The impact of Business Development Bank of Canada’s loan program for business acquisitions on the performance of acquiring firms. (001-2024)

Purpose: The goal of the project is to assess the impact of a client acquisition on the acquirer’s business performance. An increased understanding of the impact of business acquisition will help the Business Development Bank of Canada (BDC) and entrepreneurs better understand the dynamics of business acquisition and its value as a growth strategy. This will also enable the BDC to better support their clients in the future. 

To undertake this project, a list of firms receiving loans from the BDC between 2010 and 2019 for the purpose of acquiring another firm will be linked to Statistics Canada’s Business Register and the National Accounts Longitudinal Microdata File. The resulting linked data will be used to compare the performance of acquiring firms to a similar sample of firms.

Output: Statistics Canada will provide the following output to the BDC:

1. A linkage report showing the percentage of BDC supported businesses that matched with the Business Register and the National Accounts Longitudinal Microdata File. 
2. A report including a description of the methods adopted, non-confidential summary tables and non-confidential tabulations based on the estimation results (i.e., from comparing the outcomes of BDC-supported businesses to non-assisted businesses).

Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada.

Firm technology adoption, its determinants, and impacts (003-2023)

National Social and Affordable Housing Database linkage to administrative tax and immigration, Census and Canadian Housing Survey (002-2024)

Purpose: The Canada Mortgage and Housing Corporation (CMHC) has developed a research plan to understand characteristics of persons and households living in social and affordable housing. The statistics produced will help inform the evaluation and further the implementation of the National Housing Strategy. 

As part of this plan, a linkage between the National Social and Affordable Housing Database (NSAHD) to multiple data sources held by Statistics Canada will be performed. Combining information from several sources (tax, employment, immigration, housing and the Census) will contribute income and sociodemographic information, as well as housing indicators to the linked database for those living in NSAHD units. Analytical work in this domain is expected to help researchers, the general public and government to understand and to improve housing policies or programs.

Output: The analytical file, once personal identifiers are removed, along with anonymized linkage keys will be placed in the Research Data Centre (RDC) for access by CMHC as Statistics Canada deemed employees for the first four months. After this period, the files will be made available in the RDC network and access will be granted following the standard RDC approval process. Only non-confidential aggregated statistics and analyses conforming to the confidentiality provisions of the Statistics Act and any applicable requirements of the Privacy Act, will be released outside of Statistics Canada.

Linking the Level of Supervision and Official Language Variables to the ESDC Employee Wellness Survey (ESDC EWS) (004-2023)

Microdata linkage for exploring the socio-economic outcomes of individuals who have received services from British Columbia Ministry of Children and Family Development (004-2024)

Purpose: This data linkage is being carried out on a cost-recovery basis by Statistics Canada for British Columbia Ministry of Children and Family Development (BC MCFD). 

The purpose of the data-linking initiative is to analyze the socio-economic outcomes of a cohort of individuals who have received BC MCFD services, by linking BC MCFD data with Statistics Canada data for the same cohort of individuals that BC MCFD would otherwise not have access to. The inclusion of a vast array of BC MCFD data gathered through service delivery, linked with Statistics Canada economic data will allow BC MCFD analysts to explore connections and variables that could not be gleaned from BC MCFD data alone.

The outputs will allow BC MCFD to evaluate the efficacy of the services provided by BC MCFD and strengthen the assessment of services provided by BC MCFD and support future policy development.

Output: Statistical tables will be prepared at the aggregate level while meeting the confidentiality provisions of the Statistics Act. Only non-confidential aggregate statistics such as counts, percentages and sums will be provided to the client and only after appropriate suppression has been applied. A total of ten tables will be provided to the client, one for two variables of interest, giving information for data that will be used by the BC MCFD analysts. A short report on the methodology and results will also be delivered to the client. An analytical file will be made available for the exclusive access of the BC MCFD analysts as deemed employees via a Statistics Canada secure access point follow the standard approval process.

Surrey Opioid Data Collection and Community Response Project: Linking Surrey Opioids data with Census, income, health and immigration data to generate privacy-enhancing synthetic data (005-2023)

Custom data request: Student Work Placement Program linked to the Postsecondary Student Information System (PSIS), T1 Family Files (T1FF), and the 2021 Census of Population long form (006-2024)

Purpose: The Student Work Placement Project (SWPP) project will provide Employment and Social Development Canada with policy-relevant information that will allow them to evaluate the success of the SWPP. The project builds upon the information currently available to ESDC by including data pertaining to education and labour market integration and outcomes of postsecondary graduates who had a work-integrated learning placement through the SWPP, such as earnings and employment. The project will also allow disaggregation by sociodemographic characteristics with a focus on underrepresented groups, where possible. 

Output: Statistics Canada will provide aggregate, non-confidential custom tables to Employment and Social Development Canada. Only non-confidential aggregate statistics and analysis conforming to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada. If the client later wishes to access the linked source microdata files in the Research Data Centres to perform their own analysis, the source files used in the linkage with no direct identifiers will be made available to the client as Statistics Canada deemed employees through the Research Data Centre (RDC) program using the normal RDC project approval procedures.

Linkage of Canadian Employer-Employee Dynamics Database to demographic data to analyze Indigenous-owned businesses in Canada (007-2024)

Linkage of Canadian Employer-Employee Dynamics Database to demographic data to analyze Indigenous-owned businesses in Canada (007-2024)

Purpose: The purpose of the project is to link the Canadian Employer-Employee Dynamics Database to select variables from the Census of population, the National Household Survey, and the Canadian Community Health Survey to analyze Indigenous owned businesses. This data linkage will help Statistics Canada, and other Canadian government departments analyze data related to Indigenous-owned businesses in Canada to assist in policy making decisions.

Output: The resulting analytical file will be used to update CODR tables 33-10-0631-01: Private enterprises by sex and Indigenous identity of ownership, province or region and enterprise size, 33-10-0632-01: Private enterprises by sex and Indigenous identity of ownership, age group of primary owner and enterprise size on the Statistics Canada Website. The data linkage will also allow Economic Analysis Division to address several requests from other government departments related to Indigenous-owned businesses in Canada. Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada.

Diversity characteristics of full-time and part-time teaching staff at Canadian Universities. (009-2024)

Diversity characteristics of full-time and part-time teaching staff at Canadian Universities. (009-2024)

Purpose: The purpose of this microdata linkage project is to gather diversity related characteristics from other data sources and integrate them with administrative data on full-time and part-time/contract teaching staff in Canadians universities. This combined information will be used to create aggregate statistics on diverse populations within Canadian universities. These statistics may serve to report on the representation of diverse population groups among teaching personnel in Canadian universities and help measure progress toward an academic community that accurately reflects the Canadian population.

Output: The final microdata output from this linkage will be anonymized. All access to the linked microdata file will be restricted to Statistics Canada employees (including Statistics Canada deemed employees) whose work activities require access. Only non-confidential aggregate statistical outputs and insights that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada. Aggregate findings at the institutional level will be presented in a feasibility study, internal and external reporting documents, and presentations.

Diversity characteristics of full-time and part-time teaching staff at Canadian Universities. (009-2024)

Microdata Linkage for the Canadian Legal Industry Earnings Tabulation. (011-2024)

Purpose: The goal of this project is to provide the Seventh Quadrennial Judicial Compensation and Benefits Commission at the Department of Justice (DOJ) with statistics about the level and distribution of lawyers’ private practice compensation levels. The outputs will be aggregated statistical tables that will allow DOJ to evaluate and determine competitive compensation for federally appointed judges.

To undertake this project, Statistics Canada will compile and join the data for law professional corporations (LPCs) and legal partnerships with the corresponding Canadian Employer-Employee Dynamics Database and T5013 income tax fillings, for the latest year available, 2021. The work will focus only on legal corporations and partnerships under the NAICS code 541110 - Offices of lawyers. No other NAICS code will be considered.

Output: Aggregated statistical tables will be made available only to the Department of Justice, the client for this project. These will consist of tables on total incomes of private sector self-employed lawyers, law professional corporations (LPCs) and legal partnerships for the most recent year the income tax source data is available. Statistics calculated on professional income of self-employed lawyers will include the number, mean, median, standard deviation, and 75th percentile values, by age and Census Metropolitan Areas (CMAs) groups and the low-income cut-off of $80,000. Same statistics will also be calculated for legal partnerships (number, size, count of partner types), and the income generated and distributed to partners.

Linkage of the Census of the Population to the Longitudinal Administrative Databank (LAD) for Sociodemographic and Economic Analyses. (013-2024)

Purpose: This project is intended to enhance the ability of researchers and policy makers to address pressing questions in a timely manner, to help them to measure the outcome of policies and programs already in place, and possibly put in place new measures or improved public policy decisions that benefit Canadians. It allows for the creation of two linkage key files between the LAD, and the 2016 and 2021 censuses. These two files will be available in Statistics Canada’s research data centres (RDCs) to researchers with approved projects. It will fill important data gaps by providing a better understanding of the various socioeconomic backgrounds of Canadians and the long-term incomes (life cycle, income trajectories over time) and will facilitate research and recommendations on various challenges for the daily lives of Canadians.

Output: The final output will be two evergreen linkage key files updated annually when the LAD is updated: (1) linkage keys of the LAD to the 2016 Census and (2) linkage keys of the LAD to the 2021 Census. Each linkage key file will contain two fields representing non-confidential record identifiers: the first field will be the LAD identification number and the second will be the census identification number. These two fields will allow Statistics Canada deemed employees to link the LAD data with the relevant census long-form data. These linkage key files will be made available via Statistics Canada Secure Access Points (such as Research Data Centers), and access will be granted to Statistics Canada deemed employees following the standard approval process. 

Only non-confidential aggregate statistics and analyses conforming to the confidentiality provisions of the Statistics Act and any applicable requirements of the Privacy Act will be released outside Statistics Canada.

All products resulting from the linkage of the census and the LAD using the linkage key files covered by this linkage project will be disseminated in accordance with Statistics Canada's policies, guidelines and standards. Outputs from this linkage could include a wide range of analysis and standard data tables, as well as custom tabulations. Any data circulated outside of RDCs or Statistics Canada will be vetted according to the appropriate confidentiality rules in place for the products prior to circulation.

Linking the Nursing and Residential Care Facilities Survey (NRCFS) with selected employer-employee information to analyse the working conditions of long-term care workers in Canada. (016-2024) 

Purpose: The proposed linkage will bring together the Nursing and Residential Care Facilities Survey (NRCFS) with selected files from the Canadian Employer-Employee Dynamic Database (CEEDD) to support research on the working conditions of the long-term care (LTC) workforce. The objective is to provide insights on the poorer economic and health outcomes experienced by workers in the LTC sector before and during the COVID-19 outbreak. 

Output: An analytical file, with no personal identifiers, will be produced from this linkage process. Methodological and analytical findings resulting from these linked data may be used to prepare research findings for: publication in analytical reports or peer-reviewed scientific journals; presentation at conferences, workshops, or meetings; or tabular data or indicators for release on Statistics Canada’s website. Confidentiality vetting rules will be developed and applied in concordance with rules currently in place. Only non-confidential aggregated data and analytical products that conform to the confidentiality provision of the Statistics Act and any applicable requirements of the Privacy Act will be released outside of Statistics Canada. The linkage proposal is to address a current need to answer a specific research question and is not designed to provide a permanent linkage of CEEDD with NRCFS 

Linkage of Value-Added Agricultural Programs administered by Alberta’s Ministry of Agriculture and Irrigation (AGI), under the Canadian Agricultural Partnership to the Business-Linkable File Environment (B-LFE) to estimate the effect of these programs on the financial performance of the recipients. (017-2024)

Purpose: The main objective of this project is to estimate the effect of the Alberta Ministry of Agriculture and Irrigation (AGI) Value-Added Agricultural Programs:

• Emerging Opportunities and Food and Agri-Processing (EOFA)
• Emerging Opportunities Program (EOP)
• Products to Market Program (PTM)
• Value-Added Products to Markets Program (VAPM)
• On Farm Value-Added Program (VAPM)
• Value-Added Program (VAP) 

under the Canadian Agricultural Partnership on the financial performance of the recipients. The initial phase will involve preparing profiles of program participants and comparing them to eligible non-participants using the variables in the Business-Linkable File Environment (B-LFE) and the Diversity and Skills Database (DSD). The second phase will involve the use of matching to build a control group, and the use of regression models to study the effect of the programs on the financial performance of the recipients (e.g., revenues).

Output: The output will be in the form of summary tables and a fixed effects model which will examine the economic performance of businesses that received financial support from Alberta’s Ministry of Agriculture and Irrigation (AGI) Value-Added Agricultural Programs under the Canadian Agricultural Partnership to non-supported businesses. The linked Alberta Ministry of Agriculture and Irrigation (AGI) list of businesses will be housed at Statistics Canada’s Centre for Special Business Project (CSBP).

A research dataset will be produced, and the full integrated database will be used by a Statistics Canada deemed employee research team to produce an analysis and custom-designed tables of non-confidential aggregate statistics for Agriculture and Agri-Food Canada. The output will be analysed for confidentiality by CSBP employees. The output of this project will not be sent to the Canadian Centre for Data Development and Economic Research (CDER). 

Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada. 

The Role of Environmental Policies in Manufacturers' Energy Price Forecasts and Production Choices. (019-2024)

Purpose: This research project aims to measure Canadian manufacturers' responses to changes in environmental policies. The findings may contribute to public debate by evaluating the economic and environmental impacts of these policies. Additionally, it may help policymakers assess the effectiveness of environmental policies from both economic and environmental to better evaluate existing environmental policies and design better ones in the future.

Output: The anonymized data files will be made available via Statistics Canada Secure Access Points (such as Research Data Centres (RDC)), and access will only be granted to Statistics Canada deemed employees following the standard approval process. Researchers from the University of British Columbia and the Consumer Financial Protection Bureau, will use the linked anonymized data files to produce a study for submission to a peer-reviewed journal. The anonymized data files will remain in the RDCs for future projects.  Only non-confidential aggregate data that conforms to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada.  

Only non-confidential aggregate data that conforms to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada.

Indigenous Community Corrections Initiative Re-Contact Study (020-2024)

Purpose: The purpose of this project is to assess the effectiveness of the Indigenous Community Corrections Initiative (ICCI) in reducing re-contact with the criminal justice system. The ICCI supports alternatives to incarceration and reintegration projects responsive to the unique circumstances of Indigenous people in Canada. 

Output: Linked analytical files and anonymized linking keys will be used by Statistics Canada to produce non-confidential aggregate statistical tables and analytical reports for Public Safety Canada.

The impact of proposed government taxation on capital gains. (021-2024)

Purpose: The Parliamentary Budget Officer (PBO) proposes a study to examine the distributional impacts of the proposed change in the inclusion rate of realized capital gains. To facilitate PBO’s analysis, Statistics Canada will link a set of selected variables from the Longitudinal Administrative Data (LAD), T1 family files (T1FF), T2 corporation tax files as well as T4 employment remuneration files and T5013 partnership tax files, covering the years from 2013 to 2022. The results of this study will enable PBO to better understand the distribution of households most affected by the proposed tax changes to capital gains and to provide evidence based, independent and non-partisan advice to parliamentarians on the proposed change, thus promoting greater budget transparency and accountability and supporting the development of fair and efficient tax policies in the best interests of Canadians.

Output: This microdata linkage will conform to Statistics Canada’s policies and standards and business and personal identifiers will be removed from the analysis file once linkage is complete. PBO researchers as Statistics Canada deemed employees will access the analytical file via Statistics Canada Secure Access Points to conduct proposed distribution analysis on realized capital gains. Only non-confidential analysis that will not result in the identification of an individual, business or organization will be released outside of Statistics Canada. Once the analysis is complete, the PBO will publish a report that will be made publicly available on its website and tabled in both Chambers of Parliament.

The Impact of Farm Credit Canada Client Services on Firm Performance. (026-2024)

Purpose: To assess the impact of Farm Credit Canada's (FCC) Customer Services on its customers, FCC is requesting Statistics Canada's Economic Analysis Division (EAD) to conduct an impact assessment study. For the analysis, EAD will link the database of approximately 100,000 FCC customers from 2015 to 2022 to the Business Register (BR) and the Canadian Employee and Employer Dynamic Database (CEEDD), which includes incorporated and unincorporated businesses. EAD will use the linked database for this project to create a sample of matched businesses that have the same characteristics as FCC customers but did not use FCC services during the study period. The matched sample will enable EAD to assess the extent to which the performance of FCC customers differs from that of comparable companies that have not received FCC assistance.

Output: 

1. EAD will prepare a report for FCC on the linkage rate and quality of the linkage between the FCC database and the Business Register. This is a report describing the rate of companies in the FCC database that can be found, with or without precision, in the Business Register, according to matching criteria such as company name, address, etc. The report will also describe the quality of the linkage between the FCC database and the Business Register.
2. EAD will prepare tables and a brief document for FCC describing the methodology and strategy for comparison (i.e. FCC's program in relation to others).
   Access to the linked microdata will be restricted to Statistics Canada employees whose assigned work duties require such access. 
   Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada

Bridging the Gap: Analyzing Risk Capital Interactions for Canadian Startups (027-2024)

Purpose: The goal of the project is to investigate Business Development Bank of Canada (BDC) Capital’s role in the ecosystem of start-ups. By leveraging BDC’s unique position within Canada’s entrepreneurship ecosystem, this project will provide Innovation, Science and Economic Development Canada (ISED) a more comprehensive understanding of the funding landscape for startups. In particular, the project will investigate the impact of BDC Capital on company performance by comparing the growth, innovation, and competitiveness of BDC-backed firms with those receiving capital from other sources, highlighting BDC Capital's unique contributions to small and medium-sized enterprises success.

To undertake this project, ISED will provide BDC Capital client data alongside complementary information from private risk capital sources and government programs to link to Statistics Canada's Business Register and National Accounts Longitudinal Microdata File. The resulting linked data will be used to compare outcomes of firms supported by BDC Capital to a similar sample of firms.

Output: Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada. The information will be presented in the form of tables of regression results and summary statistics. The anonymized analytical file will be accessed by Statistics Canada deemed employees at the Statistics Canada Federal Research Data Centre.

Linkage of Canadian Employer-Employee Dynamics Database to Country-by-Country Report data to analyze wages, productivity, and profit shifting of Canadian Multinational Enterprises (028-2024)

Purpose: The goal of this project is to assess the extent to which Canadian multinational enterprises (MNEs) engage in profit shifting and its impact on wages, productivity, and income inequality. By linking the Canadian Employer-Employee Dynamics Database with Country-by-Country Report data from 2016 to 2022, this study will examine the following:

1. What happens to the corporate income of Canadian MNEs that is not spent on taxes?
2. Are workers' wages higher at MNEs engaged in profit shifting? 
3. How is profit shifting related to wage inequality? Are wage gains greater for high-skilled workers in high-income occupations compared to low-skilled workers at MNEs engaged in profit shifting?
4. Are workers at profit shifting MNEs more productive?
5. Are profit shifting MNEs more productive than other similar corporations in Canada?

Output: The linked data will be used to write a research paper including a description of the methodology, non-confidential summary tables and non-confidential tabulations of the estimation results. This research will be disseminated through external publications (academic journals), Statistics Canada’s internal publications, presented at conferences and potentially to other federal government agencies.

Security measures for the linking keys, identifiers and linked analysis file will conform to Statistics Canada’s policies and standards. All direct business and individual identifiers will be removed from the analysis file once linkage is complete.  

Determining the business size of suppliers to the Government of Canada in order to generate estimates of the dollar amount of procurement by Small Medium Enterprises by contract size, and department for fiscal years 2021-2024. (029-2024)

Purpose: The goal of the microdata linkage is to determine the business size of suppliers to the Government of Canada. Understanding business size of suppliers is critical to support evidence-based decisions in the context of a budget proposal to target a certain level of SME procurement that is being considered for inclusion in the 2024 Fall Economic Statement. 

Output: The aggregate tables produced for Innovation, Science and Economic Development Canada (ISED) and will be shared publicly after STATCAN Confidentiality business rules are applied.

The microdata linkage results will be made available via Statistics Canada Secure Access Points (such as RDCs), and access will only be granted to Statistics Canada deemed employees following the standard approval process.

The microdata linkage results have been requested by ISED to be made available in the Federal Research Data Centre (FRDC). 

Food insecurity and income using the Canadian Income Survey (CIS) and T1 Family Files (T1FF) (030-2024)

Purpose: The purpose of this project is to better understand what accounts for the persistently high and accelerating prevalence of food insecurity despite a falling poverty rate in order to inform the development of effective policy responses to reduce household food insecurity.  This project aims to determine the extent to which the misalignment of poverty and food insecurity rates through detailed exploration of the different reference periods for the measurement of household food insecurity and income in CIS, and the presence of households’ assets and debt.  It also aims to determine the contribution of households’ recent experiences of negative income shocks to their current probability of food insecurity, and identify the key drivers of those shocks to inform public policy interventions.

As part of this cost-recovery project, Statistics Canada will be linking the CIS to a subset of variables from the T1 Family File.  Access to the merged dataset will made be through the RDCs to employees of the PROOF research program of the University of Toronto. Only vetted tables with no identifiable information will be transferred to the researchers.

Output: The final linked analytical files without personal identifiers will be made available within Statistics Canada secure access points. Access will only be granted to researchers following the standard approval process. Research reports and presentations will be generated from the analysis files. Only non-confidential aggregate statistics and tables conforming to the confidentiality provisions of the Statistics Act and any applicable requirements of the Privacy Act will be released outside of Statistics Canada.

Microdata linkage for assessing the impact of Agriculture and Agri-Food Canada’s Trade Show program on business performance. (031-2024)

Purpose: This data linkage is being carried out on a cost-recovery basis by Statistics Canada for Agriculture and Agri-Food Canada. The goal of this project is to conduct a quantitative impact assessment of the Trade Show Program. In particular, the project seeks to compare the outcomes of businesses that participate in the program to those of similar businesses that did not participate. The ultimate objective is to probe the relationship between participation in the Trade Show Program and firm outcomes related to business performance and workforce diversity.

Output: An anonymized analytical file will be made available through Statistics Canada Secure Access Points (such as research data centres), and access will be granted to Statistics Canada deemed employees (including client) following the standard approval process.
The release of the vetted outputs will be done by Statistics Canada staff. The vetted output will be presented in the form of descriptive statistics and regression results. Descriptive statistics will provide an overview of the characteristics of program participants compared with non-participants, while regression analysis will be used to estimate the causal impact of Trade Show Program participation on business outcomes.

Only non-confidential aggregate statistical outputs and analyses that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada.

Assessing the impact of Canada Emergency Response Benefit on Opioid Overdose Crisis. (032-2024)

Purpose: The objective of this project is to assess the impact of the Canada Emergency Response Benefit (CERB) as a possible contributor to the spike in opioid overdoses during the COVID-19 pandemic. The researchers will be exploring the hypothesis that the unexpected, temporary increase in available cash provided through the CERB may have had a detrimental impact on a vulnerable segment of the population with a prescription history of opioids or mental illness leading to an increased use or addiction to opioids. 

The study will include the new linkage of the Canadian Community Health Survey (CCHS) to the CERB and Employment Insurance Status Vector (EISV).

Output: Only aggregate data that conform to the confidentiality provisions of the Statistics Act will be released outside of Statistics Canada. Analytical datasets will be placed in the Research Data Centres (RDCs) and access will be granted following the standard RDC approval process. The source datasets will be anonymized and will respect variable restrictions in effect for the source datasets. Major findings will be used to create research papers for publication in peer-reviewed journals and presentations at workshops and conferences.

Date: March 2024

Program manager:

• Director, Centre for Social Data Integration and Development
• Director General, Social Data Insights, Integration and Innovation

Reference to Personal Information Bank (PIB):

Personal information collected through the Survey Series on People and their Communities is described in Statistics Canada's "Special Surveys" Class of Personal Information. The Personal Information Bank refers to information collected through Statistics Canada's ad hoc surveys, which are not part of the regular survey taking activities of the Agency. They cover a variety of socio-economic topics including health, housing, labour market, education and literacy, as well as demographic data.

"Special Surveys" Class of Personal Information (Bank number: StatCan PPU 016) is published on the Statistics Canada website under the latest Information about Programs and Information Holdings chapter.

Description of statistical activity:

Under the authority of the Statistics Act^Footnote1, Statistics Canada is conducting Wave 5 of the voluntary Survey Series on People and their Communities^Footnote2 – Social Cohesion and Experiences of Discrimination (SSPC-SCED). This iteration of the survey will be collecting new content about the respondents’ future outlook in Canada, feelings towards others and experiences of discrimination with a focus on immigrants’ and racialized peoples’ experiences, in addition to information about self-perceived general and mental health and financial well-being. This sensitive content, in conjunction with the longitudinal nature and linkages with responses in the previous waves and the Census, warrants the development of this supplement to the Generic PIA in order to assess, describe and mitigate any potential associated privacy risks.

Reason for supplement:

As with the Supplement to the Generic PIA for the Survey Series on People and their Communities – Participation and Experiences in Community Sports (Wave 4), while the Generic Privacy Impact Assessment (PIA) addresses most of the privacy and security risks related to statistical activities conducted by Statistics Canada, this supplement was developed to address the collection and use of sensitive new content in Wave 5 of this survey series that includes information about respondents’ feelings towards others and their experiences with unfair treatment, racism and discrimination, combined with sociodemographic and other information collected from prior survey series waves. As is the case with all PIAs, Statistics Canada's privacy framework ensures that elements of privacy protection and privacy controls are documented and applied.

Necessity and Proportionality

The collection and use of personal information for Wave 5 of the Survey Series on People and their Communities – Social Cohesion and Experiences of Discrimination (SSPC-SCED) can be justified against Statistics Canada’s Necessity and Proportionality Framework^{Footnote 3}:

1. Necessity:

   The fifth wave of the SSPC addresses issues such as self-perceived general and mental health, financial well-being, satisfaction with time use and satisfaction with the local environment, which contribute to several key indicators on the well-being of people in Canada in Canada’s Quality of Life Indicator Framework and will enable the federal government to identify future policy priorities and build on previous actions to improve evidence-based decision-making.

   The fifth wave also includes several questions related to social cohesion and other similar concepts (e.g., social inclusion and social connectedness) that indicate the relative strength or weaknesses of the bonds between societal members with emphasis on inter-group or between-group interconnectedness. The between-group aspect of social cohesion is the focus of the Conference of European Statisticians’ (CES) Task Team on Social Cohesion. Cohesion represents the act or state of sticking together tightly, highlighting the bond or the distance between two units. In social terms, the distance between groups may be defined in various ways, such as trust of out-of-group members (members of different sociodemographic/economic groups than the individual), feelings toward out-of-group members, and opinions and values relative to out-of-group members. The groups between which social distance is estimated can be identified and defined in various ways:

   1. Social and demographic characteristics, such as ethnicity, language, religion and sexual orientation.
   2. Political or opinion-based lines, such as opinions/positions on issues such as climate change, abortion or vaccination, as well as traditional measures such as political affiliation and ideology.
   3. Economic terms, such as economic grievances or marginalization articulated in terms of the differentiation and opposition of ‘the people’ to institutions, elites, or ‘the establishment’.

   The questions build on the In-Depth Review of Social Cohesion prepared for the CES and endorsed at its plenary meeting in June 2023. Measures of interpersonal and between-group ties will complement efforts to disaggregate population groups under the Disaggregated Data Action Plan (DDAP) to fill the knowledge gap on the distance between groups. Interpersonal and between-group ties will be central to better understanding social cleavages as they relate to discrimination as part of Canada’s Anti-Racism Strategy to strengthen impact measurement and performance reporting.

   Finally, the fifth wave will aid in the development of a conceptual framework on racism and discrimination by Statistics Canada’s Centre for Gender, Diversity, and Inclusion Statistics in support of Canada’s Anti-Racism Strategy and the DDAP. This conceptual framework will assist in operationalizing the measurement of racism and discrimination and identifying relevant indicators and new questions to include in future Statistics Canada social surveys. Questions will examine interpersonal or individual racism including subtler forms of racism which have been absent from existing Statistics Canada data holdings and surveys.

2. Effectiveness - Working assumptions:

   The questionnaire content in Wave 5 was developed in collaboration with subject matter experts from Statistics Canada’s Centre for Social Data Insights and Innovation, Social Analysis and Modelling Division, Centre for Gender, Diversity and Inclusion Statistics and Diversity and Sociocultural Statistics division. Statistics Canada's Questionnaire Design Resource Centre (QDRC) provided input and feedback for the survey questionnaire after conducting a series of qualitative testing interviews with individuals who are representative of the sample. Overall, QDRC participants reacted positively, and the topics covered in the survey were of general interest. There were no major issues observed and there were no concerns about the proposed topics being too personal or sensitive.

3. Proportionality:

   Questions were developed with subject matter experts to fill specific data gaps and support policy and program development. While the sensitivity of these questions on experiences of unfair treatment, racism, and discrimination may cause distress for some respondents, the results from this survey are expected to provide data to help address issues related to the subtler, overlooked form of racism, “minor racism” or “everyday racism” that is not captured in other Statistics Canada social surveys. This will aid in the development of the conceptual framework on racism and discrimination at Statistics Canada and potentially lead to policy that results in a reduction in the occurrence and prevalence of racism and discrimination in Canada.

4. Alternatives:

   The same considerations for alternatives examined in the Supplement to the Generic PIA for the Survey Series on People and their Communities – Participation and Experiences in Community Sports (Wave 4) apply to SSPC-SCED.

   Specifically, alternative sources for the data to be collected are not available in any other Statistics Canada surveys or administrative data holdings. Alternative designs to a panel survey were considered for collecting this information, including a traditional survey, however, the response burden a traditional survey would place on this population would have likely compromised the ability to achieve response rates required to produce reliable statistics. Ultimately, a panel survey was deemed the optimal collection method to produce all the required disaggregated data for analysis without overburdening respondents.

   Additionally, although the General Social Survey (GSS)^{Footnote 4} program collects similar information, it does not address the analytical needs of SSPC-SCED. Moreover, the GSS does not oversample for a sufficient level of disaggregation needed to identify disparities between racialized and non-racialized groups and immigrants and non-immigrants on these topics.

Mitigation factors:

The overall risk of harm to the survey respondents has been deemed manageable with existing Statistics Canada safeguards that are described in Statistics Canada’s Generic Privacy Impact Assessment, including measures outlined in the Supplement to Statistics Canada’s Generic PIA for the Survey Series on People and their Communities – Participation and Experiences in Community Sports (Wave 4):

Resources

Some questions contained in the SSPC-SCED are considered sensitive as they relate to an individual's experience or witnessing of unfair treatment, racism, or discrimination. Relevant resources will be provided to respondents in the information for survey participants and interviewers will be instructed to provide the resources should a respondent experience any distress. Additionally, respondents will be provided with the possibility to skip questions.

Consent

Individuals selected for the SSPC-SCED will be informed in the invitation email and in the electronic questionnaire or during the telephone interview that their participation is voluntary before being asked any questions.

Confidentiality

Variables that directly identify respondents will be separated from the data files in the first stage of data processing and placed in a secure location with controlled access. Variables that might indirectly identify respondents are examined and modified as necessary in order to protect the privacy and confidentiality of respondents. Individual responses will be grouped with those of others when reporting results. Individual responses and results for very small groups will never be published or shared with government departments or agencies. Careful analysis of the data and appropriate additional disclosure controls will be performed prior to the publication and sharing of data (in RDCs, etc.) to ensure that marginalized and vulnerable communities are not disproportionally impacted.

Data linkage

The linkage of SSPC-SCED data with other sources of information will be used in statistical studies to evaluate data quality and the impact of non-response, to improve and assist with data editing and imputation, and for direct replacement of data in presence of partial non-response when the quality is deemed appropriate. The linkage files will be used only within Statistics Canada for methodological research, development and processing.

Security measures for linkage keys and administrative files respect the policies, directives and guidelines for information technology security at Statistics Canada. When linkage is required, it is done using anonymized statistical identifiers ("linkage keys") and, as a result, no linked file contains personal identifiers such as name, phone number and address (excluding postal code). These anonymized statistical identifiers are used to link to other sources of information for statistical purposes only. The personal identifiers obtained are removed from the rest of the information and securely stored with restricted access with an approved operational requirement to access them, and whose access is removed when no longer required.

Transparency

Prior to the survey, respondents will be informed that the survey is voluntary, and of the survey purpose, allowing them to decide whether they wish to participate. This information will be provided in the invitation email, the electronic questionnaire, by interviewers, as well as on the SSPC-SCED website.

Additionally, although no concerns were raised by participants in focus groups during the testing of the questionnaire, respondents will be provided with a warning about the sensitive nature of the experiences of discrimination questions, the ability to skip any questions, as well as with information on relevant resources^{Footnote 5}. Additionally, interviewers are trained to provide relevant resources to respondents who show signs of distress, and to offer to skip sensitive questions.

Conclusion:

This assessment concludes that with the existing Statistics Canada safeguards including those listed in the Supplement to the Generic PIA for the Survey Series on People and their Communities – Participation and Experiences in Community Sports (Wave 4), any remaining risks are such that Statistics Canada is prepared to continue to accept and manage the risk.

The net cash income estimates in the Net Farm Income, Agriculture Economic Statistics, are the official Statistics Canada estimates. Other estimates of net cash income (farm cash receipts minus operating expenses) can be derived from the data from the Agriculture Division of Statistics Canada, namely data from the Agriculture Taxation Data Program (ATDP), the Farm Financial Survey (FFS) and the Census of Agriculture. It is important to understand coverage and conceptual differences when comparing data collected for different purposes. Adjustments are required to make these estimates comparable.

1. Conceptual differences

Note that the text below is based on the differences that existed for the 2020 reference year.

a) Net Farm Income: Agriculture Economic Statistics

The Agriculture Economic Statistics (AES) estimates include all agricultural businesses. These data are not available by farm type, sales classes, and sub-provincial regions or at the micro level.

Receipts and expenses are estimated by calendar year. They are recorded on a cash basis when the money is paid to, or disbursed by, the farmer.

AES receipts and expenses exclude: income earned from non-agricultural use of the farm (e.g., income from tourism activities on farm); income that farm operators or their families receive from other sources (e.g., wages and salaries from non-agricultural activities, and investment income); revenue or expenses from the sale or purchase of farm capital (real estate, machinery and equipment), although the interest paid on these purchases is included as an expense; capital payments where funds do not relate to current production and transfer payments (such as training allowances) directed to individuals; unlike the ATDP, FFS and Census of Agriculture, AES estimates exclude farm-to-farm transactions, unless they occur across provincial or national borders. Within a province, sales from one farm are considered an expense to another, thus offsetting each other.

b) Agriculture Taxation Data Program

The Agriculture Taxation Data Program (ATDP) is an annual census of unincorporated and incorporated tax-filer records designed to estimate a range of financial agricultural variables.

The target population consists of all unincorporated, incorporated and communal farms in Canada. For statistical purposes, the estimates presented cover both unincorporated farms and communal farming organizations with total farm operating revenues equal to or greater than $10,000 as well as incorporated farms with total farm operating revenues of $25,000 and over.

The estimates are published on a calendar year basis but no attempt is made to adjust data from agricultural corporations reporting data on a fiscal year that may not coincide with a calendar year.

The ATDP "Total Operating Revenues" include revenues from "Custom Work and Machine Rental" and "Rental Income," which are not included in the AES farm cash receipts.

c) Census of Agriculture

The Census of Agriculture (CEAG) is a census of individuals, businesses and organizations that produce agricultural products and report revenues or expenses for tax purposes to the Canada Revenue Agency. The questionnaire continues to ask for total receipts and total operating expenses although, starting in 2016, no questions are posed about detailed expenses. Data are generally provided on a calendar year basis, or for a complete fiscal year.

Unlike the AES, CEAG data on receipts include dividends received from co-operatives, Goods and Service Tax (GST) refunds, custom work receipts, and rebates received.

2. Comparison of estimates

Conceptual and methodological differences and data collection methods can result in misleading comparisons between AES receipts or expenses series and total income or expenses derived from ATDP, FFS or Census of Agriculture data. The exclusion of farm-to-farm transactions within a province in the AES and their inclusion in the ATDP, FFS and CEAG datasets is the main reason that comparisons are difficult. However, net cash income estimates (farm cash receipts minus operating expenses) are more directly comparable since, within a province, sales from one farm are an expense to another farm, and the two offset each other.

As is the case with farm-to-farm sales, some receipt items not included in the AES receipt series would tend to cancel each other out when deriving net cash income estimates from the various sources. For example, the exclusion of custom work receipts from AES receipts is compensated to a large extent in the net income estimates by the use of a net custom work estimate (custom work expenses minus custom work receipts) in the AES expense series. The subtraction of custom work receipts from custom work expenses is done in an attempt to minimize—in the aggregate—the presence in the AES of operating costs incurred by agricultural producers in providing custom work services. In a less precise manner, one could expect the ATDP estimates for the components of "miscellaneous revenue" and "miscellaneous farm expenses" not included in the AES series to offset each other to some degree.

The ATDP publishes average receipts and expenses only for farms reporting total farm operating revenues of $10,000 or more on their income tax return and agricultural corporations reporting total farm operating revenues of $25,000 and over, and for which 50% or more of their sales come from agricultural activities. For purposes of comparisons with the AES, estimations for the unincorporated farms reporting total operating revenues below $10,000 were used internally in spite of the lower quality of these estimates.

The AES includes establishments primarily engaged in growing cannabis under glass, protective cover or in open fields while these establishments are excluded in the CEAG or ATDP data.

