Skip to content
ContentLora

    Tip: press / anywhere to search.

    Explainer

    Seasonal adjustment: why official figures are 'seasonally adjusted'

    Seasonal adjustment estimates the regular calendar patterns in a time series, such as the December jump in retail sales, and removes them so that one month or quarter can be compared fairly with the next.[1][2] Comparing the same month a year apart does not fully do the job, because holidays such as Easter move between months and turning points show up late.[3] Many headline figures, including US GDP, are published seasonally adjusted, and agencies such as the US Census Bureau maintain software such as X-13ARIMA-SEATS to do it.[4][5]

    Editor reviewedUpdated Explainer14 claims · 100% primaryStatistics and evidence, Economy and trade

    Many economic figures rise and fall on a calendar rhythm: shops sell more in December, energy use climbs in winter, and activity shifts with the school year. Seasonal adjustment strips out those regular patterns so that a change from one month to the next reflects something other than the time of year.[2][6] This page explains what is removed, how statisticians do it and what to watch for when reading adjusted numbers.

    What seasonal adjustment removes

    The US Census Bureau defines seasonal adjustment as estimating the seasonal component of a time series, and where relevant trading-day and moving-holiday effects, and then removing them. The goal is a series whose movements are easier to analyse over consecutive periods.[1] The UK’s Office for National Statistics (ONS) gives a simple example: retail sales rise each December because of Christmas.[2]

    The ONS lists several causes of seasonal effects:[6]

    • Weather, such as higher energy use with the onset of winter.
    • Administrative timing, such as the start and end of the school year.
    • Social, cultural and religious events, such as shopping before Christmas.
    • Calendar length, since months and quarters have different numbers of days.

    There are also calendar effects that do not fall in the same month every year. Trading-day effects come from months having different numbers of each weekday, and moving holidays such as Easter can fall in March or April.[7]

    Trend, seasonal and irregular

    Statisticians treat a time series as a combination of parts. The ONS describes three: calendar-related effects, irregular fluctuations, and the trend.[8] Irregular movements come from unpredictable causes such as sampling error, unseasonable weather, natural disasters or strikes.[8] A seasonally adjusted series removes only the seasonal part, so it still contains both the trend and the irregular movements.[8]

    The parts can be combined in two main ways. In an additive model the seasonal effects stay broadly the same size whatever the trend does; in a multiplicative model they change in proportion to the trend. The ONS says most economic series fit a multiplicative model best, but that model cannot be used if a series contains zero or negative values.[9]

    Why not just compare with last year?

    Comparing a month with the same month a year earlier removes some seasonality, but the Australian Bureau of Statistics (ABS) says it does not remove all of it: holidays such as Easter and Chinese New Year fall in different periods in different years. The ABS adds that year-apart comparisons lack precision and are slow to reveal turning points, when a series changes direction.[3]

    Adjustment only makes sense for series that are seasonal. The ABS warns that seasonally adjusting a series with little seasonality is impractical and will often introduce an artificial seasonal pattern.[10]

    How it is done

    The X11 method, developed by the US Census Bureau, began operation in 1965 and was adopted by many statistical agencies around the world.[11] The Census Bureau also produces and maintains X-13ARIMA-SEATS, software that offers both X-11-style filtering and model-based adjustment using a version of the SEATS software originally developed at the Bank of Spain.[5]

    Before adjusting, statisticians deal with unusual values. The ONS says outliers such as strikes or extreme weather, trend breaks and abrupt changes in the seasonal pattern need to be handled first to get a reliable estimate of the seasonal component.[12]

    Reading adjusted figures

    • Expect revisions to recent months. Filters lack enough data at the end of a series, the end-point problem. Extending the series with ARIMA forecasts reduces revisions, but recent adjusted figures can still change as new data arrive.[13]
    • Check which version you are looking at. BEA seasonally adjusts US GDP to remove patterns such as winter weather, holidays and factory production schedules, but it also publishes unadjusted data.[4]
    • Watch for annual rates. BEA reports quarterly GDP at annual rates, so a quarterly figure is expressed as a yearly pace.[4]
    • Remember the noise. Adjusted data still include irregular movements, so a single month’s change may not signal a new trend.[8][14]

    For where adjusted figures fit into the wider production of official numbers, read how official statistics are produced.

    Next in the crash course

    Developing in Science & Energy

    Questions readers ask

    What does seasonally adjusted mean?

    It means the regular seasonal pattern, and where relevant trading-day and moving-holiday effects, has been estimated and removed from the figures, so that movements between consecutive periods are easier to analyse.[1]

    What causes seasonal patterns?

