modelling of future skills demand: the implications for...

Background Research for the NSW Strategic Skills Plan 2011 ‐ 2015

Ian Watson – Freelance Researcher and Visiting Senior Research Fellow Macquarie University

Modelling of future skills demand: the implications for skills planning in NSW

Contents

Contents i

List of Tables ii

1 Introduction 11.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 How the modelling works . . . . . . . . . . . . . . . . . . 2

2 Comparing forecasts for 2025 32.1 Three global scenarios . . . . . . . . . . . . . . . . . . . . 32.2 Industry forecasts . . . . . . . . . . . . . . . . . . . . . . 52.3 Occupational forecasts . . . . . . . . . . . . . . . . . . . . 62.4 Educational qualifications forecasts . . . . . . . . . . . . . 72.5 Additional qualification requirements in 2015 . . . . . . . 92.6 The supply of qualifications . . . . . . . . . . . . . . . . . 13

3 Assumptions behind the modelling 163.1 The demographic ‘time-bomb’ . . . . . . . . . . . . . . . . 163.2 Climate change and the CPRS . . . . . . . . . . . . . . . 183.3 The global financial crisis . . . . . . . . . . . . . . . . . . 193.4 Long-term labour market changes . . . . . . . . . . . . . 20

4 Neglected issues 224.1 Other scenarios . . . . . . . . . . . . . . . . . . . . . . . . 224.2 Methodological limitations . . . . . . . . . . . . . . . . . . 25

5 The NSW dimension 285.1 New South Wales Access Economics tables . . . . . . . . 285.2 The IPART report . . . . . . . . . . . . . . . . . . . . . . . 305.3 The Foresighting Study . . . . . . . . . . . . . . . . . . . 315.4 The New South Wales State Plan . . . . . . . . . . . . . . 31

6 Other relevant publications 336.1 UK Commission for Employment

and Skills . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

7 Conclusion: planning for the medium term 357.1 Shortcomings in modelling qualifications . . . . . . . . . . 357.2 Extrapolating trends . . . . . . . . . . . . . . . . . . . . . 367.3 Key informants . . . . . . . . . . . . . . . . . . . . . . . . 377.4 Context and regions . . . . . . . . . . . . . . . . . . . . . 37

References 39

Modelling of future skills demand: the implications for skills planning in NSW i

List of Tables

2.1 Macro model parameters, CEET and Access Economics . . 42.2 Industry projections for 2025, comparisons . . . . . . . . . . 62.3 Occupational projections for 2025, comparisons . . . . . . . 72.4 Educational qualifications projections for 2025, comparisons 82.5 Additional educational qualifications required in 2015, com-

parisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.6 Skills deepening rates for 2015, comparisons . . . . . . . . . 122.7 Potential shortfall in qualifications in 2015, (thousands) . . . 142.8 Potential shortfall with net migration included, (thousands) . . 14

3.1 Intergenerational report: projections and outcomes . . . . . . 17

5.1 Additional educational qualifications Australia and NSW,comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5.2 Potential shortfall in qualifications in 2015, New SouthWales, (thousands) . . . . . . . . . . . . . . . . . . . . . . . . 29

Modelling of future skills demand: the implications for skills planning in NSW ii

1. Introduction

1.1 Overview

This report aims to provide an accessibleintroduction to some recent modelling of futureskills demand. It primarily compares tworeports, the Economic Modelling of SkillsDemand prepared by Access Economics in2009 and the Demand for qualifications andthe future labour market in Australia preparedby Chandra Shah from the Centre for theEconomics of Education and Training (CEET).The first report is referred to hereafter as the‘Access Economics report’ and the latter isreferred to as the ‘CEET report’. For ease ofexpression, Access Economics and CEET arereferred to as authors (though Shah (2010)and CEET are used interchangeably at times).A comparison of these reports forms the basisof Chapter 2, where a series of forecasts for2025 are directly compared (though 2015 isalso used for some of this). While a number ofother reports are available which also deal withthe issue of skills demand, only these twopresent their findings in ways which aredirectly comparable. In other reports, the datais presented as growth rates, rather thanlevels, the time frames differ, or the relevantpopulations are not comparable. Some ofthese other reports are considered briefly inthe later chapters of this report.As well as forecasts of employment andqualification holding in 2025, Chapter 2 alsolooks at the projections for the demand foreducation qualifications in 2015. This makes itpossible to pinpoint likely shortfalls in thesupply of qualified persons in coming years.Chapter 3 examines some of the keyassumptions behind the economic modelling.Some of these assumptions enter directly intothe modelling discussed in Chapter 2, whereasothers provide the context in which the

modelling takes place. In the case of AccessEconomics these assumptions are craftedaround three different scenarios, all of whichreflect responses to the Global Financial Crisis.However, these scenarios are far fromexhaustive, and in Chapter 4 I look at otherscenarios. These are not articulated asscenarios, per se, but they do point to anumber of issues which have not beencovered by the Access Economics and CEETmodelling. These include peak oil and theprospect of prolonged economic decline. Thischapter also looks at some of themethodological limitations in the modelling ofskills demand. These include the inability ofthese approaches to incorporate majorqualitative, as opposed to quantitative,changes into their modelling strategies.Chapter 5 discusses the New South Walesdata. While much of it is limited and thusbeyond inclusion in this report, there are someAccess Economics projections for the Statewhich can be directly related to the maindiscussions in this report. This analysis issupplemented by a discussion of theForesighting Study and mention of the AccessEconomics analysis on future VET demand(conducted for IPART). Finally, the New SouthWales State Plan is considered, and thequestion posed as to whether the COAGtargets look likely to be met.Chapter 6 looks at other related material,namely work done by the UK Commission forEmployment Skills.The final chapter draws together theinconsistencies arising from the economicmodelling examined in this report and askshow policy makers should deal with suchdilemmas. The conclusion suggests that amore balanced strategy might be possible by

Modelling of future skills demand: the implications for skills planning in NSW 1

focusing on medium term planning using twomethodologies. First, extrapolating fromcurrent trends would allow for planning for fourto five years into the future. Secondly, thiscould be supplemented by undertakingin-depth interviews with key informants, peoplewho are positioned at strategic locations in thelabour market and whose insights arelong-term and far-reaching.

1.2 How the modelling works

I discuss the modelling process in greaterdetail in the following chapter. However, it isworth summarising, in this introduction, themain steps that are involved. This provides auseful backdrop to both the CEET and AccessEconomics comparisons; even though theyused different macroeconomic models, theunderlying strategies are similar.The goal pursued by both CEET and AccessEconomics is to arrive at final estimates of thenumbers of employed persons holdingqualifications, as well as the demand foradditional qualifications. This demand can beviewed as reflecting labour market demand forskills.The starting point is a set of macroeconomicvariables, either those expected to prevail orthose allowed to vary by scenario. In the caseof the Access Economics Macro (AEM) model,these inputs are based on three differentscenarios (outlined later) while in the case ofthe CEET modelling, these inputs are basedon assumptions about current trends orforecasts provided by other parties.The steps to link these ends of the chainconsist of:

1. specifying the macroeconomic variableswhich form part of the scenarios (AccessEconomics) or are part of the currenttrends / projections (CEET); these arethe parameters for the model; somevariables are fixed at the outset; someare implied by the choice of othervariables;

2. generating estimates of industry output;3. based on these estimates, generating

industry employment forecasts; thisinvolves using input-output tables (which

specify inter-industry linkages in theeconomy) to map final components ofdemand to industry employment;

4. using these industry employmentforecasts to generate occupationalemployment forecasts; this involvesusing the 2006 Census information onthe occupational share held by eachindustry;

5. using the combined industry andoccupational data (the intersection ofwhich can be regarded as a ‘job’) togenerate forecasts of persons holdingqualifications. Australian Bureau ofStatistics survey data (such as the 2007ABS Survey of Education and Work) isused to map jobs to qualifications on theassumption that historical qualificationprofiles are likely to apply;

6. this stock of qualified persons is thenused, in combination with other data onqualification profiles, to generateforecasts of likely future demand foradditional qualifications. For AccessEconomics, the data on qualificationprofiles include assumptions from thescenarios as well as historical data.

There are many refinements along the lengthof this chain, some of which I comment on inthe following chapter. They include allowing forskills deepening, where the growth inqualifications occurs at a greater rate than thegrowth in employment, as well as for multiplequalification holding (in the case of AccessEconomics). The stock of qualified personsdiminishes over time as workers retire, so itneeds to be ‘refreshed’ by new entrants.Analysing these flows requires assumptionsabout the growth of the labour force and theretirement rates likely to prevail. This analysisthen feeds into forecasts for additionalqualifications required in the future. Finally,analysis of this labour market demand forqualifications can be supplemented withanalysis of labour market supply ofqualifications—the demand by students forqualifications. This analysis makes use of dataon student completions and can be used toforecast potential shortfalls in the supply ofqualified workers in coming years.


2. Comparing forecasts for 2025

There are a number of considerations whichhave to be taken into account when comparingeconomic forecasts. There is the actual modelitself—its system of equations and it sequenceof calculations—and there is the data requiredfor the model to work. At mentioned in the lastchapter, this includes the input-output tables,historical data on trends, recent Census tableson various proportions, and other relevantsources. Finally, there are the assumptions:expectations of economic growth, othermacroeconomic variables, and particular policysettings.The outcomes from such modelling can bemeasured as stocks: the actual numbers ofpeople in particular industries and occupationsand the qualifications they hold. They can alsobe measured as annual rates of growthbetween particular years (say 2020 to 2025).These are still stock measures, but they reflectgrowth rates, rather than levels. In thediscussion which follows, I present stocks aslevels, based on the actual counts found in theCEET and Access Economics reports. But Ialso present these levels expressed asproportions. This is a very useful way tocompare relativities and examinecompositional change. The actual numbers inthe counts don’t matter all that much, giventhey are forecasts and the time frame is so farinto the future. What matters most is whetherthere is structural change of a compositionalnature emerging from the forecasts. As weshall see shortly, one of the Access Economicsforecasts expects manufacturing to play amuch more important role than do the otherforecasts. This then flows through to adifferent occupational and qualification profilefor the workforce.As well as stocks, labour market forecastingalso incorporates flows, the recognition that

new entrants add to the stock, whilst retireesand other persons exiting the labour market,subtract from the stock. The analysis of flowsis important for understanding the likely supplyof persons with particular qualifications workingin particular occupations. Measuring thelabour market demand for qualifications is notdone directly but is based on the assumptionthat the number of persons working in anoccupation, holding certain qualifications, is astraightforward reflection of the demand forthose jobs and for those qualifications.Finally, all modelling involves stipulating therelevant population. The Access Economicsmodelling restricts analysis to employedpersons while the CEET modelling alsoproduces estimates for non-employedpersons. However, CEET also producestabulated estimates for employed persons sothat direct comparisons between AccessEconomics forecasts and CEET forecasts arefeasible. In this report I refer to employedpersons as the ‘workforce’, while the terms‘labour force’ and ‘labour market’ also includethe unemployed. Finally, ‘non-employedpersons’ can include both the unemployed andwhat the Australian Bureau of Statistics termsNILF (not in the labour force).

2.1 Three global scenarios

The Access Economics modelling providesthree different forecasts for the Australianeconomy in 2025. They are based on the threedifferent scenarios developed by SkillsAustralia (Skills Australia 2009) for thedevelopment of the National WorkforceDevelopment Strategy (Skills Australia 2010).It is worth briefly summarising the economiccharacter of these scenarios in order tounderstand how their differing assumptionsinfluence the modelling:


Table 2.1: Macro model parameters, CEET and Access Economics

CEET Access Economics

OpenDoors

Low-trust Flags

Global growth rate (%) 3.80 3.10 2.60

Assumptions

Australian growth rate1 (%) 2.70 3.93 3.00 2.20Annual net immigration1 (’000s) 180 250 200 100Annual net immigration1 (%) 1.00 0.80 0.40Labour productivity growth1 (%) 1.75 1.50 1.30Exports1 (%) 6.29 4.35 2.94Capital / labour ratio1 (%) 1.50 1.00 0.50

Implied results

Population growth1 (%) 1.70 1.47 1.02Workforce growth1 (%) 2.12 1.49 0.96Employment growth1 (%) 2.15 1.48 0.89Unemployment rate2 (%) 4.5 5.1 6.0Participation rate2 (%) 63.0 68.8 64.2 63.1Exports to GDP2 (%) 30.5 26.0 23.3Business investment to GDP2 (%) 14.0 12.2 10.5

Notes: 1. Annual growth rates for period 2010 to 2025. 2. Levels as at 2025.

Source: Taken from Access Economics (2009a, Table 2.2).

1. Open Doors: national growth pathsfollow long-term trends and globalisationcontinues with improved marketstructures;

2. Low-trust Globalisation: increasedcompetition; expansion of naturalresource industries; growth in role ofmultinational corporations; increasedtrade barriers; focus on bilateral trade;

3. Flags: global depression and expansionof bilateral and regional agreements;increased trade barriers and morenationalistic political culture.

