In the past decade online sales have grown exponentially. In 2005 there were about $100 billion online sales in the United States, an impressive number for a sector which two decades ago was non-existent. At the moment online sales still account for about 3% of retail sales in the United States, but this proportion is bound to increase significantly in the near future. Economists have therefore been looking with keen interest at the factors that may explain the growth of online sales. Furthermore, the excellent quality and richness of the data which may be obtained in this sector provide empirical answers to more general questions on consumer’s behavior; in particular, in this way various hypotheses concerning the rationality of consumers’ decisions may be tested.

One crucial issue regarding online sales revolves around their tax treatment. In the United States there is no nationwide federal sales tax as there is in European nations. However, most states collect local sales taxes, with rates ranging between 4% and 9%. In principle sales taxes should be paid for online sales too, but in practice this only occurs when an online retailer sells to customers living in the state were it is located.

A Supreme Court sentence established that a seller can be forced to collect sales taxes only if it has a “substantial presence” in the state, and a federal law established that web presence cannot be construed as substantial presence. This implies that an online retailer located in, say, California can only be forced to collect sales taxes when it sells to California residents. If an online retailer located in California sells to a New York resident then, in principle, the buyer has to report the transaction to the state tax authority and pay sales taxes; very few buyers do this, so online sales through websites to residents of another State is de facto free from the sales tax. How important is this tax advantage vis-à-vis traditional, brick-and-mortar retailers?

The authors collected an original dataset providing a number of insightful answers. They looked into the sale of computer memory modules using the Pricewatch.com search engine. The engine lists the before-tax prices of memory modules in real time in ascending order, also naming the state where the online seller is located. Since consumers look at before-tax price and state of origin, they are easily able to figure out what the after-tax price will be. This will be the same as the before-tax price if the state where the seller is located is different from the buyer’s; while the appropriate sales tax will have to be added if seller and buyer are located in the same state.

Data were collected on an hourly basis; for about one year, from May 2000 to May 2001, the list of prices advertised on Pricewatch.com was downloaded at the beginning of each hour. It was a period when memory module prices were very rapidly decreasing and quite variable. Besides the data on prices the authors collected data on the volume of sales of two websites, both California-based. Even though out of a couple of dozens only two websites were observed, the data are very useful, including as they do information on the location where the goods where shipped to. This makes the dataset ideal to test the relevance of taxes in the demand for memory modules.

The paper finds strong empirical evidence that sales taxes are important. The authors estimate that whenever a state (other than California, where the two e-retailers are based) increases the tax rate by 1%, the two websites see an increase in demand from that state of about 6%. The tax avoidance motivation appears therefore to be quite important in determining the level of demand. In order to have a better idea of the magnitude of the tax effect, it should be considered that the average sales tax rate (across all states) is 5.7%. Thus if offline sales taxes were eliminated, online demand would decrease by over 30% - an impressive amount. Since most sales of the two California-based websites are outside California, this is strong evidence that it is taxes rather than consumer preferences that are, in fact, behind this result.

The data also suggest that most of the additional demand for online sales is likely to come from consumers who would otherwise buy in traditional brick-and-mortar stores, rather than from consumers who would otherwise refrain from buying the product. Price sensitivity of demand is therefore very high, and standard economics suggests that it is unlikely that it should stem from macroeconomic effects. The authors therefore conclude that its origins are consumers who substitute online purchase for a similar product - in this case, memory modules sold in traditional retail stores.

Another intriguing result is that taxes, while important, are less important to consumers than the before-tax price. When separately estimating the impact on demand of before-tax price and sales tax, the authors find that a one dollar increase in taxes reduces demand less than a one dollar price increase. This seemingly puzzling result provides good evidence that it is difficult and costly for consumers to gather and process information about after-tax price structure, since a fully informed (or fully rational) consumer would only consider the after-tax price and would be indifferent to the source of its increase.

However, since the prices quoted on Pricewatch.com are before-tax prices, consumers have to compute the after-tax price applying their home state’s tax rate. It appears that consumers do this in a somewhat imperfect way. Probably when doing so many consumers are aware of living in states with high sales taxes, so they tend to prefer buying from out-of-state vendors. At the same time they don’t exactly compute the amount of tax that they pay, so changes in the tax rate have a smaller limited impact on demand than changes in before-tax prices.

Finally, the authors’ estimates suggest that, other things equal, consumers have a preference for buying in their home state. More precisely, consumers are prepared to pay approximately two more dollars (about 3% of the average price) if the seller is in their home state. The cause of this is interpreted to be the fact that vendors in home states have a shorter shipping time, a feature that consumers appreciate.

On the whole the results suggest that while online retailing owes much of its success to tax avoidance motives, there are still certain dimensions in which online and offline retailers may compete. Good news for consumers.