Human Resources Module of the Canadian Economy

Detailed information for 2021

The annual program of the Human Resources Module of the Canadian Economy (HRMCE) produces annual estimates of jobs, hours worked, and wages and salaries broken down by characteristics of the employees holding the jobs such as the age group, sex, education level, and immigration status.

Data release - January 17, 2024

• Description
• Data sources and methodology
• Data accuracy

Description

The aim of the Human Resources Module of the Canadian Economy (HRMCE) is to provide timely and reliable statistics on the human resource dimension of the Canadian economy. The Canadian Productivity Accounts provide the official macroeconomic estimates related to labour. They produce estimates of jobs, labour income, and hours worked by industry and geography. The HRMCE provides a breakdown of those key labour statistics by socio-demographic categories.

Since the HRMCE follows the same structure as the System of National Accounts, it complements and enhances the analytical capacity of other economic indicators, such as gross domestic product (GDP). It provides a snapshot of the job holders in Canada along with insights into trends over time by province and territory as well as industry. Using the HRMCE estimates, analyses can be made on the following key statistics: employee jobs, hours worked, and wages and salaries. Additional statistics such as hourly wages, or average annual wages per job can also be derived. These key statistics can be analyzed according to various characteristics of the persons holding those jobs (sex, age group, immigration status, visible minority status, and level of education). In addition, it should be emphasized that the HRMCE uses the number of jobs as its key measure of employment and not the number of people employed.

The estimates of the HRMCE are published at the 2-digit level from the North American Industry Classification System (NAICS) by province and territory. Those estimates are at the root of the thematic Human Resources Modules (HRM), such as the HRMs associated with the tourism and natural resources satellite accounts. Moreover, the HRMCE is used to provide context to the thematic HRM estimates. For example, users can compare the proportion of jobs held by immigrants in the tourism industries (based on the tourism HRM) with the total economy.

Reference period: Annual - calendar year

Subjects

• Economic accounts
• Labour

Data sources and methodology

Target population

The conceptual universe of the Human Resources Module of the Canadian Economy program encompasses:
- Ten provinces and three territories.
- Paid workers jobs and associated hours worked, wages and salaries received.
- 20 industries corresponding to the North American Industry Classification System aggregated to the 2-digit level.
- 16 characteristics related to the people holding the paid worker jobs.

Instrument design

This methodology type does not apply to this statistical program.

Sampling

This methodology does not apply.

Data sources

Data are collected from other Statistics Canada surveys and/or other sources.

These include the Censuses of population and the Labour Force Survey which are used to develop ratios of various characteristics. These ratios are used in conjunction with estimates of job counts, salaries and wages, and hours worked by industry from the Canadian Productivity Accounts.

Error detection

This methodology does not apply.

Imputation

When using information from the Census to derive the ratios, some imputations were necessary. No imputations were made to socio-demographic variables such as sex or age. However, some imputations were needed with regards to the province of work and the hours worked. When the value of the province of work was missing, the value of the province of residence one year prior to the Census was used. If that value was also unavailable, then the province of residence during the census reference week was used.

When deriving the ratios, the hours worked need to be annualized. The census provides the hours worked during a reference week. It also provides the number of weeks employed during the calendar year prior to the census. An assumption is made that the number of hours worked during the reference week is representative of all weeks worked in the previous year. Therefore, to get the annual number of hours worked, the number of hours worked during the reference week is multiplied by the number of weeks worked during the previous year. For observations where one of the variables used to derive annual hours worked is missing, imputations were necessary. The imputations made are based on known values of wages and salaries, weeks employed, and hours worked. Observations with all three variables known were put in the donor group. They were grouped based on the values of wages and weeks and/or hours. Observations where hours and/or weeks were missing were put in a receiver group based on the value of the know variable(s). The average of hours or weeks from the donor groups were then used to impute the observations where hours or weeks were missing.

No imputation was made to Canadian Productivity Accounts estimates.

Estimation

Step 1 - Extracting the data
This first step consists of extracting the variables of interest from the data sources. This is a relatively straightforward step.

Step 2 - Basic manipulations, validations, and adjustments
This second step consists of validating the data that are extracted and proceeding to make some basic manipulations and adjustments. This includes applying concordances, imputing for missing values, etc. For some data sources, such as the Labour Force Survey (LFS) and the Census, that might also include annualizing non-annual estimates.

Since the data extracted are usually at a higher level of detail than the published estimates, some inconsistencies or issues may be found. Thus, the need for a validation step. For example, there could be some missing values. In such cases, an imputation rule is developed and applied to address this issue. The validations and adjustments vary greatly from one data source to another.

