Source of income data

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

For the first time, 2005 estimates of income from all of the above programs were based largely on income tax data.  In order to minimize response burden and increase quality, SLID has, since 1994, asked respondents for permission to use the income data from their T1 return, rather than reporting this information directly. Approximately 80% of respondents chose to do this.  Because of the success of this approach, the Census, for the first time in 2006, also asked respondents for permission to use tax data; 82.4% gave permission¹. The primary source of income data, for the T1FF and the LAD, is also the T1 file. And, finally, the SNA's Income and Expenditure Accounts derive income estimates for the PUBS largely from administrative data such as, but not exclusively, the personal T1 Income Tax and Benefits Return.

There are advantages to using income data from tax records rather than from survey responses. One is reducing respondent burden, there is no need to engage respondents in a recollection of data that is readily available. A second advantage is improving the quality of the collected data as it is believed the tax records generally provide a better picture of a respondent's economic activity than the survey questions.

The survey mode is particularly vulnerable to response errors, either because the respondent was unwilling or unable to provide the information requested, or because the respondent misinterpreted a particular question. Illustrations of this type of error could be forgetting a small amount, reporting a rounded amount (to the closest $100, $1,000 or $5,000), not computing allowances for depreciation when reporting self-employment income, etc.

Because of the requirements of the Income Tax Act, tax filers are given precise guidelines to follow when completing their tax return. These guidelines, the presence of documentation (information slips) from the employers or payers and the review by an income tax preparer and/or CRA help to ensure that income is reported consistently.

However, the differences between the two modes are not all in favour of tax data.  Some respondents do not permit use of their income tax records when offered the option as indicated above.There may also be issues finding the correct tax record since the respondent does not provide their Social Insurance Number and probabilistic linkage of the files is required². The respondent may not have filed a tax return. There are also conceptual difficulties mapping the administrative fields to the desired income components. For example, some income components are non-taxable (or only partially taxable) and may not be required on the tax return³. Others may have an allocated space for reporting but since they are not crucial to the calculation of income tax, they are subject to less scrutiny by CRA. Some in-kind benefits offered by employers are taxable and have been accepted as income as it is difficult to separate the in-kind components on the tax returns.
 
Additionally, there is the possibility that income may not be fully disclosed on the T1 form. This may be an issue in surveys as well.

Although the use of a common input file and collection methodology for most of the records helps to minimize the differences in the estimates produced by the different programs, there are other reasons for differences, which are discussed below.

---

Notes

1. For a discussion of permission rates seen in the early years of SLID, consult Results of the Tax Permission Questionnaire in the Survey of Labour and Income Dynamics E. Abraham et al., Income Research Paper Series, Statistics Canada catalogue no. 75F0002-No. 002, March 2001.
2. For the 2006 Census, 89.1% of the consenting respondents were found in the 2005 tax file. A slightly higher rate was obtained for SLID as manual verification of uncertain links is possible with that sample size.
3. Such as child benefits, veterans' pensions and scholarships or bursaries less than $3,000 in 2005.