An important relationship in macroeconomics is the consumption function. The file consumptn.dat contains quarterly data from 1960Q1 to 2009Q4 on the percentage changes in disposable personal income and personal consumption expenditures. We describe these variables as income growth (INCGWTH) and consumption growth

(CONGWTH). To ensure that the same number of observations (197) are used for estimation in each of the models that we consider, use as your sample period 1960Q4 to 2009Q4. Where relevant, lagged variables on the right-hand side of equations can use values prior to 1960Q4.

(a) Graph the time series forCONGWTH and INCGWTH. Include a horizontal line at the mean of each series. Do the series appear to fluctuate around a constant mean?

(b) Estimate the model CONGWTHt ¼ d þ d0INCGWTHt þ vt. Interpret the estimate for d0. Check for serially correlated errors using the residual correlogram, and an LM test with two lagged errors. What do you conclude?

(c) Estimate the model CONGWTHt ¼ d þ u1CONGWTHt1þ d0INCGWTHt þ vt.

Is this model an improvement over that in part (b)? Is the estimate for u1