Internet Commerce Estimates: A Reality Check

By Kevin O'Neil(*)

From Collecting Taxes in the Cyberage
p. 23-32, published 1999

Internet commerce is booming in the United States. A University of Texas study estimated that sales over the Internet produced over $100 billion in revenue worldwide during 1998.(1) This figure, while impressive, still represents only about 1 percent of the United States’ Gross Domestic Product, and it appears that the economic potential of Internet commerce has yet to be realized.

The exact potential, however, is impossible to predict, and growth estimates vary significantly. In 1998, the Organization for Economic Co-Operation and Development compiled predictions from 12 reputable consulting firms for the annual value of Internet commerce in the years 2000 through 2002. All firms predicted high growth rates, but they still varied widely in their estimates of the total value of Internet-based commerce. The figures ranged from 6 percent to 1000 percent of the median prediction, differing by as much as $1.5 trillion.(2) More recent studies show no greater consensus about the future of Internet commerce, and frequently they do not even agree on its present value.

The growth of Internet commerce is of concern to the state because of the likelihood that sales or use taxes are not collected on most Internet purchases, even when tax is legally owed. Kentucky citizens owe sales or use tax on any purchase of a taxable good, even if the seller is located outside the state. This applies to catalog and other traditional mail-order sales, as well as Internet sales.

The state can only require out-of-state businesses to collect and remit use tax if the company has a physical presence, legally referred to as a “nexus,” in Kentucky. For instance, if an out-of-state company has stores, offices, or a distribution center in Kentucky, it has nexus. However, nexus is not established if a company’s only presence in Kentucky is via a third party it hires to deliver goods, such as UPS or Federal Express. Sellers who have nexus in Kentucky are legally required to collect sales tax on transactions and remit it to the state.

Businesses may elect to collect and remit use tax even if they have no nexus in Kentucky, but they are under no legal obligation to do so and few do. If the seller does not collect sales or use tax, consumers are legally required to report purchases on their annual income tax return and pay use tax on them. It appears that few do. Ignorance of the law and the onerous record keeping demanded ensure that this method collects little of the tax owed. Indeed, the U.S. Advisory Commission on Intergovernmental Relations estimates that only 16.5 percent of taxable business-to-consumer e-commerce sales result in sales and use tax payments.(3)

Thus, whenever a Kentucky consumer chooses to buy a taxable item from an out-of-state mail order company instead of a local storefront retailer, the state may be losing tax revenue. The method through which the order is made is not important: whether the customer uses mail, phone, fax, or the Internet to make the order, the state still may be losing sales tax. In this report, we are concerned only with commerce conducted via the Internet. Although tax can be lost on both catalog and Internet sales, more traditional forms of mail order commerce are growing only slowly, while Internet commerce is booming. We seek to assess the impact this boom will have on Kentucky’s future revenues.

Purpose

Given the difficulty of predicting how much commerce will occur over the Internet, our primary goal is not to predict precisely how much sales tax revenue will be forgone as a result of this commerce. Instead, we present a reality check, tying projected losses in sales and use tax to consumer purchasing behavior. Our study seeks to answer the question, “How much would people have to buy over the Internet for the state of Kentucky to suffer significant losses in sales tax revenue?” Specifically, we use the three scenarios in the previous chapter to determine whether the resulting changes in consumer expenditure behavior make sense.

How should we look at these results?

Our models take various levels of Internet spending by household consumers(4) and show, given a specific use tax compliance rate, the resulting loss of sales and use tax revenue. We also present the results as the amount of Internet spending households would have to do under the three scenarios presented in The Internet as a Virtual Tax-Free Zone: Implications for the State Budget.

We used Consumer Expenditure Survey (CES) data as the starting point for our models.(5) The CES provides detailed information about how much households spend annually on specific categories of goods and services, and does not incorporate business-to-business sales. We took the CES data on the average amount of money households spend in each category, projected it to the year 2003, and multiplied it by the state’s projected household population. The result enables us to estimate the level and type of consumer spending in Kentucky.(6)

The best way to use our data may be as a reality check. The results can answer the question, “How likely is it that in 2003, I and my fellow Kentuckians, including both computer lovers and technophobes, will be using the Internet to make this many of our purchases?” It may be very difficult to predict exactly, but this type of check can be used to ground our thinking on Internet commerce.

