Optimal Worker Safety One More Time

Matt Moore: It’s not that we think safety is a luxury good; rather, it’s that we think safety to be a normal good – that is, one for which the demand increases as income rises (and vice-versa). This belief contrasts with what I take to be the standard, man-on-the-street view – namely, that there is some minimally objective correct amount of safety that all workers, regardless of tastes or of incomes, should have or should demand.

I cannot and do not speak for David H. (or anyone else, of course), but one of two of the important features of a discussion such as this one is to dispel as mythical the notion that there is some minimally correct, objectively determinable level of safety that Jones ought to be empowered to force Smith to accept. No such level of safety exists.

The second important feature of a discussion such as this is to make clear that any and all attempts to impose on market arrangements levels of safety higher than those that are discovered and enforced through market competition will make workers worse off and not better off. Such attempts might well make workers more safe, but this fact does not mean that those safer workers are better off.

And because it’s a normal good I don’t think the argument changes substantially above subsistence as MikeDC’s (unpasted) third paragraph suggests.

This brings up an interesting thought experiment.

If Jones insists that Smith should have X level of safety, we calculate the income subsidy Y that would result in Smith actually demanding X.

Then send Jones the bill.

I agree with Daniel Kuehn, that Don’s first feature was right, but his second feature was wrong.

If we’re talking about not just mandates (which is indeed what I had in mind in my earlier comment) but also subsidies then of course workers could possibly be made better off. Transferring $X from Smith to Jones (in the form of a safer workplace) makes Jones better off as long as the costs that Jones must endure in order to receive $X are less than $X.

But this reality does not make the resulting level of worker safety for Jones optimal. Not only (as David H. correctly notes) might Jones prefer to receive $X worth of subsidies in some form other than increased workplace safety, but even if $X amount of increased workplace safety is precisely what Jones wants above all else, there is no reason to elevate Jones’s preferences above those of Smith.

The higher workplace safety enjoyed by Jones cannot be said to improve social welfare because Jones’s gains must be weighed against Smith’s losses.

This fact does not mean that such subsidized workplace safety is a normatively indefensible or bad idea. But it does mean that one cannot defend Jones’s subsidized higher workplace safety as being some objectively ‘right’ or ‘correct’ outcome. Such a transfer of funds from Smith to Jones necessarily involves the explicit value judgment that it is acceptable for Smith to be forced to part with $X worth of welfare in order that Jones gain $X-E worth of welfare (where E is the unavoidable cost of transferring funds from Smith to Jones). In short, such subsidies cannot be justified objectively as a correction of a market failure; they are instead justified purely normatively as a wealth transfer.

This logic, of course, is wide-open: it can be used to justify transferring money from Jones to Smith in order to help Smith build a bigger and more ornate stadium in which Smith’s professional football team will play.

One might, obviously and sensibly, protest that marginal increases in worker safety are more important than are marginal improvements in the size and opulence of sports stadia. And I’d likely agree (although I’d have no objective data or economic logic to rely upon to support my value judgment). But, again, unless one shows clearly some inability of markets to respond to worker demand for workplace safety – a showing that is very difficult in markets with low or no barriers to entry and exit – such subsidies have nothing whatsoever to do with market failures and everything to do with some people elevating their own value judgments over the revealed outcomes of market processes.

You’re correct. At the threshold of starvation the “value” of the additional safety is zero. The worker will always prefer an additional $0.25. Buying the extra safety puts him below the threshold of of starvation.
My point is that when you get close to the threshold things get very non-linear. People have very low values for safety close to the threshold, which then rise rapidly as you get away from it, and then fall again once minimal safety needs are met.

I think people have a moral objection to exchanges that occur when people are at the subsistence level, because just by giving them a little extra money – so they don’t face the threat of starvation, their valuations and their negotiating power change rapidly.

To further the analogy – if that guy’s income was bumped up to $2.50 a day, he might suddenly start valuing the safety at $0.50 a day. And the employer, it almost seems obvious, absolutely has it in his power to boost the workers wages so he isn’t at the threshold of subsistence. I’m assuming we’re not talking about these workers working for some other people making $5 a day, but for some first-world industrial company. So again it seems immoral to treat the workers calculation as “optimal” in that situation when a marginal increase in his income would change that calculation.

Again in general I suspect that safety-vs.-wage calculations are usually done correctly by the market. Just not at subsistence thresholds.

Hazel Meade writes, “Again in general I suspect that safety-vs.-wage calculations are usually done correctly by the market. Just not at subsistence thresholds”

The fifth word here is the most telling. We do not know, observe, calculate or notice the optimal bundle of safety and wages. We can only suspect this or that and we cannot tell even in hindsight if our suspicions were correct.

Those people who favor government imposed safety requirements should give this careful thought and recognize that they cannot even claim to know whether their preferred policy would improve or worsen the circumstances of the workers. This is true in every case, whether the wage level is close to subsistence or far from it.

“This whole discussion has been using at least three definitions of “optimal”. I’ve never disagreed on this point or David’s account of the private, constrained optimization decisions of Jones.”

That’s exactly the problem. When someone uses the word optimal in a very narrow sense, you are bound to get responses from people who are confused. Also, the economic argument of optimality rests on a set of assumptions that do not necessarily hold in practice. There is a large literature on market frictions/failures in the labor market. People have these frictions in the back of their mind but have trouble articulating in economic terms.

Also, the analyses and Don Boudreaux’s comment make it sound as if it’s devoid of any subjective judgment. This is not correct. You can talk of similar optimal outcomes based on coercion. Rubinstein, for instance, has an interesting model of optimal behavior in the jungle:

http://arielrubinstein.tau.ac.il/papers/eqjungle.pdf