Introduction to PcGive 10
Preliminary comments
q PcGive 10 is the latest version of the econometric software package for econometric modelling written by David Hendry and Jurgen Doornik of the University of Oxford. You can find out more about the software at Doornik’s web page at http://www.nuff.ox.ac.uk/Users/Doornik/pcgive/index.html
q One key thing to know before you start is that PcGive is one of a number of program modules all of which operate through the “front end” program called GiveWin. So when you load PcGive it will automatically load up GiveWin. You use GiveWin to enter or load your data, to undertake preliminary data transformations or construct pre-regression graphs and to view your results. You switch to the PcGive module to formulate, estimate and test your regression models.
q You can find PcGive on the University of Portsmouth network via the Start button at the bottom left hand of your screen. Select
Start | Academic Applications | PBS | Econometric | PcGive 10
q You are advised to keep your data files in Excel 2.1 file format (a single worksheet file) with cell A1 empty, each column containing a series (column titles as text in row 1) and observation numbers (or dates) in the appropriate rows of columns A.
Examples of files with the appropriate format are
· L:\PBS\LectData\JudgeG\econ303\roses.xls and
· L:\PBS\LectData\JudgeG\econ303\UKEnergy.xls
First steps
Figure 1 shows what the screen should look like when you first load PcGive. You get the initial GiveWin window.
Figure 1 Initial GiveWin window.
Step 1 Loading a data file
To load your data file select
File | Open Data File
and then move to the folder containing the file you want. For example L:\PBS\LectData\JudgeG\econ303\UKEnergy.xls
Step 2 Deriving new series using the Calculator
To derive any new series, based on those already in the database, use the Calculator. Select Tools | Calculator and then use the keys to create the new series. For example to create GDPdefl = GDPcurr/GDP95 you can click on the names of the series in the database and the calculator button for / to get the formula as shown in Figure 2.
Figure 2 Using the Calculator in GiveWin
Click on the = button and type the name you require for the new series (GPDdefl). It will be added to the database.
NOTE: Until you resave the database the new series will not permanently be added to the file. I suggest that after each set of transformations is complete you resave the file ON YOUR OWN FILE SPACE (N drive or A drive – not the L drive).
Step 3 Creating pre-regression graphs
To plot the graph of one or more series in GiveWin select Tools | Graphics and then move the required series from the right hand database pane to the left hand selection pane. You can plot several different types of graph, either one at a time or with a number of graphs on the screen at the same time. You can save or print these graphs, or copy and paste them into a Word file.
Step 4 Formulating a regression model
Select PcGive from the left hand side window where all the modules are listed.
You should see the following
Figure 3 PcGive first window.
To run a single equation regression with time series data select Model | Single Equation Dynamic Modelling.
You then select relevant variables from the right hand database pane so that they appear in the left hand Model pane (see Figure 4 below).
Figure 4 Formulating a model
Be careful to have the lag length query button set at 0 (unless you want to include a lag). You don’t need to create these series separately – PcGive will do it automatically. Note you also don’t need to create the Trend – you can take it from the “Special” pane. A Constant is automatically copied across to the Model pane.
NOTE: May sure that your dependent variable has a Y next to it (if you copy this across first, PcGive will assume that this is the Y variable).
When you are ready click the OK button.
Step 5 Estimation
Select Ordinary Least Squares. The next dialog box allows you, if you wish, to select only a subset of the data file for the regression, or to set aside the last few observations for a post-sample forecast test. See Figure 4.
Figure 4 OLS estimation dialog box.
The results will be placed in the GiveWin window. You can save them, print them or copy and paste them into Word.
Step 6 Post-regression testing
You can conduct a variety of post-regression tests. For example in PcGive select Test | Test Summary to get a summary table of diagnostic test statistics.
This completes the basic introduction. To find out more, explore the program itself, make use of the program’s Help file, or work through the tutorials on Doornik’s web site.
Guy Judge, November 2002.
Appendix
The results you should get are shown below. Note that GiveWin keeps a record of your data transformations.
---- GiveWin 2.02 session started at 13:25:27 on Thursday 28 November 2002 ----
UKEnergy.xls loaded from L:\PBS\LectData\JudgeG\Econ303\UKEnergy.xls
Ox version 3.00 (Windows) (C) J.A. Doornik, 1994-2001
---- PcGive 10.0b session started at 14:10:08 on 28-11-2002 ----
Algebra code for UKEnergy.xls:
GDPdefl = GDPcurr/GDP95;
NPEnergy = Xw/Ew;
RPEnergy = NPEnergy/GDPdefl;
LEnergy = log(Ew);
LGDP95 = log(GDP95);
LRPEnergy = log(RPEnergy);
EQ( 1) Modelling LEnergy by OLS (using UKEnergy.xls)
The estimation sample is: 1967 to 1996
Coefficient Std.Error t-value t-prob Part.R^2
Constant 3.60236 2.360 1.53 0.139 0.0852
LGDP95 0.641813 0.1861 3.45 0.002 0.3224
LRPEnergy -0.154819 0.04711 -3.29 0.003 0.3017
Tave -0.0195305 0.009442 -2.07 0.049 0.1461
Trend -0.00982097 0.004239 -2.32 0.029 0.1768
sigma 0.0228079 RSS 0.01300504
R^2 0.742597 F(4,25) = 18.03 [0.000]**
log-likelihood 73.5861 DW 0.584
no. of observations 30 no. of parameters 5
mean(LEnergy) 11.8884 var(LEnergy) 0.00168413
1-step (ex post) forecast analysis 1997 to 2001
Parameter constancy forecast tests:
Forecast Chi^2(5) = 0.36962 [0.9961]
Chow F(5,25) = 0.066857 [0.9966]
AR 1-2 test: F(2,23) = 9.0900 [0.0012]**
ARCH 1-1 test: F(1,23) = 0.84547 [0.3674]
Normality test: Chi^2(2) = 0.38170 [0.8263]
hetero test: F(8,16) = 0.90007 [0.5392]
Not enough observations for hetero-X test
RESET test: F(1,24) = 1.8452 [0.1870] |
