I left a monster comment over at Econlog and, rather than hiding it in shame, here I am displaying it for all to see. My behavior reminds me of the time when I wanted to drag a parent to the toilet to show off my tremendous poop. I believe the parent wisely declined and took my word for it but my memory is unclear.

I am responding to someone who argues (perhaps for good reason; I was not able to find empirical evidence either confirming or refuting his claim) that the empirical evidence is in and the effect of the minimum wage hike is net good (the increase in bottom wage outweighs the loss in employment). Here is my tremendous (in word count) reply:

Tom - it may be that economists (and libertarians) sometimes ask the argument to carry more weight than it really can. The basic argument is simply the point that there's a trade-off: if you raise the minimum wage, theory predicts that while the wage of some will be lifted, there will also be a countervailing tendency to increase unemployment. But this point (a) does not supply you with actual numbers (those, after all, depend on things which need to be empirically determined such as people's actual preferences, elasticity etc.), and (b) does not prove that the one effect (the increase in unemployment) "outbalances" the other effect (the rise in wages). Whether one outbalances the other is a judgment, not merely a prediction. Even if we could predict to the last penny what happens, in the end the assessment of which effect was the more important would be a judgment. You can, of course, try to mechanize the judgment by applying certain efficiency criteria, but the decision to treat the outcome of that calculation as the one to go by is itself a judgment.

However on the whole I am inclined to defend the argument against the minimum wage as far as it really does go.

However, I will say that those who support it *have* to be outside mainstream economic reasoning because mainstream economic reasoning doesn't match the actual experience of minimum wage laws (that the job loss is minimal or non-measurable).

That doesn't tell me very much because causality is very hard to measure in something like an economy. You can't do controlled experiments etc. So your claim that the job loss is non-measurable does not tell me that it is minimal.

Feel free to present me with some paper or somebody mentioning work that does the very, very hard work of trying to figure out the causality. But we remain ignorant of causality even in cases where the whole country hangs in the balance, for example the causes and therefore the cure of the current recession. I'm seeing a lot of strong opinions but I'm seeing a lot of conflicting opinions and nothing is really coming out of this as genuine knowledge. The wildly false prediction about the rise in unemployment and the apparently wildly false prediction about the effect of the stimulus do nothing to boost my confidence in our empirical tools for unraveling the causal factors in the economy. It's on point that one could argue that the stimulus may have indeed had a good effect because unemployment might have been even higher otherwise. That simply underlines the difficulty of empirically determining what the causality is.

Generally speaking, I predict that people will reliably make the error that Bastiat and Hazlitt both pointed out: people tend to see only part of the picture, in this case the easily visible boost that a minimum wage hike will have on those who are employed at minimum wage both before and after. This is the problem of the seen versus the unseen. See either Hazlitt's Economics in One Lesson or Bastiat's What is Seen and What is Unseen. Since this is a reliable bias (in favor of the seen), I am not at all surprised that the widespread experience of wage control will be strongly biased in favor of noticing the benefit while missing the cost. Theory is useful in making us keenly aware of the potential for a perverse side-effect. And a moment's thought should make us realize that this perverse side-effect is by its very nature hard to see. It is hard to see what would have been. Causality needs counterfactuals. It is hard to see individuals who would have been hired but who are not now hired. It is hard to see individuals who have given up trying to find a job specifically because of the wage hike. We try to see these things by gathering enough data that we can infer, from the numbers, what must have happened in countless individual cases, but it is hard to disentangle these things even with lots of numbers at our disposal. For instance I recall reading that the unemployment statistic compares those getting to those seeking a job, but such a statistic by its very design happens to miss those who have stopped looking out of despair at ever finding a job - those who have given up. Those too are harmed but they are not counted in a figure that excludes them by design. (Wikipedia says: "Unemployment occurs when a person is available to work and seeking work but currently without work.")

Look, all I am saying is that minimum wage arguments are a classical case of economists with physics envy. The theory is nice and beautiful, and those messy humans are fouling it up with their non-conforming behaviour.

Some people may be using the arguments that way, but there's another way to use them, and that is to overcome the above-mentioned bias in favor of the seen over the unseen. It is to let people understand precisely how a wage hike (and price controls generally) could have a perverse incentive. People's knee-jerk, un-tutored intuition completely misses the perverse effects on supply and demand of price controls. Okay, so maybe it's theory, but people come at the world already armed with a folk theory that just simply ignores effects on supply and demand. Thus, people don't come equipped with the mental tools to realize how it is that (to shift to another common example) when you tax one party of the transaction, this changes his supply curve which, in effect, causes both parties to the transaction to share the cost of the tax, and that the proportion shared by each depends on their supply and demand curves. Okay, granted, this is "merely" theory, but it's superior to the folk "theory of the stupid" which simply takes product availability for granted. "Huh, government introduced price controls on bread and gas and now the shelves are bare and there are these long lines at the gas pump, gee, who could have predicted that", or even worse, "the suppliers are evil, just look at how they're maliciously reacting to the new laws by deliberately refusing to sell products with a price ceiling, these are wreckers, let's string them up."

This means that they end up defending what much of the populace knows is 'wrong', where wrong can encompass a great deal from the simple facts to the failure to take into account human factors such as fear, status games, disgust, etc.

Much of the population "knows" it is wrong often because much of the population suffers from biases such as the tendency to be completely oblivious to that which is unseen, allowing the populace to fall for such things as the broken window fallacy.

Sorry for going off topic, but I find this tendency in economics quite distressing as it causes the general populace to devalue economics far more than they should. After all, if it's so obviously wrong about what they can see, why should they think it is right about what they can't?

I seriously question the judgment of any individual who thinks it is trivial to see such things as the full effects of a minimum wage hike. A person might go around, see people working at a new higher minimum wage, and conclude, "wow, it worked". He is not equipped to measure the less easily seen effects.