Monthly labour market data for Australia from January 1992 to April 2024. The dataset contains four related measures: the unemployment rate (\(R\)), total labour force (\(T\)), employed (\(E\)), and unemployed (\(U\)). The measures are linked by the identities $$R = 100 \times \frac{U}{T}, \qquad T = E + U.$$ In addition employment (\(E\)) and unemployment (\(U\)) follow two parallel hierarchies that:
share the national total (Australia) at the top level;
share intermediate aggregates by gender (Male and Female);
are further disaggregated by geography (
ACT,NSW,NT,QLD,SA,TAS,VIC,WA) and by age groups (15-24,25-34,35-44,45-54,55-64).
The two hierarchies partially overlap (they share upper-level aggregates), so the complete structure is more general than a hierarchical/grouped time series and can be described as a system of linearly constrained multiple time series.
Format
A data frame (or tibble) with 16296 obs. of 6 variables:
- date
First day of the reference month. Range:
1992-01-01to2024-04-01.- name
Character; geographic and age-group label.
- Rate
Numeric; the observed value of the unemployment rate (\(R\)).
- Total
Numeric; the observed value of the total labour force (\(T\)).
- Employed
Numeric; the observed value of the employed (\(E\)).
- Unemployed
Numeric; the observed value of the unemployment (\(U\)).
Source
Australian Bureau of Statistics (ABS), Labour Force, Australia — monthly time-series tables and Data Explorer extracts. Downloaded from the ABS website: https://www.abs.gov.au/statistics/labour/employment-and-unemployment/labour-force-australia.
Details
Then, this is a system of multiple time series subject to both linear constraints (across geographic and age) and non-linear relationships (the definition of \(R\)).
References
Girolimetto, D., Panagiotelis, A., Di Fonzo, T., Li, H. (2024), Forecast reconciliation with non-linear constraints, arXiv. doi:10.48550/arXiv.2510.21249