An Employment Insurance system for the 21st century: Lesson 1, Big shocks matter

By Miles Corak

This is the first commentary in a three-part series examining ideas for reforming Canada’s Employment Insurance (EI) program. This commentary argues that the program as presently constituted is not well-designed to provide adequate support for households that suffer large and enduring negative income shocks.


Canadian workers and their families have been rocked by three major shocks in just barely more than a decade, and all three times the Employment Insurance program has been found wanting. 

The Trudeau government seems to recognize the need to modernize the EI program. Specifically, the 2020 Speech from the Throne boldly claims that “This pandemic has shown that Canada needs an [Employment Insurance] system for the 21st century, including for the self-employed and those in the gig economy.”

That is a tall order. It would require a major overhaul of a complicated program in the span of the next couple of months, with little or virtually no consultation of stakeholders or engagement of experts outside of the government.

Will Minister Carla Qualtrough, her cabinet colleagues, and of course the Prime Minister, get it right? After all the need for EI reform has long been recognized, with lessons learned well before the onset of COVID19. It has, however, always been politically convenient to put off the actual implementation of reforms.

What would an EI program “for the 21st century” look like? Developments over the first twenty years of the century have taught us some big lessons about the program’s shortcomings. These lessons should be used to judge any EI reforms. This post discusses the first of three lessons and the reforms they call for: Lesson 1, Big shocks matter and need a response in real time

Lesson 1: Big shocks matter and need a response in real time

Canadian workers and their families have been rocked by three major shocks in just barely more than a decade, and all three times the Employment Insurance program has been found wanting.

During the autumn of 2008, and certainly by November of that year when the federal government issued its Economic and Fiscal Update, everyone knew, every policy maker that should have known, knew that the economies of the rich world were about to spin downward in a way not seen since the Great Depression of the 1920s and 1930s, and that this posed a major threat to Canadians. Everyone, but our Employment Insurance program.

In October of that year the number of employed Canadians began to avalanche. Employment fell by close to 300,000 during the next three months well on its way to an eventual trough 400,000 below the previous high. Over this same period the number of new Employment Insurance claims gently rose by about 35,000 during the three months between October and January only jumping to attention in the next month, rising by 45,000 in February. All of this was too little, too late.

Clearly, the EI program had not worked adequately during the 2008/09 recession. And yet six years later when Albertans saw the bottom fall out of oil prices during the second half of 2014, everyone knew, every policy maker that should have known, knew that trouble was around the corner. Everyone, but our Employment Insurance program.

The avalanche in employment losses began in September 2015 and bottomed out in June of the next year with over 66,000 fewer Albertans employed. New Employment Insurance claims? They gently rose by 5,000 between September and April, only jumping to attention in the next month, rising by 17,000 in May. Again, too little, too late.

Another six years later, in the spring of 2020, Canadian workers and their families experienced a job shock like no other in their lifetimes, or for that matter in the lifetimes of their parents and grandparents. In February and certainly by March, everyone knew, every policy maker that should have known, knew the COVID pandemic was going to have serious consequences for the economy, for jobs, and for the capacity of Canadian workers to care for their families. Everyone, but our Employment Insurance program.

The number of employed Canadians fell dramatically between February and March, by more than one million, and then further again to bottom out in April with fully 3 million fewer people employed than just two months earlier. The Employment Insurance program choked, stalled in mid gear, deer-in-the-headlight panic. Too little, too late, requiring the government to institute emergency benefits with broad eligibility and immediate delivery.

These three painful experiences demonstrate that EI was not well designed to respond to big shocks. What reforms can improve the program’s capacity to help Canadians when big shocks happen in the future?

A big shock is a big change, and so the eligibility for and generosity of Employment Insurance benefits should in some part be determined by changes in employment in a given region, not just the level. And most important the change has to be measured in as close to real time as possible.

This argues for a new benefit phase in the Employment Insurance program, one that kicks in when employment in any region suddenly falls, and that enhances the generosity of the program by allowing more weeks of eligibility and also by replacing a larger fraction of lost income.

An Employment Insurance program for the 21st century would respond to big shocks in real time. The current EI program is hard-wired to respond incompletely and with delay in part because the rule governing eligibility is tied to past unemployment rates, not current and upcoming prospects for jobs.

