Tackling disabilities in young age: Policies that work

von Simson, Kristine; Hardoy, Inés Article Tackling disabilities in young age: Policies that work IZA Journal of Labor Policy Provided in Cooperation with: IZA – Institute of Labor Economics Suggested Citation: von Simson, Kristine; Hardoy, Inés (2020) : Tackling disabilities in young age: Policies that work, IZA Journal of Labor Policy, ISSN 2193-9004, Sciendo, Warsaw, Vol. 10, Iss. 1, pp. 1-27, https://doi.org/10.2478/izajolp-2020-0013 This Version is available at: https://hdl.handle.net/10419/298756 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Kristine von Simon1,* and Inés Hardoy2 Tackling disabilities in young age—Policies that work Abstract Work impairment is an increasing concern in advanced economies, particularly among young people. Activation, rather than passively providing economic support, is often regarded as the preferred strategy for addressing this issue. However, little is known about which measures are effective for improving youth work impairment. A hazard rate competing risk model with unobserved heterogeneity applied to rich Norwegian panel data provides some insights. Wage subsidies, and to some extent education/training programs, have the intended effect. In other words, work-impaired youths who participate in these measures have a higher probability of obtaining work/starting an education and a lower probability of experiencing a transition to social security than those youths who do not participate in any measure. The impacts of follow-up initiatives and work practice programs are more mixed. Current version: October 15, 2020 Keywords: reduced work capacity, vocational rehabilitation, Timing-of-Events model, youth JEL codes: C41, I38; J08; J22; J24; J68 Corresponding author: Kristine von Simson [email protected] Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 © The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Cite as: Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13. https://doi.org/10.2478/izajolp-2020-0013 1 Institute for Social Research, Oslo 2 Institute for Social Research, Oslo, E-mail: [email protected] Page 2 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 1 Introduction Young people with health problems face several challenges in the labor market. They have a disadvantage in completing upper secondary school (Champaloux and Young, 2015) and getting a job (Maslow et al., 2011), and health problems might also affect their pay (Smith, 2009). In addition, there are indications that poor health in adolescence has negative long-term consequences for employment, especially among people with low education (Holland et al., 2011). Vocational rehabilitation programs (VRPs) have the purpose of facilitating labor market inclusion and counteracting the likelihood of labor market exclusion of people with reduced work capabilities. Moreover, the design and efficiency of public policies and the way program gatekeepers interpret the numerous policies available to them, affect the flows in and out of disability pensions (Burkhauser et al., 2016). This article investigates the impact of VRPs targeting work-impaired youths in Norway. Norway serves an interesting case. The proportion of young people aged 18–29years receiving health-related benefits in Norway increased from 1.9% in 1994, when unemployment reached its highest peak of the last decades, to 3.4% in 2000, and then further to 5.1% in 2017 (NOU, 2019:7). The majority of these youths are diagnosed with mental health problems. Combined with a low exit rate from long-term sickness/disability to work, this trend is worrying. In addition, Norway, together with the other Scandinavian countries, has a long tradition of publicly supplied welfare services and activation measures. Existing evidence points to a shift in sickness and disability policies in recent decades in most Organization for Economic Co-operation and Development (OECD) countries from passive income support to stronger employment support and benefit conditionality (Böheim and Leoni, 2018). In Norway, individuals must attempt a VRP before they can be awarded a permanent disability claim. VRPs are much more comprehensive than ordinary active labor market programs (ALMP) for the unemployed, both in terms of their cost per capita and the number of participants. There is a vast amount of literature on the impact of ordinary ALMPs. Fewer studies focus on VRPs (ALMPs for individuals with work impairment), and surprisingly few concentrate specifically on the impact of VRPs for young people. Existing evidence on ALMPs indicates that policies need to be targeted to be effective; there is a need to understand better what works for whom (Crépon and Berg, 2016). From the literature on ALMPs for ordinary unemployed individuals, it is clear that the effectiveness of programs differs substantially by age group, where young people seem to gain less from participation in ALMPs than adults do (Card et al., 2017; Hardoy et al., 2018; Kluve et al., 2019). The empirical literature on the impact of programs targeting people with reduced work capability using well-established identification strategies is scant and inconclusive. Some studies using more recent data deserve special mention. Angelov and Eliason (2018) study the effect of wage subsidies targeting job seekers with disabilities in Sweden and find both positive and negative impacts. Any positive employment effects seem to be outweighed by considerable lock-in effects. However, the participants are less likely to have a transition out of the labor force to the disability insurance program. Rehwald et al. (2018)’s Danish evidence from a randomized experiment indicates that neither vocational programs nor counseling help sick-listed workers return to work, and they might even have adverse effects. These results are in contrast to those of Holm et al. (2017), who find positive employment effects of ordinary education and Page 3 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 wage subsidies for the same group of workers in Denmark. Campolieti et al. (2014) study the effects of a vocational rehabilitation program implemented in the late 90s in Canada. Their matching estimators suggest relatively small and imprecise effects for men and larger and significant effects for women. Dean et al. (2017) apply a structural model to a rich U.S. sample from 2000 and find positive long-term effects on employment and earnings for individuals with mental health problems, particularly for employment-related services. Adamecz-Völgyi et al. (2018) in Hungary and Bewley et al. (2007) in the United Kingdom are examples of studies of comprehensive programs involving counseling, training, rehabilitation, and employment subsidies, which show positive employment effects.1 A randomized