On the sheer pointlessness of reducing hours for EI eligibility as a counter-cyclical policy

One of the points I've been making about the psychodrama over reducing the EI eligibility requirements (eg: here and here) takes the form of a question: "Just what problem is this supposed to solve? How many people are there who were hired three months ago - in the teeth of the worst labour market in a generation - and who are now unable to claim EI?"

I've not seen a convincing answer, so I decided to dig up the Labour Force Survey numbers for employment by job tenure (CANSIM Table 282-0037). I also decided to look only at the numbers for those who are 25 years old and older; the appropriate policy for people younger than that is to provide training opportunities, not passive income support. The data are not deseasonalised.

The proposal of going to 360 hours/week will matter most to someone who has accumulated 10 weeks of full-time work, or 24 weeks (at 15 hours/wk) of part-time work. The LFS has data for the number of workers with job tenures of 1-3 months and 4-6 months, so I'm going to assume that these are the relevant groups.

In October 2008, there were some 815,000 full-time workers (4.7% of all full-time workers) who had between 1 and 3 months experience. In April, that number had fallen to 501,000.  That doesn't mean that 314,000 people lost their jobs. Someone who was hired in September would have had 1 month's experience in October and 7 months' experience in April: they would be classified differently. The drop is due to the fact that hiring dropped off as the recession hit; there simply were fewer people hired between January and March.

There were another 107,000 part-time workers in October 2009 with 4-6 months' tenure, and this number had actually increased to 116,000 by April.

So that means that as of the April LFS survey, there were some 608,000 (plus or minus a certain margin of error; these are very much back-of the envelope calculations) more workers who would now be eligible for EI under the proposed rules: an increase of about 4.2% of those employed.

But how many of them would actually lose their jobs and receive benefits? So far, we've lost 2% of employment, and if the recession holds to form, we might lose another 2%. If we assume that the new rules came in today and that 2% of those newly-eligible workers lose their jobs tomorrow (i.e., before they would have become eligible under the existing rules), then that works out to about 12,000 workers - something less than 0.1% of April 2009 employment.

The EI package is being sold as a way of accelerating the fiscal stimulus. But I don't see how extending benefits to an extra 12,000 people will make much of a difference to aggregate demand.

These are crude, back-of-the-envelope numbers to be sure. But I can't see how anyone can come up with a variation on them that affects their order of magnitude.

Comments

[Leigh Caldwell]

It's likely that recently-hired employees lose their jobs much more quickly than those with longer tenure (partly because they have lower levels of firm-specific human capital, partly because of loyalty and a sense of 'natural justice' which I think prevails in the labour market, partly because new people are often on effective probation periods for several months, and partly because new posts are created with a level of uncertainty about whether there is demand for their output).

I certainly have no data to indicate the extent to which this applies, but I would honestly not be surprised to see newer hires lose their jobs at a rate three to five times higher than longer-established employees. Of course the effect is probably even greater for sub-3-month employees who would not be eligible even under the new policy, but the 4-6-month cohort will still bear a substantial impact. I don't think it would be a whole order of magnitude, but I do think it's wrong to simply apply the 2% proportion.

The above is based on my experience of the UK labour market so Canada might be different. And of course your argument may still hold even if 60,000 people are helped instead of 12,000.

[Stephen Gordon]

I had thought of that. It might also be the case that those new hires are more secure, because in the middle of a recession, employers won't hire unless the post absolutely had to be filled. I don't know which effect would be more important, I decided that it was a wash.

[Traciatim]

"The drop is due to the fact that hiring dropped off as the recession hit; there simply were fewer people hired between January and March."

Do you have any proof of this, or is it simply an assumption? Wouldn't the new hires be the first cut in the event of company troubles resulting in the need for headcount cuts? Also, many people are hired in a probationary period, if the company has it's focus changed to reduce costs by reducing the workforce wouldn't it be best not to hire these new staff full time and just let them go as their contracts expire?

I don't agree that there should be a flat national hourly requirement. I like the system that applies the requirement to the unemployment in the area of the application. Why should a place that's falling apart have the same requirements as a place that has the lowest unemployment in the country?

[Eric Eden]

To me the debate is completely moot. Who cares really about a few thousand dollars - that you can count on one hand - when unemployment is likely to last multiple months in economic environment such as today.

