Pricing in markets without money: Theory and evidence from home exchanges

A growing number of markets use token money instead of real money to allocate resources such as class seats, food donations or holiday homes. Agents earn token money by supplying and can only spend tokens by consuming within the system. A platform decides how to set goods' token prices. Should it i) clear the market, as assumed in canonical competitive equilibrium frameworks, or ii) compress prices and ration excess demand? I characterize in a two-type, two-period model under which conditions compressing prices increases supply and utilitarian welfare. Intuitively, this is possible when income effects are strong—because agents are easily satiated with tokens—and participation effects are weak. I confirm the model's predictions on a large home exchange platform, combining proprietary data on the universe of transactions with natural experiments. I find that income effects are large empirically and that a reform that reduced the price of attractive homes increased their supply and did not reduce participation. I show that many users have a strong preference against for-money platforms, allowing the exchange platform to set non-market-clearing prices. Broader insights are that lessons from traditional markets may not easily extend to token economies and that market designs without real money can have advantages even in settings where money is common.