Revisiting

In March I posted this, about employment discrimination lawsuits in the U.S. from 1978 to 2006. I received a number of good possible answers from people in the "real" world, so I s'pose it's time to figure out some kind of answer to the original problem. Check the Stolen Graphs post for a refresher...

Here's an email I received from a certain "Juliana," whose last name I will omit in order to prevent internet people finding out she associates with shady Conglomerates like this one. First check out the graph from the last post:

And here's what she wrote:

First, between 1991-1998, the economy grew tremendously, no? I like the quote I found on Wikipedia from David Greenberg, a Professor of Media Studies and History from Rutgers:

"The Clinton years were unquestionably a time of progress, especially on the economy [...] Clinton's 1992 slogan, 'Putting people first,' and his stress on 'the economy, stupid,' pitched an optimistic if still gritty populism at a middle class that had suffered under Ronald Reagan and George H.W. Bush. [...] By the end of the Clinton presidency, the numbers were uniformly impressive. Besides the record-high surpluses and the record-low poverty rates, the economy could boast the longest economic expansion in history; the lowest unemployment since the early 1970s; and the lowest poverty rates for single mothers, black Americans, and the aged."

So, it follows that on the whole people probably felt more secure in their jobs during 1991-1998 than they had during the Reagan/Bush years. Since they didn't think they'd be fired (note, single mothers, black Americans, and old people were more secure than ever), they may have been more inclined to air their grievances.

Also, they had relatively more support in terms of social policy programs. For example, a Clinton budget package included the Earned Income Tax Credit. Although it is illegal, poor and minority Americans likely had the perception that their employer would fire them indiscriminately if they complained, and that they (the poor people) would not have the resources to fight back.

Also, during a strong economy, most of these folks probably had more savings. That way, if they lost their jobs for a few months for complaining about discrimination, they were more likely to still be able cover basic expenses like heat and feeding their kids. Suddenly, they could stop worrying about survival and think about the fact that they made less than the white guy in the office next door.

Add to this a media climate that encouraged litigation and political correctness. We've also left out two important historical events that happened in 1991-1992: Rodney King and the Anita Hill hearings. These can't be left out of the picture because they altered the discourse around race and gender discrimination in the United States.

As far as the downturn in the graph, I credit that to the economic bubble bursting in 2001-2002 (tech bust, Bush tax cut for the wealthy, etc.). Suddenly, people were more worried about job security and feeding their kids again. General inequities (women making 70 cents on the dollar, for example) still mattered, but not as much on a personal level.

Very well said. So let me summarize her ideas, if I may, before I comment:

1. Higher job security meant people were "more inclined to air their grievances" with less fear.
2. The Clinton Administration provided "More support in terms of social policy programs," providing, at the very least, a psychological safety net.
3. More savings (stemming from economic growth) should have the same safety net effects.
4. A "media climate that encouraged litigation and political correctness."
5. Finally, economic stagnation helps explain the reversal in the graph.

All very good, very plausible theories.

And here's an email I received from a certain "Dave," (not my alter ego) whose last name I will omit because I didn't ask if he minded if I used his email for blog fodder (...I assume he doesn't...):

I think the graph of employment discrimination cases follows the graph of all civil cases pretty closely until 1993--the start of the Clinton administration, when workplace issues and labor discrimination became part of the public dialog. It rises disproportionate to the overall graph until 2002, when people had more important things to worry about.

Basically, it was more socially and politically acceptable to sue over workplace discrimination during the Clinton's America than ever before. 9/11 changed the social, political, and legal agenda of America so that this was no longer the case.

Both of them came to somewhat similar conclusions, I think. Of course, The Other-Dave deserves credit for writing me almost a month ago with his ideas as well as for helping me form some more thorough answers in the "real world." I didn't think I'd be revisiting this post until Juliana spoke up this week, though, so I figured I'd point this out... Also - Thanks to both of you for your thoughts.

Anyhoo...