control trial (Burns et al., 2007) carried out in six European countries among people with severe mental illness compares traditional vocational rehabilitation with individual placement and support, concluding that the latter is more effective. None of the abovementioned studies focus on youth. A recent study focusing on work-impaired individuals in Norway with reduced work capabilities deserves closer mention. Markussen and Røed (2014) use local/geographical variation in labor market offices to identify the causal effects of VRPs on labor market outcomes. They conclude that strategies focusing on early intervention and participation in measures in the ordinary labor market are better than alternative strategies that give priority to vocational education or organized work in the sheltered sector. While older individuals benefit more from work-related measures, ordinary education seems to be the most successful measure for younger work-impaired individuals.2 In this study, we compare the impact of activation to that of passively receiving welfare support. We investigate how the duration of work impairment and eventual participation in the different VRPs affect transitions to one of two outcome states: job/education and social security. To identify the impact of VRPs, we use a mixed proportional competing risk hazard model (Abbring and van den Berg, 2003), which is described in detail in Gaure et al. (2007). A special feature of the model is that it captures unobserved heterogeneity, which is crucial for separating selection effects from causal effects. Caliendo and Schmidl (2016) show that there is still no consensus regarding the effectiveness of different active labor market policies for youth in Europe. They emphasize that more evidence is required and that there is a particular need for studies including education as an outcome of success as well as studies with a long observation period. Our study adds to the literature by providing causal evidence of the effect of such policies for a particularly vulnerable group—work-impaired youth—using comprehensive administrative data covering 12 birth cohorts over a period of 13years and including several dimensions of outcome measures. Our analyses indicate that wage subsidies primarily, but also to some extent education/ training programs, increase the probability of obtaining work/starting an education and reduce the probability of experiencing a transition to social security. The impacts of follow-up initiatives and work practice programs are more mixed. During participation, these measures are 1 Positive results of VRPs are also found for Finland (Leinonen et al., 2019) and Switzerland (Hagen, 2019). 2 Salvanes et al. (2018) study the effects of ordinary education for young people with work impairment in Norway. They use a reform aimed at depriving work-impaired youths aged 22–25 years of the right to participate in ordinary education as a VRP. The analysis shows that the reform led to young people having more difficulties returning to work compared with young people who were not affected by the reform. However, the effect does not seem to be long-lasting. Page 4 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 associated with an increased probability of having a transition to job/education. After program completion, the positive effect on job/education transitions persists. However, now the youths also have a higher probability of experiencing a transition to social security. We interpret this as an indication that follow-up and work practice measures are used as a screening device and function as springboards to working life. The article proceeds as follows: first, we present the data, provide some descriptive statistics, and describe the empirical model. Next, we look into the results on the impact of different VRPs on job/education on the one hand and social security on the other. We conclude the article with a discussion of our results and a brief summary. 2 Data and sampling We have access to rich Norwegian administrative data from several individual registers covering information on employment, unemployment, income, education, social welfare, demographics, administered and merged by Statistics Norway. Our sample consists of all residents in Norway born between 1976 and 1994 registering as work impaired for at least 1 month during the period from January 2002 through December 2012, and who are between 18 and 29years old at the time of registration. Notice that the eldest we observe were 26years old in 2002, and it is only in 2005 that we first observe youths up to age 29. The data has a panel structure, making it possible to follow individuals over time and monitor their transitions in and out of the labor market. There are two possible ways to obtain the status of work impaired in the administrative registers. One is after a period of employment and sick leave and requires a certificate of ill health issued by a general practitioner (GP). The other is to be given the status of work impaired by a supervisor at the Labour and Welfare Service (NAV) office3 through a work capacity assessment. There is no lower limit on the degree of reduced working capacity. The work capacity assessment is the basis for further follow-up and labor market measures. VRPs consist of measures targeting individuals with work impairment (e.g., vocational education and work in sheltered firms4) as well as ordinary active labor market programs (e.g., training and work practice). While being registered as having work impairment, youths may receive different kinds of benefits, depending on their health status. Youths who have their work capacity reduced by at least 50% due to a GP-certified illness are entitled to a health-related benefit.5 Those who do not fulfill the health requirements may be eligible for unemployment benefits, activity support, or means-tested social assistance. The unit of analysis is spells rather than individuals. A fresh spell of work impairment is defined as a period with no occurrence of registered work impairment during the previous 6 months. The final sample consists of 130,634 unique spells beginning in the period 2002–2012, comprising 108,134 youths aged 18–29years. We follow these youths on a monthly basis until 3 NAV is all-encompassing in the sense that it provides all welfare services: social-, healthand labor market-related services. 4 Sheltered firms produce goods and/or services and are established to provide clarification, job training, or qualification to persons who have reduced their ability to work. They are financed with public resources 5 Before 2010, the temporary health-related benefits consisted of rehabilitation benefits, vocational benefits, and timelimited disability benefits. In 2010, these three benefits were merged into one benefit: the work assessment allowance. If work capacity is reduced permanently, permanent disability benefits may be granted. We remove youths who receive permanent disability benefits from our sample, as they are not likely to return to the labor market. Page 5 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 December 2014, which means that we observe all youth for a minimum of 2 years and some youths up to 13years. While youths are registered with reduced working capacity, they can participate in VRPs. Transitions to VRPs are referred to as temporary transitions. It is not uncommon to have several spells of program participation within the same spell of work impairment. However, it is problematic to model such repeated spells because previous participation in a labor market program can affect both the likelihood of future program participation and the impact of these programs. Therefore, in this study, we focus on the effect of the first VRP and censor subsequent transitions to VRPs. VRPs vary throughout analysis based on economic fluctuations and labor demand.6 Programs are grouped in such a way that they resemble the categories typically used in international studies (Kluve, 2010; Card et al., 2017). We focus on four major categories. Education/ training (EDU) refers to off-the-job classroom courses/education. Wage subsidies (WS) entail subsidized ordinary employment in the public or private sectors. Work practice (WP) is mostly on-the-job training expected to provide work experience in both the ordinary and sheltered sectors. Follow-up (FU) is supported by employment and follow-up assistance to obtain or retain work. The remaining small-scale programs are placed in a residual category, and transitions to these programs are censored. A more detailed description of the different programs is found in Appendix B. We distinguish between the effects of VRPs while participating in a program and the effects after completion of the program. A large body of research literature points to so-called lock-in effects, where the unemployed get locked into the program and spend less time searching for jobs during their participation (vanOurs, 2004; Røed and Raaum, 2006). After program completion, the likelihood of getting a job may increase again, for example, due to higher job search activity, increased formal or job-specific human capital, better information or larger networks. The work-impairment spell ends when the person is no longer registered with reduced working capacity for three consecutive months. We identify two exit states: exit to social security, which includes a permanent disability or social assistance, and exit to an ordinary job or a formal education. Appendix B describes of the definition and priority of different labor market states. Transitions to states other than social security or job/education are censored.7 In addition, we censor spells that are still on-going at the end of the observation period.8 This makes it possible to include all young people who register as work impaired without having to assume when the spell will end. 2.1 Descriptive statistics Figure 1 shows the inflow into new spells of work impairment in our data by the month of entry. There was an increase in new work-impairment spells from 2002 to 2012, particularly in the periods 2002–2004 and 2009–2010. After 2010, the inflow into work impairment 6 The target group has also been redefined in 2009/2010, and there have been changes in regulations, both of which have affected the composition of the target group and the way they are distributed in the different programs. The changes are discussed in depth in Sections 2.1 and 4.2. 7 Appendix B describes the definition and priority of different labor market states. 8 We also censor durations over 60 months, as transitions after this time are rare and complicate statistical inference. Page 6 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 decreased. The observed pattern is mainly due to changing business cycles, with recessions in 2001–2003 and 2008–2009 causing increased inflows into nonemployment. However, changes in regulations related to the follow-up of individuals with work impairment may also have affected the inflows. First, the time-limited disability benefit was introduced in 2004. This was a temporary disability benefit targeted at young people with favorable labor market prospects, to decrease inflows into permanent disability pension. Then, a new follow-up work-impaired regime was introduced during 2009–2010. It involved a workcapability assessment, along with an expansion of the eligibility criterion for receiving temporary health-related benefits to include individuals without previous sickness or labor market history. Figure 2 shows survival curves by VRP status. We see that youths spend a relatively long time in work impairment. After 36months, 23% of the youth who do not participate in any programs is still registered as work impaired. The corresponding shares among program participants are 36% for WS, and around 60% for FU, WP, and EDU. The longer durations for VRPs partly reflects the lock-in effects of participation. Figure 3 shows monthly transition rates into different VRPs. Overall, young workimpaired individuals are more likely to participate in EDU and WP and less likely to participate in FU measures and WS. We also see that EDU and WP are more frequently used early in the work-impairment spell, stabilizing around 10months, whereas the flow into the other programs is rather stable over the 5–6-year period. Figure 4 shows monthly transition rates into the outcome states work/education and social security, by VRP status. For youth who do not participate in any VRP during work impairment, the probability of having a transition to work/education decreases rapidly during Figure 1 Inflow into work impairment by the month of entry. Young people 18-26 years of age in the period 2002–2012. Note: The figure includes only youths aged 26 or under, as older youths are not equally represented over the evaluation period in our data. Page 7 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 the first 2 years and stabilizes at a low level thereafter. Again, we see indications of lock-in effects for the VRPs, with low probabilities and slightly increasing transition rates to work/ education during the first months of the spell. The pattern is quite different when it comes to social security, where the likelihood of experiencing a transition to social security drops the first year, but we also see a steep increase in the transition rate after around 40months. This partly reflects the dynamic selection problem: individuals who are still work impaired after Figure 3 Monthly transition rates into vocational rehabilitation programs (VRPs). Young people 18-29 years of age in