Jonathan Zittrain Berkman Center for Internet & Society at Harvard Law School Chairman McCain, Ranking Member Hollings, Members of the Committee:

My name is Jonathan Zittrain, and I am the executive director of the Berkman Center for Internet & Society at Harvard Law School, where I also teach on Internet-related subjects as a lecturer on law. Among my research interests is the taxation of Internet commerce, and last year I wrote an article (attached) for the National Tax Journal on the subject with Prof. Austan Goolsbee of the University of Chicago.

Today the committee considers S. 2255, which is Chairman McCain’s proposal to extend through 2006 the moratorium on certain kinds of taxes set in place by the Internet Tax Freedom Act. I will try to touch on the economic implications of S. 2255 (and thus of the Internet Tax Freedom Act), as well as on other, more significant forms of “Internet taxation” to which the Act does not speak.

My bottom line: The moratorium of the Internet Tax Freedom Act is not objectionable, because the moratorium is so limited in scope that it has little consequence for state tax revenues-it does not apply to sales tax for physical goods bought over the Internet. The moratorium may also help the Internet’s growth at an important time. However, the real issues still lie ahead, particularly because while the Internet Tax Freedom Act is silent on state sales tax for online commerce, a Supreme Court decision has itself imposed an essential moratorium on such taxes. It thus falls to Congress to decide what the boundaries of state taxation power are in this area, and what research I have done in this area suggests that this more significant moratorium may be helpful now, but should be lifted later. I will first speak to the peripheral taxes covered by the ITFA moratorium, then to the more central taxes covered by the Supreme Court’s moratorium, and finally to some guiding principles that might help sort out what the ultimate policies should be.

The scope of the Internet Tax Freedom Act’s moratorium is quite modest. It restricts states’ abilities to impose discriminatory or multiple taxes on Internet commerce, and it prohibits new taxes on Internet access.

One example of a discriminatory tax might be a surtax on products ordered through the Internet (for example, a state assessing a 10% tax on books ordered online when it only demands a 5% tax on books bought in a bookstore). Another would be claims by multiple states to collect tax for a single transaction with a buyer in one state and a seller in another, thus doubly taxing. Each of these examples is hypothetical; I know of no major attempts by states to impose discriminatory or multiple taxes on Internet commerce, and thus no substantial state money at risk if this revenue stream were clearly marked off-limits. By its very terms, this aspect of the moratorium seems at best sensible and at least unobjectionable.

Examples of the prohibition on new taxes on Internet access are taxes on monthly subscription fees for America Online,, or any other service that provides Internet access. This moratorium may impact the coffers of states that wish to tax Internet access but did not have corresponding legislation on their books before the moratorium came into force. The category of taxable commerce affected here is small compared to the revenue to be gleaned from the broad swath of traditional goods typically covered by sales tax.

While the impact on state coffers may be small, the subsidy to Internet usage and to all the economic progress that flows from it could be large. This is because the Internet is subject to positive “network externalities,” which is to say that it becomes more useful to everyone as more people use it. (This is a general phenomenon of networks; compare how useful a fax machine might be to someone when only ten others own one versus when millions of others own one.) Such networks can grow exponentially once they reach critical mass, and signing more people on at a given time-thanks to an ability to offer comparatively lower access rates-amounts to a boost to future economic activity generally, at least to the extent that new Internet commerce need not simply be drawn from competition with existing retail stores. Further, any lessening of Internet access fees might help bridge the “digital divide” by making Internet access that much more accessible. A 5 or 6% difference might not seem like a lot, but there will be some group of people on the margin for whom it would make the difference between signing on and not signing on.

To be sure, once the Internet has reached a natural saturation point among potential users, there is less reason to treat its provision any differently from any other transaction subject to sales tax. Thus the infant industry protection represented by the “no new access taxes” part of the moratorium may not need to become ensconced as established industry protection.

More notable than what the ITFA moratorium covers is what it doesn’t cover. The moratorium does not preclude the application of state sales tax for physical goods ordered through the Internet. The meat and potatoes of state sales tax revenue comes from the sale of physical goods generally, so this is the source of revenue that states are most concerned about losing to tax-free sales of goods over the Internet.

If the moratorium creates no boundary, why the worry? Because there is another boundary out there: one that separates in-state from out-of-state merchants. Sales tax is a tax technically imposed on a consumer-when we buy something we fork over a little extra money to cover the tax-but it is enforced and collected by the merchant. In its landmark Quill decision, the Supreme Court made it clear that it was Congress’s province to decide the extent to which one state could force a merchant located elsewhere to collect a sales tax, even if the buyer, located in the first state, is clearly subject to the tax. Since Congress has been silent on the issue, states can only force out-of-state merchants to collect sales tax on items they sell to people living there if the merchants have other “contacts” with the state: so-called “nexus.” In practice, out-of-state merchants can usually avoid creating that nexus, so many distant merchants (whether they receive customer orders through the Internet, mail order, or telephone) cannot be forced to, and do not, collect sales tax.

According to the Census Bureau, sales tax amounted to $193 billion of state and local tax revenues in 1998. How do transactions placed through the Internet fit in? The best estimate Prof. Goolsbee and I could make on state sales tax revenue lost to out-of-state merchants receiving taxable orders through the Internet for 1998 is $430 million on total sales of $7.3 billion, or 0.2% of the collected tax kitty. Failure to pay tax on Internet-generated sales is thus not currently significantly denting state coffers.

What makes sales tax on goods purchased through the Internet such an issue then, despite the ITFA moratorium’s silence on the subject and the relatively small revenues currently at stake?

