The Folly in Forecasting Job Numbers
At the rate in which factors such as globalisation, digitisation, the adoption of new technologies and changing demographic and social profiles are impacting the state economy, assessing labour market and future employment prospects in Queensland is no menial task.
If we look at the current state of play in the labour market, figures from the Australian Bureau of Statistics show that there were around 2,510,000 employed persons in Queensland in November 2018, consisting of 1,694,400 (67.5%) full-time and 815,600 (32.5%) part-time employees.
Quarterly data shows that between 1984 to 2018 employed persons grew by about 44,000 per year on average (2.7%); full-time jobs grew on average by around 6,350 per month (25,400 per annum) and part-time jobs 4,660 per month (18,660 per annum).
Just over the last decade, however, Queensland employment has only managed to grow by an average of 30,250 persons per year (1.3%).
The detailed ABS original labour force data for November 2018 showed that the largest employer industry in Queensland was Health Care and Social Assistance with about 13.3% of total employed, followed by Retail Trade at 10.6%, Construction at 8.9%, Education and Training at 8.7% and Accommodation and Food Services at 8%.
November 2018 Trend ABS regional employment data (SA4) compiled by ConusCBC Staff Selection shows that around half of people employed in Queensland work in Greater Brisbane, and seven in ten in South East Queensland.
Looking at education and training statistics for Queensland, despite commencement hold up relatively steady over the past five years, the number of apprentices and trainees in training in vocational education and training (VET) has fallen in line with a lower number of completions.
Across trades, the only occupational group to have seen growth over this time has been in construction trades.
Regarding non-trade trainees, the only occupational group to have grown over the five years is machinery operators and drivers.
On the other hand, commencements in the Australian higher education sector grew by 7.5% to over 457,000 students from 2014 to 2018, with major fields being Management and Commerce (20.7%), Society and Culture (20.6%) and Health (17.3%).
Potential outlooks of Queensland industry employment
Recently, Jobs Queensland released a report entitled Anticipating Future Skills: Jobs growth and alternative futures for Queensland to 2022 , which presented results of projections of employment in Queensland under various scenarios devised through a series of workshops with Jobs Queensland stakeholders.
The report provides detailed projections down to the regional, occupational and qualification level, which the report hopes can play a role in forming decision making by government in VET investment as well as skills and training policy development.
The specific scenarios examined include:
(i) Baseline or ‘business as usual’;
(ii) Scenario 1: involves technological change which increases labour productivity within all industries in Queensland to 0.25% greater than the baseline annually;
(iii) Scenario 2: involves an increase in interstate migration with a decreasing proportion of working age population; and,
(iv) Scenario 3: involves halving key commodity prices such as coal.
The baseline scenario was generated using an elaborate integration of a ‘family’ of models centred on the Victoria University Computable General Equilibrium model of the Australian economy, which brings together demographic, employment and macroeconomic data as well as forecasts from government and industry bodies.
In this article I am going to use a different approach to generate a baseline forecast using simple linear regression, which doesn’t rely on scenario planning or making assumptions about the future with any more sophistication than that it will evolve as a function of observations from the past.
Accordingly, using the previous 10 years of quarterly data from February 2009 to November 2018 (40 observations) produces the following line fit plot and regression statistics.
The regression statistics show a strong positive association with time with a correlation coefficient of 0.96 and a coefficient of determination (adjusted R2 value) of 0.91 - the mean relative error from the regression line is 0.86% and standard deviation is 0.63%, which implies a fairly tight fit of the data.
The slope of the regression line is shown to be statistically significant and indicates that jobs can be expected to be grow by 7,136 a quarter (around 28,500 a year), which further implies that Queensland employment is projected to be about 143,000 greater in 5 years.
The following table summarises the outputs of regressions performed on the quarterly original employment data across each of the ANSIC industry sectors including the association as well as the magnitude of employment changes within the various industries over time.
The statistics outline whether the relationship is a positive or negative association, the correlation coefficient (r), the coefficient of determination (R2 adjusted for sample size), as well as the coefficients describing the regression line (the intercept b0 and slope b1).
While the regression statistics indicate that the slope of the regression line is statistically significant across most of the industry divisions, the linear modelling provides a poor fit regarding Information Media and Telecommunications, Construction, Electricity, Gas, Water and Waste. Caution should also be applied to Retail Trade as well as Rental, Hiring and Real Estate, which are only significant at the 10% level (a p-value below 5% is the common level of significance to reject the null hypothesis that the slope coefficient is equal to zero).
Further, the industries in which employment is expected to experience the greatest rate of growth include Health and Social Assistance (11,120 per annum), Education and Training (5,010 p.a.), Professional, Scientific and Technical Services (3,400 p.a.), Accommodation and Food Services (3,370 p.a.) and Public Administration and Safety (3,080 p.a.).
In contrast, employment is expected to fall moderately over time in Agriculture, Forestry and Fishing (2,680 p.a.), Manufacturing (1,340 p.a.) and Wholesale Trade (1,050 p.a.).
Comparing the output from the regressions with the Jobs Queensland projections out to 2021-22 shows that there are only marginally small relative errors in aggregate terms, in fact as low as between 1-3%.
However, there are some rather significant differences across the individual industry employment projections relating to each scenario.
First, let’s look at the comparisons between the projections of the regression outputs and the Jobs Queensland baseline scenario.
The regression outputs indicate substantially larger employment growth in Education and Training, Accommodation and Food Services, Mining as well as Public Administration and moderately greater growth in Health Care and Social Services, Wholesale and Retail Trade.
While the Jobs Queensland scenarios projections indicate relatively greater growth in Professional, Scientific and Technical Services, Manufacturing, Other Services as well as Rental, Hiring and Real Estate, Agriculture, Forestry and Fisheries and Construction.
The following tables show the comparisons between projections for each of the other scenarios.
While these projections are reasonably consistent across industry categories, however, they differ in the relative magnitude of employment outcomes within industry sectors.
Below are the detailed regressions and linear forecasts of employment in Queensland for each ANZSIC industry out to 2024.
Agriculture, Forestry and Fisheries
Electricity, Gas, Water and Waste
Accommodation & Food Services
Transport, Postal & Warehousing
Information, Media & Telecommunications
Financial & Insurance Services
Rental, Hiring & Real Estate
Professional, Scientific & Technical Services
Administrative & Support Services
Public Administration & Safety
Education & Training
Health Care & Social Assistance
Arts & Recreation