With an unemployment rate of 17.1% compared to 10.2% for non-disabled people, people with disabilities face huge difficulties to enter the labour market. What are the reasons for this gap? Economists have difficulty answering this question, because behind the effect due to disability itself, discriminatory behaviour could be hidden.

Since 2000, the EU has formally recognized the equal treatment of people in the European Union (EU) in the workplace regardless of their disability. But if for other minorities we see the employment gap as unfair, the right or even legitimacy to work of people with disabilities is less consensual. Indeed, while “disability” encompasses many different conditions, what constitutes people with disabilities as a defined social group is the recognition of an inherent difficulty in participating in society : The United Nations defines a person with a disability as someone who has “[…] long-term physical, mental, intellectual or sensory impairments which, in interaction with various barriers, may hinder his/her full and effective participation in society on an equal basis with others.1[…]” And in the labour market, this difficulty to participate in society may lead to a potential lack of productivity, performance or even an inability to work.

However, the degree of disability, i.e the number of obstacles faced to take part in this or that activity of social life, is not only a matter of medical assessment but is highly related to the social environment, and another way to define people with disabilities is the individuals who face the same oppression: ableism,2 i.e. the discrimination in favour of the able-bodied. And ableism is not only about how society is organized, but also about individual behaviours toward people with disabilities, and in the labour market, this may be through outright discrimination by employers.

It is usually up to the employer to assess the ability of a job seeker based on different criteria and this is also the case for job seekers with a disability. While doctors may assess that the person is disabled, the employer is in the best position to understand whether the tasks to be performed in a job can be accomplished by a person. However, even if the choice to hire or not to hire rests with the employer, this discretionary behaviour is regulated and certain judgments are prohibited in estimating the productivity of the job applicant.

Types of discrimination are generally distinguished by economic theory into two categories: statistical and taste-based. The first is due to the imperfect information employers have about job seekers, so they judge on the basis of the group representation they have rather than on that of the individuals. The second occurs, on the other hand, when a group of individuals (employers or customers) prefers a certain group to another, based on taste rather than on any economic logic. In short, discrimination is the underestimation of a person’s productivity because of their affiliation with a specific group.

But for economists to assess whether employers are ableists or not may be even more difficult than to identify racist or homophobic employers. By taking inspiration from the Oaxaca Blinder decomposition3, we can explain an employment gap between a minority and the rest of the population by distinguishing between a proportion of the gap due to explained factors, i.e., social factors (such as the average education gap), and a residual, i.e., the proportion of the gap left unexplained. These two components are proxies for condition inequality (α) and discrimination (ε), respectively. While the former variable is reflected in the productivity gap at labour market entry between the two groups, discrimination is concerned with the average difference in getting a job between two groups of people who have the same average productivity when entering the labour market.

However, for people with disabilities, the model is more complex. One could modify the usual model by distinguishing not two, but three elements: the difference in productivity due to social components (α) (In the EU, one third of people with disabilities have a high school diploma compared to half of the rest of the population), the residual, i.e. discrimination (ε) and finally the difference in productivity due to disability (β). Indeed, while being black or a woman do not hinder your capacities to work, the disability itself can explain a part of the gap. And this third factor leads to methodological challenges. While in the first model only discrimination was unobservable, now it may be difficult to find a proxy for β, as it depends on the type of job and the disability considered. Thus, for economists, it is difficult to estimate the residual, i.e. the part of the employment gap due to discrimination.

This theoretical barrier can explain why studies on the hiring of people with disabilities face limitations in implementing effective methodologies. Estimating labor market discrimination against people with disabilities (ε) requires adapting methodologies applied with the standard model to the enhanced one characterized by the β, i. e. the nature of the disability and the way this latter prevents for working in this or that area. Indeed, due to the multiple realities that the term “disability” covers, the difficulty of access to the market depends very much on the condition of the job seeker and the job considered.

The most commonly used method for estimating discrimination in hiring is the correspondence test. It consists of sending resumes with the same characteristics to a sample of job offers, differentiating only two samples of resumes, the treated one, indicating the job seeker’s minority status, and the control. Then, if the difference in the average probability of getting an interview is statistically significant, we can assess that the average behavior of employers is discriminatory. However, the relevancy of this methodology in our case can be questioned because it focuses on discrimination at the first stage of the hiring process.

In a study published in 2020 in the Journal of Economic Inequality by Busetta et al, based on a correspondence test conducted in Italy, the authors claim that obese applicants appear to be significantly discriminated against compared to their non-obese counterparts for front office positions (exposed to tasted-based discrimination). Busetta et al [2020] justify their methodology by saying that one of the first studies on discrimination in the labour market, conducted by Allasino et al. (2004), found that 26.6% of discrimination occurs in the first stage of the hiring process. However, if it may be true for other discriminations but not for people with a disability. Indeed, it is more likely that the employer will try to evaluate the disability of an individual (β) during the interview rather than the first stage of the hiring process.

