The Canadian Consumer Price Index Reference Paper
Chapter 11 – History

11.1 The Consumer Price Index (CPI) began with a study conducted by the Department of Labour in the early1900s. The study was based on a hypothetical family budget that represented weekly expenditures of an urban working man’s family of five. Retail prices for 29 food products and 5 fuel and lighting products were collected in approximately 60 cities. In addition, information was obtained on the rent for a representative working man’s six-roomed dwelling. Using these data, indices on a 1900 index base period were calculated for Canada and the provinces. The calculation of these indices continued until August 1940.

11.2 The Department of Labour also started producing a “Cost-of-Living Index”^Note  on a 1913 index base period, with component indices for food, fuel and lighting, rent, clothing and sundries. This index was published for June and December from 1914 to 1917, for April, June, September and December from 1918 to 1926 and monthly from 1927 onward. An attempt was made to weight product classes according to their actual importance in wage-earners’ spending, even though no extensive household expenditure survey had been undertaken.

11.3 The first index of retail prices produced by the Dominion Bureau of Statistics also had a 1913 index base period. This index was calculated using prices from the Department of Labour series as well as some prices obtained directly from retailers. Basket weights used in the index were based on estimates of the total Canadian consumption of each product in 1913. The index was subsequently updated and produced on a 1926 index base period. Although the weighting system of the updated index was more refined, it was still based on the estimated total consumption in Canada. The number of product prices collected increased substantially at this time.

11.4 The index was again updated in 1940 and published on a 1935-39 index base period. The basket weights used in this index were derived from a 1938 Family Expenditure Survey (FAMEX) for urban wage-earner families with annual incomes between $450 and $2,500. The 1940 update showed that the Bureau had come to accept the Department of Labour’s view that the index should measure price changes experienced by a well-defined demographic group.

11.5 Until 1940 the CPI was a direct Laspeyres^Note  index for its entire or more recent estimation period. Any index can be calculated as a Laspeyres index when it is initiated, and until the 1947-48 basket update in 1949, it was considered acceptable to revise the CPI backward for several years at the time of a basket update.^Note 

11.6 A subsequent FAMEX covering the period 1947-48 provided the basis for the next thorough basket update of the index in 1952. At that time, 1949 became the index base period and the title was changed from “Cost-of-Living Index” to “Consumer Price Index (CPI)”.^Note  The CPI was defined as “a measure of the percentage change through time in the cost of purchasing a fixed basket of goods and services representing the consumption of a particular population group during a given period of time”. This definition remains in essence unaltered to this day.

11.7 Soon after that update, a series of small-scale biennial surveys of family expenditures were undertaken and their results were used to choose the dates of subsequent CPI basket updates. The changes in family expenditure patterns shown by the 1957, 1967 and 1974 surveys were deemed sufficiently important to justify the implementation of new baskets in January 1961, May 1973 and October 1978, respectively.

11.8 Several important changes were introduced with the 1974 basket. In particular, family size and income constraints were removed, thus broadening the target population. Also, with the updating of the 1974 basket the national indices began to be calculated as weighted averages of the corresponding indices for 59 urban centres. In addition, the New Housing Price Index replaced the Residential Building Construction Input Price Index in the CPI series measuring homeowners’ replacement cost, mortgage interest cost and homeowners’ home and mortgage insurance.

11.9 Following the adoption of the 1974 basket in the CPI, a policy of regular updating was established, with the updates tied to a four-year cycle of major FAMEX surveys. The 1978 basket was thus incorporated in April 1982, the 1982 basket in January 1985, and the 1986 basket in January 1989. The time lag between the basket reference period and the implementation of the basket was gradually reduced.

11.10 There was a six-year interval between the 1986 and the 1992 baskets, longer than the regular four-year interval. The introduction of the 1992 basket was postponed by two years to ensure it would reflect adjustments to consumption patterns resulting from the introduction of the Goods and Services Tax (GST) and from the removal of the Federal Sales Tax in January 1991. However, the time lag between the basket reference year and basket implementation remained 24 months for the 1992 basket, which was implemented with the January 1995 update.

11.11 The 1996 basket was introduced with the January 1998 update and was the last basket update that was based on FAMEX data as the survey was subsequently discontinued. The three-year lag between the introduction of the 1996 basket and the introduction of the previous basket was at the time a record for the CPI.

11.12 There was a five-year lag between the 1996 and 2001 baskets, although the basket was introduced with the January 2003 update, matching the record of the 1996 basket update’s record for the shortest lag between the end of the basket reference year and the month of implementation. The FAMEX survey was replaced in 1997 with an annual Survey of Household Spending (SHS). Starting with 1999, the SHS estimates were calculated for the provinces and territories in odd-numbered years and for the provinces only in even-numbered years. Since the CPI includes the territorial capitals Yellowknife, Whitehorse and Iqaluit as geographical strata, the year 2000 was precluded as a basket reference year. Budget problems and concerns about year 2000 bugs in computer software also contributed to the choice of 2001 for the basket update.

11.13 The 2005 and 2009 basket updates marked a return to a four year interval between baskets but also a lengthening in the implementation lag: the baskets were implemented in May 2007 and May 2011 respectively.

11.14 The 2011 basket marked the first-ever two-year interval between basket updates for the CPI. It also marked a reduction in the implementation lag to 14 months after the basket reference period, since the new basket was introduced in March 2013 with the release of February 2013 CPI data.

11.15 The 2013, 2015 and 2017 basket updates followed the two-year interval between basket updates and marked a reduction in the implementation lag to 13 month after the basket reference period, since they were respectively introduced in February 2015, February 2017 and February 2019.

11.16 Whenever a basket update takes place, the concepts and the procedures used to calculate the CPI are reviewed and revised when necessary. The CPI index base period has also been periodically changed. While not required to be implemented at the time of a basket update, for operational reasons many of the changes to the CPI index base period have taken place alongside the updating of the basket of goods and services. Changes to the index base period have usually related to changes in the base year of the Canadian System of National Accounts.

11.17 A detailed chronological history of all the basket updates for the CPI can be found in Appendix C. It includes the basket reference period; the basket link month; the basket end month; the first month published with the new basket; any major revisions made at the time of the basket update; changes to the target population and/or geographical coverage; the introduction of new products into the basket or sample; as well as notable changes that were made to the classifications, scope or methodologies.