README.TXT

 ************************************************************************************************
 * Giorgio Calzolari (2015-2017)                                                                *
 * Indirect estimation and econometrics exams: how to live a round life.                        *
 * University of Firenze. Department of Statistics, Computer Science, Applications "G. Parenti" *
 * DISIA Working Paper 2015/01                                                                  *
 ************************************************************************************************

 You have unzipped the ROUNDNUM zip file.
 Put all files into the same directory.
 If your operating system is a Windows 32 or 64-bits (from Windows XP till Windows 8),
 you can directly run one of the models simply clicking on the .EXE file.

 For instance, clicking on ROUNDNUM-REGRE1.EXE, the program will read the names of the students from the NAMES.TXT file;
 it will produce for each student one ASCII file containing the dataset and describing the linear regression model
 (to be estimated by OLS), called EXAMxxx.TXT (xxx being the order number of the student; EXAM001.TXT will be for the student
 whose name is the first in the NAMES.TXT list; EXAM002.TXT for the second, and so on).
 Also a file EXAMOUT.TEX will be produced, containing all the main results expected for each student; this file is ready
 to be processed by Latex (otherwise, it can be simply read as an ordinary ASCII file).
 The format of the input file ROUNDNUM-REGRE1.DAT should not be modified. In the third line it contains the initial seed
 of the random numbers generator (an integer number that can be modified)
 The format of the fourth line should not be changed
    60   60    1    1    4    4 0001 0040 0001 0090    0
 only 0090 can be modified (it is the maximum sample-size of each output).
 To produce datasets for a "real" list of students, their names must replace the names in NAMES.TXT (or added to the example
 names; the format must be the same as it is in the example).

 If your operating system is different, you must have a FORTRAN-77 compiler installed in your computer.
 Then you should "COMPILE, LINK and GO" the file ROUNDNUM-REGRE1.FOR

 The procedure is analogous to what described above for each model.

 ROUNDNUM-REGRE1 produces the data for cross-section linear regression models to be estimated by OLS.
 ROUNDNUM-REGRE2 produces the data for time-series linear regression models to be estimated by OLS.
 ROUNDNUM-REGRE3 produces the data for dynamic regression models, where Wald, LR and LM tests should be performed.
 ROUNDNUM-GARCH  produces the data for linear regression models with GARCH(1,1) noise, to be estimated by Gaussian ML.
 ROUNDNUM-LOGIT  produces the data for logit models, to be estimated by ML.
 ROUNDNUM-PANEL  produces the data for linear panel data models with random effects, to be estimated by Gaussian ML.
 ROUNDNUM-ARMA   produces the data for time series model (MA(1), MA(2), AR(1), AR(2) or ARMA(1,1)), to be estimated by Gaussian ML.
 ROUNDNUM-KLEIN1 produces the data for simultaneous equations models, "similar" to Klein-I model, to be estimated by OLS, or 2SLS, or 3SLS.

 Remark: All the programs "suffer" from their origin. They were born inside a more complex software structure,
 often with quite different purposes. They have simply been adapted to the purpose of producing exams datasets,
 without any "optimization". Thus, expert programmers would easily find single statements, groups of statements, or entire subroutines
 that could be modified and/or suppressed.

 Giorgio Calzolari
 Firenze, January 2015 - April 2017