Secondly, it is difficult for economists to produce replicable and generalizable results about discrimination faced by people with a disability and assess an average α. They usually limit their study to one type of disability and distinguish their analysis for different types of jobs. In article published in the International Work Review in 2012, M.Malo and R. Pagan attempted to resolve this methodological difficulty. In order to distinguish the part of the wage gap due to the lack of productivity of people with disabilities and to the discrimination they face, they compare the wage gap between people with a mild disability and other workers. In this way, the authors control for variations due to the productivity differential. They conclude that there is no statistical evidence of a wage gap due to discrimination. However, this method does not seem relevant to us because it does not consider the interaction between β and α, i.e., it does not account for the fact that discrimination depends on the disability under consideration and its severity. Indeed, people with a mild disability do not face the same discrimination than someone who has a more severe disability!

Those methodological difficulties prevent an accurate understanding of the effects of policies aimed at closing the employement gap between people with and without disabilitiess. Indeed, estimating the correlation of ε with measures such as quotas or incentives is crucial to fully understand the impact of these measures. Implementing an affirmative action may have a positive effect on β, for example by reducing taste-based discrimination by customers. But an opposite effect can occur, as Stjin Baert et al. shows in a paper published in the European Journal Health Economics in 2016. By a correspondence test, he evaluated the effect of a Belgian subsidy, the Flemish Accompanying Subsidy, to encourage the hiring of people with disabilities. His results show that the wage subsidy did not influence the probability that a disabled applicant receives a positive response to a job. The author suggests that the positive financial stimulus implied by the subsidy is offset by signaling effects (subsidies are a signal for lower productivity) and fear of bureaucracy (excessive regulation and formality that can impede productive action and decision making). In other words, the negative effect on ε cancels out the positive effect on β.

Moreover, beyond the difficulties faced during the evaluation process, This lack of distinction between the two factors β and α is also visible in the way policies aimed at reducing the employment gap between people with and without disabilities are designed. Indeed, for other minorities, policies may address the two inequality factors separately, by prohibiting and punishing discrimination and by trying to reduce the skills gap (by promoting equal access to education, for example). For people with disabilities, the policy focus is slightly different. In Europe, three types of policy regimes can be distinguished to facilitate access to the labour market for people with disabilities.4 First, some countries establish mandatory quotas for organizations. If a company or public body fails to meet an employment rate for people with disabilities, it must pay a fine. Secondly, other countries have preferred more flexible policies, based on the principle of “non-discrimination”, which promote the employment of people with disabilities by making certain adaptations to the workplace and working conditions mandatory. Finally, a third group of countries simply implements incentives without imposing specific obligations. However, while these measures appear to be quite similar to those implemented for other minority groups, they are always accompanied by a compensation mechanism that recognizes the lack of productivity that hiring a person with a disability may entail. Thus, policies usually used to tackle discrimination have here a mixed objective, tackling at the same time β and α. Quotas are notably accompanied by subsidies such as in Germany, where even if companies are required to hire at least 5% of people with disabilities, they are entitled to a subsidy of about 70% of the employee’s salary. However, as discussed above [Baert 2016], this difficulty in targeting ableism separately can undermine policy effectiveness.

On December 3rd, in a French article in the national journal Le Monde, the founders of the French union of disabled people denounced how those institutional mechanisms can fuel a “compassionate and utilitarian” vision by employers and recruiters, because even if they are endowed with the same diplomas and training as able-bodied people, people with disabilities are presumed to be less productive because of the compensation mechanisms provided for their hiring. Therefore, distinguishing between β and α could allow to evaluate this underestimation of the capacities of workers with a disability, and draft policies that prevent from nourishing the vision of the work of a person with a disability as a favour granted by the employer or the State.


1In the European Union, 87 million people between the ages of 16 and 64 live with a disability, i.e. 18% of the active population

2 Pierre-Yves Baudot in Les Politiques du handicap, Politiques sociales : l’état des savoirs (2022), pages 97 to 114 (chapter 6)

3 Alan S. Blinder, « Wage Discrimination: Reduced Form and Structural Estimates », Journal of Human Resources, vol. 8, no 4, 1973, p. 436–455 and Ronald L. Oaxaca, « Male-Female Wage Differentials in Urban Labor Markets », International Economic Review, vol. 14, no 3, 1973, p. 693–709

4 Europe: l’emploi des personnes en situation de handicap Agefiph mai 2022