CPI item aggregation

All CPI series are constructed from one or more basic indexes. A basic index is the index for one of 243 basic items in one of the 32 basic areas[1], for example apples (SEFK01) in the New York-Newark-Jersey City, NY-NJ-PA CBSA (S12A). Taking each unique combination of such items and areas results in 7,776 (243 x 32) item-area combinations that form the basic building blocks of CPI series.

As a practical matter all not seasonally adjusted (NSA) CPI series are estimated as the aggregation of a set of basics. Table 1: CPI aggregate items shows the basic items which compose each aggregate item. Note that this is not necessarily the case for seasonally adjusted aggregate CPI series. For more information, see the CPI seasonal item aggregation page.

While all NSA CPI series are built from the bottom up through the aggregation of basic indexes, it is also true that some aggregates are subsets of other aggregates, resulting in a hierarchy called the CPI Item Aggregation Tree, which can be helpful in visualizing the relationships between different CPI series.

Download an XLSX version of the CPI aggregation trees which shows how the basic items feed into expenditure classes, major groups, and selected intermediate special aggregates here:

• CPI item aggregation trees (XLSX)

CPI items can be divided into several levels of complete and mutually exclusive categories which span different levels of aggregation:

• Entry level items (ELIs) are the most granular, complete, and mutually exclusive breakdown of CPI items. ELIs are used internally for sampling, data collection, database organization purposes, and as basic items in the case of health insurance. They have five-character codes which illustrate which item stratum, expenditure class, and major group they belong to, as described below. More information about ELIs is available in CPI Handbook of Methods Appendix 2: Content of CPI Entry Level Items.

• The 211 item strata are groups of one or more ELIs which cover related items and are the most granular complete and mutually exclusive breakdown of CPI items for which data is publicly available. For most items, this level is the basic level used for aggregation. These series have six-character codes of which the last four match the first four characters of the ELIs within them. For example, the item stratum other beverage materials including tea (SEFP02) is composed of two ELIs: tea (FP021) and other beverage materials (FP022).

• The 70 expenditure classes (ECs) are groups of one or more item strata which cover related items. These series have four-letter codes which match the first four characters of item strata within them. For example, the EC beverage materials including coffee and tea (SEFP) is composed of two item strata: coffee (SEFP01) and other beverage materials including tea (SEFP02).

• The eight major groups: apparel (SAA), education and communication (SAE), food and beverages (SAF), housing (SAH), medical care (SAM), other goods and services (SAG), recreation (SAR), and transportation (SAT). Major groups are composed all of the ECs which share the third character of the item code – so beverage materials including coffee and tea (SEFP) would be contained in food and beverages (SAF).

• All items (SA0) and all items (old base) (AA0) are the top-level series which combine all basic items. (Or, alternatively, the combination of all series of any level which is complete and mutually exclusive in its coverage of basic items, like ECs or major groups.)

Note that the number of basic items (243) used for the calculation of aggregate indexes is larger than the number of item strata (211) because the entry level item (ELI) level is used for the calculation of basic indexes for health insurance retained earnings (item code SEME) rather than the item stratum level.

Download an XLSX file which shows how the basic items are grouped into expenditure classes and major groups here:

• CPI basic item aggregation (XLSX)

In addition, there are also a number of CPI series which lie outside of this aggregation hierarchy. Broadly, these series fall into two categories:

• Special aggregates are series which represent a group of items which may be of interest to certain users, but which are not grouped together in expenditure classes or major groups, such as food at home (SAF1) or energy (SA0E). Other special aggregate series exclude items which are included in other, similar aggregates – a prominent example of this is all items less food and energy (SA0L1E). Special aggregate series usually have codes which begin with SA, but this is not always the case. Exceptions to this rule include information technology commodities (SEEEC), video and audio products (SERAC), and video and audio services (SERAS).