The Consumer Price Index (CPI) produces both unadjusted and seasonally adjusted data. Seasonally adjusted data are computed using seasonal factors derived by the X-13ARIMA-SEATS Seasonal Adjustment Method. These factors are updated each February, and the new factors are used to revise the previous five years of seasonally adjusted data. For more information on data revisions and exceptions to the usual revision schedule, please see the Fact Sheet on Seasonal Adjustment and the Timeline of Seasonal Adjustment Methodological Changes.
For analyzing short-term price trends in the economy, seasonally adjusted changes are usually preferred since they eliminate the effect of changes that normally occur at the same time and in about the same magnitude every year—such as price movements resulting from changing climatic conditions, production cycles, model changeovers, holidays, and sales. This allows data users to focus on changes that are not typical for the time of year.
The unadjusted data are of primary interest to consumers concerned about the prices they actually pay. Unadjusted data are also used extensively for escalation purposes. Many collective bargaining contract agreements and pension plans, for example, tie compensation changes to the Consumer Price Index before adjustment for seasonal variation. BLS advises against the use of seasonally adjusted data in escalation agreements because seasonally adjusted series are revised annually.
The Bureau of Labor Statistics uses Intervention Analysis Seasonal Adjustment for some CPI series. Sometimes extreme values or sharp movements can distort the underlying seasonal pattern of price change. Intervention Analysis Seasonal Adjustment is a process by which the distortions caused by such unusual events are estimated and removed from the data prior to calculation of seasonal factors. The resulting seasonal factors, which more accurately represent the seasonal pattern, are then applied to the unadjusted data.
For the seasonal factors introduced in January 2020, BLS adjusted 53 series using Intervention Analysis Seasonal Adjustment, including selected food and beverage items, motor fuels, electricity, and vehicles. For example, this procedure was used for the utility (piped) gas series to offset the effects of events such as the 2014 increase in demand due to extreme winter weather in the Midwest and Northeast.
Seasonally adjusted data, including the U.S. city average all items index levels, are subject to revision for up to five years after their original release. Every year, economists in the CPI calculate new seasonal factors for seasonally adjusted series and apply them to the last five years of data. Seasonally adjusted indexes beyond the last five years of data are considered to be final and not subject to revision. In January 2020, revised seasonal factors and seasonally adjusted indexes for 2015-2019 were calculated. For directly adjusted series, the seasonal factors for 2019 will be applied to data for 2020 to produce the seasonally adjusted 2020 indexes.
Each year the seasonal status of every series is reevaluated based upon certain statistical criteria. Using these criteria, BLS economists determine whether a series should change its status: from "not seasonally adjusted" to "seasonally adjusted", or vice versa. If any of the 81 components of the U.S. city average all items index change their seasonal adjustment status from seasonally adjusted to not seasonally adjusted, not seasonally adjusted data will be used in the aggregation of the dependent series for the last five years, but the seasonally adjusted indexes before that period will not be changed. For 2020, 28 of the 81 components of the U.S. city average all items index are not seasonally adjusted.
For a more technical discussion of seasonal adjustment methodology, see the following excerpts from the BLS Handbook of Methods:
Last Modified Date: February 11, 2020