For the ninth year in a row, hkp/// group has examined top executive compensation in a global context. This article gives insights on the methodology for gender pay gap analyses as well as on the
results of this year’s special analysis: compensation differences between female and male managers in selected countries world-wide.
The gender pay gap – as well as the methodology behind its calculation – has been widely discussed in the media. But questions have been raised about the reliability, quality and accessibility of the data, as well as the validity of the corresponding results. Indeed, it is very difficult to obtain a comprehensive data set – a data set that includes information on compensation, job value, degree of part-time employment, education level, number of children or duration of parental leave. That lack of comprehensive data makes analysis of the gender pay gap rather challenging and may, in fact, lead to contradictory results and misleading conclusions.
This is a challenge companies face even when it comes to analyzing their own internal data. Because the availability of quality data is so limited, it becomes even more important to apply a robust analytical framework to the information that is available in order to draw robust conclusions.
The unique hkp/// group approach of collecting and matching compensation data for its compensation surveys provides for an extremely high level of consistency and comparability of data across companies and countries. This enables us to conduct differentiated fair pay analyses, especially regarding compensation differentials attributable to gender.
hkp/// group Gender Pay Gap Methodology
This paper analyzes pay differentials between male and female incumbents in top and middle management roles in selected countries. hkp/// group employs a regression analysis, often used in scientific applications and especially effective in analyzing differences in compensation.
Accordingly, target direct compensation1 is expressed as a linear function of explanatory variables, such as gender, age, job family, job value and company and their corresponding coefficients. If the coefficient for the variable gender is positive and statistically significant, indicating that random effects can most probably be excluded, there is statistically valid evidence of a gender pay gap.
Using this regression analysis, hkp/// group is able not only to calculate differences in reported compensation between female and male managers, but also to deliver deeper explanations on several factors which might drive these differences. In order to examine possible reasons behind the gender pay gap, a stepwise regression analysis is conducted, considering additional variables in each step of the analysis.
An example for the stepwise calculation of the gender pay gap is shown in Fig. 1. Focusing on German data2, the difference in target direct compensation between males and females appears to be 15% when accounting for gender as the sole explanatory variable (unadjusted gender pay gap).