What We Found

Two Scenarios

Scenario One: All Products and Services, 2003

How much sales tax revenue might the state lose if Kentuckians made 1 percent of all their consumer purchases on the Internet? What if they made 5 percent of their consumer purchases on the Internet? In Scenario One, we kept the percentage of expenditures made over the Internet constant across CES categories. Of course, the majority of these expenditures are not taxable and do not contribute to our estimates for forgone revenue. All taxable expenditures made on the Internet were considered a potential source of revenue loss. Modeling Internet expenditures as a fixed percentage of all expenditures is a concession to the difficulty of predicting how people will spend money there. Currently, some products and services sell much better over the Internet than others, but overall a diverse range of products and services are sold electronically, with new markets being opened constantly. In time, people may come to spend money on the Internet much as they do in traditional stores. The first scenario assumes this will be the case.

In Table 1, we see the results of Kentucky households making various percentages of all of their expenditures over the Internet for the year 2003. The results of these various buying behaviors are presented as a dollar figure for forgone revenue. This figure is also presented as a percentage of total projected sales tax receipts. For example, one could read this chart as “If Kentucky households on average made 1 percent of all expenditures over the Internet, the state would lose 11.9 million dollars (at 1 percent compliance), 10.8 million dollars (at 10 percent compliance), or 6.0 million dollars (at 50 percent compliance).” As this example illustrates, the compliance rate makes a big difference in the amount of lost sales tax.

Table 1: Results of the All Products and Services Scenario

We use this method to test the viability of the three scenarios presented in The Internet as a Virtual Tax-Free Zone: Implications for the State Budget. We presented a low, medium, and high scenario in that chapter and estimated a range of sales tax losses associated with each. We estimated a $2.2 million loss of sales tax revenue by 2003 in the low scenario, a $19 million loss in the medium scenario, and a $46.4 million loss in the high scenario.(8) Are these plausible scenarios, considering the amount of consumer expenditures on the Internet that each implies? The answer to this question is necessarily subjective, but the answer appears to be a “cautious yes.”

Figure 1 shows, for example, that in the “high scenario,” which assumes a $46.4 million sales tax loss, Kentucky household consumer expenditures on the Internet would have to equal 4 percent for all CES categories, or about $1,500 in annual expenditures. This does not appear to be an implausible figure, especially when we consider it is for 2003. However, we also have to consider that this is an average figure for all Kentucky households. One must remember that not all Kentucky households have access to the Internet, and that not all those with Internet access use it to make purchases. In 1998, about 42 percent of Kentucky adults had ever accessed the Internet. And in April 1999, a Nielsen Media study estimated that of consumers who regularly use the Internet, only 30 percent had ever purchased a product over the Internet. Results of a 1999 poll of Kentucky adults show that about 18 percent of them have bought something over the Internet. And a Forrester Research survey found that the average American who does make purchases on the Internet spends $322 a year.(9)

Figure 1: Estimated Annual Household Consumer Expenditures on the Internet Implied by the Three Scenarios

Scenario Two: Selected Products and Services, 2003

The “All Products and Services” model of Internet commerce assumes that all types of goods and services sell equally well over the Internet. This is not a realistic assumption. Currently, some markets have blossomed on the Internet, while others have failed. Products that are unique, unfamiliar or that vary piece-by-piece in quality or character are unlikely to sell well over the Net. Heavy or bulky items that cannot be easily shipped, such as furniture or major appliances, also do not represent a major threat to sales tax revenue. Even if these items were sold over the Net, a local distributor would likely be used to deliver them, giving the state nexus. The ultimate Internet commodity is of homogenous and easily certifiable quality, has a high value-to-weight ratio, and can be marketed effectively using computer graphics and sound. In addition, if the product line includes a wide selection of styles or options, an Internet retailer may have a scale advantage over smaller local shops. It should also be noted here that items such as music and software, although taxable if delivered physically on disk or CD, are not taxable under Kentucky law if they are downloaded electronically.

Music is a prime example of an Internet marketable product. An Internet user can quickly use a Web database to search thousands of albums, as opposed to browsing through a record store, which may have a smaller selection and lack a searchable database. At the same Web site, she can read online reviews, see pictures of the album liner art, and download audio samples. Once she has selected an album, she can purchase it on Compact Disc, an easily shipped format of known quality. As music increasingly becomes available by electronic download, eliminating shipping costs and time, it will become even more attractive as an Internet commodity. Further, downloaded music is not subject to Kentucky sales tax. Software and financial services are other examples of intangible items that can be delivered electronically and thus have boomed on the Internet.

To model more accurately the current state of Internet commerce and its likely avenues of growth, the “Selected Products and Services” scenario restricts Kentuckians’ Internet purchases to items that already have significant or growing online sales. We picked, from the Consumer Expenditure Survey’s 70 categories, only categories that contained goods or services that currently are sold in significant quantities over the Internet or that have demonstrated concrete potential for Internet sales. Below is a list of CES categories we consider to have high potential for Internet sales and some of our justifications for picking them.

Selected Categories, Potential Sources of Revenue Loss

These are categories that contain goods that are expected to sell well over the Internet and are taxable. The state will be unable to collect sales tax on these goods and will lose revenue due to Internet sales.

We also identified a few categories containing goods with high Internet sales potential that do not present a source of sales tax loss, either because they are exempt from the sales tax or because the particulars of the goods give the state nexus to enforce tax collection. Although these categories do not contribute to our estimates of tax loss, we use them in our calculations for comparing estimates of overall electronic commerce values. Some of these categories include financial services,(26) travel services,(27) fees and admissions,(28) food,(29) and drugs.(30)

We assumed all expenditures Kentuckians made over the Internet would be in these CES categories. Only the percentage of expenditures in these selected categories was varied. In all other categories, we assumed no Internet expenditures. Overall, the categories we identified as of high Internet sales potential contained about 25 percent of all average household expenditures. Of all the selected expenditures, about 44 percent were in the taxable categories that are potentially a source of revenue loss.

Because the types of expenditure the scenario assumes will be made over the Internet have been restricted from the All Products and Services Scenario, the revenue loss from a certain percentage of goods purchased over the Internet will be less than the same percentage applied over all the categories. However, it is far more realistic to expect people to have large percentages of expenditures in these categories over the Internet than it is to expect them to have large percentages of all categories of expenditures over the Internet.

Results of the Selected Categories Scenario

In Scenario Two, forgone revenue is again directly proportional to the percentage of selected expenditures made on the Internet and inversely proportional to the compliance rate. One could read this chart “If Kentucky households make 1 percent of their expenditures in the selected categories over the Internet, the state would lose 4.6 million dollars (1 percent compliance), 4.2 million dollars (10 percent compliance), or 2.3 million dollars (50 percent compliance).” Table 2 is directly comparable to Table 1, which presents the results of the All Categories scenario. However, the percentages of expenditures in Table 2 refer to a smaller group of expenditures than in Table 1, so the sales tax losses for any given percentage of expenditures made over the Internet are about 40 percent of those in Table 1. Nevertheless, it is far more realistic to expect high Internet expenditure rates in the Selected Categories scenario than in the All Categories scenario.

Table 2: Results of the Selected Products and Services Scenario

How much of Kentuckians’ consumer expenditures in selected categories would have to take place on the Internet in 2003 for the three scenarios presented in the previous chapter? Figure 2 shows, for example, that in the “high scenario” Kentucky household consumer expenditures on the Internet would have to equal about 10 percent for the selected CES categories. Again, we estimate that the resulting sales tax loss would equal around $46.4 million in this scenario.

Figure 2: Estimated Annual Household Consumer Expenditures on the Internet for Selected Categories Implied by the Three Scenarios

Uneven Sales

Of course, even with our selected categories, consumers will buy more of some than of others. For purposes of demonstration, we further divided our categories into high, moderate, and modest Internet sales potential. We then assumed that the percentage of expenditures made on the Internet in categories of moderate Internet growth would be half that in high growth categories, and that the percentage in the modest growth categories would be one quarter that in the high growth categories. This ratio holds true until the high growth potential categories hit 100 percent of expenditures made over the Internet, at which point the other categories continue to grow in the same fixed ratios relative to each other. We present these figures as equivalents to the flat percentages of selected categories discussed in this scenario to demonstrate a realistic example of how these types of losses might occur. The percentages of high, moderate, and modest growth categories are presented according to their equivalent in overall percentage of selected categories on the Internet in terms of the resulting tax losses. Because many non-taxable goods were classified as “high potential,” the tax losses caused by categories with uneven distribution differ somewhat from those with even distribution.

