Construction and updating of property price index series : The case of segmented markets in Stockholm

Purpose – The purpose of this paper is to construct a hedonic property price index in the segmented housing market. Design/methodology/approach – Three different research questions are investigated. The first considers how to identify separate local housing markets, with the use of a regression‐tree approach, and the second concerns when it is possible to update a price index, with a recursive regression approach. Finally, the question about seasonal adjustment is investigated. Findings – Overall, the hedonic approach is the best method to use. Moreover, the County of Stockholm is not one market in the sense that the market can be represented by only one property price index. It can be best described as a number of different sub‐markets where the property prices grow differently. The results suggest that there exist at least five different price indexes in the County of Stockholm. The results also support that recursive regressions are two appropriate methods to answer the research questions. Here, the empirical analysis suggests that the parameter estimates converge relatively fast toward the estimate using all observations. Practical implications – The paper illustrates how to derive sub‐markets in the construction of a price index and when to update a price index series. Originality/value – The introduction of new financial products, such as property derivatives, to the market has made the construction and quality of property price indexes more important. High quality price indexes are vital when it comes to, for example, pricing property derivatives. The paper facilitates this.