ABM simulations of housing portfolios in normal and bubble times

Abstract

Throughout history, the housing market have collapsed numerous times causing distraught among homeowners as well as investors, with the most prominent ones being the Asian financial crisis in 1997 and the global financial crisis in 2007. In most cases, these price inflations are driven by the wealthy. As such, understanding these individuals’ impact to the housing market could shed light on the formation of bubbles. To do so, we first need to understand how they invest, and with such an expensive asset class, how they manage their housing portfolio. In finance, the most popular method regarded for portfolio management is through Modern Portfolio Theory (MPT). Hence, our goals are to gain insight on the effectiveness of MPT strategy in the housing market and investor’s housing portfolio in both normal and bubble times. Due to the lack of information and transparency in many private housing markets, an alternative solution is to set up our own housing market. In this paper, we constructed a preliminary ABM simulation, as an Econophysics approach, to simulate interactions of multiple investors (the wealthy) in order to create an artificial housing market. For simplicity, we externally simulated the stock market as our secondary asset class and coupled it with our ABM to facilitate investors behaviour. We found that in normal stock times, a housing bubble was not formed, but if a stock market has both bubble and normal times, a housing bubble could form. We also found that MPT is effective with higher probability of positive returns in normal housing and stock times. Lastly, we found a reversion of housing portfolio weights during a housing bubble in a poor performing stock market. This implies that in an inflated housing price situation, investors portfolio allocation towards housing significantly increases but reverses in normal housing times.