Measuring and decomposing the gender pay gap: A new frontier approach

We develop an enhanced method to measure and decompose the gender pay gap. This method is based on the Data Envelopment Analysis (DEA) technique and on the Malmquist Index (MI). We use DEA and the MI to construct an index reflecting an adjusted measure of the gender pay gap, taking into account multiple productive characteristics and multiple types of pay. Furthermore, we propose a decomposition of the gender pay gap into four components: same-gender pay gap; cross-gender frontier gap magnitude; productive characteristics bias; and pay package bias. The method is applied to data relative to 15,712 men and 17,175 women working as business and administration associate professionals in the finance and insurance industry in 20 countries. The results reveal the existence of a gender pay gap in all countries and that the value of the gap, and of its components, vary considerably between countries, requiring different policies to tackle it.