the effect of the most important explanatory variable on the
dependent variable is subject to whether countries are widely
different in terms of their institutions. For countries that are
similar in their institutional structure, this explanatory variable is
not important in explaining (changes in) the dependent variable. Can
anyone suggest/point me towards proper econometric tests in this case?
If the effect of X depends on institutiional factor Z, then you
should be using an interaction of X and Z (as well as a main effect
of Z) because you want dY/dX to differ according to a country's level
of Z.
