Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a comparative analysis based on PASS survey data

Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a... This paper comparatively analyzes strategies of German Jobcenters to bring native and immigrant job seekers into employment. It focuses on clients who receive means‑tested basic income for the unemployed, based on data from the Panel Study Labour Market and Social Security (PASS) from year 2015 to 2020. By way of logistic regression, the study identifies the impact of being an immigrant on the clients’ probability of reporting different kinds of offers like job referrals or courses, controlling for a number of other influential factors. The study also looks deeper into the effects of immigrant ‑specific attributes, such as heterogeneous German language skills. We found that the likelihood of offers by Jobcenters largely depends on the amount of time since immigration. Recent immigrants have the lowest chance of reporting most of the studied measures of active labor market policies. For immigrants having stayed more than 4 years in Germany, however, we do not find a disadvantage, and some measures out of Jobcenters’ toolbox are even more often offered to the longer ‑settled immigrants than to native clients. A possible explanation for the mod‑ erately under‑average support of recent immigrants in terms of Jobcenters’ measures could be an institutional focus on improving German language skills prior to approaching the labor market. Keywords: Immigration, Integration, Jobcenter, Public employment services, Active labor market policy, ALMP, Training, Labor market, Welfare state JEL Classification: H41, I38, J08, J15, J68, J61 1 Introduction: Jobcenters as crucial actors employment is one of the most important steps, maybe for the integration of immigrants? the most important one. Yet, immigrants face a higher Germany has witnessed successive waves of immigra- risk of becoming and staying unemployed than natives tion since the World War II, in recent years especially (Integrationsbeauftragte 2019, p. 203; Kogan 2005). Fur- from Eastern and Southern Europe in the context of the thermore, with the exception of recruited immigrants enlargement of the European Union (EU) as well as from who enter Germany with an employment contract in Syria, Iraq, Afghanistan, and most recently Ukraine, due hand, most immigrants do not have a job upon arrival to wars and humanitarian crises. For successful integra- (Söhn 2019, p. 51). They are significantly overrepre - tion of immigrants in German society, finding (decent) sented among the unemployed in Germany, with con- sequences such as a high poverty risk (Bundesregierung 2021, pp. 50, 129, 210). Despite Germany’s booming *Correspondence: rene.lehwess@sofi.uni‑ goettingen.de economy, their relative share among job-seeking recipi- ents of basic-income support for the unemployed (ALG- Soziologisches Forschungsinstitut Göttingen (SOFI) e.V., an der Georg‑August ‑ Universität, Friedländer Weg 31, 37085 Göttingen, Germany II [Arbeitslosengeld II]) increased between 2015 and 2020 © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. 9 Page 2 of 24 R. Lehwess‑Litzmann , J. Söhn (BA 2020 and older versions), i.e., our episode of inves- and 2019, the average across this period being 28.4 per- tigation. Hence, public employment services (PES) is an cent; thus only one percentage point below the average institution of paramount importance for many immi- share of immigrants in all job-seeking benefit recipients grants’ labor market integration. (29.3 percent) (BA 2019c, authors’ calculation). As to The PES do much more than granting unemployment other activation measures, the share of foreigners var- benefits; they also monitor job offers and deploy a wide ies little (BA 2019a, p. Tab 1.1). Older research based on range of active labor market policy (ALMP) measures. quantitative micro data is inconclusive regarding immi- For example, they can offer subsidized courses, which grants’ access to PES-subsidized further training, show- may, for financial reasons, be the only chance to access ing a negative effect of foreign or non-EU citizenship, further education for some groups of clients, e.g., long- but a positive one of a short stay in Germany. A recent term unemployed individuals receiving means-tested vignette survey among caseworkers in German Jobcent- benefits. Arguably, the PES are even more important ers, focused on refugee clients, showed that whether for job-seeking immigrants than they are for native job caseworkers recommended employment, occupational seekers. This is due to immigrants’ lack of country-spe - training or other programs as the first option depends on cific knowledge, networks, vocational certificates, and of the socio-economic background and family situation of other resources. fictitious clients (Dietz and Osiander 2019). Boockmann When individuals register for unemployment benefits, and Scheu (2019, pp. 408, 412) find that Jobcenter staff the main objective of both social law and PES is to help consider an improvement of language skills and qualifica - them become financially independent again by bringing tion of refugees as the prerequisite for finding qualified them back into work, and this very often corresponds to work, and thus for sustainably ending benefit receipt. the clients’ aim as well. Sometimes, however, the client, However, such a long-term perspective is not system- the PES, or both, regard immediate re-employment as atically implemented, as both clients and case managers unrealistic or even undesirable under the overall circum- see also downsides of postponing contact with the labor stances. Instead, some clients may require more knowl- market by lengthy “chains of subsidized measures” (ibid.). edge of how to apply for jobs and present themselves Many refugee clients urgently wish to start working, even in job interviews. Others want to or have to profoundly in ‘bad’ jobs, and some case managers warn that receiv- re-orientate their professional goals; this may be due to ing social benefits for several years might make clients health issues or long episodes of unpaid care work in lose sight of the aim of employment (ibid., pp. 412–413). the family. Still, others consider attaining new qualifica - Boockmann and Scheu (2019, pp. 412–415) recommend tion as a necessary intermediary step to reach the final Jobcenters to combine education and training on the one goal of labor market integration or to gain employment hand with labor-market programs on the other (ibid., pp. on a higher level of qualification and income than pre - 413 et seq.). According to qualitative findings by Schnei - vious competences and qualification would allow. In all der et  al. (2008, p. 30), discriminating treatment by the those instances, the PES can offer help using their ALMP PES seems to be a rare case, while Sauer (2010) reports toolbox. such experience for some unemployed immigrants she This article investigates whether, and to what extent, interviewed, but again without comparison with native PES in Germany treat immigrant and native clients dif- clients. Experimental correspondence tests, however, ferently with regard to efforts of placing them in regular do show a lower quality of German municipalities’ and employment and channeling them into various kinds of Jobcenters’ responses to email requests of ‘Turks’ versus ALMP programs. Are there different patterns of what ‘Germans’ (Grohs et  al. 2016; Hemker and Rink 2017). PES caseworkers offer to immigrated job seekers com - Aggregate administrative statistics are descriptive only pared to German-born ones? and hardly differentiate participants according to social The percentage of non-German citizens among all or immigrant-specific characteristics (usually only gen - participants of PES-subsidized further vocational train- der, age, non-German citizenship or sometimes refugee ing increased from 19.7 to 34.1 percent between 2015 status). Finally, there are good reasons for investigating the foreign-born in a broad sense; while recent refugees have gained quite some attention in labor market research Statistics of the Federal Employment Agency (BA [Bundesagentur für Arbeit]) show that in the period from 2015 to 2019, 40.0 percent of new par- (Bonin et  al. 2020; Boockmann and Scheu 2019; Dietz ticipants in ALMP attended measures for activation and professional integra- et  al. 2018; Fendel 2019; Kasrin et  al. 2021; Kosyakova tion (854,990 thousand individuals on the yearly average), among which, e.g., 2020) in the context of the large influx of Syrian refu - 5.8 percent participated in further vocational training and 10.3 percent in job- creation schemes. 34.1 percent received classic referrals to job openings (BA gees in 2015/2017. In the long run, a comprehensive look 2021, restricted to Sozialgesetzbuch [SGB] II, authors’ calculation). on immigrants in general is called for. Quantitatively Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 3 of 24 9 important groups like EU immigrants from Southern immigrant-specific variables. Critically reflecting on and Eastern Europe have and will continue to make up a our database, we will discuss methodological challenges substantial share of immigrants. Also, naturalized immi- of the PASS-items on “vocational training and other grants disappear in administrate statistics but deserve courses” and present robustness tests (Sect.  4.3). The being included, for instance, as an immigrant group con- quantitative analysis is complemented by some insights trasting with recently arrived ones. Furthermore, as our from selected expert interviews with caseworkers and interest lies on ALMP opportunities given by Jobcent- middle-management employees of Jobcenters, which ers to their clients, rather than actual usage thereof (an were carried out in the context of a larger research pro- outcome in its own right), we need to investigate offers ject. The article finishes with a summary of the main received. Even though an offer does not equal its accept - findings and further discussions. ance, receiving an offer is still the sine qua non for being allowed to participate in a program subsidized by the 2 Conceptual considerations and previous PES. Overall, to our knowledge, there are no representa- empirical insights tive and multivariate studies on immigrant-native dif- German law distinguishes jobless individuals who obtain ferences or intragroup comparisons among immigrants unemployment insurance benefits based on prior contri - regarding Jobcenters’ strategies for individual job-seeking butions (Social Code [SGB] III) from jobless persons who clients. receive means-tested basic-income support (SGB II). Concretely, our analysis pursues the following research The latter group includes both (long-term) unemployed questions: Which ALMP measures do Jobcenter staff persons who have exhausted their entitlement to unem- offer to unemployed recipients of basic-income transfers, ployment insurance benefits and job seekers who did not depending on whether they are natives, recently arrived contribute to the insurance (long enough); among them and longer-settled immigrants? Does migration experi- recently arrived refugees who received political asylum ence play a decisive role, once other influential factors or another form of humanitarian legal status. In the (like age or family status) are considered? How are addi- German two-tiered system, the so-called Jobcenters are tional immigrant-specific factors like legal status, region the PES responsible for all job seekers without insurance of origin and German language skills linked to Jobcenters’ benefits, thus for all who receive tax-funded welfare. Our offers? empirical analysis is dedicated to Jobcenter clientele. Our contribution on German Jobcenters analyzes data The priority for Jobcenters is to help clients end or from the Panel Study Labour Market and Social Secu- reduce benefit receipt, as this is what social law requires rity (PASS), a survey administered by the Institute for (§ 1 SGB II). For most clients, this means that (non- Employment Research (IAB) of the German Federal subsidized) employment is the main goal, be it for the Employment Agency (BA). This data is nationally repre - first time or after a period of joblessness. Benefit recipi - sentative for basic-income recipients. Respondents regis- ents generally are obliged to look for potential employ- tered as job seekers during at least one of the waves, from ers; Jobcenters may also refer clients to open positions 2015 to 2020, answered the survey question whether the and encourage them to apply. Jobcenters also pursue PES referred them to specific job openings regarding the aim of “maintaining, improving, or re-establishing regular or marginal employment or offered them various the employability” of the client (ibid., authors’ transla- ALMP measures. tion). To this end, there is a range of ALMP measures Section  2 elaborates on the hypotheses based on con- which Jobcenters can implement (based on § 16 SGB II). ceptual considerations and previous research in our field. In this respect, and in contrast to direct access to non- Section  3 presents the database, operationalization, and subsidized jobs, caseworkers in Jobcenters act as crucial methods applied. After a descriptive analysis (Sect.  4.1), gate keepers. In accordance with the administration’s we will introduce Jobcenters’ offers as dependent vari - general goal of economical and effective spending, offer - ables in multivariate regression models (Sect.  4.2). In ing an ALMP program is a discretionary decision which a first step, the study analyzes natives and immigrants caseworkers as “street-level bureaucrats” (Lipsky 2010) together, in a second step, it looks at immigrants only, can opt for. They consider the degree of their clients’ as this permits extending the model by additional, Footnote on research project and funding institution as well as thanks to reviewers for their feedback. Citizens from EU-countries make up 69.4 percent of the foreign popula- tion in 2020, with 42.8, 50.9, and 43.7 percent among those with a duration of Immigrants with the precarious status of asylum seeker and with extra- stay of less than 1 year, between one and under 4 years, and between four and dition only temporarily suspended (Duldung) are the clientele of Arbe- under eight, respectively. The equivalent percentages for Syrians are 7.2, 6.3, itsagenturen (SGB III) (see Boockmann and Scheu 2019, p. 404) and not and 9.5 (see StaBa 2021, authors’ calculation). considered here. 9 Page 4 of 24 R. Lehwess‑Litzmann , J. Söhn employability, the likelihood of successfully finishing an may communicate their preferences to their caseworker educational program, with respect to the clients’ moti- and possibly influence the latter’s decision. On the one vation and learning ability, as well the probability of the hand, given the asymmetric power relation, an offer newly acquired knowledge helping to find an appropri - by the Jobcenter is a necessary condition of the clients’ ate job thereafter (Yankov 2010, pp. 8, 24–25). Having participation in an ALMP program. On the other hand, to respect their organizations’ budget restrictions, case- substantial measures like 2-year vocational re-training or workers try to identify those clients for whom a program professional coaching are too expensive to force clients seems most promising. Vocational (re-)training programs to attend. They tend to be a privilege that clients have to and long-term subsidized employment schemes are fight for and be knowledgeable and confident enough to among the most substantial and costly offers Jobcenters do so (regarding immigrants see Schneider et al. 2008, p. can make. Cheaper short-time programs do not involve 18). potential lock-in effects (standing in the way of imme - In some cases, clients’ rights limit caseworkers’ discre- diate transition into employment), which are typical for tion over ALMP measures. Participating in “integration longer programs (see the evaluation of ALMP programs’ courses”, regulated in the Residence Act (Aufenthaltsge- effects on refugee clients by Kasrin et al. 2021, p. 3). setz, §§ 43 ff.), is both a right and an obligation for certain Clients themselves also have to consider pros and cons groups of, primarily, newly arrived immigrants. These of looking for (non-subsidized) work versus attending courses mainly consist of German language classes plus (longer) programs while receiving the rather meager civic education. EU nationals may participate, too, but basic-income support. Course attendance and job search are neither obliged to nor have the right to do so. do not utterly exclude one another, but each restricts time Overall, the question whether Jobcenters make dif- and mental energy that remain for the other option. Indi- ferent offers to immigrant and native basic-income viduals more or less consciously weigh possible future recipients comes down to caseworkers’ perceptions occupational advantages of the qualification, knowledge, (Maynard-Moody and Musheno 2003) of these two or skills to be gained against their self-perceived learn- groups with regard to the necessity, needs, motivations, ing capacity and motivation, individual or family-related likelihood of success, and maybe also personal ‘worthi- obstacles or resources as well the opportunity costs of ness’, to be placed into regular employment or into some not having income from working (Becker 2019, pp. 3–5; kind of ALMP measure. In the following, we will develop Cross 1981, pp. 98, 116–117). An experimental study some hypotheses based on both theoretical arguments embedded in a survey of refugees showed that given the and related previous research. hypothetical scenario of having been offered a low paid job and vocational training, respondents were more likely 2.1 Referral to open job positions to choose the latter; the higher the future earnings after First, this study looks at the goal of direct employment the specific vocational training, the easier the exam was take-up. As for clients’ preferences, in terms of a disin- said to be, and the higher the apprenticeship wage during centive of looking for work, immigrants might consider training (Damelang and Kosyakova 2021, pp. 6–8). Again the high amount of German welfare benefits compared with regard to refugees, Boockmann and Scheu (2019, with salaries in their home country in absolute terms. p. 409) highlight the high motivation of many clients to However, living expenses in Germany is also higher than succeed in Germany, but also the potential demotivation before migration. Immigrants might actually be more by the lengthy process of overcoming various administra- eager than natives to work immediately, as many of them tive and qualification-related hurdles standing in the way feel obliged to send money to relatives back in their of job uptake. Based on interviews with Jobcenter staff, country of origin (on remittances see Sinning 2011). In they find that the wish for economic independence from addition, residence law makes their right to stay in Ger- Jobcenters can “reduce the readiness to go through quali- many beyond their fixed-term residence permit depend fication measures which would allow them enter higher on continuous pension insurance contributions (Knuth segments of the labor market in the long term” (ibid, p. 409, authors’ translation). Hence, offering and accepting ALMP programs are no isolated decisions; it rather takes place in a context where Also, short-term measures can be strategically used by the PES to verify clients’ readiness to co-operate, the absence of which can lead to sanctions, employment is the major competing alternative. Clients which case workers are more inclined to apply to ethnic-minority clients (Lin- den 2021). Non-EU family immigrants and asylum seekers recognized as refugees are For Switzerland, Liechti et  al. (2017, p. 264) show that foreigners profit no currently the largest groups. Altogether, 238 thousand immigrants attended less from ALMP programs than natives regarding their ensuing job prospects. integration courses on the yearly average between 2015 and 2019 (BAMF 2020, and previous editions; authors’ calculation). Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 5 of 24 9 2021, p. 56). In our expert interviews with Job Center with the labor market, Jobcenters might suggest applying staff, one employee put it this way: for marginal employment more often to immigrant clients (H #2). “iTh s happens quite often, when it comes to work - ing, the [migrant] clients say ‘I want work, no matter 2.2 Support for the job‑search process what kind, I will do anything’ … also the academi- Getting insight and support on how to write applications cally trained ones.” (Interview I) and how to present oneself in a job interview should be If Jobcenters shared this priority, as a theoreti- particularly relevant for newly arrived immigrants as cal assumption not tested by our data, they could refer there are most likely small or large cultural differences immigrants to open positions more often than natives (depending on the country of origin) in what is regarded (Hypothesis [H] #1a). In addition, as these newcomers are as an appropriate form. Hence, we expect unemployed less familiar with job search on the German labor mar- immigrants to be offered support with applications and ket than natives, caseworkers might not want to rely on job search more often than their native peers (H #3) (For immigrant clients’ own initiative to find employment, but our counter hypothesis regarding any type of ALMP pro- explicitly refer them to open positions. gram, see the end of this section). In contrast to this potential scenario, Jobcenters (and Looking for a job can involve monetary expenses, clients) could be aware of immigrants’ high hurdles to especially travel costs to go to a job interview. Jobcent- find employment and decide that these should be tack - ers routinely reimburse application or travel costs upon led before making efforts of placing clients into employ - job seekers’ request. Thus, such an offer by the Jobcenter ment. After all, many immigrants (and in particular hinges on the prior decision by an employer to put the recently-arrived ones) struggle with devaluation of their person on a shortlist and indicates a high level of employ- foreign qualification and/or work experience (Boock - ability. Given the obstacles immigrant job seekers face mann and Scheu 2019, p. 407; Damelang et al. 2020; Nohl on the labor market, we expect reimbursement of appli- et al. 2014, p. 24ff.), insufficient knowledge of formal and cation or travel costs to be offered less often to immigrant informal norms in the process of job search in Germany, than to native Jobcenter clients (H #4). actual or assumed German language deficits (Esser 2006, p. 400ff.), and ethnic discrimination in the German labor 2.3 PES‑subsidized (self‑) employment market (e.g., Veit and Thijsen 2019), resulting in both Some ALMP measures support employment more above-average unemployment rates, as discussed above, directly than the programs mentioned so far. They ‘cre - and below-average work quality (Gundert et  al. 2020). ate’ work experience and hence a potential bond with Caseworkers may refrain from making clients apply in an employer. In the case of a so-called “measure with the first place when success is improbable. If casework - an employer” (Maßnahme bei einem Arbeitgeber), job ers judged immigrants less employable than native job seekers work for an employer during up to 6  weeks (in seekers, they should refer immigrant clients to open job exceptional cases up to 12 weeks), who receives financial positions in regular employment less often than native (H subsidies by the PES during that period (Arbeitsagentur #1b). 2019) and ideally hires the job seeker once the program Jobcenters may also refer unemployed clients to mar- runs out. According to Kasrin et  al. (2021, p. 3), partici- ginal employment (“Mini-Jobs” in German), i.e., employ- pating companies prefer job seekers with a comparatively ment with low earnings (up 450 Euros per month), low high level of employability (Dietz et  al. 2018). Indeed, working hours, limited social insurance, and slim chances such programs bring unemployed (male) refugees most to lead to regular employment (Freier and Steiner 2008). quickly into regular non-subsidized employment com- Similar to ‘1-€-jobs’ (Arbeitsgelegenheiten) (Kasrin et  al. pared with other AMLP measures (Kasrin et al. 2021, pp. 2021, p. 3), Jobcenters might consider marginal employ- 5, with further training as the second-best option). With ment as an option mainly reserved for ‘difficult’ long- respect to the immigrant-specific need to gain employ - term unemployed clients. The latter are often faced with ers’ trust and also regarding recent immigrants’ eager- multiple employability obstacles so that Jobcenters regard ness to work, the latter might be offered a program with regular employment as out of reach. Limited personal an employer or an internship more often than native job availability for work due to child care, for instance, can seekers (H #5), provided the client is seen as sufficiently be an alternative reason for marginal employment. All qualified. these obstacles can apply to both native and immigrant Jobcenter clients. However, given the above-mentioned immigrant-specific obstacles to employment and also We do not know of any research on such cultural differences, but websites the particular need to bring refugee clients into contact like that of Just Landed (2020) imply the respective need for information. 9 Page 6 of 24 R. Lehwess‑Litzmann , J. Söhn Financial support to become self-employed could, on et  al. 2021, p. 7, regarding refugees). These arguments the one hand, be an option for immigrants as they have feed into our hypothesis that if immigrants’ special need greater difficulties to find employment (Integrations - for further vocational training and the effectiveness beauftragte 2019, p. 203; Kogan 2005); on the other hand, of such programs were decisive—again, a theoretical successful foundation of one’s own business requires assumption grounded in previous insights but not tested knowledge about administrative procedures and markets here—Jobcenters should offer it to immigrants more often that recent immigrants tend not to have. Jobcenters offer than to natives (H #7a). The opposite expectation (H #7b) self-employment support too rarely to statistically test a is favored by the time factor of attending a longer program respective hypothesis with our data. like a 2-year vocational re-training course, which might turn this option prohibitive for immigrants who already 2.4 C ounselling on occupational goals and qualification spent numerous months in German-language classes pathways and now urgently want to earn their living rather than Many immigrants are in need of (re-)orientation regard- continue learning with only the modest welfare benefits ing their occupational future when faced with devalu- available. Refugees’ case managers at times fear adverse ation of their previous qualification and/or work effects of too-long (chains of ) measures, possibly lead - experience (Damelang et  al. 2020; Nohl et  al. 2014), or ing immigrants to become used to welfare transfers as a when they have never worked beyond their household or durable substitute for income from employment (Boock- the informal sector in their country of origin (Bonin et al. mann and Scheu 2019, p. 412). 