--Sebastian
For some reason I was reading about the Ledbetter case, and recent Congressional attempts to reverse the Supreme Court ruling. The thing the struck me most, was how odd it was that Congress was letting itself get stuck in the narrow role of directly addressing the Supreme Court's ruling in a 'yes' or 'no' fashion.
The actual history of Title VII in this respect is pretty clear. Congress passed it with a surprisingly short statute of limitations: you have to sue within 6 months of the discriminatory act in order be covered under the statute. As most statutes of limitations in civil matters are in the 1-3 year range, this seems a bit severe. But as Title VII seemed most likely at the time to deal with hirings as the discriminatory act in question, it might not have been totally crazy. Quickly the most lucrative cases for plaintiff's attorneys turned out to be the firing cases and later the cases involving raises. This may not have been the original intended effect of Title VII but it is clearly within the language of the statute so I have no problem with it.
But at some point, someone in the Justice Department clearly thought the 6 month window wasn't fair for cases involving raises. From a policy perspective that may well be true (I suspect that discrimination in pay might be trickier to discover than at point of hiring, and that the 6 month window was too short in any case). So they developed a lawyerly workaround by which the discriminatory act wasn't the decision at the time of the raise, but rather the event of issuing each paycheck. The Supreme Court essentially ruled that such an approach was too lawyerly a workaround from the pretty clear Congressional time limits and rolled it back to 6 months from the discriminatory act, where that counts from the time of the actual discriminatory decision.
This was a statutory ruling, not a Constitutional ruling. Therefore Congress can change the statute if they think that the policy implications of the ruling aren't to their liking.
The thing that is odd to me, is that the Congressional attempts to deal with changing the policy implications are stuck in the frame of the original lawyerly workaround. The Lilly Ledbetter Fair Pay Act just ratifies the workaround, and says that the 6 month statute of limitations starts again with every new paycheck. But statutes of limitations exist in law for very good reasons and just legislating the exact content of the lawyerly workaround ignores those policy issues. First exculpatory evidence can be lost or destroyed at some point. The lawyerly workaround makes it very possible that an allegedly discriminatory act could be made 15 years before, the person who made it left immediately, and that no further discriminatory choices were made, but that the company is on the hook for so long as they pay the employee payroll checks. Even much more serious criminal cases don't get 15 or 20 year statutes, largely because of evidentiary issues. (Murder is almost the only crime that has an unlimited statute of limitations. Rape does in some jurisdictions, but not very many). Republicans who note that there is something odd about that, are correct.
But why get stuck in that frame at all? If the statute of limitations is too short, lengthen it in a way that still takes the idea of a statute of limitations and the policy implications of it seriously. Make the statute for pay raise cases 2 or 3 years. Maybe even 5. But making it effectively unlimited seems like you are getting caught in the frame of the original case, when that was all about a lawyerly workaround. Or it makes it seem like you are being unfairly punitive by extending the statute far beyond what we do for even very serious crimes. Why do that when you are revisiting the policy anyway and getting rid of the need for a lawyerly workaround?
Update: I want to be clear about the distinctions I'm making. In a normal statute of limitations case (criminal or civil) we would think it was ridiculous to try to bring a case (either through prosecution or civil proceeding) of a 20 year old pay raise decision if the person quit that job the next year. There would be 19 years for exculpatory documents to get destroyed, 19 years of memories about the incident getting stale, 19 years of other reasons why that person didn't get the raise being misconstrued in the plaintiff's memory, etc. Even for most very serious crimes we don't typically allow proceedings that old to go forward because of such factors. Bank robbery for instance typically has a statute of 5-7 years. Federal kidnapping has a statute of 5 years. These exist not because we think it is fantastic for bank robbers and kidnappers to get away with it after 5 years, but because we understand that it is very difficult to have a fair proceeding on it so much later. Such considerations are even more true in a discrimination case where much of the evidence is necessarily going to have a subjective cast to it. (Was she really underperforming? Did other legitimate factors come into play?)
If you agree that a statute of limitations makes sense in the case of a 20 year old discriminatory act that took place if the person quit the job 19 years ago, it doesn't make sense to prosecute a discriminatory act that took place 20 years ago if the person did not quit the job.
Republicans who note that there is something odd about that
think that once a crime has been going on for long enough, they should be allowed to get away with it.
but that the company is on the hook for so long as they pay the employee payroll checks
The company is on the hook for so long as they pay a black employee less than white employees, pay women less than men, pay Jews less than Christians, etc.
How does it seem unfair to you that the ability to sue for discrimination continues for as long as the period of discrimination, plus six months?
Posted by: Jesurgislac | July 12, 2008 at 01:35 PM
Quite frankly, the "lawyerly workaround" strikes me as the right policy. First for practical reasons: it creates a strong incentive for employers to make sure they are not engaging in discriminatory pay practices. Second for moral reasons: it just seems obvious to me that in such cases the discrimination is on-going, rather than isolated at the point at which pay was set.
Posted by: jdkbrown | July 12, 2008 at 01:46 PM
I've always found this whole business confusing.
First, the notion that issuing a paycheck is a discriminatory act does not seem to me to be a "lawyerly workaround" but rather common sense. If Jack and Jill do the same work in April it seems to me to be plainly discriminatory if Jill's April paycheck is smaller than Jack's, regardless of when their pay levels were decided. The Court's finding to the contrary, if I understand it correctly, seems just silly. It's not as if the payroll system, once set in motion, is uncontrollable by management.
Second, if the new six month period begins with each paycheck, is the company's liability also limited to the pay difference over the previous six months? If so the company isn't really "on the hook" for very much.
Finally, I don't agree that it's
Employees' pay is normally reviewed at annual or semi-annual intervals. Even if you disagree with the idea that issuing a discriminatory paycheck is a discriminatory act, surely it is not unreasonable to expect whoever does the reviewing to notice the discrimination. Why wouldn't failure to do so, and correct it, be a discriminatory act?
