Sales Taxes: Electronic Commerce Growth Presents Challenges; Revenue
Losses Are Uncertain (Letter Report, 06/30/2000, GAO/GGD/OCE-00-165).
Pursuant to a congressional request, GAO provided information on the
impact of electronic commerce (e-commerce) growth on state and local
government sales tax collections, focusing on: (1) how taxes associated
with the sale of goods and services over the Internet differ from taxes
associated with sales by other remote sellers and in-store sellers; (2)
the extent to which each state relies on sales and use tax revenues to
fund the services they provide; (3) how much revenue state and local
governments are losing this year by not being able to collect sales and
use taxes on sales made by all remote sellers and, particularly, by
Internet sellers; and (4) how much revenue state and local governments
would likely lose in 2003 under various growth scenarios for all remote
and Internet sales.
GAO noted that: (1) in-store, Internet, and other remote sales are
generally taxed at the same rate by a state or local government; (2)
however, compliance rates differ significantly depending on nexus; (3)
in-store and remote sellers (including Internet sellers) with a
substantial presence, or nexus, with the state are legally required to
collect and remit the tax; (4) for sales without nexus, purchasers are
themselves legally required to remit the tax, but purchaser compliance
is generally much lower than seller compliance; (5) the continued growth
of e-commerce is likely to magnify existing compliance problems and, as
new types of digital goods and transactions are developed, create new
ones, such as identifying the location of a sale; (6) such compliance
challenges have led some observers to question the long-term viability
of sales and use taxes; (7) states' reliance on general sales
taxes--whether measured as a percentage of tax revenues, own-source
revenues, or total general revenues--varies considerably across states;
(8) little empirical data exist on the key factors needed to calculate
the amount of sales and use tax revenues that state and local
governments lose on Internet and other remote sales; (9) what
information does exist is often of unknown accuracy; (10) GAO
constructed scenarios representing different assumptions about the
important determinants of the loss; (11) under all of GAO's scenarios,
the size of the tax loss from Internet sales for 2000 is less than 2
percent of aggregate general sales tax revenues; (12) under all of GAO's
scenarios, the size of the loss from all remote sales is less than 5
percent of aggregate sales tax revenues; (13) the rapid change in the
Internet economy makes projections of revenue losses from Internet and
total remote sales for future years even more uncertain than they are
for 2000; (14) under the scenarios GAO constructed for 2003, the size of
the tax loss from Internet sales ranged from less than 1 percent to
about 5 percent of projected sales tax revenues; (15) for all remote
sales, the corresponding loss ranged from about 1 percent to about 8
percent; (16) the results of GAO's scenarios highlight the importance of
developing better data about Internet tax losses and understanding the
limits of such data; and (17) some of GAO's scenarios show tax losses
that by 2003 could present significant revenue challenges for state and
local government officials, while other scenarios produce smaller
revenue losses.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: GGD/OCE-00-165
TITLE: Sales Taxes: Electronic Commerce Growth Presents
Challenges; Revenue Losses Are Uncertain
DATE: 06/30/2000
SUBJECT: State governments
Economic analysis
Tax expenditures
Sales taxes
Computer networks
Municipal governments
Interstate commerce
IDENTIFIER: Internet
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
GAO/GGD/OCE-00-165
United States General Accounting Office
GAO
Report to Congressional Requesters
June 2000
GAO/GGD/OCE-00-165
SALES TAXES
Electronic Commerce Growth Presents Challenges;
Revenue Losses Are Uncertain
Ordering Copies of GAO Reports
The first copy of each GAO report and testimony
is free. Additional copies are $2 each. Orders
should be sent to the following address,
accompanied by a check or money order made out
to the Superintendent of Documents, when
necessary. VISA and MasterCard credit cards are
accepted, also. Orders for 100 or more copies to
be mailed to a single address are discounted 25
percent.
Order by mail:
U.S. General Accounting Office
P.O. Box 37050
Washington, DC 20013
or visit:
Room 1100
700 4th St. NW (corner of 4th and G Sts. NW)
U.S. General Accounting Office
Washington, DC
Orders may also be placed by calling (202) 512-
6000 or by using fax number (202) 512-6061, or
TDD (202) 512-2537.
Each day, GAO issues a list of newly available
reports and testimony. To receive facsimile
copies of the daily list or any list from the
past 30 days, please call (202) 512-6000 using a
touch-tone phone. A recorded menu will provide
information on how to obtain these lists.
Viewing GAO Reports on the Internet
For information on how to access GAO reports on
the INTERNET, send e-mail message with "info" in
the body to:
[email protected]
or visit GAO's World Wide Web Home Page at:
http://www.gao.gov
Reporting Fraud, Waste, and Abuse in Federal
Programs
To contact GAO FraudNET use:
Web site:
http://www.gao.gov/fraudnet/fraudnet.htm
E-Mail: [email protected]
Telephone: 1-800-424-5454 (automated answering
system)
United States
General Accounting Office
Washington, D.C. 20548-0001
Official Business
Penalty for Private Use $300
Address Correction Requested
Bulk Rate
Postage & Fees Paid
GAO
Permit No. G100
(268892)
Contents
Page 241GAO/GGD/OCE-00-165 Sales Taxes on e-Commerce
Letter 1
Appendix I 28
Methodology for the
Revenue Loss Scenarios
The Volume and Composition of Internet 28
and Other Remote Sales
The Taxability of Remote Sales 30
The Extent to Which Remote Sellers 32
Already Collect Tax
The Extent to Which Purchasers Already 34
Pay Tax
Auction Sales 35
Behavioral Response 35
Sensitivity Analysis 36
Appendix II 37
Constitutional
Restrictions on State
Authority to Impose
Sales and Use Taxes
Commerce and Due Process Clauses 37
Dual Entity Arrangements 42
Appendix III 44
General Sales Tax
Provisions
Appendix IV 56
State and Local
Government Reliance on
Sales Tax Revenue
Appendix V 59
State and Local
Government Sales and Use
Tax Losses
Appendix VI 62
List of Contacts
Private Sector Research Groups 62
National Organizations Representing 62
the Public Sector
Industry Representatives 62
Department of Commerce 62
Department of the Treasury 62
States 62
Academics and Others 63
Appendix VII 64
GAO Contacts and Staff
Acknowledgments
Tables Table 1: State and Local Government 14
Reliance on General Sales Taxes
Table 2: State and Local Sales Tax 19
Losses for All Remote Sales and
Internet Sales Only in 2000
Table 3: Sensitivity of Higher 19
Scenario Revenue Losses in 2000 to
Changes in Key Assumptions
Table 4: State and Local Sales Tax 21
Losses for All Remote Sales and
Internet Sales Only in 2003
Table 5: Sensitivity of Higher 21
Scenario Revenue Losses in 2003 to
Changes in Key Assumptions
Table I.1: Choices of Sales Estimates 30
Used in the Scenarios
Table I.2: Seller Collection Rate 33
Estimates for Selected Categories of
All Business-to-Consumer Remote Sales
Table I.3: Seller Collection Rate 33
Estimates for Selected Categories of
Business-to-Consumer Internet Sales
Table I.4: State and Local Sales Tax 36
Losses for All Remote Sales and
Internet Sales Alone in 2000
Table I.5: State and Local Sales Tax 36
Losses for All Remote Sales and
Internet Sales Alone in 2003
Table III.2: General Sales and Use Tax 48
Treatment for Selected Services That
Can be Sold by Remote Sellers, 1996
Table III.2: (cont.) 49
Table III.2: (cont.) 50
Table III.2: (cont.) 51
Table III.3: State and Local General 53
Sales Tax Rates and Combined Rates
for Selected Cities, March 1999
Table IV.1: State Government Reliance 56
on Sales Tax Revenue, 1998
Table IV.2: Local Government Reliance 57
on Sales Tax Revenue, 1995-96
Table V.1: State and Local Sales and 59
Use Tax Losses for All Remote Sales
and Internet Sales Alone in 2000
Table V.2: State and Local Sales and 60
Use Tax Losses for All Remote Sales
and Internet Sales Alone in 2003
Figures Figure 1: Steps Involved in Computing 8
Revenue Losses
Figure 2: Responsibility for Sales and 12
Use Tax Collection and Remittance
Figure III.1: Sales and Use Tax 44
Treatment of Selected Goods That Can
Be Sold by Remote Sellers, 1999
B-284955
Page 4GAO/GGD/OCE-00-165 Sales Taxes on e-Commerce
B-284955
June 30, 2000
The Honorable George V. Voinovich, Chairman
Subcommittee on Oversight of Government
Management,
Restructuring, and the District of Columbia
Committee on Governmental Affairs
United States Senate
The Honorable Bob Graham
United States Senate
The rapid growth of electronic commerce (e-
commerce), especially the sale of goods and
services over the Internet, has fueled a debate
about the taxation of such commerce. On the one
hand, there are concerns about the impact of e-
commerce growth on state and local government
sales tax collections. These concerns arise
because, while states can impose a tax on
residents' purchases from out-of-state vendors,
they cannot impose an obligation on those vendors
to collect the tax unless the vendor has a
substantial presence, or nexus, in the state.1
Without collection by sellers, and absent
intrusive and costly collection actions aimed at
purchasers, portions of sales and use taxes can be
avoided.2
On the other hand, there are concerns that the
taxation of Internet sales could slow innovation
and growth in the economy. E-commerce and the
Internet are viewed as part of a productivity-
enhancing "information technology revolution."
Taxation of e-commerce, it is feared, could
discourage such innovation.
Congress has recognized the need for more
information about the implications of taxing e-
commerce. For example, in 1998, Congress passed
the Internet Tax Freedom Act,3 which, among other
actions, established the Advisory Commission on
Electronic Commerce to study the tax treatment of
Internet transactions.4 The act also temporarily
banned the imposition of certain types of taxes on
e-commerce, but not the collection of existing
taxes, such as sales and use taxes.
Given your interest in the taxation of e-commerce,
and particularly the impact of e-commerce growth
on state and local government sales tax
collections, you asked us for information specific
to sales and use tax collections for Internet
sales as well as for all remote sales. In response
to your request, this report addresses the
following questions:
1. How do the taxes associated with the sale of
goods and services over the Internet differ from
taxes associated with sales by other remote
sellers and in-store sellers?
2. To what extent does each state rely on sales
and use tax revenues to fund the services they
provide?
3. How much revenue are state and local
governments losing this year by not being able to
collect sales and use taxes on sales made by all
remote sellers and, particularly, by Internet
sellers?
4. How much revenue would state and local
governments likely lose in 2003 under various
growth scenarios for all remote and Internet
sales?
In light of the considerable uncertainty
surrounding the volume of Internet and all remote
sales and any resulting tax losses, we agreed with
your office to model different possible scenarios.5
The scenarios are based on different assumptions
about the volume of Internet and remote sales, the
proportion of sales that are taxable, the
proportion in different taxing jurisdictions, the
proportion of taxes actually collected, and other
factors that affect tax revenue. We developed
lower and higher scenarios to demonstrate an
overall range of uncertainty and the potential
effects on revenue loss. We also performed a
sensitivity analysis to show the revenue loss
effects due to uncertainty about specific
assumptions. Because of the uncertainty
surrounding the assumptions, the scenarios are not
estimates but, rather, are illustrations of the
importance of the various assumptions. The data
and specific assumptions that we used in
developing our scenarios are described further in
the methodology section of this letter and in
appendix I.
Results in Brief
In-store, Internet, and other remote sales are
generally taxed at the same rate by a state or
local government. However, compliance rates differ
significantly depending on nexus. In-store and
remote sellers (including Internet sellers) with a
substantial presence, or nexus, with the state are
legally required to collect and remit the tax. For
sales without nexus, purchasers are themselves
legally required to remit the tax, but purchaser
compliance is generally much lower than seller
compliance. The continued growth of e-commerce is
likely to magnify existing compliance problems
and, as new types of digital goods and
transactions are developed, create new ones, such
as identifying the location of a sale. Such
compliance challenges have led some observers to
question the long-term viability of sales and use
taxes.
States' reliance on general sales taxes-whether
measured as a percentage of tax revenues, own-
source revenues, or total general revenues-varies
considerably across states.6 For example, in
Delaware, Montana, New Hampshire, and Oregon,
neither state nor local governments collect such
taxes. In contrast, state governments in Florida,
Nevada, South Dakota, Tennessee, Texas, and
Washington and local governments in Louisiana
obtain over 50 percent of their tax revenues from
general sales taxes. In 1999, state and local
governments collected $203 billion in general
sales tax revenues. On average, general sales
taxes account for 33 percent of state and 11
percent of local tax revenues.
Little empirical data exist on the key factors
needed to calculate the amount of sales and use
tax revenues that state and local governments lose
on Internet and other remote sales. What
information does exist is often of unknown
accuracy. Consequently, we constructed scenarios
representing different assumptions about the
important determinants of the loss. Under all of
our scenarios, the size of the tax loss from
Internet sales for 2000 is less than 2 percent of
aggregate general sales tax revenues. Under all of
our scenarios, the size of the loss from all
remote sales is less than 5 percent of aggregate
sales tax revenues.
The rapid change in the Internet economy makes
projections of revenue losses from Internet and
total remote sales for future years even more
uncertain than they are for 2000. Under the
scenarios we constructed for 2003, the size of the
tax loss from Internet sales ranged from less than
1 percent to about 5 percent of projected sales
tax revenues (see p. 21). For all remote sales,
the corresponding loss ranged from about 1 percent
to about 8 percent.
The results of our scenarios highlight the
importance of developing better data about
Internet tax losses and understanding the limits
of such data. Some of our scenarios show tax
losses that by 2003 could present significant
revenue challenges for state and local government
officials, while other scenarios produce smaller
revenue losses. Better data, from efforts such as
one by the Bureau of the Census, could reduce the
uncertainty. However, even with better data, the
rapid and fundamental nature of innovations in e-
commerce means that policymaking regarding the tax
treatment of Internet sales will be done in an
environment of significant uncertainty.
Background
Sales and use taxes are imposed on specific sales
transactions. Generally, states require that in-
state sellers collect sales tax on the goods and
services they sell at the time of sale, based on
the price or value of the goods or services sold.
States require that out-of-state remote sellers
collect a use tax on the sale of goods and
services if the sellers have a substantial
presence, or nexus, with the state.7 The use tax,
which complements the sales tax, is imposed on the
purchaser for the privilege of use, ownership, or
possession of taxable goods or services. If the
out-of-state remote seller does not collect the
use tax, the purchaser is required to remit the
tax.
Based on case law interpreting the constitutional
requirements, out-of-state remote sellers
generally meet the nexus standards if they have an
office or place of business, agent, or property in
the taxing state. Nexus is not established if the
seller's property is insignificant. The Supreme
Court has ruled that contact with in-state
purchasers by mail or common carrier, only, does
not constitute nexus.8 Although a business can
establish dual entity operations to minimize tax
liabilities, the extent to which Internet and in-
store operations may interact and retain their
distinction has not been resolved.
