Publications & Articles
GAPM Articles > > Elective Modules > Internet Taxation in North America
Internet Taxation in North America
By CBO Government Research
Economic Issues in Taxing Internet and Mail-Order Sales
Preface
The collection of state and local sales and use taxes when individuals purchase items over the Internet or from other remote sources is an issue that is now before the Congress. The Internet Tax Freedom Act of 1998 established an Advisory Commission on Electronic Commerce to make recommendations about Internet taxation, including whether to require retailers to collect sales taxes on Internet purchases. But the commission made no formal recommendations, and the issue has remained unresolved. This Congressional Budget Office (CBO) paper--prepared at the request of the Senate Finance Committee--uses an economic framework to evaluate the various arguments that have been advanced by proponents and opponents of remote collection of sales taxes. In keeping with CBO's mandate to provide objective analysis, the report makes no recommendations.
Dennis Zimmerman of CBO's Tax Analysis Division prepared the paper under the supervision of G. Thomas Woodward and Roberton Williams. The paper benefited from comments by reviewers outside of CBO, including Thomas Neubig of Ernst & Young and William Fox of the University of Tennessee. Within CBO, Robert Dennis, Theresa Gullo, Roger Hitchner, Arlene Holen, Deborah Lucas, David Moore, Robert Murphy, Jennifer Smith, and John Sturrock provided helpful suggestions.
Leah Mazade edited the manuscript, and Christine Bogusz proofread it. Denise Jordan-Williams and Simone Thomas typed early drafts of the paper. Lenny Skutnik produced the printed copies, and Annette Kalicki prepared the electronic versions for CBO's Web site.
Douglas Holtz-Eakin
Director
October 2003
CONTENTS
Summary An Overview of the Remote Sales Taxation Issue States' Inability to Collect Tax on Remote Sales The Growth of E-Commerce The Magnitude of Uncollected State and Local Use Tax Revenue The Debate: Whether to Require Remote Sellers to Collect Use Taxes The Economic Trade-Off That Remote Collection Presents Nonneutral Taxation of Commodities Administrative and Compliance Costs Current Efforts to Reduce Compliance Costs Other Economic Issues in the Remote Collection Debate Network Externalities and Market Failures The Size of the Public Sector International Considerations Remote Collection and Fairness Fiscal Autonomy of States and Local Governments
Tables 1. State and Local General Sales Tax Rates and State General Sales Tax Revenue2. Sales Tax Revenue for Selected Local Governments with Populations of 200,000 or More, 2000 Boxes 1. Origin-Based Versus Destination-Based Sales Tax Systems2. Taxing Digital Goods
Summary
All of the 45 states, as well as the District of Columbia, that levy retail sales taxes on purchases made within their borders (local sales) also impose an equivalent "use" tax on their residents' purchases of items out of state (remote sales). Local governments also impose sales and sometimes use taxes. (However, because states in many cases administer those taxes, the general discussion that follows uses "states" for simplicity's sake to mean both states and local governments.) Previously, those out-of-state purchases were mainly catalog and telephone sales. But the growth of on-line, or "e" (electronic), commerce over the past several years has increased remote sales by billions of dollars.
States rely on retailers to collect sales taxes from consumers on their local purchases and to remit them to the states, an approach with relatively low administrative costs and a relatively high rate of collections. But decisions by the Supreme Court have established that states do not have the authority to require out-of-state sellers to collect and remit use taxes on remote sales unless the seller has established a connection, or "nexus," with the purchaser's state--in this instance, a physical presence (an office or other place of business, property, or an agent). As a result, states rely on the use tax, which is paid directly by purchasers and whose administrative costs are relatively high and rate of collections quite low. Estimates of uncollected use taxes from all remote sales in 2003 range from $2.5 billion to $20.4 billion. Projections for 2011 of uncollected taxes from on-line commerce also vary widely, ranging from $4.5 billion to $54.8 billion.
In addition to its holding on nexus, the Court stated that the Congress could permit states to require remote sellers to collect use taxes, a conclusion that has led many jurisdictions to look to the federal government to confer that authority on them. A federal advisory committee studied the question but was unable to reach a consensus and thus issued no formal recommendations. Opponents and proponents of granting such authority to states have advanced a variety of arguments; however, the central economic issue is the magnitude of two kinds of costs related to taxation: the distortions (referred to as excess burden) that come from taxing goods unevenly and the administrative and compliance costs of collecting revenue.
Remote-seller collection of use taxes would eliminate the uneven taxation of identical goods purchased from a local seller and a remote seller and thereby reduce the loss of national income that results when such tax differentials cause people to make purely tax-motivated decisions about consumption and production. Consumers may be willing to purchase a good remotely even if the total cost of production and delivery exceeds the comparable in-state cost because the money they save in taxes compensates them for the money they pay in shipping costs. Similarly, producers may be willing to construct facilities in locations where production and shipping costs are high to avoid nexus and the need to charge their customers sales taxes. The more unevenly a tax is applied, the more producers and consumers waste resources in efforts to avoid it--thereby reducing economic efficiency. And if a greater fraction of sales escapes taxation over time, states may seek to maintain the same level of receipts by raising tax rates, which would increase the tax system's excess burden.
While taxing remote sales could permit lower overall tax rates and reduced administrative costs, it is also likely to increase the burden of compliance costs imposed on retailers doing business in many states compared with the burden borne by local sellers doing business in one state. In addition to the 45 states that impose sales taxes, local governments in most of them may impose such levies. A firm engaged in business nationwide could thus be dealing with more than 6,000 tax regimes, many with different tax bases (the goods and services to which the tax applies), tax rates, and rules about who is subject to tax. Evidence suggests that the cost of complying with that multiplicity of tax systems, particularly for smaller firms, will exceed compliance costs for local sellers dealing with a single sales tax system.
