Do-It-Yourself Recovery

Here is my take on today’s Labour Force Survey:

Self-Employment Surge

April’s apparent gain in employment was entirely due to increased self-employment. Specifically, total employment rose by 36,000 while self-employment rose by 37,000, meaning that 1,000 fewer Canadians were paid by employers last month.

One must ask whether more Canadians are becoming self-employed voluntarily or because they cannot find jobs paid by an employer. The fact that self-employment is surging amid a severe economic downturn suggests that workers are turning to this option by necessity rather than by choice.

Indeed, Canadian self-employment reached its highest level ever in April, both in both raw numbers and adjusted for seasonality. (Labour Force Survey figures on self-employment go back to January 1976.)

Youth Leave Labour Force

In April, the number of Canadians aged 15 to 24 in the labour force declined by 20,000. Unemployment among this age group also declined by 21,000, suggesting that young unemployed workers are abandoning the labour market because of the lack of job opportunities. The threshold for new workers to qualify for Employment Insurance (EI) benefits is especially high: 910 work hours.

Masked Unemployment

Because most unemployed workers do not receive EI benefits, many are pushed into marginal forms of self-employment or out of the labour force altogether. Due to more self-employment and youth leaving the labour force, Canada’s official unemployment rate did not increase between March and April. However, the accessibility and duration of EI benefits are explicitly based on the official unemployment rate.

Therefore, the unavailability of EI benefits threatens to become a self-perpetuating problem. The solution, of course, is for the Government of Canada to reform EI. Specifically, it should reduce the eligibility standard to 360 hours and extend the maximum duration of benefits to 50 weeks across the country.

Even though much unemployment was masked as increased self-employment and youth leaving the workforce, official unemployment still rose by 8,000 people in April. The total now stands at 1,464,600, its highest level since November 1996.

Of course, the increase in official unemployment was greater if youth are factored out. Among workers aged 25 and older, official unemployment rose by 29,000 in April.

UPDATE (May 9): My title is quoted in The Toronto Star’s headline and text. It also appears in the opening line of The Financial Post’s report.


  • Good analysis Erin.

    I would add the following: the LFS question on Self Employment is not black and white. That is, it is very easy for a respondent to say they are self-employed, and there is not much in the way of ensuring they are actually self-employed in a manner that one would qualify as a generally accepted job. It would be interesting to unpack the data a bit further on this group.

    Considering large in self- employed, there is without a doubt some kind of anomaly happening here in the measurement definition.

    I wonder what is happening with the R8.

    The US had another dismal jobs report, which throws another suspicious light on ours.


  • Yep this follows the time honoured pattern of discouragement (dropping out of labour force) scramble (finding odd jobs or in this case seasonal contracts) then despair. I think the trick here is to take the average LF participation rate over the say the last three years and the average level of self-employment over the same period and then use those averages to calculate the unemployment rate. Yet the fact does remain that workers declaring themselves as newly self-employed does point to at the very least a desperate optimism.

  • Over time employment and self-employment will both increase along with overall output, but there are shorter run compensating movements, in which declines in paid employment can lead to an increase in self-employment. Conversely, if paid jobs are rising in number and pay, one can see a drop in self-employment as people who were working for themselves find a good job offer.

    I have heard it said that the self-employed generally earn less and work longer hours than paid employees with the same human capital, a compensating differential I suppose for “being one’s own boss”.

    Looking ahead a couple of years, if self-employment keeps growing throughout the recession it may be a propitious time to consider the issue of unemployment insurance for these people.

  • If you look at the details, just about the only major number that is statistically significant from this month’s labour force survey numbers was the increase in self employment.

    According to the sampling errors, there’s a 90% chance that there was an increase of self-employment of between zero and about 70,000. For the overall employment numbers, there was about a 25% chance that there could have been some job loss or an increase of over 70,000.

    Most of the other monthly numbers for the April survey are considerably less statistically significant, with many changes quite a bit less than the monthly sampling error. The change needs to be twice the size of the sampling error to significant at the 95% level.

    I don’t usually pay much attention to Statscan’s monthly labour force numbers because I’ve found that they seem to bounce around a lot more than they used to. (Instead, I usually look at changes over a number of months or where they are confirmed by other sources.)

    Have others found that as well?

    I think they’ve made a number of changes to the methodology that has made the LF survey less reliable, perhaps even less than these reported sampling errors. For instance, households are kept in for 6 months, with a turnover of only 1/6 of the sample per month, and less in-depth interview in successive months. I haven’t been part of the survey, but people I know who were in it found these successive interviews to be very cursory and so presumably less reliable. I don’t know whether these aspects of the survey are reflected in the sampling errors or not.

    I took a look at the LFS methodology document
    ( ) and found that they reduced the sample size by 3% starting in 2004 and did this redesign in a partial (and so presumably less accurate) manner. They also moved to a telephone first contact in urban areas instead of initial in-person interviews. (And then there are the populations that are not included in the sample.)

    Perhaps this explains my experience as a relatively lay observer that the LF survey has become less accurate over the years.

    According to the guide, these changes were made because of budget constraints at that time. I don’t know to what extent they have or are planning to improve the labour force survey and fix some of these problems.

    Considering the effort that goes into the labour force survey and the attention that it gets, I think developing a more accurate survey should be a priority.

  • You mean it *isn’t* a good thing when workers take control of the means of production?

  • It’s a very good thing when tenured academics become politically active, if and only if they declare their partisan alignment in an open, honest and straightforward manner.

  • Ah. Another candidate for Minister Who Decides Who May Speak. A hotly-contested post, that.

  • Good news…increased innovation and entrepreneurship will lead to a stronger economy as we move out of this recession.

  • “You mean it *isn’t* a good thing when workers take control of the means of production?”

    Socialist theory considers small business people (aka the petty bourgeois) to be very much in the same boat as the working class and not in control of the means of production, despite the fact that most of them consider themselves to be members of the bourgeois.

  • A very hotly contested post, with candidates from the media punditocracy and elsewhere, is the position of “most convincing job of persuading the public that they are an independent, knowledgable commentator when they are in fact pushing a very particular party or interest group case”. That title is a bit of a mouthful, but the idea behind it is actually remarkably simple.

  • Bah-ha-hah!


    Unintentional irony is one of my guilty pleasures, I’m afraid.

    And I’m also sorry to say that I’ve seen any number of attempts to divert attention away from what I say to whether or not I am Pure. I’m unimpressed.

  • I would think that labour would be in favour of increased self-employment with the marginal contractors making less money than salaried staff. Haven’t any of you worked on a research contract for a labour union? Just joking.

    In the consulting sector, there is a long term pendulum which moves work back and forth between in-house and contracted-out staff, depending on the labour market. When the labour market is tight, billing rates are high and providers are hard to find, so there is a gradual shift towards hiring in-house staff who are cheaper. Then unemployment goes up, billing rates drop, and it becomes cheaper to hire people on contract. And it’s symmetrical on the worker’s side.

    It’s true that often self-employed people are sometimes in the same boat as workers. But they vote differently.

Leave a Reply

Your email address will not be published. Required fields are marked *