01
Working Paper
11
·
September 2026
Weak-signal
detection in regional labour data.
We test whether administrative labour series carry detectable signals of sectoral shift earlier than survey instruments. Across 14 years of New Brunswick data, three of eleven series lead by more than two quarters.
Survey instruments are the standard way to read a labour market. They are also slow: fielded, collected and published on a cycle that lags the thing they measure.
The test
We took fourteen years of New Brunswick administrative labour data and asked whether any series moved before the surveys did. Eleven series were in scope.
The result
Three of the eleven lead the survey instruments by more than two quarters. The remaining eight do not lead reliably enough to act on.