Hedonic Regression

The hedonic model of price measurement is based on the assumption that an asset's value derives from the value of its different characteristics. The price of a house will therefore depend on the value the buyer places on both qualitative (e.g. heating type) and quantitative attributes (e.g. number of bedrooms).

Since the prices of these characterictics cannot simply be observed, hedonic regression estimates the implicit market value of a unit of each attribute by comparing sample house prices with the associated characteristics (Thwaites and Wood, 2003).

Example: In order to avoid recording a rise in average property prices only because the number of detached houses changing hands in a certain period increases temporarily, the hedonic regression model will estimate the value of "terracedness" relative to "detachedness". If correctly specified, the price of each characteristic, and thus the price of a property with these fixed characteristics, will not be influenced by a change in the composition of overall property transactions. The result would be the desired standardisation to neutralise the effect of variations in the sample data.

Problems

• The relevance of a certain characteristic to the value of a property cannot be assessed objectively.
• It is often unclear how a certain attribute influences the price of a house.
• Virtually all the attributes of a property that may affect its value have to be taken into account.