Say a firm screws over a worker on ten dimensions at once, subject to some constraint of the worker leaving. With those ten depredations, the firm tries to extract from the worker as efficiently as possible, but again so the worker does not leave or leaves only with some lower probability.
Let’s then say the law alters one of those dimensions to favor the worker, for instance the firm cannot force the worker to consent to a body search. The firm may then increase one of the other depredations. Perhaps not for individual workers on the spot (there is “depredation stickiness”), but over time firms will fill the niches of new depredations to the maximum degree possible.
The new depredations will be less efficient ways of extracting rents than the old, which may hurt the worker. At the same time, the firm’s rate of return on enforcing depredations may decrease, which may help the worker.
The net effect is indeterminate. Note also that the lower efficiency may in the longer run limit firm entry and production, which will hurt workers and consumers (and possibly shareholders). Again, the net effect still is indeterminate but the more you think about this model the less you will see it as an effective way to help workers. You might try “regulating all ten depredations” but for that to work you also must fix the wage, and so on.
Maybe some of the model variants here will help the workers, but come on, let’s be realistic. Is it the case that the commentators have firm beliefs about these models for well-argued reasons? I don’t think so. It is more likely the case that there is a core belief something should be done, with not too much concern for the systemic effects nor with the “not totally sound but still better than what the critics are serving up literature on compensating differentials.” I am worried by the common tendency to first cite a lack of perfect competition and then assume the proverbial pony.
Fortunately there is a rather smooth path forward. Raise the utility of unemployment to workers. This could be a guaranteed annual income, better unemployment insurance, more food stamps, whatever. Call it the welfare state. Improving the welfare state will improve worker bargaining across virtually all workplace dimensions and in the longer run limit the scope of all the employer depredations.
We’re back to the point that what helps is to give people cash, or something cash-like, including when it comes to the dimensions of workplace quality. It is also a huge help to institute policies which will raise rather than lower worker productivity.
As I said before, the criticisms haven’t even yet dented the traditional economic point of view on this issue. Those criticisms are operating within the current frontier of analysis, not on it much less beyond it.