Virtual Intermediaries: Consumption Tax Problems In Japan, Europe, And The United States – The Case Of The Virtual Travel Agent


Marketplace technology is (inadvertently) chipping away at the effectiveness of consumption taxes – the Japanese Consumption Tax (CT), the European value added tax (VAT), and the American sales tax (ST) are all affected. Frequently a technology-patch or a law change can repair the tax-damage, but sometimes even though a patch or a change is known the design of the levy (or the politics behind the design) impedes application.

All consumption taxes are not the same. Although theoretically similar – the stated policy objective of each is to tax final consumption – structural differences among them make the CT highly vulnerable to virtual intermediaries, whereas the VAT is nearly immune. The American ST, caught between these extremes, seems to be in the worst position – disputes in the hotel sector with virtual intermediaries have spawned litigation coast-to-coast.

This paper reinforces an old observation – new business models (virtual intermediaries) may be good for commerce, but they can be very disruptive to a non-adaptive tax system.


Perhaps the classic examples of the tax-threat posed by technology were the internet retailers or e-tailers of the 1990s. During the “dot com” era tax administrations feared that distant internet sellers would replace Main Street, and would not collect consumption taxes.[1] Concerns encompassed business-to-business (B2B) as well as business-to-consumer (B2C) transactions. The CT, the VAT and the ST all made adjustments, and even though the solutions are different it is reasonably clear that the tax due on most internet sales is being collected today (although the results appear to be better under the VAT than under the CT2 or the ST3).

This article moves beyond an assessment of this classic tax-threat (the problem of “what if” the e-tailer simply refuses to collect the tax). Instead it looks at a single technology-induced business change – virtual intermediaries in the travel industry – and follows this development through to its consumption tax consequences. The point underscored is that digital intermediaries are engaged in far more than disintermediation – they do more than simply pass on savings by eliminating middlemen. They are forcing fundamental changes in business processes, and by extension are forcing changes in consumption taxes.

No one – not the crafters of the ST, the creators of the VAT, nor the designers of the CT foresaw the impact of modern technology – much less the specific impact of virtual intermediaries. It appears that the CT and the ST are struggling more with these changes than is the VAT. This suggests that technology is pushing American ST jurisdictions[4] and the Japanese[5] to consider replacing their systems with a European style VAT.

But before we take up this topic, there is something else to consider. The current recession is battering consumption taxes. Consumers are buying less; tax revenue is down sharply, and tax authorities are making every effort to secure (enhance) revenue streams. The simple solution is a rate increase, and there have been rate increases in both ST and VAT.[6] Japan has considered a rate increase, but that option was rejected by the prior government and continues to be unacceptable.[7] In jurisdictions where rate increases have occurred, they have sometimes gone into effect so fast that businesses have found it hard to adjust.[8]

There are limits to what can be done with rate increases. As those limits are approached, the continuing search for revenue exposes the strengths and weaknesses of the ST, the VAT and the CT. One can almost hear revenue authorities asking: Where else (other than in a rate increase) can we find more revenue? Where does the consumption tax itself tell us to look for more money? The answers are revealing.

The US ST and EU VAT are looking in entirely different places. VAT authorities have been focusing heavily on fraud detection.[9] ST authorities have been trying to expand the tax base adding taxes on services,[10] and extending nexus[11] to bring more businesses into the net. The Japanese CT is even more constrained, because the tax base is derivative. The corporate income tax determines that base, so other than more aggressively pursuing income tax audits, it appears that nothing more can be done. However, is there something that Japan can learn from the ST and the VAT?

An example helps. In Europe major steps have been taken all through the recession to curb technology-facilitated sales suppression frauds – skimming revenue from cash sales in B2C transactions. An EU-wide committee has been established,[12] technology-based solutions are being developed,[13] audit techniques and enforcement approaches are being shared. All of this has lead to litigation,[14] new regulations, and significant statutory revisions.[15]

The ST is just as vulnerable to skimming technology, as is the VAT. American tax losses are estimated to exceed $20 billion in the restaurant industry alone.[16] However, no US State is engaged in an enforcement effort even remotely comparable to that in the EU.[17] The same is true in Japan.

Two US cases have been uncovered. Japan has uncovered none. Both of the US cases were the result of IRS (income tax) investigations. Neither case arose from a technology audit (that is, the IRS found something – technology fraud – that they were not looking for, and no one – state or federal – followed the technology problem back into the marketplace).[18] Why is this, and why does Japan appear to be following the US?

The answer in the US is structural. It’s not that ST authorities do not believe that there could be automated fraud programs embedded in American electronic cash registers (ECRs) – it’s just that when these administrators look carefully at the ST, the tax itself seems to suggest that the “low hanging fruit” or the “big and easy returns” are in expanding the tax base. Examining internal factors seems much less productive – intensive technology audits of businesses that are already filing is not “where the big money is.”

The answer in Japan is also structural. The CT and the income tax are tied so closely that there is very little independent CT policy. Japan could be overlooking significant revenue by seeing itself as primarily an income tax jurisdiction. With a strong CT, independent of the income tax, Japan’s CT could be more like the VAT, and recover tax on more of the consumed value added in the country.


Intermediaries – Generally. All consumption taxes have problems with intermediaries when sales cross jurisdictional boundaries. The most challenging problems involve drop shipments – transactions where an intermediary is contractually involved in two legs of a three- party transaction, but not directly (physically) involved in making the supply.

Suppose “A” agrees to supply “B” at an agreed price (100), and then “B,” acting as an intermediary, marks-up the supply (to 150) and re-sells to “C.” Suppose further that “C” is a final consumer and consumes in its domestic jurisdiction. This three-party transaction is a drop shipment transaction if “B” instructs “A” to deliver directly to “C.”

Problems arise because the second sales contract (“B” to “C”) is uncoupled from the delivery (“A” to “C”). Said another way, both “A” and “B” know half the information. “A” knows when the supply is made, but does not know the price that “C” pays. “B” knows the price

“C” paid, but does not know when delivery is made.

The tax problem is to align substantive jurisdiction with enforcement jurisdiction. How can we align “the right to tax,” determined by the place of consumption, with the right to compel the taxpayer (or a withholding agent) to collect the tax?


Intermediaries – EU background. The EU VAT always reaches a taxable result in drop shipment transactions. Substantive tax jurisdiction aligns with enforcement jurisdiction (tax is collected in the jurisdiction that has enforcement jurisdiction). Mandatory EU-wide rules control the outcome.[19]

The VAT is premised on the supplier (not the customer) being the taxpayer.[20] Thus, if “C” is a final consumer, then “B” (the intermediary) is obliged to collect VAT based on the invoice price and is required to remit the funds to “C’s” jurisdiction. An exception applies for sales of goods below a threshold (100,000 or 35,000 euro), and some additional refinements, but none of them result in the tax being unreachable.

