On economics as method

Via Mark Thoma’s twitter feed came this gem on “neo-classical economics”.  It strongly supports my view on what economics is so let’s have a peak at the conclusion:

I sometimes argue that the main version of economics,  at its best, is that  it is a disciplined way of thinking and arguing that makes clear where people disagree. If two economists disagree, they can unpack each other’s arguments. Do they disagree in their underlying assumptions? In their model of how those assumptions fit together? In their arguments over cause and effect? In their beliefs about what data to use? In the statistical methods they use? Even when economist end up disagreeing, they should be able to pinpoint the sources of their disagreement–and thus to agree on what issues need to be further researched and resolved. From this process, provisional truths (and is there really any other kind?) do emerge.

This is completely true, and the way a vast vast majority of economists see economics and developing economic knowledge.

However, the sentence prior to this description also raises an interesting point for me:

For noneconomists, I guess the obvious question is: “If economics doesn’t give a correct and clear answer most of of the time, what good is it?”

Indeed, I get this a lot from people.  If economics isn’t telling us what the “right” answer is off the bat, and if economists can disagree, then their is a presumption the entire disciple is a complete waste of time.

For a long time I did not understand this in the slightest, and assumed that when people talked like that it was because they wanted to use different assumptions, and value judgments, without wanting to evaluate them!  It frustrated me – but my underlying presumption was that the method of discourse and the type of knowledge that we CAN have (provisional, or conditional, knowledge) was obvious to everyone.

Over time my views have changed.  Non-economists do have underlying models, but they also have a stronger view on underlying truth – which is why claims that are made “irrationally firm” are seen as being more compelling.  Understanding the model building process of non-economists, and helping to make economic arguments accessible in a form that works with this process as a result has use.

The easy “answer” is to exaggerate our faith in our own results – but this makes arguments vulnerable to things we know will go wrong, conditional conclusions will be wrong as we never define an entire system.  Fundamentally, we can’t forecast, and reasonable economists will disagree on (at least) the basis of normative judgments, and if we exaggerate the strength of our results we just end up in a situation where a single failure can undermine a broader set of useful economic ideas and principles!

For me it is about reframing and counterfactuals.  When people say “the answer is obviously this”, make the assumptions that result requires TRANSPARENT.  If the result is based on ridiculous assumptions about behaviour, or is only partial and is missing trade-offs, say it.

Finding real conditional results is hard.  Making these results into analogies that are useful and not misleading is hard.  I’m asking economists to do EVEN MORE by making sure the analogies they use can also be used to point out flaws in other arguments.  And this is why I’m an entirely unreasonable person :)

Note:  This is loosely related to the debate between Noah Smith and Lars Syll.  I agree with Noah – although Lars post is indeed a good read.  My main issues in Lars post are.

  1. As Noah says, economics is broader than the neo-classical method.
  2. Calling ‘neo-classical economics’ deductive is a stretch – many of the assumptions being called “core” in this sense, are seen as simplifying assumptions that will be loosened when it is possible.  If we only call the deductive elements of economics neo-classical, then we are looking at a subset of what economists do – and a subset of the method of how a “question is answered”.
  3. To do normative, or policy, economics you NEED to make assumptions and have them sitting out transparently.  ‘Neo-classical’ economics does that – strict empiricism certainly does not.  People may say “lets just describe and not do policy then” – and that is indeed what many empirical economists have moved towards ;)
  4. Criticising assumptions is an easy target and always will be.  But assumptions depend on the question of interest.  This is hardly an ignored issue in the discipline (Maki, Maki, SugdenREP).  A true “replacement” for neo-classical economics would provide tractable and useful other assumptions – behavioural, experimental, and neuroeconomics are particularly exciting for the discipline for a good reason ;)

The real question is, does the discipline let data influence theory sufficiently – I’d note that it is possible to answer that it has too much sway as well as too little.  Data is an imperfect window on the assumed objective reality that exists, and in uninterpretable without a starting point (apriori theory) – an iterative process between theory and data should lead the discipline somewhere, but whether that is the “right” equilibrium for the discipline isn’t even necessarily a given.  Fun.

Note 2:  I wrote this like a week before I it came up on the site (scheduled posting), but I suspect I used the thought process when I wrote this post about Mankiw and tax.  Weird.