Monday 4 March 2013

Theoreticians or empiricists?

Martin Wolf and Simon Wren-Lewis have recently mentioned the distinction between engineers and theoreticians as a metaphor for how economics should be done. It is an important, and interesting, distinction, which got me thinking about my own field, and how wrongly the distinction is often perceived.

I delve into what is pompously known as Operational Excellence. Performance improvement, if you will, but not for a sportsman: for professional organisations.

In this field you frequently come across things called Lean and Six Sigma. I'm not going to explain in great detail here what they are, but if we accept a caricature, you could say that Six Sigma describes a project-based approach to situation where you first go through a data collection phase before isolating the problem, fixing it and sustaining the improvement, whereas Lean will give you a set of principles to apply in all situations and instruct to treat any deviation (aka problem) immediately.

This is crude, and anyway much of the use of experience in the field serves to know what of either approach to use when and how, but that will do for what I'll try to describe.

In applying for projects, or in training, I often get to describe how I would start by a diagnosis involving collecting data (which can be from a database, or by direct observations, or by collecting verbatims...), and then how the project will follow depending on what was identified. In short, a methodological framework.

Well, of course, you win some and lose some, actually I can't complain about my winning ratio so far. But when I've lost (and even sometimes despite having won), I've often been described as being seen as theoretical when the alternative felt more practical.

Now this has always baffled me. By training, I'm an engineer, a very practical trade. And I really don't believe I've ever had a clear idea of how something should work before studying it -on occasions I did not have any idea how it worked at all. But that's the point of having a framework to study things.
I believe that I am an empiricist -which is why I so need data to make up my mind. And this was confirmed when I was able to get a hint of what the alternative, the one deemed practical, had suggested.

To summarise, where I would say that I would make observations and collect data so I would know where to look and what to look for, and based on that be able to decide what I would do, they had said "This is what I will do" (for example, setting-up a Kanban). They had recommended a course of action without having had a chance to even look at the situation.

Now, this is rather close to what Wren-Lewis was describing. An empiricist will constantly seek new information to test what he thought, and adjust his stories accordingly. A theoretician (at least in economics) will often merely seek to check internal validity, that is, making sure that the conclusions are consistent with the axioms in the theory. Keynesians tend to be empiricists, with recommendations that differ greatly with circumstances and therefore the need for data to evaluate those circumstances and quantify impacts. Freshwater and Austrians are typically theoreticians -arguing with maths (Freshwater) or without (Austrians) that you should deregulate and cut top income tax rates.

So how could empiricism be seen as theory? Well, one thing is that properly done empiricism is hard. It uses data, but cannot process it to get an immediate definitive conclusion, because it knows that it is imperfect data, and that there are a bunch of factors at play, not just one. So it must dwell into uncertainties and probabilities. Look at medicine, now the ultimate empiric field (many treatments were proven to be effective long before we had any idea why they worked -which is not to suggest that it was not a worthy endeavour to then find out why they did, and this is where theoreticians came in). It is very complex, devising measurements can be a challenge, you often need meta-studies as individual ones aren't enough due to residual variation, methodology needs to be very clearly described... But back when it was theoretical, it would say things like "bleed the patient".

And one common trend of organisations is that, people above a certain level often treat things they don't understand as irrelevant, unimportant, or easy (and yes Dilbert lovers, I do believe that I've first seen such a description made by Scott Adams). And so someone deploying a methodological framework can be dismissed as theoretical (or even dogmatic), where paradoxically someone confidently giving unchanging recommendations without much or any information about the situation may be called practical and consistent. In the same vein, we see economic empiricists being decried as ideologues while Austrians (who refuse the very idea of testing their prescriptions with data) are called responsible and pragmatic.

There can sometimes be a need for expediency. But beware: therein lies the road to bleeding patients. Or supply-siding your economy.

No comments:

Post a Comment