So far this year, Wisconsin’s total employment level peaked in July with close to three million employed workers. The employment total will likely fall off during the last months of 2003, but that is the normal pattern. Like waves rolling toward a beach, fluctuations of employment occur regularly. From August through the end of the year and through the following January, employment decreases. However, that decrease is offset by employment growth from February through July.
The spring and early summer growth overcomes the previous year’s decrease when the economy is healthy and expanding. If the economy is in a recession, seasonal growth is too meager to reach the previous peak and the underlying trend in employment is down.
This seasonal routine of growth and loss shows up in employment and unemployment, in urban areas and rural areas and in some industries more than others. What’s more, seasonal employment growth in some industries may offset loss in others.
What causes the seasonal phenomenon? Influences like the regular shifts in weather, the opening and closing of schools, holidays, retooling of factories for annual model changes, and even annual sales events held by retailers. Many employers in Wisconsin shut-down operations in November during deer hunting season, causing sharp increases of unemployment and unemployment compensation claims.
So what? How does an awareness of seasonal change help a job-seeker, an employer, or anyone concerned with the labor market? Seasonal events can distort judgment about the current labor market and can hide underlying trends in employment. Employment in the construction industry, say, may decrease throughout the state by 2500 jobs from August to September. But, construction is a seasonal activity that normally slows in fall, and the long-term trend may be substantial growth. The August to September decrease gives a false indication of the industry’s health.
Because seasonal change is regular and predictable, analysts interested in underlying trends rely on seasonally adjusted data. Seasonally adjustment smoothes off the high points and fills in the valleys of a year’s worth of monthly data. It is a process that functions within a year ... all the monthly seasonally adjusted data in a year add up to all the monthly unadjusted data for that year.
Averaging all the monthly data for a year, producing annual averages, also deletes the seasonal influences, but a complete year’s worth of data must be averaged. Thus, the underlying trend for the current year becomes available only after December.
Comparing unadjusted to seasonally adjusted data is not valid.
National employment, unemployment, and unemployment rate numbers are usually seasonally adjusted.
Releases of Wisconsin statewide and metropolitan area employment, unemployment, and unemployment rates contain both unadjusted and seasonally adjusted data.
County data is not seasonally adjusted, unless a single county is one of the state’s designated metropolitan areas.
Employment in specific industries is not seasonally adjusted, so month-to-month changes of employment can be misleading.
Written by Joe Tumpach, Research Analyst, Department of Workforce Development, Office of Economic Advisors, November 2003.