2020, p. 73, regarding refugees). Courses which inform The arguments put forth so far theoretically indicate a participants about different occupational fields in Ger - higher need of (recent) immigrants for ALMP programs many or individual coaching could, as a first step, support than there is among native job seekers, and such needs such individuals in deciding whether to apply for jobs, could turn into actual choices on part of Jobcenters. apprenticeships, internships, vocational re-training or However, there are two general reasons to expect fewer courses leading to full recognition of their foreign qualifi - offers to immigrants: Jobcenters may, rightfully or not, cation. If such needs were paramount for the Jobcenters’ consider immigrants’ linguistic ability too low to success- decision, they should offer courses on occupational orien - fully follow a mainstream ALMP program in German. tation or respective vouchers more often to immigrant cli- Furthermore, immigrants could be less knowledgeable, ents than to native ones. (H #6). confident, or vocal about their preferences in favor of a suitable ALMP program in their communicative interac- 2.5 Further training tion with caseworkers (Holzinger 2020). Schneider et  al. Some immigrants have special needs for further voca- (2008, p. 30) report “situations in which immigrants feel tional training. First, insufficient educational infrastruc - disadvantaged and sense that their qualifications, com - ture, war, or discrimination in the country of origin could petencies and career plans cannot be met” by the PES have prevented them from realizing their educational (Sauer 2010, p. 157). Relatedly, field experiments revealed goals there. Second, insufficient transferability of immi - some discrimination of German municipal govern- grants’ foreign qualification can serve as a push factor to ments (Grohs et  al. 2016) and Jobcenters (Hemker and gain new or complementary qualification in the immigra - Rink 2017) regarding the response quality in reacting tion country in order to prevent long-term social down- to email-requests from persons with typical Turkish (or ward mobility (for Canada, the USA, Germany, and the Rumanian) names. Against this backdrop, the counter- Netherlands repectively see Adamuti-Trache 2011, pp. hypothesis of fewer offers to immigrants could hold for 75–76; Hashmi Khan 1997, p. 287; van Tubergen and De some or all of the above-mentioned ALMP measures. Werfhorst 2007, p. 885). In addition, caseworkers could be aware of the fact and communicate it to their clients 2.6 Immigrant‑specific influences that attending educational programs is indeed an effec - Taking a deeper look at migration-specific factors, we tive, though strenuous and time-consuming, strategy to presume a shorter duration of stay in Germany to lead to a raise long-term chances of finding qualified work (Deeke larger difference in the treatment by Jobcenters compared and Baas 2013, regarding immigrants’ employment after with that of native clients (H #8). It remains to be seen further training financed by German PES; Kanas and van whether this assumption holds once the level of German Tubergen 2009; Lancee and Bol 2017). Jobcenter-subsi- language skills is considered. Good German skills would dized further training is even more successful in raising make language classes superuou fl s and should make both formerly unemployed persons’ incomes than subsidized referrals to (regular and subsidized) employment and to employer-programs, while both participant groups fare ALMP courses held in German more likely (H #9). As far far better than job seekers receiving no measure (Kasrin as legal status and nationality are concerned, Jobcenters’ Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 7 of 24 9 referrals to job vacancies (and reimbursement of travel calculate the duration of stay as well as to identify natu- cost) could be indirectly influenced by employers’ pos - ralized immigrants. sible recruiting preference for white, non-Muslim and/ We primarily use the first of the two samples the PASS or European immigrants (Koopmans et al. 2019; Veit and consists of, i.e., the one representing households which Thijsen 2019), as well as for those with German citizen- receive basic-income support (the other one is sampled ship (Steinhardt 2011) or permanent residence status (H on the German population without further restrictions) #10). (Hohmeyer and Wolff 2015, p. 13). Our investigation The potential effects of migration status, hypothesized is limited to the year 2015 to 2020, as the item which is in this section, could be mediated by other personal and most crucial for our analysis is only included in the sur- structural factors known to influence Jobcenters’ deci - vey as of wave 9. Our target population consists of adults, sions more generally. Such potential composition effects aged 18 to 64, who were registered at a Jobcenter among could be based on, e.g., immigrants being younger on those looking for work and receiving basic-income sup- average than native job seekers, with higher age being port. The sample was restricted to survey participants negatively associated with participation in further train- registered as job seekers, as only they are asked about ing (Osiander 2019, p. 70). Furthermore, higher levels of Jobcenters’ offers during the last year. Furthermore, we education are a component of positive (self-)selection excluded secondary-school students as well as recipients into employment and (further) education (Kruppe 2009, who were not in regular contact with the Jobcenter: Such pp. 11, 14; Osiander 2019, p. 76), also among refugee contact is a precondition of a potential ALMP offer. How - clients (Dietz and Osiander 2019, p. table  3). Depend- ever, this restrictions disproportionally excludes female ent children in one’s household are a possible hurdle as refugees (Bähr et al. 2017). well (Kruppe 2009, p. 14). Finally, there are good reasons In our unbalanced panel, a minority of participants of for investigating the foreign-born in a broad sense: while the PASS survey fulfilled our sample criteria for more recent refugees have gained quite some scientific atten - than one wave in our observation period. We choose tion (e.g., Bonin et al. 2020; Boockmann and Scheu 2019; person-waves (instead of persons) as our unit of obser- Dietz et al. 2018; Fendel 2019; Kasrin et al. 2021; Kosya- vation because each year that a person fulfills the crite - kova 2020) following the large influx of Syrian refugees ria of receiving basic income and looking for a job, there in 2015/2017, a comprehensive look on immigrants in is a new opportunity for offers to be made by Jobcenters general is called for. Quantitatively important groups to the client. We thus have a hierarchical data structure, like EU immigrants from Southern and Eastern Europe with 8275 person-waves nested in 4954 interviewees. have and will continue to make up a substantial share of 61.8 percent of the observed persons appear in our sam- immigrants. Also, naturalized immigrants disappear in ple only once (Table 5). administrate statistics but deserve being included, e.g., We define immigrants as foreign-born having arrived as an immigrant group contrasting with recently arrived in Germany as adults. According to this definition, ones. The following sections shows our empirical study is 45.3 percent of persons (and 39.1 percent of person- hence adequately inclusive. waves) in the observed population are immigrants. We further distinguish immigrants by their duration of stay, 3 Methods: database, operationalization, strategy i.e., whether they have arrived in Germany up to 4 years of analysis ago or at least 5 years ago. Note that immigrants tend to 3.1 Data and sample be observed for a smaller number of survey waves than The study uses the Panel Study Labour Market and Social natives: 69.8 percent of immigrants and only 55.2 percent Security (PASS) (Berg et  al. 2020) as it contains a ques- of natives are part of our sample during 1 year only. tion on having received an offer of various ALMP pro - grams by the PES (see below for details). In addition, the 3.2 Items for target variables PASS (rather than administrative data) is well suited for For operationalizing Jobcenters’ offers, we draw on the this research as it contains important information nec- survey item “Since your household has obtained Unem- essary to study immigrant integration, namely the year ployment Benefit II [since our last interview, respec - of immigration for any foreign born, which allows to tively], have you ever been offered the following by the Jobcenter?” (PTK1701). There is a battery of items to Citizens from EU-countries make up 69.4 percent of the foreign population in 2020, with 42.8, 50.9, and 43.7 percent among those with a duration of stay Immigrants having arrived as minors are left out because schooling abroad of less than 1 year, between 1 and under 4 years, and between four and under is a major contributor to migration-related problems on the labor market eight, respectively. The equivalent percentages for Syrians are 7.2, 6.3, and 9.5 (Kalter and Granato 2018, on differences between first- and second-genera - (see StaBa 2021, authors’ calculation). tion immigrants). 9 Page 8 of 24 R. Lehwess‑Litzmann , J. Söhn choose from, multiple answers possible. In our analysis, thirds of the vouchers actually redeemed do so (BA we consider the following offers (IAB’s translation of 2019a, table 5, authors’ calculation). On the PES web- the survey items into English) and we connect them with site, courses which the vouchers are meant for refer the hypotheses in Sect. 2: to, e.g. assistance in job applications, self-marketing strategies for academically trained job seekers, immi- • A part-time or full-time job or an apprenticeship: Job- grant-specific courses supporting economic integra - centers refer clients to open job positions in the form tion, one-on-one application coaching, or business of regular employment, subject to social insurance English. (H #6-plus) contributions, and encourage them to apply. Appren- • Vocational training, retraining or a course: This offer ticeships refer to the German dual vocational train- includes both courses lasting a few months and full ing in which learning takes place in a company and in vocational re-training taking 2 years. (H #7) a vocational school alternately. Apprenticeships are • An integration course or another German course: also subject to social security contributions. While This item should exclusively apply to immigrant Job - the previous section formulated H #1a/b only com- center clients. It has been included in the survey only prises regular employment, its arguments also apply as of wave 10 (year 2016), so the case numbers are to apprenticeships in firms as far as caseworkers may smaller in our models which include this item. Ger- view clients as (not) well-enough prepared to com- man courses other than those which are part of the pete with other applicants. integration course are, e.g., classes for occupation- • Minor employment, e.g., a mini-job (H #2) specific German targeting more advanced language • Support for your applications, e.g., assistance with learners. We study this target variable only with the preparation and compilation of application docu- regard to the intra-group comparison among immi- ments (H #3). This item partly overlaps with the one grant clients. on vouchers, see below, as it can be implemented by private providers rather than by the PES. Caseworkers may offer several measures (or nothing at • Reimbursement of application costs or travel all) during one counselling appointment or in consecu- expenses: though this is something Jobcenters can tive ones. The survey, however, asks respondents whether offer (H #4), it actually mirrors the “success” of hav - the Jobcenter has offered them any kind of offers since ing been invited to a job interview. they became a job seeker or since the last time they took • A program with an employer or internship (H #5) part in the yearly survey, respectively. During that time, • An activation or placement voucher with which you more than one measure of the same type could have can choose a program yourself: Financed by the PES, been offered and the sequence of those offers remains these vouchers give access to services by licensed unknown. Hence, we measure whether a type of program private local providers of various kinds of programs. was offered at least once during the reference period This item is not optimal for operationalizing our according to the respondents’ memory and willingness to hypotheses as it refers to the way Jobcenter offer answer correctly. programs, rather than their content. Support with Another source of potential bias is that the system of applications, occupational orientation, or further labor market services also has a “memory”: a specific vocational training, of which only two are response offer may be made only once (in consequence, persons categories of their own, can be granted via vouchers. observed for a smaller number of years could tend to The respective hypotheses (H #3, #6, #7) connected have a higher chance of receiving an offer in a given year to these measures also apply to vouchers. Jobcenter of observation). Yet, offers could also become more prob - statistics tell us that 56 percent of vouchers pertain able the longer the client is in the situation of receiving to the policy goal of giving clients orientation about benefits and looking for a job. As a test of robustness employment and vocational training (Heranführen (4.3), we will do an extra analysis where persons are con- an Ausbildung und Arbeitsmarkt) and more than two sidered only once across the whole observation period. Our findings on shorter- vs. longer-settled immigrants could be biased in so far as some of the immigrants who We do not test a hypothesis on the item of financial support to become self- arrived earlier could have left the country in the mean- employed, as the number of positive cases is too small for a multivariate anal- time, leading to some kind of social selectivity of the ysis. For the same reason, we refrain from using the item “other offers”, which remaining longer-settled immigrants. Yet, this methodo- only 1.9 percent say to have received. Hence, the other categories in the PASS survey are almost exhaustive. logical problem cannot be overcome with available data. It remains unknown whether respondents also subsumed ‘1-€-jobs’ Regarding one item, it is necessary to consider that under this item, as the item battery does not explicitly mention this meas- questions are asked in a certain order in the context of ure. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 9 of 24 9 the CATI/CAPI interviews (Jesske and Schulz 2018, p. and Möller 2008, p. 418). The under-employment quota 68), in which interviewers usually read out loud battery (individuals registered as unemployed plus those enrolled items one after the other. In case of items that overlap in PES programs; BA 2019b) of the respective regional to a certain degree, the ordering can be consequential. state controls for local employment opportunities. It is thus unfortunate that the item Vocational training, Some predictor variables are inherently related to each retraining or a course is proposed to respondents before other, e.g. duration of stay and respondents’ age (Pear- that of Integration course or another German course: it is son’s rho = 0.53): given our definition of immigrants as likely that some of the immigrants who had been offered having arrived as adults, those with longer duration of a language course in the reference year reported “yes” stay cannot be very young adults at the time of the inter- already when they read or heard of the item on vocational view. We use only broad categories of years since arrival (re-)training and the undefined “course” mentioned here. in order to avoid empty cells and hence attenuate the Indeed, out of the 818 immigrants who reported the offer challenge of collinearity. Duration of stay is also linked of Vocational training, retraining or a course in waves 10 to region of origin due to successive migration waves to 13, 648 also reported to have been offered an integra - from different regions (e.g., most persons from the for - tion/German course. There is no way to find out in ret - mer USSR arrived in Germany between 20 and 5  years rospect how many of those with two positive answers before the survey interview). Due to naturalization being have actually been offered both kinds of courses, and granted after several years of residence and due to more how many were only offered an integration or language generous rules applying to German resettlers, regions of course. However, chapter  4.3 presents a robustness test origin are also strongly correlated with legal status (Cra- controlling for the described issue and further critical mér’s V = 0.58). All other independent variables’ corre- arguments. lations are inconspicuous. 3.3 P redictor and control variables3.4 Methods With regard to our hypotheses on immigrant-specific fea - Seeking to analyze several different but not mutually tures, the duration of stay in Germany (the time between exclusive outcomes, we estimate separate binary mod- immigration and the survey interview) is adequate to els for each dependent variable. The main interest of test H #8 and level of linguistic competence in German our analysis is on differences in “treatment” by Jobcent - (respondents’ subjective evaluation) regarding H #9. ers between groups of persons, that is natives and immi- Both legal status (temporary residence permit being the grants. Individuals’ membership in these groups does not status of recognized asylum seekers and most recently change over time in our sample (or only rarely, regarding arrived non-EU citizens) and world regions of origin the categorized duration of stay). As the majority of per- (broad categories considering the number of observa- sons does not stay in the sample for more than one survey tions) operationalize H #10. Note, however, that legal sta- wave, the variance in the data which drives our results tus and duration of stay do not provide any information is derived from the differences between person-waves, on the time when German authorities grant the status of rather than within persons across waves. Therefore, we a legally recognized refugee and thus a (fixed-term) resi - choose simple logistic regression analysis (a robustness dence permit. check with logistic random-effects models yield very sim - Several other individual and household-related char- ilar results, cp. Sect. 4.3). By applying clustered standard acteristics of the observed population (including natives) errors at the person level, our models take account of the serve as control variables in our multivariate analyses, non-independence of repeated observations of the same including age, level of education, duration of the current person, where applicable. Results are presented as aver- unemployment episode, subjective health, and partner age marginal effects (AME), which can be read as the pre - household. In addition, we assume gender to become rel- dicted average difference in percentage points regarding evant (only) when viewed in its interaction with own chil- the probability of the dependent variable for a one-unit dren to care for, also depending on children’s age (Leber Further correlations pertain to age and duration of unemployment (Person’s rho = 0.32) and a fairly but not prohibitively high correlation between having a German language skills were coded as “very good” if respondents said that partner and a combined variable which captures a persons’ sex and the fact of German was their first language; missing and implausible values were sorted having children in the household (Cramer’s V = 0.55). into the medium category. Beyond the sampling-induced clustering of person-years within persons, As the survey is offered in various languages, Turkish, Russian, and since one could also control for treatment-induced clustering at the level of Job- wave 10 Arabic (Jesske and Schulz 2018, p. 74), its participation rate and centers or case managers. However, such information is not provided in the validity regarding immigrants should be better than in German-only sur- data. If it could be included in the models, this might slightly change the veys. estimated standard errors and significance of coefficients. 9 Page 10 of 24 R. Lehwess‑Litzmann , J. Söhn Table 1 Jobcenters’ offers to job ‑seeking basic‑income recipients: natives and immigrants, by duration of stay in Germany All job‑seekers Thereof… Natives Migrants Thereof … Up to 4 years At least of stay 5 years of stay Regular employment 31.8 31.8 31.7 24.5 38.1 Marginal employment 17.0 17.2 17.2 13.7 20.4 Assistance in applications 28.3 28.0 29.3 29.9 29.2 Reimbursement of application costs or travel expenses 43.0 47.6 31.9 29.0 32.5 Program with employer or internship 12.1 11.8 13.0 16.3 10.2 Fin. support to become self‑ employed 2.3 2.3 2.2 1.6 3.0 Activation or placement voucher 14.9 15.4 13.8 10.4 18.1 Vocational (re‑)training or a course 19.6 14.2 31.9 39.8 26.8 Other offers 1.8 2.0 1.3 0.9 2.0 No offer, excl. integration or language course 29.5 29.0 31.4 31.2 30.4 Integration or language course 14.7 0.9 46.2 69.6 23.8 No offer, incl. integration or language course 26.6 29.0 22.3 15.8 27.9 Source: PASS waves 9–14, own calculations. Weighted mean yearly shares of clients who received at least one of the respective offers in the period 2015–2020 change of an independent variable, with all other predic- always refer to the weighted yearly average across the tors held constant. 6 years studied. While the multivariate analysis is based on unweighted Compared with the natives in our observed popula- data, descriptive results are always weighted (pertain- tion, the share of men is higher among immigrants (59.8 ing to person-waves). Weighting the data evens out dis- vs. 52.5 percent), in particular among those immigrants proportional numbers in the sample, e.g., persons form having arrived during the recent 4  years (73.4 percent Syria and Iraq, who were oversampled in wave 10 (year men). Immigrants without any professional qualification 2017) of the PASS survey (Jesske et  al. 2019, preface). are clearly overrepresented (54.9 vs. 39.4 percent among The weighting parameters included in the PASS always natives), with 61.2 percent among the recent immigrants. refer to one specific survey wave. In order to determine Yet, the share with an academic qualification is also con - values for the whole period of observation, we first cal - siderably higher among immigrants studied here (16.8 culate weighted values for each of the five panel waves vs. 3.6 percent among natives). In the native group, a separately and then average those values across all waves. majority (57.0 percent) has a degree of non-academic As a small qualitative add-on, we insert few selected vocational training (vs. 28.3 percent of immigrants). The excerpts of expert interviews we conducted during the foreign-born Jobcenter-clients live far more often with larger mixed-methods research projects into which the a partner (61.7 vs. 28.8 percent) and with children than analyses presented here was embedded. The interviews, their native peers. anonymized in this article, with mid-level managers of Among the immigrants, 13.2 percent are German citi- Jobcenters and Employment Agencies in a mid-sized zens and 19.4 percent citizens of another EU country West-German city and in a more rural community took (Appendix Table  7). Among third country nationals, the place in 2016/2017. share with a temporary residence permit is higher than with a permanent one. This holds in particular for immi - grants living in Germany for less than 5  years (65.0 per- cent with temporary residence permit). Among those 4 Results and discussion who give valid answers to the respective survey question, 4.1 Description of the r eference population a high share of the recent immigrants say they came to and Jobcenters’ ALMP offers Germany as asylum-seekers: 86.4 percent in 2018 (wave The first part of this section describes our reference 12) and 69.3 percent in 2019 (wave 13). Persons from population by some personal and household features the former Soviet Union and Eastern Europe constitute of native and immigrant Jobcenter clients (Appendix the largest shares of non-recent immigrants (30.8 and Tables 6 and 7) as well as offers made to them by Jobcent - 26.0 percent, respectively). The majority of those having ers between 2015 and 2020 (Table 1). The shares reported Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 11 of 24 9 stayed in Germany for less than 5 years stem from (West- category “clean” of integration and language classes (see ern) Asian countries (62.5 percent), including countries Sect.  4.3 below). The integration courses are “offered” to many refugees fled from. Looking at immigrants’ Ger- 69.6 percent of recent immigrants, they are obligatory in man language skills, 36.4 percent among recent immi- many cases. Among those staying in Germany already grants and 54.4 percent among the longer-settled ones more than 4  years, 23.8 percent reported the offer to consider their own skills as good or very good. attend integration and language classes. Regarding the different measures which the PES can Leaving aside such measures not applicable to natives, offer to job-seeking clients, Table  1 shows no major dif- we find that 29.0 percent of native and 31.4 percent of ference between the proportions of native and (all) immigrant clients (all in regular contact with their Job- immigrant clients referred to job vacancies pertaining to center) do not receive any offer in an average year of regular employment (31.8 vs. 31.7 percent). Yet, distin- observation. Once we include integration and language guishing immigrants by duration of residence reveals a classes, the share of recent immigrants without offer substantial gap. In accordance with some of our expec- from the Jobcenter is only 15.8 percent. tations, fewer recent immigrants (24.5 percent) report such referrals, but 38.1 percent of the longer-settled 4.2 Determinants of Jobcenters’ offers: multivariate immigrants do. The latter also received referrals to mar - analyses ginal employment more often (20.4 percent) than natives The presentation of multivariate results has two parts: (17.2 percent) or recent immigrants (13.7 percent). By In the first, we analyze for the whole sample, natives contrast, Jobcenters offered support with job applications and immigrants, in how far personal and context fac- most often to recent immigrants (29.9 percent). Natives tors explain the clients’ chance of the Jobcenters offering feature the highest share of those offered reimbursement them a specific type of employment policy measure and of application costs and travel expenses (47.6 vs. 29.0 whether being an immigrant plays a significant role. In and 32.5 percent among recent and longer-settled immi- the second part, we run our model for immigrants only grants respectively). Probably, native Jobcenter clients and add immigrant-specific variables in order to differ - are invited more often to job interviews than immigrant entiate this heterogeneous group. This helps us to look clients. deeper into the conditions under which Jobcenters make With the exception of assistance in applying for jobs, offers to immigrants. the measures reviewed so far depend strongly on labor market conditions and employers’ decisions. In contrast, 4.2.1 Nativ es and immigrants compared caseworkers’ own considerations are more decisive for The joint model for immigrants and natives in Table  2 offers regarding the other elements of the ALMP tool - confirms most of the above descriptive findings (cp. box. Jobcenters offered a subsidized program with an Table 1). It is again paramount to distinguish immigrants employer or an internship most often to recent immi- with shorter and longer duration of stay because effects grants (16.3 percent). Financial support to become self- sometimes point in opposite directions. For those immi- employed is rarely offered (2.3 percent of the observed grants who have arrived up to 4  years before the inter- population, and most often non-recent immigrants, view, the estimated probability to be referred to a regular 3.0 percent). Activation or placement vouchers, often job or an apprenticeship is on average 12.3 percentage meant for personal occupational orientation but also points (p.p.) lower than that of native clients, while there for further training or support in the job search, show a is no significant difference between natives and longer- similar pattern as referrals to (regular or minor) employ- settled immigrants. H #1a—the expectation that the PES ment. Recent immigrants receive them least often (10.4 offer more jobs to immigrant clients in general—is thus percent) and longer settled immigrants more often (18.1 not supported by our findings, while there is evidence percent), with native clients in between (15.4 percent). which corroborates H #1b. Possibly, Jobcenter staffs do Concerning occupational (re-)training “or other not consider recent immigrants yet fit for employment or courses”, by contrast, recently arrived immigrants report they anticipate that it is harder for them to meet employ- such offers most often (39.8 percent), natives least often ers’ requirements (e.g., German language skills). With (14.2 percent), the other immigrants positioned in regard to minor employment, caseworkers are most likely between (26.8 percent). As mentioned above, however, to refer longer-settled immigrants to job vacancies (6.0 we cannot be sure that all respondents kept this response p.p. more than native clients), with no difference between natives and recently arrived immigrants. H #2 is thus supported only for part of the immigrant clientele. In the multivariate analyses, we sort the small groups of clients from Turkey With regard to ALMP programs, the pattern varies. to the Middle East/Asian category, while those from Northern, Western and Contrary to the descriptive results, the model with all Southern Europe are aggregated with Eastern Europeans. 