Behavioural heterogeneity in the housing market : a comparison between Singapore and Hong Kong

Abstract

Unlike the Benchmark Rational Expectations Model (BRM), the Heterogeneous Agent Model (HAM) recognises that the assumption of market players having full rationality is often unrealistic. Instead, it claims that players have heterogeneous behaviours and bounded rationality. Over the years, HAM has been proven to be one of the more successful frameworks in forecasting future asset prices, explaining its rising popularity amongst researchers studying on asset markets. Specifically, we use the Markov Regime-Switching Heterogeneous Agent Model (HAM-MS) in our research as we acknowledge the possibility of regime-dependent behaviours. Meanwhile, Singapore and Hong Kong are Asia’s two key financial hubs and they share numerous similarities. Yet, there were prominent differences in their housing market performances. Therefore, we start with testing HAM-MS and BRM without breakpoints (BRM-NB) to determine if between and within-group behavioural heterogeneity are present in Singapore’s and Hong Kong’s housing markets. Subsequently, we compare the results of the two countries and explain their similarities and differences. Generally, there are evidences of between and within-group heterogeneity in both countries. Consequently, HAM-MS outperforms BRM-NB for in-sample estimation efficiency and out-of-sample prediction. Comparatively, HAM-MS has better accuracy in Singapore than in Hong Kong and we attribute it to the greater stability in the former’s housing market.