3. Results

Table 1 — Net cash income comparison — Total

	2020
	AESTable 1 footnote1	ATDPTable 1 footnote 2	CEAG
	thousands of dollars
Newfoundland and Labrador	5,743	23,408	26,567
Prince Edward Island	99,770	122,948	120,589
Nova Scotia	10,293	88,248	84,858
New Brunswick	171,379	109,839	107,078
Quebec	1,807,904	2,200,891	2,115,720
Ontario	3,432,776	3,354,301	3,266,198
Manitoba	1,505,872	1,215,510	1,184,141
Saskatchewan	5,583,907	4,194,316	4,096,507
Alberta	2,907,566	3,092,450	3,193,317
British Columbia	412,778	591,372	588,458
Canada	15,937,987	14,993,281	14,783,433
Table 1 footnote 1 The Agriculture economic statistics (AES) covers all agricultural holdings including cannabis operations. For comparison purposes, the Canada total is the sum of the provincial totals. Return to Table 1 footnote 1 referrer Table 1 footnote 2 Adjusted to include unincorporated farms with total farm operating revenues lower than $10,000. Does not include agricultural corporations with total operating revenues below $25,000. Return to Table 1 footnote 2 referrer

Table 2 — Net cash income comparison — Absolute Difference

	2020
	ATDP - AESTable 2 footnote 1	CEAG - AESTable 2 Footnote 2
	thousands of dollars
Newfoundland and Labrador	17,665	20,824
Prince Edward Island	23,178	20,819
Nova Scotia	77,955	74,565
New Brunswick	-61,540	-64,301
Quebec	392,987	307,816
Ontario	-78,475	-166,578
Manitoba	-290,362	-321,731
Saskatchewan	-1,389,591	-1,487,400
Alberta	184,884	285,751
British Columbia	178,594	175,680
Canada	-944,707	-1,154,554
Footnote 1 Agriculture Taxation Data Program (ATDP) minus the Agriculture Economic Statistics (AES). Return to Table 2 footnote 1 referrer Footnote 2 Census of Agriculture (CEAG) minus the AES. Return to Table 2 footnote 2 referrer

Table 3 — Net cash income comparison — Relative Difference

	2020
	(ATDP - AES) / AESTable 3 Footnote 1	(CEAG - AES) / AESTable 3 Footnote 2
	percent
Newfoundland and Labrador	307.6	362.6
Prince Edward Island	23.2	20.9
Nova Scotia	757.4	724.4
New Brunswick	-35.9	-37.5
Quebec	21.7	17.0
Ontario	-2.3	-4.9
Manitoba	-19.3	-21.4
Saskatchewan	-24.9	-26.6
Alberta	6.4	9.8
British Columbia	43.3	42.6
Canada	-5.9	-7.2
Footnote 1 Agriculture Taxation Data Program (ATDP) minus the Agriculture Economic Statistics (AES) divided by the AES. Return to Table 3 footnote 1 referrer Footnote 2 Census of Agriculture (CEAG) minus the AES divided by the AES. Return to first Table 3 footnote 2 referrer

4. Conclusion

Comparing data collected for different purposes is not an easy task. It is extremely difficult to identify precisely what causes the discrepancies at the aggregate level. There will be always discrepancies due to differences in coverage, accounting methods, fiscal years as well as the edit, imputation and estimation methods of the survey, census or administrative data. These differences are often compounded in an estimate, such as net cash income, that is measured residually.

Contents

• Section 1: Overview
• Section 2: Risk Area Identification and Categorization
• Section 3: Analysis of the personal information elements for the program or activity
• Section 4: Flow of personal information for the program or activity
• Section 5: Privacy compliance analysis
  • Principle 1: Accountability
  • Principle 2: Limiting Collection
  • Principle 3: Direct Collection and Purpose Identification
  • Principle 4: Retention
  • Principle 5: Accuracy
  • Principle 6: Disposal
  • Principle 7: Limiting use
  • Principle 8: Limiting Disclosure
  • Principle 9: Safeguards
  • Principle 10: Openness
  • Principle 11: Individual Access
• Section 6: Threat and risk assessment
• Section 7: Summary of analysis and recommendations
• Section 8: Breach protocol
• Section 9: Supplementary Documents List
• Appendix 1 – PIA Summary
• Appendix 2 – Privacy Notice

Section 1: Overview

Responsible department: Statistics Canada

Chief Privacy Officer: Director, Office of Privacy Management and Information Coordination
Subject-matter manager: Director, Census Communications
Senior official: Assistant Chief Statistician, Corporate Strategy and Management
Legal authority: Financial Administration Act
Reference to Personal Information Bank (PIB): Standard PIB – Public Communications, PRN 939, PSU 914

Project Description:

To improve respondent services, Statistics Canada is introducing a chatbot on its Census website as a new communications method. This automated communication and service channel aims to help Census respondents anonymously submit questions and obtain more timely responses in a positive and secure online experience. This technology will improve services and assist respondents in their questionnaire needs, in addition to creating a positive and secure online experience.

While Statistics Canada already provides a Contact us form^{Footnote 1} where users can submit questions, the chatbot will automatically provide answers instead of waiting the general 5 to 10 business days to receive a response.

The chatbot will be of particular value when Statistics Canada conducts census activities. During the 2021 Census collection cycle, the Census Help Line (CHL) received over 2 million calls. This unprecedented volume led to 1.2M not being answered due to limited resources. Equally, Statistics Canada's Respondent Relations Unit received 63,000 emails and 7,000 pieces of mail, five times the amount received in the previous census, and the social media (Web2Social) team received numerous complaints regarding the lack of immediate response, long phone queues, or dropped calls.

The chatbot does not collect any personal information and a privacy notice (see Appendix 2) statement will be provided to users at the beginning of the chat session instructing them not to share any personal information and advising users who nevertheless include PI, that it will be manually removed from chatbot transcripts. The chatbot will use Artificial Intelligence (AI) to detect keywords to answer a user's question. If a user requires further assistance, the chatbot will transfer the conversation to a live chat agent. The chatbot was initially trained on extracted content from 2021 Census Questions and Answers (e.g., What is a census letter? Where can I find my secure access code?) and will keep training on new content that will be added as needed. As the chatbot does not make any administrative decisions on individuals, an Algorithmic Impact Assessment is not required.^{Footnote 2}

The chatbot software and licenses have been acquired by Statistics Canada from a service provider. It has been integrated into Statistics Canada's website structure in a development environment prior to being moved to its production environment on January 15, 2024. A full security assessment was completed on January 9, 2024, with a risk management input (RMI) assessed as very low.

Section 2: Risk Area Identification and Categorization

The following table below evaluates the aggregate risk of the proposed initiative against a suit of standard dimensions applicable to most Statistics Canada's programs and activities. The numbered risk scale is presented in ascending order: level 1 represents the lowest level of potential risk for the risk dimension the fourth level (4) represents the highest level of potential risk for the given risk dimension.

Applicable risk level for each dimension is in BOLD.

a) Type of program or activity	Risk scale
Program or activity that does NOT involve a decision about an identifiable individual	1
Administration of program or activity and services	2
Compliance or regulatory investigations and enforcement	3
Criminal investigation and enforcement or national security	4
b) Type of personal information involved and context	Risk scale
Only personal information, with no contextual sensitivities, collected directly from the individual or provided with the consent of the individual for disclosure under an authorized program.	1
Personal information, with no contextual sensitivities after the time of collection, provided by the individual with consent to also use personal information held by another source.	2
Social Insurance Number, medical, financial or other sensitive personal information or the context surrounding the personal information is sensitive. Personal information of minors, legally incompetent individuals or a representative acting on behalf of the individual.	3
Sensitive personal information, including detailed profiles, allegations or suspicions and bodily samples, or the context surrounding the personal information is particularly sensitive.	4
c) Program or activity partners and private sector involvement	Risk scale
Within the institution (among one or more programs within the same institution)	1
With other government institutions	2
With other institutions or a combination of federal, provincial or territorial, and municipal governments	3
With foreign governments, international organizations and/or private sector organizations	4
d) Duration of the program or activity	Risk scale
One-time program or activity	1
Short-term program (January 16, 2024 to June 28, 2024 for Behavioral Testing and from January 2026 to August 2026 for the 2026 Census of Population collection period).	2
Long-term program (long-term date / ongoing / no end date)	3
e) Program population*	Risk scale
The program's use of personal information for internal administrative purposes affects certain employees.	1
The program's use of personal information for internal administrative purposes affects all employees.	2
he program's use of personal information for external administrative purposes affects certain individuals.	3
The program's use of personal information for external administrative purposes affects all individuals for external administrative purposes.	4
* The program's use of personal information is not for administrative purposes. Information is collected for statistical purposes, under the authority of the Statistics Act.	N/A
f) Personal information transmission	Risk scale
The personal information is used within a closed system (i.e., no connections to the Internet, Intranet or any other system and the circulation of hardcopy documents is controlled).	1
The personal information is used in a system that has connections to at least one other system.	2
The personal information is transferred to portable devices or printed (USB key, diskette, laptop computer), transferred to a different medium or is printed.	3
The personal information is transmitted using wireless technologies.	4
g) Technology and privacy
Does the new or substantially modified program or activity involve implementation of a new electronic system or the use of a new application or software, including collaborative software (or groupware), to support the program or activity in terms of the creation, collection or handling of personal information? Yes. Power Microsoft Agent (chatbot) from Microsoft Dynamics is being implemented to facilitate a new communication method via chatbot to improve Statistic Canada's services and promote direct communication with data users and survey respondents.
Does the new or significantly modified activity or program require changes to IT legacy systems? Yes. Integrating Power Microsoft Agent (chatbot) from Microsoft Dynamics code into the website will require changes to our IT systems to enable integration of the code in the infrastructure.
Specific technological issues and privacy Does the new or substantially modified program or activity involve implementation of one or more of the following technologies: enhanced identification methods (e.g., biometric technology) surveillance automated personal information analysis, personal information comparison and knowledge discovery techniques No.
h) Potential risk that in the event of a privacy breach, there will be an impact on the individual or employee
Users will be advised to not provide any personal information when a chat session is opened. If provided, it will not be used and will be destroyed after three months. Therefore, the risk of some personal information being disclosed without proper authorization is very low, and the impact on the individual would be low.
i) Potential risk that in the event of a privacy breach, there will be an impact on the institution.
Users will be advised to not provide any personal information. If provided, it will not be used and will be destroyed after three months. Therefore, the risk of certain personal information being disclosed without proper authorization is very low, and the impact on the institution would be low.

Section 3: Analysis of the personal information elements for the program or activity

A privacy notice will inform users to not provide personal information when the chat starts. However, users might still provide personal information voluntarily. Such information could be anonymously retained by Statistics Canada for a three-month period in the form of transcripts for the purpose of assessing quality of the service and to meet user needs.

If users voluntarily share personal information despite the privacy notice, deemed employees from Respondent Relations will manually remove personal information from the transcript before being saved on Statistics Canada servers.

Necessity and Proportionality

While the chatbot is not intended to collect or use personal information and users are advised not to include any, some personal information could inadvertently be provided. The implementation of the chatbot can be justified against Statistics Canada's Necessity and Proportionality Framework:

1. Necessity: The chatbot will be a new communications method introduced on Statistics Canada website to improve respondent services by providing automated communication and service channels that will increase the agency's Census Program's efficiency by reducing the number of correspondence pieces to be treated by Statistics Canada employees. These improvements will also benefit Canadians by creating a positive, secure online experience and assist respondents in their 2026 Census of Population questionnaire needs.
2. Effectiveness - Working assumptions: The chatbot will allow live agents to focus on answering more complex questions in a timely manner and improve service standards. Moreover, the chatbot will help users fulfill their legal census obligation more quickly by submitting a question anonymously to the chatbot. Then, the chatbot will detect keywords and propose answers to users without using any personal information. The chatbot will create efficiency for Canadians and the agency.
3. Proportionality: While there is a potential for the agency to collect personal information, the benefits of having quick and efficient access to information to help Canadians complete the Census are highly valuable proportional to any privacy risks. The chatbot aims to improve and supplement the existing services provided to Canadians by creating a positive, secure online experience and assist respondents in their 2026 Census of Population questionnaire needs.
4. Alternatives: Unfortunately, there are no alternatives to automated services such as a chatbot. The method used in the past resulted in a high number of incoming calls and emails for the 2021 Census collection cycle to the CHL and Respondent Relations Unit of many of which were left unanswered due to lack of resources and unprecedented volumes. The chatbot will enable automation of Census program's services and efficiently provide answers to Canadian users. In the event the chatbot is not able to provide users with the information they need to fulfill their obligation towards the Census program, their request will be transferred to our Live Chat (CHL) to provide them with further assistance. Requests that require research or a personalized answer, such as complaints, will be transferred to a Respondent Relations agent that will work on a personalized response.

Section 4: Flow of personal information for the program or activity

Identify the source(s) of the personal information collected and/or how the personal information will be created.

When a user opens a chat session, they will be presented with a privacy notice statement and informed not to share any personal information before being prompted to ask a question. Once the question is submitted in the conversation window, the chatbot will analyze keywords to find a suitable answer to the user's question. Although users are advised to not provide any identifying personal information, they could voluntarily provide some.

Identify the internal and external uses and disclosures. Specifically, identify the areas, groups and individuals who will have access to or handle the personal information and to whom it will be provided or disclosed.

Information that users may voluntarily provide will not be used or disclosed, either internally or externally. Only deemed employees from Respondent Relations, Census Help Line (CHL) and the social media (Web2Social) are viewing and handling the data.

Identify where the personal information will transit and will be stored or retained.

During a chatbot session, any personal information that may be provided voluntarily by the user will transit through Statistics Canada's Microsoft Dynamics portal. The exchange will then be manually striped of any personal information and be stored on Statistics Canada's Cloud for a three-month period for the purpose of assessing quality of the service and to meet user needs.

Identify where areas, groups and individuals can access the personal information

When a chatbot user is redirected to a live chat CHL, Respondent Relations, Web2Social or recruitment agent, the agent in question and their supervisor will have access to that information.

During the three-month storage period, only supervisors and managers (four to six people) who need to consult the information for evaluation purposes will have access to the data.

Section 5: Privacy compliance analysis

As recommended by the Office of the Privacy Commissioner's Guide to the Privacy Impact Assessment Process, this chatbot has been assessed against the following principles that are based on the Organization for Economic Co-operation and Development's (OECD) Guidelines on the Protection of Privacy and Transborder Flows of Personal Data.

Principle 1: Accountability

Statistics Canada is responsible for all personal confidential information collected and used under authority of the Financial Administration Act. The agency is also responsible for all personal information under its control and has designated individuals who are accountable for the agency's compliance with the obligations of federal departments to respect privacy rights as described in sections 4 to 8 of the Privacy Act.

The Director of the Office of Privacy Management and Information Coordination is Statistics Canada's Chief Privacy Officer (CPO) and is accountable for the department's compliance with the principles contained in this document. The CPO is also responsible for the development of Statistics Canada's policies related to information, including all aspects of information classification, control, and access and for providing advice, guidance, and assistance in the implementation of information security measures.

The Chief Security Officer (CSO) is responsible for the day-to-day operations of the Departmental Security Office, and for the development and administration of the security program for Statistics Canada.

The Respondent Relations, CHL and Web2Social teams and its managers, as directed by the Census Communications Director (Communications Branch), are responsible for applying all central agency and Statistics Canada policies related to the protection of privacy and personal information for this activity.

The Communications and Dissemination Branch is responsible for the entire Statistics Canada website.

An individual can address a challenge concerning compliance by Statistics Canada with the above principles.

Complaints may be addressed to:

Chief Privacy Officer
Statistics Canada
R.H. Coats Building, 2nd floor
100 Tunney's Pasture Driveway
Ottawa, Ontario K1A 0T6
Telephone: 613-951-0466
E-mail: statcan.atip-aiprp.statcan@statcan.gc.ca

In addition, under the Privacy Act, individuals may make a complaint to the Office of the Privacy Commissioner of Canada, who will undertake an investigation.

Principle 2: Limiting Collection

The authority to collect this information falls under the Financial Administration Act. Personal information will be not collected, but any that might be voluntarily provided by users will be protected under the Privacy Act. That information is included in the "Public Communications" Personal Information Bank.

The chatbot does not require the collection of personal information to provide is service to Canadians. As mentioned in previous section, users might voluntarily share personal information even if they have been informed not to at the beginning of the chat session.

Principle 3: Direct Collection and Purpose Identification

The purpose of the chatbot is to provide Canadians with a service that will increase Statistic Canada's Census Program's efficiency by helping respondents fulfill their legal census obligation more quickly and efficiently by anonymously submitting a question to the chatbot.

No personal information is asked or required from users to use the chatbot. Statistics Canada will not disclose personal information that might be provided by chatbot users. A privacy notice statement (Appendix 2) will be given to users instructing them to not share personal information when the chat session starts.

The conversation history will be used to improve Statistics Canada's services and to make information about the Census more accessible to the Canadian population. The data is used only by those individuals with a need-to-know including Respondent Relations, Social Media and recruitment agents and their supervisors to ensure the quality of the service provided and to meet user needs.

Consent will not be asked for because no personal information is collected, used or disclosed. Any personal information that might be voluntarily provided by a user will not be used, and the personal information will be manually removed from the conversation transcripts. Such transcripts will only be retained by Statistics Canada for a three-month period.

Principle 4: Retention

Personal information inputted by respondents will not be retained since the nature of the chatbot does not require the use of such information to provide users with an answer to their posed question. If users voluntarily include personal information, the information will be manually deleted from the conversation transcripts. These redacted transcripts will then be retained by Statistics Canada for a three-month period to improve services and to make information about the Census more accessible to the Canadian population.

Additionally, the Power Virtual Agent system (by Microsoft) will keep the conversation transcripts for a period of 28 days after which it will automatically be deleted. Microsoft has confirmed that it does not use the information to train, retrain, or improve their models.

Principle 5: Accuracy

Participants submit their own information electronically. Statistics Canada will not modify any of the submitted information, so inaccuracies should not occur. However, any errors that are brought to the attention of Statistics Canada by participants will be rectified immediately. If the chatbot provides the wrong answers based on faulty users' information, the chatbot will prompt the user to reformulate or specify his concerns in order to provide the proper information.

Statistics Canada will not alter any information that may be voluntarily submitted by users, and that information will be retained only for three months. Because that information is not required and will not be used or disclosed, its accuracy in the context of this activity is irrelevant.

Principle 6: Disposal

Voluntarily shared personal information will be in digital format. Disposal of such information will be done directly from the chatbot system.

Technology and privacy issues

• There will be no change to the business requirements. The same principles and guidelines applicable to email communication via Respondent Relations, the existing communication service, will be applied to chatbot.
• The existing communication tools, as well as the addition of chatbot, comply with privacy obligations.
• A privacy notice before using chatbot will notify users that their information will remain protected under the Privacy Act.

Principle 7: Limiting use

The information will be used to improve Statistics Canada's services and to make data more accessible to the Canadian population. The information is used only by Respondent Relations and Social Media agents and their supervisors to ensure the quality of the service provided and to meet user needs. However, the use of personal information is not required for the chatbot to operate.

Once the chatbot is launched, an automated message will appear to inform users not to share personal information.

Principle 8: Limiting Disclosure

Statistics Canada will not disclose personal information from the chatbot, without the consent of the respondent or unless as permitted by the Statistics Act. Access to any information obtained under the authority of the Statistics Act is restricted to employees who must swear an oath of confidentiality under the Act and who also have a need-to-know as part of their job duties.

Principle 9: Safeguards

Statistics Canada takes seriously its legal obligation to safeguard the personal information of all Canadians. The agency has had in place a framework of policies, directives, procedures and practices to safeguard protected information, including personal information, against loss, theft, unauthorized access or disclosure; they are supported by physical, organizational and technological measures that protect all the personal information that Statistics Canada holds.

Access will be restricted solely to Respondent Relations, CHL and Web2Social agents and supervisors located in the National Capital Region office. The information will be protected by existing mechanisms, such as the following:

• When logging into a work session on the computer, agents and supervisors must enter their protected, confidential username and password.
• Personal information voluntarily received from clients will pass through the site only to measure performance, productivity, and the quality of the service provided by Respondent Relations, CHL and Web2Social agents, and to answer users' questions. Briefly, the information will help measure accuracy of answers and it will allow us to remain at the dawn of new topics emerging to rehabilitate our content.
• The information will be stored only on Statistics Canada's internal platform; the information will not be disseminated or printed.
• Information passing over the Internet will be secured by HTTPS. It will then be stored on Statistics Canada's Cloud that meets the stronger IT security requirements. In addition, vulnerability analyses of the Census Chatbot have been conducted by Statistics Canada's Cyber Security Division. They completed a Security Assessment and Authorization (SA&A), and an interim Authority to Operate (iATO) was granted for a one-year period (Jan 2024-Jan 2025).
• The software used contains an option to hide the IP address, and that option will be used to prevent the storage of IP addresses.

Upon discovery of an actual or suspected privacy breach (however unlikely), the steps described in Section 8 would be taken.

Principle 10: Openness

Statistics Canada produces specific, easily accessible information about its policies and practices on the management and protection of personal information. Information on the use of personal information, in the form of a Privacy notice, can be read on the agency's website at www.statcan.gc.ca.

Summaries of approved privacy impact assessments are also available from the website, under "About Statistics Canada – Privacy Impact Assessments"

Information on the Statistics Canada website related to policies and procedures

The agency's Privacy Notice can be found under "About us – Terms and conditions and Privacy" on the Statistics Canada website, where there is also information about:

• the agency's Privacy Framework
• protecting confidential information and privacy at Statistics Canada;
• privacy-related policies and practices at Statistics Canada;
• what record linkage is and how it is used at Statistics Canada;
• Statistics Canada's Directive on Microdata Linkage;
• approved linkages at Statistics Canada, with their purpose, description and output

Statistics Canada's Trust Center also provides answers to questions on the security, privacy and confidentiality of personal information.

Contacts for further information

For further information about NetSupport, the contact person is:
Program Manager of the Business Integration, Collection, Planning, and Research Division
Statistics Canada
150 Tunney's Pasture Driveway
Ottawa, Ontario K1A 0T6
Telephone: 902-220-9153
Email: Patrick.Ellis@statcan.gc.ca

Principle 11: Individual Access

The collection of personal information through the Census communications chatbot is described in the Personal Information Bank (Public Communications – PSU 914), which is published in Statistics Canada's Information about Programs and Information Holdings chapter.

Upon request, Statistics Canada will provide users with access to their personal information held by the agency.

If users wish to submit a formal request for access to their personal information under the Privacy Act, the contact person at Statistics Canada is:

Access to Information and Privacy Coordinator
Statistics Canada
R.H. Coats Building, 2nd floor
100 Tunney's Pasture Driveway
Ottawa, Ontario K1A 0T6
Telephone: 613-951-0466
E-mail: statcan.atip-aiprp.statcan@statcan.gc.ca

Section 6: Threat and risk assessment

The purpose of this section is to assess the consultation for potential threats and risks that could compromise privacy. It outlines Statistics Canada's existing safeguards, the probability of occurrence of the threat, and the severity of the impact as it relates to the privacy and protection of participants' information.

Statistics Canada currently uses numerous safeguards to reduce threat probabilities. These safeguards are described in agency policies, practices, tools and/or techniques.

The ratings for threat probability, impact and residual risk are defined and presented as follows

Threat: An undesirable event with the potential to compromise privacy or breach data confidentiality.

Threat probability: The likelihood that the threat will occur, given Statistics Canada's existing safeguards. The threat probability is rated numerically.

• Level 1: The threat can only come about through the use of very specialized knowledge and/or costly specialized facilities and/or a sustained effort. The threat is unlikely to occur.
• Level 2: The threat requires some specialized knowledge and/or facilities and/or a special endeavour to create or take advantage of the threat opportunity. The threat is somewhat likely to occur.
• Level 3: The threat opportunity is widely available and can occur either intentionally or accidentally with little or no specialized knowledge and/or facilities. The threat is very likely to occur.

Impact: The effect on a participant's privacy in the event that a threat comes to fruition and their information is compromised. The level or degree of impact is expressed in terms of outcome severity as it relates to individual privacy.

• Level 1: Minor injury with no or minimal harm or embarrassment to the individual.
• Level 2: Moderate injury causing some harm or embarrassment to the individual, but with no direct negative effects.
• Level 3: Severe injury such as lasting harm or embarrassment that will have direct negative effects on an individual's career, reputation, financial position, safety, health or well-being.

Residual risk: A numeric rating is given following an evaluation and comparison of the threat probability and the impact on privacy.

Threat and Risk Assessment Grid

Chatbot on Statistics Canada's website

While the Chatbot does not collect personal information and users are instructed not to include any personal information in the chat, the following addresses any risks should any personal information be inadvertently or voluntarily included and compromised before it gets manually removed by deemed Respondent Relations employees.

Threats	Existing Statistics Canada Safeguards	Probability	Impact	Residual Risk	Assessment of Residual Risk
Environment: Risk associated with users' privacy – within Statistics Canada
Activity: access to chatbot
1. There is unauthorized access to NetSupport by a non-Statistics Canada employee.	Access to the Cloud on which chatbot is hosted is restricted to a small number of Statistics Canada employees and deemed employees of Shared Services Canada who provide services to Statistics Canada. IT security measures include protection by firewall, configuration and access via Statistics Canada's internal network on local Network B. Access to the system requires a username and password reserved for authorized employees of the Respondent Relations, CHL and Web2Social teams.	1	1	1	Acceptable
2. There is unauthorized access to NetSupport by a Statistics Canada employee.	IT security measures include access via Statistics Canada's internal network on local Network B. Access to the system requires a username and password reserved for authorized employees of the Respondent Relations, CHL and Web2Social teams. Also, employees are trained to lock their computers before leaving their offices.	1	1	1	Acceptable
3. There is unauthorized use or disclosure of personal information	Employees have been made aware of the importance of protecting personal information and must comply with the Directive on Access to Information and Privacy. They have a legal obligation to comply with the legislative requirements of the Access to Information Act and the Privacy Act. If they fail to do comply, they are subject to the penalties set out in the legislation. Any employee who uses or discloses personal information in an unauthorized manner would receive additional training and may be subject to disciplinary follow-up.	1	1	1	Acceptable

Section 7: Summary of Analysis and Recommendations

A privacy impact assessment for the chatbot was conducted to determine if there were any privacy, confidentiality and security issues associated with the program, and if so, to make recommendations for their resolution or mitigation.

This document summarizes Statistics Canada's assessment of the privacy implications of the chatbot. It includes a review of the privacy principles as they apply to the program. Also included is an assessment of the risks to the privacy, confidentiality, and security of users' information.

This assessment did not identify any privacy risks that cannot be managed using existing safeguards.

Section 8: Breach protocol

The chatbot meets agency standards for both IT and physical security. It includes password protection for access to the database/tool, configuration and use of a firewall. For this reason, the threat and risk assessment (TRA) grid rates unauthorized access by either Statistics Canada employees or individuals outside Statistics Canada as low probability.

Upon discovery of an actual or suspected privacy breach (however unlikely), the following steps, in accordance with the Statistics Canada Information and Privacy Breach Protocol, would be taken:

• Immediate notification of the Chief Security Officer and the Chief Privacy Officer. Response could include suspending operation of the chatbot activities.
• In collaboration with Departmental Security and IT Security, there would be an internal investigation that would include recommendations to prevent any recurrence. Any investigation would document in detail the circumstances that gave rise to the privacy breach, and determine what information may have been breached, the impact of the breach, and what measures have been introduced to eliminate the risk of any subsequent breach.
• In the case of a "material privacy breach", in accordance with the TBS Directive on Privacy Practices, Statistics Canada would notify the Office of the Privacy Commissioner (OPC) and the Treasury Board Secretariat (TBS). "Material breaches" are those involving sensitive personal information and that could reasonably be expected to cause serious injury or harm to the individual.
• Depending on the nature of the breach, impacted individuals would be provided with an explanation of the situation and the steps being taken to remove the information from the possession of those not authorized to have it. Individuals would also be informed that they have the right to file a complaint with the Office of the Privacy Commissioner (OPC). The OPC and TBS would be informed of the individual(s) whose information was disclosed, the investigation and what actions have been taken to prevent a re-occurrence.

Section 9: Supplementary Documents List

Appendix 1 – PIA Summary
Appendix 2 – Privacy notice

Appendix 1 – PIA Summary

Census Chatbot
Privacy Impact Assessment Summary

Introduction

To improve respondent services, Statistics Canada is introducing a chatbot on its Census website as a new communications method. This automated communication and service channel aims to help Census respondents anonymously submit questions and obtain more timely responses in a positive and secure online experience. This technology will improve services and assist respondents in their questionnaire needs, in addition to creating a positive and secure online experience.

While Statistics Canada already provides a Contact us form^{Footnote 3} where users can submit questions, the chatbot will automatically provide answers instead of waiting the general 5 to 10 business days to receive a response.

Objective

A privacy impact assessment for Chatbot was conducted to determine if there were any privacy, confidentiality or security issues with this program and, if so, to make recommendations for their resolution or mitigation.

Description

The chatbot will be of particular value when Statistics Canada conducts census activities. During the 2021 Census collection cycle, the Census Help Line (CHL) received over 2 million calls. This unprecedented volume led to 1.2M not being answered due to limited resources. Equally, Statistics Canada's Respondent Relations Unit received 63,000 emails and 7,000 pieces of mail, five times the amount received in the previous census, and the social media (Web2Social) team received numerous complaints regarding the lack of immediate response, long phone queues, or dropped calls.

The chatbot does not collect any personal information and a privacy notice (see Appendix 2) statement will be provided to users at the beginning of the chat session instructing them not to share any personal information and advising users who nevertheless include PI, what will happen in such instances. The chatbot will use Artificial Intelligence (AI) to detect keywords to answer a user's question. If a user requires further assistance, the chatbot will transfer the conversation to a live chat agent. The chatbot was initially trained on extracted content from 2021 Census Questions and Answers (e.g., What is a census letter? Where can I find my secure access code?) and will keep training on new content that will be added as needed. As the chatbot does not make any administrative decisions on individuals, an Algorithmic Impact Assessment is not required.

The chatbot software and licenses have been acquired by Statistics Canada from a service provider. It has been integrated into Statistics Canada's website structure in a development environment prior to being moved to its production environment on January 15, 2024. A full security assessment was completed on January 9, 2024 with a risk management input (RMI) assessed as a very low.

Risk Area Identification and Categorization

The PIA identifies the level of potential risk (level 1 is the lowest level of potential risk and level 4 is the highest) associated with the following risk areas:

a) Type of program or activity	Risk scale
Program or activity that does not involve a decision about an identifiable individual.	1
b) Type of personal information involved and context
Only personal information, with no contextual sensitivities, collected directly from the individual or provided with the consent of the individual for disclosure under an authorized program.	1
c) Program or activity partners and private sector involvement
Within the institution (among one or more programs within the same institution)	1
d) Duration of the program or activity
Long-term program or activity.	3
e) Program population
The program's use of personal information for external administrative purposes affects certain individuals.	3
f) Personal information transmission
The personal information is transmitted using wireless technologies.	4
g) Technology and privacy
Power Microsoft Agent (chatbot) from Microsoft Dynamics is being implemented to facilitate a new communication method via chatbot to improve Statistic Canada's services and promote direct communication with data users and survey respondents. Integrating Power Microsoft Agent (chatbot) from Microsoft Dynamics code into the website will require changes to our IT systems to enable integration of the code in the infrastructure. Specific technological issues and privacy
h) Potential risk that in the event of a privacy breach, there will be an impact on the individual or employee.
Users will be advised to not provide any personal information when a chat session is opened. If provided, it will not be used and will be destroyed after three months. Therefore, the risk of some personal information being disclosed without proper authorization is very low, and the impact on the individual would be low.
i) Potential risk that in the event of a privacy breach, there will be an impact on the institution.
Users will be advised to not provide any personal information. If provided, it will not be used and will be destroyed after three months. Therefore, the risk of certain personal information being disclosed without proper authorization is very low, and the impact on the institution would be low.

Conclusion

This assessment of the Chatbot did not identify any privacy risks that cannot be managed using existing safeguards.

Appendix 2 – Privacy Notice Statement

When the chatbot box opens a privacy notice will appear in a banner at the top of the chat window.

Shorter version of Privacy Notice for chatbot banner (600-character limit (with space)):

The chatbot is designed to assist you with general inquiries related to Census. Please do not include any personal information, such as your name, address, or other identifying information.

If you have any concerns or specific issues that require personalized assistance, we recommend reaching out to our dedicated support channels, where trained professionals will be happy to help you.

For more details, visit the Privacy section of (webpage name).

Detailed Privacy Notice available to respondents on (hyperlink to StatCan webpage):

The chatbot is designed to assist you with general inquiries related to Census. Please do not include any personal information, such as your name, address, or other identifying information, as it is not required for general inquiries related to the Census.

If nevertheless users decide to share personal information, such information would be manually removed from chatbot transcripts. These transcripts could be anonymously retained for a three-month period for service evaluation purposes in the form of a compilation for the purpose of assessing quality of the service and to meet user needs.

If you have any concerns or specific issues that require personalized assistance, we recommend reaching out to our dedicated support channels, where trained professionals will be happy to help you.

If you have questions, please visit our website at www.census.gc.ca where you can also chat online with us, or contact us at 1-833-835-2024. Respondents with access to TTY (for persons who have a hearing or speech impairment) should call 1-833-830-3109. Video relay services (VRS) can also be used.

Date: February 2024

Program manager: Director, Centre for Special Business Projects
Director General, Agriculture, Energy and Environment Statistics

Reference to Personal Information Bank (PIB):

In accordance with the Privacy Act, Statistics Canada is submitting an amendment to the institutional personal information bank (PIB) StatCan PPU 116, Sociodemographic Information on Business Primary Decision Makers to describe any personal information about individuals that work for, own or operate businesses, non-profits and other organizations as volunteers, employees, and/or primary decision-makers, collected in surveys or acquired administrative data for use in Statistics Canada’s Economic and Environmental Statistics programs operating under the authority of the Statistics Act.

Sociodemographic Information on Business Owners, Primary Decision Makers, Employees, and Volunteers

Description: This bank describes personal information that relates to individuals that work for, own or operate businesses, non-profits and other organizations as volunteers, employees, and/or primary decision-makers, collected in surveys or acquired administrative data for use in Statistics Canada’s Economic and Environmental Statistics programs operating under the authority of the Statistics Act. Personal information may include gender, sexual and gender diversity, First Nations (North American Indian), Métis, or Inuk (Inuit), visible minority, persons with a disability, citizenship and immigration status, education and age.
Class of Individuals: Business owners, primary decision makers, employees and volunteers of non-profits and other organizations or businesses that are included in Statistics Canada’s Economic and Environmental Statistics programs.
Purpose: The personal information is used to produce statistical data that help provide insight into various gaps in the economy for a variety of minority groups, and serves to inform evidence-based decisions on funding and support for specific groups of businesses, non-profits or other organizations. Personal information is collected pursuant to the Statistics Act (Sections 3, 7, 8).
Consistent Uses: When collected directly and with informed consent, identifiable sociodemographic personal information may be shared with provincial and territorial statistical agencies and other government organizations that have demonstrated a requirement to use the data, and as permitted under the provisions of Sections 11 or 12 of the Statistics Act.
Retention and Disposal Standards: Information is retained until it is no longer required for statistical purposes and then it is destroyed.
RDA Number: 2007/001
Related Record Number: To be assigned by Statistics Canada
TBS Registration: To be assigned by TBS
Bank Number: StatCan PPU 166

Description of statistical activity:

Under the authority of the Statistics Act^{Footnote 1}, Statistics Canada conducts the Canadian Survey on the Provision of Child Care Services (CSPCCS). This irregular voluntary survey is funded by Employment and Social Development Canada (ESDC) and its purpose is to collect information on the provision of child care services in Canada for children ages 12 and under at the national, provincial, and territorial levels. Information about staffing levels and training, services provided, enrollment and daily fees is collected from licensed and unlicensed home-based and centre-based child care providers. 

Starting with the 2024 survey cycle, four new sociodemographic questions pertaining to only home-based child care operators will be asked: 

• Gender
• First Nations (North American Indian), Métis, or Inuk (Inuit)
• Landed immigrant within the last 10 years 
• Person with a disability

Home-based child care operators make up approximately 69% of child care providers in Canada. They are often run and operated by the same individual and thus the sociodemographic questions will typically be answered directly by the individual to whom they pertain. 

Respondents have the option to select the response “Prefer not to answer” or “Don’t know” for the questions pertaining to the gender of the home operator, whether the home operator is First Nations, Métis, or Inuk (Inuit), whether the home operator is a landed immigrant to Canada in the last 10 years, and whether the home operator is a person with a disability. Further, the respondent may leave any of these questions unanswered and continue the survey. As such, the respondent may choose to answer these questions to the best of their knowledge or not. Furthermore, the survey respondent is instructed in the beginning of the questionnaire that the owner or operator of the child care service is the best person to respond to the survey.

Reason for supplement:

While the Generic Privacy Impact Assessment (PIA) addresses most of the privacy and security risks related to statistical activities conducted by Statistics Canada, this supplement was conducted due to the potentially sensitive nature of the sociodemographic information about home child care operators. As is the case with all PIAs, Statistics Canada’s privacy framework ensures that elements of privacy protection and privacy controls are documented and applied.

Necessity and Proportionality

The collection of personal information for the Canadian Survey on the Provision of Child Care Services can be justified against Statistics Canada’s Necessity and Proportionality Framework: 

1. Necessity: Child care was included as part of Budget 2021 to support and create more high quality, affordable child care spaces across the country. The CSPCCS is funded by the $95 million that was previously allocated to child care data collection and research in Budget 2017^{Footnote 2} as data gaps exist in Canada in terms of profiles of home-based child care providers. Home-based providers service a wide range of child care needs, and it is important to understand whether they reflect the demographics of the population to whom their services are being provided. It is expected that building the capacity of the national child care sector will necessitate supporting diverse groups to develop their skills to take on greater leadership roles, building the entrepreneurial capacity and business skills of under-represented groups (such as Indigenous communities, women, immigrants and persons with disabilities), facilitating the sharing of industry experience, best practices and knowledge to help under-represented groups manage transformation, and reinforcing the child care sector by incorporating the views of a more diverse set of industry players. Better data, analysis and information can highlight inequities and promote fairness and inclusion in the delivery of programs, services, and policy decisions. 

   Collecting this data allows for baseline metrics of participation by underrepresented groups (i.e., to what extent are each of the groups represented among home-based child care providers and how does this compare to representation of the children served in the population?). 

   Employment and Social Development Canada (ESDC) and other Government of Canada departments will use the resulting information to better understand the child care sector in order to make recommendations to policy departments regarding the development and support of child care programs in Canada. The information will also allow national, provincial and territorial governments as well as researchers to further support and foster collaboration in the child care sector. 

   Specifically, the sociodemographic information collected may help strengthen the sector by leading to the creation of policies that better reflect the diversity of Canadian communities, enhance collaboration across different jurisdictions, secure and support public trust in the sector, and improve client services. In addition, the data may be used to inform upcoming child care workforce strategies. 

   Questions on immigrant status and Indigenous identity of home-based child care providers contribute to an understanding of the Early Learning and Child Care Framework (ELCC) home-based workforce. The CSPCCS can provide insights into the intersection between low remuneration in this nearly entirely female field and its demographic characteristics of which nearly half are immigrants or non-permanent residents. Further, comparatively little information has been collected on Indigenous home-based child care providers. It is possible that this information may be relevant to gaining a better understanding of the supply of child care services across Canada. For example, the Indigenous Early Learning and Child Care Framework (IELCC) represents the Government of Canada and Indigenous peoples’ work to co-develop a transformative Indigenous framework that reflects the unique cultures, aspirations and needs of First Nations, Inuit and Métis children across Canada. Having information about the supply of Indigenous owned home-based child care services across Canada would be crucial to evaluate the effectiveness of the IELCC.

   Questions on gender and persons with disabilities provide a comprehensive view of the population of home-based child care operators. Understanding the gender distribution and prevalence of persons with disabilities among child care operators provides valuable insights on any disparities in the sector. For example, to what extent are each of these groups represented among home-based child care providers and how does this compare to representation of the children served in the population.

2. Effectiveness - Working assumptions: Although the survey is voluntary and the respondents have the option to select the response “Prefer not to answer” or “Don’t know”, it is expected that those that choose to provide the information will do so as accurately as possible. Further, the survey instructions identify the owner or operator of the child care service to be the best person to respond to the survey, increasing the likelihood of direct collection of personal information. When these questions are received by proxy, the small size of daycares is a factor, as it would be reasonably expected that any others at the daycare would be aware of this information, or would decline to respond or select “Prefer not to answer” or “Don’t know” in cases where they were not. 

3. Proportionality: The information will be collected to better understand participation in the sector by underrepresented population groups, allowing for the implementation of evidence-based policies in the child care sector to further encourage more widespread participation. The five sociodemographic variables represent the minimum required to best represent Canada’s diverse communities as determined by subject matter experts at Statistics Canada and ESDC and are only collected from home-based child care services which make up approximately 69% of child care providers in Canada. Practices in the child care sector vary significantly across the provinces and territories, necessitating good information for each jurisdiction. As such, the sample has been designed to support the production of quality estimates at the provincial and territorial level. Further, linking to Statistics Canada’s only linkable dataset that contains these types of variables – the Census of Population long-form, to which only a portion of Canadians were asked to respond – would result in a linkage rate that would be too low to meet Statistics Canada’s quality guidelines. 

   This effort ultimately aims to address inequalities and disparities affecting underrepresented population groups, allowing for a broader reach in participation across the child care sector and to foster future generations of child care operators. As such, the collection of the new sociodemographic information is considered proportional to the potential benefits.

4. Alternatives: The following alternative modes of collecting the sociodemographic information were considered, and due to their limitations, the voluntary collection of sociodemographic information about operators was identified as being the most efficient and effective method to achieve the identified needs.

   Direct collection from operators through additional screening questions: Despite potentially being more accurate, the addition of screening questions to validate that the survey respondent is also the business operator could be more burdensome on respondents, limiting the response rate and compromising the program’s ability to produce statistics about these minority groups that comply with Statistics Canada’s quality guidelines. 

   Linking the sociodemographic variables from other surveys or administrative files: Statistics Canada’s only linkable dataset that contains these types of variables is the Census of Population long-from to which only a portion of Canadians were asked to respond. As such, the linkage rate would be too low to meet Statistics Canada’s quality guidelines.

Mitigation factors:

The overall risk of harm to the survey respondents has been deemed manageable with existing Statistics Canada safeguards that are described in Statistics Canada’s Generic Privacy Impact Assessment, such as with the following measures:

• Respondents will be informed at the start of the questionnaire that it is voluntary, and that they have the right to refuse to provide any information. 
• Respondents have the option to select the response “Prefer not to answer” or “Don’t know” for the questions pertaining to gender, First Nations, Métis, or Inuk (Inuit), landed immigrant to Canada in the last 10 years, and person with a disability. 
• Respondents can leave the questions blank and proceed with the rest of the survey. 
• Respondents are instructed at the beginning of the questionnaire that the owner or operator of the child care service is the best person to respond to the survey.
• Statistics Canada will not attempt to re-identify any individuals from the personal information collected. 
• Statistics Canada will not publish any information that could allow the identification of any individuals. Additional suppressions will be performed on outputs to ensure any risk of re-identification is mitigated.

Conclusion:

This assessment concludes that the overall risk of harm has been deemed manageable with the abovementioned mitigations and existing Statistics Canada safeguards described in Statistics Canada's Generic Privacy Impact Assessment.

• Introduction and document purpose
• Context: What is the Business Register?
• Data Integration Infrastructure Division business data products
  • 1. Canadian business counts (formerly Canadian business patterns)
  • 2. Custom aggregate data tables
    • Employment size ranges
    • Revenue ranges
    • Profit and non-profit data (December only)
    • Business type and public and private data (December only)
  • 3. Business Register microdata
• Other relevant information about the Business Register
  • Data variations caused by methodological changes, by year
• Data quality and limitations
  • Time series
  • Creations
  • Inactivation
  • North American Industry Classification System
• Key definitions found in Business Register Data
  • Statistical entities
    • Statistical enterprise
    • Statistical establishment
    • Statistical location
  • Employment
  • Employment size ranges
• Geography
  • Structure of the Standard Geographical Classification
    • (1) Province and territory
    • (2) Census division
    • (3) Census subdivision
  • Statistical Area Classification
  • Census metropolitan area and census agglomeration
  • Other geographies
  • Economic region
  • Census tract
  • Federal electoral district
  • Dissemination area
  • Forward sortation area
• “000” residue
• Industry codes: North American Industry Classification System
  • Revenue
• Contact us

Introduction and document purpose

This document is intended as a guide for external clients who use or seek to use information provided by the Data Integration Infrastructure Division (DIID) from Statistics Canada’s Business Register (BR).

These products include counts of active businesses by industry, geography and employment range, which can be obtained through statcan.gc.ca or through custom tabulations from the DIID on request.

Other products provided by DIID are the listings of businesses provided to provincial or territorial and federal departments under disclosure orders by the chief statistician authorized under the Statistics Act.

Please note that other Statistics Canada divisions also produce business demographic statistics.Notably, these include information from the Economic Analysis Division and the Centre for Special Business Projects.

DIID experts are available to provide further information on these products as well,or to refer to contacts in these divisions. However, they are not covered in this document.

Context: What is the Business Register?

The BR is Statistics Canada’s internal repository that keeps the baseline information about Canada’s businesses and institutions that is required for the agency’s statistical work.

The BR conceptualizes and maintains information for businesses and their individual operating entities as statistical units of observation. It also provides related statistical attributes, including the industrial activity classification; geographic location; and key measures of business size, such as the number of employees and annual revenues.

The BR is key to the compilation of coordinated, consistent and reliable information from Statistics Canada’s many economic data programs. It is the source of sampling frames for virtually all of the agency’s business and institutional surveys. Also, the BR provides the reference data and linkage keys needed to integrate and enable the use of business data from survey, administrative and other data sources across the Statistics Canada data ecosystem.

The BR can be used to produce business demography statistics, such as counts of active businesses, openings and closures, and other business dynamics.

Data sources and confidentiality

The BR is compiled and maintained primarily from administrative sources, through data Statistics Canada receives from the Canada Revenue Agency (CRA). Through its own processes, including statistical enterprise profiling, survey data collection and data analysis, Statistics Canada builds on the CRA-provided data to provision a fully representative and up-to-date statistical frame of business entities and their operating locations.

All BR data, whether sourced from tax information or Statistics Canada’s own processes, are obtained and maintained by Statistics Canada under the authority of the Statistics Act. Under the act,information related to individual businesses or information that can identify a specific business is strictly confidential and cannot be divulged, other than to a person employed or deemed to be employed by Statistics Canada. Under certain circumstances, specific variables are provided to federal and provincial government partners, as per the conditions outlined in the Statistics Act.

Data Integration Infrastructure Division business data products

Three categories of data products related to businesses are provided by DIID. In most cases, the coverage is restricted to businesses or government institutions meeting any of the following criteria:

• have an employee workforce for which they submit payroll remittances to the CRA.
• are incorporated under a federal or provincial act and have filed a federal corporate income tax form within the past three years.
• have at least $30,000 in annual revenues.

1. Canadian business counts (formerly Canadian business patterns)

The Canadian business counts (CBCs) are produced biannually every June and December and are released in August and February. This product can be used to compare the number of businesses across North American Industry Classification System (NAICS) categories, employment size ranges, provinces and territories, and census metropolitan areas (CMAs) and the census subdivisions (CSDs) within them, as well as all CSDs in Canada with 10 or more active businesses with employees. The following data tables are available to the public on the Statistics Canada website:

• business location counts with employees for Canada, by province, NAICS category and employment size range.
• business location counts without employees for Canada, by province and NAICS category.
• business location counts with employees by CMA, CSD, NAICS category and employment size range.

Website link: The Daily - In the news: Statistics Canada’s official release bulletin.

You can find all iterations of the CBC data product from the link above or through your web browser and search engine of choice using the following terms: “Canadian Business Counts” or “Canadian Business Patterns” (for historic counts).

2. Custom aggregate data tables

For all custom aggregate data table questions or orders, please email statcan.statisticalregisters-registresstatistiques.statcan@statcan.gc.ca.

Employment size ranges

• Units: Location, establishment or enterprise counts
• Geography: All geography
• Industry: All levels of NAICS
• Employment size ranges: Standard 9 ranges or custom 13 or 21 ranges
• Confidentiality measures: None

Revenue ranges

• Units: Location, establishment or enterprise counts
• Geography: Province and census agglomeration (CA) or CMA
• Industry: NAICS-2 and NAICS-3
• Confidentiality measures: Rounding to the nearest five counts

Profit and non-profit data (December only)

• Units: Establishment counts
• Geography: Province
• Industry: NAICS-2
• Confidentiality measures: Suppression

Business type and public and private data (December only)

• Units: Enterprise counts
• Geography: Province and CMA (14)
• Industry: NAICS-2
• Confidentiality measures: Suppression

3. Business Register microdata

BR microdata files are provided to specific federal, provincial and territorial government departments or agencies under specific orders authorized by the Statistics Act. Under the terms of these orders, the use of this information is restricted to statistical or research purposes, and recipients must agree to strict protection of confidential information. Further information is available upon request.

For BR microdata questions or orders, please email statcan.statisticalregisters-registresstatistiques.statcan@statcan.gc.ca.

Other relevant information about the Business Register

Data variations caused by methodological changes, by year

Data are affected by not only business data, but also multiple conceptual and methodological factors, such as ingestion of new data sources; changes to activation, inactivation and reactivation rules; batch fixes and updates for NAICS; new NAICS classification vintages; and new geography vintages.  NAICS and geography are continuously updated, with major revisions every 5 years. These changes will take between 6-18 months to be reflected in our counts. Most recently in June 2022 counts, 2021 Census geographies replaced the 2016 geographies. NAICS 2022 replaced NAICS 2017 for the December 2022 counts.

Below is a list of major methodological changes in reverse chronological order. Please note that many smaller methodological changes occur on a regular basis.  Starting with the June 2022 counts, 2021 Census geographies replaced the 2016 geographies. NAICS 2017 was implemented with the December 2017 counts and has been replaced by NAICS 2022 for the December 2022 counts.

• With the December 2021 counts, refinements to the geographic coding of business locations were made, which may change the geographic classification of some businesses at low levels of geography. Generally, the reclassification is to nearby zones, resulting in only minor impacts on counts for CSDs. There are no impacts on the overall counts for CMAs and CSDs and, by extension, for provinces and territories.
• The December 2019 counts reflect a downward correction to the number of businesses, especially those without employees, because of new criteria for identifying businesses that had become inactive. Approximately 140,000 units are affected by this correction.
• Starting in December 2014, businesses without employees now cover all enterprises that meet one of the following criteria: is incorporated or shows at least $30,000 in revenue (non-taxable or taxable). This change affects businesses that didn’t have $30,000 in taxable revenue in previous years but did have at least $30,000 combined in non-taxable and taxable revenue. These businesses will now be included and represent approximately 600,000 units. Business counts in NAICS sectors 53 (real estate and rental and leasing) and 62 (health care and social assistance) have the largest increases.
• In December 2014, a revision of the employer status on all units of the BR resulted in approximately 70,000 businesses with employees shifting to the businesses without employees category. This change is mostly noticeable in the smaller employment size ranges. Business counts in NAICS sectors 72 (accommodation and food services), 62 (health care and social assistance), 31–33 (manufacturing) and 44–45 (retail trade) see the largest decreases.
• There are two industrial classification categories introduced in 2014: unclassified, which is a new category for businesses that haven’t received a NAICS code, and classified, a category for businesses that have received a NAICS code. The impact of adding the unclassified category is an additional 78,718 locations with employees and 313,107 locations without employees. These counts can be easily identified, because they’re in a separate category.
• A small portion of the increase in businesses in December 2013 is attributable to new rules regarding the acceptance of auto-coded NAICS codes, which resulted in these businesses being included in the data. The impact wasn’t as widespread as the initial NAICS auto-code increase in June 2013—it mostly affected non-employers across most sectors.
• A large increase in the June 2013 reference period is attributable to incorporated businesses that are now required to auto-code a NAICS code to record their tax form information with the CRA. The increase represents an accumulation of about two years of auto-coding. This change affects almost every sector and accounts for most of the growth in the data from December 2012 to June 2013.
• For the first time, the December 2010 reference period includes all unincorporated (T1) businesses with sales of at least $30,000. This integration of T1 businesses is intended to create a more comprehensive representation of the business population on our register. Specifically, this change has mainly affected the following NAICS sectors: 53 (real estate and rental and leasing), 44–45 (retail trade) and 62 (health care and social assistance). The introduction of these units hasn’t had a significant impact on total business counts and represents 1.6% of all locations in December 2010.
• The December 2008 and June 2009 reference periods show a decrease in the number of businesses. This can be attributed to the introduction of new inactivation rules that expand the ability to identify units that aren’t reporting any economic activity.
• The December 2008 reference period introduced the use of “statistical location” counts, besides the usual establishment counts. The use of location counts provides a better measurement of business units. Definitions of the statistical establishment and location are provided later in this document under the “Statistical establishment” and “Statistical location” sections.
• The December 2007 reference period is based on the redesigned BR. The statistical structure (including establishments) has been simplified to better reflect the operating structure of the business. The decrease in the number of establishments is the result of our continuous efforts to detect inactive businesses as early as possible.
• The June 2006 reference period shows an increase in the number of businesses because of a methodological change. There is a new way of identifying newcomers on the BR. The following NAICS sectors have been affected: 48–49 (transportation and warehousing), 53 (real estate and rental and leasing) and 54 (professional, scientific and technical services).
• In December 2000 and June 2005, the number of smaller businesses declined. The BR has analyzed new administrative sources to detect business closures more rapidly and accurately. This has resulted in the use of new signals that are now part of the processes to update the BR.

Data quality and limitations

The BR is largely based on the Business Number (BN) registration source as collected by the CRA.

Time series

Changes to the BR’s methodology or to business industrial classification strategies can cause increases or decreases in the number of active businesses. As a result, the data do not represent changes in the business population over time. Statistics Canada recommends that users not use the data as a time series.

To view recent methodological changes, please refer to the Data variations caused by methodological changes, by year section above.

Creations

Generally, the creation of an entity on the BR occurs shortly after the business registers with CRA. The BN registrations are used to update the BR database weekly. Businesses with multiple locations may also be contacted or profiled to obtain the necessary information for the creation of location entities.

Inactivation

Businesses are assigned an inactive status on the BR when neither a tax payment nor payroll remittance has been made by these businesses for some time or following the closure of CRA tax accounts.

North American Industry Classification System

For newly created businesses, the primary industrial coding is initially processed using automated coding software. This software evaluates the activity description indicated by the business and assigns the appropriate industry classification coding to about 50% of new business records. Activity descriptions lacking precision are subjected to a manual coding process.

Key definitions found in Business Register Data

Statistical entities

Statistical enterprise

An enterprise is the legal operating entity at the top of the operating structure. There is only one enterprise per operating structure. It’s associated with a complete set of financial statements.

Statistical establishment

A statistical establishment is the production entity or the smallest grouping of production entities that

1. produces a homogeneous set of goods or services.
2. doesn’t cross provincial boundaries.
3. provides data on the value of output, together with the cost of principal intermediate inputs used, along with the cost and quantity of labour resources used to produce the output.

Statistical location

The location is an operating entity, specifically a production entity that

1. conducts economic activity at, or from, a single physical location or group of locations.
2. resides within the smallest standardized geographical area.
3. provides employment data at a minimum.

Employment

Source

Employment is based on both corporations’ payroll remittance and profiling and survey data. These data are edited and imputed before being used as input for other processes.

For simple units, attached to only one legal entity, the employment is derived from payroll deductions using the second maximum input within the last 12 months of data. For the complex units, aggregated employment, obtained from profiling, is first determined at the enterprise level. This value is then distributed at the establishment and location levels based on the profiled employment distribution from the BR.

Employment size ranges

The following are the standard employment size ranges (nine) available in the BR:

• 0 (without employees)
• 1 to 4
• 5 to 9
• 10 to 19
• 20 to 49
• 50 to 99
• 100 to 199
• 200 to 499
• 500 and over.

Employment size ranges of 13 and 21 are also available upon request.

Employment size range 13:

• 0 (without employees)
• 1 to 4
• 5 to 9
• 10 to 19
• 20 to 49
• 50 to 99
• 100 to 199
• 200 to 499
• 500 to 999
• 1,000 to 1,499
• 1,500 to 2,499
• 2,500 to 4,999
• 5,000 and over.

Employment size range 21:

• 0 (without employees)
• 1 to 4
• 5 to 9
• 10 to 19
• 20 to 29
• 30 to 49
• 50 to 99
• 100 to 149
• 150 to 199
• 200 to 249
• 250 to 299
• 300 to 399
• 400 to 499
• 500 to 999
• 1,000 to 1,499
• 1,500 to 1,999
• 2,000 to 2,499
• 2,500 to 2,999
• 3,000 to 3,999
• 4,000 to 4,999
• 5,000 and over.

Locations without employees include the self-employed (i.e., those who don’t maintain an employee payroll but may have a workforce that consists of contracted workers, family members or business owners). They also include employers who didn’t report employees in the last 12 months.