Forty-five states and the District of Columbia
have general sales tax programs under which they
administer the sales and use tax provisions.9
About 7,600 local jurisdictions have general sales
tax programs authorized by 34 states. Generally,
state governments administer the state and local
sales taxes.10
In 1999, the combined state and local general
sales and use tax rates ranged from about 5 to 8
percent in most states. State general sales tax
rates were about 4 or 5 percent in most states.
Local general sales tax rates varied more and
ranged from 0.5 percent to about 4 percent in some
jurisdictions.
A number of prior studies have made nationwide
estimates of the amount of sales and use tax
revenues that state and local governments lose on
Internet and other remote sales.11 The Advisory
Commission on Intergovernmental Relations (ACIR)
published a series of studies from 1986 through
1994 estimating revenue losses from mail-order
sales. ACIR estimated that in 1994, before the
recent growth in Internet use, the state and local
revenue loss was about $3.3 billion.12 In more
recent years, there have been efforts to estimate
the lost tax revenue from Internet sales. A study
by Ernst & Young for the eCommerce Coalition13
concluded that the sales and use taxes not
collected from the increase in remote sales due
the Internet was less than $170 million in 1998.
The authors of that study did not estimate losses
on business-to-business Internet sales, but they
suggested that these losses would be very small.
Researchers, Goolsbee and Zittrain,14 assumed zero
revenue losses from business-to-business Internet
sales when they estimated that tax losses from
Internet sales in 1998 ranged from $210 million to
$430 million and that losses would be about $3.5
billion in 2003.
In contrast, a recent study by researchers, Bruce
and Fox,15 produced much larger revenue loss
estimates because the authors assumed that more
than half of business-to-business Internet sales
are taxable and that compliance on the part of
purchasers is well below 100 percent. Bruce and
Fox estimated that the revenue loss from Internet
sales will grow from $1.23 billion in 1999 to
$10.8 billion in 2003. Finally, a study by
Forrester Research, Inc.,16 which focused only on
business-to-consumer sales, estimated that sales
tax revenue losses from those sales were $525
million in 1999. The authors of most of these
studies acknowledged that there is a limited
empirical basis for many of the assumptions that
need to be made when making such estimates.
The Internet Tax Freedom Act established the
Advisory Commission on Electronic Commerce to
study "Federal, State and local, and international
taxation and tariff treatment of transactions
using the Internet and Internet access and other
comparable intrastate, interstate or international
sales activities."17 The majority of the Commission
issued its report to Congress in April 2000. The
Commission voted in favor of a policy proposal
relating to state and local government taxation of
Internet sales that, among other things, would:
� extend the current moratorium on multiple and
discriminatory taxation of e-commerce;
� encourage state and local governments to make
their sales and use taxes more uniform;
� prohibit taxation of sales of digitized goods
and their nondigitized equivalents; and
� modify the definition of nexus in order to
allow out-of-state vendors to conduct additional
operations in a state, such as allowing for the
return of merchandise or for repairs, without
subjecting the vendor to the requirement of
remitting sales taxes to the state.
Those voting for the proposal argued that the it
would foster innovation and growth of the Internet
and e-commerce while recognizing the role of state
and local governments to continue providing needed
services to their citizens. Those who voted
against or abstained were particularly concerned
that it would result in large revenue losses for
state and local governments, impairing their
ability to provide needed services to their
citizens. Since these proposals did not receive
the two-thirds vote required by the Internet Tax
Freedom Act, they were not given the status of
formal findings or recommendations of the
Commission.
Scope and Methodology
To determine how taxes associated with the sale of
goods and services by Internet sellers, other
remote sellers, and in-store sellers differ, we
reviewed information relating to (1) the federal,
state, and local taxes that apply to sales goods
and services and to the businesses that sell them
and (2) the conditions under which sellers are
required to collect state and local sales and use
taxes. We reviewed published tax guides, conducted
legal research of precedent-setting court cases,
and interviewed officials from state tax agencies,
the Department of the Treasury, and national
organizations representing sellers and state and
local governments. We also attended numerous
conferences addressing tax issues and the
Internet, including the meetings of the Advisory
Commission on Electronic Commerce.
To determine the extent that state and local
governments rely on sales and use tax revenues, we
analyzed data from the Census Bureau relating to
U.S. totals for those revenues in calendar year
1999. We also analyzed Census data on state
government revenues for fiscal year 1998 and local
government revenues for fiscal year 1996, the
latest years for which state-by-state data were
available.
To model different scenarios for the state and
local government sales and use tax revenue losses,
we obtained estimates of the total amount of sales
that will be transacted remotely in 2000 and
subjected them to a series of computations that
reflect (1) details of state sales tax systems and
(2) assumptions relating to the various factors
that determine the size of the revenue losses. In
addition to the revenue loss associated with all
remote sales, we modeled different scenarios for
the loss that was attributable to Internet sales
alone. Figure 1 summarizes the steps in our
revenue loss computations.
To approximate the amount of remote sales that
will be taxable, we apportioned the sales data
among individual states and then subtracted state-
specific exemptions for particular types of
products, services, purchasers, and uses. We then
multiplied the taxable sales in each state by the
appropriate tax rate to obtain an approximation of
the sales or use tax owed to each state. To
compute the amount of revenue that each state
government is unable to collect, we made
assumptions regarding the amount of the tax owed
on remote sales that would be paid to each state
by either sellers or purchasers. We then
subtracted that amount from the amount owed to the
state to obtain the state-level revenue loss.
Figure 1: Steps Involved in Computing Revenue
Losses
Source: GAO methodology.
We report high and low estimates for all remote
and Internet only sales for the years 2000 and
2003. To calculate the potential sales and use tax
losses for the higher scenario, we use the
endpoint of the range for each of our assumptions
that leads to a higher revenue loss. For example,
we use the high estimate of sales, a low estimate
of nexus for sellers, a low rate of purchaser
compliance, and a low rate of product and
purchaser exemptions. We use the other endpoints
of our estimated ranges to calculate the sales tax
losses for our low tax loss scenario. Combining
assumptions in this way increases the likelihood
that the actual tax losses fall between the high
tax and low tax scenario results.
We obtained the local government revenue loss in
each state by multiplying the state government
loss by the ratio of local sales tax collections
to state sales tax collections in each state.18 We
also modeled the amounts of revenue that state and
local governments would potentially lose on
Internet and other remote sales in 2003 under
alternative scenarios for the growth of those
sales.
There were few reliable data sources on which to
base the calculations and adjustments summarized
above. The growth of on-line sales has been so
rapid that the economic data available from
federal and state governments have not been
modified to provide this kind of information, and
those that are collected are not well suited for
this purpose. Most of the sales estimates that are
available are from private-sector sources, and
some of these providers view their data sources
and details as proprietary. Finally, projections
of sales are particularly difficult to make given
the rapidly changing environment and the
importance of decisions yet to be made by
consumers, businesses, and policymakers that will
determine the ultimate level of those sales. We
were not able to assess the accuracy of any of the
available estimates and projections of sales.
In addition to the uncertainty regarding the
magnitude of remote sales, there is considerable
uncertainty about the amount of tax that state and
local governments are already collecting from
these remote sales and the extent to which
Internet sales replace other forms of remote
sales. Little empirical data exist to reduce these
uncertainties. To ensure that we did not overlook
any important data, we reviewed the existing
literature and spoke with numerous experts in
academia, the private sector, and in government,
including officials from 17 states.19 In certain
cases, we collected our own data on important
parameters where we believed we had an opportunity
to improve upon the information that prior
analysts had used. For example, we gathered
information from 150 large remote retailers
regarding the specific states for which they were
already collecting sales taxes. We also used
Department of Commerce data as a basis for our
assumptions relating to the proportions of
business-to-business remote sales that are sold to
various types of tax-exempt purchasers. We also
performed a sensitivity analysis to show the
revenue loss effects due to uncertainty about
specific assumptions.
We also subjected our work to peer review by noted
experts in the field of tax policy. These experts
agreed with the general approach that we followed
in making our estimates, but they provided
different estimates about specific factors that
determine the size of the revenue loss, such as
the extent to which purchasers are currently
complying with their use tax obligations. The
experts confirmed that uncertainty surrounds many
of these factors incorporated into the model. Our
approach reflects their suggestions and comments,
particularly the use of ranges of estimates for
key determinants of the revenue loss.
The definition of revenue loss that we use in our
scenarios is the amount of sales or use tax owed
on remote sales, minus any amount already being
paid by sellers or purchasers. There are two
reasons why this amount is likely to be higher
than the amount that state and local governments
would receive if all remote retailers were
required to collect and remit taxes on their
sales. First, even if all remote sellers were
required to collect the taxes due on their sales,
compliance is not likely to be 100 percent.
Second, the total volume of taxable sales may
decline in response to a higher rate of tax
collection on these sales. In computing the
revenue loss attributable solely to the advent of
Internet sales, we excluded losses associated with
the portion of Internet sales that would have been
transacted by other remote means, such as mail
order, in the absence of the Internet.
Detailed information about our methodology,
including the data sources that we used, are
provided in appendix I. We conducted our work from
June 1999 to May 2000 in accordance with generally
accepted government auditing standards.
Tax Liabilities for Internet and Other Sales Are
Generally the Same, but Compliance Can Differ
For a particular good or service and taxing
jurisdiction, remote sales, Internet sales, and in-
store sales are generally subject to the same rate
of sales or use tax. However, tax compliance
differs by type of sale, with nexus being an
important influence. For example, remote sellers
with nexus are required to collect the tax but
sellers without nexus are not. E-commerce presents
compliance challenges for sales and use tax
administration beyond those created by other
remote sales.
Tax Liabilities for Internet, Other Remote, and In-
store Sales Are Generally the Same
After reviewing published information and talking
to state tax officials, officials from several
national organizations representing state
governments, and private-sector representatives,
we were unable to identify significant differences
in the tax rates on in-store sales, Internet
sales, and other remote sales. Although states
vary in which goods and services they tax and in
their tax rates for a given good or service in a
particular location, the rate does not depend on
whether the sale is in-store, Internet, or other
remote.20
The type of goods and services included in the
sales and use tax base vary by taxing
jurisdiction. In states with sales and use taxes,
retail goods are taxed unless exempted. The list
of exempt goods varies by state. For example, most
but not all states exempt groceries. Unlike goods,
services are generally untaxed, although there are
exceptions. Tables III.1 and III.2 in appendix III
provide more detailed information about the tax
treatment of goods and services by state.
Sales and use tax rates also vary by taxing
jurisdiction. Five states do not impose a state-
level general sales or use tax. The 1999 combined
state, county, and city tax rates for selected
jurisdictions ranged from 4 to 9 percent. Table
III.3 in appendix III provides more details on
1999 general sales tax rates for each state and
selected local jurisdictions.
Collection Responsibilities for Remote Retailers
Depend Upon Nexus
Whether a remote retailer is legally required to
collect a sales tax depends on whether the
retailer has substantial presence or nexus with
the taxing jurisdiction. As defined by case law,
remote sellers generally meet the nexus standard
if they have an office or other place of business,
property, or agent in the taxing state. Remote
sellers, including Internet sellers, that have
nexus with a taxing state are responsible for
collecting the use tax from purchasers at the time
of sale and remitting the tax to the taxing
jurisdiction. Remote sellers with nexus have the
same tax collection responsibilities as an in-
store seller. Figure 2 summarizes tax collection
and remittance responsibilities for in-store,
Internet, and other remote sellers.
Court decisions interpreting the provisions of the
Commerce and Due Process Clauses of the
Constitution preclude the states from requiring a
remote seller without nexus to collect the use
tax. If the remote seller does not collect a use
tax, then the purchaser is responsible for paying
the tax to the taxing state where they use,
consume, or store the purchased goods or service.
Appendix II contains a more detailed discussion of
the constitutional restrictions on state authority
to require a remote retailer to collect the use
tax.
Figure 2: Responsibility for Sales and Use Tax
Collection and Remittance
Source: GAO analysis.
Sales and Use Tax Compliance Differs by Type of
Sale
While reliable national estimates of sales and use
tax compliance do not exist, state officials and
other observers believe that compliance is highest
for in-store sales, next highest for remote sales
with nexus, and lowest for remote sales without
nexus.21 Their belief rests on three facts. First,
in-store sellers are more visible to the states
than remote sellers, leaving the states better
positioned to enforce compliance through audits
and other actions. Second, the states have legal
authority to enforce sales and use tax collection
by in-store sellers and remote sellers with nexus.
Third, because of enforcement costs, the states
generally rely on purchasers to voluntarily comply
with the use tax when there is no nexus. The
differences in compliance thus depend on whether
the sale is in-store or remote and, for remote
sales, on whether the remote seller has nexus.
Electronic Commerce Presents Challenges for Sales
and Use Tax Systems
Electronic commerce and the related changes in
technology present challenges for the
administration of sales and use taxes. One
challenge is presented by continued growth in the
volume of Internet sales. To the extent that such
growth occurs, it increases remote sales where
compliance is already most problematic. Another
challenge is that the expanding variety of e-
commerce transactions and products may create new
types of compliance problems, such as identifying
the location and nature of a sale. Such challenges
have led some observers to question the long-term
viability of the sales and use tax system.
Although the future growth rate of Internet sales
is not known, certain characteristics favor the
rapid growth of Internet sales. For example,
Commerce has reported that e-commerce not only
reduces the cost and time of doing business but
also provides alternative shopping sites, expands
existing markets, and creates new markets. E-
commerce also frees some sellers from the
"geographic confines and the costs of running
actual stores." These characteristics have the
potential to increase the number of remote sellers
and purchasers as well as increase the volume of
remote sales. To the extent that such sales growth
occurs, it will magnify the existing sales and use
tax compliance problems associated with remote
sales, such as the difficulty of enforcing
compliance by purchasers in the case of remote
sales without nexus.
The expanding variety of electronic transactions
may also create new compliance challenges. Shifts
from traditional forms of sales to Internet sales
can make it more difficult to identify the
location of the buyer and the seller, the status
(business, individual, other) of the buyer or
seller, and the nature of the product itself. In
terms of the location, both sellers and purchasers
may have multiple locations, and the Internet
makes it easier for these firms to conduct their
transactions from the location that offers the
greatest tax advantages. Businesses may also
choose to establish a presence in certain
jurisdictions in order to maximize these
advantages. As a result, determining the location
of buyers and the sellers' activities for nexus
purposes, which is important for the collection of
sales and use taxes, is more difficult in an
environment with Internet sales.
A related challenge for the collection of sales
and use taxes is determining the status of the
buyer and seller in Internet transactions. The
status of the seller, for example, is relevant
since certain sales by individuals are not subject
to sales and use taxes. However, the development
of new markets, such as Internet auctions, has
created a new opportunity for businesses as well
as individuals to avoid sales and use taxes. To
the extent that businesses are using these new
markets to make sales, it would be necessary for
tax authorities to be able to identify those
sellers as businesses rather than as individuals
in order to assess the appropriate taxes.
