![Sample our Health and Social Care journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5938/TOC-Subject-Sample-2021-Health-Social-Care.jpg"})
Abstract
Recognising the rapid increase in housing prices and the presence of socio-economic and demographic disparities that often characterise a metropolitan city, we adopted a sub-city approach and deployed an array of methods to detect bubbles in the different regions of Greater Sydney – western, inner-west, southern, eastern and northern – over 1991 to 2016, using Westerlund error correction-based panel cointegration, backward supremum augmented Dickey–Fuller (BSADF) procedure, and dynamic ordinary least square (DOLS) tests. Our cointegration results show no evidence of cointegration between real house price and rent in the western region. However, there is strong evidence of cointegration in the eastern and northern regions. This confirms the existence of housing submarkets in Greater Sydney, and an indication of housing price bubbles in Western Sydney. Further, the formal BSADF bubble tests reveal strong evidence of explosive price bubbles in Western Sydney, while there is no comparable evidence for the other regions of Sydney, which further highlights the importance of submarket analysis. Importantly, the DOLS results suggest that housing investment plays a major role in the build-up of housing bubbles in Western Sydney, supporting Shiller's Psychological Theory of bubbles which posits that bubbles occur via the speculative behaviour of investors. The implications of the findings are also discussed.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 See section 2 for the definition of submarket
2 See section 4 for the detail discussion of the hypothesis formation
3 This is confirmed by the results from the cointegration tests. The results from Appendix 3 suggest that there is a long run equilibrium in house price for all LGAs that makeup each region.
4 Our cointegration tests in Appendix 4 clearly show that housing prices in different regions vary significantly. Similar results are obtained for income and rent each but they are not reported for brevity.
5 SEIFA is the Socio-Economic Indexes for Areas computed by the Australian Bureau of Statistics (ABS) to measure relative socio-economic advantages and disadvantages in the municipalities of Australia. The index of relative socio-economic advantage and disadvantage is one of such indexes that compares relative socio-economic advantages and disadvantages in areas. It is computed by incorporating household income, participation in the work force, education, family dynamics, access to resources, and housing arrangement. A score below 1000 indicates relative socio-economic disadvantage (ABS Citation2016c). See Appendix 5.
6 Even though owner occupiers are also interested in wealth-related factors, they are mainly driven by consumption motives, and they are more likely to consider their homes as a long-term investment. However, investors are mainly driven by wealth-related factors and more likely to outweigh potential gains than market fundamentals; thereby they tend to speculate more than owner occupiers.
7 Private dwelling as defined by the ABS is usually a house, flat, or even a room. For a more accurate estimate of rental properties, the rental properties in the central business district of Sydney were excluded from the Eastern region.
8 We would like to thank the referee to suggest this approach. A simple ratio could be an effective way to demonstrate some early sign of a housing bubble.
9 The LGAs that make up these regions are reported in Appendix 1.
10 This CPI subgroup excludes housing costs.
11 All variables are I(1) stationary. DOLS model deals with both potential simultaneity bias and small-sample bias among the explanatory variables. This is done by incorporating lagged and lead values of first difference explanatory variables (Narayan & Narayan, Citation2005).
12 First homebuyers are often owner-occupiers, while investors are landlords.
13 To ensure the robustness of our studies, we introduced housing supply and population in our Westerlund (Citation2007) error-correction panel cointegration test (that involves ARDL and ECM). We regress house price, income, rent, housing supply and population each across regions. While there is cointegration in housing supply and population each across regions, no similar evidence is found in house price, rent and income. This suggests that the baseline results are robust.
14 Following Lee & Reed (Citation2014) and Lee et al. (Citation2018), a formal GARCH model is also performed to model the volatility of housing prices. The results show that the volatility of the inner-west region did increase over this period. The results are consistent with the standard rolling 3-year risk analysis. The results are not reported for brevity.