In the case of a sale of goods, if “C” was a business then the cross-border (intra- community) supply becomes a zero-rated sale by “B,”[21] and “C” is required to self-assess (reverse charge) the VAT and remit the tax to “C’s” jurisdiction.[22]

If services (instead of goods) are involved the EU VAT still assures collection of the VAT. Depending on the nature of the services, and the status of the customer (business or final consumer) proxy rules source the transaction either to the buyer’s or seller’s jurisdiction.[23]

Depending on the facts either the seller is required to collect the VAT, or the buyer is obligated to perform a reverse charge. However, no transaction escapes tax.

Intermediaries – US background. The American ST does not always reach a taxable result in a drop shipment transaction. It is easy to see why. At its core, the ST relies on a jurisdictional mismatch. The customer (not the supplier) is the taxpayer under the ST, and substantive jurisdiction is premised on consumption occurring within the jurisdiction. However, the vendor is the collection agent for the ST, and when the vendor and the customer are not located within the taxing jurisdiction there are difficulties.

The constitutional linchpin in this misalignment is the “substantial nexus” requirement locked in place by Quill Corp. v. North Dakota.[24] Quill makes it clear that US states cannot enforce collection of the ST unless the vendor is “physically present” in the state. Imposing a collection obligation on an out-of-state business making $10 million in sales into a state (as was the case in Quill) is unconstitutional (it unduly burdens interstate commerce) if the vendor lacks physical presence.[25]

This is a particular problem for drop shipment transactions. Assume that consumer (“C”) orders goods from an out-of-state intermediary (“B”) that has no physical presence in “C’s” jurisdiction. Assume further that “B” does not keep these goods in inventory, but secures them on an “as needed” basis from a wholesaler that also has no physical presence in “C’s” jurisdiction. In this case taxing jurisdiction is in “C’s” state. “C” is a final consumer. However, there is no third party enforcement jurisdiction. All that can be hoped is for “C” to volunteer payment.

This outcome encourages states to stretch nexus rules. One approach has been to apply common law agency principles and attribute nexus to out of state retailers (“B”) based on the activities of a related party within the state.[26] In other instances states have transferred collection responsibility to the drop shipper (“A”) making deliveries.[27] Recently, a “flash title” argument has been developed that endeavors to satisfy Quill’s “substantial” nexus test by finding “physical presence” of inventory within the state for a moment (a flash). The end result has been to make drop shipments a difficult area of ST compliance.

Intermediaries – Japanese background. Like the VAT the CT aligns substantive and enforcement jurisdiction. The taxpayer under the CT is the seller, but the seller is present for CT purposes only if they are present for income tax purposes. Thus, like the ST, the CT becomes concerned about the “physical presence” of the intermediary.

In the Amazon case referenced above, the NTA attributes taxable presence based on the activities of related parties. If Amazon is within the scope of the income tax it is within the ambit of the CT.[28] But if the CT were more like the VAT than the ST more transactions would be captured. Could the CT be assessed simply based on Amazon making sales into Japan, regardless of physical presence?


By the end of 2000 Harvard Business School (HBS) was dissecting the internet’s impact on prices and market-making. One HBS Note “prepared as the basis of class discussion” assessed seven different market-making/ internet-sensitive mechanisms and argued that business professionals should move their thinking away from a “commodity selling mentality” in the Internet era.[29]

When business professionals change their thinking, tax officials need to follow suit.

The thrust of the HBS Note was that internet-based market mechanisms were doing more than disintermediation. Virtual intermediaries were offering new products (not just lower prices on existing commodities). Teaching Notes indicated that, “[t]he ‘ah-ha’ [moment] for most [students came] from seeing that this is not merely a pricing story … but rather a product story … [a story that is] of great interest to suppliers.”[30] HBS saw this change of thinking most clearly in the travel industry.

The travel reservation industry is dominated by four internet-based intermediaries. The largest, and are independent of the major airlines; the others, and, began as wholly owned airline subsidiaries. Where and began as disintermediation agents (cutting out independent ticket agents) this was not the and business model. used a “buyer driven” business model. sold a new product units of demand. Traditionally units of supply are offered (a specific seat on an airline going from Boston to New York, on a specified day and at a specific time).

Under the model consumers offer a unit of demand – the demand for a specific seat. The customer specifies the price. With multiple airlines traveling the same routes, one may have an unsold seat on the designated day. If there is a match and the price is acceptable, the consumer’s demand will be satisfied.

This is a “reverse auction.” characterizes its business process as a “demand collection system.”[31] It works because there are an estimated 500,000 unfilled airline seats each day that have a near zero marginal cost. collects demand for these seats. If a consumer makes a low, but irrevocable offer to purchase one of these seats – airlines listen.[32]’s business process is patented.[33] The patent attracted academic attention when sued Microsoft and for infringement when rolled out a demand collection system for hotel reservations – Hotel Match Maker. The suit was settled. was allowed to continue to use Hotel Match Maker, but it had to pay a royalty back to

Virtual (travel) intermediaries and the EU VAT. From its earliest days the EU VAT anticipated that travel agents, functioning as intermediaries, would need to be treated under a special scheme. In 1977 there were two concerns: (1) simplification and (2) sourcing revenue to the place of consumption.

The solution was to make the travel agent a final consumer of tickets and reservations. VAT paid by the agent was not deductible (not allowed as input credit on the agent’s return, nor did it qualify for refund in any Member State).[34 ] When tickets and reservations were resold the agent was required to determine its margin and remit VAT based on this amount. The margin was defined as the total cost of tickets purchased (plus VAT and related operational costs) less the total price paid by the traveler (without VAT). Although VAT on this margin is not stated on the traveler’s invoice, it is (most likely) “passed through” to the traveler. There is an assumption that the “traveler” is not traveling for business.[35]

An example is helpful. Assume a travel agent in the UK purchases the following items for a package tour: round trip train tickets from Brussels to Copenhagen (50); lodging reservations in Copenhagen (30); and meals at various German restaurants (20). The total cost is 100. The package tour is sold to travelers for 150 (the VAT exclusive amount). Operational expenses are 5. The agent will pay VAT of 22.[36]

Additionally, there is UK VAT on travel services of 15%, so the agent must remit 3.45 VAT on each package sold. 3.45 is 15% of the agent’s margin of 23 (150 – 127).[37] Assuming the VAT is “passed through” to the traveler, the invoice will be for 153.45 (150 + 3.45).[38 ] The invoice will not specify the VAT. VAT is fully collected, but the markup is not revealed.

This scheme simplifies compliance (only one return is filed domestically) and we have sourced VAT on travel,39 lodging[40] and meals[41] to the jurisdictions where consumption occurs. The details can be diagramed as follows:

Presently, this scheme is not functioning optimally. There are three areas of concern: derogations, vertical integration of the industry, and the Internet. Most notably, the EU has permitted a large number of derogations that were expected to be temporary. The derogations have remained, and have made the scheme uneven among the Member States.

