Testing for housing bubbles across London housing markets

This thesis comprises three papers in which I investigate the existence, nature, and effects of bubbles in London housing markets. In the first paper, I applied a recursive unit root test following Phillips et al. (2011) and Phillips et al. (2015) (PSY) to four housing affordability ratios (price to rent and income ratios, and the user cost to rent ratio, calculated using the five year average and an optimised forecast for expected price growth) across London boroughs for the time period 2001- 2019. With all of these, housing bubbles are identified around the Global Financial Crisis (GFC) period. With the first two ratios, but not the latter two, housing bubbles are identified in the post-GFC period in the mid to late 2010s. The key result of this is that the price to income and rent ratios can misidentify bubble periods due to them ignoring other fundamentals (i.e. interest rates). In the second paper, I tested the rational bubble model across London housing markets using the coexplosive framework developed by Nielsen (2010) and Engsted and Nielsen (2012). I found that the data did not fit the rational bubble model for all London boroughs, which I argue is evidence against the existence of rational bubbles. In the third paper, I tested for ripple effects across London boroughs using a spatio-temporal model following Holly et al. (2011). I found that there is evidence of price convergence and dependence, however there is not evidence of Central London acting as a dominant region. I further tested the London housing markets with a coexplosive model of bubble transmission following Evripidou et al. (2022). I found that several regions showed signs of coexplosivity. I argue that this means the ripple effect could be explained by the spatio-temporal transmission of information.