Introduction
Health and unemployment are interrelated (Olesen et al., Reference Olesen, Butterworth, Leach, Kelaher and Pirkis2013). Health problems are a major determinant of entries into (Riphahn, Reference Riphahn1999) and exits from unemployment or benefit receipt (Beste et al., Reference Beste, Coban and Trappmann2023; Loetters et al., Reference Loetters, Carlier, Bakker, Borgers, Schuring and Burdorf2013), and unemployment has adverse health effects (Krug and Eberl, Reference Krug and Eberl2018). Poor health also increases the likelihood of leaving the labour force completely (Riphahn, Reference Riphahn1999). Consequently, there is a strong overlap of health impairments and non-employment. From a policy perspective, this is reflected in high international diversity of access conditions to disability benefits versus unemployment insurance.
In many countries such as the UK, the Netherlands, Sweden, and Finland, there has been concern over high rates of disability benefit receipt among the working-age population (Alanko & Outinen, Reference Alanko and Outinen2016; Lindsay & Houston, Reference Lindsay and Houston2013a). Germany faces the opposite situation, with low accessibility of incapacity pensions and high proportions of people with health impairments among the unemployed (Brussig & Knuth, Reference Brussig, Knuth, Lindsay and Houston2013; Konle-Seidl et al., Reference Konle-Seidl, Rhein and Trübswetter2014). The status of unemployment benefit recipient exposes many non-employed people with health impairments in Germany to activation policy and jobcentre counselling. Raising employment is central to EU strategies (Bakker & van Vliet, Reference Bakker and van Vliet2022) and is seen as a means toward social integration and participation, as reflected in Germany’s activation policy (Rauch & Dornette, Reference Rauch and Dornette2010). Indeed, societal participation of people with disabilities has become an international political goal, as discussed for instance by Hästbacka et al. (Reference Hästbacka, Nygård and Nyqvist2016), with reference to the UN Convention on the Rights of Persons with Disabilities.
While employment can give access to important prerequisites of societal participation, such as financial resources and access to social insurance schemes, not all jobs are well paid nor provide benefits (Brucker & Henly, Reference Brucker and Henly2019). In contexts that prioritise the employment integration of people with health impairments or disabilities, access to good quality jobs is decisive, if employment is to contribute to their societal participation.
The analyses in this paper focus not only on entries into regular socially insured jobs, but also provide evidence on the relationship between health impairments and entries into uninsured minijobs, as an important aspect of employment quality. These findings contribute to an understanding of how activation for people with disabilities affects their societal participation. Moreover, the paper offers findings on how specific types of health impairments affect job entry rates. Further analyses add to knowledge on the jobcentre relationship of benefit recipients with health impairments, which has previously been found to vary strongly internationally (Geiger, Reference Geiger2017). Specifically, findings show how job search requirements depend on health impairments, and, on the basis of timing-of-events hazard models, how counselling affects employment opportunities. The analyses draw on survey data from the Panel Study Labour Market and Social Security, linked to administrative employment and benefit spell data.
The next sections discuss social policies for non-employed people with health impairments in international comparison, the relevance of employment quality for societal participation, and jobcentre experiences and counselling effects for people with health impairments. After discussing the data and method, the empirical findings are presented. The final section sets these findings into context and discusses their policy implications.
Social policies for non-employed people with health impairments in international comparison
It is not always possible to draw a clear line between disability and unemployment. Geiger et al. (Reference Geiger, Garthwaite, Warren and Bambra2018) observe that disability assessment is at the heart of social security, dividing disability benefit claimants from those receiving unemployment benefits. In this context, Alanko and Outinen (Reference Alanko and Outinen2016) point out that policy discussions have often focused on whether unemployed people have been miscategorised as disabled, and vice versa. They suggest that instead of focusing on such a polarisation, it is more helpful to conceive of groups and categorisations of labour market inactivity as forming a continuum between disability and unemployment.
Internationally, access conditions to disability versus unemployment benefits as well as the resulting proportions of the non-employed population receiving the respective types of benefits vary greatly. Konle-Seidl et al. (Reference Konle-Seidl, Rhein and Trübswetter2014) find that in Sweden and Denmark, approximately half, in the Netherlands and Great Britain, approximately a quarter, in Germany, 16 per cent, and in Spain 10 per cent of the overall non-employed population are long-term invalidity benefit recipients. High rates in disability pensions have led to discussion of a ‘disability crisis’ in many countries, such as the UK, the Netherlands, Sweden, and Finland (Alanko & Outinen, Reference Alanko and Outinen2016; Lindsay & Houston, Reference Lindsay, Houston, Lindsay and Houston2013b; Ulmestig, Reference Ulmestig, Lindsay and Houston2013; van Berkel, Reference van Berkel, Lindsay and Houston2013). Ensuing policy reforms have tightened access conditions to disability benefits, which were perceived to have been too lenient and to have resulted in hidden unemployment, long-term inactivity, and lost employment potential. In Australia and the United States, reforms have likewise focused on encouraging the labour market integration of people with disabilities (Collie et al., Reference Collie, Sheehan and Lane2022; Giordono, Reference Giordono2022).
