Suppose that we want to evaluate the e ect of several variables on annual saving and that we have a panel data set on individuals collected in January 1990 and January 1992. If we include a year dummy variable for 1992 and use rst di erencing, can we also include age in the original model ? Explain