Electronic commerce, state sales taxation, and intergovernmental fiscal relations.

"Our generation stands on the cusp of the greatest technological revolution that mankind has ever faced. Some compare this age of electronic communications with the arrival of the Gutenberg press, or with the industrial revolution.... Even our notions of sovereignty and governance could be profoundly

INTRODUCTION

Electronic commerce, which can usefully be defined as "the use of computer networks to facilitate transactions involving the production, distribution, and sale and delivery of goods and services in the marketplace,"(1) has burst on the fiscal scene almost as suddenly and as brilliantly as the Hale-Bopp comet.(2) It has created a flurry of activity among state tax administrators intent on sharing in the spoils.(3) This, in turn, has produced panic among potential taxpayers and their advocates--visions of doom reminiscent of the fears comets produced long ago. For example, Congressman Christopher Cox (1997) has said, "It is not hyperbole to say that taxes might drive the Internet to an early grave." The problem is not merely that electronic commerce will be subject to unreasonable taxation; part of the problem is uncertainty about how such commerce will be taxed. Senator Ron Wyden has said (1997b), "Allowing a chaotic, helter-skelter approach to taxing electronic commerce could harm thousands of businesses and millions of consumers" and (1997a) "The crazy quilt of state and local taxes could do irreparable harm to the Internet, killing the goose that could lay billions of dollars in golden eggs."

Unlike the comet, electronic commerce will not go away. Rather, to continue the analogy, electronic commerce is on a collision course with state tax systems, as we know them.(4) If we are lucky, the collision will bring about reforms of these systems that are long overdue. But these reforms may occur only in the context of a radical change in intergovernmental fiscal relations. Two developments underscore this fact: Cox and Wyden have proposed an "Internet Tax Freedom Act," federal legislation that would impose a moratorium on state and local taxation of the Internet, and the White House (1997) has included state and local taxation in its call for no new taxes on the Internet, after first being silent on the issue.

This paper discusses the implications of electronic commerce in the sales tax area. The second section contrasts the digital revolution that underlies electronic commerce with revolutions in the storage and transmittal of information that preceded it and draws implications for tax policy and administration. The third section describes in broad terms the current irrational "system" of state sales taxation and explains why a piecemeal approach that deals only with taxation of electronic commerce cannot produce a satisfactory solution. The fourth section proposes a rational solution to the problems we face. While this solution is quite unlike what prevails today, it is amazingly simple, at least in concept. But it may require federal legislation to authorize or mandate its implementation. The final section comments on possible effects on IFR.

The paper is essentially normative, not prophetic; it expresses strongly held views on what should be done, and why, not what will be done. If the intent were prophesy, the conclusion might be that we will muddle through, rather than adopting the reforms proposed here. But the prophesy would probably be quite fuzzy about how the muddling would be done.

PATTERNS OF COMMERCE

Six interrelated developments are crucial to the arguments presented here: (1) the increasing role of services and intangible products, (2) developments in technologies for storing and communicating information, (3) resulting changes in techniques of marketing, (4) the evolution of techniques for paying for mail-order and similar purchases, (5) changes in the structure of telecommunication industries, and (6) implications for taxation. While the third section describes the salient features and major defects of state sales taxes in some detail, the following brief summary of issues may be useful, so that the reader can see the forest while passing through the intervening trees.

Tax Issues Raised by Electronic Commerce

Current state sales taxes were designed for a world in which local merchants sell manufactured products. Thus, they apply primarily to tangible products; in general, they do not apply to services, except on a selective basis, or to intangible products. The first and most obvious issue in the debate over taxation of electronic commerce is whether sales taxes should apply to services and intangible products delivered over the Internet (and presumably to similar products delivered in other ways). If not, the burgeoning of electronic commerce raises a second question: how to distinguish taxable and exempt products in a world in which traditional distinctions are becoming blurred?

If the tax net is widened to include services and intangibles, a third question arises: whether these products should be exempt when bought by businesses? Historically, exemptions for business purchases have been limited primarily to tangible products bought for resale, for incorporation into goods produced for sale, or for direct use in production, but also fuels and electric utilities; purchases of many otherwise taxable products (e.g., office equipment) generally are not exempt. Given that most electronic commerce involves sales to business firms, this is an enormously important issue because of the risk of multiple taxation.

With the advent of mail-order sales came the possibility that significant amounts of sales would not be made by local merchants. The U.S. Supreme Court has ruled that states where purchasers are located cannot force out-of-state vendors to collect sales tax; nor, because of the administrative burden that would be involved, can they force such vendors to collect use taxes (a tax on the purchaser intended to compensate for the lost sales tax) due from their customers. This places local merchants at a competitive disadvantage, relative to mail-order firms, and raises two additional questions: whether interstate electronic commerce should benefit from the same protection as mail-order sales between states and whether the present protection of out-of-state vendors should be restricted, perhaps by exempting only those whose sales do not exceed a de minimis amount.

The Increasing Role of Services and Intangible Products

Throughout most of history, commerce has involved primarily trade in tangible products. This characterization has accurately described business purchases, as well as household consumption. The essentials of life have been tangible products such as food, water, clothing, and housing.(5) (Only economists talk about the "services" of housing.) Similarly, while businesses used the "services" of employees, they purchased relatively few services from other businesses. (Transportation was among the notable exceptions.) Economic production and distribution involved primarily tangible products, not "goods and services." Only as we have moved past the industrial age and into the age of the service and information economy has this characterization become invalid.

The discovery of ways to generate, transmit, and use electricity added an important and radically new element to the market basket of households and the input matrix of businesses--an intangible product. (Before then, the term "intangible product" would have been essentially an oxymoron, unless it was merely another name for services.) It also underlay the development of dramatic improvements in the capacity to store and transmit information. Electronic transmission of information via telephone, radio, and television also involved intangible products.

Storage and Transmittal of Information

Through much of history, information could not easily be stored or transported. Thus, information-intensive services were consumed when and where they were produced (e.g., as lectures, theater performances and concerts, sporting events, visits to the doctor, or tourism). Discovery of efficient methods of storing and transmitting information has changed the nature of commerce in information, and thus in many services. It has also had important impacts on techniques of marketing and paying for products. The following describes some of these developments, concentrating on those of particular relevance for sales taxation.

Printing, photography, and cinematography

The printing press made the publication and mass distribution of books, newspapers, and magazines possible. Photography made it possible to publish pictures, as well as text and drawings, Various types of information-intensive services (e.g., lectures, storytelling, medical expertise, and various types of advice) could thus be incorporated in tangible media, and consumption of information could be separated from production in both time and place.(6) Cinematography added another element, storage of moving pictures, creating an entirely new form of entertainment. (But for "movies" to become "talkies" required recorded sound, considered below.)

Analog transmission

The successive invention of the telegraph, telephone, radio, and television each added new and important dimensions to the separation of the production and distribution of information.(7) The telegraph introduced electronic--and thus more rapid--transmission, and the telephone added transmission of sounds. Whereas telegraph and telephone service was limited by the use of copper wire and the need for fixed locations, radio and television receivers took broadcasts from the air and were much more mobile. While the telephone was used until recently primarily for transmittal of conversation between two parties, radio and television, which added transmission of images, were used primarily for entertainment. They allowed entertainment produced in one place to be consumed elsewhere--that is, the transportation of a service (entertainment) in the form of an intangible product.

Analog recording

The capacity to record electronic signals allowed the production and distribution of sounds and images to be separated in time--and in space, since recordings could be transported. Thus, radio and television no longer needed to be "live," and they did not need to be heard or watched in "real time." Moreover, sounds recorded on phonograph records, audio tapes, and compact discs and images recorded on video tapes and disks could be played back and experienced independently of radio and television transmission.

Digitization

The digital revolution is making possible the recording and high-speed interactive transmission of high-quality sounds and images in multimedia form (that is, inter alia., as text, graphics, pictures, video, voice, and music).(8) This has had--or will increasingly have--several dramatic effects on commerce. First, there will be substantial electronic commerce in content--digitized images and sounds.(9) Rather than buying content in tangible forms, for example, as compact disks, video tapes, games, shrink-wrapped software, or even printed matter, consumers will be able to download it from the Internet in an intangible form. Thus, boundaries between tangible products, intangible products, and services are becoming blurred. As Negroponte (1996, p. 71) states, "The medium is not the message in a digital world." Moreover, vendors will use the Internet, especially the World Wide Web ("the Web"), to sell tangible products. Finally, the structure of telecommunications industries is changing rapidly. All these developments have important implications for taxation.

Summary statement

Developments in the storage and transmission of information have led to the creation of large and important new industries. Among the obvious examples are printing and publishing, advertising, telecommunications (including radio, television, and, more recently, the Internet), motion pictures, the recording industry, professional sports, the manufacture of computers and peripherals, and software development. Such developments will also cause a decline in some industries (for example, musical instruction) and dramatic changes in others (for example, publishing).

Techniques of Marketing

Throughout history, most sales to households have been made by local merchants. (By comparison, businesses have bought a higher percentage of their purchases from vendors lacking a local physical presence.) Cheap transportation and developments in the storage and transmission of information have led to the geographic expansion of markets and have facilitated what has recently been called "disintermediation," the elimination of middlemen (wholesalers and retailers), many of them local, from the distribution chain that links producers and consumers.(10) Table 1 summarizes the salient features of the developments described in the remainder of this subsection.

TABLE 1
SALIENT FEATURES OF PROTOTYPICAL TYPES OF COMMERCE

                                                        Role of
Activity                   Product      Publicity       Customer

1. Traditional commerce:
   a. Retailing            tangible     local           active
   b. Services             services     local           active
2. Disintermediated
   commerce:
   a. Mail order           tangible     mail            passive

   b. TV/Telemarketing     tangible     TV/phone        passive
   c. Web order            tangible     on-line         active
   d. Electric commerce
      in content           intangible   on-line         active

Activity                   Purchase         Payment

1. Traditional commerce:
   a. Retailing            face-to-face     face-to-face
   b. Services             face-to-face     face-to-face
2. Disintermediated
   commerce:
   a. Mail order           mail/phone/fax   COD/MO
                                            check/card
   b. TV/Telemarketing     mail/phone/fax   check/card

   c. Web order            phone/on-line    on-line
   d. Electric commerce
      in content           on-line          on-line

Activity                   Delivery

1. Traditional commerce:
   a. Retailing            face-to-face
   b. Services             face-to-face
2. Disintermediated
   commerce:
   a. Mail order           common carrier

   b. TV/Telemarketing     common carrier

   c. Web order            common carrier
   d. Electric commerce
      in content           on-line

Key to abbreviations: COD--cash on delivery; MO--money order; card-credit or debit card; and italics--items that differ from earlier practice.

Mail-order

The possibility of using print media in marketing, via mail-order catalogs, began the process of disintermediation. This process has been encouraged by public subsidies (postage rates for mailing catalogs that do not cover postal costs) and, more recently, by the favorable state sales tax treatment of interstate sales described below. Much of what is bought by mail order is delivered by the U.S. Postal Service or by common carriers, such as United Parcel Service or Federal Express.(11)

Electronic marketing

Telemarketing and TV shopping channels use electronic media to extend the traditional mail-order business in tangible products and have led to further disintermediation.

Web-Order(12)

The development of the World Wide Web brings several new dimensions to the disintermediation that began with mail-order marketing and continued with telemarketing and TV shopping. Vendors of tangible products, including manufacturers, can establish Web sites on which they advertise their products, providing more information than is easily included in standard printed catalogs (for example, information that is updated frequently, reviews by other customers, and hyperlinks that allow shoppers to "click" on products of related interest) and allowing customers to place orders simply by selecting items from an on-screen menu and providing credit card information. Among the tangible products that are successfully being sold over the Internet are computers and related equipment, software, books, automobiles, and wine.(13) While the best known Web sites of this type are those of large vendors, this technology makes it possible for small merchants to gain access to niche markets (for example, for wine) nationwide--indeed, worldwide--at very low cost, something that would be impossible using any prior technology.

Electronic commerce in content

Electronic commerce comes into its own with commerce in content (intangible products and services). Most obvious, the convenience of purchasing products such as music, videos, games, and software in intangible form, by downloading them from the Internet, threatens to raise disintermediation to a new level. Negroponte (1996, p. 173) predicts that "videocassette-rental stores will go out of business in less than ten years." The publishing and recording industries are also likely to be adversely affected, or at least changed dramatically. Moreover, it is (or will be) possible to transmit certain services electronically (e.g., those traditionally performed by travel agents); surgeons using remote robots to perform operations is an oft-cited and dramatic example. While most of the attention in the popular press focuses on applications such as these, it should be recognized that most electronic commerce in content involves business applications. The majority of that is conducted over proprietary networks, commonly called intranets. This creates an enormous risk of double taxation (see the third section).

Methods of Procurement and Payment

Throughout much of history, most commerce has involved face-to-face exchange of goods and services for cash or credit extended by the seller (or, more recently, for checks written by customers known to the seller). The development of mail order sales required innovations in techniques of paying for purchases; one ordinarily would not send cash through the mail or extend credit to a stranger. "Cash on delivery" (COD) was one early standard; vendors relied on the good faith of the purchaser, and both vendors and buyers trusted the U.S. Post Office to remit funds entrusted to it. Another technique, the use of money orders, required that buyers trust sellers and the issuer of the money order. Personal checks from unknown customers could be held for payment. More recently, credit cards have become the dominant means of paying for mail-order merchandise. Credit card information is commonly transmitted with mail-order purchases and their analog electronic counterparts (telemarketing and TV shopping). Debit cards are likely to replace credit cards, because they eliminate the risk of loss from bad debts.

Web-order and electronic commerce in content provide the possibility of variations on these themes. First, those who shop on the Web can transmit credit card information with their order. While some may prefer to use more traditional means, such as phone and fax, because of fears of credit card fraud (e.g., by hackers searching the Web for card numbers), it appears that encryption technology will soon render these fears groundless. Second, some purchases of electronic content may not lend themselves to traditional means of payment, either because of the time required for approval of credit card purchases or because transactions are so small that credit/debit card charges are not cost-effective. Among the ways these problems may be solved are the development of "smart cards" (cards with magnetic stripes that can "contain" money, similar to prepaid telephone calling cards), "electronic cash" to be used in small transactions, "bundling" of content with Internet access so it can be covered by a fixed fee, and using advertising revenues to cover the cost of content. While bundling creates problems for a tax system that attempts to differentiate between taxable communications services and exempt content, it facilitates uniform taxation of all consumer purchases on a destination basis and exemption of all business purchases.

One of the key features of the Internet is that its users do not have a readily identified location. For example, someone resident in one state may log on to the Internet via a server located in another state or (especially as the cost of telecommunications continues to fall) even in another country. Thus, it may be virtually impossible, for technical reasons, to apply the sales tax of the state of residence to purchases of electronic content. (Doing so might also require federal legislation, as discussed below.) Hellerstein (1997) proposes to address this problem by assuming such sales are made in the state where the customer's mailing address is located. This could, of course, create an opportunity for abuse through the use of fake mailing addresses. Moreover, both encryption and electronic cash potentially pose further problems for tax administration. Electronic cash makes it possible to make purchases that are difficult to trace--in effect eliminating mailing addresses. It would be virtually impossible to collect tax on transmission of content (especially if sent from abroad and paid for with digital cash, a smart card, or encrypted credit card). Hammond (1996, p. 165) writes, "if the use of anonymous digital cash and encrypted communications reaches critical mass on the Net, it has the potential to cripple and ultimately destroy the notion of both local and national taxation." The Economist (1997, p. 13) notes, however, that, "For all the beauty of digital cash, it has won precious few takers."

Effects on Telecommunications

Concurrent with the digital revolution in marketing described above have been revolutionary developments in the structure of several industries that transmit information electronically.

The revolution in telecommunications

The telephone industry, long seen as a "natural monopoly" for technological reasons, was traditionally organized as a heavily regulated or state-owned monopoly in most countries. Where in private hands, as in the United States, it was commonly taxed separately from other industries, often on the grounds that it enjoyed special privileges, such as the use of local right-of-way and a monopoly position. More recently, there has been a revolution in the way telecommunications is structured in many countries, including the United States. First, many state-owned monopolies have been privatized. Second, monopoly structures have been replaced by competitive (or at least oligopolistic) structures. Third, there have been international agreements to allow carriers to enter foreign markets. Satellite, microwave, and cellular technology have contributed to the last two developments. Fourth, we now see the beginning of the use of the Internet to transmit telephone calls. This development will further increase competition and dramatically reduce the cost of telephone service, especially internationally. Finally, telephone service no longer involves merely "basic" service, the "vanilla" transmission of messages. Increasingly, it includes "enhanced" services, such as paging and voice mail.

Television via cable/direct broadcast satellites

There have been equally far-reaching developments in the television industry. Initially, both radio and television were broadcast to listeners and viewers. This severely limited the number of available channels or stations in a given area, especially for television. Moreover, because of an important "public goods" characteristic of broadcasting--the inability to exclude those who do not pay--radio and television broadcasts were traditionally financed by advertising, by public subsidies, or by contributions, and not by charging listeners or viewers.(14) Thus, they could not be subject to sales taxation. More recently, the use of cables and direct broadcast satellites (DBS) and the development of technology for scrambling and descrambling signals have created the possibility of many more TV channels and have allowed charging for the receipt of signal, creating the phenomenon of "pay-per-view" and opening the technological possibility of taxing cable and satellite service.(15)

Internet access

Most users gain access to the Internet through either Internet Service Providers (ISPs) or On-line Service Providers (OSPs). While distinctions are becoming blurred, the former are analogous to common carriers, simply using leased capacity to transmit signals for others, while the latter use proprietary networks and provide electronic content (e.g., databases) and some services (e.g., billing) for their clients. The ISPs tend to be smaller and more numerous than OSPs, but consolidation of ISPs may render that generalization invalid.

Implications for Tax Policy and Administration

The digital revolution is having dramatic effects on taxation, some of which have been noted already and some of which are described in the next section.

It is relatively easy to levy sales taxes on local merchants and for merchants to comply with the tax laws of the states where they operate, even if such laws differ from state to state. But when vendors operate across state borders, both compliance and administration become more difficult, especially if state tax laws are not uniform. Thus, even in principle, the process of disintermediation described above raises a difficult policy problem: how to subject interstate sales to tax in the state where the consumer resides, without creating overwhelming compliance problems. In the American context, there is an important additional issue: the constitutionality of taxing sales by out-of-state vendors. The next two sections describe this problem in greater detail and offer a solution to it.

