I have just heard about a toolbox in STATA, but what about programs in R, matlab, eviews?
STATA is really good and user-friendly for panel data models. For panel VAR models in particular, I think there is a recommendable program that you can download and use in your estimations. Check this: http://www.stata.com/statalist/archive/2010-08/msg01369.html
Many thanks for your answer Sabri, could you give more details about the way to estimate it with Eviews? I have version 7 installed, what is the command for estimating the panel var, is there a tutorial on the web?
STATA is really good and user-friendly for panel data models. For panel VAR models in particular, I think there is a recommendable program that you can download and use in your estimations. Check this: http://www.stata.com/statalist/archive/2010-08/msg01369.html
R and Matlab are programing hectic (though much better) software vis-a-vis Stata and Eviews. Eviews is the most user friendly software. However, the major draw-back for its use is that it makes you addicted to pressing buttons and getting things done without understanding them. This is quickly becoming an important issue as Eviews GUI usage wont be able to allow you to create a reproducible code which is quickly becoming a necessary requirement in research. Plus of course, you can't manipulate the things to customized designs. Stata offers the best of both worlds as it would provides you with not only an adequate GUI experience but would also make you acquire adequate programming skills. So I would suggest Eviews for a starter, Stata for intermediate and R and Matlab for advanced level programming users.
@Muhammad Hasnain Yousaf -- I agree about the drawback on the flip side of EViews GUI based ease of use. I will mention that one product with a very nice feature is OxMetrics. After you run a series of GUI operations you can select Alt+o and retrieve the batch code used for the prior set of computations. It fills in all of the options you specify, includes helpful commented sections, and is a good way to learn how to program in Ox much faster. Many functions are provided in code, as opposed to just binaries -- so you can see under the hood a bit.
Hi;
Herewith the R code I have written for VAR model in R:
library(vars)
library(timeSeries)
library(fAssets)
library(Rcmdr) ## input Dataset using “Rcmdr” package
library(timeSeries)
john=as.timeSeries(Dataset)##
###You can also use the following code to read the data:
## select the data from excel, copy it (using "control+shift+copy")
john=read.delim("clipboard")
###LAG SELECTION: min AIC value (of all lags), HQ, …
> VARselect(john, lag.max = 8, type = "both") ## we’ll use 2 lags
modelvar
In a case of Matlab, please go to http://www.paneldatatoolbox.com/
Yes, Eviews 8 can be the best alternative; it has recently included the Bayesian VAR model.
For panel data I recommend Stata. There are many good tools for panels (sorting etc.). Matlab is better for time series modelling and unusual techniques such as wavelets etc.
Eviews is for beginners, it is useful only if you program parts of the codes. Otherwise, it takes a time to go through the data and repeat estimations with different settings.
The advantage of R is in bayesian approach. There are many tools in this field...
R project: install.packages("panelvar")
Article Panel Vector Autoregression in R with the Package panelvar
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software