PROS & CONS OF RATS AND TSP

Pros and Cons of RATS and TSP.

is a good Time Series Analysis software which include popular Time Series Analysis Techniques, eg. Vector Autoregressions, Box-Jenkins analysis, Spectral Analysis, Estimating/Forecasting ARIMA models, transfer functions and intervention models
DOS version (RATS386) contains user-friendly graphics options, an excellent data utility called "RATSDATA" which can import (export) from (to) ascii/spreadsheet/RATS format data file easily and interactive sessions which can use mouse to select options from the menus easily
provides high quality graphs which can be easily export to other popular windows applications like Corel Draw, Microsoft Word, or WordPerfect. It can exports graphs to PostScript, HPGL, PIC, Windows Metafile, and Macintosh PLCT.
abundant supply of RATS programs and procedures available in various public ftp sites
There is no limitation in the number of characters of the series labels for both DOS and UNIX version of RATS. Underscore is also allowed.
There is excellent on-line help in DOS RATS.

lack of general matrix programming options/functions (eg. have to use a very clumsy way to program how to stack/concatenate each vectors horizontally or vertically. This cons will not affect if one is not an advanced applied econometrics user (i.e. programing the tools by yourself)
not good at structural estimation when comparing with TSP, eg. only has one numerical iteration algorithm (option) to compute the standard errors (Hessian and Gradient) when conducting Maximum Likelihood estimation.
Poor options/applications in Kalman Filter Analysis when comparing with TSP.
RATS is not available in LINUX at this moment.
There is no on-line help in UNIX RATS.

is a good Time Series Analysis and Structural Estimation software, eg. Kalman Filter Analysis, Monte Carlo Analysis, Maximum Likelihood Estimation, Nonlinear Least Squares
more general matrix programming options/functions than RATS
sufficient data utility to read/export an ascii data file
has a nice "dot" programming feature which can be very convenient if working on Monte Carlo Analysis or G7 countries
good at doing structural estimation when comparing with RATS, eg. has several numerical iteration algorithms (option) to compute Hessian and Gradient.
has better options/applications in Kalman Filter Analysis than RATS.
QED also provides LINUX version of TSP. This can allow one to run a moderate size job in LINUX instead of crowded QED server. However, it is recommended to run a huge job in Sargan machine since LINUX's speed is slower than UNIX.
There is on-line help in all of the versions of TSP.

do not have good and user-friendly graphics options like RATS386
do not have Spectral Analysis
limited supply of sample programs or procedures available in public ftp sites
weak at Vector Autoregression analysis applications
There is a limitation in the number of characters (maximum 8) of the series labels for both DOS and UNIX version.

Copyright @1996 Irene Chan
Last Modified: November 28, 1996
If you have any comments, please feel free to email me chani@qed.econ.queensu.ca