I am working on long panel with N=365 and T=8. My model is that DV is a function of Independent variable, lagged value of Independent variable and lagged value of Dependent variable. DV= f(DV(t-1), IV, IV(t-1)). Should I use ARDL model or GMM?
In this situation, you should use a dynamic panel data model, such as the Arellano-Bond or Blundell-Bond estimators. These estimators are specifically designed for panel data with lagged dependent variables, and they also account for potential autocorrelation. The ARDL (autoregressive distributed lag) model is more suitable for cross-sectional data, while the GMM (generalized method of moments) is typically used for panel data without lagged dependent variables.
It depends on the nature of your data and the research question you are trying to answer. Both methods have their own advantages and disadvantages. ARDL is a cointegration technique that can be used to estimate long-run relationships between variables. It is useful when you have non-stationary variables and you want to estimate a long-run relationship between them. GMM is a method of moments estimator that can be used to estimate parameters in models with endogeneity. It is useful when you have endogenous variables and you want to estimate the parameters of the model.
The choice between Panel ARDL and GMM depends on the nature of your data and the research question you are trying to answer.
Panel ARDL is used when you have a panel data set with both stationary and non-stationary variables.
GMM is used when you have a dynamic panel data set with endogenous variables
- Badges
- Science topic
- Similar topics
- Thesis Research