Economics as a soft science

What we’re covering right now in class is not something I’ll cover here in detail – it’s very technical stuff. A few excerpts from today’s lecture notes (click to view full size):

Stuff like this is why I actually get a bit annoyed by people who state that their impression is that economics is a relatively ‘soft’ science, and ask questions like ‘the math you guys make use of isn’t all that hard, is it?’ (I’ve been asked this question a few times in the past) It’s actually true that a lot of it isn’t – we spend a lot of time calculating derivatives and finding the signs of those derivatives and similar stuff. And economics is a reasonably heterogenous field, so surely there’s a lot of variation – for example, in Denmark business graduates often call themselves economists too even though a business graduates’ background, in terms of what we’ve learned during our education, would most often be reasonably different from e.g. my own.

What I’ll just say here is that the statistics stuff generally is not easy (if you think it is, you’ve spent way too little time on that stuff*). And yeah, the above excerpt is from what I consider my ‘easy course’ this semester – most of it is not like that, but some of it sure is.

Incidentally I should just comment in advance here, before people start talking about physics envy (mostly related to macro, IMO (and remember again the field heterogeneity; many, perhaps a majority of, economists don’t specialize in that stuff and don’t really know all that much about it…)), that the complexity economists deal with when they work with statistics – which is also economics – is the same kind of complexity that’s dealt with in all other subject areas where people need to analyze data to reach conclusions about what the data can tell us. Much of the complexity is in the data – the complexity relates to the fact that the real world is complex, and if we want to model it right and get results that make sense, we need to think very hard about which tools to use and how we use them. The economists who decide to work with that kind of stuff, more than they absolutely have to in order to get their degrees that is, are economists who are taught how to analyze data and do it the right way, and how what is the right way may depend upon what kind of data you’re working with and the questions you want to answer. This also involves learning what an Epanechnikov kernel is and what it implies that the error terms of a model are m-dependent.

(*…or (Plamus?) way too much time…)

October 30, 2012 - Posted by | Econometrics, Economics


  1. Excellent post – how can it not be – it mentions me 🙂

    I have not had your experience with people short-changing economics and economic applications of statistics, but that’s probably because I do not sell myself as an economist. I have only an undergraduate degree in economics, plus a fair bit of reading done on my own. Likewise, I am no statistician. I have taken several courses in stats, but a good bit of it was taught by math/stats guys, and thus was too academical – I know for sure I’ll never need to transform an alpha-distributed variable into a beta-distributed one (had that as a homework assignment in one course), and if I ever need to solve a 5×5 transportation problem, I would not do it manually (had that as an assignment too, and messed the calculations up), but plug it into a software package.

    I would never knock down statistics as easy – hell, I have my intellectual scars from tussling with it. Nor would I knock down economics as “soft”. In fact, my big beef with economics is exactly the opposite – it’s way, way too “hard”. Whole sub-fields of economics have turned into a (pardon my French) intellectual circle-jerk, where economists try to out-wow each others with mathematically elegant models with at best zero, and but often negative practical value – negative because their findings make their way into public policy without proper attention to assumptions, omitted variables, and common sense. A model’s purpose (ideally) should be to tell us something useful – not to get you published and invited to conferences. I could go on and on, but I’d rather leave Tim Hartford’s TED talk on the God complex and Tyler Cowen’s TED talk on stories here – I suspect, you have seen them, but your readers may benefit from them. Greatly. I count these two talks among the most valuable learning moments of my life, since they both taught me a lot and managed to crystallize many gut feelings I had had. I think most (no, not all) of modern economics is severely (irreparably?) infected by the God complex and the desire to tell a compelling story.

    You say “The economists who decide to work with that kind of stuff, more than they absolutely have to in order to get their degrees that is, are economists who are taught how to analyze data and do it the right way, and how what is the right way may depend upon what kind of data you’re working with and the questions you want to answer.” Very true. But even the honest ones (I am not gonna mention Krugman, Stigitz, DeLong – or did I just do that?) seem to mostly forget that they are not God, that they are talking to mostly non-economists, that they have made a crapload of assumptions along the way, and that for the n variables they use and k variables they control for there are (infinity – n – k) others.

    Finally, to wrap up my rant, I’ll note that I know what an Epanechnikov kernel is, but that’s not important. What, IMHO, is important, is why I know it. My job description is, to a large degree, to find investment models that work for the company I work for. I am as lazy as one gets, and proud of it, so I found a software package that lets you run some of the tests I wanted to run on our data a lot faster than I could get Excel to do it (what can I say, I’ll always love Excel, much as I hate Microsoft). I taught myself how to use that package, and lo and behold, among the analyses you could run there was this thing called support vector machines. Well, in order to get those to work for you, you have to give them some parameters to work with, and among those was choice of a kernel – dot, radial, polynomial, neural, anova, Epanechnikov, Gaussian combination, or multiquadric. [Side note: there are many, MANY more options, many contingent on your choice of kernel.] So I spent a vodka-fueled weekend reading up on SVMs and kernels – what they are [including what m-dependence is :)], what they are good for, what they are bad for, etc. I still am miles from being able to comprehend the full theory of kernels – smarter men than I am make a living researching that. But I was able, building on their advice, to build the best investment model I ever have – and trust me, I have tested it inside and out for robustness, and I manage a good bit of company money, and my own retirement savings based on it. It was up 44% last year, and is up 13% this year so far (tough year, what can I say). It took about a month of playing with the options and adding in and taking out variables. But the takeaway is that I managed it without being a statistician. I did it through trial and error. And I have plenty of room to improve it – I have only scratched the surface of the possible combinations. I have not the flimsiest idea whether the variables I use are m-dependent. I could test for that, but it’d take time and effort, and I am impatient and lazy. Instead, I try, err, learn from it, adjust, and try to improve. And it’s paying off for me.

    Rant off, apologies for the incoherent ramblings (vodka, my inspiration and undoing!), and cheers!

    Comment by Plamus | November 2, 2012 | Reply

  2. i. Your returns most likely imply that your risk profile is very different from my own (be careful).

    ii. You should probably know that the post was at least to some extent motivated by two recent discussions I’ve had – so those kinds of ideas do surface from time to time. I like to think that I’m trying to be careful not to assume that I know what other fields are like and so to be skeptical about any kinds of comparisons probably mainly because of the self-serving bias – I’m always going to try to convince myself that the stuff I learn is harder than the stuff other people learn because that’ll make me look good. But it’s a bit silly to go too far in that direction. If less than 1 in 8 Danes get the equivalent of a Master’s degree or higher (the stat for people aged 30-34 is 12% with Master’s degree or higher – I’ll probably blog these numbers in more detail later), there’s no way economics is in a ‘soft/easy’ category; people well within the top 5% (2%?) in terms of IQ are struggling just to pass some of the courses. And not all smart people get a university education.

    iii. One problem is that people who don’t know much about it will judge the field based on what they consider the field’s ‘representatives’ – i.e. people like Krugman (or bank spokesmen). a) Some of those people are one step, and a very small one at that, away from simply being astrologers (with a god complex). b) More than a few of them aren’t actually economists (rather they have business degrees). c) A big part of what those representatives do have nothing to do with economics – it’s PR or politics. Other fields surely have similar problems but it’s worth having them in mind just the same. If you haven’t worked in the field but only/mainly know economics from these sources, it’s very easy to convince yourself that economics can’t possibly be all that hard.

    iv. I’m planning on posting ‘something on models’ before long, so I’d rather save that stuff for later rather than comment in much detail here. But as I said, physics envy is in my impression mostly macro-related.

    Comment by US | November 2, 2012 | Reply

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: