Imbens on DAGs, and the Pedagogy of Causal Inference

Guido Imbens has an interesting new essay on the graphical causal modeling approach pioneered by Judea Pearl, which uses directed acyclic graphs (DAGs) to understand how to infer causal relationships from data. Imbens is a pioneer in applying the potential outcomes (POs) framework in economics to study causal questions. Although authors such as Morgan and Winship see DAGs and POs as complementary, many proponents of DAGs hold that DAGs are superior to POs, for purposes of clarity of exposition, generality, and others. This debate can get pretty heated, with allegations that economists (in particular) are stupid, arrogant, or somehow afraid of DAGs.

Imbens’s new essay presents an alternative view, on which which is charitable to DAGs but does not concede the point that they are generally superior to POs. His discussion of questions of manipulability (essential in POs, not so in DAGs) raises familiar questions from the philosophy of causality that go back to Lewis (1973) on counterfactuals and causation. So too with simultaneous causation. But his discussion of techniques such as instrumental variables from the DAG-vs-PO perspectives is, to my knowledge, novel. He writes

some of the key assumptions in instrumental variables settings are not naturally captured in DAGs, whereas they are easily articulated in the PO framework. This extends to other shape restrictions that play an important role in economic theory. Second, one of the modern results in instrumental variables settings, the identification of the Local Average Treatment Effect (LATE, Imbens and Angrist [1994], Angrist et al. [1996]) is not easily derived in a DAG approach

This is important because it directly challenges the argument that “DAGs are better are whatever POs do” (the DAG > PO argument). It also threatens the more general argument that “DAGs can do whatever POs can do” (the DAG == PO argument).

Imbens’s conclusions resonate with my own from my experiences teaching causal inference to beginners. I insist that DAGs are wonderfully helpful for teaching causal inference in a context that focuses on identification rather than estimation and inference. They are, in fact, absolutely essential for illustrating concepts like collider bias, which I cannot even express properly without a DAG. The benefit of DAGs is that they allow students to understand the logic of causality without requiring statistical foundations.

But in my experience, and again consistent with Imbens’s argument, this does not extend to instrumental variables and related identification strategies, largely developed in the field of applied economics. Here, I find DAGs often unhelpful. Take Imbens’s Figure 3:

I remain unable to articulate to my own satisfaction why this DAG is a solution to the problem of identifying the effect of X on Y without either invoking a range of additional concepts drawn from the PO framework or making other assumptions, as in Morgan and Winship, pp. 297-299.* And certainly I cannot explain how instrumental variables work or what effect they identify without such concepts drawn from POs. Even Pearl cites Imbens when presenting what one can learn from instrumental variables in his canonical text on DAGs (see Causality, Second edition, p. 90), or cites economists who managed to derive the IV formula without the assistance of DAGs back in the 1920s (see pp. 152-153), although a more typical expression of the DAG > PO argument is on pp. 247-248, especially note 29.

Imbens notes that others working in the graphical causal modeling tradition have recently proposed DAGs for instrumental variables designs, e.g. Steiner et al. (2017).

Imbens “do[es] not find these DAGs particularly illuminating” (challenging the DAG > PO argument). I will take a stronger position: I suspect that Steiner et al. could not have derived their Figure 5 without knowledge of the PO framework. This challenges the DAG == PO argument. Similar arguments about the expository limits of DAGs are probably also valid for designs such as regression discontinuity and others, although the instrumental variable case is the one that springs to mind given personal experience.

The conclusion I draw from this reflection on Imbens on DAGs is that students need many tools. DAGs are a remarkable nonparametric identification algorithm, and for some pedagogical purposes they are irreplaceable. But not for all, and at least some of the conceptual insights derived from POs are irreplaceable.

NOTE

* I will note that once students have a firm command of instrumental variables, DAGs are helpful for representing the assumptions that are implied by such designs. They are also very useful for explicating how “conditional” instrumental variables designs work, such as Morgan and Winship’s Figure 9.2(b).

