CIBC on Employment Quality

Benjamin Tal of CIBC produces a quarterly Canadian Employment Quality Index. The releases from today (July 11) and February 11 provide amazingly different spins on amazingly similar figures. The basic facts are virtually unchanged:

- Most new employment has been self-employment as opposed to jobs paid by an employer.

– Most new employment has been full-time as opposed to part-time.

– Low-paying industries have grown faster than high-paying industries.

– In Alberta, job quality is slightly better now than in the mid-1990s.

– In Saskatchewan/Manitoba, job quality is about the same now as in the mid-1990s.

– Everywhere else, job quality is worse now than in the mid-1990s.

The February 11 release took the long view, noting that “Our index of employment quality is now at its lowest level since the early 1990s.” Although this conclusion remains accurate, today’s release is entitled “The Best of Both Worlds” and focuses almost exclusively on a slight, recent up-tick in the index.

While it makes sense to highlight recent developments, one wonders whether Tal’s completely new tone reflects a change in CIBC’s editorial policy. His February 11 release explicitly endorsed the view that lost manufacturing jobs are undermining overall employment quality. Today’s release is more consistent with the June 15 commentary by Jeff Rubin, CIBC’s Chief Economist, who argued that new jobs in construction and resource extraction have sustained overall employment quality despite the manufacturing crisis.

UPDATE (July 12): A handful of newspapers covered this employment-quality debate today.


  • I find it difficult to understand precisely how CIBC translate the first half of 2007 labour force data into the small uptick in their index.

    124,000 of the 197,000 new jobs created in the first half of the year were in the form of self-employment (which is seen by CIBC as an inferior form of employment compared to paid jobs.)

    And fully half of the new jobs – 101,000 out of the 197,000 – were in the two lowest paid parts of the private services sector – namely trade and accommodation and food services.

    Perhaps the other half were in particularly well-paid sectors?

  • Since writing the above I spoke to Bennie Tal. It turns out that in his index high paid jobs are weighted more heavily than low paid jobs – which is reasonable if the objective is to get at the implications of changes in the stock of jobs for labour income. He sees in the data some shift in the first half of 2007 to better paid jobs within various sectors which are not captured in net changes by broad sector eg a shift to better paid jobs defined by sub sector within trade and financial services. Bennie observes that the slight uptick in his index does not change the broad trend, as observed by Erin.

  • Most new full-time jobs in Canada high-paying and high-quality: CIBC report today.

    This is a headline coming out today from the CIBC job quality index which suggests that 400,000 new fulltime, high quality jobs were created in Canada last year, making up for the 130,000 jobs lost in the manufacturing sector. Suggesting that the manufacturing sector doesn’t matter and continuing on this bone headed notion that somehow Canada doesn’t need manufacturing.

    I am currently writiing a review of precarious work and the measurement of job quality within the Canadian employment. I am amazed at how these people at the CIBC can crank out such amazing artifacts from the LFS and SEPH surveys and mix them up with their own data sources and somehow come up with such bold conclusions. One would think they were accurate given every major media outlet picks up the story.

    I will say right off their methodology is highly questionable, given that I have worked with those files for over 12 years and I find it impossible to duplicate their numbers and come up with similar results. That is with using any notion of generally accepted statistical measures for validity, accuracy and bias. (actually they do not produce any form of statistical quality measures for this index, no standard errors or CV)

    There is a lot at stake here, as this report and the mileage it is getting within the media is making the claim that the job loss in manufacturing is not something to be overly worried about. This is quite a leap to take on such a flimsy, quasi-scientific, no error estimates approach to survey methodology as that employed by the CIBC. I mean they have what 2 people maybe working part-time on this vehicle. We have the whole labour statistics division at Statistics Canada employing over 150 or so highly trained specialists, who, if such a claim had any merit, or even close to being accurate would have been publishing something within this realm of notions.

    In fact about the only thing Statistics Canada had to say about fulltime job growth for 2007 was, if anything, it was more precarious than most years, as a majority of the full-time job growth was in self-employed or the public sector. Private sector non-self employed employment growth was anemic in 2007.

    We know that self employment, if you look at various reports, paint a picture of more precarious work than employed workers and many dimensions. I would say that public sector employment from a variety of indicators has definitely risen in 2007 and it is indeed of higher quality than the average. However, it is not a sustainable growth vehicle for job quality or more importantly the economy.

    This would seem to fly in the face of what came out of the CIBC report and the general conclusions made. Why they would make these broad based spurious conclusions is beyond me.

    Where is the future of our economic primary wealth generation coming from if it is not in the manufacturing sector. Please don’t say the knowledge economy, because it just is not registering within the survey vehicles to make these conclusions. Public sector knowledge economy maybe, but again we need something other than the public sector to power the economy.

    Eventually I will be putting forth a fairly substantial critique of the CIBC job quality index. But I will say this, in order to have a fairly robust job quality index one must have a survey that has an occupational dimension to it. Unlike the United States we do not have this. Not one of our national surveys are occupationally based. so I am unsure of how one can have a job quality index given this mission prerequisite. The heterogeneity of the variance in job quality based upon an industry estimate (as they must be using in the CIBC index) produces estimates and Coefficients of Variation that make these measures next to useless. Industry based estimates using NAICS are okay for some dimensions of employment, but job quality is not one of them. That is why the US in their labour cost Index uses a occupationally based survey. This is one of the main reasons why we do not have a labour cost index in Canada, no occupationally based survey vehicle. In my experience, and I have been in this field for 12 years, implementing an LCI without $15- 20 million in your pocket, you will never have a quality job quality measure that can stand up to any notion of producing statistically acceptable outputs.

    Paul T.

Leave a Reply

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