So I already wrote about right-to-shirk (aka right-to-work) laws, and made the basic freedom of contract argument. By its functional operation, right-to-shirk laws prevent two parties (a union and an employer) from agreeing to certain terms. I am not actually interested in the argument itself; I just raise it to mock the hypocrisy of libertarians on this issue. Since then, Yglesias has written a similar post, and the odious Reason magazine has written a post conceding the point. Meanwhile Adam Ozimek has a winding piece of a much different sort. I am somewhat fascinated by Ozimek’s piece, and so I thought I might muse on it a bit to explain why.

It deserves pointing out first that Ozimek does not define freedom in the post. That is a big task, no doubt, but when we start delving into really fine detail, whether something violates freedom will often turn on the definition being used. I suspect from how he writes here and his other writings that freedom basically means “whatever would result under laissez-faire.” That’s obviously problematic, but I wont tease that out yet again here. I cannot escape this question-begging zombie of an idea no matter how hard I try, and I can’t keep complaining endlessly about it.

The basic structure of Ozimek’s analysis, although somewhat obscured by the analogical style he uses, is straightforward. He does not believe that right-to-shirk laws increase or decrease freedom categorically; rather, we have to do a case-by-case inquiry. We are supposed to look at each given union contract and ask the following: would this union’s contract and the security clause contained therein have come into existence in a hypothetical world where the National Labor Relations Act did not exist? If the answer is that such a contract would still exist in the hypothetical no-NLRA world, then in that case, right-to-shirk laws decrease freedom. If the answer is that the contract would not still exist, then right-to-shirk laws increase freedom in that case.

As should be pretty obvious, this is not a terribly helpful analysis. It does not give us an up or down on whether we should be passing such laws. Presumably, we should calculate net freedom gain using the approach above, and then that will tell us whether the laws are good. Of course, you cannot possibly know what the no-NLRA baseline looks like. Ozimek imagines that there would be less union contracts in the no-NLRA baseline than there is in the status quo, but this is a big assumption.

The NLRA is not a big win for unions. It has the weak duty to bargain in it along with exclusive representation (both things Ozimek seizes upon), but it also outlaws secondary boycotts, which is one of the most devastating Labor tactics. If that were resurrected as legal, a relatively small number of strategically located workers could so disrupt production that masses of employers would be hard-pressed not to sign contracts. So quite plausibly we would have massively higher union density without the NLRA. This does not harm Ozimek’s framework as much as it shows how impossible it would be to apply it to the world and generate answers on where freedom is being increased or decreased. It is not as simple as asking: would the company have agreed if it weren’t for exclusive representation and duty to bargain? Under the no-NLRA baseline, we have to also ask: would the company agree if all the things restricting unions in the NLRA were also repealed. In practice, that’s an impossible question to answer.

And that’s just the tip of the iceberg really. Why just compare the status quo to a hypothetical (and unknowable) no-NLRA baseline? Shouldn’t the true analysis be that we have to figure out what union agreements might look like against the full blown laissez-faire baseline? For instance, consider the Occupational Safety and Health Administration. It protects workers from workplace dangers. It is a violation of laissez-faire. So we need to strip it out of our laissez-faire baseline, and ask what would happen. Well, quite plausibly if the federal government were not protecting workers from workplace dangers, more workers would want to join unions to keep themselves protected that way. So under an even more laissez-faire baseline than the no-NLRA baseline Ozimek uses, we may have even more unions.

But why even stop there? Maybe we should go with a full counterfactual history and ask ourselves what the present reality would look like if the NLRA never existed. How might history have changed and in what way might those changes have compounded on each other to generate a different present-day society? Surely we should model that into our calculations as well and ask under a full-fledged counterfactual history whether, in a given case, there would be a union contract with a security clause.

As is probably clear, I find this type of analysis a tad bit silly. If the only way you can know whether a particular bill increases or decreases freedom is by modeling some sort of hypothetical world where various sorts of other legislation doesn’t exist either, then you really can’t know at all. All your argument really reduces down to is “I think laissez-faire is freedom and I cannot meaningfully comment on anything once the economy has taken more than a single step away from laissez-faire.” I find this an especially odd way to analyze the situation because this hypothetical counterfactual laissez-faire world has never even existed. That means we are supposed to compare given policies against a baseline that is at its best a historical imagination and at its worst an undefined mass of libertarian hand-waiving.