Hi! I wonder if there is a program to apply the principal components technique to an unbalanced panel of data. I have 45 monthly series which some start at different dates. Stock and Watson (see Stock, J and M Watson (2002): “Macroeconomic forecasting using diffusion indexes”, Journal of Business and Economic Statistics, vol 20, no 2, April, pp 147–62) propose the use of an Expected Maximization algorithm that estimate the factors iteratively. As I understand, what they basically do is to make a regression of the variables with missing observations on principal components estimated from the balanced sample. Then a projection of the series with missing data is made for the sample with missing observations. Next, a new set of principal components is calculated on the basis of the complete and the projected series.This process goes on and on until it eventually converges and the princial components estimated are close enough.
Thanks in advance
Cristian