Posted by: Bernard Yomtov | July 12, 2008 at 01:57 PM
I suspect that is because you don't value statutes of limitations. If you don't think there are any justice concerns in allowing lawsuits (or prosecutions) to be open indefinitely, then limiting it won't make sense. But if you are going to disregard the decision-making mark, why not just lift the statute of limitations entirely and make it analogous to murder?
Posted by: Sebastian | July 12, 2008 at 01:59 PM
"Second, if the new six month period begins with each paycheck, is the company's liability also limited to the pay difference over the previous six months? If so the company isn't really "on the hook" for very much."
No. You are forgetting about punitive damages. And that is precisely why I think an effectively unlimited statute of limitations doesn't make sense. If you have a discriminatory act 20 years ago and then standard raises for each of 19 years since then, it doesn't make sense to exercise punitive damage awards for a 20 year-old act that you wouldn't have punished if the person had left the company.
If you don't think it is right to allow lawsuits with punitive damages if they had left the company 15 years ago for a discriminatory pay raise 20 years ago, it makes no more sense to change that policy just because they continued working there.
Posted by: Sebastian | July 12, 2008 at 02:04 PM
The answer to the question of limiting liability depends on who you think should have the burden of correcting the harm. The disparate paycheck is an ongoing disparity. It not only doesn't stop at the time of the decision, but it replicates when raises are given.
If you think the burden should be on the employer to correct the harm - i.e. by having an ongoing duty to review their pay scheme and ensure it's fairness, then the renewing SOL makes sense.
If you think that it is the duty of the employee to complain and attempt correction, then the shorter SOL makes sense.
The problem with your approach is that pay rates are often secret and the employee has no way to know of the discrimination. By the time s/he realizes that they are being paid less for the same job, they may be years from the decision point.
Given the information disparity, it makes sense that the employer, who knows everyone's comparative pay, and work duties, and has control of the decision makers, should bear the burden. They also should, just as good business practice, review their pay scheme from time to time to correct any disparities.
It would be inequitable to limit the employees SOL, unless you include a "knew or should have known" requirement also.
Posted by: plainbrown1 | July 12, 2008 at 02:16 PM
"Given the information disparity, it makes sense that the employer, who knows everyone's comparative pay, and work duties, and has control of the decision makers, should bear the burden. They also should, just as good business practice, review their pay scheme from time to time to correct any disparities."
Unless the level of discrimination is enormous or unless the file says "I gave her a lower raise because she was a woman" that isn't going to turn it up.
"The problem with your approach is that pay rates are often secret and the employee has no way to know of the discrimination. By the time s/he realizes that they are being paid less for the same job, they may be years from the decision point."
Again this is true even if the person quit the job 19 years ago and the decision point was 20 years ago.
Posted by: Sebastian | July 12, 2008 at 02:27 PM
Sebastian,
I do appreciate the need for statutes of limitations. But if punitive damages are the issue then why not address them rather than the SOL?
Besides, treating only the decision as a discriminatory act seems to lead to some oddities. Suppose a company awards year-end bonuses to employees - a not uncommon practice. In January management decides that female employees, all else equal, will receive 10% smaller bonuses than males. In December the checks are issued. Well, more than six months has passed. Should there be no valid discrimination claim?
Posted by: Bernard Yomtov | July 12, 2008 at 02:28 PM
Well first, I think the 6 month window is too short. So if you want to deal with that make it 2 years like most civil issues. That gets rid of the bonus problem right there.
My whole point is that it is silly to get stuck in the narrow workaround that the Justice Department created to avoid the really short statute of limitations when instead you could just make it a perfectly reasonable statute of limitations with all of the built in balances that entails.
Posted by: Sebastian | July 12, 2008 at 02:37 PM
Sebastian: If you have a discriminatory act 20 years ago and then standard raises for each of 19 years since then
You're still being discriminated against.
If I'm hired at 30K a year, and a man is hired at the same time for the same job with no better qualifications and experience than me but he gets 40K a year because the man who decides salaries is like that, then for the next 20 years, as each of us gets a 3-5% pay rise each year decided based solely on our achievements in the past year without any discrimination, the discriminatory pay differential will still exist.
And, as noted, even if the person originally responsible for the decision has long since left the company, it would have been possible for the company to figure they were being discriminatory and give the woman who was hired at 30K a substantial pay raise to equalise her with the man, at any time in the 20 years.
If they never do, they are perpetuating the discrimination, and they still deserve to get sued.
Why do you feel that once someone has been discriminated against for a few years, it ought to be legal to maintain that discrimination in perpetuity?
Posted by: Jesurgislac | July 12, 2008 at 02:45 PM
I'd be happy to keep a strictly limited statute of limitation if businesses were required to publish their pay rates for each employee. Otherwise, no, this is just another example of using rules to screw the designated class of victims again.
Posted by: Bruce Baugh | July 12, 2008 at 03:01 PM
"even if the person originally responsible for the decision has long since left the company, it would have been possible for the company to figure they were being discriminatory"
This isn't a true statement for any except the most expansive definitions of 'possible' unless the pay decision at the time is obviously discriminatory (memo saying "I pay women less").
Posted by: Sebastian | July 12, 2008 at 03:21 PM
Oh, baloney, Sebastian. Establishing gross patterns of discrimination in pay is a matter of minutes' looking at spreadsheets, given competent accounting, even without any access at all to hiring decisions. Spotting anomalies is a few minutes' more work. At least it would have been with the kind of accounting systems I worked with when office temping in the early '90s, and the technology's improved just then.
Get the list of employees in a given category. Sort and graph, and see if there are patterns that suggest discrimination. Investigate if so. Throw in additional data to resolve ambiguous cases. Proceed with a policy of pay review that makes a particular point of valuing comparable work comparably, and see what happens to the anomalies then. Yeah, the details are more complicated than that, but the kind of thing that often leads to charges can well be that flagrant and avoidable, if anybody cares. From what I understand, the Ledbetter case is one of those - the disparity wasn't subtle.
