Ethno-Religious Discrimination in Access to Public Services: Does Organizational Ownership Make a Difference?

In most European countries, public welfare services ought to be delivered without consideration to clients' religion or ethnicity. However, while legislation and institutional guidelines prohibit discrimination, organizational design provides some wiggle room. Indeed, employees of street-level organizations have discretion in how they treat applicants, given them the opportunity to increase administrative burdens. For instance, they can choose whether to encourage a client to apply or not, withhold useful or necessary information, or simply increase the psychological costs to enroll. Burdens disproportionally increased for members of minority groups constitute discrimination. In line with the street-level bureaucracy literature, discretionary biases in the allocation of opportunities and resources can result in unequal treatment of minority groups – especially in the presence of adverse organizational incentives.

But why would frontline employees discriminate against ethno-religious minorities? Drawing on the theory of statistical discrimination, we argue that at the street-level discrimination can be conceptualized as a means of cream-skimming. This means that frontline employees would have an inherent incentive to focus on clients who are less costly to their organization. However, knowledge about the future costliness of individual prospective clients is often not available. Consequently, frontline employees use imperfect signals of future costs, like ethno-religion, that can be inferred from client’s names. In particular, absent other meaningful signals, administrative burdens are likely to be more severe for prospective welfare service clients with minority names.

We explore two implications of this framework. First, if frontline employees infer organizational costs from minority names, countering information should reduce discrimination. In particular, signals from prospective minority clients that they are not costly should reduce the administrative burdens imposed on them. Second, organizational incentive structure matters. In particular, organizations more able to impose diverging administrative burdens or with more financial reason to do so are likely to discriminate more. They are also more likely to be affected by countering signals. Consequently, inferred cost signals should matter more in organizations that operate under a profit maximizing maxim compared to public or nonprofit organizations that are bounded by a ‘non-distribution constraint’ limiting profits to be extracted for private gains in the organizations.

We test these propositions in a large-scale field experiment among 2,000 Danish childcare facilities employing an audit study type design. We send requests for information about admittance from fictitious prospective clients to those facilities, experimentally varying request sender names between a majority (Viktor) and a minority (Mohammed) condition. In addition, we experimentally include a countering signal of the future costliness of a client. We expect less administrative burdens - in the form of (helpful) replies to our requests - in the majority than in the minority condition, that this difference is countered by our direct cost signal, and that both effects are larger among private for-profit and non-profit organizations than in public organizations.