We propose a measure of the industrial gender wage gap which is free from an identification problem by using inter-industry wage differentials, or industrial wage premia. We draw on a recent literature showing that a normalized regression equation can be used to resolve the identification problem in detailed Oaxaca decompositions of wage differentials. By identifying the constant and the coefficients of dummy variables, including the reference category, the normalized equation can resolve the two key identification problems that arise in studying wage gaps: one in detailed Oaxaca decompositions; the other measuring industrial gender wage gaps.