On measuring comparative advantage: A note on Bowen’s indices