Survey of Financial Security (SFS)

Status:
Active
Frequency:
Occasional
Record number:
2620

This survey provides information of the net worth (wealth) of Canadian families, that is, the value of their assets less their debts.

Detailed information for 2012

Data release - February 25, 2014

Description

The purpose of the survey is to collect information from a sample of Canadian families on their assets, debts, employment, income and education. This helps in understanding how family finances change because of economic pressures.

The SFS provides a comprehensive picture of the net worth of Canadians. Information is collected on the value of all major financial and non-financial assets and on the money owing on mortgages, vehicles, credit cards, student loans and other debts. A family's net worth can be thought of as the amount of money they would be left with if they sold all of their assets and paid off all of their debts.

The survey data are used by government departments to help formulate policy, the private sector and by individuals and families to compare their wealth with those of similar types of families.

Subjects

  • Household assets, debts and wealth
  • Household spending and savings
  • Income, pensions, spending and wealth

Data sources and methodology

Target population

The Survey of Financial Security covers the population living in the ten provinces. Excluded from the survey's coverage are: persons living on reserves and other Aboriginal settlements in the provinces, official representatives of foreign countries living in Canada and their families, members of religious and other communal colonies, members of the Canadian Forces living on military bases or in military camps, persons living full-time in institutions, for example, inmates of penal institutions and chronic care patients living in hospitals and nursing homes. Altogether these exclusions represent approximately 2% of the population.

Instrument design

The questionnaire content is determined in consultation with Employment and Social Development Canada, the Bank of Canada, Canada Mortgage and Housing Corporation, Finance Canada and Industry Canada.

Sampling

This is a sample survey with a cross-sectional design.

The 2012 SFS has a stratified multi-stage dual frame design. The overall initial sample size was 20,000 dwellings. The sample was selected as two independent samples from two overlapping frames: the Labour Force Survey (LFS) area frame and a frame constructed from the urban portion, that is Census Metropolitan Areas (CMAs) and Census Agglomerations (CAs), of the 2009 T1 family file (T1FF).

The LFS area frame strata were grouped into urban and rural strata within each province. A sample of 3,860 dwellings was selected from the urban strata and 7,731 from the rural strata.

The urban T1FF frame was stratified by province and four levels of predicted household net worth and a sample of 8,409 dwellings was selected.

All families residing in the selected dwellings were included in the sample.

Data sources

Data collection for this reference period: 2012-09-04 to 2012-11-30

Responding to this survey is voluntary.

Data are collected directly from survey respondents, extracted from administrative files and derived from other Statistics Canada surveys and/or other sources.

Interviews are conducted using a Computer-Assisted Personal Interviewing (CAPI) application. Respondents are encouraged to use bills and other material to provide a more accurate response. The average interview lasts approximately 45 minutes.

Information is not gathered from persons temporarily living away from their families (for example, students at university), because it would be gathered from their families if selected. In this way, double counting of such individuals is avoided.

The interview is conducted with the family member most knowledgeable of the family's financial situation. If necessary, follow-up is done with other family members. Proxy response is accepted. This allows one family member to answer questions on behalf of any or all other members of the family, providing he or she is willing and able to do so.

Use of administrative data:

To reduce response burden, information from personal tax data and the Pension Plans in Canada survey are used.

View the Questionnaire(s) and reporting guide(s) .

Error detection

Treatment of Large Values

For any sample, estimates can be affected disproportionately by the presence or absence of extreme values from the population. In an asset and debt survey, a few extreme values are expected in the sample, as valid extreme values do exist in the population. Values outside defined bounds are identified and reviewed in relation to other information reported for that respondent. If the value is judged to be the result of a reporting or processing error, it is adjusted. Otherwise, it is retained.

Imputation

For income data, all respondents are matched to the tax data file unless they refuse to have their information linked. Data obtained from the tax file are complete and do not require imputation. Only in the absence of tax data are income figures imputed. Donor imputation by the nearest neighbour method is generally used and is performed primarily with Statistics Canada's Census Edit and Imputation System (CANCEIS). However, amounts received through certain government programs such as the universal child care benefit and child tax benefits are derived from other information (i.e. number of children in the household) using a deductive imputation method.