the period 2002–2012. Figure 2 Survival curves by vocational rehabilitation program (VRP) status. Young people 18-29 years of age in the period 2002–2012. Page 8 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 40months constitute a highly selected group with a high probability of entering social security and a low probability of entering employment or education. Descriptive statistics of the observed characteristics of work-impaired youths are presented in Table 1. Only 46% of work-impairment spells contain VRP participation; the remaining 54% comprise the comparison group. Many young people spend considerable time in work impairment, with an average of 20months for spells of no participation and slightly longer for VRP participation spells.9 On an average, youths spend around 8 months in VRPs, with educational measures lasting the longest. There are signs of considerable selection into different VRPs. Women participate more often in EDU, whereas they are strongly underrepresented among participants in WS. There is also a relatively low proportion of non-European youths in WS compared with participants in other programs. Participants in EDU tend to be positively selected in that they are more educated, have more recent labor market experience and also have parents with higher education and income. Participants in WP, on the other hand, seem to be negatively selected. Moreover, while nearly 44% of participants in EDU receive health-related benefits10 at the start of the work-impairment spell, only 25% of participants in WS do.11 There are also notable differences in outcomes states, depending on VRP participation. WS and EDU seem to be more successful in terms of outcome states than WP and FU. The 9 Duration is measured including time spent in VRP. 10 Health-related benefits include all benefits requiring a doctor-certificated medical condition: rehabilitation benefits, vocational benefits, time-limited disability benefits (before 2010), and work assessment allowance (after 2010). The first month is chosen for practical reasons. 11 We, unfortunately, do not have information about other kinds of transfers the youth receive during work impairment. Bragstad and Sørbø (2014) find in their study of young work impaired that 23% of the youth receive social assistance during the first month of work impairment, while 32% receive no benefit at all. Their sample is, however, not directly comparable to our sample, as they focus only on youth entering work impairment in 2011. Figure 4 Monthly transition into work/education and social security, by vocational rehabilitation program (VRP) status. Young people 18-29 years of age during the period 2002–2012. Page 15 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Figure 6 The effect of the vocational rehabilitation program (VRP) on the transition to job/ education. FU, follow-up; WP, work practice; EDU, education/training; WS, Wages Subsidies. Notes: The dashed line shows the transition probability for a reference person who does not participate in any VRP, and who has been registered as work impaired for 6–10months. The reference person is a male, aged 22–25, living in Eastern Norway, native-born, with no completed upper secondary education, with average parental background, by average youth unemployment. Figure 5 The effect of vocational rehabilitation program (VRP) on the transition to social security. Notes: FU, follow-up; WP, work practice; EDU, education/training; WS, wages subsidies. The dashed line shows the transition probability for a reference person who did not participate in any VRP, and who has been registered as work impaired for 6–10months. The reference person is a male, aged 22–25 years, living in Eastern Norway, native-born, with no completed upper secondary education, with average parental background, by average youth unemployment. Page 16 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 The figure on the right shows the positive effects of all programs after participation. WS is associated with the largest positive effects, in line with most studies of both ordinary ALMPs and VRPs, also in an international context. The impact is an increase of about 4% points, from 2% to 6%. Meanwhile, EDU shows an increase of nearly 2% points compared with the nonparticipation alternative. The likelihood of getting an ordinary job or starting an education is nearly three times as high after participation in WS relative to not participating in any program. The objective of WS is that participants continue to work for the firm that receives the subsidy after the subsidy is removed, which can partly explain the positive effect. However, for those participants who return to work impairment after the subsidy period, there is still a significantly increased likelihood of a transition to work or education. This indicates that the impact of wage subsidies is not exclusively a deadweight effect (i.e., it is not the case that employers only hire people they would have employed anyway).16 4.2 Robustness tests As mentioned earlier, the ToE model makes some assumptions that are difficult to test. In order to investigate the sensitivity of our results, we run three robustness tests. The first two concern unobserved heterogeneity, while the last test concerns the sample composition. The ToE model assumes that unobserved heterogeneity is time invariant. However, during the spell of work impairment, health may change. If changes in health status influence the probability of treatment as well as the probability of having a transition to one of the outcomes, our effect estimates may be biased. For instance, youths with deteriorating health statuses may be less likely to participate in VRPs because they need to get better in order to benefit from participation. At the same time, deteriorating health may be associated with an increased probability of having a transition to social security and a decreased probability of having a transition to job/education. Although our data do not contain any direct information about individual health status, we observe whether the youths receive health-related benefits during the spell of work impairment. A medical-certified reduced work capacity of at least 50% is required in order to be eligible for a health-related benefit. Receipt of a health-related benefit may thus serve as a signal of the gravity of the health condition. We include a time-varying indicator equal to 1 with the receipt of health-related benefits and 0 otherwise as a proxy to changes in health status.17 The second robustness test investigates the sensitivity of our results to the choice of information criterion used to select the number of mass points in the distribution of unobserved heterogeneity. Lombardi et al. (2019) show that selecting too few or too many mass points for the distribution of unobserved heterogeneity may seriously bias the treatment effects. They compare the performance of the Maximum Likelihood (ML) criterion (i.e., choose the number of mass points where there is no further improvement in the log likelihood) to information criteria penalizing parameter abundance: the AIC, the Bayesian Information Criterion (BIC), and 16 The literature often argues that WS has greater deadweight and displacement effects than the other types of measures. Caliendo et al. (2017) point out that the selection to WS might be more complex than for other programs because it involves more active participation on the part of the employed during the hiring. Attempts to fully control for positive selection might not be altogether successful. 