Dear Stata User,

I am conducting a research which I try to explain the relationship between trade openness and social protection expenditure. Trade variable is a sum of import and export as a share of GDP. Social protection expenditure is a share of GDP as well. I run my fixed model as level-level form.

However I confused a little interpreting of coefficents.

After run xtreg social protection trade x2 x3 x4, fe

I get the following estimated coefficent for trade: -0.02243 with 0.015 p value.

I know that for level- level model we can interpret coefficient like this: “If you change x by one, we’d expect y to change by β1".

However, what "one" term refer to in my study? Does one unit change means one percentage point change in my study since my data is a share of GDP? But I do not run model log-log, that why I can not be sure about it!

How can I adapt this (“If you change x by one, we’d expect y to change by β1" ) logic to my research.

Thank you in advance any help.