Over the course of a year, the size of the labor force, employment levels, unemployment rates, and other measures of labor market activity fluctuate according to known seasonal patterns. Seasonal events such as changes in the weather, major holidays, and school schedules influence labor force availability and demand for services. In a state like Montana, which is known for its long, cold winters, the construction industry is highly seasonal and exerts a huge influence on the state’s employment numbers.
When the data points are graphed, the result can look something like this:
While an overall trend can be seen over time, it is difficult to tell how much of the month-to-month change represents an underlying trend, and how much is due to seasonal factors.
This is where seasonal adjustment comes in.
Seasonal adjustment is a statistical technique that removes the influences of predictable seasonal patterns to reveal how employment numbers change from month to month.
The seasonally adjusted data reveals the same basic trend, but makes it possible to measure non-seasonal change in employment patterns:
When events occur that cause significant changes in the labor force supply, the unadjusted numbers sometimes disguise or downplay the change. Using seasonal adjustment allows us to identify these events.
For example, this spike in employment that occurred from May to July 2006 is hidden in the unadjusted line, but shows prominently when seasonal adjustment is applied.
As a general rule, the monthly employment and unemployment numbers reported in the news are seasonally adjusted data. When annual average estimates are calculated, unadjusted numbers are used to ensure the greatest accuracy.