Posted by: Bruce Baugh | July 12, 2008 at 03:27 PM
Sebastian: This isn't a true statement for any except the most expansive definitions of 'possible' unless the pay decision at the time is obviously discriminatory
What Bruce said.
With access to the kind of information that any company has on their employees, a company can very easily find out if discriminatory pay decisions are being made on the basis of gender.
If they want to, that is. But if they know that if they can get away with it for six months - or five years - then the employees they're paying less money will never even be allowed to sue them, why should they care about finding out if they are being discriminatory - still less, about fixing it?
Posted by: Jesurgislac | July 12, 2008 at 03:33 PM
Or, put another way:
For a company to decide they're going to pay some employees less than others because of their gender, race, religion, sexual orientation, gender identity, or disability, is just plain wrong.
For a company to escape retribution for so doing merely because they can pin the blame for the decision on an individual employee who no longer works for the company, is also wrong.
If a company does not want to be sued for discriminatory rates of pay, their proper solution is to ensure that they do not have discriminatory rates of pay, not to mess with the law so that it makes it more difficult - or impossible - for employees whom they have exploited for years to sue them for doing so.
End of story. "Lawyerly" forsooth.
Posted by: Jesurgislac | July 12, 2008 at 03:55 PM
I recall when first hearing the case that the simplest workaround would be to not start the clock until the disparity in pay (or other discriminatory action) was revealed to the employee.
I particularly like that solution because I think it would encourage companies to share payroll information (even in the general sense of 'managers at X level average Y per paycheck) to start the clock as soon as possible. Whereas the current regimen encourages corporate secrecy, and the proposed solution doesn't seem to fix that very much (balancing the advantage of uncovering issues earlier v the advantage of keeping them from ever seeing the light).
Maybe also add a one-year window where companies can come clean & grant all appropriate back pay etc without facing punitive damages. On the one hand, it forgives (somewhat) bad behavior, otoh it would maybe get a large number of people money that they deserved but never received.
These exist not because we think it is fantastic for bank robbers and kidnappers to get away with it after 5 years, but because we understand that it is very difficult to have a fair proceeding on it so much later.
That doesn't seem right; it suggests that it's Ok to have less fair proceedings (ie further from the events) for more serious crimes. I always thought of this as a factor of the state's concern with the events- solving a robbery from 30 years ago is just not that important. Of course, that could be handled by policy without explicitly making the crime unprosecutable.
Or maybe it's related to some older legal traditions (eg a serf who left his estate in medieval Europe could be free from his overlord after a year and a day without recapture)- for a small crime committed at 20, should a man always have to look over his shoulder?
Posted by: Carleton Wu | July 12, 2008 at 05:25 PM
"Get the list of employees in a given category. Sort and graph, and see if there are patterns that suggest discrimination. Investigate if so. Throw in additional data to resolve ambiguous cases. Proceed with a policy of pay review that makes a particular point of valuing comparable work comparably, and see what happens to the anomalies then. Yeah, the details are more complicated than that, but the kind of thing that often leads to charges can well be that flagrant and avoidable, if anybody cares."
If you are talking about 30-40% difference sure. Most cases turn on far less than that and people in the same category often have somewhat different pay, so we aren't talking about that typically. Forcing a company to constantly investigate all of those differences every six months is ridiculous. This is not just about the Ledbetter case. Her case may have been flagrant enough to turn up with a normal review. Lots of other ones wouldn't, and we are talking about making a rule for all of them. People in the same gross category don't all get paid exactly the same amount. A spreadsheet isn't going to detect a 20 year-old discrepency like that that has carried forward unless it is enormous.
Posted by: Sebastian | July 12, 2008 at 06:07 PM
"That doesn't seem right; it suggests that it's Ok to have less fair proceedings (ie further from the events) for more serious crimes. I always thought of this as a factor of the state's concern with the events- solving a robbery from 30 years ago is just not that important. Of course, that could be handled by policy without explicitly making the crime unprosecutable."
I'm pretty sure most statute of limitations are a balance of importance of the crime and fairness. A murder is important enough that there is typically no statute of limitations. Almost anything else gets to a 5-7 year statute very quickly. Kidnapping, bank robbery, assault, etc. The risk of unfairness in those cases isn't worth their importance. In a discrimination case the risk of unfairness is even higher than a bank robbery (statements and actions of people 20 years ago--possibly people who are dead--being construed without the hope of finding context) and the importance isn't considered as high as murder. Most civil cases have a statute in the 1-3 zone. That includes things like accidental poisoning to death and the like.
Heck if we wanted to have the tried and true discovery rule (when you knew or should have known) start the statute I'd be fine with that. I'm not fine with an essentially unending statute of limitations for this one thing when we have statutes of limitations for lots of other important things. The reasons we have statutes of limitations are important. Sweeping all that away without seriously looking at it doesn't seem wise.
Posted by: Sebastian | July 12, 2008 at 06:15 PM
I think that Sebastian's basic point -- that it's not obvious that the best way to address this is via the statute of limitations - is right. The basic problem, after all, is that employees might not know about the discrimination. If there were a way to ascertain when they found out, then possibly the best fix would be just to start the clock there.
(I have no idea how the statute of limitations works in cases that are analogous, in terms of the victim probably not knowing about the crime. For instance, if you dump some carcinogen into the water supply in secret, and no one finds out until cancer rates shot up years later, are you off the hook if the statute of limitations has expired?)
I think it's a policy question whether the right answer to this is: (a) start the clock when the victim finds out, or (b) set something up (e.g., mandatory revealing of salaries without identifying information, but with gender, race, etc.) that would enable anyone who wanted to know whether or not s/he was being discriminated against to check, or (c) taking each paycheck to be a new discriminatory act. All would address the basic problem, and it's not clear that (c) is the best one.
That said, it might very well be the best. How do you tell when someone really finds out? Would it be possible to post the salary info with all the categories most likely to be a basis for discrimination, but in such a way that individuals could not be identified? What if there is, in fact, only one female VP in a company -- do we want to mandate that her salary info, and only hers, will be identifiable as the salary info of a specific individual? Etc., etc., etc.
For this reason, while I'm quite prepared to believe that there's a better fix out there, and while (b) might be that fix, depending on how you balance risks to individual privacy vs. giving companies a clear incentive not to discriminate, (c) strikes me as a pretty good one. Certainly, it does a good job of ensuring that companies cannot just keep lawbreaking secret for 6 months and then get off scot free, if their lawbreaking is continuing to harm people, and they could (but w/o (c) would have no incentive to) fix it.
That's what I want most to ensure: that they can't do this and say: well ha ha, we kept it secret for 6 months, and there was no realistic way you could possibly have known: too bad!
Posted by: hilzoy | July 12, 2008 at 06:19 PM
I don't think that the statute of limitations originally had anything to do with the fairness of the proceedings as proceedings are automatically assumed to have the imprimatur of justice under God. My thought is that the statute of limitations imagined that people would make amends somehow for lesser crimes, but more serious crimes, the weight of the crime requires that making amends is only accomplished by some public admission, so the case that Hilzoy suggests, if brought as a manslaughter case, would not be limited. While America took the lack of limitations from English common law, the lifting of the statute of limitations for murder only took place in France and Germany after WWII and then only for crimes against humanity. I think that Germany has discarded it for common murder as well, but for France, it is now 10 years after the investigation is closed which seems appropriate for a country that gave us Inspector Javert.
Posted by: liberal japonicus | July 12, 2008 at 06:23 PM
If you are talking about 30-40% difference sure. Most cases turn on far less than that and people in the same category often have somewhat different pay, so we aren't talking about that typically. Forcing a company to constantly investigate all of those differences every six months is ridiculous.
I don't think it takes a 30-40% discrepancy. Even a much smaller one showing up consistently would be enough. That would certainly be true in a case where a number of people are doing essentially the same job. Plus, additional information is available - performance reviews at least - so there's some sort of record to go by.
I doubt it is necessary to do this every six months, and in any case merely generating and looking at the kind of report Bruce describes isn't that hard. You're talking about a few pages generated by the payroll system using data that's likely there already.
This is not just about the Ledbetter case. Her case may have been flagrant enough to turn up with a normal review. Lots of other ones wouldn't, and we are talking about making a rule for all of them. People in the same gross category don't all get paid exactly the same amount. A spreadsheet isn't going to detect a 20 year-old discrepency like that that has carried forward unless it is enormous.
I'm not sure what the objection is here. You seem to be saying that it's very hard to prove discrimination, so we should restrict discrimination suits.
Posted by: Bernard Yomtov | July 12, 2008 at 07:08 PM
kidnapping, bank robbery, assault, etc. The risk of unfairness in those cases isn't worth their importance.
It seems to me that the risk of unfairness is the same, if we assume that society is balancing crime and punishment equitably. I am curious how this sort of thing works in real life- could a prosecutor delay charging a murder until an elderly alibi witness had died?
[From Wikipedia, Canada apparently has no statue of limitations for serious crimes (eg murder, arson, stealing non-American beer). I wonder whether this makes much difference in the practice of criminal law.]
Posted by: Carleton Wu | July 12, 2008 at 08:18 PM
We argued about this the last time this came up, and I just have to say that you simply do not understand the law or policy on this subject. I will only try to cover a few points.
The accrual theory in question is not some oddball "workaround." It is a basic rule for statute of limitations in those class of cases known as continuing wrongs (and I have researched this subject to death in connection with appeal work I have done on the question, which I am proud to say resulted in a published decision affirming this rule in another context). There was nothing in the statutory language that mandated the result here.
The result in Ledbetter was based on the assumption that in a discriminatory pay case, there is only one discriminatory act, and that is the initial pay decision. Allegedly, each subsequent underpayment was not a discriminatory act -- just some sort of neutral act. Frankly, that is crappy sophistry not mandated by law, evidence or policy -- just a niggardliness about screwing employees who try to make such claims.
However, when Ledbetter came up on appeal in the Circuit Court, the plaintiff conceded without argument the contention that the subsequent pay acts were not accompanied with any discriminatory intent. On its facts, the result in Ledbetter was correct given this concession, and note how the conservative majority when out of its way to fashion a more general rule even thought the actual factual question was not before it.
Frankly, that was just stupid lawyering by Ledbetter's counsel. If you have adequate proof of an initial intentional act of discrimination in pay, what happens for the next paycheck -- the employer just forgot the reason for underpayment? Gee, I don't know why we are paying her less, but I just have to follow the pay chart? By what logic can you defend the reasoning that the first payment was an act of discrimination, but subsequent underpayments are just business as usual? THAT is the logic of this decision -- it is not grounded on any statutory language but that nonsensical logic.
What the 180 day statute of limitation does is cut off damage recovery for older time periods, unless there is a delayed accrual (which I do not think there is in this type of case, but I am uncertain). In other words, the discrimination may have been going on for years, but the recovery is limited to recent pay periods. THAT is the traditional and proper application of the statute of limitations in this instance.
What is an employer to do to protect himself from old acts of discrimination carried forward into the present? Periodically peruse the payroll to see that it is fair and not grounded in past discrimination? The correct rule encourages that policy -- the actual decision lets the employer off and allows it to perpetuate discrimination without concern. That is a very crappy policy.
Ironically, another woman just hired who was treated similarly to Ledbetter in the same time period in question could make the claim, but Ledbetter must continue to suffer discrimination without remedy, unless you consider quitting the proper remedy. Think about the nonsense in that, some employees are lawfully discriminated against when others are not, and the only difference is that newer employees can sue but older ones cannot FOR THE SAME MISCONDUCT?
Also, in the common law, delayed accrual of the claim is the traditional rule when the wrongdoing is an intentional act of misconduct such as discrimination. The 180 day statutory rule here is actually extremely short in relation to other intentional wrongs. The classic example of this is fraud, in which the clock does not even start to tick until you knew or should have known of the misconduct. There is no delayed accrual for equal pay cases , but we have for a very long time (over a century) had a far less restrictive view of the proper application of statute of limitations than you advocate is proper.
