Estimates of employment for the period 2000 to the present quarter shown in the SkillsInsight are generated at DoL from a mix of underlying data sources. DoL has developed a framework that offers quarterly estimates of employment for detailed (5 digit ANZSIC) industries and (5 digit NZSCO) occupations. The purpose of this framework is to consolidate a range of existing data sources that each present some but not all the required information into a single framework.

The reason for bringing together data sources into a single framework is to fill the gaps due to time lags, coverage and definition. A single framework of national level trends in employment for detailed industry and detailed occupation could also form the basis for further work. It may be used as benchmarks for regional breakdowns and could be important for the DoL’s ability to forecast employment by industry and occupation, both at the national and potentially at the regional level.

Output

1. Measure: Quarterly updated estimates of employment
2. Period: December 2000 to the latest quarter for which official (HLFS) employment estimates are available.
3. Industries: 472 ANZSIC industries
4. Occupations: 562 NZSCO occupations
5. 5. Analysis: decomposition of change in employment for an occupation in an industry over the last four years (national growth effect, industry composition effect, occupational competitiveness effect)

Methodology

Currently, there are a number of sources for indicators of employment by industry and employment by occupation which are not always consistent. This is mainly due to differences in measurement and definition, but also due to some sources containing sample/non-sample errors. At the aggregate level, official employment and unemployment estimates are available from the Household Labour Force Survey (HLFS) on a quarterly basis. These estimates are considered by many labour market policy makers and practitioners to be the overall aggregate benchmark of employment in New Zealand.

Lower level employment estimates are also released on a quarterly basis from the HLFS for 1-digit industry and occupation, while 2-digit industry and 2 and 3-digit occupation are available in unpublished format. The general consensus at DoL is, however, that due to sampling errors and coding errors, the 2 and even 1-digit industry as well as the 3 and 2-digit occupation data sourced from the HLFS is less suitable than measures from other data sources. Data from the Linked Employer-Employee Dataset (LEED) is now considered to be more superior sources for industry level employment data, given their administrative nature as opposed to relying on samples.

Estimates of employment for detailed industries primarily utilise data from annual and quarterly person level LEED data benchmarked on aggregate employment from the HLFS. The time gap between the latest quarterly LEED data and the latest HLFS publication of 4 quarters is bridged by means of forecasting LEED employment shares (of lower disaggregation level control totals) using QES (where available), time and seasonal dummies as explanatory variables.

Mapping industry employment estimates from detailed ANZSIC industries to detailed NZSCO occupations is based on interpolated and extrapolated occupational shares of industry employment available from the 2001 and 2006 Population Census.