City Research Online

Abstract

This thesis develops econometric methods for detecting and interpreting explosive dynamics in financial time series, with particular emphasis on real-time monitoring, financial stability, and investment applications.

The first chapter develops an adaptive common bubble detection framework based on Incremental Principal Component Analysis with Recursive Least Squares (IPCA--RLS). The method sequentially updates the common factor and backward sup-ADF statistic using only information available up to each period, generating non-revisable signals suitable for real-time monitoring. Monte Carlo experiments show substantial reductions in computational complexity with negligible loss of numerical accuracy. An application to U.S. artificial-intelligence stocks demonstrates its ability to identify common explosive behaviour in a timely manner.

The second chapter applies this framework to the U.S. green equity market using daily price-to-earnings ratios for nine clean energy, clean technology, and green energy indices. The analysis identifies multiple episodes of common explosiveness, with the most prominent occurring in March--April 2021. These episodes show limited overlap with broad equity market bubbles and are associated with macro-financial conditions, particularly credit spreads, and physical climate risk.

The third chapter examines whether explosiveness contains predictive information for investment decisions. Using 153 U.S. equity factors, the chapter documents widespread explosive behaviour and develops factor strategies based on right-tailed unit root statistics. The time-series factor explosiveness strategy achieves an annualised excess return of 5.29% and a Sharpe ratio of 0.92, outperforming conventional time-series factor momentum.

Overall, the thesis shows that explosive dynamics contain information relevant not only for bubble detection and financial stability monitoring, but also for investment decision-making.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HG Finance
Departments:	Bayes Business School > Bayes Business School Doctoral Theses Bayes Business School > Faculty of Finance Doctoral Theses

Export

Downloads