Each Thursday morning, economic commentators get excited about new data for initial unemployment claims. Today, for example, we learned that number was 450,000 -- the lowest number since early July. But what does this number really mean? And how does it compare to the monthly unemployment numbers that accompany changes in the national unemployment rate? There's a lot of confusion out there, so let's clear it up.

Unemployment Insurance Weekly Claims

First, let's talk about the weekly number. It is compiled by the Department of Labor. Each week, it gets in statistics on how many Americans filed for unemployment insurance. That provides these numbers. The DOL counts the number of Americans that actually request unemployment assistance from the government.

So the first statistic is called "initial claims." But this term is misleading. This isn't necessary the first time an American has filed for unemployment insurance. There are a number of reasons why a claim might be considered "initial," even if it isn't in the strict sense of the word. For example, if someone stopped getting insurance for a time, when they file again, it may be considered an initial claim. So really, this number isn't terribly useful, because lots of variables can effect whether a claim is considered initial or not. As mentioned, during the week ending September 11, 2010, 450,000 Americans filed initial unemployment insurance claims.

The next term is "continuing" claims. Again, here someone requests unemployment insurance, but in such a way that the DOL's definition of "initial" is not met. So those Americans whose filings are not considered initial and have been in the program for 26 weeks or fewer are grouped under continuing claims. This number is a little bit better of a gauge of labor market health, but is still incomplete. During the week ending September 4, 2010, 4,485,000 filed continuing unemployment insurance claims. This includes initial claims.

Finally, some Americans are grouped under the "Emergency Unemployment Compensation" (EUC) heading. Congress has extended the time period for which people can file for unemployment insurance past the usual 26-week period. Once you exhaust the first 26-week phase of the program, you are considered to be in the emergency portion -- until after 99 weeks, which is the current limit set by Congress. During the week ending August 28th, 4,108,529 people filed under the EUC program.

There are a few things to note about claims. First, benefits aren't always awarded. This is just a measure of how many people try to get money for being unemployed. Second, people who don't qualify for unemployment insurance -- like if they quit their job instead of getting laid off -- often aren't included, though unemployed. Third, some people have their benefit period run out. Congress didn't extend eligibility forever, so some Americans can't get anymore jobless benefits if they have been out of work for a very long time. By now, it should be clear that this data provides a pretty incomplete picture of the labor market.

Unemployed Persons

The other measure is provided monthly. It is compiled by the Bureau of Labor Statistics. Each month, BLS surveys households and asks them questions about their employment status. From its sample, it determines how the labor market has changed during the month.

There are two headline numbers that reporters love with the BLS report. The first is the number of jobs added or lost. This is an estimate, based on BLS's survey, of the net number of workers to added or subtracted from the labor market each month. So this doesn't provide information about how many gross jobs firms added or eliminated during the month -- it just provides the net change.

The second number is the unemployment rate. This is essentially the number of unemployed persons as defined by BLS who were unemployed during the month divided by the total labor force size. There's some play there, however, as an unemployed person who is discouraged about finding work is not considered a part of the workforce and left out of the calculation, even though other people might consider him or her unemployed. The measure has other shortcomings as well, like its failure to differentiate between part-time and full-time work. But other BLS estimates can help to fill in those gaps.

Comparing the Two

Each of these statistical methods will provide a number of unemployed Americans. For example, if you add continuing claims and EUC, you find that 8.2 million people filed for unemployment insurance during the week ending August 28th. But the Bureau of Labor Statistics report for August says 14.9 million Americans were unemployed. As you can see, these are very different measures.

This shouldn't be surprising. You will always find the total number of unemployed Americans reported by BLS is higher than those claiming benefits. As described above, not everyone who is unemployed qualifies for benefits. So while it might be fun and/or interesting to see how unemployment claims change each week, they shouldn't be taken as a broad indicator of labor market health. In fact, as unemployment duration becomes more severe, and people are forced to drop out of the program, even claims numbers that include the EUC population will become less and less useful in determining just how many Americans are jobless.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.