A note on error components models, and, Further evidenece on the estimation of dynamic economic relations from a time series of cross sections