The increasing variety of digital products also
creates challenges for sales and use taxes.
Currently, purchasers can buy many digital
products, such as books, music, software, and
videos, that were only available as tangible
products a few years ago. The sale of digital
products often makes it more difficult for states
to determine if there was a sale, the point of
sale, and the cost or value of the products sold.
Further complicating state tax requirements and
compliance efforts is the taxability of digital
products and services, which can be questionable.
Case law defining the conditions that must exist
before a state can require a remote seller to
collect a use tax refers to the taxed goods as
"tangible personal property." States may not be
able to require the remote sellers to collect the
use tax on the sale of intangible digital products
unless they categorize these products as
intangible services.
General Sales Taxes Account for 33 Percent of
State Government and 12 Percent of Local
Government Tax Revenues
On average, general sales taxes account for 33
percent of the state government tax revenue and 11
percent of local government tax revenue. However,
reliance on general sales taxes varies
considerably across states. Table 1 shows the
reliance of state and local governments on sales
taxes whether measured as a percentage of tax
revenues, own-source revenues, or total general
revenues.
Table 1: State and Local Government Reliance on
General Sales Taxes
General sales tax revenue State Local
as a percentage of:
Percen Yeara Perce Yeara
t nt
Total tax revenue 33 1999 11 1999
Total general own-source 25 1998 7 1996
revenue
Total general revenue 18 1998 4 1996
aMost recent year that data were available.
Source: GAO analysis based on Bureau of the Census
data.
Overall, state and local governments collected
$203 billion in general sales and use taxes in
1999. In 34 states, both state and local
governments collect at least some revenue from
general sales taxes; while in Delaware, New
Hampshire, Montana, and Oregon, neither level of
government collects general sales tax revenue. In
the remaining 12 states, only one level of
government collects a general sales tax.
Tennessee, Florida, and Nevada are the three state
governments that rely most heavily on general
sales tax revenues, whether measured as a share of
tax revenues, own-source revenues, or all general
revenues.22 Each of these states obtains at least
53 percent of their tax revenues, 44 percent of
their own-source general revenues, and 28 percent
of all their general government revenues from
general sales taxes. Local governments in
Louisiana, Oklahoma, and Alabama rely on general
sales taxes for greater shares of their tax
revenues (at least 40 percent each) and own-source
revenues (at least 19 percent each) than do local
governments in other states. Local governments in
Louisiana, Oklahoma and Colorado rely on general
sales taxes for the greatest shares of total
general revenues (at least 12 percent each). (See
app. IV for state and local revenue information.)
State and Local Revenue Losses From Internet and
Other Remote Sales Are Small Under Most Scenarios
for 2000
The amount of sales and use tax revenues that
state and local governments may lose by not being
able to collect those taxes on Internet and other
remote sales is difficult to model because
considerable uncertainty surrounds the factors
that determine the loss. However, under our
scenarios representing different assumptions about
the important determinants of the loss, the size
of the loss from Internet sales for 2000 is less
than 2 percent of aggregate sales and use tax
revenues. The size of the loss from all remote
sales is less than 5 percent of aggregate sales
and use tax revenues.
Considerable Uncertainty Surrounds the
Determinants of State and Local Revenue Losses
Important factors that determine the tax loss on
Internet and remote sales are the volume of
Internet and other remote sales, the portion of
the sales subject to tax, the extent of compliance
by sellers or purchasers, and the extent to which
Internet sales displace other types of remote
sales. However, as discussed below, little data
exist on these factors and the accuracy of the
information that exists is often unknown.
The Volume of Remote Sales
No statistical agency compiles data on the total
value or composition of Internet or other remote
sales. Although the Census Bureau has collected
data on mail-order sales, their figures do not
include the mail-order sales of any firms whose
primary business is not mail order. As a result,
their figures significantly understate total mail-
order sales. The Bureau has plans to produce
comprehensive data on Internet sales based on its
annual surveys of firms in the manufacturing,
wholesale, retail, and other sectors, but much of
these data will not be available until 2001. A
number of private-sector consulting firms make
estimates and projections of Internet and other
remote sales. It is difficult to assess the
accuracy of any of their estimates. Given the
uncertainty surrounding available projections of
total remote and Internet sales for 2000 and 2003,
we use a broad range of projections in the
scenarios that we present below. Appendix 1
describes the projections that we use.
The Taxability of Remote Sales
The rate at which a remote sale is taxed can vary
depending on the state of residence of the
purchaser, the nature of the product or service
being sold, the nature of the purchaser, and the
use that the purchaser makes of the product or
service. For example, a computer purchased by a
state agency or by a business that uses it in a
manufacturing process may not be taxed in a
particular state, even though a computer purchased
by a law firm in that same state would be taxed.23
We were able to obtain some estimates of Internet
and total remote sales that were disaggregated by
broad categories of purchasers (businesses versus
individuals) and by broad categories of products
and services, but we could not determine the
accuracy of these estimated disaggregations. We
were unable to identify any estimates of sales by
taxable versus tax-exempt purchaser or by taxable
versus tax-exempt use.
The Extent to Which Remote Sellers Already Collect
Tax
When remote sellers have nexus in states in which
they make sales, they are required to collect any
sales taxes that apply to those sales. In
addition, some remote sellers collect taxes
voluntarily, even when they do not have nexus. The
proportion of tax already being collected by
sellers (which we call the "seller collection
rate") varies by type of product. For example, a
very high proportion of the taxes due on cable
television services and utilities are likely to be
collected because a large proportion of those
services are provided by businesses with in-state
physical infrastructure. In contrast, a relatively
low proportion of the taxes due on remote sales of
computers are likely to be collected from sellers
because a large proportion of these sales are made
by sellers who have nexus in only a few states.
As a result of this variation across products and
services, the overall seller collection rate will
change over time as the composition of remote
sales changes. The collection rate within
particular product categories will also change as
the market shares, physical locations, and
organizational form of particular businesses
change. Such change is particularly rapid in the
Internet economy. Policy changes that affect the
determination of nexus could also have substantial
impacts on seller collection rates.
We found two studies containing empirical
estimates of the seller collection rate, but they
were either dated or limited in scope. Other
analysts who have estimated the revenue losses
associated with Internet and other remote sales
have used a variety of assumptions regarding the
proportion of tax already being paid. We based our
own assumptions on information we obtained from
businesses that account for large proportions of
Internet and other remote sales as well as on
discussions with state tax officials (see app. I).
The Extent to Which Purchasers Already Pay Tax
No comprehensive data are available relating to
the rate of use tax compliance on the part of
purchasers. Only 2 of the 17 states we contacted
provided empirically based estimates of use tax
compliance on the part of business purchasers, and
those estimates date from the early 1990s-before
the widespread use of the Internet. The two
states, Washington and Wisconsin, used the results
of state use tax audits to estimate that
approximately 80 percent of the use tax owed by
business purchasers in their states was paid
voluntarily. An official from the Michigan
Department of Treasury believed that nearly all
businesses in the state comply with the use tax
because it is covered in single business tax
audits. That official noted that compliance rates
are likely to vary from state to state, depending
on enforcement efforts. An official from
Connecticut's Department of Revenue Services
believed that businesses pay about 65 percent of
their use tax liabilities, while an official from
Ohio's Department of Taxation believed that
businesses' rate of compliance is 75 percent.
Other states were unable to provide estimates but
some experts believe the rate could be as low as
50 percent for non-motor-vehicle purchases.24 Our
scenarios reflect the broad range of opinions that
state officials and other analysts have regarding
this compliance rate.
In contrast to the wide range of opinion on the
compliance of business purchasers was a wide
consensus among the state officials and other
experts who provided estimates that use tax
compliance by individual purchasers was extremely
low-on the order of 0 to 5 percent. However, there
was also a wide consensus that compliance with the
use tax on motor vehicles is close to 100 percent
because the taxes must be paid before those
vehicles can be registered.
The Extent to Which Internet Sales Displace Other
Types of Sales
A portion of total Internet sales displaces sales
that would have been transacted in stores; another
portion displaces sales that would have been
transacted through other remote channels, such as
mail orders; and a final portion represents sales
that would not have occurred in the absence of the
Internet. In the results that we present below, we
exclude the revenue losses on Internet sales that
displace other remote sales. We found no empirical
evidence indicating what proportion of Internet
sales replace other forms of remote sales. Few
researchers or other experts that we contacted
offered estimates of this proportion, and none of
those were empirically based. Given this high
degree of uncertainty, we use a broad range of
assumptions in our scenarios.
The distinction between Internet and other remote
sales has become less meaningful as more
businesses offer both Internet and other remote
transaction options. Some purchasers may order and
pay for products by mail after obtaining the
necessary information over the Internet; other
purchasers may place orders over the Internet
after obtaining information from a mail-order
catalogue. The tax treatment of the product
purchased does not differ between orders placed by
mail and orders placed over the Internet.
Under Our Scenarios, the Current Loss on All
Remote Sales Is Less Than 5 Percent of Sales Tax
Revenue and the Loss on Internet Sales Is Less
Than 2 Percent
We developed two basic scenarios to illustrate the
uncertainty surrounding the revenue loss on
Internet and other remote sales in 2000. Each
scenario consists of a set of assumptions about
the factors that determine the size of that loss.
We identified a range of plausible assumptions for
each factor, based on available data, estimates,
and expert opinion. In our lower scenario, the
assumptions that we selected for each factor were
those that tended to reduce the revenue loss. In
our higher scenario, we used assumptions that
tended to increase the revenue loss.25 However,
although the results from our higher scenario are
based on a combination of assumptions that tend to
increase the revenue loss, we cannot be certain
that those results represent an upper bound to the
revenue loss. We also examined how the results of
each scenario changed when individual assumptions
were altered. Table 2 presents the results of the
two scenarios.26
The assumptions used in the lower scenario result
in a revenue loss on all remote sales of $1.6
billion. The loss attributable to Internet sales
that did not simply replace other remote sales is
$0.3 billion. In contrast, the assumptions used in
the higher scenario yield a revenue loss of $9.1
billion on all remote sales-less than 5 percent of
state and local general sales tax revenues.27 The
loss attributable to Internet sales is $3.8
billion-less than 2 percent of revenues. The
revenue losses associated with business-to-
business sales vary more widely across scenarios
than do the losses associated with business-to-
consumer sales for several reasons. First, there
is a greater variance across scenarios in the
underlying business-to-business sales estimates.
Second, we used a wider range of assumptions
regarding the proportion of business-to-business
sales that are taxable. Finally, we used a wider
range of assumptions for the business purchaser
compliance rate.
Table 2: State and Local Sales Tax Losses for All
Remote Sales and Internet Sales Only in 2000
Dollars in
billions
Revenue losses Lower scenario Higher scenario
All remote sales
Business-to- $1.5 $4.2
consumer
Business-to- 0.1 4.9
business
Total $1.6 $9.1
Internet sales
Business-to- $0.2 $0.8
consumer
Business-to- a 2.9
business
Total $0.3 $3.8
Note: Columns may not add due to rounding.
aAn amount less than $50 million.
Source: GAO analysis.
Table 3 shows how sensitive our higher scenario
results are to changes in important assumptions.
Table 3: Sensitivity of Higher Scenario Revenue
Losses in 2000 to Changes in Key Assumptions
Dollars in
billions
Revenue Higher Using Using Using
losses scenari lower higher higher
o from sales business displacem
table estimate purchaser ent rate
2 s compliance
rate
All remote
sales
Business-to- $4.2 $2.8 $2.8 No change
consumer
Business-to- 4.9 3.3 0.3 No change
business
Total $9.1 $6.1 $3.1 No change
Internet
sales
Business-to- $0.8 $0.5 $0.3 $0.6
consumer
Business-to- 2.9 2.0 0.2 1.0
business
Total $3.8 $2.5 $0.5 $1.5
Note: Columns may not add due to rounding.
Source: GAO analysis.
As expected, the underlying sales estimates that
we use have a great influence on the revenue loss.
For example, if we kept all of the assumptions for
our higher scenario unchanged, except for
switching to the lower sales estimates, the
revenue loss on all remote sales for 2000 would
fall from $9.1 billion to $6.1 billion. However,
even for a given sales figure, we obtained a
considerable range of revenue estimates by varying
assumptions for other factors. For example, if we
changed the assumed rate of business purchaser
compliance in the higher scenario from 50 percent
to 95 percent and left everything else the same,
the revenue loss on all remote sales would fall to
$3.1 billion.
The assumptions that we made about the proportion
of Internet sales that displace other remote sales
do not affect the revenue loss on all remote
sales; however, they do have a significant effect
on the loss attributed to Internet sales. For
example, if we kept all of the assumptions for our
higher scenario unchanged, except for changing the
displacement proportion from 25 percent to 50
percent for sales to consumers and from 40 percent
to 80 percent for sales to businesses, the revenue
loss on Internet sales would fall from $3.8
billion to $1.5 billion. Appendix I presents
variations of our two basic scenarios that show
the sensitivity of our results to other changes in
assumptions.
Each state's share of the revenue loss is
primarily a function of its share of total sales,
the scope of its sales tax exemptions, its rate of
tax, and the extent to which remote sellers have a
substantial connection with it. State-by-state
revenue losses under our two scenarios are
presented in appendix V. These state-by-state
results are more sensitive than our national
results are to some of our assumptions (e.g., how
aggregate sales are distributed across states).
The Size of Future State and Local Revenue Losses
Is Very Uncertain
We developed lower and higher revenue loss
scenarios for 2003 that are the same as the
scenarios presented above, except for the
underlying sales projections (which are described
in app. I). The rapid change in the Internet
economy makes projections of Internet and total
remote sales for future years considerably more
uncertain than they are for 2000. The rate of
growth and fundamental changes in the patterns of
buying and selling from one year to the next
suggest that historical information is not
particularly useful in making estimates of future
growth. Table 4 presents the results of the two
scenarios for 2003. The assumptions used in the
lower scenario result in a revenue loss on all
remote sales of $2.5 billion-equivalent to about 1
percent of projected general sales tax revenues.
The loss attributable to Internet sales that did
not simply replace other remote sales is $1.0
billion-less than 1 percent of projected general
sales tax revenues. In contrast, the assumptions
used in the higher scenario yield a revenue loss
of $20.4 billion on all remote sales-about 8
percent of projected revenues. The loss
attributable to Internet sales is $12.4
billion-about 5 percent of projected revenues.
Table 4: State and Local Sales Tax Losses for All
Remote Sales and Internet Sales Only in 2003
Dollars in
billions
Revenue losses Lower scenario Higher scenario
All remote sales
Business-to- $2.1 $5.9
consumer
Business-to- 0.4 14.5
business
Total $2.5 $20.4
Internet sales
Business-to- $0.9 $3.7
consumer
Business-to- 0.1 8.7
business
Total $1.0 $12.4
Note: Columns may not add due to rounding.
