Hidden Biases in European Public Administrations

We are in need of a much deeper conversation about hidden biases within European public administrations. New data may be able to help.

If you are an analyst or a decision maker, you will have noticed that a critical part of your job may be disentangling signals from noise in public debates.  By signal I mean the newly produced knowledge, ideas and data that help uncover structural aspects of a problem, as opposed to anecdotal ones.  One of the most critical  topics where this ability is important is that of diversity and inclusion in the workforce.      

How a society manages diversity may determine core aspects of its social fabric, such as how inequalities multiply, what opportunities are available to young people, and the extent to which social trust and collaboration is possible. If we carefully trace the signal, we are able to observe that many important debates have moved forward significantly in recent years. Both the #MeToo and #BlackLivesMatter movements pushed many organisations and industries to move away from ‘performative’ diversity management into more authentic and detailed discussions about the disadvantages faced by particular social groups.  While the industry of diversity management (where separating form from substance also comes in handy) has been flourishing for years, the latest discussions have prompted or given visibility to at least three important developments.

 First, more companies are voluntarily reporting yearly statistics on diversity in their workforces, which enables a more informed discussion about obstacles and possibilities (see the example of Big Tech). Second, a greater variety of initiatives are being implemented or trialled to alleviate organisational biases (such as allyship programmes, targeted recruitment, unconscious bias training, grievance procedures, etc.), albeit with different levels of success.  Third, there has been a revamp of the discussion on the so-called ‘business case’ for diversity, which now more clearly focuses on the empirical evidence on impacts and its different interpretations.

The public sector has also seen increased focus in diversity management. In June 2021, Joe Biden passed an ambitious executive order defining  sixteen ‘underserved communities’ (see Sec. 2) which require  policies and targets for  career development within the federal government. In the  latest Government at a Glance Report from 2021, the OECD systematically looked  at cross-country initiatives for including nine different disadvantaged groups in the public administration for the first time.  Furthermore, the Sustainable Development Goal 16 stresses the need to ‘build effective, accountable and inclusive institutions at all levels’ with diversity playing a centralrole.

On the academic front, the study of representative bureaucracies is the subfield of political science which more directly tackles issues of how diversity in  society is reflected in the public workforce. The theory of representative bureaucracy was originally developed alongside pluralist theories of representative democracies through the works of Frederick Mosher and Samuel Krislov among others. Its main argument is straightforward: a political system is better at considering multiple interests if all major social groups are represented in power arenas, including non-elective ones such as bureaucracy.  Significant empirical data has been produced on the impact of more representative bureaucracies, often concluding that under certain conditions, representation leads to more equitable policies and higher bureaucratic performance (see, for example, the meta-review by Ding, Lu & Riccucci 2021). 

However, the field is heavily impaired by one remarkable factor: the near inexistence of quantitative cross-country analyses on representative bureaucracies (exceptions are An, Song & Meier 2021 and Campbell 2021; see the comprehensive dataset of representative bureaucracy developed by Prof. Kenneth Meier and colleagues). 

In my latest article titled “Representative Bureaucracy and Perceptions of Social Exclusion in Europe”,  I make two substantial contributions to the existing discussion. First, I introduce an index of Bureaucratic Under-representation (IBU) that has unprecedented potential for studying cross-country patterns of social compositions in the civil service. Second, the index is multidimensional, covering five different aspects: gender, ethnicity, nationality, age and disability. The source data for a total of 37 countries between 2008 and 2018 (unbalanced) and a short explanation of the index are openly accessible. 

Table 1 shows the aggregate index and its subcomponents for the year 2018. The first column ranks 30 European countries with Sweden scoring as the only country where no social group (of the five) is under-represented in the civil service. The figures regarding individual components show that young people within the  population are heavily under-represented in most European countries while ethnicity and nationality tend to present more problematic challenges  to under-representation than gender (although, it should be noted that the index does not consider how positions of power have been allocated).

Table 1: Aggregate IBU index and individual components (2018)

Source: Cingolani (2022). Notes: green=overrepresented, red=underrepresented, white=balanced representation. For Poland, the survey did not include any non-national with information on all relevant variables. The index is then computed with the remaining four dimensions. For Hungary, Slovenia and Slovakia, 1 represents a scenario where the public sector has no non-nationals, in this case, the maximum level of underrepresentation in the individual components. 

Figure 1 also illustrates different country profiles over time, with Scandinavian countries having the most balanced representation while Mediterranean states tend to be more unbalanced. 

Figure 1: Multidimensional Index of Bureaucratic Underrepresentation (selected countries, 2008-2018)

Source: Cingolani (2022).

If patterns of bureaucratic representation can be measured more precisely, they can also be better managed. They invite fundamental questions that practitioners and social scientists interested in representation and diversity can further disentangle, such as: how does the social fabric of the public administration impact how policies are designed and for whom? Which policies or social mobility dynamics explain a more balanced bureaucratic representation? How are the capacity and efficiency of governments affected by different levels of diversity in their workforces?

In a global context where democratic principles are wavering, zooming in on the lesser-studied aspects of representation becomes critical. Well-thought indices and data can be a solid start. Democracy, after all, is also in the details.

Luciana Cingolani is Assistant Professor of Public Administration at Hertie School. Her research interest is in the interfaces between public administration and technology, in particular the expansion of administrative capacities, transparency and accountability as well as the role of e-services and open data. Luciana also works as a guest lecturer in the Master’s degree program Public Policy and Human Development at the United Nations University, UNU-MERIT, in Maastricht and is involved in the Horizon2020 project “Digiwhist” on the quality of public procurement in Europe. In addition to her scientific work, Luciana worked as a consultant for the United Nations Development Program, the French Development Agency and the OECD Sigma Project. She studied at the San Andrés University in Argentina and received her doctorate from the University of Maastricht.

Views expressed by the author(s) do not represent the Hertie School.