HowTo Run CG2

1. The runcg2.ksh script will call cg2_4v3 for you and setup everything the program needs to run.  The first line of the script should point to either the bash or ksh shell (by default /bin/bash).  Open the script in a text editor and make sure that the CG variable points to the correct location of the cg2_4v3 binary on your system.  The PROGDIR variable should contain the name of the directory containing the correct 02_size_calc.lst file.
2. At the prompt type ./runcg2.ksh or ./runcg2.ksh & to put the process in the background.
3. There are two log files.  Runcg.log contains output from the script, such as when CG2 started and finished, while cg.cglog contains output from CG2.  You can monitor the progress of CG2 by looking at cg.cglog periodically.  The size of the problem and memory allocated should match what you expect.  The first step is to read the data, the second stage preconditions the data, the third stage searches for the optimal paramater estimates, and in the final step the results are output.
4. Optional.  The cgout directory contains the synthetic data CG2 output files from the reference system.  Compare these with your cg.betas and cg.means file to test your installation.  

1. cg.betas - Text file containing your parameter estimates.  The covariates or right hand side variables are listed first, followed by the fixed person effects, and finally the firm effects minus one.  Since all of the person effects are estimated no constant is included in the model.  The covariates are correct as reported by CG2, but the person and firm effects are not identified at this point.  The post-processing imposes the identification rules in the next stage.
2. cg.means - The mean of the dependent variable and each covariate.  Can be used to check if CG2 read in the data correctly and are used in the post-processing.