Assessing the gender gap in labour market index: volatility of results and reliability

Purpose – The construction and use of composite indicators has recently increased considerably because such indicators permit the comparison and ranking of countries with respect to complex phenomena of global importance. Together with the diffusion of these composite indicators, a major debate has emerged over their real capacity to produce objective and reliable results. The paper aims to discuss these issues. Design/methodology/approach – In this paper, an analysis of the results of the Gender Gap Labour Market Index (GGLMI) – an ad hoc composite indicator constructed by the authors to analyse the gender gap in European labour markets – is proposed, in order to identify the most appropriate choices in the construction of composite indicators, to obtain reliable results. This is a sensible approach, especially when various aspects (some of them controversial) have to be synthesised. Findings – In addition to highlighting the greater robustness of non-compensatory methods, this analysis confirms the variability of countries’ rankings on graded positional lists, and suggests the need to use a procedure based on a set of alternative methods conveniently combined in each step of the composite indicator construction. Research limitations/implications – The research identifies the weak steps in composite indicator construction and the data characteristics which increase the volatility. Only an approach based on various alternative paths can control for it. Practical implications – It offers some reflections on the extent to which gender disparities in labour market still persist in European countries and the different ways in which it manifests. Social implications – The analysis of characteristics and policies activated in countries at the top of the ranking can suggest initiatives to promote gender equality. Originality/value – Besides of testing the robustness of results, an effort in order to identify the best way to obtain a synthetic and reliable single rank is made. Further, the results produced by the GGLMI for the year 2011 are presented and discussed.