Problem 6.13 The data needed for this problem is contained in the text files pepsiqeps.asc and ibmqeps.asc. Open them in S-Plus and run them as a scripts. You should get two timeSeries objects. The first one gives the quarterly earnings per share of Pepsi Co. and IBM.

1. Use the Pepsi data and reproduce the filtering analysis done in the text for the General Electric quarterly earnings.

2. The goal of this second question is to perform the same analysis with the IBM data contained in the file ibmqeps.asc.

2.1. Fit a state space model and use it to predict the last two years of quarterly earnings using only the data up to the last quarter of 2000.

2.2. Fit a state space model to the data of the quarterly earnings over the period starting with the first quarter of 1996, and ending with the third quarter of 2000. Use this model and Kalman filtering as above to predict the last two years of quarterly earnings using only the data used to fit the model.

2.3. Fit a model to the data up to the last quarter of 1990, and use it to predict the quarterly earnings of the period 1991-1995 using only the data up to the last quarter of 1990.

2.4. Comment on the performance of the prediction in each case.