Source: GAO analysis.
The choice of sales estimates that one uses and
the choice of purchaser compliance rates on sales
to businesses have particularly large effects on
the results (see table 5). If we kept all of the
assumptions for our higher scenario unchanged,
except for switching to the lower sales estimates,
the revenue loss on all remote sales would fall
from $20.4 billion to $13.6 billion. If we changed
the assumed rate of business purchaser compliance
in the higher scenario from 50 percent to 95
percent and left everything else the same, the
revenue loss on all remote sales would fall to
$4.9 billion.
Table 5: Sensitivity of Higher Scenario Revenue
Losses in 2003 to Changes in Key Assumptions
Dollars in
billions
Revenue Higher Using Using Using
losses scenari lower higher higher
o from sales business displace
table estimate purchaser ment
4 s compliance rate
rate
All remote
sales
Business-to- $5.9 $3.9 $3.9 No
consumer change
Business-to- 14.5 9.7 1.0 No
business change
Total $20.4 $13.6 $4.9 No
change
Internet
sales
Business-to- $3.7 $2.0 $2.0 $2.5
consumer
Business-to- 8.7 5.8 0.6 2.9
business
Total $12.4 $7.8 $2.6 $5.4
Note: Columns may not add due to rounding.
Source: GAO analysis.
Concluding Observations
There are three themes that run through this
report. One is that under most of our scenarios,
the tax loss associated with Internet sales is
small for 2000, but under some scenarios, could be
much larger by 2003. Another is that continued
growth in Internet sales is likely to present
major compliance challenges for sales and use tax
administrators-to the extent that some have
questioned the long-term viability of such taxes.
Finally, there is tremendous uncertainty about all
of the major determinants of the tax loss.
When combined, these three themes highlight the
importance of efforts to get better data about the
determinants of the tax loss, such as the Census
Bureau's program to measure Internet sales.
Current economic data are not well suited to
tracking rapidly evolving Internet activity. With
better data, policymakers would be better
positioned to confront the challenges presented by
e-commerce to sales and use tax administration.
One benefit of such data would be more accurate
estimates of sales and use tax losses to state and
local governments. Perhaps more importantly, such
data could provide more of a basis for evaluating
alternative policy choices.
Understanding the limits of data, however, in an
environment as dynamic as the Internet is
important. Innovations in Internet sales-with new
types of goods, services, and transactions-are
rapid and unpredictable. Further, many of the key
decisions by consumers, businesses, and
policymakers that will determine the extent of
Internet tax losses in the near future have not
yet been made and will not be reflected in data
that are necessarily historical. As a consequence,
even with improved data, policymaking regarding
Internet sales will be done in an environment of
significant uncertainty.
As agreed, unless you announce the contents of
this report earlier, we plan no further
distribution until 14 days from the date of this
letter. At that time, we will send copies of this
report to Senator Richard J. Durbin, Ranking
Minority Member of the Subcommittee; the Honorable
Lawrence H. Summers, Secretary of the Treasury;
the Honorable William M. Daley, Secretary of
Commerce; the Honorable Charles O. Rossotti,
Commissioner of Internal Revenue and other
interested congressional parties. We will also
make copies available to others on request.
Please contact me or James A. Wozny at (202) 512-
9110 if you have any questions. Key contributors
to this report are acknowledged in appendix VII.
Sincerely yours,
James R. White
Director, Tax Policy
and Administration Issues
_______________________________
1Based on case law, out-of-state remote sellers
generally meet the nexus standards if they have an
office or place of business, agent, or significant
property in the taxing state.
2A use tax, generally imposed on the purchaser
when a sales tax has not been paid, is imposed on
the privilege of ownership, possession, or use of
a taxable good or service.
3P.L. 105-277, Div. C, Title XI.
4The Commission reported in April 2000. The
Background section of our report contains more
details about the Commission's report.
5Revenue loss is calculated as the amount of tax
liability minus the amount already being paid.
6General revenues include all revenues except the
non-tax revenues generated by government-owned
liquor stores or utilities and insurance trust
fund revenues (contributions to and investment
earnings of public employee retirement and social
insurance systems). In addition to tax revenues,
own-source general revenues include charges for
specific general government services, such as
tuition at state universities, and miscellaneous
general revenues, such as interest earnings and
proceeds from the sale of property. Total general
revenues equal own-source revenues plus transfers
from other levels of government.
7A "remote seller" can be located in the same
state as the purchaser; we use the term "out-of-
state remote seller" when the remote seller is not
located in the same state as the purchaser.
8See appendix II for discussion of National Bellas
Hess Inc. v. Department of Revenue of Illinois,
386 U.S. 753 (1967) (addresses Due Process and
Commerce Clause nexus standards for mail-order
sellers); Quill Corp. v. North Dakota, 504 U.S.
298 (1992) (draws distinction between Due Process
Clause and Commerce Clause requirements);and other
precedent-setting decisions.
9Alaska, Delaware, Montana, New Hampshire, and
Oregon do not have general sales tax programs.
Delaware does, however, impose a gross receipts
tax.
10Local jurisdictions in some states, such as
Alabama, Colorado, and Alaska, administer local
sales tax programs.
11Appendix I identifies specific assumptions and
data sources used in these past studies.
12U.S. Advisory Commission on Intergovernmental
Relations, Taxation of Interstate Mail Order
Sales: 1994 Revenue Estimates (1994).
13Robert J. Cline and Thomas S. Neubig, The Sky Is
Not Falling: Why State and Local Revenues Were Not
Significantly Impacted by the Internet in 1998,
Ernst & Young, Economics Consulting and
Quantitative Analysis (June 18, 1999).
14Austan Goolsbee and Jonathan Zittrain,
"Evaluating the Costs and Benefits of Taxing
Internet Commerce," National Tax Journal, 52(3),
Sept. 1999, pp. 413-28.
15Donald Bruce and William F. Fox, "E-Commerce in
the Context of Declining State Sales Tax Bases,"
mimeo, University of Tennessee Center for Business
and Economic Research (Apr. 2000).
16James L. McQuivey, with Gillian DeMoulin, States
Lose Half A Billion In Taxes To Web Retail, A
Technographics Brief, Forrester (Cambridge, MA,
Feb. 24, 2000).
17P.L. 105-277, Div C, Title XI, Oct. 21, 1998.
18In the case of Alaska, where local governments
collect general sales taxes but the state
government does not, we assumed that the state's
share of the nationwide local government revenue
loss was proportionate to its share of nationwide
local government sales tax collections.
19We selected the states to contact on the basis of
referrals from national organizations, including
the Multistate Tax Commission and the Federation
of Tax Administrators, which indicated that these
states were conducting studies on the issue of
remote sales or had cutting-edge compliance
programs.
20We asked officials from state revenue departments
and national associations, such as the Multistate
Tax Commission, National Governors' Association,
and National Retail Federation, to identify
specific examples of different tax requirements
for in-store and out-of-state remote sales. None
identified any significant different sales,
excise, or income tax requirements, but several
referred to Connecticut's tax on the on-line sale
of a newspaper that purchasers could buy untaxed
at the newsstand. A Connecticut official advised
that the state taxes paid-for digital services
that include newspapers sold on-line. Connecticut
expects to phase out its on-line newspaper tax by
2002.
21Available evidence suggests that compliance among
businesses is also highest for in-store sales and
lowest for remote sales without nexus. However,
the rate of business purchaser compliance for
remote sales without nexus is believed to be
considerably higher than consumer purchaser
compliance.
22The Census general sales tax figures for
Washington State and Indiana include,
respectively, $1,854 million and $548 million from
gross receipt taxes that are closer to business
taxes than they are to general sales taxes. If
those revenues are disregarded, Washington State
is not one of the top three states in terms of
reliance on general sales taxes. There are
additional reasons why the Census data for general
sales tax revenues are not strictly comparable
across states (see Due and Mikesell, 1994). For
example, certain products and services in some
states are not subject to the general sales tax;
however, they are subject to equivalent special
sales taxes. Those special taxes are not included
in the Census general sales tax figures, but they
are covered in our revenue loss computations.
23The success that a state has in collecting the
tax due on a sale also can vary by type of
purchaser. Business purchasers are more likely
than individual purchasers to comply with their
use tax obligations because they face a much
higher probability of being audited by state tax
agencies.
24Some authors noted that use tax compliance was
low even before the advent of e-commerce.
25The assumptions that we used for each scenario
are provided in appendix I.
26Appendix V provides results for each state.
27These percentages were computed using the
assumption that the annual rate of growth in
collections from 1999 to 2003 would be the same as
the annual rate of growth from 1990 through 1999.
Appendix I
Methodology for the Revenue Loss Scenarios
Page 36GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
This appendix provides further details on the
methodology we used to compute the various revenue
loss scenarios presented in the letter in tables 2
through 5. It also provides additional analyses of
the sensitivity of our results to changes in
specific assumptions.
The Volume and Composition of Internet and Other
Remote Sales
To obtain sales estimates, we reviewed academic,
government, and private-sector studies, including
those published on the Internet. We also contacted
the authors and other experts in this field to
identify other potential sources of sales
estimates. There were many estimates available for
business-to-consumer Internet sales but only a
limited number for business-to-business Internet
sales and all remote sales. We were not able to
assess the accuracy of any of the available
estimates. However, given the difficulty of making
such estimates, particularly for future years, we
believe that they all have substantial margins of
error. In our revenue loss scenarios, we try to
represent the broad range of estimates for
business-to-consumer Internet sales that have been
made for 2000 and 2003. For other categories of
sales, where we had to rely on a single source of
estimates, we applied a margin of error of 20
percent to acknowledge the fact that there is
considerable uncertainty around the estimates.
Business-to-Consumer Internet Sales
For business-to-consumer Internet sales, the range
of estimates that we identified is similar to the
list reported in eMarketer's "The eCommerce: B2C
Report," April 2000.1 We chose our range of
estimates from a subset of sources (all private
research firms) that used the same definition of
Internet sales-transactions placed and paid for
over the Internet. From among that group,
Forrester Research, Inc., represented the higher
end of the range, with sales estimates of $39
billion for 2000 and $143 billion for 2003.2
Jupiter Communications represented the lower end
of the range, with sales estimates of $23 billion
for 2000 and $78 billion for 2003. Both Forrester
and Jupiter base their estimates on surveys of
consumers and surveys and interviews of
businesses. Bruce and Fox, and Goolsbee and
Zittrain both used business-to-consumer sales
estimates by Forrester in their studies. Cline and
Neubig used a sales estimate that they said was on
the high end of available estimates at the time of
their analysis. When making its own state revenue
loss estimate, Forrester relied on two surveys of
consumers, one of which asked buyers how often
they paid taxes when shopping on-line.3
Business-to-Business Internet Sales
Forrester Research's estimates of business-to-
business Internet sales are the only ones that we
found that were made within the last year. 4 We
used Forrester's estimates with a 20-percent
margin of error for our scenarios, giving us lower
end estimates of $325 billion for 2000 and $1,459
billion for 2003, and higher bound estimates of
$487 billion for 2000 and $2,188 billion for 2003.
Forrester made separate estimates of business-to-
business sales for service industries, which we
also incorporated into our scenarios with 20-
percent error bounds. The lower sales estimates
for these services are $35 billion for 2000 and
$176 billion for 2003. The higher estimates are
$52 billion for 2000 and $264 billion for 2003. In
their study, Bruce and Fox used an earlier set of
Forrester business-to-business sales estimates.
Total Business-to-Consumer Remote Sales
We decided that the best available estimates of
total remote business-to-consumer sales for our
analysis were those produced by Marketing
Logistics Inc., a research firm that has produced
such estimates for 18 years.5 Its estimates are
disaggregated by detailed product and service
categories.
Marketing Logistics measures orders that are
placed by mail, phone, or electronically and that
do not require the purchaser to visit the seller's
premises, or vice versa. Its estimates are based
on "micro" and "macro" techniques. The micro
technique involves identifying the mail order
businesses in an industry and estimating the
remote sales for each business. Marketing
Logistics has a proprietary database that contains
information on over 10,000 companies. The macro
technique uses estimates of aggregate remote sales
for certain sales segments. The Advisory
Commission on Intergovernmental Relations (ACIR)
used the Marketing Logistics data as the basis for
making its past estimates of revenue losses on
remote sales.
After applying a 20-percent margin of error to the
Marketing Logistics estimates, the higher and
lower estimates that we used for business-to-
consumer sales in 2000 were $278 billion and $186
billion. The higher and lower estimates for 2003
were $391 and $261 billion.
Total Business-to-Business Remote Sales
Marketing Logistics' estimates of total business-
to-consumer remote sales were significantly higher
than the highest estimates that we used for
business-to-consumer Internet sales. However, its
estimates of total business-to-business remote
sales were much lower than the Forrester estimates
of business-to-business Internet sales (even after
we reduced the latter by 20 percent). We decided
to use the adjusted Forrester estimates as the
estimates for both Internet business-to-business
sales and total business-to-business remote sales
because the Forrester estimates were considerably
more recent than the Marketing Logistics estimates
and because the e-commerce environment is changing
so rapidly. This decision prevents a logical
inconsistency between the Internet and the total
remote sales estimates that we use. However, the
fact that the inconsistency exists between the
original estimates is further evidence that these
estimates should be used with caution.
Table I.1 summarizes the choices we made among
sources of estimates for each major category of
sales.
Table I.1: Choices of Sales Estimates Used in the
Scenarios
Sales Business-to-consumer Business-to-
business
Internet
High end Forrester Forrester x 120
percent
Low end Jupiter Forrester x 80
percent
All remote
High end Marketing Logistics x Forrestera x 120
120 percent percent
Low end Marketing Logistics x Forrestera x 80
80 percent percent
aForrester estimates of business-to-business
Internet sales.
Source: GAO.
The Taxability of Remote Sales
To estimate the amount of tax due on remote sales,
we apportioned a share of total remote and
Internet sales to each state and then applied each
state's tax exemptions and rates to those sales.
We allocated sales across states by assuming that
each state's share of sales to individuals is
proportionate to the state's share of total
disposable income and that each state's share of
sales to businesses is proportionate to the
state's share of total state product.6 We made
this allocation for each product and service
category. We then determined which categories of
products and services are taxed by each state. Our
main sources for state exemptions were CCH's 1999
US Master Sales and Use Tax Guide for goods, the
Federation of Tax Administrators' 1996 Update for
the Sales Taxation of Services for services, and
the Research Institute of America's 1999 All
States Tax Handbook. We supplemented these sources
with information obtained from individual state
revenue departments.