There is no wisdom or logic to this decision other than creating high burdens to make a claim.
Posted by: dmbeaster | July 12, 2008 at 10:39 PM
Hilzoy: "I have no idea how the statute of limitations works in cases that are analogous, in terms of the victim probably not knowing about the crime. For instance, if you dump some carcinogen into the water supply in secret, and no one finds out until cancer rates shot up years later, are you off the hook if the statute of limitations has expired?"
This is normally done with the discovery rule. As soon as you discover or should have discovered the harm the statute begins even if the original act is far in the past. In a carcinogen case, usually the statute begins to run on the day of diagnosis--you have a year or two from then to figure out that some company or other may have caused the problem. It is the typical way of dealing with a statute of limitations in hard-to-discover wrongs.
My understanding of the Ledbetter case is that she may have found out about the discrimination more than 6 months before her retirement, thus she couldn't try to use a discovery rule type of workaround. But the record is not 100% clear on that because it wasn't argued.
dmbeaster: "The 180 day statutory rule here is actually extremely short in relation to other intentional wrongs. The classic example of this is fraud, in which the clock does not even start to tick until you knew or should have known of the misconduct. There is no delayed accrual for equal pay cases , but we have for a very long time (over a century) had a far less restrictive view of the proper application of statute of limitations than you advocate is proper."
No that isn't the case. I have no problem with the discovery rule. But that isn't the same as the Ledbetter rule at all. Under the Ledbetter rule you could have a discriminatory act 20 years ago, discovery 19 years ago and a lawsuit now. That isn't typical at all.
Posted by: Sebastian | July 13, 2008 at 12:01 AM
Under the Ledbetter rule you could have a discriminatory act 20 years ago, discovery 19 years ago and a lawsuit now. That isn't typical at all.
And this is where you are 100% wrong. The first discriminatory act was 20 years ago, but under prior law, the claim can be made now only as to the discriminatory acts that took place in the last 180 days. Ledbetter says you cannot even do that because the repetition of the discriminatory act is not a new wrong.
And this is typical for most every other type of claim involving continuing wrongs. If once a year I steal from you for 20 years, the past wrongs do not make the most recent episode OK. It is simply another wrong, but the statute cuts off the right to recover for older wrongs.
The problem is pretending that discriminatory pay is only one wrong committed once when the first act of underpayment occurs. That is the perverse holding in Ledbetter. There is nothing in the statute or policy concerning the law that supports that logic.
Ledbetter's claim was only for the last two years of her employment in 1997 and 1998. In that time period, based on the ongoing discrimination in pay, she was earning $3,727 per month. Men similarly situated (many with less experience) were paid from $4,286 per month to $5,236 per month. The difference in pay was based on sex discrimination.
Under your Ledbetter rule, that discriminatory pay can continue forever without recourse becuase the subsequent unequal payments are allegedly not acts of discrimination. The decision is a perversion of logic and justice that was not required by any statutory language, but a policy choice based on what?
Posted by: dmbeaster | July 13, 2008 at 12:28 AM
It's worth noting, by the way, that the 15% (approximately) gap between Ledbetter's pay and that of her least-paid male colleague is exactly the sort of thing that would show up in any very simple review of compensation rates. A difference of 3% might be hard to spot (though I think not, with good basic graphing), but 15% is really noticeable. Ignorance is an excuse only insofar as anybody refused to look.
Posted by: Bruce Baugh | July 13, 2008 at 02:02 AM
"Under the Ledbetter rule you could have a discriminatory act 20 years ago, discovery 19 years ago and a lawsuit now. That isn't typical at all.
And this is where you are 100% wrong. The first discriminatory act was 20 years ago, but under prior law, the claim can be made now only as to the discriminatory acts that took place in the last 180 days. Ledbetter says you cannot even do that because the repetition of the discriminatory act is not a new wrong."
The punitives however.......
Posted by: Sebastian | July 13, 2008 at 03:16 AM
For large companies are statistical tests that can be done as well to detect discrimination. In theory the average salary of employees (doing the same job) of different races or genders should be the same. Using statistical tests one can determine whether there is a real difference in those salaries that can be explained by gender or race.
If any CEOs want to hire me to do this for them on a regular basis, e-mail me. ;-)
Seriously, anyone who remembers Stats 101 could do this, there is no reason for large employers to not do it every six months.
Posted by: Shinobi | July 14, 2008 at 10:47 AM
"In theory the average salary of employees (doing the same job) of different races or genders should be the same. Using statistical tests one can determine whether there is a real difference in those salaries that can be explained by gender or race."
This would end up being a complete misuse of statistical analysis in all but the very largest companies or for the very lowest level of jobs. Even if you have 5 employees (doing the 'same' job--which for lots of the type of jobs we're talking about would be more than expected) the expected average variation over such a small sample is likely to be large enough to swamp the data. This is especially true if there is only one woman in the five, or one man (which even without any discrimination at all would be expected in 1/16 of all cases and there are a lot more than 16 companies and 16 job functions in the United States).
Posted by: Sebastian | July 14, 2008 at 11:17 AM
For any company that wasn't both very large AND unless the job in question was very common, this isn't true. The greatest majority of companies won't have even so many as five people with the exact 'same job'. At a sample of five instances, you are going to have serious trouble getting useful statistical information of the kind we want--with any variation at all you are likely to have at least one and probably two people noticeably far from the average, and your expected variation is likely to be very high. This is especially true if the sample has exactly one woman or exactly one man (something that will take place 1/16 of the time).
This kind of analysis might work ok for a call center where your 'N' is going to be in the hundreds and the job output is easily measurable. For almost any other job with a limited 'N' under 10, not so much.
Even for very low level jobs (entry level secretaries, mail room clerks, etc.) you have to be in a very large company to have 10 statistically interchangable jobs to analyze (or even 5). This is even more true as you go up the ladder.
Posted by: Sebastian | July 14, 2008 at 11:28 AM
Hooray, Typepad tricked me. Lol.