The digital revolution raises new issues that are worth noting. One aspect of predigital revolutions involved storage and transmittal of information in tangible forms (for example, as printed matter, phonograph records, audio tapes, compact disks, and video tapes). Sales taxes on tangible products would seemingly apply to such products. But the digital revolution blurs distinctions between tangible products, services, and intangible products. For example, is a floppy disk containing software a tangible product (the disk), a service (the services of the programmer), or an intangible product (the software)? What if the software is downloaded from the Internet, so there is no tangible medium? Does it matter if the software is "canned" or is custom-made for the customer? While generalization is risky, it seems that most states tax canned (shrink-wrapped) software as a tangible product, but exempt custom software as intangible property. See Frieden and Porter (1996). State sales taxes draw distinctions that will become increasingly untenable.

Whereas it was relatively simple and perhaps reasonable, if not defensible on policy grounds, to subject telephone service to special levies in the days when the industry was well-defined and a monopoly, the revolution in telecommunications makes such a policy quite indefensible. First, it is impossible to separate basic telecommunications from enhanced service, which might be taxed separately and perhaps not at all. Bundling of Internet access and content raises similar problems, unless both are taxed in the same way. Second, under current laws, it is likely that telecommunications services provided to OSPs and ISPs will be taxed twice, once as telecommunications and again as part of Internet access. (See also the discussion of business exemptions below.)

THE CURRENT SALES TAX

Current state sales taxes differ markedly from the type of tax that most tax economists would probably regard as appropriate. This section describes a benchmark against which to evaluate these taxes, describes the major deviations from the benchmark tax structure, and explains why a piecemeal approach that deals only with electronic commerce will not work.

The Benchmark: A Uniform Consumption-Based, Destination-Principle Sales Tax

A conceptually ideal sales tax has several dimensions.(16) Each is considered in turn.

Consumption basis

Ideally, sales taxation would apply only to sales to final consumers. In theory, this objective can be achieved in several ways. One is to exempt all purchases by businesses, taxing only sales to households. Another is to tax all sales, but allow business taxpayers credits for tax paid on their purchases, as under the credit method value-added tax (VAT) used in almost all advanced countries and many important developing countries.(17) Under a subtraction-method VAT, firms would be allowed a deduction for all purchases in calculating taxable value added. Failure to allow exemptions, credits, or deductions for business purchases produces a gross receipts or turnover tax. A primary advantage of the VAT is that it is administratively superior to a retail sales tax (RST) as a means of eliminating tax on business purchases. Whereas the VAT operates by allowing credits for tax on all such purchases (or deductions for all such purchases), the RST must implement exemption at the point of sale. It is often difficult to differentiate sales to businesses and sales to consumers.

Uniformity

Ideally, all consumption by households would be taxed at a uniform rate.(18) If exemptions were thought justified on distributional grounds, despite their manifest administrative disadvantages and adverse effects on horizontal equity and on revenues or tax rates, eligibility should be based on the relative importance of various items in the market basket of the poor.(19) If business purchases are taxed, consumer products will not be taxed uniformly.

Destination basis

In principle, taxation can consistently treat trade between jurisdictions in either of two ways: products can be taxed according to the destination principle, by taxing imports and exempting exports, or they can be taxed under the origin principle, by taxing exports but not imports. Stated differently, and assuming a consumption-based tax, a destination-based sales tax is a tax on consumption occurring in the taxing jurisdiction, while an origin-based tax is a tax on the jurisdiction's production of consumer goods. While arguments can be made for either principle, I believe the strongest justifies reliance on the destination principle. In this view, private consumption is a better proxy for consumption of the benefits of public services than is production. This probably also accords with most people's view of fairness. Certainly, most sophisticated sales tax systems are based on the destination principle. If business products are taxed, it,is impossible to implement the destination principle, which requires a rebate of the tax on exports and collection of an equivalent tax on imports.

Deviations from the Benchmark

The sales taxes of the American states deviate from the benchmark sales tax in several important ways.(20)

Business exemptions

State sales taxes do not allow exemptions for all purchases by businesses. All states allow exemptions for goods bought for resale. Beyond that, practice is quite heterogeneous. (I describe only the general patterns; for details, see Due and Mikesell, 1994; for a discussion that focuses on the tax treatment of electronic content, see Frieden and Porter, 1996.) Most states exempt sales of tangible products that will be incorporated in goods to be sold; many exempt those for "direct use" in production of such products; and exemptions for fuels and electric power are common. By comparison, few states allow exemption of business purchases of products that are not directly used in production or of services that are otherwise subject to tax.

Taxation of services and intangible products

Most states tax most tangible products bought by households (and many bought by businesses). By comparison, few apply comprehensive taxation to services, and there is hardly any taxation of intangible products, aside from electricity, telephone, and cable TV service. This last group of products has traditionally been subject to separate taxation. Thus, state sales taxes generally fail miserably to meet the objective of taxing all consumption uniformly. Exemption of services and intangible products, however, does have the advantage that at least these products are not subject to double taxation when bought by business.

Taxation of interstate trade

In general, state retail sales taxes correspond to the destination principle; they exempt out-of-state sales, but are applied to retail sales, whether the taxed item is produced in the taxing state or elsewhere. There is, however, a glaring exception to this generalization (in addition to the fact that RST on business purchases is embodied in the price of many sales made to out-of-state buyers). The U.S. Supreme Court has ruled that, under the Commerce Clause of the U.S. Constitution, states cannot impose sales taxes on out-of-state vendors and cannot require such vendors to collect use tax, unless they have a physical presence in the state: see National Bellas Hess (386 U.S. 753, 1967) and Quill (504 U.S. 298, 1992). This seemingly idiotic ruling has been justified by the massive compliance burden that out-of-state vendors would face if they were forced to comply with the extremely complicated and heterogeneous tax laws of all states and localities in which they might potentially do business. Thus, in the interest of protecting interstate trade, the Court has placed local merchants at a distinct disadvantage, relative to out-of-state competitors.(21)

Implications for Electronic Commerce

The (sales) taxation of electronic commerce is subject to substantial uncertainty. Some ways of reducing the uncertainty would adversely affect the achievement of the goals outlined above. It is unlikely that a piecemeal solution that deals only with electronic commerce will produce satisfactory results.