The Commission has proposed a Directive, and a Regulation that will resolve these issues.[42] The Directive would eliminate derogations and extend the margin scheme to business customers. Although VAT-inclusive invoicing of travelers would remain, an opt-out rule allows businesses to receive a pass-through of the VAT charges on a service-by-service basis. There are also changes in the place of supply for transactions involving non-EU travel agents.

This last change targets virtual intermediaries. The place of supply will move from the place where the agent is established or has a fixed establishment out of which services are provided [43] to the place where the customer is established.

To see how this will work, consider the earlier example. Under the current margin scheme a 3.45 UK VAT is due on the travel agent’s services. If the agent performing these services moves to the US, these services will not be subject to VAT. Thus, if the package tour sells for 153.45 (with a margin of 23), it now has a profit of 26.45.

But, what happens if instead of moving to the US the agency moves to a low cost (as well as no tax) jurisdiction? Operating costs will be cut. Suppose operating expenses are cut by 40% to 3. The profit margin increases to 28.85. Is this all?

clip_image004Perhaps a better way of seeing this is to imagine that the travel agent is back in the UK. The “traditional” travel agent would pay UK VAT of 3.45. A travel agent following the business model would pay UK VAT of 8.94.[49]

In other words, under current rules the EU VAT captures the entire margin of the business model but only if the travel agent is established in the EU. Reductions in the VAT base for travel, lodging and meals is matched by an increase in the VAT base for travel agent services.

Now consider the Commission’s proposals. Once again the entire margin is taxed, but this time it is taxed even if a of travel agent sells from a no tax/ low cost jurisdiction. The rate on the margin would vary depending on the place of establishment of the traveler. If a UK traveler purchased the package VAT of 9.7575 would be due.[50] If a Swedish traveler purchased the same package, VAT of 16.2625 would be payable.[51] Under the Commission’s proposal the travel packages must be for travel within the EU,[52] but the problem posed by the virtual intermediary in the EU VAT is solved.

Virtual (travel) intermediaries and the US ST. It is safe to say that the US consumption tax never anticipated that it would have problems with travel agents. After all, travel agents are service providers who primarily resell services (tickets for passenger travel or admission to entertainment venues), and as a general rule services are not subject to ST in the US.[53]

In fact, the biggest ticket item for travel agents – passenger travel – is substantially off limits. Federal legislation prohibits states and localities from imposing taxes on the sale or resale of inter-state passenger tickets.[54]

Travel agents, particularly when they operate virtually, impact American consumption taxes when they (a) bundle into package tours non-travel transactions – hotels, car rentals and meals, and (b) sell bundled package from outside the jurisdiction of consumption. Hotel accommodations, frequently the second largest expense in a travel package, are normally taxed through a special ST rather than the general retail sales tax. Virtual intermediaries appear to be taking advantage of ambiguities in these laws to improve margins.

Recall Professor Dolan’s “ah-ha moment.”[55] Virtual intermediaries are selling new products. Rather than just offering the same products to the public at a lower price, they re-mix goods and services into something new. In VAT (where both goods and services are taxed) this is not significant, however in the ST shifting sales from tangible property to services is very significant. The following diagram adapts the earlier example:


The problem begins when virtual intermediaries disrupt the alignment of substantive and enforcement tax jurisdiction. Rather than having the hotel operator collect hotel tax from the guest, the virtual intermediary collects it remotely. Full payment is made in advance (online with a credit card).[56]

Merchants bill the intermediary after consumption. At this point the difference between a general sales tax and the hotel tax is important. Under most general STs an intermediary’s markups (commission, fee or other service charge) are included in the sale price.

Because most [but not all[57]] hotel taxes place collection responsibility on the hotel, not the intermediary, the tax base is (arguably) the discounted price the hotel charges for the room – not the full price charged the traveler. Thus (using’s “reasonable” discount of 30%), the tax base would be 21 not 30 for lodging, although it would remain 20 not 14 for meals. Assuming a uniform 10% tax rate,[58] the tax-inclusive hotel bill to the intermediary would be 23.1 (21 + 2.1).

The merchant (hotel) files returns and remit tax. States are concerned that the hotel tax is being underreported. They would argue in our example that the hotel tax should be 3 not 2.1, (10% of 30, not 10% of 21).[59]

The VAT never has this problem. If a virtual intermediary reduces prices through a “demand collection system,” the full tax is always collected.

Services are not generally taxed in the US (and a virtual intermediary can always locate itself in a jurisdiction that does not tax services). As a result, the business model easily threatens the tax base when taxable transactions become non-taxable (service) transactions.

Virtual intermediaries however do not have the best defenses when tax authorities come looking. By injecting itself into in-state transactions as a voluntary tax collector virtual intermediary concede nexus. In other words, the virtual intermediary cannot resist compliance responsibilities under Quill, and must respond.[60] Nevertheless, virtual intermediaries are resisting and there is coast-to-coast litigation.

American states cannot resolve this problem in an EU manner – they cannot tax intermediary services where they are performed. As a result, the American approach has been to undo what the virtual intermediary has done – to try to re-incorporate services back into the property, and taxing these amounts in the jurisdiction where the property is located rather than where the services are rendered.

There are very good business reasons (on both the intermediary’s and the merchant’s side) for designing things this way. In summary form, the business reasons for this arrangement are:

– The intermediary does not want the merchant to collect tax directly. The intermediary increases sales by creating the impression (in the mind of the traveler) that it provides the best bargains. Revealing the discounted price through the tax system would tell the traveler that a better price was possible.

– The intermediary does not want the traveler to know its margin. Disclosing this would encourage customers to bargain harder.

– The intermediary wants to pay tax on the discounted price not the full price. Keeping taxes low increases margins.

– Merchants do not want to collect tax from the customers based on their discounted prices, because they do not want to make their discounts public. They receive shields:

– a price shield;

– a brand shield.[61]

US Litigation. Early on State litigation was ineffective. Recently tax authorities have become more successful. Million dollar judgments are being collected.[62]

The resource commitment is huge. For example, in the State of Texas as of the end of 2002 there were 417 jurisdictions that levied a hotel tax.[63] The city of San Antonio, brought suit against (and other web-base hotel booking companies[64]) in 2007 for unpaid occupancy taxes,[65] and the class certified by the court contains 175 cities divided into two sub- classes of 47 and 128. The 242 other jurisdictions face the same issues, but they cannot be joined until local remedies are exhausted. As a result, there is the distinct possibility that we will see up to 200 more law suits on this issue in Texas alone.[66]

Virtual (travel) intermediaries and the Japanese CT. A quick check of the major internet travel agencies shows that each are re-selling hotel rooms in Japan. Are these virtual intermediaries operating in a manner that minimizes CT revenue? Because there have been no court cases or official statement on this issue, there is no way to know from the outside.

However, if we assume that virtual travel intermediaries apply the same business model in Japan that they use in the US, and if they are structured to have no Japanese taxable presence (no permanent establishment), then the CT on the service element of a room reservation will be avoided. In aggregate these amounts could be significant. One would expect that these intermediaries would avoid the 5% CT. It is a direct cost of doing business.