17 Holm et al. (2017) conduct a similar robustness test in their evaluation of active labor market programs for sick-listed workers in Denmark. Page 17 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Figure 7 Yearly inflow into vocational rehabilitation program (VRP), as share of all ongoing work-impairment spells. Young people 18-29 years of age old. Note: FU, Follow-up; WP, work practice; EDU, education/training; WS wage subsidies. the Hannan–Quinn Information Criterion (HQIC) in a ToE framework. They conclude that all information criteria perform better than the ML criterion, but no single criterion performs better in all settings. They thus recommend using all three criteria and report the results from the different criteria as a robustness check. However, they also show that the risk of overcorrection is larger in small samples than in large samples, which implies using a less restrictive criterion, such as the AIC, in our case. The last robustness test concerns changes to the composition of our sample related to the introduction of the new work-impairment regime in 2009–2010. Importantly, the regime introduced obligatory work capability assessments and expanded the target group to include individuals without previous labor market or sickness histories. This change is particularly relevant for our target group, as young people often lack labor market experience. In addition, the regime emphasized intensified follow-up and early activation. As shown in Figure 7, the reform led to a steep rise in VRP participation. We investigate whether the implementation of the WAA reform affects our results by estimating the model solely on work-impairment spells starting before 2009.18 Results from the robustness tests are presented in Table 3. The first column shows results from the original model of Table 2, where health is assumed to be time invariant, the number of mass points is chosen according to the AIC and the whole sample is used. The next column then introduces time-varying health, as explained above. The third and fourth column presents results using the two other information criteria, whereas the last column shows results for spells starting before the work-impairment reform in 2009. As shown in Table 3, the results seem to be largely robust to the inclusion of time-varying health. However, some of the positive 18 Ideally, we would also like to estimate the model on spells starting after 2009. However, the time span is too short for many to experience a permanent transition. Hence, a large portion of the observations is censored. Page 18 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Table 3 Effect of vocational rehabilitation programs (VRPs) during and after program participation on the transition to social security and job/education: Robustness tests To social security To job/education Original model Time-varying health BIC HQIC Before 2009 Original model Time-varying health BIC HQIC Before 2009 Under treatment FU 0.034 (0.067) −0.030 (0.064) 0.070 (0.063) 0.082 (0.064) 0.354*** (0.074) 0.470*** (0.034) 0.330*** (0.036) 0.480*** (0.033) 0.469*** (0.034) 0.559*** (0.042) WP −0.061 (0.040) −0.098** (0.040) −0.071* (0.040) −0.063 (0.040) 0.333*** (0.046) 0.251*** (0.024) 0.119*** (0.025) 0.264*** (0.024) 0.258*** (0.024) 0.546*** (0.029) EDU −0.736*** (0.054) −0.895*** (0.053) −0.660*** (0.052) −0.722*** (0.054) −0.534*** (0.061) 0.088*** (0.021) −0.107*** (0.021) 0.072*** (0.020) 0.089*** (0.021) 0.220*** (0.024) WS −0.598*** (0.117) −0.684*** (0.115) −0.523*** (0.113) −0.575*** (0.115) −0.528*** (0.123) 0.691*** (0.043) 0.808*** (0.046) 0.683*** (0.042) 0.693*** (0.043) 0.797*** (0.046) After treatment FU 0.237*** (0.074) 0.212*** (0.071) 0.246*** (0.064) 0.263*** (0.071) 0.301*** (0.095) 0.367*** (0.043) 0.443*** (0.045) 0.377*** (0.043) 0.366*** (0.043) 0.018 (0.070) WP 0.136*** (0.045) 0.119*** (0.045) 0.115*** (0.040) 0.127*** (0.045) 0.162*** (0.060) 0.273*** (0.028) 0.307*** (0.029) 0.281*** (0.028) 0.277*** (0.028) 0.031 (0.044) EDU −0.083 (0.057) −0.182*** (0.056) 0.005 (0.054) −0.072 (0.057) −0.066 (0.074) 0.491*** (0.027) 0.574*** (0.028) 0.474*** (0.027) 0.493*** (0.027) 0.455*** (0.035) WS −0.242 (0.169) −0.297* (0.169) −0.146 (0.115) −0.217 (0.170) −0.284 (0.195) 0.918*** (0.069) 1.129*** (0.073) 0.900*** (0.069) 0.913*** (0.069) 0.554*** (0.096) BIC, Bayesian information criterion; HQIC, Hannan–Quinn information criterion; FU, follow-up; WP, work practice; EDU, education/training; WS, wage subsidies. Notes: * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level. In addition to the treatment effects, the estimations include controls for age, gender, immigrant background, education level, activity before work impairment, previous income, parental background, region of residence, indicator for health-related benefit receipt, local unemployment rate, duration dependence, and calendar variables. The original model has eight mass points in the heterogeneity distribution, whereas the model with time-varying health has nine mass points. The BIC and HQIQ models have five and six mass points, respectively, while the model with spells starting before 2009 has nine mass points. Page 19 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 impacts become more prevalent, and EDU and WS now significantly reduce the likelihood of a transition to social security. We also see that the results are robust to the information criterion used; both the BIC and the HQIC give very similar results as the original model using the AIC. When estimating the model using only spells starting before 2009, two interesting differences appear regarding FU and WP. The results indicate that participants in FU and WP are more likely to have a transition to social security while on the program when we disregard the period after 2009. This may be associated with the use of VRPs as a screening device. Rules are such that activation (work-oriented measures) is compulsory before permanent disability benefits can be granted. As the eligibility criteria for receiving health-related benefits became less strict after the WAA reform, it is likely that the sample was comprised of relatively more youths with serious health conditions before the implementation of WAA than after it. FU and WP are more likely than EDU and WS to have been more frequently offered to youth with unclear work capability prospects in the period before 2010. The other noticeable change is that the positive job/education effect of having participated in WP and FU disappears. This suggests that the positive effects of the period after 2009 are driving the positive results in the original model. Recall that inflow into VRPs increased dramatically with WAA, with VRP participants likely to have better employment prospects since the rules for entitlement became less strict. The observed pattern seems to reflect this change in the composition of the target group. 