9 Page 12 of 24 R. Lehwess‑Litzmann , J. Söhn Table 2 Determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support ( joint model) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program with Activation or Vocational No offer employment employment applications of application or employer or placement (re‑)training travel costs internship voucher or course Person category (reference: native basic income recipients looking for a job) Basic income − 0.123*** − 0.007 − 0.053*** − 0.201*** 0.015 − 0.069*** 0.139*** 0.101*** recipients looking for a job who immigrated during the past 4 years Basic income 0.039 0.060*** 0.057** − 0.114*** 0.012 − 0.013 0.115*** 0.008 recipients looking for a job who immigrated at least 5 years ago Age (reference: 35 to 44 years) 18 to 0.035 0.053* 0.103*** 0.048 0.131*** 0.009 0.005 − 0.075*** 24 years 25 to 0.015 0.024 0.032 0.046* 0.030* 0.011 0.014 − 0.035* 34 years 45 to − 0.032 0.003 − 0.050** − 0.056** − 0.015 − 0.035** − 0.054*** 0.070*** 54 years 55 to − 0.089*** 0.005 − 0.089*** − 0.090*** − 0.038** − 0.063*** − 0.126*** 0.142*** 64 years State of health: − 0.028 − 0.008 − 0.035 − 0.069*** − 0.025 − 0.034* − 0.056** 0.043* bad (reference: very good to less good) Professional qualification (reference: none, lower ‑secondary school‑leaving certificate at most) None, but 0.031 − 0.033* 0.003 0.012 0.000 0.025 0.048** − 0.002 upper or intermediate secondary school‑ leaving certificate Non‑ 0.056*** − 0.018 0.031* 0.068*** 0.002 0.032** 0.021 − 0.031* academic professional qualification Academic 0.049* − 0.070*** 0.084*** 0.101*** 0.008 0.074*** 0.046** − 0.064*** qualification (university or technical/ teacher training col‑ lege) Duration of current unemployment so far (reference: 12 to 23 months) 0 to − 0.063*** − 0.019 − 0.057*** − 0.093*** − 0.041*** − 0.027* − 0.068*** 0.105*** 2 months 3 to 0.034 − 0.005 0.017 − 0.012 − 0.025 0.010 0.001 − 0.015 11 months 24 months − 0.057*** 0.008 − 0.011 − 0.021 0.000 0.003 − 0.019 0.014 and more Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 13 of 24 9 Table 2 (continued) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program with Activation or Vocational No offer employment employment applications of application or employer or placement (re‑)training travel costs internship voucher or course Gender and youngest child in household (reference: woman w/o children in household) Mother with − 0.136*** − 0.067*** − 0.078** − 0.210*** − 0.044** − 0.081*** − 0.125*** 0.288*** child aged 0 to 2 years Mother with − 0.020 0.007 − 0.009 − 0.046* 0.005 − 0.009 − 0.012 0.021 child aged 3 to 17 years Father with 0.066* − 0.022 0.081** − 0.037 0.049* 0.031 0.019 − 0.018 child aged 0 to 2 years Father with 0.022 − 0.013 0.059* − 0.012 0.037* 0.016 0.008 − 0.032 child aged 3 to 17 years Man without 0.015 − 0.010 0.048** − 0.002 0.033** 0.023 0.010 − 0.014 children in household Partner in − 0.057*** − 0.045*** − 0.044** 0.011 − 0.007 − 0.033** − 0.030* 0.041** household (reference: none) Underem‑ − 0.010*** − 0.004* − 0.015*** − 0.015*** − 0.004* 0.008*** − 0.004* 0.012*** ployment rate in fed‑ eral state Year (reference: 2015) 2016 − 0.018 0.005 0.005 − 0.001 0.012 0.003 0.006 0.010 2017 − 0.020 0.002 0.009 − 0.013 0.018 0.008 0.001 0.025 2018 − 0.024 − 0.013 − 0.024 − 0.057** 0.006 0.000 0.008 0.027 2019 − 0.004 − 0.027* − 0.023 − 0.050** 0.030* 0.011 0.019 0.029 2020 0.034 0.000 0.042* − 0.017 0.050*** 0.052*** 0.072*** − 0.027 Pseudo‑R 0.032 0.019 0.029 0.040 0.034 0.034 0.064 0.041 N (person‑ 8226 8224 8221 8207 8228 8200 8225 8234 waves) Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations *p < 0.05; **p < 0.01; ***p < 0.001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter a  “No offer” means “financial support to become self‑ employed” and “other offers” have also not been granted, neither integration or language courses control variables shows that non-recent immigrants are Probably, native clients receive invitations for job inter- most likely to report support with their applications, views more often than immigrants (and then have the while recent immigrants are least probable to do so costs reimbursed), in spite of receiving fewer referrals (+ 5.7 p.p. respectively − 5.3 p.p. compared with native and less support with applications from Jobcenters than clients), in part falsifying H #3. In line with H #4, both non-recent immigrants. recent and non-recent immigrants have a significantly Subsidized programs with employers or internships lower likelihood of Jobcenters reimbursing application could help jobless immigrants gain employers’ trust, or travel costs than natives (− 20.1 p.p. and − 11.4 p.p.). which they possibly do not enjoy as much as native appli- cants. However, multivariate regression yields no dif- ference between the three client groups for this costly ALMP measure. Hence, despite the bivariate result that This difference between these gross effects (see Additional file  1: Table S1) the recently immigrated reported this offer above aver - and net effects is partly due to the relatively high share of academics among age, H #5 is not confirmed when controlling for the other recent arrivals, as high education makes reporting support for applications more likely, independently of migration status (see Table 2). 9 Page 14 of 24 R. Lehwess‑Litzmann , J. Söhn covariates in the model. But there is no inequity in access (e.g., the large share without professional qualification) to either. a larger extent than by migration-related factors. By con- Regarding activation or placement vouchers (for, e.g., trast, support with applications does not differ between occupational orientation, further vocational training), the two groups at first sight, but turns out statistically newly settled immigrants face lower chances (− 6.9 p.p.) significant in the full model, such that recently-arrived of receiving respective offers compared with natives, immigrants receive fewer offers of this kind than natives, despite their hypothesized higher need for such measures given their social composition. Overall, the explanatory due to lacking knowledge about the local labor market power of our models with regard to Jobcenters’ offers is and job culture or because of non-recognized foreign cre- not very high. This hints at the importance of unobserved dentials. We can only speculate whether these offers are factors, probably linked to the dynamics of the service crowded out by other programs (e.g., integration courses interaction, clients’ personal preferences, and case man- might partly cover labor market topics) or whether agers’ professional attitudes (Dietz and Osiander 2019) as caseworker consider recent immigrants’ knowledge of well as the particular “cultures” of individual Jobcenters. German insufficient for following programs offered in One main finding of this section is that recent immi - German. As longer-settled immigrants face those obsta- grants generally receive fewer offers from out of the cles to a lesser extent than new arrivals, it appears plausi- ALMP toolbox than native clients. As for the longer-set- ble that they are not treated differently from natives. Yet, tled immigrants, the pattern of Jobcenter activity in some H #6-plus still needs to be dismissed. respects looks like an attempt to compensate the labor Jobcenters offer the more extensive programs of “voca - market obstacles that still confront this group, possibly tional (re-)training or a course” comparatively often to because their linguistic and cultural knowledge is already recent immigrants, with a likelihood 13.9 p.p. higher higher than that of recently arrived immigrants, making compared with natives and non-recent immigrants’ prob- non-migrant specific offers more feasible in the eye of the ability increased by 11.5 p.p., respectively. This finding Jobcenter. chimes with H #7a, which supposes an especially high need and motivation on the part of immigrants (however, 4.2.2 Comparisons among immigrants see chapter 4.3.1). In order to find out more about how Jobcenters react to Finally, recent immigrants (but not the longer-settled) the different profiles of immigrant clients, this section remain significantly more often (+ 10.1 p.p.) without any will explore the pattern of proposed ALMP measures ALMP offer than natives in a given year of observation within the immigrant subgroup of our sample. Restrict- (integration and language classes excluded). ing our estimations to immigrants allows us to extend Table  2 reports the effects of control variables which our estimation models by some variables which are only are, however, of no central interest here. Let us just men- meaningful for this group: legal status, duration of stay tion that most ALMP measures are more likely offered by in Germany, and the self-perceived level of German lan- Jobcenters to clients with a higher level of qualification guage skills. As for the dependent variables, we now rather than a low level. Jobcenters do not try to compen- also look at offers reserved for immigrants, that is inte - sate the largest gaps in education by their measures, oth- gration courses and other German language classes. erwise, those with only lower-secondary education would A long duration of stay has a significant impact: Refer - be more in their focus. Control variables contribute to rals to regular employment or apprenticeships are more the statistical explanation of the dependent variables’ var- likely (+ 11.1 p.p.) if immigrants have lived more than 19 2 iance, increasing the Pseudo-R compared to the gross 10  years (rather than less long) in the country (Table  3). models with migration experience as the only predic- For them, offers of minor employment also have a prob - tor (see Additional file  1: Table  S1). Overall, when there ability (7.0 p.p.) higher than that to more recently arrived already is a significant bivariate correlation, the effects of migration status on Jobcenter offers remain statisti - cally significant with regard to most outcome variables There is no information on this in the PASS data. Matching the ADIAB when the full multivariate models consider the composi- extension to the PASS could provide us with the information on the respon- tion of the groups. The less frequent referrals to marginal sible Jobcenter. We did not opt for this due to a loss of cases and a lack of employment of recently arrived immigrants (compared administrative data in the case of recently-arrived migrants. The information on world regions of origin would also be suited as a to natives) are explained by the composition of that group variable, but it is too highly correlated with the legal status to be simultane- ously included. If regions are included instead of legal status, we do not find any effects of the former on ALMP offers, except that clients from Africa and Middle East/Asia are more often proposed integration or language Other possible predictors like the partner’s employment status and house- courses. This, of course, should have to do with their legal status, which is hold income turned out to have no influence. omitted in these alternative models. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 15 of 24 9 individuals. By contrast, longer-settled immigrants face Middle East, are also offered an integration or language a higher risk of not reporting vocational (re-)training course significantly more often than all other clients. and other courses (− 13.2 p.p.), and, of course, receive All the remaining variables are identical to the ones significantly fewer offers to participate in integration or used in Sect.  4.1, Table  2. We observe that the impact of language courses. Interestingly, there is no significant dif - clients’ qualifications differs from the full sample model ference between the most-recently arrived and the immi- (including natives) with regard to two measures: in the grants staying between 5 and 10  years, apart from the immigrant-only model, academics are not referred to more frequent offer of integration or language courses to regular employment more often than persons with very the former group. It is only due to these courses that the low education, and neither are academics significantly most recently immigrated have a lower probability (− 6.6 more likely of being offered activation or placement p.p.) than less recent arrivals of reporting no offer at all. vouchers than less educated immigrants. Also, for other Appendix Table 7 shows that self-assessed German lan- factors captured in variables, like bad health, age, or a guage skills differ widely between immigrant Jobcenter partner in the household, we find fewer or weaker effects clients. Corroborating H #9, support with applications is in the migrant-only model. Yet, this might be due to the less probable for persons with bad or very bad German smaller sample size. skills (− 10 p.p. compared with clients with “reasonable” German skills). There is no significant limitation of refer -4.3 Robustness tests rals to minor jobs in case of clients with weak German Given the limitations of the data already addressed in language skills, but a reduced likelihood of reimbursing the methods section, we put our results to three tests job-searching costs points to employers’ under-average of robustness. They include using a different regression response to those persons’ applications. In addition, cli- model (4.3.2), a different sample (4.3.3), and an attempt ents with “bad” or “very bad” German skills are less likely to correct potential false responses to one survey item (− 4.7 p.p. and − 13.3 p.p.) to report being offered pro - (4.3.1). Statistically speaking, our results presented above grams with employers or internships than clients with turn out to be trustworthy. In the case of the item “voca- “reasonable” German. As for vocational (re-)training tional (re-)training or a course”, a doubt remains in the and other courses, the probability of offers diminishes, case of immigrants in our PASS-sample. too (− 6.2 p.p. and − 13.0 p.p.). It seems that casework- ers prioritize integration and language courses if clients 4.3.1 R eassessing offers of vocational (re‑)training and other hardly speak German, and contemplate ALMP options courses in the light of integration and language classes as soon as a certain level of German language has been In chapter  3, we already addressed the issue of a pos- reached. Possibly, Jobcenters consider the possession of sible overlap of the response category vocational (re-) some German skills as the condition of being able to par- training or a course and the migrant-specific integra - ticipate in most ALMP programs. The results on integra - tion or language courses. Due to the order of interview tion and language courses also reflect the regulation that items, clients who were only offered an integration or recent immigrants with good knowledge of German do language course are likely to say “yes” when they are first not need to attend them. In addition, referrals to regular asked about offers of Vocational (re-)training or a course”. jobs are significantly less often made to immigrant clients Potentially, these individuals then tick “yes” a second with “bad” German skills (− 5.1 p.p.), but more often to time when they are concretely asked about an Integration those with “very good” ones (+ 6.1 p.p.). All in all, results or language course later on. As a test of robustness, we on the impact of German language skills are consistent therefore eliminate all person-waves from our sample in with H #9 regarding both ALMP measures and referrals which both offers are reported and re-run our regression to job vacancies. model on this reduced sample, which now consists of Legal status, being a German citizen or not, having a 7449 person-waves, among which 2414 are migrants (820 permanent or fixed-term residence permit, does not person-waves less than the full sample). The results seem to affect most Jobcenters’ ALMP offers, as soon as other immigrant-specific factors are controlled for. The Also, the additional variables in the second model forbid a direct compari- possibility that the client might (have to) leave Germany son of control variables’ impacts with the above model for the whole sample. in the near future does not seem to discourage casework- Additional file  1: Table  S2 displays some differences of the characteris - ers. Counter to H #10, German nationals, EU citizens and tics of respondents, grouped by the combination of offers they reported: third-country nationals with a permanent residence sta- respondents who report both offers are more similar to respondents who tus are offered a program with an employer or an intern - only reported “integration or language course” in terms of legal status, duration of stay and German language skills, but more similar to respond- ship less often than clients with a temporary residence ents who only said yes to having been offered “vocational (re-)training or a permit. The latter, mostly stemming from Africa or the course” in terms of professional qualification. 9 Page 16 of 24 R. Lehwess‑Litzmann , J. Söhn Table 3 Determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support (immigrants only) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program Activation or Vocational Integration No offer employment employment applications of application or with placement (re‑)training or language travel costs employer or voucher course course internship Legal status (reference: not EU citizen, fixed‑term residence) German 0.028 0.028 0.04 0.035 − 0.056* 0.032 − 0.036 − 0.154*** 0.024 citizen EU citizen, 0.046 0.018 0.04 0.047 − 0.047* 0.046* 0.000 − 0.174*** 0.036 non‑ German Not EU − 0.013 − 0.02 0.110* 0.000 − 0.082** 0.048 0.038 − 0.146** 0.042 citizen, open‑ ended residence Duration of stay in Germany so far (reference: 5 to below 10 years) 0 to below − 0.038 0.000 − 0.044 − 0.03 − 0.014 − 0.011 − 0.018 0.188*** − 0.066** 5 years 10 years and 0.111** 0.070* 0.013 0.008 − 0.042 0.044 − 0.132*** − 0.302*** 0.044 more German language skills (reference: reasonable) Very good 0.061* 0.032 0.025 0.077* − 0.018 0.013 − 0.031 − 0.127*** 0.015 Good 0.037 − 0.005 0.039 0.076*** 0.005 0.006 0.004 − 0.102*** 0.017 Bad − 0.051* − 0.020 − 0.101*** − 0.100*** − 0.047* − 0.042* − 0.062* − 0.022 0.036 Very bad − 0.067 − 0.031 − 0.100* − 0.172*** − 0.133*** − 0.028 − 0.130* − 0.012 0.033 Age (reference: 35 to 44 years) 18 to 24 years − 0.016 0.019 0.013 − 0.008 0.049 0.027 0.013 0.027 − 0.008 25 to 34 years 0.000 0.005 − 0.004 − 0.01 0.003 0.009 − 0.018 − 0.021 0.020 45 to 54 years − 0.014 0.01 − 0.024 − 0.031 − 0.004 − 0.039* − 0.056* − 0.034 0.025 55 to 64 years − 0.104*** − 0.010 − 0.053 − 0.057 0.003 − 0.050* − 0.123*** − 0.049 0.092*** State of health: − 0.004 − 0.008 − 0.073 − 0.048 − 0.037 − 0.021 − 0.123** − 0.049 0.015 bad (reference: very good to less good) Professional qualification (reference: none, lower ‑secondary school‑leaving certificate at most) None, but 0.013 − 0.030 0.036 0.018 0.026 0.005 0.052* − 0.048* 0.021 upper or intermediate secondary school‑ leaving certificate Non‑ 0.056* − 0.023 0.048 0.01 0.045* 0.023 0.057* − 0.043 0.016 academic professional qualification Academic 0.008 − 0.060** 0.082*** 0.077** 0.038* 0.035* 0.045 − 0.048* 0.002 qualification (university or technical/ teacher train‑ ing college) Duration of current unemployment so far (reference: 12 to 23 months) 0 to − 0.001 − 0.002 − 0.024 − 0.001 − 0.028 − 0.008 − 0.048 − 0.128*** 0.087*** 2 months 3 to 0.054 0.009 0.043 0.008 − 0.021 0.006 − 0.003 − 0.048 0.026 11 months 24 months 0.014 0.025 0.027 0.018 0.008 0.011 0.004 − 0.006 0.025 and more Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 17 of 24 9 Table 3 (continued) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program Activation or Vocational Integration No offer employment employment applications of application or with placement (re‑)training or language travel costs employer or voucher course course internship Gender and youngest child in household (reference: woman w/o children in household) Mother with − 0.144*** − 0.057 − 0.071 − 0.123** − 0.039 − 0.039 − 0.123* − 0.099 0.187*** child aged 0 to 2 years Mother with − 0.061 0.017 − 0.017 − 0.037 0.001 0.007 0.000 0.028 0.006 child aged 3 to 17 years Father with 0.035 0.004 0.124** 0.004 0.086** 0.050 0.034 − 0.005 − 0.034 child aged 0 to 2 years Father with 0.003 0.007 0.065 − 0.001 0.046 0.033 0.019 − 0.009 − 0.018 child aged 3 to 17 years Man without 0.023 0.012 0.055 0.016 0.061** 0.022 0.006 0.000 0.003 children in household Partner in − 0.037 − 0.039* − 0.056* − 0.005 − 0.023 − 0.036* − 0.070** − 0.004 0.047** household (reference: none) Underem‑ − 0.003 − 0.002 − 0.009** − 0.012*** − 0.005 0.006* 0.002 − 0.005 0.002 ployment rate in federal state Year (reference: 2017) 2015 0.046 − 0.024 − 0.030 − 0.052 − 0.021 − 0.009 0.075* (base) 0.042 2016 0.030 − 0.008 − 0.023 0.012 − 0.007 0.007 0.063* 0.023 0.015 2018 0.018 − 0.031 − 0.006 − 0.048 − 0.024 0.014 0.048 0.044 − 0.014 2019 0.049 − 0.053** − 0.047 − 0.056* − 0.006 0.025 0.041 − 0.012 0.014 2020 0.143*** 0.005 0.056 − 0.003 0.036 0.079*** 0.119*** 0.027 − 0.033 Pseudo‑R 0.049 0.034 0.037 0.034 0.047 0.036 0.043 0.247 0.104 N 3191 3187 3185 3176 3193 3172 3190 2886 3195 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations *p < .05; **p < .01; ***p < .001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter “No offer” means “financial support to become self‑ employed” and “other offers” have also not been granted dramatically change concerning Jobcenters’ offer of integration/language course. Yet, there are some reasons Vocational (re-)training or a course (Table  4): instead of to expect a high number of immigrant clients who were prioritizing newly arrived immigrant clients by 13.9 p.p. only offered integration classes, regarding both timing as compared to native clients (Sect.  4.1), they are now and language preconditions. An integration class usu- at a disadvantage of − 10.4 p.p. As for non-recent immi- ally takes six to seven month on average (Goethe Insti- grants, the coefficient turns insignificant, i.e., Jobcenters tut 2020). Depending on when it starts and how often do not treat them differently form native clients. clients and case managers talk to each other, the chance But how convincing are these alternative results, given of being offered a subsequent vocational (re-)training in that a part of the observations which have been elimi- the same reference year of the survey should be limited. nated may really have been offered both types of courses, Concerning language skills, integration courses lead to in which case ticking both survey items with a “yes” the language level A2 or B1 (Goethe Institut 2020), which would be the adequate answer? There is no way of know - is usually not sufficient to follow the complex teach - ing the exact share of respondents who says yes to voca- ing in vocational training attended by native speakers tional (re-)training or a course” but in fact meant only an as well. Furthermore, official statistics do not imply an 9 Page 18 of 24 R. Lehwess‑Litzmann , J. Söhn Table 4 Robustness test (I): determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support ( joint model), modified responses to item “vocational training or a course” Independent variables Dependent variable: vocational (re‑)training or a course All sample persons Only sample persons who did not also tick the offer of integration or language class Person category (reference: native basic income recipients looking for a job) Basic income recipients looking for a job who immigrated during the past 0.139*** − 0.104*** 4 years Basic income recipients looking for a job who immigrated at least 5 years ago 0.115*** 0.017 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations ***p < 0.001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter Repetition of results from Table 2 for comparison. Same control variables as in Table 2 (coefficients not shown due to their similarity to those in Table 2) 4.3.3 O nly one wave per person overrepresentation of immigrants (restricted to those On average, immigrant persons remain in our sample for with non-German citizenship) among participants of a smaller number of waves. In order to check whether PES-financed further vocational training (see introduc - this could (partly) explain our results, we perform an tion; BA 2019c). A final reason to expect the share of cor - extra regression analysis where each person is observed rect “double-yes answers” to be small can be derived from only once across the whole observation period, no mat- the research of Kasrin et  al. (2021, p. 4). They find that ter if they satisfied the sampling criteria in 1, 2, or even the number of refugee Jobcenter-clients who attended more years of observation. We find that the results do further vocational training was only close to half of those not differ much from the full model, no matter whether attending a program with an employer. As we know from we use the first or last wave in which the sample person our data that recently-immigrated PASS respondents is observed (Additional file  1: Table  S4). Regarding the mention the latter offer relatively rarely (16.3 percent), impact of migrations status as seen in Table 2, significant the share of those receiving vocational (re-)training offers coefficients are widely reproduced. The only exception is should be much smaller than the 39.8 percent reported that longer-settled immigrants get more referrals to regu- above in Table  1. These arguments speak in favor of our lar jobs in their first year of observation, whereas there argumentation that among respondents reporting both is no significant difference compared with natives in the (re-)training and language classes many were in fact only person-wave model (Table 2). As for effect sizes, they are offered the latter, which will have a bearing on the con - also very close in the alternative models, with the mod- clusions drawn below. els using the first observation of each person yielding slightly bigger effect sizes than the models using the last 4.3.2 Random‑effects model observation, especially for the newly arrived immigrants. As mentioned above in the methods section, our empiri- This could mean that the effect of migration background cal findings are derived from the variance of phenomena on Jobcenters’ offers is higher in the beginning of a job between observations, i.e., person-waves. Theoretically, it search. Overall, the findings of our robustness checks could also be possible that Jobcenters’ offers are triggered underscore that the results in Sect.  4.2 are not driven by changes that happen in the lives of clients. In this case, by a difference within persons over time, but differences a model that takes the longitudinal dimension of the between persons. panel data into account would be more adequate. As a robustness test, we apply a random effects model, which 5 Conclusion builds on the variance both between and within persons. Based on survey answers given by job-seeking recipients The results do not change compared to our main model: of basic income, our analysis of the PASS survey in year it is the same coefficients that turn out statistically sig - 2015 to 2020 sought to identify determinants of ALMP nificant, and they have the same sign (Additional file  1: offers made by German Jobcenters to its clients. Our Table  S3). The between component seems be dominant special attention was on possible differences between compared with within component. This may also be due native and immigrant clients as well as on the effect of to the fact that only a minority of respondents is observed duration of stay among the latter. Our joint model for for more than 1 year. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 19 of 24 9 all job-seeking basic income recipients included native specific way one applies for jobs in Germany. An alter - clients, recent immigrants (having arrived in Germany native explanation of recent immigrants’ high propen- up to 4  years before the interview) and longer-settled sity to tick “yes” on the survey item “vocational training, immigrants (having stayed for 5 years or more). Whereas retraining or a course” could be its overlap, possibly aggregate administrative data, treating non-German caused by the sequence of the respective survey items, citizen as one single group, suggested no relevant immi- with “integration course or another German course”. This grant-native gaps (see BA 2019c as cited in the introduc- could erroneously make the occurrence of “vocational tion of our contribution), our distinction by time since training, retraining or a course” appear as relatively fre- migration proved crucial. quent among recent immigrants. There are indeed good Our multivariate analyses showed that recent immi- reasons to conclude that a high proportion of persons grants generally receive fewer referrals to job vacancies reporting both items were actually only offered integra - and fewer offers from out of the ALMP toolbox than tion or language courses (see Sect.  4.3.1). Our alterna- natives in a given year. For example, the likelihood of a tive model, excluding respondents with double positive caseworker suggesting such immigrants to apply for an answers, implies that Jobcenters are indeed likely to offer advertised regular job is 12.3 percentage points (p.p.) occupational (re-)training less often to recent immigrants lower, and the likelihood of offering an activation or than to natives or longer-settled immigrants, with no placement voucher is 6.9 p.p. lower than regarding native significant difference between the two latter groups. But clients. These results chime with some older, less differ - certainly, future research should use better data, e.g., on entiated findings on a negative effect of foreign or non- the realization of Jobcenter-sponsored (re-)training, to EU citizenship on benefitting from PES-sponsored (re-) verify this preliminary result. training or vouchers (Kruppe 2009, p. 14; Möller and Is there an alternative reading of our overall results Walwei 2009, pp. 305, 308–309; Osikominu 2005, p. 61) with regard to the more recent immigrants? The low as well as with previous research indicating some degree probability of referrals to open job positions to this group of ethnic discrimination by German administration could be based on their preference not to work. We judge (Grohs et al. 2016; Hemker and Rink 2017). Our interpre- this as implausible because Jobcenters tend not to allow tation is that many caseworkers, and probably employers, “laziness” on part of the clients, their main goal being to often perceive newly-arrived immigrants as not yet ready bring people into employment and to end benefit receipt. for entering the labor market, even though they do not Moreover, migration research (e.g. Sinning 2011) and our question their eagerness and even assert that some barri- own interviews with Jobcenter staff suggest that earning ers to employment which burden native Jobcenter clients money and sending part of it back to family members in are absent in the case of refugees (Boockmann and Scheu their home country is a major motivation and social obli- 2019, p. 409). Still, Jobcenters primarily offer integration gation among immigrants. Similar to findings by Boock - and language courses, which often constitute both a right mann and Scheu (2019, p. 409), one of the interviewed and an obligation, to recent immigrants. experts put it this way: There is one category that recently-immigrated “Those who already gained certificates strive for hav - respondents report significantly more often (+ 13.9 p.p.) ing them recognized, but this is not their main focus. than native Jobcenter clients. In the questionnaire, it is Their main focus is earning money.” (Interview II) called “vocational training, retraining or a course”. On the one hand, it might be convincing that recent immi- Integration and language classes, i.e., the kinds of grants received this kind of offer particularly often, as measures primarily proposed to recent immigrants, may this would reflect the special needs of a group that often be the first or second choice from the viewpoint of social lacks qualifications and certificates relevant for and/or law, Jobcenters, or immigrants themselves; they are often recognized on the German labor market (Knuth 2021, a necessary interim step to their successful labor market p. 56). Case managers explicitly highlight refugees’ need integration in the longer term. for further training (Boockmann and Scheu 2019, p. 408). With regard to longer-settled immigrants, Jobcent- On the other hand, most other measures were offered ers seem to use the full breadth of their toolbox to bring less often to recent immigrants, even support with appli- their clients into employment. As regards assistance with cations (− 5.3 p.p. compared with natives), despite such their job search, immigrants who have been staying in clients being probably little familiar with the culturally One can only speculate whether a confounding of language courses with One should keep in mind that the survey item for referrals to open job vocational training might also be a reason why Kruppe (2009, p. 14) finds a positions also mentions dual vocational training. Future research should try to small positive effect for recent immigration on the issue of training vouchers. distinguish these options. 9 Page 20 of 24 R. Lehwess‑Litzmann , J. Söhn Germany already for 5 years or more have a significantly individual migrants willing to learn, up to 4  years since higher probability of being referred to marginal (+ 6.0 arrival without training-related support can feel like a p.p.) employment as well as being offered support with long time and might cool off respective educational aspi - their applications (+ 5.7 p.p.), compared to native cli- rations. In order for Jobcenters to institutionally meas- ents, with no statistically significant difference regarding ure up to recent immigrants’ high motivation to succeed open positions of regular employment. By contrast, they in their host country, they should use an adequate mix are less often (− 11.4 p.p.) offered a reimbursement of of long-term oriented investment in human capital and application or travel costs, which conveys that they are short-term support to find work. ALMP measures com - less often invited for job interviews by potential employ- bining employment, occupational qualification and ers—an indicator of their disadvantages on the German advanced German language training seem promising labor market. The preferential support with applications (Boockmann and Scheu 2019, p. 418). offered to longer-settled migrants and the statistical simi - A desirable extension of our study would be to con- larity with native clients regarding most other ALMP trast ALMP measures offered with ALMP measures that measures suggests that there is no negative discrimina- are actually implemented. While our analysis of the for- tion against these migrants by Jobcenters. mer predominantly dealt with phenomena of selection, This reading is further backed by our finding that, as in an extended analysis could analyze the determinants of accordance with the principle of equity, migrants’ world self-selection that influence whether a client accepts an region of origin makes no significant difference for their offer. Furthermore, we do not know when offers are used probability of receiving ALMP offers, except for inte - strategically by Jobcenters to merely test the availability gration or language courses. The latter are significantly of clients for the labor market. This would, of course, more often proposed to clients from Africa and Middle change the interpretation of our results as such activating East/Asia, who overwhelmingly belong to the immigrant ‘offers’ mean nothing positive from the clients’ viewpoint. groups targeted by integration policy. Regarding ALMP However, we find it more plausible that Jobcenters’ offers measures which are not language-related, we do not find are intended as supportive, especially regarding compar- major differences by migrants’ legal status. The following atively long and extensive ALMP programs. statement of a caseworker interviewed for our project Our analysis was restricted to years when refugees underlines this: from Asia formed a large group among PES’ immigrant clients in Germany. It is up to future research to test Actually, as soon as the language part has more or whether our results hold once the structure of immigra- less been dealt with, we treat immigrant clients just tion changes again, e.g., if immigrants from European like the rest of our clients. We ask ourselves where countries (like Ukraine) and family migrants from out- the clients are heading, what would fit to them, what side of Europe will again become the majority of recently is reasonable, what are they able to achieve. (Inter- arrived immigrants. view II) In addition, our time window analyzed is a historical This statement also points to the importance of suf - boom phase of the German economy with low unem- ficient knowledge of German as a facilitator of further ployment rates. Relatively few job seekers competed for ALMP measures, and in particular for chances of finding the public resources dedicated to ALMP. If unemploy- employment. Our deeper analysis of which immigrant- ment should rise again in the years to come, possibly as specific factors are linked to Jobcenters’ offers shows a consequence of geopolitical conflicts, ALMP spending that referrals to regular employment and reimbursement per person will probably shrink even if aggregate expend- of application expenses are indeed offered more often iture grows (Lehwess-Litzmann 2018). No one can know to immigrant clients the better their German skills are. at present whether the German labor market will deal Immigrant job-seekers with weaker German skills are with this as well as with the COVID-19 crisis. Jobcent- also least likely to gain assistance in applications or pro- ers’ future immigrant clients will certainly need new grams with an employer or internships. occupational skills and certificates as well as employment Overall, our analysis gives reason to believe that, experience in the German labor market. compared with native clients, job-seeking immigrants receive overall equal treatment by the German PES, but only regarding longer-settled immigrants or those However, a particular consequence of the pandemic for migrants and refu- with (assumed or actual) good German language skills. gees is revealed by Brücker et al. (2021, 26 and 30): These groups experienced Whether Jobcenters’ activities are actually sufficient and a stronger rise of unemployment in the course of the year 2020, not primarily due to layoffs (net employment rates remained rather stable), but as a conse - how closely they correspond to immigrant clients’ prefer- quence of interrupted or cancelled integration, language and vocational train- ences deserves further investigation. At least in the eye of ing courses. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 21 of 24 9 Appendix See Tables 5, 6 and 7. Table 5 Basic‑income recipients looking for a job: persons and person‑ waves in the sample by migration status Person category Number of Number of times the person figures in the sample Number persons of person‑ 1 2 3 4 5 6 waves Natives 2711 1496 610 301 146 110 48 5041 100.0% 55.2% 22.5% 11.1% 5.4% 4.1% 1.8% Immigrants 2243 1.565 459 143 60 14 2 3234 100.0% 69.8% 20.5% 6.4% 2.7% 0.6% 0.1% Total 4954 3061 1069 444 206 124 50 8275 100.0% 61.8% 21.6% 9.0% 4.2% 2.5% 1.0% Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations Table 6 Basic‑income recipient looking for a job: socio ‑ demographic characteristics, by migration status (as % of the observed population) Attributes Natives All immigrants …Up to 4 years of stay …At least 5 years of stay Gender Male 52.5 59.8 73.4 46.7 Female 47.5 40.2 26.6 53.3 Age 18–24 10.9 5.4 11.1 0.5 25–34 30.8 31.3 43.6 21.9 35–44 20.2 29.1 32.0 25.0 45–54 21.0 20.4 10.9 30.0 55–64 17.0 13.8 2.4 22.8 Professional qualification No professional qualification 39.4 54.9 61.2 47.8 … With lower‑secondary school‑leaving certifi‑ 29.7 40.6 40.3 38.6 cate at most … With upper or intermediate secondary school‑ 9.7 14.3 20.9 9.2 leaving certificate Non‑academic professional training 57.0 28.3 21.2 36.0 academic qualification 3.6 16.8 17.6 16.2 Self‑reported (very) bad health 9.7 8.5 2.9 12.9 With partner in household 28.8 61.7 71.2 55.6 Children (by age) in household No children in household 62.3 38.8 37.6 40.1 At least one child aged 0–2 8.4 18.4 26.3 10.7 At least one child aged 3–17 29.4 42.7 36.1 49.1 Duration of current unemployment episode 0–2 months 20.6 26.8 23.1 30.4 3–11 months 9.2 9.7 12.4 8.0 12–23 months 10.3 12.1 20.1 7.2 24 months and more 59.8 51.5 44.4 54.3 N (person‑ wave) 5041 3234 2202 1032 Source: IAB, PASS, Welle 14 v1. 2015–2020. Own calculations. The figures represent the mean over the weighted values for each year of observation 9 Page 22 of 24 R. Lehwess‑Litzmann , J. Söhn Table 7 Immigrant basic‑income recipients looking for a job: legal status and countries of origin, by duration of stay (as % of the observed population of immigrants) Attributes All immigrants …Up to 4 years of stay …At least 5 years of stay Legal status German citizen 13.2 3.0 21.8 Non‑ German EU citizen 19.4 18.6 22.4 Third country national with permanent residence permit 18.8 13.4 22.0 Third country national with temporary residence permit 48.6 65.0 33.7 Duration of stay (mean in years) 9.8 3.2 14.9 Region of origin Northern, Western and Southern Europe 3.2 3.1 3.8 Eastern Europe 20.1 16.7 26.0 Former Soviet Union 18.7 7.7 30.8 Other Asian country 45.0 62.5 24.8 Turkey 3.1 2.5 3.7 Africa 8.4 6.6 9.4 Other countries/unknown 1.5 1.4 1.6 German language skills (own account) Very good 12.2 7.7 15.4 Good 33.3 28.7 39.0 Satisfactory 38.6 46.6 30.8 Bad 13.4 13.5 12.3 Very bad 2.5 3.5 2.5 N (person‑ wave) 3234 2202 1032 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations. The figures represent the mean over the weighted values for each year of observation Abbreviations Author contributions ALMP: Active labor market policy; BA: German Federal Employment Agency The authors have produced all parts of the paper collaboratively, with an (Bundesagentur für Arbeit); EU: European Union; IAB: Institute for Employment emphasis of JS on the theoretical framing and of RLL on the empirical analysis. Research; PASS: Panel Study Labour Market and Social Security; PES: Public Both authors read and approved the final manuscript. employment services; SGB: Social Code (Sozialgesetzbuch). Authors’ information Dr. René Lehwess‑Litzmann is a senior researcher at the Sociological Research Supplementary Information Institute (SOFI) in Göttingen. He analyzes labor‑market and social policy, labor ‑ The online version contains supplementary material available at https:// doi. market trends, individual career choices and employment trajectories in the org/ 10. 1186/ s12651‑ 022‑ 00313‑8. context of societal challenges. Dr. Janina Söhn is a senior researcher at the SOFI in Göttingen. Within the broad field of social inequality, she has published on immigrant integration, Additional file 1: Table S1. Gross model with migration experience legal status and integration policy, education, vocational training, employ‑ as the only determinants of various ALMP measures by Jobcenters to ment trajectories, and retirement. job‑seeking recipients of basic‑income support ( joint model). Table S2. Characteristics of immigrant basic‑income recipients looking for a job, by Funding combination of offers reported. Table S3. Robustness test (II): Random‑ The research was funded by Deutsche Forschungsgemeinschaft (DFG) under effects mode (Odds ratios). Determinants of various ALMP measures the sign SO1286/2‑1 and contract number 616892. The authors conducted by Jobcenters to job‑seeking recipients of basic‑income support ( joint the research under conditions of scientific independence. model). Table S4. Robustness test (III): Only one person‑ wave per person. Determinants of various ALMP measures by Jobcenters to job‑seeking Availability of data and materials recipients of basic‑income support ( joint model). PASS data can be obtained from the IAB’s research data center under certain conditions, see https:// fdz. iab. de/ en/ FDZ_ Data_ Access. aspx. Acknowledgements We would like to thank participants of the Colloquium of the Sociological Declarations Institute of University of Duisburg‑Essen, where we presented preliminary results on 3rd June 2020. Many thanks goes also to the PASS team for some Ethics approval and consent to participate valuable pieces of information and last but not least to our anonymous The article mainly uses anonymized and aggregated secondary data, drawing reviewers for their commitment and their constructive feedback on earlier on the IAB’s PASS survey. Consent of respondents was assured by the makers versions of our article. of the survey. As for the expert interviews performed in the framework of our specific project, the interview partners were informed about rules and Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 23 of 24 9 procedures of privacy protection, upon which their consent to participate was survey experiment among employers. Soc. Forces 99(2), 648–671 (2020). obtained.https:// doi. org/ 10. 1093/ sf/ soz154 Deeke, A., Baas, M.: Abbau oder Reproduktion von Ungleichheit? – Erträge der Consent for publication beruflichen Weiterbildung arbeitsloser Migranten. Sozialer Fortschritt The article mainly uses anonymized and aggregated secondary data, drawing 62(1), 23–32 (2013) on the IAB’s PASS survey. As for the three extracts from expert interviews, Dietz, M., Osiander, C.: Labor market integration of refugees: the role of which are presented in a highly anonymized way in the text, consent to pub‑ caseworkers. In: Paper presented at the 31st SASE annual conference, The lish was obtained from the interview partners. New School, New York, June 27–29, 2019 (2019) Dietz, M., Osiander, C., Stobbe, H.: Arbeitsmarktintegration von Geflüchteten Competing interests aus Sicht der Vermittler. Online‑Befragung in Arbeitsagenturen und The authors declare that they have no competing interests. Jobcentern. IAB‑Kurzbericht 25|2018. Institut für Arbeitsmarkt ‑ und Berufsforschung, Nürnberg (2018) Received: 4 June 2021 Accepted: 21 June 2022 Esser, H.: Sprache und Integration. Die sozialen Bedingungen und Folgen des Spracherwerbs von Migranten. Campus, Frankfurt/New York (2006) Fendel, T.: Die Arbeitsmarktintegration geflüchteter Frauen. WISO Direkt 02/2019. Friedrich‑Ebert ‑stiftung, Bonn (2019) Freier, R., Steiner, V.: ‘Marginal employment’: stepping stone or dead end? Eval‑ uating the German experience. J. Labor Mark. Res. 2/3, 223–243 (2008) References Goethe Institut.: Integation course. https:// www. goethe. de/ prj/ mwd/ en/ idl/ Adamuti‑ Trache, M.: First 4 years in Canada: post‑secondary education path‑ deu/ iku. html (2020) ways of highly educated immigrants. J. Int. Migr. Integr. 12, 61–83 (2011). Grohs, S., Adam, C., Knill, C.: Are some citizens more equal than others? Evi‑ https:// doi. org/ 10. 1007/ s12134‑ 010‑ 0164‑2 dence from a field experiment. Public Adm. Rev. 76(1), 155–164 (2016) Arbeitsagentur.: Maßnahmen bei einem Arbeitgeber (MAG) nach § 45 SGB III Gundert, S., Kosyakova, Y., Fendel, T.: Migrantinnen und Migranten am im Rahmen der Maßnahmen zur Aktivierung und beruflichen Einglieder ‑ deutschen Arbeitsmarkt. Qualität der Arbeitsplätze als wichtiger ung. Fachliche Weisungen zur Durchführung des § 45 SGB III. https:// Gradmesser einer gelungenen integration. IAB‑Kurzbericht 25. Institut für www. arbei tsage ntur. de/ datei/ ba146 324. pdf (2019) Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2020) BA, B. f. A.: Aktivierung und beruflichen Eingliederung (MAbE) nach § 45 SGB III Hashmi Khan, A.: Post‑migration investment in education by immigrants in the (Jahreszahlen). Nürnberg (2019a) United States. Q. Rev. Econ. Finance 37(Special Issue), 285–313 (1997) BA, B. f. A.: Arbeitslosigkeit und Unterbeschäftigung. Deutschland und Länder Hemker, J., Rink, A.: Multiple dimensions of bureaucratic discrimination: 2018. Nürnberg (2019b) evidence from German welfare offices. Am. J. Polit. Sci. 61(64), 786–803 BA, B. f. A.: Förderung der beruflichen Weiterbildung (Monatszahlen). (2017) Deutschland, Länder und Regionaldirektionen. Nürnberg (2019c) Hohmeyer, K., Wolff, J.: Selektivität von Ein‑Euro ‑ Job‑Ankündigungen. IAB‑ BA, B. f. A.: Arbeitslose nach Rechtskreisen (Jahreszahlen). Deutschland und Forschungsbericht 2|2015. Institut für Arbeitsmarkt‑ und Berufsforschung, Länder. Berichtsjahr 2019. Nürnberg (2020) Nürnberg (2015) BA, B. f. A.: Arbeitsmarktpolitische Instrumente (Zeitreihe Jahreszahlen) Holzinger, C.: ‘We don’t worry that much about language’: street‑level 2000–2019. Nürnberg (2021) bureaucracy in the context of linguistic diversity. J. Ethn. Migr. Stud. 46(9), Bähr, S., Beste, J., Wenzig, C.: Arbeitsmarktintegration von Geflüchteten im 1792–1808 (2020). https:// doi. org/ 10. 1080/ 13691 83X. 2019. 16103 65 SGB IIHemmnisse abbauen und Potenziale nutzen. IAB‑Kurzbericht Integrationsbeauftragte, B. d. B. f. M., Flüchtlinge und Integration.: Deutschland 23/2017. Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg kann Integration: Potenziale fördern, Integration fordern, Zusammenhalt (2017) stärken.12. Bericht der Beauftragten der Bundesregierung für Migration, BAMF, B. f. M. u. F.: Integrationskursgeschäftsstatistik für das Jahr 2019 (Bun‑ Flüchtlinge und Integration in Deutschland. Berlin (2019) desländer). Nürnberg (2020) Jesske, B., Schulz, S.: Methodenbericht Panel Arbeitsmarkt und Soziale Becker, R.: Economic change and continuous vocational training in the work Sicherung PASS 11. Erhebungswelle – 2017. 01/2009 FDZ‑Methodenre ‑ history: a longitudinal multilevel analysis of the employees’ participa‑ port 13/2018. Forschungsdatenzentrum (FDZ) des Bundesagentur für tion in further training and the effects on their occupational careers Arbeit im Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2018) in Germany, 1970–2008. Empir. Res. Vocat. Educ. Train. 11(1), 4 (2019). Jesske, B., Schulz, S., Quandt, S.: Methodenbericht Panel Arbeitsmarkt und Sozi‑ https:// doi. org/ 10. 1186/ s40461‑ 019‑ 0079‑x ale Sicherung PASS‑Erhebungswelle 12 – 2018 (Haupterhebung). Institut Berg, M., et al.: Codebook and documentation of the panel study “Labour für angewandte Sozialwissenschaft GmbH (infas), Bonn (2019) Market and Social Security” (PASS)—Datenreport wave 12. FDZ‑ Just Landed.: Job applications. How to apply for a job in Germany. https:// Datenreport, 09/2019 (en). Research Data Centre (FDZ ) of the Federal www. justl anded. com/ engli sh/ Germa ny/ Germa ny‑ Guide/ Jobs/ Job‑ appli Employment Agency (BA) in the Institute for Employment Research catio ns (2020) (IAB), Nürnberg (2020) Kalter, F., Granato, N.: Migration und ethnische Ungleichheit auf dem Arbeits‑ Bonin, H., Boockmann, B., Brändle, T., Bredtmann, J., Brussig, M., Dorner, markt. In: Abraham, M., Hinz, T. (eds.) Arbeitsmarktsoziologie: Probleme, M., et al.: Begleitevaluation der arbeitsmarktpolitischen Integrations‑ Theorien, empirische Befunde, pp. 355–387. Springer Fachmedien maßnahmen für Geflüchtete. Zweiter Zwischenbericht. Forschungsber ‑ Wiesbaden, Wiesbaden (2018) icht 546. Bundesministerium für Arbeit und Soziales, Berlin (2020) Kanas, A., van Tubergen, F.: The impact of origin and host country schooling Boockmann, B., Scheu, T.: Integration der Geflüchteten in den Arbeitsmarkt: on the economic performance of immigrants. Soc. Forces 88(2), 893–915 Ziele, Strategien und Hemmnisse aus Sicht der Jobcenter. Zeitschrift Für (2009). https:// doi. org/ 10. 1353/ sof.0. 0269 Sozialreform 65(4), 393–426 (2019) Kasrin, Z., Stockinger, B., Tübbicke, S.: Aktive Arbeitsmarktpolitik für arbeit‑ Brücker, H., Gundacker, L., Hauptmann, A., Jaschke, Ph.: Die Arbeitsmark‑ slose Geflüchtete im SGB II. Der Großteil der Maßnahmen erhöht den twirkungen der COVID‑19‑Pandemie auf Geflüchtete und andere Arbeitsmarkterfolg. IAB‑Kurzbericht 7|2021. Institut für Arbeitsmarkt ‑ und Migrantinnen und Migranten. IABForschungsbericht, 05/2021. Institut für Berufsforschung, Nürnberg (2021) Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2021) Knuth, M.: "Transformative Arbeitsmarktpolitik. Herausforderungen der Arbe‑ Bundesregierung.: Lebenslagen in Deutschland. Der 6. Armuts‑ und Reich‑ itsmarktpolitik unter den Bedingungen der, konfluenten Digitalisierung’". tumsbericht der Bundesregierung. Berlin (2021) Working Paper Forschungsförderung. Berlin: Forschungsstelle "Arbeit der Cross, K.P.: Adults as Learners. Increasing Participation and Facilitating Learning. Zukunft" der Hans Böckler Stiftung (2021) Jossey‑Bass, San Francisco (1981) Kogan, I.: Last hired, first fired? The unemployment dynamics of male immi‑ Damelang, A., Kosyakova, Y.: To work or to study? Postmigration educational grants in Germany. Eur. Sociol. Rev. 20(5), 445–461 (2005) investments of adult refugees in Germany—evidence from a choice Koopmans, R., Veit, S., Yemane, R.: Taste or statistics? A correspondence study experiment. Res. Comp. Int. Educ. 73(online‑first), 1–12 (2021) of ethnic, racial and religious labour‑market discrimination in Germany. Damelang, A., Ebensperger, S., Stumpf, F.: Foreign credential recognition and Ethn. Racial Stud. 42(16), 233–252 (2019) immigrants’ chances of being hired for skilled jobs—evidence from a 9 Page 24 of 24 R. Lehwess‑Litzmann , J. Söhn Kosyakova, Y.: Mehr als nur ein Job: die qualitative Dimension der Integration Yankova, K.: Der Selektionsprozess in Maßnahmen der aktiven Arbeitsmarktpo‑ in Arbeit von Geflüchteten in Deutschland. WISO DISKURS 09/2020. litik. Eine explorative Untersuchung für die deutsche Arbeitsvermittlung. Friedrich‑Ebert ‑Stiftung, Bonn (2020) IAB‑ discussion paper 11/2010. Institut für Arbeitsmarkt‑ und Berufs‑ Kruppe, T.: Bildungsgutscheine in der aktiven Arbeitsmarktpolitik. Sozialer forschung, Nürnberg (2010) Fortschritt 58(1), 9–19 (2009) Lancee, B., Bol, T.: The transferability of skills and degrees: why the place of Publisher’s Note education affects immigrant earnings. Soc. Forces 96(2), 691–716 (2017). Springer Nature remains neutral with regard to jurisdictional claims in pub‑ https:// doi. org/ 10. 1093/ sf/ sox058 lished maps and institutional affiliations. Leber, U., Möller, I.: Weiterbildungsbeteiligung Ausgewählter Personengrup‑ pen. Schmollers Jahr. 128(3), 405–429 (2008) Lehwess‑Litzmann, R.: Is labour market policy heading for social investment in European countries? An analysis of spending trends between 2006 and 2015. Zeitschrift Für Sozialreform 64(3), 333–362 (2018). https:// doi. org/ 10. 1515/ zsr‑ 2018‑ 0018 Liechti, F., Fossati, F., Bonoli, G., Auer, D.: The signaling value of labor market programs. Eur. Sociol. Rev. 33(257–274) (2017) Linden, P.: Wie viel Geld ist angemessen? Eine Vignettenstudie zur Akzeptanz von Sanktionen im SGB II. WSI Mittelungen (2021). https:// doi. org/ 10. 5771/ 0342‑ 300X‑ 2021‑6‑ 454 Lipsky, M.: Street‑level Bureaucrats. The Dilemmas of Individuals in Public Services, Updated Russell Sage Foundation, New York (2010) Maynard‑Moody, S., Musheno, M.: Cops, Teachers, Counselors: Stories From the Front Lines of Public Service. University of Michigan Press, Ann Arbor (2003) Möller, J., Walwei, U.: Handbuch Arbeitsmarkt. Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2009) Nohl, A.‑M., Schittenhelm, K., Schmidtke, O., Weiß, A. (eds.): Work in Transition. Cultural Capital and Highly Skilled Migrants’ Passages into the Labour Market. University of Toronto Press, Toronto (2014) Osiander, C.: Zugang zu beruflicher Weiterbildung ‑ Rechtlicher Rahmen, Teilnahme an und Wirkung von beruflicher Weiterbildung, Weiter ‑ bildungshemmnisse. In: Goth, G.G., Kretschmer, S., Pfeiffer, I. (eds.) Bildungsinnovationen für nicht formal Qualifizierte. Zielgruppengerechte Weiterbildungssettings in der Bildungspraxis, pp. 61–88. wbv Media, Bielefeld (2019) Osikominu, A.: Eine Analyse der Teilnehmerselektion in die berufliche Weiterbildung auf Basis der Integrierten Erwerbsbiografien (IEB). IAB‑ Forschungsbericht 23/2005. Nürnberg Institut für Arbeitsmarkt‑ und Berufsforschung [Institute for Employment Research, Nuremberg, Germany] (2005) Sauer, M.: Was Migrant/innen mit Jobcentern erlebt haben: Fallstudien in Bedarfsgemeinschaften mit Migrationshintergrund. In: Knuth, M. (ed.) Arbeitsmarktintegration und Integrationspolitik ‑ zur notwendigen Verknüpfung zweier Politikfelder: eine Untersuchung über SGB II‑ Leistungsbeziehende mit Migrationshintergrund, pp. 151–160. Nomos, Baden‑Baden (2010) Schneider, J., Fischer, M., Kovacheva, V.: Migrants in the Job Centre Qualitative Findings on Migrants’ Experiences with Public Employment Support Ser‑ vices in Germany. Hamburg Institute of International Economics (HWWI), Hamburg (2008) Sinning, M.G.: Determinants of savings and remittances: empirical evidence from immigrants to Germany. Rev. Econ. Househ. 9(1), 45–67 (2011). https:// doi. org/ 10. 1007/ s11150‑ 009‑ 9082‑5 Söhn, J.: Initial employment pathways of immigrants in Germany. Why legal contexts of reception matter—an analysis of life‑ course data. Transf. Eur. Rev. Labour Res. 25(1), 43–62 (2019). https:// doi. org/ 10. 1177/ 10242 58918 StaBa, S. B.: Foreign population by duration of stay and selected citizenships. https:// www. desta tis. de/ EN/ Themes/ Socie ty‑ Envir onment/ Popul ation/ Migra tion‑ Integ ration/ Tables/ forei gner‑ durat ion‑ of‑ stay. html (2021) Steinhardt, M.F.: The impact of naturalisation on immigran labour market integration in Germany and Switzerland. In: OECD (ed.) Naturalisation: A Passport for the Better Integration of Immigrants, pp. 145–151. OECD Publishing, Paris (2011) van Tubergen, F., De Werfhorst, H.V.: Postimmigration investments in educa‑ tion: a study of immigrants in the Netherlands. Demography 44(4), 285–313 (2007) Veit, S., Thijsen, L.: Almost identical but still treated differently: hiring discrimi‑ nation against foreign‑born and domestic‑born minorities. J. Ethn. Migr. Stud. (2019). https:// doi. org/ 10. 1080/ 13691 83X. 2019. 16228 25 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal for Labour Market Research Springer Journals

Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a comparative analysis based on PASS survey data

Loading next page...