The situation in Germany is rather the opposite to that in the UK, the Netherlands, and Sweden. As pointed out by Brussig and Knuth (Reference Brussig, Knuth, Lindsay and Houston2013), there is strict gatekeeping for incapacity pensions in Germany. Thus, there has never been a discussion of a ‘disability crisis’. Incapacity pensions are part of the general pension system, which is contribution-based. Eligibility and benefit rates depend on contributions while in employment. To qualify for incapacity pension, applicants must pass a medical examination by a physician commissioned by the pension fund (Brussig & Knuth, Reference Brussig, Knuth, Lindsay and Houston2013). In addition, Brussig and Knuth (Reference Brussig, Knuth, Lindsay and Houston2013) find that incapacity pensions are financially unattractive, with payments at a similar level or lower than basic income support for jobseekers.
Instead, basic income support for jobseekers is the main benefit for people with no or low income in Germany. Basic income support for jobseekers is a means-tested benefit for people without unemployment insurance or who have run out of their unemployment insurance benefits. To be eligible for basic income support, one needs to have a household income below the eligibility threshold. One also needs to be capable of working for at least three hours a day, as determined by the employment agency (Social Code II). Those with a working capability of less than three hours a day, certified by the pension insurance, can apply for social assistance or incapacity pension instead.
People who are only temporarily ill or injured, but can be expected to become capable of employment again in the foreseeable future, qualify for basic income support for jobseekers (Social Code II). However, they are generally not expected to be available for employment or activation measures until their health improves again. On the basis of survey data, Trappmann et al. (Reference Trappmann, Ramos Lobato, Unger and Lietzmann2019) estimate that approximately 184,000 basic income support recipients in the year 2016 have not been capable of employment for at least two years. Thus, it appears that people with severe health impairments remain in the system of basic income support for jobseekers for extended periods of time, even when they are not able to work at all.
The total group of benefit recipients with severe health impairments includes not only those who are (temporarily) not able to work at all, but also those with a reduced working capability or who report severe health impairments despite being able to work. Brussig and Knuth (Reference Brussig, Knuth, Lindsay and Houston2013) find for 2007/08 that 37 per cent of benefit recipients report a daily working capability of less than eight hours, and (Stockinger & Zabel, Reference Stockinger and Zabel2020) find an increase of the proportion with severe health impairments from 38 per cent in 2007/08 to 53 per cent in 2018. Standardised for the age distribution of the general population, Eggs et al. (Reference Eggs, Trappmann and Unger2014) show for 2012 that 55 per cent of male and 45 per cent of female unemployed basic income support recipients report severe health impairments or a recognised disability. These figures are much higher than for the working population without benefit receipt (19 per cent for both men and women).
Thus, work capability conditions for basic income support for jobseekers are often interpreted quite broadly. This is in line with the activation policy paradigm. Since the labour market reforms of the early 2000s, policies have emphasised comprehensive activation. Even people with very low employment capability are intended to have access to the labour market and to Active Labour Market Policy (ALMP) (Rauch & Dornette, Reference Rauch and Dornette2010).
Job quality and societal participation of people with health impairments
In aiming to counteract social exclusion of people with disabilities by raising their labour market participation rates, German social and employment policy follows EU employment strategies (Annesley, Reference Annesley2007; Bakker & van Vliet, Reference Bakker and van Vliet2022; Rauch & Dornette, Reference Rauch and Dornette2010). These policies conceive societal participation and labour market participation to be closely linked. Several theories describe mechanisms by which unemployment can have detrimental effects on social integration and psychological well-being. Jahoda’s (Reference Jahoda1982) theory of latent deprivation indicates that, apart from its manifest functions, such as providing income, employment also has latent functions, providing time structure, social contacts, common goals, status or identity, and regular activity. Jahoda argues that these functions are lost by unemployment, while Fryer (Reference Fryer1986) particularly emphasises the loss of agency, and Ezzy (Reference Ezzy1993) the loss of social status (Sage, Reference Sage2018). Sage (Reference Sage2018) observes that academics have particularly drawn on Jahoda’s theory of latent deprivation when making policy recommendations. Thus, contemporary policies are often based on the assumption that employment is generally conducive for health and well-being, and therefore any employment preferable to no employment.
Reemployment has indeed been shown to have positive effects on quality of life and health (Carlier et al., Reference Carlier, Schuring, Lötters, Bakker, Borgers and Burdorf2013). However, lower quality employment, such as fixed-term employment, has been shown to be related to worse health outcomes than permanent employment (Hammarström et al., Reference Hammarström, Virtanen and Janlert2011). Moreover, intrinsic aspects of work quality connected to societal participation, such as tasks perceived to make a meaningful contribution to society, can be important for well-being (Fisher, Reference Fisher2022). Financial resources as well as access to benefits such as health or pension insurance brought about by good quality employment can likewise be beneficial for societal participation.