Table 3 simply shows how losses might occur if various types of goods do not sell evenly. For instance, looking at Table 2, we see that if 10 percent of expenditures in every one of the selected categories is sold over the Internet, the state will lose between $23 and $46 million, depending on the compliance rate. Table 3 shows that the same tax loss would occur if 18 percent of expenditures were made in the “high potential” category, 9 percent in the “moderate potential” category, and 5 percent in the “modest potential” category. We include this chart in our discussion to show a more sophisticated model of consumer behavior. Based on current sales and trends, it may be more realistic to expect people to buy 19 percent of such goods as books and music and only 5 percent of such goods as small appliances, as Table 3 indicates, rather than expecting them to buy 10 percent of all of these goods, the equivalent in Table 2, in which all of the selected categories are presumed to sell at the same rate on the Internet.

Table 3: Equivalent Tax Losses Assuming Uneven Sales on the Internet

Conclusion

The analysis in this chapter gives us an idea of the plausibility of the three scenarios presented in the previous chapter, and it would seem that even the “High Scenario” is plausible. For example, in the worst case scenario (i.e., high), Kentuckians would have to purchase, on average, only 4 percent of all their consumer products on the Internet, or 10.4 percent of selected products (e.g., books, software, music CDs), by 2003. Remember, though, that these are averages, and a significant number of Kentuckians do not have access to the Internet.

Our results show that a tax loss equivalent to 1 percent of expected sales tax (about $26 million) could result from Kentucky households using the Internet to make as little as 2.4 percent of all their expenditures (or 6.2 percent of only the type of products most frequently purchased online).(31) It seems less likely, then, that the state would experience sales and use tax losses that would be equal to or greater than 5 percent of projected receipts in 2003. To result in a 5 percent sales tax loss, Kentuckians would have to make 13 percent of all expenditures, or 30 percent of the selected category expenditures, over the Internet. This level of spending represents a fairly aggressive extension of Internet commerce in 4 years and greatly exceeds estimates by both Forrester and the DMA. Still, Internet commerce is currently growing rapidly.

Beyond 2003, lost tax revenue due to Internet commerce could expand enormously, and the state does not have strong policy options to improve compliance. The state should take what steps it can to increase payment of use tax. Our models show that increasing the “compliance rate” decreases losses to Internet commerce. For example, in the All Products and Services Scenario, if people make 10 percent of their purchases over the Internet, the state loses $119 million if it collects 1 percent of the tax owed, but only $60 million if it collects 50 percent of the tax owed, and of course, loses nothing if all the tax is collected. The most effective way to increase compliance is to require sellers to collect and remit sales tax. This is why the compliance rate from storefront expenditures is essentially 100 percent and why transactions where nexus exists account for most of what little tax is collected from Internet and other mail-order purchases.

Unfortunately, the state presently has no legal power over out-of-state sellers without nexus in the state. Congress, however, can enforce sales taxes on sellers with no nexus. Congress has charged the Advisory Commission on Electronic Commerce with the responsibility of recommending policy on sales tax and electronic commerce. After their first meeting, several members of the Commission expressed the opinion that sales tax will eventually be applied to Internet purchases. One likely solution is to make state sales tax enforceable on out-of-state vendors and to require these vendors to use a database to match buyers with their local tax jurisdictions. Another proposed solution is to impose a national sales tax on interstate purchases that would be collected by the federal government and then distributed to the states. In the long term, Kentucky’s best policy tool is simply to monitor national progress in this area and to lobby for legislation favorable to the state. Kentucky is well-prepared for national sales tax reform, because it has no local option sales taxes, which might have to be eliminated to simplify a national sales tax system.

Our results suggest that the threat of Internet commerce is not sufficient to justify drastic fiscal changes within the next few years. The state can take some steps to improve the compliance rate, but without legislation by Congress, the policy tools at hand are of limited effectiveness. Vendors with no nexus in the state are not legally responsible for the tax, and it is difficult to enforce the tax on individual consumers. The state participates in the Southeastern States Exchange of Information Agreement. Participating states share information on high-value interstate sales and encourage out-of-state vendors to remit sales tax voluntarily. In practice, however, vendors have little incentive to collect sales tax, and the states only record a small segment of sales. The state could, in theory, try to audit taxpayers’ reports by purchasing lists of customers from out-of-state vendors, but such a solution would be prohibitively costly and raise serious privacy issues. Educating the public about the law may be the state’s best strategy for raising the compliance rate. A recent poll by the Kentucky Long-Term Policy Research Center indicated that knowledge of the law also increases a person’s willingness to pay the tax (refer to A New Way to Shop: Kentuckians Find the Web). Efforts to improve the convenience of reporting use tax might increase revenues enough to justify their cost.

  Back to The Internet as a Virtual Tax-Free Zone: Implications for the State Budget

  Ahead to Issues in Use Tax Administration: Increasing the Compliance Rate

Footnotes

* Kevin O'Neil got to know the world of Internet commerce while a summer intern at the Kentucky Long-Term Policy Research Center. He currently studies Economics and English Literature at Swarthmore College in Pennsylvania. Kevin’s other experiences include study abroad in Spain, summer work with the Lexington-Fayette Urban County Government, and volunteer work as a firefighter and teacher of English as a second language.  Return to text.

  1. Barua, A. and Whinston, A., The Internet Economy Indicators. (Austin, TX: University of Texas 1999): online, The Internet Economy Indicators Web site http://www.internetindicators.com/features.html, Internet, 21 July 1999.   Return to text.

  2. Organization for Economic Co-Operation and Development (OECD), The Economic and Social Impacts of Electronic Commerce: Preliminary Findings and Research Agenda. (OECD, Paris.1999) 27: online, OECD Web site http://www.oecd.org/subject/e_commerce/ebooks/027-054.pdf, Internet, 21 July, 1999.   Return to text.

  3. United States Advisory Commission on Intergovernmental Affairs, Taxation of Interstate Mail Order Sales: 1994 Revenue Estimates. Return to text.

  4. We assumed that consumer purchases were the primary source of sales tax loss. Business-to-business sales make up the bulk of online commerce, and sales tax collected from such sales composes much of Kentucky’s sales tax receipts. But businesses generally have better tax information, keep better records, and face higher standards of accountability than do household consumers. Individual consumers are more difficult to audit and less familiar with use tax law. Thus, consumer purchases represent the greatest potential loss of sales tax owed on traditional mail-order and Internet purchases. Return to text.

  5. Data from the 1997 Consumer Expenditure Survey (CES) were used to model Kentucky consumer spending habits. The survey is conducted by the federal government’s Bureau of Labor Statistics and is considered the most comprehensive and accurate data available on consumer spending behavior. The CES presents average expenditures per household, with a household defined as either a family living in the same house, a financially interdependent couple, or a financially independent individual. The CES reports household expenditures in over 70 distinct categories, ranging from purchases of tangible goods such as “Cereals and cereal products,” to purely financial expenditures such as “Retirement, pension, and Social Security contributions.” CES results specific to Kentucky are not available, so we used regional data for the South as a proxy. Although no data are available for household expenditures at a state level, we feel expenditure data for the South are a close enough approximation for our purposes. Kentucky’s average household expenditures are probably slightly smaller than the South’s average, however. Average household expenditures for the South in 1997 were 93 percent of the national average. In comparison, Kentucky in 1996 had per capita retail sales 94 percent of the national average (Marketing Statistics, 1997 Statistical Abstract of the United States) and a per capita disposable income of 82 percent of the national average (US Census, 1997 Statistical Abstract of the United States). Return to text.

  6. To extend the model to years after 1997, the CES data were corrected for growth in household expenditures over time. Average household expenditures have consistently risen from year to year as income and living standards have risen. Our model assumes that the average household will increase its expenditures at a constant annually compounded percentage rate based on the average growth in CES expenditures over the past five years. In our projections, expenditures in all categories grow at an annual rate of 4.37 percent. Household population projections are KLTPRC interpolations of projections made by the Kentucky State Data Center based on the 1990 US Census. Return to text.

  7. We assume a 10 percent compliance rate with the sales and use tax for these projections. Return to text.

  8. We also assumed a 16.5 percent use tax compliance rate in the low scenario, 10 percent in the medium scenario, and 4 percent in the high scenario. See Peter Schirmer, Kevin O’Neil, and Michael Childress, “The Internet as a Virtual Tax-Free Zone,” Collecting Taxes in the Cyberage, (Frankfort, KY: Kentucky Long-Term Policy Research Center, 1999). Return to text.

  9. Austan Goolsbee, “In a World Without Borders: The Impact of Taxes on Internet Commerce,” revision of NBER working paper #6863, 1999: online, University of Chicago Web site, http://gsbwww.uchicago.edu/fac/austan.goolsbee/research/intertax.pdf, Internet, 21 July 1999: 6. Return to text.