As noted in the first chapter, variousassumptions are fed into the economic modelsas parameters and various estimates ofindustry employment emerge. Theseparameters can be varied for differentscenarios; they can be constrained to fixedvalues; and they can be determined by otherparameters. In the case of the forecasts doneby Access Economics which are examinedhere, these scenarios are entered into themodel as assumptions about economicgrowth, migration, productivity, export intensityand the capital labour ratio. These parametersin turn generate estimates for population

growth, employment growth, the participationrate, and so forth. These parameters, and theirintermediate estimates are shown in the lastthree columns of Table 2.1, and these aretaken directly from the Access Economicsreport. In the case of the CEET modelling,which makes use of the MONASH employmentforecasting model, the parameters are notexplicitly presented. However, in discussingthe macroeconomic context for the CEETreport Shah (2010, p. 4) provides some of thekey projections from the Treasury modelling forthe Intergenerational Report 2010. One canonly assume that these are the parameters forthe MONASH model and they are included inTable 2.1 in the first column.This table shows that the scenarios differ quiteconsiderably across these important economicvariables, with several of the Flagssettings—such as economic growth andimmigration—between 40 per cent and 60 percent lower than for the Open Doors scenario.Favourable outcomes—such as lowunemployment and high participation—arealso considerably worse under the Flagsscenario. The CEET parameters place theirmodel closer to the Low-trust globalisationscenario when it comes to economic growthand net immigration, but closer to the Flags


model in terms of the assumption about theparticipation rate.Scenarios are a useful device to canvas arange of possibilities, but they are somewhatinflexible. For example, Open Doors is anexpansive scenario in every respect, Flags is aconservative, inward-looking scenario for allsettings, and Low-trust is simply amiddle-position. What is missing is therecognition that some elements of onescenario might be mixed with those of another.For example, in the next few years globalgrowth might limp along at 2.6 per cent, assuggested by Flags, but the Australian growthrate might move along at 3.0 per cent, assuggested by Low-trust globalisation. Thismay happen because Australia’s economiclinks with China may shelter the domesticeconomy from the continuing downturn inEurope and North America. The currentlyrevived population debate might seeimmigration reduced, in line with Flags, butlabour productivity might achieve 1.5 per cent,in line with the Low-trust scenario, as a resultof large-scale investment in infrastructure,such as the National Broadband Network.This hybrid scenario is probably more realisticin the current climate than the three scenariosexplored by Skills Australia. And it is not as ifthat exploration is dated—these scenarioswere prepared as recently as June 2009. Whatthis suggests is that continuing uncertainty inthe global economy, and the speed ofdevelopments on the Australian politicallandscape, render much current speculationabout future developments unrealistic. Tocountenance various hybrid models in themodelling would involve permutations on thesethree scenarios and make the exercisecumbersome. One might end up, for example,with six or nine sets of forecasts instead of justthree.What are the implications of this? From aforecasting point of view, not a lot. Morerealistic mixed scenarios, like these hybrids,might simply result in some of the parameterscancelling each other out. Without more detailon the equations used in this modelling, it’shard to be precise but it seems likely that thefinal forecasts would simply end upsomewhere in the middle of the range between

the most expansionist scenario (Open Doors)and the least (Flags.)

2.2 Industry forecasts

What do these different scenarios imply foremployment growth in 2025? Do the outcomesdiffer all that much? Table 2.2 certainly showsquite different outcomes in terms of the size ofthe workforce, with a spread between 12.5million and 15.3 million. However, this tablealso suggests that in a number of key areas,the structure of employment in the economywill not differ all that much irrespective of whichscenario eventuates. Large public sectoremployment industries, such as education,health and government differ by less than 1percentage point across the differentscenarios. Core private sector areas likefinance, real estate, construction and miningdiffer by very small amounts—generally about0.5 percentage points—while largeemployment sectors like retail differ by about 1percentage points. The most striking change inforecasts across the three scenarios concernsmanufacturing. Under the Open Doors andLow-trust globalisation scenariosmanufacturing employment is expected toconstitute about 6.7 per cent of the workforce.Under the Flags scenario it will make up nearlytwice this proportion: 12.6 per cent. This isconsistent with the more closed-economycharacteristics of that scenario.How do these Access Economics forecastscompare with the CEET forecasts? As notedearlier, the latter are not based on scenarioprojections, but an extrapolation of currenttrends which take into account the context ofthe forecasts made by the IntergenerationalReport 2010 and the possible impact of theproposed Carbon Pollution Reduction Scheme(CPRS) (Shah 2010, pp.8–9). Shah (2010)does not explicitly tabulate the parameterswhich were fed into the MONASH model toproduce the employment forecasts, but onecan surmise that this macroeconomic contextinfluenced the parameters chosen.In terms of the size of the workforce, the CEETforecasts place it quite close to the Low-trustglobalisation projections. However, in relativeterms the forecasts for the CEET study areoften closer to the Open Doors scenario for


Table 2.2: Industry projections for 2025, comparisons


Open Doors Low-trust Flags

’000s % ’000s % ’000s % ’000s %

Agriculture, forestry & fishing 297 2.2 455 3.0 361 2.6 277 2.2Mining 229 1.7 241 1.6 163 1.2 133 1.1Manufacturing 886 6.6 1,032 6.7 914 6.7 1,573 12.6Electricity, gas, water & waste 215 1.6 144 0.9 117 0.9 217 1.7Construction 1,207 9.0 1,325 8.6 1,134 8.3 1,019 8.1Wholesale trade 525 3.9 486 3.2 423 3.1 332 2.7Retail trade 1,550 11.5 1,800 11.7 1,629 11.9 1,362 10.9Accommodation & food 863 6.4 1,016 6.6 936 6.8 779 6.2Transport, postal etc 781 5.8 1,003 6.5 821 6.0 673 5.4Info media & telecom 250 1.9 332 2.2 352 2.6 318 2.5Financial & insurance services 478 3.6 549 3.6 535 3.9 443 3.5Rental, hiring & real estate 246 1.8 345 2.3 305 2.2 253 2.0Prof, scientific & tech services 1,179 8.8 1,337 8.7 1,179 8.6 979 7.8Admin & support services 469 3.5 531 3.5 536 3.9 449 3.6Public admin & safety 835 6.2 992 6.5 911 6.6 762 6.1Education & training 1,112 8.3 1,096 7.2 1,011 7.4 929 7.4Health care & social assistance 1,547 11.5 1,733 11.3 1,615 11.8 1,359 10.9Art & recreation services 292 2.2 286 1.9 255 1.9 180 1.4Other services 497 3.7 619 4.0 545 4.0 468 3.7Total 13,458 100.0 15,322 100.0 13,742 100.0 12,504 100.0

Notes: For explanation of Access Economics scenarios see Skills Australia (2009).

Source: Shah (2010, Table 22); Access Economics (2009a, Tables 6.1, 6.2, 6.3).

Population: Employed persons, Australia.

particular industry sectors. This applies toareas like mining, manufacturing, construction,wholesale and retail, accommodation and food,information and media, finance, professionalservices, health and arts. Where the CEETprojections do come closer to Low-trustglobalisation in relative terms is in sectors liketransport and education. The CEET forecastsare closer to the Flags scenario for sectors likeagriculture, electricity etc, real estate andgovernment. The CEET study matches theAccess Economics projections formanufacturing for both the Open Doors andLow-trust globalisation scenarios but not forthe Flags scenario where this extraordinaryfigure of 12.6 per cent remains a stark outlier.Employment in the education and trainingsector are somewhat higher for CEET than forany of the Access Economics forecasts. Thisis presumably based on the expectation thatthe current trend of employment shifting tohigh-skill occupations continues and therebyreinforces a continuing expansion in educationand training provision. It is unclear what roleoverseas student enrolments play in thesescenarios. The Global Financial Crisis (GFC)

has played a role in subduing suchenrolments, and this might influence outcomesfor the two lower-growth Access Economicsscenarios. However, a much bigger impact islikely to flow from immigration policy changesaround permanent residency which wereannounced in early 2010. These changes areexpected to have a major impact on overseasenrolments, particularly among privateeducation providers. It is unlikely that this hasbeen factored into any of the modelling.

2.3 Occupational forecasts

As noted earlier the occupational forecastsflow from the industry projections, usingoccupational employment shares from the2006 Census. In the case of AccessEconomics the conversion from industry tooccupation makes use of the older ASCOcoding system at the 3 digit level, beforeconverting to the current ANZSCO codingsystem (at the 1 digit level). In the case of theCEET projections, it is not clear how theconversions are done, but the presentation ofoccupational results at the ANZSCO 2 digit


Table 2.3: Occupational projections for 2025, comparisons



’000s % ’000s % ’000s % ’000s %

Managers 1,947 14.5 1,940 12.7 1,711 12.4 1,559 12.5Professionals 3,116 23.2 3,247 21.2 2,934 21.3 2,608 20.9Technicians & trades workers 1,888 14.0 2,188 14.3 1,945 14.2 1,891 15.1Community & personal service 1,307 9.7 1,414 9.2 1,279 9.3 1,123 9.0Clerical & administrative workers 1,929 14.3 2,381 15.5 2,149 15.6 1,885 15.1Sales workers 1,166 8.7 1,574 10.3 1,431 10.4 1,219 9.7Machinery operators & drivers 864 6.4 1,005 6.6 892 6.5 859 6.9Labourers 1,240 9.2 1,573 10.3 1,402 10.2 1,361 10.9Total 13,457 100.0 15,322 100.0 13,743 100.0 12,504 100.0




level suggests that the CEET modelling stayswith the ANZSCO system for the entireprocess.The most interesting aspect to theoccupational projections is the similaritybetween the three Access Economicsforecasts and the sharp difference with theCEET forecasts. This applies, however, only tothe top and bottom of the occupational ladder.CEET provides very optimistic projections formanagers and professionals, expecting themto constitute 37.7 per cent of the workforce in2025. By comparison, even the mostexpansionist Access Economics scenario(Open Doors) expects that these twooccupations will constitute 33.9 per cent. Onthe other hand, Access Economics expectslabourers to make up between 10.2 and 10.9per cent of the workforce while CEET hasthem at just 9.2 per cent. A similar pattern isevident with salesworkers: Access Economicsexpects them to constitute between 9.7 and10.4 per cent, while CEET puts the figure atjust 8.7 per cent. CEET clearly foresees amore highly-skilled future for the Australianworkforce than does Access Economics.The stronger position of manufacturing in theAccess Economics industry forecasts for theFlags scenario is evident in the occupationalprojections. Technicians and trades workers,machinery operators and drivers, andlabourers all feature strongly in the Flagsprojections, where they total 32.9 per cent ofthe workforce. The CEET projections, with

their expectation of a more highly-skilledfuture, put the figure for these threeoccupational groups at 29.6 per cent. It’simportant to keep in mind that a difference of2.3 percentage points represents about310,000 workers (in terms of the CEETworkforce total).

2.4 Educational qualifications forecasts

Projections of future educational qualificationsare based on extrapolating from currentoccupational qualification profiles, using theoccupational forecasts just discussed. Inaddition, allowance is made for ‘skillsdeepening’ within occupations. This meansthat one does not just extrapolate with a staticratio which reflects the qualifications profile ofan occupation as that pertained at a particulartime. Rather, those occupations wherequalification-holding is growing at a faster ratethan the occupation itself (the definition ofskills deepening) make use of a higher ratiowhich reflects this trend. Both the CEETforecasts and the Access Economics forecastsincorporate trends in skills deepening into theirprojections.The forecasts for educational qualifications in2025 are shown in Table 2.4 and follow quitelogically from the occupational forecasts. Withits expectation of a high-skills future, it is nosurprise to see that CEET expects theproportion of the workforce who holdqualifications to be quite high (76.6 per cent), afigure in excess of the most expansionist of the


Table 2.4: Educational qualifications projections for 2025, comparisons



’000s % ’000s % ’000s % ’000s %

Higher education 4,324 32.1 5,238 34.2 4,631 33.7 4,039 32.3VET sub-total 5,985 44.5 6,332 41.3 5,500 40.0 4,934 39.5

Diploma / Advanced Diploma 2,391 17.8 2,001 13.1 1,745 12.7 1,552 12.4Certificate III / IV 3,184 23.7 3,425 22.4 2,970 21.6 2,685 21.5Certificate I / II 410 3.0 906 5.9 786 5.7 698 5.6

Qualifications sub-total 10,309 76.6 11,570 75.5 10,131 73.7 8,973 71.8No qualifications 3,148 23.4 3,752 24.5 3,612 26.3 3,531 28.2

Total 13,457 100.0 15,322 100.0 13,743 100.0 12,504 100.0




Access Economics forecasts of 75.5 per cent.At the other end of the scale—the AccessEconomics Flags scenario—this figure is just71.8 per cent. We saw earlier that theoccupational forecasts for the Flags scenarioexpected higher proportions of blue-collarworkers—following on from the elevated rolefor manufacturing—and this is likely to seelower proportions of qualified workers. Bycontrast, the Open Doors scenario foresawmore managers and professionals, and thisflows into the qualifications forecasts whichthey make. Higher education qualifications, forexample, make up 34.2 per cent of the total inthe Access Economics projections for thisexpansionist scenario.What is interesting, however, is that while theCEET projections for a highly-skilled futurehave particularly high expectations for moremanagers and professionals, this does nottranslate into particularly large forecasts forhigher education qualifications. Indeed, allAccess Economics forecasts in this regard arehigher than the CEET forecasts.Where the higher qualifications profileexpected by CEET—the 76.6 per centfigure—shows up most notably is in VETqualifications, particularly Diploma andAdvanced Diploma qualifications. The gapshere are quite remarkable: the 17.8 per centfigure compares with 13.1 per cent under OpenDoors and just 12.4 per cent under Flags.There are also notable differences withCertificate III / IV qualifications, but not of the

same magnitude. CEET expects 23.7 per centof the workforce to have this level ofqualifications, whereas the Access Economicsfigures range from 21.5 per to 22.4 per cent.Finally, at the lowest level of VET the pattern isreversed, with CEET expecting only 3 per centof the workforce to hold Certificate I / IIqualifications and Access Economicsexpecting proportions which are nearly doublethis.Clearly, the CEET vision of a high-skills futuretranslates within the VET field into anexpectation of greater proportions of the moreadvanced qualifications than is the case withany of the Access Economics projections.As mentioned earlier CEET incorporates theshifting qualifications profile of the Australianpopulation in its modelling and Shah (2010,p. 5) observes that the growth in Diplomaholding has been increasing at a rate of 10 percent per annum while the number holdingCertificate I qualifications has been declining.These trends are clearly influencing the CEETprojections more strongly than is the case forAccess Economics. However, as Shah (2010,p. 49) also notes in his conclusion, it isuncertain how much of the growth in Diplomaholding is driven by labour market demandbecause of the possibility of ‘supply-induceddemand’. He suggests that more informationwould be needed to resolve this.


Table 2.5: Additional educational qualifications required in 2015, comparisons



’000s % ’000s % ’000s % ’000s %

Higher education 162 34.0 351 45.5 302 46.8 253 46.8VET sub-total 315 66.0 420 54.5 344 53.2 287 53.2

Diploma / Advanced Diploma 123 25.8 138 17.9 116 18.0 97 17.9Certificate III / IV 163 34.2 190 24.7 156 24.1 133 24.7Certificate I / II 29 6.1 91 11.8 72 11.2 57 10.6

Total 477 100.0 771 100.0 646 100.0 540 100.0



Population: Additional number of qualifications required by the workforce, Australia.

2.5 Additional qualificationrequirements in 2015

The discussion in this section changes in twoimportant ways: the time frame differs and theperspective shifts from stocks to flows. Theforecasts are now for 2015 because the CEETprojections are tabulated in such a way thatthis is the only common year between theirforecasts and those from Access Economics.Secondly, one needs a flows perspectivewhich recognises that the stock ofqualifications is diminished as people retire,but is refreshed by new entrants to the labourmarket, either from the educational system orthrough net migration (that is, additionaleducated migrants less departing educatedlocals). If the stock itself is expected togrow—as it does in all the forecasts discussedin the last section—then the rate of inflow mustexceed the outflow. Finally, if skills deepeningis taking place, then additional qualificationsare needed, independently of the growth instocks. Access Economics provides a usefulsummary of these concepts (AccessEconomics 2009a, p. 58) and uses a simpledecomposition framework to present itsforecasts. This consists of:

1. employment growth overall;2. additional requirements to offset

retirements; and3. the influence of skills deepening.

Using this framework, Access Economicsprovides estimates for annual requirements in2015 and totals based on these are shown inTable 2.5 along with percentage breakdowns.

In the case of the CEET report, the data ispresented differently showing just new entrantsand existing workers. It is possible that thelatter is equivalent to skills deepening, butCEET does not present it as such. The totalsfrom the CEET report, along with percentagebreakdowns, are shown in Table 2.5. Despiteslightly different phrasing—Access Economicsuses the term ‘additional’—both reportspresent comparable data at the aggregatelevel for future qualification requirements for2015 and this forms the basis of thecomparison in the following discussion.There is also an issue of the relevantpopulation. In the data being compared thecommon population is the workforce, that is,the employed population. CEET also presentsforecasts which include the non-employedpopulation which are considerably higher thanthe numbers discussed below. However,Access Economics does not discuss thispopulation but restricts its forecasts to theemployed population only. Consequently, inthe discussion below, the population referredto remains the workforce. A furthercomplication is the fact that the AccessEconomics forecasts include multiplequalification holding, something which CEETdoes not incorporate. (I go into more detailsabout this issue on page 11 below). Apart fromthis, the two different sets of forecasts are ascomparable as one can get.Turning to the comparisons, the differencesbetween CEET and Access Economics arequite stark. By 2015 CEET is expecting thedemand for additional qualifications to numberabout 477,000. By contrast, none of the


Access Economics forecasts come close tothis: they range from a conservative 540,000to an expansionist 771,000. As noted earlier,the CEET assumptions come closer to theOpen Doors and Low-trust scenarios, so thefact that these projections are so far fromthese more relevant scenarios suggests quitedifferent assumptions are entering themodelling. I explore this issue shortly.Looking at relativities, the breakdowns withinthese additional qualifications are also quiterevealing. CEET expects that VETrequirements will make up 66 per cent of thetotal. It’s worth noting that while this figurerefers to 2015, the discussion in the CEETreport makes it clear that this figure is close tothe average across the whole period (whichcannot be easily tabulated here because ofincompatible presentations between CEET andAccess Economics). The Access Economicsforecasts for VET are considerably below this:ranging from 53.2 per cent to 54.5 per cent.The pattern we saw earlier, in which higherlevel VET qualifications are emphasised byCEET, is also evident here. CEET expectsDiplomas and Advanced Diplomas to make upnearly 26 per cent of the total; AccessEconomics puts this figure at about 18 percent, a difference of 8 percentage points. Thestarkest difference, however, is at theCertificate III / IV level: CEET expects just overa third of the total to be at this level; AccessEconomics expects demand at this level to beonly a quarter of the total. This is a differenceof nearly 10 percentage points. Notsurprisingly, at the other end of the scale,additional demand for Certificate I / IIqualifications is forecast by CEET to be verylow, at just 6.1 per cent. The AccessEconomics figures range from 10.6 per cent to11.8 per cent. Finally, the higher educationfigures also reflect the earlier pattern: CEETdownplays this while Access Economicsemphasises it. The difference is considerable:between 11 and 13 percentage points.As mentioned above, the issue of differencesin magnitude does require further discussion.To begin, it’s worth bringing these differencesinto relief by comparing Tables 2.4 and 2.5above (and rounding off the numbers forconvenience). Looking first at Table 2.4 the

CEET projections and the Low-trust projectionsare quite close: around 13.5 million persons.The expectation that between 4.3 and 4.6million workers will have higher educationalqualifications is also in reasonable agreement.Similarly, the difference between 3.2 millionand 3 million workers with Certificate III / IVqualifications does not stretch credibility.However, the size of the discrepancy in theDiploma / Advanced Diploma differences doesraise eyebrows: CEET expects 2.4 million,Access Economics expects 1.7 million. Whilea difference of this size is likely to affect furthercomparisons it is still fair to say that theamount of overlap between CEET and AccessEconomics is not unreasonable, particularlygiven the magnitudes involved (measurementsin the millions) and the time frame of 2025.Let us now turn to Table 2.5, the Table whichhas greater policy and planning implicationsbecause it constitutes the estimates of futuredemand. Here the magnitudes are in thehundreds of thousands and the time frame ismuch closer, 2015. Yet the overlaps are fewand far between. The closest matchingforecasts are between CEET and the Flagsscenario, but even this puts the totals at oddsby more than 60,000. The differences forhigher education are in the order of 90,000,though the differences for VET are in the orderof about 30,000. To be consistent, however,we should compare the CEET forecasts withthe Low-trust scenario, particularly given thatmost of the earlier assumptions suggest acloser match with this comparison. In thiscase, the overall gap is about 170,000; thehigher education difference is about 140,000and the VET difference is again 30,000. Thisdifference, however, lies in the oppositedirection. In the Flags scenario—despite alarger overall total—Access Economicsexpected that VET qualifications would amountto an additional 287,000 workers, 28,000 lessthan CEET. For the Low-trust scenario, AccessEconomics expected that VET would amountto 344,000 additional workers, 29,000 morethan CEET.Why are there such large discrepancies? Asmentioned earlier, forecasts for Diploma /Advanced Diploma levels differ considerablybetween the two reports, but this is insufficient


to explain these differences in additionalrequirements.One likely explanation is multiple qualificationholding which is likely to boost the overallmagnitude in the Access Economics forecastsby a considerable amount. As the CEET reportnotes,

The total number of qualificationsthat would need to be completed inthis period is likely to be higherthan estimates provided in thisreport because some people do[hold] multiple qualifications. Lowerlevel qualifications are sometimesstepping stones to higher levelqualifications (Shah 2010, p. 49).

Some indication of the size of this difference isevident if we compare the Access Economicsestimates for highest level qualifications(discussed in the last section) with those formultiple qualification holding. The latter hasnot been presented so far but the data are tobe found in Tables 8.10, 8.11 and 8.12 (AccessEconomics 2009a, p. 57). For 2015 the highestlevel of qualification numbers are at 8.6million, 8.1 million and 7.7 million for the threescenarios (Open Doors, Low-trust and Flags).The equivalent figures for multiple qualificationholding are 12.9 million, 12.1 million and 11.4million. The ratio between these figures isalmost constant across scenarios, at about1.5. In other words, using the assumption thatmultiple qualification holding should be takeninto account produces estimates of populationtotals which are 1.5 times greater than wouldbe the case if this assumption were ignored.Of course, this set of estimates is for the totalstock of qualifications, not for the additionalqualifications required. This means that wecan’t just expect the latter to be magnified by1.5 in the Access Economics forecasts. This isbecause assumptions about retirements andskills deepening also contribute to the finalforecasts.

Skills deepening

An obvious question which arises at this pointis whether there is overlap between multiplequalification holding and skills deepening. It is

quite likely that some of the skillsdeepening—that is, growth in qualificationholding at a faster rate than the growth in theworkforce—involves existing qualificationholders taking on additional qualifications, aswell as the trend in which qualification holdingspreads more extensively within the workforce.While Access Economics explain theirmethodology (Access Economics 2009a,p. 58) for calculating additional qualifications,the detail is insufficient to know how they havedealt with this potential for overlap andwhether it might constitute a form ofdouble-counting. In the case of CEET, there islittle discussion of this issue of multiplequalification holding, though assumptionsabout skills deepening play an important role intheir forecasts (Shah 2010, p. 46).Is it possible to compare the assumptionsregarding skills deepening? Shah (2010, p. 46)notes that the CEET estimates for skillsdeepening within occupations use ‘recentlyobserved trends’ so they incorporate theexpectation that skills deepening will increaseover time. Access Economics takes a similarapproach, in which ‘rules of thumb related tohistoric trends’ are used rather than anymeasure of ‘underlying requirements’ (AccessEconomics 2009a, p. 77).Because the Access Economics and CEETreports report their breakdowns differently, it’snot easy to directly compare estimates of thecontribution made by skills deepening. If weassume that the qualifications required for theexisting workforce in the CEET report areequivalent to the skills deepening componentof the Access Economics report, then we cantentatively compare these figures. These skillsdeepening rates are expressed as proportionsin Table 2.6 and are labelled ‘skills deepening’(with the caveat that this is only assumed to bean appropriate term for the CEET figures).It appears that CEET expects much higherrates of skills deepening in 2015, somewherein the order of 0.40. (That is, 40 per cent of theadditional qualifications will be accounted forby skills deepening.) By comparison, the figurefor Access Economics across all scenarios is0.29. The differences for higher education areless marked: CEET expects a figure of 0.31while the Access Economics figures range


Table 2.6: Skills deepening rates for 2015, comparisons


OpenDoors

Low-trust

Flags

Higher education 0.31 0.33 0.34 0.36VET sub-total 0.46 0.25 0.25 0.24

Diploma / Advanced Diploma 0.56 0.33 0.33 0.34Certificate III / IV 0.40 0.29 0.29 0.28Certificate I / II 0.34 0.05 0.01 -0.03

Total 0.40 0.29 0.29 0.29


Source: Calculations based on Shah (2010, Table 36); Access Economics (2009a, Tables 8.13, 8.14,

8.15).


from 0.33 to 0.36 and all are greater than theCEET figure. By contrast, in the VET areaCEET expects much higher skills deepening,an average of 0.46 and a figure as high as0.56 for Diplomas and Advanced Diplomas.Access Economics, on the other hand, not onlyputs the VET figures much lower than those forhigher education, but considerably lower thanthe equivalent CEET figures. The average forVET is just 0.25 and for Diplomas andAdvanced Diplomas it is 0.33.These CEET expectations for higher levels ofskills deepening among the VET qualifiedworkforce partly explain the significant rolegiven to VET in the CEET forecasts. As notedearlier, CEET expects 66 per cent of additionaleducational qualifications in 2015 to come fromthe VET sector, a figure some 11 percentagepoints higher than what Access Economicsexpects. Apart from the Flags scenario, theactual numbers forecast by CEET are lowerthan for Access Economics but this is becauseCEET forecasts such low totals to begin with(the 477,000 figure).

Retirements

The other important issue which influencesfuture requirements for qualifications is theassumption made about retirements. AccessEconomics puts the replacement rates whichtake account of people leaving the labour force(or moving permanently to another occupation)at 0.30, 0.35 and 0.41 for the three scenarios(Open Doors, Low-trust and Flags). In otherwords, somewhere between one third and twofifths of the future demand for educational

requirements arises because of departuresfrom the labour force due to retirement. Theseestimates are for the near future—2015—whilethose for the more distant future—2025—aresomewhat higher, reflecting the demographictime-bomb assumptions: 0.33, 0.40 and 0.50.What is interesting about these figures is thatthe highest retirement rates apply to the mostconservative scenario, the inward-lookingscenario of Flags in which immigration levelsare more subdued. In this scenario retirementsnot only constitute the largest ratio, but onewhich increases considerably over time. InOpen Doors, for example, the increase is from0.30 to 0.33. For Flags the increase is from0.41 to 0.50. With reduced immigration, theageing of the population results in considerablyhigher retirement rates because the agestructure of the population is not beingbalanced by as many migrants, a group whoseaverage age is younger than the population asa whole.Access Economics calculates its estimates forretirements using five year age cohorts andthree-digit ASCO counts from the 2006Census (Access Economics 2009a,pp. 75–76). These retirements reflect thenumber of retirements to be expected in eachoccupation in each forward period. In this way,the ‘baby-boomer bulge’, for example, movesthrough the forecasts producing differentretirement rates at different points in time.Because the CEET projections are tabulateddifferently it’s not possible to compareretirement rates. CEET uses the category ‘newentrants’ to cover both the growth in the labour


force and the need to replace departingworkers. In assessing the numbers involved,CEET includes a number of causes of labourforce departure (emigration, death, retirement)as well as taking account of ‘occupationalturnover’ (Shah 2010, pp. 58, 2). Shah (2010,p. 58) does discuss the assumptions made inthe calculations for retirement rates, whichmake use of age, sex and occupational trendsfor the last few years. He estimatesreplacement rates of between 2.0 per cent and2.3 per cent. These figures don’t relate to theAccess Economics figures because theserepresent the rate at which the stock ofqualified persons will decline in coming years.Fortunately, the CEET report does include atable showing the relative importance ofreplacement needs compared to employmentgrowth and these can be used to calculate aratio which is equivalent to the ratio of the firsttwo components (employment growth andretirements) in the Access Economics tables.CEET puts these two rates at 1.4 per cent foreconomic growth and 2.3 per cent for netreplacement (Shah 2010, p. 60), a ratio of1.65. It is important to note that these rates arefor the period 2010 to 2025. The economicgrowth components for the Access Economicsforecasts are: 0.37, 0.29 and 0.19. Thesefigures have been selected for 2025, in orderto approach some degree of comparability withthe CEET forecasts. Expressing these growthfigures as a ratio to the retirement ratesmentioned earlier produces ratios of: 0.89,1.38 and 2.63.In summary, only the Low-trust scenariocomes close to the CEET assumptions aboutthe likely ratio between how much futurequalification demand will be driven byreplacement needs, and how much it will bedriven by economic growth. CEET expectsreplacements to be 1.65 times greater thanwhat economic growth will require, AccessEconomics expects it to be about 1.38 timesgreater. However, in the Flags scenario, whichreflects a low growth future for Australia, thisfigure is as high as 2.63. By contrast, in theOpen Doors scenario, with its higher rates ofGDP growth and its high immigration levels,economic growth will actually contribute moreto future demand for qualifications than will theneed to replace retiring workers.