5 Discussion and conclusion Results from OECD’s Better Life Index show that Norway ranks very high on most measures of well-being, with short working hours, low-income inequality and high life satisfaction.19 This suggests that workers in Norway should, in principle, enjoy just as good health as workers in comparable countries. Nevertheless, Norway has high sickness absence rates and is among the countries with the greatest proportion of temporary or permanent health-related benefits.20 Of particular concern is the large number of youths receiving health-related benefits, of which mental disorders are the primary cause. This trend has evolved over the last 30years despite recent reforms and has proven to be difficult to mitigate. While in the early 1990s there were twice as many recipients of unemployment benefits as recipients of temporary health-related benefits, the relation today is three to one in favor of recipients of temporary health-related benefits (Fevang et al., 2017). The most recent figures from Statistics Norway indicate that these numbers are continuing to rise. Public expenditure in Norway on social security benefits amounted to 20% of GDP in 2018, over a third of which covers health-related benefits. Activation of work-impaired individuals through VRPs is a major goal of the labor market authorities. Despite this, we know little about how VRPs function, particularly when it comes to work-impaired youths. There are clear indications that youths react differently to activation than adults, and they face quite different challenges. Many young people with reduced work capabilities have little or no work experience. This means that VRPs are highly important as a means to gain the skills and labor market experience necessary to improve and facilitate their labor market attachment. Economic 19 http://www.oecdbetterlifeindex.org/countries/norway. 20 From an international perspective, Norway has very low unemployment rates, and there are indications that some of the disability claims may be “unemployment in disguise” (Bratsberg et al., 2013). Page 20 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 fluctuations clearly affect labor market attachment, more so for young people than for adults and disadvantaged youths than for ordinary unemployed (Barth and von Simson, 2012). Moreover, the distinction between unemployment and disability is rather blurry (Røed, 2012). The heterogeneity of the target group is further exacerbated by the extra uncertainty related to their health, both with respect to the type of diagnosis and the degree of reasonable disability/work capability. When it comes to youth, mental health problems are by far the most important factor related to work impairment, the causes of which can be rather indistinct. WP is the program with the greatest scope in Norway, both today and in the past 30years. Earlier Norwegian studies suggest that educational and training measures work relatively better for people with physical disorders, while those with mental health problems benefit more from participating in work-oriented measures (Børing, 2002; Møller, 2005). Markussen and Røed (2014) study of work-impaired individuals is in line with the above. They recommend early intervention and participation in measures in the ordinary labor market. The exception is youths, who seem to benefit more from ordinary education. We can draw several interesting policy-relevant findings from our analysis. The results show that WS, and to some extent EDU, have the intended effect: work-impaired youths who participate in these measures have a higher probability of obtaining work or starting an education and a lower probability of experiencing a transition to social security than youths who do not participate in any measure. For FU and WP, the results are more mixed. During participation, these two measures are associated with an increased probability of having a transition to work or education and a decreased probability of having a transition to social security. After completion of FU or WP, the increased likelihood of getting a job or starting an education persists. However, the participants are also more likely to have a transition to social security. Recall that activation is a prerequisite for being considered for permanent disability benefits. Our results indicate that these measures work primarily as a screening device to sort work-impaired individuals into those in need of thorough assistance and those in need of a ‘nudge’. The counseling/motivation/mapping on the part of the social worker as well as the experience from the program might be motivating factors driving the youth to search more actively for an ordinary job. Employers may also use the opportunity to sort the people they may eventually want to keep. Hence, push and pull factors may be at play, initiated by the youths, the social worker, and/or the employer. For some workimpaired youths, activation seems to help clarify the need for prolonged rehabilitation or work incapacity. For others, it can effectively counteract the moral hazard problem inherent in social insurance. Mental health problems are the most prevalent condition of the work impaired. The apparent rise of mental health problems among youth in Norway (Bakken, 2019) can also be seen across most OECD countries (OECD, 2018). It is widely documented that mental health problems early in life are detrimental for overall well-being, health, and education, both in the short and long runs (Collishaw, 2015). For instance, estimates show that the Danish state could save ca. 2.8 billion Euros annually (NOK 29.5 billion) if vulnerable children and young people are quickly helped back to a normal life course (Rambøll, 2012). School interventions targeted at preventing and alleviating mental health problems could spare considerable economic and societal resources of youth who are at risk of not completing school and experiencing