David R. Henderson  

Does Market Power Affect Optimal Worker Safety?

PRINT
NGDP futures targeting means n... The Unfair Labor Training Act...

The possibly surprising answer is "No."

In an earlier post, I showed that in a free market and under certain assumptions, even workers who are poor will get what is, from their viewpoint, the optimal amount of safety.

One commenter seemed to suggest (I'm not sure) that if an employer has market power, employees will get less than the optimal amount of safety. In one sense, this is true. An employer with market power will pay less than the optimal amount of compensation. So if, as I have assumed earlier and as is amply justified by historical and current data, safety is a normal good, the lower compensation will mean that workers will "demand" less safety. In other words, safety is less than it would be if the employer were not a monopsony.

But the safety component of the compensation package is still optimal given the amount of overall compensation. That is, with the compensation that workers get from a monopsony employer, they want the amount of safety they get. Were the employer to increase safety but decrease monetary pay, the workers would be less satisfied than they would be with less safety and more pay.

Why does this matter?

It's not just a theoretical nicety. It matters because if people believe that workers get less than the optimal amount of safety, their inclination is to advocate that a government mandate an increased level of safety. Ignore all the real problems about how the government could do this and whether such a mandate would be effective. If the mandate is effective, it will make people who work for a monopsony employer worse off.

Update:
Daniel Kuehn argues below:

I do maintain as I did in the other comment thread, though, that we need to distinguish between what economists call "optimal" and other perspectives on it. Restricting ourselves to economic optimality I largely agree with you, but I don't think that's what people have in mind when they're advocating minimal standards consistent with their view of human dignity.

The problem is, as noted above, that minimum standards above the economists' optimum make people worse off. I don't think that's particularly dignified.

Now back to my vacation at my cottage, far from Wi-Fi.


Comments and Sharing






COMMENTS (18 to date)
Daniel Kuehn writes:

I was responding to _NL's claim that the "workers are paying for it"... "because obviously it comes out of the same pool of money that could be otherwise used to compensate workers".

I don't think it's right, and I think we'd expect workers and firms to both pay for it. The other point I brought up besides market power is subsidies which would also shift around the balance.

I agree with the rest of your post here.

I do maintain as I did in the other comment thread, though, that we need to distinguish between what economists call "optimal" and other perspectives on it. Restricting ourselves to economic optimality I largely agree with you, but I don't think that's what people have in mind when they're advocating minimal standards consistent with their view of human dignity.

Daniel Kuehn writes:

Even restricting to the economist's version of "optimal", though, I think we need to be careful with the monopsony case.

Yes, in the monopsony case everyone involved is "optimizing" and at their "optimal" position but this case is not typically considered socially optimal because of the benchmarks we (for better or worse) typically use to think about social optimality.

In either case, though, I don't think that's what people are generally thinking of when they talk about the right amount of safety.

Kevin Erdmann writes:

"Restricting ourselves to economic optimality I largely agree with you, but I don't think that's what people have in mind when they're advocating minimal standards consistent with their view of human dignity."

The problem is that there is a huge gap between the sweatshop immigrant thanking Jesus for the pair of protective goggles they couldn't afford to get because their wages all go toward grandma's hospice care and 5 guys sitting around for 30 minutes waiting on the person who's authorized to flip a switch, and there is little evidence that our governing institutions can differentiate between them.

Daniel Kuehn writes:

Kevin -
Governing institutions are a whole different can of worms. I just don't want to confuse the economic analysis with the moral analysis here. There is definitely plenty more to tackle before actually advocating anything.

Daniel Kuehn writes:

David -
re: "The problem is, as noted above, that minimum standards above the economists' optimum make people worse off. I don't think that's particularly dignified."

By "people" do you mean the workers? This point is true *if* the policy solution is just a mandate. But there are lots of ways to think about how to do these minimum standards.

Thomas Strenge writes:

Please let me add some insight from the business world. Yes, often good businessmen make bad economists, but here the obverse may be true. ;-) A business makes money as long as it's up and running and satisfying customers. When a machine breaks down in a factory, then you can no longer make product, which means you're not making money. The "reliability-centric maintenance" mode of thinking is merging with lean management. As we business people try to streamline processes to make them more efficient and produce better quality, the outcome is also a safer process. In a sense, don't think of safety as a part of worker's compensation, but as a byproduct of an efficient business process. In summary, any sort of work stoppage, including accidents, is inefficient and costs money!

Matt writes:

The "who pays for it" question is surprisingly subtle. Suppose that we start at the unregulated equilibrium, where employers and workers have found the privately optimal mix of pay and safety. Now suppose that the government perturbs this equilibrium, forcing a somewhat higher level of safety through regulation. Does pay go down by an offsetting amount?

In principle, yes - at least if the change is small. Here is the classical analysis. Suppose we draw supply and demand curves for the industry, with quantity of employment on the horizontal axis and effective hourly compensation (including the combined value of pay and safety, from the worker's point of view) on the vertical axis. Because of how I've defined compensation, safety regulation does not change the labor supply curve, but it does move the demand curve a little, by forcing a certain level of safety on the employer and thus making it more costly to achieve a given amount of effective compensation.