Note:

BR employment data should be used with caution. The methodology on the BR used to derive the number of employees for a given business is to select the second-highest monthly value from the last 12 months to reduce volatility for survey sampling. As well, the file contains employment size ranges, which can affect the ability to compile totals. For these reasons, employment data should be used with caveats when attempting to calculate an employment total for any given NAICS category or geography. It is more straightforward to use the data to compile aggregate counts of businesses by NAICS category, geography and employment size class.

Geography

The Standard Geographical Classification (SGC) is Statistics Canada’s official classification for the geographical areas in Canada. It was developed to facilitate the analysis of statistical data using a uniform geographical area definition. It produces a range of geographical areas that are useful for analysis, data collection and compilation on this basis. It is intended primarily for the classification of statistical units such as locations.

A business is geolocated using its available address information. This geolocation process will aim to code the business at the most precise level possible. When insufficient address information exists, the postal code is used as a last resort. Since the postal code is designed by Canada Post to target the efficient delivery of the mail, there are situations where one postal code may not align exactly to the boundaries of a single SGC geographic unit. In such cases, a default SGC is selected for the business. The smaller and rural geographic units are more likely to be subject to this possibility.

For more detailed information, please visit Geographic classifications.

Structure of the Standard Geographical Classification

Each of the three sets of areas covers all of Canada. They are hierarchical: a CSD aggregates to a census division (CD), which in turn aggregates to a province or territory.

(1) Province and territory

The terms “province” and “territory” refer to the major political units of Canada. From a statistical point of view, province and territory are basic areas for which data are tabulated. Canada is divided into 10 provinces and 3 territories.

(2) Census division

“Census division” (CD) is the general term for provincially legislated areas, such as counties and regional districts, or their equivalents. CDs are intermediate geographic areas between the province or territory level and the municipality (CSD).

Usually, they are groups of neighbouring municipalities joined together for the purposes of regional planning and managing common services (such as police or ambulance services). These groupings are established under laws in certain provinces of Canada.

(3) Census subdivision

“Census subdivision” (CSD) is the general term for municipalities (as determined by provincial or territorial legislation) or areas treated as municipal equivalents for statistical purposes (e.g., Indian reserves, Indian settlements and unorganized territories).

Please take note, when using the CSD, of the volatility of the counts between the different reference periods. Units move from one CSD to another, not because of actual changes in physical location, but because of changes in linkages between a specific CSD and a postal code.

Statistical Area Classification

The Statistical Area Classification (SAC) groups CSDs according to whether they are a component of a CMA, CA, or CMA and CA influenced zone (MIZ). The MIZ categorizes all CSDs in provinces and territories that are outside CMAs and CAs. CSDs within provinces that are outside CMAs and CAs are assigned to one of four categories according to the degree of influence (strong, moderate, weak or no influence) that the CMAs or CAs have on them. CSDs within territories that are outside CAs are assigned to a separate category.

The SAC is a variant of the SGC. CSDs form the lowest level of the classification variant. The next level consists of CMAs, CAs and MIZs, including the territories. The highest level consists of two categories that cover all Canada’s land mass:

• inside CMAs and CAs
• outside CMAs and CAs.

The SAC provides unique numeric identification (codes) for these hierarchically related geographic areas. It was established for the purpose of reporting statistics.

Census metropolitan area and census agglomeration

A CMA or CA is formed by one or more adjacent municipalities centred on a population centre (known as the core). A CMA must have a total population of at least 100,000, of which 50,000 or more must live in the core. A CA must have a core population of at least 10,000. To be included in a CMA or CA, other adjacent municipalities must have a high degree of integration with the core as measured by commuting flows derived from previous census place of work data.

If the population of the core of a CA declines below 10,000, the CA is retired. However, once an area becomes a CMA, it is retained as a CMA even if its total population declines below 100,000 or the population of its core falls below 50,000. All areas inside the CMA or CA that aren’t population centres are deemed rural areas.

Other geographies

Economic region

An economic region is a grouping of complete CDs, with one exception in Ontario, created as a standard geographic unit for analysis of regional economic activity.

Census tract

A census tract is an area that is small and relatively stable. Census tracts usually have a population of 2,500 to 8,000. They are in large urban centres that must have an urban core population of 50,000 or more.

Federal electoral district

A federal electoral district is an area represented by a Member of Parliament elected to the House of Commons.

Dissemination area

A dissemination area is a small area composed of one or more neighbouring blocks, with a population of 400 to 700 people. All of Canada is divided into dissemination areas.

Forward sortation area

A forward sortation area is an area composed of the first three digits of the postal code, which is a six-character code defined and maintained by Canada Post for the purpose of sorting and delivering mail.

“000” residue

Please note that codes have been created for residues. They consist of the province or territory code followed by zeroes. This residual category reflects statistical units in Canada where there is insufficient information to precisely locate the locations within a CD or CSD as determined by the SGC.

Industry codes: North American Industry Classification System

NAICS is an industry classification system developed by the statistical agencies of Canada, Mexico and the United States. It’s designed to provide common definitions of the industrial structure of the three countries and a common statistical framework to facilitate the analysis of the three economies. NAICS is based on supply or production-oriented principles to ensure that industrial data, classified to NAICS, are suitable for the analysis of production-related issues such as industrial performance.

For more detailed information, please visit North American Industry Classification System (NAICS) Canada.

NAICS is a system encompassing all economic activities. It has a hierarchical structure.

NAICS is a system encompassing all economic activities. It has a hierarchical structure.

North American Industry Classification System
Sectors	two digits
Sub-sectors	three digits
Industry groups	four digits
Industries	five digits
National industries	six digits

Revenue

These revenues are derived mostly from administrative files from the CRA. They are based on both corporations’ income tax revenues and goods and services tax (GST) sales remittances. These data are at first edited and imputed before being used as input for other processes. For simple units, attached to only one legal entity, the revenue is derived from a regression model using the GST sales as the independent variable, the income tax revenue being the dependent variable. For the complex units, aggregated revenue is first determined at the enterprise level. This value is then distributed at the establishment and location levels based on the profiled revenue distribution from the BR.

Contact us

Dissemination Unit
Data Integration Infrastructure Division
Statistics Canada
Tunney’s Pasture
Ottawa, Ontario
K1A 0T6

statcan.statisticalregisters-registresstatistiques.statcan@statcan.gc.ca

By: Uchenna Mgbaja, Md Mahbub Mishu, Maryam Zamani, Sumithra Balamurugan, and Aya Heba; NorQuest College

As per the 2021 census, there were 5-million rental households in Canada, which means roughly one-third of Canadian households are renters. However, much of this rental activity occurs privately, leading to limited and inconsistent data. To bridge this knowledge gap, we acquired processed, analyzed, and visualized rental listings from the stakeholder – Community Data Program, for Ontario. This dataset offers new insights into spatial trends in metropolitan and small community housing markets, which surpasses other available sources in detail and granularity. Notably, cities like Toronto, Brampton, and Mississauga exhibit high rental prices per square foot, reflecting the region's economic dynamics. We also analyzed areas in Ontario where the population is less than 10,000.

This study aims to address three main objectives:

• interpret trends in datasets and their implications for the housing market,
• apply ML models to the datasets so that the model can forecast future trends, and
• deployment of the best model.

Methodology

We acquired a robust dataset from our client which comprised of 18 columns detailing the regions, bedrooms, addresses, and other pertinent information.

To extract valuable insights, we employed coding techniques and visual representations such as charts and graphs. This helps us to successfully unearth key patterns in housing dynamics, particularly identifying regions with noteworthy differences in housing expenses and density of listings.

Exploratory data analysis

For the exploratory data analysis (EDA), we selected small communities based on their population count. This approach helped us narrow our focus and gain a better understanding of the housing dynamics in these specific regions. However, the "Price" column of our dataset contained inconsistencies such as dollar signs and commas, making it difficult to analyze. To clean this, we removed special characters and converted the column to a numerical format. This enabled us to perform numerical operations and visualize the data, effectively.

Next, we identified that some records in the "Bedrooms" and "Bathroom" columns contained complex entries like "2+ Den," where the regex function only captured the numbers, ignoring the additional "Den." This led to inaccuracies in the representation of bedroom and bathroom counts. To address this issue, we created a temporary column to identify "+ Den" entries, converted 'Bedrooms' and 'Bathrooms' to numeric values, and adjusted the counts to account for the "Den" part. Afterward, we dropped the temporary column, ensuring accurate bedroom counts for each property listing.

The "Size" column contained non-numeric values such as "Not Available" which caused errors when attempting to convert the column to a float data type. To address this issue, we replaced non-numeric values like "Not Available" with NaN (Not a Number) using pandas' replace() function.

Entries in the "Size" column that were below 200 or above 9000 square feet were considered outliers and did not make sense in the context of property sizes. These outliers could skew analysis and visualization results if not addressed appropriately.

Geographical map of rental listings in Ontario

In this section, we used Google's Looker Studio to generate graphs, charts, maps etc., as well as Plotly Express in Python for the visualizations of the dataset.

We created a scatter map (shown in Figure 1 above) using Plotly Express. Each point on the map represents a property listing. We chose an open street map style for clarity and simplicity.

Histogram representing the distribution of rental prices

The histogram (shown in Figure 2) allows users to explore the distribution of rental prices. To ensure that the visualization is intuitive, we maintained clear labels for axes, title, and provide a concise explanation of what the histogram represents.

Scatter plot representing the size and price with the number of bedrooms

The scatter plot (shown in Figure 3) allowed users to explore the relationship between size, price, and the number of bedrooms in a rental property. Users can identify trends, such as how prices vary with the size and numbers of bedrooms.

Box plot representing price distribution based on number of bedrooms

The boxplot (shown in Figure 4) allows users to identify trends in price distribution based on the number of bedrooms. Analyzing outliers can provide insights into exceptional properties and market trends, helping users make informed decisions regarding rental or investment properties.

Bar charts for the top ten most expensive and least expensive regions in Ontario

The bar charts (shown in Figures 5 and 6) provide insights into regional price distributions, highlighting the top ten most expensive and least expensive regions. Users can identify trends, such as regional disparities in rental prices, and make informed decisions by targeting regions with lower average prices for investment opportunities.

Pie chart provides insights into the distribution of property types

The pie chart (shown in Figure 7) provides insights into the distribution of property types in the rental market. Users can identify the most prevalent property type, such as apartments, which accounts for the largest percentage (37.6%). This information can help users make informed decisions, such as best property types for investment or rental opportunities.

Applying Machine Learning to cleaned data

To predict the rental prices, we applied machine learning (ML) models to our dataset. Our data is not time series, as the listings cover various dates without a consistent reference period. Instead, we focused on predictive regression models to predict rental prices which is our target variable. These models helped us analyze and predict price movements based on various features like location, property type, and amenities.

We trained various Machine Learning Models as shown below.

• Regression Models:
  • First, we split the dataset into training and test sets. 80% for training and remining 20% for testing. This approach ensured that the model was trained on a substantial portion of the data while retaining a significant portion for testing.
  • Next, we trained the model using Regressor models- Random Forest, Linear Regression & Gradient Boosting models to predict rental prices – which was our target variable/label.
  • Our next step was to carry out cross-validation and for this we used k-fold cross validation technique to assess model performance and generalization.
  • Finally, we evaluated the models based on the following performance metrics:
    • Root Mean Squared Error (RMSE): This metric measures the average magnitude of the errors between the predicted values and the actual values. The lower the RMSE value, the better the model.
    • R-squared (R2) Score: This metric indicates how well the regression model's predictions fit the actual data. The higher this value, the better the model's prediction.
• Classification Models

To enhance our analysis, we transformed the regression problem into a classification problem by setting price thresholds. Specifically, we categorized rental prices into three distinct groups: low, medium, and high. The thresholds were selected based on the distribution of prices in the dataset, with bins set at 0-1500, 1500-2500, and above 2500. This categorization allowed us to apply classification models, such as Random Forest (RF) and Decision Tree (DT), to predict the rental price categories. This approach is subject to the user's perspective of what is considered a high or low price.

We also developed classification models to predict the rental property type based on the given features. The goal was to recommend a suitable property type based on the user's specifications.

Results

Regression Models

After training and evaluating multiple ML models, it was determined that based on comparative evaluation, the linear regression exhibited superior performance compared to other regression models.

Table 1: Performance from ML models.

Model name	RMSE	R2
Random Forest Regressor	483.05	0.6120
Linear Regression	467.54	0.6568
Gradient Boosting	488.56	0.6372

Description - Table 1: Performance from ML models.

This table compares the performance of three different ML models using two metrics: RMSE (root mean square error) and R^2 (coefficient of determination). The table lists the following models: Random Forest regressor, linear regression, and gradient boosting. The RMSE and R^2 values are provided for each model to evaluate their accuracy and predictive power, respectively. The linear regression model exhibits the lowest RMSE at 467.54 and the highest R^2 value at 0.6568, indicating it performs better compared to the others.

Classification Models

The table below provides a detailed comparison of three machine learning models—Logistic Regression, Decision Tree, and Random Forest—used to classify rental property prices into different categories based on their features. The metrics considered for comparison are the accuracy, precision, and recall scores achieved by each model. These metrics are crucial in evaluating the effectiveness and reliability of the models in predicting rental property prices.

Table 2

Model Name	Accuracy	Precision	Recall
Logistic Regression	0.73	0.81	0.81
Decision Tree	0.73	0.77	0.80
Random Forest	0.74	0.79	0.80

Description - Table 2

This table highlights the performance of different machine learning models in classifying rental property prices. The random forest model outperformed the other models in terms of accuracy, achieving a score of 0.74. Both logistic regression and decision tree models achieved the same accuracy score of 0.73. In terms of precision and recall, the logistic regression model achieved the highest score of 0.81, making it slightly better in identifying true positive instances. This comparison provides valuable insights into the effectiveness of these models in predicting rental property prices and helps in selecting the most suitable model for this task.

Feature selection methods: P-values

The concept of p-values in statistical analysis is fundamental for determining the significance of observed results. In hypothesis testing, particularly in the context of feature selection for ML models, p-values help assess the strength of evidence against a null hypothesis. A low p-value typically indicates that the observed data is unlikely under the assumption that the null hypothesis is true, leading to the rejection of the null hypothesis in favour of an alternative hypothesis.

In the above output, the data frame showcases feature names alongside their corresponding p-values derived from the ANOVA F-test. This statistical technique assesses the significance of individual features concerning the target variable "Price". A smaller p-value signifies a stronger association between the feature and the target variable, indicating a higher likelihood that the feature is relevant for predicting housing prices.

Notably, features such as "Hydro_N" "Hydro_Y" "Size" and various geographical indicators exhibit extremely low p-values, underlining their substantial impact on determining housing prices.

Correlation method

Correlation analysis is a statistical technique used to measure the strength and direction of the linear relationship between two variables. In the context of feature selection for ML, correlation analysis helps identify which features are highly correlated with the target variable and have a significant impact on predicting the target. A correlation coefficient ranges from -1 to 1, where:

• A correlation coefficient of 1 indicates a perfect positive linear relationship, meaning that as one variable increases, the other variable also increases proportionally.
• A correlation coefficient of -1 indicates a perfect negative linear relationship, meaning that as one variable increases, the other variable decreases proportionally.
• A correlation coefficient close to 0 suggests little to no linear relationship between the variables.

In above output, the correlation coefficients between the "Price" (target variable) and other features are listed. "Bedrooms" and "Bathrooms" have relatively high positive correlations with "Price" (0.63 and 0.63, respectively), indicating that as the number of bedrooms or bathrooms in a property increases, the price tends to increase as well.

"Water_N" and "Water_Y" have the same correlation coefficient of 0.35 with "Price" suggesting that the presence or absence of water access might influence property prices to some extent.

Features such as "CSDNAME_South Frontenac" "CSDNAME_Norwich" and "CSDNAME_Chatsworth" have very low positive correlations with "Price" (close to 0), indicating weak linear relationships between these geographic indicators and property prices. We made informed decisions to retain only the most relevant geographic features based on domain expertise. This meticulous feature selection process contributed to a more robust and effective machine-learning model for predicting property prices.

Creating a Rental Housing Insights Application

The creation of the Rental Housing Insights Application represents a comprehensive effort to leverage data science techniques in analysing rental housing data. This section outlines the development process, key features, and the potential impact of the application on stakeholders and the community.

Application Development

The application is developed using the Streamlit framework, using Python. The development process involves several key steps:

• Data preprocessing: Cleaning and formatting the rental housing dataset to ensure data quality and consistency.
• Feature engineering:Creating new features and transforming existing ones to enhance model performance and interpretability.
• ML modeling: Training and evaluating predictive models to forecast rental prices and property types.
• User Interface Design: Designing an intuitive and user-friendly interface for seamless navigation and interaction.

Features

The Rental Housing Insights Application offers the following key features:

• Dashboard: Provides an overview of the project objectives and key findings.
• Exploratory data analysis (EDA): Allows users to explore rental housing data through visualizations.

  

• ML modeling: Enables users to predict rental prices and property types based on input parameters.

  

• Community mapping: Displays rental housing listings on maps, providing spatial insights into market trends.

  

• Google's Looker Studio Integration: Embeds additional insights and reports for enhanced analysis and visualization.

  

User Experience

The application prioritizes user experience by offering an intuitive interface, interactive features, and real-time insights. Users can easily navigate between different sections, customize input parameters, and visualize results in a dynamic and engaging manner.

Impact and Benefits of the App

The Rental Housing Insights Application has the potential to make a significant impact on stakeholders and the community by:

• Providing valuable insights into rental housing trends and patterns.
• Supporting informed decision-making in real estate investments and property management.
• Empowering users with predictive analytics capabilities for strategic planning and resource allocation.
• Enhancing transparency and accessibility of rental housing data for policymakers, researchers, and community organizations.

Future work

The study offers a comprehensive exploration into the rental housing market in Ontario, Canada. By employing an EDA and ML techniques, the authors provide valuable insights into spatial trends, housing dynamics, and rental price forecasts, benefiting both metropolitan and small community markets.

Through meticulous data cleaning, feature engineering, and the application of various ML models, the study sheds light on crucial aspects such as price distributions, geographic influences, and the impact of housing attributes on rental prices. The development of an interactive Rental Housing Insights Application further enhances data exploration, predictive modeling, and spatial visualization, thereby empowering stakeholders with actionable insights and supporting informed decision-making in the rental housing market.

Overall, the study highlights the transformative potential of data-driven approaches in addressing complex societal challenges, such as affordable housing, and emphasizes the importance of collaboration between academia, industry, and government stakeholders to drive positive change in the rental housing landscape.

