Directorates
·	Director General's Office
·	Data Development
·	Strategic Evaluations
·	Partnership and Continuous Evaluations
Publications
·	Evaluation Reports
·	Evaluation Briefs
·	Labour Market Development Agreement Series
·	Lessons Learned Series
·	Evaluation Tool Kit Series
·	Evaluation Forum Newsletter
·	From Theory To Practice

This section provides a profile of subscribers, examines the characteristics of their RESPs and beneficiaries, and looks at the levels of grants provided under the CESG.

5.1 Socio-Demographic Traits of RESP Subscribers

RESP subscribers are compared to non-subscribers based on results of the RESP subscriber survey and the non-subscriber survey, supplemented by the 1999 SAEP data and the CESG/LAD1999 administrative data,¹⁷ noting again that the latter data are for taxpayers who are not necessarily representative of all households.

Multivariate analysis is also presented to corroborate key findings from the surveys. This type of analysis permits identification of the unique role played by a particular variable on RESP contributions, while controlling for the effect of other variables. The multivariate analyses uses the SAEP 1999 data¹⁸ and CESG/LAD 1999 administrative data.¹⁹ The focus in this section is on RESP subscribers and non-subscribers who are parents:

• The subscriber survey data indicate that most subscribers are parents (87 percent), only five percent are grandparents, and six percent are unrelated to their beneficiary; and
• The non-subscriber survey included parents and grandparents.²⁰ Grandparents are over-represented because the relative proportions of grandparents and parents in the non-subscriber population were not known at the time of the survey (so quotas could not be set).

The analysis of subscriber and non-subscriber survey data presented here is limited to parents,²¹ to maximize the comparability of findings from the two surveys.

Perceived Targets of the Program

The general view among informants is that the CESG is targeted to all Canadians.

The informant analysis indicates that five (out of six) program officials and eight (out of 12) non-government financial or research experts consider the CESG to be directed to all Canadians. This corresponds to what the documentation indicates is the target of the program. Five of the 20 informants suggested the program is directed at low to middle-income Canadians (i.e. one program official, two non-program government officials, and two financial experts).

Just over half of the informants (11 of 20) believe the program is benefiting those it is intending to benefit to a large extent (i.e. responding with 6 or 7 on a 7-point rating scale rating of the program, where 1= to no extent and 7= to a great extent).

Parents' Age and Family Composition

Parents who are RESP subscribers are more likely to be age 35 or older, in two-parent families, and with two children.

Comparing subscribers to non-subscribers indicates that subscribers are somewhat older (37 percent versus 24 percent are 45 years and older), as shown in Table 5.1. Parents contributing to RESPs are more likely to be married or common-law couples (90 percent) compared to their counterparts who do not save using RESPs (74 percent). Contributing parents are more likely to have two children, while non-subscribers are more likely to have one child.

These findings are generally confirmed by the merged CESG/LAD data. RESP take-up rates (the proportion of all 1999 taxpayers who contributed to RESPs between 1998 and 2001) are higher among married/common-law parents compared to single parents and in families with two children compared to those with one child or three or more children (as shown in Table 5.1). Also, the take-up rate peaks in the 25 to 44 age bracket. Data not shown indicate that the take-up rate falls to 7.1 percent for the 54 to 64 age group.

These findings are also confirmed by the multivariate analysis based on the SAEP data. After controlling for other factors, the SAEP analysis indicates that being a dual parent, having two children, and being older than 35 years of age increases the likelihood of RESP savings. On the other hand, heading a larger family, especially with more than three children, decreases the likelihood of RESP contributions.

Characteristics of Children

Children more likely to have parents contributing to an RESP on their behalf are younger, performing well in school and expected to attend university.

Results of analysis of SAEP data indicate that the characteristics of the child play a role in RESP contributions.²² The SAEP data indicate that the percentage of children who have RESP savings on their behalf falls with their age (i.e. from 19 percent for children under five years of age, to 10 percent for those 15 years and over).

The multivariate analysis confirms this finding. Controlling for the influence of other variables, the likelihood of a child having parents contributing to a RESP for him/her declines with the child's age. The multivariate analysis also indicates the powerful role played by other child-based characteristics, particularly parental aspirations for the child's future PSE and the child's performance in school. Specifically, the higher the level of education expected of the child and the better the child has performed in school, the greater the likelihood of RESP contributions.

