So now that I'm at Brookings this blog (for the forseeable future) will just point to my research on the Brookings site.  In that regard, I'm happy to report that I've put out my first piece at Brookings--a brief post on our group blog that finds a longer-term problem with labor markets predating the recession.  Enjoy, and thanks for your interest!
I keep seeing that chart that shows how employment declines in the current recession are so much worse than in past ones.  You know, this one:
On many dimensions, of course, the current recession is much worse, but this chart has always seemed funny to me.  And after reading Paul Krugman mock the idea that the recessions of the 1970s and 1980s were at all comparable, I decided to make my own damn chart.  Because the above chart looks at employment levels, which are affected by labor force growth, I decided to look at employment rates instead (subtracting the unemployment rate for each month from 100).  Because the composition of the labor force has also changed over time (lots more married women, most notably), I decided to confine to white men ages 20 and up.  And because it's unclear to me what "peak" is used in this chart (see the vague note at the bottom of Rampell's chart) and since the relationship of the NBER business cycle peak to the unemployment rate involves a lag, I decided to measure from the peak employment level.  Got all that?  Here's my chart:
I've labeled the lines the same way that Rampell's chart is labeled, by the recessions that followed each employment rate peak.  The figures are from BLS and are based on their seasonally adjusted series.

This approach makes clear why people were disappointed by the "jobless" recoveries from the recessions of the early 1990s and 2000s, which were no faster than after the much more severe recession of the early 1970s (though of course, the declines in employment were much smaller to begin with).  More to the point, it also shows that while the current recession still looks bad, bad, bad, the decline in employment is comparable to the decline during the double-dip recession, which is apparent from the "1980" line.  That's not the most fantastic news of course, but it's worth noting.  Unfortunately, I doubt this is the chart you'll see others use and update as things evolve in the next few months.

(cross-posted at and

When it comes to economic conditions, I'm generally a glass-three-quarters-full kind of guy.  Take unemployment.  Quick—what was the risk in 2008 that an American worker would experience at least one bout of unemployment?  Chances are you thought that that risk was higher than one in eight.*  But figures from government surveys indeed suggest that thirteen out of fifteen workers (or would-be workers) had not a single day unemployed during the first year of the "Great Recession".** (Incidentally, the recessions of the mid-1970s and the early 1980s were also called the "Great Recession" by some commentators.)

The 2009 data won't be out until later in the year, but if last year ends up comparable to the depths of the early 1980s recession, then the average worker will "only" have had a seven in nine chance of avoiding unemployment.***  But these figures overstate economic risk because some unemployment is voluntary and much of it is brief.  According to the Congressional Budget Office, the chance that a worker experienced an unemployment spell lasting more than two weeks during the three years from 2001 to 2003 was just one in thirteen—a period covering the last recession.

So as I've been following the debate about unemployment insurance and whether it actually worsens the unemployment rate, I've actually been open to the idea that being able to receive benefits for up to two years might create perverse incentives.  The research is not as uniformly dismissive of the idea as some liberal assessments have implied (go to NBER's website and search the working papers for "unemployment" if you want to check this out yourself).

In particular, the idea that there were 5 people looking for work for every job opening struck me as sounding overly alarmist.  So I started looking into the numbers to determine whether I thought they were reliable.  The figures folks are  using rely on a survey from the Bureau of Labor Statistics called the Job Openings and Labor Turnover Survey, which unfortunately only goes back to December of 2000.  But the Conference Board has put out estimates of the number of help wanted ads since the 1950s.  Through mid-2005, the estimates were based on print ads, as far as I can tell, but the Conference Board then switched to monitoring online ads.  You can find the monthly figures for print ads here and the ones for online ads here.  The JOLT and unemployment figures are relatively easy to find at BLS's website.

When I graphed the two Conference Board series (which requires some indexing to make them consistent--the print ad series being an index pegged to 1987 while the online series gives the actual number of ads) against the number of unemployed, and then the JOLT series against the unemployed, here's what I found:

I'll just say I was shocked and that I am much more sympathetic to extension of unemployment insurance than I was yesterday.

*The post originally said one in ten, which was wrong (the result of mistakenly using a figure I had computed  for an older age range).  Technically, the the figure was 13.2%, or 1 in 7.6.
** The original post said nine out of ten.
*** The original post said that if it reaches the depths of the 1990s recession, then the average worker will have had a five in six chance of unemployment.  I located data for the early 1980s recession, which is a better comparison to the current one.