on audit. To prevent
the complications, retail sales taxes
characteristically exclude difficult service
transactions from the base—and, for
good measure, exclude the simple ones
as well.
Finally, limiting the tax to commodities
causes difficult problems when goods
and services are sold together. Examples
include computer programs (purchase of
the physical disk or the service of the
program?) and glasses (the physical
eyeglasses or the services of the
optometrist?), although there are also
problems with repair charges that
involve both materials and labor.
Excluding Business Purchases
Taxing services does, however, present
puzzles. First, firms selling services are
frequently small. The firms are often
highly specialized and there are few
economies from large size because the
labor content of the value of the
The business purchase exemption
maintains the logic of a tax on household consumption, keeps the tax from
discouraging capital investment by
being imposed on purchases of machinery and equipment, and reduces a
157
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
barrier to state economic development.
The exemption is critical to prevent the
tax from applying to multiple stages in
the production and distribution chain
and the tax pyramiding that results. The
truth of the state retail sales taxes does
not follow that model.
cost, no less than the cost of labor,
supplies, or other purchases of the firm.
The refrigerator, although physically
finished, is not economically final
because it contributes to the production
of a good or service bought by a
household. The critical element for tax
policy is not whether the product is
finished but rather whether the purchase is household consumption or
business cost.
In practice, retail sales taxes exempt
items purchased for resale and components or ingredients of items to be
produced for sale. Many also exempt
industrial and agricultural machinery
purchased for production, but not all
do. A few states, including California,
Nevada, South Dakota, Washington,
Wyoming, and Hawaii, fully tax industrial machinery purchases and a number
of states exempt only machinery
purchased for new and expanded
production or tax such purchases at a
preferential rate. The states also vary
dramatically in the extent to which their
taxes exempt consumables, fuels, and
utilities, even when clearly used in
production.10 A number of states turn
the concept of consumption tax entirely
on its head by taxing business purchases
of fuels and utilities, while exempting
their purchase by households. State
sales taxes by no means remove
business purchases from the tax.
Including business purchases creates
two problems. First, these extra costs
get hidden from customers, being
reflected in final prices by varying
amounts that depend on market
conditions and on the extent to which
businesses purchase taxed inputs. Thus,
embedded sales tax paid by households
would differ according to the sort of
purchases made, rather than being a
uniform percent of consumption.
Second, taxing inputs distorts production decisions. The tax discourages
investment in production and distribution equipment by reducing the net
return from their acquisition and
depressing the prospects for economic
development. Joulfaian and Mackie
(1992, p. 102) present the problem:
“Sales taxes on purchases of capital
assets can be fairly large. For some
assets, the combined state and local
sales tax rate can be nearly as high as 6
percent of the sales price. Furthermore,
sales tax rates vary dramatically across
investments, averaging 4.2 percent for
equipment, 1.6 percent for structures,
and 0 percent for inventories and land.
Thus, sales taxes can affect both the
overall average incentive to save and
invest, as well as the allocation of
investment across assets.” In other
words, American retail sales taxes, as
they are structured, do not have the
benign (or even encouraging) effect on
investment that a consumption tax
should to have. Furthermore, vertically
States face considerable temptation to
include business purchases in their retail
sales tax base. Goods purchased for
business use are just as physically
“finished” or final as those purchased
by households. By that standard, there is
no difference between a refrigerator
purchased by a restaurant and a
refrigerator purchased by a family,
except possibly the size. Both customers
are “final.” Taxing both purchases
increases revenue from any given
statutory rate, and including both seems
to avoid a business tax preference.
However, the tax on the restaurant
purchase becomes part of operating
158
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
integrated firms able to avoid purchasing
from outside suppliers have an
extra competitive advantage, and distortions are likely to reduce domestic capital
investment and economic growth.
A NATIONAL RETAIL SALES TAX?
Retail sales taxes and VATs offer two
economically equivalent alternatives for
collecting a broad indirect, transactionbased tax on consumption expenditure.