For example, assume a Tokyo hotel has a room that is:

– Regularly offered to the public at ¥50,000 (per night);

– Secretly offered to a virtual intermediary at a 50% discount (or ¥25,000); and then

– Resold to a guest at a 30% discount off the published price (or ¥35,000). The hotel’s CT base will be ¥25,000. When the guest checks into the hotel, the hotel will invoice the intermediary for ¥25,000 (inclusive of the CT). The CT is ¥1,190 of this amount.

However, the guest has paid ¥35,000. This is the actual measure of consumption. In the typical American transaction the online reservation engine of the intermediary indicates that all taxes and processing fees must be paid (along with the room charge) in advance. Intermediaries are notorious for not itemizing amounts. If a guest needed to determine the amount of CT paid (as a business expense or for other reasons) he would probably assume that a CT of ¥1,667 had been paid on a base of ¥35,000. However, this amount exceeds the actually remission by ¥477.

Is this significant? Aggregate CT lost could be ¥26,532 million. This figure is arrived at in the following manner. Assume that 20% of Japanese hotel rooms are purchased through online intermediaries,[67] and that average price reductions are achieved. If we apply these assumptions to total annual hotel sales (2004 figures are the most recent) we get the following:

– Total annual hotel sales = ¥6.63 trillion (¥6,632,920 million)[68]

– 20% of these sales = ¥1,326,584 million

– CT reported on these sales = ¥66,329 million[69]

– Amount paid by guests for these hotel rooms = ¥3,714,435 million[70]

– CT collected by virtual intermediaries on these sales = ¥92,861 million[71]

– Underreported CT = ¥26,532 million.[72]


CT losses of ¥26,532 million are significant. These losses follow directly from the virtual travel intermediary’s standard business model.[73] It is a model that has brought considerable efficiency to the travel industry, but it is also a model that may have contributed to CT shortfalls like this in each of the past eight years.

The essential question facing the Japanese CT is: should substantive CT jurisdiction over the seller continue to be derived from substantive income tax jurisdiction over the seller? In other words, should the CT continue as before or should it follow the VAT and assert jurisdiction over any enterprises making sales sourced in Japan?

The VAT suggests that Japan should sever the CT from the corporate income tax, if only for the purpose of determining whether or not a business has an obligation to collect and remit the CT. Substantive CT jurisdiction should be found by applying “place of supply rules” to transactions, rather than by applying “physical presence” rules to the entities making supplies.

If not, then the ST suggests several other avenues for Japan. First, virtual intermediaries could be found directly subject to the CT because they are conducting business in Japan – they are re-selling rooms from their Japanese inventory.[74] Second, virtual intermediaries could be found subject to the CT because they have conceded substantive jurisdiction. They collect the CT, and only enterprises doing business in Japan are allowed to collect the CT. The intermediary is a voluntary tax collection agent.[75] Third, the virtual intermediary could be found unjustly enriched. They are under social (or moral) obligation to remit the full CT. The full tax is what the intermediary has collected from the guest, and they should not be permitted to keep part of it for themselves.[76]

Virtual travel intermediaries are only one of many technology-intensive businesses that are increasing business efficiency, but are making consumption taxes more difficult to collect. Virtual intermediaries are making these changes necessary. This is especially true in Japan.

The VAT, ST and CT are all impacted, but lessons learned in one tax can be applied in the others. Exactly how Japan proceeds on this issue may depend on how it resolves some administrative concerns – how will it determine the full tax base on each room that is rented through an on-line intermediary; how will it compel on-line intermediaries to file returns; will the hotel be deemed the agent of the on-line intermediary, or will the intermediary be treated independently? Fortunately, many of these questions (and more) have been worked out within the VAT and the ST.

[1 ]In the US a flurry of white papers appeared that were frequently republished in State Tax Notes. A representative sample includes: Interactive Services Association, Logging On to Cyberspace Tax Policy: An Interactive Services Association Task Force White Paper, 97 STN 209 (Jan. 20, 1997); Information Highway State and Local Tax Study Group, Supporting the Information Highway: A Framework for State and Local Taxation of Telecommunications and Information Services, 95 STN 57 (July 3, 1995); Karl A. Frieden and Michael E. Porter, The Taxation of Cyberspace: State Tax Issues Related to the Internet and Electronic Commerce, 96 STN 1363 (Nov. 11, 1996); Multistate Tax Commission, Statement of Direction on Electronic Commerce Issues (Jan. 17, 1997); Department of the Treasury, Office of Tax Policy, Selected Tax Policy Implications of Global Electronic Commerce (Nov. 1996); White House Interagency Task Force, A Framework for Global Electronic Commerce, Draft #9 (Dec. 11, 1996).

Two elements are needed to solve the problem presented by the e-commerce business model: (1) clear set of rules that determine the taxability of transactions involving distant digital sales and sales of digital products, and (2) a digital compliance regime to facilitate accurate and efficient tax collection under this new business model. The technology patch for the e-commerce business model in the US is the Streamlined Sales and Use Tax Agreement (SSUTA) (adopted November 12, 2002, amended November 19, 2003 and further amended November 16, 2004) available at

The same e-commerce concerns in the EU were met with mandatory changes in the place of supply rules throughout the EU. Council Directive 2002/38/EC, art. 1, 2002 O.J. (L 128) 41, available at:

[2] is engaged in a classic e-commerce dispute with the Japanese National Tax Administration over distance sales of goods made over the internet with delivery by common carrier. Because the CT is dependent on the corporate income tax, an income tax audit easily rolls into a CT assessment. One can be sure that the NTA is also looking for CT revenue in the following:

In Amazon’s 2008 annual report, released April 17, the company said Japanese tax officials in 2007 assessed income tax of about $119 million, including penalties and interest, against a U.S. subsidiary for tax years 2003-2005. …

According to media reports, International Sales, a Seattle-based affiliate of Amazon, concluded direct contracts with customers in Japan and received payments for goods. Two Japanese subsidiaries of Amazon — Amazon Japan and Amazon Japan Logistics – managed sales, merchandise distribution, and logistics operations in Japan. The sales income from the Japanese business was reported in the United States by International Sales. Kristen A. Parillo, Japanese Tax Officials Seeking $119 Million in Back Taxes From Amazon, TAX NOTES INT’L (July 8, 2009) Doc 2009-15254; 2009 WTD 128-3.

