Appendix A – Operationali­sation of barriers

In this section we further describe how we have operationalized the barriers from the framework over employment barriers in the Nordic countries. The structure of the section follows the four barrier categories from the framework.

One important employment barrier from the is health issues. In general, we have seen a rise in self-reported sickness or disability among the working population in both EU- and OECD-countries, which both can be an objective increase or due to greater awareness or better diagnostics (Fernandez et al., 2016). Irrespective of the reason, health issues do constitute a barrier to the labour market and probably to a greater extent compared to 20 years ago.

This barrier can be operationalized and measured in several ways in EU-SILC. For example, the individuals in the survey are asked about their self-perceived health status, whether they have any chronic health conditions and whether they are limited in daily activities due to on-going physical or mental health problems. Depending on the measure adopted the results can vary considerably. For example, chronic health conditions include hay fever, which to a very little degree affects people’s ability to work and therefore this does not capture what we are interested in measuring. What we are interested in measuring are health problems that hamper the individual’s ability to do activities that people usually do. Therefore, we follow Fernandez et al. (2016) and define individuals to be facing a health-related barrier if the individual is either limited or strongly limited in daily activities due to an on-going physical or mental health issues. Unfortunately, EU-SILC do not allow us to distinguish between mental and physical health barriers, which could have strengthened the analysis, since the literature (see for example Bredgaard & Shamshiri-Pedersen, 2018) points to the fact that individuals with mental health problems do have the lowest employment and labour force participation rate among persons with disabilities. This indicates that individuals with mental health problems might face other barriers compared to individuals with physical health problems due to for example different perceptions or lack of knowledge about mental health problems compared to physical health problems.

Lack of general skills have been identified as an employment barrier in our framework in various ways and can also be a result of different factors. Lack of skills can be the result of no or very little education, which to some degree limits individuals to participate in today’s demanding labour market. A typical way to define no or very little education, which we also adopt in this report, is to define an individual to be facing an educational barrier if the individual has no competence-giving education. We use the question about highest attained education, and if the individual has no more than upper secondary education according to the ISCED-2011 standards the individual is said to face an employment barrier related to lack of education.

Besides lack of education, lack of skills can also be the result of little work experience since work experience enhances and maintains work-related skills (both technical and social) and plays an important role in explaining different labour market outcomes among individuals with the same educational attainment (Fernandez et al., 2016). In EU-SILC experience in the labour market can be measured in a number of ways and we define two barriers related to lacking experience in the labour market.

The first barrier related to experience in the labour market refers to recent experience. If an individual has not worked one month during either the reference year or at the moment the interview took place the individual is said to face lack of recent work experience, since the individual has been out of the labour market for more than one year.

The second barrier related to experience in the labour market refers to whether the individual has any work experience. In EU-SILC the respondents are asked whether they have ever worked. Note that this does not include e.g., vacation jobs taken by students, casual jobs undertaken from time to time, etc. If the respondent answer that they have never worked, they face the experience related barrier never worked.

Lastly, lack of skills can simply be a result of having low work-related skills that to a lesser extent e.g., are demanded in the Nordic labour markets due to for instance automation. The respondents in EU-SILC are asked the skill level of either their current job or their most recent job. The skill level is measured using the ISCO-08-taxonomy, which is an internationally recognized classification system that categorizes jobs based on the skill level and skill specialization required for the job. We define individuals to be facing this barrier if their most recent job was in one of the lowest two ISCO-08 occupation categories, which refers to plant and machine operators and assemblers as well as elementary occupations.

Care responsibilities is another barrier that are identified as a barrier to participate on the labour market. The level of detail in EU-SILC allows us to construct an indicator, which convincingly can identify individuals who face a barrier related to care responsibilities. We can identify households where children below 13 years old are present and whether these children are enrolled in formal care services (e.g., kindergarten) for at least 30 hours per week. If the children are not in day care for at least 30 hours per week the child needs care at home. Based on this, we define individuals to be facing a barrier related to care responsibilities if a) there is only one potential caregiver in the household, or b) there is more than one potential caregiver, but one of them reports to be inactive or working part-time because of care responsibilities.

Weak work incentives can arise when income gain from taking up a job or working more is limited, because net wages are low or because generous out-of-work benefits are withdrawn as people start to work. In addition, people may decide to limit their work effort because they have access to other income sources.

