The expectation of market clearing for programmer wages is completely implausible here.
Going independent requires additional skills beyond the programming productivity that we're supposedly trying to measure, is perceived as riskier (and it's generally understood that most people are irrationally risk-averse), and may involve additional practical complications, such as if our hypothetical programmer relies on salaried employment to provide health insurance for a family. So "would this programmer make more independently" is pretty much an irrelevant question.
As for moving between "regular" jobs, there's little evidence that most hiring processes can distinguish programmers by productivity, nor that most organizations measure the productivity of employed programmers in a useful way. Switching jobs also has practical obstacles, such as the disruption of relocating, and avoiding the perception of "job hopping". So as far as we can tell, salaries have little relationship with productivity, which is kind of the whole point.
Fundamentally, the article makes the mistake of taking principles that are simplifying assumptions in economics and then arguing as if they're universally correct. This is roughly akin to an argument like "Things tend to fall down, so it's a fallacy to ask why helium balloons float."
Of course, you can simply define someone's productivity as the market value of their time and thus "prove" that order of magnitude differences in productivity between programmers don't generally exist, but proof-by-redefining-the-terms doesn't really make for a persuasive argument.
Going independent requires additional skills beyond the programming productivity that we're supposedly trying to measure, is perceived as riskier (and it's generally understood that most people are irrationally risk-averse), and may involve additional practical complications, such as if our hypothetical programmer relies on salaried employment to provide health insurance for a family. So "would this programmer make more independently" is pretty much an irrelevant question.
As for moving between "regular" jobs, there's little evidence that most hiring processes can distinguish programmers by productivity, nor that most organizations measure the productivity of employed programmers in a useful way. Switching jobs also has practical obstacles, such as the disruption of relocating, and avoiding the perception of "job hopping". So as far as we can tell, salaries have little relationship with productivity, which is kind of the whole point.
Fundamentally, the article makes the mistake of taking principles that are simplifying assumptions in economics and then arguing as if they're universally correct. This is roughly akin to an argument like "Things tend to fall down, so it's a fallacy to ask why helium balloons float."
Of course, you can simply define someone's productivity as the market value of their time and thus "prove" that order of magnitude differences in productivity between programmers don't generally exist, but proof-by-redefining-the-terms doesn't really make for a persuasive argument.