Measuring discrimination with a questionnaire?

A new report from Timbro, the Swedish right-wing think tank, has caused some commentary. The author argues that discrimination does not explain why immigrants face challenges on the Swedish labour market. This argument rest on the observation that employers in other countries (Canada, Denmark, Germany) answer a question from World Value Survey concerning privileging native born job seekers over foreign born such more negatively than Swedish employers do.

While I have to admit that I was surprised that Canadian employers would be presenting such answers to the question, the report, in my opinion, has a fundamental methodological flaw, and that is its reliance on the World Value Survey for its conclusions. This survey is a quantitative survey of values, taken in from across the world (as the name implies). It could be argued the aggregate data of the Survey gives us some idea about cultural differences across the world (though it is possible to question that, too, to some extent - after all, how do you develop distinctly separate categories of "traditional" and "secular-rational" values?) but I'm not sure that it is a useful in a smaller sample studie like this.

More importantly, it is inherently quantative, and I'm not at all convinced that quantitative methods can reveal a whole lot about discrimination in practice. At best, the report manages to show that Swedish employers have a self-image of not being discriminatory, but from there, it's a bit of a leap to conclude that they will not act in a discriminatory fashion in practice. Indeed, unconscious discrimation, i.e. acts that are not intended as discriminatory, but result in discriminatory outcomes because of ignorare, or prejudice that is assumed to be common sense, cannot be easily probed using a questionnaire.

Qualitative methods, like in-depth interviews, participant observation and so on would be more appropriate, and every time such methods have been employed in Sweden, significant discriminatory behaviour has been observed.

Timbro-seminar 2: Subsidy, the road to work?

The second Timbro seminar I watched concerned the significant discrepancy in employment between immigrants and the native born population. This discrepancy is hardly new, but has persisted for at least as long as I've been actively studying the field (i.e. since 1998). To the best of my knowledge, it started sometime in the mid 80s. All the same, it is good to give it continuous attention, since it is a solid indicator of segregation.

Von Bahr, the author of the report presented in the seminar, discusses different explanatory variables that have been addressed in the public debate. She argues that discrimination is hardly a convincing such, since a survey conducted for the report showed that Swedes are more tolerant than many other populations (including, for instance, the Norwegians). Nor is there a lack of interventions. In fact, Sweden is known for spending heavily on programmes and training to integrate newcomers into the labour market. Rather, one should look to general labour market policies, like wage levels and work security. Since these are both very high, internationally speaking, there are some signifcant thresolds for immigrants to get over before they can enter the labour market. I.e, the entry level sustenance jobs that are found in many other countries (including Canada) do not really exist in Sweden, which creates an impediment for labour market integration.

In short, Timbro is saying pretty much the same as it has been saying for the past decades - reform work security legislation and lower pay roll taxes, and integration will improve.

Now, there is some truth to these arguments, I'd say. For instance, in a labour market where "the last one in, the first one out" is a regulatory regime at lay offs, new people (i.e. immigrants, youth and possibly women returning from maternity leave) in the labour market will, on a structural basis, find it harder to get entrenched there. This much I can agree with.

However, there are several problems with the presentation, as well (again, I have yet to get around to read the full report):

1) Discounting discrimination as problem is difficult to do based on the presented material. Opinion surveys, it seems to me, are probably a weak instrument at best to capture prevalence of discriminatory practices on the ground. This was wisely commented on by one of the commentators on the report, Ardalan Shekarabi, who pointed out that if you ask people if they are tolerant, they'll likely answer "yes", which says little about how they will actually conduct themselves when confronted with a situation where stereotypes and essentialist ideas about "others" come into play. This is probably particularly true in Sweden, where being tolerant is very much a deeply entrenched part of the Swedish self-image. What is needed, instead, is participatory observation and situational testing, and I'm not aware that anyone has conducted such studies in a comparative fashion between countries. Those that have been carried out in Sweden during the past decade consistently show that discriminatory behaviour does exist, both in employment situations, as well as when seeking housing, for instance.

2) Large sums of money are devoted to providing newcomers with language training and labour market interventions, this much is true, but being satisfied with this assumes that outcomes can be directly correlated to spending levels, and I'm not aware of any public management literature who would find this assumption convincing. In fact, the Swedish public administration has been consistently criticized for not knowing if the labour market policy administration delivers any measurable outcomes at all, both for the population in general and for immigrants. The past decade has even seen a lot of critique directed towards Swedish as a Second Language courses. I even believe that Timbro as sometime was part of that choir. It's strange that this dimension seems to have been left out in the report.

3) Entry level jobs are often a necessary first step for newcomers, this is true. But Swedish labour market statistics seem to stop measuring labour market success at that point, seemingly content as long as an immigrant has gotten a job, any job, regardless of the immigrant's previous work experience or educational level. That is hardly very satisfactory. If a great number of highly skilled people are doing work below their levels of competence, there is still a significant integration problem, and I'm not at all convinced that this would be addressed by simply lowering the payroll taxes.

Timbro-seminar 1: Who should be able to become a Swedish citizen

Timbro, the Swedish equivalent of the Canadian Frasier Institute (i.e. a right-wing think tank), is holding a series of seminars on integration, presenting reports they have solicitied on the subject. Dr. Johansson Heinö was the first presenter (followers of this blog will recall that he also presented on the Axess seminar).  This was a follow up on his previous report "Integration or Assimilation".

This time, Johansson Heinö presented a proposal to improve integration, through addressing the status of citizenship. Sweden has, by international standards, a very low requirement for citizenship. In practice, all you need to do is to fulfill a certain period of residence. There are no tests, nor any oath of allegiance. This, evidently, has some concerned about the social cohesion of the country, arguing that this devalues the status of citizenship and undermines pride and a sense of belonging. At this point, I can't recall that I've seen any research to support that claim.

Johansson Heinö's remedy is to introduce a language test and make citizenship dependent on the applicant being fluent in Swedish.

While I'm not as opposed to this proposal as some others might be, there are still a number of reservations that should be brought up (note that this is based on watching the webcast seminar, I'll have to read Johansson Heinö's report to make further commentary):

1) The connection to improved integration is not very well established. MIPEX ranks Sweden as very succesful in the field of integrating immigrants (which I take exception to because it only studies formal processes) and does so specifically because citizenship is so easy to access. By this rationale, if citizenship is harder to attain, the integration process will actually become impeded. Since research shows that the level of integration is correlated to processes of naturalization, this is actually quite a convincing argument.

2) The connection between how difficult the citizenship is to attain and the sense of belonging is also not established. From what I recall from other studies, sense of belonging tends to be more correlated to immigrants' experience of how just society is in terms of equal opportunities and protection from discrimination as well as accumulated social capital within immigrant communities, so this suggestion might simply be addressing the wrong dimension of the process.

3) What do to with elderly immigrants? I seem to recall that older people have a harder time learning a new language, and it's quite established that many elderly immigrants never learn the language of the new country at all. This suggestion might very well become a barrier for that category of people.

4) Why addressing language, particularly? I'm asking, because the citizenship test in Canada is based on learning civics and history, which seems more conducive to creating the capacity to orient in the new country, including topics like how to vote and the basics of the democratic system, which are functions a citizen will need to know about to be able to participate fully in civic life. Learning about the common points of reference of the host society just seems, on an intuitive level, more conducive to increasing a sense of belonging than a test of grammar. Note that the Canadian test is written in one of the two official languages, so some level of linguistic ability is implicit, but testing history knowledge is quite different from testing vocabulary.