Key Factors Affecting CESG Take-Up

Key factors affecting take-up include parent's education, age and school aspirations for their children, the child's performance in school, and province of residence. RESP contributions rise with income and are significantly lower than the population share for parents with low household income and higher for those with high household income.

Multivariate analysis identified certain characteristics of subscribers as being influential in predicting RESP take-up, after controlling for the influence of other factors. Among the strongest predictors of whether or not parents will contribute to an RESP on behalf of the child are the child's school performance and their parents' expectations that they will attend university. Other predictive factors include parents' education, their age (over 35 years) and not living in Quebec. The lower RESP take-up rate in Quebec is likely attributable to the province publicly funded college system (CEGEP) and relatively low university tuition fees for Quebec residents. The effect of income, although significant, is weaker than these other factors.

As shown in Table 5.2 below, the share of RESP contributing households during the period of 1998 to 2001, was very low for households with $20,000 or less of pre-tax income (8.6 percent) in comparison to their share of all households (33.6 percent). Households in the $20,000 - $39,999 income category have a share of RESP contribution that is slightly less than their share of households while households in the $40,000 - $79,999 categories have a modestly higher share of RESP contribution than their share of all households. For households above $80,000, their share of RESP contributing households is much higher (36.2 percent) than their share of all households (16.5 percent).

The foregoing analysis indicates the importance in a summative evaluation to undertake segmented analysis of the factors affecting take-up for various characteristics of subscribers (e.g. education, age, income). This analysis would possibly reveal important factors that predict RESP take-up among different subscriber groups.

Parents' Education

RESP subscribers tend to have much higher educational attainment levels than non-subscribers.

The survey data show that 81 percent of parents who save using RESPs have some form of PSE, with almost half (47 percent) having a university degree. In contrast, 64 percent of non-subscriber parents have some form of PSE, with 25 percent having a university degree (as shown in Table 5.3).

Once again, the SAEP data (both descriptive and multivariate analysis) confirm the survey results. Parents with a university education were considerably more likely to save using RESPs compared to parents who save for their child's PSE using other means and those who do not save for their child's PSE at all.

The multivariate analysis indicates that the role played by parents' education in the decision to have an RESP is stronger than the influence of the parents' income, except when they are in the lowest income category. In other words, even within different income groups (except the lowest income category), the higher the level of education the parents have attained, the greater the likelihood the parent has RESPs on behalf of their children.

Location

The RESP take-up rate is higher for parents living in Ontario and British Columbia, and lower for parents living in Quebec.

The survey data also indicate regional differences between RESP contributors and non-contributors (as shown in Table 5.3). Residents from Ontario represent a much higher proportion of subscribers than non-subscribers (42 percent and 31 percent respectively), while Quebec parents represent a much lower proportion of subscribers than non-subscribers (16 percent and 26 percent respectively).

The same patterns are observed in results based on the SAEP and CESG/LAD data. Table 5.3 shows that RESP take-up is lower in Quebec (6.6 percent) and higher in Ontario (10.1 percent) and British Columbia (10.2 percent). Quebec's effect on the incidence of RESP contributions is further confirmed by multivariate analyses.

The lower RESP take-up rate in Quebec is likely a reflection of the province's publicly funded college system and relatively low university tuition fees.

Parents living in urban areas are more likely to contribute to RESPs than those living in rural areas.

Other results based on the tax data indicate that urban residents are more likely to contribute to RESPs than rural residents (9.8 percent and 7.0 percent respectively). In the multivariate analysis, the urban factor maintains its effect on the likelihood of having an RESP, when controlling for the effect of income characteristics, but disappears when parents' education and their attitudes towards their child's future PSE are taken into account.

Equity Group

Eleven percent of RESP subscribers consider themselves to be members of equity groups, compared to six percent of non-subscribers.

The survey evidence indicates the proportion of respondents considering themselves to be a member of visible minority is somewhat higher (11 percent) among RESP subscribers than non-subscribers (6 percent), as shown in Table 5.3. There is little difference in Aboriginal status between the two groups. The CESG/LAD data indicate considerably higher RESP take-up among immigrants (16.1 percent) and somewhat higher take-up among those with a disability (10.1 percent) compared to Canadians overall.