It is also in this step that the appropriate weights are applied to the observations.

Step 3 - Deriving ratios
Once adjusted annual data are available, it is possible to derive the ratios of interest. The ratios are developed and applied in layers. Each socio-demographic characteristic is considered a layer.

Step 4 - Validating and adjusting ratios
Once the ratios are derived, it is important to validate them. Most of these validations stem from the comparison of ratios, identifying inconsistencies, and if needed, the appropriate adjustment is applied. For example, the ratios can be compared across geography and time. While some fluctuations are expected, it is unlikely that the profile of an industry changes significantly in between census years. When such situations happen, there are many validations in place. The ratios can be compared with the ratios from another time series to see if such a break exists there as well. If it does, then it is important to investigate what might have created it. Was there a new legislation impacting the participation of older people in the labour force? Was there an international crisis that brought large numbers of immigrants into the country? Many events can impact the labour force. If an event can explain the shift, then no adjustment should be made.

If the break observed in one series cannot be observed in another, then it is important to first validate the work done in previous steps. This ensures that the source data were properly extracted, and that concordances, imputations, and other preparatory steps were properly applied. Once it is confirmed that the data are not the source of significantly different ratios, the calculations used to derive the ratio must be reviewed. If nothing can explain the break beyond sample effects in the Census, then adjustments should be made. The adjustments made will depend on the situation. It is important to remember that the long form Censuses only have sample sizes around a quarter of the population. This means that from one Census to another, it is not always the same respondents answering the survey. Other data sources with a sample like the Labour Force Survey can denote a similar phenomenon. This can lead to misleading results. Therefore, when trying to measure very precise categories, or small industries in smaller geographies, the impact of the sample fluctuations can necessitate adjustments.

Step 5 - Deriving time series of final ratios
Once the ratios have been derived and validated for all data sources, it is important to calculate ratios for any missing years in the timeseries. Up to this point, only ratios for Census reference years have been obtained and the intercensal years need to be estimated. Other sources such as the LFS already have a full-time series, and no gaps will need to be filled.

Ratios derived for census years are used as anchor points. By default, the interpolation in between is presumed to be a linear relationship. In other words, the ratios will grow or decline at a uniform rate depending on the difference between the two anchoring census years. This also means that the further away a reference year is from an anchor point, the less influence this anchor point has on the value.

In some situations, such as a change in legislation or a major world event having acute impacts, a linear interpolation may be inappropriate. In such situations, indicator series are used for the interpolation. Indicator series are used to validate and, when needed, interpolate between census years. Currently, LFS is used. This means that the use of indicator series to interpolate is not systematic. For example, in the event of a new legislation introduced between census years, significantly impacting the participation of women in the labour force, ratios from LFS could be used to properly introduce the break in the time series.

For years beyond the latest census, the ratios from the latest census are kept constant. However, ratios from LFS are derived and if those ratios show major break(s), then further validation is done. If needed, the ratios are adjusted to reflect the break.

Step 6 - Applying the ratios to Canadian Productivity Accounts estimates
Once the ratios have been validated and adjustments implemented, a dataset is created with the most detailed ratios available. Those ratios are then applied to the Canadian Productivity Accounts estimates. The ratios are applied by geography and industry.

Quality evaluation

The Human Resources Module of the Canadian Economy is based on a wide array of related and comparable data. Labour-related estimates generally follow the trends of Canadian Productivity Accounts while the socio-demographic estimates follow the trends observed in the Census.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which 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.

Revisions and seasonal adjustment

Revisions to the Human Resources Module of the Canadian Economy (HRMCE) are mainly due to revisions to the underlying source data, namely the Canadian Productivity Accounts (CPA) and the introduction of a new Census. Statistical revisions are carried out regularly in the CPA and more broadly in the Canadian System of Macroeconomic Accounts (CSMA), to incorporate the most current information from censuses, annual surveys, administrative statistics, public accounts, and other sources.

Periodically, more comprehensive revisions are conducted. These provide an opportunity to improve estimation methods, incorporate improved data sources, introduce conceptual changes and adopt new international standards into the CSMA. Such revisions usually do not impact the ratios of each socio-demographic characteristic in the HRMCE but rather the values of the labour-related variables to which those ratios are applied.

Ratios are revised when a new census is available.

Seasonal adjustment is not necessary given that the HRMCE by industry at the provincial and territorial level is only available on an annual basis.

Data accuracy

No direct measures of the margin of error in the estimates can be calculated. The quality of the estimates can be inferred from an analysis of revisions and from a subjective assessment of the data sources and methodology used in the preparation of the estimates.