social and economic exclusion. Page 21 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Abbreviations AIC, Akaike information criterion; ALMP, Active labor market programs; BA, Basic amount; BIC, Bayesian information criterion; GP, General practitioner; HQIC, Hannan-Quinn information criterion; ML, Maximum likelihood; NAV, Labor and Welfare Service; OECD, Organization of Economic Co-operation and Development; ToE, Timing-of-Events; VRP, Vocational rehabilitation programs; EDU, Education/training; WS, Wage subsidies; WP, Work practice; FU, Follow-up; WAA, Work assessment allowance. Declarations Availability of data and materials Restrictions apply to the availability of the data used under license from Statistics Norway for this study. Data are available from the authors upon reasonable request and with permission of Statistics Norway only. Competing interests The authors declare that they have no competing interests. Funding This research is funded by the Norwegian Research Council (grant number 247996). Authors’ contributions The authors have shared responsibility and work equally. KS has estimated the models. Acknowledgments We thank two anonymous referees, seminar participants at the Ministry of Labor and Social Affairs, Oslo 2018, and WEAI, San Francisco 2019 for their helpful comments. References Abbring, Jaap H.; Gerard J. van den Berg (2003): The Nonparametric Identification of Treatment Effects in Duration Models. Econometrica 71, 1491-517. Adamecz-Völgyi, Anna; Petra Zsuzsa Lévay; Katalin Bördős; Ágota Scharle (2018): Impact of A Personalised Active Labour Market Programme for Persons with Disabilities. Scandinavian Journal of Public Health 46, 32-48. Angelov, Nikolay; Marcus Eliason (2018): Wage Subsidies Targeted to Jobseekers with Disabilities: Subsequent Employment and Disability Retirement. IZA Journal of Labor Policy 7, 12. Bakken, Anders (2019): Ungdata 2019. Nasjonale resultater. NOVA rapport 9/19. Barth, Erling; Kristine von Simson (2012): Ungdomsarbeidsledighet og konjunkturer. Økonomiske analyser 5, 2012. Bewley, H.; R. Dorsett; G. A. Haile (2007): The Impact of Pathways to Work. DWP Research Report no 435. Leeds, UK. Brage, Søren; Torunn Bragstad (2011): Unge på arbeidsog helserelaterte ordninger. In NAV-rapport, edited by Arbeidsog velferdsetaten. Bragstad, Torunn; Johannes Sørbø (2014): Hvem er de unge med nedsatt arbeidsevne? Arbeid og velferd nr. 1-2014: 51-63. Bratsberg, Bernt; Elisabeth Fevang; Knut Røed (2013): Job Loss and Disability Insurance. Labour Economics 24, 137-150. Brinch, Christian N. (2007): Nonparametric Identification of the Mixed Hazards Model with Time-Varying Covariates. Econometric Theory 23, 349-354. Burkhauser, Richard V.; Mary C. Daly; Nicolas R. Ziebarth (2016): Protecting Working-Age People with Disabilities: Experiences of Four Industrialized Nations. Journal for Labour Market Research 49, 367-386. Burns, Tom; Jocelyn Catty; Thomas Becker; Robert E. Drake; Angelo Fioritti; Martin Knapp; Christoph Lauber; Wulf Rössler; Toma Tomov; Jooske van Busschbach; Sarah White; Durk Wiersma (2007): The Effectiveness of Supported Employment for People with Severe Mental Illness: A Randomised Controlled Trial. The Lancet 370, 1146-1152. Böheim, René; Thomas Leoni (2018): Sickness and Disability Policies: Reform Paths in OECD Countries Between 1990 and 2014. International Journal of Social Welfare 27, 168-185. Børing, Pål (2002) Varighet av yrkesrettet attføring: Kommer yrkeshemmede arbeidssøkere i jobb? Søkelys på arbeidsmarkedet, 19, 157-167. Page 22 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Caliendo, Marco; Robert Mahlstedt; Oscar A. Mitnik (2017): Unobservable, but Unimportant? The Relevance of Usually Unobserved Variables for the Evaluation of Labor Market Policies. Labour Economics 46, 14-25. Caliendo, Marco; Ricarda Schmidl (2016): Youth unemployment and active labor market policies in Europe. IZA Journal of Labor Policy 5, 1. Campolieti, Michele; Morley K. L. Gunderson; Jeffrey A. Smith (2014): The Effect of Vocational Rehabilitation on the Employment Outcomes of Disability Insurance Beneficiaries: New Evidence from Canada. IZA Journal of Labor Policy 3, 10. Card, David; Jochen Kluve; Andrea Weber (2017): What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations. Journal of the European Economic Association, 16, 894-931. Champaloux, Steven W.; Deborah R. Young (2015): Childhood Chronic Health Conditions and Educational Attainment: A Social Ecological Approach. Journal of Adolescent Health 56, 98-105. Clausen, Jens; Eskil Heinesen; Hans Hummelgaard; Leif Husted; Michael Rosholm (2009): The Effect of Integration Policies on the Time Until Regular Employment of Newly Arrived Immigrants: Evidence from Denmark. Labour Economics 16, 409-417. Collishaw, S. (2015): Annual research review: Secular trends in child and adolescent mental health. J Child Psychol Psychiatry, 56(3), 370-393. doi:10.1111/jcpp.12372. Crépon, Bruno; Gerard J. van den Berg (2016): Active Labor Market Policies. Annual Review of Economics 8, 521-546. Dean, David; John V. Pepper; Robert Schmidt; Steven Stern (2017): The Effects of Vocational Rehabilitation Services for People with Mental Illness. Journal of Human Resources 52, 826-858. Fevang, Elisabeth; Inés Hardoy; Knut Røed (2017): Temporary Disability and Economic Incentives. The Economic Journal 127, 1410-1432. Gaure, S.; K. Røed; T. Zhang (2007): Time and Causality: A Monte Carlo Assessment of the Timing-of-Events Approach. Journal of Econometrics 141, 1159-1195. Hagen, Tobias. (2019): Evaluation of a Placement Coaching Program for Recipients of Disability Insurance Benefits in Switzerland. Journal of Occupational Rehabilitation 29, 72-90. Hardoy, Ines; Knut Røed; Kristine Von Simson; Tao Zhang (2018): Initiatives to Combat the Labour Market Exclusion of Youth in Northern Europe: A Meta-Analysis, in: Malo, M. A.; A. Moreno Minguz (eds.), European Youth Labour Markets (Springer, Cham). Heckman, J.; B. Singer. (1984): A Method for Minimizing the Impact of Distributional Assumptions in Econometric Models for Duration Data. Econometrica 52, 271-320. Heinesen, Eskil; Leif Husted; Michael Rosholm (2013): The Effects of Active Labour Market Policies for Immigrants Receiving Social Assistance in Denmark. IZA Journal of Migration 2, 15. Holland, Paula; Bo Burström; Margaret Whitehead; Finn Diderichsen; Espen Dahl; Ben Barr; Lotta Nylén; Wen-Hao Chen; Karsten Thielen; Kjetil A. van der Wel; Stephen Clayton; Sharanjit Uppal (2011): How Do Macro-Level Contexts and Policies Affect the Employment Chances of Chronically Ill and Disabled People? Part I: The Impact of Recession and Deindustrialization. International Journal of Health Services 41, 395-413. Holm, Anders; Jan Høgelund; Mette Gørtz; Kristin Storck Rasmussen; Helle Sofie Bøje Houlberg (2017): Employment Effects of Active Labor Market Programs for Sick-Listed Workers. Journal of Health Economics 52, 33-44. Kluve, Jochen (2010): The Effectiveness of European Active Labor Market Programs. Labour Economics 17, 904-918. Kluve, Jochen; Susana Puerto; David Robalino; Jose M. Romero; Friederike Rother; Jonathan Stöterau, Felix Weidenkaff; Marc Witte (2019): Do Youth Employment Programs Improve Labor Market Outcomes? A Quantitative Review. World Development 114, 237-253. Kyyrä, Tomi; José M. Arranz; Carlos García-Serrano (2019): Does Subsidized Part-Time Employment Help Unemployed Workers to Find Full-Time Employment? Labour Economics 56, 68-83. Lande, Sigrid; Christine Selnes (2017): Årsaker til sen oppstart i første tiltak for NAV-brukere med nedsatt arbeidsevne. Arbeid og Velferd nr. 2-2017: 49-64. Leinonen, T.; E. Viikari-Juntura; K. Husgafvel-Pursiainen; P. Juvonen-Posti; M. Laaksonen; S. Solovieva (2019): The Effectiveness of Vocational Rehabilitation on Work Participation: A Propensity Score Matched Analysis Using Nationwide Register Data. Scand J Work Environ Health 45, 651-660. Lombardi, Stefano; Gerard Johannes van den Berg; Johan Vikström (2019): Empirical Monte Carlo Evidence on Estimation of Timing-of-Events Models, in: Stefano Lombardi (ed.), Essays on Event History Analysis and the Effects of Social Programs on Individuals and Firms (Department of Economics: Uppsala). Maibom Pedersen, Jonas; Michael Rosholm; Michael Svarer (2014): Can Active Labour Market Policies Combat Youth Unemployment? In IZA Discussion Papers, edited by Institute for the Study of Labor (IZA). Page 23 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Markussen, Simen; Knut Røed (2014): The Impacts of Vocational Rehabilitation. Labour Economics 31, 1-13. Maslow, Gary R.; Abigail Haydon; Annie-Laurie McRee; Carol A. Ford; Carolyn T. Halpern (2011). Growing Up With a Chronic Illness: Social Success, Educational/Vocational Distress. Journal of Adolescent Health 49, 206-212. Muller, Paul; Bas van der Klaauw; Arjan Heyma (2020): Comparing Econometric Methods to Empirically Evaluate Job-Search Assistance. Journal of Applied Econometrics 35(5), 526-547. Møller, Geir (2005): Yrkeshemmede med psykiske lidelser. Tiltaksbruk og effekter. Arbeidsrapport 6/2004. Telemarksforskning. NOU 2019:7. (2019): Arbeid og inntektssikring - Tiltak for økt sysselsetting. Arbeidsog sosialdepartementet. Oslo. OECD (2018): Investing in Youth: Norway. Retrieved from https://doi.org/10.1787/9789264283671-en. Rambøll, (2012): Analyse af de økonomiske konsekvenser på området for udsatte børn og unge. Rapport, Marts 2012. Rehwald, Kai; Michael Rosholm; Bénédicte Rouland (2018): Labour Market Effects of Activating Sick-Listed Workers. Labour Economics 53, 15-32. Riksrevisjonen (2018): Undersøkelse av forvaltning og bruk av arbeidsmarkedstiltak i NAV. Dokument 3:5 (2017–2018). Røed, Knut (2012): Active Social Insurance. IZA Journal of Labor Policy, 1, 8. Røed, Knut; Oddbjørn Raaum (2006): Do Labour Market Programmes Speed up the Return to Work? Oxford Bulletin of Economics and Statistics 68, 541-568. Salvanes, Kari Vea; Rune Borgan Reiling; Astrid Marie Jorde Sandsør (2018): Utdanning som arbeidsrettet tiltak for ungdom med redusert arbeidsevne. Søkelys på arbeidslivet 35, 23-42. Smith, James P. (2009): The Impact of Childhood Health on Adult Labor Market Outcomes. The Review of Economics and Statistics 91, 478-489. Sutterud, Lars (2017): Personer med nedsatt arbeidsevne og personer med rett til arbeidsavklaringspenger. Notat. Arbeidsog velferdsdirektoratet (NAV). van den Berg, Gerard J. (2001): Duration Models: Specification, Identification and Multiple Durations, in: Heckman, J. J.; E. Leamer (eds.), Handbook of Econometrics (Elsevier). van Ours, Jan C. (2004): The Locking-in Effect of Subsidized Jobs. Journal of Comparative Economics, 32, 37-55. Page 24 of 27 Simon and Hardoy. IZA Journal of Labor Policy (2020) 10:13 Appendix A Table A1 Complete estimation results: Transitions to vocational rehabilitation programs (VRP) FU WP EDU WS Coefficient (SE) Coefficient (SE) Coefficient (SE) Coefficient (SE) Completed upper section education −0.048 (0.034) −0.159 (0.022) 0.443 (0.018) 0.142 (0.052) Receiving health-related benefits −0.576 (0.037) −0.461 (0.023) −0.096 (0.023) −2.273 (0.061) Non-European immigrant −0.184 (0.062) −0.210 (0.038) 0.293 (0.035) −0.501 (0.110) Female −0.350 (0.028) −0.168 (0.017) −0.065 (0.016) −1.102 (0.053) Income last 3 years prior to work impairment Nontaxable transfers −0.180 (0.033) −0.059 (0.019) 0.075 (0.018) −0.212 (0.059) Taxable transfers 0.095 (0.029) 0.014 (0.019) 0.002 (0.018) 0.232 (0.047) Labor income −0.075 (0.015) −0.049 (0.009) 0.143 (0.007) 0.058 (0.020) Parents’ income average (7–17 years) Transfers 0.004 (0.007) −0.027 (0.005) −0.016 (0.004) −0.032 (0.012) Labor income −0.004 (0.004) −0.015 (0.002) 0.003 (0.002) 0.010 (0.006) Higher-educated parents −0.219 (0.036) −0.200 (0.022) 0.098 (0.019) −0.330 (0.060) Local youth unemployment rate 0.009 (0.011) 0.022 (0.006) 0.015 (0.006) −0.021 (0.017) Activity year prior to work impairment In education 0.189 (0.032) 0.115 (0.020) 0.273 (0.018) 0.100 (0.053) Employed 0.007 (0.037) −0.103 (0.024) 0.133 (0.021) 0.402 (0.059) Year dummies (2002 reference) Year 2003 −0.372 (0.190) −0.431 (0.092) 0.179 (0.067) −0.693 (0.211) Year 2004 −0.129 (0.174) −0.194 (0.084) 0.051 (0.066) −0.629 (0.192) Year 2005 −0.005 (0.167) −0.359 (0.084) −0.061 (0.066) −0.109 (0.177) Year 2006 0.028 (0.166) −0.252 (0.083) −0.105 (0.066) −0.282 (0.178) Year 2007 0.259 (0.165) −0.355 (0.085) −0.056 (0.068) −0.197 (0.179) Year 2008 0.517 (0.164) −0.126 (0.084) 0.077 (0.067) 0.026 (0.178) Year 2009 1.571 (0.158) 1.073 (0.077) 1.063 (0.062) 0.557 (0.174) Year 2010 1.625 (0.158) 1.326 (0.076) 0.892 (0.063) −0.208 (0.186) Year 2011 1.794 (0.158) 1.464 (0.077) 0.817 (0.063) 0.290 (0.180) Year 2012 1.824 (0.159) 1.586 (0.077) 0.738 (0.064) 0.428 (0.180) Year 2013 1.916 (0.162) 1.609 (0.080) 0.553 (0.069) 0.656 (0.186) Year 2014 2.175 (0.168) 1.690 (0.087) 0.562 (0.078) 0.655 (0.207) Age category (21–24 reference) Age 18–20 −0.177 (0.039) 0.108 (0.024) −0.001 (0.025) −0.273 (0.070) Age 25–29 −0.184 (0.034) −0.204 (0.022) −0.033 (0.019) −0.119 (0.053) Duration of work-impairment spell (1–2 months reference) 3–4 months −0.062 (0.057) −0.100 (0.031) 0.036 (0.028) 0.277 (0.104) 5–6 months −0.081 (0.060) −0.145 (0.033) −0.087 (0.030) 0.397 (0.108) 7–10 months −0.036 (0.053) −0.301 (0.031) −0.148 (0.027) 0.598 (0.096) 11–14 months −0.038 (0.058) −0.286 (0.033) −0.080 (0.029) 0.773 (0.102) 15–18 months 0.160 (0.059) −0.273 (0.036) −0.116 (0.032) 0.966 (0.107) 19–24 months 0.140 (0.059) −0.256 (0.036) −0.149 (0.032) 1.159 (0.106) 25–30 months 0.262 (0.064) −0.303 (0.040) −0.126 (0.036) 1.261 (0.115) (Continued)