What’s in a Name: The Case of ‘Metrics

While it is hard to think of an econometrics book being popular in any traditional sense of the term, the best candidate for a popular econometrics textbook is Josh Angrist and Jörn-Steffen Pischke’s Mostly Harmless Econometrics (MHE). The title is a deliberate reference to The Hitchhiker’s Guide to the Galaxy, and communicates the irreverent tone taken in the handbook. While it came out a bit before I was aware of such discussions, it seems like the Spring 2010 issue of the Journal of Economic Perspectives (free to read online) seems to be dedicated at least in part to the approach the book takes (though almost certainly these were in practice before the book) and responses to it. While MHE does not concern itself too much with theory, it is a good companion to an existing textbook, or may be of interest to someone who is interested in applied work, though it demands a certain level of sophistication from the readers (I’d say it’s probably targeted to an advanced undergraduate or introductory graduate level).

This year saw the release of the pair’s second book called Mastering ‘Metrics (MM) which is geared more towards an undergraduate level. On the first page of the introduction they say “Economists’ use of data to answer cause-and-effect questions constitutes the field of applied econometrics, known to students and masters alike as ‘metrics.” When the book was released I was taking my 3rd course in econometrics and found that it was the first I had ever heard of the abbreviation. Now in the graduate program and taking my 4th course in econometrics I find my classmates referring to it as ‘metrics, though, curiously, those who use the term most readily (and originally from my understanding) were my classmates at UBC who received the same econometrics education that I did and had not used the term before. Reading MHE or MM will tell you how I could design a project to test whether or not the release of the book led to the adoption by my classmates, but I will state without evidence that I believe the release of MM marks the coinage and widespread adoption of the term ‘metrics. (Edit: not entirely accurate. See below)

Is there anything we can make out of this? As always, I’m interested in reading too far into things, but I think the title of the book indicates positive things for the social sciences and Angrist and Pischke’s audience, though does not communicate anything favourable about my peers. MHE contains an endorsement from James Robinson (probably best known for his collaboration with Daron Acemoglu on Why Nations Fail) declaring it a “must-read… [for] political scientists, sociologists, historians, geographers, and anthropologists.” He himself seems to stand astride economics and political science, though his collaborations with Acemoglu unquestionably fall under economics. My interest in the abbreviation to ‘metrics in the undergraduate level text (MM) is that I think it communicates to open up these extremely helpful tools for causal inference to a wider range of disciplines. I am too early in my academic journey to be able to speak with much authority on this, but econometrics appears to offer one of the best sets of tools to answer these kinds of questions, and they may find ready applications outside of economics itself. For instance, my undergraduate thesis involved taking a data set from a political science paper and in a first pass I simply replicated the results then went in and adjusted where I felt the methods were inconsistent with my understanding of how to work with such data, which immediately resulted in stronger results (later, it also identified some areas in which the other paper was weak and possibly the results were heavily worked to promote a particular conclusion). My own result wasn’t necessarily impressive in any respect (for instance, there are likely endogeneity problems and, well, I really don’t know what the hell I’m talking about), but I think it at least some evidence that good questions can find good answers if researchers are willing to look at the economist’s toolbox which seems to hold some of the most advanced techniques.

The idea here is that the abbreviation is a good one because it removes the ‘econo’ element and communicates that the techniques (the ‘Furious Five’ as Angrist and Pischke call them) are not limited to questions in economics, but have more general applications in conducting any kind of causal inference. This isn’t to limit economic inquiry (the great appeal to me at least is the great flexibility economic analysis affords me), but rather to be more inclusive in the terminology. It doesn’t really change anything, but it does avoid the hang up of, say, a sociologist taking advantage of something like quantile regression, explaining it’s a helpful econometric technique and then having to answer “what does economics have to do with any of this?” (or worse, dealing with the assumption that economics is somehow tainted by unreasonable assumptions and thus the technique is invalid). Basically, if we have social scientists using the best tools available (at least so far as I’m aware), then we, as a whole, benefit. A common language between disciplines will allow for easier collaboration, and, rather than hoping that economists have all the good ideas, disciplines with other interests can take advantage of these tools to improve their research (this, of course, makes it more difficult for researchers if the overall quality of work improves, but this is a nice problem to have. I won’t cry if there are fewer papers with results that collapse with a minor change in assumptions). Of course, to gain the full benefit of the tools available, researchers should take up MHE, and maybe move on to something like the Handbook of Econometrics (behind a paywall but possibly available to you if you went to a university that gives alumni access to academic journals) which requires getting over the red herring of what we call things, but I don’t mind dropping ‘econo’ in the introductory material if it means we’ll benefit from better research.

My enthusiasm for the rechristening to the term ‘metrics is somewhat diminished when I hear my peers use it. There are a couple of PhD students who use this term, but I notice a fairly high rate of adoption amongst the MA students. I can get the impulse, the hope to communicate that one is hip to the latest trends in the profession. It’s a bit problematic when you consider that the term seems to have gained currency when an undergraduate level book has been released, but never mind that, only squares who use words like ‘hip’ (and square) call it econometrics, all the cool kids call it ‘metrics. My problem is that I actually think taking the econ out of the MA economics cohort is actually a fairly accurate assessment of the situation. Here’s an example: I spent a few minutes this afternoon writing in the discussion for wiki entry on the protestant work ethic because I noticed a claim made about Schumpeter’s account of the origins of capitalism that I’d not read before. Following the reference did not present any support for the claim in the entry, and so it raised a few interesting questions for me: Did Schumpeter ever write on the origins of capitalism? If so, where? If not, where might this impression have come from (it seems similar to Marx’s account, but not enough for me to want to make an edit)? What I would like to be able to do is to raise any one of these topics with a classmate and investigate it (it might be a short conversation because I’ve only read a little of Capitalism, Socialism and Democracy but I’d find it preferable to find myself understudied for conversations with my peers). The most common conversation will be either how to find problem set solutions or how to land a good job after grad school. In fact today a student more or less said they haven’t really understood what’s been said in the last two econometrics courses (not classes) and have just focused on how to solve the problems. It’s hard to see the class as particularly interested in economics so much as building the appropriate ‘signal’ to employers that they’re worthy of a high paying job.

Again, I sort of understand the impulse. Everyone has to make a living, and we prefer a high paying job to a low paying one. Likewise, econ is a decidedly employable degree, and a BA has been reduced to the point that it has become a requirement to rent cars to people at Budget. But that’s not the reason why I’m in the program, and in the end, I think advanced study in any subject should be more than just finding a good job. To me it’s a problem that I cannot have a conversation with an MA student about a topic in economics if it is not directly related to the grade they will be getting in the end. It gotten to a point that I argued with a classmate who was complaining that we weren’t permitted a formula sheet for the mathematics final as being ‘unfair’ because they were planning on writing the solutions to past finals. In addition to pointing out the dubious application of the term ‘unfair’ my position was that the purpose of the course is to teach us not just mathematical techniques, but reasoning (ie. how to do proofs), and that rewriting a past answer (which is actually more successful than you’d think) is simply imitation, not understanding. For this I was perceived as being the unreasonable one because the material was ‘hard.’ The problem is that I know it’s hard, because I struggled with it and had to write the same exam. Topology defines open and closed sets in a way that allows sets to be both open and closed at the same time, confirming my suspicion that they are deliberately trying to make the subject even more difficult than it already is. But it’s economics, and mathematics is the tool and language we use to bring clarity to the problems we work on. We don’t study topology to impress the ladies (though, ladies, you know where to find me), we study this because that branch of mathematics allows us to prove certain propositions fundamental to our analysis. A working understanding of mathematics not only allows us to understand the fundamentals of our discipline, but equips us to handle our ultimate goal: answering questions that nobody has gotten the answer to yet (or, ideally, haven’t even asked yet!). You can’t imitate your way to that, and it requires a lot of hard thinking both mathematically and creatively. A technically perfect solution to an uninteresting question is at least useless as a poorly formed but interesting question (although, in theory, the latter can be picked up by someone with the capacity to do the heavy lifting). Economics problems are worth taking the time to do right, and nobody ‘owes’ us our degrees. Either we have the tools or we don’t. It sucks for me, because I know I’d love to have free time to stream, and read for pleasure, but I also want to be able to formulate and answer these questions head on without hiding behind “well I haven’t learned <topic> yet…”

In the end, while I think the general abbreviation to ‘metrics is inclusive and encouraging, its specific application is a tragically accurate representation of the priorities of my peers (at least most. Obviously there are exceptions in any case): concerned with appearances, and not too interested in economics.

Edit: It dawned on me I could actually find out of Mastering ‘Metrics was the origin of the term by checking out economics resources. I’ve never really seen it used in any blogs, but I noticed that on EJMR there are references to ‘metrics going back 4 years (possibly more). While I still think its application in my cohort comes from the use in the book (I have the advantage of seeing them before and after), it’s definitely not Angrist and Pischke’s coinage.