Sunday, March 6, 2016

Lars P. Syll’s Blog — Ceteris paribus — an alibi fairy


There is nothing wrong with using restrictive assumptions in model building, for example, for heuristic use rather than representational. Teaching generally begins with simple models and builds on them. 

For example, in a basic course in physics, Newton's law of gravitation is taught assuming a vacuum. Clearly, this is a restrictive assumption regarding the real world where friction must be taken into account. But the basic model serves as a useful pedagogical tool, and students are told that things are more complicated than that. The teacher might even demonstrate with an apple and a feather. But for the test, students will be instructed to assume a vacuum.

Similarly in basic economics two variable models are used "holding everything else equal." There is nothing wrong with using simplified models heuristically as long it is not conflated with representation of the real where all things are almost never equal.

The problems arise not from the modeling heuristically but from the logic involved in arguing from heuristics to representation. This involves an illogical jump owing to a category error.

It is comparable to an engineer using only rules of thumb to design an intricate project. No engineer would be foolish enough to do that, owing to reputational risk at minimum and possible incurring of legal liabilities. 

But economists do this frequently, for example, in formulating and selling policy arguments based on low representational models. The difference seems to be that reputational risk for being wrong is low in economics while professional (ethical) and legal liabilities are non-existent, even for misrepresentation and negligence.

There are other similar types of error that economists make. Daniel Kahneman and Amos Tversky investigated this in behavioral economics, and Kahneman reports on it in Thinking Fast and Slow.

So I don't think it is necessary to banish the use of restrictive modeling in teaching. But good teaching also needs to include the logical restrictions, too. Conventional economics doesn't seem to pay enough attention to this.

For example, teaching basic economics is often tailored toward an ideological outcome: Scarcity exists while notional demand is infinite., and the free market is the optimal (most efficient and effective) way to address this. Simple models are adduced to demonstrate this foregone conclusion. This is misuse of restrictive assumptions like cet. par.

The consequences are huge.

Another category error is conflating household and firm finance, with which most everyone is familiar, with government finance, where the correct relationship is inverse rather than direct. 

The illogic of such moves is so obvious that one wonders how such mistakes get made in the first place, survive for so long, and are so hard to correct. It boggles the mind.

Lars P. Syll’s Blog
Ceteris paribus — an alibi fairy
Lars P. Syll | Professor, Malmo University

2 comments:

Peter Pan said...

Cui bono?

Matt Franko said...

"one wonders how such mistakes get made in the first place,"

Because.... they.... are..... I-N-C-O-M-P-E-T-E-N-T....