People with health impairments may have difficulties finding better-quality employment that enables participation, due to discrimination or because they are restricted concerning the number of hours or types of tasks they can carry out, which may lead to competitive disadvantages. For instance, Brucker and Henly (Reference Brucker and Henly2019) and Jetha et al. (Reference Jetha, Martin Ginis, Ibrahim and Gignac2020) find lower job quality for workers with disabilities in the United States and Canada, respectively, and find that this is driven by lower self-rated health. The analyses in the present paper aim to contribute findings on effects of health impairments on opportunities of obtaining insured and uninsured jobs, as a measure of job quality that is relevant for societal participation.
Jobcentre counselling for people with health impairments
Jobcentre counselling plays a key role in implementing policy goals of employment integration, but the actual jobcentre experiences of benefit recipients with health impairments can vary greatly across countries. Findings for the UK point towards high levels of anxiety experienced by disabled benefit recipients in the context of Work Capability Assessments (WCA) following reforms aiming to reduce levels of disability benefit receipt. WCAs were perceived to inadequately reflect actual work capacity, and were often delayed, exposing recipients to intensive conditionality during the waiting period (Irvine & Haggar, Reference Irvine and Haggar2023; Wright et al., Reference Wright, Robertson and Stewart2022). At the same time, work coaches have been found to use their discretion to reduce conditionality for benefit recipients with mental health problems (Wright et al., Reference Wright, Robertson and Stewart2022).
In a typology of the implementation of benefit conditionality for disabled people, Geiger (Reference Geiger2017) accordingly classifies the UK as a compliance-based system with high conditionality and a weak link to rehabilitation. In contrast, Denmark and the Netherlands are classified as demanding systems with high conditionality strongly linked to rehabilitation, and Norway and Germany as passive systems with low conditionality weakly linked to rehabilitation. In Denmark, interdisciplinary rehabilitation teams in jobcentres refer claimants either to work tests, long-term rehabilitation programmes, or subsidised employment. In the interdisciplinary rehabilitation teams, the health professional competencies of occupational therapists are found to play a significant role (Christensen et al., Reference Christensen, Hald, Nielsen, Nagaraj-Sithamparanathan and Ilvig2021).
Kupka et al. (Reference Kupka, Oschmiansky and Popp2017) find on the basis of a qualitative study for Germany that many benefit recipients with mental health restrictions would welcome job search support by the jobcentre, but feel their potential is underestimated. Some report no job offers by the jobcentre at all or only offers that do not fit their capabilities. As Kupka et al. (Reference Kupka, Oschmiansky and Popp2017) note, jobcentre performance quotas represent disincentives for investing in the employment opportunities of benefit recipients with low initial employment prospects. Brussig and Knuth (Reference Brussig, Knuth, Lindsay and Houston2013) find a similar amount of counselling sessions, but less ALMPs for benefit recipients with poorer than for those with better health. This may be an explanation for the lower satisfaction with jobcentre counselling that (Stockinger & Zabel, Reference Stockinger and Zabel2020) find for benefit recipients with health restrictions. When those with mental health impairments actually participate in ALMPs, they profit more than others (Tübbicke & Schiele, Reference Tübbicke and Schiele2023).
Kupka et al. (Reference Kupka, Oschmiansky and Popp2017) find that benefit recipients with mental health restrictions who were referred to case management, a comprehensive counselling that additionally addresses issues such as housing problems, debt, or health, reported positive experiences in terms of support and offered measures. However, funding for case management is restricted, such that not all benefit recipients with health restrictions could participate, and it is not offered by all jobcentres.
Occupational rehabilitation is a further option for unemployed people with a disability or risk of disability in Germany. It aims to identify individual training and/ or work-place adjustment needs and to provide such measures. Nivorozhkin (Reference Nivorozhkin2019) finds positive effects of being accepted for occupational rehabilitation on unemployed rehabilitation applicants’ employment outcomes. Reims and Tisch (Reference Reims and Tisch2022) find that rehabilitation status enhances the effectiveness of vocational training for unemployed people with health impairments.
However, Rauch and Dornette (Reference Rauch and Dornette2010) show that despite the greater emphasis on activation brought about by the labour market reforms of the early 2000s, obstacles to employment integration have actually increased for people with disabilities. In particular, the number of participants in rehabilitation programmes authorised by the Federal Employment Agency declined between 2002 and 2008. This is attributed to frictions between increased demands for efficiency in activation policy on the one hand and the time- and cost-intensive processes of occupational rehabilitation on the other hand. Caseworkers’ high caseloads and insufficient time for identifying rehabilitation needs, as well as the elevated complexity of coordination between various institutions, are further reasons named for declining rehabilitation referrals (Rauch & Dornette, Reference Rauch and Dornette2010). Rehabilitation counselling, which accompanies rehabilitation programmes, is one specialised form of counselling in jobcentres. In the present article, descriptive statistics for rehabilitation counselling are shown. However, case numbers are too small to evaluate effects of rehabilitation counselling on employment entries. This article contributes findings on employment effects of overall counselling frequency for benefit recipients with health impairments. The subsequent section reviews previous findings on counselling effects for unemployed people in general.
