What are we assuming when using inequality measures to quantify geographic concentration? An axiomatic approach