As the ranks of Canada’s unemployed swelled in the midst of the recession earlier this year, the adequacy of employment insurance (EI) was called into question from many quarters. Disagreement about just what is right or wrong with the program could well have precipitated a summer election had it not been for the agreement by the Conservatives and the Liberals to create a bipartisan working group to examine reform options.
But much of the debate thus far has focused on whether to reduce the number of hours required to qualify: the opposition parties are in favour of reducing the work requirement to 360 hours (the equivalent of about 11 35hour weeks) in all regions of Canada from the current scheme, under which minimum hours required to qualify range from 420 to 700, depending on regional economic conditions. The ruling Conservatives contend that the status quo is appropriate, preferring instead to temporarily extend benefits for five weeks.
The narrow focus on minimum work requirements and length of benefits is unfortunate, because any changes in this regard will have fairly small effects on unemployment trends and EI coverage. The more serious issues — including how to deal with the large number of long-term unemployed who are no longer eligible for EI and how to adjust EI funding to encourage firms to use layoffs as a last rather than first resort for dealing with economic recessions — have been largely absent in public debate. It would be unfortunate if the working group does not reflect on these challenges and offer some pragmatic solutions.
The basic concern voiced in a number of quarters is that EI has not been responsive enough to rising unemployment, meaning that many newly laid-off workers suddenly find themselves without a safety net. The reason seems obvious: prior to the recession, it was widely pointed out, only about 40 percent of unemployed Canadians received EI benefits, and this ratio increased only marginally as the recession took hold. Thus, it would seem to follow that EI is “not responsive” to changes in business conditions.
But at the national level, this assertion is plain wrong. A better way to judge the responsiveness of EI to deteriorating economic conditions is to look at changes in unemployment and EI beneficiaries during the recession, and this picture tells quite a different story. From the beginning of the recession last October through April — the worst of the economic crisis — the number of unemployed Canadians increased by 567,000, and 462,000 Canadians were added to the EI rolls over the same period. While not a strict apples-to-apples comparison (the data reflect net rather than gross labour force changes), this implies that slightly more than 80 percent of the net increase in unemployment was “covered” by a corresponding increase in EI beneficiaries. This improves to nearly 90 percent if we focus our attention on the 25-and-over demographic group, which comprises the vast bulk of families.
The fact that fewer than half of unemployed Canadians receive EI benefits is nothing new (see figure 1). The gap between unemployed individuals and EI recipients widened gradually between 1990 and 1997, largely because of tighter eligibility requirements embodied in reforms adopted in the 1990s. But before jumping to conclusions, it is important to recall what these changes were. One of the most important was to make workers who quit or were fired with cause completely ineligible for benefits (prior to 1993, they could qualify after a several-week waiting period). Going back to the fundamental role of EI as insurance against temporary unemployment due to factors outside the worker’s control, it is hard to justify paying benefits under these circumstances.
Other changes increased the amount of time required to work before qualifying for benefits, such that the minimum number of weeks increased from 10 to 14 (depending on the regional unemployment rate) to 12 to 20 weeks from 1990 to 1996. Starting in 1996, work requirements were calculated on the basis of hours worked rather than weeks worked, in order to eliminate the inequity between the eligibility of full-time and part-time workers. This had the effect of tightening requirements for part-time workers, with the explicit goal of encouraging greater attachment to the labour force. These reforms (combined with a strong economy) had dramatic effects on the average duration of unemployment spells, which dropped steadily from 25 weeks in the mid-1990s to less than 15 weeks in the 2000s.
It is also important to understand that only a small proportion of potential EI recipients lack sufficient hours to qualify. Table 1 shows the structure of Canada’s unemployed population with respect to EI eligibility and confirms that in 2008 about 40 percent of unemployed Canadians received EI benefits. Of the remaining 60 percent, well over two-thirds did not receive benefits because they either had not worked in the past 12 months or had separated from their job for an ineligible reason (generally because they left voluntarily or were fired due to misconduct). Less than 10 percent of unemployed workers were unable to qualify for EI because of insufficient work hours.
