Thursday, January 28, 2021

F**K, Marry, Kill

 I'm really exhausted by this quixotic attempt at the elevation of (price) theory. As Boettke, Coyne and Leeson (2003) themselves point out, theory has evolved to the point where any proposition is provable, making good empirical work more important than ever.

Take the minimum wage for example. There's the basic supply and demand model that says one thing and there are monopsony and other models that say something else. IT'S AN EMPIRICAL QUESTION!
In the absence of experiments, modern methods that try to get at identifying causal effects are super-important.
If your theory is so right, you should be able to find empirical support in models using the best available methods for the problem.
None of this is to say that empirical work doesn't suffer from file drawer bias or p-hacking, or cannot be influenced by the ideology of the researcher. Of course it can. BUT SO CAN THEORY!!
I really feel like to be a scientist, you have to have at least some tiny part of your brain be willing to consider the possibility that your preferred theory may be wrong.

5 years ago I would have said that raising the minimum wage was a laughably horrible way to try and help people, but today I'm not so sure how bad it really is. Thanks to the research of Dube and company.

5 comments:

George said...

Just a single story that adds a tiny amount of actual data (empiricism!). My son knows a man who owns several fast food establishments. A few months after Seattle raised the minimum wage, McDonalds came out with their kiosks for ordering, which resulted in job losses all over Seattle, at least where the kiosks were deployed. Now Georgia, where we live, has not raised the minimum wage. But the kiosks were such money savers that this owner tried out some in one of his stores. He was able to lay of 3 people. He has trotted out the kiosks to the rest of the restaurants that he owns. He told my son that about 15 people had been let go.

Now if an economist wishes to do a comparison of Seattle and Atlanta to identify if there was a different outcome, this economist might not see a significant difference in employment, after proper accounting for the differing trends in the cities, etc. The conclusion might be that there wasn't a labor loss effect, but the right answer is that the effect was twice as bad: people lost their jobs in both cities.

Sometimes economists have to actually survey businesses across the country to get actual data and context, because just downloading data from the BLS website and other sources will not yield meaningful answers and conclusions. I hope they eventually try this out, because I think a lot of information is missing in most econ studies.

Frank said...

Yeah, we need an experiment. Let's raise the minimum wage to $30 per hour and see what happens.

Blissex said...

«quixotic attempt at the elevation of (price) theory. [...] Take the minimum wage for example. There's the basic supply and demand model that says one thing and there are monopsony and other models that say something else. IT'S AN EMPIRICAL QUESTION!»

Oh please such naivety is so sad, consider:

#1 In Economics it is never an empirical question, because the validity of a proposition in Economics is based on *internal* (with the core assumptions) consistency, not with external (with much overrated reality) consistency.

#2 In political economy pretty much everything is an empirical question. JM Keynes argued powerfully against eternal theories, because political economies change shape in the medium and especially the long term. D Glyn argued that political economy studies should be about *mechanisms* more then overall theories.

In case #1 the core assumptions are JB Clark's "three parables" (so called by P Smuelson), of which the most important one is that factor incomes are uniquely and exactly determined by their contribution; any theory of Economics that is incompatible with that is "internally inconsistent".

In case #2 we can say that political economy studies are intermediate between a natural science and the arts. I call that case a "discipline", that is a study of something that does not immutable natural laws, is not entirely based on taste and judgement like the arts, but it based on a systematic approach.

«I really feel like to be a scientist, you have to have at least some tiny part of your brain be willing to consider the possibility that your preferred theory may be wrong. 5 years ago I would have said that raising the minimum wage was a laughably horrible way to try and help people, but today I'm not so sure how bad it really is. Thanks to the research of Dube and company.»

It is not just just being able to consider “the possibility that your preferred theory may be wrong”, but also the ability to question "obvious" received wisdom.

In this case my guess is that the misconception that “raising the minimum wage was a laughably horrible way to try and help people” was probably based on a naive acceptance of receive wisdom of microeconomics, where supply and demand schedules are drawn as smooth lines. But every businessperson in "external consistency" reality knows that they are *bands* (and not smooth ones), so the "market clearing price point" (and quantity) is actually not a point but a region, and where the actual price falls within it depends on non-market factors.

Another possible naive misconception may be that there are two approaches to the "tatonnement" that clears markets, the walrasian and marshallian ones, where in one markets are cleared by adjusting price, with given quantities, and in the other by adjusting quantity, with given prices, and most assume that it is prices that necessarily adjust. But every businessman knows that both adjust, and in the case of wages, within a large band, it may be that an increase in the minimum wages leads to a less than proportional reduction in *hours* per person, so that an increase in wage by 10% may lead to a decrease in hours worked by 5%, leaving workers still ahead (and by more than 5% in terms of "ofelimity" if the decrease in hours worked has a leisure value for them).

Blissex said...

«in "external consistency" reality knows that they are *bands*»

Part of my points above are implied by the notion of "consumer surplus", which neoclassical Economists who care about "internal consistency" cannot apply to the labor markets: if there is "consumer surplus" when businesses employ labor, an increase in wages up to a point will only to transfer some of the relevant "consumer surplus" from employers to workers, but it is still going to leave the employer with some "consumer surplus" that makes the business still viable.

Now if capital were perfectly fluid some capital in the sectors affected by a rise in the minimum wage would move to other sectors with higher "consumer surplus", but that capital is perfectly fluid is a critical assumption for "internal consistency" and is not compatible with "external consistency" (see the Cambridges Capital Controversy and "leets").

Blissex said...

«Let's raise the minimum wage to $30 per hour and see what happens»

That's a straw-man exxxxxxxaggeration, and thus just vacuous handwaving.

However, there is a notable case of something similar happening in so-called "reality": several years ago the Singapore government raised wages by decree a lot (IIRC 20% or 30%), because they wanted to price singaporean workers out of low added-value work. Their reasons was that Singapore has limited land and population, and therefore should specialize in high value-added work rather than "bulk" low value-added work. Not being stupid they had prepared the move by investing a lot in training and education so the limited population *could* do the higher value-added work, but the move was very disruptive at the same time, but it all worked out in the end.

To show how committed the singaporean government was to that, their immigration policy is pretty much the reverse of that of most "western" governments, in that immigration for low-wage jobs is pretty much unrestricted, but it is a lot more difficult to get visas for high-wage jobs, unless it is for workers in training or education, or for workers in trade where singaporean co-workers can learn that trade by working with foreigners. The reason for this policy is to reserve well-paid middle and upper class jobs for singaporean voters, and to increase the competition for low-pay servant-class jobs among non-voting immigrants (most of them malays BTW).