2.6 The supply of qualifications

The discussion of skills demand up to this pointhas relied on the assumption that the demandfor skills can be measured by the employmentof workers holding certain qualifications. Asthe CEET report puts it: ‘[The modelling] alsoassumes that the number holding qualificationsrepresents the true demand for qualifications.’(Shah 2010, p. 46). As I suggest in the nextchapter, this assumption can be misleadingbecause it unproblematically equatesqualifications with skills and it ignores theunder-utilisation of skills in the workplace.Nevertheless, in the absence of other datawhich might more directly measure skillsdemand, this assumption is a reasonableoperational necessity.As well as looking at qualifications in terms oflabour market demand, it can also be looked atin terms of student demand. This is bestmeasured by student completions and can beregarded as one measure for the potentialsupply of skilled workers in the future. Netmigration is another source of future supply.While CEET does not provide information oncourse completions, Access Economics does.It also provides information on the contributionmade by net migration. Combining these datawith the additional qualifications requirementsdiscussed in the last section allows AccessEconomics to estimate possible shortfalls inthe supply of qualified persons in comingyears. As noted earlier, the requirementforecasts are based on assumptions aboutemployment growth, skills deepening andretirement rates and these are used to produce‘implied’ labour market demand for variouslevels of qualification (Access Economics2009a, p. 62).By way of contrast, the estimates for futurestudent course completions are based onequations which combine projections fromYear 12 completion rates, unemployment ratesand wages. The first of these is a structuralcomponent, which is largely independent of thebusiness cycle and reflects ‘underlyingchanges in the demand for education(Australia’s move to becoming a higher skilleconomy)’ (Access Economics 2009a, p. 25).Educational policy over the last two decades


Table 2.7: Potential shortfall in qualifications in 2015, (thousands)


Supply Demand Difference Supply Demand Difference Supply Demand Difference

Higher education 203 351 -148 199 302 -103 191 253 -61VET sub-total 330 420 -90 325 344 -19 315 287 27

Diploma / Advanced Diploma 51 138 -88 50 116 -66 48 97 -49Certificate III / IV 188 190 -3 185 156 29 179 133 45Certificate I / II 92 91 1 91 72 19 88 57 31

Total 533 771 -238 524 646 -121 506 540 -34

Notes: Supply refers to projected student demand (supply of students); demand refers to implied labour market demand; difference is supply less demand and

represents the shortfall.

For explanation of Access Economics scenarios see Skills Australia (2009).

Source: Access Economics (2009a, Tables 9.1, 9.2, 9.3).


Table 2.8: Potential shortfall with net migration included, (thousands)


Higher education -75 -44 -31VET sub-total -51 12 44

Diploma / Advanced Diploma -67 -50 -40Certificate III / IV 15 43 53Certificate I / II 1 19 31

Total -127 -32 13

Notes: For 2015. Figures are the equivalent of the ‘Difference’ column shown in Table 2.7. For ex-

planation of Access Economics scenarios see Skills Australia (2009).

Source: Access Economics (2009a, Tables 9.4).


which has focussed on raising Year 12retention rates is part of this framework.The latter two components—unemploymentrates and wages—are more cyclical in theireffects. They shape decisions students makebetween studying and entering the labourmarket. A classic instance of this was evidentduring the recessions of the early 1980s and1990s, when school retention rates rosedramatically as young people realised theirlabour market options were very limited.Relative wage rates also influence suchdecisions, as well as shaping the enrolmentpatterns between different sectors, such as thechoice between higher education versusvocational education.Access Economics presents forecasts showingthe balance between the implied labour marketdemand for qualifications and the supplybased on student completions. Forconsistency with the last section, I present onlythe 2015 forecasts (though Access Economicsalso provides figures for 2020 and 2025). As

Table 2.7 makes clear, Access Economicsexpects shortfalls in the supply of students forhigher education across all scenarios, thoughthe magnitude is considerably reduced in theFlags scenario. When it comes to VET, noneof the scenarios envisage shortfalls forCertificate I and II students, but all scenariosdo show shortfalls for Diplomas and AdvancedDiplomas, though the magnitudes do not differall that much (unlike the higher educationdifferences).When it comes to Certificate III and IVstudents, there is some variation according tothe scenario. In the high-growth Open Doorsscenario shortfalls are expected, though themagnitude is very small in relative terms.Neither the Low-trust nor Flags scenariosexpect shortfalls at the Certificate III and IVlevels. In other words, essentially all of theshortfalls in VET student completions areexpected to occur in the Diploma andAdvanced Diploma stream, with the differences


between scenarios far less pronounced than isthe case for higher education.What is striking about Table 2.7 is theassumption that the supply of students willbarely change at all across these differingscenarios—even though they imply quitedifferent assumptions about population growth(as was evident in Table 2.1). Whether ashortfall eventuates is almost entirely driven bythe labour market demand for qualifications.For example, comparing overall VET figures,the differences between Open Doors andFlags entails a reduction of just 4.5 per cent instudent completions. On the other hand, thecomparable figure for labour market demand isa reduction of 31.6 per cent as one movesfrom Open Doors to Flags.Access Economics also goes on to calculatewhat the shortfall in qualifications might bewhen net migration is incorporated. Thismeans taking account of the differencebetween losses to the labour market fromqualified Australians emigrating and the gainsto the labour market from qualified foreignersimmigrating to Australia. This modified versionof the shortfalls table is shown in Table 2.8.The most notable feature of this table is thatthere is an improvement in the situation acrossthe board, but the improvement is muchgreater in higher education than in VET. Theeffect of net migration is to almost halve theshortfall in higher education. By contrast, inthe VET sector, the shortfalls for Diplomas andAdvanced Diplomas is reduced by only aboutone quarter. In terms of Certificate levelcourses, the impact of immigration is anexpected over-supply of qualifications—andthis applies across all scenarios, though it isparticularly pronounced for the Low-trust andFlags scenarios.What is particularly interesting about theseVET forecasts is that in the earlier projectionsfor additional educational qualifications (seeTable 2.5) Access Economics expected muchlower relative demand for Diplomas andAdvanced Diplomas than did CEET. The latterexpected them to constitute nearly 26 per centof all additional educational qualificationrequirements; the former put the figure ataround 18 per cent. And yet in these shortfallforecasts, the Access Economics figures for

VET highlight Diplomas and AdvancedDiplomas as the areas of concern. If CEET’s(forthcoming) modelling of the supply ofeducational qualifications is in the vicinity ofthe Access Economics projections—which aswe’ve just seen, don’t vary all that much bydifferent scenarios—then one could assumethat the CEET forecasts for shortfalls ofDiplomas and Advanced Diplomas may beconsiderably higher than these AccessEconomics figures.


3. Assumptions behind the modelling

There are a number of assumptions whichform the backdrop to the forecasts discussedin the last chapter. Some are long termstructural changes, others are more recentpolicy-oriented changes, while others reflectthe recent contemporary economic situation.Some of these issues enter directly into themodel parameters, and others are simplydiscussed as part of the context.

3.1 The demographic ‘time-bomb’

Over the last decade the Australiangovernment has produced threeIntergenerational Reports, generallyabbreviated to IGR 2002–03, IGR 2007 andIGR 2010 (Commonwealth Treasury 2002;Commonwealth Treasury 2007;Commonwealth Treasury 2010). While themost recent Intergenerational Report (IGR2010) sparked a public debate on immigrationand the potential size of Australia’s population,it was the first one (IGR 2002–03) which hadthe most impact on public policy in terms ofconfronting the challenges posed bydemographic change.The ageing of the population, with its potentialhealth costs, and the impending departurefrom the labour force of the baby boomergeneration, with the potential loss of taxationrevenue and workforce skills, were centralfeatures of the analysis. When combined, theimpact of these were seen as a majorchallenge for Commonwealth finances andwere often discussed in terms of a rising‘dependency ratio’. More broadly, theeconomic implications of an ageing populationfor labour supply were also canvassed.The policy responses which have beendiscussed since 2003 have emphasisedincreased immigration, higher levels of labour

force participation, and efforts to increaseproductivity. A prosperous future was seen todepend on achieving good outcomes in theseareas to offset the demographic ‘time bomb’.This policy framework is a constant backdropfor both the CEET and Access Economicsmodelling and enters directly into their modelparameters. For example, assumptions aboutdiffering levels of immigration and populationgrowth are key inputs in all three AccessEconomics scenarios. In the case of the CEETmodelling, assumptions about immigrationlevels based on the latest IGR appear to bepart of the parameters for the MONASH model.In its discussion of the macroeconomiccontext, the CEET report draws on the mostrecent IGR to emphasis the challenge of anageing population, particularly the need to liftparticipation rates (Shah 2010, pp. 1–2).Access Economics observes that the threescenarios imply different outcomes in terms ofpopulation growth, with the Open Doorsscenario more likely to ensure no long termdecline in population. However, in terms of theage composition of the population, all threescenarios point in the same direction. Inparticular, projected growth in the primestudent cohort (aged 18 to 22) sees a majordecline for the period 2010 to 2020, beforereviving and moving upward to reach levelsstill below those of 2010. Access Economicsemphasises the ‘demographic challenge facingAustralia’ with the impending retirement ofbaby boomers and the risk that inadequateparticipation rates will not be able tocompensate for their departure from the labourmarket (Access Economics 2009a, p. ii–iii).While the projections from the IGR 2002–03have been revised in the more recent reports,it is worth looking at these early projectionsbecause we are now in a position to evaluate


Table 3.1: Intergenerational report: projections and outcomes

Intergenerational Report 2002-03 (millions) ABS

Age range 2002 2012 2022 2032 2042 2009

0 to 14 3.9 3.8 3.8 3.8 3.7 4.215 to 64 13.2 14.6 15.1 15.3 15.4 14.965 to 84 2.2 2.7 3.8 4.7 5.1 2.585+ 0.3 0.4 0.5 0.8 1.1 0.4Persons 19.6 21.5 23.2 24.5 25.3 22.0

Working age as % 67.3 67.9 65.1 62.4 60.1 67.7Older to working age as % 18.9 21.2 28.5 35.9 40.3 19.5

Notes: IGR 2002 and ABS figures for 2009 are actual outcomes; all others are projections.

Source: IGR from Commonwealth Treasury (2002, Table 3, p. 22); ABS from ABS (2009b, Table 9).

Population: Australian population, all persons.

their accuracy. Table 3.1 shows the populationprojections from that report and the relevantAustralian Bureau of Statistics populationfigures for 2009. The closest forecast year forthese actual figures is 2012, so thecomparisons are made with that year.The first notable feature of Table 3.1 is theconsistent trend in the projected numbers foryoung people (0 to 14) as they slide graduallydownward. At the same time, the numbers ofolder people (those 65 and older) rises quitesharply. Indeed, by 2042 this group isprojected to have more than doubled in size(from 2.5 million to 6.2 million). Where theconcerns about dependency ratios surfaces isin the shrinking ratio of the working agepopulation to the remainder of the population.This is projected to fall from 67.3 per cent in2002 to 60.1 per cent by 2042. While youngpeople make up part of the dependent sectorof the population, it is mainly the older agegroups which are the cause for policy concern.As Table 3.1 shows, this group is projected toincrease from 18.9 per cent to 40.3 per centover this period. In other words, where therewere over 5 workers generating nationalincome to support one retired person in 2002,by 2042 there will only be 2.5 workers carryingthis load. This kind of scenario is the basis formuch of the policy angst which seeks toencourage greater workforce participationamong all groups, and to delay retirementamong older age groups.The actual outcomes in 2009 show aninteresting break from these projections. For astart, the gradual decline in young people isreversed, with an increase in their numbers to

4.2 million rather than a decline to 3.8 million.This reflects recent changes in the birth ratewhich have begun to depart from long termtrends. At the same time, the rate of growth inthe working age population is greater than thatforeseen in the IGR. By 2012 the IGRexpected this group to number 14.6 million, upfrom 13.2 million. By 2009 they had alreadyreached 14.9 million, well ahead ofexpectations. This boost mostly reflects recordlevels of immigration over recent years, levelsnot anticipated during the late 1990s.The rise in the percentage of the populationwho are of working age may still reachexpectations—in 2009 it was 67.7 per cent andit was expected to reach 67.9 per cent by2012. After that, it is expected to decline. Ofcourse, the unexpected boost in the youngerage group (0 to 14) may alter this trajectory,since as they mature in coming decades theywill enter the labour force and boost thenumbers of working age persons.As for the concern about older persons relyingon the working age population, this ratio hasnot increased at the rate expected. It was 18.9per cent in 2002 and had only grown to 19.5per cent by 2009. It seems unlikely to reach21.2 per cent by 2012.So is the demographic time-bomb lookingmore or less likely to eventuate? These figuressuggest its ultimate trajectory is unchanged,but the rates of change implied in the IGR2002–03 look unlikely. The closest matchoccurs with the figure of the working ageproportion: 67.9 compared with 67.7 per cent.But the reason this figure is close, given the


above-expected increase in the working agepopulation, is because of the influx of youngpeople (0 to 14) into the dependent sector,thereby pulling this rate further down thanmight otherwise be the case. However, thisyounger cohort will eventually join the labourforce and thus improve this ratio of workers tonon-workers. The dependency issue whichhas most troubled policy-makers—the ratio ofolder persons to working age persons—doesappear to be moderating, partly because netmigration levels have boosted the numbers ofworking age persons more than wasanticipated back in 2002.

