Didn’t anticipate writing anything further on the topic, as it’s been covered numerous times on these pages already,..
Readers may recall a post last summer comparing the Canadian and U.S. unemployment and labour underutilisation measures. It was shared with the Canadian Labour Congress, and followed up with an Economy Lab post co-authored with one of its economists.
Which made reading this on the Progressive Economics Forum (and a related CLC ‘report‘) all the more disappointing. Apparently someone forgot what they seemingly co-wrote, about why the US measure was more widely reported and discussed than its Canadian counterpart:
The BLS (US Bureau of Labour Statistics) U-6 is easy to calculate and understand; Statscan’s R8 is more obscure and harder to replicate.
The KISS Rule
The elegance in the BLS U-6 labour underutilization rate lies not only in its potential as a more realistic representation of household labour market attachment, but in the ease with which it could be explained: It would represent all those who wanted work (involuntary part-timers + job-seekers + those who gave up searching) as a share of the total labour force (all workers + all those who wanted work).
(‘Potential’ because, unfortunately, the BLS does impose some arbitrary restrictions of its own: a 12-month window for the marginally attached, and ‘economic reasons’ for involuntary part-time workers. See here for more details.)
The proposed CLC measure (which we neither agree with nor were made aware of until recently) on the other hand is a hodge-podge of both the U-6 and R8, making it more abstract and difficult to explain than either measure alone. The key is the CLC’s decision to wedge in a rather peculiar StatsCan classification that neither falls under employment, unemployment nor marginal attachment.
Long-term future starts (LTFS)
As explained here, future starts are
the unemployed… who haven’t looked for work in the last four weeks but will start a new job in the upcoming four weeks.
Whereas long-term future starts (LTFS) are
people who did not look for work but who had a job to start in five weeks or more. These individuals have not looked for work and are considered not available to be part of the current supply of labour because they do not express a near-term desire to work.
Counting future starts as unemployed is a good idea, and not just for the reason stated. StatsCan also asks the unemployed what they were doing the month prior to becoming so; apparently many of those anticipated ‘future starts’ never start.
Not to overdo it with the definitions, but the majority of LTFS are workers with some level of reassurance they’ve got a job to return to in an upcoming season. Hence their reluctance to either look for or want work (makes sense). Not a good fit for a measure that’s supposed to represent the share of the population who wanted work.
Lastly, the number of LTFS is quite small relative to the unemployed, marginally attached and involuntary part-timers. Adding the peculiar category to the numerator and denominator of the Canadian version of U-6 barely registers as a rounding error (increases the 2012 rate of 14.2% by ~0.2%).
Adding LTFS not only makes the labour underutilisation measure more difficult to explain, but it neither fits well with nor has a significant impact on the measure. Perhaps CLC wanted to distinguish its measure from the one proposed here – which itself is not original (then again, originality wasn’t the objective).
Curiously missing demographic
The linked ‘Progressive’ Economics Forum post was categorised under the following keywords: labour market, unemployment, women, young workers. StatsCan’s Labour Force Survey only queries three broad demographic traits: age, gender and immigration status (interestingly, not race).
Neither the referenced CLC ‘report’ nor related PEF post address immigrant labour underutilisation. Disappointing, all things considered. Perhaps not unexpected from Labour. Certainly cynical.