  10. OECD: 47, 52. Return to text.

  11. Clinton Wilder, “Myths and Realities” Information Week, December 7, 1998: online, Information Week Web site http://www/informationweek.com/712/12iumyt.htm, Internet, 21 July, 1999. Return to text.

  12. Robert Hof and Linda Himmelstein, “eBay vs. Amazon.com,” Business Week 31 May 1999: 129. Return to text.

  13. Nielsen Media Research, “CommerceNet and Nielsen Media Research Issue Results of Spring 1999 Internet Demographic Survey,” press release (1999): online, CommerceNet Web site http://www.commerce.net/news/press/fact0699.html, Internet, 22 July 1998. Return to text.

  14. Austan Goolsbee, “In a World Without Borders: The Impact of Taxes on Internet Commerce,” Table 2B. Return to text.

  15. Austan Goolsbee, “In a World Without Borders: The Impact of Taxes on Internet Commerce,” Table 2B. Return to text.

  16. Ernst & Young LLP, “Second Annual Ernst & Young Internet Shopping Study,” (1999): online, Ernst & Young Web site, http://www.ey.com/global/vault.nsf/US/Internet_Shopping_Study_1999/$file/internetshopping.pdf, Internet, 21 July 1999, 13. Return to text.

  17. David Leonhardt, “Checking out the Corner Cyberstore,” Business Week 10 May 1999: 130.  Return to text.

  18. Nua Ltd., “1998 US Consumer Spending Online Chart,” online, Nua Ltd. Web site, http://www.nua.ie/surveys/graphs_charts/1998graphs/consumer_spending_us.html, Internet, 21 July, 1999. Return to text.

  19. Ernst & Young LLP, 13. Return to text.

  20. For more information see: Austan Goolsbee, “Evaluating the Costs and Benefits of Taxing Internet Commerce.” Return to text.

  21. Ernst & Young LLP, 13. Return to text.

  22. Austan Goolsbee, “In a World Without Borders: The Impact of Taxes on Internet Commerce,” Table 2B. Return to text.

  23. Austan Goolsbee, “In a World Without Borders: The Impact of Taxes on Internet Commerce,” Table 2B. Return to text.

  24. Michelle Rafter, “Pets.com Fetches Investment From Amazon.com” The Industry Standard 29 March 1999, online, http://www.thestandard.com/articles/display/0%2C1449%2C4021%2C00.html, Internet, 22 July 1999. Return to text.

  25. Nua Ltd., “1997 US Consumer Spending Online Chart,” online, Nua Ltd. Web site, http://www.nua.ie/surveys/graphs_charts/1997graphs/consumer_spending_us.html, Internet, 21 July, 1999. Return to text.

  26. Boston Consulting estimated that $1.4 billion was spent over the Internet on financial services in just the first half of 1998, second only to computer goods in total consumer sales. See Austan Goolsbee, “Evaluating the Costs and Benefits of Taxing Internet Commerce,” Table 2A. Also, the OECD predicts that 60 percent of the retail stock trade will be online by 2005. Refer to OECD, 44.  Return to text.

  27. Travel expenses by both consumers and businesses accounted for $3 billion in sales over the Internet, according to Forrester Associates, and are predicted to reach $29 billion in Internet sales by 2003. See Austan Goolsbee, “Evaluating the Costs and Benefits of Taxing Internet Commerce,” Table 2B.  Return to text.

  28. Ticketmaster.com reported online sales of event tickets had risen 270 percent in 1997 to be worth over $19.8 million per quarter. Refer to OECD, 41.  Return to text.

  29. Online grocers have reported increasing sales. Online grocery sales were estimated at $235 million in 1998 and are projected to reach $11 billion in 2003, significant but still less than 2 percent of the overall grocery market. Refer to Linda Himmelstein, “Can You Sell Groceries Like Books?” Business Week, 26 July, 1999, EB44. Grocery sales are in general exempt from sales tax. A small number of grocery purchases (candy and soft drinks) are taxable. We decided to not include these purchases as possible sources of revenue loss because successful online grocers such as Peapod or Netvan are tending to use local distribution centers rather than mail delivery. A local distribution center gives the state nexus to require tax collection.  Return to text.

  30. This category includes prescription drugs that are exempt from sales tax (although there is a separate prescription tax). Thus, although prescription drugs have a high potential for Internet sales, we consider a portion of this category not to be a source of lost sales tax revenue. Return to text.

  31. We assume a 10 percent compliance rate with the sales and use tax for these projections.  Return to text.