Yesterday, the initial Job Vacancy and Wages Survey (JVWS) release was announced by Statscan. The timing of its release, just a couple of weeks into a three-month long general election campaign, is questionable at best. That the new survey happened to find a previously undiscovered 180,000 job vacancies, 82 percent more than Statscan’s current Job Vacancy Survey (JVS), is even more so. There isn’t much to go on at this point; nevertheless it’s worth taking a closer look at what little there is so far.
First, the release is titled Job vacancies and payroll employment by province and territory, first quarter 2015. Yet the note to readers indicates that data collection only began in February 2015; the so-called ‘first quarter 2015’ JVWS release obviously doesn’t cover the first quarter. Nor for that matter does it cover the ‘W’ part, since the same notes indicate data collection for the wage component is scheduled to start in 2016. Clearly, this was a premature release. The question is why.
That there wasn’t a quarter’s worth of data to support the supposed Q1 2015 JVWS release would only be an issue if the release specified how its estimated 399.900 job vacancies were calculated: Is it some mean of monthly results, the last reported month’s result. or something else? Neither the release text, reader notes, table footnotes nor brief JVWS survey description – all the publicly available info so far – indicate which.
The closest point of comparison for the remarkable job vacancy figure is the number of vacancies reported in the February 2015 JVS, based on a 3-month moving average ending that month, which was only 220,000. As noted, that’s a 180,000 vacancy, 82 percent discrepancy.
The JVWS note to readers skimps over accounting for the discrepancy, writing it off as “differences in methodology” between the two surveys. Specifically, five differences are briefly touched on, listed below with explanatory notes / comments:
The JVWS quarterly sample includes 100,000 business locations, compared with a monthly sample of 15,000 from the survey component of the SEPH.
It’s unclear how the reader is supposed to interpret this. Does the much larger sample size provide a more accurate estimate, or does simply surveying more locations yield more jobs? In any case, the figures here also appear somewhat misrepresented. A note further down indicates that only 67,000 locations were sampled for the supposed Q1 2015 release – not 100,000 as indicated further up – because it only covered two months (February and March). The JVWS appears to have a rolling sample of 33,000 locations per month. Statscan should be careful with what it’s suggesting here, since it’s effectively implying that its payroll and other labour market surveys may be fundamentally flawed by virtue of their sampling size and scope.
The JVWS covers the entire agriculture, forestry, fishing and hunting sector, whereas the SEPH covers select subsectors of this industry group.
This is a notable difference, since the JVS (see brief survey description, as well as sample table) explicitly excludes the agriculture sector.That’s because the sector is largely comprised of casual, self-employed and foreign workers. See info on the Seasonal Agricultural Worker Program (SAWP), presented by Employment and Social Development Canada (ESDC) – coincidentally, the same agency funding this new JVWS survey that includes the agricultural sector. By law, SAWs are contracted employees covered by collective agreements and entitled to “the same wages and benefits as those provided to their Canadian and permanent resident employees working in the same occupation.” The latest estimates available from Citizenship and Immigration Canada (CIC) indicate there were 39,548 SAWs in Canada last year.
In the JVWS, the sampling unit is the location, while it is the establishment in the SEPH. For example, the JVWS surveys individual stores or restaurants, whereas the SEPH generally surveys the head offices of large retailers or restaurant chains.
Again, it’s unclear how the reader is supposed to interpret this. Do establishments routinely and remarkably underestimate workforce figures for their respective individual locations? If true, this would have significant economic consequences beyond simply skewing Statscan payroll survey results.
JVWS respondents tend to be responsible for human resources with a good understanding of both current and emerging job vacancies, while the SEPH respondents are often responsible for the company payroll.
This expands on the previous point, suggesting the JVWS also includes future job vacancies not otherwise accounted for in the JVS – like the CFIB job vacancy survey criterion, which “does not require that a specific position exists, only a need.” But that’s simply false. The JVWS questionnaire clearly indicates that only positions for which the employer was “doing active external recruitment” count as vacancies – the same criterion as the JVS.
JVWS respondents are asked to report jobs that are vacant on the first day of the month as well as those that will become vacant during the month. In contrast, the two questions on job vacancies in the SEPH refer to jobs that are vacant on the last day of the month.
This one’s just plain silly. Since there’s no info on which period the Q1 2015 JVWS vacancy figure covers, let’s assume it covered some period up to March 1st, 2015. The February 2015 JVS provides a three-month moving average ending on February 28, 2015. So somehow, between February 28th and March 1st, there was a near doubling of job vacancies available (220,000 vs 399,900)?
At this point there are way more questions than answers regarding the JVWS release.
Statscan has indicated that industry-level figures for the first JVWS release should be available by the end of August. The industry-level data and additional feedback provided by the agency will be posted accordingly.