 
/lp/springer-journals/jobcenters-strategies-to-promoting-the-inclusion-of-immigrant-and-kCySP8TPXh
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
ISSN
1614-3485
eISSN
2510-5027
DOI
10.1186/s12651-022-00313-8
Publisher site
See Article on Publisher Site

Abstract

This paper comparatively analyzes strategies of German Jobcenters to bring native and immigrant job seekers into employment. It focuses on clients who receive means‑tested basic income for the unemployed, based on data from the Panel Study Labour Market and Social Security (PASS) from year 2015 to 2020. By way of logistic regression, the study identifies the impact of being an immigrant on the clients’ probability of reporting different kinds of offers like job referrals or courses, controlling for a number of other influential factors. The study also looks deeper into the effects of immigrant ‑specific attributes, such as heterogeneous German language skills. We found that the likelihood of offers by Jobcenters largely depends on the amount of time since immigration. Recent immigrants have the lowest chance of reporting most of the studied measures of active labor market policies. For immigrants having stayed more than 4 years in Germany, however, we do not find a disadvantage, and some measures out of Jobcenters’ toolbox are even more often offered to the longer ‑settled immigrants than to native clients. A possible explanation for the mod‑ erately under‑average support of recent immigrants in terms of Jobcenters’ measures could be an institutional focus on improving German language skills prior to approaching the labor market. Keywords: Immigration, Integration, Jobcenter, Public employment services, Active labor market policy, ALMP, Training, Labor market, Welfare state JEL Classification: H41, I38, J08, J15, J68, J61 1 Introduction: Jobcenters as crucial actors employment is one of the most important steps, maybe for the integration of immigrants? the most important one. Yet, immigrants face a higher Germany has witnessed successive waves of immigra- risk of becoming and staying unemployed than natives tion since the World War II, in recent years especially (Integrationsbeauftragte 2019, p. 203; Kogan 2005). Fur- from Eastern and Southern Europe in the context of the thermore, with the exception of recruited immigrants enlargement of the European Union (EU) as well as from who enter Germany with an employment contract in Syria, Iraq, Afghanistan, and most recently Ukraine, due hand, most immigrants do not have a job upon arrival to wars and humanitarian crises. For successful integra- (Söhn 2019, p. 51). They are significantly overrepre - tion of immigrants in German society, finding (decent) sented among the unemployed in Germany, with con- sequences such as a high poverty risk (Bundesregierung 2021, pp. 50, 129, 210). Despite Germany’s booming *Correspondence: rene.lehwess@sofi.uni‑ goettingen.de economy, their relative share among job-seeking recipi- ents of basic-income support for the unemployed (ALG- Soziologisches Forschungsinstitut Göttingen (SOFI) e.V., an der Georg‑August ‑ Universität, Friedländer Weg 31, 37085 Göttingen, Germany II [Arbeitslosengeld II]) increased between 2015 and 2020 © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. 9 Page 2 of 24 R. Lehwess‑Litzmann , J. Söhn (BA 2020 and older versions), i.e., our episode of inves- and 2019, the average across this period being 28.4 per- tigation. Hence, public employment services (PES) is an cent; thus only one percentage point below the average institution of paramount importance for many immi- share of immigrants in all job-seeking benefit recipients grants’ labor market integration. (29.3 percent) (BA 2019c, authors’ calculation). As to The PES do much more than granting unemployment other activation measures, the share of foreigners var- benefits; they also monitor job offers and deploy a wide ies little (BA 2019a, p. Tab 1.1). Older research based on range of active labor market policy (ALMP) measures. quantitative micro data is inconclusive regarding immi- For example, they can offer subsidized courses, which grants’ access to PES-subsidized further training, show- may, for financial reasons, be the only chance to access ing a negative effect of foreign or non-EU citizenship, further education for some groups of clients, e.g., long- but a positive one of a short stay in Germany. A recent term unemployed individuals receiving means-tested vignette survey among caseworkers in German Jobcent- benefits. Arguably, the PES are even more important ers, focused on refugee clients, showed that whether for job-seeking immigrants than they are for native job caseworkers recommended employment, occupational seekers. This is due to immigrants’ lack of country-spe - training or other programs as the first option depends on cific knowledge, networks, vocational certificates, and of the socio-economic background and family situation of other resources. fictitious clients (Dietz and Osiander 2019). Boockmann When individuals register for unemployment benefits, and Scheu (2019, pp. 408, 412) find that Jobcenter staff the main objective of both social law and PES is to help consider an improvement of language skills and qualifica - them become financially independent again by bringing tion of refugees as the prerequisite for finding qualified them back into work, and this very often corresponds to work, and thus for sustainably ending benefit receipt. the clients’ aim as well. Sometimes, however, the client, However, such a long-term perspective is not system- the PES, or both, regard immediate re-employment as atically implemented, as both clients and case managers unrealistic or even undesirable under the overall circum- see also downsides of postponing contact with the labor stances. Instead, some clients may require more knowl- market by lengthy “chains of subsidized measures” (ibid.). edge of how to apply for jobs and present themselves Many refugee clients urgently wish to start working, even in job interviews. Others want to or have to profoundly in ‘bad’ jobs, and some case managers warn that receiv- re-orientate their professional goals; this may be due to ing social benefits for several years might make clients health issues or long episodes of unpaid care work in lose sight of the aim of employment (ibid., pp. 412–413). the family. Still, others consider attaining new qualifica - Boockmann and Scheu (2019, pp. 412–415) recommend tion as a necessary intermediary step to reach the final Jobcenters to combine education and training on the one goal of labor market integration or to gain employment hand with labor-market programs on the other (ibid., pp. on a higher level of qualification and income than pre - 413 et seq.). According to qualitative findings by Schnei - vious competences and qualification would allow. In all der et  al. (2008, p. 30), discriminating treatment by the those instances, the PES can offer help using their ALMP PES seems to be a rare case, while Sauer (2010) reports toolbox. such experience for some unemployed immigrants she This article investigates whether, and to what extent, interviewed, but again without comparison with native PES in Germany treat immigrant and native clients dif- clients. Experimental correspondence tests, however, ferently with regard to efforts of placing them in regular do show a lower quality of German municipalities’ and employment and channeling them into various kinds of Jobcenters’ responses to email requests of ‘Turks’ versus ALMP programs. Are there different patterns of what ‘Germans’ (Grohs et  al. 2016; Hemker and Rink 2017). PES caseworkers offer to immigrated job seekers com - Aggregate administrative statistics are descriptive only pared to German-born ones? and hardly differentiate participants according to social The percentage of non-German citizens among all or immigrant-specific characteristics (usually only gen - participants of PES-subsidized further vocational train- der, age, non-German citizenship or sometimes refugee ing increased from 19.7 to 34.1 percent between 2015 status). Finally, there are good reasons for investigating the foreign-born in a broad sense; while recent refugees have gained quite some attention in labor market research Statistics of the Federal Employment Agency (BA [Bundesagentur für Arbeit]) show that in the period from 2015 to 2019, 40.0 percent of new par- (Bonin et  al. 2020; Boockmann and Scheu 2019; Dietz ticipants in ALMP attended measures for activation and professional integra- et  al. 2018; Fendel 2019; Kasrin et  al. 2021; Kosyakova tion (854,990 thousand individuals on the yearly average), among which, e.g., 2020) in the context of the large influx of Syrian refu - 5.8 percent participated in further vocational training and 10.3 percent in job- creation schemes. 34.1 percent received classic referrals to job openings (BA gees in 2015/2017. In the long run, a comprehensive look 2021, restricted to Sozialgesetzbuch [SGB] II, authors’ calculation). on immigrants in general is called for. Quantitatively Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 3 of 24 9 important groups like EU immigrants from Southern immigrant-specific variables. Critically reflecting on and Eastern Europe have and will continue to make up a our database, we will discuss methodological challenges substantial share of immigrants. Also, naturalized immi- of the PASS-items on “vocational training and other grants disappear in administrate statistics but deserve courses” and present robustness tests (Sect.  4.3). The being included, for instance, as an immigrant group con- quantitative analysis is complemented by some insights trasting with recently arrived ones. Furthermore, as our from selected expert interviews with caseworkers and interest lies on ALMP opportunities given by Jobcent- middle-management employees of Jobcenters, which ers to their clients, rather than actual usage thereof (an were carried out in the context of a larger research pro- outcome in its own right), we need to investigate offers ject. The article finishes with a summary of the main received. Even though an offer does not equal its accept - findings and further discussions. ance, receiving an offer is still the sine qua non for being allowed to participate in a program subsidized by the 2 Conceptual considerations and previous PES. Overall, to our knowledge, there are no representa- empirical insights tive and multivariate studies on immigrant-native dif- German law distinguishes jobless individuals who obtain ferences or intragroup comparisons among immigrants unemployment insurance benefits based on prior contri - regarding Jobcenters’ strategies for individual job-seeking butions (Social Code [SGB] III) from jobless persons who clients. receive means-tested basic-income support (SGB II). Concretely, our analysis pursues the following research The latter group includes both (long-term) unemployed questions: Which ALMP measures do Jobcenter staff persons who have exhausted their entitlement to unem- offer to unemployed recipients of basic-income transfers, ployment insurance benefits and job seekers who did not depending on whether they are natives, recently arrived contribute to the insurance (long enough); among them and longer-settled immigrants? Does migration experi- recently arrived refugees who received political asylum ence play a decisive role, once other influential factors or another form of humanitarian legal status. In the (like age or family status) are considered? How are addi- German two-tiered system, the so-called Jobcenters are tional immigrant-specific factors like legal status, region the PES responsible for all job seekers without insurance of origin and German language skills linked to Jobcenters’ benefits, thus for all who receive tax-funded welfare. Our offers? empirical analysis is dedicated to Jobcenter clientele. Our contribution on German Jobcenters analyzes data The priority for Jobcenters is to help clients end or from the Panel Study Labour Market and Social Secu- reduce benefit receipt, as this is what social law requires rity (PASS), a survey administered by the Institute for (§ 1 SGB II). For most clients, this means that (non- Employment Research (IAB) of the German Federal subsidized) employment is the main goal, be it for the Employment Agency (BA). This data is nationally repre - first time or after a period of joblessness. Benefit recipi - sentative for basic-income recipients. Respondents regis- ents generally are obliged to look for potential employ- tered as job seekers during at least one of the waves, from ers; Jobcenters may also refer clients to open positions 2015 to 2020, answered the survey question whether the and encourage them to apply. Jobcenters also pursue PES referred them to specific job openings regarding the aim of “maintaining, improving, or re-establishing regular or marginal employment or offered them various the employability” of the client (ibid., authors’ transla- ALMP measures. tion). To this end, there is a range of ALMP measures Section  2 elaborates on the hypotheses based on con- which Jobcenters can implement (based on § 16 SGB II). ceptual considerations and previous research in our field. In this respect, and in contrast to direct access to non- Section  3 presents the database, operationalization, and subsidized jobs, caseworkers in Jobcenters act as crucial methods applied. After a descriptive analysis (Sect.  4.1), gate keepers. In accordance with the administration’s we will introduce Jobcenters’ offers as dependent vari - general goal of economical and effective spending, offer - ables in multivariate regression models (Sect.  4.2). In ing an ALMP program is a discretionary decision which a first step, the study analyzes natives and immigrants caseworkers as “street-level bureaucrats” (Lipsky 2010) together, in a second step, it looks at immigrants only, can opt for. They consider the degree of their clients’ as this permits extending the model by additional, Footnote on research project and funding institution as well as thanks to reviewers for their feedback. Citizens from EU-countries make up 69.4 percent of the foreign popula- tion in 2020, with 42.8, 50.9, and 43.7 percent among those with a duration of Immigrants with the precarious status of asylum seeker and with extra- stay of less than 1 year, between one and under 4 years, and between four and dition only temporarily suspended (Duldung) are the clientele of Arbe- under eight, respectively. The equivalent percentages for Syrians are 7.2, 6.3, itsagenturen (SGB III) (see Boockmann and Scheu 2019, p. 404) and not and 9.5 (see StaBa 2021, authors’ calculation). considered here. 9 Page 4 of 24 R. Lehwess‑Litzmann , J. Söhn employability, the likelihood of successfully finishing an may communicate their preferences to their caseworker educational program, with respect to the clients’ moti- and possibly influence the latter’s decision. On the one vation and learning ability, as well the probability of the hand, given the asymmetric power relation, an offer newly acquired knowledge helping to find an appropri - by the Jobcenter is a necessary condition of the clients’ ate job thereafter (Yankov 2010, pp. 8, 24–25). Having participation in an ALMP program. On the other hand, to respect their organizations’ budget restrictions, case- substantial measures like 2-year vocational re-training or workers try to identify those clients for whom a program professional coaching are too expensive to force clients seems most promising. Vocational (re-)training programs to attend. They tend to be a privilege that clients have to and long-term subsidized employment schemes are fight for and be knowledgeable and confident enough to among the most substantial and costly offers Jobcenters do so (regarding immigrants see Schneider et al. 2008, p. can make. Cheaper short-time programs do not involve 18). potential lock-in effects (standing in the way of imme - In some cases, clients’ rights limit caseworkers’ discre- diate transition into employment), which are typical for tion over ALMP measures. Participating in “integration longer programs (see the evaluation of ALMP programs’ courses”, regulated in the Residence Act (Aufenthaltsge- effects on refugee clients by Kasrin et al. 2021, p. 3). setz, §§ 43 ff.), is both a right and an obligation for certain Clients themselves also have to consider pros and cons groups of, primarily, newly arrived immigrants. These of looking for (non-subsidized) work versus attending courses mainly consist of German language classes plus (longer) programs while receiving the rather meager civic education. EU nationals may participate, too, but basic-income support. Course attendance and job search are neither obliged to nor have the right to do so. do not utterly exclude one another, but each restricts time Overall, the question whether Jobcenters make dif- and mental energy that remain for the other option. Indi- ferent offers to immigrant and native basic-income viduals more or less consciously weigh possible future recipients comes down to caseworkers’ perceptions occupational advantages of the qualification, knowledge, (Maynard-Moody and Musheno 2003) of these two or skills to be gained against their self-perceived learn- groups with regard to the necessity, needs, motivations, ing capacity and motivation, individual or family-related likelihood of success, and maybe also personal ‘worthi- obstacles or resources as well the opportunity costs of ness’, to be placed into regular employment or into some not having income from working (Becker 2019, pp. 3–5; kind of ALMP measure. In the following, we will develop Cross 1981, pp. 98, 116–117). An experimental study some hypotheses based on both theoretical arguments embedded in a survey of refugees showed that given the and related previous research. hypothetical scenario of having been offered a low paid job and vocational training, respondents were more likely 2.1 Referral to open job positions to choose the latter; the higher the future earnings after First, this study looks at the goal of direct employment the specific vocational training, the easier the exam was take-up. As for clients’ preferences, in terms of a disin- said to be, and the higher the apprenticeship wage during centive of looking for work, immigrants might consider training (Damelang and Kosyakova 2021, pp. 6–8). Again the high amount of German welfare benefits compared with regard to refugees, Boockmann and Scheu (2019, with salaries in their home country in absolute terms. p. 409) highlight the high motivation of many clients to However, living expenses in Germany is also higher than succeed in Germany, but also the potential demotivation before migration. Immigrants might actually be more by the lengthy process of overcoming various administra- eager than natives to work immediately, as many of them tive and qualification-related hurdles standing in the way feel obliged to send money to relatives back in their of job uptake. Based on interviews with Jobcenter staff, country of origin (on remittances see Sinning 2011). In they find that the wish for economic independence from addition, residence law makes their right to stay in Ger- Jobcenters can “reduce the readiness to go through quali- many beyond their fixed-term residence permit depend fication measures which would allow them enter higher on continuous pension insurance contributions (Knuth segments of the labor market in the long term” (ibid, p. 409, authors’ translation). Hence, offering and accepting ALMP programs are no isolated decisions; it rather takes place in a context where Also, short-term measures can be strategically used by the PES to verify clients’ readiness to co-operate, the absence of which can lead to sanctions, employment is the major competing alternative. Clients which case workers are more inclined to apply to ethnic-minority clients (Lin- den 2021). Non-EU family immigrants and asylum seekers recognized as refugees are For Switzerland, Liechti et  al. (2017, p. 264) show that foreigners profit no currently the largest groups. Altogether, 238 thousand immigrants attended less from ALMP programs than natives regarding their ensuing job prospects. integration courses on the yearly average between 2015 and 2019 (BAMF 2020, and previous editions; authors’ calculation). Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 5 of 24 9 2021, p. 56). In our expert interviews with Job Center with the labor market, Jobcenters might suggest applying staff, one employee put it this way: for marginal employment more often to immigrant clients (H #2). “iTh s happens quite often, when it comes to work - ing, the [migrant] clients say ‘I want work, no matter 2.2 Support for the job‑search process what kind, I will do anything’ … also the academi- Getting insight and support on how to write applications cally trained ones.” (Interview I) and how to present oneself in a job interview should be If Jobcenters shared this priority, as a theoreti- particularly relevant for newly arrived immigrants as cal assumption not tested by our data, they could refer there are most likely small or large cultural differences immigrants to open positions more often than natives (depending on the country of origin) in what is regarded (Hypothesis [H] #1a). In addition, as these newcomers are as an appropriate form. Hence, we expect unemployed less familiar with job search on the German labor mar- immigrants to be offered support with applications and ket than natives, caseworkers might not want to rely on job search more often than their native peers (H #3) (For immigrant clients’ own initiative to find employment, but our counter hypothesis regarding any type of ALMP pro- explicitly refer them to open positions. gram, see the end of this section). In contrast to this potential scenario, Jobcenters (and Looking for a job can involve monetary expenses, clients) could be aware of immigrants’ high hurdles to especially travel costs to go to a job interview. Jobcent- find employment and decide that these should be tack - ers routinely reimburse application or travel costs upon led before making efforts of placing clients into employ - job seekers’ request. Thus, such an offer by the Jobcenter ment. After all, many immigrants (and in particular hinges on the prior decision by an employer to put the recently-arrived ones) struggle with devaluation of their person on a shortlist and indicates a high level of employ- foreign qualification and/or work experience (Boock - ability. Given the obstacles immigrant job seekers face mann and Scheu 2019, p. 407; Damelang et al. 2020; Nohl on the labor market, we expect reimbursement of appli- et al. 2014, p. 24ff.), insufficient knowledge of formal and cation or travel costs to be offered less often to immigrant informal norms in the process of job search in Germany, than to native Jobcenter clients (H #4). actual or assumed German language deficits (Esser 2006, p. 400ff.), and ethnic discrimination in the German labor 2.3 PES‑subsidized (self‑) employment market (e.g., Veit and Thijsen 2019), resulting in both Some ALMP measures support employment more above-average unemployment rates, as discussed above, directly than the programs mentioned so far. They ‘cre - and below-average work quality (Gundert et  al. 2020). ate’ work experience and hence a potential bond with Caseworkers may refrain from making clients apply in an employer. In the case of a so-called “measure with the first place when success is improbable. If casework - an employer” (Maßnahme bei einem Arbeitgeber), job ers judged immigrants less employable than native job seekers work for an employer during up to 6  weeks (in seekers, they should refer immigrant clients to open job exceptional cases up to 12 weeks), who receives financial positions in regular employment less often than native (H subsidies by the PES during that period (Arbeitsagentur #1b). 2019) and ideally hires the job seeker once the program Jobcenters may also refer unemployed clients to mar- runs out. According to Kasrin et  al. (2021, p. 3), partici- ginal employment (“Mini-Jobs” in German), i.e., employ- pating companies prefer job seekers with a comparatively ment with low earnings (up 450 Euros per month), low high level of employability (Dietz et  al. 2018). Indeed, working hours, limited social insurance, and slim chances such programs bring unemployed (male) refugees most to lead to regular employment (Freier and Steiner 2008). quickly into regular non-subsidized employment com- Similar to ‘1-€-jobs’ (Arbeitsgelegenheiten) (Kasrin et  al. pared with other AMLP measures (Kasrin et al. 2021, pp. 2021, p. 3), Jobcenters might consider marginal employ- 5, with further training as the second-best option). With ment as an option mainly reserved for ‘difficult’ long- respect to the immigrant-specific need to gain employ - term unemployed clients. The latter are often faced with ers’ trust and also regarding recent immigrants’ eager- multiple employability obstacles so that Jobcenters regard ness to work, the latter might be offered a program with regular employment as out of reach. Limited personal an employer or an internship more often than native job availability for work due to child care, for instance, can seekers (H #5), provided the client is seen as sufficiently be an alternative reason for marginal employment. All qualified. these obstacles can apply to both native and immigrant Jobcenter clients. However, given the above-mentioned immigrant-specific obstacles to employment and also We do not know of any research on such cultural differences, but websites the particular need to bring refugee clients into contact like that of Just Landed (2020) imply the respective need for information. 9 Page 6 of 24 R. Lehwess‑Litzmann , J. Söhn Financial support to become self-employed could, on et  al. 2021, p. 7, regarding refugees). These arguments the one hand, be an option for immigrants as they have feed into our hypothesis that if immigrants’ special need greater difficulties to find employment (Integrations - for further vocational training and the effectiveness beauftragte 2019, p. 203; Kogan 2005); on the other hand, of such programs were decisive—again, a theoretical successful foundation of one’s own business requires assumption grounded in previous insights but not tested knowledge about administrative procedures and markets here—Jobcenters should offer it to immigrants more often that recent immigrants tend not to have. Jobcenters offer than to natives (H #7a). The opposite expectation (H #7b) self-employment support too rarely to statistically test a is favored by the time factor of attending a longer program respective hypothesis with our data. like a 2-year vocational re-training course, which might turn this option prohibitive for immigrants who already 2.4 C ounselling on occupational goals and qualification spent numerous months in German-language classes pathways and now urgently want to earn their living rather than Many immigrants are in need of (re-)orientation regard- continue learning with only the modest welfare benefits ing their occupational future when faced with devalu- available. Refugees’ case managers at times fear adverse ation of their previous qualification and/or work effects of too-long (chains of ) measures, possibly lead - experience (Damelang et  al. 2020; Nohl et  al. 2014), or ing immigrants to become used to welfare transfers as a when they have never worked beyond their household or durable substitute for income from employment (Boock- the informal sector in their country of origin (Bonin et al. mann and Scheu 2019, p. 412). 