CESifo Press Service - Article on Two-Sided Markets

The Quirks of Two-Sided Markets

Double-edged: Handle with care Two-sided markets resemble two-sided swords: they must be handled with care, can cut both ways, and if well managed can be highly effective. They are also tricky to set up: in order to succeed, they have to get two different groups of customers on board. But each of these two groups expects the other one to be already committed before they step in: a sort of chicken-and-egg conundrum.

A case in point are video game consoles: nobody is going to buy them if there are no games to be played on them. And no game developers will write games for that console if nobody buys it. Further examples are credit card systems (banks won't jump aboard unless there are lots of users willing to pay with them and lots of stores and restaurants willing to accept them as a means of payment), dating clubs (luring men and women), newspapers (courting advertisers and readers), TV stations (advertisers and viewers), real-estate brokers (attracting buyers and sellers), and so on.

But two-sided markets have a further singularity: consumers may be ready to buy more of a taxed good if it is sold by a two-sided platform firm, as posited by Hans Jarle Kind, Marko Koethenbuerger and Guttorm Schjelderup in their latest CESifo Working Paper.

The mechanism in question works so that an increase in the ad valorem tax—i.e. a tax as a percentage of value—may end up lowering the price to consumers and thereby boosting sales.
To illustrate this, consider a newspaper company: it derives income from selling papers and advertisements, and income from advertising rises with higher paper sales. If the tax on newspaper sales is increased, the media firm may decide to lower the newspaper price, reducing thus its tax burden and simultaneously attracting more readers, enabling it then to increase the price charged to advertisers. In an extreme case, a very high tax on paper sales may prompt the firm to provide the paper for free and rely solely on advertising to generate income. In this, it is as if advertisers pay readers for their attention.

In order to substantiate their assertion, the researchers set up a general model of a two-sided market, showing the exact conditions for when a tax increase causes the end-user price to fall and, subsequently, demand to rise. Not surprisingly, they opted for a media company to explain their results, although the model has a general application.

In most countries newspapers are subject to reduced value-added tax rates, such as 7% in Germany, or even exempt from it altogether, as in the UK or Denmark. The reason for this is that governments consider newspapers to be essential for the dissemination of information, and try to make it easier for readers to buy the paper by slapping a lower tax rate on its sales. That, however, is the thinking behind a one-sided market. If the reasoning of our authors is correct, it might turn out that a lower VAT may actually reduce the sales of newspapers.

Another interesting case is credit card services, which are VAT-exempt in the EU. The use of credit cards has become so widespread that by now it is no longer necessary for governments to foster their adoption by providing tax breaks. Abolishing the preferential VAT treatment would presumably exert no negative impact on credit card use, while imposing a VAT on these services may even expand the size of both sides of the market —card holders and merchants alike.

In other words, in sharp contrast to conventional markets, where higher taxation leads to lower sales, in two-sided markets—under certain conditions—higher ad valorem taxes can have the opposite effect.

Still, imposing a VAT on credit card services would probably not fill credit card companies with joy. But for cash-strapped governments, it is a thought worth entertaining.

Hans Jarle Kind, Marko Koethenbuerger and Guttorm Schjelderup: Do Consumers Buy Less of a Taxed Good?, CESifo Working Paper No.1635

Note: This text is the responsibility of the writer (Julio C. Saavedra) and does not necessarily reflect the opinion of either the CESifo Working Paper author(s) cited or of the CESifo Group Munich.