Canadians and their U.S. neighbours have watched official unemployment rates drop in recent months to near pre-recession levels. Predictably, the discussion has turned to how well the official rates reflect underlying post-recession labour market dynamics.
Both the Canadian and U.S. labour survey programs produce alternative unemployment measures. While the broadest U.S. measure of unemployment has recently received a great deal of attention, its Canadian counterpart has not. In addition to being more complicated, the narrow scope of the Canadian measure significantly understates labour underutilization – defeating its intended purpose.
The unemployment rate is supposed to measure the extent of underutilized labour resources in an economy. National labour survey programs significantly influence their official unemployment rates, by defining acceptable levels of willingness and availability to work. Underemployment, the share of the employed who are willing and available to work “better” or “more adequately” (as the International Labour Organization puts it), is excluded from official unemployment. Marginal attachment to the labour force – the share of the economically inactive wanting to work but otherwise classified as unavailable or unwilling – is likewise excluded. As such, official unemployment tends to understate labour underutilization.
To address this shortcoming, The U.S. Bureau of Labor Statistics (BLS) in 1976 introduced a series of alternative unemployment measures to its Current Population Survey (CPS). Following suit, in 1979 Statistics Canada introduced supplementary unemployment measures to its Labour Force Survey (LFS). In response to the changing nature of employment, the early 1990s saw both agencies revise these measures, along with their respective labour force surveys.
The broadest of these new unemployment measures – BLS U-6 in the United States, and Statscan R8 in Canada – include elements of both underemployment and marginal attachment.
Both the BLS and Statscan consider a labour force survey respondent underemployed if he or she worked part-time, despite being willing and available to work full-time. (This is known as involuntary part-time employment.) The BLS includes each such worker as underemployed in U-6.
But Statscan counts just the hours of relative underemployment for each worker – and each such worker is counted as a fraction of a full-time employee, with those fractions combined into a count of “full-time equivalents” (FTE) in R8. There were 897,000 people last year who were involuntary part-timers, according to Statscan’s survey; collectively, they were accounted for as 467,000 “full-time equivalent” workers, effectively underestimating the number of underemployed by 430,000.
(Statscan makes use of the FTE approach in its supplementary unemployment measures, but not its officially reported unemployment rate. Adding the FTE share of unused hours to unemployment would significantly raise the official unemployment rate.)
Both the BLS and Statscan determine a respondent’s marginal attachment to the labour force based on whether he or she wanted to work, and the primary reason for not actively searching for work (over the past four weeks in CPS, over the past week in LFS). The BLS counts all responses indicating a reason for marginal attachment to the labour force in U-6. But despite conceding that the marginally attached behave much like the unemployed, Statscan only counts in R8 those “marginally attached to the labour force because of economic reasons (discouragement, waiting for recall or reply from prospective employers).” Those indicating other reasons for not actively searching for work – due to illness, injury or disability, child or elderly care, other household responsibilities, school or training – are simply considered to be out of the work force.
Last year, there were 484,000 such people who indicated any of these reasons for wanting to work yet not actively searching, according to Statscan’s survey. Statscan’s R8 only counted 64,000 of them as “marginally attached”; the others were excluded entirely from the labour force count.
Add it all up, and Statscan’s definitions of underemployment and marginal attachment understated labour underutilization by 850,000 in 2012. The full count of underemployment and marginal attachment (1,381,000) exceeded official unemployment (1,368,000) last year. This also occurred in 2011, the first year since revised LFS data was published in 1997 that it happened; 2013 will make it three years in a row. The labour underutilization rate, using a full count of underemployed and marginally attached, would have been 14.2 per cent in 2012, double the official unemployment rate of 7.2 per cent.
Labour underutilization is increasingly being referenced by the media and policy makers in the United States. In Canada, however, it is rarely mentioned. The BLS U-6 is easy to calculate and understand; Statscan’s R8 is more obscure and harder to replicate.
It is important for policy makers and citizens to have multiple tools at their disposal to help them understand what’s happening in the economy, and make informed decisions. The official unemployment rate has neither captured the scale nor scope of unemployment since the Great Recession of 2008-09. A better-designed labour underutilization rate would provide greater insight into the state of the Canadian labour market.
Sam Boshra is an independent Montreal-based economist, and editor for EconomicJustice.ca. Angella MacEwen is a senior economist with the Canadian Labour Congress and research fellow at the Broadbent Institute.