Retail Services Price Index (RSPI)
Detailed information for second quarter 2023
The purpose of the survey is to collect and compile price data to measure the monthly change in the movement of the price of retail services over time. These price data are combined and the component indexes chained to form a price index. The estimates are produced on a quarterly basis.
Data release - September 22, 2023
The Retail Services Price Index is constructed from price data collected by the Retail Trade Price Report. These price data are combined to estimate a price index for retail services that can be joined with other business services indexes to provide better estimates of real output and productivity, monitor inflation and feed an important research agenda at Statistics Canada.
These indexes are a part of the Services Producer Price Index program (SPPI) at Statistics Canada.
The SPPI program develops and produces price indexes for a wide range of business service categories. This initiative fills an important data gap in the area of economic statistics and has resulted in a more comprehensive set of service price indexes. It also allows Statistics Canada to produce more accurate estimates of real value added of the Gross Domestic Product and changes in productivity.
Reference period: The period for which the Retail Services Price Index equals 100; currently this is the year 2013.
Collection period: Collection takes place during the quarter following the reference quarter.
- Prices and price indexes
- Retail and wholesale
- Service price indexes
Data sources and methodology
For the Retail Services Price Index, the target population consists of all enterprises with at least one establishment classified to the retail sector (codes 44-45 of the North American Industry Classification System [NAICS]) on the Business Register, excluding Automobile and Other Motor Vehicle Dealers and Non-Store Retailers (codes 4411, 4412 and 454 of the NAICS).
Pricing methodologies and questionnaire design were researched and based on internationally accepted practice. The Retail Trade Price Report (survey questionnaire) was developed and tested through consultation and collaboration with retailers, industry experts and the Questionnaire Design Resource Centre at Statistics Canada.
The questionnaire, last revised in 2014, is designed to collect information on, among other variables, the average monthly purchase price (amount paid for the acquisition of a given product) and the average monthly selling price (amount received for selling the same product), excluding taxes.
This is a sample survey with a longitudinal design.
The frame is part of Statistics Canada's Business Register classified to the Retail sector (codes 44-45 of the North American Industry Classification System [NAICS]).
The sampling unit is at the enterprise level. The enterprise, as a statistical unit, is defined as the organizational unit of a business that directs and controls the allocation of resources relating to its domestic operations, and for which consolidated financial and balance sheet accounts are maintained from which international transactions, an international investment position and a consolidated financial position for the unit can be derived.
The population is stratified by 5-digit NAICS and by size based on revenue. The largest enterprises were selected into the sample with a probability of 100% ("Take-All method"). The remaining enterprises were selected according to the "Take-Some" method, where the probability of selection was proportional to the revenue of the enterprise for the reference year.
Sampling and subsampling
The sample is allocated to strata by revenue and selected based on probability proportional to size using sequential Poisson sampling.
Every enterprise selected for the Retail Trade Price Report survey reports a maximum of twelve products chosen based on a judgmental sampling approach. Respondents select products that are representative of their retailing activity, chosen based on either contribution to annual sales or sold regularly throughout the year.
Responding to this survey is mandatory.
Data are collected directly from survey respondents.
Following sample selection, new survey participants (respondents) are introduced to the survey through telephone calls. During this initial phase of data collection, respondents are guided through the process of selecting representative products for which prices and their characteristics (specifications) will be monitored. This process typically spans several collection cycles until respondents become conversant with the survey.
In subsequent cycles, monthly data are collected on a quarterly basis via electronic questionnaire and by mail out mail back paper questionnaires, while telephone communication (computer-assisted telephone interviewing) is used for non-response and data follow-up. Several follow-up contacts can be made, including sending out a reminder fax or letter in order to collect data.
Information about the time it takes respondents to complete the survey is collected and monitored closely. Currently, it takes respondents an average of 30 minutes to complete the survey questionnaire.
View the Questionnaire(s) and reporting guide(s) .
Error detection is conducted at the time of data collection and also during post collection processing, using a set of systematized error detection procedures to identify outliers and possible reporting anomalies. Records that fail these edits are reviewed for editing and correction when necessary or edit failure may trigger a follow-up with the respondent.
Time and effort are devoted to keeping the specifications constant such that only the pure changes in price are tracked. Some information is also collected in order to ensure, as much as possible, that the collected data correspond to the same specifications over time. This constant quality price then feeds into the Canadian System of Macroeconomic Accounts estimates of constant dollars Gross Domestic Product.