In the neighborhood of the initial equilibrium, however, the distortion imposed by the government's safety requirement is only second-order - so the demand curve barely moves, and we're more or less at the same equilibrium in (quantity,effective compensation) space. The only sizable change is that effective compensation now consists more of safety and less of wages, and the only entity experiencing first-order losses is therefore the taxman.

So in this case, workers do almost exactly "pay the cost" of increased safety with wages - but this is only true because the distortion is second-order, so that it's not a big deal anyway.

If the distortion becomes large, because safety is pushed far above its privately optimal level, then the demand curve moves inward more substantially. The incidence of the costs depends on whether labor supply or demand is more elastic. In most cases, labor supply is almost certainly much more elastic at the industry level, so the costs are mainly borne by the demand side.

Looking at the full distribution of costs in general equilibrium, things become more complicated - are the "demand side" costs borne by other factors in the industry, profits in the industry, or consumers? It depends a lot on the situation, the elasticity of substitution between factors, consumers' elasticity of demand, etc., but the best first approximation is usually that consumers pay the cost. (This is almost exactly true if all elasticities of substitution are 1.)

To summarize: first-order costs to taxman, second-order cost to consumers, much smaller second-order net cost to workers (with nearly offsetting falls in pay and rises in safety).

[Side notes: obviously I don't intend this to be a complete analysis. On the libertarian side, one could complain that I'm ignoring the administrative costs, etc. of the regulation, as well as the fact that regulation quite frequently pushes us very far indeed from the private optimum. On the progressive side, one could justly complain that I am ignoring all the subtleties about job quality and regulation that I raised yesterday in a comment on the earlier post! Still, I think it's valuable to get the econ 101 analysis clear as a benchmark.]

john hare writes:

I might have a less cynical view of safety regs if they were written by people that had some idea of what they are talking about. During one MSHA class, required to enter the mine property to work, we were instructed to leave poisonous snakes alone until they left the area. Apparently the instructor had no idea that said snake was far more likely to be under a tool or board when we went back which is far more dangerous than a snake in view.

There are numerous instances where safety regs make things more dangerous for us, not less. That doesn't even get into the cost of compliance and the advantage to those dodging the regs.

Mk writes:

Wouldn't the issue be one of potential asymmetric information on safety risks?

I think this is the intuition many people are using regarding worker safety.

Andrew_FL writes:

Daniel's argument seems to boil down to me, to arguing that people's preferences can be objectively wrong.

Either that, or that his subjective preferences are a superior substitute for those of other individual human beings. That doesn't make sense, though, so I have to assume it's the former.

For myself as well, and I would think David as well, the "right" ie optimum amount of workplace safety is determined by individual's subjective valuations of marginal additional potential workplace safey, compared to their subjective valuations of what they'd have to sacrifice on the margin to get that additional safety. For Daniel, the "right" amount of safety is something else entirely. I'd be very interested to know how he would determine that "right" amount.

Daniel Kuehn writes:

Andrew - not arguing either of those things but I'm not sure exactly what the disconnect is.

The heart of it is that individual optimism are always constrained optimums. Constrained optimums are fine for the economist explaining human behavior but not a strong basis for making moral claims unless you want to argue that people's marginal value products are a metric of their moral worth. In that sense it's a bit like Sen's approach.

MikeDC writes:
Constrained optimums are fine for the economist explaining human behavior but not a strong basis for making moral claims unless you want to argue that people's marginal value products are a metric of their moral worth.

I'd say we can make claims about morality subject to the constraint. :)

Daniel Kuehn writes:

Andrew - one more thing, I'm not asserting a "right" answer the way you appear to be. That may be a source of the confusion as well. I am asserting that your answer is not the right answer. I'm not clear there even is a right answer and I certainly don't know it.

Daniel Kuehn writes:

Talking about miners in a developed countries muddies the waters a bit on the argument I'm making. We feel a certain freedom to shrug our shoulders at that.

Consider a hungry kid riddled with disease in the developing world. The kid can't make much money for an employer. Presumably though the kid is optimizing as is his employer. We can say all that with a fair bit of confidence I think. And that's fine - as an economist I can say all that. But why in the world should I feel obligated to give any kind of moral imprimatur to that situation? How does the fact that the kid has chosen his "right" level given his constraints say anything about whether that's right in a broader sense?

As I've said a couple times now, what to do about that gets tricky. For the reasons that David outlined (and I agree with) a simple mandate is going to be problematic in most (though perhaps not all) cases. And then of course also problematic is finding what the "right" level is.

But it's definitely not the private or social optimum as economists define it - or if it is the private or social optimum, then that's purely accidental.

Andrew_FL writes:

Daniel, with respect, I can't make heads or tails of what you're trying to argue, it's getting dangerously close to solipsism now.

Daniel Kuehn writes:

Andrew -
I think I'm being clear - what are you finding confusing exactly? It's hard to clarify things when you're vague like that.

- Economic optimums are constrained optimums
- We don't have good reason to think the constraints that structure economic optimums have a lot of moral significance

Where does the confusion come in? Which concept or argument?

John hare writes:

My reference to first world mine regs was about the ability of regulators to even know how to produce safety.

Andrew_FL writes:

Daniel, I'm confused how you can judge the economic optimum and find it wanting, if you admit you have no standard to judge it against.

Comments for this entry have been closed
Return to top