The thing about seasonal adjustments

This post was originally published in the Washington Business Journal’s WBJBizBeat Blog. | Washington Business Journal

The May jobs report that only 69,000 jobs were created during the month was viewed as bad news and markets sold off. While the apparent slowing might have been expected to cause the Federal Reserve to take action, in testimony to Congress on June 8, Fed Chairman Bernanke noted that: “This apparent slowing in the labor market may have been exaggerated by issues related to seasonal adjustment…. the estimated seasonal factors for some economic indicators may have been influenced by the timing of the steepest part of the decline in activity during the 2008-09 winter months….” So the first question to ask is whether seasonal adjustments matter? And the answer is yes. All we have to do is look at the May jobs report without the seasonal adjustment. Looking at the raw data, April 2012 nonfarm payrolls totaled 132,938,000, while May 2012 nonfarm payrolls were 133,727,000. In actuality, there were 789,000 jobs created in May, a far different picture than the seasonally adjusted 69,000 jobs.

The second question is what are seasonal adjustments? According to the Bureau of Labor Statistics, the seasonal adjustment is a statistical technique that attempts to measure and remove the influences of predictable seasonal patterns to reveal how employment and unemployment change from month to month. Because these seasonal events follow a more or less regular pattern each year, their influence on statistical trends can be eliminated by seasonally adjusting the statistics from month to month. These seasonal adjustments make it easier to observe the cyclical, underlying trend.

The third question is how are the seasonal adjustments created? The technical details are that they use X-12 Arima software developed by the Census Bureau to develop the seasonal adjustments. A simplified description is that the seasonal adjustment process used by the Bureau of Labor Statistics looks back in time to create a “smoothing” factor. However, such a look-back may create problems for the seasonal adjustment because the historical data includes data from the financial crisis. While seasonal adjustments are supposed to be normalizing, there was nothing normal about the financial crisis data being used to calculate current seasonal adjustments.

Bottom line: Monthly data like the employment report that have seasonal adjustments can be volatile and are not to be used in isolation when making decisions.