Those two economic costs arising from taxation--distortion costs and compliance costs--represent a loss to society. That loss constitutes the social costs of taxation (costs in excess of the revenue collected). Requiring remote sellers to collect and remit use taxes would have unclear effects on social costs: distortion costs would probably be reduced, but compliance costs would probably increase. The decision by policymakers to either grant or withhold from states the authority to collect use taxes on remote sales involves a trade-off between those two costs. Several proposals to grant that authority have been introduced in the 108th Congress.
Federal action is not the only source of policy changes that could affect that trade-off. As of August 2003, more than 30 states had developed and 20 adopted the Streamlined Sales and Use Tax Agreement (SSUTA). The SSUTA is intended to reduce the compliance costs imposed on remote sellers by implementing three strategies: (1) adopting computer technology and institutional arrangements to enable vendors to comply more cheaply with multiple tax laws and taxing authorities; (2) adopting common tax-base definitions for categories of consumption, similar tax bases, and similar tax-filing and audit procedures; and (3) compensating vendors for the compliance costs that remain after the first two strategies have been implemented.
Proponents and opponents of remote collection raise other issues as part of the debate. Opponents have argued that effectively exempting Internet purchases from sales taxes is a means of inducing more people to use the Internet and stimulating the growth of e-commerce. The growth of Internet use is desirable because the Internet exhibits "network externalities"--a person's joining the network benefits not only him- or herself but also other participants, by adding to the total number of participants in the network. Too few people use the Internet when those external benefits are not reflected in the price of access to it. Providing a subsidy would thus increase use of the Internet and benefit society.
Network externalities, however, exist mainly in the early stages of a network's growth. For that reason, it is difficult to justify incentives for a network such as the Internet, which was used by 56 percent of the U.S. population above the age of eight in 2001. And even if such externalities are present, not collecting use taxes on remote purchases is an indirect and unevenly targeted means of encouraging more people to use the Internet (and in states that have no sales tax, no incentive at all). Thus, such an approach may not increase overall efficiency, whereas a more direct incentive for Internet use or access might.
Some opponents of remote-seller use tax collection agree that nonneutral, or uneven, taxation of commodities can lead to higher tax rates and excess burden. But they also argue that collecting use taxes on remote sales will nonetheless increase the social costs of taxation if, as some people believe, governments have an inherent tendency to be too large. If taxes are costly to collect, that cost can help constrain government spending. Reducing the costs of collecting use taxes, therefore, may not result in lower rates on other taxes and a lessening of distortion costs but instead will encourage larger government and increase the social costs of taxation. Those analytic possibilities notwithstanding, the empirical evidence regarding such claims is mixed. And regardless of the issue of remote collection, concerns about the magnitude of government spending could be addressed through the political process with a variety of tax and spending controls that have been adopted by some states and local governments.
Tax equity, another issue in the remote collection debate, has been used as the basis for some proponents' arguments. They contend that not collecting use taxes violates principles of fairness in that two individuals who purchase identical goods may pay different taxes because one of them buys the good over the Internet. Proponents further maintain that because Internet access--as well as the use of credit cards and other instruments typically used to shop on-line--tends to rise with income, the effective exemption of e-commerce from taxation causes state sales tax systems to be more regressive than they would be without such an exemption.
Both arguments represent value judgments about how taxes should be levied. As a result, one cannot evaluate the effect that the violation of those equity norms has on economic well-being as one can evaluate the effect of violating the efficiency standard. Consumers who buy the same goods through different channels choose to do so and in other ways are not identical; comparing them to assess fairness is thus problematic. As for the differential effects on consumers with different income, the potential for e-commerce to make the sales tax more regressive diminishes as Internet access becomes more nearly universal.
Opponents of requiring remote sellers to collect use taxes raise a different fairness issue--one related to businesses that receive no benefits or services from a jurisdiction because they are located elsewhere. Those opponents argue that such businesses should not have to pay taxes in support of those benefits and services. But the "tax" in question is a levy on consumers that is merely collected by firms, and any impact on those businesses appears as compliance costs, a burden that could be lightened by compensation. To the extent that issues of tax equity arise, they are centered on the effects on residents of a taxing jurisdiction who purchase goods and services remotely.
Another concern expressed by opponents of remote-seller tax collection is the question of states' fiscal autonomy. Granting the authority to collect use taxes on remote sales in exchange for simplifying sales tax regimes, they argue, would limit states that have sales taxes in tailoring their tax to their citizens' preferences. To the extent that states choose well on behalf of their residents, those choices presumably increase well-being. As a general matter, however, the issue is beyond the scope of economic analysis.
Of the various arguments related to efficiency and fairness, nonneutral taxation and compliance costs appear to be central to the issue of remote sales tax collection. However, even the gains from eliminating nonneutral taxation would depend on the degree to which international differences in taxing remote sales persisted and the ability to track and tax digital goods (for example, digitized music). The gains from reducing compliance costs have yet to be demonstrated.
Economic Issues in Taxing
Internet and Mail-Order Sales
States and local governments that levy taxes on retail sales have long faced the issue of the cost of collecting that revenue when their residents purchase items by mail or telephone from retailers that are located out of their jurisdictions--so-called remote sellers. Retail sales over the Internet, known as electronic, or "e," commerce, have boosted remote sales and increased states' and local governments' interest in collecting taxes on those purchases. The uncollected revenue from the e-commerce share of such sales is expected to continue to increase, but projections of its magnitude vary widely--for example, for 2011, they range from $4.5 billion to $54.8 billion.(1)
States and local governments require merchants to collect sales taxes at the point of sale and remit them to the government. That system works well when the sale takes place within the taxing jurisdiction (local sales). But those governments have limited authority to require sellers outside their jurisdictions to collect tax on purchases made by their residents. As a result, all of the states that levy sales taxes have adopted a companion "use" tax, to be collected from residents, that applies to goods purchased from out-of-state vendors that are "used" at home. The use tax is intended to reduce the incentive for residents to purchase goods from retailers located in lower-tax (or no-tax) jurisdictions. But the administrative costs of collecting use taxes are relatively high. As a result, governments tend to avoid those costs and rely primarily on residents to voluntarily pay the taxes. Relatively little revenue is collected, however.(2) Consequently, the underlying economic issue is not whether sales over the Internet (or by mail or telephone) are to be subject to tax--they already are. The issue is rather the cost of collecting such taxes.