Posted by: Sebastian | July 14, 2008 at 11:28 AM
Even if you have 5 employees (doing the 'same' job--which for lots of the type of jobs we're talking about would be more than expected) the expected average variation over such a small sample is likely to be large enough to swamp the data.
Shinobi specified large employers, Sebastian. Is there some reason you didn't notice this?
This kind of analysis might work ok for a call center where your 'N' is going to be in the hundreds and the job output is easily measurable. For almost any other job with a limited 'N' under 10, not so much.
Huh. I guess I must just have imagined working for Digital, Compaq and IBM, then. Because I imagined that those companies I imagined I worked for had technical authors, software engineers, and software testers, by the hundreds. Nor did I imagine that those three companies were singularly isolated in having such large numbers of employees.
And for all your claims that "output" is what is used to judge, Sebastian, the point is: if en masse a company discovers that female employees in the same role as male employees are invariably being paid less, the company can then take a look at the salaries that individual female employees are being paid and confirm that they don't have a case for discrimination. Or that they do, in which case, promptly up the discriminated-against employee's pay to a fairer rate.
Or, of course, not, and run the risk of someday being sued for 20 years of discriminatory pay.
What is your problem with this, Sebastian? Anyone would think that you felt it was wrong that companies who pay female employees unfairly should get sued for doing so - your protectiveness always seems to apply to the company, not the employees who got paid less for not being male.
Posted by: Jesurgislac | July 14, 2008 at 11:59 AM
Sebastian -
Good to see you on the front page again!
If you agree that a statute of limitations makes sense in the case of a 20 year old discriminatory act that took place if the person quit the job 19 years ago, it doesn't make sense to prosecute a discriminatory act that took place 20 years ago if the person did not quit the job.
I disagree with this. As others have also pointed out, the harm is ongoing -- Ledbetter continued to be paid less -- so it seems like the clock should keep running.
I write software for a living. Imagine this:
25 years ago, while employed at a major insurance company, I inserted some code into their claims payment software that would pay $10,000 a month into my Swiss bank account. Not a lot of money to them, so they'll probably never notice. A significant amount to me, however. Kind of a win-win.
They discover it today.
Can I be prosecuted? My action was committed 25 years ago. Ever since then I've just been watching the $$$ roll in.
Not my fault that it took them so long to notice, is it?
Thanks -
Posted by: russell | July 14, 2008 at 12:28 PM
Russell if you did it, told your wife the money going into the bank was a pension, died and then 20 years later the software company wanted both their money back and triple damages from any money she might have earned elsewhere, would that be fair?
I think not.
Thanks-
Posted by: Sebastian | July 14, 2008 at 12:41 PM
You worked for DEC, Jesurgislac? I have an aunt that worked for them, prior to the merger/Borgification. Very cool lady, but not all that likely you ever ran into her.
Posted by: Slartibartfast | July 14, 2008 at 12:53 PM
Actually, come to think of it, she's not my aunt, she's my great-aunt's daughter.
Posted by: Slartibartfast | July 14, 2008 at 12:56 PM
"Huh. I guess I must just have imagined working for Digital, Compaq and IBM, then. Because I imagined that those companies I imagined I worked for had technical authors, software engineers, and software testers, by the hundreds. Nor did I imagine that those three companies were singularly isolated in having such large numbers of employees."
First, I seriously doubt all of them were really the same positions. I strongly suspect you had senior engineers, junior software testers, specialty technical authors, non-specialty user help, etc. VERY few companies are going to have more than ten people in exactly the same position and VERY VERY few companies are going to have more than 20. The number of companies that will have 30 or 40 people in exactly the same position for positions much above call center guy isn't worth setting national policy over. 'Programmers' are not all the same. And I think you are very well aware of that.
Second, even among each class I seriously doubt they were all getting paid the same amount. Some were better than others. Some had more experience than others. Some didn't work well with others. For any group below about 15 or 20 these factors are going to statistically swamp gender or frankly any other single factor you are trying to control for unless the discrimination is outrageously huge. The idea that statistical analysis for the typical 'N' of 2-5 is going to solve the problem is an enormous misunderstanding of how you can use statistical analysis. This is ESPECIALLY true in any case where the discrimination is going on below a company-wide level. If you have a department leader who is discriminating, it is very unlikely that is going to show up on a pure statistical analysis.
It may show up through complaints or by paying attention to how the discriminator deals with people on a day to day basis, but it won't show up on statistical analysis for most cases. The 'N' for comparable jobs is just too small. I work in a medium-sized company right now and I don't think we could easily get to an 'N' of more than 3 for any job at all. You might be able to do it for companies that have 20 floors of a skyscraper (AIG maybe?) but for almost anything smaller than that, and even with them for jobs once you get a couple of years past entry-level, you just don't have enough comparable people to resolve that kind of a question on a statistical basis.
Now, if you get reports that some one is discriminating, you could use statistics to bolster the case further. But the idea that companies who aren't using such analysis on a regular basis to automatically uncover discrimination is to misunderstand the enormous false positive rate that would be generated by such a policy.
(See for example lots of anti-terrorism procedures around airports. The false positive rate is enormously high. So high that I have no confidence that anyone is adequately sorting the false positives from the real ones in any effective manner that is also fair to the people who are getting hit with the false positives.)
The same problem happens with # of people in a field of study. You can probably suggest discrimination if in an entire field of work women are massively underrepresented (though still you have to wonder about choice issues) but with a small department with 'N' of 5, the chance that only one will be a woman even if women are equally qualifed and there is no discrimination is still 1 in 32. And there are a lot more than 32 departments with 5 people in the US. So there will be a lot of non-discriminatory ones that have only 1 woman out of 5 people.
Posted by: Sebastian | July 14, 2008 at 01:01 PM
would that be fair?
No, but not because too much time had elapsed. It would be unfair because what you describe is not the normal penalty for the kind of embezzlement I described.
If I was still alive, I would fully expect to be prosecuted and sent to jail. If there are punishments and/or legal remedies that apply to folks who inherit ill-gotten gains, I'd expect them to apply to my widow, were I dead.
Likewise, I'd expect the punishments and/or remedies that apply to illegal compensation practices to apply as long as the harm from those practices was ongoing.
IANAL so my expectations may or may not reflect what the law actually states. I'm just offering a common-sense example of what seems reasonable and fair.
Thanks -
Posted by: russell | July 14, 2008 at 01:05 PM
"If I was still alive, I would fully expect to be prosecuted and sent to jail. If there are punishments and/or legal remedies that apply to folks who inherit ill-gotten gains, I'd expect them to apply to my widow, were I dead."
But not punitive damages against your wife, right? And in point of fact not the 20 year old money either.
What if you hacked the payroll computer 20 years and gave yourself a bigish bump in pay. You took no further illegal action as the computer system changed and you didn't have access. You would benefit from the bump on an ongoing basis for any percentage based raises you got over the next 19 years. Can you be put in jail for it if they discover it 20 years later? Almost certainly not. Would the increased pay you got last year because of the increased base from 20 years ago count as a new criminal act? You might find a hang-him high judge who would say yes, but typically the answer would be no.
Posted by: Sebastian | July 14, 2008 at 02:07 PM
I'm missing the analogy to my wife. Unless I'm mistaken, the organization that discriminated against Ledbetter is the same one that has been paying her. Am I mistaken in that?
Regarding the one-time bump vs an ongoing embezzlement, is Ledbetter's claim that she was discriminated against once only, or that she was subject to a pattern of discrimination over her career?
Thanks -
Posted by: russell | July 14, 2008 at 02:13 PM
I don't know what Ledbetter's specific claim is, I'm talking about how the proposed rule would operate. A one time discriminatory decision made by a long gone supervisor would still count (for full punitive damage purposes) under the proposed rule change no matter how far in the past it was.
I'm not sure what you're missing about your wife. She is a party to a social unit which works together for common gain. She didn't know what you did, but may have received a benefit from it. She may have to pay certain things back, but not if past the statute of limitations. She won't typcially go to jail for your actions.
The analogy isn't strict, I think that a corporation is more responsible for its employees than a wife is for a husband, but on the other hand they aren't IDENTICAL which is how it seems people above want to treat it.
A corporation can have an ongoing pattern of discrimination (with many people acting in a discriminatory fashion) and/or a supervisor can. The propositions are similar but not identical.
Ongoing processing of payroll is not the same as choosing to discriminate against someone. If I outsource the payroll to ADP are they liable for discrimination?
Posted by: Sebastian | July 14, 2008 at 02:51 PM
Ongoing processing of payroll is not the same as choosing to discriminate against someone.
I agree with this.
I should probably acquaint myself more thoroughly with the particulars of the Ledbetter case before saying much more here, but I would think that if she worked for the firm for any length of time, there was more than one instance of discrimination involved.
Virtually all corporations above the mom and pop level have regular review cycles, and virtually all maintain fairly detailed profiles of who gets paid what, for what. If Ledbetter was being paid less than her peers for any length of time, active discrimination was likely occuring throughout that period.
I agree that extending the statute of limitations is a better solution than relying on the "each paycheck is a new violation" principle. It more clearly ties the violation to the intentional act of discrimination.
Thanks -
Posted by: russell | July 14, 2008 at 08:40 PM
VERY few companies are going to have more than ten people in exactly the same position and VERY VERY few companies are going to have more than 20.
....I'm not quite sure what to say to this, Sebastian.
I mean, if you want to argue that most employers aren't the size of IBM or Compaq or, hell, Halliburton... well, I'd agree to that.
But the employers that are that size, exist. And the notion that it's just impossible to find many employees in the same position across the company... huh. You are completely wrong as far as my experience goes.
Ongoing processing of payroll is not the same as choosing to discriminate against someone. If I outsource the payroll to ADP are they liable for discrimination?
No. Silly question. Why do you even ask it? Oh, because it evades the issue?
Again, Sebastian - why does it appear you are more concerned with protecting the company against justified suits for discriminatory pay?
Posted by: Jesurgislac | July 14, 2008 at 09:09 PM
russel: Regarding the one-time bump vs an ongoing embezzlement, is Ledbetter's claim that she was discriminated against once only, or that she was subject to a pattern of discrimination over her career?
I don't know the particulars of Ledbetter's case, but if she was initially paid 20% less than her male colleagues, and thereafter raises and bonuses came as a percentage of salary, she would never ever achieve equality with male colleagues - she simply couldn't.
What Sebastian seems to think is that because the initial decision to pay her less was made 20 years ago, it's wrong to hold the company in any way responsible for the 20 years in which the company profited by having an employee whom they paid significantly less than they would have paid her if she'd been male.
In fact, Sebastian seems to think it's wholly, totally, completely unfair to hold the company in any way responsible - just because they had an employee whom they underpaid for 20 years because she was female.
Because the initial decision to discriminate was 20 years ago, and the company profited by that act of discrimination for 20 years, it would be unreasonable to hold the company criminally responsible for those 20 years profit, since the company can't be held to blame for a criminal decision made 20 years ago from which the company profited ever since. No indeed. Companies are never held collectively responsible for bad decisions from which the company profits. Bad decisions are always the responsibility of individuals who have always left the company by that time. The company, however, retains the profits of that bad decision, and it would unfair to make the company disgorge them.
I understand now.
Posted by: Jesurgislac | July 14, 2008 at 09:17 PM
Sebastian,
You seem to not know what punitive damages are. They are not the ordinary damages one gets from proving the violation, those are compensatory damages.These are designed to put you in the position you'd be if Punitive damages are on top of that, designed to punish the offender. They are only appropriate when the litigant can prove malice or reckless indifference on the part of the employer.
Posted by: RY | July 15, 2008 at 08:12 AM
"You seem to not know what punitive damages are. They are not the ordinary damages one gets from proving the violation, those are compensatory damages.These are designed to put you in the position you'd be if Punitive damages are on top of that, designed to punish the offender. They are only appropriate when the litigant can prove malice or reckless indifference on the part of the employer."