I can offer two reasons. First, Quill provides its own effective moratorium on sales tax on most Internet-driven sales unless Congress rescinds it, and everyone has big expectations for the growth of the proportion of sales taking place through Internet commerce. The most recent figures from the Census Bureau estimate $5.3 billion in online commerce sales for the fourth quarter of 1999, and Forrester Research estimates $108 billion per year in online retail sales by 2003. Predictions beyond 2003 are due to be quite speculative; perhaps between 2004 and 2007 revenue loss from online, inter-state transactions for which Quill blocks sales tax collection could amount to ten percent of total sales tax revenue, if more interstate, sales tax-exempt trade happens overall thanks to the added ease of Internet ordering.

A second reason for the current worry over an inability to apply sales tax to goods purchased across state lines has to do with a desire not to unduly distort markets with arbitrarily applied taxes. Tax experts may have differing personal views as to whether taxes should be raised or lowered generally, but they tend to be in agreement over the idea that one should tinker with rate rather than scope when seeking to adjust the public’s tax burden. Over the short term, at least, Quill’s moratorium on out-of-state tax collection will likely encourage more people to use the Internet for shopping, just as the ITFA’s moratorium on new access charges will encourage more people to sign up for Internet access in the first instance. But as the use of the Internet matures and the benefits of the network externalities I discussed earlier are reaped, distinctions such as “in-state/out-of-state” or “ordered through Internet/ordered in a store” become truly arbitrary. Differences in tax rates should be made on the basis of the substance of a sales transaction, not on where or through what medium it takes place. Local merchants, themselves in many instances limited to margins of 4 or 5% on their wares, should not forever pay a sales tax while their online and/or out-of-state counterparts do not.

Further, the enforcement costs of imposing taxes on goods ordered and paid for over the Internet could drop over time. Thanks to authentication and encryption technologies under development in the private sector-technologies to ensure that when one orders a dozen pizzas through the Internet, one cannot repudiate the bargain by saying someone else actually placed the order-it may become quite easy to know who is buying what from whom, to know where the buyer is, and then to collect the appropriate tax. This raises serious privacy issues, particularly if the scope of state sales tax varies so much that one must know and verify the nature of the item purchased in order to actually assess and account for the tax. But in a simplified scheme where the various states can agree on common definitions if not rates-something sorely needed and long overdue-one could actually imagine the collection of sales tax as second nature in online transactions, far easier than the corresponding calculation, collection, and remittance by local merchants in a traditional transaction. Indeed, structured properly, the collection of tax could come straight from the user, converting “sales tax” collected from the merchant into a corresponding “use tax” collected from the buyer, and in such a way that the buyer would not revolt. (Current use taxes, owed by consumers whenever they have managed to avoid having their merchants collect sales taxes thanks to Quill, remain largely uncollected, presumably because consumers would not take well to having to maintain accounts of everything they have purchased and what tax they might owe on it.)

In the current political climate it seems difficult to imagine Congress enabling states’ collection of sales tax from out-of-state merchants, so the revenues will only be obtained-if at all-through creative electronic collection schemes that can manage to only minimally burden both seller and buyer, or through reciprocal state tax collection agreements, through which New York, say, could ask a New York merchant to collect and remit New Jersey sales tax for its New Jersey customers. Neither of these solutions is particularly appealing, nor are they easy to implement, though they well emerge as alternatives to Congressional action to allow states to collect sales taxes across state borders.

Again, over the long run, state boundaries seem odd and unhelpful lines to draw on sales tax collection, as do boundaries between electronic and physical means of ordering. The legal and technical status quo whereby some transactions avoid the tax while others do not should, in the long run, be traded in for a more comprehensive tax reform that offers uniform tax relief (perhaps in tax rates) while enabling or maintaining other revenue streams in as simple and direct a way as possible.

So far I have interpreted “Internet commerce” to cover the purchase of physical goods ordered via a computer network instead of a telephone call or visit to a store. I do see this as the core of the Internet tax controversy, because a lot of money will sooner or later be at stake through such channels, and because there exist bricks-and-mortar merchants who sell identical products and for whom differential tax treatment seems, over the long run, unfair.

But the Internet also enables the sale of digital goods: e-books and software, for example. Depending on one’s reading of the Internet Tax Freedom Act, these purchases may not be taxed by the states, whether the purchase is inter-state or intra-state. Since the distribution of wholly digital goods is especially in its infancy, even more so than the online purchase of physical goods, this would be an auspicious time for a moratorium on taxes of such goods.

Indeed, we may see the creation of new markets where individuals can sell cookie recipes or bedtime stories one at a time, for 25 or 50 cents each. To insist on collection and calculation of sales tax on such transactions might produce an administrative barrier that would preclude the development of such a “small fry” sellers’ market.

It is also important to ensure that other countries to not impose onerous or discriminatory taxes on digital Internet commerce, especially as they might perceive that digital merchants on the Internet are disproportionately American vendors.

The Internet Tax Freedom Act does not speak to the taxes that really fill state coffers-and hit consumers’ pocketbooks. What few taxes it does preclude deserve to be precluded, and thus an extension of the Act’s moratorium seems perfectly appropriate, if not particularly efficacious. But in passing this Act, it’s important to note that much more work remains to be done. In particular, the convening of the commission that the Act chartered has helped focus attention on a long-simmering issue for which the growth of the Internet is turning up the heat: the fact that states require Congress’s formal assent before they can readily collect most of the taxes they wish to on goods purchased by in-state consumers from out-of-state merchants. A long-term compromise might be the easing of states’ ability to collect such taxes in exchange for a serious simplification and harmonization of the substantive scope and administrative burden associated with the respective state sales tax regimes.

An extension of the moratorium should be accompanied by efforts to broaden the difficult conversation begun in earnest by the Advisory Commission on Electronic Commerce and among officials representing state and local governments, attempting to agree on the fairest and most practical ways to enjoy economic growth and freedom while paying the piper for the common services from which we benefit.