Thus, despite the acrimonious political debate, reducing the number of hours to qualify for EI would have a fairly small effect on access to benefits. Even if the number of qualifying hours were reduced to zero — which would be pure folly — EI coverage would increase by less than 10 percentage points nationally.
The 2008-09 recession has nonetheless underscored some regional inequities under current eligibility rules, which tie hours required to qualify to the regional unemployment rate. Under these so-called “variable entrance requirements,” or VER, the number of hours required to qualify and the benefit period are both a function of the average unemployment rate over the preceding three months. Required hours range from 420 in high-unemployment regions to 700 in low-unemployment regions. This is designed to equalize access to EI benefits based on a worker’s probability of unemployment, but has a secondary effect (of unknown magnitude) of discouraging migration from high-unemployment to low-unemployment regions. Because they had relatively low unemployment rates coming into the recession, the areas of the country hardest hit by the current recession — the Alberta oil patch and southern Ontario — had until recently the stiffest eligibility requirements. As of April 2009, increases in UI beneficiaries amounted to less than 70 percent of the increase in unemployment in Alberta and Ontario since the beginning of the recession, compared to 83 percent in Canada as a whole.
This inequity points to the fact that the backward-looking average unemployment rate does not capture current labour market conditions. The fact that it is based on a three-month average means looser requirements do not kick in immediately when the economy turns down, penalizing people laid off at the leading edge of a recession.
But a more fundamental problem is that the unemployment rate is a poor indicator of the probability of finding work, particularly in recessions. As noted in an April 2009 analysis by TD Economics, what matters is changes in the availability of jobs. It is straightforward to show that when firms are not hiring (as is typically the case during economic downturns), availability of jobs decreases, making it as hard to find work in a region with 5 percent unemployment as it is in one with 15 percent unemployment. However, information on job vacancies is not currently available on a national basis, which precludes effective calibration of qualifying hours to regional economic conditions.
The fact that the dynamics of job search change dramatically in downturns suggests that flattening the VER to, for instance, 420 hours in periods of recession would be elegant in its simplicity and would improve regional equity. It would increase aggregate EI coverage by only a few percentage points, but would have a discernible effect in regions with low (but rising) unemployment rates. However, an EI recession “trigger” should be formulaic rather than based on ad hoc political decisions and working groups. As an example, flattening of the VER could kick in after two straight quarters of net job losses and expire after two straight quarters of net growth. Without such a trigger, we could find ourselves in the situation we are in now — the recession may well have ended by the time there is any agreement on how to change EI to address it. (The unemployment rate may increase after the “official” end of a recession, but this is typically due to more people entering the labour force rather than continuing net job losses.)
It is too late for EI reform to help combat the 2008-09 recession, which may already be in the past. But politicians and policy-makers should not allow this window of opportunity for more fundamental change to close. EI reform has always been a political hot potato, but the stars may be as well aligned as they can be.
In practice, a better EI program would involve reinstating and reinforcing many of the provisions enacted in the Employment Insurance Act of 1996, which sought to discourage use of the program as a regular income supplement as opposed to insurance against unpredictable job loss. The Act brought modest experience rating to benefit levels, reducing the replacement rate by 1 percentage point for each group of 20-plus weeks of benefits over the prior five years — though this measure was rescinded in 2000. Most changes to the EI regime since 2000 have involved increasing the generosity of EI benefits paid to seasonal workers, which has simply encouraged more seasonal employment and long-term dependence on EI.