We made additional adjustments to reflect the fact
that some sales are exempted on the basis of the
type of purchaser or the type of use. These
purchaser and use exemptions are important for
estimating what proportion of business-to-business
remote sales are taxable. Our sources of sales
estimates did not disaggregate them by type of
purchaser or types of use. In order to estimate
the percentage of business-to-business sales that
were exempt, we used the input-output tables
compiled by the U.S. Department of Commerce.7 The
Use Table shows the interindustry transactions of
the U.S. economy for 1996 and provides detailed
information on the composition of inputs and the
distribution of outputs of all major U.S.
industries. We used parts of the "Intermediate
Use" column of this table to estimate the share of
inputs for key industries that are exempted as raw
materials, or inputs that are exempted because
they are incorporated into the final product. We
used parts of the "Gross Private Fixed Investment"
column to estimate the share of industry inputs
that are purchases of machinery and other
equipment that are also exempted from sales and
use taxes by many states. The input-output tables
also provide detailed information on the
percentage of each industry's output that are
exported, sold to federal and state governments,
and sold for consumption. We used these data to
estimate the share of each major industry's output
that should be included in calculations of
business-to-business sales that are relevant for
sales and use tax purposes.
On the basis of our analysis of the input-output
data, we used the following ranges of assumptions
in our scenarios for each product category in the
Forrester business-to-business Internet sales
estimates. The numbers in parentheses indicate the
percent of sales assumed to be not taxed.
Purchases by tax-exempt entities, such as
governments and charitable organizations, are
reflected in the percentages.
� Computing and electronics (35-55 percent not
taxed)
� Motor vehicles (70-80 percent not taxed)
� Paper and office products (50-65 percent not
taxed)
� Industrial equipment (65-85 percent not taxed
in states that exempt machinery used for
manufacturing; 50-70 percent not taxed in the
remaining states)
� Petrochemicals, shipping and warehousing,
pharmaceutical and medical products, construction,
and heavy industries (90-100 percent not taxed)
� Food and agriculture and consumer goods (95-
100 percent not taxed)
� Aerospace and defense (85-100 percent not
taxed)
� Utilities (100 percent not taxed in the
states that completely exempt utilities; 60-90
percent not taxed in the states that exempt only
utilities purchased by manufacturers; 40-60
percent not taxed in states that do not exempt
utilities at all).
In addition we assumed that between zero and 6
percent of business-to-consumer sales were to tax-
exempt entities. These assumptions are based on
our analysis of the Commerce Department data and
opinions obtained from state revenue officials.
The Extent to Which Remote Sellers Already Collect
Tax
To estimate seller collection rates for selected
categories of Internet and other remote sales, we
followed an approach similar to that used in
earlier studies by ACIR and the Pennsylvania
Department of Revenue. We made separate estimates
for all remote sales and for Internet sellers
because a somewhat different population of firms
dominates in each case. To make our estimate for
all remote sales, we obtained information from 96
of the largest remote sellers from Marketing
Logistics' 1998 list of leading sellers to
determine the states in which they collect sales
taxes. We grouped the companies that responded by
product category and distributed their sales
across states. We were then able to estimate what
percent of sales in each product category in each
state was made by taxpaying sellers. We used the
ratios of the respondents' sales to Marketing
Logistics' estimates of total remote sales in each
product category to determine how reliable our
estimates were. For the five categories in which
the respondents' sales represented at least 40
percent of total sales, we used our specific
collection rate estimate in our model. We grouped
the remaining categories into a separate "other
consumer goods" category and estimated one rate
for that category. In order to reflect the
significant margin of error around our estimates,
in our lower scenario we used collection rates
that were 25 percent below our estimates; in the
higher scenario we assumed rates that were 25
percent above our estimates. Table I.2 shows the
nexus assumptions we used for selected categories
of all remote sales.
Table I.2: Seller Collection Rate Estimates for
Selected Categories of All Business-to-Consumer
Remote Sales
Category Seller collection rate
for
Higher Lower
scenario scenario
Apparel 29 % 48 %
Auto service clubs 73 100
Department stores and broad 59 98
range catalogues
Computer hardware 14 24
Computer software 11 19
Other consumer goods 35 58
Source: GAO estimates.
For other categories of sales, where we could not
obtain sufficient data from companies, but where
discussions with tax officials led us to believe
that collection rates would be high, we assigned
collection rates. Specifically, for the following
categories, we assumed a 100-percent seller
collection rate in our lower scenario and a 75-
percent collection rate in our higher scenario:
cable television, direct broadcasting satellite
services, and insurance.
We followed a similar approach for estimating
seller collection rates for business-to-consumer
Internet sales. We obtained information from 86 of
the companies on Stores.org's 1999 list of the top
100 Internet retailers. We used Forrester's
estimates of business-to-consumer Internet sales
in 1999 to estimate the product category shares
represented by our respondents. Table I.3 shows
the seller collection rate assumptions for our
scenarios that were based on our empirical
estimates.
Table I.3: Seller Collection Rate Estimates for
Selected Categories of Business-to-Consumer
Internet Sales
Category Seller collection rate
for
Higher Lower
scenario scenario
Apparel 34 % 57 %
Books/videos/music 13 22
Computer software 28 46
Computer hardware 14 23
Other consumer goods 22 36
Source: GAO estimates.
We could not find appropriate data to estimate
collection rates for the business-to-business
Internet sales categories used by Forrester. We
assumed the following ranges of seller collection
rates for each of the Forrester industrial
categories (after examining the more detailed
subcategories contained in each):
� Computing and electronics, consumer goods,
food and agriculture, industrial equipment, paper
and office products, and pharmaceuticals (25-50
percent);
� Aerospace and defense, construction, and
heavy industries (70-90 percent);
� Petrochemicals, shipping and warehousing, and
utilities (75-95 percent);
� Motor vehicles (85-100 percent).
Other analysts who have estimated the revenue
losses associated with Internet and other remote
sales have used a variety of assumptions regarding
the proportion of tax already being paid. However,
they have not always specified what portion
sellers pay and what portion purchasers pay. Bruce
and Fox assumed that 100 percent of the tax due on
the Internet sales of automobiles to businesses
and consumers are being collected. They assume
that 50 percent of the tax due on all other
Internet sales to businesses and 10 percent of the
tax due on all other Internet sales to individual
consumers are being collected either from the
sellers or from the purchasers. Cline and Neubig
assumed that no significant amount of tax is going
uncollected on Internet sales to businesses, while
nearly 11 percent of the tax due on Internet sales
to consumers is being collected from either
sellers or purchasers. Goolsbee and Zittrain also
assumed that no significant losses result from
business-to-business Internet sales, but they
assumed that no tax is collected on sales to
consumers. None of these authors cited empirical
evidence for these assumptions, though authors of
the first two studies referred to discussions with
private sector tax experts.
The Extent to Which Purchasers Already Pay Tax
There is a wide range of opinion regarding the
compliance rate of business purchasers. One state
official we contacted believes the rate is close
to 100 percent; another believes it to be around
65 percent. The range of assumptions used in
previous studies is similarly wide. As noted
above, Cline and Neubig, and Goolsbee and Zittrain
have assumed that almost all taxes owed on
business-to-business sales are being paid. In
contrast, Bruce and Fox assume that only 50
percent is being paid on most purchases, while 100
percent is being paid on purchases of motor
vehicles. Neither the Pennsylvania Department of
Revenue nor the ACIR estimated the rate of
compliance on the part of purchasers, although
both acknowledged that there would be some
compliance on the part of business purchasers.
In contrast to the wide range of opinion that
exists with respect to the compliance of business
purchasers, there was a wide consensus among
previous studies and the state officials, who
provided us estimates that use tax compliance by
individual purchasers was extremely low. However,
there was also a wide consensus that compliance
with the use tax on motor vehicles was close to
100 percent because the taxes must be paid before
those vehicles can be registered. 8
For our scenarios, we used assumptions for the
business purchaser compliance rate that range from
50 to 95 percent on all products, with the
previously stated exception for motor vehicles,
where the range is from 85 to 100 percent. We used
assumptions for the consumer purchaser compliance
rate that range from zero to 5 percent. The
business-to-consumer estimates that we used for
Internet and all remote sales did not show motor
vehicle sales separately.
Auction Sales
The Forrester Internet sales estimates that we use
in our scenarios include sales that businesses
make through auction sites, but they exclude
person-to-person auction sales. We exclude those
sales under the assumption that state tax
authorities would not have any more success
collecting from individuals who sell over the
Internet than they do from individuals who
purchase over the Internet. The latter already
have an responsibility to pay any tax that is owed
on auction purchases. Boston Consulting Group
estimates that person-to-person auction sales of
collectibles will total $6 billion in 2000. If all
of these sales were taxable, the total state and
local sales tax owed on them would be about $0.4
billion. However, almost all states exempt
"occasional" sales (other than motor vehicles,
vessels, and aircraft) made by persons not engaged
in business.
Behavioral Response
Our estimates are based on the assumption that the
same volume of Internet and remote sales would
have occurred even if taxes were collected on all
of those sales. This assumption ignores consumer's
response to the reduction in taxes paid when
purchases shift to the Internet, and leads to an
upward bias in our estimates of revenue loss. In
addition, there are several ways in which the
existence of and growth in the Internet has
changed the behavior of consumers and businesses.
First, given the added convenience and lower
prices associated with purchasing certain goods
over the Internet, it is likely that the volume of
those sales has increased over what would have
occurred in the absence of Internet and other
remote outlets. In addition, the rapid growth of
the Internet has also changed the geographic
pattern if not the overall level of retail
activity in the United States. To the extent that
these effects have increased retail sales over
what would have occurred in the absence of the
Internet, this would lead to an upward bias in our
results. We were unable to find sufficient
evidence to allow us to adjust for these different
changes.
Sensitivity Analysis
Tables I.4 and I.5 show the revenue losses for
2000 and 2003 that we calculated using various
combinations of the assumptions and sales
estimates described above. They illustrate the
effects of changing individual assumptions while
holding everything else constant. For example, the
third cell of the first row shows the $9.1 billion
loss that we obtained by combining a higher end
estimate for total remote sales and assumptions
that all tended to increase the revenue loss. The
third cell of the second row shows that if we kept
everything the same as in the prior scenario, with
the exception of using our lower rather than
higher set of seller collection rates, the revenue
loss would fall to $5.6 billion.
Table I.4: State and Local Sales Tax Losses for
All Remote Sales and Internet Sales Alone in 2000
Dollars in billions
Scenario Lower sales Higher sales
All Intern All Inter
remote et remote net
All higher scenario $6.1 $2.5 $9.1 $3.8
assumptions
All higher scenario
assumptions except for
Seller collection 3.7 1.7 5.6 2.5
rates
Purchaser compliance 3.0 0.7 4.5 1.1
Purchaser exemptions 4.8 1.8 7.2 2.8
Displacement rates No 1.0 No 1.5
change change
All lower scenario 1.6 0.3 2.4 0.5
assumptions
All lower scenario
assumptions except for
Seller collection 2.7 0.3 4.1 0.6
rates
Purchaser compliance 2.9 0.5 4.4 0.8
Purchaser exemptions 1.8 0.3 2.7 0.5
Displacement rates No 0.4 No 0.7
change change
Source: GAO estimates.
Table I.5: State and Local Sales Tax Losses for
All Remote Sales and Internet Sales Alone in 2003
Dollars in billions
Scenario Lower sales Higher sales
All Intern All Inter
remote et remote net
All higher scenario $13.6 $7.8 $20.4 $12.4
assumptions
All higher scenario
assumptions except for
Seller collection 8.0 5.0 12.0 8.0
rates
Purchaser compliance 4.7 2.5 7.1 4.4
Purchaser exemptions 9.8 5.6 14.8 9.0
Displacement rates No 3.3 No 5.4
change change
All lower scenario 2.5 1.0 3.6 1.9
assumptions
All lower scenario
assumptions except for
Seller collection 4.1 1.3 6.2 2.4
rates
Purchaser compliance 5.8 1.7 8.8 2.9
Purchaser exemptions 2.8 1.1 4.1 2.0
Displacement rates No 1.6 No 2.9
change change
Source: GAO estimates.
_______________________________
1"The eCommerce: B2C Report," eMarketer (New York,
Apr. 2000), p. 30.
2The business-to-consumer sales estimate for 2000
contained in the April 2000 shop.org/Boston
Consulting Group study was higher than the
Forrester estimate, but after removing non-U.S.
sales, financial services, and nonsales revenues
contained in the Boston Consulting Group estimate,
the Forrester estimate was higher.
3Forrester's general approach differed from those
of the other prior studies (which were broadly
similar to our approach). Among the significant
differences are that Forrester did not attempt to
distinguish between taxable and nontaxable sales
(except for travel) and did not make an adjustment
for Internet sales that displace other remote
sales.
4Forrester's primary report on business-to-
business Internet sales, "eMarketplaces Boost B2B
Trade," is dates February 2000; its separate
report on business-to-business services, "Business
Services On the Net," is dated January 1999.
Forrester based both of these studies on
interviews with officials from large firms. For
the first study, it also relied on discussions
with industry experts, strategy consultants,
technology vendors, and eMarketplaces.
5Direct Marketing Association estimates a
category of sale-"direct order sales"-that is
roughly equivalent to total remote sales; however,
we believe the estimating techniques and the
amount of detail were preferable in the Marketing
Logistics data.
6The distribution across states of personal
income, disposable income, and state product is so
highly correlated that our aggregate results would
be the same regardless of which of these factors
we used to apportion sales.
7Table 2, Survey of Current Business (Jan. 2000),
pp. 56-65.
8The reason why we assume less than 100-percent
compliance in our higher scenario is that the
motor vehicles category in the Forrester data
includes some sales of automotive parts.
Appendix II
Constitutional Restrictions on State Authority to
Impose Sales and Use Taxes
Page 38GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
Commerce and Due Process Clauses
The authority of the states to impose sales and
use taxes is limited by the U. S. Constitution.
The Commerce Clause of Article I and the Due
Process Clause of the 14th Amendment are the
principal constitutional challenges to these
taxes.1 These two provisions directly impact the
ability of the states to tax nonresidents and
interstate commerce. Both provisions require a
sufficient connection between the state and the
taxpayer it seeks to tax or the seller on which
the state seeks to impose a responsibility to
collect a use tax in order for the tax to be
upheld.
Under the Commerce Clause, Congress has sole
authority to regulate commerce with foreign
nations, among the states, and with the Indian
tribes. Accordingly, the Commerce Clause prevents
the states from interfering with or unduly
burdening interstate commerce through the use of
its taxing authority. The Supreme Court's
interpretation of this restriction (the Complete
Auto test) provides that a state tax does not
unduly burden interstate commerce if it is applied
to an activity with a substantial connection or
"nexus" with the taxing state, is fairly
apportioned, does not discriminate against
interstate commerce, and is fairly related to
services provided by the state.2
Under the Due Process Clause, states may not
deprive any person of life, liberty, or property
without due process of law. This restriction
limits the territorial reach of the states' taxing
authority to persons, property, and business
transactions within their jurisdictions. The
Supreme Court's interpretation of this restriction
requires some definite link, some minimum
connection or "nexus," between a state and the
person, property, or transaction it seeks to tax.3
Considerable case law has evolved addressing the
differing constitutional requirements. Two
Supreme Court cases are particularly relevant to
the discussion of the Commerce Clause and Due
Process Clause challenges to state imposition of
use tax collection the responsibility on out-of-
state sellers. These two cases, National Bellas
Hess v. Department of Revenue of Illinois , 386
U.S. 753 (1967), and, more recently, Quill Corp.
v. North Dakota, 504 U.S. 298 (1992), address the
"nexus" requirements for taxation of interstate
transactions.
