The ubiquitous mantra surrounding business intelligence is "the right information, at the right time, in the right format".

There is nothing wrong with this phrasing. Information put in the right hands and in an immediately usable format can be very empowering. One of the finer points is around "right information" — how do we know what is right?

An example of this issue (and to pick on something in the news a lot right now), are the U.S. unemployment figures batted about in the news. Are we better than the Great Depression era? Where does the current economy sit in relation to the Great Depression? Where are we in relation to other recessions or depressions?

On the surface when we attempt to answer this, one can say that the U.S. employment metric of 8.1% from February 2009 is still far below the 1933 peak of 24.9%. Even removing seasonal adjustments on the current number puts the "official" unemployment metric at 8.9%. Does this mean we are better off? What relief can one take from this measure?

As with any measure, it is important to know how it is calculated – this becomes even more important when trying to compare across time period, and not just one quarter from the next, but in this instance, decades. Here are a few quick facts to keep in mind when looking at and trying to compare these unemployment measures:

  • Before 1940, unemployment statistics were gathered in a variety of decentralized means, with various ways to measure employed versus unemployed
  • The Works Progress Administration owned the survey between 1940 and 1942, before transferring to the Census Bureau
  • In 1959, the Bureau of Labor Stats took over analyzing and publishing the data, but the Census Bureau still collects it
  • In 1994, the CPS (Current population survey) chose a different measure for the "official" unemployment metric. Article here

What do these various facts tell us?

Our current unemployment measures are broken into several categories, U1-U6. U3 is the current official metric that is quoted in the news and plastered across news papers. U3 is the unemployment measure mentioned above, currently sitting at 8.1% for end of February 2009. Historical economists however, recognize that the surveys that were done during the 1930s are closer to our current U6 unemployment measure. Given this, can we really compare our current official metric to the official metric of the Great Depression era? To be accurate, no.

Compounding the issue in looking at even more recent recessions in the 1970s and 1980s is the fourth bullet point: the BLS switched officially reported unemployment metrics from 1993 to the 1994 year. Pre-1994, a modified version of U5 was the official measure, while post 1993 and forward, the BLS has switched to U3.

Luckily for the savvy numbers enthusiast, the Bureau of Labor Stats still published all six unemployment measures. So how do we really compare? Still a tricky answer; the accuracy of the pre-1940 surveys are still called into question, however, if one works off of the assumption that our current U6 unemployment measure is the closest, then we are at 14.8%!

Seasonally adjusted of course.

References: