Thumbnail Image

Testing for bubbles in time series data using long historical series

The concept of asset price bubble has drawn a large amount of academic attention. A bubble is commonly defined as where the asset price is not justified by expectations of the fundamentals. The empirical characteristics of a bubble are widely regarded as a dramatic rise in asset prices in the early periods along with a sudden collapse, which have been observed in a wide range of financial markets including commodity, real estate and stock. This thesis attempts to examine several well-noticed bubble episodes in history, for example, the Mississippi Bubble in 1719, the South Sea Bubble in 1719-20, Japan’s asset price bubble during the 1980s and the more recent US real estate bubble in the early 2000s. One of the greatest contribution of this thesis is that it has provided econometric-based evidence to support the existence of these famous episodes for the first time using recently developed econometric tests. The origination and collapse dates for these identified episodes are coincided with the traditional view of famous bubbles. This thesis not only tests for the existence of bubbles but also seeks to provide some empirical evidence to support the existence of bubble spillovers/contagion using the most-documented bubble episodes. The first example considered in this thesis is whether financial bubbles/crises spill-over to housing markets in Amsterdam, Norway and France. The second example is considered for a number of British financial organisations during the South Sea episode in 1719-20 (e.g., the South Sea Company, the Bank of England, East India Company, London Assurance, Million Bank, the Royal African Company and the Royal Exchange Assurance). The South Sea Bubble seems to migrate to other organisations as the British share market was speculative for that period. The timings of some identified episodes can be viewed as possible evidence of spillovers or contagion for this important period in England. The third example looks at the Japan’s real estate and stock markets during the 1980-90s. Based on econometric tools, the findings indicate signs of bubble migration from Japan’s stock market to its real estate market for the first time. Even if this phenomenon has been observed in the literature, none of existing studies provide any solid evidences to support this claim. This thesis investigates the long-run equilibrium relationship between house prices and economic fundamentals in the US based on quantile cointegration approaches for the first time. The presence of cointegration between house prices and fundamentals would imply convergence to a stable equilibrium relationship, suggesting a temporary run-up or drop in house prices would eventually come back to the equilibrium. By making use of quantile autoregressive distributed lag model (QARDL), this research explores the cointegration relationship across the whole conditional distribution.
Type of thesis
Hu, Y. (2019). Testing for bubbles in time series data using long historical series (Thesis, Doctor of Philosophy (PhD)). The University of Waikato, Hamilton, New Zealand. Retrieved from https://hdl.handle.net/10289/12308
The University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.