3.2 Climate change and the CPRS

Recent Australian research, such as the reportby the Australian Academy of Science (2010),outline the likely effects of climate change inAustralia. These range from highertemperatures, more extreme weather events(such as floods and cyclones), rising sea-waterlevels and coastal flooding, higher levels ofocean acidity, declining biodiversity andincreased migration flows (‘climate refugees’).The direct impact of these on the economy islikely to be profound, particularly in sectors likeagriculture, tourism, insurance and propertydevelopment. These impacts will flow throughto businesses and households, and particularlyto state and local governments.Despite this reality, in a short space of time thecentrality of climate change to political debatehas receded. It was central to the politicaldebates in the 2007 election, but marginal tothe debates in 2010. The failure of theCopenhagen process, and the renewed effortsof climate change skeptics, has seen publicconcern about climate change wane in recentyears (as tracked by the Lowy Poll, forexample).Prior to the current impasse, public policy inAustralia had moved in recent years to directlyaddress climate change, with Australia’sGarnaut Report echoing the UK’s Stern Reportin arguing that the long-term economic costs ofignoring climate change would outweigh theburden of dealing with the problem in theimmediate future (Garnaut 2008; Stern 2006).Australia’s policy position, until quite recently,entailed a Carbon Pollution Reduction Scheme

centred on an emissions trading framework.This policy response, and the actual risks ofclimate change, have therefore been animportant backdrop for much Governmentcommissioned research in recent years. BothAccess Economics and the CEET reportemphasise the importance of climate changeand the CPRS in their discussions of thebroader economic context.In the case of the CEET report, the challengeof climate change forms part of the context forthe forecasting but it does not enter themodelling process per se. In discussing theemployment forecasts for 2010 to 2025, theCEET report observes:

MONASH forecasts do not includethe impact of climate mitigation andadaption policies that are likely tobe implemented over the next fewyears, although the impact of somelarge infrastructure projects that arerelated to climate change, such asin the water industry, are likely tohave been factored in.Nevertheless, the model is reallyabout a ‘business as usual’ casebut one which ignores the costs ofnot tackling climate change, coststhat Stern (2006) and Garnaut(2008) believe will be incurred.Future research in modelling theeconomy does need to considerthe impact of climatic events thatare predicted to increase infrequency. The modelling will needto incorporate the risk associatedwith these events. These risks aremeasurable and have already beencalculated by climate scientists(Shah 2010, p. 25).

In other contexts, the MONASH model hasbeen used to model the employment effects ofresponding to climate change (such as movingmore decisively to renewables) but this doesnot appear to have been incorporated into theCEET modelling. The CEET report stresses,however, that policies like the CPRS will haveimplications for skills formation, and foreseesthat additional skills would be needed to dealwith new tasks, new tools and new materials.


It also sees the requirement to reduce carbonpollution leading to innovation in industry and adiffusion of new knowledge into the widerpopulation (Shah 2010, p. 2).In the case of Access Economics, the OpenDoors scenario includes the prospect of aglobal agreement to mitigate climate change.Earlier modelling by Concept Economics in2008 examined the effect on industryemployment of implementing the CPRS andAccess Economics incorporates the findings ofthat modelling into its own forecasts for theOpen Doors scenario. The ConceptEconomics modelling forecast an 18 per centreduction in coal mining employment by 2025,a 14 per cent reduction in employment inelectricity and gas supply and a 7 per centreduction in oil and gas extractionemployment. These reductions were thenapplied, in the Access Economics modelling, tothe industry aggregates (with the otherindustries adjusted so that the totals stillmatched the model’s targets) (AccessEconomics 2009a, p. 35). Using thismethodology, Access Economics doesmanage to incorporate some of the potentialemployment effects of the CPRS into the OpenDoors scenario. The other two scenarios donot include any responses to climate change.

3.3 The global financial crisis

Both the Access Economics and CEETmodelling were undertaken in the wake of the2008 Global Financial Crisis (GFC). The latterphrase has become common terminology forthe credit crisis of September 2008, whichfollowed the collapse of Lehman Brothers inthe United States, and the subsequent crash inshare market prices around the world. Withglobal liquidity drying up, and facing the risk ofmajor recession as economic transactionsground to a halt, governments around theworld stepped in to safeguard the banking andcredit systems. In some cases, this involvedguaranteeing bad debts held by banks andother financial institutions, guaranteeing bankdeposits, and even nationalising banks.As an economic problem, the crisis had itsimmediate roots in the subprime mortgagecrisis of 2007 and the ‘credit crunch’ of2007–08, whilst its longer term genesis lay in

the deregulation of investment banking datingback to at least the 1990s and thesecuritisation of debt, which flourished duringthe 2000s. For a useful overview see Mason(2009). The broader context was a long-termdecline in US economic hegemony and therise of the Chinese economy, something Ireturn to in the next chapter.In Australia, a combination of shrewd stimulusmeasures engineered by the CommonwealthTreasury, rapid falls in interest rates initiatedby the Reserve Bank, and the unexpectedcontinuation of strong demand for rawmaterials from China, helped avoid recession.Against this backdrop the Access Economicsand CEET reports were written, in October2009 and March 2010 respectively. Thescenarios used by Access Economics werealso constructed in mid-2009. Not surprisingly,the GFC is a central element in all threescenarios, and one can regard each of themas a different response to the problems foreconomic growth thrown up by the GFC. TheOpen Doors scenario envisaged a quickrecovery from the GFC and a resumption ofeconomic globalisation, with improvements inthe governance of the international financialsystem. The Low-trust globalisation scenariowas less optimistic: a global depression wasavoided but countries moved to become moreinsular and distrustful and economic behaviourwas expected to become dominated by‘carrots and sticks’ rather than improvedsystems of governance. Finally, the Flagsscenario foresaw the world economy slidinginto a depression as a result of the GFC, anoutcome which might last as long as three tofive years (Skills Australia 2009, pp. 6 ff.).It is interesting to note that despite the SkillsAustralia depiction of the Flags scenario as aglobal depression, when Access Economicsoperationalised this scenario in their modellingof skills demand, the parameters were basedon assumptions of growth. Global growth wasexpected to average 2.6 per cent per annumand Australian domestic growth was expectedto average 2.2 per cent. Given that the Flagsscenario foresaw the global depression lastingthree to five years, and the time frame forAccess Economics is 2010 to 2025, thesegrowth rates may be realistic. This would


mean, however, considerably higher growthrates in the period 2015 to 2025 in order tocompensate for negative growth rates in theperiod 2010 to 2015. Whether this assumptionis made, and how these higher growth ratesmight occur, particularly in a climate of inward,nationalistic economic management, is notdiscussed.The CEET report discusses the GFC in anumber of places. It notes, at the outset, thatAustralia largely avoided the GFC andattributes this to the role of its Asian tradingpartners (China, Korea, Japan and India)continuing to buy raw materials (2010, p. 1).CEET also notes that the GFC had less impacton the Australian labour market because of theunexpected behaviour of employers, whoreacted to the crisis by reducing hours ratherthan retrenching workers (2010, p. 9). Anumber of economic commentators have alsonoted this behaviour, suggesting that the skillshortages of the previous few years had madethem wary of losing staff, particularly if theGFC proved to be short-lived.As far as the modelling itself, the CEET reportnotes that the MONASH model incorporatedAccess Economics five-year macro forecastswhich had taken account of the impact of theGFC, ‘though not necessarily the speed of therecovery so far’ (Shah 2010, p. 25).1 In termsof outcomes, the modelling indicated a minimalimpact on aggregate employment flowing fromthe GFC, with most of the negativeconsequences restricted to the period from2010 to 2011. In terms of sectoral effects,manufacturing and construction were expectedto suffer, whilst the occupations most affectedwere technicians, tradesworkers and labourers(Shah 2010, pp. 26–27)

3.4 Long-term labour market changes

The key labour market issues addressed inboth the Access Economics and CEET reportsinclude participation rates, impending skillslosses with the retirement of the baby boomersand the changing skills composition of theworkforce. The first two issues have beencanvassed in the earlier discussion on the

demographic time-bomb. In this section I lookat the last issue.The CEET report, in particular, discerns along-term trend in the labour market, namely afundamental shift to high-skill occupations.CEET attributes this to industrial restructuring,primarily driven by international competitionand new technology. Low skill manufacturingjobs were an early casualty of this process, but‘service and high-skill production jobs areincreasingly being exposed to competition’(Shah 2010, p. 3). Some insulatedindustries—such as aged care—are likely toavoid competition and we can expect growth inlow-skilled service jobs in sectors like these.As shown in the last chapter, the CEET reportis strongly attuned to a high-skills future,particularly for vocational skills such asDiplomas and Certificate III and IVqualifications. There is no doubt thatassumptions about this long-term trend haveshaped the modelling for the CEET report.In the case of Access Economics both theOpen Doors and Low-trust scenarios share theCEET perspective, whereby future skillsdemand is at the ‘higher end of the skillsspectrum’ (Access Economics 2009a, p. 63).By contrast, the Flags scenario envisages an‘excess supply’ of qualifications in the decadefollowing 2015, a forecast which implies alower demand for high skills in the economy.Despite this variation, the modelling forecastsof additional educational requirementsdiscussed in the last chapter (Table 2.5)showed they actually differed very little. Thedifferences between the Flags scenario andthe other two were minor when it came toforecasts for additional higher leveleducational qualifications. Rather, it was thesharp differences between the CEETprojections and all of the Access Economicsprojections which was one of the most strikingfindings in that chapter.There is little doubt that long-term occupationalchange has seen a solid expansion in theshare of jobs held by the most skilled group,namely professionals (Wooden 2000, p. 192).However, in looking at jobs growth during the1990s, (Watson et al. 2003, p. 57) argued that

1 The CEET report also noted that the MONASH occupational employment forecasts had themselves been produced in2007, which meant they did not include the strong employment growth of 2008 nor the impact of the GFC in 2009 (Shah2010, p. 25).


we need to look beyond the aggregatedoccupational categories, because the story atlower levels in the occupational scheme issomewhat different: in the 1990s ‘there wasremarkable growth in lower skilled jobs,particularly sales assistants and checkoutoperators’. Watson points to the findings ofCully (1999) and Cully (2002) who argues for a‘hollowing-out thesis’. While it’s true that theshare held by highly-skilled jobs has beenincreasing, so too has the share held bylow-skilled jobs. The jobs in decline are thosewith middle-level skills.While this thesis has been contested byWooden (2000, p. 197), that refutation waspartly based on an hours analysis ofoccupational growth, rather than a headcountmeasure. From the perspective of educationalprovision, the headcount measure is theappropriate approach. This remains acontroversial area of labour market research,but it is important to note that the assumptionof a secular trend towards a uniformlyhigher-skilled future, which is particularlyimportant for the CEET analysis, is notuncontested.


4. Neglected issues

In this section I discuss some of the issueswhich are neglected in the forecasting modelsused by CEET and Access Economics. I alsolook at issues which are mentioned by thesereports, but which don’t appear to influencetheir modelling strategy. Some of these issuesoverlap with those in the last chapter, but takea different direction, such as differentresponses to climate change. There are twobroad dimensions to this discussion: thinkingabout different scenarios and recognisingmethodological shortcomings in the modelling.To some extent these themes overlap but Ideal with each in turn.

4.1 Other scenarios

In this section I consider two of the mainomissions from the forecasting modelling:peak oil and prolonged economic malaise.Elements of both are present in the AccessEconomics and CEET modelling, but their fullimplications are never acknowledged.

Peak oil

As noted in the last chapter, both AccessEconomics and CEET discuss the challenge ofclimate change, but none of their modellingfully incorporates this recognition. The kind oftransformation of the economy, particularly inthe areas of energy and transport, required tofully meet the challenge of climate change issimply not envisaged in these reports. As Isuggest in the methodological discussionbelow, the modelling approach does notappear suited to such a task.In the case of peak oil there is noacknowledgement of its presence and thechallenges it poses. Peak oil refers to the

situation when oil extraction reaches itsmaximum and thereafter begins to decline.This is not the same as depletion or runningout of oil. Rather it refers to decliningaggregate production over time, even with newreservoirs being opened up. The moreoptimistic forecasts are for global peak oil to bereached by about 2020; the more pessimisticsuggest it may already have arrived. TheHirsch Report, written in 2005, expected peakoil to occur within 20 years(Hirsch & Bezdek & Wendling 2005, p. 5).The demand for oil at a global level is expectedto grow by 50 per cent during the period 2005to 2025. To meet such a demand, oilproduction must also expand, something whichrequires increased production from existingreservoirs and the opening of new reservoirs.Hirsch and his colleagues observe: ‘Whenworld oil production peaks, there will still belarge reserves remaining. Peaking means thatthe rate of world oil production cannotincrease; it also means that production willthereafter decrease with time’ (2005, p. 12).The expectation that the inevitable rise in theprice of oil will spur the opening of newreservoirs and provide an incentive for moreintensive extraction using new technology—ashas happened with other minerals—appearsunlikely. Geographical constraints, in terms ofhow oil reservoirs are laid down, appear toplace fundamental limits on this process:‘Higher prices and improved technology areunlikely to yield dramatically higherconventional oil production’(Hirsch & Bezdek & Wendling 2005, p. 17).1

Hirsch and his colleagues summarise theimplications of global peak oil and note: ‘The

1 It is important to distinguish conventional oil (high quality, light oil) and unconventional oil (heavy, tar-like oil) (Hirsch &Bezdek & Wendling 2005, p. 13). Most production over the last century has concentrated on the former because the easeof extraction and costs of production have been much more favourable.


peaking of world oil production could createenormous economic disruption, as onlyglimpsed during the 1973 oil embargo and the1979 Iranian oil cut-off’ (2005, p. 13).Estimates of the cost to OECD countries of the1979 interruption has been put at 3 per cent ofGDP in 1979, rising to 4.25 percent in 1981‘and accounted for much of the decline ineconomic growth and the increase in inflationand unemployment in the OECD in 1981–82’(2005, p. 29). Oil production is not only centralto energy and transport but is used extensivelyas an input into other products, particularlychemicals. The full implications of a decline inthe long-term availability of oil do not appear tobe part of any serious public policydiscussions.One reason the lessons of these earlier criseswere misread was that the restrictions in oilsupply were short-lived and oil prices stabilisedin the ensuing period, leading to theassumption that oil price rises were likely to becyclical. The prospect of long-term restrictionsin oil supply have not been fully recognised(Hirsch & Bezdek & Wendling 2005, p. 27). Inthe wake of the Global Financial Crisis (GFC)most of the bleak scenarios which arise havefocussed on financial instability and the term‘crisis’ is invariably linked to debt, rather thanto something as fundamental as peak oil. Inpractice, serious efforts have been underwayto reduce dependence on oil, but so far theyrepresent only a small beginning. In Australia,they include LPG gas and ethanol for transport(Davidson 2004, p. 1).Thus despite a general recognition that oilprices will rise, the implications of peak oilremain a blind spot for public policy. Forexample, the Skills Australia scenarios whichformed the basis of the Access Economicsmodelling did contemplate higher oil prices inthe Low-trust globalisation scenario,‘uncertainty and price variability’ in the OpenDoors scenario, and restricted access to oilresources in the Flags scenario (though this isoffset by a slowing in the demand for oil)(Skills Australia 2009, pp. 9, 19, 28). Thesescenarios acknowledged the prospects of oilprice rises, as well as major shortages ofsupply, but the full implications of peak oil werenever canvassed. Alternative energy sourceswere discussed, suggesting that the builders of

these scenarios envisaged a straightforwardtransition to a post-oil economy. In the case ofthe CEET report, there are no references to oil,not even in the general contextual discussions.

