May 2014 LFS: An alternative view. How Canada defines ‘part-time’, and why it matters (updated June 9, 2014)

Chart 1 Full-time and part-time status by actual hours for employeesChart 1 Full-time and part-time status by actual hours for employees

Source: CANSIM Table 282-0022 – Labour force survey estimates (LFS), by actual hours worked, class of worker, North American Industry Classification System (NAICS) and sex, annual (persons unless otherwise noted), Statistics Canada

(No, it’s not ‘usual’ hours – more on that below.)

Going by the media reports today’s May 2014 Labour Force Survey (LFS) release has prompted interest in the rise in part-time work since the Great Recession. (The continued self-employment ‘noise’ shouldn’t be ignored – LFS estimated a 40,400 drop in self-employed last month.)

Ever wonder why Canada defines full-time work as starting at 30 ‘usual’ hours per week? That’s a bit presumptuous, as readers may not be aware that Canada defines it as such. Understandable, given that it’s not the norm in most industrialised countries.

OECD Labour Market and Social Policy Occasional Papers, No. 22: The Definition of Part-Time Work for the Purpose of International Comparisons
Alois van Bastelaer, Georges Lemaître, Pascal Marianna, Organization for Economic Co-operation and Development 1997

While both the Census (work activity) and LFS (type of work) dictionaries currently define full-time work as mostly/usually working at least 30 hours per week, neither indicate why that particular number of hours was selected as the cut-off, nor when the decision was made to go with that particular cut-off. The linked OECD report notes

A national hours cut-off is used to differentiate between part-time and full-time work in Australia (35 hours), Austria (35), Canada (30), Finland (30), Hungary (36), Iceland (35), Japan (35), New Zealand (30), Norway (37), Sweden (35), Turkey (36) and the United States (35)…

Canada: Data are based on averages of monthly estimates from the Labour Force Survey published in The Labour Force, Statistics Canada, catalogue No. 71-001. Part-time is defined on the basis of total usual hours for the main job, less than 35 for the years prior to 1975, less than 30 thereafter. Estimates were available for both definitions for 1975, and estimates for years prior to then have been adjusted using a ratio of new-to-old estimates calculated for 1975.

If France, known for its 4-day-work-week, or Finland, known for its generous social welfare system, uses 30 hours, that may make sense given the context. Neither situation characterises present-day Canada.

What’s more, the U.S.A., Canada’s primary trade partner – and the country most Canadians compare their economy’s relative performance to – defines full-time as starting at 35 ‘usual’ hours worked per week. Mexico, Canada’s other North American Free Trade Agreement partner takes it further, defining full-time as 35 ‘actual’ hours per week.


As Chart 1 shows, defining full-time at 35 ‘actual’ hours per week for Canada shows a more pronounced trend toward part-time employment. And the greatest jump in part-time work using that definition occurs neither following the 1990 nor 2008 recessions, but rather between 2000 and 2003.

So what changed in the Canadian labour market during that period to explain the sudden and sustained jump in actual (> 35 hours) part-time work?

This is more general observation / theory than analysis, but it looks like a sudden jump in female labour market participation – by 2.5% (59.4 to 61.9%) in just 3 years (2000-2003) – had something to do with it. Female LM participation had been flat over the preceding decade (58.5% in 1990), and remained flat over the proceeding decade (62.1% in 2013).

The source table for Chart 1 also includes data by industry, which likewise points to female LM participation as the source. While the gap between 35 and 30 hour part-time work increased in both goods and service producing sectors, production sector part-time employment (either definition) stagnated / declined starting in 2000, whereas service sector employment consistently increased over the same period (temporary drop in 2008-2009).

Disproportionately more women work in the service sector, and practically all the increase in female employment between 2000 and 2003 (493,000 out of 530,000) was in the service sector. Half of that increase was in part-time employment.

So what caused the sudden jump in female LM participation around that time? Likely not something easily derived from the LFS alone. One possible explanation is the heavily subsidised day-care program Canada’s second largest province (Québec) introduced in 1997. Unfortunately, the LFS only started collecting limited data on child-care challenges that same year, with its revised ‘Reason not looking for work’ question.

If the preceding is indeed true, then it begs the question: Is it really progress if heavily subsidized day care leads to more women working lower-paying, part-time service sector jobs? That research suggests children enrolled in the subsidized day care program suffer significantly poorer early childhood education outcomes further calls into question the merit of such programs.

… Then again, that’s all speculation based on a handful of variables and annualised data. Certainly would be an interesting topic for empirical research, if it hasn’t been done already.

‘Actual’ vs ‘Usual’ hours

Chart 1 shows ‘actual’ instead of ‘usual’ hours, the latter being the official measure of full-time / part-time employment.

 Chart 2 Full-time and part-time status by usual hours for employees
Chart 2 Full-time and part-time status by usual hours for employees

Source: CANSIM Table 282-0019 – Labour force survey estimates (LFS), by usual hours worked, class of worker, North American Industry Classification System (NAICS) and sex, annual (persons unless otherwise noted), Statistics Canada

The one question for which there’s no clear answer is the large and widened gap starting in 2000 between ‘actual’ and ‘usual’ hours (Chart 2) – for both part-time definitions, but in particular for 35-hours plus.

Whither the workweek?
Diane Galarneau, Jean-Pierre Maynard, Jin Lee, Perspectives on Labour and Income Vol. 6, no. 6 (Statistics Canada) June 2005

It looks like StatsCan also noticed the widened gap between LFS reported usual and actual hours over the period 2000 and 2003. Remarkably, the StatsCan analysis overlooks the spike in female LM participation. It also attributes the drop in hours to more men than women taking up part-time work over that period, despite the data – at least the most recently revised – showing the opposite.

The article writes off the 2000-2003 decline in hours / jump in actual part-time work as a combination of underestimation of actual hours in 2000 (“More than half of the decrease was in fact due to survey methodology”), increased vacation time taken by older workers and increased maternity leave stemming from a change in EI policy.

To be fair, that analysis was published in 2005, without the benefit of time. That the jump in actual part-time work wasn’t a one-off, but rather persisted after 2003 suggests it’s worth a second look. What happened between 2000 and 2003 may be a peculiar coincidence, as the StatsCan analysis suggests. Or it may help shed light on the increasingly part-time nature of the Canadian labour market and the widening gap between actual and usual part-time work.

Leave a Reply

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