Extract: Seasonally Adjusting Unemployment Time Series
13 June 2018
Topic
Unemployment
This technical note provides an analysis on the revision to Singapore's methodology for seasonally adjusting unemployment time series, explaining the shift from forward to concurrent seasonal adjustment from Q1 2017, the rationale behind the change, and the associated revision policy for unemployment data.
Read the Seasonally Adjusting Unemployment Time Series Extract
KEY INSIGHTS
Unemployment time series are influenced by seasonality — periodic fluctuations from recurring calendar events — and seasonal adjustment removes these influences to enable more meaningful trend analysis. MRSD moved to concurrent seasonal adjustment from Q1 2017 after noticing unstable seasonal patterns in 2016 citizen unemployment data, which had caused the forward approach to produce more erratic results.
Concurrent adjustment is more effective as time series components can be better estimated as new data becomes available.
Under MRSD's revision policy, previous quarters' estimates are not revised during the year; instead, revisions are done once annually based on final December data.