National Bellas Hess Addresses Nexus Standards for
Mail-order Sellers
The National Bellas Hess company was a mail-order
house with its principal place of business in
Missouri. It had neither outlets (nor any tangible
property, real or personal) in Illinois nor sales
representatives physically located there to sell
or take orders. Twice-a-year catalogs were mailed
to the company's customers throughout the United
States, including Illinois. Customers mailed
orders for the goods to the National Bellas Hess
plant in Missouri. The ordered goods were then
sent to the customers either by mail or common
carrier.
The State of Illinois obtained a judgment from its
highest court requiring National Bellas Hess to
collect and pay to the state a use tax imposed
upon its consumers who purchased goods for use
within Illinois. National Bellas Hess argued that
imposition of the responsibility to collect a use
tax collection violated the Due Process Clause and
created an unconstitutional burden upon interstate
commerce.
The Supreme Court reversed the ruling of Illinois'
highest court, noting, first, that National Bellas
Hess' two constitutional challenges were closely
related. According to the Court, the test for
whether a particular state tax invades the
exclusive authority of Congress to regulate
commerce among the states and the test for a
state's compliance with the requirements of due
process in this area are similar. The Court
pointed to its previous holding that state
taxation falling on interstate commerce can only
be justified to bear a fair share of the cost of
the local government whose protection it enjoys.
In determining whether a state tax falls within
the confines of the Due Process Clause, the Court
noted its previous holding that the controlling
question is whether the state has given anything
for which it can ask a return. According to the
Court, the same principles had been held
applicable in determining the power of a state to
impose the burdens of collecting use taxes upon
interstate sales. There, too, the Court noted, the
Constitution requires some definite link, some
minimum connection, between a state and the
person, property, or transaction it seeks to tax.4
The Court then noted that in applying these
principles it had upheld the power of a state to
impose liability upon an out-of-state seller to
collect a local use tax in many circumstances, but
it had never upheld the power to impose this duty
upon a seller whose only connection with customers
in the state was by common carrier or the U.S.
mail. The Court refused to repudiate here the
distinction it had previously drawn between mail
order sellers with retail outlets, solicitors, or
property within a state and those sellers who do
no more than communicate with customers in the
state by mail or common carrier. Accordingly, the
Court concluded that imposition on National Bellas
Hess of the responsibility for use tax collecting
a use tax, in fact, unconstitutional on both
grounds.
Quill Draws Distinction Between Due Process Clause
and Commerce Clause Requirements
In Quill v. North Dakota, the Court reviewed its
earlier decision in National Bellas Hess. The
Court used this opportunity to draw a clearer
distinction between the Due Process Clause and
Commerce Clause nexus requirements.
The Quill Corporation was a mail order house with
offices and warehouse in Illinois, California, and
Georgia. It had neither outlets nor tangible
property in North Dakota, nor did any of its
employees work or reside there. Quill sold office
equipment and supplies through catalogs and
flyers, advertisements in national periodicals,
and telephone calls. Its annual national sales
exceeded $200 million of which almost $1 million
was made from about 3,000 customers in North
Dakota. Quill delivered all of its merchandise to
its North Dakota customers by mail or by common
carrier from its out-of-state locations.
Quill took the position that North Dakota did not
have the power to compel it to collect a use tax
from its North Dakota customers. A North Dakota
trial court agreed with Quill finding the case
indistinguishable from the Supreme Court's
decision in National Bellas Hess.
North Dakota's highest court reversed the trial
court, concluding that wholesale changes in both
the economy and the law made it inappropriate to
follow the National Bellas Hess decision. The
principal economic change noted by the court was
the remarkable growth of the mail-order business
from a relatively inconsequential market in 1967
to a "goliath" with annual sales that reached
$183.3 billion in 1989. Equally important in the
court's view were changes it perceived in the
legal landscape. The court maintained that the
Supreme Court's subsequent four-part Commerce
Clause analysis (the Complete Auto test) indicated
that the Commerce Clause no longer mandated the
sort of physical presence nexus suggested in
National Bellas Hess. The North Dakota court
further concluded that the Due Process requirement
of a minimum connection to establish nexus was no
longer a separate requirement but was encompassed
within the Complete Auto test. According to the
court, the relevant inquiry was whether the state
had provided some protection, opportunities, or
benefit from which it could expect a return. With
regard to the case at hand, the court emphasized
that North Dakota had created an economic climate
that fostered demand for Quill's products,
maintained a legal infrastructure that protected
that market, and disposed of 24 tons of catalogs
and flyers mailed by Quill each year into the
state.
The U.S. Supreme Court reversed the ruling of
North Dakota's highest court. The Court agreed
with the North Dakota court's conclusion that the
Due Process Clause did not bar enforcement of that
state's use tax against Quill. The Court
concluded, however, that the state's enforcement
of the use tax against Quill placed an
unconstitutional burden on interstate commerce.
The Court noted that although it had not always
been precise in distinguishing between the two,
the Due Process Clause and Commerce Clause reflect
different constitutional concerns and are
analytically distinct.
The Supreme Court agreed with the North Dakota
court that nexus is not synonymous with physical
presence for due process purposes and overruled
its previous holdings to that effect. The Court
noted that its due process jurisprudence had
evolved substantially in the 25 years since
National Bellas Hess and that the relevant inquiry
was whether a defendant had minimum contacts with
a jurisdiction such that maintenance of the suit
did not offend traditional notions of fair play
and substantial justice. The Court concluded that
Quill's widespread and continuous solicitation in
North Dakota made the magnitude of its contacts
more than sufficient for due process purposes.
In contrast, the Court upheld its previous holding
in National Bellas Hess to the extent that it
required physical presence in the Commerce Clause
context. The Court first concluded that its
decision in National Bellas Hess is not
inconsistent with Complete Auto and other recent
cases. The Court noted that under Complete Auto's
four-part test, a tax will be sustained against a
Commerce Clause challenge so long as the tax is
applied to an activity with a substantial nexus
with the taxing state, is fairly apportioned, does
not discriminate against interstate commerce, and
is fairly related to the services provided by the
state. According to the Court, National Bellas
Hess concerns the first of these tests and stands
for the proposition that a vendor whose only
contacts with the taxing state are by mail or
common carrier lacks the "substantial nexus"
required by the Commerce Clause. Using this bright-
line, physical presence, rule the Court then
concluded that the imposition of the
responsibility to collect the use on Quill placed
an unconstitutional burden on interstate commerce.
After concluding its decision on the case, the
Supreme Court noted in Quill that Congress may not
only be better qualified to resolve the underlying
issue in the case, but also is the one with the
ultimate power to do so. The Court stated that no
matter how it evaluated the burdens that use taxes
impose on interstate commerce, Congress remains
free to disagree. The Court further noted that in
recent years, Congress had in fact considered
legislation that would legislatively overrule the
National Bellas Hess decision. The Court surmised
that Congress' decision not to take action in that
direction may have been dictated by its holding in
National Bellas Hess that the Due Process Clause
prohibits states from imposing such use tax
collection responsibilities.5 The Court noted that
since the Quill decision overruled that aspect of
National Bellas Hess, Congress, with the sole
authority to regulate commerce among the states,
could freely decide whether, when, and to what
extent the states could burden interstate mail-
order concerns with a duty to collect use taxes.
National Geographic Holds That Nexus Need Not
Relate to Taxed Activity
The Supreme Court, in ruling on the National
Geographic Society case, held that the activity or
physical presence that established a company's
nexus did not have to be related to the taxed
activity.6 National Geographic Society's mail-
order office that made merchandise sales to
customers in California was separate from the
Society's magazine sales and advertising office
that maintained offices in the state. The Court
held that the maintenance of the two magazine
sales offices in California with advertising copy
in the range of $1 million annually adequately
established a relationship of nexus between the
Society and the State of California. This
connection was sufficient for California to
require National Geographic to collect the
California use tax. In so holding, the Court
rejected the Society's argument that there must be
a relationship between the taxed activity and the
seller's activity within the state.
State Court Holds Deminimis Contact Insufficient
to Establish Nexus
While National Geographic held that the activity
that established the company's nexus did not have
to relate to the taxed activity, a state court has
ruled on circumstances that do not constitute
sufficient nexus. For example, the Connecticut
Supreme Court ruled that insignificant property in
a state does not necessarily establish nexus.
Cally Curtis, a California firm, rented film to
customers in Connecticut for a 3-day preview
period before purchase. The court ruled that the
presence of film for the preview period was de
minimis contact and insufficient to support a
nexus relationship between Cally Curtis and
Connecticut.7 The U.S. Supreme Court declined
review of this case.8
Dual Entity Arrangements
Several cases have examined the use of dual entity
arrangements and whether nexus can be imputed to a
vendor that does not appear to have sufficient
nexus to support a state sales and use tax
collection responsibility because of its
affiliation, through a parent-subsidiary or
brother-sister relationship, with another vendor
that does have nexus with the state. The issue has
generally turned on whether the two affiliated
companies are separate and distinct entities and
whether the affiliated company that has sufficient
nexus with the state has acted as an agent for the
company that does not have nexus. Two case
examples follow.
In SFA Folio Collections, Inc. v. Bannon, 217
Conn. 220 (1991), Saks and Company, a New York
Corporation, owned both Folio, a New York
Corporation whose mail-order business sold to
Connecticut customers but had no physical presence
in that state, and Saks-Stamford, a separate
corporation operating a retail store in
Connecticut. The Connecticut Supreme Court
rejected the argument of Connecticut's Revenue
Commission that because these separate entities
were linked by their common parent, Saks and
Company, their separate existence should be
disregarded and that they should be treated as one
enterprise for the purpose of establishing nexus.
The Connecticut court noted that the
commissioner's argument demonstrated a
misunderstanding of a fundamental principle
underlying our system of taxation, which is that
taxpayers may arrange their affairs to minimize
their tax liabilities. According to the court,
this included careful planning of both
transactions and corporate structure. The Supreme
Court declined review of this case.9
Similarly, in Bloomingdale's v. Department of
Revenue, 527 Pa. 347 (1991), the Pennsylvania
Court found that there was not sufficient nexus
between an out-of-state mail-order company,
Bloomingdale's By Mail, which did mail-order
business in the state but had no physical presence
there, even though its parent company,
Bloomingdale's, did own and operate retail stores
in the state. In that case, the Department of
Revenue argued that Bloomingdale's By Mail's
separate corporate existence from Bloomingdale's
department stores was a mere legal formality. The
court pointed to previous court holdings of a
parent/subsidiary relationship with nothing more
would not justify disregarding the separate
corporate identity. According to the court, the
issue turned on whether the Bloomingdale's
department stores had acted as an agent or
representative for Bloomingdale's By Mail.10 The
Pennsylvania court concluded though that the
revenue department had not established the
existence of an agency relationship between
Bloomingdale's department stores and
Bloomingdale's By Mail. The Supreme Court declined
review of this case.11
_______________________________
1Commerce Clause, Sec. 8, Cl. 3, Art. I and Due
Process Clause, Sec. 1, amend. XIV. Additional
constitutional restraints on state taxation
include the Import-Export Clause that prevents
states from imposing duties on imports or exports
without congressional consent; the Privileges and
Immunities Clause that prevents states from
imposing greater burdens on nonresidents than on
residents; the Supremacy Clause that prevents
state taxing statutes from contravening federal
laws, regulations, or treaties; the First
Amendment that prevents states from discriminating
against free speech or freedom of religion; and
the Equal Protection Clause of the 14th Amendment
that prevents states from making unfair
classifications.
2Complete Auto Transit Inc. v. Brady, 430 U.S. 274
(1977).
3Miller Brothers Co. v. Maryland, 347 U.S. 347
(1954).
4There was no question of the connection or link
between the State and the person it sought to tax,
i.e. Illinois residents who used the goods
purchased from National Bellas Hess. Although
National Bellas Hess was not the person being
directly taxed (but rather it was asked to collect
the tax from the user), it was, however, made
directly liable for the payment of the tax whether
collected or not. Ill. Rev. Stat. C. 120, sec.
439.8 (1965).
5While Congress has plenary power to regulate
commerce among the states and thus may authorize
state actions that burden interstate commerce, it
does not similarly have the power to authorize
violations of the Due Process Clause.
6National Geographic Society v. State Board of
Equalization, 430 U.S. 551 (1977).
7Cally Curtis Co. v. Groppo, 214 Conn 292 (1990).
8Writ of certiorari denied, Commissioner of
Revenue Services v. Cally Curtis Co., 498 U.S. 824
(1990).
9Writ of certiorari denied, Commissioner of
Revenue Services v.SFA Folio Collections, 501 U.S.
1223 (1991).
10In response to the Department of Revenue's
argument that catalog purchasers had been allowed
to return merchandise directly to the local
department store, the court found that such
returns appeared to be "an aberration from normal
practice," so it did not reach a conclusion as to
whether nexus could have been established if such
returns had been a regular practice.
11Writ of certiorari denied, Pennsylvania
Department of Revenue v. Bloomingdale's By Mail,
504 U.S. 955 (1992).
Appendix III
General Sales Tax Provisions
Page 51GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
Figure III.1: Sales and Use Tax Treatment of
Selected Goods That Can Be Sold by Remote Sellers,
1999
Sales tax programs vary state-by-state in the
treatment of goods and services sold to customers
in their jurisdictions. Generally, state and
local governments tax the sale of goods unless the
state specifically exempts the sale. On the other
hand, sales of services are generally untaxed
unless the state specifically includes these sales
in its tax base. Tables III.1 and III.2 show the
variance in tax treatment for some goods and
services that can be sold by remote sellers. The
general sales tax rates also vary by state and by
local jurisdiction. The combined state and local
sales tax rate can vary for certain jurisdictions
within the same state. Table III.3 shows the
state general sales tax rate for states with sales
tax programs and the rates for some of the larger
cities in these states. Four states-Delaware,
Oregon, New Hampshire, and Montana-do not have
general sales tax program. Alaska has no state
sales tax program, but local jurisdictions may
impose sales and use taxes.
