2 Data
The data used for this project come from New Zealand’s 1986, 1991, 1996 and 2001 censuses of population and dwellings. The data were provided by Statistics New Zealand (SNZ) in cell-format, where the population cells are stratified by individuals’ highest educational qualification, gender, age, regional location, labour force status, and industry and occupation for those employed. The data obtained are in the form of cross-tabulations of these variables, and include cell sizes and annual income characteristics of the tabulated cells. In order to protect the confidentiality of individual responses, all cell counts have been randomly rounded to base-3.
The measure of qualifications used is a categorical measure of the highest qualification attained. Although there have been changes in the form of the census qualifications questions over the period of our study, it is possible to derive more or less consistent measures. In this paper we group the highest qualification reported into nine categories: higher degree, bachelor’s degree, other (non-University) post-school qualifications, higher-school qualification, sixth-form qualification, school-certificate, other-school qualifications, no qualification, and highest qualification not specified.[7] Employment in industry and occupation cells is measured as a count of all persons gainfully employed, irrespective of full-time or part-time status. The income data are collected in categorical form in each census.
The raw data extract includes the numbers of individuals in each of the cells, the numbers in each categorical income band and those not specifying an income band, together with summary statistics on the income distribution within the cell (mean, mean log(income), and the 10th, 25th, 50th, 75th and 90th percentiles of the distribution). In order to derive the summary statistics pertaining to the income distribution for each cell, SNZ assigned to each cell the corresponding average income estimated from the Household Economic Survey (HES), where income is recorded as a continuous variable.
The analyses of changes in qualifications, employment outcomes and incomes that follow are based on a variety of populations. In the next section we provide descriptive analyses of qualification changes for employed workers, the labour force, and the whole population; together with various subgroups of each of these populations. The following section focuses on employment changes across industry and occupation groups for the employed population. We supplement this analysis by adding in, first, the unemployed and then also those not in the labour force to consider the broader labour force and whole working-age populations. Our analysis of the qualifications-based changes human capital in Section 5 is conducted separately for the employed, labour force, and whole populations. In addition, our analyses of income concentrate on the mean income data. Finally, analyses requiring income data use the number of individuals, including those that did not specify an income, as the relevant cell size.
Notes
- [7]We also suspect that the recording of the highest qualification as “other post-school” versus one of the secondary school qualifications (“higher school”, “6th form”, and “school certificate”) was not applied using the same criteria in 1991 as in the other years. For example, table 1 shows that there was a marked increase in the fraction of the population reporting other post-school as their highest qualification from 1986 to 1991, which was subsequently reversed in 1996 and 2001; while the fractions in the school groups (especially 6th form and school certificate) show the opposite trend.
