Trends and Prospects of the U.S. Housing Market Using the Markov Switching Model

Abstract

Accurate projection of the economic conditions in a country can enable the government to establish appropriate policies in a timely manner. This also applies to enterprises and individuals in terms of decision-making processes, such as investing and production planning and household consumption and saving. The U.S. housing market is no exception to this practice. The prompt and accurate assessment of the trends in the U.S. housing market enables consumers to make quick decisions and to come up with the corresponding measures, thus minimizing risks associated with market uncertainty. The monthly indices of the U.S. housing market indicators are released at the end of the following month, creating a month-long standstill in making a judgment regarding the housing market. Consequently, it is not possible to predict the current month’s market status. Therefore, in this study, various “U.S. housing market-related” indicators were calculated as average month-to-month changes using the composite index methodology of the National Bureau of Economic Research (NBER). The trends in the U.S. housing market were analyzed using the Markov switching models: the Markov switching random walk (MS-RW) model and the Markov switching autoregressive (MS-AR) model. Results showed that the methods can accurately determine the trends in the U.S. housing market. Findings from the forecasting performance test made it possible to predict or forecast the prospects of the U.S. housing market within the month-long standstill period.