'Tis the Season to Seasonally Adjust
I have a long commute. (This is Atlanta—who doesn’t?) During my trek into the office last week, I heard an alarming account on the radio about Georgia’s unemployment rate. The report noted that the state’s unemployment rate had jumped to 9.3 percent in June—its highest level in nearly a year. Thankfully, I’m kind of a data geek and I reckoned that this figure was the nonseasonally adjusted number. The radio report didn’t mention that detail, but it’s a pretty important one. Here’s why.
Economists refrain from reading too much into month-to-month fluctuations in nonseasonally adjusted data. Rather, they tend to look at data that are seasonally adjusted. It’s not a trick or a way to spin the data. It is a statistically sound method that, according to the U.S. Bureau of Labor Statistics, “eliminates the influences of weather, holidays, the opening and closing of schools, and other recurring seasonal events from economic time series. This permits easier observation and analysis of cyclical, trend, and other nonseasonal movements in the data. By eliminating seasonal fluctuations, the series becomes smoother and it is easier to compare data from month to month.”
Labor market data are particularly susceptible to the influence of seasonal factors during the summer months. Think of the school year: New graduates entering the labor force, teachers off for the summer, school administrators and maintenance workers scaling back, etc. These workers return later in the summer when school reopens (something my kids are dreading, by the way). Other seasonal factors affect the data at other points in the year, such as retailers that step up hiring over the holidays, then return to pre-holiday staffing levels in January.
The chart below compares shows Georgia’s nonseasonally adjusted and seasonally adjusted unemployment rates from January 2011 through June 2013.
The nonseasonally adjusted data show much greater volatility, most notably during the summer months and over the holidays, than do the seasonally adjusted data. The spike in June 2013 is particularly notable.
The table below shows the difference in unemployment rates among states in the Southeast. All experienced much larger jumps in the nonseasonally adjusted measure in June than they did for the seasonally adjusted gauge.
Another way to account for seasonal factors is to look at the year-over-year percent change in the nonseasonally adjusted data, which we do in the chart below. Compared with last June, Georgia has fewer total unemployed—2.3 percent fewer, to be exact.
You’ve probably noticed a couple of things in the charts above. In the first chart, even the seasonally adjusted data show an increase in Georgia’s unemployment rate. In the second chart, while the year-over-year percent change in nonseasonally adjusted level of unemployed is negative, it is decidedly less negative that it has been. The message here is that regardless of which unemployment rate measurement you look at, they continue to reflect high levels of unemployment. But it’s important to recognize that seasonally adjusted data better reflect the underlying trends in labor market data.
By Mike Chriszt, a vice president in the Atlanta Fed’s public affairs department
TrackBack URL for this entry:
Listed below are links to blogs that reference 'Tis the Season to Seasonally Adjust: