Data on claims of sexual harassment … are used to calculate the first measures of sexual harassment risks by industry, age group, and sex. Female workers face far higher sexual harassment risks. On balance, workers receive a compensating wage differential for exposure to the risk of sexual harassment. … The … wage difference between a job with zero sexual harassment risk and a job with the mean sexual harassment risk is … about 25 cents per hour for women, and … about 50 cents per hour for men. (
As with anti-discrimination laws, it is only employers who are constrained.
That's because sexual harassment is a power issue: one feels less power to defend oneself from unwanted advances when one's economic situation depends directly on the harasser. When such pressure exists, the decision of the harassee is made, to some extent, under duress. I presume you understand why coercion is and should be illegal.
If firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures the firm would internalize the problem. Thus it would make economic sense to let each firm’s management decide whether or not to allow harassment in their firm.
This presumes that sexual harassment would show up on the balance sheet in a way that couldn't be explained some other way, for which there would be plenty of incentive for both the harasser and the firm itself. It also presumes that businessmen make their decisions based entirely on economic factors, which is also not true. Businessmen are human, and some humans really want to work in a place where they won't get in trouble for playing grabass with the secretary, or they care more about keeping a talented but horny exec on board than they do about whatever economic losses are due to harassment. That might even be a wise decision, in terms of the bottom line.
For the prospective harassee, it's not like the invisible hand can be all that effective in this case: there is no information on the sexual harassment environment at a workplace in the Classifieds. This is an enormous information gap, only reliably bridgeable by taking a job and then seeing if you get harassed or not. Given that frequent job changes have a negative effect on one's potential earnings, one can see how merely taking two or three wrong jobs in a row, then quitting because of harassment, could scuttle an otherwise promising career. Can you not see the legitimate role of government in preventing such a reduction in a person's hirability because they refused someone's sexual advances?
Robin finds it plausible that the reported differential of 25-50 cents/hour is in effect a reasonable fee for being harrassed. Suppose the study had found a much larger differential; presumably Robin would have said the same. Suppose, on the other hand, the differential had been zero.
Case 1. (I find this the more plausible, for what it's worth.) Robin would have said: it seems that people don't actually mind harrassment all that much, since they evidently aren't prepared to sacrifice much wealth to avoid it. -- In this case, Robin's conclusion is basically unresponsive to the actual data, and the actual logic goes something like: "By definition, the wage anyone's prepared to accept is fair compensation for their actual work and working conditions; therefore there's no reason to oppose anything that goes on in the workplace, and anyone proposing to legislate such things is merely signalling."
Case 2. Robin would then have said: Hmm, apparently differences in pay don't correspond to differences in employees' quality of life when it comes to harrassment. -- In this case, surely there's nothing magic about the exact figure of $0/hour; any small enough differential should lead to much the same conclusion. Why isn't 25-50 cents/hour small enough?
A couple of other remarks.
1. As Unnamed says: If these numbers really mean what Robin takes them to mean, isn't it curious that women (who apparently suffer much more harrassment, and certainly report it more often) value not being harrassed *less* than men?
2. At present, there *are* laws, and workplace policies, and social norms, that prohibit harrassment. Presumably they have some effect. If so, then in Robin's hypothetical richer freer world with no anti-harrassment rules, the differentials (which Robin interprets as fees for harrassment) would have to be much larger than the ones found by this study.
Let's see... women require a $0.25 per hour compensating differential to have a job with the mean level of sexual harassment (8.6 filed cases of sexual harassment per 100,000 workers) rather than zero sexual harassment. Assuming 2000 hours worked per year, that means $500 per year, or $5.8 million per filed case of sexual harassment. Men require a $0.50/hr compensating differential and have a mean level of 1.4 filed cases of sexual harassment per 100,000 workers, which means $1,000 per year or $74 million per filed case of sexual harassment.
These numbers don't represent the cost of one instance of sexual harassment, since many instances of sexual harassment never get reported or get handled internally and don't result in filed cases. But why the big (12x) difference for women and men? I'd guess that sexual harassment of males is wildly underreported, or something funny is going on with the regression (sexual harassment rates are correlated with other bad workplace conditions?).
This is interesting, if maybe not cogent. Simply because there is a premium does not mean that the premium fully compensates the employee. There are probably numerous factors correlated with level of sexual harassment. It is entirely possible that many employees take these jobs, but would willing forsake all or more than all of the differential if it made their risk of harassment zero.
