How Do Different Policies Jointly Affect the Employment Prospects of Disabled Individuals? A Review of the Literature

1 Introduction

Disabled people include those who have long-term physical, mental, intellectual or sensory impairments which in interaction with various barriers may hinder their full and effective participation in society on an equal basis with others.^[1] Thus, disability itself stems from the interplay between individuals with a health condition, for example cerebral palsy, Down syndrome or depression, with personal and environmental factors including negative attitudes, inaccessible infrastructure, and limited social support.^[2] While some health conditions associated with disability result in poor health, others do not.^[3]

Over a billion people (15%) worldwide are disabled. Disability prevalence is higher among women than men, it increases with age and decreases with income. Disabled people face a higher risk of poverty and social exclusion.^[4] Concerning labour market outcomes, disability has a negative impact on employment prospects (Webber and Bjelland 2015), and earnings (Jones 2008). On average, disabled people (most of which are not employed^[5]) are less educated than nondisabled individuals (World Health Organization & World Bank 2011). However, disabled workers are more educated than their nondisabled counterparts on average (Jones 2010) and often overeducated with respect to the requirements of their jobs (Jones and Sloane 2010). This is consistent with disability onset leading to downward occupational movement (Jones et al. 2014). Thus, disability is associated with reduced odds of employment and lower job quality.

2 Current Situation

Disability policy must achieve two goals at the same time: labour market integration and income security. Disabled people must be empowered to engage in gainful employment, but they must also be provided with means to achieve an adequate standard of living, even if they are less productive than their nondisabled counterparts or unable to work (OECD 2003). Disability policies have been converging since 1990. In general, benefits have been getting progressively less generous, with a tightening of eligibility criteria. Most countries have placed an increased emphasis on active labour market policies (OECD 2010; Scharle et al. 2015), even though investments in such policies as percentage of GDP remain relatively low (Pignatti and Van Belle 2021). While demand-side activation policies have received little attention and funding (Adams et al. 2019), many countries have made benefit recipience conditional on participation in job training programs (Waddington et al. 2016).

Generous disability benefits have a small but statistically significant negative impact on employment among disabled people, while tightening eligibility criteria has no effect on employment (Barr et al. 2010; McHale et al. 2020). Disabled people no longer eligible for permanent benefits do not engage in the labour market, and end up with temporary benefits or no benefits at all (Jensen et al. 2019).

Regarding active labour market policies, it must be acknowledged that they are very heterogeneous. However, a broad distinction can be made between supply-side and demand-side policies. Supply-side policies aim at making disabled individuals more employable, for example through job training programs. Demand-side activation policies aim at stimulating the demand for disabled workers. They include (but are not limited to) employment quotas, wage subsidies, funding for workplace accommodations, and sheltered employment. Clayton et al. (2012) review several activation policies in four countries (Canada, Denmark, Norway, and the UK). They find no convincing evidence that job training programs are effective, noting that some interventions favour more advantaged disabled people who are close to the labour market. On the contrary, generous wage subsidies increase employment among disabled people. The same holds for workplace accommodations and involving employers in return to work planning, which are found to be effective even though uptake is low. Regarding employment quotas, their effectiveness depends crucially on the amount of the non-compliance penalty In Austria, increasing non-compliance penalties led to better employment outcomes for disabled people (Lalive et al. 2013; Wuellrich 2010), while in Italy (Agovino et al. 2019) and France (Barnay et al. 2019) non-compliance penalties are too low to affect hiring decisions. If anything, employers may be dissuaded from hiring disabled candidates by the cost of workplace accommodations (Fuchs 2014), despite the availability of financial support to face such expenses (Bellemare et al. 2018). Moreover, employment quotas often are concentrated in low quality jobs (Fuchs 2014). The OECD (2010) considers sheltered employment an effective activation policy only if disabled individuals can easily transition to the open labour market. People employed by sheltered workshops are economically inactive, have no right to a minimum wage and cannot join trade unions, even if the workshop provides goods or services for the open market (Visier 1998). There is some evidence that sheltered workshops protect employment during economic recessions, when disabled people are more likely to lose their jobs in the open labour market (Álvarez 2012).