Under the Constitution, the Congress could reduce the administrative costs of collecting sales and use taxes by giving states and local governments the authority to require remote sellers to collect them. Such a requirement would eliminate some tax differentials among commodities and reduce the loss of national income that results when that uneven taxation causes people to decide whether to consume or produce something on the basis of the taxes that apply. The costs arising from the "distortion" of those decisions are sometimes referred to as excess burden. But the requirement would also impose higher compliance costs on remote sellers, who would have to keep track of a complex system of tax bases (see the discussion below) and rates imposed by the 45 states and thousands of local jurisdictions that levy sales and use taxes.
Those two effects--distortion costs and compliance costs--represent a loss to society and constitute the social costs of taxation (costs in excess of the revenue collected). A decision to require remote sellers to collect and remit use taxes would have conflicting effects on social costs: it would reduce excess burden but increase administrative and compliance costs. As a result, federal policymakers face a difficult choice. In the Internet Tax Freedom Act of 1998 (Public Law 105-277), the Congress established the Advisory Commission on Electronic Commerce to develop recommendations about taxation related to the Internet, including recommendations on expanding the duty of retailers to collect taxes on remote sales. But the commission was unable to reach a consensus and issued no official recommendations.
Thus, the Congress still faces the question of whether to act on this issue and, if so, what action to take. This Congressional Budget Office paper examines in more detail the trade-off between distortion costs and administrative and compliance costs.(3) It also evaluates several other claims that have been advanced as part of the debate about whether to grant states the power to require remote sellers to collect use taxes. Those claims concern the effect of that power on the growth of the Internet and on the growth of government. In addition, the claims raise issues of fairness (who pays taxes) and of states' fiscal autonomy (the freedom to determine their tax policy without interference from the federal government).
An Overview of the Remote Sales Taxation Issue
General sales taxes were first adopted in the 1930s. Today, 45 states, the District of Columbia, and many local governments levy sales and use taxes, imposing their own tax rates and choosing their own definitions of the "base"of goods and services to which the tax applies. Sales taxes were originally conceived as a levy on all retail sales with sellers acting as collection agents; however, the base of the tax in most jurisdictions is neither comprehensive nor uniform. For example, sales tax bases typically exclude most services and intangible products, such as medical, legal, financial, and education services. Many jurisdictions also exempt some categories of tangible goods, such as food and medicine, to lighten the burden of the tax on lower-income taxpayers. In contrast, the base includes a substantial amount of goods purchased by businesses--which means that those commodities are effectively taxed twice under the sales tax, once as an input in a production process and once as part of the final good that is produced.(4)
The general sales tax finances a substantial portion of states' public services, accounting in 2002 for 33.5 percent of states' total tax revenue (see Table 1). Among the states, that share varies considerably: it accounts for more than 50 percent of tax revenue in Arizona, Florida, Nevada, South Dakota, Tennessee, Texas, and Washington and, as expected, for none of the revenue in the five no-sales-tax states--Alaska, Delaware, Montana, New Hampshire, and Oregon. Thirty-five states also allow local governments to levy sales taxes. (However, in many cases, the states administer those local taxes.)
States' Inability to Collect Tax on Remote Sales
As mail-order businesses have expanded, states have sought ways to maintain their revenue base in the face of a growing volume of out-of-jurisdiction purchases. In that pursuit, they tested their authority to require remote sellers to collect the use tax applicable in the customer's state. But the Supreme Court invalidated those statutes as unconstitutional constraints on interstate commerce in cases in which the remote sellers had an insufficient connection to the taxing states.
For instance, in 1967, in National Bellas Hess, Incorporated, v. Department of Revenue for the State of Illinois, the Court struck down an Illinois statute that required remote sellers to collect use taxes on their sales to Illinois customers.(5) National Bellas Hess was a Missouri mail-order business whose only connection with customers in the state of Illinois was through the U.S. Postal Service or common carriers (for example, United Parcel Service). The Court held that under the Constitution's commerce and due process clauses, the firm lacked the requisite physical presence in Illinois--or "nexus" with the state--either to justify the burden that the statute imposed on interstate commerce or to overcome the statute's "offense" to due process.(6)
The Court used the firm's costs for complying with such statutes to illustrate the burden that the Illinois statute imposed on interstate commerce. It concluded that if the Illinois statute was upheld, the decision could potentially subject National Bellas Hess's interstate business to the tax rates, exemptions, and recordkeeping requirements of every political subdivision in the country. The result, according to the Court, would be to "entangle National's interstate business in a virtual welter of complicated obligation to local jurisdictions who would have no legitimate claim that they were imposing 'a fair share of the cost' of government."
In a more recent case, the Court signaled its view of where resolution of the issue of remote sales taxation might be pursued. In Quill Corporation v. North Dakota, the Court was again faced with a state statute that required remote sellers to collect use taxes on their sales to out-of-state customers.(7) The Quill Corporation, a Delaware-based remote seller, was connected to customers in North Dakota only by common carrier or the mail. The Court found that Quill's "minimum contacts" with the state satisfied the requirements of due process but that the firm nevertheless lacked what the judges now termed a "substantial nexus" with the state as required under the commerce clause. The Court thus upheld the standard it had expressed in National Bellas Hess and concluded that Quill lacked the requisite physical presence in North Dakota to justify the burden that the state statute imposed on interstate commerce. The Court noted that the underlying issue of taxing remote sales was "not only one that Congress may be better qualified to resolve but also one that Congress has the ultimate power [under the Constitution] to resolve."
The Growth of E-Commerce
Determining the rate of growth of e-commerce over more than a few years is difficult. The federal government first published complete data on total e-commerce for 1999. The value of Internet sales grew by more than half over the next two years, rising from $660 billion in 1999 to just over $1 trillion in 2001. More than 93 percent of that latter amount was attributable to business-to-business transactions for "manufacturing shipments" and "merchant wholesale trade sales," categories of economic activity that the Bureau of the Census uses in its e-commerce surveys.(8) Business-to-consumer transactions appear in two other categories: "retail trade sales" and "service revenue." (The term "total e-commerce" covers both business-to-business and business-to-consumer transactions.)
Retail sales statistics give some indication of the growth of e-commerce over the past decade. The Census Bureau collects time-series data on the retail component of total e-commerce in its "Monthly Retail Survey"; the category "electronic shopping and mail-order houses"covers all remote retail sales (mail order, telephone, and e-commerce). Those sales have grown rapidly, rising from $35.1 billion in 1992 to $109.7 billion in 2001 and $116.6 billion in 2002. The e-commerce share of those sales cannot be identified for 1992, but data for 1999 to 2001 show the share almost doubling, from 16.6 percent to 31.3 percent.
The Magnitude of Uncollected State and Local Use Tax Revenue
The concerns of states and local governments about uncollected use tax revenue (from both business-to-consumer and business-to-business transactions) predate the rise of sales on the Internet. Uncollected revenue from remote sales in 1994 reached $3.3 billion according to the Advisory Commission on Intergovernmental Relations, or 2.5 percent of states' sales and use tax collections and 0.8 percent of their total tax revenue.(9) Perhaps because many observers expected evolving computer technology to increase remote sales and uncollected use tax revenue, the policy debate is often described as being about taxation of sales over the Internet. But, as noted earlier, the relevant issue is the cost of collecting tax revenue from all remote sales rather than the increase in such uncollected revenue caused by e-commerce. Any policy requiring remote sellers to collect and remit use taxes levied on e-commerce would probably be applied to mail-order and telephone sales as well (assuming that all types of remote sellers would bear similar compliance costs).
The General Accounting Office in 2000 projected upper and lower bounds for uncollected revenue from remote sales for 2003.(10) Its upper-bound estimate was $20.4 billion, or 7.9 percent of projected state and local general sales tax revenue of $256.4 billion, and its lower-bound estimate was $2.5 billion. Those projections are probably too high because they predated the recent recession (total sales tax revenue for the year ending June 2003 was only $228 billion). But the wide range of the estimates reflects the uncertainty arising from shortcomings in the data available to quantify uncollected revenue from remote sales. Factors that must be estimated include the rate of growth of e-commerce; the proportion of e-commerce that is not part of the sales tax base; the share of e-commerce that is part of the tax base but that represents purchases by exempt entities; and the proportion of taxable e-commerce on which tax is already being collected and that replaces other forms of remote sales. A series of assumptions must be made to adjust for those factors in estimating total sales over the Internet.
More-recent estimates of uncollected tax revenue from all remote sales are not available. But two forecasts of such revenue from the e-commerce share of remote sales for 2006 and 2011 assign different values to each of the factors contributing to overall uncertainty and consequently reach very different conclusions. According to one set of estimates, uncollected state and local revenue would be $45.2 billion in 2006 and $54.8 billion in 2011, losses that would represent 5.6 percent of state tax revenue in 2006 and 5.4 percent in 2011.(11) The share of local tax revenue lost would be 1.8 percent in both years. Another set of estimates is considerably lower--$3.2 billion in 2006 and $4.5 billion in 2011, or less than 1 percent of state and local tax revenue in both years.(12)
The Debate: Whether to Require Remote Sellers to Collect Use Taxes
The potential loss of revenue stemming from states' and local governments' inability to efficiently collect use taxes on remote purchases has generated a policy debate over whether vendors should be required to collect use taxes in their stead. The Supreme Court's decision in Quill Corporation v. North Dakota makes it clear that only the Congress can give states the authority to require remote sellers to collect use taxes. In fact, that policy role is the federal government's only stake in the Internet sales tax debate; the issue has no direct federal budgetary implications.
The proponents and opponents of such regulation call on principles related to economic efficiency, fairness, and states' fiscal autonomy to support their respective positions. According to one side or the other in the debate, requiring remote sellers to collect use taxes will:
Have opposing effects on the social costs of taxation--decrease the loss of national income that results when the differential taxation of commodities causes tax-motivated decisions about consumption and production (excess burden) and increase the compliance costs that would be imposed on remote sellers to collect and remit use taxes;Increase the size of government and eliminate a tax advantage that is helping the Internet grow to its economically desirable size;
Distribute the burden of sales taxes more equitably and treat people in equal circumstances equally;
Impose a tax burden on remote sellers who, unlike local sellers, receive no compensating public service benefits (for example, fire and police protection); or
Compromise the fiscal autonomy of states and local governments, which is guaranteed by the Constitution, if standardization of tax bases and rates is required to reduce compliance costs.
Faced with competing demands involving e-commerce, the Congress considered legislation in 1998 to tax Internet access (the fees paid to be connected to the Internet) and require remote sellers to collect taxes on Internet sales. Policymakers eventually enacted the Internet Tax Freedom Act (ITFA), which allowed existing taxes on Internet access to remain in effect but imposed a three-year moratorium on new federal, state, and local access levies. In addition, it allowed existing taxes on sales over the Internet to remain in effect and permitted governments to impose new taxes on such sales as long as they applied equally to sales made by other means, but it prohibited discriminatory taxes on Internet sales. The law did not give states and local governments the authority to require that remote sellers collect sales taxes.
The ITFA, as noted earlier, also established the Advisory Commission on Electronic Commerce, which was to report to the Congress in April 2000 with recommendations about Internet taxation, including sales taxes. However, the commission was unable to achieve the two-thirds majority (13 of 19 members) required to issue official findings and recommendations about remote sellers' expanded duty to collect sales taxes.(13) Policymakers were split into two major camps: those who would make the Internet a tax-free zone and those who would allow collection of use taxes on remote sales provided state and local governments met certain requirements for simplifying and standardizing their tax bases or rates (or both). In November 2001, a compromise of sorts was reached by policymakers' extension of the ITFA for two years. The extension's pending expiration in 2003 makes it likely that the current Congress will address the issue.
