October 1, 2014

How the Unemployment Rate Declined in January

The unemployment rate declined 0.4% in January, bringing the rate to 9.0%, according to the latest statistics [.pdf] from the Labor Department. The unemployment rate similarly declined 0.4% in December. This 0.8% drop in the last two months is among the largest declines in the nation’s history. Since unemployment record keeping began in 1948, the unemployment rate has dropped 0.8% in two months on only one other occasion, which was in 1958. In other words, the January & December decline is the greatest in over fifty years. Despite the pleasing face value of this fact, payrolls in January only added 36,000 workers. This relatively small amount of new workers, which was the fewest in four months, could not have caused the unemployment rate to drop 0.4%, but three statistical explanations clarify how this drop occurred.

First, the Department of Labor adjusts the population of the country in January of each year. Population adjustments in January resulted with a reduction of 457,000 individuals [.pdf]. To be clear, this reduction is not from the 2010 Census data, but is instead the last stage of the incremental implementation of the population shift from the 2000 Census.

The Labor Department utilizes an incremental adjustment of population data spread over ten years because a sudden shift of all the data at once would significantly distort the historical record of unemployment data. In light of this statistical method, the unemployment results during January of each year is problematic for data analysis. With a population reduction of 457,000 individuals in 2011, this contributed to the 0.4% decline in the unemployment rate.

A second methodological contribution to the January decline is a result of a growing number of individuals who are marginally attached to the work force. Marginal attachment refers to individuals who are not employed and also have not actively looked for a job in the past 4 weeks. Most importantly, the Labor Department excludes individuals among the marginally attached from the unemployment rate. In January alone, there was a growth of 191,000 individuals to those in the marginally attached. As a result of the January growth in the marginally attached, this also contributed to the 0.4% decline in the unemployment rate.

This brings the total number of the marginally attached to 2.8 million – a record high. In January 2010, there were 2.5 million among the marginally attached, reflecting a rise of 300,000 over the past 12 months.

A third methodological contribution to the January decline is a result seasonal adjustments. Each year, the Labor Department utilizes what it refers to as “benchmark revisions,” which are revisions of the labor force based on tax records to control for errors in past estimates. The January report revised the employment level for every month in 2010, with eight of those months being reductions.

For instance, the original reading for January 2010 was a gain of 14,000 jobs, but the benchmark revision changed this to a decline of 39,000 jobs. Altogether, the benchmark revision resulted with a reduction of 215,000 workers from 2010 job gains. This means there was not as much job growth in 2010 as the Labor Department originally reported. Despite this shift in employment numbers, this data as a whole does not contribute to the unemployment rate because the Labor Department measures the unemployment rate through a monthly survey of 60,000 households, known as the Current Population Survey, which excludes payroll revisions.

A separate survey measures the number of individuals on payrolls, which is known as the Current Employment Statistics Survey. While the Labor Department reports the results of these surveys at the same time each month, the CPS survey is the sole measurement for the unemployment rate. Still, even though the benchmark revision does not change the unemployment rate, the revision does change the calculation for seasonal adjustments in the unemployment rate.

Seasonal adjustments involve the comparison of a month to the patterns of that month in the past. Since January 2010 results were negative, as mentioned above, whereas January 2010 results were positive, this positively affects the seasonal adjustment for January 2011 and contributes to the 0.4% decline.

In the end, the unemployment rate declined 0.4% in January mostly because of the statistical methodology of the Labor Department, not because of a surge in job growth. Regardless of any methodological adjustments, the unemployment rate has now been 9% or higher for 21 consecutive months – the worst streak since the Great Depression.

Speak Your Mind