Evaluations of non-marketed goods

Using “The Urban Housing Market” database, we are studying questions about how individuals value things normally not priced on a market, such as parks, proximity to transport infrastructure, ethnically mixed populations or risks for nuclear power accidents.

The database has information on prices and characteristics for the majority of all residential properties sold by estate agents in Sweden. The database also includes the date the property was put up for sale, how long it was on the market and detailed information on where the property is located.

The studies use the hedonistic price method (see Rosen, 1974). The underlying assumption of the hedonic price method as applied to housing is that a home can be seen as a collection of factors that, when combined, determine the property value. In a traditional hedonistic price estimate, the property’s value is estimated as a function of a series of different characteristics to identify the implicit value for the various characteristics.

Estimating the implicit value in this way for different area characteristics and other spatial characteristics is often difficult. There can be non-observable characteristics correlated with the characteristics you want to study the value of which impact the area’s housing prices (i.e., problem with missing variables), and the housing prices in an area can directly or indirectly impact what characteristics an area has (i.e., problem with reverse causality). These issues can be mitigated by combining the hedonistic price method with quasi-experimental approaches.

Project start



Jan Wallanders och Tom Hedelius stiftelse, Tore Browaldhs stiftelse 


Matz Dahlberg (project manager), professor in economics

Last modified: 2022-12-06