The taxes differ only in how they are
administered. The VAT applies to each
transaction in the flow of output to the
consumer; in its most common variant,
each business pays tax on its purchases
but is reimbursed from tax collected on
its sales. By that means, only the
household consumer bears the tax,
business purchases having been
removed from tax by the refunding
process. The retail sales tax applies only
to the final consumer purchase, business
purchases having been exempt by the
suspension certificate. The two apparently dissimilar taxes differ only (but
importantly) in the mechanism they use
to relieve business purchases from tax,
allowing the tax to apply to consumption.
Unfortunately, exempting business
purchases appears to favor households
over businesses, so removing intermediate goods from tax has proven to be a
difficult legislative challenge. Legislatures have had difficulty understanding
that exempting business purchases
represents a refinement to apply the tax
to the defensible base, not an unwarranted tax advantage for business. And,
at least as important, it is an argument
that the voting public and the media
have difficulty understanding. Therefore,
the typical state sales tax base includes a
sizable chunk of business purchases that
simply do not belong in a consumption
tax.
There is no entirely satisfactory estimate
of the extent to which business purchasers bear the impact of state sales taxes.
All evidence does indicate that the
portion is substantial, although it
obviously differs according to the
structure of each state tax and the
nature of the state economy. One study
(Ring, 1989) estimated the national
average to be 41 percent of the tax paid
by business purchasers, with a range
from around 20 percent (Maryland and
Virginia) to 65 percent (Wyoming and
Louisiana).11 Individual state studies also
estimate the business share to be
considerable. For instance, Texas
estimated the share there to be 58
percent while Iowa estimated its share
to be 39 percent (Due and Mikesell,
1994). A sizable component of the tax
base is business purchases, and much of
the tax paid by these businesses will be
reflected in higher prices of goods made
by these inputs, although the impact on
prices will undoubtedly not be uniform.
Most national governments have
selected the VAT, rather than the retail
sales tax, as their general consumption
tax. More than 100 countries around
the world, including all those in Europe,
all of Latin America, Japan, and Canada
levy the tax; only the United States and
Australia of the large industrial nations
do not levy it. Countries have generally
adopted the VAT for administrative
advantages, especially when rates are
high: the VAT more easily and completely excludes producers goods from
tax, more completely includes consumer
services in the base, and appears to
provide a better defense against evasion
(OECD, 1995). Nevertheless, the retail
sales tax is familiar to Americans. It is
reasonable to consider that collection
method if the federal tax system is to
move toward greater reliance on taxes
on consumption.12 However, reflections
from state experience should enlighten
159
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
and caution Congress as it weighs the
options.
rates would almost certainly produce
untenable economic distortions and
increase compliance problems. One
observer of retail sales taxation guesses:
“At 5 percent, the incentive to evade
tax is probably not worth the penalties
of prosecution; at 10 percent, evasion is
more attractive, and at 15–20 percent,
becomes extremely tempting” (Tait,
1988, p. 18). At the rates speculated
here, evasion would be remarkably
profitable and administration would
become a greater challenge.
Higher Reliance and Higher Rates
To replace the federal income taxes with
a retail sales tax would mean much
heavier use of the tax than the United
States has experienced: the federal
individual income tax alone produced
over $590 billion in fiscal 1995, roughly
four and one-half times the $132 billion
generated by all state sales taxes, and
the corporate income tax produced
$157 billion more (OMB, 1996; Bureau
of Census, 1996). A national retail sales
tax sufficient to replace existing federal
individual and corporate income taxes
would require a remarkably high rate. If
the base of the national tax mirrored the
median state sales tax (in fiscal 1994,
Minnesota, with each one percent of tax
rate generating 0.467 percent of
personal income), then a national rate
of 25.5 percent would be needed to
replace the federal taxes. Copying the
broadest tax would require a rate of 11
percent; copying the narrowest, a rate
of 44 percent.13 The message is clear: a
national sales tax to replace the federal
income taxes would require a rate
considerably higher than the rates states
now levy. Furthermore, the national tax
would be on top of existing state and
local rates—presumably these governments would not be forced to find other
revenues when the federal government
stepped in—and the combined rates
would almost certainly be global records
for such taxes!