2020, p. 73, regarding refugees). Courses which inform The arguments put forth so far theoretically indicate a participants about different occupational fields in Ger - higher need of (recent) immigrants for ALMP programs many or individual coaching could, as a first step, support than there is among native job seekers, and such needs such individuals in deciding whether to apply for jobs, could turn into actual choices on part of Jobcenters. apprenticeships, internships, vocational re-training or However, there are two general reasons to expect fewer courses leading to full recognition of their foreign qualifi - offers to immigrants: Jobcenters may, rightfully or not, cation. If such needs were paramount for the Jobcenters’ consider immigrants’ linguistic ability too low to success- decision, they should offer courses on occupational orien - fully follow a mainstream ALMP program in German. tation or respective vouchers more often to immigrant cli- Furthermore, immigrants could be less knowledgeable, ents than to native ones. (H #6). confident, or vocal about their preferences in favor of a suitable ALMP program in their communicative interac- 2.5 Further training tion with caseworkers (Holzinger 2020). Schneider et  al. Some immigrants have special needs for further voca- (2008, p. 30) report “situations in which immigrants feel tional training. First, insufficient educational infrastruc - disadvantaged and sense that their qualifications, com - ture, war, or discrimination in the country of origin could petencies and career plans cannot be met” by the PES have prevented them from realizing their educational (Sauer 2010, p. 157). Relatedly, field experiments revealed goals there. Second, insufficient transferability of immi - some discrimination of German municipal govern- grants’ foreign qualification can serve as a push factor to ments (Grohs et  al. 2016) and Jobcenters (Hemker and gain new or complementary qualification in the immigra - Rink 2017) regarding the response quality in reacting tion country in order to prevent long-term social down- to email-requests from persons with typical Turkish (or ward mobility (for Canada, the USA, Germany, and the Rumanian) names. Against this backdrop, the counter- Netherlands repectively see Adamuti-Trache 2011, pp. hypothesis of fewer offers to immigrants could hold for 75–76; Hashmi Khan 1997, p. 287; van Tubergen and De some or all of the above-mentioned ALMP measures. Werfhorst 2007, p. 885). In addition, caseworkers could be aware of the fact and communicate it to their clients 2.6 Immigrant‑specific influences that attending educational programs is indeed an effec - Taking a deeper look at migration-specific factors, we tive, though strenuous and time-consuming, strategy to presume a shorter duration of stay in Germany to lead to a raise long-term chances of finding qualified work (Deeke larger difference in the treatment by Jobcenters compared and Baas 2013, regarding immigrants’ employment after with that of native clients (H #8). It remains to be seen further training financed by German PES; Kanas and van whether this assumption holds once the level of German Tubergen 2009; Lancee and Bol 2017). Jobcenter-subsi- language skills is considered. Good German skills would dized further training is even more successful in raising make language classes superuou fl s and should make both formerly unemployed persons’ incomes than subsidized referrals to (regular and subsidized) employment and to employer-programs, while both participant groups fare ALMP courses held in German more likely (H #9). As far far better than job seekers receiving no measure (Kasrin as legal status and nationality are concerned, Jobcenters’ Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 7 of 24 9 referrals to job vacancies (and reimbursement of travel calculate the duration of stay as well as to identify natu- cost) could be indirectly influenced by employers’ pos - ralized immigrants. sible recruiting preference for white, non-Muslim and/ We primarily use the first of the two samples the PASS or European immigrants (Koopmans et al. 2019; Veit and consists of, i.e., the one representing households which Thijsen 2019), as well as for those with German citizen- receive basic-income support (the other one is sampled ship (Steinhardt 2011) or permanent residence status (H on the German population without further restrictions) #10). (Hohmeyer and Wolff 2015, p. 13). Our investigation The potential effects of migration status, hypothesized is limited to the year 2015 to 2020, as the item which is in this section, could be mediated by other personal and most crucial for our analysis is only included in the sur- structural factors known to influence Jobcenters’ deci - vey as of wave 9. Our target population consists of adults, sions more generally. Such potential composition effects aged 18 to 64, who were registered at a Jobcenter among could be based on, e.g., immigrants being younger on those looking for work and receiving basic-income sup- average than native job seekers, with higher age being port. The sample was restricted to survey participants negatively associated with participation in further train- registered as job seekers, as only they are asked about ing (Osiander 2019, p. 70). Furthermore, higher levels of Jobcenters’ offers during the last year. Furthermore, we education are a component of positive (self-)selection excluded secondary-school students as well as recipients into employment and (further) education (Kruppe 2009, who were not in regular contact with the Jobcenter: Such pp. 11, 14; Osiander 2019, p. 76), also among refugee contact is a precondition of a potential ALMP offer. How - clients (Dietz and Osiander 2019, p. table  3). Depend- ever, this restrictions disproportionally excludes female ent children in one’s household are a possible hurdle as refugees (Bähr et al. 2017). well (Kruppe 2009, p. 14). Finally, there are good reasons In our unbalanced panel, a minority of participants of for investigating the foreign-born in a broad sense: while the PASS survey fulfilled our sample criteria for more recent refugees have gained quite some scientific atten - than one wave in our observation period. We choose tion (e.g., Bonin et al. 2020; Boockmann and Scheu 2019; person-waves (instead of persons) as our unit of obser- Dietz et al. 2018; Fendel 2019; Kasrin et al. 2021; Kosya- vation because each year that a person fulfills the crite - kova 2020) following the large influx of Syrian refugees ria of receiving basic income and looking for a job, there in 2015/2017, a comprehensive look on immigrants in is a new opportunity for offers to be made by Jobcenters general is called for. Quantitatively important groups to the client. We thus have a hierarchical data structure, like EU immigrants from Southern and Eastern Europe with 8275 person-waves nested in 4954 interviewees. have and will continue to make up a substantial share of 61.8 percent of the observed persons appear in our sam- immigrants. Also, naturalized immigrants disappear in ple only once (Table 5). administrate statistics but deserve being included, e.g., We define immigrants as foreign-born having arrived as an immigrant group contrasting with recently arrived in Germany as adults. According to this definition, ones. The following sections shows our empirical study is 45.3 percent of persons (and 39.1 percent of person- hence adequately inclusive. waves) in the observed population are immigrants. We further distinguish immigrants by their duration of stay, 3 Methods: database, operationalization, strategy i.e., whether they have arrived in Germany up to 4 years of analysis ago or at least 5 years ago. Note that immigrants tend to 3.1 Data and sample be observed for a smaller number of survey waves than The study uses the Panel Study Labour Market and Social natives: 69.8 percent of immigrants and only 55.2 percent Security (PASS) (Berg et  al. 2020) as it contains a ques- of natives are part of our sample during 1 year only. tion on having received an offer of various ALMP pro - grams by the PES (see below for details). In addition, the 3.2 Items for target variables PASS (rather than administrative data) is well suited for For operationalizing Jobcenters’ offers, we draw on the this research as it contains important information nec- survey item “Since your household has obtained Unem- essary to study immigrant integration, namely the year ployment Benefit II [since our last interview, respec - of immigration for any foreign born, which allows to tively], have you ever been offered the following by the Jobcenter?” (PTK1701). There is a battery of items to Citizens from EU-countries make up 69.4 percent of the foreign population in 2020, with 42.8, 50.9, and 43.7 percent among those with a duration of stay Immigrants having arrived as minors are left out because schooling abroad of less than 1 year, between 1 and under 4 years, and between four and under is a major contributor to migration-related problems on the labor market eight, respectively. The equivalent percentages for Syrians are 7.2, 6.3, and 9.5 (Kalter and Granato 2018, on differences between first- and second-genera - (see StaBa 2021, authors’ calculation). tion immigrants). 9 Page 8 of 24 R. Lehwess‑Litzmann , J. Söhn choose from, multiple answers possible. In our analysis, thirds of the vouchers actually redeemed do so (BA we consider the following offers (IAB’s translation of 2019a, table 5, authors’ calculation). On the PES web- the survey items into English) and we connect them with site, courses which the vouchers are meant for refer the hypotheses in Sect. 2: to, e.g. assistance in job applications, self-marketing strategies for academically trained job seekers, immi- • A part-time or full-time job or an apprenticeship: Job- grant-specific courses supporting economic integra - centers refer clients to open job positions in the form tion, one-on-one application coaching, or business of regular employment, subject to social insurance English. (H #6-plus) contributions, and encourage them to apply. Appren- • Vocational training, retraining or a course: This offer ticeships refer to the German dual vocational train- includes both courses lasting a few months and full ing in which learning takes place in a company and in vocational re-training taking 2 years. (H #7) a vocational school alternately. Apprenticeships are • An integration course or another German course: also subject to social security contributions. While This item should exclusively apply to immigrant Job - the previous section formulated H #1a/b only com- center clients. It has been included in the survey only prises regular employment, its arguments also apply as of wave 10 (year 2016), so the case numbers are to apprenticeships in firms as far as caseworkers may smaller in our models which include this item. Ger- view clients as (not) well-enough prepared to com- man courses other than those which are part of the pete with other applicants. integration course are, e.g., classes for occupation- • Minor employment, e.g., a mini-job (H #2) specific German targeting more advanced language • Support for your applications, e.g., assistance with learners. We study this target variable only with the preparation and compilation of application docu- regard to the intra-group comparison among immi- ments (H #3). This item partly overlaps with the one grant clients. on vouchers, see below, as it can be implemented by private providers rather than by the PES. Caseworkers may offer several measures (or nothing at • Reimbursement of application costs or travel all) during one counselling appointment or in consecu- expenses: though this is something Jobcenters can tive ones. The survey, however, asks respondents whether offer (H #4), it actually mirrors the “success” of hav - the Jobcenter has offered them any kind of offers since ing been invited to a job interview. they became a job seeker or since the last time they took • A program with an employer or internship (H #5) part in the yearly survey, respectively. During that time, • An activation or placement voucher with which you more than one measure of the same type could have can choose a program yourself: Financed by the PES, been offered and the sequence of those offers remains these vouchers give access to services by licensed unknown. Hence, we measure whether a type of program private local providers of various kinds of programs. was offered at least once during the reference period This item is not optimal for operationalizing our according to the respondents’ memory and willingness to hypotheses as it refers to the way Jobcenter offer answer correctly. programs, rather than their content. Support with Another source of potential bias is that the system of applications, occupational orientation, or further labor market services also has a “memory”: a specific vocational training, of which only two are response offer may be made only once (in consequence, persons categories of their own, can be granted via vouchers. observed for a smaller number of years could tend to The respective hypotheses (H #3, #6, #7) connected have a higher chance of receiving an offer in a given year to these measures also apply to vouchers. Jobcenter of observation). Yet, offers could also become more prob - statistics tell us that 56 percent of vouchers pertain able the longer the client is in the situation of receiving to the policy goal of giving clients orientation about benefits and looking for a job. As a test of robustness employment and vocational training (Heranführen (4.3), we will do an extra analysis where persons are con- an Ausbildung und Arbeitsmarkt) and more than two sidered only once across the whole observation period. Our findings on shorter- vs. longer-settled immigrants could be biased in so far as some of the immigrants who We do not test a hypothesis on the item of financial support to become self- arrived earlier could have left the country in the mean- employed, as the number of positive cases is too small for a multivariate anal- time, leading to some kind of social selectivity of the ysis. For the same reason, we refrain from using the item “other offers”, which remaining longer-settled immigrants. Yet, this methodo- only 1.9 percent say to have received. Hence, the other categories in the PASS survey are almost exhaustive. logical problem cannot be overcome with available data. It remains unknown whether respondents also subsumed ‘1-€-jobs’ Regarding one item, it is necessary to consider that under this item, as the item battery does not explicitly mention this meas- questions are asked in a certain order in the context of ure. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 9 of 24 9 the CATI/CAPI interviews (Jesske and Schulz 2018, p. and Möller 2008, p. 418). The under-employment quota 68), in which interviewers usually read out loud battery (individuals registered as unemployed plus those enrolled items one after the other. In case of items that overlap in PES programs; BA 2019b) of the respective regional to a certain degree, the ordering can be consequential. state controls for local employment opportunities. It is thus unfortunate that the item Vocational training, Some predictor variables are inherently related to each retraining or a course is proposed to respondents before other, e.g. duration of stay and respondents’ age (Pear- that of Integration course or another German course: it is son’s rho = 0.53): given our definition of immigrants as likely that some of the immigrants who had been offered having arrived as adults, those with longer duration of a language course in the reference year reported “yes” stay cannot be very young adults at the time of the inter- already when they read or heard of the item on vocational view. We use only broad categories of years since arrival (re-)training and the undefined “course” mentioned here. in order to avoid empty cells and hence attenuate the Indeed, out of the 818 immigrants who reported the offer challenge of collinearity. Duration of stay is also linked of Vocational training, retraining or a course in waves 10 to region of origin due to successive migration waves to 13, 648 also reported to have been offered an integra - from different regions (e.g., most persons from the for - tion/German course. There is no way to find out in ret - mer USSR arrived in Germany between 20 and 5  years rospect how many of those with two positive answers before the survey interview). Due to naturalization being have actually been offered both kinds of courses, and granted after several years of residence and due to more how many were only offered an integration or language generous rules applying to German resettlers, regions of course. However, chapter  4.3 presents a robustness test origin are also strongly correlated with legal status (Cra- controlling for the described issue and further critical mér’s V = 0.58). All other independent variables’ corre- arguments. lations are inconspicuous. 3.3 P redictor and control variables3.4 Methods With regard to our hypotheses on immigrant-specific fea - Seeking to analyze several different but not mutually tures, the duration of stay in Germany (the time between exclusive outcomes, we estimate separate binary mod- immigration and the survey interview) is adequate to els for each dependent variable. The main interest of test H #8 and level of linguistic competence in German our analysis is on differences in “treatment” by Jobcent - (respondents’ subjective evaluation) regarding H #9. ers between groups of persons, that is natives and immi- Both legal status (temporary residence permit being the grants. Individuals’ membership in these groups does not status of recognized asylum seekers and most recently change over time in our sample (or only rarely, regarding arrived non-EU citizens) and world regions of origin the categorized duration of stay). As the majority of per- (broad categories considering the number of observa- sons does not stay in the sample for more than one survey tions) operationalize H #10. Note, however, that legal sta- wave, the variance in the data which drives our results tus and duration of stay do not provide any information is derived from the differences between person-waves, on the time when German authorities grant the status of rather than within persons across waves. Therefore, we a legally recognized refugee and thus a (fixed-term) resi - choose simple logistic regression analysis (a robustness dence permit. check with logistic random-effects models yield very sim - Several other individual and household-related char- ilar results, cp. Sect. 4.3). By applying clustered standard acteristics of the observed population (including natives) errors at the person level, our models take account of the serve as control variables in our multivariate analyses, non-independence of repeated observations of the same including age, level of education, duration of the current person, where applicable. Results are presented as aver- unemployment episode, subjective health, and partner age marginal effects (AME), which can be read as the pre - household. In addition, we assume gender to become rel- dicted average difference in percentage points regarding evant (only) when viewed in its interaction with own chil- the probability of the dependent variable for a one-unit dren to care for, also depending on children’s age (Leber Further correlations pertain to age and duration of unemployment (Person’s rho = 0.32) and a fairly but not prohibitively high correlation between having a German language skills were coded as “very good” if respondents said that partner and a combined variable which captures a persons’ sex and the fact of German was their first language; missing and implausible values were sorted having children in the household (Cramer’s V = 0.55). into the medium category. Beyond the sampling-induced clustering of person-years within persons, As the survey is offered in various languages, Turkish, Russian, and since one could also control for treatment-induced clustering at the level of Job- wave 10 Arabic (Jesske and Schulz 2018, p. 74), its participation rate and centers or case managers. However, such information is not provided in the validity regarding immigrants should be better than in German-only sur- data. If it could be included in the models, this might slightly change the veys. estimated standard errors and significance of coefficients. 9 Page 10 of 24 R. Lehwess‑Litzmann , J. Söhn Table 1 Jobcenters’ offers to job ‑seeking basic‑income recipients: natives and immigrants, by duration of stay in Germany All job‑seekers Thereof… Natives Migrants Thereof … Up to 4 years At least of stay 5 years of stay Regular employment 31.8 31.8 31.7 24.5 38.1 Marginal employment 17.0 17.2 17.2 13.7 20.4 Assistance in applications 28.3 28.0 29.3 29.9 29.2 Reimbursement of application costs or travel expenses 43.0 47.6 31.9 29.0 32.5 Program with employer or internship 12.1 11.8 13.0 16.3 10.2 Fin. support to become self‑ employed 2.3 2.3 2.2 1.6 3.0 Activation or placement voucher 14.9 15.4 13.8 10.4 18.1 Vocational (re‑)training or a course 19.6 14.2 31.9 39.8 26.8 Other offers 1.8 2.0 1.3 0.9 2.0 No offer, excl. integration or language course 29.5 29.0 31.4 31.2 30.4 Integration or language course 14.7 0.9 46.2 69.6 23.8 No offer, incl. integration or language course 26.6 29.0 22.3 15.8 27.9 Source: PASS waves 9–14, own calculations. Weighted mean yearly shares of clients who received at least one of the respective offers in the period 2015–2020 change of an independent variable, with all other predic- always refer to the weighted yearly average across the tors held constant. 6 years studied. While the multivariate analysis is based on unweighted Compared with the natives in our observed popula- data, descriptive results are always weighted (pertain- tion, the share of men is higher among immigrants (59.8 ing to person-waves). Weighting the data evens out dis- vs. 52.5 percent), in particular among those immigrants proportional numbers in the sample, e.g., persons form having arrived during the recent 4  years (73.4 percent Syria and Iraq, who were oversampled in wave 10 (year men). Immigrants without any professional qualification 2017) of the PASS survey (Jesske et  al. 2019, preface). are clearly overrepresented (54.9 vs. 39.4 percent among The weighting parameters included in the PASS always natives), with 61.2 percent among the recent immigrants. refer to one specific survey wave. In order to determine Yet, the share with an academic qualification is also con - values for the whole period of observation, we first cal - siderably higher among immigrants studied here (16.8 culate weighted values for each of the five panel waves vs. 3.6 percent among natives). In the native group, a separately and then average those values across all waves. majority (57.0 percent) has a degree of non-academic As a small qualitative add-on, we insert few selected vocational training (vs. 28.3 percent of immigrants). The excerpts of expert interviews we conducted during the foreign-born Jobcenter-clients live far more often with larger mixed-methods research projects into which the a partner (61.7 vs. 28.8 percent) and with children than analyses presented here was embedded. The interviews, their native peers. anonymized in this article, with mid-level managers of Among the immigrants, 13.2 percent are German citi- Jobcenters and Employment Agencies in a mid-sized zens and 19.4 percent citizens of another EU country West-German city and in a more rural community took (Appendix Table  7). Among third country nationals, the place in 2016/2017. share with a temporary residence permit is higher than with a permanent one. This holds in particular for immi - grants living in Germany for less than 5  years (65.0 per- cent with temporary residence permit). Among those 4 Results and discussion who give valid answers to the respective survey question, 4.1 Description of the r eference population a high share of the recent immigrants say they came to and Jobcenters’ ALMP offers Germany as asylum-seekers: 86.4 percent in 2018 (wave The first part of this section describes our reference 12) and 69.3 percent in 2019 (wave 13). Persons from population by some personal and household features the former Soviet Union and Eastern Europe constitute of native and immigrant Jobcenter clients (Appendix the largest shares of non-recent immigrants (30.8 and Tables 6 and 7) as well as offers made to them by Jobcent - 26.0 percent, respectively). The majority of those having ers between 2015 and 2020 (Table 1). The shares reported Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 11 of 24 9 stayed in Germany for less than 5 years stem from (West- category “clean” of integration and language classes (see ern) Asian countries (62.5 percent), including countries Sect.  4.3 below). The integration courses are “offered” to many refugees fled from. Looking at immigrants’ Ger- 69.6 percent of recent immigrants, they are obligatory in man language skills, 36.4 percent among recent immi- many cases. Among those staying in Germany already grants and 54.4 percent among the longer-settled ones more than 4  years, 23.8 percent reported the offer to consider their own skills as good or very good. attend integration and language classes. Regarding the different measures which the PES can Leaving aside such measures not applicable to natives, offer to job-seeking clients, Table  1 shows no major dif- we find that 29.0 percent of native and 31.4 percent of ference between the proportions of native and (all) immigrant clients (all in regular contact with their Job- immigrant clients referred to job vacancies pertaining to center) do not receive any offer in an average year of regular employment (31.8 vs. 31.7 percent). Yet, distin- observation. Once we include integration and language guishing immigrants by duration of residence reveals a classes, the share of recent immigrants without offer substantial gap. In accordance with some of our expec- from the Jobcenter is only 15.8 percent. tations, fewer recent immigrants (24.5 percent) report such referrals, but 38.1 percent of the longer-settled 4.2 Determinants of Jobcenters’ offers: multivariate immigrants do. The latter also received referrals to mar - analyses ginal employment more often (20.4 percent) than natives The presentation of multivariate results has two parts: (17.2 percent) or recent immigrants (13.7 percent). By In the first, we analyze for the whole sample, natives contrast, Jobcenters offered support with job applications and immigrants, in how far personal and context fac- most often to recent immigrants (29.9 percent). Natives tors explain the clients’ chance of the Jobcenters offering feature the highest share of those offered reimbursement them a specific type of employment policy measure and of application costs and travel expenses (47.6 vs. 29.0 whether being an immigrant plays a significant role. In and 32.5 percent among recent and longer-settled immi- the second part, we run our model for immigrants only grants respectively). Probably, native Jobcenter clients and add immigrant-specific variables in order to differ - are invited more often to job interviews than immigrant entiate this heterogeneous group. This helps us to look clients. deeper into the conditions under which Jobcenters make With the exception of assistance in applying for jobs, offers to immigrants. the measures reviewed so far depend strongly on labor market conditions and employers’ decisions. In contrast, 4.2.1 Nativ es and immigrants compared caseworkers’ own considerations are more decisive for The joint model for immigrants and natives in Table  2 offers regarding the other elements of the ALMP tool - confirms most of the above descriptive findings (cp. box. Jobcenters offered a subsidized program with an Table 1). It is again paramount to distinguish immigrants employer or an internship most often to recent immi- with shorter and longer duration of stay because effects grants (16.3 percent). Financial support to become self- sometimes point in opposite directions. For those immi- employed is rarely offered (2.3 percent of the observed grants who have arrived up to 4  years before the inter- population, and most often non-recent immigrants, view, the estimated probability to be referred to a regular 3.0 percent). Activation or placement vouchers, often job or an apprenticeship is on average 12.3 percentage meant for personal occupational orientation but also points (p.p.) lower than that of native clients, while there for further training or support in the job search, show a is no significant difference between natives and longer- similar pattern as referrals to (regular or minor) employ- settled immigrants. H #1a—the expectation that the PES ment. Recent immigrants receive them least often (10.4 offer more jobs to immigrant clients in general—is thus percent) and longer settled immigrants more often (18.1 not supported by our findings, while there is evidence percent), with native clients in between (15.4 percent). which corroborates H #1b. Possibly, Jobcenter staffs do Concerning occupational (re-)training “or other not consider recent immigrants yet fit for employment or courses”, by contrast, recently arrived immigrants report they anticipate that it is harder for them to meet employ- such offers most often (39.8 percent), natives least often ers’ requirements (e.g., German language skills). With (14.2 percent), the other immigrants positioned in regard to minor employment, caseworkers are most likely between (26.8 percent). As mentioned above, however, to refer longer-settled immigrants to job vacancies (6.0 we cannot be sure that all respondents kept this response p.p. more than native clients), with no difference between natives and recently arrived immigrants. H #2 is thus supported only for part of the immigrant clientele. In the multivariate analyses, we sort the small groups of clients from Turkey With regard to ALMP programs, the pattern varies. to the Middle East/Asian category, while those from Northern, Western and Contrary to the descriptive results, the model with all Southern Europe are aggregated with Eastern Europeans. 