A proxy for the latter can be derived straightforwardly from the EU-SILC variable ‘gross household income’ (which includes pre-tax income from labour and capital plus government transfers) minus own income that depends on the person’s own work efforts (i.e., employment income and

earnings-replacement benefits, such as unemployment benefits, of the person of interest) and minus a share, depending on the number of adults in the household, of social transfers awarded at household level (for instance, social assistance or rent allowances). We derive this indicator in equivalised terms to account for differences in household size.

The income independent on own work effort are then used to create two group of individuals using a threshold of 1.6 times the median in the reference population. If an individual has a value above this threshold the individuals are defined to have a high partner or non-labour income.

Another incentive related barrier that can be derived from EU-SILC is to an approximate measure of the extent of benefit reductions that an individual is likely to experience when taking up full-time employment. To do this, we use the ratio of the amount of earnings-replacement benefits received at the individual level in the numerator, and own potential wage in the denominator.

The amount of earnings-replacement benefits at the individual level is constructed as the sum of individual earning-replacement benefits plus the individual’s part in any earnings-replacement benefits received at household level.

The shadow labour income is estimated using a Mincer-equation with the two-step Heckman approach to account for endogenous sample selection. The variables that enter the Mincer-equation are education, age, age squared, gender, health limitations and degree of urbanization. The participation equation uses a number of additional variables including family structure, number of children in the household, household income that are independent on own work effort. The estimated parameters are used to derive estimates of shadow labour incomes.

The resulting variable is discretised into a binary indicator that takes value 1 for ratios equal or higher than 60 pct. In other words, if more than 60 pct. of an individual’s potential in-work earnings are “taxed away” when taking up employment the individual is said to face an employment barrier related to high earnings replacement (benefits).

In the framework over employment barriers one barrier that hinders individuals from participating on the labour market is the lack of local employment opportunities and further the state of the economy can also constitute a barrier for some vulnerable groups. For example, a paper from Sweden shows that differences in the employment structure of the local labour market contributed to differences in the native-refugee employment gap (Engdahl & Liljeberg, 2022).

However, such a variable can be challenging to define at the micro-level, since this variable essentially must describe the availability of vacancies in labour-market segments that are relevant for each individual given their skills, location, age, etc. Following Fernandez et al. (2016), we operationalise this variable by estimating a linear risk of demand-side constraints in labour market segments described by age, gender and education on long-lasting unemployment and persistent involuntary part-time activity, which to some degree work as proxies for the labour market tightness in a given area. For example, if there are many individuals with specific characteristics who are involuntarily unemployed in an area, it means that individuals with those specific characteristics in that area have a higher probability of being demand side constrained. Note that we use Eurostat’s Labour Force Survey in the respective Nordic countries to get the parameters and then we use the parameters to estimate the probability of being demand side constrained for the population in EU-SILC. Those with a risk higher than 1.6 times the median in the reference population is defined as facing the scarce employment opportunity-barrier.

In this section we will describe and operationalize barriers related to public employment services. As already mentioned, these barriers relate to how successful the public employment services are towards their clients. For example, vulnerable young people are often in risk of being left out/fall between chairs due to lack of coherence and coordination among the many different institutions involved in assisting the youths (Bolvig et al., 2019)

Variables related to the public employment services can be difficult to identify on the micro-level (which is required to make a meaningful Latent Class Analysis), since the public employment services inherently/typically are founded on the macro-level. For example, ‘cream-skimming’ (i.e., services prioritizing clients with the best employment prospects) can be the result of limited resources and wrong-designed incentive structures at the national level. On the other hand, cream-skimming could also be the result of a single caseworker maximizing own performance. Neither the reason, it is difficult to identify this sort of barrier in data, but it should still be recognized as an important employment barrier and especially among the most vulnerable on the labour market.

Therefore, we have used questions from Eurostat’s Labour Force Survey to form a barrier related to the contact or lack thereof with the public employment services. We have formed a variable that reflects the extent to which job-seeking individuals have been in contact with PES. Job-seeking individuals is asked how they so far have searched for a job and we use the ones related to contact with the public employment services (MethodA and MethodK in LFS) to form a dummy variable for whether job-seeking individuals have been in contact with the public employment services.

Next, we estimate an equation that uses various background characteristics (age, gender, education) to predict the probability of an individual having been in contact with public employment services. If an individual's expected contact with PES is below the 20th percentile, the individual is defined to face a barrier related to low contact with the public employment services (PES).