Consider the case of aspirin. Suppose, for the sake of argument, that Bayer aspirin and store brand aspirin are identical products, but that Bayer costs twice as much because some consumers believe (falsely, in this example) that Bayer is better. Would we expect markets to eradicate this price difference? Since the Bayer brand name is trademarked, it is not (legally) possible to go into the business of buying the store brand aspirin and repackaging it in Bayer bottles. This inability to transform the store brand into Bayer prevents one method arbitrageurs might use to drive the two prices to equality. Another possibility for arbitrageurs would be to try to sell the more expensive Bayer aspirin short today, betting that the price discrepancy will narrow once the buyers of Bayer “come to their senses.” Short selling works like this: an arbitrageur would borrow some bottles from a cooperative owner, sell the bottles today and promise the owner to replace the borrowed bottles with equivalent Bayer bottles in the future. Notice that two problems impede this strategy. First, there is no practical way to sell a consumer product short, and second, there is no way to predict when consumers will see the error in their ways. These problems create limits to the forces of arbitrage, and in most consumer goods markets, the Law may be violated quite dramatically.

The aspirin example illustrates the essential ingredients to violations of the law of one price. First, some agents have to believe falsely that there are real differences between two identical goods, and second, there have to be some impediments to prevent rational arbitrageurs from restoring the equality of prices that rationality predicts. Can these conditions hold in ? financial markets, where transactions costs are small, short selling is permitted and competition is fierce?