[3 ]For example, it would be unheard of under the EU VAT for a major retailer to threaten to leave a jurisdiction based simply on a requirement that it collect the VAT on sales made to domestic customers. However, this is exactly what is happening under the RST. It is an effective way to prevent adverse legislation from being enacted. See: Lenny Golberg, California Governor Vetoes Tax Enforcement Bill and Its Nexus Provisions (July 6, 2009) 2009 STT 126- 8; Doc 2009-15139 (indicating that and’s threat to terminate business relationships with internet affiliates in California if the sales tax law was amended to make them liable to collect the sales tax on sales into California was the reason the governor vetoed this bill). These threats are effective in large part because they are not “idle threats.” Both Amazon and Overstock have made good on similar threats, most recently in North

Carolina. On August 7, 2009 Governor Bev Perdue of North Carolina signed “affiliate nexus” legislation, and as a result both and terminated their North Carolina affiliate program. See: Geoffrey A. Fowler, Amazon Cuts North Carolina Affiliates to Avoid Tax, WSJ (Jun. 27, 2009) available at:; John Buhl, Discontinues Use of Advertiser Affiliates in Four States, State Tax Notes 12 (Jul. 6, 2009) Doc 2009-15078 2009 STT 125-1

[4 ]For example, the State of California is currently considering replacing the state level RST and the Corporate Income Tax with a net business receipts tax (NBRT). The NBRT was adopted in Michigan, and is similar to an addition method VAT determined on an annual basis. COMMISSION ON THE 21ST CENTURY ECONOMY, REPORT (September 2009) available at:

[5] In 2003 the Japanese Tax Commission characterized the Japanese fiscal situation as a “state of crisis … [brought about by the current system where] only about half of Japan’s annual expenditure is covered by tax revenue,” and recommended three very significant CT changes: (1) the rate should double; (2) multiple rates should be employed; and (3) the “bookkeeping method” of accounting should be abandoned and replaced with the significantly more complex “invoice method” that is used in Europe. JAPANESE TAX COMMISSION, A SUSTAINABLE TAX SYSTEM FOR JAPAN’S AGING SOCIETY (Jun. 2003) 4 & 9 available (in an “informal English translation”) at

[6] Both Americans and Europeans have raised consumption tax rates during the heart of the current recession. For example, in the US there have been 761 rate increases from November 1, 2008 through June 1, 2009. Among the 27 Member States of the EU there have been 5: Among the 27 Member States of the EU there have been 6: Estonia increased its reduced rates from 5% to 9%, effective January 1, 2009 and has increased the standard rate from 18% to 20% effective July 1, 2009; Ireland increased the standard rate from 21% to 21.5%, effective December 1, 2008; Latvia increase the reduced rate from 5% to 10%, and increased the standard rate from 18% to 21%; Lithuania increased the standard rate from 18% to 19% and eliminated almost all of the reduced rates; Hungary increased the standard rate from 20% to 25% and instituted a second reduced rate of 18% while leaving some transactions at the older reduced rate of 5%. Personal e-mail communication from Matthew Walsh, Director of Tax Research ADP, June 8, 2009 (on file with author). Rate reductions during the recession have been rare. None have been reported in the US, and only the UK has reduced the standard rate, from 17.5% to 15% in an effort to stimulate the economy.

Tanzania seems poised to follow the UK, reducing its rate from 20% to 18%. Costantine Sebastian, Budget 2009/10: Package to curb impact of global crisis, THE CITIZEN (Jun. 12, 2009) available at:

[7 ]Daily Yomiyuri Online, New Panel on Tax Holds First Meeting – Body to Study DPJ Election Vows’ Feasibility, (Oct. 9, 2009) available at (“[T]he panel will address the feasibility of an environmental tax to combat global warming as well as review alcohol and tobacco taxes. To finance measures for pension reforms, discussions focusing on the eventual raising of the consumption tax rate also will be launched. Hatoyama has said his government will not raise the consumption tax for the next four years. However, Naoki Minezaki, senior vice finance minister, indicated the consumption tax issue would become

one of the new tax commission’s key issues. “)

[8]For example, the Estonian Parliament decided on June 18th to raise the standard VAT rate from 18% to 20% (effective July 1, 2009). Chancellor of Justice Indrek Teder told Parliament that the amendments became valid so quickly that their rapidity probably violated several articles of the Estonian Constitution. As a result the Parliament met twice more, extended the rate increase by one month, and offered a compensation system for businesses hurt by the change.

[9 ]The current anti-fraud effort had its genesis in a set of proposals issued by the Commission June 18, 2001. COM(2001)294 final available at: http://eurlex. The proposals on administrative

cooperation and mutual assistance were adopted as Commission Regulation 1798/2003, 2003 O.J. (L 264) 1 available at: and

Council Directive 2003/93/EC, 2003 O.J. (L 264) 23 available at: http://eurlex.

10 For example, effective January 1, 2010 Maine will expand the sales tax base to include (1) amusement, entertainment, and recreation services; (2) installation, repair, and maintenance services; (3) transportation and courier services; and (4) personal property services. In addition, “taxable service” will also include the rental or lease of tangible personal property with respect to leases that are entered into, extended, or renewed on or after April 1, 2010.

[11] Andrew W. Swain & Helder A. Pinto, State Interest Grows in Tapping Out-of-State Sellers for Revenue, 53 STATE TAX NOTES 505 (Aug. 24, 2009) (indicating that at least ten states have considered or are considering following New York’s extension of nexus in the area of affiliate relationships).

[12] Fiscalis Committee Project Group 12, Cash Register Project Group, Cash Register Good Practice Guide, (Dec.2006) (unpublished report on file with author). See also: Working Group on Cash Registers: Interim Report (Mar.16, 2005) (Ger.) (on file with author).

[13 ]The German solution involves encrypting critical data from the ECR on smart cards securely embedded in ECRs. The German National Metrology Institute (PTB: Physikalisch-Technische Bundesanstalt) is the home of the INSIKA project (Integrierte Sicherheitslösung für Kassensysteme – Integrated Security Solutions for Cash Registers). INSIKA began work on prototypes of the solution in 2008. Luigi Lo Iacono, Christoph Rulans & Norbert Zisky, Secure Transfer of Measurement Data in Open Systems, 28 COMP. STANDARDS & INTERFACES 311 (Jan. 2006); SELMA Project (in German) (last visited Feb. 12, 2008).

[14]See: District Court of Rotterdam, LJN: AX6802 (Jun 2, 2006) available at: (in Dutch) (translation on file with author); appealed to the District Court of The Hague where the judgment is upheld LJN: BC5500 (Feb. 29, 2008)(involving a café, Dudok, that skimmed cash sales with the assistance of pre-programmed fraud systems installed in Straight Systems ECRs) available at: (in Dutch) (translation on file with author); LJN: AT 5876, District Court of Arnhem (Jul. 27, 2005) (Microcraft Software which developed Analyse (aka, CX Analyse and Retail) as a management information system for grocery stores, butchers and bakers, but which also had a secret combination of key stokes with which the user could then indicate a percentage of turnover that would be skimmed) (in Dutch) (translation on file with author).

[15]Effective January 1, 2010 all cash registers in Sweden must be certified by the government as tamper-proof. Legislation was enacted by the Swedish Parliament (2007:592). On 12 January 2009 the Tax Agency enacted a series of regulations on the requirements for cash registers (SKVFS 2009:1) on the use of cash registers (SKVFS 2009:3), and on the control unit for cash registers (SKVFS 2009:2).

