Not seasonally adjusted: Meaning, Criticisms & Real-World Uses

What Is Not Seasonally Adjusted?

"Not seasonally adjusted" data refers to raw economic figures that have not undergone statistical adjustments to remove predictable, recurring patterns tied to the calendar year. These patterns, known as seasonality, are common in various economic indicators and typically result from factors such as holidays, weather, or academic calendars. In economic statistics, presenting data as not seasonally adjusted means that the monthly or quarterly variations still reflect these regular, calendar-driven fluctuations.

Understanding data that is not seasonally adjusted is crucial for a complete data analysis, as it provides an unvarnished view of economic activity before any mathematical smoothing. This raw data highlights the actual observed numbers, including all their natural, cyclical surges and dips. Analysts often compare not seasonally adjusted figures year-over-year to bypass the need for complex adjustments while still observing underlying trends.

History and Origin

The practice of accounting for seasonal patterns in economic data gained prominence as modern economic policy and analysis matured. In the early 20th century, institutions like the Federal Reserve began to realize that raw economic numbers contained significant seasonal swings that could obscure underlying economic trends. For instance, a sharp drop in employment every January might simply reflect the end of holiday retail hiring, rather than a genuine economic downturn. To better gauge fundamental economic shifts, the Federal Reserve started developing methods to produce seasonally adjusted data based on raw figures.¹⁰

Initially, these adjustments were labor-intensive, often performed manually with mechanical calculators and graph paper.⁹ However, as computing power advanced, the U.S. Bureau of Labor Statistics (BLS) and other statistical agencies took on the responsibility of performing and publishing seasonal adjustments. By the 1960s, the BLS began directly adjusting and releasing key economic series, making it easier for economists and the public to analyze underlying trends without being misled by predictable seasonal "noise."⁸ Nevertheless, the original, not seasonally adjusted figures continue to be published and remain vital for specific analytical purposes.

Key Takeaways

"Not seasonally adjusted" data represents raw economic figures, reflecting all observed fluctuations, including predictable seasonal patterns.
These figures show the actual magnitude of changes influenced by calendar events like holidays, weather, or school cycles.
Comparing not seasonally adjusted data year-over-year helps analysts identify trends while sidestepping seasonal variations.
It is essential for understanding the true, unadjusted volume of activity in sectors with strong seasonal components, such as retail or construction.
Government agencies often release both seasonally adjusted and not seasonally adjusted versions of economic indicators to serve different analytical needs.

Interpreting the Not Seasonally Adjusted

Interpreting data that is not seasonally adjusted requires an awareness of the inherent calendar-driven patterns. When examining a series such as retail sales or employment rate, a month-over-month comparison of not seasonally adjusted data can be misleading because it will often show expected seasonal spikes or troughs. For example, holiday hiring traditionally boosts employment in November and December, followed by a decline in January. Looking at January's not seasonally adjusted unemployment rate alone might suggest a significant economic deterioration, when it merely reflects a normal seasonal unwinding of temporary jobs.

To derive meaningful insights from not seasonally adjusted figures, analysts frequently compare a specific month's data to the same month in previous years. This year-over-year comparison effectively removes the seasonal component by looking at similar calendar periods, allowing for clearer identification of underlying growth or contraction. For instance, comparing July's not seasonally adjusted unemployment rate this year to last July's rate can reveal genuine shifts in the labor market. The Bureau of Labor Statistics (BLS) provides both not seasonally adjusted and seasonally adjusted data for many of its series, allowing users to choose the appropriate measure for their analytical needs.⁷

Hypothetical Example

Consider a small island economy heavily reliant on tourism. Its central statistical office publishes monthly tourism arrival figures.

January: 10,000 visitors
February: 11,000 visitors
March: 15,000 visitors (Spring break begins)
April: 18,000 visitors
May: 22,000 visitors (Summer holiday season starts)
June: 25,000 visitors
July: 26,000 visitors
August: 24,000 visitors (School restarts elsewhere)
September: 16,000 visitors
October: 14,000 visitors
November: 12,000 visitors
December: 20,000 visitors (Winter holiday rush)

These are the not seasonally adjusted figures. If you compared February (11,000) to March (15,000), it would show a 36% jump. However, this large increase is primarily due to the predictable influx of spring break tourists. Similarly, the drop from July (26,000) to September (16,000) is a 38% decline, but it reflects the end of the peak summer season.

To understand the island's true tourism health beyond seasonal swings, an analyst would look at, for example, July's 26,000 visitors this year versus July's visitors from the previous year. If last July had 24,000 visitors, the 8.3% increase would indicate a genuine growth in the island's tourism appeal, irrespective of the annual summer peak. This year-over-year comparison uses the not seasonally adjusted data to strip out the calendar's influence and pinpoint underlying trends relevant for financial markets and local businesses.

Practical Applications

Not seasonally adjusted data is critical in several practical applications, particularly when understanding the true magnitude of an event or the actual volume of activity occurring during specific periods.

One primary use is in industries with strong seasonal patterns, such as retail, construction, or agriculture. For example, businesses in the retail sector closely monitor not seasonally adjusted retail sales data to assess the actual impact of holiday shopping seasons or back-to-school periods on their revenues and staffing needs. They need to know the true number of transactions and units sold, rather than an adjusted figure that smooths out these crucial peak times.

Similarly, the construction industry tracks not seasonally adjusted manufacturing data related to housing starts and building permits to understand the actual pace of activity, which naturally fluctuates with weather conditions across different seasons. The Bureau of Labor Statistics (BLS) regularly publishes not seasonally adjusted employment and unemployment rate figures for specific sectors or geographic areas, which are vital for granular analysis of labor market dynamics.⁶ For instance, when analyzing the impact of temporary hiring for the holiday season, one would look at the not seasonally adjusted figures to observe the real, unadjusted changes in employment levels.⁵ This raw data provides the necessary context for operational planning and policy evaluation, particularly when assessing short-term, actual changes in economic activity.

Limitations and Criticisms

While not seasonally adjusted data offers a raw, unfiltered view of economic activity, it comes with notable limitations, especially for general economic analysis and forecasting. The most significant drawback is the presence of inherent seasonal fluctuations, which can obscure underlying trends and cycles in a time series. Month-over-month comparisons of not seasonally adjusted data can be misleading, as a rise or fall might simply reflect a predictable seasonal pattern rather than a genuine shift in the business cycle. This can lead to misinterpretations of the economy's direction, potentially triggering unwarranted reactions from policymakers or market participants.

Critics argue that focusing solely on not seasonally adjusted data makes it difficult to ascertain the true state of the economy or identify the turning points of a business cycle.⁴ Large, non-seasonal shocks, like the COVID-19 pandemic, can also complicate the interpretation of both seasonally adjusted and not seasonally adjusted data. In such unprecedented events, the usual seasonal patterns may be drastically altered, making historical comparisons less reliable even for not seasonally adjusted data, and challenging the assumptions behind seasonal adjustment methodologies.³ For analysts seeking to understand the cyclical or trend components of an economic series, the raw, not seasonally adjusted figures often require additional statistical methods or year-over-year comparisons to filter out seasonal noise effectively.

Not Seasonally Adjusted vs. Seasonally Adjusted

The key difference between "not seasonally adjusted" and "Seasonally adjusted" data lies in the statistical treatment applied to the raw economic figures.

Not Seasonally Adjusted (NSA) data represents the original, observed values for an economic series. It includes all fluctuations, both those due to underlying economic trends and cycles, and those that occur regularly at the same time each year due to seasonal influences like holidays, weather patterns, or academic schedules. For instance, the not seasonally adjusted unemployment rate in a given month will show the actual number of people unemployed, including the typical spike in new graduates entering the job market in summer or a drop in retail employment after the winter holidays. This raw data is useful for understanding the absolute magnitude of economic activity during specific periods.

Seasonally Adjusted (SA) data, on the other hand, has undergone a statistical process to remove these predictable seasonal variations. The goal is to isolate the underlying trend and cyclical components of a time series, making it easier to compare data from one month or quarter to the next and identify significant non-seasonal shifts in economic activity. When economists and policymakers discuss monthly changes in key indicators like gross domestic product or the Consumer Price Index, they are almost always referring to seasonally adjusted figures to avoid misinterpretations caused by normal seasonal patterns. While not seasonally adjusted data provides the real, unmanipulated numbers, seasonally adjusted data offers a clearer view of the economy's fundamental direction.

FAQs

Why do government agencies publish data that is not seasonally adjusted?

Government agencies, such as the Bureau of Labor Statistics, publish data that is not seasonally adjusted to provide a complete picture of economic activity. This raw data is essential for understanding the actual, unadjusted counts or volumes for a specific period, including all seasonal effects. For certain analyses, like comparing a month's performance to the same month in previous years, or understanding the real demand for temporary workers during holidays, the not seasonally adjusted data is more appropriate.²

Can I use not seasonally adjusted data for month-over-month comparisons?

While you can technically compare not seasonally adjusted data month-over-month, it is generally not recommended for assessing underlying economic trends. Such comparisons will be heavily influenced by predictable seasonal patterns, making it difficult to discern true changes in the business cycle or other non-seasonal movements. For assessing trends, it is usually better to compare not seasonally adjusted data year-over-year or use seasonally adjusted data.

When is not seasonally adjusted data more useful than seasonally adjusted data?

Not seasonally adjusted data is more useful when the seasonal pattern itself is the focus of the analysis, or when you need to understand the absolute volume or impact of an event during a specific calendar period. This applies to industries with strong seasonal variations like retail, tourism, or construction, where businesses need to track actual sales, employment, or production figures to manage operations, staffing, and inventory. It is also important when looking at very recent data that might not yet have reliable seasonal adjustments, or when evaluating the direct, unadjusted impact of short-term events.¹

Does "not seasonally adjusted" mean the data is inaccurate?

No, "not seasonally adjusted" does not mean the data is inaccurate. It simply means the data has not been processed to remove the regular, predictable fluctuations that occur at the same time each year. It is the raw, observed data. Its accuracy depends on the original collection methods, not on whether it has been seasonally adjusted or not. Both not seasonally adjusted and Seasonally adjusted data are valid and provide different perspectives for economic analysis.