May 2013 LFS: Great headline, statistical artifact (again) + proposed new metric

Why ‘again’? A post las summer touched on why the media should refrain from trumpeting Statscan’s notoriously unreliable monthly Labour Force Survey (LFS). Unfortinately, since that post was written, it’s been discovered that Stascan’s other major labour market survey, the Survey of Employment, Payroll and Hours (SEPH), has increasingly become unreliable since the great recession.

And the worst is yet to come, as the now completely unreliable National Household Survey (NHS) will soon form the base for all of Statscan’s social surveys, including the LFS. This was mentioned during the summer of 2010 after the long-form Census cancellation was announced, but didn’t get much traction in the media (exception).

The Canadian federal government will be happy to leave it that way, for reasons to be discussed in an upcoming post. For the time being, Canadians are left without a reliable (publicly available) source of national labour market data. Which would be bad enough, but for the fact the metrics used to report that data are likewise questionable. We’ve previously touched on the issue with the ’official’ reported unemployment rate.

While a solution to the questionable survey methodology and resulting questionable data it produces is unlikely in the foreseeable future, a change in metrics to better reflect labour market reality (instead of simply ‘a fast way to lower jobless rate’, courtesy of Miles Corak) is fairly straight-forward:

The estimate of population is a quarterly intercensal update of the Census counts. There’s no better estimate nor easy alternative measure, so it will be taken as given (although readers may note the signficant ‘adjustments’ in Census years). For context, the estimated Canadian population, non-institutional, age 15+, has grown ~6% between 2007 and 2012.

Labour Force
This is probably the most contentious defintion, although it looks fairly straight-forward at first glance. The Guide to the Labour Force Survey has a section that discusses how labour force status is defined for the survey.  It is suggested the term should refer to total available labour supply (non-institutional, age 15+). However, currently almost half a million Canadians who wanted to work but were unable to find it are defined away as Not in the labour force but wanted work. For context, the number of Canadians in this excluded group has grown ~30% between 2007 and 2012. A reasonable estimator of the labour force would include all those who wanted to work irrespective of subjective assessment of their availability, as well as those currently employed.

This is the other big headline grabber (after the unemployment rate). The LFS counts unpaid workers in family businesses and individuals who identify themselves as self-employed in an unincorporated business without paid help as employed. The LFS doesn’t ask respondents who identify as self-employed about their earnings or hours. The SEPH is a better indicator of actual paid employment. While the SEPH only enumerates businesses with payrolls, this would include self-employed in incorporated businesses that reported payroll, as well as the self-employed who are incorporated (this would include regular temp agency contractors). Unfortunately, as mentioned in the preceding, the SEPH has increasingly been including ‘unclassified business’ jobs, for which it reports no pay or hours. The number of ‘unclassifed business’ jobs has fluctuated between 363 and 440 thousand over the last year. Among the possible estimators for employment from the two labour market surveys, the estimate of jobs reported for classified businesses in SEPH best reflects actual paid employment.

As touched on in the preceding discussion of the labour force definition, the LFS only counts those deemed to be ‘available’ for work as unemployed. Again, this seems straight-forward, but in reality is rather subjective, erring on the side of discounting the unemployed. Those who were discouraged in their job search due to economic conditions, say something like  “the worst recession since the Great Depression“, and stayed home without work, either tending to family responsibilities, personal issues, and/or continuing their education while holding out for work are excluded as Not in the labour force but wanted work. A reasonable estimator of unemployment would include all those who wanted to work but could not secure employment, irrespective of subjective assessment of their availability.

Participation rate
The participation rate, reported with the monthly LFS release, is the estimated total labour force in the survey divided by the monthly population estimate (non-institutional, age 15+). The proposed estimator would use the SEPH jobs reported for classified businesses as the employment estimator, and would also include those classified as Not in the labour force but wanted work in the unemployment estimator.

Employment rate
The employment rate, reported with the monthly LFS release, is the estimated total employment in the survey divided by the monthly population estimate (non-institutional, age 15+). The proposed estimator would use the SEPH jobs reported for classified businesses, jobs for which wages and hours are known, as the employment estimator.

Unemployment rate
This is the big headline grabber. The unemployment rate, reported with the monthly LFS release, only counts those who were not working AND met a number of other conditions in the numerator. The denominator is the sum of the employed and unemployed, again within the limited definition of unemployment. The proposed estimator would include those classified as Not in the labour force but wanted work as unemployed; as such they would be included in both the numerator and denomiator.

Involuntary under-employment (part-time)
Nearly 1 million Canadians currently working part-time (1/3 of all part-time workers) are doing so involuntarily, for lack of available full-time opportunities. Unfortunately, cross-tabs showing just how involuntarily part-time these workers are (hours, pay, industry/occupation for the referenced workers) are not provided. One suspects most are in retail, food & accomodation and other low-paying, precarious service sector jobs, where many employees are hired, each assigned very limited, irregular hours (helping their employers limit vacation, sick leave, health and other benefits that might otherwise accrue to regular/full-time employees).

Table 282-0013
Labour force survey estimates (LFS), part-time employment by reason for part-time work, sex and age group, unadjusted for seasonality monthly (persons x 1,000)

There are no questions designed specifically to measure under-employment by experience/education, though it can be infered from specific cross-tabs (not possible from the limited publicly available data).

(Current and historical figures and rates using the proposed estimates to be posted)

Leave a Reply

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