Econometric Analysis of the Financial Crisis

Yale University, New Haven CT

Investigators

Abstract

As events during the 2008 financial crisis and its aftermath have shown, extreme movements in financial time series can have wide socio-economic impact. The immediate effects involve huge swings in wealth, which in turn can lead to financial insecurity, institutional insolvency, misallocation of economic resources, and global consequences that threaten the credibility of major economic institutions. In addition, economic globalization has led to strong financial linkages that can increase risk exposure in the event of a large common shock. Understanding this phenomenon, exploring its causes, mapping its evolution over financial markets, and studying its effects on the real economy all present major challenges to the economics profession. This project aims to develop quantitative methods that will contribute to our understanding of some of these issues. Beyond the immediate concerns that have confronted policy makers during the recent financial crisis lie questions that relate to the emergence and detection of the bubble phenomena and the evolutionary course of the resulting crisis through the financial and economic systems. These issues form the focus of the primary program of proposed research, which seeks to design new econometric methodology to enable early detection of bubble behaviour and to provide empirical date stamping technology that quantifies dynamic timing and helps to map the evolutionary course of a financial crisis. The project will provide a rigorous econometric approach to dating financial bubbles, deriving a limit theory for empirical estimates of the origination and collapse dates of bubble episodes, and using a new model of financial bubble activity that is based on the successive conjunction of ?efficient market? models and ?mildly explosive? models that can capture episodes of financial exuberance. An extensive empirical implementation of this technology to recent financial crisis data is envisaged, yielding empirical information on the temporal extent, the magnitude, and the course of the various bubble phenomena that have involved financial and commodity markets, exchange rates and real economic activity. In addition to this primary research program, two secondary projects will be pursued. The first deals with co-movement in economic and financial data where there is some uncertainty about the persistence characteristics in the data, as is common in empirical research. The second program of research will explore new econometric methodology for dynamic panel modelling in the presence of individual effects. The importance of these fields is reflected in the vast empirical literatures in the social sciences that utilize cointegrating methods to study long run linkages among time series and dynamic panel regression to study the effects of individual decision making over time.