Table III.2: General Sales and Use Tax Treatment
for Selected Services That Can be Sold by Remote
Sellers, 1996
States AL AK AZ AR CA CO CT DE DC FL GA
State rate 4 0 5 4.5 6 3 6 0.4 6 6 4
Travel services E nd E E E E E E E E E
Utility services
Interstate E nd E E E E d E
telephone
Cellar telephone E nd E 4.25 7
Finance servicesa E nd E E E E E b Ec E E
Personal services
Debt counseling E nd E E E E E E E
900 number E nd E E E E E E
services
Tax return E nd E E E E E E E E
preparation
Business services
Marketing & E nd E E E E E E E E
advertising
Credit E nd E E E E E
information
Employment E nd E E E E E E E
agencies
Lobbying/consulti E nd E E E E E E E
ng
Photocopy nd 7.25 E E
services
Photo finishing nd 7.25
Printing nd 7.25 nd
Typesetting E nd E 7.25 E E E E
(industrial)
Public relations, E nd E E E E i E E E
consulting,
contract
telemarketing
Telephone E nd E E E E E E
answering service
Computers
software,
services, and
access
Canned software E nd 7.25 nd
Modified software E nd E 7.25 E
Custom (material) nd E E E E
Custom (services) E nd E E E E E E
Data processing E nd E E j E k E
and information
Mainframe access E nd E E E E E
Admissions &
amusements
Pari-mutuel E nd 6.5 E E 10 E N/A
racing
School sports nd E E E E 10 E E
Professional nd E E 10
sports
Cultural nd E E 10 E
Films/tapes E nd E E E E 2.3 E E E
(theaters)
Video tapes nd 5.5 7.25 2.3
(home)
Cable TV services E E E E 4.25 9.7 E
Service contracts E E 7.25 E
Table III.2: (cont.)
HI ID IL IN IA KS KY LA ME MD MA MI MN MS MO
4 5 6.25 5 5 4.9 6 4 6 5 5 4 6.5 7 4.225
E E E E E E E E E E E E E E
E 5 E E E E E E 6 E E
5.885 E 5 3 6
e E E E f E E E E E g E E E h
E E E E E E E E E E E E E E
E 5 E E E E E E E
E E E E E E E E E E E E E E
E E E E E E E E E E E E E E
E E E E E E E E E E E E E
E E E E E E E E E E E E E
E E E E E E E E E E E E E E
E E 6 E
E 6
0.5 E 6
E E E E E E E 6 E E
E E E E E E E E E E E E E E
E E E E E E E E E E
6
E E E E E
E E E E E E 6 E
E E E E E E E E E E E
E E E E E E E E E E E E E E
E E E E E E E E E E E E E E
n/a E E 15 E E 10 E E
E E E E E 10 E E
E E E 10 E E
E E E 10 E E E
E E E E E E E E E E E E
E 6 E
E E E E E E E
E E E E E E E E
Table III.2: (cont.)
States MT NE NV NH NJ NM NY NC ND OH OK
State rate 0 5 6.5 0 6 5 4 4 5 5 4.5
Travel services nd E E nd E E E E E E E
Utility services
Interstate E E E 6 4.25 E E
telephone
Cellar telephone 1.8 E 6 3 nd
Finance servicesa E E nd E m E E E E E
Personal services
Debt counseling nd E E nd E E E E E E
900 number nd E 6 E E
services
Tax return nd E E nd E E E E E E
preparation
Business services
Marketing & nd E E nd E n E E E E E
advertising
Credit information nd E E nd E E E E E
Employment nd E E nd E E E E E
agencies
Lobbying/consultin nd E E nd E E E E E E
g
Photocopy services nd nd
Photo finishing nd nd
Printing nd nd
Typesetting nd E nd E E E E E E
(industrial)
Public relations, nd E E nd E E E E E E
consulting,
contract
telemarketing
Telephone nd E E nd E E E E
answering service
Computers
software,
services, and
access
Canned software nd nd
Modified software nd E nd E E
Custom (material) nd E nd E E E E
Custom (services) nd E nd E E E E E E
Data processing nd E E nd E E E
and information
Mainframe access nd E E nd E E nd E E E
Admissions &
amusements
Pari-mutuel racing 1 E nd E E 3 E 14.5
School sports nd E nd 3 E E
Professional 5 E nd 3 E
sports
Cultural nd E nd E 3 E
Films/tapes E E E nd E E E E E E
(theaters)
Video tapes (home) E nd
Cable TV services nd E nd E E E E E E
Service contracts nd E nd E E E
Table III.2: (cont.)
OR PA RI SC SD TN TX UT VT VA WA WV WI WY
0 6 7 5 4 6 6.25 5 5 4.5 6.5 6 5 4
na E E E E E E E E E 0.287 E E
na
na E E E E E E E
na o p E E E E
na E E E q E r E E E s Et E E
na
na E E E E E E E E 2 E E
na 10 E E E 2 E E
na E E E E E E E E 2 E E
na
na E E E u E E E E E 2 E E E
na E E E E E E E
na E E E E E E E 1.829 E E
na E E E E E E E 2 E E E
na 4.875
na E 4.875 E
na 4.875
na E E E E 2 E
E E E E E E E E v E E
na
na E E E E 1.829 E E
na
na 4.875
na E E E
na E E E
na E E E E 2 E E E
E w E E E E 2 Ex E E
na
na E E E E E 2 E E E
na
na E E 4.875 E 1.829
na E E E E 4.875 E E 2
na E E E 4.875 E 2
na E E 4.875 E
na E E E E E E E E 2 E E E
na 4.875 4.5
na E E E 1.829 E
na E 4.875 E E
Note 1: The space is blank if the jurisdiction
taxes the sale at the general rate.
Note 2: Exempt indicated by "E."
No data available indicated by "nd."
Not applicable indicated by "na."
aFinancial services include service charges of
banking institutions, insurance services,
investment counseling, loan broker fees, and
ticker tape financial reporting.
bDelaware exempts bank service charges and
insurance services but taxes investment
counseling, loan broker fees, and ticker tape
financial reporting.
cD.C. exempts bank service charges, insurance
services, investment counseling and loan broker
fees but taxes ticker tape financial reporting.
dFlorida exempts residential interstate telephone
services but taxes business interstate telephone
services.
eHawaii exempts banking service charges but taxes
the other financial services.
fIowa taxes banking service charges and investment
counseling but exempts the other financial
services.
gMassachusetts taxes ticker tape financial
reporting.
hMontana taxes insurance services at 2.75 percent.
iConnecticut taxes public relations services but
exempts telemarketing service contracts.
jCalifornia taxes data processing services and
exempts information services.
kFlorida taxes information services and exempts
data processing service.
lMontana has no general sales tax but taxes
insurance services at 2.75 percent.
mNew Mexico exempts insurance services and loan
broker fees but taxes bank service charges,
investment counseling, and ticker tape financial
reporting.
nNew Mexico exempts national radio and television
advertising but tax other advertising services.
oPennsylvania reported no data for residential
cellular telephone services but that it taxes
industrial cellular telephone services.
pUtah taxes cellular telephone services for
industrial use at 4.875 percent and for
residential use at 5 percent.
qSouth Dakota exempts bank service charges
rTexas exempts bank service charges, investment
counseling, loan broker fees but taxes insurance
services and ticker tape financial services.
sWashington taxes financial services,
telemarketing, and public relations at lower rates
(2 percent to 1.829 percent).
tWest Virginia exempts banking service charges,
investment counseling; and taxes insurance
services at 3 percent, loan broker fees and ticker
tape financial reporting at 6 percent.
uSouth Dakota exempts advertising time and space
but taxes agency fees.
vWashington taxes marketing and advertising at
reduced rates from .506 percent to 1.829.
wSouth Carolina exempts data processing services
but taxes information services.
xWest Virginia exempts data processing services
but taxes information services.
Source: Federation of Tax Administrators, Sales
Taxation of Services: 1996 Update, April 1997.
Table III.3: State and Local General Sales Tax
Rates and Combined Rates for Selected Cities,
March 1999
Statea City (county) Rates
State County City Other Combined
Alabama Birmingham (Jefferson) 4.000 1.000 3.000 8.000
Huntsville (Madison) 4.000 1.000 3.500 8.500
Mobile (Mobile) 4.000 1.000 4.000 9.000
Montgomery (Montgomery) 4.000 1.500 2.500 8.000
Tuscaloosa (Tuscaloosa) 4.000 2.000 2.000 8.000
Alaska Juneau 5.000 5.000
Arizona Phoenix (Maricopa) 5.000 0.700 1.300 7.000
Tucson (Pima) 5.000 2.000 7.000
Yuma (Yuma) 5.000 1.000 1.700 7.700
Arkansas Fort Smith (Sebastian) 4.625 1.000 1.500 7.125
Little Rock (Pulaski) 4.625 1.000 0.500 6.125
North Little Rock 4.625 1.000 5.625
(Pulaski)
California Bakersfield (Kern) 6.000 1.250 7.250
Los Angeles (Los 6.000 1.250 1.00 8.250
Angeles)
Sacramento (Sacramento) 6.000 1.250 0.50 7.750
San Diego (San Diego) 6.000 1.250 0.50 7.750
San Francisco (San 6.000 1.250 1.25 8.500
Francisco)
San Jose (Santa Clara) 6.000 1.250 1.00 8.250
Colorado Aurora (Arapahoe) 3.000 3.750 0.80 7.550
Boulder (Boulder) 3.000 0.400 3.260 0.80 7.460
Colorado Springs (El 3.000 1.000 2.100 6.100
Paso)
Denver (Denver) 3.000 3.500 0.80 7.300
Fort Collins (Larimer) 3.000 0.750 3.000 6.750
Connecticut No local general sales 6.000 6.000
taxes
Delaware No state or local 0
general sales taxes
District of 5.750 5.750
Columbia
Florida Fort Lauderdale 6.000 6.000
(Broward)
Jacksonville (Duval) 6.000 0.500 6.500
Miami (Dade) 6.000 0.500 6.500
Miami Beach (Dade) 6.000 0.500 6.500
Orlando (Orange) 6.000 6.000
Florida St. Petersburg 6.000 1.000 7.000
(Pinellas)
Tallahassee (Leon) 6.000 1.000 7.000
Tampa (Hillsborough) 6.000 0.750 0.250 7.000
Georgia Atlanta (Fulton) 4.000 1.000 2.00 7.000
Columbus (Muscogee) 4.000 1.000 2.00 7.000
Savannah (Chatham) 4.000 1.000 1.000 6.000
Hawaii No local general sales 4.000 4.000
taxes
Idaho Boise 5.000 5.000
Ketchum 5.000 1.000 6.000
Sun Valley 5.000 2.000 7.000
Illinois Chicago (Cook) 6.250 0.750 1.000 0.75 8.750
Decatur (Macon) 6.250 1.250 7.500
Peoria (Peoria) 6.250 1.000 7.250
Rockford (Winnebago) 6.250 6.250
Indiana No local general sales 5.000 5.000
taxes
Iowa Cedar Rapids (Linn) 5.000 5.000
Davenport (Scott) 5.000 1.000 6.000
Des Moines (Polk) 5.000 5.000
Dubuque (Dubuque) 5.000 1.000 6.000
Kansas Kansas City (Wyandotte) 4.900 1.000 1.000 6.900
Topeka (Shawnee) 4.900 0.250 1.000 6.150
Wichita (Sedgwick) 4.900 1.000 5.900
Kentucky No local general sales 6.000 6.000
taxes
Louisiana Baton Rouge (East Baton 4.000 4.940 8.940
Rouge)
Monroe (Ouachita) 4.000 4.500 8.500
New Orleans (Orleans) 4.000 5.000 9.000
Shreveport (Caddo) 4.000 4.250 8.250
Maine No local general sales 5.500 5.500
taxes
Maryland No local general sales 5.000 5.000
taxes
Massachusetts No local general sales 5.000 5.000
taxes
Michigan No local general sales 6.000 6.000
taxes
Minnesota Duluth (St. Louis) 6.500 1.000 7.500
Minneapolis (Hennepin) 6.500 0.500 7.000
Rochester (Olmsted) 6.500 0.500 7.000
St. Paul (Ramsey) 6.500 0.500 7.000
Mississippi No local general sales 7.000 7.000
taxes
Missouri Independence (Jackson) 4.225 0.875 1.000 6.100
Kansas City (Jackson) 4.225 0.875 1.500 0.50 7.100
St. Louis 4.225 1.875 0.75 6.850
Springfield (Greene) 4.225 0.875 1.250 0.125 6.475
Montana No state or local 0
general sales taxes
Nebraska Lincoln (Lancaster) 5.000 1.500 6.500
Omaha (Douglas) 5.000 1.500 6.500
Nevada Las Vegas (Clark) 6.500 0.500 7.000
Reno (Washoe) 6.500 0.500 7.000
New Hampshire No state or local 0
general sales taxes
New Jersey No local general sales 6.000 6.000
taxes
New Mexico Albuquerque 5.000 0.250 0.813 5.563
(Bernalillo)
New Mexico Santa Fe (Santa Fe) 5.000 0.375 1.438 6.313
New York Albany (Albany) 4.000 4.000 8.000
Buffalo (Erie) 4.000 4.000 8.000
New York 4.000 4.000 0.25 8.250
Rochester (Monroe) 4.000 4.000 8.000
Syracuse (Onondaga) 4.000 3.000 7.000
Yonkers (Westchester) 4.000 1.500 2.500 0.25 8.250
North Carolina Charlotte (Mecklenburg) 4.000 2.000 6.000
Durham (Durham) 4.000 2.000 6.000
Raleigh (Wake) 4.000 2.000 6.000
Winston-Salem (Forsyth) 4.000 2.000 6.000
North Dakota Fargo (Cass) 5.000 5.000
Ohio Akron (Summit) 5.000 0.750 5.750
Cincinnati (Hamilton) 5.000 1.000 6.000
Cleveland (Cuyahoga) 5.000 2.000 7.000
Columbus (Franklin) 5.000 0.750 5.750
Dayton (Montgomery) 5.000 1.500 6.500
Toledo (Lucas) 5.000 1.250 6.250
Youngstown (Mahoning) 5.000 0.500 5.500
Oklahoma Oklahoma City (Oklahoma 4.500 3.875 8.375
City)
Tulsa (Tulsa) 4.500 0.417 3.000 7.917
Oregon No state or local 0
general sales taxes
Pennsylvania Philadelphia (City and 6.000 1.000 7.000
County)
No other local sales 0
taxes
Rhode Island No local general sales 7.000 7.000
taxes
South Carolina Charleston (Charleston) 5.000 1.000 6.000
South Dakota Rapid City (Pennington) 4.000 2.000 6.000
Sioux Falls (Minnehaha) 4.000 2.000 6.000
Tennessee Chattanooga (Hamilton) 6.000 2.250 8.250
Knoxville (Knox) 6.000 2.250 8.250
Memphis (Shelby) 6.000 2.250 8.250
Nashville (Davidson) 6.000 2.250 8.250
Texas Austin (Travis) 6.250 1.000 1.00 8.250
Corpus Christi (Nueces) 6.250 0.125 1.000 0.50 7.875
Dallas (Dallas) 6.250 1.000 1.00 8.250
Fort Worth (Tarrant) 6.250 0.500 1.000 0.50 8.250
Houston (Harris) 6.250 1.000 1.00 8.250
San Antonio (Bexar) 6.250 1.000 0.50 7.750
Wichita Falls (Wichita) 6.250 2.000 8.250
Utah Ogden (Weber) 4.750 1.000 0.25 6.000
Provo (Utah) 4.750 1.000 0.25 6.000
Salt Lake City (Salt 4.750 1.000 0.60 6.350
Lake)
Vermont No local general sales 5.000 5.000
taxes
Virginia Alexandria 3.500 1.000 4.500
Fairfax County 3.500 1.000 4.500
Newport News 3.500 1.000 4.500
Norfolk 3.500 1.000 4.500
Richmond 3.500 1.000 4.500
Washington Seattle (King) 6.500 1.700 0.40 8.600
Spokane (Spokane) 6.500 1.600 8.100
Tacoma (Pierce) 6.500 1.500 0.40 8.400
West Virginia No local general sales 6.000 6.000
taxes
Wisconsin Madison (Dane) 5.000 0.500 5.500
Milwaukee (Milwaukee) 5.000 0.500 0.10 5.600
Racine (Racine) 5.000 0.10 5.100
Wyoming Cheyenne (Laramie) 4.000 1.000 5.000
Lincoln 4.000 1.000 5.000
aIncludes the 50 states and the District of
Columbia.
