I've updated and expanded the data archived on Project Link, my attempt to take the fragments of data published by Statistics Canada and piece them together into a coherent whole.
In my post introducing Project Link, I made note of a chart I came across while putting together the headline data from the Labour Force Survey going back to 1953. I had always more-or-less assumed - admittedly without looking at the data or even thinking very much about it - that the surge in female labour force participation rates started sometime in the 1960s or 1970s. But it turns out that the trend goes back to at least to 1953, when the LFS started being conducted on a monthly basis.
- Labour force
- Employment
- Unemployment rates
- Employment rates
- Participation rates
These are available for men, women and both sexes, for the age groups 15 and over as well as 25-54.
Here are a couple of charts. The one I find most interesting is for prime-age (25-54) employment rates:
As I said, I had always sort of thought that the surge of women entering the labour force was a consequence of the social and cultural transformations of the 1960s and 1970s. But in the comments in that post, Frances pointed out that there was lots of evidence pointing towards technology-based explanations: the introduction of household labour-saving devices and more sophisticated birth-control techniques freed up women to work outside the home. And indeed, prime-age female employment rates increased more-or-less uniformly for roughly 40 years between the early 1950s to the early 1990s. The rate of increase slowed during the 1990s, and the female prime-age employment rate plateaued 10 or 15 years ago. Looking back, you almost have to wonder why The Mary Tyler Moore Show was so revolutionary: it was reflecting a trend that was was already two decades old.
The other chart is for unemployment rates:
Here the story is the differing degrees to which male and female unemployment rates vary over the business cycle. There was a certain amount of commentary during the last downturn on the 'he-session': because men are disproportionately represented in the goods sector - which is also the sector most susceptible to business cycle fluctuations - the spike in male unemployment rates was much more pronounced than it was for women. It's a story you'd think would be generalizable across recessions, but not as much as I would have thought. You see the same pattern in the late 1950s and - to a lesser extent - in the early 1990s, but it's not a general rule. The recession of the early 1980s - which still holds the title for Worst Post-War Canadian Recession - hit men and women equally, and women suffered more than men during the 1970s.
A time series that starts with a trend always bothers me: when did that trend start? I'm now thinking about pushing the LFS back before 1953. Starting in (I think) 1947, the LFS was done on a quarterly basis, but with the same basic structure of the modern LFS: a survey based on a activity in a given reference week. Unresolved issues include interpolation and seasonality - but I think it's worth digging deeper here.
Well done Stephen! It's great for me to see these charts. Like you, I was looking at the recessions, and the cyclicality of male/female unemployment.
Posted by: Nick Rowe | July 10, 2017 at 08:05 PM
I'd caution the straightforward interpretation of the 1950s recession as a 'he'-cession.
Look at the combined labour force participation rate (employed + unemployed): in the 1950s men were almost universally in the labour force, whereas only 26-32% of women were. If the "in the labour force" designation was sticky, then we'd expect a priori a labour demand shock to affect the same fraction of men and women who were in the labour force.
That's closer to what we see. Around 1960, from these charts the male population was 90% employed and 7% unemployed, whereas the female population was 30% (ish) employed and 4% unemployed. 7.2% of men in the labour force were unemployed, and 11.7% of women in the labour force were unemployed.
Posted by: Majromax | July 11, 2017 at 09:18 AM
I feel like the unemployment rate can be misleading. The employment rates show smoother recessions for women throughout the series, so it seems like recessions that feature equal or greater female unemployment might also indicate women entering the workforce to make up for lost male income. Notably, the biggest gap in unemployment is the late '70s, which isn't a recession at all. That effect has now mostly played out, so empoloyment and unemployment rates tell the same story toward the end of the series.
Posted by: Neil | July 11, 2017 at 10:32 AM
Is there any historical labour force data available which separates out fathers from male non-fathers and mothers from female non-mothers?
We have lots of data showing that at the present time being a parent is correlated with men working more and women working less (in paid employment, that is), so graphs like the first one here make me wonder if we're largely ignoring an effect from changes in the birth rate: If women have 0-2 children each, they'll have far more years of "not caring for young children" than if they have 5-10 children each. How much of the increase in female employment rates from 1955 to 2010 can be explained by a population shift from "mostly with children" to "mostly without children"?
Posted by: Colin Percival | July 11, 2017 at 07:31 PM
This is fantastic, Stephen.
I'd second the comments about being careful about the interpretation of unemployment figures for women in old data. My guess is that the measures of unemployment would be quite gender-biased, so women who had some kind of household responsibilities wouldn't be considered to be unemployed on the grounds that they were not available for work. I find the employment rate is a nicer way of measuring female's participation in the paid labour market.
Posted by: Frances Woolley | July 12, 2017 at 02:55 PM
Great piece! Sylvia Ostry's 'The Female Worker in Canada', published in 1968, discusses FLFP trends dating back to 1911 and is archived at https://archive.org/details/1961995531968eng . Its main focus is the 1961 census.
Posted by: Twinklenomics | July 12, 2017 at 03:21 PM
Thanks *so* much for the reference! Already downloaded!
Posted by: Stephen Gordon | July 12, 2017 at 04:16 PM
Frances: as Québec humorist Yvon Deschamps once said "Moman travaille pas, a trop d'ouvrage" "Mom doesn't work. She's too busy"
https://amecq.ca/2017/05/16/moman-travaille-pas/ (the link is not to his monologue but to a real article...)
Posted by: Jacques René Giguère | July 12, 2017 at 05:58 PM
Jacques Rene - For sure!That's an interesting article, thanks for the link.
Posted by: Frances Woolley | July 14, 2017 at 09:44 AM