References

• Min, H., Wood, R., Seong-Jong, J. (2023). Machine Learning Methods and Predictive Modeling to Identify Failures in the Military Aircraft. International Journal of Industrial Engineering, 30(5), 1273-1283. 10.23055/ijietap.2023.30.5.8659
• Belcastro, L., Carbone, D., Cosentino, C., Marozzo, F., Trunfio, P. (2023). Enhancing Cryptocurrency Price Forecasting by Integrating Machine Learning with Social Media and Market Data. Algorithms 2023, 16, 542. 10.3390/a16120542
• Chaudhuri, T., Ghosh, I., Singh, P. (2017). Application of Machine Learning Tools in Predictive Modeling of Pairs Trade in Indian Stock Market. The IUP Journal of Applied Finance, Vol. 23, No. 1.
• Paul RK, Yeasin M., Kumar P, Kumar P, Balasubramanian M, Roy HS, et al. (2022). Machine learning techniques for forecasting agricultural prices: A case of brinjal in Odisha, India. PLoS ONE 17(7): e0270553.
• Pyo S, Lee J, Cha M, Jang H. (2017). Predictability of machine learning techniques to forecast the trends of market index prices: Hypothesis testing for the Korean stock markets. PLoS ONE 12(11): e0188107.
• Rowe, W. (n.d.). Mean Square Error & R2 Score Clearly Explained. BMC Blogs.
• R, V. (2018, September 11). Feature selection — Correlation and P-value. Machine Learning - The Science, The Engineering, and The Ops.
• sklearn.metrics.r2_score — scikit-learn 0.24.1 documentation. (n.d.). Scikit-Learn.org.
• panData. (2023, April 8). Exploratory Data Analysis (EDA): Techniques and Methods for Effective ML Models. Medium.
• Streamlit • The fastest way to build and share data apps. (n.d.). Streamlit.io.
• Looker business intelligence platform embedded analytics. (n.d.). Google Cloud.
• Ashraf, A. (2023, September 22). Correlation in machine learning — All you need to know. Medium.
• P-value in Machine Learning. (2020, July 9). GeeksforGeeks.
• Ray, S. (2019, March 7). 7 Types of Regression Techniques you should know. Analytics Vidhya.
• Brownlee, J. (2023, October 4). A Gentle Introduction to k-fold Cross-Validation. Machine Learning Mastery.

Diversity and Sociocultural Statistics
Statistics Canada

Contents

• Part 1: Introduction
  • 1.1. Background
  • 1.2. The engagement initiative
    • 1.2.1 Activities
    • 1.2.2 Topics
• Part 2: Findings and recommendations
  • Topic 1: Terminology
    • Recommendation on the terminology
  • Topic 2: Groups measured (categories)
    • Recommendation for groups measured (categories)
  • Topic 3: Purpose of the data
    • Recommendation for the purpose of the data
  • Topic 4: Presenting the data
    • Recommendation for presenting the data
• Part 3: Next steps

Part 1: Introduction

1.1. Background

Canada is known for its ethnocultural and religious diversity, which is reflected in the data collected through a national census since 1871. Considering the changing and complex nature of diversity in the country, Statistics Canada has conducted extensive engagement and research to improve the collection and measurement of the ethnocultural and religious diversity of the population and the terminology used to describe it.

The term "visible minority" was first coined by African-Canadian activist Kay Livingstone in 1975, who notably worked to organize a national conference of visible minority women. The term became an organizing tool to challenge unfair institutional practices in education, policing and immigration, and it was soon picked up in the media.^{Footnote 1} The term gained further recognition when it was used by Justice Rosalie Silberman Abella in her report for the federal Royal Commission on Equality in Employment (1984).^{Footnote 2} The document explored the systemic barriers that had, "whether by design or impact, the effect of limiting an individual's or a group's right to the opportunities generally available because of attributed rather than actual characteristics."^{Footnote 3} Abella identified four groups that would require targeted policy measures for employment equity, notably women, Indigenous peoples, people with disabilities and visible minorities. She contended that addressing the systemic barriers experienced by "visible minorities, (…) must begin with an attack on racism."^{Footnote 4}

The report led to the adoption of the Employment Equity Act (EEA) in 1985, which defined visible minorities as "persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour." The law was amended in 1996, but the definition for visible minority remains in effect to this day.

An interdepartmental working group led by the Labour Program at Employment and Social Development Canada (then called Employment and Immigration Canada) was tasked with developing a statistical program. To comply with the EEA, Statistics Canada was mandated to provide benchmarking data, which is collected through the national census. Groups designated as "visible minorities" were identified in 1987 by Employment and Immigration Canada and included in Employment Equity: Technical reference papers. They included, among others, South Asian, Chinese, Black, Filipino, Arab, Latin American, Southeast Asian, West Asian, Korean and Japanese.

Canadian society has evolved since the 1980s, and so have data needs. While data on visible minorities were specifically collected for the EEA, these data are also relevant for developing policies to fight racism and discrimination and to provide equal opportunities for all. They are currently used by governments, businesses, communities, health care providers, researchers and various organizations across the country.

In recent years, Statistics Canada has received feedback on the use of the term "visible minority" when disseminating data, and various stakeholders are calling for the agency to retire this term. There is also a growing demand for more detailed data on the population groups currently designated as "visible minorities" to highlight the diversity of these populations. In other words, policy and program development increasingly requires information on the different groups that make up the visible minority population.

At the same time, many data users have stressed the importance of comparability between census cycles and different data sources, as this might be affected if the census question is changed. These considerations should be taken into account when revising the current standard.

Over time, Statistics Canada has explored various ways of presenting relevant data that reflect the diversity of the population. One approach is to present data on distinct population groups and avoid presenting data on visible minorities as a whole. Furthermore, the agency is increasingly providing intersectional data and analyses to further break down population groups by ethnicity, religious affiliation, place of birth, generation status, language, gender and sexual orientation.

Several factors must be considered when changing or updating a statistical standard used with the census. These include legislative changes, new data needs, findings from consultative engagements, and the results of robust testing to assess the effect of changes to terminology and categories.

In the case of revising the visible minority standard, future amendments to the EEA should be factored in. The EEA Review Task Force (EEARTF) was established in 2021 with a mandate to review the EEA and advise the Minister of Labour on how to modernize and strengthen the federal equity framework.^{Footnote 5} The final report was published in December 2023, along with recommendations.^{Footnote 6} While Statistics Canada is catering to a vast array of stakeholders and recognizes the need to provide data that are well adapted to a range of purposes beyond employment equity, the specific recommendations of the EEARTF will be taken into serious consideration.

Furthermore, data standards used in the national census are developed to work adequately across a wide range of demographics throughout Canada, respecting the highest standards of data quality and confidentiality. The gaps in population characteristics between provinces, notable differences between rural and urban areas, and the requirements to produce and define data in both official languages pose challenges in revisiting the visible minority standard.

Statistics Canada is currently testing different options to update the visible minority data standard in preparation for the 2026 Census. In addition to this targeted engagement initiative, additional consultations were held regarding the content of the previous census (2026 Census of Population Content Consultation). This feedback was reflected in the approaches for the 2024 Census Test. Results of the census test will be used to provide recommendations for the final content included in the 2026 Census questionnaire. While Statistics Canada is following the census timeline, the overall objective is to determine the agency standard that can be used beyond the census, notably in various surveys and in other federal, provincial and municipal departments.

1.2. The engagement initiative

Statistics Canada has committed to consulting with partners, stakeholders and the general public to establish suitable terminology for describing the population, review the categories to reflect its increasing diversity, and offer adequate and flexible ways to present the data for a wide range of purposes.

Over the past several years, there has been a continuous effort to interact with data users on ethnocultural data through informal discussions with government and community stakeholders. Additionally, a focused engagement with government stakeholders took place in the spring of 2021, complemented by the establishment of the Expert Advisory Committee on Immigration and Ethnocultural Statistics.

These initiatives are led in parallel to an extensive consultation program that allows data users and interested parties across Canada to share their views on both the content and dissemination strategy of the Census. The report to the 2026 Census of population content consultation was published in May 2024.

To refine the focus on the visible minority standard and broaden the discourse to include a diverse array of stakeholders, Statistics Canada initiated a consultative engagement on the visible minority concept in October 2022 (Visible minority concept consultative engagement). The aim of this consultation was to solicit input on the visible minority data standard from a wide spectrum of participants, including data users; representatives of civil society organizations; government entities at the federal, provincial and local levels; academics; researchers; and all other interested parties, such as the general public. A range of activities was planned, comprising targeted meetings, two phases of discussion groups and a publicly accessible feedback form. The overarching goal was to leverage this collective input in proposing modifications to the visible minority standard in preparation for the 2026 Census.

The consultation was publicized through public announcements on the Statistics Canada web page and social media outlets. The announcements listed the types of input sought, provided a timeline for the consultation, and gave contact information for interested parties to make submissions or reach out to Statistics Canada with questions and comments.

In addition, stakeholders and partners, including civil society organizations and a number of researchers in the field of ethnocultural diversity, were invited by email to participate and were encouraged to share the consultation invitation with others within their network. Interested parties were invited to participate in group discussions and to submit written proposals to Statistics Canada.

The engagement focused on various aspects of the visible minority standard, notably the terminology used for dissemination, the categories, the purpose of the data, and the derivation of responses leading to the final data dissemination.

1.2.1 Activities

The engagement took place from October 2022 to November 2023. In the fall of 2022, Statistics Canada organized discussion groups comprising academics and key stakeholders in federal and provincial governments. This activity gathered 48 participants.

In May 2023, Statistics Canada organized virtual information sessions with representatives from various federal departments to share preliminary results from the engagement and to collect feedback. This activity gathered 178 participants, who were also invited to share feedback through an online form. Over 70 submissions were received.

A second phase of discussion groups was held in June 2023 to collect feedback from representatives of organizations involved in equity, diversity and inclusion initiatives, including various not-for-profit organizations supporting diverse communities. This activity gathered 57 participants.

Finally, early findings were shared with the general public in November 2023 and promoted on various social media platforms. At that time, any interested party was invited to provide feedback on these preliminary results, informing the conclusions and recommendations of the current report. Statistics Canada received over 200 responses.

In sum, over 500 people participated in the engagement activities, including academics; representatives from federal, provincial and municipal governments; not-for-profit organizations; community-based organizations; and the general public.

1.2.2 Topics

Terminology

The United Nations (UN) Committee on the Elimination of Racial Discrimination has called on Canada (in 2007, 2012 and 2017 (International Convention on the Elimination of All Forms of Racial Discrimination)) to reconsider using the term "visible minority." The committee stated that the term's "lack of precision may pose a barrier to effectively addressing the socio-economic gaps of different ethnic groups." The UN Independent Expert on Minority Issues (Statement by the United Nations Independent Expert on minority issues, Ms Gay McDougall, on the conclusion of her official visit to Canada) has also stated that Canada's use of the term "has served to obscure and dilute the differences and distinct experiences of diverse minority groups." The UN Working Group of Experts on People of African Descent (Statement to the media by the United Nations’ Working Group of Experts on People of African Descent, on the conclusion of its official visit to Canada, 17-21 October 2016) has expressed similar concerns. While some of these considerations were addressed when data were presented for distinct groups that make up the visible minority population, the term itself continues to be criticized by many stakeholders.

As previously mentioned, the Government of Canada announced in 2021 the creation of a task force to review the EEA (Government of Canada launches Task Force to review the Employment Equity Act). The terms of reference recognized some challenges to the federal employment equity framework, including renewed attention to systemic racism and "calls by stakeholders to retire the term 'visible minorities' and rethink the category."^{Footnote 7} In its final report, the task force recommended replacing the term "visible minority" with "racialized worker," noting that "the language of 'visible minority' has been almost universally criticized, and the task force received substantial requests to have it changed."^{Footnote 8} However, while the term "visible minority" is vastly criticized, the results of the current consultative engagement suggest that there is no consensus regarding the most appropriate term with which to replace it.

Categories

As previously mentioned, the current categories (Classification of visible minority) were listed in Employment Equity: Technical reference papers (1987) by Employment and Immigration Canada. While minor changes were made over the years, the categories remained mostly intact to keep comparability between census cycles.

In recent years, various stakeholders have proposed changing the response categories to ensure their coherence, collect more granular data or meet the data needs of specific communities.^{Footnote 9} In the end, different data users have different data needs, and no single list of categories can meet everyone's expectations.

Some stakeholders have pointed out that the current list overrepresents certain groups. For instance, some proposed reducing the number of groups to combine the "Chinese," "Korean" and "Japanese" categories into an "East Asian" category, or to include the "Filipino" category in the "Southeast Asian" category. Another common proposition was to combine the "Arab" and "West Asian" categories into a "Middle Eastern" category. Furthermore, combining categories appears preferable to some data users when the survey sample prevents the presentation of data for smaller groups.

Also, according to employment equity definitions, Indigenous peoples and visible minorities are mutually exclusive equity-deserving groups. Therefore, to avoid response burden, people who identify as Indigenous in the census and in other Statistics Canada surveys have been counted as part of the "non-visible minority" population and skip the census question altogether. Many stakeholders have mentioned that a considerable number of Indigenous people are of mixed background (i.e., Indigenous and non-Indigenous) and that this important dimension of Indigenous identity should be reflected in the data.

One final consideration is that there is currently no single definition or recommended classification for measuring ethnocultural characteristics at the international level.^{Footnote 10} Rather, approaches and criteria vary widely between countries. The choice of terms and classifications depends on various factors, reflecting historical and political developments and legislative requirements specific to each country.

Purpose of the data

The primary objective of the data on visible minorities collected in the Canadian census is to provide benchmarks to meet the requirements of the EEA. That said, data on visible minorities also support anti-racism strategy programs and the measurement of equity and diversity in the labour, social, health, education and justice fields.

Given the recent initiatives by federal^{Footnote 11} and provincial^{Footnote 12} governments to develop anti-racism programs and legislation, stakeholders are increasingly expressing their need for race-based data. While the EEA defines visible minorities as people "other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour," the data collected and disseminated in the census and in other surveys extend beyond racial groups (e.g., Black, White) to include ethnic (e.g., Arab, Latin American), region-based (e.g., South Asian, West Asian) and country-based (e.g., Chinese, Filipino, Korean) groups.

During the current consultative engagement, Statistics Canada gathered feedback on the data needs of participants beyond employment equity to evaluate how the future standard could be more adaptable to various research purposes, including anti-racism, health, education and justice.

Presenting the data

The current approach to presenting the data on visible minorities collected in the census is to derive each respondent in one group, even when they select more than one category (for more information on the derivation method, see Visible minority of person). For instance, a respondent identifying as a Black and White person is included only in the "Black" population category, and a respondent identifying as a Black and Chinese person is included in the "multiple visible minorities" population category. This approach has been criticized by some stakeholders, and alternative options were presented to participants during the consultative engagement.

Another issue is that data on visible minorities divide the population into large categories that do not reflect the diversity of groups (e.g., Black, Latin American, South Asian). Statistics Canada sought feedback on options to provide more granular and flexible data to users and communities.

While the visible minority variable has been the most commonly disseminated variable from the census question since 1996, more detailed data are available through the population group variable^{Footnote 13} contrast to the visible minority variable, the population group variable provides data that more closely match the responses provided by respondents on the census questionnaire, including single and multiple responses. This variable includes 12 single-response categories, counting respondents who reported a single population group. Additionally, it presents a large number of multiple-response categories (67 in total), corresponding to each of the possible combinations of two population groups (e.g., "White and South Asian," "Chinese and Black"). Finally, it includes categories for those who reported three or more population groups ("White and multiple population groups" and "multiple population groups, not included elsewhere"). However, while these data offer granular detail on multiple responses, they are not always suitable for analytical purposes because of the large number of categories.

Part 2: Findings and draft recommendations

Topic 1: Terminology

During the consultative engagement, various terms were presented to participants as potential substitutes for "visible minority," including "racialized group," "population group," "Black, Indigenous and People of Colour," and "ethnocultural group." These terms were presented because they were most frequently reported by various stakeholders before the engagement. No clear consensus emerged from participants, and various considerations were raised.

Several participants preferred the term "racialized group" to replace "visible minority." They noted that the term "racialized" is already used by various federal departments, by provincial and municipal governments, and in the media. They also argued that the term more accurately presents race as a social construct by emphasizing the process of racialization.

However, the term "racialized" was also the most controversial option. Many participants were offended when they were described as belonging to a racialized group. Furthermore, most francophone participants did not think that Statistics Canada should adopt race-based terminology because it is more generally considered to be offensive in French. Regardless of their official language, many participants felt that labelling all non-White people as "racialized" reinforces the White population as the dominant group. Finally, participants also noted the various definitions of "racialization" related to skin colour, culture, religion, ethnicity, language, etc. currently in use.

"Population group" was the second most preferred term. Participants argued that it is sufficiently broad and flexible to apply to a number of situations and to be defined differently according to the needs of different organizations and programs. It was also considered to be a more neutral term that would likely have a longer lifespan, considering the sensitivity of this topic. Participants also noted that the term could include the White population, without making this population either the reference or the norm. On the other hand, some participants opposed this term because of its vagueness.

Recommendation on the terminology

1. Align the terminology in accordance with future amendments to the EEA

In its final report to the Minister of Labour, the EEARTF recommends changing the term "visible minority" to "racialized workers."^{Footnote 14} Because the EEA has enshrined the definition and terminology for data on visible minorities since the 1980s, the recommendation of the EEARTF weighs heavily in the balance. If the recommendation is adopted in a future amendment to the EEA, Statistics Canada should align its terminology and definitions to provide benchmarking data.

While the term "racialized" did not yield a consensus for a term to replace "visible minority," a majority of participants preferred the term over other options. It appears preferable because it acknowledges the effect of racialization on the lived experiences of many Canadians while emphasizing the social and subjective construction of racial categories.

Statistics Canada recognizes that using this term when referring to a specific set of groups is limiting. For instance, these data cannot encompass every group that experiences racialization based on its religion, language or ethnicity.

Also, while Indigenous peoples experience racialization in Canadian society, data on these populations should be separate from other racialized groups to reflect the mandate of the EEA. Similarly, the EEARTF has recommended separating the Black population from other racialized groups. A detailed communication plan will be developed to explain and define the new standard to the public and its relation to the terms and definitions of the EEA.

Topic 2: Groups measured (categories)

During the consultative engagement, no clear consensus emerged on a list of categories to measure groups. Some participants suggested that combining certain categories, as seen in option 2, would be more useful for anti-racism purposes because the resulting data would better reflect how others perceive individuals rather than the respondent's personal identity, which can often be quite specific.

Other participants argued that more detail is always preferable and saw no advantage in reducing the number of categories. Moreover, these participants noted that reducing the number of categories would mean that details for certain groups would be lost (e.g., Chinese, Japanese, Korean, Filipino, Arab, West Asian).

Various participants believed that Indigenous peoples should also be able to identify with a non-Indigenous group because this population is increasingly diverse.

One common criticism was that the categories on both lists are incoherent because they straddle race, ethnicity, nationality and geographical descent.

That said, most respondents felt that historical comparability is important for their data needs and were concerned with the potential effects caused by changing the categories in the questionnaire.

Most respondents believed that some categories are too broad and should be more granular. For instance, various participants expressed that the "Black" category should be further disaggregated. They contended that increased emphasis on this population is justified given the disproportionate effect of racism on Black people. Different subcategories were suggested (e.g., "Black Canadian," "Black Caribbean," "Black East African," "Black West African"). These comments were echoed for the Latin American and South Asian groups, which encompass diverse populations that should be further broken down.

