Detecting asset price bubbles with time-series methods

To promote the financial stability, there is a need for an early warning system to signal the formation of asset price misalignments. This research provides two novel methods to accomplish this task. Results in this research shows that the conventional unit root tests in modified forms can be used to construct early warning indicators for bubbles in financial markets. More precisely, the conventional augmented Dickey-Fuller unit root test is shown to provide a basis for two novel bubble indicators. These new indicators are tested via MC simulations to analyze their ability to signal emerging unit roots in time series and to compare their power with standard stability and unit root tests. Simulation results concerning these two new stability tests are promising: they seem to be more robust and to have more power in the presence of changing persistence than the standard stability and unit root tests. When these new tests are applied to real US stock market data starting from 1871, they are able to signal most of the consensus bubbles, defined as stock market booms for example by the IMF, and they also flash warning signals far ahead of a crash. Also encouraging are the results with these methods in practical applications using equity prices in the UK, Finland and China as the methods seem to be able to signal most of the consensus bubbles from the data. Finally, these early warning indicators are applied to data for several housing markets. In most of the cases the indicators seem to work relatively well, indicating bubbles before the periods which, according to the consensus literature, are seen as periods of sizeable upward or downward movements. The scope of application of these early warning indicators could be wide. They could be used eg to help determine the right timing for the start of a monetary tightening cycle or for an increase in countercyclical capital buffers.