There are sixty-two EI regions with the number of hours of work required to qualify for benefits determined by the prevailing unemployment rate. For instance, in a region with an unemployment rate of over 13 per cent an individual must have completed 420 hours of insured work, amounting to 10 and a half weeks of full time employment. In a region with an unemployment rate that is 6 per cent or lower, 700 hours are required for to become EI eligible.

Source: Employment and Social Development Canada,

So theoretically when the unemployment rate in your region jumps you need fewer hours to qualify. But it is not as simple as that. More than sixty regions tile the entire country, each requiring a monthly unemployment rate to be calculated. This stretches the limits of Statistic Canada’s workhorse survey, the Labour Force Survey, beyond statistical credibility.

The 50,000-odd households captured in this survey is a big enough sample to offer a quick and reasonably clear picture of employment and unemployment for the country and the provinces, but not for smaller areas. Statistical noise doesn’t cloud the signal of reality when you are interested in Quebec or in Alberta in their entirety, but it is a dense fog when you want information for Gaspésie-Îles-de-la-Madeleine or Northern Alberta outside of Edmonton.

Statistics Canada resorts to a three month rolling average to try to boost the signal and minimize the noise when calculating unemployment rates for each of the sixty-two EI regions. When the Agency tells us that the unemployment rate in the EI region of Toronto in February 2020 was 5.5%, what it is saying is that the average of the unemployment rates calculated from the Labour Force Survey in December, January, and February is 5.5%.

So it is no surprise that this rate, and the number of hours required to qualify for Employment Insurance benefits, did not change in March nor in April of 2020 even as the country entered into a lockdown, and job prospects evaporated.

Only in May, as the three month rolling average began to take into account the COVID crisis, did the unemployment rate for Toronto jump to 8%; and finally by June all the numbers feeding into the average no longer reflected pre-COVID and the regional unemployment rate for the city jumped further to 11 per cent.

This means that EI eligibility rules respond with a significant lag when things change suddenly and sharply, when hardship is obvious to all except the program meant to address it. This needs to be reformed if the program is to meet the challenges of big shocks, the very kind of challenges that lie at the heart of its mandate.

The University of Ottawa economist David Gray and his co-author Colin Busby recognized this years ago. Writing in 2011 they suggested:

“New benefit criteria should be linked to the national unemployment rate, or better yet, the rate of growth or decline in national employment.”

My view is a nuanced version of this. I propose putting the focus on changes in provincial employment rates.

The provincial data is as reliable and rapid as the national data, being released every month by Statistics Canada. And provincial statistics might be preferred to national numbers since big shocks may be province-specific or impact provinces differently. Finally, employment rates might be better suited than unemployment rates  because they are more closely tied to job prospects and the demand-side of the labour market.

Either way, these reflections ask that reforms toward a 21st century Employment Insurance should enable the program to do, in a rules-based way, what COVID era policy-makers were forced to do in a discretionary way. In short, EI needs to be redesigned to provide more generous benefits that are easier to access when job prospects suddenly disappear.

The direction for reform should be to add an extra benefit phase to the program, a Canada Recovery Benefit type of benefit phase that everyone would have access to, when the provincial employment rate falls by a given amount. For instance, the threshold could be a decline of percentage point or more below a province’s previous 12 month high. This recovery benefit would then fade away when employment rates returned to normal.

This program phase should also be more generous. During big shocks jobs disappear, and this means joblessness is involuntary and not influenced by work incentives. When jobs suddenly disappear in large numbers, the trade-off between program generosity and positive employment incentives that are embedded in EI become less important, and the program can be more generous without meaningful negative consequences related to work incentives.


In just over a decade, Canadian households have been hit by three major economic shocks. All three times, the performance of the EI program has been found wanting. Put simply, the program was not and still is not adequately designed to react quickly or adequately when big shocks occur.

A 21st century EI program that is equipped to respond adequately to big shocks would offer a higher replacement of lost income and longer duration of benefits when such shocks occur.

This is a slightly revised version of a commentary that was originally published on Miles Corak’s personal website.