To the extent business purchases
remained in the tax, vertically integrated
firms able to acquire production assets
without purchasing them from another
business would enjoy considerable
competitive advantage, and the tax
applied to the purchase of capital assets
would discourage investment. Furthermore, the embedded tax on business
purchases would create difficult
problems in international trade. It is
impossible to know to what extent the
sales tax is included in business cost at
the point of export. Therefore, trade
rules do not allow rebate of the tax.
Adding a high but embedded and nonrefundable national sales tax to American products would create challenging
competitive adjustments on international markets. Relatively low retail sales
tax rates now do not present a major
problem. That would change with the
much higher national retail sales tax
added on.
Taxing Consumption?
States can operate their sales taxes with
their many imperfections as consumption levies because the statutory rates
are relatively low. Purchasers and
vendors generally comply, but the rates
are low enough to keep the stakes from
evasion relatively low. Imposing an
American retail sales tax at replacement
A national retail sales tax generally
following the pattern of the state taxes
would not be a general and uniform tax
on consumption. State retail sales taxes
are partial consumption taxes and
partial business purchase taxes. They
exempt many household purchases of
goods, are narrowly selective in their
160
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
coverage of household service purchases, and apply to a wide array of
business inputs, including the purchase
of capital assets. This hybrid tax would
not have the performance influences on
savings, investment, and economic
development that a uniform and
general consumption tax would have.
State experience shows the political
difficulty in getting the retail sales tax
structured as a tax on consumption.
Legislatures see many options for
excluding consumption expenditure
from tax in the pursuit of distributional
equity or to boost regional causes.
income taxes—be less susceptible to the
pressures that have created the American state sales taxes than have the state
legislatures?
Concerns About Compliance and
Administration
A federal tax piggybacked on the
existing state taxes would be completely
inappropriate for a national tax because
of the considerable state-by-state
differences in the tax. For instance, a
shopper buying a suit in New Jersey
would pay the national sales tax; the
shopper’s cousin buying an identical suit
in Pennsylvania would not pay the tax.
Such differences would create major
disruptions in interstate trade (adding
the federal tax rate to existing state
differentials would make cross-border
shopping remarkably attractive on highticket items), as well as cause substantial
differences in effective federal sales tax
rates among the states. Determining the
tax in the five states without a sales tax
would be an interesting matter for
Congress. Should such piggybacking be
attempted, states would find it in the
best interests of their citizens to adjust
their sales taxes to the narrowest base
possible as a means of minimizing their
national tax burden.
However, the problem may be fundamental. The technique of tax suspension
probably has inherent administrative
limits in the completeness of business
purchase exclusion it can achieve.
Suspension places the burden of parsing
between business purchases and
household purchases on the vendor, not
a tax authority, and the vendor has an
inherent competitive incentive to grant
the suspension. Hence, tax administrators face a continual need to police
acceptance of suspension certificates.
Authorities tighten to prevent evasion
through misuse of suspension authority,
but tightening makes tasks more
difficult for vendors. Loosening makes
evasion easier.
Congress could prevent differences in
effective rates across states by adopting
its own national tax base. A distinct
federally designed national retail sales
tax might be a uniform and nondistorting consumption tax, avoiding all
the imperfections that have emerged
from the state legislatures. Experience
with other federal taxes should cause
some considerable skepticism about
Congressional ability to pursue such
purity in taxation, but even if the
remarkable were to happen, there
would still be problems. Vendors, the
gatekeepers in any retail sales tax,
Could Congress devise a politically
feasible approach to taxing household
consumption so that the base closely
approaches the consumption ideal,
meets contemporary ideas about
fairness in taxation, and keeps up with
change in the economy? Could Congress manage to suspend more completely the tax on intermediate goods to
protect investment, avoid economic
distortion, and accommodate international competitiveness? Would Congress—the collective author of the
current federal individual and corporate
161
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
would have to sort purchases as to
taxability between the state and federal
systems, and tax authorities would have
to ensure that each sort was accurate.
Of course, the federal government
could force states to use the new
federal base, but that would mean
considerable loss of fiscal sovereignty.
States have considerable experience in
simultaneous administration of multiple
taxes, so the primary new problem
would be with compliance rather than
administration.
government. Structural problems,
however, suggest that they would not
be the best sort of general consumption
tax if the federal government were to
replace its income taxes. At the high
rates required to yield enough revenue
to replace the income taxes, problems
of evasion and higher reward from
avoidance, distortion caused by incomplete exclusion of business purchases,
and narrow coverage of consumer
services diminish the attractiveness of
the retail sales tax. The taxes are not
general taxes on consumption expenditure, and they do not have all the
economic advantages associated with
such taxes.
As noted earlier, the extremely high
combined national sales tax rate would
change the stakes for taxpayer compliance. Most national governments have
concluded that the clearer audit trail
offered by VATs administered by the
credit-invoice approach is important as
an enforcement tool for high rate
consumption taxes, and they reject the
retail sales tax administrative approach.
Congress needs to be prepared with
considerable enforcement resources and
zeal if it wants to prevent tax cheats
from having a competitive advantage,
given the much higher rewards to
evasion with the high national tax rate.
The compliance experience with the
much lower state rates is probably
irrelevant.14
What are the conclusions for the
American tax system? State and local
governments should perfect their retail
sales taxes as consumption expenditure
levies. Surely the case for taxation of
consumption being made at the
national level has considerable validity
for subnational taxes as well. The typical
retail sales tax would need to be
extended to professional services sold to
households, personal services, and
repair-and-installation sorts of services
associated with goods whose purchase
would be taxable. Producer purchase
exemptions would need to be extended
broadly to business inputs. These policy
changes would improve the “capitalinvestment” aspect of tax climate and
would dramatically reduce the number
of sales tax cases in the judicial system.
Neither broadening to services nor
narrowing to exclude business purchases would be politically easy—if so,
states would surely have made the
changes years ago. Sales taxes are also
challenged by marketing systems that
lack a physical presence within the state
—sales by catalog, telephone, computer
network, etc. Constitutional restrictions
prevent states from requiring such
Conclusions
Retail sales and use taxes have been a
vital component of state and local
government revenue systems since their
inception during the Great Depression.
They took over the slack left by the
collapse of the real property tax and
supported the considerable growth in
state governments in the last halfcentury. They continue as the most
important single tax levied by state
governments, and they yield considerable revenue for support of local
162
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
vendors to register for the tax, even as
this marketing technique becomes more
prevalent. Although some such merchants voluntarily agree to collect tax,
Congress certainly could help states by
devising some mechanism for enforcing
this tax.
3
Attempting to levy a national tax as a
supplement to state sales taxes would
be folly. There is simply insufficient
uniformity in what states tax and
exclude to allow a linked federal tax to
be fair and efficient. On the other hand,
a separate national retail sales tax would
complicate the work of state tax
collection, increase the problems that
businesses face in complying with
multiple tax bases, and tax on a base
not equal to household consumption.
Other countries probably have it right
when they select VATs to raise considerable revenue at the national level,
because, in practice, such taxes using
the credit-invoice method more fully
exclude business purchases from tax,
more completely tax household purchases than do retail sales taxes, and
leave a better trail for enforcement.15
5
4
6
7
8
9
10
11
ENDNOTES
1
2
John F. Due, Matthew Murray, and Joel Slemrod
made helpful suggestions on an earlier draft.
It is now generally an American tax: among the
industrialized countries, only state and local
governments in the United States and provincial
governments in Canada levy the retail sales tax.
Switzerland and some Scandinavian countries did
use the tax, but all have now replaced it with VATs.
Other countries with some sort of retail sales tax
include Namibia, some states in India, and parts of
the former Soviet Union (for instance, the Kyrgyz
Republic).
The Washington study estimated compliance of
only 59.7 percent for the compensating use tax,
however. The compensating use tax, a companion
to all state and many local sales taxes, is applied to
the use of taxable items bought without payment
of sales tax, normally because the vendor was
out-of-state. Successful administration remains a
puzzle.
12
163
Its close relative, the VAT, removes business
purchases from the tax by refunding, through
credit or rebate, tax paid on their purchases.
The taxes in Hawaii (the general excise tax) and
New Mexico (the gross receipts tax) are considerably different from the other taxes considered
here. Both will be included in this analysis,
however, because each has the basic features of a
retail sales tax, including being accompanied by a
compensating use tax.
Other differences include the taxability of nonprofit organization purchases and sales. States
differ in terms of whether organizations will be
exempt specifically or by category, what
organizations or categories will be exempt, and
whether sales as well as purchases will be exempt
(Mikesell, 1992). Local retail sales taxes generally,
but not always, follow the base of their state.
Therefore, in the discussion here, comments about
state sales taxes should be understood to apply
also to local taxes within the state.
Regressivity is far less if the horizon of measurement expands from the current year to lifetime or
permanent income (Poterba, 1989). State
legislatures, however, seem more concerned with
annual patterns.
The estimate is based on “Selected NIPA
Tables”(1996) and data from the Consumer
Expenditure Survey.
The Georgia reduced rate is part of a phase-in to
full exemption on October 1, 1998.
Not all service purchases are fast growing,
however, and business purchases of services, those
not suitable for inclusion in the tax, grow most
rapidly (Dye and McGuire, 1991).
For more detail on the taxation of production
inputs, see Due and Mikesell (1994).
On the basis of Consumer Expenditure Survey data,
Ring (1989) estimates what consumer sales tax
payments should be for each state. He then
attributes the difference between this estimate and
actual collection to tax paid by producers. The
approach has many flaws—CES categories do not
align well with state tax laws, CES data do not
reflect state-by-state differences in household
expenditure behavior, any flaws in tax administration artificially reduce the business share, etc.—
but these estimates are the national standard by
default, despite the fact that they often conflict
with evidence from detailed individual state
studies.
This section will consider a federal retail sales tax
patterned generally after the American sales
taxes. The National Retail Sales Tax Act of 1996,
H.R. 3039 (introduced March 6, 1996), by
Representatives Schaefer, Tauzin, Chrysler, Bono,
Hefley, Linder, and Stump is radically different; a
critical analysis of that proposal appears in Mikesell
(1996a).
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
NATIONAL TAX JOURNAL VOL. L NO. 1
13
14
15
Bartlett (1995) uses a different estimating
approach, but his range from 15 to 30 percent
is similar.
See Murray in this volume for some thoughts on
this question.
See Mikesell (forthcoming) for a comparison of the
fundamental differences between the retail sales
and value-added approaches to general
consumption taxation.
1986.” National Tax Journal 45 No. 1 (March,
1992): 89–105.
Kaldor, Nicholas. An Expenditure Tax. London:
Allen and Unwin, 1955.
Lugar, Dick. “The Lugar Plan to End the Income
Tax.” Statement delivered to the Cato Institute,
Washington, D.C., April 5, 1995.
Mikesell, John L. “Sales Taxation of Nonprofit
Organizations: Purchases and Sales.” In Sales
Taxation, Critical Issues in Policy and Administration, edited by William F. Fox. Westport, CT:
Praeger, 1992.
REFERENCES
Mikesell, John L. “A National Retail Sales Tax?
Some Thoughts on Taxing Consumption the
American Way.” State Tax Notes 10 (July 8,
1996a): 105–9.
Adams, Virginia N., and Kirk L. Johnson.
“Estimating the Sales Tax Gap.” Tennessee
Department of Revenue 1989 Workshop on Tax
Administration. Nashville: Tennessee Department of Revenue, 1989.
Mikesell, John L. “Should Grocery Food
Purchases Bear a Sales Tax Burden?” State Tax
Notes 11 (September 9, 1996b): 751–4.
