Employment Financial security Governance Poverty Transparency

June 2013 LFS: Comparing the Canadian and U.S. unemployment and labour underutilisation metrics

As repeatedly mentioned on these pages, analysing the monthly movements of the Labour Force Survey (LFS) is a fool’s errand. The monthly LFS data is notoriously unreliable. Now that the standard errors are included with the LFS monthly releases, readers can evaluate it themselves. Notably, the June 2013 industry-level data is useless, the standard errors exceeding the monthly changes pretty much all the way down the list.

This semi-regular LFS blurb will touch on general concepts and historical trends. This month, given the U.S. Federal Reserve Chairman’s recent comments on using unemployment as a benchmark in the FMOC decision to turn the corporate welfare (err… ‘quantitative easing’) tap off and the subsequent reaction on both sides of the border, a comparison of the U.S. and Canadian unemployment data may be worthwhile.


Over the last 30 years, innovation in automation, competition in international freight, reduction in trade protection and little progress harmonising international labour (incl health, safety, human rights) standards has encouraged Western manufacturers to off-shore their operations to lower-wage locales.  Over the last two decades, advances in information and communication technology (ICT) have encouraged the off-shoring of not only low-skill call centre jobs, but just about every high-skill professional service imaginable, from software development to financial and radiological imaging services. As Paul Krugman recently realised, the ‘re-shoring’ trend is bringing back mostly automated operations with few if any jobs.

This has led to what one could (and will, in a future write-up) argue is a permanent structural shift in many Western industrialised economies. Hundreds of thousands of jobs lost in manufacturing have largely been replaced with retail jobs, professional service jobs largely replaced by sales and support services. All of these new jobs can be broadly characterised as more precarious, providing lower pay with little to no benefits, few opportunities for advancement and no job security, relative to the jobs they’ve replaced.

Labour market survey redesign

Labour market surveys were slow to adapt to the changing labour market structure and dynamic. Both the Bureau of Labour Statistics (BLS) and Statistics Canada updated their alternative / supplementary measures of unemployment, aka labour underutilisation, in the mid-90s.


The preceding docs detail the revisions to the U.S. Current Population Survey and the Canadian Labour Force Survey. The broadest measure of unemployment from the BLS is U-6; from the LFS, R8. To summarise, the key differences between the two measures are:

U-6 counts respondents who indicated they were either involuntarily working part-time or those who were marginally attached to the labour force for a variety of reasons. However, R8 only counts involuntary part-timers’ hours expressed as ‘full-time equivalents’, or those who were “marginally attached to the labour force because of economic reasons”.

Different approaches, different trends

Using the publicly available data from the CANSIM 282 series for the LFS, it is possible to recreate a Canadian equivalent of U-6. The results are surprising, to say the least:


A quick comparison of the BLS U-6 and the LFS U-6 equivalent shows that while labour underutilisation in the U.S. increased dramatically with the onset of The Great Recession, it’s been consistently high in Canada since the survey redesign – and likely had been for some time before that. In fact, in 1997, the first year data was available for both the CPS and LFS following their respective redesigns, Canada’s U-6 equivalent was nearly double that of the U.S. U-6, 17.0% to 8.9%.

A few other brief notes:

– The LFS R8 dramatically underestimates underemployment and marginal attachment relative to LFS U-6 equivalent. The gap between the official unemployment rate R4 and R8 is about half the size of that between R4 and Canada’s U-6 equivalent.

– Contrary to Miles Corak’s observation, measuring Canadian unemployment using the BLS equivalent (U-3) yields the same official LFS (R4) annualised unemployment rate. Unfortunately, Mr. Corak was looking at the seasonally adjusted data; had he used the unadjusted data for August 2012, he would have noticed a much smaller difference (8.2% vs 7.8%, respectively).

That said, the surveys, in addition to using slightly different populations (age 16+ for BLS, age 15+ for LFS), also ask questions slightly differently. For details, readers can consult the CPS and LFS questionnaires. That said, the differences are not that significant, as evidenced from the identical annualised unemployment rates obtained from the LFS using R4 and the U-3 equivalent.

Labour force surveys offer incomplete picture

While the BLS effort to broaden its definition of labour underutilisation (relative to Statistics Canada) is commendable, it would be impossible for a labour survey to fully capture the impact The Great Recession had on households.

Between October 2008 and December 2011, the LFS recorded an increase in Ontario unemployment of 68K (up 16.1%). However, it also recorded an additional 367K as no longer in the Ontario labour force, (up 10.8%). Only 50K of those were reclassified as ‘Not in the labour force but wanted work’. Over the same period, the number of Ontario residents receiving regular social assistance increased by 105K (up 28.4%), while the number receiving disability support increased 65K (up 19.3%) – both increases well above the population growth rate.

There are still a few figures needed to complete the picture. It’s clear, however, that many Ontarians were left far worse off than either the official or broader unemployment figures would suggest. It’s not too far a stretch to assume the same malaise in many parts of the country (if any readers are interested in helping fill in the national picture, please hit the CONTACT page).

Public policy, moving forward

In such an environment, public policies cutting social assistance, further limiting employment insurance benefits, raising the qualifying age for OAS/GIS and curtailing other social transfers would be ill-advised, to say the least.

But that’s what governments can get away with when the public is provided an incomplete picture. The loss of the long-form Census will only further obscure the extent of the problem.

Unfortunately, it’s easier to kick the can down the road than to deal openly and honestly with a growing structural problem. Since the industrial revolution, income distribution for the majority has come through the exchange of labour for a wage or salary. That system is increasingly eroding. The first post on this site showed the increasing gap between salaries/wages and corporate profits as shares of GDP between 1990 and 2008. The drop in corporate profits during the Great Recession likely caused a break in the series, but the long-term trend is obvious. Not surprisingly, consumer spending as a share of GDP dropped over the same period, as household debt soared (posts two and three on here). It doesn’t take much to figure out why, at least in Canada, there’s yet to be a sustained economic recovery: The Great Recession capped nearly two decades of unsustainable growth.

Leave a Reply

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