I understand that quite well. You might note that I brought it up. Punitive damages would be inappropriate for a 20 year old act of discrimination which is being bootstrapped past the statute of limitations by 'every paycheck a discriminatory act'. And I brought up that distinction repeatedly because at least two people upthread think that Ledbetter could only get back pay for six months even with the work-around rule, which is strictly true ONLY if you ignore punitive damages. Punitive damages allow you to bootstrap the punishment all the way to the 20 year-old act and ignore the monetary limitations of the 6 month rule. This may be a good or bad thing, however you want to argue about it. But it shows that the 6 month 'limitation' effectively doesn't do anything in such cases--and my point all along is that it would be better to have a reasonable statute of limitations that really operates under normal statute of limitations principles than it is to have a really short one that you work around which then creates an unlimited statute of limitations.
Again, the proposed rule would allow for punitive (punishment) damages for acts decades old if the person still worked for the company (but very strangely not if they quit sometime in the last 6 months because then they wouldn't have received a paycheck). The statute of limitations for bank robbery and almost any other very serious crime doesn't last that long, so it seems odd that you would do so for pay discrimination.
Posted by: Sebastian | July 15, 2008 at 11:55 AM
Well given the fact that one needs to prove more to get punitive damages, it seems much more appropriate to hold the employer liable than in cases of simple damages.
In any event, statutes of limitations exist in part to encourage individuals to file claims when the information is fresh.
But there is significant asymmetrical information in equal pay cases. A bank knows when it's been robbed, even if it doesn't by whom. The police should investigate at that time. But an individual does not necessarily know that she is being paid less than her peers of the different sex. Therefore the analogy isn't be to a bank robbery, but a physician who has left a foreign object in a patient after surgery. In those cases, the statute of limitations doesn't start to run until discovery. Such a rule seems much more appropriate than "sorry you should have stopped us from screwing you over then, so we get to keep doing it now."
Posted by: RY | July 15, 2008 at 05:04 PM
"In those cases, the statute of limitations doesn't start to run until discovery."
And I'm ok with the discovery rule too. But that still allows for a tight statute of limitations, unlike the proposed rule.
My problem with the proposed rule is that it throws out everything we normally do in all sorts of other cases. Statutes of Limitations exist for a reason and so does the discovery rule. Together they can easily handle Ledbetter-like cases without resorting to other unnecessary workarounds. (My guess is that Ledbetter herself actually knew about the discrimination long enough not to be saved by a discovery rule but that is conjecture).
Posted by: Sebastian | July 16, 2008 at 11:45 AM
Sebastian: Punitive damages would be inappropriate for a 20 year old act of discrimination which is being bootstrapped past the statute of limitations by 'every paycheck a discriminatory act'.
You do keep focussing very tightly on the idea that once a company has got away with paying a woman less than a man, for enough years, it's inappropriate to penalize the company for continuing to discriminate.
Of course every paycheck is a discriminatory act. Why do you have such trouble accepting this? On some level, Sebastian, do you think women deserve to be paid less than men?
Posted by: Jesurgislac | July 16, 2008 at 12:15 PM
Oh hooray, snide insinuations that I don't like women. Again. On every topic.
My thoughts on statutes of limitations apply in non-women situations too. I think they make sense for bank robberies (usually not women) and pretty much any situation where evidence is likely to get stale. For discrimination suits where the paper trail isn't crystal clear (i.e. cases other than: I hate women, will pay them less and will admit it in this memo) memories about relative performance, legitimate or illegtimate issues with supervisors, and all sorts of other things come to the fore which aren't going to become clearer by adding 19 years to everyone's memory of the events.
Posted by: Sebastian | July 16, 2008 at 12:33 PM
and pretty much any situation where evidence is likely to get stale.
So why are you bringing this up for this situation, where the evidence is plain and fresh and the paper trail is clear?
Hm?
Oh hooray, snide insinuations that I don't like women. Again. On every topic.
Only when you seem to be arguing that women deserve less pay or fewer human rights than men do, Sebastian.
Posted by: Jesurgislac | July 16, 2008 at 01:05 PM
"So why are you bringing this up for this situation, where the evidence is plain and fresh and the paper trail is clear?
Hm?"
Because the proposed rule generalizes far beyond THIS CASE.
"Only when you seem to be arguing that women deserve less pay or fewer human rights than men do, Sebastian."
Please quote the portion that you believe is me arguing that women deserve less pay."
Until then, please refrain from saying such lies.
Thanks.
Posted by: Sebastian | July 16, 2008 at 01:15 PM
To make a more appropriate analogy than the embezzlement case that Russell brought up, what if company X was a contractor to some company Y and during 20 years of business company X had padded their invoices for services/goods that were never received.
The CEO that originally was the head of company X at the time the first occurrance of the padded invoices dies and a new CEO takes over. Company Y finally discovers this fraud.
Is company X still liable for the 20 years of fraud? Are they only liable for the last padded invoice? Or do they get away with it completely because the statute of limitations for fraud ran out for the initial act of fraud 20 years ago ?
It seems you are arguing the latter with respect to the Ledbetter case, yet I am reasonably certain that company X (in my example) would be liable for the entire twenty years of fraudulent invoices.
Posted by: Dr. Morpheus | July 16, 2008 at 06:04 PM
"It seems you are arguing the latter with respect to the Ledbetter case, yet I am reasonably certain that company X (in my example) would be liable for the entire twenty years of fraudulent invoices."
No I'm almost completely certain that it would only be liable for invoices done within the statute of limitations unless the discovery rule saved it, but I'm almost certain that it wouldn't in this kind of case (tough to be completely certain because the rules can be very different in different jurisdictions). The closest universal policy (Uniform Commerical Code) on a similar topic (check fraud) is fairly strict about the statute of limitations (3 years from the date of negotiation of the check).
Posted by: Sebastian | July 17, 2008 at 12:06 PM