Content area

Full text

Introduction

The real estate sector plays a vital role in the economy of any nation. Several investigators suggest that a strong link exists between the property market and the economy (Pholphirul and Rukumnuaykit, 2009; Jiang et al., 2013). At the microeconomic level, stakeholders are motivated to invest in real estate assets for three main reasons:

  • hedge against inflation;

  • regular flow of future income; and

  • durability, amongst other qualities (Shapiro et al., 2012).

However, the real estate market (cycle) usually experiences booms and bubbles (Malpezzi and Wachter, 2005). These booms and bubbles are commonly triggered due to the interplay between macroeconomic variables and property prices (Quigley, 2001). Booms and bubbles could have an effect on the portfolio of real estate investors (Crowe et al., 2013). For instance, changes in the socioeconomic environment (such as the 2007 global financial crisis and the 1997 Asian financial crisis) were linked to activities in the real estate sector (Mera and Renaud, 2000; Jiang et al., 2013). To reduce the uncertainty associated with such fluctuations, it is important to develop models that can produce reliable predictions of property price index (PPI).

Previous research has shown that traditional approaches (i.e. regression-based models) can be used for property price prediction. However, there is a need to identify new methods which provide reliable predictions. This is because accurate and reliable prediction of smooth changes in house prices could help to achieve economic growth (Ge and Lam, 2002). Most property markets around the world are not immune from property price bubbles (Case and Shiller, 2003), and Hong Kong is not an exception. The Hong Kong property market has experienced a number of booms and bubbles over the decades (Teng et al., 2013). Hui et al. (2011) reported that bubbles exist in residential mass and residential properties in Hong Kong, which could have devastating impact on housing prices stability and government policies. Therefore, the present paper evaluates the efficacy of using the artificial intelligence (AI) techniques for modeling and predicting of PPI in Hong Kong. This was achieved by:

  • identification of macroeconomic indicators that influence the prices of properties;

  • collection of relevant data relating to identified macroeconomic indicators and PPI;

  • fitting the collected data into three modeling techniques (artificial neural network...

Show less