More on the “Do It Yourself” Recovery

Erin has already posted a comment on last month’s surge in self employment .

My colleague Sylvain Schetagne extracted some data from the Labour Force micro data file to get a better handle on the change last month – note these are NOT seasonally adjusted data so they should be interpreted with a bit of caution.

The not seasonally adjusted jump in self employment from March to April (not counting unpaid family workers) was 48,358 – of which 38,844 (80%) came in the form of self-employment in non incorporated enterprises with no employees. This category made up half (51%) of self-employment in March, so the increase from March to April was disproportionate.

Some self employed positions – notably those who are incorporated and have employees – offer decent incomes – but research by Leah Vosko and others have shown that those who are employed in unincorporated businesses with no employees generally have extremely low incomes and are extremely insecure. (See Leah Vosko (Ed) Precarious Work. McGill Queen’s University Press, especially Chapter 3.)

In short, not only was there a big jump in self employment, there seems to have been an even bigger jump in the worst form of self employment.


  • If only we could get Statscan to apply the fourfold test from Montreal Locomotive when asking self-employed workers about their business relationship with their biggest client. Then we would know for sure.

  • Great work Sylvain.

    the interpretations of self employment measured by the LFS, as you make note of Andrew, comes in many folded up dimensions. Given the rotational basis of the survey, and the extra focus on the longitudinal aspect versus the cross sectional, I find this jump in the more precarious type of self- employment extremely suspicious or basically down right odd.

    The different types of self employed that you are going into this measure are basically, in archetype terms: owner-managers versus temp workers.

    So your numbers definitely lends support to the view that was suggested in the field, in which some respondents recently out of work would rather respond to the survey that they are self-employed rather than unemployed. That is my hunch on this one.

    Again this goes back to the very real gap in the LFS. We have no measure of the quality of employment. If we had such a measure I am sure it would have picked up on this precariousness within the self employment, as that is one of the component that would have been feeding into such an index.

    I find it a bit suspicious seeing a jump in more casual work within the ranks of the self employed say in the form of casual temp workers. This due to a reference to a US survey, I foget which one, but I read it the other day,m which stated that Temp agency hirings were quite low. Typically the temp firms are busier in growth periods as firms tend to ramp up rather cautiously making use of temp agency workers.(the whole core periphery argument)

    I am very surprised Statcan did not release any kind of follow up into this question of self emplyment.

    Bottom line, the LFS numbers are basically funky again during a time of great need, because of bad design, or potentially design and dissemination. There should have been a follow up release done by the LFS and maybe there will be. Any responsible statistical agency, given these times would have provided a lot more investigation into such blips and bumps. It really must be something that is looked at for future work.

Leave a Reply

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