Missing data are generally estimated by a systematized imputation process. In any given period, price data may not be available for estimation. In such cases, missing data are imputed using the average price movement of remaining units within the same stratum (overall mean or targeted mean imputation method).
The price of the retail service is defined as the margin price, which is the difference between the average vendor price and the average retail price of the retail product being priced. It is not a retail selling price.
The Retail Services Price Index (RSPI) uses enterprise revenues from the Business Register and industry gross margins from Annual Retail Trade Survey (record number 2447) as its weighting sources. The weights reference period is currently 2013. Weights are updated during a sample/basket update which typically occurs every five years. The last update occurred with the release of the fourth quarter 2015 data.
Estimates are produced by calculating a weighted average of price relatives by industry, which are chained together to form an index series. The RSPI is a Laspeyres chain linked index, available at the Canada level only.
Linking of indexes
With the introduction of a new basket, historical estimates are linked to the new basket by maintaining the same historical monthly changes. This is done by calculating a link factor for each index series as the ratio of the new basket index series (2013=100) in the overlap period to the historical index series (2008=100). This link factor is applied to the historical index series to bring it up or down to the level of the new index.
The overlap period for RSPI is currently January 2014.
An in-depth assessment of quality is conducted prior to the dissemination of estimates. This assessment is based on two key elements of quality (accuracy and coherence), as defined in Statistics Canada's guidelines for the validation of statistical outputs.
The survey's data collection strategy is designed to ensure that targeted response rates are met every cycle. Analysts pay close attention to this metric and take the appropriate measures to ensure that the survey's coverage is thorough. Particular attention is also given to ensuring that sampled products or services are representative of actual transactions taking place in the market. These two activities, fundamental to the overall quality of the estimates, are done consistently.
Analysts also undertake additional validation activities every cycle to ensure the coherence of survey estimates. These activities include the analysis of price changes period-over-period and the analysis of trends at the business/company, industry, subsector and sector levels, the certification of key contributors to price change as well as the confrontation of estimates against other related data sources. Contextual analysis of survey results is also performed in light of prevailing economic conditions.
Engagements with relevant stakeholders are also undertaken periodically. Forums involving other Statistics Canada analysts, industry stakeholders and partners at other national and international statistical agencies provide valuable insight into the development and research agenda of the program.
Statistics Canada is prohibited by law from releasing any information it collects that could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.
Various transformations are applied to collected price data in the calculation of price indexes, such that it is not possible to identify the raw price data obtained from any survey participant. Confidentiality rules are also applied to price indexes that are released or published to prevent the publication or disclosure of any information deemed confidential.
Collected data are converted to price indexes and data are released as such, so that it is not possible to identify the suppliers of raw prices.
Revisions and seasonal adjustment
Data for the most recent quarter are preliminary. The previous quarter of the series data are subject to revision. The series is also subject to an annual revision released with the second quarter data of the following reference year. The indexes are not seasonally adjusted.
The statistical accuracy of this index depends on price and weight data obtained from sample surveys. Each type of input data is subject to its own errors. Processing procedures for editing and imputation are in place to ensure the quality of data.
Data-quality indicators for each index are based on measures of survey precision (standard errors of the annual percentage change), estimates of market coverage (based on the Retail Services Price Index [RSPI] sample), and the number of reported price quotes.
- Code D = Acceptable (market coverage equal to or greater than 25%, and standard error greater than 2%, but less than or equal to 5%).
- Code E = Use with caution (market coverage less than 25% or standard error greater than 5% or fewer than 5 reported price quotes).
Index series without data-quality indicators are considered to be statistically reliable.
Users are encouraged to take into account quality indicators when using RSPI data. These indicators will be updated with every second quarter data release to reflect the current data quality ratings for individual data points.
The survey achieves about an 80% response rate and uses a survey methodology designed to control errors and reduce their effect on estimates. However, the survey results remain subject to sampling and non-sampling error. Sampling errors occur when observations are made only on a sample and not on the entire population. All other errors that arise from the various survey phases are referred to as non-sampling errors.
For example, non-sampling errors can occur when a respondent provides incorrect information or does not answer certain questions; when a unit in the target population is omitted or covered more than once; when an out of scope unit is included by mistake or when errors occur in data processing, such as coding or capture errors.
A systematized imputation process is used to impute for the non-response portion of the sample, achieving an effective 100% coverage. Non-response bias is also minimized during the same process.