I would like to thank you for providing me with the opportunity to present my views on the employment insurance program.
I plan to address two questions here. First, and very briefly, given access to benefits, what are the effects of various EI provisions on women in terms of the labour market behaviour of their outcomes? Secondly, should we be concerned about whether men and women have equitable access to EI benefits?
These questions should be examined within the context of the objectives of the EI program. EI's objectives do not appear to be clearly defined within the legislation. However, I would suggest there are three goals here. First, EI protects individuals from unexpected earnings losses associated with unemployment. Regular EI benefits provide this type of income support. Secondly, the programs help unemployed persons be productive participants in the labour force. To this end, for example, we have provisions that allow some individuals to receive support for training and education. Thirdly, EI promotes continuity in employment and attachment to the labour force. To this end, the provisions for maternity and parental leave, the allowances for seasonal employment and temporary layoff, may facilitate continued employment with the same employer.
Most often, economists are concerned with the extent to which policy will distort the decisions made by individuals. In the literature you can find several examples of how particular EI provisions change the labour market decisions and outcomes of individuals. For example, a study by David Green and Craig Riddell at the University of British Columbia demonstrated that when you extend the qualifying period for unemployment insurance, individuals will remain employed longer. David Gray's work at the University of Ottawa suggests that frequent EI users are sensitive to changes in benefit calculation formulas such that expansions of the program to cover non-standard employment arrangements may have unintended repercussions.
To my knowledge, we do not have a great amount of evidence on whether there exists a gender difference in the extent to which individuals will adjust their work arrangements to qualify for regular EI benefits. Obviously, we expect to see gender differences in the effects of EI maternity leave provisions. Michael Baker has already spoken to that extensively, where his research with Kevin Milligan has shown that the expansion of maternity and parental leave allowances in the late 1990s led to a substantial increase in the time mothers got to spend at home. More importantly, the provision of maternity leave benefits of any length will increase job continuity with the same employer.
Consider that any policy that increases women's labour force attachment and continuity with employers will have important long-run implications. In my recent research I have shown that the most recent groups of women entering retirement have greater access than previous groups to employer-provided pension plans and Canada Pension Plan benefits. This came with a greater attachment to the labour force in career employment, which in turn can be attributed in part to the legalization of the birth control pill. EI maternity and parental leave policy should lead to even greater access to public and private pensions among women.
Consider, then, the second question: Do men and women have equitable access to EI benefits? To examine access to regular EI benefits, I took a sample from the employment insurance coverage survey and measured the number of individuals who had received EI benefits since their last job as a portion of involuntarily unemployed individuals aged 25 to 44. Here in the sample, to note, I've excluded any mothers of infants and anyone who was self-employed.
A gender gap in the likelihood of receiving EI benefits when unemployed clearly exists. In my sample, 68% of unemployed women and 75% of unemployed men had received EI benefits since their last job.
I undertook a simple decomposition of that gender gap, and not surprisingly found that nearly 40% of the gap could be attributed to gender differences in work arrangements. Note that roughly 80% of employed women work full-time. Nearly 10% of unemployed women and only 2% of the unemployed men in my sample worked in permanent part-time jobs. Thirty-two percent of women work in non-permanent positions—this is not including the seasonal jobs—compared to 23% of men in such non-permanent positions.
Clearly, we expect that any individual in a full-time, permanent position will be more likely eligible for EI benefits than those in part-time, casual employment. This clearly fits EI objectives to provide income support for unexpected earnings losses associated with unexpected periods of unemployment. From the decomposition I also found that more that one-fifth of the gender gap in access can be attributed to gender differences in career paths, as characterized by the industry and occupation of their previous employment. While men tend to work in construction and manufacturing industries, particularly in positions of transport or equipment operations, women are more likely to work in retail, accommodation, and food services, in sales and services positions.
The resulting gap in EI access might be attributed to the industry differences in average hours within full-time or part-time classifications. I suspect there are also some industries and occupations where casual workers' hours are less likely insured.
Should we, then, modify EI to close the gender gap in EI access?
Overall, if you were to compare access to EI among comparable men and women—that is, men and women with the same work arrangements, same career paths, and so on—you would find that men and women have equal access to EI benefits. The differences in access are easily explained by the different labour market choices made by men and women.
Let us assume that on average women are rational, with a basic understanding of probability. Then we can say that when a woman chooses to take on part-time or casual employment instead of full-time permanent employment, she is aware of her lower likelihood to have access to many employment benefits. She is unlikely to have a pension plan or health benefits, she will not contribute as much to her Canada Pension Plan, and she will be less likely to qualify for EI benefits if she is out of work. These are the costs associated with her choice.
The benefits of her work arrangement will include more flexibility with her time, in many cases allowing her to take on home-production activities, such as child care and other family responsibilities. This choice is optimal in the sense that it is best for that individual.
Note that only a small portion of women working part-time jobs do so involuntarily. Within my sample of unemployed individuals in the Employment Insurance Coverage Survey, only 9% of those women are involuntarily part-time.
To suggest that we need to change the EI program so that the women working part-time or casually are more likely to qualify suggests one of two things to me. First, it might suggest that we do not believe women are able to act rationally or understand the implications of their choices. I do not think we have any reason to believe that and I would personally take offence at that suggestion. Second, it might suggest that Canadians want to encourage home-production activities by subsidizing the choice to take part-time and casual employment. This is certainly not a policy objective of the EI program, and achieving such objectives is best left to other policy levers.
A more appropriate policy action would be to ensure that women have equal access to the full-time permanent jobs in those careers likely to lead to EI benefit access. This has generally been the goal of affirmative action programs and even of pay equity legislation. Again, this does not require modifications to the EI program.
Over the years we have seen a greater tendency for women to take on full-time employment and have seen some reductions in occupational gender segregation. If we expect these trends to continue, we should expect the gender gap to narrow over time. As long as women are primarily responsible for home production, however, some gap will remain as women choose to take part-time employment. Further reductions of that gap could be achieved, for example, through the implementation of national child care programs.
It is important to remark on the fact that modifications to EI to accommodate non-standard work arrangements will have important impacts on the choices made by men and women. Any accommodation will likely create an incentive for men and women to take on less secure employment and develop a long-term dependency on the EI program. This will make the program much more expensive, will make its role as insurance questionable, and certainly does not meet the objectives of promoting labour force attachment or job continuity. Furthermore, as all modifications to the program must apply to men and women equally, the gender gap in EI may persist regardless.
Overall, I would strongly recommend against modifying the EI program to accommodate non-standard work arrangements, as it is not clear that such modifications will support the objectives of the program, and other policy options appear more desirable to support the interests of women.
As a final note, I would like to point out that there are several people, male and female, who will never qualify for EI benefits yet who have to make contributions. It seems unreasonable to require that people pay insurance premiums toward insurance they could never benefit from.
For example, to be eligible for regular EI benefits in a high-unemployment region, a typical person must have had 420 insurable hours in the previous 52 weeks. On average, this person must have worked more than eight hours per week at all jobs combined. Introducing a year’s basic exemption to EI premiums based on an individual’s annual earnings with all employers, not a single employer, might be a reasonable solution to consider here. The provision of benefits to these individuals would otherwise require benefit qualifications to depend on weeks worked rather than on hours, and it is not clear how the number of weeks should be defined.
That is all I have to say for now.
Thanks.