A Case Study on House Price Index Based on Generalized Repeat Sales Model

Abstract

Conventional repeat sales model for constructing house price indices (HPIs) based on arm s-length transaction data has a limitation in that it cannot produce stable indices when the sales pairs are not sufficient, which is common in a small area. To tackle the problem a generalized repeat sales model can be utilized. The model was proposed to estimate originally repeat sales HPIs of several heterogeneous sub-markets and the total market simultaneously. In this paper, we conducted a case analysis on the house price indices generated from the generalized repeat sales model for row houses of 5 metropolitan cities in Korea. It was shown that the repeat sales index for each city, which has not been produced officially up to now because of the lack of sufficient sales pairs, can be obtained by the generalized repeat sales model with the index of the whole market of the 5 cities. Furthermore, the result provides some valuable information about how different the house price trends among the 5 cities are.