    The ONS lists weather, such as higher energy use in winter; administrative timing, such as school year dates; social, cultural and religious events such as Christmas shopping; and the differing lengths of months and quarters.[6][2]

    Why not just compare with the same month last year?

    The Australian Bureau of Statistics says this does not fully remove seasonal effects, because holidays such as Easter and Chinese New Year fall in different periods each year, and it delays spotting turning points.[3]

    Why do seasonally adjusted figures change after they are published?

    Adjustment filters lack enough data at the end of a series. Extending the series with ARIMA forecasts reduces revisions, but the newest adjusted figures can still change as later data arrive.[13]

    Is US GDP seasonally adjusted?

    Yes. BEA seasonally adjusts GDP to remove yearly patterns such as winter weather, holidays and factory production schedules, reports quarterly figures at annual rates, and also publishes unadjusted data.[4]

    Sources

    · 14 claims from 5 publishers

    Each numbered claim is a statement we checked against the sources listed with it. Tap a number in the text to see its sources in place.

    1. [1]

      The Census Bureau defines seasonal adjustment as estimating the seasonal component of a time series, and where relevant trading-day and moving-holiday effects, and removing them, so that movements are easier to analyse over consecutive periods. confirmedas of 2026-11-14

    2. [2]

      The ONS says seasonal adjustment removes systematic calendar-related variation so that consecutive periods can be compared; for example, retail sales rise each December because of Christmas. confirmedas of 2026-11-14

    3. [3]

      The Australian Bureau of Statistics says comparing the same month in different years does not fully remove seasonal effects, because holidays such as Easter and Chinese New Year move between periods, and that it delays spotting turning points. confirmedas of 2026-11-14

    4. [4]

      BEA seasonally adjusts GDP to remove yearly patterns such as winter weather, holidays and factory production schedules, also publishes unadjusted GDP data, and reports quarterly figures at annual rates. confirmedas of 2026-11-14

    5. [5]

      The Census Bureau produces and maintains X-13ARIMA-SEATS, seasonal adjustment software that offers both model-based adjustment using a version of the SEATS software originally developed at the Bank of Spain and nonparametric X-11 adjustments. confirmedas of 2026-11-14

    6. [6]

      The ONS lists causes of seasonal effects including weather (such as higher energy use in winter), administrative timing (such as school year dates), social, cultural and religious events, and differing lengths of months and quarters. confirmedas of 2026-11-14

    7. [7]

      Other calendar effects include trading-day effects, from months having different numbers of each weekday, and moving holidays such as Easter, which can fall in March or April. confirmedas of 2026-11-14

    8. [8]

      The ONS treats a time series as a combination of calendar-related effects, irregular fluctuations (from causes such as sampling error, unseasonable weather, natural disasters or strikes) and the trend; removing the seasonal component leaves the trend and irregular parts in the seasonally adjusted series. confirmedas of 2026-11-14

    9. [9]

      Seasonal effects can be modelled as additive (roughly constant in size) or multiplicative (changing in proportion to the trend); the ONS says most economic series are multiplicative, but that model cannot be used if a series has zero or negative values. confirmedas of 2026-11-14

    10. [10]

      The ABS warns that seasonally adjusting a series with little seasonality is impractical and will often introduce an artificial seasonal pattern. confirmedas of 2026-11-14

    11. [11]

      The X11 method, developed by the US Census Bureau, began operation in the United States in 1965 and was adopted by many statistical agencies around the world. confirmedas of 2026-11-14

    12. [12]

      The ONS says outliers (such as strikes or extreme weather), trend breaks and abrupt changes in the seasonal pattern need to be dealt with before seasonal adjustment to get a reliable seasonal estimate. confirmedas of 2026-11-14

    13. [13]

      Moving-average filters lack enough data at the end of a series (the end-point problem), and extending the series with ARIMA forecasts helps reduce later revisions to the seasonally adjusted figures. confirmedas of 2026-11-14

    14. [14]

      UK government guidance on communicating uncertainty recommends showing confidence intervals around estimates where applicable, and warns that changes between periods can fall within survey error and random variation, especially in weekly, monthly or quarterly series. confirmedas of 2026-11-14

    Revision history (1)
    1. Page created.

    Created Nov 14, 2026. Last reviewed by an editor on Nov 14, 2026. Next scheduled review: Sep 8, 2027.

    Cite this page

    "Seasonal adjustment: why official figures are 'seasonally adjusted'." ContentLora, updated Nov 14, 2026. https://contentlora.com/explain/seasonal-adjustment-explained

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.