Here's my commentary:

First off, I agree with everything that both of these folks have said. There is no reason to doubt that increased public attention, the effects of economic growth, certain policies enacted by the Clinton Administration and the media's effect on culture (e.g. political correctness) would bring to light discrimination by U.S. employers.

There are, however, a few problems:

1. The beginning of the rapid growth in employment discrimination suits in the graph was around the first part of 1991. In fact, from 1991 to the beginning of 1993 (before Clinton even took office), lawsuits rose from about 8,000 to about 12,000, a 50% rise. That's pretty significant, and can't be explained by Clinton-related policies. It can, however, still be explained in part by public attention, media focus, expectations, et al.

2. These theories operate under the assumption that employer discrimination levels remained unchanged over the years. Instead of people reporting it more, it's possible that employers were actually discriminating more. Is it not likely that the rapid growth of discrimination cases from 1991 to 1998 is due in large part to the growth of business, new jobs and discrimination rather than just the growth of discrimination reporting? Take a look at total job growth from 1990 to 2007:

With such dramatic job growth, you could expect the sheer number of employment discrimination suits to rise dramatically as well, right?


3. More relevant, I think, is the type of new jobs that were being added to the economy. Technological growth and innovation was the major driver of economic growth in the '90s, and businesses in this industry were young, inexperienced and, perhaps, not focused much on proper hiring/firing practices. It's even quite possible that it took four or five years for these businesses to train and/or hire the right kind of human resources staff to prevent such discrimination. (Thanks to Other-Dave, again, for helping me formulate this idea.) To give you an idea of tech jobs growth by itself, here's some data from the BLS on the Computer Services sector:
Notice that the beginning of the curve is around the start of 1991, just as in the discrimination graph.

4. And the statistical problem: The data set is very small. If we had data from the '50s, '60s and '70s, perhaps we might discover that dramatic changes in the number of employment discrimination suits and a low correlation with civil suits are not unique occurrences. It's possible that all periods of economic growth see such dramatic rises, and we just happen to have a graph that exaggerates the effect in the '90s. Maybe.

5. But the biggest problem of all, in my opinion, is the failure of most of these theories to explain the latter part of the graph: The decrease in lawsuits from 1998, the temporary increase in 2001-2003 (during a recessionary period, no less!), and then the big drop starting in 2004.

Even an economic/business growth theory, by itself, wouldn't account for the 2000-2006 period. Here's a GDP graph to help show why:

(We could also just use the job growth graph from above.)
If discrimination suits are directly correlated to GDP growth, we'd see the decrease starting in the third quarter of 2000. In theory. Also, from Q3 2000 to the following year we'd see a consistent decrease, but that happens to be the one time in the period from 1999-2006 that discrimination cases are increasing in the graph (figured I'd post it again):

So, if I had to come up with a theory that used the fewest numbers of most-likely explanations, it would be this:

The growth of tech-sector jobs from 1990 to 2000 was the primary contributor to discrimination cases during that period, mostly due to the fact that they were new, inexperienced, operating in a non-traditional sector, and - most importantly - they sprouted up faster than the supply of trained hiring mangers/lawyers/human resources staff. Growth of these companies didn't slow in 1998, but the supply of trained human resources employees started to overtake tech-company growth around that time. The longer-term decline in discrimination lawsuits followed, as tech-sector growth was finally in line with human resources growth. When tech-sector growth started lagging behind other sectors, the number of lawsuits started to regress even more rapidly to the mean. The one blip in the graph - the small upward trend from the latter half of 2000 to the end of 2002 - can be explained by substantial job losses nationwide, from which there are bound to be a slightly increased number of wrongful firing/discrimination suits.

---

But, in reality, I think the answer might have a lot more to do with specific legal policies. As I am both lazy and obsessed with economics, though, my theory has everything to do with economics and nothing to do with things that involve reading law history.

This is a pretty good start, I think. If you combine the theories from Juliana, Other-Dave, Gretchen and myself - provided there were no specific and overly obvious legal policies that correspond precisely with the graph - then we've probably got an answer as to how to explain the employment discrimination graph.

Right?

Thank you guys, a ton, for playing along. Let me know if I've missed anything...