Which indicators are most useful for comparing citizenship policies? - Thomas Janoski: What we need citizenship indicators for depends on who are “we”

Thomas Janoski: What we need citizenship indicators for depends on who are “we”

January 14, 2011

I’d like to thank the four contributors to the citizenship indicators forum. For many years I felt like I was doing this type of data collection and research in a vacuum. Fortunately, things have changed due to your efforts. Based on the four discussion papers on the EUDO website, I’d like to discuss four issues related to citizenship: (1) naturalization rates, (2) citizenship indexes, (3) explanations of policy or integration, and (4) the of/for distinction or “whom do we want to influence?” question.

First, on the issue of naturalization, what is it exactly? Nationality law covers how strangers become citizens, but many would like to restrict ‘naturalization’ to a specific process. However, depending on the country, many people do not go through the standard process. How do orphans or foundlings become citizens? They do not go through a residence process with a test and so forth. In many but not all countries, spouses do not go through much of a naturalization process. Foreign members of the armed forces and refugees get a highly abbreviated naturalization process. Many states have/had special provisions for neighboring countries (e.g., the Nordic countries gave their neighbors special privileges, and the British seem to give the Irish considerable leeway on citizenship). Technically, naturalization should apply only to adult immigrants coming from what is considered to be a foreign (not friendly neighboring) country. Should one drop the spouses, adoptions, and special cases? This would cut the field of naturalizing citizens down quite a bit, but we would have people who went through the same “pure naturalization process.”

On the other hand, we could expand the range of ways foreigners can become citizens. One rarely encounters a discussion of naturalization that does not mention ius soli. Yet many would drop ius soli citizenship from the discussion of ‘foreigners becoming citizens.’ Maarten Vink finds good reasons for this. In the publication process I went round and round with reviewers of an AJS article about what to call the rate with ius soli adjustments. Many different names were suggested for this wider version of naturalization rates (“citizen acquisition rate” was one somewhat cumbersome suggestions), but for the article, we finally decided on “nationality rate.” In the book, the syndicate did not want an artificial name and asked me to use ‘naturalization rate.’ But in the end, I’m not sure the ‘nominal process of naming a rate’ is all that important. The rate is defined in the text and anyone who reads the paper or book knows what it includes and excludes. The important point to me is “what percentage of those foreigners who acquire citizenship” do you want to talk about? Only those who go through the “full process of naturalization” or “as many of those who acquire citizenship as possible?” In my mind, the explanation of the larger number of “foreigners who become citizens” is more important.

However, I do think that an article or book is needed on the methodology of naturalization rates. Such an article would lay out the different ways that naturalization rates can be computed and their methodological difficulties. It would discuss data limitations. For instance, it is difficult to get the birth rates of foreign populations, and nearly impossible to get the births of foreigners who themselves were born in the country. On this latter point, calculating double ius soli adjustments for a naturalization rate would be extremely difficult for a large number of countries. For some countries it is difficult to get basic naturalization data (e.g., Ireland and Italy). One country ignored numerous requests for naturalization data over a 12 year period (but data did appear from other sources). Such an article or book need not take a position on which measurement is preferable, but would instead outline the ways that many different measures could be obtained, estimated, and/or calculated [1].

Second, in the “kickoff” Marc Helbling discusses new developments in the citizenship literature concerning “citizenship indicators.” He asks whether we need so many, whether they measure the same thing and whether they measure what they intend to measure. On the first question, he concludes that they largely measure the same concepts. On the second question, he indicates that although they largely measure the same concepts, we would do better to separate policy aspects from naturalization rates. And on the third question, he indicates that most studies focus on naturalization rates as outcomes, but he thinks we should focus on rejection rates.

