Appendix C: The Fair-Parke Program for the Estimation and Analysis of Nonlinear Econometric Models

The Fair-Parke program allows all the techniques discussed in this book to be used automatically once the necessary information on a model has been read by the program. The necessary information consists of a few FORTRAN subroutines and a few data sets to present the stochastic equations and the identities. Once the program has this information. almost all the techniques can be used with no further programming. (In a few cases, such as FIML estimation, the user must supply additional information. In the FIML case, for example, information on the Jacobian must be supplied. These exceptions are discussed later in this appendix.) This has the obvious advantage of allowing many things to be done with only one setup, and it also means that the model only needs to be debugged once. It is quite easy, as will be seen, to check coding errors, and once these errors have been corrected, one need not worry about further coding errors for any of the techniques. The model is represented by (6.1), which is repeated here: