“Sometimes the Magic Works. Sometimes It Doesn’t”: A Comment on Chilton and Versteeg

Mark Tushnet1William Nelson Cromwell Professor of Law, Emeritus, Harvard Law School.

* * *

A part of the series, Measuring Impact in Constitutional Law.

Chilton and Versteeg tell an interesting story. I use the word “story” deliberately, to locate the genre of the work in narrative—in their terms, qualitative—analysis, rather than in the scientistic genre. Another way to put it is that they give us a short story in the style of an article in the American Economic Review.2I think it clear that the quantitative analysis is the book’s heart, with brief case studies used to illustrate and sometimes motivate the quantitative analysis. Put another way, Chilton and Versteeg’s reference to Mixed Methods Research is, in my view, largely decorative. The story they think they tell in their quantitative analysis is about the relation between constitutional rights and de facto freedoms, where “constitutional rights” has some substantive referents.3I say this because the discussion of the substantive content of the rights in each chapter would be irrelevant otherwise. The story they actually tell is a bit different, though I believe that their core distinction between individual rights and organizational rights fits into the actual story. That story—that is, the story they actually tell—is about how the presence of magic words in a constitutional document has effects in the real world.

I’ll make a few points about the book itself, then address something I think implicit in the book’s argument.

The first point identifies the actual story by looking at what the quantitative analysis shows us. I have questions about the coding of the dependent variables, but here I’ll focus only on the independent variable. Chilton and Versteeg expressly ignore the presence of limitations clauses, whether general or specific to particular rights. They do so for good methodological reasons,4Limitations clauses have various verbal formulations, and different formulations might (or might not) affect substantive rights differently. Worse, even constitutional provisions that lack accompanying limitations clauses are in fact qualified in many ways, typically through “internal” limitations on the interpretation of the magic words or through balancing the magic words against other magic words. (I note that they treat the provision in the Constitution of Chile, that everyone has “the right to unionize in the cases and manner provided by the law” as unqualified, which seems odd without some discussion of how the term “provided by the law” has been interpreted. And they code Malaysia as protecting religious freedom because of the provision stating, “Every person has the right to profess and practise his religion” without taking into account the qualification attached to that provision, nor the provision, “Islam is the religion of the Federation, but other religions may be practiced in peace and harmony.”) but the choice constrains the stories available to them. Ignore limitations clauses and you’re not measuring the presence in a constitution of constitutional rights with substantive content but rather the presence in the constitution of certain words.

For expository purposes, take an easy example: A constitution guarantees a right to housing within available resources. If you try to connect that guarantee to what Chilton and Versteeg refer to as de facto constitutional guarantees by measuring something about the availability of housing—whether housing budgets, the number of housing units, or anything else—you can’t know anything about the relation, because, for all you know, the “within available resources” qualification means that the substantive right is 100% complied with.5I’m not a methodologist, but I don’t think that including the “within available resources,” “minimum core,” and other doctrinal variants in the robustness checks addresses my concern. These are checks on the back end (I think), whereas the doctrinal qualifications have to be built in on the front end (I think).

The fact that the independent variable is the presence of words, not anything of substance, is the hook for my treatment of the book as a narrative. When do words alone have effects in the world, which words have such effects, and why do they have them? The book’s argument is that words that refer to organizations have effects in the real world, while words that don’t generally don’t. The argument is that words referring to organizations help someone solve coordination and collective action problems, basically because the organizations’ leaders can take advantage of the magic words more effectively than the beneficiaries of individual rights can.

I don’t quarrel with the basic logic here, but I think it worthwhile to bring out the fundamentally formalist popular jurisprudence that drives it. The argument starts with the proposition that organization leaders can capitalize upon saying, “See, the constitution says we have a right to organize (or whatever),” to motivate members to contribute additional resources to organizational self-defense.6“Additional,” because of the “by-product” argument developed by Mancur Olson, about which I want to flag one puzzle: Why do members allow themselves to be exploited by paying more than the organization’s core product is worth to them? I think there are answers—agency slack, mostly—but they are rarely addressed in this literature. The lawyer in me wants to say, “Hold on, you don’t really know that because you aren’t paying attention to the limitations clause.” And the leaders’ response is, basically, “What part of ‘right to organize’ don’t you understand?” That is, words identifying rights define the rights for the organization’s audience even if they don’t do that for lawyers. I’ve called this a formalist jurisprudence, but it’s actually more like a super-formalist jurisprudence, in which a subset of all the words in a constitution have motivating force.

Chilton and Versteeg’s discussion of political parties suggests a slightly different perspective. They write that the existence of the magic words in the constitution signals to political entrepreneurs that organizing a party might be “profitable.”7They don’t put the point in precisely those terms, but that’s the logic of their argument. So, perhaps we should think about religions and unions in the same way: as merely accidentally pre-existing the constitution, but “induced” into a more stable existence by the magic words.8The modern literature on the economics of religious competition seems relevant here. And the introduction of political entrepreneurs raises the possibility that another set of entrepreneurs—call them “public interest” entrepreneurs—could similarly exploit the super-formalist popular jurisprudence to create organizations that would advance nonorganizational rights.9Chilton and Versteeg do note this possibility, but discount its likely effectiveness for reasons, I think, associated with the Mancur-Olsen-based argument that people support organizations that provide them with material benefits unavailable to nonmembers, about which I have some questions noted below.

My next point comes in connection with some of Chilton and Versteeg’s case studies, but it could be raised in connection with codings of dependent variables based upon experts’ substantive judgments. I use the Polish case study because it’s the one I’m most familiar with. The case study is shot through with claims about the unconstitutionality of various actions; at one point the term “plainly unconstitutional” is used.10Similar phrasings are used in the description of the Bulgarian law on freedom of religion. The support for those claims are the opinions of informed Polish insiders—but only a few of them, and as far as I can tell all of whom were opponents of PiS before the actions they’re evaluating were taken. It’s akin to saying, “The Affordable Care Act was plainly unconstitutional” and citing authors affiliated with the Heritage Foundation. At the very least, we need to hear from the government lawyers who developed the arguments that the contested provisions were indeed constitutional—or have some evidence demonstrating that the government believed its actions were unconstitutional and took them anyway.11Chilton and Versteeg note that one of their Polish research assistants was a PiS member, but as far as I could tell we don’t hear anything from that research assistant in the case study.

The general point is that experts about domestic constitutional law almost always have views about the constitution’s meaning that are closely connected to the views of domestic political forces. We know that’s true in the United States and in our scholarly work we typically (I hope) try to take that into account. As I’ve suggested, the difficulty goes beyond the case studies. Codings about de facto constitutional protections based upon expert evaluations suffer from the same problem: The experts upon whom the coding systems rely have political views that almost certainly infect their judgments.12The same problem might occur in connection with the Turkish survey experiment. Respondents were told that “legal experts” had identified views. Perhaps some respondents believed (not mistakenly) that legal experts tended to read their political views (and, in particular, their opposition to the Erdogan government) into their constitutional ones. Those respondents’ answers wouldn’t tell us about the respondents’ views about the implications of what they regarded as constitutional violations. (I suspect that most of the experts are generally sympathetic to the international human rights agenda and tend to take expansive views about the scope of core fundamental rights like freedom of expression.) For several reasons, including personal affiliations, lack of linguistic skills, and unfamiliarity with legal argumentation in foreign legal systems, scholars tend to overlook the point when relying on domestic constitutional lawyers for their assessments of domestic constitutionality. 

Finally, politics. Explaining the point of including nonjusticiable directive principles in the Irish Constitution, one participant in the debates said that they would provide criteria for the public to evaluate the legislature’s performance. Or, more accurately, they would provide criteria for political parties to use in describing their own and their opponents’ programs and performance.

So, consider an issue touched on at the book’s conclusion, in its suggestions for further research. Think of yourself first as someone interested in substantive outcomes—actual freedom of movement (as you define it), housing availability to all, and the like. I think we have a lot of evidence that policies are more effectively implemented when they have support from important political actors, and in particular when the policies are the objects of attention from political parties. So, I think, from a policy point of view you probably should care more about whether a political party has one of your priorities on its agenda than about whether your constitution includes a guarantee of that priority.

The next thing to think about, then, is whether or when a political entrepreneur will build a constitution’s magic words into his or her program. And, although I need to think more about this, my initial intuition is that, from an entrepreneur’s point of view, it’s not that important whether the magic words refer to individual or organizational rights. The political party becomes the organizational vehicle for advancing the individual rights. And, without more, it’s not clear to me why magic words about organizations motivate more effectively than magic words about speech.13It’s not that you’re getting more from the organization when it protects its own rights; indeed, I think the logic of Mancur Olson’s argument is that you’re actually getting less than you could because some of your contributions are being diverted from paying for the stuff you get into paying for the defense work. Think of it this way: An organization that charges a membership fee and uses that fee for two things —a nice magazine and political activism supporting the organization’s goals, such as unionization—should face competition from an organization that charges the same fee and provides a nicer magazine, or from an organization that charges a lower fee and provides just as nice a magazine. Chilton has recounted an anecdote about the luxurious offices of a labor union compared to the spare ones of a rights-oriented NGO. From one perspective, the union’s members are being ripped off by the union’s leaders; from the same perspective they are being ripped off when the union uses membership dues for political action.

Octavio Luiz Motta Ferraz’s study of the effects of constitutionalizing a right to health care in the Brazilian Constitution, probably the most detailed we have, confirms this story. Including the magic words appears to have been enough to create a constituency supporting politicians who promised to improve the health care system and delivered—to the degree that politicians do—on that promise. Motta Ferraz’s story does include organizations like those Chilton and Versteeg discuss: NGOs and interest groups associated with the health care system. They aren’t the primary actors, though; they work with and through political parties to approach their goals.

In this light, the fact that effects of magic words might be indirect and slow isn’t surprising.14See Adam Chilton & Mila Versteeg, How Constitutional Rights Matter 172 (2020) (“[T]he Colombian case . . . is one of slow and indirect progress.”). I note that the Russian case study about religion seems to be one of slow and indirect regression. The book might show that including magic words in a constitution actually does matter—just not right away and with a complicated causal path: not enforcement in the courts nor direct confrontation or negotiation with the government in power, but through political parties or factions within the government in power.15I note that for the examination of social rights, Chilton and Versteeg chose a dependent variable— social spending—rather than outcome measures specifically because “full realization of social rights is a long-term process that is affected by many factors beyond a government’s control.” Id. at 185. I think this could be read to show that the book’s findings result from the choice of what to measure rather than from the properties of a constitution. And in that case, the relevant constitutional provisions might not even include the right to form a political party, but rather the creation of an electoral system that induces a reasonably stable system of reasonably enduring political parties: probably parliamentarism rather than presidentialism, and probably legislative elections under either first-past-the-post with plurality winners, as inducing somewhere between two and four parties,16Or a single-transferable vote system, which appears to have the same effect. or proportional representation with a moderately high threshold, as inducing somewhere between five and ten main parties.17The empirical evidence suggests that 3% is too low and 10% too high.

I end with a thought experiment. Suppose a constitution said only that the legislature will be elected with seats allocated according to a proportional representation system with a 5% threshold. It contains no other rights provisions. Under what circumstances would parties that were committed to rights emerge, and if so, what rights?

* * *

Mark Tushnet is the William Nelson Cromwell Professor of Law, Emeritus, at Harvard Law School.

* * *

Click here to return to the Measuring Impact in Constitutional Law series main page.

Leave a Reply