Bottom Watch

Last week, James Hamilton noted that historically, a peak in initial unemployment claims has been closely followed -- like within a matter of weeks -- by an end to output contraction. He mentioned this datum based on Robert Gordon's observation that claims had flattened in recent weeks (albeit at a really high level), and ticked downward slightly during the first week of April. But Calculated Risk cautioned that it might be a blip, quoting Goldman Sachs economist Seamus Smyth:

Roughly speaking, a 20,000 decline in the 4-week moving average corresponds to a 50% probability that the peak has already been reached, and a 40,000 improvement to a 90% probability.

The longer trend needs to come down before we can feel good, in other words. So what does today's release betoken? Claims were expected to rise but instead fell by 53,000. Even taking into consideration that revisions to the previous week's data basically eliminated the tiny improvement that originally interested Hamilton, that's a pretty substantial drop. The four-week moving average is down, too, by 8,500. That doesn't mean we're on the other side of the trough, but it's clearly good news.

The rub is that continuing claims hit another record, above 6 million. And that indicates how long and ugly recovery will be. Long after claims have begun falling, we'll still be bleeding jobs. And long after the economy has begun adding, rather than losing, jobs, total employment will be below the long-term trend. The ship will right, but it will take a while.