Methodology
Every rule that turns two national surveys into one comparable series.
The five harmonization rules
01
Quarterly aggregation
JOLTS and the CES employment series are monthly. Each quarter is the mean of its three months. JVWS is already quarterly, so no aggregation is needed on the Canadian side.
02
Scope cut
The US series drops government openings (JOLTS series JTS9000JOL) and government employment (CES series USGOVT). The published US total includes them; the harmonized series does not, because the comparison targets the business labour market on both sides.
03
Identical rate formula
Vacancy rate equals openings divided by employment plus openings, times 100. Both agencies publish this formula. We recompute it from the quarterly levels after the scope cut instead of averaging published rates.
04
NAICS crosswalk
JOLTS industries map to JVWS NAICS sectors. Where one JOLTS industry spans several JVWS sectors, the Canadian side is composited by summing vacancies and payroll employees first, then recomputing the rate. Ten pairs survive the crosswalk.
05
Seasonal adjustment kept
Both series are used seasonally adjusted as published. No additional seasonal model is applied.
The reference recipe
The scope recipe follows Bauer, K. (2015), Comparison of U.S. and international labor turnover statistics, Monthly Labor Review, doi:10.21916/mlr.2015.26, which documents the definitional deltas between JOLTS and peer surveys including Canada's JVWS.
Bauer, K. (2015). Comparison of U.S. and international labor turnover statistics. Monthly Labor Review. https://doi.org/10.21916/mlr.2015.26
Honest limits
- The BLS comparison article could not be fetched from this build environment, so the recipe was implemented from the published survey definitions on both sides. Check it against Bauer (2015) before citing the recipe itself.
- Canada's published total includes public administration; the US harmonized series excludes government. Public administration ran at a 2.0% vacancy rate in Canada in 2026Q2, so the residual scope gap is small but nonzero.
- JVWS includes agriculture, forestry, fishing and hunting. JOLTS covers nonfarm establishments only. Agriculture is roughly 2% of Canadian vacancies.
- JOLTS counts openings on the last business day of the month. JVWS counts vacancies on the first day of the month. Both are point-in-time stocks, one month apart in reference timing.
- The sector chart covers the JOLTS industries available as FRED series. Wholesale trade, transportation and warehousing, information, finance and insurance, professional services, and other services have no FRED JOLTS series, so those Canadian sectors have no US counterpart here.
- The US education and health series cover the private sector only. Canada's include public schools and hospitals, which narrows the measured gap in that pair.
- 2026Q3 is excluded because the quarter is incomplete in both sources.
- 2020Q2 and 2020Q3 are absent from the series: StatCan suspended JVWS collection during the first COVID wave, so no Canadian observation exists for those quarters.
What was wrong with the raw data
Nothing was broken in either source. The problem was comparability: monthly versus quarterly, all-nonfarm-including-government versus all-industries, industry cuts that do not line up. The data-fixes entry for this project records the before and after of each harmonization step.
Download the harmonized series →