Counselling effects on employment outcomes
Several international studies give insight as to how caseworker counselling affects the employment prospects of people who are unemployed. For Switzerland, Schiprowski (Reference Schiprowski2020) finds significantly positive effects of caseworker meetings on exiting unemployment, exploiting data on unplanned caseworker absences. Van den Berg et al. (Reference van den Berg, Kjærsgaard and Rosholm2012) similarly find positive effects of caseworker meetings on transitions from unemployment to employment for Denmark. Cheung et al. (Reference Cheung, Egebark, Forslund, Laun, Rödin and Vikström2019) show that more frequent caseworker meetings decreased unemployment among the treated, on the basis of an experiment in Sweden. For Germany, Boockmann et al. (Reference Boockmann, Osiander and Stops2014) find positive, but only partially significant, effects of more frequent caseworker meetings on exiting unemployment. Several studies also demonstrate that a lower client-to-caseworker ratio positively affects employment outcomes, e.g. Hainmueller et al. (Reference Hainmueller, Hofmann, Krug and Wolf2016) for Germany, Böheim et al. (Reference Böheim, Eppel and Mahringer2022) for Austria, and Ravn and Nielsen (Reference Ravn and Nielsen2019) for Denmark. In contrast, a study for a group of young benefit recipients with low formal qualifications in Vienna shows that more intensive counselling involving more referrals to training instead of quick job placement had no significant employment effects within a three-year follow-up period (Eppel & Mahringer, Reference Eppel and Mahringer2022). Also, McGuinness et al. (Reference McGuinness, O’Connell and Kelly2019) find for Ireland that, in the context of an absence of sanctions and monitoring, employment office interviews had negative effects on employment outcomes.
Data and method
The analyses are based on data from the panel study ‘Labour Market and Social Security’ (PASS) (Altschul et al., Reference Altschul, Bähr, Beste, Collischon, Coban, Dummert, Frodermann, Gleiser, Gundert, Küfner, Mackeben, Malich, Müller, Schwarz, Stegmaier, Teichler, Trappmann, Unger, Wenzig, Berg, Cramer, Dickmann, Gilberg, Jesske and Kleudgen2023) Footnote 1 , linked with administrative data. The PASS is a yearly population survey that is representative of the resident population in Germany, and which oversamples basic income support recipients in order to provide sufficient sample sizes for differentiated analyses. To date, the PASS has run for fifteen waves (Berg et al., Reference Berg, Cramer, Dickmann, Gilberg, Jesske, Kleudgen, Beste, Dummert, Frodermann, Schwarz, Wenzig, Trappmann, Altschul, Bähr, Collischon, Coban, Gleiser, Gundert, Küfner, Mackeben, Malich, Müller, Stegmaier, Teichler and Unger2022). PASS items used for the present analyses include information on jobcentre counselling, as well as respondent characteristics such as health, age, level of education, partnership status, number and age of children, and place of birth. A general indicator of severe health impairments was surveyed in all fifteen waves. Further detailed health items were included or altered from wave 9 onwards. Thus, descriptive analyses for most health indicators are shown for waves 9–15.
Analyses on employment entries make use of data linkages to administrative data that were provided up to wave 11, for respondents who consented to the data linkage. The administrative data used are the Integrated Employment Biographies (IEB) and Basic Income Support Histories (LHG), which have been made available for scientific analysis and are based on Federal Employment Agency data originating from employment offices and notifications sent by employers to health and pension insurance funds Footnote 2 .
For the analyses, spells in which respondents received basic income support and were not employed are prepared on the basis of the administrative data. These spells are merged to the PASS survey data. Spells are left-censored at the time of interview of a given wave if a person had no interview in the previous wave. Spells are right-censored one year after an interview if the person had no interview in the subsequent wave. Spells are split at each interview date. PASS characteristics are set at their values as surveyed at a given wave up to the subsequent wave or up to the time point of right-censoring, respectively. People who were in education or training, self-employed, or were retired or receiving incapacity pensions at a given interview were excluded from the sample.
Hazard models estimate entry rates into socially insured jobs and uninsured minijobs as competing risks. Minijobs are jobs with earnings of no more than 520 euros (until 2022: 450 euros, until 2012: 400 euros) a month or that are limited to no more than three months (until 2014: two months) duration within a calendar year. People employed in minijobs do not have health, unemployment, or nursing care insurance via their employment and can opt out of pension insurance payments.
