Analysis Of Breakpoints In Means And Variances For Panel Data

Panel data, which they are N series(variables), and each series has T observations. Panel data is a common research object in the economic model, such as the GDP in different countries, different regions of the oil price and so on. Break point, also known as change point, is a change which has taken place in each series at an unknown common point, referred to as the common break point in which all series have the same break point. Prior to this, there have been some articles (Bai,2010) gives the estimation and distribution of breaks pints for panel data, his research mainly focus on the breakpoint in mean or variance. In this paper, the study area is enlarged, the breakpoints in mean and variance may not be consistent, and the original conclusion is extended to the case.In this paper, we focus on two problems, one is the estimation of the breakpoint, the other is giving the non-degenerate distribution of the breakpoint under certain conditions. When the panel data has the common break in mean and variance, we need to estimate the two parameters. To estimate the two breakpoints, we firstly estimate the break in mean and then give the estimation of break in variance. We use the least square method to estimate the breakpoint in mean, the quasi likelihood estimation method to take the breakpoint in variance. The conclusions show that under certain conditions, using these methods, we can make the estimations be consistency with actual breakpoints. In order to further validate our conclusions, we use Monte Carlo method to simulate the experiment. The simulation results are consistent with the conclusion. In the earlier analysis, we found that the estimation value is a degenerate distribution, in order to give the non-degenerate distribution, we modify the previous conditions, and give the breakpoint distribution in mean and variance.