No leading indicator is perfect, but initial claims for unemployment insurance is fairly good. Here's the latest chart (data available here):
It looks like the worst is behind us, but that the improvement has been very, very mild. In other words, consistent with much of the other data coming out. So I expect things to get better, but at a very gradual pace.
The rest of this post is about data issues, which most readers will want to skip.
Why is initial claims the better leading indicator than total claims? First, historical evidence. You look back for several decades and you can clearly see that this series tends to lead the contemporaneous measures of the economy.
Second, initial claims is a more consistent series than total claims. Total claims is affected by changes in rules regarding the total duration of benefit eligibility. In most states, 26 weeks is the maximum period of time that you can collect UI. But there are automatic triggers that can extend the maximum to 39 weeks. (The triggers involve a high and rising unemployment rate.) In addition to these automatic triggers, Congress enacts a special extension of benefits most recessions. Why does Congress do this if there is already an automatic trigger? Because it enable Congress to say that they've done something. The bottom line is that total claims is based on less consistent rules than initial claims.
However, anyone comparing today's level of initial claims with historic data should be aware of two issues. One is that we have a larger labor force today than in past cycels. Mark Perry over at Carpe Diem blog has adjusted the data; here's his latest look.
The second issue is that rules for UI eligibility have broadened over time, and especially this year. It's complicated arithmetic dealing with this issue: to be eligible, you have to have earned a certain amount in the past year. The old way of calculating this involved a long time lag, due to paper records taking a long time to evaluate. The new way involves a much shorter time lag, which is enabled by electronic technology. However, it increases the number of people eligible for UI, particularly among those with only six to 12 months of work experience. As a result, initial claims are today higher than it would have been if the old rules had remained in force.