The true starting points of spells of non-employment and benefit receipt are known from the administrative data, and are used to correctly specify models using the left-censored spells. The models control for individual-level unobserved heterogeneity, across multiple spells of individuals. A first set of hazard models estimates effects of various detailed health indicators on employment entries. These models use data from waves 9–11, for which detailed health information as well as linked administrative data was available. Hazard models for effects of jobcentre counselling frequency on employment entries make use of PASS waves 4–6 and 8–11. Analyses are restricted to these waves because jobcentre counselling information from waves 1–3 is not directly comparable with information from wave 4 onward when filters and reference time frames were altered. In wave 7, questions on jobcentre counselling were omitted from the questionnaire. Jobcentre counselling frequency is operationalised as the number of counselling sessions named by respondents for the reference time frame – the time since the last interview, within the last year or since people started collecting basic income support, depending on whether people had been interviewed before. On the basis of this information, counselling sessions per year were calculated. These were categorised as no counselling, one to four counselling sessions per year, and five or more counselling sessions per year. Preliminary analyses showed little effect differences on employment entries within the range of one to four counselling sessions, but an effect of having five or more rather than fewer counselling sessions.
Models for effects of jobcentre counselling are restricted to people who self-identify as active jobseekers, as only they were given the questions on counselling frequency. An ordered logit model is estimated to control for selectivity into jobcentre counselling frequency. The joint models for counselling frequency and entries into insured jobs and uninsured minijobs control for the correlations of unobserved heterogeneity between all three processes. In this manner, effects of jobcentre counselling on job entry hazards can be obtained net of selectivity of counselling frequency. Models for effects of jobcentre counselling on employment entries are estimated separately for people with and without severe health impairments. A basic indicator for severe health impairments is available for all waves, such that these models were estimated for waves 4–6 and 8–11 for which both jobcentre counselling as well as administrative employment spell data was available. Severe health impairments are operationalised as either a legally recognised disability, having applied for recognition of a disability, or having a different severe health impairment.
For the unobserved heterogeneity distribution, a normal distribution with three dimensions was assumed, estimated via numerical integration with twelve support points per dimension. In the sample with health impairments, correlation of unobserved heterogeneity between minijob and insured employment entries was very high and negative. Convergence was achieved by assuming a common heterogeneity distribution for the two processes, allowing for interaction with a parameter.
Finally, additional analyses look into effects of case management versus regular job-placement counselling on employment entries for people with and without severe health impairments. Information on counselling type is drawn from administrative data. These models likewise control for selectivity in take-up of case management, by controlling for the correlation of unobserved heterogeneity in the competing risk employment model and the logit model for case management take-up. Information on type of counselling was not available for jobcentres run by municipalities alone, without cooperation with the Federal Employment Agency, leading to an exclusion of 20 per cent of the original sample for these models.
Findings
Prevalence and type of health impairments among non-employed benefit recipients in Germany
Figure 1 shows that the proportion of non-employed basic income support recipients with severe health impairments (a legally recognised disability, application for recognition of a disability, or other severe health impairment) increased over time. In 2021, 51 per cent of non-employed benefit recipients reported severe health impairments. A likely explanation for this development is that, due to economic upturns in recent years, many jobseekers without health impairments were able to find employment, leaving those with greater job-finding difficulties behind. General population aging further contributes to an increasing average age of benefit recipients, and in turn, to higher proportions with health impairments.
Proportion of non-employed benefit recipients reporting severe health impairments.
Data source: PASS, waves 1-15. N=32,826 observations, weighted.
Notes: Non-employed basic income support recipients aged 15-64. Excluding students and people receiving retirement or disability pensions.
The most frequent types of severe health impairments reported by basic income support recipients are musculoskeletal impairments, followed by mental disorders, and internal and cardiovascular impairments (Figure 2). These types of health impairments are likely to restrict the types of jobs people can take up, thus potentially leading to longer job search durations.
Percentage of non-employed basic income support recipients with given type of severe health impairment. Multiple responses allowed.
Data source: PASS, waves 9-15 (2015-2021), N=12,603 observations, weighted.
Notes: Non-employed basic income support recipients aged 15-64. Excluding students, people receiving retirement or disability pensions and missings.
Table 1 gives information on legal recognition and severity of disability. 19.7 per cent of non-employed benefit recipients reported a legally recognised disability, and a further 4.7 per cent had applied for recognition of their disability. Among those with a legally recognised disability, the distribution of the recognised severity of the disability on a scale with the values 20, 30, 40, 50, 60, 70, 80, 90 and 100 is shown. When people apply for legal recognition of their disability, its degree is assessed on the basis of medical records, following legally prescribed guidelines. Depending on type and degree of their legally recognised disability, people are eligible for various types of benefits, such as tax deductions, work, and transportation benefits. Of the sample members with a recognised disability, 3.3 per cent gave a severity value that was between the points on the scale. These were recoded to the next lower scale value. Over half of those with a recognised disability had a severity value of 50 or above. Overall, 43.2 per cent of non-employed benefit recipients reported a severe health impairment that was not a legally recognised disability. This could either be in addition to or other than a recognised disability.
