1)Figure 1 above plots the variable consumption over a period of 15 years. The graph suggests that there may have been a change in the consumption function after year 5. How would you account for this structural break when modelling the consumption within a time-series regression framework?  If you ignore the structural break what would the effect be on the forecast accuracy of the model?
2)Suppose, now, that the shift in the parameters of the consumption equation is more gradual than the one in Figure 1 and, therefore, the precise break point is not known: how would you test for parameters stability in this case?