Prolonged economic decline

We saw in the last chapter that the GlobalFinancial Crisis (GFC) featured prominently inboth the Access Economics modelling and theCEET report. Indeed, the scenarios which theAccess Economics modelling addressed werelargely constructed as likely outcomes orresponses to the GFC. The GFC was seenlargely as a dampener on economic growth, apossible impediment to further globalisationand a challenge for governance of the financialsector. It was not, however, seen assymptomatic of structural weaknesses in thecapitalist system and the sign of a prolongedeconomic decline in Europe and NorthAmerica. There is a growing body of literaturewhich suggests that the period of Americaneconomic hegemony is coming to an end andthat these structural weaknesses are movingincreasingly beyond the control of nationstates. This section draws on both journalisticcommentary in the business press and thefollowing literature: Wood (2005); Stiglitz(2002); Brenner (2002); Brenner (2006);Henwood (2003); Arrighi (2007); Huang(2008).There is certainly an awareness in both theAccess Economics and CEET reports ofchanging economic fortunes on the globalstage, particularly the rise of China and Indiaand the general importance of North Asia toAustralia’s prosperity. But there is littlediscussion about Europe or North America,even though developments during 2010 haveemphasised the fragile nature of economicrecovery in both these continents.In the case of the United States, the subprimemortgage crisis may have passed, butexpectations of another property marketcollapse have been recently raised. While thebanks have returned to profitability, the baddebts off-loaded onto the public purse havecontributed so much to government debt thatfiscal measures (ie. budget spending) tostimulate the United States economy havebecome politically unacceptable to the


Democrats. At the same time, near zerointerest rates have left monetary policy withnowhere to go. Whether the United Statesenters a double-dip recession, as somecommentators are speculating, is largely adefinitional question. The main point is that theUnited States economy has entered a periodof prolonged decline. Only a series ofspeculative bubbles over the last decade (inproperty and in technology stocks), and aperiod of household consumption based onever-increasing levels of debt, have shelteredAmericans from the realisation that their timein the sun has passed.Europe’s predicament is similar. Bad loanswere also off-loaded onto the governmentbooks, already burdened with large publicdebts. Austerity measures to contain thefurther expansion of this debt were introducedduring 2010, sparking political turmoil incountries like Greece. The main dilemma, formost European countries, is that such austerityundermines the potential for anyKeynesian-style revival of economic activityand it is hard to see a revitalised private sectorleading them out of their current malaise. Forexample, a large proportion of Britain’seconomic buoyancy over the last decade wasbased on the finance sector in the City ofLondon. This has now been left considerablyweakened by the GFC. In the case of bothWall Street and the City of London, thereremains the prospect of a second round of baddebts surfacing in the near future.As in Australia, the Keynesian strategy of theChinese government appeared to haveprotected the Chinese economy from thewash-up of the GFC. With its heavy relianceon export trade with the West, China wasexpected to suffer badly when recessionunfolded in its main markets. However, amajor shift in major economic activity frommanufactured exports to domesticconsumption and to domestic infrastructureprojects appears to have allowed the Chineseeconomy to continue growing at exceptionalrates. While some commentators cast doubton the veracity of Chinese statistics, there islittle doubt that China has escaped recessionand has continued to grow its economy at arate commensurate with the growth inpopulation. Indeed, labour shortages and

significant wage increases became a feature ofthe Chinese labour market during 2010.This is not to say China’s structural problemshave been resolved. These largely centre onthe downside of being a successful exporteconomy. Domestic consumption has beenconstrained for a long period, as the strategyof using cheap labour to build export marketshas been pursued relentlessly, particularly inthe coastal regions of southern China. Whencombined with rural discontent over localgovernment corruption—particularly aroundland expropriations—this exploitation of cheaplabour has led to considerable political unrestin recent years. This is well documented butnot often discussed in the West (see, forexample, Lee 2007). The imbalances broughtabout by rural under-development and rapidurban expansion point to a major structuralweakness in the Chinese economy (see, forexample, Huang 2008).The other negative aspect to China’s exportsuccess has been its trade imbalances. Itsmassive trade surpluses have been invested inthe West, predominantly in United Statesbonds and securities. This has given China alarge stake in the fortunes of the United Stateseconomy and its currency. Moves to unseatthe US dollar as the global reserve currencyhave been pursued more forcefully by Chinaand Russia over the last few years as onestrategy to break this vulnerability. Anotherstrategy has seen China diversifying itsoverseas investments by buying stakes inresource companies, a move which also aimsto ensure future access to raw materials.Australian press coverage of China in recentyears, such as that by the Herald ’s JohnGaunaut, coalesce around two main themes.First, there are the networks of patronage andthe accompanying spread of corruption, whichappear to accompany many economictransactions involving government agencies,particularly at the local government level.These not only inflame discontent among theChinese people, but they also impose majorburdens on economic activity. Secondly, theChinese state is capable of decisive action inpursuit of its economic and social goals.Decisions to close down small and dangerouscoal mines are taken without regard to the


local employment consequences; decisions tomassively expand renewable energy are takenwithout being hampered by vested interests,such as fossil fuel industries; and decisions tobuild major rail infrastructure throughout theinterior of the country are taken without regardto the size of public debt.National strategic decision-making on thisscale has not been evident in Westerncountries since the 1960s. The ascendancy ofneo-liberalism in the West from the 1980sonwards pushed major public infrastructureinvestments off the agenda, as governmentsincreasingly recast their role as stewards ofeconomic growth rather than ‘nation-building’states (Pusey 1991). The Australiangovernment’s recent decision to embark onbuilding the National Broadband Networkmarks a sharp departure from this tradition,though the intention to privatise the systemwithin a few years shows that neo-liberalsentiments are still firmly in place.It is worth noting, in passing, that much of thediscussion in both the CEET and AccessEconomics reports emphasises the importanceof raising productivity, as well as raisingparticipation rates, yet there is only theslightest acknowledgement of the importanceof large-scale public investment ininfrastructure for achieving this goal.While many of the problems outlined aboveinvolve problems for nation states, thereremains an over-arching dilemma foreconomies which derives from the nature ofcapitalist production. This is the problem ofexcess capacity, the tendency for firms tocontinue producing long past that period whena declining rate of profit dictates that theircapital should exit from that sector ofproduction. This problem has been analysedextensively by Brenner (2002) and Brenner(2006) and is the basis for a theory of thelong-term decline in the rate of profit inWestern economies. Two decades of bubbleeconomies—based on a series of speculativebooms—has allowed this long term decline tobe masked, but its implications are now beingconfronted directly in Europe and NorthAmerica.In summary, while most current economiccommentary acknowledges the problems of

economic stagnation in Europe and NorthAmerica, the Asian story makes manycommentators more optimistic. The fact thatAustralia’s fortunes are increasingly linked withChina and India, in particular, is seen as usefulinsurance against the storm clouds hoveringover Europe and North America. However,such optimism overlooks the deeper problemsin the Chinese economy, the structuralimbalances in the global economy, and theproblems of excess capacity and declininglong term profitability.

4.2 Methodological limitations

Modelling qualitative change

As mentioned in the opening chapter, theapproaches used for forecasting introduceparameters into the models which largely varyin quantitative terms: GDP growth rates,immigration levels, population growth, labourforce participation rates, productivity growth,export intensity and capital labour ratios.There is very little scope for variables whichreflect qualitative change. This is one reasonwhy all of the Access Economics forecasts areso similar in terms of their estimates for thecomposition of employment, even though theirscenarios are quite disparate and theirestimates of total employment are quitedifferent. The two parameters which shape thecomposition of employment—as opposed to itsoverall magnitude—are the share held byexports and the capital to labour ratio in theeconomy. Differing assumptions in theseareas can result in qualitative changes in theeconomy’s industry mix, but the scale of thesechanges is modest.The most notable qualitative change in theAccess Economics report is the enhanced roleof manufacturing in the Flags scenario, anoutcome which reflects more nationalistic andprotectionist policy settings. This is achieved inthe Access Economics modelling by ‘scaling’the industry employment projections to meetthese policy targets (Access Economics2009a, p. 73).Is there scope for other policy settings toinfluence forecasts in a more fundamentalway? It appears that the input-output tableswhich transform the policy inputs into industry


projections appear to be largely taken asgiven.2 There is, in the case of AccessEconomics, an adjustment made to thesecalculations which makes use of historical datato correct for changes in worker productivity orstructural changes in the economy (AccessEconomics 2009a, p. 73). However, in terms ofpolicy options, these input-output tables arelargely fixed, and these inter-industry linkagesare not regarded as variables in the same waythat the macro parameters are.As noted above, the Access Economicsmodelling can accommodate policy responsesto climate change and their use of theemployment effects of the CPRS using theConcept Economics projections provided anexample of this. However, this was a fairlysimple quantitative adjustment to the industryaggregates which had already been producedfrom the application of the input-output tables.It did not constitute a reworking of therelationships between different industrieswhich might arise with the kind of restructuringimplied by a major response to climate change.The inter-industry linkages represented in theinput-output tables have historically evolvedand are expected to continue largelyundisturbed. Take electricity supply, forexample. As an input into the economy(expressed in value terms) it is primarilyconsumed by industry and private households,with the former consuming about two-thirds ofoutput and latter consuming one third. By wayof comparison, the consumption of gas byindustry is about 72 per cent while householdsconsume about 28 per cent. Looking at theinputs into electricity supply, these are drawnfrom a large number of sectors, but in terms ofthe raw materials needed for electricitygeneration, coal provides two and a half timesas much input as does oil and gas (ABS2005–06, year 2005–06).Now, one of the expected responses to climatechange at a policy level is a movement awayfrom coal to gas. Though the latter is still afossil fuel, it is viewed as less carbon polluting.However, one could argue that a brighterfuture for gas is still only a quantitativechange—a shift in relative shares between gas

and coal. What about a truly qualitativechange, such as a movement to a 100 per centrenewable energy sector, as envisaged by theZero Carbon Australia Stationary Energy Plan(Wright & Hearps 2010)? Such a shift wouldsee all of the current inter-industry linkagesbetween coal, gas and electricity becomelargely redundant.The implications of a large network of regionalthermal solar power stations and wind farmswould have a considerable impact on a rangeof industries, particularly manufacturing andtransport. The composition of employment—interms of occupational and skillsrequirements—arising from such a scenariowould be quite different to a future basedprimarily on extrapolating historical trendswhich assume central roles for coal or gas.The current Access Economics approachwould not appear to be capable of modellingsuch a change, since it introduces itsadjustments with the input-output tables takenas a given. Access Economics points tocurrent research by the CommonwealthTreasury which might provide a foundation for‘better understanding the likely employmenteffects of climate change mitigation’, therebyproviding them with possible data for futuremodelling (Access Economics 2009a, p. 77).This is not to say that current modelling cannotaccommodate the climate change challenge.As well as the Concept Economics modellingmentioned above, there is the modelling reportby Hatfield-Dodds & Turner & Schandl (2008).This makes use of the Monash UniversityMMRF-Green model, which is a computablegeneral equilibrium (CGE) model of theAustralian economy ‘with enhanced detail ofelectricity generation, other energy products,and greenhouse emissions accounting’ (2008,p. 4). This modelling is able to produceprojections for future labour demand indifferent industries which flow from variousscenarios for responding to climate change.While not specifically modelled, the skillsdimensions of a green-collar economy arediscussed at length.

2 The Australian Bureau of Statistics prepares input-output tables as part of the Australian national accounts. As the ABSputs it ‘They provide detailed information about the supply and use of products in the Australian economy and about thestructure of and inter–relationships between Australian industries.’


Assuming qualifications equal skills

In discussing both the context and the resultsof forecasting, both CEET and AccessEconomics assume that the concepts of skillsand qualifications are equivalent. To someextent, this is a necessary evil, since onlyqualifications are quantified in a way whichlends itself to the modelling exercise. Thereare, however, four major issues which thissimplification does not address:

1. ‘over-qualified’ workers;2. employer under-use of skills;3. literacy and numeracy issues; and4. informal and on-the-job learning.

In surveying these issues I do not propose togo into detail. There is a considerableliterature which discusses all these issues andis well known among policy makers.Recent research (Linsley 2005; Cully et al.2006; Watson 2008) has highlighted the extentto which workers have educationalqualifications in excess of the requirements oftheir jobs. One estimate is that 21 per cent ofworkers with a degree are working in jobswhich don’t require that level of education,while about 46 per cent of workers with a VETqualification are in jobs that don’t require thatlevel (Linsley 2005, p. 121).Allied to this problem is the research whichshows that under-use of skills by employers isa more widespread problem than skillshortages or skill gaps (Watson 2008;Mavromaras et al. 2007;Mavromaras & McGuinness & Fok 2009).Skills shortages and gaps are usually assumedto be major constraints on economic growthand much of the skills formation policy debatepresumes that a more qualified workforce isneeded to deal with this. Yet this recentresearch reinforces the observation that agreater utilisation of existing skills requires abroad workforce development strategy, not anarrow educational qualification strategy.The National Workforce Development Strategydevotes considerable attention to the problemsof literacy and numeracy within the Australianworkforce. In 2006 some 46 per cent ofAustralians ‘had literacy scores below the

minimum level needed to function fully in lifeand work’ (Skills Australia 2010, p. 35). I donot intend to rehearse these findings in anydetail but it is worth observing that the Strategyemphasises the impact of poor levels ofliteracy and numeracy on workforceproductivity and participation, educationalachievement and social inclusion (2010,pp. 36–37). Data on literacy and numeracy isimportant because it highlights the potential‘disconnect’ between skills and qualifications.Some 20 per cent of persons with a bachelordegree fall into the two lowest literacy levels(on the prose literacy scale), levels which areregarded as inadequate from a skills formationpoint of view (ABS 2006, Table 10).The final issue, informal and on-the-job trainingis particularly important. It is linked to theliteracy and numeracy issue in a fundamentalway; it is an important dimension of workplaceproductivity which does not show up in anyprojections of qualifications—though it ismentioned in passing—and it is fundamental tothe very concept of skill formation.To illustrate the problem it’s worth brieflyconsidering some findings from the mostrecent Australian Bureau of Statistics survey ofliterary and numeracy. Looking just at thosewith Level 1 prose literacy (the lowest level),some 17 per cent participated in an adultlearning course. The equivalent figure forthose with Level 4/5 (the highest) was 65 percent. Of those with Level 1 literacy, some 65per cent took part in informal learning. Theequivalent figure for the Level 4/5 group was97 per cent (ABS 2006, Table 11). In otherwords, reasonable literacy and numeracy is thebasis for entry into further workplace learning,and being blocked in this regard meansrelegation to a skills formation backwater.


