Using the data given below, estimate the model Y t = £]1 + £]2 X t + U t
Given that Y =  inventories and X = sales, both measured in billions of dollars.

Download:  1232.dat ,  1232.xls

(a) Estimate the preceding regression. (Eviews)

(b) From the estimated residual find out if there is positive autocorrelation using
     (i) the Durbin-Watson test and (ii) the large-sample normality test. (Suggested Answer)

(c) If p is positive, apply the Berenblutt-Webb test to test the hypothesis that p = 1.
     (Suggested Answer)

(d) If you suspect that the autoregressive error structure is of order p, use the Breusch-Godfrey test
     to verify this. How would you choose the order of p?  (Suggested Answer)

(e) Based on the results of this test, how would you transform the data to remove
     autocorrelation? (Eviews)

(f) Test your model for the ARCH effect. If an ARCH effect is observed, would you modify your
    conclusion about autocorrelation reached previously? (Suggested Answer)

(g) Repeat the preceding steps using the following model: (Eviews)(Rats)
      lnYt = £]1 + £]2lnXt + Ut

      (b*) (Suggested Answer)
      (c*) (Suggested Answer)
        (d*) (Suggested Answer)
      (e*) (Eviews)
      (f *) (Suggested Answer)

(h) How would you decide between the linear and log-linear specification.(Suggested Answer)