Legally recognised disability and degree of disability
Data source: PASS, waves 9–15 (2015–2021), N = 12,728 observations, weighted.
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions and missings.
Respondents were also asked to assess their capability of employment. Table 2 shows that 36.9 per cent of non-employed benefit recipients reported they were fully capable of employment without any restrictions for health reasons. The remaining 63.1 per cent reported some form of employment restrictions due to health reasons: 11 per cent reported restrictions concerning tasks, 7.7 per cent concerning work hours, 19.3 per cent concerning both tasks and hours, and 25.1 per cent reported that they were not capable of employment at all. Thus, the majority of non-employed benefit recipients report some form of health restrictions on their employment capabilities.
Self-assessed health restrictions on capability of employment
Data source: PASS, waves 9–15 (2015–2021), N = 12,423 observations, weighted.
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions and missings.
To determine how health impairments affect actual employment outcomes, competing risks hazard models were estimated (Table 3). The estimates show effects of health impairments on entering insured jobs and uninsured minijobs. The first model is for effects of different types of health impairments. Findings for entries into insured employment are that visual and hearing impairments as well as mental disorders have the strongest negative effects on employment entry rates, followed by musculoskeletal and internal diseases. Entries into uninsured minijobs seem much less affected by health impairments, with hardly any significant effects. An explanation may be that competition for minijobs is lower, making them accessible for people with lower productivity. It is interesting that the types of health impairments representing the greatest obstacles for regular employment (visual and hearing, mental, and musculoskeletal impairments) do not hinder minijob entry.
Competing risks hazard models for effects of different health impairment measures on employment entry rates. Log relative risks
The second model shown in Table 3 looks into effects of severity of health impairments. Entry rates into socially insured employment are shown to be significantly lower for people with legally recognised disabilities and other health impairments than for people without health impairments. The degree of severity of the legally recognised disability does not seem to matter much for employment entry rates. Effect sizes are even slightly larger for people who have applied for recognition of their disability or have a different severe health impairment that has not been legally recognised as a disability. Again, entry rates into minijobs are not significantly affected.
The third model shows effects of self-assessed work restrictions on subsequent employment entries (Table 3). While restrictions on types of tasks significantly lower entry rates both into socially insured jobs and minijobs, restrictions on hours or both tasks and hours lower entry rates into socially insured jobs only. This makes sense, as minijobs are necessarily short part-time jobs, as implied by minimum wage regulations. People who report that they cannot work at all have much lower employment entry rates, especially into insured employment. Altogether, people’s self-rated work capabilities appear to be quite good predictors of their actual subsequent employment outcomes.
Jobcentre relationship
As described in the previous section then, health impairments are very common among benefit recipients, and significantly affect their employment opportunities. The jobcentres can exempt benefit recipients from job search requirements if they are temporarily unavailable for work on the basis of health reasons, childcare, or elderly care, for instance. Moreover, even for people who are required to search for a job, requirements have been found to be more lenient when caseworkers perceive benefit recipients to have less control over their situation, for instance due to health restrictions (Senghaas, Reference Senghaas2021).
Table 4 shows that overall, 40 per cent of non-employed benefit recipients report that the jobcentre requires of them to search for a job. Differentiating by self-assessed health restrictions, those without health restrictions and those with restrictions concerning tasks are most frequently (52.7 per cent and 57.1 per cent) required to search for a job. Those who report that they are not capable of employment at all are more seldom required to search for a job (15.7 per cent).
Job search requirement by health restrictions on capability of employment
Data source: PASS, waves 9–15 (2015–2021), weighted.
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions and missings.
People who are not required to search for a job can be voluntary jobseekers. In fact, this is more common among people who report health restrictions on their employment capabilities (22.2–24.2 per cent) than among those who do not (18.8 per cent) (Table 4). Thus, an important proportion of benefit recipients with health impairments views themselves as jobseekers, in opposition to the jobcentre’s appraisal. Among those who are not capable of employment at all, in contrast, voluntary job search is the exception (8.3 per cent), as would be expected.
The sample in Table 5 is restricted to (mandatory and voluntary) jobseekers. Both mandatory and voluntary jobseekers can receive jobcentre counselling, addressing benefit recipients’ employment options and/ or their personal situation. Table 5 shows the frequency of such jobcentre counselling sessions. Overall, 28 per cent of jobseekers reported not having received any such counselling in the reference period (since last interview wave or during the last year for first-time interviewees). Half of the jobseekers received one to four counselling sessions per year. A further 22 per cent report five or more counselling sessions a year. Among jobseekers, differences in counselling intensity between those with and without health impairments are quite small (Table 5).