Because EI has evolved into a multi-purpose social program (encompassing parental leave, compassionate care benefits and training programs, among others) that is trying to be all things to all constituencies, it cannot in its current form effectively accomplish its original mission: provide income support to individuals who become unemployed for relatively short periods through no fault of their own. In order for social insurance schemes to function effectively and provide the right incentives, the premiums paid must to some extent reflect the probability of needing benefits. A perfect example is auto insurance, where the premium paid is a function of factors related to the probability of making a claim (such as age group, number of kilometres driven, driving record, previous claims, etc.). Such experience rating, used in state unemployment insurance throughout the United States, is completely absent from Canada’s EI system: firms pay the same premiums irrespective of their history of layoffs, and workers pay the same premiums (and receive the same benefit levels) irrespective of the frequency with which they make claims.
Thus, firms have no incentive to minimize layoffs in economic hard times, because they pay no penalty by not doing so. Similarly, repeat users of EI have little incentive to increase their labour force attachment above the minimum qualification, because working more hours would have no effect on the benefits they can potentially receive. The net result is that unemployment spells are more numerous and longer than they would be otherwise, and this is one of the reasons (though certainly not the only one) why average unemployment rates and durations have tended to be lower in the United States than in Canada over the past several decades.
A perennial complaint about experience rating is that it unfairly penalizes seasonal workers and employers, because the nature of their industry implies regular and more frequent layoffs. But this complaint hides a more fundamental inequity, which is in fact the contrary: under the status quo funding arrangement, seasonal workers receive by far more EI benefits than workers in other industries, relative to the premiums they pay. This is not true insurance; rather, it is the subsidization of seasonal industries by other workers. Such support may be warranted and desirable in Canada, but it should take the form of an explicit occupation-specific wage subsidy (funded by all taxpayers rather than just EI contributors) so that its costs are transparent rather than buried implicitly within EI.
The other programs under the EI umbrella (which now amount to almost half of total EI expenditures) are not strictly social insurance programs, and thus should be divorced from EI. The largest of them is parental benefits, which costs about $3 billion per year. But it cannot reasonably be argued that the birth of a child is unforeseen (nor are birth rates particularly sensitive to the business cycle), so the notion of experience rating as a basis for these funding formulas falls apart. These programs are in fact social programs rather than insurance schemes, and should be funded from general revenues (to which upper-income families contribute proportionately more) rather than from a regressive payroll tax (which is capped, effectively exempting much of the income of wealthier Canadians).
Another large piece of EI that should be available to (and funded by) all Canadians, not just those who have previously held an EI-insured job, is training and other programs that go beyond short-term income insurance. The 2009 federal budget provided $500 million in additional training funds to all unemployed Canadians, not just EI contributors. As a matter of equity, all such EI initiatives should be divorced from EI and funded via general revenues, hammering home the principle that access to employment training, like access to health care, should be a right of citizenship. This would go a long way to improving the employment prospects of the 25 percent of the unemployed who do not qualify for EI because they have not worked in the prior 12 months.
A final issue relates to the structural changes that the 2008-09 recession is likely to bring to the economy. The reduction of once mighty industrial giants General Motors and Chrysler to shadows of their former selves is illustrative: it has dramatically increased the ranks of laid-off workers whose jobs are unlikely to come back in the recovery. Many of these workers have worked for decades in a specific company with minimal recourse to EI, yet they are most likely to suffer the most from the effects of the recession well after it is over. For them, the largest cost is not the loss of income during a spell of unemployment, but the wage loss when they become re-employed. There are many promising ideas from the Expert Panel on Older Displaced Workers (which went largely unnoticed when it reported last summer), including wage insurance and targeted EI benefit extensions, that suddenly seem prescient.
Employment insurance has been unfairly maligned as ineffective in responding to the 2008-09 recession, and the debate about hours required to qualify is a tempest in a teapot. But the heightened attention to the program provides a golden opportunity for policy-makers in general and the bipartisan working group in particular to make broader improvements to the safety net, including bringing EI back to its core mission of income support during unanticipated job loss, as well as beefing up programs to improve the employment prospects of the many Canadians who do not qualify.