5. The NSW dimension

Some of the forecasts produced by AccessEconomics are also available for New SouthWales, though there is little in the way ofcommentary or analysis for the state-leveldata. In the case of the CEET modelling thereare no state-level forecasts at this stage. Inthis chapter I consider the New South Walesdimension in two ways: I look at these NewSouth Wales forecasts and compare them tothe national figures to see what light thesedifferences might shed on the ways in whichNew South Wales is distinctive. Secondly, Ilook at the Access Economics ForesightingStudy which deals specifically with thesituation of the New South Wales economy in2020 and I look at an earlier study by AccessEconomics for the Independent Pricing andRegulatory Tribunal (IPART). I conclude bylooking at the Council of AustralianGovernments (COAG) targets in the NewSouth Wales State Plan to see how far thesegoals are likely to be met.

5.1 New South Wales Access Economicstables

The state results produced by the AccessEconomics modelling of skills demand (AccessEconomics 2009b) are mostly presented asgrowth rates. Consequently, it is not possibleto compare them with the results of thenational forecasts discussed earlier in Chapter2. However, when it comes to additionaleducational qualifications and studentcompletions, the state results provideestimates for levels and this makes themdirectly comparable with the data in the earlierchapter.The earlier forecast for additional educationalqualifications by Access Economics at thenational level (Table 2.5) are shown here in

Table 5.1, with the equivalent New SouthWales tables below them for comparison. (TheCEET projections are missing because there isno New South Wales data available.) The NewSouth Wales estimates are nearly all about 31per cent of the national figure, and thisproportion is almost constant across allscenarios and across all levels of qualification,ranging from a low of about 28 per cent to ahigh of 32 per cent. This is somewhat unusualand is probably the result of the uniformity ofthe modelling approach. The commentary onthe state tables notes that the methodologyused is ‘the same as at the national level’(Access Economics 2009b, p. 3). Without moredetails it is hard to know if relevant state-leveldata has been used throughout the modellingchain, or only for the key macroeconomicparameters. For example, the input-outputtables may only be available at a national level.In summary, the forecasts for educationalqualifications in New South Wales follow thenational pattern very closely. Demand forhigher education is projected to be slightlyhigher in New South Wales than at the nationallevel, particularly in the Flags scenario.Diploma level requirements and Certificate Iand II levels are almost identical, while thedemand for Certificate III and IV levelqualifications is slightly lower in New SouthWales than nationally.Access Economics also produced estimates ofstudent completions at the state level,comparable to those at the national level whichwere shown earlier in Table 2.7. Again, theNew South Wales figures constitute about 31per cent of the national figures, which is againunusual, perhaps more so. The variationacross the different scenarios and levels is notas great as with labour demand: ranging from31 to 32 per cent. One would not expect this


Table 5.1: Additional educational qualifications Australia and NSW, comparisons


’000s % ’000s % ’000s %

AustraliaHigher education 351 45.5 302 46.8 253 46.8VET sub-total 420 54.5 344 53.2 287 53.2

Diploma / Advanced Diploma 138 17.9 116 18.0 97 17.9Certificate III / IV 190 24.7 156 24.1 133 24.7Certificate I / II 91 11.8 72 11.2 57 10.6

Total 771 100.0 646 100.0 540 100.0

New South WalesHigher education 111 46.7 95 48.0 79 48.5VET sub-total 127 53.3 103 52.0 84 51.5

Diploma / Advanced Diploma 43 18.0 36 17.9 29 17.8Certificate III / IV 55 23.3 45 22.7 38 23.1Certificate I / II 29 12.1 23 11.4 17 10.6

Total 238 100.0 198 100.0 163 100.0


Source: Access Economics (2009a, Tables 8.13, 8.14, 8.15) and Access Economics (2009b, Table 4.1).

Population: Additional number of qualifications required by the workforce, Australia (top panel) and New South Wales

(bottom panel).

Table 5.2: Potential shortfall in qualifications in 2015, New South Wales, (thousands)


Supply Demand Difference Supply Demand Difference Supply Demand Difference

Higher education 64 111 -47 63 95 -32 60 79 -19VET sub-total 104 127 -23 102 103 -1 99 84 15

Diploma / Advanced Diploma 16 43 -27 15 36 -20 15 29 -14Certificate III / IV 60 55 4 59 45 14 57 38 19Certificate I / II 29 29 0 28 23 5 27 17 10

Total 168 238 -70 165 198 -33 159 163 -4

Notes: Supply refers to projected student demand (supply of students); demand refers to implied labour market demand; difference is supply less demand and

represents the shortfall.

For explanation of Access Economics scenarios see Skills Australia (2009).

Source: Access Economics (2009b, Tables 2.2, 4.1).

Population: Employed persons, New South Wales.

ratio to be so constant in all thesecomparisons. It seems reasonable to assumethat some aspects of school completion anddifferent qualification enrolment patterns woulddiffer more substantially and this would showup as different estimates of student supply.As with the national data, there is hardly anydifference in the absolute numbers of studentdemand for qualifications (supply) across allthree scenarios. As noted earlier, thisconsistency is somewhat unusual (see thecomments made earlier in Section 2.6).In terms of the difference between impliedlabour market demand and potential supply,Table 5.2 confirms the national picture. These

data point towards larger shortfalls for highereducation rather than VET. The shortfallsrange from a low of 19,000 in the Flagsscenario to a high of 47,000 in the Open Doorsscenario. Within the VET sector, the shortfallswhich occur are only with respect to Diplomasand Advanced Diplomas, and range from a lowof 14,000 (Flags) to a high of 27,000 (Opendoors). In New South Wales there are noprojected shortfalls for certificate level courses.Indeed, there is an expectation of surplusCertificate III and IV qualification holders ofnearly 20,000 in the Flags scenario and about14,000 in the Low-trust scenario.


Unlike the national reporting by AccessEconomics, there is no information for NewSouth Wales on the effect of net migration onqualification shortfalls. As we saw in Section2.6 this factor made a considerable differenceto the extent of the shortfalls, reducing theirimpact considerably. There is little doubt thattaking account of net migration for New SouthWales would see similar improvements,though there is no way of quantifying this asthere is at the national level. As we shall see inthe next section, among the various scenariosused in the Access Economics modelling, NewSouth Wales appears to come closest to theFlags scenario, so one might expect that theshortfalls likely to emerge in New South Waleswill be at the bottom end of these estimates.When we combine this with the (unquantified)reduction in the shortfall due to net migration, itseems likely that the VET sector in New SouthWales does not face any serious potentialshortfalls.

5.2 The IPART report

As part of a modelling exercise in 2006,Access Economics prepared projections for theIndependent Pricing and Regulatory Tribunal(IPART) on the future demand for vocationaleducation and training in New South Wales(Access Economics 2006). Unfortunately, mostof the data in the IPART report is presented asannual growth rates, with no substantialinformation on levels. This makes it impossibleto compare the projections in the IPART reportwith those presented above in Chapter 2.Unlike the 2009 modelling, the projections forIPART did not incorporate a number of fictionalscenarios. Instead, one set of projections wasbased on a baseline scenario, in line with theIGR 2002–03 (discussed earlier in Section3.1), and the other set was based on ascenario in which the participation rate wasraised so that the more dire predictions of thedemographic time-bomb were avoided.For the baseline scenario, Access Economicsexpects annual employment growth in theperiod from 2005 to 2025 to average 0.6 percent for New South Wales and 0.9 per cent forAustralia (Access Economics 2006, p. 20). It isworth noting that this New South Wales figureis well below the Access Economics national

figures used in the 2009 report. The morepessimistic Flags scenario which anticipatedemployment to grow at 0.89 per cent, while themore optimistic Open Doors scenario expectedgrowth of 2.15 per cent.The IPART report anticipates that the keydrivers of the employment growth in NewSouth Wales will be property and businessservices, recreational services and communityservices, while the big losers are expected tobe agriculture, utilities and manufacturing(Access Economics 2006, p. 24). AccessEconomics anticipates that the average annualgrowth in VET students will be about 0.98 percent, and the growth in VET hours will average1.08 per cent (Access Economics 2006, p. 41).For the target scenario—of higher participationrates—Access Economics expects annualemployment growth in New South Wales overthe period 2005 to 2025 to be much higher, at1.1 percent (Access Economics 2006, p. 48).Not surprisingly, VET activity is expected to bemuch greater: annual average growth instudents is put at 2.37 per cent per annum andVET hours are expected to average 2.61 percent growth (Access Economics 2006, p. 56).In terms of assessing the future of VET, theIPART report makes a useful distinctionbetween the short term and the long term:

changes in the composition of theworkforce are a clear negative forVET demand initially – more so inNSW than Australia as a whole.This reflects a relative movement injobs away from VET-intensiveoccupations and towardsoccupations which are lessVET-intensive. Most notably, this iscaused by a cyclical downturn indemand for trade skills over theshort term. That downturn hasarguably already begun in NewSouth Wales. Beyond the shortterm, such trends are a positive forVET demand, led by strongdemand for service sector workersover the longer term, such as fornurses and care staff, businesssupport workers and hospitality(Access Economics 2006, p. 45).


5.3 The Foresighting Study

The Foresighting Study produced by AccessEconomics is primarily a broad rangingdocument dealing with the New South Waleseconomy over the next 10 years. It containslittle in the way of specific forecasts forqualifications or skills but it does contain someemployment projections which can be relatedto the national-level forecasts discussed earlierin this report. For example, Access Economics(2010, Table 3.3, p. 21) predicts annualemployment growth of 1.3 per cent for NewSouth Wales in the period from 2010 to 2020.The comparable national figures (though forthe period 2010 to 2025) are 2.12 per cent,1.49 per cent and 0.96 per cent for the OpenDoors, Low-trust and Flags scenariosrespectively. In the case of productivity growth,the NSW projection is also for 1.3 per centannual growth. By contrast, the nationalprojections are for 1.75 per cent, 1.5 per centand 1.3 per cent for the three scenarios.In summary, Access Economics expects NewSouth Wales to be somewhere between theLow-trust and Flags scenarios when it comesto employment growth—though closer to thelatter—and squarely within the Flags scenariowhen it comes to productivity growth.If we compare the Foresighting Study figureswith the IPART projections, produced nearlyfive years earlier, it’s clear that they are muchmore optimistic. IPART put annualemployment growth at just 0.6 per cent for itsbaseline scenario compared with the figurehere of 1.3 per cent. The IPART figure for the2020 projections (to make them morecomparable with the Foresighting Study ) arejust 0.5 per cent.As we have just seen, the IPART reportidentified the key drivers for employmentgrowth to be located in property and businessservices, recreational services and communityservices. The Foresighting Study endorsesthis assessment but places mining at the top ofthe list.1 Where the IPART study had miningemployment in decline (-1.4 per cent), the

Foresighting Study expects it to boom (3.7 percent). Similarly, but at a more modest level,the IPART study expected agriculture to be indecline (-2.2 per cent), but the ForesightingStudy expects a positive story (1.1 per cent).Both reports expect a bleak outcome formanufacturing, with the Foresighting Studyeven more pessimistic (-2.8 per cent comparedto -1.4)The Foresighting Study covers many of theissues touched on in earlier sections of thisreport. It takes climate change seriously andalso looks closely at the situation in China andIndia. It also deals more thoroughly than doCEET or the other Access Economics reportswith the likely future role of innovation anddevelopments in the information economy.However, despite this openness, theForesighting Study also repeats somecommonly held views concerning futureenergy security. While admitting that carboncapture and storage is an unproventechnology, it does not discuss its viability andsimply recommends more research anddevelopment in the area (Access Economics2010, p. 88). On the other hand, the feasibilityof solar thermal for base-load electricity—asemphasised by Wright & Hearps (2010)—isignored in the repetition of the view thatrenewable energy cannot provide forbase-load demand.2 In both respects, theeconomic importance of the coal industry inNew South Wales is evident.

5.4 The New South Wales State Plan

In its discussion of the ‘Clever State’ the NewSouth Wales State Plan refers to the COAGVET targets (NSW 2010, p. 25). These include:

1. By 2015, 90% of 20–24 year olds inNSW will have achieved Year 12 or aCertificate II qualification or above;

2. By 2020, 90% of 20–24 year olds inNSW will have achieved Year 12 or aCertificate III qualification or above;

1 The Foresighting Study uses a more recent industry classification scheme to that used by IPART, which makes exactcomparisons difficult. For the comparisons considered here, the 2020 projections from Table 3 of the IPART report arecompared with the projections in Table 3.3 of the Foresighting Study.

2 ‘Apart from wind generation, most of the activity on renewable energy for electricity is likely to be of relatively small scaleand not useful for baseload generation.’ (Access Economics 2010, p. 87).