Jobcentre counselling frequency by health impairment. Jobseekers only
Data source: PASS, waves 4–6, 8–11 (2010–2012, 2014–2017), weighted.
To study effects of jobcentre counselling on employment outcomes, competing risks hazard models for entries into insured employment and uninsured minijobs were estimated. An ordered probit model for selection into a counselling-intensity category (no counselling, one to four counsellings per year, five or more counsellings per year) was estimated as well. The standard deviations of the unobserved heterogeneity distributions for each of the three processes as well as their correlations are estimated to control for selectivity in counselling intensity. The complete model estimates are shown in Tables A4 and A5.
Table 6 shows the effects of jobcentre counselling on employment entry from these models. For people without severe health impairments, five or more counselling sessions significantly raise entry rates into insured employment. Counselling has no significant effects on entries into minijobs. For people with health impairments, no significant employment effects of counselling are found.
Competing risks hazard models for effects of jobcentre counselling frequency on employment entries. Controlling for selectivity of jobcentre counselling frequency. Jobseekers only. Log relative risks
* p < 0.1.
** p < 0.05.
*** p < 0.01.
Data source: PASS, waves 4–6, 8–11 (2010–2012, 2014–2017) linked to administrative data (IEB, LHG).
Non-employed basic income support recipients aged 15–64. Excluding students and people receiving retirement or disability pensions.
See Tables A4 and A5 for sample sizes, control variables, and complete jobcentre counselling selectivity model.
A possible explanation for the lack of employment effects of jobcentre counselling for people with health impairments may be that they receive less offers or support during these counselling sessions. Table 7 shows that proportions of jobseekers reporting training programme offers are indeed significantly lower among those with than without severe health impairments. Proportions reporting job offers are likewise somewhat lower among those with health impairments, but this difference is not significant. Thus, these findings partially support those by Brussig and Knuth (Reference Brussig, Knuth, Lindsay and Houston2013), who also report less jobcentre offers for benefit recipients with health impairments.
Jobcentre offers by health impairment. Jobseekers only
* p < 0.1.
** p < 0.05.
*** p < 0.01.
Data source: PASS, waves 9–11 (2015–2017), N = 2,793 observations, weighted.
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions, missings, and cases with only one psu per strata.
A further explanation from qualitative research (Kupka et al., Reference Kupka, Oschmiansky and Popp2017) is that job referrals provided by jobcentres often do not fit the needs of people with (mental) health impairments. Possibly, this is a result of high caseloads and pressure to meet performance quotas within standard job-placement counselling. Kupka et al. (Reference Kupka, Oschmiansky and Popp2017) report that benefit recipients with mental health impairments find case management, which covers further areas such as health or housing and is not restricted to job placement, more helpful. Thus, further analyses look into the frequency and effects of case management for benefit recipients with and without severe health impairments.
Table 8 shows percentages of jobseekers with and without health restrictions receiving different types of counselling. Of jobseekers without severe health impairments, 20.8 per cent receive case management. For jobseekers with severe health restrictions, this figure is somewhat higher, at 24.1 per cent. Of those with health restrictions, 9.9 per cent receive rehabilitation counselling, and 2 per cent disability counselling. It is possible for benefit recipients to be allocated to several different counselling types at the same time. Among people with health impairments, the percentage receiving rehabilitation counselling is the same (10 per cent) among those additionally receiving and not receiving case management. Thus, there should be no differences in proportions of people in rehabilitation between the case management and comparison group that might affect the estimates of the employment effects of case management.
Type of jobcentre counselling by health impairment. Multiple parallel types of counselling possible. Jobseekers only
* p < 0.1.
** p < 0.05.
*** p < 0.01.
Data source: PASS, waves 4–6, 8–11 (2010–2012, 2014–2017) linked to administrative data (DWH), N = 6,465 observations, weighted.
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions, missings, and jobcentres run by municipalities alone.
Table 9 shows effects of case management on entries into insured and uninsured jobs within the one-year period following each survey wave. The models control for selectivity in case management take-up. No significant effects of case management versus receiving other types of counselling only on employment entries are found, neither for people with nor without health impairments. Future research could study effects of case management using a longer observation period. The present analyses exploit the panel nature of the data as best possible. Yet, the drawback is a relatively short observation period.
Competing risks hazard models for effect of case management on employment entries, controlling for selectivity of case management. Jobseekers only. Log relative risks
* p < 0.1.
** p < 0.05.
*** p < 0.01
Data source: PASS, waves 4–6, 8–11 (2010–2012, 2014–2017) linked to administrative data (IEB, LHG, DWH).
Non-employed basic income support recipients aged 15–64. Excluding students, people receiving retirement or disability pensions and jobcentres run by municipalities alone.
See Tables A6 and A7 for sample sizes, control variables, and complete case management selectivity model.