[16] The $20 billion figure is an extrapolation based on a GDP (purchasing power parity) comparison between the Canadian Province of Quebec and each of the States that have a RST. It assumes that the 7% QST is roughly equivalent to the RST rate in each of the States, and that the tax bases are roughly equivalent. The Quebec study examined only fraud in the restaurant sector, and detected $425 million in fraud.

[17]See: Richard T. Ainsworth, Zappers and Phantomware: Are State Tax Administrators Listening Now? 49 State Tax Notes 103 (July 14, 2008); Richard T. Ainsworth, Zappers: Technology-assisted Tax Fraud and the SSUTA and Encryption Solutions 61 The Tax Lawyer – State and Local Edition 1075 (2008).

[18]The two cases are: (1) Stew Leonard’s Dairy (a local grocery chain associated at one time with a dairy farm) in Norwalk Connecticut skimmed an estimated $17 million in receipts over a ten year period. See, DEPT. OF THE TREAS., I. R. S. 75 YEARS OF CRIMINAL INVESTIGATION HISTORY (1919 – 1994) 146, available at (last visited Aug. 28, 2009). (2) The La Shish restaurant chain in Detroit, Michigan skimmed more than $20 million in cash sales over a four year period and allegedly sent the funds in small denomination cashier’s checks to Hezbollah in Lebanon. See, Press Release, U.S. Dept of Justice, Eastern District of Michigan, Superseding Indictment returned Against LaShish Owner (May 30, 2007) available at: (last visited Aug. 28, 2009).

[19]RECAST VAT DIRECTIVE (RVD), Council Directive 2006/112/EC on the Common system of value added tax, O.J. (L 347) 1.

[20] RVD Art. 9(1). ‘Taxable person’ shall mean any person who, independently, carries out in any place any economic activity, whatever the purpose or results of that activity.

[21] RVD Art. 138(1).

[22] RVD Art. 197(1)(b).

[23] RVD Art. 43 sets out the main rule for all services. RVD Art. 56(1) contains exceptions for intangible services – there are additional exceptions applicable to other services. The main (and therefore fall-back) rule deems all services to be supplied at the seller’s location. RVD 56(1) isolates miscellaneous services (sometimes called intangible services) and provides a different proxy. Intangible services are deemed to be supplied at the buyer’s location.

[24] 504 U.S. 298 (1992).

[25] Id., at 317.

[26] Attributional (agency) nexus takes many forms, but all center on the degree of control the out-of-state seller exercises over an in-state entity. Simply common ownership is not enough. RESTATEMENT (SECOND) CONFLICT OF LAWS, §52 at comment b. (“[J]uridical jurisdiction over a subsidiary corporation does not of itself give a state juridical jurisdiction over the parent corporation. [However], juridical jurisdiction over a subsidiary corporation will … give the state juridical jurisdiction over the parent corporation if the parent so controls and dominates the subsidiary as in effect to disregard the latter’s independent legal existence.”)

[27]Requiring the drop shipper to collect tax is problematical. It is unlikely that the manufacturer or wholesaler would know the tax base (the price charged the in-state customer by the out-of-state retailer). Nevertheless, a number of states have taken this approach: WIS. STAT. § 77.51(14)(d); WIS. ADMIN. CODE TAX 11.94(1)(e); CONN. GEN. STAT. § 12-407(a)(3); CONN. LEGAL RUL. 2003-2 (May 30, 2003); CONN. LEGAL RUL No. 90-16, (Feb. 5, 1990); RI Gen. Laws §44-18-8; KAN. REV. DEP. PUB. NOTICE No. 03-09 (June 25, 2003); MASS. DEPT. REV. DIR. 86-5; CAL. SALES & USE TAX REG.§ 1706(c)(2); NEV. REV. STAT. § 372.050(2);NEV. REV. STAT. § 374.055(2).

[28] Kristen A. Parillo, Japanese Tax Officials Seeking $119 Million in Back Taxes From Amazon, TAX NOTES INT’L (July 8, 2009) Doc 2009-15254; 2009 WTD 128-3. Two Japanese subsidiaries of Amazon – Amazon Japan and Amazon Japan Logistics – managed sales, merchandise distribution, and logistics operations in Japan. The sales income from the Japanese business was reported in the United States by International Sales.

The Tokyo Regional Taxation Bureau reportedly determined that the Japanese operations constituted a permanent establishment of International Sales in Japan under the Japan-U.S. income tax treaty and that the income should have been reported in Japan. Japanese tax officials found that employees of Amazon Japan Logistics received instructions from International Sales via e-mails at a distribution center in Ichikawa, Chiba Prefecture, and that approval from the U.S. company was required when the distribution center assigned its employees to new posts, according to media reports.

[29] Robert J. Dolan & Youngme Moon, Pricing and Market Making on the Internet, HARVARD BUSINESS SCHOOL NOTE 20 (No. 9-500-065, Nov. 9, 2000).

[30]Robert J. Dolan, Name Your Own Price, HARVARD BUSINESS SCHOOL TEACHING NOTE 2 (No. 5- 501-046, Jan. 5, 2001).

[31] Jay Walker, the founder and vice chairman of explains what a “demand collection system” means as follows: In the traditional model of commerce, a seller advertises a unit of supply in the marketplace at a specified price, and a buyer takes it or leaves it. Priceline turns that model around. We allow a buyer to advertise a unit of demand to a group of sellers. The seller can then decide whether to fill that demand or not. In effect, we provide a mechanism for collecting and forwarding units of demand to interested sellers – a demand collection system. N. Carr, Forethought – Redesigning Business: A Conversation with Jay Walker, HARVARD BUSINESS REVIEW 19 (Nov.-Dec. 1999).

[32] Under the system consumers could only offer to purchase round-trip tickets, and could only specify the day of travel. The consumer would provide a credit card which would be charged immediately (without possibility of refund) when a qualifying flight was found. Consumers are only allowed to make one offer in a seven day period. These rules forced the customer to make a serious “best offer.” It standardized demand. Because the seller (airline) is anonymous it gets two clear benefits (in addition to incremental sales). First, it gets a brand shield. If it had publicly advertised a lower price for its product, it would have eroded brand value. Secondly, it gets a price shield. It can maintain the integrity of its established prices because it never advertises that a lower price can be secured.

[33] Press Release, issued U.S. patent #5,794,207 for the world’s first buyer-driven e-commerce system, available at: (last visited, Jun. 27, 2009); U.S. Patent No. 5,794,207 (issued Aug. 11, 1998) Method and apparatus for a cryptographically assisted commercial network system designed to facilitate buyer-driven conditional purchase offers (filed Sept. 4, 1996) available at:


[34] RVD Art. 310.