Much of this issue depends upon the question: “Who is “we”?” For a meager attempt at humor, it reminds me of Tonto’s response to the Lone Ranger who says “We’re surrounded by hostile Indians.” And Tonto, who is an Indian, responds: “What do you mean “we,” Kimosabe?” In the case of citizenship, “we” could mean a EUDO “citizenship project” or the EU in general. Should EUDO, OECD, EUROSTAT or the EU in general develop their own citizenship measures or citizenship indicators? In the US, I always have to struggle with the “we” that centers around Hispanic immigration and naturalization as the most important problem that “we” all should be concerned with. Further, should the disciplines of political science or sociology have their own citizenship indicators? Helbling has done some interesting comparisons of the different indicators and they often have relatively high correlations. He is right that more rigorous indexing methods could be applied to these measures. But the type of citizenship indicator will always vary according to the questions a researcher asks and his/her audience (i.e., the “we”). For those who want a general indicator of the integration of foreigners, naturalization and unemployment of immigrants would go together. For those who want to explain naturalization rates, a more restricted measure that only looks at naturalization laws would fit. For those who want to explain how foreigners become citizens, a measure that includes all types of naturalization law including ius soli might do. And some only want to measure the law itself and not correlate it (or assume its relation) to naturalization rates. Further, the data set may vary and this influences what measure one might want to use. Some may look at the historical 18 OECD countries (advanced industrialized countries) and others might want to look at the expanded EU with 33 countries (including the accession candidates and EEA countries but not the Americas or antipodes). Others may increasingly include countries from Central and South America and Asia including Korea, Singapore and Indonesia. These purposes are going to differ according to each research question. I’m not a fan of trying to standardize this measure except where purposes are largely similar. Lawyers see nationality law as infinitely more complex than social scientists do, but on the other hand, lawyers intend to apply the law and not to explain long term changes in countries or regions. Any index like this is an abstraction and selection of various aspects of law. A legal index may infinitely regress toward particularity, but that generally doesn’t help much in the explanation of some outcome. Concerning these questions, I agree with David Reichel that researchers will always need different indicators for different research projects. And this may also vary according to the selection of countries for the data set.
Third, more attention needs to be paid to the explanation of various naturalization rates and policies. I say this even though I think scholarly attention toward data issues is severely undervalued. I can’t help noting that some of the most important evaluators and analyzers of data seem to leave academia for private employment after completing their projects. Nonetheless, explanations are the ultimate result. We should also discuss a variety of research techniques that can be used to explain naturalization rates and laws. With regards to this I was a little puzzled by Reichel’s weak correlation of the MIPEX nationality index and naturalization rates. In my own research with a different combination of countries and years, the naturalization law index (I call it a barrier index) was the strongest and most significant factor in explaining naturalization rates. But despite left party power leading to more open naturalization index numbers, left party power had its own independent effect, which can be interpreted in a way analogous to Bloemraad’s work in the US and Canada (i.e., the other things that state does when it is controlled or influenced by the left). But different measurement approaches will yield different results.

Bringing more rigor to the explanation requires (1) measuring various regime types rather than simply naming them, (2) extending the time period under study and doing it as much as possible in a yearly manner, and (3) trying to combine explanatory analysis using individuals in survey research studies with more national oriented studies. A further challenge suggested by the cosmopolitanism literature might be to collect data by Provinces/Länder/départements and then see how much national policies do or do not explain. Helbling mentions a short test, which I would like to see more of, that uses structural equation models. While I was not convinced by his related discussion that rejection rates were more important than naturalization rates (I don’t deny that they can be quite insightful), the use of unobserved variables from multiple indicators using more components could be quite useful. For instance, LISREL or AMOS could construct an unobserved dependent variable consisting of (a) pure and narrow naturalization rates, (b) best estimates of ius soli births and double ius soli if they can be obtained, and (c) rejection rates. This also suggests an analogy from the strike literature in industrial relations. They look at the “shape” of strikes in terms of frequency (numbers), duration or length of strikes, and intensity in terms of how many people participate. In a similar way, a “shape of naturalization” measure could include the duration of naturalization processes (see Reichel’s discussion of duration), the number of people who are naturalized, and the number of people rejected. The immigration rate could also be included because higher naturalization rates may induce greater immigration. These are some suggestions that are not fully worked out. But they do point to some approaches to explanation and naturalization that could be helpful to research in the future. Finally, one can more fully use pooled time-series analysis for this type of data, while continually working on sensitivity analysis. I use this term to mean testing relationships while removing/including one country at a time to see what their impact is on the results one tends to find.

Fourth, Jan Willem Duyvendak, Rogier van Reekum, Peter Scholten and Christophe Bertossi discuss the of/for distinction. I take their statement that “the rationale of sociology for policy is to generate knowledge that can be rationally applied to policy development in order to reach more desirable results in terms of certain indicators” to mean that this approach takes a value position on the policy. Often this might mean more integration is desirable as opposed to segregation, or in the US, more immigration of Hispanics is desirable. Or it could be to expound upon the virtues of one country vs. another (e.g., methodological nationalism). The sociology of policy seems to be a more dispassionate explanation of the causes of policy with less regard for how these results might be perceived by policy-makers. I agree that mixing up these two approaches can result in some confusion. They conclude that the sociology of policy should be the focus of academic pursuits. Their point is well taken, but we will continue to see both approaches in different research projects, and also mixed up in other projects. In some ways this is unfortunate but inevitable. We just need to be careful in distinguishing the two approaches in the literature and at conferences.

In conclusion, I would like to encourage working out these first two data issues in articles or perhaps a conference.

Notes

[1] One side note, Vink’s example of bias in calculating rates due to ius soli could be generalized to the overall naturalization process. For instance: Two countries have 1,000 resident foreigners, and A has double the naturalization rate of B (200 vs. 100). After year one, with an additional immigration of 100 immigrants, country A would have a foreign population of 900 (1000-200+100) and country B would have 1,000 (1000-100+100). If both countries A and B naturalized the same number of people (say, 100) in the next year, then country A would have a higher naturalization rate than country B. Thus, more open countries have a bias toward higher naturalization rates, and more closed countries have a bias toward lower rates, mainly due to developments in their foreign populations.