I do not follow " Random effects model does not consider the group effects such as change due to technical policy change, industry effects."

RE can handle multiple and cross-classified structures

see Structures at this site

http://www.bristol.ac.uk/cmm/software/mlwin/mlwin-resources.html#mlmbkgrd

and the papers at this project

https://www.researchgate.net/project/Fixed-and-Random-effects

Thanks Kelvyn Jones

If you are not satisfied with the result of your analysis, despite the reliability of your data. you can further increase the size of your sample and compare the results obtained. The adequacy between the type of panel (cylindrical or not) is important for some estimates.

Capturing the individual effect of each group, the fixed-effect model or the Within estimators are considered to be effective. On the other hand, the Hausman test offers us the appropriate model for our sample.

There is also the double fixed effect model with the double within estimators where both the individual and temporal effects are taken into account.

Without excluding, you can estimate by dummy variable too. The case of panel VAR, that supposes first that your variables are stationary preliminary, it is that happens rarely in the analyzes.

Abdi-Basid Adan ,

Thanks for your response

my sample size is 285 data points.

Can double fixed effect model be applied?

For sure You can estimate and then compare the results obtained between the fixed effect model, the temporal effect model and the dual fixed effect model.

Once you have fixed effects at a level then you cannot include group level variables as all the degrees of freedom at that level are used up , that is why random effects are so useful and the within-between specification or the closely related Mundlak specification allows this , and you get thhe same results as using a fixed effects model

Have a look at

https://en.wikipedia.org/wiki/Kelvyn_Jones

the line beginning

"He (with colleagues) has challenged the 'gold standard' that fixed effects should be the standard approach to the analysis of Panel data and that a Hausman test is an appropriate way of choosing between a Fixed effects model and a Random effects model."