[35] If a business traveler was involved then the travel agent would not make purchases in its own name. Instead the agent would disclose that they were purchasing on behalf of another (RVD Art. 309), set up a suspense account for those purchases (RVD Art. 153), and then pass these invoices to the business traveler along with a separate invoice (and VAT) for its services.

36 Assuming that the train travels 30% in Belgium, 40% in Germany, and 30% in Denmark, then the VAT on the 50 euro train ticket is 10.70 (50 x 30% x 21% = 3.15 Belgian VAT + 50 x 40% x 19% = 3.80 German VAT + 50 x 30% x 25% = 3.75 Danish VAT); the VAT on the 30 euro hotel reservation is 7.50 (30 x 25% = 7.50); the VAT on the 20 euro meals is 3.80 (20 x 19% = 3.80).

[37] 127 is the sum of 100 in direct ticket costs, plus 22 in non-deductible, non-refundable VAT charged to the travel agent on these tickets, plus 5 in office operating expenses.

[38] RVD Art. 308.

[39] Intra-community transportation is subject to VAT in proportion to the distance traveled, RVD Art. 46.

[40] Lodging is sourced to the location of the hotel, RVD Art. 47.

[41] Meals are sourced to the place of establishment of the restaurant, RVD Art. 43. This rule is scheduled to change on January 1, 2010 so that the place of supply will be the place where the restaurant is located COUNCIL DIRECTIVE 2008/8/EC amending Directive 2006/112/EC as regards the place of supply of services, Sec. 7, (new Art. 57) available at:

[42] Commission proposal for a Council Directive amending Directive 77/388/EEC as regards the special scheme for travel agents, 2002 O.J. (C 126 E) May 28, 2002 [not published in Official Journal]; COM(2002)64 final.

[43] RVD Art. 43.

[44] 30% is the figure used by Robert J. Dolan, in the HBS case study Name Your Own Price, at 7, referencing the Form 10-Q (Nov. 15, 1999) at the Overview, which stated: [W]e also analyze the percentage of “reasonable” offers that we are able to fill. We consider an offer for an airline ticket to be “reasonable” when it is no more than 30% lower than the lowest generally available advance-purchase fare for the same route. Using this standard, the overall percentage of offers considered reasonable for the nine-month period ended September 30, 1999 was approximately 50.0%. now advertises 40% reductions in airfare, 50% reductions in hotels, and 30% reduction in car rentals.

[45]Assuming that the train travels 30% in Belgium, 40% in Germany, and 30% in Denmark, then the VAT on the 35 euro train ticket is 7.49 (35 x 30% x 21% = 2.205 Belgian VAT + 35 x 40% x 19% = 2.660 German VAT + 35 x 30% x 25% = 2.625 Danish VAT); the VAT on the 21 euro hotel reservation is 5.25 (21 x 25% = 5.25); the VAT on the 14 euro meals is 2.66 (14 x 19% = 2.66).

[46 ]38.6 = 30 (tickets and reservations savings) + 6.60 (VAT savings on these purchases) + 2 (savings in the cost of operations).

[47] Full selling price of 153.45 reduced by the cost of ticket and reservation (35 + 21 + 14), reduced by the VAT charge (15.40), and reduced by the cost of office operations (3).

[48] 28.45 is the sum of the original margin 23, plus the cost savings from moving to a low cost third-country jurisdiction of 2, plus cost saving of the UK VAT that is no longer due on the margin of 3.45.

[49]15% x 59.6 = 8.94. 59.6 (150 – 90.4) is the result of subtracting all costs (70 for tickets and reservation + 15.40 VAT on those purchases + 5 UK operating costs) from the VAT exclusive cost of the travel package.

[50] The UK VAT is higher in this example because the UK is effectively able to tax the benefit of the lower cost structure in the third country. In other words, assuming the maximum margin (selling the package tour for the market rate of 153.45) the 15% UK VAT would require a payment of 9.7575 based on a margin of 65.05. It seems something of an anomaly, but the UK will collect more VAT (not less) if its travel agents adopt a business model and move off shore (the difference is between 9.7575 and 8.94).

51 Under the same assumptions, a margin of 65.05 would return a Swedish VAT obligation of 16.2626.

[52] Id., at Art.1, proposed Art. 26(2)(b) (applying an EU-wide “use and enjoyment” rule, such that the purchase of travel services from a non-EU travel agent that would be used and enjoyed in any EU jurisdiction would sufficient to require VAT to be applied to the travel agent’s margin at the rate applicable in the traveler’s jurisdiction).

[53] The normal rule in the RST is that sales of tangible personal property to final consumers are taxed – unless specifically excluded, but that sales of services are excluded – unless specifically included. JEROME R. HELLERSTEIN &WALTER HELLERSTEIN, STATE TAXATION, available at 2009 WL ST TAXN WGL ¶ 12.05;

[54] Interstate Commerce Commission Termination Act of 1995, 49 USCA § 14505 (prohibiting states and governments from levying a sales tax on inter-state passenger travel after the Supreme Court had determined that such a tax would be constitutional in Jefferson Lines Inc. v. Oklahoma State Tax Commission, 514 U.S. 1135 (1995)).

[55 ]Robert J. Dolan,

[56 ]There are significant cash flow advantages to this structure for the intermediary. Payments are made by the traveler considerably in advance of the invoice from the retail merchant, the due date of the related tax returns and the remission of taxes collected.

[57] Texas is a notable exception. At the state level collecting and remitting the hotel occupancy tax is frequently the responsibility of both the hotel and the virtual intermediary – the hotel’s responsibility is for the tax on the discounted price and the virtual intermediary for the margin. See: Texas Comptroller of Public Accounts, Texas Policy Letter Rulings 200603503L (Mar. 20, 2006); 200310132L (Oct. 7, 2003); 200307995L (Jul. 10, 2003); 200212648L (Dec. 19, 2002); 200209424L (Sept. 12, 2002); 200208379l (Aug. 22, 2002).

[58] Because rates are in fact not uniform, and because some parts of a package tour are not taxable (either because they are inter-state passenger travel or non-taxable services) states and localities will face a further problem with the allocation of the margin among the various parts of the whole package. Internet travel companies will be inclined to allocate more of the margin to the no-tax/low-tax elements.

[59] This is where the analogy to drop shipments is most appropriate. The hotel operator only knows the discounted price of 21. It has no access to the markup. Only the virtual intermediary has this information. In the instance of the restaurant selling meals for discounted amounts, it is likely that the bill sent to the intermediary will determine the tax based on the retail price of the meals.

[60] In certifying a class action against all the major virtual intermediaries in the travel agent industry Judge Orlando L. Garcia characterized the nexus argument based in Quill as a “red herring.” He stated:

Defendants contend they cannot be “taxed” on activities in cities with which they have no substantial nexus. Again, this argument is a red herring because the occupant of the room (who is the taxpayer) is already being taxed, and the Defendants have already been collecting and remitting taxes on the rooms they sell. The only question in this lawsuit is whether the Defendants have an obligation to collect and remit occupancy tax on the higher sell rate, rather than on the lower net rate. If Defendants believed that they had no obligation whatsoever to collect and remit occupancy taxes, they would not have been doing so. City of San Antonio v., 2008 U.S. Dist. LEXIS 58793 (W.D. Tex., May 27, 2008) at *55..

