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

5.1 Ex Ante Model

The final question raised by the study was the following: to what extent were new entrants able to secure additional weeks to meet the higher entrance requirement? The question was addressed by conducting the following type of ex ante analysis: the 1995-Q4 data were analyzed to see if the distribution of the insured weeks of new entrants showed a “spike” at around 20 weeks (the old entrance requirement for new entrants). Such a spike would indicate that some individuals are able to extend their weeks of work in order to meet the entrance requirement. The statistical significance of a spike was tested using logit regression analysis.

Chart 5a shows the distribution of insured weeks during the qualifying period of new entrants (both beneficiaries and non beneficiaries). The above distribution was compared to the distribution of insured weeks of non new entrants.⁶

The data show that there was a “spike” at 20-to-22 insured weeks, and a “drop” at 17-to-19 insured weeks preceding the entrance requirement. A similar spike, but less pronounced, was also observed among non-new entrants.

Chart 5b shows that among individuals with insured weeks near the entrance requirement, there was a pronounced difference between new entrants and non-new entrants: new entrants were much more likely than non-new entrants to have 20 or just above 20 insured weeks, than just below 20 insured weeks.

The presence of the above pronounced “spike” and “drop” among new entrants provides initial evidence that some new entrants were able to secure additional insured weeks to meet the entrance requirement.

The above initial evidence was further tested using the following logit regression model among all individuals (new entrants and not new entrants) with 17 to 22 insured weeks:

where ABOVE is the probability that an individual had 20-to-22 insured weeks, rather than 17-to-19; and NEWENTR is a dummy that takes the value 1 if the individual was a new entrant and zero otherwise.

The logit regression results confirmed that new entrants with insured weeks near the entrance requirement were more likely than non new entrants to be above the entrance requirement (Table 5).

The regression results suggest that about one-fifth of new entrants were successful in increasing their insured weeks and moving from just below the entrance requirement to above the entrance requirement.

Behavioural effects, therefore, may lower the predicted above reduction in the average monthly number of regular beneficiaries from 45,600 down to 36,500. Correspondingly, behavioural effects may lower the expected annual savings in regular benefit payments in 1997 from $650,000 down to $520,000. This adjustment, however, may overstate the actual behavioural response if new entrants are less successful in adjusting their insured weeks, than as the case in the past at the lower entrance requirement.

5.2 Corroborating Ex Post Evidence

There is corroborating evidence of the above estimate for 1997. Estimates, based on the new Employment Insurance Coverage Survey (EICS), indicate that the average monthly number of new entrants, who were adversely affected by Bill C-12 in 1997, were approximately 35,000 — very close to the Canadian Out of Employment Panel (COEP) ex ante estimate of 36,500, after behavioural effects are factored in.⁷

More specifically, the results of the EICS analysis show that:

• The average monthly number of jobless (unemployed and out-of-the-labour force) individuals who had a paid job in the last 12 months, and worked at least 700 hours in 1997 was 1,107,000.
• Of these jobless individuals, 665,000 already received or will receive benefits, while 442,000 did not receive Employment Insurance (EI) and neither do they expect to receive benefits.
• Among the latter group of non-beneficiaries, 86,000 worked 700-910 hours, i.e. within the range of increase of the entrance requirement for new-entrants/re-entrants.
• Of them, 35,000 gave as a reason for not receiving benefits that they “did not have enough work.” This figure provides an estimate of the average monthly number of individuals that were affected adversely by the higher entrance requirement for new entrants/re-entrants.

Footnotes

6	Chart 5a excludes job terminations with 52 insured weeks to make it easier to compare new entrants to non-new entrants. The reason is that among non-new entrants, there is a big spike at 52 insured weeks (21 percent), while among new entrants the corresponding percentage is only 2 percent.
7	The EICS survey is a new HRDC/Statistics Canada quarterly survey that started in 1997. The survey data are collected through a follow up survey of a targeted group of respondents to the regular Labour Force Survey. A strength of the EICS is that its sample is representative of the labour force. A weakness of the EICS is that there is no link to HRDC administrative files.