How to calculate un(der)employment

For my day job, I wrote a thing about underemployment in Canada. I thought that it might be useful to post my method here so that other interested parties could calculate it for themselves.

The headline unemployment rate counts all those who are unemployed, available to start work, and actively looking for a job. The internationally accepted measure of the unmet need for employment includes those who are unemployed and adds those who are partially employed but want more hours, those who aren’t able to start work right away, and those who are discouraged from actively looking, but would accept work.

Fortunately, it’s much easier to calculate this statistic than it is to explain it.


unemployment + 

All involuntary part time (CANSIM Table 282-0014) +

All “Not in the labour force but wanted work” (CANSIM Table 282-0219)


Divide by:

employed + unemployed +

All “Not in the labour force but wanted work” (CANSIM Table 282-0219)


This gets you pretty close.

There is one tiny component that isn’t directly available on CANSIM, and that is Long-Term Future Starts (LTFS).

You can estimate LTFS from R6, which is (unemployed + recall, replies + long-term future starts) / (employed +unemployed +recall, replies +  long-term future starts).

The number of recall and replies is available in CANSIM Table 282-0219. Or you can request that data from Statistics Canada directly.

LTFS are not counted as part of the labour force, so just like the marginally attached group, you need to add LTFS to both the numerator and the denominator.

R8, Statistics Canada’s most comprehensive underemployment rate only counts about half of the time-related underemployment, and an even smaller portion of the marginally attached group.

R8 = [(unemployed + discouraged searchers + waiting for recall, replies + long-term future starts + involuntary part-timers * (1 – average hours of involuntary part-timers at main job / average hours of full-time workers at main job))
/ (employed + unemployed + discouraged searchers + waiting for recall, replies + long-term future starts)]

(Definitions for all of Statistics Canada’s supplementary unemployment measures can be found in “Inside the Labour Market Downturn”, Perspectives on Labour and Income, 2011).

Statistics Canada already collects this information, and could very easily begin reporting on this measure in its monthly LFS release. Australia, for example, already releases underemployment reports on a quarterly basis, and will soon be releasing seasonally adjusted data monthly, broken down by age, sex, and region.

It’s time to update how we measure the unmet need for employment. Before you can fix a thing, you must first understand it.


  • R8 is something I do think should be reported more as it is what essentially is under utilized labour and provides an estimate of how many people could be used in society for producing goods and services on the one hand, and on the other it represents those workers, in most cases without adequate means of employment to sustain themselves- which when control over the means of production is within the hands of capital becomes quite problematic at the individual and the social level.

    However- I think a measure that is even more in demand and Statistics Canada does very little to monitor is employment quality. Without any notion of employment quality- we can look at any of the measure between R1 to R8 and still have very little idea on the quality of employment that is in place.

    It is indeed the new measure of the future of labour markets- especially when the thesis of capitals becomes low quality- barely enough jobs. Just as the BLS in the US is redefining how notions such as manufacturing jobs are measured in the US, progressives must also ramp up the pressure to change these official Statistics. Essentially many jobs that are service industry jobs in the US will now be included within the manufacturing statistics. Apparently an much easier way of on-shoring jobs than actually really on-shoring jobs.

    Currently working as a volunteer with the CCPA to produce a measure that gets us down the pathway to developing a measure of employment quality. I must say the process has been long and under funded- like many things I guess it will take the passion to put this measure through- cannot seem convince enough people that this effort is worthy. Pretty messed up and disorganized on the left is all I can say- hell in a hand basket- is one descriptive that comes to mind.

  • just frustrated today I think- too much crunching is going to my head

  • Paul, you’re right, this only measures labour underutilization – the unmet need for work. It doesn’t measure the unmet need for *good* jobs – secure and well-paid work. My hope is that talking about labour underutilization opens up more room to talk about job quality too!

  • Since ‘our’ government seems likely to lie or equivocate to us more or less forever and since R8 appears to be a number that can be calculated automatically is there no possibility that we could do so and place it somewhere that the general public is likely to see it?

    To Paul Tulloch I would say, “What left?”

  • Absolutely, I’ll be doing that here, and in the CLC responses to the Labour Force Survey monthly release.

Leave a Reply

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