Sources: Significant Features of Fiscal
Federalism, Volume 1-Budget Processes and Tax
Systems (American Council on International
Relations, 1995), p. x and 1999 U.S. Master Sales
and Use Tax Guide (CCH Incorporated, Mar. 1999).
Appendix IV
State and Local Government Reliance on Sales Tax
Revenue
Page 58GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
Table IV.1: State Government Reliance on Sales Tax
Revenue, 1998
State General sales tax revenue as a
percentage of
Dollars in Total Total general Total
millions tax revenue own-source general
revenue revenue
United States 155,971 32.9 25.0 18.0
Alabama 1,571 27.4 18.7 12.6
Alaska 0 0 0 0
Arizona 3,050 43.9 36.0 25.8
Arkansas 1,514 37.3 28.3 19.6
California 21,302 31.5 26.6 19.2
Colorado 1,531 26.0 18.7 14.0
Connecticut 3,032 32.3 26.5 21.0
Delaware 0 0 0 0
Florida 12,924 57.4 45.4 35.1
Georgia 3,993 34.5 27.6 19.8
Hawaii 1,425 44.9 33.2 26.0
Idaho 653 31.7 23.9 18.2
Illinois 5,596 28.3 22.5 16.6
Indiana 3,156a 32.4 24.0 18.4
Iowa 1,529 31.8 23.1 17.3
Kansas 1,619 34.7 27.3 20.8
Kentucky 1,981 27.8 21.2 15.3
Louisiana 1,981 32.6 20.6 14.5
Maine 831 35.1 26.3 18.2
Maryland 2,161 23.5 17.9 13.9
Massachusetts 2,963 20.4 15.3 11.5
Michigan 7,573 35.7 27.5 21.0
Minnesota 3,244 28.2 23.3 18.2
Mississippi 2,035 48.0 37.3 24.2
Missouri 2,628 32.0 24.7 17.7
Montana 0 0 0 0
Nebraska 920 34.9 25.9 19.0
Nevada 1,657 53.2 44.7 35.9
New Hampshire 0 0 0 0
New Jersey 4,766 30.5 21.7 16.8
New Mexico 1,455 40.7 27.6 20.4
New York 7,615 21.1 16.2 9.4
North Carolina 3,273 23.6 19.1 13.7
North Dakota 309 28.7 18.8 12.2
Ohio 5,531 31.4 23.7 17.1
Oklahoma 1,328 25.1 19.3 14.1
Oregon 0 0 0 0
Pennsylvania 6,313 30.6 23.2 17.1
Rhode Island 526 28.9 20 13.9
South Carolina 2,163 38.1 27.1 18.9
South Dakota 443 53.1 33.2 21.1
Tennessee 4,028 57.6 45.7 28.6
Texas 12,474 50.6 37.3 26.0
Utah 1,312 37.5 26.6 19.8
Vermont 195 20.3 13.3 8.9
Virginia 2,225 21.1 14.4 11.5
Washington 6,909b 58.5 46.6 36.2
West Virginia 856 28.4 20.8 13.8
Wisconsin 3,047 27.3 21.2 16.8
Wyoming 335 39.2 22.4 14.4
aIndiana figures include $547 million in
corporate gross income tax revenue.
bWashington figures include $1,854 million in
business and occupation gross receipt tax revenue.
Source: GAO based on Bureau of the Census data.
Table IV.2: Local Government Reliance on Sales Tax
Revenue, 1995-96
State General sales tax revenue as a
percentage of
Dollars in Total Total general Total
millions tax own-source general
revenue revenue revenue
United States 29,709 11.0 6.8 4.2
Alabama 961 40.5 19.0 11.5
Alaska 109 13.9 7.7 4.7
Arizona 731 19.5 12.3 7.0
Arkansas 301 26.2 13.7 7.9
California 4,315 15.2 7.8 4.2
Colorado 1,289 29.1 18.0 12.3
Connecticut 0 0.0 0.0 0.0
Delaware 0 0.0 0.0 0.0
District of Columbia 468 18.8 15.4 9.5
Florida 356 2.6 1.3 0.9
Georgia 1,525 21.7 12.1 8.2
Hawaii 0 0.0 0.0 0.0
Idaho 0 0.0 0.0 0.0
Illinois 1,135 7.5 5.3 3.6
Indiana 0 0.0 0.0 0.0
Iowa 60 2.4 1.4 0.9
Kansas 289 12.1 7.1 4.6
Kentucky 0 0.0 0.0 0.0
Louisiana 1,928 54.2 31.8 20.5
Maine 0 0.0 0.0 0.0
Maryland 0 0.0 0.0 0.0
Massachusetts 0 0.0 0.0 0.0
Michigan 0 0.0 0.0 0.0
Minnesota 25 0.6 0.3 0.2
Mississippi 1 0.1 0.0 0.0
Missouri 1,015 22.7 14.7 9.5
Montana 0 0.0 0.0 0.0
Nebraska 143 7.9 5.1 3.7
Nevada 83 6.1 3.0 1.8
New Hampshire 0 0.0 0.0 0.0
New Jersey 0 0.0 0.0 0.0
New Mexico 302 37.1 18.7 8.1
New York 6,171 16.1 11.5 7.5
North Carolina 884 19.2 9.0 5.3
North Dakota 30 6.6 4.0 2.4
Ohio 932 7.6 5.2 3.3
Oklahoma 813 41.9 21.7 13.1
Oregon 0 0.0 0.0 0.0
Pennsylvania 100 0.8 0.6 0.3
Rhode Island 0 0.0 0.0 0.0
South Carolina 54 2.4 1.1 0.7
South Dakota 121 17.1 12.6 9.1
Tennessee 1,110 29.1 15.2 10.3
Texas 2,340 12.3 7.5 5.2
Utah 247 17.8 10.1 6.0
Vermont 0 0.0 0.0 0.0
Virginia 594 8.8 6.3 4.2
Washington 1,058 21.7 11.6 6.8
West Virginia 0 0.0 0.0 0.0
Wisconsin 144 2.6 1.7 1.0
Wyoming 75 16.5 7.8 4.7
Source: GAO based on Bureau of the Census data.
Appendix V
State and Local Government Sales and Use Tax
Losses
Page 60GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
There are additional sources of uncertainty
related to state estimates. For example, we assume
that purchaser compliance rates are the same
across states.
Table V.1: State and Local Sales and Use Tax
Losses for All Remote Sales and Internet Sales
Alone in 2000
Dollars in
millions
Lower Higher scenario
scenario
All remote Internet All remote Internet
Alabama 58 5 167 54
Alaska 1 a 6 2
Arizona 52 5 169 58
Arkansas 29 2 84 25
California 298 23 1446 533
Colorado 52 5 159 56
Connecticut 61 5 191 62
Delaware 0 0 0 0
District of 13 2 48 21
Columbia
Florida 120 13 503 179
Georgia 80 7 270 95
Hawaii 12 1 38 12
Idaho 11 1 33 11
Illinois 117 13 545 212
Indiana 52 5 177 62
Iowa 26 2 94 31
Kansas 31 3 103 33
Kentucky 43 4 135 45
Louisiana 77 7 237 81
Maine 13 1 41 14
Maryland 65 5 199 60
Massachuset 66 6 221 83
ts
Michigan 109 10 343 125
Minnesota 49 5 192 72
Mississippi 34 3 99 32
Missouri 65 6 205 69
Montana 0 0 0 0
Nebraska 22 2 67 22
Nevada 21 2 77 29
New 0 0 0 0
Hampshire
New Jersey 101 10 346 130
New Mexico 21 2 65 21
New York 196 22 889 357
North 62 6 231 84
Carolina
North 7 1 21 7
Dakota
Ohio 108 11 375 141
Oklahoma 48 4 137 45
Oregon 0 0 0 0
Pennsylvani 102 12 381 156
a
Rhode 12 1 40 15
Island
South 36 3 114 36
Carolina
Lower scenario Higher scenario
All remote Internet All remote Internet
South 7 1 26 8
Dakota
Tennessee 50 6 239 85
Texas 252 26 992 342
Utah 18 2 65 23
Vermont 7 1 19 6
Virginia 47 5 175 69
Washington 82 8 284 98
West 21 2 62 18
Virginia
Wisconsin 51 5 173 58
Wyoming 6 a 18 6
Note: Estimates are combined state and local
losses.
aLess than 500,000.
Source: GAO estimates.
Table V.2: State and Local Sales and Use Tax
Losses for All Remote Sales and Internet Sales
Alone in 2003
Dollars in
millions
Lower scenario Higher scenario
All remote Internet All Internet
remote
Alabama 144 19 415 184
Alaska 2 1 13 8
Arizona 130 18 420 191
Arkansas 67 8 200 85
California 686 86 3650 1720
Colorado 130 18 394 181
Connecticut 150 20 466 205
Delaware 0 0 0 0
District of 37 6 128 66
Columbia
Florida 321 48 1279 595
Georgia 200 28 675 312
Hawaii 30 4 91 39
Idaho 28 4 82 36
Illinois 298 44 1389 671
Indiana 134 19 444 204
Iowa 64 9 230 103
Kansas 81 11 253 111
Kentucky 105 14 333 150
Louisiana 191 26 593 270
Maine 34 5 103 46
Maryland 154 20 472 199
Massachuset 172 25 574 274
ts
Michigan 276 39 882 415
Minnesota 129 19 489 232
Mississippi 86 12 246 109
Missouri 164 23 512 232
Montana 0 0 0 0
Nebraska 52 7 164 73
Lower scenario Higher scenario
All remote Internet All Internet
remote
Nevada 57 8 199 95
New 0 0 0 0
Hampshire
New Jersey 256 37 879 419
New Mexico 51 7 158 70
New York 521 81 2339 1,155
North 166 25 593 279
Carolina
North 17 2 50 22
Dakota
Ohio 286 43 955 454
Oklahoma 121 16 343 154
Oregon 0 0 0 0
Pennsylvani 281 45 1,012 505
a
Rhode 30 5 101 48
Island
South 93 13 276 120
Carolina
South 19 3 62 27
Dakota
Tennessee 139 22 606 282
Texas 655 96 2466 1125
Utah 47 7 162 75
Vermont 16 2 46 20
Virginia 123 18 458 224
Washington 213 30 712 326
West 52 7 147 62
Virginia
Wisconsin 126 17 424 190
Wyoming 14 2 45 21
Note: Estimates are combined state and local
losses.
Source: GAO estimates.
Appendix VI
List of Contacts
Page 63GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
Private Sector Research Groups
ActiveMedia
Boston Consulting Group
Cyber Dialogue
Direct Marketing Association
Emarketer
Ernst & Young
Forrester
Giga
IDC
Jupiter
Yankee Group
National Organizations Representing the Public
Sector
Federation of Tax Administrators
Government Finance Officers Association
National Association of Counties
National Association of State Budget Officers
National Conference of State Legislatures
National Governors' Association
National League of Cities
National Tax Association
Multistate Tax Commission
Industry Representatives
Census, Retail and Wholesale Indicator Program
CommerceNet
Direct Marketing Association
Economic and Statistics Administration
National Retail Federation
National Association of Manufacturers
Sears, Roebuck and Company
The Internet Tax Fairness Coalition
Department of Commerce
Census Bureau
Secretary of Electronic Commerce
Department of the Treasury
Office of International Tax Counsel
States
California, State Board of Equalization
Connecticut, Department of Revenue Services
Florida, Department of Revenue, Research and
Analysis Division
Georgia, Department of Revenue, Research and
Analysis Division
Kansas, Department of Revenue, Policy and Research
Michigan, Bureau of Revenue, Revenue and Tax
Analysis
Minnesota, Department of Revenue, Sales and Use
Tax
Nebraska, Department of Revenue, Research and
Audit Divisions
New Mexico, Taxation and Revenue Department
New York, State Department of Taxation and
Finance, Tax Policy
and Analysis
North Carolina, Department of Revenue, Tax
Research Division
Ohio, Department of Taxation, Tax Analysis and
Local Government
Distribution Division
Pennsylvania, Department of Revenue, Fiscal Policy
and Analysis
Texas, Comptroller of Public Accounts, Fiscal
Management
Utah, State Tax Commission, Economic and
Statistical Unit
Washington, Department of Revenue, Taxpayer
Services
Wisconsin, Department of Revenue, Division of
Income, Sales and
Excise Taxes
Academics and Others
Lynda McDonald Applegate, Economic Council on
Information
Management and Technology (ECIMT), and Harvard
Business School
Bradley S. Dugger, ECIMT, and Chief of Information
Systems, Tennessee
Paul E. Rummell, ECIMT and President's Chief
Executive Officer, RLG
Netperformance
Bill Fox, University of Tennessee
Austan Goolsbee, University of Chicago
Charles McClure, Jr., Hoover Institution
Tom Neubig and Robert Cline, Ernst & Young
Holley Ulbrich, ACIR studies on Taxation of Out-of-
State Mail Order Sales
Appendix VII
GAO Contacts and Staff Acknowledgments
Page 64GAO/GGD/OCE-00-165 Sales Taxes on e-Commerc
e
GAO Contacts
James R. White, (202) 512-9110
James A. Wozny, (202) 512-9110
Acknowledgments
In addition to those named above, Helen
Fauntleroy-Branch, Tara Carter, Daniel E. Coates,
Shirley A. Jones, Kirsten Landeryou, MacDonald R.
Phillips, Loren Yager, Jena Sinkfield, Jennifer
Trombley, Katherine Wheeler Raheb made key
contributions to this report.
*** End of Document ***