More than a statistical analysis... Before all that we will do, we need an (or more than one) economic theory (theories). That is true that statistics will always help us, but not will be alone to help. All results can be significant, but unless there is not a connection with an economic theory.
So, an econometrician may not be a specialist on each economic theory, but at least has to know the recent literature of own research. But, in my opinion, to know a subject is more important than to do numerous analyses without knowing exactly the economic theory behind the research subject.
It is clear that, while explaining an economic relation, ignoring the economic theory behind this economic relation, is not reasonable.
Our knowledge about the theory is helpful and guiding about our expectations from our estimations. To know, to interpret and to use our findings, always there exists a need of an economic theory.