ABSTRACT
A transaction-based price index uses actual property transactions to calculate the general change in prices over time. An active and transparent market is necessary for the construction of transaction-based property price indices that can reflect general changes in the price levels of properties. This chapter examines four different statistical methods that can be used to construct a transaction-based index: mean and median; hedonic price; repeat sales; and a hybrid model. Each method deals with the problem of quality heterogeneity differently and has its own advantages and disadvantages. One of the sets of indices that adopt the quality-adjusted mean is the official Hong Kong Property Price Indices, which are based on prices per floor area adjusted by appraised rental values for tax purposes. The origin of hedonic-based price indices for heterogeneous goods can be traced back to the constant-quality price indices for automobiles and computers. Later on, the same approach became popular for the construction of real estate price indices.