What Does it Mean to be “Western” in the Netherlands? [Updated]

Some time ago I wrote about the curious case of Japanese colonial subjects in the Dutch East Indies. The interesting observation is that

After concluding a Treaty of Trade and Navigation in 1896, the Netherlands and Japan recognised each other as most favoured nations, and subsequently the Tokyo government pressed the Dutch to accord its migrants in their colonies the same legal status as Europeans. In 1899, Japanese citizens in the Indies acquired European status…

Today I learned two more interesting facts about the regulation of identity in the Netherlands today, and how the Netherlands government conceptualize what it means to be from the West.

First, when the Netherlands government conceptualizes migrants as being from the West or not, it continues to follow the precedent set in the Indies back at the turn of the last century.

Due to their socioeconomic and cultural position, people from Indonesia and Japan residing in the Netherlands are considered as having a ‘western’ migration background. These are mainly people born in the former Dutch East Indies and expatriates employed by Japanese companies with their families.

Not only does the Netherlands government maintain the idea that Japanese are “not like” other non-Europeans, it also applies that standard to anyone from Indonesia. It is somehow meaningful that this principle of “the former colonial subject is Western, like us” does not apply to, say, Suriname, although I’ll leave it to a specialist on Netherlands politics to explain how that has come to be and what it means.

Second, what it means to be a cultural minority in the Netherlands:

Student from a cultural minority defined by the Dutch Ministry of Education as someone meeting one of the following criteria:

  • belongs to a Moluccan group;
  • at least one parent or guardian originally comes from Greece, Italy, former Yugoslavia, Cape Verde, Morocco, Portugal, Spain, Tunisia or Turkey
  • at least one parent or guardian originally comes from Suriname, Aruba or the Netherlands Antilles;
  • at least one parent or guardian originally comes from another non-English speaking country outside Europe, except Indonesia
  • at least one parent or guardian was admitted as a foreigner under article 15 of the Aliens Act.

Here we see a bunch of knotty problems in distinguishing among former colonial subjects. Someone from the Moluccas (Maluku) is from a cultural minority background, but not someone from elsewhere in what is today Indonesia, like someone from Java. Unless that Javanese person went to Suriname first after slavery was abolished there, in which case that person would be from a cultural minority background.

In the Netherlands, as in elsewhere, we learn a lot about politics and society from the categories that we use to regulate identity.

UPDATE

Another day, another thing learned. There are plenty other ways that peoplehood and citizenship are regulated in the Netherlands (and, of course, elsewhere too, but the Netherlands example is the one that fascinates this Indonesianist). Take, for example, the requirements for obtaining a temporary residence permit to live in the Netherlands. Many migrants have to take a Basic Civic Integration Exam. But not all of them!

You do not have to take the exam in one of the following situations:

  • You are under 18 or you have reached your AOW pension age. The AOW pension age differs per person.
  • You have a valid residence permit and you want to change the purpose of stay.
  • You have lived in the Netherlands for at least 8 years during the compulsory school age.
  • You have the nationality of an EU/EEA country, Australia, Canada, Japan, New Zealand, South Korea, Switzerland, United States of America, Vatican City.
  • You have the status of a long-term resident EC in another EU country.
  • You have the Turkish nationality or you are a family member of a Turkish national that has lawful stay in the Netherlands…
  • You have the Surinamese nationality and have at least finished primary school in the Dutch language in Surinam or the Netherlands…

Whereas the examples above show that Indonesians are viewed as having a Western background rather than a non-Western one, when it comes to civic integration, Indonesians are treated like other non-Western migrants—but Surinamese who are educated in Dutch are not!

There is a fascinating book to be written on the politics of citizenship in the Netherlands and Indonesia. The mutual co-constitution of Dutch and Indonesian ideas of citizenship, from the colonial period through today.