Bartlett, Bruce. “Replacing Federal Taxes with a
Sales Tax.” Tax Notes 68 (August 21, 1995):
997–1003.
Mikesell, John L. “Is the Retail Sales Tax Really
Inferior to the Value Added Tax?” In The Sales
Tax in the 21st Century, edited by Matthew
Murray and William Fox. Westport, CT:
Greenwood Publishing (forthcoming).
Bradford, David F.
“The Case for a Personal
Consumption Tax.” In What Should Be Taxed:
Income or Expenditure?, edited by Joseph A.
Pechman. Washington, D.C.: The Brookings
Institution, 1980.
Office of Management and Budget.
Historical
Tables, Budget of the United States Government,
Fiscal Year 1997. Washington, D.C.: Government Printing Office, 1996.
Case, Bradford, and Robert Ebel.
“Using
State Consumer Tax Credits for Achieving
Equity.” National Tax Journal 42 No. 3
(September, 1989): 323–37.
Organisation for Economic Co-operation and
Development. Consumption Tax Trends. Paris:
OECD, 1995.
Cnossen, Sijbren, and Cedric Sanford.
Taxing
Consumption. Paris: Organisation for Economic
Co-operation and Development, 1988.
Poterba, James M. “Lifetime Incidence and the
Distributional Burden of Excise Taxes.” American
Economic Review 59 No. 4 (May, 1989): 325–
30.
Due, John F. Sales Taxation. Urbana, IL:
University of Illinois Press, 1957.
Research Institute of America.
1996 Guide to
Sales and Use Taxes. New York: Research
Institute of America, 1996.
Due, John F., and John L. Mikesell.
Sales
Taxation, State and Local Structure and
Administration. 2d ed. Washington, D.C.:
Urban Institute Press, 1994.
Ring, Raymond. “The Proportion of Consumers’ and Producers’ Goods in the General Sales
Tax.” National Tax Journal 37 No. 2 (June,
1989): 167–79.
Duncombe, William.
“Economic Change and
the Evolving State Tax Structure: The Case of the
Sales Tax.” National Tax Journal 45 No. 3
(September, 1992): 299–314.
Schaefer, Jeffrey M.
“Clothing Exemptions and
Sales Tax Regressivity.” American Economic
Review 59 No. 4 (September, 1969): 596–9.
Dye, Richard, and Therese McGuire.
“Growth
and Variability of State Individual Income and
General Sales Taxes.” National Tax Journal 44
No. 1 (March, 1992): 55–66.
“Selected NIPA Tables.” Survey of Current
Business 76 (July, 1996): 5–28.
Fox, William, and Matthew Murray.
“Economic Aspects of Taxing Services.” National
Tax Journal 41 No. 1 (March, 1988): 19–36.
Siegfried, John, and Paul Smith.
“The
Distributional Effects of a Sales Tax on Services.”
National Tax Journal 44 No. 1 (March, 1991): 41–
53.
Joulfaian, David, and James Mackie.
“Sales
Taxes, Investment, and the Tax Reform Act of
Simons, Henry C. Federal Tax Reform.
Chicago: University of Chicago Press, 1950.
164
National Tax Journal
Vol 50 no. 1 (March 1997) pp. 149-65
SYMPOSIUM ON A NATIONAL RETAIL SALES TAX
Tait, Alan A. Value Added Tax: International
Practices and Problems. Washington, D.C.:
International Monetary Fund, 1988.
U.S. Bureau of the Census.
1995 State Tax
Collection Data by State [Homepage of U.S.
Census Bureau], [Online]. Available: http://
www.census.gov/ftp/pub/govs/www/sttax95.
html [1995, Access Date: December, 1996].
U.S. Bureau of the Census.
1994 State Tax
Collection Data by State [Homepage of U.S.
Census Bureau], [Online]. Available: http://
www.census.gov/ftp/pub/govs/www/sttax94.
html [1995, Access Date: December, 1996].
Washington Department of Revenue.
Washington State Excise Tax Noncompliance
Study. Olympia: Department of Revenue, 1990.
165