Forecasting the Performance of Commercial Property Market: Beyond the Primary Reliance on Econometric Models
This study seeks to explore extreme downside risks in the commercial property market and attempts to integrate the information into real estate forecast models.
As real estate forms a major part of the asset portfolio, it is critical that analysts and institutions employ wide-ranged techniques for modelling and forecasting the future performance of real estate assets. Compared to alternative asset classes, reversal of a property decision can take significant time and destroy valuable capital. This uncertainty in the real estate investments holds considerable risks for investors. Further, the known, unknown and unknowable risks (KuU) are unique across the spectrum of commercial real estate asset classes which appears to have a better performance measurement. The performance of commercial property is principally measured by mathematical models to choose optimal levels of both risk and return, and the statistical techniques employed in the real estate market is recognised as real estate econometrics. For instance, Regression analysis, Time series models, Multi-equation structural models, and Vector autoregressive models are some of the quantitative approaches that facilitate modelling and forecasting real estate data. (Adair & Hutchison 2005; Bardhan & Edelstein 2010; Brooks & Tsolacos 2010; Brown 1991; Hargitay & Yu 1993; Higgins 2013b).
However, past researchers have demonstrated a limited predictive capacity of estimated real estate performance where key underlying macroeconomic indicators have been the driving forces behind the trend in real estate performance. Within the Australian context, macro-economic factors such as inflation, industrial production, interest rates and employment are being significant in office, retail, industrial and listed property trust returns. These indicators are used as warning signals for retrospective and prospective performance appraisal of the real estate (Hargitay & Yu 1993; Higgins 2014; Huang & Wang 2005; Worthington 2006). However, standard assumptions of mainstream economics derived from historic data can fail when stable assumptions cease to hold and extreme volatility occurs, as featured by the recent severe price swings associated with Global Financial Crisis (Hargitay & Yu 1993; Higgins 2014; Mandelbrot & Hudson 2004). Therefore, modelling of the long-run trends and short-run ?uctuations in the property market has become a great challenge. In addition to the focus of econometrics relying on underlying macroeconomic indicators, short-term black swan events and long-term structural changes and transformational forces are another two drivers that may impact on real estate modelling and forecasting. Hence, all three forces directly and indirectly challenge core economic activity and are intertwined in the determination of real estate outcomes (Higgins 2014; Trahan & Krantz 2011).
However, black swan events and structural changes have been a lacuna in the modelling and forecasting real estate performance. Taleb (2008) coined the term 'black swan' to describe random events that form part of our lives. These events have the following three key characteristics: outlier, being outside the realm of regular expectations; carries an extreme impact; and explanations for the occurrence are concocted after the fact, making it explainable and predictable. It has also been identified as ‘noise in the market’. Further, these shocks extended into a broader depth that cover natural, man-made and hybrid disasters (Higgins 2014; Miller, Engemann, & Yage 2015; Parker 1992; Shaluf 2007; Turner & Pidgeon 1997). As the next emerging untouched pitfall in the real estate modelling and forecasting, structural changes often initiated by policy decisions and innovation appear to have long-term economic implications. Further, structural breaks that affect the performance permanently resulted in deviating the market behavior from the historical pattern. Moreover, 'the rise in digital business innovation means we need innovation in economic metrics. If we are looking at wrong gauges we will end up in wrong decisions' (Brooks & Tsolacos 2010; Brynjolfsson & McAfee 2014, p. 122; Higgins 2014).
As an identified demerit, economists focus on normal conditions, while unpredictable shocks are often outside the scope of the economists’ models. Further, such models have been developed based on smoothed and filtered data to remove outliers that are contaminated by noise. These extreme events are commonly referred to as statistical outliers (outside ± 2 standard deviations) and therefore follows a ‘power law’ distribution with fat tails, which leaves room for many big price swings than a normal bell curved distribution. In greater emphasis, Mandelbrot strongly rejected normality as a distributional model for asset returns where the normality can lead to an underestimation of the true risk of direct property and so an overestimation of the associated performance (Higgins 2013b; Lee & Higgins 2009; Mandelbrot & Hudson 2004; Rachev, Menn, & Fabozzi 2005; Thompson 2011).
Therefore, this study shall seek to answer the following research question: How do we model extreme downside risk drivers in forecasting the performance of the commercial property market? The main aim of this research is to develop a conceptual model of integrating extreme downside risk drivers in forecasting the performance of real estate assets. This can be constructed through the identification of pitfalls in common forecasting models for predicting the outliers in real estate performance and then determine solutions to integrate them in modelling. The current research study follows Pragmatism assumption in which the attention is given to the research problem and then uses pluralistic approaches to derive knowledge about the problem (Creswell 2003). Hence, the study begins with quantitative methods to be followed by a qualitative method. Secondary data of real estate performance ought to be collected to determine the accuracy econometric forecasts to determine extreme downside risks. The research then moves forward to obtain primary data from a semi structured interview survey to develop a conceptual model of integrating extreme downside risk drivers to forecast the future performance of the commercial property market.