Hello;
I work on panel data, the Hsiao test resulted in total heterogeneity and I do not have sufficient data to make an estimate per individual. what to do?
is this test necessary?
first try to expand your sample, if not possible depending on your problem, you will use hasio test or not
, but if you have a heterogeneous total model you have to estimate equation by equation
Hello Charrass,
the objective is to test if your panel is homogenous. I am not sure what software you are using.
if your sample has no sufficient data per person, i suggest to first present some descriptive look for example on kurtosis..
Then at a second stage, you can run multiple models and compare RSS by hand and do the test manually.
In most of cases you are invited to control for individual dummies. I am not sure about a decision for your case as I don’t know exactly your data
Heterogeneity, like cross sectional dependence test is crucial when estimating panel models. Otherwise, the model will fail some underlying assumptions. Thus, the estmiates produced from such analysis. The result you have does not change or affect your work significantly. Having found total heterogeneity, the next step is to control for heterogeneity inherent in your panel data. This implies you have to use a method that accounts for heterogeneity.
- Badges
- Science topic
- Similar topics
- Mathematics