Moreover, it's entirely unclear the degree to which this functions as a salary for the harassers. A lot of sexual harassment, as I understand it, is basically maladroit sexual advances, which are of little value to anyone. It's also likely that the marginal value of being able to harass is close to zero, in that, if people were completely forbidden from harassing, they would not change their occupational behaviour at all. But employee costs would go down. That would seem like a superior equilibrium for both the harassed employee and the owner who has to pay the differential.
Voting is anonymous (I recall John Lott saying voting rates dropped a lot after secret ballots reduced the feasibility of credible bribes). How effective a signal can it be?
Again I'll link to Gelman on Peter Dorman's critique of the theory of compensating differentials (or at least one application of it).
The company and shareholders are deriving no benefit from the sexual harassment of employees, only those doing the harassing are benefiting.
Do the managers reimburse the employer for the sexual harassment differential? Otherwise they are stealing that harassment differential from the company.
The authors find a wage difference of 25 to 50 cents/hr for the mean harrassment level. Does a wage cut of 25 to 50 cents/hr (for a zero-harrassment workplace) raise big redistributional concerns?
“If these wage changes were the only effect of harassment, there would be no economic reason to oppose harassment…What if others were effected?…If firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures the firm would internalize the problem.”
No matter what, there is no “economic” (i.e. efficiency) reason to change anything other than what one observes. If externalities can be efficiently corrected, the affected parties pay to have them corrected; if they can’t be efficiently corrected, then the affected parties don’t.
"If these wage changes were the only effect of harassment, there would be no economic reason to oppose harassment – harassers would be paying the harassed an agreeable fee, and no one else would be effected."
Adverse selection is a big problem here: not all males are would-be harrassers and not all females are willing to be harrassed, and sometimes females are harrassers and males are harrassed. There is moral hazard as well: if there is no compensation for marginal increases in harrassment the incentives will be misaligned, and a lot of harrassment will happen that could otherwise be cheaply avoided.
The link goes to a paper studying fertility, not the harassment paper by Joni Hersch.
Also, the Coase theorem only implies efficient internalization; it doesn't say anything about distributional effects vis a vis harassers and the harassed. Presumably one might desire that the harassers pay the harassed a wage premium for the privilege, not that the harassed put up with harassment for employment at the prevailing wage.
And atop that, "if firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures..." is a nontrivial condition. Even if it held 'most of the time', there would be edge cases where it does not and the individuals involved may feel extremely aggrieved.
As with anti-discrimination laws, it is only employers who are constrained.
That's because sexual harassment is a power issue: one feels less power to defend oneself from unwanted advances when one's economic situation depends directly on the harasser. When such pressure exists, the decision of the harassee is made, to some extent, under duress. I presume you understand why coercion is and should be illegal.
If firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures the firm would internalize the problem. Thus it would make economic sense to let each firm’s management decide whether or not to allow harassment in their firm.
This presumes that sexual harassment would show up on the balance sheet in a way that couldn't be explained some other way, for which there would be plenty of incentive for both the harasser and the firm itself. It also presumes that businessmen make their decisions based entirely on economic factors, which is also not true. Businessmen are human, and some humans really want to work in a place where they won't get in trouble for playing grabass with the secretary, or they care more about keeping a talented but horny exec on board than they do about whatever economic losses are due to harassment. That might even be a wise decision, in terms of the bottom line.
For the prospective harassee, it's not like the invisible hand can be all that effective in this case: there is no information on the sexual harassment environment at a workplace in the Classifieds. This is an enormous information gap, only reliably bridgeable by taking a job and then seeing if you get harassed or not. Given that frequent job changes have a negative effect on one's potential earnings, one can see how merely taking two or three wrong jobs in a row, then quitting because of harassment, could scuttle an otherwise promising career. Can you not see the legitimate role of government in preventing such a reduction in a person's hirability because they refused someone's sexual advances?
it's only flirting if it's a two way street.
you say "harassment", i say "flirting"
See my update to this post.
I'll defer to others on the reliability of the statistics. The AER published it, which I took as a measure of quality of the analysis.
Robin finds it plausible that the reported differential of 25-50 cents/hour is in effect a reasonable fee for being harrassed. Suppose the study had found a much larger differential; presumably Robin would have said the same. Suppose, on the other hand, the differential had been zero.
Case 1. (I find this the more plausible, for what it's worth.) Robin would have said: it seems that people don't actually mind harrassment all that much, since they evidently aren't prepared to sacrifice much wealth to avoid it. -- In this case, Robin's conclusion is basically unresponsive to the actual data, and the actual logic goes something like: "By definition, the wage anyone's prepared to accept is fair compensation for their actual work and working conditions; therefore there's no reason to oppose anything that goes on in the workplace, and anyone proposing to legislate such things is merely signalling."
