Methodology

• Previous page
• Table of contents
• Next page

Data sources

There are several data sources for this project, including the SMEP, census, and administrative data:

• The Survey of Maintenance Enforcement programs 2005/2006-2015/2016. Only data on payers and recipients of child support were included in the data, not those who paid spousal support.
• The T1 Family File 2006-2016. The T1FF is an administrative data file constructed from Personal Tax files, as well as the T4 annual employer records. The T1FF file is considered a Census of all Canadian tax filers and is based on the Census family concept.^{Footnote 4}
• The Census of Population (long form) 2006, 2011,^{Footnote 5} 2016. The Census is conducted every five years to collect socio-demographic information on individuals living in Canada. The long-form Census is administered to 25% of the Canadian population and includes questions on geographic, economic, and social characteristics such as family structure, immigration, and ethnic and racial identity.^{Footnote 6} With respect to Indigenous status, the data only apply to those residing on reserves that participated in the Census, and participation can vary across census years. However, the analysis does include those individuals who live off reserve and self-identify as Indigenous.
• The Longitudinal Administrative Databank 2006-2016.^{Footnote 7} The Longitudinal Administrative Databank (LAD) is a 20% random sample of the T1FF tax filers and their families that can be linked yearly. Only individuals with a Social Insurance Number are included in the sample. The LAD contains more detailed tax, employment, and geographical information than the T1FF and has been used to study the effects of social policy changes, as well as to better understand social program outcomes and employment/unemployment trends across various geographic levels in Canada. The LAD does not contain information on education levels or demographic characteristics such as Indigenous or visible minority status.

This project was run as a pilot; three jurisdictions signed on to participate:

• Nova Scotia (2005/2006 to 2015/2016)
• Prince Edward Island (2007/2008 to 2015/2016)
• Newfoundland and Labrador (2010/2011 to 2015/2016)

Linkage rates

Through the Social Data linkage Environment,^{Footnote 8} the SMEP data (2005/2006 to 2015/2016) for those who paid or received child support were linked to T1FF (2006 to 2015).

High linkage rates were achieved between SMEP and T1FF. For example, the linkage rate from SMEP to T1FF (tax) was 92%. The linkage rates between SMEP to Census varied from 52% to 88% depending on the Census year and whether it was linked to payers or recipients. For instance, the linkage rates between SMEP and the 2011 Census long form were 52% for payers and 66% for recipients. The lower linkage rates for the census were likely due to the fact that there was only a 25% sample of Canadians reporting on the long form.

The SMEP data were linked separately to the T1FF and Census (long form) to increase sample sizes as the Census data only included 25% of Canadians. Merging across the three files (SMEP, T1FF and Census) would limit the T1FF information only to those who were also included in the Census long form. Table 1 shows the number of records linked between SMEP and T1FF for the three participating jurisdictions from 2006 to 2015. The linked records are for individuals over 15 years old. In addition, Table 1 illustrates that more recipients than payers from SMEP were linked to T1FF.

Table 2 shows the number of records linked^{Footnote 9} between SMEP and the Census long form for 2006, 2011 and 2016. The long form Census is designed as a cross-sectional survey, including information on, for example, education levels, income, and details about home ownership and dwellings.

The non-SMEP comparison group

The SMEP comparison group included persons who were not enrolled in a MEP, but who indicated in the personal tax (T1FF) that they paid or received child support, or spousal support, or both. The SMEP comparison group was created using information from the LAD.^{Footnote 10} To create a SMEP comparison group, members of the SMEP cohort^{Footnote 11} were first removed from each of the eligible T1FF years (2006 to 2015). After the removal of SMEP cohort individuals, those who had paid or received support at any point between 2006 and 2015 were included in the comparison group. Members of the comparison group who filed their taxes in Newfoundland and Labrador, Nova Scotia, or Prince Edward Island at some point over the course of the ten-year period were included. Those who reported having paid or received support were included for the entire ten-year reference period, even if they paid or received in only one of the ten years. Table 3 shows the number of payers and recipients who were not enrolled in a MEP, but who indicated in T1FF that they paid support at any point between 2006 and 2015.

Limitations of the data

Due to the small sample sizes in the Atlantic provinces, along with the Census being a 25% sample of the population, some information produced from the SMEP to Census linkages could not be released because of risks related to information disclosures. Those that could be released, often required response categories to be collapsed to protect confidentiality and to control the risk of disclosure.

For the SMEP group of payers and recipients:

• Only three small jurisdictions were included, which limits the generalizability of the data to the larger Canadian population
• Due to small sample sizes in certain cases the results had to a suppressed or combined with other years or categories of data. These actions will make it more difficult to interpret the results
• For the linked SMEP data payers and recipients were not necessarily related due to
• Not all records in the SMEP could be linked and included in the database, so there is not a 1:1 correspondence in payers and recipients
• Some payers could have had multiple recipients and vice versa

For non-SMEP data that included those claiming to pay or receive support but were not registered with a MEP (in T1FF and LAD), it should be noted that:

• Cases where there was indication of paying or receiving support in at least one year between 2006 and 2016 were included. The results may be less accurate due to this inclusion criteria; even if there was only one year paying or receiving support the case was included
• The payers’ and recipients’ relationship cannot be established for the SMEP comparison group since there were not unique identifiers to bring records together
• The accuracy of the data is unknown, as information on paying or receiving support cannot be verified
• Child support payments are not deductible for payers and do not need to be included in recipients’ income. It is unknown what factors might be associated with reporting child support on income tax forms versus not reporting it. Conversely spousal support payments made under a court order or written agreements are taxable and deductible in most cases^{Footnote 12}

Due to these data limitations, these data cannot be generalized to the rest of Canada, and caution should be used in comparing these groups as there are many unknowns regarding the characteristics of these groups.

• Previous page
• Table of contents
• Next page