Weighted average least squares (WALS)

October 19 2009

The zipfile mpp.zip contains 3 folders and 5 files, as follows:

folders:

1. data (two excel files)
2. mpp_results (output of twelve runs of the program)
3. results (empty, your results are written to this folder)

files:

1. est_bma.m (matlab code for BMA program)
2. est_wals.m (matlab code for WALS program)
3. est_gts.m (matlab code for GTS program)
4. mpp_est.m (matlab code to run the previous codes)
5. readme.txt (explanatory notes)

GETTING STARTED

1. Unzip all files and folders to a new folder "mpp_averaging".
2. Start Matlab.
3. Change the current directory to "mpp_averaging" (top of the screen, just above the Command Window).
4. Run "mpp_est.m" (type "mpp_est" in the Command Window).

The estimation results from the paper for WALS, BMA, or GTS will now be displayed to the screen and saved in ".txt" files in the "mpp_averaging\results" folder.

To see that your results are the same as ours, compare with the output in the mpp_results folder.

See the comments in "mpp_est.m" for further information.

This code and the codes est_wals.m, est_bma.m, and est_gts may be used freely, but reference should be made to:

Magnus, J.R., O. Powell, and P. Prüfer (2009), A comparison of two model averaging techniques with an application to growth empirics, Journal of Econometrics, 154, 139-153.

The authors accept no responsibility for any errors or malfunctions in the program.

Tested with Matlab 7.4.0 (R2007a).

Comments and suggestions are welcome and should be addressed to Jan Magnus.

See also:

• CentER Applied Research
• CentERdata
• Faculty of Economics & Business Administration
• Rankings & Research

Contact:

CentER
PO Box 90153
5000 LE Tilburg
The Netherlands