Trickytreex

Image

Trickytreex is a unique econometric software developed by Dr. Iman Sugema of the International Center for Applied Finance and Economics (Inter-CAFE) and the Department of Economics, Bogor Agricultural University. Benny Kapri acts as codes writer of the software.

Unlike any other softwares, Trickytreex runs on Microsoft Excel, seemingly as an add-in. However, it is not really an Excel add-in, in the sense that most of the features are run outside. But as users, you may simply think it as just an add-in.

There are two parts of Trickytreex; the pull down menu and the short-command (SC). The latter is provided as short of “programming language” in Excel. With SC you will be able to estimate almost all kind of estimation methods in econometrics such as GLS, GMM, ML, and NLS. SC is provided for advanced users but with less demanding knowledge of programming language. Anyone familiar with GAUSS or Matlab will find SC as a very interesting and flexible alternative. We have included 256 SC’s in the Excel and to use them you have to read the user guide. Unfortunately, the user guide for SC will be available by first January of 2014. So you have to wait for that.

The second part is the “user friendly” device: click on the menu and then run a regression. No programming is required. It is suitable for frequent operation. Under the current version, there are four features: single equation, binary choice, time series analysis, and panel data analysis. For the next version (updated by March 2014) you will also have ARIMA and Garch.

The single equation technique includes OLS, GLS (WLS) and IV (GMM). You can also run various post-estimation tests especially tests for heteroskedasticity, serial correlation and normality of the error terms. Moreover, you can adjust the covariance matrix by means of White HC or Newey-West HAC.

In the binary choice menu, you can do Probit and Logit estimation. We use Newton-Raphson non-linear procedure for the estimation. Under the current version you can only have “single equation” Probit and Logit. In the next version, we will include Panel Probit and Panel Logit. Make sure you update the software by March 2014.

With “time series” menu, you can do a lot of things, from data specification, to model estimation and tests on the error terms. The menu includes:

  • Unit root (ADF & PP)
  • Cointegration tests (Engle-Granger, Johansen)
  • VAR (estimation, impulse, FEVD, residual test)
  • Johansen Cointegrating VAR (Rank test, VECM, residual test, impulse, FEVD)
  • Two-step Engle-Granger VECM (more flexible modeling, VECM, residual test, impulse, FEVD)

In the “panel” data analysis you may have the following:

  • Fixed effect model (FEM)
  • Random effect model (REM), Fuller-Battesse procedure
  • Random coefficient model (RCM)
  • Non-stationary panel:
    • Unit root (Im-Pessaran-Shin and Breitungs)
    • Cointegrating panel (Pedroni)