The Economic Trade-Off That Remote Collection Presents
In general, the economic issue involved in collecting use taxes on remote sales (in short, remote collection) is fairly straightforward, notwithstanding the variety of economic arguments advanced by both sides in the policy debate. Remote collection reflects a trade-off between two kinds of social costs that arise from taxation: the loss of national income from nonuniform taxation and the loss from incurring administrative and compliance costs for collection. The desirability of remote collection depends on the magnitudes of those two kinds of costs.
Nonneutral Taxation of Commodities
Virtually all taxes impose costs in the form of reduced economic efficiency. That is, the loss to taxpayers paying the tax is greater than the value of the taxes collected because taxes distort decisions about private consumption and production and lead to less efficient use of national economic resources. Research indicates that to raise a given amount of tax revenue, those distortions, or excess burden, can be minimized by taxing all commodities uniformly up to the point that taxing an additional commodity will increase the costs of administering the tax by more than it reduces excess burden.(14) Within that framework, a larger public sector will require a higher tax rate to raise more revenue. Because the loss of economic well-being from a tax increases more than proportionately with the rise in the tax rate, the outcome will be added inefficiency from the higher rate on taxed goods.
It is the violation of the standard of uniform tax rates that supporters of remote-seller tax collection cite as part of their concern about e-commerce. Under current arrangements, the tax-inclusive price that a consumer pays for a good varies depending on whether the consumer purchases it locally and pays tax on it or buys it over the Internet without complying with the use tax. That kind of price differential causes some decisions about both consumption and producers' locations to be motivated by taxes rather than by the cost of the resources used to produce and sell the good.
For example, a consumer might choose to purchase books over the Internet for $100 inclusive of the shipping cost, pay no sales tax, and fail to comply with the use tax rather than purchase the same books at a local bookstore for $102 inclusive of a local $5 sales tax. The real resource cost of the books (including profit) purchased from the Internet seller is $100; that is, the "market" values the resources that are used to produce and deliver those books at $100. The real resource cost of the same books (including profit) available for sale from the local bookseller is $97; the portion of the books' cost that is sales tax ($5) is a transfer from the consumer to the government and uses no resources. Thus, the tax differential that results from the consumer's noncompliance with the use tax causes this consumer to make a choice that increases the production cost of books by $3. That money represents a loss of economic well-being to society because those $3 worth of resources could have been used to produce $3 worth of other goods or services.
The same circumstances can be used to show how a tax differential may affect decisions about production. Suppose the Internet bookseller located its East Coast distribution center in jurisdiction A, a small state that has few potential customers and in which distribution costs for the book order discussed above are $10, rather than in jurisdiction B, a large state that has many potential customers and in which distribution costs are $7. The bookseller made that decision because by choosing jurisdiction A, it avoided a physical presence in jurisdiction B and the requirement to collect jurisdiction B's $5 sales tax from customers who live there. The bookseller thus uses $10 of resources for distribution and charges a jurisdiction B customer $100 (with no sales tax) rather than using $7 of resources for distribution and charging $102 (inclusive of sales tax).
Regardless of whether the effect comes through decisions about consumption or production, the potential for distortion costs arises when differing taxation causes the price of a good to vary according to whether it is purchased locally or remotely. In addition, empirical work confirms that tax differentials do matter. Studies have shown that retail prices rise when a sales tax is imposed;(15) research also indicates that differences in sales tax rates along state borders cause consumers to switch their purchases from the higher-tax to the lower-tax jurisdiction.(16) Some evidence even suggests that taxing tangible goods but not services may have contributed to growth in the consumption of services relative to goods.(17) Other research shows that reduced economic well-being can result when consumers alter their choices of what to buy because of differential taxes on commodities.(18)
In sum, avoiding a situation in which residents of a given state face differential taxation on the same good tends to minimize distortions. Not collecting use taxes on e-commerce amplifies the differential taxation that is already occurring through mail-order, telephone, and cross-border sales. Short of eliminating all sales taxes, policymakers cannot correct the losses from such nonneutral taxation by adopting an alternative approach.
Administrative and Compliance Costs
Every tax system must be administered; tax laws must be enforced, and taxpayers--both individuals and businesses--must spend time and money to comply with those laws. Administrative costs can be thought of as the expenses incurred by the tax authorities, the state and local agencies charged with collecting taxes. Compliance costs are borne by others, typically those who pay a tax, such as the individual taxpayer in the case of an income tax. In the case of the general sales tax, however, those costs fall not on consumers directly but rather on retailers who must collect the tax and remit it to the state or to the local government. Of course, some compliance costs may be passed on to the consumer in the form of higher prices for retailers' goods.(19) (As discussed later, consumers would also bear compliance costs indirectly in their role as taxpayers if tax authorities compensated retailers for those costs out of collected revenue.)
Administrative and compliance costs, therefore, represent the other major social cost of taxation. Like distortion costs, they are a cost above and beyond what is collected and transferred from the private sector to governments for public use. Consequently, all else being the same, it is desirable to minimize the costs of administering and complying with a tax. Although precise information on such costs for different taxes is largely unavailable, states and local governments nonetheless appear to design their tax structures to account for them. For example, some services are excluded from the existing sales tax base because they are thought to entail higher-than-usual administrative costs per dollar of revenue collected.(20)
High administrative costs help explain why states and local governments seldom enforce use taxes for most business-to-consumer remote sales. Consider the example of Connecticut's Operation Equity program. The program pays other states 50 percent of the first year's use tax that Connecticut collects as a result of information provided by those states' sales tax audits. The size of that percentage suggests that the cost of achieving compliance by other means--for example, by pursuing individual Connecticut taxpayers--is high.
