Standardized coefficient with panel data

[Mounira NAKAA]

Dear members,

I am running a rregression with panel data and I standardized all the variables ( dependeant and independant).
My independant variables are interaction terms.
Here is my code

xtreg std_exports std_institution std_capital std_skill i.year, fe vce(cluster ctrind)

I have two questions:
- Is it possible to standardize the coefficients in this case
- Is it normal that the constant is not zero? Usually there is no intercept when using standardised coefficient...

Many thanks for your help

[Phil Bromiley]

You'll increase your changes of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

Some on this listserve would recommend not standardizing variables that have meaningful metrics - it just makes things harder to interpret.

That said, you can do this and I guess you can talk about a predicted change in exports measured in standard deviations of exports for a standard deviation change in x. As for intercepts, they are not particularly interesting in fixed effects - you're assuming different intercepts for each panel anyway.

[Mounira NAKAA]

Many thanks for your answer Phil!