Uniformity

Some favor complete sales tax exemption for electronic commerce in content, and perhaps what I have called Web order as well. This clearly would violate the principle that all sales to consumers should be treated in the same way.

The digital revolution is blurring the lines between tangible products (usually taxed), services (sometimes taxed, but commonly not), and intangible products (virtually never taxed). The possibility of providing essentially the same product in more than one way increases the likelihood that the artificial distinctions currently found in state tax laws will become increasingly troublesome and untenable. What is needed is more uniform taxation of such sales, not more exemptions.

Business exemptions

To the extent that taxation is extended to more services and to intangible products, a policy many state tax authorities favor, multiple taxation will increase, since many such services and products would not be eligible for business exemptions, as currently structured. Such a development would move the RST even further from the conceptual ideal of a consumption-based tax. This risk is particularly great in the case of providers of Internet access, which might not be allowed resale exemptions for their bulk purchases of telecommunications services or business use exemptions for their purchases of computers and related equipment for use as servers. All purchases by business should be exempt.

Interstate commerce

Whether and how the Supreme Court would apply the "bright-line physical presence" test of Quill to interstate electronic commerce in services and intangible products is unclear.(22) One of the rationales used to justify the Quill decision, reliance of a substantial mail-order industry on the Court's prior decision in National Bellas Hess, arguably is not applicable to electronic commerce, especially that in content. Two recent court cases, Goldberg (488 U.S. 252, 1989) and Jefferson Lines (115 S. Ct. 1331, 1995), suggest that the Court might take a different tack, applying what amounts to an "economic presence" test of nexus to electronic commerce in content. That is, the Court may find a firm to have nexus in a state where it provides services, if the services are billed to an address in that state. Either way, the outcome is not likely to be satisfactory, unless legislative action is taken to rationalize the taxation of interstate commerce. If Quill is applied to electronic commerce, the results could be devastating for local merchants. If it is not, vendors' need to comply with the diverse tax laws of all states where they have customers could seriously hinder the growth of electronic commerce.

Electronic commerce raises an interesting twist in this area that shows why many engaged in electronic commerce are clamoring for protective federal legislation. The Supreme Court has ruled that a vendor can have taxable nexus in a state where it uses agents, for example, to make sales or provide after-sales service to its customers. Some have argued that OSPs, and perhaps even ISPs, act as agents for those whose Web sites they carry, especially if they provide services such as billing. See Summers (1996) and MTC (1995). Under this view, vendors would have nexus in any state where their Internet access provider has a physical presence. Further, it is argued that telecommunications firms, which undoubtedly have physical presence in most states, act as agents for ISPs and OSPs that use their facilities. This reasoning implies that virtually all vendors engaged in commerce over the Internet would have nexus in all states. See Grierson (1996a, b). Others argue that this carries the theory of agency too far--that providers of Internet access, not to mention telecommunications firms, are more like common carriers. It is unclear how successful the states will be in applying concepts of agency to electronic commerce.

A RATIONAL SOLUTION

Most of the elements of a rational solution to the taxation of electronic commerce are implicit in the description of the benchmark sales tax at the beginning of the previous section.(23) First, all distinctions between tangible products, services, intangible products, telecommunications, and Internet access should be abolished. They lack both economic substance and basis in sound policy, especially in the world of electronic commerce, and they create administrative headaches, economic distortions, and inequities. If a sale is to a consumer, it should be taxed.

Second, in principle, all sales to business should be exempt. The only exception to this rule would be motivated by practical considerations, namely, whether it is possible to achieve the conceptual objective without opening the way for abuse. (Thus, for example, it might be appropriate to define some types of software as never exempt, even if bought by a business, and some always exempt, even if bought by a consumer, in order to simplify implementation of an inevitably difficult principle.) With these two changes, it would be relatively simple to have a uniform definition of the sales tax base throughout the country.

Third, in principle, sales by out-of-state vendors should be subject to the same sales tax as sales by local merchants.(24) To avoid unacceptable compliance burdens, there should be a safe-harbor exception for vendors who lack physical presence in the taxing state and also have sales in the state that fall below a de minimis amount. (Thus, those lacking physical presence in the state and making only minimal sales would not be subject to use tax, but large mail-order firms would be subject to tax in all states where sales exceed the statutory minimum. The question of agency would become largely irrelevant in such a regime, providing it were not used to pull small out-of-state vendors into the tax net.) In addition, it would be desirable to have exemption certificates for business purchasers that would be uniform and recognized throughout the country. It would also be useful to have a minimum sales tax on non-nexus sales, in order to prevent them from going untaxed. Salient features of existing law and of these proposals are summarized in Table 2.

TABLE 2
NEXUS RULES AND TAX TREATMENT. PRESENT AND PROPOSED LAW(a)

                              Present Law            Proposed Law

Tangible products
  Taxable nexus        Quill (physical presence)   physical presence
                                                   or sales > de
                                                   minimis
  Sales to households  taxable                     taxable
  Sales to business    varies                      exempt

Intangible products
  Taxable nexus        unclear                     physical presence
                                                   or sales > de
                                                   minimis
  Sales to households  generally exempt            taxable
  Sales to business    generally exempt            exempt

Services
  Taxable nexus        unclear                     physical presence
                                                   or sales > de
                                                   minimis
  Sales to households  selectively taxed           taxable
  Sales to business    selectively taxed           exempt

Telecommunications
  Nexus                presumably                  presumably
  Sales to households  general/selective           taxable
  Sales to business    general/selective           exempt

OSPs and ISPs
  Nexus                unclear                     physical presence
                                                   or sales > de
                                                   minimis
  Sales to households  varies                      taxable
  Sales to business    varies                      exempt

Salient features       discrimination,             uniformity,
                       distortions,                equity, and
                       inequities, and             simplicity
                       complexity

(a) States exempt Various household purchases; I do not address this issue.