Case 2. Robin would then have said: Hmm, apparently differences in pay don't correspond to differences in employees' quality of life when it comes to harrassment. -- In this case, surely there's nothing magic about the exact figure of $0/hour; any small enough differential should lead to much the same conclusion. Why isn't 25-50 cents/hour small enough?
A couple of other remarks.
1. As Unnamed says: If these numbers really mean what Robin takes them to mean, isn't it curious that women (who apparently suffer much more harrassment, and certainly report it more often) value not being harrassed *less* than men?
2. At present, there *are* laws, and workplace policies, and social norms, that prohibit harrassment. Presumably they have some effect. If so, then in Robin's hypothetical richer freer world with no anti-harrassment rules, the differentials (which Robin interprets as fees for harrassment) would have to be much larger than the ones found by this study.
Let's see... women require a $0.25 per hour compensating differential to have a job with the mean level of sexual harassment (8.6 filed cases of sexual harassment per 100,000 workers) rather than zero sexual harassment. Assuming 2000 hours worked per year, that means $500 per year, or $5.8 million per filed case of sexual harassment. Men require a $0.50/hr compensating differential and have a mean level of 1.4 filed cases of sexual harassment per 100,000 workers, which means $1,000 per year or $74 million per filed case of sexual harassment.
These numbers don't represent the cost of one instance of sexual harassment, since many instances of sexual harassment never get reported or get handled internally and don't result in filed cases. But why the big (12x) difference for women and men? I'd guess that sexual harassment of males is wildly underreported, or something funny is going on with the regression (sexual harassment rates are correlated with other bad workplace conditions?).
This is interesting, if maybe not cogent. Simply because there is a premium does not mean that the premium fully compensates the employee. There are probably numerous factors correlated with level of sexual harassment. It is entirely possible that many employees take these jobs, but would willing forsake all or more than all of the differential if it made their risk of harassment zero.
Moreover, it's entirely unclear the degree to which this functions as a salary for the harassers. A lot of sexual harassment, as I understand it, is basically maladroit sexual advances, which are of little value to anyone. It's also likely that the marginal value of being able to harass is close to zero, in that, if people were completely forbidden from harassing, they would not change their occupational behaviour at all. But employee costs would go down. That would seem like a superior equilibrium for both the harassed employee and the owner who has to pay the differential.
Voting is anonymous (I recall John Lott saying voting rates dropped a lot after secret ballots reduced the feasibility of credible bribes). How effective a signal can it be?
Again I'll link to Gelman on Peter Dorman's critique of the theory of compensating differentials (or at least one application of it).
Quantifying Richard Posner's rape license! Nice.
The company and shareholders are deriving no benefit from the sexual harassment of employees, only those doing the harassing are benefiting.
Do the managers reimburse the employer for the sexual harassment differential? Otherwise they are stealing that harassment differential from the company.
The authors find a wage difference of 25 to 50 cents/hr for the mean harrassment level. Does a wage cut of 25 to 50 cents/hr (for a zero-harrassment workplace) raise big redistributional concerns?
“If these wage changes were the only effect of harassment, there would be no economic reason to oppose harassment…What if others were effected?…If firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures the firm would internalize the problem.”
No matter what, there is no “economic” (i.e. efficiency) reason to change anything other than what one observes. If externalities can be efficiently corrected, the affected parties pay to have them corrected; if they can’t be efficiently corrected, then the affected parties don’t.
Inefficiency cannot be observed.
"If these wage changes were the only effect of harassment, there would be no economic reason to oppose harassment – harassers would be paying the harassed an agreeable fee, and no one else would be effected."
Adverse selection is a big problem here: not all males are would-be harrassers and not all females are willing to be harrassed, and sometimes females are harrassers and males are harrassed. There is moral hazard as well: if there is no compensation for marginal increases in harrassment the incentives will be misaligned, and a lot of harrassment will happen that could otherwise be cheaply avoided.
The link goes to a paper studying fertility, not the harassment paper by Joni Hersch.
Also, the Coase theorem only implies efficient internalization; it doesn't say anything about distributional effects vis a vis harassers and the harassed. Presumably one might desire that the harassers pay the harassed a wage premium for the privilege, not that the harassed put up with harassment for employment at the prevailing wage.
And atop that, "if firm managers had strong enough incentives to maximize profits, then in the absence of other relevant market failures..." is a nontrivial condition. Even if it held 'most of the time', there would be edge cases where it does not and the individuals involved may feel extremely aggrieved.