One way to reduce administrative costs is for the federal government to give states the authority to require remote sellers to collect use taxes on those sales and remit them to the purchaser's state of residence, a mandate that would eliminate costly collection efforts by the states. From the perspective of state authorities, such a policy would unambiguously reduce the social costs of taxation--administrative costs and perhaps distortion costs as well. From the perspective of federal policymakers, the effect on social costs is less certain. Under that approach, the reduction in a state's or local government's administrative costs would be accompanied by an increase in remote retailers' compliance costs. Those costs would probably be higher for remote sales than for local sales. Vendors making remote sales would be dealing with the varied tax bases, rates, and audit procedures of as many as 45 states (those that levy sales taxes) and an unknown number of local jurisdictions that impose their own sales tax and do not rely on the state to administer it. In contrast, vendors making only local sales usually face one definition of the base to which the sales tax applies and at most two rates (one state and one local).(21)
Thus, the size of vendors' compliance costs relative to the distortion costs of sales taxes appears to be central to the question of whether it is efficient to require remote sellers to collect use taxes. Yet few data exist for measuring those costs. A recent analysis for the state of Washington found that compliance costs ranged from 0.97 percent of the state's sales tax revenue collected by large firms (defined as having at least $1.5 million in annual gross sales) to 6.47 percent of the revenue collected by small firms (firms with at least $150,000 but less than $400,000 in annual gross sales).(22) Robert J. Cline and Thomas S. Neubig, in a study for Ernst & Young, used Washington's figures to estimate the compliance costs that remote sellers would incur for collecting and remitting the use tax in three cases: one state, 15 states, and 45 states plus the District of Columbia.(23) They concluded that compliance costs as a share of revenue collections would rise as the number of states in which a firm did business increased and the size of the firm decreased; their estimates of costs ranged from 1 percent of tax revenue for a large firm collecting taxes in one state to 48 percent or 87 percent of revenue for a medium-sized or small firm doing business in 45 states. Lorrie Jo Brown, the author of the Washington state study, has suggested that those adjustments to the Washington state estimates grossly overstate the likely compliance costs, but she has provided no comparable estimate for multistate retailers.(24)
Despite the lack of specific estimates, the available evidence taken as a whole suggests that the compliance costs from collecting use taxes will constitute a substantially greater burden for vendors' remote sales than the costs attached to local sales. Moreover, such costs figure prominently in the opposition to remote collection. Opponents of requiring vendors to collect use taxes on remote sales suggest that those increased costs would outweigh the reduction of distortion costs from eliminating differential taxation. Consequently, federal policy that is focused on reducing the social costs of taxation must consider compliance costs carefully, and state tax authorities must find a way to reduce those costs if they are to obtain the authority to collect taxes remotely.
Current Efforts to Reduce Compliance Costs
Federal legislation to permit states and local governments to collect use taxes on remote sales would impose high compliance costs on remote sellers, costs that would tend to offset the gains in economic efficiency from reducing distortion costs. However, states and local governments are working to lower those compliance costs, specifically through the Streamlined Sales Tax Project (SSTP), set up in 2000. More than 30 states have joined together to design a voluntary sales and use tax system. By encouraging businesses to voluntarily collect use taxes, the project may help states and local governments increase their revenue even in the absence of federal legislation. The system is intended to demonstrate how tax simplification, the adoption of computer technology, and compensation can reduce costs and thereby increase the likelihood of Congressional action to require remote collection.
By November 2002, 34 states and the District of Columbia had agreed on the administrative characteristics of such a system and submitted the Streamlined Sales and Use Tax Agreement (SSUTA) to states for adoption by their legislatures and governors.(25) The agreement was to go into effect for adopting states when it was enacted by at least 10 states that together constituted at least 20 percent of the total population of the states imposing a sales tax. Both thresholds were achieved by August 2003; at that point, 20 states had enacted legislation that aligned their sales tax systems with the provisions of the agreement.(26) Most of those changes are to take effect by July 1, 2004.
The states adopting the SSUTA will now review each state's law to ensure that it complies with the agreement. Once the 10-state and 20-percent-population thresholds have been verified, the agreement will go into effect, and a governing board with one vote per state will be established. That board, which will include participation by businesses, will respond to questions and resolve disputes, judge states' compliance with the agreement, and consider amendments to it.
Under the SSUTA, remote sellers' compliance costs would be reduced in several ways. The agreement would lessen the complexity and diversity of the sales tax structures faced by sellers by simplifying bases and rates and imposing uniform sourcing rules (for identifying the purchaser's residence). It would require the use of computer technology to simplify the process by which sales and use taxes are computed, remitted, and audited. And it would require states to compensate remote sellers for many of the remaining compliance costs.
Although states' compensation of retailers' costs for collecting sales and use taxes would remove a principal objection to requiring remote collection, compensation is not a substitute for the other two cost-reducing options. Instead of diminishing the resources devoted to collecting taxes that would result from improvements in technology and simplification, compensation would shift those costs from retailers and consumers to the state's taxpayers. From a federal policy perspective, compensation would not reduce the social costs of taxation.
Simplifying Sales Tax Bases and Rates and Handling Other Administrative Matters. The SSUTA would simplify but not standardize the structures of the various states' sales taxes, allowing considerable differences in tax bases among but not within the states that impose such taxes. Under the system, the states would jointly define the items included in major categories of goods and services (for example, food, clothing, automobiles, and service groups) that were subject to sales taxation, and each state would compose its base from among those categories. Local governments that levied sales taxes would have to use the same base as the state's. The result would be a system that was currently limited to 46 tax bases (assuming that all states with existing sales taxes plus the District of Columbia adopted it); the bases would differ by the sales categories they included, but for any given category, the items it covered would be identical among all the states.
Under the terms of the SSUTA, all items in a state's tax base would be subject to the same sales tax rate--with the exception of food and drugs, to which a different rate (possibly zero) might apply. The agreement would permit no more than one local rate, and all local sales taxes would be administered by the state. The SSUTA would establish uniform rules for the frequency of tax filings and for changes to tax bases and rates, and would simplify the procedures by which purchasers could obtain tax exemptions. It would also shift the responsibility for validating those exemptions from the seller to the state. Furthermore, the agreement would impose uniform sourcing rules that relied first on the destination of the good--the purchaser's location--which would be determined by using shipping instructions or an address gleaned from the seller's business records (including the address associated with the purchaser's payment instrument, such as a credit card). If that information was lacking, sourcing would revert to the origin of the good, determined by the address from which tangible property was shipped or, in the case of a digital good, the location at which it was first available for transmission. (Box 1 discusses origin-based versus destination-based sales taxes and the social costs of taxation.) Devising a reliable and simple system for sourcing digital goods remains a thorny issue for states and local governments (see Box 2).
