The article touches upon the topical issues of the residential real estate market, which are proposed to be solved by means of time series cointegration. The study aims to assess the structure of the housing market by types of apartments using price dynamics per one square meter of apartments' total area. The objectives of the study are to develop a methodology of determination of time series cointegration for the data with structural relationships; to analyze the average prices for the types of apartments on the primary and secondary housing market; to study the housing market in the Russian Federation by quarterly data of state statistics for the period 2000–2020 based on the developed methodology.The results of the research showed that the prices at the primary and secondary housing market by types of apartments do not always represent an integrated process of the frst order and cannot be used for building a co-integration equation. This necessitated additional analysis and, as a consequence, the correction of the time period. It was proposed to ensure stationarity of linear combination of nonstationary data corresponding to the integrated process of the frst order by using the generalized least squares method (GLS). The sum of the elements of the cointegrating vector obtained this way tends to unity, and the elements themselves are estimates of the relative indi cators of the structure by types of apartments on the primary and secondary housing markets respectively. Thus, the suggested methodology allows estimating, on average, the share of the sold apartments of each type in the period under consideration, both in the regional context and in the country as a whole.The proposed methodology can be used for the estimation of relative indicators of the structure according to temporal data in different applications.
Read full abstract