ARIMA exercise in Soritec - Computer lab session

11/14/2000

Part I. Reproduce the graphs on Tables 7-4 and 7-5 in text (pp. 308 and 311, respectively). Use the command "Open Data File" under the File menu to read the c7t2.sal data file in the Soritec DATASETS directory. In order to create these graphs the command

on plot

needs to be executed. The command "on plot" is necessary to see the ACF and PACF when using the "Time Series" menu option "Examine autocorrelations."

 

Part II. Find the series Balance of Payments: Imports of Services; SA in the Economagic site and generate both ACF and PACF graphs for this series. To do this bring the data onto the screen and look for a command (in Economagic) called acf. Click on it to make Economagic generate the two graphs. Keep this screen open to compare these with the graphs to be generated in Soritec (in Part III).

 

Part III. Download the file economagic1.txt and read it in Soritec. Click on the link, and save it to a local file by choosing "Save As" in the File menu (in Netscape, or Explorer). The following information is stored in this file (SA stands for seasonally adjusted, SAAR stands for seasonally adjusted, annual rate):

  1. Column 1: year
  2. Column 2: month
  3. Series #1: Balance of Payments: Imports of Services; SA//bopimp
  4. Series #2: Spanish Pesetas to one U.S. Dollar; Exchange Rate//exspus
  5. Series #3: Inventory/Sales Ratio: Total Business SA//isratio
  6. Series #4: Real Personal Consumption Expenditures: Durable Goods; Billions of Chained 1996 Dollars, SAAR//pcedg
  7. Series #5: 3-Month Treasury Bill Rate, Auction Average//tb3ma
  8. Series #6: Unemployment Rate; SA//unrate

Use the following Soritec commands to read the data:

use 1992m1 2000m6
read ('c:\economagic1.txt') year month imp exspus isratio pcedg tb3ma nrate
on plot

Generate ACF (autocorrelation function) and PACF (partial autocorrelation function) for each series and try to identify an approximate ARIMA model to estimate.