Conclusion
To improve societal participation of people with health impairments or disabilities, social policy in Germany follows EU strategies aiming to increase their employment rates (Annesley, Reference Annesley2007; Rauch & Dornette, Reference Rauch and Dornette2010). Access to incapacity pensions in Germany is quite restrictive in international comparison (Konle-Seidl et al., Reference Konle-Seidl, Rhein and Trübswetter2014). Instead, high proportions of people with health restrictions receive basic income support for jobseekers. Analyses in this paper show that of all non-employed recipients of basic income support for jobseekers, 63 per cent report some form of restriction of their employment capabilities for health reasons, 51 per cent report severe health impairments, and 25 per cent that they cannot work at all.
Policies for recipients of basic income support for jobseekers have focused on activation and work-first strategies (Eichhorst et al., Reference Eichhorst, Grienberger-Zingerle and Konle-Seidl2010). Whether employment integration can, however, succeed in advancing societal participation as intended depends on the quality of the jobs taken up, for instance on whether they provide an adequate income and social insurance benefits (Hammarström et al., Reference Hammarström, Virtanen and Janlert2011; Brucker & Henly, Reference Brucker and Henly2019).
This paper contributes findings on effects of health impairments not only on entering regular insured employment, but also on minijob entries, as an important measure of employment quality. Results from competing-risk hazard models show that health impairments significantly decrease benefit recipients’ entry rates into socially insured jobs, but not into minijobs. Minijobs are jobs that pay no more than 520 euros a month and make no contributions to health or unemployment and only electively to pension insurance funds, thus limiting employees’ opportunities for societal participation in these areas. Due to minimum wage regulations, minijobs are necessarily jobs with very short working hours (currently amounting to less than two hours a day for a five-day workweek).
Findings indicate that jobcentres use their discretion to reduce job search requirements for people with health restrictions. This corresponds to Geiger’s (Reference Geiger2017) classification of implementation of benefit conditionality for disabled people in Germany as passive in international comparison, based on low conditionality and a weak link to rehabilitation measures. Further findings in this paper show that among jobseekers, people with health impairments obtain a similar amount of counselling as those without. However, while timing-of-events hazard models show positive effects of higher-frequency counselling for benefit recipients without health impairments, no such effects are found for those with health impairments over a one-year follow-up period after each panel wave. This indicates the need for a longer-term perspective for benefit recipients with health impairments.
Thus, altogether, the analyses in this paper have shown that health impairments are widespread among recipients of basic income support for jobseekers, that health impairments reduce entry rates into socially insured jobs but much less affect entry rates into uninsured minijobs, and that jobcentre counselling frequency has short-term employment effects only for those without health impairments. Access to better quality employment is needed to enhance societal participation opportunities for benefit recipients with health impairments. To this end, it would be important to more explicitly acknowledge that the system of basic income support for jobseekers in Germany is in fact to a large extent a benefit system for people with health impairments.
One approach could be to raise caseworkers’ awareness for the potentials and ambitions of benefit recipients with health impairments, in order to generate suitable job offers (Kupka et al., Reference Kupka, Oschmiansky and Popp2017). Reasons that benefit recipients with health impairments take up uninsured minijobs, apart from lower working hours, may include less competition for these jobs. Thus, more use of subsidies for employers who employ benefit recipients with health impairments in socially insured jobs might be an option. Further policy responses could include more investments into rehabilitation and retraining, which have been shown to improve opportunities for socially insured employment (Nivorozhkin, Reference Nivorozhkin2019; Reims & Tisch, Reference Reims and Tisch2022).
As mentioned earlier, in international comparison, access to incapacity pensions is quite restrictive in Germany, while the proportion of people with health impairments among recipients of basic income support for jobseekers is quite high (Konle-Seidl et al., Reference Konle-Seidl, Rhein and Trübswetter2014). In contrast, many other countries have faced high levels of disability benefit claims, with policy reforms aiming to limit access (Alanko & Outinen, Reference Alanko and Outinen2016; Lindsay & Houston, Reference Lindsay, Houston, Lindsay and Houston2013b; Ulmestig, Reference Ulmestig, Lindsay and Houston2013; van Berkel, Reference van Berkel, Lindsay and Houston2013). Such reforms can lead to a situation such as that in Germany, with high rates of people with health impairments in other benefit systems, such as those for jobseekers (Wright et al., Reference Wright, Robertson and Stewart2022). Policy implications based on the findings in this paper, which call for adaptions to the situation of jobseekers with health impairments, particularly to enable better quality employment, may thus be relevant for other countries as well.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0047279424000059
Acknowledgements
I would like to thank Monika Senghaas, Markus Wolf, Joachim Wolff, the two anonymous reviewers, as well as the participants of the ESPANET 2022 Conference, the 2022 Salzburg Conference in Interdisciplinary Poverty Research, the SLLS 2023 Conference, and the IAB seminar on labour market policy for very helpful comments and advice.
Competing interests
The author declares none.
Open access
