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

To develop a profile of LMDA program and service participants, this chapter outlines sociodemographic data, for both participants overall and for users of individual programs, based on the most part on administrative data supplied by Human Resources Development Canada (HRDC). Comparisons are also drawn between active EI claimants on the one hand and reachback and comparison groups on the other.

3.1 Comprehensiveness of LMDA Participant Data

As shown in Exhibit 3.1, HRDC administrative data provided to Ekos for purposes of this evaluation were relatively complete for participants' gender and age, and to a lesser extent, spoken language. There was far less information available (information for only about one in four individual participants) regarding those who may have been disabled, a visible minority, a social assistance recipient, or Aboriginal. Education data were available for only one of every seven participants. Because of concerns over measurement bias, profile results for variables other than age, sex, language and income (see down) will not be reported based on the administrative data.

3.2 PBM Activity

By a significant margin, the intervention most likely to be used was Skills Loans and Grants (SLG). A total of 54.5 per cent of LMDA participants were involved in SLG, compared to 12 per cent for Partners (Exhibit 3.2). Other benefits and measures were used much less frequently: six per cent participated in Job Action; three per cent participated in Employment Assistance Services; and two per cent participated in the Entrepreneur benefit. Twenty-two per cent participated in Rural Experience, which although it is not an LMDA Provincial Benefit or Measure, is included here because it is partly funded under the Research and Innovations component of the LMDA. Results on activity by type of intervention are presented in Appendix B.

Most (about nine in 10) LMDA participants engaged in only one intervention (Exhibit 3.3). The vast majority of others (9.4 per cent of all LMDA participants) participated in two interventions, meaning that only one per cent of all participants had three or more interventions.

3.3 Profile of LMDA Participants Overall

About twice as many men than women participated in Cda/New Brunswick LMDA benefits and measures (Exhibit 3.4). Almost two-thirds (64.1 per cent) of participants were men, compared to one third (35.9 per cent) who were women. It is important to note that the proportion of women accessing benefits and measures is similar to the overall proportion of women who made new claims for EI in 1997 and 1998⁸ (36 and 40 per cent, respectively).

Participants tended to be relatively young — under one in five (17.2 per cent) were 45 or older. Moreover, a similar number of participants (41.2 per cent) were between 30 and 44 as were under 30. The youth participants were just as likely to be under 25 (21.8 per cent) as between 25 and 29 (19.8 per cent) years. A solid majority (72.1 per cent) of participants identified themselves as English-speaking. Just over one quarter (27.8 per cent) indicated that they were French-speaking.

The mean total income of participants decreased in the period leading up to the intervention (Exhibit 3.5).⁹ The average value of Employment Insurance (EI)period benefits and social assistance benefits received fluctuated, and generally declined, during that time period. EI benefits were an average of $650 lower in 1996 than in 1994, while social assistance benefits were an average of $430 lower over the same time period.

3.4 Profile of Participants by PBM

Sociodemographic differences between participants by program type are detailed in Exhibit 3.6 based on HRDC administrative data. Information is presented with these two variables because of deficiencies in the other variables as noted in Exhibit 3.1. The total column of this exhibit repeats some of the results presented in Exhibit 3.4. A more comprehensive profile of participants in each benefit and measure, based on the survey data, is presented in Chapter Six.

A description of participation in each benefit and measure incorporating these results follows.

• Partners participants were more likely to be male (59.6 per cent versus 40.4 per cent female), although somewhat less so than the overall mean (Exhibit 3.6). Just under half of program participants (46.3 per cent) were between 30 and 44 years of age.
• As shown in Exhibit 3.6, Entrepreneur participants tended to be older (three quarters were 30 years of age or older) than the average LMDA participant. Participants' gender mix was, however, close to the overall distribution (61.1 per cent male, versus 38.9 per cent female).
• Job Action recorded the lowest male participation (46.8 per cent male versus 53.2 per cent female) (Exhibit 3.6). Participants' mean age tended to be near the overall mean age.
• The most widely used program, Skills Loans and Grants (SLG), had almost twice as many male participants as female (65.7 per cent male versus 34.3 per cent female). This program had the highest proportion under 25 years of age (24 per cent).
• Employment Assistance Services interventions were used almost equally by men and women (51.5 per cent men versus 48.5 per cent women) (Exhibit 3.4). They tended to be older than the overall population of participants.
• Rural Experience, although it is not an LMDA Provincial Benefit or Measure, is included here because it is partly funded under the Research and Innovations component of the LMDA. The program had the highest proportion of male participants (69.9 per cent male versus 30.1 per cent female) (Exhibit 3.4). This clientele was slightly older than the average age of other participants.

View EXHIBIT 3.6

3.5 Profile of EI Claimant, Reachback, and Comparison Groups

As mentioned in the program description, there are two client groups eligible for PBMs: individuals who are current EI claimants and reachback clients (those who have claimed regular EI within the past three years or claimed maternity/paternal benefits within the last five years). The administrative data file could not provide a reliable indicator of EI claimant/reachback status of participants and so an indicator was constructed based on participants' EI profile.

In Exhibit 3.7, we observe the extent to which the reachback share of participants varies across programs, based on the administrative data for the full population. This exhibit also presents the distribution of reachbacks according to the length of time between when the EI claim ended and the current intervention began. Note that excluded are those whose claimant status is unknown. The results in row one indicate, first, that reachbacks represent over one-half (56 per cent) of all participants. Second, the vast majority (87 per cent) of Entrepreneur participants is represented by reachbacks. Third, there is little variation in reachback share for the other interventions, with the share being the smallest for SLG participants (47 per cent). Fourth, the largest share of reachbacks participated in their intervention within three months of the end of their EI claim, with over one-half (56 per cent) reachbacks being in this group.

Exhibit 3.8 presents an overview of the sociodemographic characteristics of the claimant, reachback, and comparison groups based on survey data.¹⁰ Note that the comparison group is made up of only active EI claimants (who did not participate in the PBMs). The data suggest, first, that females represent a greater percentage of reachback participants than EI claimants. Women represent an even larger share of the comparison group (49 per cent). As for the age mix, the results indicate that, at 15 per cent, youth (under 25 years of age) represent a greater share of EI claimant participants than of reachbacks (10 per cent) and the comparison group (six per cent). The language profile of EI claimants, reachbacks and comparison group members is very similar, with over 70 per cent of all three groups being anglophone. The educational profile of the three groups is also similar, though EI claimants are somewhat more likely to have completed a college diploma than their counterparts in the reachback group, while members of the comparison group were somewhat more likely to have a university degree.

As for marital status, reachbacks were more likely than EI claimants to have once been married (13 versus eight per cent), while comparison group members were considerably more likely to be married or in a common-law relationship 67 versus 52 per cent). Finally, the representation of equity group members is similar among the three groups.

Exhibit 3.9 presents historical data on EI claimants' and reachback clients' income and earnings in the three years prior to their intervention and also their use of EI and social assistance during this time period. These figures are based on administrative data. It should be noted that EI claimant and reachback results are based on the full population of program participants to obtain an accurate picture of these individuals' earnings and income history. However, for the comparison group members, the results are based on just survey respondents because of difficulties in establishing the comparative reference period for the full population of non-participants. (Recall that individuals were selected into the comparison group on the basis of having an EI claim that coincided with participants in the PBMs). It should be further noted that, because of this, comparisons between participants and non-participants should be made in terms of patterns over time rather than absolute levels of earnings and benefits.

It should also be pointed out that in Exhibit 3.9 the EI figures for EI claimant and reachback participant groups were computed somewhat differently from the other figures presented in the table. The former were based on a 12-month period basis rather than a calendar year period on which the other figures were based (including the EI usage figures for the comparison group). This approach yielded a more accurate representation of EI use based on monthly usage, particularly for reachbacks. For example, on a calendar year basis, a reachback participant whose EI claim ended in March 1996 and whose LMDA intervention began in December 1997 would have been counted as having a claim one year before the intervention. This is despite the fact that the person would have been classified as a "21-month reachback", i.e., whose EI claim ended 21 months prior to the intervention. Since EI monthly EI data were available, we were able to compute EI usage rates and levels for 12-month periods and this person would have been more accurately counted as receiving EI benefits in the second "year" (12-month period) prior to the intervention but no EI benefits one "year" before. However, since, as noted above, it was difficult to establish (hypothetical) intervention dates for the full population of non-participants, we did not perform this calculation for the comparison group. And, because only annual and not monthly data were available for gross earnings and social assistance (SA), the same calculation could not be performed for the earnings and SA. Once again, as noted above, because of the different methodologies, comparisons between the groups should be made according to patterns not levels.

Exhibit 3.9 indicates, first, that, while in the third year prior to program entry EI claimants and reachbacks were in similar earnings brackets, the pattern in the intervening years indicates rising earnings among the former and declining earnings among the latter. For comparison group members, earnings too were declining over the pre-intervention period. As for employment insurance (EI) benefits, panel 2 of Exhibit 3.9 indicates that, as expected, EI claimants were less likely than reachbacks to have received EI benefits in the period leading up to the intervention and somewhat less likely than the comparison group in the 12 month period prior to an intervention. For comparison group members, EI incidence appears to have been rising over time at levels similar to participants. As for levels of EI benefits (panel 3), we observe more or less stationary annual levels for EI claimants and rising levels for reachbacks leading up to program entry. (Note that the means and medians are for the entire population of participants including those with zero benefits in the respective year.) For the comparison group, mean EI benefit levels declined in the last two years prior to (hypothetical) program entry, while the median rose in the last three years.

Finally, the social assistance (SA) incidence results (panel 4 of Exhibit 3.9) indicate, first, that SA incidence was similar in each of the years prior to program entry for EI claimants but rising for reachbacks. By the last year prior to program participation, SA incidence was more than twice as high for reachbacks as for EI claimants (19 versus eight per cent). For the comparison group, SA incidence was even lower (two per cent). As for SA levels (panel 5 of Exhibit 3.9), we observe that mean SA benefit levels were falling for EI claimant participants but rising for reachback participants. Mean SA levels were negligible for comparison group members and falling over time.

Footnotes

8	Source: 1998 Employment Insurance Monitoring and Assessment Report.
9	Note that the means shown are for those who had the respective type of income in the year, not for all participants.
10	Note that no reachback clients are included in the comparison group survey.