Type of Income/Investment

The RESP take-up rate is higher for parents with investment and/or rental income and those donating to charities.

Other analysis based on the CESG/LAD data indicates that the RESP take-up rate among those with different sources of investment income is higher than those without such income. This suggests that contributors to RESPs generally invest in the market. The RESP take-up rate is much higher for those with interest, rental and dividend income compared to the overall incidence. It is also higher for those who make RRSP contributions, as well as those who make charitable donations. While Canadians in receipt of a public pension (CPP/QPP) are much less likely to contribute to an RESP, Canadians receiving pensions from their employer (pensions or superannuation) are slightly more likely to be contributors.

Multivariate analysis of RESP incidence based on the CESG/LAD database confirmed that having investment and rental income and donating to a charity play an important positive role in the decision to contribute to RESPs. Total household income loses much of its explanatory power when controlling for the effect of all explanatory variables included in the model, including family type (two parents versus single parent), not living in Quebec, and, to a lesser degree, parents' age.

5.2 Characteristics of Plans and Beneficiaries

In this section a number of basic characteristics of plans and beneficiaries are presented. Evidence is taken from the Survey of Subscribers and the CESG administrative data. The evidence is presented for all subscribers (i.e. parents and non-parents).

RESP subscribers have been contributing to RESPs for an average of six years. Almost all have one or two plans and deal with only one promoter or trustee (financial institution). Most have one or two beneficiaries. Average contributions to group plans are the lowest among the three types of plans, reflecting the lower income of those holding these types of plans.

Duration: On average, subscribers have been saving using RESPs for six years. Almost one-half (47 percent) have been saving for 3 to 5 years, with almost two-thirds saving for five years or less, and 15 percent have been saving for longer than 10 years. Parents have been saving longer than non-parents. The proportion of subscribers saving for a relatively short period of time falls with the number of children, and with the age of the parents and the children. The proportion saving for a longer period of time is higher for university-educated and higher income subscribers, for subscribers using other savings vehicles for PSE, and for family plan holders.

Number of plans: Two-thirds of RESP subscribers have only one plan, and another one-quarter (24 percent) have two plans. The remaining 10 percent have three or more plans. On average subscribers have 1.6 plans each.

Type of plan: Table 5.4 indicates that almost half (45 percent) of all plans are individual family plans. One-fifth (19 percent) are individual non-family plans. The survey of RESP subscribers also indicates the following:

• Individual non-family plans are more likely to be found among residents of Quebec and households with older children;
• Group plans are more likely to be found in Quebec and the Atlantic provinces, and among those with lower levels of education and income; and
• Individual family plans are more common among residents of the Prairies and Ontario, university degree holders, and those with a total household income of $120,000 or more.

Type of promoter/trustee (financial institution): Table 5.4 indicates that there are three main types of financial institutions among RESP promoters/trustees: banks (33 percent); scholarship foundations (30 percent); and investment management companies (21 percent). Other results indicate that banks and fund management companies are favoured among holders of individual non-family and family plans. Not surprisingly, scholarship foundations are greatly favoured among the promoters/trustees of group plans (otherwise known as scholarship funds).

Number of promoters/trustees: 91 percent of subscribers deal with only one promoter/trustee (based on the administrative data).

Number of beneficiaries: 89 percent of subscribers have one or two beneficiaries (i.e. 53 percent of subscribers have one beneficiary and 36 percent have two beneficiaries).

• It is interesting to note that parents do not always establish RESPs for all their children. Forty-one percent of parents with two children have one RESP beneficiary (average=1.43 beneficiaries), and only 32 percent of those with three or more children have three or more beneficiaries (average=1.85 beneficiaries).
• Children without RESPs in families where siblings have RESPs would likely include those recently born or already in PSE, as well as those who may not be performing well in school and/or those whom the parent does not expect to go to PSE.

