ABSTRACT
The most disaggregate panel data are panel of household survey data. National longitudinal survey conducted by Bureau of Labor Statistics is a well-known example. The Panel Study of Income Dynamics by the University of Michigan has been used popularly as well. British Household Panel Study, Household Income and Labor Dynamics in Australia, Korea Labor Income Panel Study, Canadian Survey of Labor and Income Dynamics, and German Socioeconomic Panel have provided household survey data. Once the factor number and the common factors are estimated, it is always useful to check how much the estimated common factors explain the variation of the panel data of interest. It is important to note that the idiosyncratic term must be stationary. Obviously, the first eigenvalue is larger than the second eigenvalue so that the accuracy of the principal component estimator for the first factor is higher than that of the second factor.