As can be seen, the existing literature reviews on disability policy examine either benefits or activation policies, never both at the same time. Furthermore, the role of different welfare state systems is completely neglected by reviews of literature on disability and the labour market, and so is that of flexicurity. Finally, most literature reviews focus on employment itself, with no consideration about job quality. I aim to fill these gaps in the literature.

3 Theoretical Framework

Europe is characterised by five different welfare state systems (or simply “welfare systems”), which are relatively stable over time (Esping-Andersen 1990; Ferrera 1996). The liberal (or Anglo-Saxon) welfare system characterises all English-speaking countries and Israel (Doron 2001). It has means-tested assistance schemes, with low benefits and strict eligibility criteria. Moreover, employment protection is minimal, resulting in a very flexible labour market. The conservative (corporatist or Bismarckian) welfare system has been adopted by German and French speaking European countries, where redistribution is low and benefit entitlement is determined by one’s earnings, as benefits are often provided by employers. However, the freedom of the market is curtailed, resulting a certain degree of employment protection. The social-democratic regime is characteristic of Northern European countries, which promote full employment and income protection through a high level of decommodification (i.e. generous and universalistic benefits), high investments in activation policies as a percentage of GDP, and strong employment protection.

Welfare sceptics believe that generous benefits lead to the emergence of “dependency cultures”, weakening people’s motivation to engage in employment and making it socially acceptable to live on benefits (Heinemann 2008). On the contrary, the social investment perspective posits that benefits provide individuals with the resources they would otherwise lack in order to participate in the labour market (Midgley 1999). Therefore, the welfare scepticism view would predict worse employment outcomes for disabled people living in countries characterised by generous welfare systems and high employment protection, ceteris paribus, while the social investment perspective would predict the opposite. On the one hand, if disabled workers are “last hired, first fired” (Kruse and Schur 2003), greater flexibility will result in lower employment odds, especially during economic downturns (Reeves et al. 2014). On the other hand, potential employers may be more willing to recruit disabled job applicants if can fire them easily, should they not be a good fit for the job (Heggebø 2015). The welfare scepticism view and the social investment perspective have some common ground, as they both posit that high investments in activation policies foster employment.

Flexicurity has emerged as a way to reconcile flexibility (which leads to job creation) and economic security. The concept originated in the Netherlands, where employment protection for part-time workers was increased in the mid-90s. The Danish flexicurity model, on the other hand, combines reduced employment protection with generous benefits and an increased emphasis on activation policies (Bekker and Mailand 2019). A skill component is introduced as well, with high-skilled workers being more protected than their low-skilled counterparts (Danquah 2018; Heggebø 2016). Flexicurity may provide disabled people with options for partial work while keeping them economically safe, thus improving their employment prospects (Greve 2009).

4 Methodology

I used Google scholar to find literature on welfare systems, disability, health and employment by entering relevant keywords, and found works on flexicurity through snowballing. I then screened the literature for relevance. I selected works which use quantitative and/or qualitative methods to analyse how employment commitment and different employment outcomes among disabled people are affected by one or more of the following factors:

1. Disability benefits and activation policies.

2. Flexibility/Flexicurity.

3. Prevalence of temporary employment in the national labour market.

There is some overlap in the literature, with eight papers analysing the impact of benefits and activation policies, three of which consider flexibility as well. One paper analyses the effect of labour market flexibility by comparing Denmark to neighbouring countries (Heggebø 2015). Temporary employment is only investigated in two papers, both focused on Scandinavia.

The literature is summarized in Table 1 and ordered chronologically. It spans thirteen years, from 2007 to 2020, with the greatest number of articles published in the years 2015, 2016 and 2019. While a few scholars only contributed one work (e.g. Blekesaune 2007), most contributed more, often with co-authors.

The bar chart in Figure 1 lists all the countries considered in every work reviewed. As can be seen, while Scandinavian countries appear in most papers (Denmark in all of them), followed by the UK, other countries are mentioned only in one paper. This is the case for all non-European countries, as well as Malta. Furthermore, all the works reviewed focus on the Global North. Since the selected literature is very heterogeneous, it is important to illustrate how key variables are measured.

Figure 1:

Geographical distribution of the literature.