Box 1. Origin-Based Versus Destination-Based Sales Tax Systems
Use taxes on remote sales are difficult to collect because sales and use tax systems are "destination based." That is, a state imposes its tax on the basis of the destination of the purchased item; if the purchaser resides in the state, the tax is incurred. By contrast, in an "origin-based" system, the tax is levied at the source or origin of the item being sold; if the seller is in the state, the tax is incurred. Thus, under an origin-based sales and use tax system, the difficulty of collecting taxes on remote sales does not arise because remote sales are not taxable.
Some observers have argued that shifting to an origin-based system could improve sales tax compliance.1For merchants, the advantage of an origin-based system is that retailers deal only with the sales tax system in the jurisdiction in which their firm is located. The advantage for states and local governments is that an origin-based system does not depend as heavily as a destination-based system does on the voluntary compliance of individuals; under an origin-based framework, all the use taxes paid under a destination-based system would be collected as sales taxes by vendors whose businesses would all lie within the taxing jurisdiction.
The effect on states' and local governments' revenue of moving to an origin-based system would vary by jurisdiction. A jurisdiction would lose the revenue from remote sales to its residents but gain the revenue from sales by its merchants to out-of-state buyers. As a result, implementing an origin-based system could increase interstate tax competition. Lower-tax states would have some advantage in attracting retailers who marketed their goods and services in other, higher-tax states. However, the net consequences for efficiency are not clear. Tax competition could tend to reduce jurisdictions' revenue from the sales tax relative to what they would derive from a destination-based system, an outcome that would affect either the composition of revenue or the jurisdiction's overall spending, or both.
Under an origin-based system, consumers could avoid sales taxes by buying from an out-of-state merchant. Merchants thus would have incentives to locate in places that allowed their customers to avoid the sales tax. The overall economic efficiency that resulted would depend on the efficiency of the sales tax relative to the efficiency of the kind of tax that replaced any reduction in sales tax revenue, and on the overall size of the public sector. How retailers reacted to the differences in taxes and how states and local governments responded to retailers' decisions about where to locate their businesses would help determine an origin-based system's impact in each state.
1. An origin-based system is discussed in Aaron Lukas, Tax Bytes: A Primer on the Taxation of Electronic Commerce (Washington, D.C.: Cato Institute, December 17, 1999), pp. 19-20; and Wade Anderson and Andrew Wagner, "Guidelines for Establishing an Origin-Based Sales Tax," State Tax Notes, vol. 18, no. 12 (March 20, 2000), pp. 915-918. The article by Anderson and Wagner suggests using an interstate revenue-sharing agreement to distribute the revenue collected in conformance with a destination-based tax. Box 2. Taxing Digital Goods
If states are given the authority to require remote sellers to collect and remit use taxes, each sale must be "sourced"; that is, the seller must determine where the good will be used. (Most often, in the case of a purchase by a consumer, that use will occur at the residence of the buyer.) For tangible goods, sourcing is a manageable problem. The good must be shipped to a location, which is a reasonable approximation of where it will be used, and the opportunities for businesses and consumers to behave in ways that minimize their taxes are not that great. A buyer located on the border of two states with very different tax rates might be able to have the purchase shipped to the lower-tax jurisdiction (using a post office box, for example, or the address of a friend or relative) and not incur costs for transportation that exceeded what he or she saved in taxes. That type of strategy would be particularly useful for expensive goods, but such opportunities are limited.
Digital goods are not subject to similar constraints. For example, Apple Music's iTunes Music Store does not collect use tax on its on-line sales of digitized music. Even if the sellers of such goods decided to collect it, buyers (particularly consumers, whose purchases are less easily tracked than businesses' purchases) could conceivably have the digital product "shipped" to a computer in any location and pay for the product with a credit card whose billing address listed a state without a sales tax. Indeed, the possibility of avoiding a destination-based tax on digital goods appears to be substantial. One approach proposed by the Streamlined Sales and Use Tax Agreement would be to revert to origin-based taxation (as described in Box 1, a tax assessed on the basis of the seller's location) in cases in which a shipping address and the customer's financial information were not adequate to determine the customer's residence for a transaction. Of course, the digital nature of the goods makes it entirely possible that the identifiable seller's location may be chosen to minimize sales and use taxation.
The taxing of digital goods thus presents difficulties, and no technical solution grounded in current technologies appears to be imminent. As a result, requiring remote sellers to collect and remit use taxes is likely to generate less revenue from digital goods than many people expect. To the extent that taxes on such goods were avoided or minimized, the gains and losses in efficiency discussed in this analysis would be moderated.
Simplifying the Process of Computing and Remitting Taxes. The very force that has increased uncollected use tax revenue from e-commerce--rapid technological growth in the computer industry--may also have created tools to reverse that trend. Software purchased by vendors or operated at off-site centralized computer centers could significantly reduce remote sellers' costs of coping with the complexity of numerous tax bases, rates, and auditing procedures. Those programs could determine, in real time (during the transaction), the sales tax status of a product on the basis of the state and local tax law in the purchaser's jurisdiction. Information on the amount of the tax and the jurisdiction to which it was to be remitted could be stored (again, on or off site) in an audit file set up for each remote seller.
Under the computer strategy spelled out in the SSUTA, remote sellers could choose to contract with a certified service provider (CSP) that would handle all the seller's sales and use tax functions, including filing all tax returns and being the focal point for audits. CSPs would be certified by the states and would have to use a certified automated system (CAS) for those functions. (The states would certify that the software met all requirements for identifying the taxable status of the purchaser, the tax base and rate, the good's source, and other necessary factors.) Remote sellers could also deal with their sales and use taxes on their own, using a CAS. Alternatively, sellers with sales in at least five states and totaling a minimum of $500 million could use their own system for that task, provided that it had been approved by the project's member states.
