Mixed Dynamic Factor Models----With Application to Explosive House Prices
Speaker: Prof. Ye Chen (Capital University of Economics and Business)
|
Topic: |
Mixed Dynamic Factor Models----With Application to Explosive House Prices |
|
|
Time&Date: |
15:00-16:15 pm, 2019/3/15 (Friday) |
|
|
Venue: |
Room 619, Teaching A |
|
|
Speaker: |
Prof. Ye Chen (Capital University of Economics and Business) |
|
| Abstract: | The standard asymptotics for the dynamic factor models are built on the assumption of the common persistency among factors. However, the empirical study of the financial exuberance requires models with the distinct persistent factors. In particular, when observing the prevailed bubble behavior of the asset prices in different markets, such phenomenon may be driven by the explosive factors from the bubble component, by the I(1) factors from the fundamental value component, and by the stationary factors from the economic environment. In this paper, we consider the dynamic factors models with both mildly explosive factors, I(1) factors and stationary factors in the system. First, we estimate the common factors and factor loadings by the method of principal component. Then, we establish the associated asymptotic theory within the framework of both large $N$ and $T$. We show that, the convergence rates of estimated factor loadings are different due to the different persistency of the factors. The estimated factors are asymptotically normal with the convergence rate $\sqrt{N}$. Information criterions are proposed to select the number of factors in our mixed dynamic factor models. Simulation studies show that the proposed information criterion are able to consistently select the number of factors. By applying our proposed model, we investigate the housing market in China and Australia, finding that the dominant factors for driving up the house prices are the money supply and GDP respectively during the sample period. |