Age of beneficiaries: The survey data indicate that the proportion of subscribers (parents and grandparents) with children in different age groups rises from eight percent for subscribers with children two years and younger; is 17 percent in the 3 to 4 and 18 percent in the 5 to 6 year age categories; is 19 percent to 21 percent in the 7 to 9, 10 to 12 and 13 to 15 year age categories; falls to 14 percent in the 16 to 17 age group; and is nine percent of those with children in the 18 years and over age group.

• SAEP data corroborate the drop off in RESP incidence from age 15 and on. The SAEP data further indicate that the proportion of children with RESP savings rises somewhat between the ages of one and two. This suggests that parents wait for a period of time after birth, possibly to apply for the child's SIN, which is a requirement for establishing an RESP for a child.
• The administrative data indicate that the average age of beneficiaries is 10.3 years old and varies slightly across all plans, with beneficiaries of individual non-family plans being the oldest (10.7 years old) and their counterparts with group plans being the youngest (9.5 years old).

5.3 Grant and RESP Contribution Levels

The evidence suggests that the introduction of the CESG increased RESP contribution levels.

Table 5.5 presents data on grants paid and RESP contributions at the subscriber level (i.e., across all plans held by a subscriber) based on CESG administrative data.²³ The results cover the years 1998 to 2001, with the latter representing only a partial year of data. Table 5.5 indicates that average assisted contributions (attracting a grant) rose steeply from the initial year of the program (1998) to the second year when the program was fully operational (rising from $1,640 to $2,105). This suggests the implementation of the CESG had an impact on contribution levels, although average contributions fell somewhat to $1,945 in 2000. The average grant allotted in the last complete year for which CESG data were available (2000) was $389 per subscriber. The time series information on grants paid mirrors that of RESP contributions, since the grant paid is equal to 20 percent of the contribution.

The foregoing provides evidence of the potential that CESG led to an increase in savings of PSE. However, the study methodology does not allow for a determination of the extent to which increases in contribution levels are attributable to the program.

In 2000, the average contributions per subscriber were lowest (by a large amount) for those who held group plans ($1,037), reflecting the fact that the incidence of this type of plan is highest in low-income groups. The opposite is true for individual family plans ($2,840), which are most likely to be found in higher income groups.

Subscribers from Ontario were, on average, the biggest contributors in 2000 ($2,096), although only somewhat higher than the average contribution of residents of the Prairies and British Columbia. Subscribers in the Atlantic Provinces contributed the least on average ($1,445), with Quebec at about the same average ($1,576).

Footnotes

17	Data provided in the merged dataset (i.e. CESG administrative data linked to 1999 taxpayer data) did not permit the computation of annual RESP incidence rates for population sub-groups (e.g., in different income categories), just the proportion contributing to an RESP at any time over the 1998-2001 period.
18	It should be noted that when conducting multivariate analysis of the SAEP data, a wide range of different factors were included in the model used to explain parents' decision to save for an RESP. The model incorporated not only socio-demographic characteristics such as parents' education and household income, but also the child's characteristics (age, gender, and performance at school), parental expectations and attitudes regarding child's education, and parental awareness and attitudes with respect to financing PSE.
19	It should be noted that, although the LAD/CESG database had basic household characteristics and extensive information on income over the last five years, it lacked vital information such as parents' education and parental expectations and attitudes regarding their child's education.
20	It was determined that it would be difficult to find non-subscribers who would be the equivalent to subscribers who are not related to their beneficiaries by blood (e.g., aunts, uncles, godparents).
21	While the exclusion of non-parents decreased the average age of non-subscribers by about five years and reduced the size of the samples of subscribers and non-subscribers by 13 percent (n=1755) and 25 percent (n=755) respectively, it had little effect on overall distributions of other socio-demographic characteristics such as income, education, region and equity group status.
22	The evaluation survey of subscribers asked only about the age of RESP beneficiaries, not the age of all children in a family, whereas the non-subscribers survey asked about the age of all children. This is why children's age is not compared between subscribers and non-subscribers. The age of beneficiaries is considered in the next section on beneficiaries.
23	Based only on plans where non-zero contributions were made. This computation does not control for the number of plans a subscriber might hold (e.g. contribution levels would be expected to rise with the number of plans). The fact that subscribers could have several plans per beneficiary renders the computation of per-plan contributions difficult and the results difficult to interpret.