Unemployment Data

I've come to appreciate one aspect of our huge government: there is a lot of data which is made available to the public. Sometimes it is difficult to find, but if you know where to look, you can usually find what you want. I trust most of the data on face value, but some must be taken with a grain of salt - especially projections.

I've been casually watching the weekly reports on initial claims for unemployment. However, each week they (the Department of Labor) adjust the number which was reported the previous week. Without actually studying the data, it had seemed to me that the numbers were systematically under-reported in the initial report. I decided to look at the reports to see if my hypothesis was valid.

This first graph below shows the reported volumes versus the adjusted volumes for initial claims. As a matter of process, each Friday, the DOL releases the report at 8:30 AM providing the initial report of claims for the previous week. Additionally, they provide an adjustment to the previous week's report.

Source: Department of Labor

As you can see, the adjusted volume tend to always be higher than the reported volume. Since Obama's inauguration, the adjusted volume has been higher sixteen of eighteen times. The difference has usually been about 4,000 claims. This could be random, due to bias, or a tool of propaganda. I can't really prove this; however, based on what I see in the data, I would chalk it up to randomness with only a slight bit of bias.

When the report comes out, a key measure is the change from week to week. This is seen as an indicator of whether or not the economy is (in our current condition) getting worse at a slower or faster pace. In this comparison, the newly reported claims are usually compared to the previous week's adjusted claims. This is suspect to me unless the adjustment is truly random (no bias or propaganda). However, regardless of the purity of randomness, the difference does not seem to be too large. See this next graph.

Source: Department of Labor

In summary, I think my suspicions are generally unfounded. Most of the recent adjustments have indeed exhibited a potential bias. However, it has had little impact. If this bias continues, then there might be something wrong.