Compensation. Simplifying tax structures and adopting computer technology would each reduce compliance costs to some degree, although estimates of those reductions are not available.(27) Regardless, under the SSUTA, some compensation would be provided for those costs. A CSP's costs would be paid from the sales tax revenue it collected. In addition, the CSP would receive a bonus each year for two years based on a percentage of the tax revenue generated by a seller, with the expectation that the CSP would share that revenue with the seller as an inducement to participate in the system. Sellers who decided to handle their own sales and use tax functions by employing a CAS or proprietary software would receive a percentage of the revenue collected for a period of two years plus any compensation currently provided by the seller's state. After two years, compensation would come only from the seller's state.
An industry monitoring group, the Council on State Taxation (COST), has praised the SSUTA as a move in the right direction.(28) However, COST suggests that the agreement lacks enforcement provisions, is vague about how closely compensation would be tied to actual compliance costs, and is missing a federal monitoring mechanism that should be incorporated before the authority to require remote sellers to collect and remit use taxes is provided by the Congress.
Other Economic Issues in the Remote Collection Debate
Both opponents and proponents of remote collection cite additional economic considerations beyond distortion costs and administrative and compliance costs in their arguments--in particular, network externalities, the size of government, horizontal and vertical equity, unfair taxation of remote sellers, and states' fiscal autonomy. In general, however, reflection suggests that those issues are less central to the debate over remote collection of use taxes than are social costs.
Network Externalities and Market Failures
One argument advanced by opponents of remote collection is that excluding Internet sales from taxation provides a desirable stimulus to the growth of the Internet and e-commerce. Underlying that contention is the notion that the Internet is characterized by "externalities" that cause too few people to become part of the network and that leave it at less than its desirable (that is, efficient) size. Such a contention is consistent with economic theory but does not appear to be consistent with people's current use of the Internet. More than 136 million of the 241 million people in the United States above the age of eight (56 percent) used the Internet in 2001.(29) That degree of use suggests that the Internet and e-commerce have probably expanded beyond the network's critical-mass phase (see the discussion below).
Externalities occur when a transaction imposes economic effects on third parties that are not reflected in the price established between the buyer and seller of a good. Not taking those effects into account leads to the incorrect amount of the good being bought and sold--or, more accurately, to the incorrect amount of the external effect being provided. If there are positive externalities, an incentive that reduces the price of a good by an amount equal to the external benefit that society receives from the transaction will increase output to a level that more nearly reflects both the benefits received by private citizens (the buyers and sellers) and the external benefits received by all citizens. In that case, the market's failure to take all benefits into account would be corrected.
In the context of taxing purchases made over the Internet, opponents of such taxes have cited "network externalities" as part of the rationale for their position.(30) The Internet is, of course, a communications network and, like all such structures, tends to be characterized by network externalities; that is, when a person joins a network, he or she receives benefits from being able to communicate with its other members. If the network's services are provided competitively, the price that a new member pays accurately reflects the value of those benefits. But the other members of the network each receive a benefit as well--that of being able to communicate with the new participant. Because potential members do not take those network (external) benefits into account when deciding whether or not to join, the possibility arises that too few people will participate and the network will be too small. The market in that case is inefficient because the benefits derived from adding more members to the network still exceed the cost of the resources that would be expended to add them. Opponents of requiring remote sellers to collect use taxes argue that exempting e-commerce from that collection responsibility is desirable because it acts as an incentive that can help correct that inefficiency.
Any economic advantage from providing a tax-based subsidy, or preference (in this instance, the "preferential" treatment of not collecting use taxes), to e-commerce sales depends, among other things, on the extent of the network's development relative to its optimal size.(31) No network can exist until the price of joining it falls to a level that induces a critical mass of people to purchase the network's services. Those initial participants tend to be "high-demand individuals" who not only value the services highly but whose intense interest in the technological wizardry of the new network makes them provide high levels of benefits (in the form of advice and troubleshooting) to the network's other participants. As costs continue to fall over time, people with successively less demand for network services and possessing less technological expertise are induced to join. Those new entrants generate ever-smaller benefits for other participants, for two reasons.
First, the probability that existing network members will communicate with one more entrant decreases; even if each new member is just as technically competent as those who came before, the new entrant's value as a communications partner falls because so many others are already available. Second, the new entrant's decreased technological ability relative to previous entrants makes him or her less valuable as a source of advice. A network reaches an economically efficient size when the cost of joining exceeds the value placed on participating in the network by those who have yet to join. Those always-potential entrants would have provided relatively few benefits to other participants--and they might have imposed costs in the form of congestion on a network that had become stable in size.
An incentive during the early stages of the Internet's development and the rise of e-commerce could arguably have increased economic efficiency. But permanently exempting e-commerce from sales taxation when more than 50 percent of the population who conceivably might use the Internet already does so risks providing an incentive in exchange for few or no offsetting external benefits, thereby increasing rather than reducing a distortion.(32) Even if external benefits from new entrants remained significant, the object of an incentive would be to stimulate the growth of the network, not of e-commerce per se.(33)
Opponents of remote collection proffer another argument to support their position: that it is not the size of the network that needs to be subsidized but rather the e-commerce business model. In effect, the argument here is that e-commerce is an infant industry and needs to be subsidized to enable it to compete with traditional forms of commerce until its scale increases to a size that allows its production costs to decline enough to make it competitive. Yet the industry's sales exceeded $1 trillion in 2001 (including both business-to-business and business-to-consumer sales), and more and more so-called bricks-and-mortar firms are expanding their use of e-commerce to complement their traditional mode of selling. The e-commerce model thus hardly seems endangered. Furthermore, a standard problem with subsidizing an infant industry
About the Authors
Governmental Research