5 Results

Hypothesis 1 is not supported, while the social investment perspective is partially supported. Six papers (Blekesaune 2007; Holland et al. 2011; Kuznetsova et al. 2017; Reinders Folmer et al. 2020; Van der Wel et al. 2011, 2012) find that generous benefits are associated with better employment outcomes among disabled people, and one finds no effect (Geiger et al. 2019). Van der Wel and Halvorsen (2015) find that employment commitment is higher in more generous and activating welfare systems, although the difference in employment commitment between healthy people and people in poor health is greater. Therefore, the only two works which contradict the welfare investment perspective on the issue of benefits are the ones which conflate disability and poor health. This illustrates how the choice of disability model (Mont 2007) can affect results.

Evidence on hypothesis 2 is inconclusive, with half of the literature finding a positive effect of activation policies on the employment chances of disabled people, three papers finding no effect and one paper (Reinders Folmer et al. 2020) finding a negative association. Holland et al. (2011) remark that differences in employment rate between disabled and nondisabled people are smaller in Norway than in other Scandinavian countries, despite negligible investments in activation policies. Reinders Folmer et al. (2020) find that both demand-side and supply-side activation policies (taken separately) do not affect the employment odds of disabled individuals, while sheltered employment has a positive impact.^[6] Thus, they argue that supply-side activation policies, which make up the majority of activation measures, are ineffective.

Hypothesis 3 is not supported. In fact, Blekesaune (2007), Van der Wel et al. (2011) and Heggebø (2015) find that labour market flexibility has a negative impact on the employment prospects of disabled individuals, while Holland et al. (2011) find it has no effect. Heggebø (2016) notes that disabled people with high education are more likely to be recruited in Denmark than in either Norway or Sweden. However, this cannot be due to a more flexible labour market, as high skill workers benefit from high employment protection. Moreover, the literature on flexicurity that considers Europe as a whole, rather than focusing on Scandinavia, agrees that flexicurity has no effect on either the employment chances of disabled people (Heggebø and Buffel 2019; McAllister et al. 2015) or the likelihood of them retaining their job after disability onset (Backhans et al. 2016).

Evidence on hypothesis 4 is inconclusive. Heggebø (2016) finds that disabled workers are twice as likely to be in temporary employment in Denmark, and the same pattern is present (to a lesser extent) in Sweden, but not in Norway. He notes that temporary employment is more prevalent in Sweden than in other Scandinavian countries. However, Danquah (2018) finds that disabled workers are more likely to be in temporary employment only in Sweden, not in neighbouring countries.

6 Conclusion

The present work reviews the literature investigating the joint effect of different disability and labour market policies on the employment outcomes of disabled individuals. I formulate four hypotheses: three encapsulate the welfare scepticism perspective, while the fourth is derived from the social investment perspective.

When disability is measured by limitation in daily activities due to health problems, generous benefits have a positive impact on employment outcomes. Regarding activation policies, evidence is mixed. Neither labour market flexibility nor flexicurity foster employment among disabled individuals. In fact, there is partial evidence that flexibility has a negative impact on the employment chances of disabled individuals. There is no conclusive evidence linking disability to a higher likelihood of temporary employment. Overall, the welfare scepticism perspective is not supported, while the social investment perspective is partially supported.

To sum up, considering different policies separately can be misleading (Heggebø and Buffel 2019; Reinders Folmer et al. 2020). Moreover, the different implementation of flexicurity across countries may make it difficult to evaluate its effectiveness (Viebrock and Clasen 2009). Furthermore, activation policies are very heterogeneous both within and especially across countries (Fuchs 2014; Visier 1998), which might explain why evidence about them is mixed.

The strength of the present work lies in the inclusion of papers using different methodologies, which lends robustness to the results. Its main limitation is the fact that three out of the fourteen papers reviewed (Holland et al. 2011; Kuznetsova et al. 2017; McAllister et al. 2015) perform cross-country comparisons between countries with different institutional characteristics, adding no country level controls. Thus, the methodology used in those particular works is rather weak. However, the results they obtain align with those obtained using a robust methodology (e.g. Blekesaune 2007, Van der Wel et al. 2011).

Future research could further explore the effect of disability policies on job quality (Danquah 2018; Heggebø 2016) by considering the earnings differential between disabled and nondisabled workers across countries and/or welfare systems. Moreover, the literature on the impact of different policies specifically on recruitment of disabled workers (Heggebø 2016) and return to work after disability onset (Backhans et al. 2016) is still scarce.