Comparing job losses in recessions and a depression

  • Share
  • Read Later

I haven’t done this in a few months:
The chart shows payroll job losses from the employment peak of November 2007 through this November, expressed as a percentage of the working-age population. Sebastian Dartevelle suggested measuring the job losses this way (he called it “labor-able” population, which is why it says that on the chart), because if you simply measure the percentage decline in payroll employment from the start of the downturn (as I do in the chart of post-World War II recessions below), you overstate the severity of Depression-era job losses because a much larger percentage of the workforce was on the farm then. Those farmers don’t show up in payroll employment data, and presumably they were less likely to lose their jobs in the Depression than factory workers were. Then again, many people were thrown off their farms, so measuring things this way probably understates the severity of the Depression-era job losses. But as a rough approximation it’s at least a start.

Plukasiak suggested measuring from the month of the market panic that started things off (October 1929 and October 2008) to get a better comparison. Another advantage of doing it this way is that the numbers aren’t seasonally adjusted, so if both charts are covering the same times of the year they’re more comparable. Here’s the result:

So the severity of job losses is comparable. Until the past few months, that is, when the Great Recession started to ease. The Great Depression just kept getting worse. How much worse? Here, at the suggestion of Mrs. Curious Capitalist, is a chart showing the next 11 months of the Depression:

It actually got even worse after that, with job losses eventually maxing out at more than 10% of the working-age population. It looks like we’re on a trajectory to avoid that kind of disaster. But no guarantees. What’s been clear for a while is that this recession is far more severe than the five that preceded it. This time I decided to throw in the sharp but brief 1957-58 recession too. This one’s worse than that (these numbers are seasonally adjusted, by the way).

Update: I actually didn’t mean to post this Saturday. I thought I was just saving a draft that I would revise and then post on Monday but, uh, I guess I hit “publish.” I’ve now added a couple of links, but I guess I’ll leave it alone otherwise. I was thinking of prettying up the charts, but that’ll have to wait.

Also, because this question often comes up: None of these charts have anything to do with the unemployment rate, so the various debates over how that is measured aren’t relevant here. I’m simply using data from from the monthly establishment survey of non-farm employers to compare the pace of job losses in different economic downturns. These numbers are far from perfect—and all they tell you is the change in non-farm payroll employment, not how many people are out of work or how hard it is to find a job. But … they are what they are.