Preying on their populations’ lingering post-recession employment uncertainty and financial insecurity, governments of many Western industrialized economies have taken to framing just about every public policy in terms of jobs.
Last month, for the first time, Canada’s Department of Finance released a separate Jobs Report along with the federal budget. Problem: Canada’s job creation record, as indicated in Statistics Canada’s Job Vacancy Statistics (JVS), has been dismal. Solution: Ignore the JVS, use something else and hope no one notices. Unfortunately, as The Globe and Mail reported, that last part didn’t work out.
The Jobs Report released on budget day (Feb. 11) noted a job vacancy rate of 4.2 per cent, as of January. The Statscan JVS, released one week later, noted a 1.5-per-cent job vacancy rate for the three-month moving average ended last November. (The job vacancy rate represents vacancies as a percentage of total employment positions, filled and unfilled.) Expressed as jobs, Finance reported 662,000 vacancies for January; Statscan reported 220,000 vacancies for November.
That’s quite a spread. But on closer inspection, not only was Finance not citing a job vacancy survey, it wasn’t even measuring job vacancies.
Finance noted (in the information for a chart on page 30 of the Jobs Report) that several organizations publish estimates of job vacancies: Specifically, Statistics Canada, the Canadian Federation of Independent Business (CFIB) and Wanted Analytics. It indicated that it favoured the data from Wanted – a Quebec-based company that looks at online job postings – because it “provides highly accurate job vacancy rates as the job boards surveyed capture 85 per cent of all online job postings in Canada. Wanted Analytics data are available on a monthly basis for a four-year period, cover all sectors of the Canadian economy and are available at the occupational level.”
In response to queries, a Finance official also explained that Wanted has data going back to 2005, while Statscan’s JVS was only introduced in 2011 – thus lacking data surrounding the recession.
Let’s consider the differences among these three alternative sources.
The CFIB’s job vacancy survey uses a sample that only covers CFIB members. It also defines job vacancy quite loosely: “CFIB’s definition … does not require that a specific position exists, only a need.” Even with this biased sample and questionable definition, the latest CFIB Help Wanted report for the third quarter of 2013 could only muster a vacancy rate of 2.4 per cent.
While Statscan’s JVS could use some work, its survey sample does cover most Canadian employers and industrial sectors. The survey asks employers how many specific positions they had vacant at the end of each month.
Finance’s job vacancy rate did not come from Statscan, the government’s official statistical agency. Representatives from Statscan’s Labour Statistics Division and from Finance indicated that there was no consultation or collaboration between the two on the vacancy data in the Finance report.
How did Finance come up with its remarkable rate? As its report indicates, “The job vacancy rate is defined as the number of online job postings (based on Wanted Analytics data) divided by labour demand, that is, total job postings plus occupied positions (total employment, from Statistics Canada).”
Unlike the CFIB and Statscan, Finance didn’t count job vacancies as the number of positions vacant with employers. It counted the number of ads posted online, as provided by Wanted’s data.
Wanted Analytics, justifiably, distances itself from any research conducted and/or conclusions drawn by clients using its data. As a representative of Wanted Analytics said: “We simply provide insight into what the employers are placing online … Our customers infer what they would like.”
Broadly, there are two major issues with using job listings, especially online postings, as a proxy for job vacancies. One is duplicates – the same job posted on one or more employment websites. The other is phantom listings – vacancies that have either expired, or never existed (for example, when companies advertise in order to recruit a database of applicants for general future opportunities, but they have no specific vacancy).
Statscan used to have a survey based on job postings, the Help Wanted Index (HWI), but it was discontinued in 2003. The official reason: “Many users have expressed concern over its performance in recent years, especially in light of the growing use of the Internet by employers as a means of posting job openings.”
As to why Statscan didn’t simply integrate online job postings into its then-newspaper-based HWI, a Statscan analyst said: “It would have been very complex at the time to sort out which Internet job posters should be the focus, how to eliminate duplication within and across job posting sites, etc.”
Despite employing complex algorithms to identify and remove them, Wanted Analytics concedes there are numerous situations where duplicate and phantom posts can’t be automatically filtered out; duplicate ads often don’t identify the employer, and don’t use identical wording. (Wanted doesn’t conduct a manual verification.)
There is simply no plausible rationale for Finance substituting online job ads for available positions with employers in its so-called job vacancy rate. There is, however, one possible, less rational reason: Politics.
A 7-per-cent official unemployment rate, with a 4.2-per-cent job vacancy rate, speaks to a strong labour market suffering a skills shortage. On the other hand, a more realistic 14.2-per-cent labour underutilization rate and a 1.5-per-cent job vacancy rate speaks to a weaker labour market suffering a shortage of jobs.
Any sitting government would obviously prefer the former narrative, especially one heavily invested in “Jobs, Growth, Prosperity” sloganeering. Unfortunately, politics may be getting in the way of transparency and well-informed debate.
Sam Boshra is an independent Montreal-based economist, and editor for EconomicJustice.ca.