IBM says 'several thousand' layoffs on the way

I work on a product that's pulled in >$2bn in revenue over the last 20 years. All with an average of 25 developers working on it. It's still going strong.

The model is a simple deductive truth only when its premises -- which include actors that behave strictly according to the rational choice model (including perfect information) and the absence of externalities.

Neither of these is generally true in the real world.

> I think you read my sentence as "(following the principal of maximizing profits is what is most socially beneficial) in the absence of externalities" where what I meant was "following the principal of (maximizing profits (in the absence of externalities)) is what is most socially beneficial".

Both have the failing when being applied to the real world that they silently assume the assumptions of the rational actor model, including perfect information, and either way they present problems when used as a statement about a real-world decision in which externalities are present (though the exact nature of the problem differs between the two.)

> Classical economics with optimal "Pigovian" taxes can be thought of as a first order approximation to reality.

Not justifiably. I'd agree that approximating optimal Pigovian taxes is a worthy goal for government policy, but I don't think that there's any justification for assuming that actual government policies do that.

(In fact, given the distribution of power over government policy that would have to occur for that to be true, there's a pretty good reason to assume that its not even approximately true.)

Further, the behavioral model underlying classical economics are a tolerable first order approximation of reality in select markets, and useful baseline from which, via different circumstances in other markets which explain variation, to explain the behavior that occurs in other markets were they aren't good as such an approximation, which (aside from the prominence of ideologies which are justified by giving that model too much weight) is why its still taught.

> Also, you can produce any Pareto efficient outcome from free markets + redistribution.

Making many of the same assumptions with limited and occasional connections to real behavior that underlie your previous statements, starting with the rational choice model (including perfect information), this is true.

Not sure what your point is with it, or how it is supposed to be germane to the discussion.

Almost certainly, it is an externality in the strict economic sense, in that the "thing that decreases the value of labor" is almost certainly the product of investment decisions made by actors that are not the same set of people who are impacted by the reduction in the value of labor.

> Redistribution can counteract this effect.

Right, but in practice rarely does not, because what redistribution occurs is controlled by who has power over government, and power over government is disproportionately in the hands of those who have gained the most benefit from the economy, so those harmed by externalities and who would be most inclined, on a self-interested level, to seek redistribution are also the least likely to see their wishes reflected in government policy.

> If productivity increases but it also increases inequality, then there is some level of redistribution that will correct the inequality while making everyone better off (in the sense that the number of people earning more than X increases for all X). Not exactly a theorem, but a rough consequence of general equilibrium theory.

Its not really a "rough consequence of general equilibrium theory", whereas general equilibrium theory holds that without externalities (and with rational choice) a pareto-efficient result will be achieved, your conclusion requires the assumption (which general equilibrium theory does not support) that with externalities, a pareto-efficient result will not be reached, and further that the actual result will be such that there will exist an alternative result reachable by redistribution which features less inequality by whatever the relevant measure of inequality is, and is closer to being pareto-efficient.

But general equilibrium theory does not guarantee pareto-inefficiency with externalities, and if a pareto-efficient result is attained prior to redistribution, no redistribution can "make everyone better off" (since the definition of pareto-efficiency is that no one can gain without someone losing.)