Over at Crooked Timber, there's a great post on the minimum wage where Kathy argues that 1) there are plenty of empirical reasons why increasing a minimum wage may not lead to higher unemployment and 2) that you'll get into trouble with the economic 'fundamentalists' if you try and work on this issue without concluding that it does.

This piece, and the winding comments, are a fine testament to the problem of applying models to the real world. Here's a single issue - increasing minimum wage - with plenty of data, many studies and lots of theory, and there's still little in the way of consensus. Kathy points to (unconscious) selection bias among researchers, finding data that supports their pet model. She - and others in the comments - argue that simple supply and demand is of no use. Of one economist's attachment to supply-demand models she notes he...

... is so enamored of them that he sees them, I think, not as tools for understanding, but as God’s revealed truth, handed down to Moses on stone tablets.

This fundamentalism, she argues, blinds the believer from all the evidence for other forces at work:

...the perfect competition model assumes a frictionless job market, with complete information, no search costs, no mobility costs, no heterogeneous preferences, no firm-specific human capital. Indeed, they assume a job market so frictionless that if an individual employer cuts wages by even one cent, every employee at that firm will immediately quit.

Do all these factors make the supply-demand model useless? Oddly, I find I'm with Tyler Cowan on this - he comments:

You either think that the demand curve for labor slopes downward at the relevant margin or not. If you don't think it does, you need to argue for that directly. Arguing that the labor market is not perfectly competitive — while true — does not get you there. And if you admit the demand curve slopes downward you are already very close to Kevin Murphy’s point of view [the 'Gods revealed truth' guy]. It is correct that there isn’t always a negative employment effect from raising the minimum wage. The simplest model to explain that result is this: the employer responds by lowering non-money components of the wage, such as working conditions, to keep the real net wage constant. That's not such a welcome conclusion for what you are trying to argue.

Or, in the UK, perhaps the simplest explanation can be found in the huge expansion of black market labour. With an issue as complex as this, Occam may not be the best guide. But at least it makes one look for the simpler, obvious answers.

Most important for me is that there's no getting away from the sums. Despite the various 'frictions' mentioned, a firm paying the same number of workers more money will have to find the money. Quickly, though, you have to leave math-world and start speculating. My speculation: firms being what they are, shareholders will likely apply pressure to wages and benefits (or turn a blind eye to dodgy subcontractors). But then, perhaps a model can offer other simple math insights: what impact do those extra wages have on the wider economy? Hell, but then it gets all tangled again: too much of a wage increase and inflation goes wampy.

On a different tack, another good comment [9] points out:

We are all deontologists under the skin and consequentialist arguments are of limited utility. I think most people of a libertarianish persuasion (myself included), economist or not, ultimately oppose the minimum wage on deontological freedom-of-contract grounds: we think that employers have a natural right to offer jobs on whatever terms they damn well please, and interference with that freedom in the name of 'social justice' is just plain evil. And I think most lefty supporters of the minimum wage support it on the deontological grounds that workers have a natural right to a living wage, and allowing them to get paid less than that is just plain evil.

Well put - this is exactly my problem. If the consequence of a social-justice policy is its opposite, what good is the policy? There are, without doubt, cases where laws (describable in models) can show what might happen. On the other hand, why should anyone listen to people who are clearly using arguments about immutable social reality on behalf of the rich and powerful?

Oo, its a tangle, and no mistake. Perhaps the answer's to not worry about the consequences, get on with being deontologists and let God sort out the rest.