What difference is there estimated the coefficients year by year or by using both years of observation in a single regression. I do not understand for this estimate what hypotheses I am making regarding the values of the regression parameters from one individual to another (heterogeneity) and the variance of the error term? Thanks