Various participants were also in favour of changing the labels of certain categories, arguing that they should be more relevant and reflective of the self-perception of the concerned communities. Some participants asked to change the label "Latin American" to "Latino," "Latino or Latina," or "Latino, Latina or Latinx." The label "West Asian" was considered inapt to refer to the descendants of this region of Asia, as respondents would not necessarily recognize themselves as "West Asian." The term "Middle Eastern" was preferred, even if other participants denoted a colonial undertone in this terminology. Finally, the term "Arab" was criticized for not adequately representing non-Arab populations of North Africa (e.g., Berber, Coptic). The term "North African" is preferred.

Recommendation for groups measured (categories)

1. Modify the categories to ensure relevance while retaining comparability and data quality

Many participants in the consultative engagement voiced the opinion that categories should be modified to better reflect the diversity of the population. At the same time, most emphasized the importance of keeping historical comparability between census cycles. Meeting both these objectives requires a balanced and measured approach.

To keep historical comparability with past census cycles, the categories in the census questionnaire should remain the same until enough evidence is gathered regarding the effects of potential modifications. A testing strategy will be put in place to evaluate the effect of modifying certain categories, notably "West Asian," "Arab" and "Latin American."

That said, Statistics Canada is currently testing the removal of the "skip" between the Indigenous group and population group questions. If results are positive, Indigenous respondents would be able to identify with one or many non-Indigenous population groups, reflecting the increasing diversity of this population.

Even if the categories remain mostly intact on the questionnaire, different approaches can be adopted at the stage of presenting the data. For instance, a "White" category should be presented instead of the "not a visible minority" (or non-racialized) category. Furthermore, different variables can be created to accommodate a variety of data needs and to further disaggregate the data (see Topic 4: Presenting the data).

Topic 3: Purpose of the data

Participants expressed a wide range of applications for the data collected on visible minorities that are not reflected in the current emphasis on the EEA. Government organizations, academic researchers and community organizations use data on visible minorities in the fields of health, education and justice, and several participants use the current categories as a proxy for "race" data. For many users, data are used for benchmarking specific populations in a context where anti-racism policies are being developed by the federal government and by provincial and municipal governments.

Participants highlighted the need to provide data for a wide range of purposes. In this context, it was recognized that Statistics Canada will face challenges in meeting the needs of different data users. For instance, data users mentioned the challenges of fulfilling the mandates of anti-racism policies with the current standard given the discrimination experienced by certain religious communities (i.e., Muslim and Jewish). Consequently, Statistics Canada heard from various participants that data on religious groups are crucial to understanding the various dimensions of racialization in Canada.

Participants recognized the wealth of data collected through the national census, including data on ethnic and cultural origins, immigration, religion, and language. However, a recurring challenge is that the number of questions included in the census and other household surveys is limited.

Some data users called for Statistics Canada to develop a "race" data standard, with explicit reference to this concept in the question. They contended that Statistics Canada is evading the reality of racism by adopting euphemistic terminologies and has a duty as the national statistical agency to tackle the phenomenon of racism by providing reliable data. However, other participants believed this approach could potentially reinforce racism in society by reifying the concept of "race." As previously mentioned, this important debate is also reflected in conversations on the terminology and the categories.

Various participants questioned whether these data should reflect the respondents' perception of their identity or how they are being perceived by others in society. While self-identification is the primary purpose of the question, participants who are looking to evaluate the effects of racialization in society want the question to measure how people are being perceived and would prefer fewer categories (e.g., "Black," "Asian," "Latino," "Middle Eastern," "Indigenous"). At the same time, many participants advocated for a more open-ended question (i.e., a write-in field without check boxes) allowing the respondent to identify according to their self-perception.

Participants also expressed a clear need for data on racialization and the experience of racism that is not addressed by the current question in the census. Meeting this need would require the development of a larger statistical program, including a sample survey.

Recommendation for the purpose of the data

1. Expand statistical programs to measure racism and discrimination

There are still gaps in capturing the experiences, perceptions and awareness of racism and hate that cannot be addressed through current surveys and administrative data. Filling these gaps will require the development of new questions for Statistics Canada's current social surveys and the advancement of a conceptual framework for operationalizing a measure of racism and discrimination with relevant indicators to better understand equity, diversity and inclusion.

Statistics Canada aims to develop tools to understand and monitor racism, including the process of racialization and its effect on Canadians. Different means to this end are currently being considered and tested, notably new questions to better measure how a person is perceived by others in society (i.e., how the person is "racialized") and the effects experienced because of this process of racialization. While Statistics Canada collects data on experiences of discrimination in the General Social Survey, a more targeted approach could expand the understanding of racism. Also, socioeconomic indicators (labour, education, housing, health) should be factored in to understand and measure the effects of systemic racism in Canadian society.

Topic 4: Presenting the data

In general, participants in the engagement expressed a need for more flexible and disaggregated data. Different options were presented to the participants for their feedback, including multiple response variables and cross-tabulations.

One important consideration is that according to the technical guidelines from the EEA, individuals of mixed backgrounds who select more than one response category are currently derived in one category (see the detailed derivation approach (Visible minority of person)). While the detail for various combinations of multiple responses is available in large data tables (Visible minority and population group by generation status: Canada, provinces and territories, census metropolitan areas and census agglomerations with parts), most data users rely on the visible minority variable. One way to mitigate these limitations is to present the data as a multiple response variable, displaying multiple and single responses. Most participants agreed that this approach would be useful.

Different stakeholders had previously expressed a need to obtain more details for certain groups. Alternate approaches were presented to the participants, offering more granular and flexible data. The first option was a data table cross-tabulating visible minorities by region of birth of parents. Most participants were enthusiastic with this proposition, even if some considered the level of disaggregation to be insufficient.

Recommendation for presenting the data

1. Provide more options to meet various data needs

To provide more disaggregated data, Statistics Canada should leverage the data collected in the census and provide different options to data users. To that aim, more than one variable could be derived to present the data collected through the census question and in other surveys.

While data for the purpose of the EEA would match the operational definition of the legislation (i.e., terminology and categories), a more disaggregated variable could be available to data users. Inversely, a more aggregated variable could be derived when samples are not sufficient to present data for all groups (i.e., combining Chinese, Japanese and Korean into an "East Asian" category; Filipino and Southeast Asian into a single category; and West Asian and Arab into a single category).

Also, in response to data users who want to see the total responses of certain population groups, a multiple response variable could be derived. A multiple response variable presents the sum of single and multiple responses for each group. Total response counts indicate the number of people who reported a specified group, either as their only response or in addition to one or more groups. For an example, see data for ethnic and cultural origins.

Finally, cross-tabulating data with other variables such as the region of birth of parents would provide more granular data. Other possibilities are to cross them with the ethnic and cultural origin, language, or religion variables.

Part 3: Next steps

The public consultation summarized in this report was one of several activities that informed the work involved in the development and updating of ethnocultural data. In addition to the public consultation, Statistics Canada regularly meets with an advisory committee, composed of leading researchers and academics, to discuss its data programs. Ongoing conversations with various stakeholders also inform the development and updating of the visible minority data standard.

Statistics Canada is currently undertaking 2024 Census Test, which will provide recommendations for the 2026 Census. All the comments received during this consultation and other engagement activities were considered, and many are reflected in the recommendations of this report.

Research continues in preparation for the disseminating of data from the 2026 Census and other household surveys. As previously mentioned, modifying the question is only one step toward modernizing the data standard.

Changes to the derivation method, including processing multiple responses, combining response categories when relevant and combining different variables to obtain more disaggregated data, will be investigated in preparation for the disseminating of data from the 2026 Census.

The Canadian farm balance sheet has been designed to: record the value of farm business assets; record the value of farm business liabilities; record the value of equity for farm businesses; display standard financial ratios which are based on estimates from the balance sheet and the value-added account; be based on the establishment concept; display the information at December 31, by province.

Four different balance sheet accounts have been developed in order to separate the assets and liabilities of farm businesses from those of farm operator households and non-operator landlords. Non-operator landlords are individuals or businesses not engaged in the activity of farming who lease assets to farm operators.

In this series, data are provided only for the Balance Sheet of the Agricultural Sector (set 2). This is because set 2 most closely reflects the assets employed in the production of agricultural products. The other sets of balance sheet accounts are available on request. The four sets of aggregate balance sheets produced for Canadian agriculture are as follows:

Set 1 - The Balance Sheet of the Agricultural Sector and Farm Operator Households includes all farm sector assets and liabilities regardless of ownership. It treats the farm operator households and farm businesses as a single entity. It also includes farm real estate assets leased from non-operator landlords and the liabilities outstanding on these assets.

Set 2 - The Balance Sheet of the Agricultural Sector is designed to account for only those farm assets and liabilities used in the production of agricultural products. This set includes farm real estate assets leased from non-operator landlords and the liabilities outstanding on these assets. It also includes automobiles, trucks and farm machinery leased to farm operators. It treats the farm operator households and farm businesses as separate entities, so the personal portion of farm households’ assets and liabilities is excluded.

Set 3 - The Balance Sheet of Farm Businesses and Farm Operator Households reflects the position of farm operators and includes the assets and liabilities of both farm businesses and farm operator households. Thus, the value of farm real estate leased to farm operators by non-operator landlords is not part of this balance sheet. Similarly, the liabilities related to these leased assets are excluded.

Set 4 - The Balance Sheet of Farm Businesses of Farm Operators includes only the farm business portion of assets and liabilities. This account excludes the assets and liabilities of both non-operator landlords and farm operator households.

The layout of the balance sheet follows that recommended in the Farm Accounting Standardization Manual, published by the Farm Accounting Standardization Review Committee, Farm Credit Corporation, 1991. Some definitions and comments on the concepts and ratios have also been extracted from this manual.

Although ratio analysis can assist in managing and analyzing a business, a proper financial analysis of the business requires more tools than just ratio analysis. Consequently, complete reliance upon such financial measures is a very unsound business practice. The four major types of financial ratios which have been calculated include: liquidity, solvency, profitability, and financial efficiency.

The ratios calculated in the tables reflect an aggregate ratio for the agricultural sector. Users should note that the desired and actual value of the ratios will vary significantly according to the type of farming activity (livestock, crop, horticulture, etc.).

When developing and interpreting financial ratios, many limitations must be kept in mind, such as the method of asset valuation; the type, size, and cycle of the business; and the information used to prepare them. Ratios are most meaningful when compared between years. For further information on developing and interpreting financial ratios, refer to the Farm Accounting Standardization Manual.

Users should be especially cautious in using estimates of accounts receivable and cash, bonds and savings data in the provinces of Manitoba, Saskatchewan, Alberta and British Columbia. The estimates of cash, bonds and savings in these provinces may include the value of deferred grain receipts whereas these receipts are generally reported under "accounts receivable". This should not affect the aggregate estimates of current assets.

Current assets are assets which in the normal course of operations are expected to be converted into cash or consumed in the production process within one year. The three components include: cash, bonds and savings; accounts receivable; and inventories.

Accounts receivable are amounts owed to farm businesses, usually arising from the sale of goods or services. Examples include uncollected receipts for grain or livestock sales, or custom work performed.

Inventories are items of tangible property which are held for sale in the ordinary course of business or are in the process of production for such sale, or are to be directly consumed in the production of goods or services. The three categories of inventories are: poultry and market livestock; crops; and inputs.

Poultry and market livestock include chickens, turkeys, slaughter beef heifers, steers, calves, pigs other than boars or sows, and market lambs.

Crops include wheat excluding durum, durum wheat, oats, barley, rye, corn, flaxseed, canola, soybeans, mustard seed, canary seed, sunflower seed, dry peas, chickpeas, lentils, tobacco, and potatoes.

Inputs include feed, seed, fertilizer, chemicals, fuel, and other supplies used for farm businesses.

Prior to 1991, household contents include assets such as furniture and appliances. The farm business portion of household contents refers to office fixtures, equipment and supplies.

Quota is essentially a license, or a right, to sell a certain amount of a specific commodity. This right is regulated by marketing boards. Some quotas (such as milk) are transferable and therefore have a value associated with them. In provinces where quotas are traded, quota values reflect current market values. In provinces where quota trading is prohibited, quota values are implicitly reflected in the value of fixed assets of agricultural holdings because some of the value of quota is normally capitalized into fixed assets.

Breeding livestock consists of animals acquired or raised for the production of progeny, or for the production of a livestock product. Breeding livestock includes bulls, dairy cows, beef cows, dairy heifers, beef replacement heifers, boars, sows, rams, ewes and replacement lambs. Also included are all animals on fur farms at December 31 because all or the great majority of these are breeding stock.

There are three components of machinery: autos, trucks, and other machinery. Other machinery includes equipment for tillage, planting, fertilizing, chemical application, harvesting and haying; dairy, poultry, and other livestock equipment; and other miscellaneous items. As of 1991, computers and other office equipment used for the farm business are included with other machinery.

Farm real estate includes land, service buildings and homes (owned and leased). The value of land includes all farmland operated by farm operators. The value of service buildings includes all buildings except homes. The value of homes includes all farm dwellings occupied by farm operators.

Other long-term assets include long-term investments, as well as AgriInvest balances (beginning in 2008). Prior to the end of the programs in 2007, and the subsequent closure of all the producer accounts in 2009, Net Income Stabilization Account (NISA) balances and, in Québec, balances in the “Compte de stabilisation du revenu agricole” (CSRA), were also included. This series starts in 1991.

Total assets include all tangible and intangible items of value at December 31. It is the sum of current assets, quota, breeding livestock, machinery, farm real estate and other long-term assets.

Current liabilities are payable within the current year. Examples include accounts and notes payable within the year.

Long-term liabilities have a maturity beyond one year from the date of the balance sheet. Examples include mortgages and equipment loans owed by farm operators to various lenders. These lenders include: chartered banks; Farm Credit Canada; the Business Development Bank of Canada; credit unions; treasury branches; federal and provincial agencies; insurance, trust and loan companies; supply companies and private individuals, Veterans Affairs and advance payment programs.

Total liabilities include all obligations of a business arising from past transactions that are to be paid in the future. Total liabilities are the sum of current and long-term liabilities.

Equity refers to the ownership interest in the business. Equity equals total assets minus total liabilities and could be considered to be the owners’ claim against the assets of the business. Equity is increased by the owners’ net contribution of assets to the business and the accumulated net income of the business. As equity is derived residually, by definition any change in the value of assets and liabilities will cause a proportionately larger change in the value of equity.

Liquidity refers to the ability of a business to meet financial obligations as they come due in the ordinary course of business. Three liquidity ratios are calculated using balance sheet values: the current ratio, the acid-test or quick ratio and the debt structure ratio.

The current ratio measures a business’ ability to meet financial obligations as they come due, without disrupting normal operations. If the current ratio is greater than 1, the business is considered to be liquid. A ratio of less than 1 may indicate a potential liquidity problem. Users should note that a favourable liquidity position may be a misleading indicator of the ability of current assets to cover current liabilities because a significant portion of the current assets may be comprised of inventories which may not be easily converted to cash. Also, the value of the ratio may vary depending upon the production cycle, (eg. the ratio may obtain a significantly different result if calculated in the fall when inventories are typically high than in spring when inventories are usually depleted). The ratio is also limited in that it does not predict the timing or the adequacy of future cash flows.

The acid-test (quick) ratio is a variation of the current ratio, and is defined as the ratio of cash, marketable securities, and accounts receivable to current liabilities. The exclusion of inventories in the calculation allows for an assessment of the "immediate" liquidity position of farm businesses. An acid-test ratio of 1 indicates that there are just enough assets of a very liquid nature to cover current liabilities. The desired value of the ratio varies according to type of farming activity. For example, the desired value of the ratio for a dairy operation will be different than for a grain operation. The ratio is also limited in that it does not predict the timing or the adequacy of future cash flows.

The debt structure ratio measures the proportion of current liabilities to total liabilities. This ratio, in conjunction with the current ratio, will provide information on the relative solvency of a business. A high debt structure ratio may indicate solvency problems. However, this may not always be the case, especially for farm businesses with a relatively low value of long-term liabilities. In this case, businesses may have no solvency problems. Thus, it is important to interpret this ratio in conjunction with the value of liabilities and cash flow from farming operations.

Solvency refers to the financial measures that gauge the amount of debt of a business relative to the amount of capital invested in the business. Three solvency ratios are calculated using values from the balance sheet: leverage, equity, and debt. These ratios are indicators of the risk involved in investing in the operation: the higher the debt, the greater the risk to all investors.

The leverage ratio is the value of total liabilities per dollar of equity. The ratio is a measure of the degree to which the creditors have financed the business as compared to the owners. The higher the ratio, the greater is the financing of the farm business by creditors. A leverage ratio of 0.5, for example, indicates that the farm operators have twice as much equity as debt. The higher the value of the leverage ratio, the greater the creditors have financed the farm businesses and thus the higher the risk. The desired value of the ratio will depend upon the income variability of farm businesses and other factors such as the risk associated with production, farm businesses with high income variability or business risk would desire a lower ratio.

The equity ratio is the value of equity per dollar of total assets. The ratio measures the proportion of total assets financed by the owners, as opposed to that financed by creditors. The higher the ratio the more resources are supplied by the owners.

The debt ratio is a measure of the extent of leverage being used by a business, or the proportion of total assets financed by debt. The higher the ratio, the higher is the financial risk.

Profitability refers to the extent to which a business is able to generate profit from the utilization of the business resources. Profitability ratios are calculated using values from the balance sheet and the value-added account because the two series are conceptually and methodologically related. The three calculated profitability ratios are: capital turnover, return on assets and return on equity.

The capital turnover ratio indicates the extent to which a business efficiently utilizes its assets to generate revenue. The higher the ratio the more efficiently assets are being used to generate revenue. The desired value of the capital turnover ratio will vary significantly by type of farming activity. Users should be aware that the ratio is a comparison of flows over stocks, that is, revenues cover an accounting period while total assets refer to a specific point in time. Therefore, the ratio may be misleading in the event that total assets fluctuate significantly in one direction (either up or down) in the accounting period.

The return on assets ratio is a measure of return on investment; it reflects earnings per dollar of both owned and borrowed capital. The higher the ratio, the greater is the return on assets.

The return on equity ratio provides a measure of the return to the owner on the owner’s investment in the business, as it reflects only the return per dollar of owned capital.

Because the value of unpaid family and operator labour is not estimated, the usefulness and the interpretation of return on assets and return on equity may be influenced. Comparisons of these ratios to other return on assets and return on equity ratios should not be made unless the method of calculating the ratios is the same. These ratios do not consider the unrealized capital gains that may be present in the value of assets such as farm land. The higher the value of return on equity, the greater is the return on investment. However, a high value for this ratio may signify a highly leveraged business. Therefore, interpretation of the significance of this ratio should be made in conjunction with other ratios.

The return on assets ratio and the return on equity ratio reflect the different balance sheets. In sets 1 and 2, which include non-operator landlords, the returns include rent to non-operator landlords. In sets 1 and 3, which include the personal share of households, the returns include the family wages.

Financial efficiency refers to the extent to which a business is able to efficiently utilize the businesses resources.

The interest coverage ratio is one of the most widely used financial efficiency ratios for analyzing the ability of a business to pay the interest on debt. Similar to the return on assets or equity ratios, the interest coverage ratio reflects the inclusion or exclusion of non-operator landlords and the personal share of households.