Imputation is also performed for other key variables when information is missing. The imputation of the non-income variables was also done primarily using the nearest neighbour imputation method. In some cases deductive imputation was also used. The imputation is performed in several steps. Variables in the same sections of the questionnaire or in related sections of the questionnaire are often imputed together. An appropriate set of matching variables and imputation classes are determined for each group of variables processed together. Statistic Canada's CANCEIS and BANFF (generalized edit and imputation system) software, as well as custom SAS programs were used to perform the imputation of the non-income variables.

Imputation is also used to handle item nonresponse for many, but not all, variables. Variables not subject to imputation retain missing values in the form of "Don't Know", "Refused" and "Not Stated" reserve codes. Imputation rates for questions which applied to the majority of respondents tended to be quite low, but items that applied only to a small percentage of respondents sometimes had fairly high imputation rates indicating that the respondents were having much more difficulty answering or were less willing to provide answers to these questions.

Estimation

An integrated weight, meaning that all household members are given the same weight, is produced for SFS. The weighting process begins by calculating design weights separately for the two samples. The weights are then adjusted for non-response separately within each sample. In order to combine the two samples, the weights must be adjusted to take into account the fact that the dwellings in the overlap of the two frames have a chance of being selected in both samples. The weights are then calibrated to known population totals. The totals include demographic projections produced by Statistics Canada's Demography Division based on the 2006 Census, as well as the number of wage and salary earners by 7 wage classes by province based on the Canada Revenue Agency's T4 file. The demographic totals for each province include age/sex counts as well as household size and family size counts. Influential observations are then identified and weights are reduced for a small number of extreme observations. The weights of the remaining observations are then adjusted with another round of calibration.

A corresponding set of 1,000 bootstrap weights is produced which can be used to estimate sampling variance.

Quality evaluation

Data are compared to the results of other data sources: Census, administrative and other Statistics Canada surveys.

Direct comparisons with outside sources, such as the Financial and Wealth Accounts of the System of National Accounts, are difficult to make due to definitional, coverage and treatment differences. However, based on rough comparisons the following general conclusions can be drawn:

(a) SFS appears to underestimate some net worth components, particularly financial assets and consumer debt.

(b) The quality of estimates of real assets (e.g., owner-occupied homes, vehicles) is much better than that of financial assets.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

Data accuracy

The reliability of estimates produced from survey data depend on both sampling and non-sampling errors.

Sampling errors occur because data from only a sample of the population, rather than the entire population, are used to draw inferences about the population as a whole. The magnitude of the sampling error is affected by the sample design, the variability of the characteristic being measured and the sample size. The sampling error is often reported through the use of standard errors or coefficients of variation (CVs). The CV of an estimate is expressed as a percentage and is defined as the standard error of an estimate divided by the estimate itself. Examples of CVs for three important SFS variables at the national level are: total assets 1.5%, total debts 2.4% and net worth 1.7%.

Non-sampling errors occur in the form of coverage errors, non-response errors, response errors, other forms of collection and processing errors. Non-sampling errors that occur randomly have minimal impact on survey estimates. Systematically occurring errors could however lead to biases in survey estimates.

Coverage errors (both over and under-coverage) arise when the survey frame does not accurately cover the target population. The slippage rate, a measure of coverage error, is defined as the relative difference between the pre-calibration weighted counts and a set of population totals originating from an outside source. The 2012 SFS observed an overall slippage rate of 11.2%. Slippage rates range from -10.1% in Newfoundland and Labrador to 18.0% in Alberta. Calibration of the survey weights to control totals was employed to reduce the coverage error. Within each province, the control totals included demographic projections of age/sex counts and counts by family and household size as well as counts of wage and salary earners by earnings classes.

Both total and partial nonresponse can also impact the quality of survey estimates. The lower the response rate the greater the potential for nonresponse bias. Total nonresponse is dealt with by adjusting the weights of the respondents to account for the nonrespondents. The overall SFS response rate was 68.6%.