[61] Jay Walker, the founder and vice chairman of explains the brand shield and the price shield of a virtual intermediary as follows:

First, it gets a brand shield. If it had publicly advertised a lower price for its product or service, it would have eroded its brand. But since it can accept the unit of demand without letting the buyer know the brand in advance, it suffers no such reduction.

Second, the seller gets a price shield. It can maintain the integrity of its established prices because it never advertises that a lower price is being filled. The seller avoids the problem of free riders – people who take a discounted price even though they would have been willing to pay the full price.

N. Carr, Forethought – Redesigning Business, at 19. See the HBS Case discussion of this concept at Dolan, Name Your Own Price, at 3, and the HBS Teaching Notes at Dolan, Name Your Own Price, at 3.

[62] Karen Setze, Expedia Pays San Francisco Hotel Tax, 2009 STT134-2, DOC 2009-16097 (July 16, 2009)

(indicating that Expedia paid “about $35 million in disputed hotel taxes after a the California Second District Court of Appeals refused to stay a trial order demanding the company pay the taxes before further appeal); Expedia, Inc v. Superior Court of Los Angeles, case number B217215 (July 9, 2009) record available at: But there are many more jurisdiction waiting. See Jonathan O’Connell, D.C. Could be Losing Hotel Taxes to Online Companies, WASH. BUS J. (May 29, 2009) available at:

D.C., as one of the nation’s top tourist destinations, could be owed more than $100 million in back taxes and penalties but — despite an anticipated budget deficit of $967 million in fiscal 2011 — it has yet to join the fray.

[63] Research Division, Texas Legislative Council, Overview of Local Taxes in Texas (Nov. 2002) at 5 & Table 1 (indicating that in addition to the state hotel tax of 6% there are 20 counties, 392 cities and 5 special districts that impose a hotel tax with county rates ranging from 1% to 8%; city rates ranging 2% to 13%. Because these jurisdictions overlap there is additional complexity in the Texas system. With one exception there is no maximum combined rate. The exception involves the extraterritorial jurisdiction of cities with a population base under 35,000.

In this case the combined rate may not exceed 15%) available at:

[64] Defendants in the suit were:, L.P.;, G.P., LLC, Hotwire, Inc., Expedia Inc.,, Inc., Cheap Tickets Inc., Interwork Publishing Corp. d/b/a, Orbitz, Inc., Orbitz, LLC, Lowest Fare.comr, Inc., Inc. Travelweb, LLC, Manupintour Holding, LLC, Site,, Inc. and, LP.

[65] City of San Antonio, Tx. v., 2007 U.S. Dist. LEXIS 39757 (W.D. Tex., Mar. 20, 2007).

[66] The other counties, cities, and special districts with occupancy taxes cannot be joined because of local exhaustion of remedies issues. The class specified makes this clear:

Texas cities whose ordinances contain language that requires every person owning, operating, managing or controlling any hotel to collect and remit hotel occupancy taxes and whose ordinances do not contain administrative prerequisites to filing a lawsuit for the failure to collect or remit hotel occupancy taxes. 2008 U.S. Dist. LEXIS 58793 (W.D. Tex., May 27, 2008) at *13.

[67] John A. Swain, Inernet Travel Companies – Taxing the Middleman, State Tax Notes 477 (Feb. 14, 2005) (indicating that 20% was the number for the US in 2003).

[68 ]JAPAN STATISTICAL YEARBOOK 2009, Ch.13, Table 13-10 Number Of Establishments By Sales Group and Sales in Service Industries (Individual Proprietorships and Companies) (2004) available at.

69 5% x ¥1,326,584 = ¥66,329.

[70] Assumes (a) rooms were discounted when released to the virtual intermediaries by 50% from the published price, and (b) that guest paid 30% less than the published price for these rooms. 2 x ¥1,326,584 million = ¥2,653,168 million and 70% of ¥2,653,168 = ¥1,857,218 million.

[71] 5% x ¥1,857,218 million = ¥92,861 million.

[72 ]¥92,861 million less ¥66,329 million = ¥26,532 million (roughly $292 million).

[73] In correspondence with National Tax Administration officers the following solution has been suggested. Article 6(2)-7 of the Cabinet Order of the Consumption Tax denies a deduction for services where the place of supply is unclear. This may be the case here. If guests pay the hotel directly (instead of the virtual intermediary), and if the virtual intermediary is paid by the hotel (much like an “introduction fee” would be paid) then it is likely that the CT will tax this service. See: JAPAN’S REVISED CONSUMPTION TAX LAW (SHOUHIZEIHOU), LAW NO. 108, 1988, and APPENDIXES; CABINET ORDER (SHOUHIZEIHOU SEKOUREI) NO. 360, 1988 available at:

idxsearch.cgi (in Japanese). The difficultly with this solution is that virtual intermediaries (not the hotels) normally collect the full room charge. All fees are collected in advance. The hotel bills the virtual intermediary only for the agreed discounted price, and the hotel has no knowledge of the price actually paid by the guest. Other discussions with Japanese tax professionals have suggested that (working entirely within the current system) administrative rules could be adopted (and may be necessary) requiring hotels to collect data (from the guest or from the virtual intermediary) that would specify the amount actually paid for the accommodations. Depending on the location of the virtual intermediary exchange of information provisions in treaties could be utilized secure this data.

[74] Travelscape, LLC v. South Carolina Department of Revenue, Docket No. 08-ALJ-17-0076-CC (Feb. 12, 2009) available at Tax Analysts Document Service Doc 2009-3711and The aggregate assessment (July 1, 2000 – June 30, 2006) for a single virtual intermediary was $6,376,454.71 ($4,120,542.24 in tax + $666,736.41 in interest + $1,589,176.06 in penalties).

[75 ]Expedia, Inc. v. City of Columbus, GA, S09A0567, 2009 Ga. LEXUS 315 (June 15, 2009); The City of

Goodlettsville, TN v., 605 F. Supp. 2d 982 (M.D. TN 2009).

[76] City of Charleston, South Carolina v., 570 F. Supp. 2d 757 (2007); Town of Mount Pleasant, South Carolina v., 586 F. Supp. 2d 538 (2008); City of Findlay, Ohio v., 441 F. Supp. 2d 855 (N.D. Ohio 2006); 561 F. Supp. 2d 917 (N.D. Ohio 2008); City of Gallup, NM v., 2007 U.S. Dist. LEXIS 86720 (D. N.M., 2007).

Previously published by the Boston University School of Law, October 2009

Adjunct Faculty, Boston University - School of Law, Boston, MA, USA