Most of Idaho experiences below freezing temperatures and is often covered under a blanket of snow between November and February. When the weather turns colder, employment levels decline in industries such as construction, mining/logging and leisure/hospitality. The holiday season and winter breaks result in fluctuations in educational services and retail. Despite the Farmer’s Almanac predicting Idaho’s 2022-2023 winter season to be dry and calm, it will still see patterns in our overall employment levels where the direction and magnitude can often be forecasted. This reduction in employment does not represent a permanent trend for analyzing the labor force but reverses in a predictable manner throughout the calendar year. Data that is seasonally adjusted helps reduce the noise of recurring seasonal fluctuation and show the true underlying employment trends present in Idaho’s labor force.
Idaho’s employment often peaks in October and reaches its lowest point in January
For seven of the past 10 years (2012-2021), Idaho’s employment levels have reached the highest annual point during the month of October. Why? October’s weather is generally mild. Seasonal payrolls in fair weather industries, such as construction and logging, can still operate at near full employment. Other sectors (retail, transportation, warehousing, leisure/hospitality) are beginning to prepare for the busy upcoming holiday season and are increasing payrolls to keep up with a boost in consumer demand. In line with the beginning of the school year, September and October are also key months where educational services tend to have employment increases compared to low summer break employment levels in July and August.
When looking at data that is not seasonally adjusted for October 2017 through December 2019 (blue line in Figure 1), the data shows repeating patterns of declining employment from November through January and steady employment gains for February through June, followed by a slight dip in July and August before peaking in October. However, when we compare this pattern to seasonally adjusted data (orange line in Figure 1), we finish at the same endpoint for each data series; however, the growth is smoothed, and the extreme peaks and valleys are spread throughout the year to show underlying trends of increasing overall employment. Seasonally adjusted data allows us to realize fundamental growth in a time series without the noise of job gains or losses from external, predictable seasonal factors. Without seasonally adjusted data, it may not be useful to compare employment patterns from one month to the next as seasonal factors need to be considered.
The table below shows predicted seasonal high and low points for selected Idaho industries based on the Bureau of Labor Statistics data from 2012-2021. To demonstrate how significant seasonal factors can be for a given industry, the last column shows the average change in employment between the maximum and minimum calendar months each year from 2012-2021.
Industries in Idaho with the highest employment levels during the winter
Although overall employment in Idaho declines when winter hits, a few industries move in a counter direction and realize their highest demand periods of the year.
Winter recreation – Idaho’s ski resort employment peaks at 2,000-2,500 jobs statewide during the first quarter of each calendar year, but then plummets by two thirds to levels of 700-800 during the seasonal warmth of the second and third quarters. For the third year in a row (2021-2023), readers of Ski Magazine have named Sun Valley Resort as the Top Ski Resort in the West 2023. On average, nearly 70% of Idaho’s statewide ski resort employment occurs in the first and fourth quarters of each calendar year.
Retail sales – Retail’s peak employment months of November and December coincide with the end of the calendar year (thanks to Christmas being six days from year end) and first quarter is typically the lowest point of employment all year (with the exception of the extreme second quarter of 2020). Idaho often sees the retail sector first increase employment by 1,500-3,000 temporary seasonal positions in November and December and then sheds them in January when consumer demand slows to more normal levels. With retail employment accounting for approximately 12% of Idaho’s nonfarm jobs and nearly 15% of total private employment, this seasonal bump would have large effects on overall employment patterns if not seasonally adjusted. When comparing fourth quarter average employment for 2019, 2020 and 2021, retail employment is on an upward trend year over year despite cyclical ups and downs during the year. General merchandise and department stores typically show highest annual employment during November and December while motor vehicles and building materials tend to peak in the summer and early fall. Grocery store employment tends to be fairly consistent year round.
The lower retail employment chart compares seasonally adjusted data (orange line) with the not seasonally adjusted data (blue line). Whereas the not seasonally adjusted data shows large drops in employment from 4Q to 1Q, the seasonally adjusted data spreads these effects and smooths the series over the entire year.
Transportation and warehousing – For each of the past 10 years (2012-2021), the transportation and warehousing industry has realized peak annual employment in the month of December and hit the lowest point of employment from April through July, with June being the most common low in five of the past 10 years. Following the trend of retail industry demand being strongest at the end of the year, the flow of merchandise and other goods naturally increases demand for heavy truck drivers, warehouse workers and local delivery drivers (FedEx, UPS, etc.). Average employment for this sector is often 5-10% higher in November through January than during the remainder of the year from February through October.
Benefits and challenges of seasonally adjusted data
Seasonal adjustment is the ongoing process of estimating and accounting for seasonal variation within a time series to determine underlying patterns. The majority of published employment data is seasonally adjusted in order to remove cyclical fluctuations as well as short-term differences such as number of work days, weekends, holidays and employee strikes in any given month. The main goal is to provide a clearer view of net trends and turning points in the data that are not seasonal, while simultaneously allowing different time periods to be analyzed together. The main advantage of seasonal adjustment allows easier comparisons over time as calendar effects, major outliers and one-time events are eliminated.
Despite its advantages, a seasonally adjusted model may not turn out to be an accurate picture of what is truly happening in the moment as the model blends multiple years of historical data (including the most recent) and can be influenced by extraordinary, unpredictable and non-recurring events for any given period. The process is complex and combines both art and science in its fundamental process. Once every year, the Bureau of Labor Statistics analyzes and recalculates seasonal factors for the previous five years using the newest model specifications. This results in a dynamic process that can lead to large revisions in the data up to five years after its original publish date. Seasonal adjustment works well when seasonal patterns are fairly consistent year to year and may incorrectly attribute movement to seasonal factors when it actually resulted from a nonseasonal event. Unfortunately, the model is not always able to adapt to significant external events, such as COVID-19’s drastic and lingering effect on the labor market during 2020 and beyond. The effects from the pandemic have resulted in more active intervention by the statistics team, where previous intervention occurred after only major weather related or natural disaster events (hurricane, earthquake, etc.).
Adjusting for seasonal data is a dynamic process with many moving parts and subject to long-term revision. Despite the drawbacks of this method, data that is seasonally adjusted will provide the user with more information about current trends and growth rates while making different time periods more comparable as cyclical noise is reduced.
 Bureau of Labor Statistics, “The challenges of seasonal adjustment for the Current Employment Statistics survey during the COVID-19 pandemic”, May 2022