Instead - when unemployment is low, pay very little. When unemployment raises, pay more. Maintain the same qualification levels or adjust them it doesn't matter to me, as the point is to add stability in the job market. When job prospects are reasonable - typical - EI only needs to bridge and personal savings buffer can be used. When job prospects have decayed and the general economy is suffering, and will further suffer extended job loss, EI becomes stimulus, as well as being needed funding for people's lives.

The amount of EI that is collectable is below self-insurable levels for the same premiums. 10 years paying something like $700, gave me 5200$ in EI benefits.

Therefore the system provides no pooling of risk for the consistent worker. What it seems to do is fund short-term workers where eligibility becomes a critical characteristic. To me this means that it doesn't satisfy its prime goal, and so there is no reason to debate a detail of it.

[westslope]

The Liberal campaign to open the EI flood gates gets particularly embarassing when the Liberal Finance Critic John McCallum, former professor of economics at McGill University, argues in favour of the program (with a straight face). I paste from McCallum's web-site:

Before entering politics, Mr. McCallum was senior vice-president and chief economist of the Royal Bank of Canada between 1994 and 2000. Prior to that he worked as a professor of economics at McGill University (1987-94), Université du Québec à Montréal (1982-87), Simon Fraser University (1978-82), and the University of Manitoba (1976-78). Prior to joining the Royal Bank, he was dean of the faculty of arts at McGill University .

A native of Montreal, Mr. McCallum obtained a bachelor of arts from Cambridge University, a diplôme d'études supérieures from Université de Paris and a doctorate in economics from McGill University . Mr. McCallum and his wife, Nancy Lim, have three sons.

[ramster]

But if the effect is so marginal, then what's the big deal? I agree that it's all political theater but from an economists POV, isn't a bad (but negligibly impactful) policy less of an issue than a bad (but significantly impactful) policy . If the number is 12000 people, assuming they get about $10K a year, that's $120 million per year. The GST cuts added $12 billion a year to the deficit.

[Stephen Gordon]

The concern is that it screws up the incentives over the longer term; see this earlier post

[Frances Woolley]

One real problem that moving to a shorter number of hours will solve is the ineligibility of part-time workers, especially long-term part-time workers for benefits.

The switch to an hours based model was actually supposed to help workers with under 15 hours a week who hadn't been covered under the old system, but studies done around the time of the switch towards hours based entitlement (there's a paper by Phipps MacDonald and MacPhail in CPP) found that a disproportionate number of part-time workers lost eligibility.

Would it be totally impractical to refund the EI premiums paid by/on behalf of workers who find themselves ineligible to benefit due to lack of hours?

Frances

[Nick Rowe]

"Would it be totally impractical to refund the EI premiums paid by/on behalf of workers who find themselves ineligible to benefit due to lack of hours?"

Practicalities aside, there does seem to be a certain logic, as well as fairness, to that proposal. I can't see any moral hazard problem. Iw might increase efficiency?

Thinking of Frances' point, together with the point made in a comment on another post (was it by Andrew?) about tapering off the benefits, there does seem to be something weird about these sudden sharp cutoffs. It's a bit like having marginal tax rates suddenly changing or going over 100%. Work 9.9 weeks get nothing; work 10.1 weeks and you are suddenly eligible for a big insurance payout.

The marginal tax rate on hours of work from EI is high up until 9.9 weeks, then suddenly goes strongly negative, then rises again.

Similarly, when you are about to run out of benefits, your marginal tax rate for getting job is high, then suddenly drops on the week you run out.

Shouldn't marginal tax rates always be smooth in decision variables?

[truemuse]

It is when you have been absent from the workforce for a period of years, that the 720 hours rule comes in. This is the barrier to watch. Anyone returning from child-rearing, mental illness, you-name-it, is held to a higher standard in order to claim. These are the same people who will likely need a few 'start-up' jobs before they find their feet again in the workplace. These are the same people most likely to fail in a new job. Coverage for someone who is fired is as important as lowering the threshold of hours. I don't know anything about marginal taxrates, but I would like to see people qualify for coverage of one month after one month of work, two months for two, three months for three. EI easily becomes a crutch if the payouts are too high over too long. So give max payouts (70% of income) for any work longer than a month. If you have 10 months of work, you can choose max payments for three months or reduced payments for 35 weeks (putting more control in the workers hands to decide how to use the support). The main idea of EI (forget about recession demands) is to keep people off of welfare and attached to the workplace. It makes sense, then to offer a choice about what type of benefit to accept. The gov't can win, depending on the choices people make. The system could evolve into something much better with those changes.
ty,
karen