Overheated behaviours and spread effects: an analysis of London’s housing market

ABSTRACT

This paper takes the London housing market as an example to discuss whether housing markets with more significant overheating effects have a greater influence on other housing markets. The spillover effect of London’s housing market has been the subject of much literature, but the empirical results of the relevant literature are inconsistent. This paper proposed a new perspective to explain these inconsistent conclusions: Whether the housing market is overheated or not can affect the spread effect of a regional housing market. The paper estimated the indicators of real estate overheated behaviours and market connectedness. Further, it revealed that when real estate overheated behaviours became more severe, the irrationally high house prices resulted in spread effects. The empirical results also showed that when the overheating effect occurred in London’s housing market, the rental yield of its adjacent regions significantly increased; specifically, the rental yield of a region increased with its closeness to London. The results of this study provide an explanation for the variation in spread effects. This study also indicated that when studying the influence of the relatively booming housing markets on other regions, it is necessary to test the prosperity status of these housing markets to more accurately estimate the spread effects of these markets.

KEYWORDS:

• London’s housing market
• real estate overheated behaviours
• overheating effects
• market connectedness
• spread effect

Acknowledgments

I am immensely grateful to Professor Myounggu Kang (Editor) and the two anonymous referees for the constructive comments of this paper. Funding from the Ministry of Science and Technology (Taiwan) under Project No. MOST 110-2410-H-390-008-MY3 has enabled the continuation of this research and the dissemination of these results.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Phillips et al. (2015) proposed a dynamic unit root test to overcome the problem of multiple bubbles and structural changes in asset price series and used the S&P 500 series for empirical analysis. Whether an asset has a bubble component can be examined by several different definitions. What Phillips et al. (2015) explore is whether the market is over-prosperous, that is, whether the asset price has divergence from the fundamentals. Many follow-up studies, such as Pavlidis et al. (2016), used the unit root test method proposed by Phillips et al. (2015) to detect whether the housing market has excessive prosperity or bubbles. Their testing is based on the ratio of house prices to rental prices to determine whether house prices deviate from fundamental variables.

2 Zhang et al. (2017) explored the house price spillover effects of 35 cities in China and discovered that house prices spilled over from cities with high economic development to regions with low levels of development. Lee and Lee (2019) explored the Korean housing market and indicated that Seoul had the most influential housing market in Korea.

3 For more information, please visit https://www.ons.gov.uk/.

4 When estimating the dynamic unit root tests, we need to select the number of periods for rolling windows. According to Phillips et al. (2015), the minimum window size is computed by $T(0.01+\frac{1.8}{\sqrt{T}})$, where $T$ is the number of samples. Hence, this paper used 26 as the number of periods for rolling windows.

5 When estimating dynamic connectedness indices, we need to select the number of periods for rolling windows, which in this paper is based on the shortest period for estimating a complete set of total connectedness indices. Antonakakis et al. (2018) also analyzed the connectedness of the UK regional housing markets and used the quarterly data to measure the connectedness of 13 regions of the United Kingdom. They used 60-quarter rolling windows to estimate. Because the VAR model for two lag periods of 9 variables (regions) is estimated in this current paper, the paper begins estimation by using 40-month rolling windows. In addition, with 40-month rolling windows, we are able to estimate and observe changes in the correlation between regional markets starting in July 2008. Therefore, we can observe the changes in market relevance under the global financial turmoil.

Additional information

Funding

This work was supported by Ministry of Science and Technology, Taiwan [grant number: MOST 110-2410-H-390-008-MY3].