9 Page 12 of 24 R. Lehwess‑Litzmann , J. Söhn Table 2 Determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support ( joint model) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program with Activation or Vocational No offer employment employment applications of application or employer or placement (re‑)training travel costs internship voucher or course Person category (reference: native basic income recipients looking for a job) Basic income − 0.123*** − 0.007 − 0.053*** − 0.201*** 0.015 − 0.069*** 0.139*** 0.101*** recipients looking for a job who immigrated during the past 4 years Basic income 0.039 0.060*** 0.057** − 0.114*** 0.012 − 0.013 0.115*** 0.008 recipients looking for a job who immigrated at least 5 years ago Age (reference: 35 to 44 years) 18 to 0.035 0.053* 0.103*** 0.048 0.131*** 0.009 0.005 − 0.075*** 24 years 25 to 0.015 0.024 0.032 0.046* 0.030* 0.011 0.014 − 0.035* 34 years 45 to − 0.032 0.003 − 0.050** − 0.056** − 0.015 − 0.035** − 0.054*** 0.070*** 54 years 55 to − 0.089*** 0.005 − 0.089*** − 0.090*** − 0.038** − 0.063*** − 0.126*** 0.142*** 64 years State of health: − 0.028 − 0.008 − 0.035 − 0.069*** − 0.025 − 0.034* − 0.056** 0.043* bad (reference: very good to less good) Professional qualification (reference: none, lower ‑secondary school‑leaving certificate at most) None, but 0.031 − 0.033* 0.003 0.012 0.000 0.025 0.048** − 0.002 upper or intermediate secondary school‑ leaving certificate Non‑ 0.056*** − 0.018 0.031* 0.068*** 0.002 0.032** 0.021 − 0.031* academic professional qualification Academic 0.049* − 0.070*** 0.084*** 0.101*** 0.008 0.074*** 0.046** − 0.064*** qualification (university or technical/ teacher training col‑ lege) Duration of current unemployment so far (reference: 12 to 23 months) 0 to − 0.063*** − 0.019 − 0.057*** − 0.093*** − 0.041*** − 0.027* − 0.068*** 0.105*** 2 months 3 to 0.034 − 0.005 0.017 − 0.012 − 0.025 0.010 0.001 − 0.015 11 months 24 months − 0.057*** 0.008 − 0.011 − 0.021 0.000 0.003 − 0.019 0.014 and more Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 13 of 24 9 Table 2 (continued) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program with Activation or Vocational No offer employment employment applications of application or employer or placement (re‑)training travel costs internship voucher or course Gender and youngest child in household (reference: woman w/o children in household) Mother with − 0.136*** − 0.067*** − 0.078** − 0.210*** − 0.044** − 0.081*** − 0.125*** 0.288*** child aged 0 to 2 years Mother with − 0.020 0.007 − 0.009 − 0.046* 0.005 − 0.009 − 0.012 0.021 child aged 3 to 17 years Father with 0.066* − 0.022 0.081** − 0.037 0.049* 0.031 0.019 − 0.018 child aged 0 to 2 years Father with 0.022 − 0.013 0.059* − 0.012 0.037* 0.016 0.008 − 0.032 child aged 3 to 17 years Man without 0.015 − 0.010 0.048** − 0.002 0.033** 0.023 0.010 − 0.014 children in household Partner in − 0.057*** − 0.045*** − 0.044** 0.011 − 0.007 − 0.033** − 0.030* 0.041** household (reference: none) Underem‑ − 0.010*** − 0.004* − 0.015*** − 0.015*** − 0.004* 0.008*** − 0.004* 0.012*** ployment rate in fed‑ eral state Year (reference: 2015) 2016 − 0.018 0.005 0.005 − 0.001 0.012 0.003 0.006 0.010 2017 − 0.020 0.002 0.009 − 0.013 0.018 0.008 0.001 0.025 2018 − 0.024 − 0.013 − 0.024 − 0.057** 0.006 0.000 0.008 0.027 2019 − 0.004 − 0.027* − 0.023 − 0.050** 0.030* 0.011 0.019 0.029 2020 0.034 0.000 0.042* − 0.017 0.050*** 0.052*** 0.072*** − 0.027 Pseudo‑R 0.032 0.019 0.029 0.040 0.034 0.034 0.064 0.041 N (person‑ 8226 8224 8221 8207 8228 8200 8225 8234 waves) Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations *p < 0.05; **p < 0.01; ***p < 0.001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter a  “No offer” means “financial support to become self‑ employed” and “other offers” have also not been granted, neither integration or language courses control variables shows that non-recent immigrants are Probably, native clients receive invitations for job inter- most likely to report support with their applications, views more often than immigrants (and then have the while recent immigrants are least probable to do so costs reimbursed), in spite of receiving fewer referrals (+ 5.7 p.p. respectively − 5.3 p.p. compared with native and less support with applications from Jobcenters than clients), in part falsifying H #3. In line with H #4, both non-recent immigrants. recent and non-recent immigrants have a significantly Subsidized programs with employers or internships lower likelihood of Jobcenters reimbursing application could help jobless immigrants gain employers’ trust, or travel costs than natives (− 20.1 p.p. and − 11.4 p.p.). which they possibly do not enjoy as much as native appli- cants. However, multivariate regression yields no dif- ference between the three client groups for this costly ALMP measure. Hence, despite the bivariate result that This difference between these gross effects (see Additional file  1: Table S1) the recently immigrated reported this offer above aver - and net effects is partly due to the relatively high share of academics among age, H #5 is not confirmed when controlling for the other recent arrivals, as high education makes reporting support for applications more likely, independently of migration status (see Table 2). 9 Page 14 of 24 R. Lehwess‑Litzmann , J. Söhn covariates in the model. But there is no inequity in access (e.g., the large share without professional qualification) to either. a larger extent than by migration-related factors. By con- Regarding activation or placement vouchers (for, e.g., trast, support with applications does not differ between occupational orientation, further vocational training), the two groups at first sight, but turns out statistically newly settled immigrants face lower chances (− 6.9 p.p.) significant in the full model, such that recently-arrived of receiving respective offers compared with natives, immigrants receive fewer offers of this kind than natives, despite their hypothesized higher need for such measures given their social composition. Overall, the explanatory due to lacking knowledge about the local labor market power of our models with regard to Jobcenters’ offers is and job culture or because of non-recognized foreign cre- not very high. This hints at the importance of unobserved dentials. We can only speculate whether these offers are factors, probably linked to the dynamics of the service crowded out by other programs (e.g., integration courses interaction, clients’ personal preferences, and case man- might partly cover labor market topics) or whether agers’ professional attitudes (Dietz and Osiander 2019) as caseworker consider recent immigrants’ knowledge of well as the particular “cultures” of individual Jobcenters. German insufficient for following programs offered in One main finding of this section is that recent immi - German. As longer-settled immigrants face those obsta- grants generally receive fewer offers from out of the cles to a lesser extent than new arrivals, it appears plausi- ALMP toolbox than native clients. As for the longer-set- ble that they are not treated differently from natives. Yet, tled immigrants, the pattern of Jobcenter activity in some H #6-plus still needs to be dismissed. respects looks like an attempt to compensate the labor Jobcenters offer the more extensive programs of “voca - market obstacles that still confront this group, possibly tional (re-)training or a course” comparatively often to because their linguistic and cultural knowledge is already recent immigrants, with a likelihood 13.9 p.p. higher higher than that of recently arrived immigrants, making compared with natives and non-recent immigrants’ prob- non-migrant specific offers more feasible in the eye of the ability increased by 11.5 p.p., respectively. This finding Jobcenter. chimes with H #7a, which supposes an especially high need and motivation on the part of immigrants (however, 4.2.2 Comparisons among immigrants see chapter 4.3.1). In order to find out more about how Jobcenters react to Finally, recent immigrants (but not the longer-settled) the different profiles of immigrant clients, this section remain significantly more often (+ 10.1 p.p.) without any will explore the pattern of proposed ALMP measures ALMP offer than natives in a given year of observation within the immigrant subgroup of our sample. Restrict- (integration and language classes excluded). ing our estimations to immigrants allows us to extend Table  2 reports the effects of control variables which our estimation models by some variables which are only are, however, of no central interest here. Let us just men- meaningful for this group: legal status, duration of stay tion that most ALMP measures are more likely offered by in Germany, and the self-perceived level of German lan- Jobcenters to clients with a higher level of qualification guage skills. As for the dependent variables, we now rather than a low level. Jobcenters do not try to compen- also look at offers reserved for immigrants, that is inte - sate the largest gaps in education by their measures, oth- gration courses and other German language classes. erwise, those with only lower-secondary education would A long duration of stay has a significant impact: Refer - be more in their focus. Control variables contribute to rals to regular employment or apprenticeships are more the statistical explanation of the dependent variables’ var- likely (+ 11.1 p.p.) if immigrants have lived more than 19 2 iance, increasing the Pseudo-R compared to the gross 10  years (rather than less long) in the country (Table  3). models with migration experience as the only predic- For them, offers of minor employment also have a prob - tor (see Additional file  1: Table  S1). Overall, when there ability (7.0 p.p.) higher than that to more recently arrived already is a significant bivariate correlation, the effects of migration status on Jobcenter offers remain statisti - cally significant with regard to most outcome variables There is no information on this in the PASS data. Matching the ADIAB when the full multivariate models consider the composi- extension to the PASS could provide us with the information on the respon- tion of the groups. The less frequent referrals to marginal sible Jobcenter. We did not opt for this due to a loss of cases and a lack of employment of recently arrived immigrants (compared administrative data in the case of recently-arrived migrants. The information on world regions of origin would also be suited as a to natives) are explained by the composition of that group variable, but it is too highly correlated with the legal status to be simultane- ously included. If regions are included instead of legal status, we do not find any effects of the former on ALMP offers, except that clients from Africa and Middle East/Asia are more often proposed integration or language Other possible predictors like the partner’s employment status and house- courses. This, of course, should have to do with their legal status, which is hold income turned out to have no influence. omitted in these alternative models. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 15 of 24 9 individuals. By contrast, longer-settled immigrants face Middle East, are also offered an integration or language a higher risk of not reporting vocational (re-)training course significantly more often than all other clients. and other courses (− 13.2 p.p.), and, of course, receive All the remaining variables are identical to the ones significantly fewer offers to participate in integration or used in Sect.  4.1, Table  2. We observe that the impact of language courses. Interestingly, there is no significant dif - clients’ qualifications differs from the full sample model ference between the most-recently arrived and the immi- (including natives) with regard to two measures: in the grants staying between 5 and 10  years, apart from the immigrant-only model, academics are not referred to more frequent offer of integration or language courses to regular employment more often than persons with very the former group. It is only due to these courses that the low education, and neither are academics significantly most recently immigrated have a lower probability (− 6.6 more likely of being offered activation or placement p.p.) than less recent arrivals of reporting no offer at all. vouchers than less educated immigrants. Also, for other Appendix Table 7 shows that self-assessed German lan- factors captured in variables, like bad health, age, or a guage skills differ widely between immigrant Jobcenter partner in the household, we find fewer or weaker effects clients. Corroborating H #9, support with applications is in the migrant-only model. Yet, this might be due to the less probable for persons with bad or very bad German smaller sample size. skills (− 10 p.p. compared with clients with “reasonable” German skills). There is no significant limitation of refer -4.3 Robustness tests rals to minor jobs in case of clients with weak German Given the limitations of the data already addressed in language skills, but a reduced likelihood of reimbursing the methods section, we put our results to three tests job-searching costs points to employers’ under-average of robustness. They include using a different regression response to those persons’ applications. In addition, cli- model (4.3.2), a different sample (4.3.3), and an attempt ents with “bad” or “very bad” German skills are less likely to correct potential false responses to one survey item (− 4.7 p.p. and − 13.3 p.p.) to report being offered pro - (4.3.1). Statistically speaking, our results presented above grams with employers or internships than clients with turn out to be trustworthy. In the case of the item “voca- “reasonable” German. As for vocational (re-)training tional (re-)training or a course”, a doubt remains in the and other courses, the probability of offers diminishes, case of immigrants in our PASS-sample. too (− 6.2 p.p. and − 13.0 p.p.). It seems that casework- ers prioritize integration and language courses if clients 4.3.1 R eassessing offers of vocational (re‑)training and other hardly speak German, and contemplate ALMP options courses in the light of integration and language classes as soon as a certain level of German language has been In chapter  3, we already addressed the issue of a pos- reached. Possibly, Jobcenters consider the possession of sible overlap of the response category vocational (re-) some German skills as the condition of being able to par- training or a course and the migrant-specific integra - ticipate in most ALMP programs. The results on integra - tion or language courses. Due to the order of interview tion and language courses also reflect the regulation that items, clients who were only offered an integration or recent immigrants with good knowledge of German do language course are likely to say “yes” when they are first not need to attend them. In addition, referrals to regular asked about offers of Vocational (re-)training or a course”. jobs are significantly less often made to immigrant clients Potentially, these individuals then tick “yes” a second with “bad” German skills (− 5.1 p.p.), but more often to time when they are concretely asked about an Integration those with “very good” ones (+ 6.1 p.p.). All in all, results or language course later on. As a test of robustness, we on the impact of German language skills are consistent therefore eliminate all person-waves from our sample in with H #9 regarding both ALMP measures and referrals which both offers are reported and re-run our regression to job vacancies. model on this reduced sample, which now consists of Legal status, being a German citizen or not, having a 7449 person-waves, among which 2414 are migrants (820 permanent or fixed-term residence permit, does not person-waves less than the full sample). The results seem to affect most Jobcenters’ ALMP offers, as soon as other immigrant-specific factors are controlled for. The Also, the additional variables in the second model forbid a direct compari- possibility that the client might (have to) leave Germany son of control variables’ impacts with the above model for the whole sample. in the near future does not seem to discourage casework- Additional file  1: Table  S2 displays some differences of the characteris - ers. Counter to H #10, German nationals, EU citizens and tics of respondents, grouped by the combination of offers they reported: third-country nationals with a permanent residence sta- respondents who report both offers are more similar to respondents who tus are offered a program with an employer or an intern - only reported “integration or language course” in terms of legal status, duration of stay and German language skills, but more similar to respond- ship less often than clients with a temporary residence ents who only said yes to having been offered “vocational (re-)training or a permit. The latter, mostly stemming from Africa or the course” in terms of professional qualification. 9 Page 16 of 24 R. Lehwess‑Litzmann , J. Söhn Table 3 Determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support (immigrants only) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program Activation or Vocational Integration No offer employment employment applications of application or with placement (re‑)training or language travel costs employer or voucher course course internship Legal status (reference: not EU citizen, fixed‑term residence) German 0.028 0.028 0.04 0.035 − 0.056* 0.032 − 0.036 − 0.154*** 0.024 citizen EU citizen, 0.046 0.018 0.04 0.047 − 0.047* 0.046* 0.000 − 0.174*** 0.036 non‑ German Not EU − 0.013 − 0.02 0.110* 0.000 − 0.082** 0.048 0.038 − 0.146** 0.042 citizen, open‑ ended residence Duration of stay in Germany so far (reference: 5 to below 10 years) 0 to below − 0.038 0.000 − 0.044 − 0.03 − 0.014 − 0.011 − 0.018 0.188*** − 0.066** 5 years 10 years and 0.111** 0.070* 0.013 0.008 − 0.042 0.044 − 0.132*** − 0.302*** 0.044 more German language skills (reference: reasonable) Very good 0.061* 0.032 0.025 0.077* − 0.018 0.013 − 0.031 − 0.127*** 0.015 Good 0.037 − 0.005 0.039 0.076*** 0.005 0.006 0.004 − 0.102*** 0.017 Bad − 0.051* − 0.020 − 0.101*** − 0.100*** − 0.047* − 0.042* − 0.062* − 0.022 0.036 Very bad − 0.067 − 0.031 − 0.100* − 0.172*** − 0.133*** − 0.028 − 0.130* − 0.012 0.033 Age (reference: 35 to 44 years) 18 to 24 years − 0.016 0.019 0.013 − 0.008 0.049 0.027 0.013 0.027 − 0.008 25 to 34 years 0.000 0.005 − 0.004 − 0.01 0.003 0.009 − 0.018 − 0.021 0.020 45 to 54 years − 0.014 0.01 − 0.024 − 0.031 − 0.004 − 0.039* − 0.056* − 0.034 0.025 55 to 64 years − 0.104*** − 0.010 − 0.053 − 0.057 0.003 − 0.050* − 0.123*** − 0.049 0.092*** State of health: − 0.004 − 0.008 − 0.073 − 0.048 − 0.037 − 0.021 − 0.123** − 0.049 0.015 bad (reference: very good to less good) Professional qualification (reference: none, lower ‑secondary school‑leaving certificate at most) None, but 0.013 − 0.030 0.036 0.018 0.026 0.005 0.052* − 0.048* 0.021 upper or intermediate secondary school‑ leaving certificate Non‑ 0.056* − 0.023 0.048 0.01 0.045* 0.023 0.057* − 0.043 0.016 academic professional qualification Academic 0.008 − 0.060** 0.082*** 0.077** 0.038* 0.035* 0.045 − 0.048* 0.002 qualification (university or technical/ teacher train‑ ing college) Duration of current unemployment so far (reference: 12 to 23 months) 0 to − 0.001 − 0.002 − 0.024 − 0.001 − 0.028 − 0.008 − 0.048 − 0.128*** 0.087*** 2 months 3 to 0.054 0.009 0.043 0.008 − 0.021 0.006 − 0.003 − 0.048 0.026 11 months 24 months 0.014 0.025 0.027 0.018 0.008 0.011 0.004 − 0.006 0.025 and more Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 17 of 24 9 Table 3 (continued) Independent Dependent variable: referral/offer made by the Jobcenter variables Regular Marginal Assistance in Reimbursement Program Activation or Vocational Integration No offer employment employment applications of application or with placement (re‑)training or language travel costs employer or voucher course course internship Gender and youngest child in household (reference: woman w/o children in household) Mother with − 0.144*** − 0.057 − 0.071 − 0.123** − 0.039 − 0.039 − 0.123* − 0.099 0.187*** child aged 0 to 2 years Mother with − 0.061 0.017 − 0.017 − 0.037 0.001 0.007 0.000 0.028 0.006 child aged 3 to 17 years Father with 0.035 0.004 0.124** 0.004 0.086** 0.050 0.034 − 0.005 − 0.034 child aged 0 to 2 years Father with 0.003 0.007 0.065 − 0.001 0.046 0.033 0.019 − 0.009 − 0.018 child aged 3 to 17 years Man without 0.023 0.012 0.055 0.016 0.061** 0.022 0.006 0.000 0.003 children in household Partner in − 0.037 − 0.039* − 0.056* − 0.005 − 0.023 − 0.036* − 0.070** − 0.004 0.047** household (reference: none) Underem‑ − 0.003 − 0.002 − 0.009** − 0.012*** − 0.005 0.006* 0.002 − 0.005 0.002 ployment rate in federal state Year (reference: 2017) 2015 0.046 − 0.024 − 0.030 − 0.052 − 0.021 − 0.009 0.075* (base) 0.042 2016 0.030 − 0.008 − 0.023 0.012 − 0.007 0.007 0.063* 0.023 0.015 2018 0.018 − 0.031 − 0.006 − 0.048 − 0.024 0.014 0.048 0.044 − 0.014 2019 0.049 − 0.053** − 0.047 − 0.056* − 0.006 0.025 0.041 − 0.012 0.014 2020 0.143*** 0.005 0.056 − 0.003 0.036 0.079*** 0.119*** 0.027 − 0.033 Pseudo‑R 0.049 0.034 0.037 0.034 0.047 0.036 0.043 0.247 0.104 N 3191 3187 3185 3176 3193 3172 3190 2886 3195 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations *p < .05; **p < .01; ***p < .001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter “No offer” means “financial support to become self‑ employed” and “other offers” have also not been granted dramatically change concerning Jobcenters’ offer of integration/language course. Yet, there are some reasons Vocational (re-)training or a course (Table  4): instead of to expect a high number of immigrant clients who were prioritizing newly arrived immigrant clients by 13.9 p.p. only offered integration classes, regarding both timing as compared to native clients (Sect.  4.1), they are now and language preconditions. An integration class usu- at a disadvantage of − 10.4 p.p. As for non-recent immi- ally takes six to seven month on average (Goethe Insti- grants, the coefficient turns insignificant, i.e., Jobcenters tut 2020). Depending on when it starts and how often do not treat them differently form native clients. clients and case managers talk to each other, the chance But how convincing are these alternative results, given of being offered a subsequent vocational (re-)training in that a part of the observations which have been elimi- the same reference year of the survey should be limited. nated may really have been offered both types of courses, Concerning language skills, integration courses lead to in which case ticking both survey items with a “yes” the language level A2 or B1 (Goethe Institut 2020), which would be the adequate answer? There is no way of know - is usually not sufficient to follow the complex teach - ing the exact share of respondents who says yes to voca- ing in vocational training attended by native speakers tional (re-)training or a course” but in fact meant only an as well. Furthermore, official statistics do not imply an 9 Page 18 of 24 R. Lehwess‑Litzmann , J. Söhn Table 4 Robustness test (I): determinants of various ALMP measures by Jobcenters to job‑seeking recipients of basic‑income support ( joint model), modified responses to item “vocational training or a course” Independent variables Dependent variable: vocational (re‑)training or a course All sample persons Only sample persons who did not also tick the offer of integration or language class Person category (reference: native basic income recipients looking for a job) Basic income recipients looking for a job who immigrated during the past 0.139*** − 0.104*** 4 years Basic income recipients looking for a job who immigrated at least 5 years ago 0.115*** 0.017 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations ***p < 0.001. Logit model, reported as average marginal effects. The significant coefficients can be read as the impact of a one ‑unit change of the independent variable on the estimated probability of receiving the offer by the Jobcenter Repetition of results from Table 2 for comparison. Same control variables as in Table 2 (coefficients not shown due to their similarity to those in Table 2) 4.3.3 O nly one wave per person overrepresentation of immigrants (restricted to those On average, immigrant persons remain in our sample for with non-German citizenship) among participants of a smaller number of waves. In order to check whether PES-financed further vocational training (see introduc - this could (partly) explain our results, we perform an tion; BA 2019c). A final reason to expect the share of cor - extra regression analysis where each person is observed rect “double-yes answers” to be small can be derived from only once across the whole observation period, no mat- the research of Kasrin et  al. (2021, p. 4). They find that ter if they satisfied the sampling criteria in 1, 2, or even the number of refugee Jobcenter-clients who attended more years of observation. We find that the results do further vocational training was only close to half of those not differ much from the full model, no matter whether attending a program with an employer. As we know from we use the first or last wave in which the sample person our data that recently-immigrated PASS respondents is observed (Additional file  1: Table  S4). Regarding the mention the latter offer relatively rarely (16.3 percent), impact of migrations status as seen in Table 2, significant the share of those receiving vocational (re-)training offers coefficients are widely reproduced. The only exception is should be much smaller than the 39.8 percent reported that longer-settled immigrants get more referrals to regu- above in Table  1. These arguments speak in favor of our lar jobs in their first year of observation, whereas there argumentation that among respondents reporting both is no significant difference compared with natives in the (re-)training and language classes many were in fact only person-wave model (Table 2). As for effect sizes, they are offered the latter, which will have a bearing on the con - also very close in the alternative models, with the mod- clusions drawn below. els using the first observation of each person yielding slightly bigger effect sizes than the models using the last 4.3.2 Random‑effects model observation, especially for the newly arrived immigrants. As mentioned above in the methods section, our empiri- This could mean that the effect of migration background cal findings are derived from the variance of phenomena on Jobcenters’ offers is higher in the beginning of a job between observations, i.e., person-waves. Theoretically, it search. Overall, the findings of our robustness checks could also be possible that Jobcenters’ offers are triggered underscore that the results in Sect.  4.2 are not driven by changes that happen in the lives of clients. In this case, by a difference within persons over time, but differences a model that takes the longitudinal dimension of the between persons. panel data into account would be more adequate. As a robustness test, we apply a random effects model, which 5 Conclusion builds on the variance both between and within persons. Based on survey answers given by job-seeking recipients The results do not change compared to our main model: of basic income, our analysis of the PASS survey in year it is the same coefficients that turn out statistically sig - 2015 to 2020 sought to identify determinants of ALMP nificant, and they have the same sign (Additional file  1: offers made by German Jobcenters to its clients. Our Table  S3). The between component seems be dominant special attention was on possible differences between compared with within component. This may also be due native and immigrant clients as well as on the effect of to the fact that only a minority of respondents is observed duration of stay among the latter. Our joint model for for more than 1 year. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 19 of 24 9 all job-seeking basic income recipients included native specific way one applies for jobs in Germany. An alter - clients, recent immigrants (having arrived in Germany native explanation of recent immigrants’ high propen- up to 4  years before the interview) and longer-settled sity to tick “yes” on the survey item “vocational training, immigrants (having stayed for 5 years or more). Whereas retraining or a course” could be its overlap, possibly aggregate administrative data, treating non-German caused by the sequence of the respective survey items, citizen as one single group, suggested no relevant immi- with “integration course or another German course”. This grant-native gaps (see BA 2019c as cited in the introduc- could erroneously make the occurrence of “vocational tion of our contribution), our distinction by time since training, retraining or a course” appear as relatively fre- migration proved crucial. quent among recent immigrants. There are indeed good Our multivariate analyses showed that recent immi- reasons to conclude that a high proportion of persons grants generally receive fewer referrals to job vacancies reporting both items were actually only offered integra - and fewer offers from out of the ALMP toolbox than tion or language courses (see Sect.  4.3.1). Our alterna- natives in a given year. For example, the likelihood of a tive model, excluding respondents with double positive caseworker suggesting such immigrants to apply for an answers, implies that Jobcenters are indeed likely to offer advertised regular job is 12.3 percentage points (p.p.) occupational (re-)training less often to recent immigrants lower, and the likelihood of offering an activation or than to natives or longer-settled immigrants, with no placement voucher is 6.9 p.p. lower than regarding native significant difference between the two latter groups. But clients. These results chime with some older, less differ - certainly, future research should use better data, e.g., on entiated findings on a negative effect of foreign or non- the realization of Jobcenter-sponsored (re-)training, to EU citizenship on benefitting from PES-sponsored (re-) verify this preliminary result. training or vouchers (Kruppe 2009, p. 14; Möller and Is there an alternative reading of our overall results Walwei 2009, pp. 305, 308–309; Osikominu 2005, p. 61) with regard to the more recent immigrants? The low as well as with previous research indicating some degree probability of referrals to open job positions to this group of ethnic discrimination by German administration could be based on their preference not to work. We judge (Grohs et al. 2016; Hemker and Rink 2017). Our interpre- this as implausible because Jobcenters tend not to allow tation is that many caseworkers, and probably employers, “laziness” on part of the clients, their main goal being to often perceive newly-arrived immigrants as not yet ready bring people into employment and to end benefit receipt. for entering the labor market, even though they do not Moreover, migration research (e.g. Sinning 2011) and our question their eagerness and even assert that some barri- own interviews with Jobcenter staff suggest that earning ers to employment which burden native Jobcenter clients money and sending part of it back to family members in are absent in the case of refugees (Boockmann and Scheu their home country is a major motivation and social obli- 2019, p. 409). Still, Jobcenters primarily offer integration gation among immigrants. Similar to findings by Boock - and language courses, which often constitute both a right mann and Scheu (2019, p. 409), one of the interviewed and an obligation, to recent immigrants. experts put it this way: There is one category that recently-immigrated “Those who already gained certificates strive for hav - respondents report significantly more often (+ 13.9 p.p.) ing them recognized, but this is not their main focus. than native Jobcenter clients. In the questionnaire, it is Their main focus is earning money.” (Interview II) called “vocational training, retraining or a course”. On the one hand, it might be convincing that recent immi- Integration and language classes, i.e., the kinds of grants received this kind of offer particularly often, as measures primarily proposed to recent immigrants, may this would reflect the special needs of a group that often be the first or second choice from the viewpoint of social lacks qualifications and certificates relevant for and/or law, Jobcenters, or immigrants themselves; they are often recognized on the German labor market (Knuth 2021, a necessary interim step to their successful labor market p. 56). Case managers explicitly highlight refugees’ need integration in the longer term. for further training (Boockmann and Scheu 2019, p. 408). With regard to longer-settled immigrants, Jobcent- On the other hand, most other measures were offered ers seem to use the full breadth of their toolbox to bring less often to recent immigrants, even support with appli- their clients into employment. As regards assistance with cations (− 5.3 p.p. compared with natives), despite such their job search, immigrants who have been staying in clients being probably little familiar with the culturally One can only speculate whether a confounding of language courses with One should keep in mind that the survey item for referrals to open job vocational training might also be a reason why Kruppe (2009, p. 14) finds a positions also mentions dual vocational training. Future research should try to small positive effect for recent immigration on the issue of training vouchers. distinguish these options. 9 Page 20 of 24 R. Lehwess‑Litzmann , J. Söhn Germany already for 5 years or more have a significantly individual migrants willing to learn, up to 4  years since higher probability of being referred to marginal (+ 6.0 arrival without training-related support can feel like a p.p.) employment as well as being offered support with long time and might cool off respective educational aspi - their applications (+ 5.7 p.p.), compared to native cli- rations. In order for Jobcenters to institutionally meas- ents, with no statistically significant difference regarding ure up to recent immigrants’ high motivation to succeed open positions of regular employment. By contrast, they in their host country, they should use an adequate mix are less often (− 11.4 p.p.) offered a reimbursement of of long-term oriented investment in human capital and application or travel costs, which conveys that they are short-term support to find work. ALMP measures com - less often invited for job interviews by potential employ- bining employment, occupational qualification and ers—an indicator of their disadvantages on the German advanced German language training seem promising labor market. The preferential support with applications (Boockmann and Scheu 2019, p. 418). offered to longer-settled migrants and the statistical simi - A desirable extension of our study would be to con- larity with native clients regarding most other ALMP trast ALMP measures offered with ALMP measures that measures suggests that there is no negative discrimina- are actually implemented. While our analysis of the for- tion against these migrants by Jobcenters. mer predominantly dealt with phenomena of selection, This reading is further backed by our finding that, as in an extended analysis could analyze the determinants of accordance with the principle of equity, migrants’ world self-selection that influence whether a client accepts an region of origin makes no significant difference for their offer. Furthermore, we do not know when offers are used probability of receiving ALMP offers, except for inte - strategically by Jobcenters to merely test the availability gration or language courses. The latter are significantly of clients for the labor market. This would, of course, more often proposed to clients from Africa and Middle change the interpretation of our results as such activating East/Asia, who overwhelmingly belong to the immigrant ‘offers’ mean nothing positive from the clients’ viewpoint. groups targeted by integration policy. Regarding ALMP However, we find it more plausible that Jobcenters’ offers measures which are not language-related, we do not find are intended as supportive, especially regarding compar- major differences by migrants’ legal status. The following atively long and extensive ALMP programs. statement of a caseworker interviewed for our project Our analysis was restricted to years when refugees underlines this: from Asia formed a large group among PES’ immigrant clients in Germany. It is up to future research to test Actually, as soon as the language part has more or whether our results hold once the structure of immigra- less been dealt with, we treat immigrant clients just tion changes again, e.g., if immigrants from European like the rest of our clients. We ask ourselves where countries (like Ukraine) and family migrants from out- the clients are heading, what would fit to them, what side of Europe will again become the majority of recently is reasonable, what are they able to achieve. (Inter- arrived immigrants. view II) In addition, our time window analyzed is a historical This statement also points to the importance of suf - boom phase of the German economy with low unem- ficient knowledge of German as a facilitator of further ployment rates. Relatively few job seekers competed for ALMP measures, and in particular for chances of finding the public resources dedicated to ALMP. If unemploy- employment. Our deeper analysis of which immigrant- ment should rise again in the years to come, possibly as specific factors are linked to Jobcenters’ offers shows a consequence of geopolitical conflicts, ALMP spending that referrals to regular employment and reimbursement per person will probably shrink even if aggregate expend- of application expenses are indeed offered more often iture grows (Lehwess-Litzmann 2018). No one can know to immigrant clients the better their German skills are. at present whether the German labor market will deal Immigrant job-seekers with weaker German skills are with this as well as with the COVID-19 crisis. Jobcent- also least likely to gain assistance in applications or pro- ers’ future immigrant clients will certainly need new grams with an employer or internships. occupational skills and certificates as well as employment Overall, our analysis gives reason to believe that, experience in the German labor market. compared with native clients, job-seeking immigrants receive overall equal treatment by the German PES, but only regarding longer-settled immigrants or those However, a particular consequence of the pandemic for migrants and refu- with (assumed or actual) good German language skills. gees is revealed by Brücker et al. (2021, 26 and 30): These groups experienced Whether Jobcenters’ activities are actually sufficient and a stronger rise of unemployment in the course of the year 2020, not primarily due to layoffs (net employment rates remained rather stable), but as a conse - how closely they correspond to immigrant clients’ prefer- quence of interrupted or cancelled integration, language and vocational train- ences deserves further investigation. At least in the eye of ing courses. Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 21 of 24 9 Appendix See Tables 5, 6 and 7. Table 5 Basic‑income recipients looking for a job: persons and person‑ waves in the sample by migration status Person category Number of Number of times the person figures in the sample Number persons of person‑ 1 2 3 4 5 6 waves Natives 2711 1496 610 301 146 110 48 5041 100.0% 55.2% 22.5% 11.1% 5.4% 4.1% 1.8% Immigrants 2243 1.565 459 143 60 14 2 3234 100.0% 69.8% 20.5% 6.4% 2.7% 0.6% 0.1% Total 4954 3061 1069 444 206 124 50 8275 100.0% 61.8% 21.6% 9.0% 4.2% 2.5% 1.0% Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations Table 6 Basic‑income recipient looking for a job: socio ‑ demographic characteristics, by migration status (as % of the observed population) Attributes Natives All immigrants …Up to 4 years of stay …At least 5 years of stay Gender Male 52.5 59.8 73.4 46.7 Female 47.5 40.2 26.6 53.3 Age 18–24 10.9 5.4 11.1 0.5 25–34 30.8 31.3 43.6 21.9 35–44 20.2 29.1 32.0 25.0 45–54 21.0 20.4 10.9 30.0 55–64 17.0 13.8 2.4 22.8 Professional qualification No professional qualification 39.4 54.9 61.2 47.8 … With lower‑secondary school‑leaving certifi‑ 29.7 40.6 40.3 38.6 cate at most … With upper or intermediate secondary school‑ 9.7 14.3 20.9 9.2 leaving certificate Non‑academic professional training 57.0 28.3 21.2 36.0 academic qualification 3.6 16.8 17.6 16.2 Self‑reported (very) bad health 9.7 8.5 2.9 12.9 With partner in household 28.8 61.7 71.2 55.6 Children (by age) in household No children in household 62.3 38.8 37.6 40.1 At least one child aged 0–2 8.4 18.4 26.3 10.7 At least one child aged 3–17 29.4 42.7 36.1 49.1 Duration of current unemployment episode 0–2 months 20.6 26.8 23.1 30.4 3–11 months 9.2 9.7 12.4 8.0 12–23 months 10.3 12.1 20.1 7.2 24 months and more 59.8 51.5 44.4 54.3 N (person‑ wave) 5041 3234 2202 1032 Source: IAB, PASS, Welle 14 v1. 2015–2020. Own calculations. The figures represent the mean over the weighted values for each year of observation 9 Page 22 of 24 R. Lehwess‑Litzmann , J. Söhn Table 7 Immigrant basic‑income recipients looking for a job: legal status and countries of origin, by duration of stay (as % of the observed population of immigrants) Attributes All immigrants …Up to 4 years of stay …At least 5 years of stay Legal status German citizen 13.2 3.0 21.8 Non‑ German EU citizen 19.4 18.6 22.4 Third country national with permanent residence permit 18.8 13.4 22.0 Third country national with temporary residence permit 48.6 65.0 33.7 Duration of stay (mean in years) 9.8 3.2 14.9 Region of origin Northern, Western and Southern Europe 3.2 3.1 3.8 Eastern Europe 20.1 16.7 26.0 Former Soviet Union 18.7 7.7 30.8 Other Asian country 45.0 62.5 24.8 Turkey 3.1 2.5 3.7 Africa 8.4 6.6 9.4 Other countries/unknown 1.5 1.4 1.6 German language skills (own account) Very good 12.2 7.7 15.4 Good 33.3 28.7 39.0 Satisfactory 38.6 46.6 30.8 Bad 13.4 13.5 12.3 Very bad 2.5 3.5 2.5 N (person‑ wave) 3234 2202 1032 Source: IAB, PASS, Welle 14 v1, 2015–2020. Own calculations. The figures represent the mean over the weighted values for each year of observation Abbreviations Author contributions ALMP: Active labor market policy; BA: German Federal Employment Agency The authors have produced all parts of the paper collaboratively, with an (Bundesagentur für Arbeit); EU: European Union; IAB: Institute for Employment emphasis of JS on the theoretical framing and of RLL on the empirical analysis. Research; PASS: Panel Study Labour Market and Social Security; PES: Public Both authors read and approved the final manuscript. employment services; SGB: Social Code (Sozialgesetzbuch). Authors’ information Dr. René Lehwess‑Litzmann is a senior researcher at the Sociological Research Supplementary Information Institute (SOFI) in Göttingen. He analyzes labor‑market and social policy, labor ‑ The online version contains supplementary material available at https:// doi. market trends, individual career choices and employment trajectories in the org/ 10. 1186/ s12651‑ 022‑ 00313‑8. context of societal challenges. Dr. Janina Söhn is a senior researcher at the SOFI in Göttingen. Within the broad field of social inequality, she has published on immigrant integration, Additional file 1: Table S1. Gross model with migration experience legal status and integration policy, education, vocational training, employ‑ as the only determinants of various ALMP measures by Jobcenters to ment trajectories, and retirement. job‑seeking recipients of basic‑income support ( joint model). Table S2. Characteristics of immigrant basic‑income recipients looking for a job, by Funding combination of offers reported. Table S3. Robustness test (II): Random‑ The research was funded by Deutsche Forschungsgemeinschaft (DFG) under effects mode (Odds ratios). Determinants of various ALMP measures the sign SO1286/2‑1 and contract number 616892. The authors conducted by Jobcenters to job‑seeking recipients of basic‑income support ( joint the research under conditions of scientific independence. model). Table S4. Robustness test (III): Only one person‑ wave per person. Determinants of various ALMP measures by Jobcenters to job‑seeking Availability of data and materials recipients of basic‑income support ( joint model). PASS data can be obtained from the IAB’s research data center under certain conditions, see https:// fdz. iab. de/ en/ FDZ_ Data_ Access. aspx. Acknowledgements We would like to thank participants of the Colloquium of the Sociological Declarations Institute of University of Duisburg‑Essen, where we presented preliminary results on 3rd June 2020. Many thanks goes also to the PASS team for some Ethics approval and consent to participate valuable pieces of information and last but not least to our anonymous The article mainly uses anonymized and aggregated secondary data, drawing reviewers for their commitment and their constructive feedback on earlier on the IAB’s PASS survey. Consent of respondents was assured by the makers versions of our article. of the survey. As for the expert interviews performed in the framework of our specific project, the interview partners were informed about rules and Jobcenters’ strategies to promoting the inclusion of immigrant and native job seekers: a… Page 23 of 24 9 procedures of privacy protection, upon which their consent to participate was survey experiment among employers. Soc. Forces 99(2), 648–671 (2020). obtained.https:// doi. org/ 10. 1093/ sf/ soz154 Deeke, A., Baas, M.: Abbau oder Reproduktion von Ungleichheit? – Erträge der Consent for publication beruflichen Weiterbildung arbeitsloser Migranten. Sozialer Fortschritt The article mainly uses anonymized and aggregated secondary data, drawing 62(1), 23–32 (2013) on the IAB’s PASS survey. As for the three extracts from expert interviews, Dietz, M., Osiander, C.: Labor market integration of refugees: the role of which are presented in a highly anonymized way in the text, consent to pub‑ caseworkers. In: Paper presented at the 31st SASE annual conference, The lish was obtained from the interview partners. New School, New York, June 27–29, 2019 (2019) Dietz, M., Osiander, C., Stobbe, H.: Arbeitsmarktintegration von Geflüchteten Competing interests aus Sicht der Vermittler. Online‑Befragung in Arbeitsagenturen und The authors declare that they have no competing interests. Jobcentern. IAB‑Kurzbericht 25|2018. Institut für Arbeitsmarkt ‑ und Berufsforschung, Nürnberg (2018) Received: 4 June 2021 Accepted: 21 June 2022 Esser, H.: Sprache und Integration. Die sozialen Bedingungen und Folgen des Spracherwerbs von Migranten. Campus, Frankfurt/New York (2006) Fendel, T.: Die Arbeitsmarktintegration geflüchteter Frauen. WISO Direkt 02/2019. Friedrich‑Ebert ‑stiftung, Bonn (2019) Freier, R., Steiner, V.: ‘Marginal employment’: stepping stone or dead end? Eval‑ uating the German experience. J. Labor Mark. Res. 2/3, 223–243 (2008) References Goethe Institut.: Integation course. https:// www. goethe. de/ prj/ mwd/ en/ idl/ Adamuti‑ Trache, M.: First 4 years in Canada: post‑secondary education path‑ deu/ iku. html (2020) ways of highly educated immigrants. J. Int. Migr. Integr. 12, 61–83 (2011). Grohs, S., Adam, C., Knill, C.: Are some citizens more equal than others? Evi‑ https:// doi. org/ 10. 1007/ s12134‑ 010‑ 0164‑2 dence from a field experiment. Public Adm. Rev. 76(1), 155–164 (2016) Arbeitsagentur.: Maßnahmen bei einem Arbeitgeber (MAG) nach § 45 SGB III Gundert, S., Kosyakova, Y., Fendel, T.: Migrantinnen und Migranten am im Rahmen der Maßnahmen zur Aktivierung und beruflichen Einglieder ‑ deutschen Arbeitsmarkt. Qualität der Arbeitsplätze als wichtiger ung. Fachliche Weisungen zur Durchführung des § 45 SGB III. https:// Gradmesser einer gelungenen integration. IAB‑Kurzbericht 25. Institut für www. arbei tsage ntur. de/ datei/ ba146 324. pdf (2019) Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2020) BA, B. f. A.: Aktivierung und beruflichen Eingliederung (MAbE) nach § 45 SGB III Hashmi Khan, A.: Post‑migration investment in education by immigrants in the (Jahreszahlen). Nürnberg (2019a) United States. Q. Rev. Econ. Finance 37(Special Issue), 285–313 (1997) BA, B. f. A.: Arbeitslosigkeit und Unterbeschäftigung. Deutschland und Länder Hemker, J., Rink, A.: Multiple dimensions of bureaucratic discrimination: 2018. Nürnberg (2019b) evidence from German welfare offices. Am. J. Polit. Sci. 61(64), 786–803 BA, B. f. A.: Förderung der beruflichen Weiterbildung (Monatszahlen). (2017) Deutschland, Länder und Regionaldirektionen. Nürnberg (2019c) Hohmeyer, K., Wolff, J.: Selektivität von Ein‑Euro ‑ Job‑Ankündigungen. IAB‑ BA, B. f. A.: Arbeitslose nach Rechtskreisen (Jahreszahlen). Deutschland und Forschungsbericht 2|2015. Institut für Arbeitsmarkt‑ und Berufsforschung, Länder. Berichtsjahr 2019. Nürnberg (2020) Nürnberg (2015) BA, B. f. A.: Arbeitsmarktpolitische Instrumente (Zeitreihe Jahreszahlen) Holzinger, C.: ‘We don’t worry that much about language’: street‑level 2000–2019. Nürnberg (2021) bureaucracy in the context of linguistic diversity. J. Ethn. Migr. Stud. 46(9), Bähr, S., Beste, J., Wenzig, C.: Arbeitsmarktintegration von Geflüchteten im 1792–1808 (2020). https:// doi. org/ 10. 1080/ 13691 83X. 2019. 16103 65 SGB IIHemmnisse abbauen und Potenziale nutzen. IAB‑Kurzbericht Integrationsbeauftragte, B. d. B. f. M., Flüchtlinge und Integration.: Deutschland 23/2017. Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg kann Integration: Potenziale fördern, Integration fordern, Zusammenhalt (2017) stärken.12. Bericht der Beauftragten der Bundesregierung für Migration, BAMF, B. f. M. u. F.: Integrationskursgeschäftsstatistik für das Jahr 2019 (Bun‑ Flüchtlinge und Integration in Deutschland. Berlin (2019) desländer). Nürnberg (2020) Jesske, B., Schulz, S.: Methodenbericht Panel Arbeitsmarkt und Soziale Becker, R.: Economic change and continuous vocational training in the work Sicherung PASS 11. Erhebungswelle – 2017. 01/2009 FDZ‑Methodenre ‑ history: a longitudinal multilevel analysis of the employees’ participa‑ port 13/2018. Forschungsdatenzentrum (FDZ) des Bundesagentur für tion in further training and the effects on their occupational careers Arbeit im Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2018) in Germany, 1970–2008. Empir. Res. Vocat. Educ. Train. 11(1), 4 (2019). Jesske, B., Schulz, S., Quandt, S.: Methodenbericht Panel Arbeitsmarkt und Sozi‑ https:// doi. org/ 10. 1186/ s40461‑ 019‑ 0079‑x ale Sicherung PASS‑Erhebungswelle 12 – 2018 (Haupterhebung). Institut Berg, M., et al.: Codebook and documentation of the panel study “Labour für angewandte Sozialwissenschaft GmbH (infas), Bonn (2019) Market and Social Security” (PASS)—Datenreport wave 12. FDZ‑ Just Landed.: Job applications. How to apply for a job in Germany. https:// Datenreport, 09/2019 (en). Research Data Centre (FDZ ) of the Federal www. justl anded. com/ engli sh/ Germa ny/ Germa ny‑ Guide/ Jobs/ Job‑ appli Employment Agency (BA) in the Institute for Employment Research catio ns (2020) (IAB), Nürnberg (2020) Kalter, F., Granato, N.: Migration und ethnische Ungleichheit auf dem Arbeits‑ Bonin, H., Boockmann, B., Brändle, T., Bredtmann, J., Brussig, M., Dorner, markt. In: Abraham, M., Hinz, T. (eds.) Arbeitsmarktsoziologie: Probleme, M., et al.: Begleitevaluation der arbeitsmarktpolitischen Integrations‑ Theorien, empirische Befunde, pp. 355–387. Springer Fachmedien maßnahmen für Geflüchtete. Zweiter Zwischenbericht. Forschungsber ‑ Wiesbaden, Wiesbaden (2018) icht 546. Bundesministerium für Arbeit und Soziales, Berlin (2020) Kanas, A., van Tubergen, F.: The impact of origin and host country schooling Boockmann, B., Scheu, T.: Integration der Geflüchteten in den Arbeitsmarkt: on the economic performance of immigrants. Soc. Forces 88(2), 893–915 Ziele, Strategien und Hemmnisse aus Sicht der Jobcenter. Zeitschrift Für (2009). https:// doi. org/ 10. 1353/ sof.0. 0269 Sozialreform 65(4), 393–426 (2019) Kasrin, Z., Stockinger, B., Tübbicke, S.: Aktive Arbeitsmarktpolitik für arbeit‑ Brücker, H., Gundacker, L., Hauptmann, A., Jaschke, Ph.: Die Arbeitsmark‑ slose Geflüchtete im SGB II. Der Großteil der Maßnahmen erhöht den twirkungen der COVID‑19‑Pandemie auf Geflüchtete und andere Arbeitsmarkterfolg. IAB‑Kurzbericht 7|2021. Institut für Arbeitsmarkt ‑ und Migrantinnen und Migranten. IABForschungsbericht, 05/2021. Institut für Berufsforschung, Nürnberg (2021) Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2021) Knuth, M.: "Transformative Arbeitsmarktpolitik. Herausforderungen der Arbe‑ Bundesregierung.: Lebenslagen in Deutschland. Der 6. Armuts‑ und Reich‑ itsmarktpolitik unter den Bedingungen der, konfluenten Digitalisierung’". tumsbericht der Bundesregierung. Berlin (2021) Working Paper Forschungsförderung. Berlin: Forschungsstelle "Arbeit der Cross, K.P.: Adults as Learners. Increasing Participation and Facilitating Learning. Zukunft" der Hans Böckler Stiftung (2021) Jossey‑Bass, San Francisco (1981) Kogan, I.: Last hired, first fired? The unemployment dynamics of male immi‑ Damelang, A., Kosyakova, Y.: To work or to study? Postmigration educational grants in Germany. Eur. Sociol. Rev. 20(5), 445–461 (2005) investments of adult refugees in Germany—evidence from a choice Koopmans, R., Veit, S., Yemane, R.: Taste or statistics? A correspondence study experiment. Res. Comp. Int. Educ. 73(online‑first), 1–12 (2021) of ethnic, racial and religious labour‑market discrimination in Germany. Damelang, A., Ebensperger, S., Stumpf, F.: Foreign credential recognition and Ethn. Racial Stud. 42(16), 233–252 (2019) immigrants’ chances of being hired for skilled jobs—evidence from a 9 Page 24 of 24 R. Lehwess‑Litzmann , J. Söhn Kosyakova, Y.: Mehr als nur ein Job: die qualitative Dimension der Integration Yankova, K.: Der Selektionsprozess in Maßnahmen der aktiven Arbeitsmarktpo‑ in Arbeit von Geflüchteten in Deutschland. WISO DISKURS 09/2020. litik. Eine explorative Untersuchung für die deutsche Arbeitsvermittlung. Friedrich‑Ebert ‑Stiftung, Bonn (2020) IAB‑ discussion paper 11/2010. Institut für Arbeitsmarkt‑ und Berufs‑ Kruppe, T.: Bildungsgutscheine in der aktiven Arbeitsmarktpolitik. Sozialer forschung, Nürnberg (2010) Fortschritt 58(1), 9–19 (2009) Lancee, B., Bol, T.: The transferability of skills and degrees: why the place of Publisher’s Note education affects immigrant earnings. Soc. Forces 96(2), 691–716 (2017). Springer Nature remains neutral with regard to jurisdictional claims in pub‑ https:// doi. org/ 10. 1093/ sf/ sox058 lished maps and institutional affiliations. Leber, U., Möller, I.: Weiterbildungsbeteiligung Ausgewählter Personengrup‑ pen. Schmollers Jahr. 128(3), 405–429 (2008) Lehwess‑Litzmann, R.: Is labour market policy heading for social investment in European countries? An analysis of spending trends between 2006 and 2015. Zeitschrift Für Sozialreform 64(3), 333–362 (2018). https:// doi. org/ 10. 1515/ zsr‑ 2018‑ 0018 Liechti, F., Fossati, F., Bonoli, G., Auer, D.: The signaling value of labor market programs. Eur. Sociol. Rev. 33(257–274) (2017) Linden, P.: Wie viel Geld ist angemessen? Eine Vignettenstudie zur Akzeptanz von Sanktionen im SGB II. WSI Mittelungen (2021). https:// doi. org/ 10. 5771/ 0342‑ 300X‑ 2021‑6‑ 454 Lipsky, M.: Street‑level Bureaucrats. The Dilemmas of Individuals in Public Services, Updated Russell Sage Foundation, New York (2010) Maynard‑Moody, S., Musheno, M.: Cops, Teachers, Counselors: Stories From the Front Lines of Public Service. University of Michigan Press, Ann Arbor (2003) Möller, J., Walwei, U.: Handbuch Arbeitsmarkt. Institut für Arbeitsmarkt‑ und Berufsforschung, Nürnberg (2009) Nohl, A.‑M., Schittenhelm, K., Schmidtke, O., Weiß, A. (eds.): Work in Transition. Cultural Capital and Highly Skilled Migrants’ Passages into the Labour Market. University of Toronto Press, Toronto (2014) Osiander, C.: Zugang zu beruflicher Weiterbildung ‑ Rechtlicher Rahmen, Teilnahme an und Wirkung von beruflicher Weiterbildung, Weiter ‑ bildungshemmnisse. In: Goth, G.G., Kretschmer, S., Pfeiffer, I. (eds.) Bildungsinnovationen für nicht formal Qualifizierte. Zielgruppengerechte Weiterbildungssettings in der Bildungspraxis, pp. 61–88. wbv Media, Bielefeld (2019) Osikominu, A.: Eine Analyse der Teilnehmerselektion in die berufliche Weiterbildung auf Basis der Integrierten Erwerbsbiografien (IEB). IAB‑ Forschungsbericht 23/2005. Nürnberg Institut für Arbeitsmarkt‑ und Berufsforschung [Institute for Employment Research, Nuremberg, Germany] (2005) Sauer, M.: Was Migrant/innen mit Jobcentern erlebt haben: Fallstudien in Bedarfsgemeinschaften mit Migrationshintergrund. In: Knuth, M. (ed.) Arbeitsmarktintegration und Integrationspolitik ‑ zur notwendigen Verknüpfung zweier Politikfelder: eine Untersuchung über SGB II‑ Leistungsbeziehende mit Migrationshintergrund, pp. 151–160. Nomos, Baden‑Baden (2010) Schneider, J., Fischer, M., Kovacheva, V.: Migrants in the Job Centre Qualitative Findings on Migrants’ Experiences with Public Employment Support Ser‑ vices in Germany. Hamburg Institute of International Economics (HWWI), Hamburg (2008) Sinning, M.G.: Determinants of savings and remittances: empirical evidence from immigrants to Germany. Rev. Econ. Househ. 9(1), 45–67 (2011). https:// doi. org/ 10. 1007/ s11150‑ 009‑ 9082‑5 Söhn, J.: Initial employment pathways of immigrants in Germany. Why legal contexts of reception matter—an analysis of life‑ course data. Transf. Eur. Rev. Labour Res. 25(1), 43–62 (2019). https:// doi. org/ 10. 1177/ 10242 58918 StaBa, S. B.: Foreign population by duration of stay and selected citizenships. https:// www. desta tis. de/ EN/ Themes/ Socie ty‑ Envir onment/ Popul ation/ Migra tion‑ Integ ration/ Tables/ forei gner‑ durat ion‑ of‑ stay. html (2021) Steinhardt, M.F.: The impact of naturalisation on immigran labour market integration in Germany and Switzerland. In: OECD (ed.) Naturalisation: A Passport for the Better Integration of Immigrants, pp. 145–151. OECD Publishing, Paris (2011) van Tubergen, F., De Werfhorst, H.V.: Postimmigration investments in educa‑ tion: a study of immigrants in the Netherlands. Demography 44(4), 285–313 (2007) Veit, S., Thijsen, L.: Almost identical but still treated differently: hiring discrimi‑ nation against foreign‑born and domestic‑born minorities. J. Ethn. Migr. Stud. (2019). https:// doi. org/ 10. 1080/ 13691 83X. 2019. 16228 25

Journal

Journal for Labour Market ResearchSpringer Journals

Published: Dec 1, 2022

Keywords: Immigration; Integration; Jobcenter; Public employment services; Active labor market policy; ALMP; Training; Labor market; Welfare state; H41; I38; J08; J15; J68; J61

References