It does not appear to be feasible to craft a satisfactory solution that deals only with electronic commerce and leaves the rest of the irrational system of sales taxation intact.(25) First, exemption of all sales made over the Internet would be highly non-neutral and unfair; such sales should be taxed when made to individuals. On the other hand, it would be inappropriate to tax all sales made over the Internet; such sales should be exempt when made to business--a result not likely to occur unless rules for business exemptions are radically revised. Finally, it clearly would be inadequate simply to extend the bright-line physical presence test of Quill to electronic commerce. It is the wrong answer, because it aggravates the present discrimination against local vendors, and it would leave unanswered troubling questions of interpretation, such as the meaning of agency in an electronic world. But it is equally inappropriate not to provide some protection for small out-of-state vendors. The appropriate response is a de minimis test, not the blanket exemption of Quill.

IMPLICATIONS FOR INTERGOVERNMENTAL FISCAL RELATIONS

A satisfactory solution to the problems described here is likely to require wrenching changes in intergovernmental fiscal relations in the United States. As the passage from OECD (1997, p. 3) quoted above says more generally of the effects of electronic commerce, "our notions of sovereignty and governance could be profoundly affected." This section comments on this issue, first at a theoretical level and then at a practical political level.

The theory of tax assignment

The conventional wisdom on tax assignment assigns subnational governments (the state level in the United States) a prominent role in sales taxation. See Musgrave (1983). This simple answer is not adequate, however, even in a world without electronic commerce, because there is no guarantee that, left to their own devices, the states will adopt sufficiently uniform and appropriate definitions of the tax base to avoid serious distortions of interstate commerce. As the European Common Market, the predecessor of the European Union, understood long ago, for a common market to function, there must be at least a minimal level of coordination of sales taxation. Electronic commerce makes this even more abundantly clear.

I find it useful to decompose tax assignment into four subsidiary questions: which taxes to levy, the definition of tax bases, the setting of tax rates, and tax administration.(26) The setting of rates is clearly the most important of these in guaranteeing the fiscal autonomy of subnational governments. By comparison, total subnational autonomy in the choice of tax base and tax administration can cause needless complexity, duplication of compliance and administrative efforts, economic distortions, and inequities. It would be desirable to have greater uniformity in sales tax bases, even at the cost of some loss of state fiscal autonomy.

The sales tax regime applied to interstate mail-order sales is further evidence that simply assigning sales taxation to subnational governments may not work. The U.S. Supreme Court had no choice but to impose the bright-line physical presence test of Quill, because the states had failed to act responsibly by providing neither sufficient uniformity nor de minimis exemptions for out-of-state vendors. Since the mail-order firms now have the legal high ground, federal legislation may now be the only way to link taxable nexus to economic presence; current state attempts to ignore or wriggle around Quill are clearly not the way to solve the problem.

Political musings

Two recent developments are quite extraordinary. First, in its 1997 paper on electronic commerce, the White House urged that state and local governments impose no new taxes on the Internet--after first releasing a draft in November 1996 that made no mention of subnational taxes. Second, the Cox/Wyden "Internet Tax Freedom Act," which would impose a moratorium on state and local taxes on the Internet, had more than 50 cosponsors when it was introduced in the House. This is rather remarkable, in light of the historic reluctance of the U.S. Congress to exercise its constitutional prerogative to regulate interstate and foreign commerce or to impose restrictions on the taxing powers of the states.(27) The Clinton administration has testified in favor of the principles underlying the Cox/Wyden bill. See Summers (1997).

Several scenarios are possible. The states may be able to ignore the threat of restrictive federal legislation and go their own way; frankly, I doubt it. While the states may stave off the Cox/Wyden initiative, they can hardly ignore it or the concerns it signifies. Alternatively, a federal mandate for greater uniformity may be thrust upon the states; that also seems relatively unlikely. As Fox and Murray (1997) opine, "The general absence of such sweeping intervention in the past, coupled with questions of federalism and states' rights, will likely forestall such initiatives." What seems more likely is that the states will respond to pressure for federal legislation by moving at least partway toward a more rational system (and thus to muddle through).

Some will decry any reduction in the taxing power of the states as an usurpation of power by the federal government. This seems to be based on a misreading of both the U.S. Constitution and economic reality. First, the Founding Fathers were clearly concerned that the states' pursuit of parochial interests could undermine freedom of internal trade. Recent developments in the taxation of electronic commerce provide ample evidence that these fears were justified. As Senator Ron Wyden (1997a) has stated, "If there was ever a need to invoke the Interstate Commerce Clause of the Constitution, this is it."

Second, the type of anarchic system described above clearly is not appropriate for the information age; the individual states can no longer go their own ways, disregarding effects on a national marketplace that is becoming more and more closely integrated each day. In an earlier, primarily industrial, age, when there was relatively little disintermediation, it might have been satisfactory to allow individual states to define sales taxes as they wished. Differential treatment of tangible products, services, intangible services, and telecommunications was tolerable, if not sensible, and exemptions for resale, component parts, direct use, and fuels might have been adequate to avoid gross double taxation. This is no longer true. With the blurring of boundaries between industries and the rise of nationwide markets, artificial distinctions in tax treatment of similar products and multiple taxation of business inputs are becoming increasingly distortionary and unacceptable. There is thus much to be said for the position of Dean Andal (1997), a member of California's State Board of Equalization:

The lack of a uniform resolution to the

taxation of electronic commerce is a substantial

impediment to the growth of

electronic commerce. Congress must act,

as is should have long ago, to clearly identify

the boundaries of state taxation of

electronic commerce....The global implications

of electronic commerce raise

just the type of dilemma the Founding

Fathers envisioned being resolved by Congress.

Of course, as emphasized above, this sentiment has greater applicability; it generally is not possible to deal adequately with the problem by restricting attention to electronic commerce.

A final note

I am not naive enough to believe that the U.S. Congress or the states will enact the type of system I have proposed. Some might thus say it is best to leave well enough alone. But the present "system" is not "well enough"; it is very nearly a disaster--and can only get worse. Thus, I am willing to bet that the threat of a roll of the legislative dice--and federal legislation, if necessary--will make things better.

ENDNOTES

This paper draws heavily on ideas first developed for presentation, and presented more fully, in McLure (forthcoming, 1997a), which provides more extensive references to literature. The author wishes to thank Deborah Schenk for asking him to write that paper and Joel Slemrod for asking him to write this one. Errors are, of course, the author's responsibility.

(1) This definition, from Abrams and Doernberg 1997), seems more useful than the definition in U.S. Department of Treasury (1996), the crucial part of which is, "... the exchange of goods or services ... using electronic tools and techniques." Abrams and Doernberg (1997) provide a particularly useful background for the issues discussed here. On electronic commerce in general, see also The Economist (1997), Gates (1996), Hammond (1996), Negroponte (1996), and Tapscott (1996).

(2) This is evidenced by the fact that so many references to the literature carry dates of 1996, 1997, or "forthcoming."

(3) Wyden (1997a) refers to "... state and local tax authorities chomping at the bit to milk the Internet cash cow...." Cox (1997) provides a description of some of the developments in this area.

(4) Newman (1995) uses a different analogy, "Road Kill on the Information Superhighway," and Murray (1997) asks, "Will Technology Break the Back of the Sales Tax?"

(5) While moderns may consider health care and education to be essential and entertainment almost so, earlier generations did not, in part because these services were not available on the present scale and in part because ordinary people could not have afforded them. Health care as we know it could not have been "essential," because it did not exist; often, when people got sick, they died.

(6) We are accustomed to thinking of the growth of the service economy as a relatively recent phenomenon. In fact, its birth might better be dated from Gutenberg's invention of movable print, which allowed both the ubiquitous separation of consumption and production of information-intensive services and the mass consumption of "imprinted" services. Since services embodied in print take a tangible form, many have generally not traditionally been classified as services. But consider such obvious examples as arm-chair tourism, books on self-improvement, and "how-to" books and magazines on an enormous range of subjects; all can be considered to embody services of one type or another.

(7) I do not discuss the development of fax machines, which also involve digital transmission of images, since it is a technological cul-de-sac--what Negroponte (1996, p. 187) calls "a serious blemish on the information landscape, a step backwards." Since fax messages cannot be read by computers, they cannot be incorporated in multimedia applications.

(8) The signals involved are called digital, because they are composed of ones and zeros (representing "on" and "off"). For an elementary exposition, see, for example, Abrams and Doernberg (1997), which draws on Gates (1996). Analog devices transmit, receive, record, or play back signals represented by electronic impulses that induce magnets to cause diaphragms to vibrate. During the analog age, modems ("modulator-demodulators") were invented to allow digital signals to be sent over analog communications media. Electronic data interchange was an early forerunner of electronic commerce; its best known use is in retailing, where it is used to link inventory management with automated purchasing.

(9) There are presently limits on the feasibility of electronic transmittal of content such as movies that require large amounts of "bandwidth," the carrying capacity of the transmission medium. Since it is expected that these will soon be overcome, I do not dwell on them. On this, see Negroponte (1996).

(10) Disintermediation can occur locally or it can transcend political boundaries. Because of U.S. constitutional law, it is primarily the latter that interests us (see the third section). These developments have also created intermediation. For example, information created by storytellers or musicians can be provided through books and records produced and sold by intermediaries, instead of being consumed solely in real time in the presence of storytellers or musicians. More recently, the development of electronic commerce has seen the growth of a particularly important form of "reintermediation," the provision of information; see The Economist (1997).

(11) This is the mail-order business that is of interest to us. Some mail-order merchandise is delivered to local retail outlets for pickup by customers. It does not benefit from favorable sales tax treatment.

(12) I have not seen a set of terms that satisfactorily describes and differentiates the various forms of electronic commerce. I use the term "Web-order" to describe commerce conducted via Web sites in analogy with "mail order" and to distinguish it from telemarketing and TV shopping and from electronic commerce in content.

(13) See Hammond (1996). Wine merchants and microbreweries have been so successful in exploiting out-of-state markets that many states have recently enacted (or begun strictly enforcing) laws prohibiting mail-order importation of alcoholic beverages. See San Francisco Chronicle, August 21, 1997.

(14) The other characteristic of pure public goods, "jointness" in supply ("more for you does not mean less for me"), does not preclude the use of charges to finance broadcasting; if exclusion is possible, jointness simply means that consumption will be suboptimal.

(15) Paul Samuelson, in his best-selling textbook, has long used the example of scramblers and descramblers of radio signals to illustrate the difference in the two properties of public goods. Only recently has the example become realistic. In the Telecommunications Act of 1996, the federal government preempted the right of local governments to tax DBS services.

(16) Neither space nor sound judgment allows going over this well-plowed ground in detail. See any good undergraduate textbook.

(17) For more complete descriptions of the way various types of sales taxes function, see McLure (1987).

(18) Some might favor a highly differentiated tax structured along the lines of the theory of optimal taxation. This is not the place for an extended discussion of why uniform tax rates may be preferable as a practical matter. See Slemrod (1990).

(19) Some may think that business exemptions should be limited to inputs used to produce goods that are subject to sales tax. This view makes no sense, if exemptions of consumer purchases are predicated on distributional grounds. Exemption of all business inputs, like zero-rating under a VAT (instead of exemption), is justified in such cases and is easier to implement.

(20) Murray (1997) identifies the same three problems.

(21) The disadvantage is not so great for sales to business, especially in states with income taxes. While the strictures of Quill are equally protective of sales by out-of-state competitors, it is substantially simpler to force business purchasers to comply with use tax requirements by auditing their income tax deductions.

(22) If so, the exemption of many such products under current laws would reduce the competitive advantage that out-of-state suppliers would otherwise enjoy. This "two wrongs" approach is hardly a rational way to go about getting things right. Moreover, elimination of exemptions for these products would put local suppliers at a competitive disadvantage.

(23) Murray (1997) reaches the same conclusions: "First, nexus should be defined based on a vendor's economic rather than legal contacts with a state. Accordingly, any business making sales contacts in a state should be required to collect and remit use tax on its final sales. Second, all final consumption should be subject to tax, whether consumption is of a tangible or intangible item. Finally, all inputs and sales for resale should be exempt from sales taxation."

(24) This is essentially the concept of "economic nexus" favored by the states and endorsed by Murray (1997). To satisfy the artificial strictures created by judicial interpretation of the U.S. Constitution, it might be necessary to call the tax on purchases from out-of-state vendors a use tax. I ignore this distinction, which presumably would not be necessary if federal legislation were to be enacted to allow taxation of such sales.

(25) Again, Murray (1997) agrees: "The solution is not incremental reform, nor policy changes targeted to specific market transactions. Instead, sweeping reform of the entire structure of the sales tax will be needed to ensure its long-term viability."

(26) See, for example, McLure (1997b, c) and references cited therein.

(27) P.L. 86-272 prohibits a state from levying income tax on a seller whose only business activities in the state are solicitation of orders for sales of tangible personal property to be filled by shipment from outside the state. I discuss the implications of this law for electronic commerce in McLure (1 997a). More recently local governments have been prohibited from applying sales taxes to DBS service and from applying income tax to pension income attributable to income earned in the state during the working life of the taxpayer, but received after retirement while a resident of another state.

REFERENCES

Abrams, Howard E., and Richard L. Doernberg. "How Electronic Commerce Works." State Tax Notes (May, 1997).

Andall, Dean F. "State and Local Taxation of Electronic Commerce: Read My E-mail, No New Taxes." Paper presented at the symposium on Multi-Jurisdictional Taxation of Electronic Commerce, Harvard Law School, Cambridge, MA, April 5, 1997.

Cox, Christopher. "Opening Statement in Support of the Internet Tax Freedom Act." Testimony Before the Subcommittee on Telecommunications, Trade, and Consumer Protection of the House Commerce Committee. Washington, D.C., July 11, 1997.

Due, John F., and John L. Mikesell. Sales Taxation: State and Local Structure and Administration. 2nd ed. Washington, D.C.: Urban Institute Press, 1994.

The Economist. "Electronic Commerce: In Search of the Perfect Market." 343 No. 8016 (May 10, 1997): 13.

Fox, William F., and Matthew N. Murray. "The Sales Tax and Electronic Commerce: So What's New?" National Tax Journal 50 No. 3 (September, 1997): 573-606.

Frieden, Karl A., and Michael E. Porter. "The Taxation of Cyberspace: State Issues Related to the Internet and Electronic Commerce." State Tax Notes (November 11, 1996): 1363-94.

Gates, Bill. The Road Ahead. 2nd ed. New York: Penguin Books, 1996.

Grierson, R. Scot. "Constitutional Limitations on State Taxation: Sales and Use Tax Nexus on the information Highway." State Tax Notes (February 19, 1996a): 589-97.

Grierson, R. Scot. "State Taxation of the Information Superhighway: A Proposal for Taxation of Information Services." Loyola of Los Angeles Entertainment Law Journal 4 (1996b): 603-65.

Hammond, Ray. Digital Business: Surviving and Thriving in an On-Line World. London: Hodder and Stoughton, 1996.

Hellerstein, Walter. "State Taxation of Electronic Commerce." Paper presented at the New York University Conference on Taxation of Electronic Commerce, New York, May 19, 1997.

McLure, Charles E., Jr. The Value Added Tax: Key to Deficit Reduction? Washington, D.C.: American Enterprise Institute, 1987.

McLure, Charles E., Jr. "Taxation of Electronic Commerce: Economic Objectives, Technological Constraints, and Tax Law." Tax Law Review (forthcoming, 1997a).

McLure, Charles E., Jr. "Topics in the Theory of Revenue Assignment: Gaps, Traps, and Nuances." In Macroeconomic Dimensions of Public Finance: Essays in Honour of Vito Tanzi, edited by Mario I. Blejer and Teresa Ter-Minassian, 94-109. London: Routledge, 1997b (forthcoming).

McLure, Charles E., Jr. "The Tax Assignment Problem: Ends, Means, and Constraints." Public Budgeting and Financial Management (forthcoming, 1997c).

Multistate Tax Commission. "Computer Company's Provision of In-State Repair Services Creates Nexus." Nexus Program Bulletin, NB 951. Washington, D.C.: Multistate Tax Commission, 1995.

Murray, Matthew N. "Telecommunications Services and Electronic Commerce: Will Technology Break the Back of the Sales Tax?" State Tax Notes (January 27, 1997): 273-80.

Musgrave, Richard A. "Who Should Tax, Where, and What?" In Tax Assignment in Federal Countries, edited by Charles E. McLure, Jr., 2-19. Canberra: Centre for Research on Federal Financial Relations, 1983.

Negroponte, Nicholas. Being Digital. Paperback ed. New York: Vintage Books, 1996.

Newman, Nathan. "Prop 13 Meets the Internet: How State and Local Government Finances are Becoming Road Kill on the Information Superhighway." Economic Democracy Information Network Report, University of California. Berkeley: Center for Community Economic Research, 1995.

Organisation for Economic Co-operation and Development. Electronic Commerce: Opportunities and Challenges for Government. Paris: OECD, 1997.

Slemrod, Joel. "Optimal Taxation and Optimal Tax Systems." Journal of Economic Perspectives 4 No. 1 (Winter, 1990): 157-78.

Summers, June T. "Electronic Commerce: What Are the State Tax Implications?" Paper presented at the Vanderbilt School of Law Third Annual Paul J. Hartman State and Local Tax Forum, Nashville, October 17-18, 1996.

Summers, Lawrence H. Testimony before United States Senate, Committee on Commerce Subcommittee on Communications, on the Internet Freedom Act, S. 442. Washington, D.C., May 22, 1997.

Tapscott, Don. The Digital Economy: Promise and Peril in the Age of Networked Intelligence. New York: McGraw-Hill, 1996.

U.S. Department of the Treasury. Office of Tax Policy. Selected Tax Policy Implications of Global Electronic Commerce. Washington, D.C., November, 1996.

The White House. The Framework for Global Electronic Commerce. [Homepage of the White House], [Online]. Available: http:// www.whitehouse.gov/WH/New/Commerce/ read.html [Access Date: September, 1997].

Wyden, Ron. Statement on the Internet Tax Freedom Act. [Homepage of The United States Senate], [Online]. Available: http://www. senate.gov/~wyden/leg/cybstate. htm [Access Date: September, 1997a].

Wyden, Ron. Stopping `Cyber' Tax Chaos. [Homepage of The United States Senate], [Online]. Available: http://www. senate. gov/ ~wyden/leg/cybstate.htm [Access Date: September, 1997b].

CHARLES E. MCLURE, JR., Hoover Institution, Stanford University, Stanford, CA 94305.