3. A 50% drop in 20–64 year olds withoutCertificate III level or above qualificationsbetween 2009 and 2020;

4. A 100% increase in people achievingDiploma and Advanced Diplomaqualifications between 2009 and 2020.

Unfortunately, the state results for the AccessEconomics modelling do not contain anyprojections which would allow us to assess thelikely progress towards some of these targets.The national results, on the other hand, docontain useful material but it is not availablewith age group breakdowns. The AccessEconomics report does, however, contain anexplicit discussion of whether some of thesetargets will be met.The CEET forecasts suggest that in 2010about 39 per cent of employed persons wouldbe in the category of ‘no post-schoolqualifications’ (Shah 2010, Table 24). By 2020this figure was forecast to fall to 29 per cent.Access Economics does not provide estimatesfor 2010, but its 2015 estimates put thecomparable figures at 32 per cent (OpenDoors), 33 per cent (Low-trust) and 34 per cent(Flags). In 2020 Access Economics expectsthese three figures to have dropped to 28 percent, 29 per cent and 31 per cent respectively.Clearly there is a close alignment betweenthese projections, particularly for the Low-trustscenario. The CEET projections, in particular,show a substantial improvement in increasingthe qualifications profile of the workforce in theorder of 10 percentage points. How thistranslates into targets for 20–24 year olds inNew South Wales (targets 1 and 2) isimpossible to say.With respect to targets 3 and 4, AccessEconomics summarises its findings and I drawdirectly on these here. With regard to target 3,taking account of differing age definitions ofthe population, Access Economics notes thatreductions of the order of 31 per cent can beexpected by 2020 (that is, from 48.4 per centto 33.6 per cent). This takes place under theOpen Doors scenario. As Access Economicsobserves, this is a ‘substantial reduction’, butnot quite halving the proportion. To meet theCOAG target, Access Economics notes that anadditional 1.3 million people would need to

obtain Certificate III or above qualifications by2020.Access Economics further notes that thesupply shortfalls (discussed earlier in Section2.6) imply that there would be insufficientstudent completions to meet this target ‘so afurther condition to meet the target would be asignificant contribution to qualificationsresulting from net migration and/or additionalresources to encourage higher studentparticipation’ (Access Economics 2009a,p. 67).As for target 4, the 100 per cent increase inpersons holding Diplomas and AdvancedDiplomas, Access Economics expects that thistarget will not be met under any of thescenarios (Access Economics 2009a, p. 68).This is based on the student completionprojections. Access Economics does note,however, that if one looks at the implied labourmarket demand, there is scope for thesetargets to be met. There are two caveats to thisexpectation: significant skills deepening overtime and multiple qualification holding wherebythose with higher qualifications also holdDiplomas or Advanced Diplomas. Whether thislast caveat is consistent with the spirit of theCOAG targets is perhaps a moot point.In summary, while there are no data in eitherthe CEET or Access Economics reports bywhich to judge how well New South Wales isfaring in meeting these targets, it seemsreasonable to assume that the national picturealso applies at the state level. If anything,NSW may face additional challenges given theForesighting Study predictions of employmentgrowth to 2020.


6. Other relevant publications

6.1 UK Commission for Employmentand Skills

Targets for 2020

The UK Commission for Employment andSkills (UKCES) outlines two kinds of targets forskills formation in the UK (UKCES 2010). Oneset of targets concerns the percentage ofadults with certain levels of qualifications to beachieved by 2020. The other target is a goal ofcertain level of OECD ranking (based on thefirst set of targets). The latter is an elusivegoal, in so far as the first set of targets mightbe achieved, but improvements by othercountries might see the UK fail to gain thedesired rankings. In other words, achievingdomestic goals in skills formation might nottranslate into adequate league table outcomes.The first set of UKCES targets are for:

• 95 per cent of adult to have functionalliteracy and numeracy (termed ‘Basicskills’);

• 90 per cent of adults being qualified to atleast level 2 (termed ‘Low level skills’);

• 68 per cent of adults being qualified to atleast level 3 (termed ‘Intermediate levelskills’)

• 40 per cent of adults being qualified to atleast level 4 (termed ‘High level skills’)

Because higher level skills imply lower levelskills, these targets translate into the followingfigures at each level for 2020:

• Level 4+ (high): 40 per cent;• Level 3 (intermediate): 28 per cent;• Level 2 (low): 22 percent;• Below Level 2: 6 per cent;• No qualifications: 4 per cent.

The projections in 2010 suggested that thehigh level target would be exceeded (42 percent), the intermediate would be well belowtarget (19 per cent); the low level would beslightly below target (20 per cent). As for thebottom levels, the 10 per cent target would bebadly missed, with a forecast of 19 per cent.In terms of ranking, the UKCES aimed to be inthe top 8 of the international rankings for thethree main levels of skills. Its actual ranking in2007 was 19th for low skills, 21st forintermediate skills and 12th for high level skills.It expected that these rankings would be 20th,21st and 11th respectively by 2020. In otherwords, the goals would not be achieved, andimprovement was only expected for high levelskills. For both basic and intermediate skillsUKCES expected the UK to remain the sameor fall back one place (UKCES 2010, p. 5).Clearly, whether using rankings or percentageoutcomes, the forecasts outlined by UKCES in2010 suggest that the UK is considerablybehind at current rates of progress on where ithopes to be by 2020.

The UKCES methodology

The UKCES forecasting model is on basedprojecting existing trends into the future, withsome account taken of population changes(including migration) and age-cohort effects.For example, as people progress through theirworking lives their qualifications change andthe modelling takes this into account. By wayof contrast, the Australian modelling reviewedearlier begins with assumptions concerningeconomic growth, derives industry projectionsfrom these, then moves onto occupationalprojections, before arriving finally atqualification forecasts. The upshot of this is


that UKCES takes the current rate of progressand examines where the UK will be in 2020.The UKCES Technical Report also notes,however, an alternative approach to skillsforecasting. The Department for Business,Innovation and Skills (BIS) use a methodwhich combines planned public funding ofskills formation with expected population flows(including migration) and ‘private upskilling’ toarrive at a predicted level of skills at a futuredate (in this case 2020). In other words,‘assuming …planned levels of investment areimplemented’ then the skills position in 2002will look like this. In this respect, the BISapproach captures how much investment isneeded to arrive at future targets(Bosworth & Kik 2009, p. 12).The UKCES Technical Report suggests thatthe two approaches are complementary andthat the forecasts should converge as 2020approaches, assuming of course that theplanned investments actually take place.The UKCES approach is both more expansiveand more limited than Australian approaches.It is more limited in that it relies entirely onprojecting future trends, without attempting tomodel industry and occupational change. Inaddition, it does not incorporate differentgrowth scenarios into its forecasts. On theother hand, the UKCES approach is locatedwithin a conceptual framework which is moreexpansive because it considers issues such asinequality and social justice, both of which areoff the radar in the Australian studies. Both theUKCES and the Australian studies emphasiseproductivity and participation as key drivers ofeconomic prosperity, but the UKCES studiesappear more aware of the social dimensions ofeconomic growth. This does not influence theactual modelling methodology—which is reallyquite simply an extrapolation technique—but itdoes influence the policy framework in whichthe whole exercise takes place.

Unresolved issues

Like the Australian studies the UKCES teamunproblematically equate qualifications withskills when it comes to quantifying theirprojected outcomes. They are, however,aware that qualifications are an inadequate

measure of skills but argue along pragmaticlines:

Traditionally however skills havebeen measured by qualifications,despite their shortcomings as aproxy for skills; qualifications allowrelatively simple andstraightforward comparisons overtime, between sub-groups, andinternationally (Bosworth & Kik2009, p. 15).


7. Conclusion: planning for the medium term

7.1 Shortcomings in modellingqualifications

The earlier chapters have highlighted anumber of shortcomings in modellingeducational qualifications in Australia. Despitereasonably close alignment in projections forindustry employment, different economicmodels produce quite different policy-relevantoutcomes. At each link in the chain—frommacro-settings to industry employment tooccupational employment to educationalqualifications—the divergence betweendifferent models increases. By the time onearrives at the key outcomes from a policy pointof view, one confronts a number of perplexingchoices:

1. is the likely demand for additionaleducational qualifications at the nationallevel likely to number about 480,000, oris it more likely to be as high as 770,000(see Table 2.5)?

2. is there going to be a national shortfall inVET qualifications of over 50,000, or asurplus of over 40,000 (see Table 2.8)?

3. in the case of New South Wales, is theregoing to be a shortfall in VETqualifications of over 20,000, or a surplusof about 15,000 (see Table 5.2)?

4. in more general terms, is the labourmarket in coming years likely to becharacterised as largely high-skilled, or isthere a more polarised skills futureawaiting us?

5. in other words, is a large expansion inthe provision of educational qualificationsrequired, or are current arrangements byand large adequate?

It is clear that all of the models considered inthis report make use of detailed empiricalmaterial, drawn from the Census and othersurvey data, and their macro-models appear tobe robust and comprehensive. However, whatappears to happen is that key assumptions aremade at each link in the modelling chain andthis leads to divergence. In particular, differentassumptions about skills deepening andmultiple qualification holding result in importantdivergences between the CEET modelling andthe Access Economics modelling. In thissense, despite the richness of the dataemployed, outcomes are basicallyassumption-driven.Even the uniformity in the macro-settings is tosome extent premised on shared assumptions.While there may be sharp differences inexpectations about economic growth, andeven differences around compositionalchange—such as the role ofmanufacturing—the overall framework used byall the models does not depart fromconventional assumptions. As noted inChapter 4, this means that some key blindspots are ignored, particularly around criticalenvironmental and economic issues. Again,this suggests that even the coremacro-modelling is to some extentassumption-driven. A truly low-carbon future,for example, is just not part of any modellingframework.So what are policy makers to do when facedwith these dilemmas? Should it come down tosiding with those projections which accord withone’s own assumptions, or is there a morebalanced strategy available? The option ofvacating the terrain of planning is not anoption; it simply leads to reactive policyformulation. Even if state governments were toabdicate any role in managing their local


economies, leaving the private sector to sortout its own future, they would still haveunavoidable obligations in the labour market.As the primary employers of the workforce inthe health, community and educationalsectors, not to mention police and publictransport, the state governments in Australiamust undertake some form of workforceplanning if they are to fulfil their basicfunctions. In the case of educational provision,state governments are directly implicated inthe provision of future qualifications, at least atthe secondary and VET levels.As a minimum then, some kind of frameworkfor medium-term planning seems unavoidable,even for policy-makers who may be completelyjaded with the confusing perplexity ofeconomic projections. What options areavailable that avoid some of this confusion, butwhich go beyond mere short-term reactivepolicy-making?

7.2 Extrapolating trends

One of the most reliable approaches forplanning in the medium term—say from four tofive years out—is to simply extrapolate fromcurrent trends. Most labour market trends areglacial. Only a sharp economic contraction ora major shift in policy settings disrupts thesetrends. Examples of these include the majortariff cuts of the 1970s and the recessions ofthe early 1980s and 1990s. All had majorramifications for the composition ofemployment in Australia, particularly formanufacturing, and the time-frames involvedwere quite short-term. But many of the othertrends which characterise the labourmarket—under-employment, casualisation, lowparticipation rates, work intensification, limitedworkplace training—have evolved slowly overthe last thirty years. Similarly, both industryand occupational restructuring have evolvedslowly, and have been largely predictable.What about skills shortages? Surely, it can beargued that the resources boom and thedemands it has placed on skilled labour havelargely come out of the blue? It’s important torealise that mining and construction areindustries which are very prone tostop-and-start conditions. With the onset of theGFC, the resources boom in Western Australia

contracted sharply, only to restart within ashort period once China’s own recovery wasguaranteed. The imminent collapse ofconstruction jobs at the start of the GFC waslargely averted by Federal Governmentstimulus measures. Both examples highlightthe fragility of the boom and bustcharacteristics of these sectors.What is the relevance of this for the widereconomy? Writing in 2008, Watson arguedthat the ‘skills crisis’ was over-rated, and thatWestern Australian shortages did notadequately reflect the national picture. TheDEWR skilled vacancy index at the nationallevel was largely flat from 1994 through to2005, with a small rise around 1999–2000which was followed by a steady decline(Watson 2008, p. 7). Employer associationshave, of course, emphasised skills shortages,but this needs to be seen in the light of twoimportant issues. The first is an employerpreference for importing skilled labour, ratherthan investing in domestic training, and ahighly vocal lobbying presence follows fromthis. Secondly, confusion between recruitmentdifficulties and actual skills shortages iswidespread, leading many employers tomistake the former for the latter. The overallconclusion of Watson (2008) was that therewere undoubtedly major skills shortages insome sectors of the economy, but the overallskills crisis was exaggerated.Returning to the theme of extrapolating trends,it seems clear that one cannot allow anomalieslike the resources boom or the constructionindustry to dictate an overall methodology forthe labour market as a whole. What is requiredfor the latter is simply good data. In thisrespect, policy makers are well placed, withone major exception. On the positive side:

1. the five-yearly Australian Bureau ofStatistics Census of Population andHousing provides very rich data onindustry, occupation, education and age.Moreover, it is available in great detail atthe regional level. The information onage, in particular, allows for usefulretirement planning. The recentintroduction of the TableBuilder productmeans that researchers can nowproduce their own customised Census


tables, including detailed regional tablesand data items at a very disaggregatedlevel (see ABS 2009a).

2. the surveys of student outcomes,conducted by the National Centre forVocational and Educational Research(NCVER), provides rich detail on whichqualifications are in high demand byindustry (see, for example, NCVER2008).

3. the Household Income and LabourDynamics in Australia (HILDA) surveyprovides the best longitudinal data ondetailed labour market outcomes overtime. While its sample size is adequate atthe national level, it is somewhat limitedat the state level. (See http://www.melbourneinstitute.com/hilda/.)

4. public sector employee databases, suchas the NSW Public Sector WorkforceProfile, provide planners with extensiveand detailed information on all publicsector occupations. In the case of thosekey groups likely to be affected by theexit of the baby boomers, such asteachers and nurses, the data exists forpredicting these outcomes quiteprecisely.

On the negative side, no adequate large-scaleworkplace surveys have been conducted inAustralia to rival the Australian WorkplaceIndustrial Relations Survey (AWIRS)endeavours of the 1990s. Such surveysprovide core information on the behaviour ofemployers—how they make recruitmentdecisions, how they engage their workforceand then deploy them, how they utilise theirworkers’ skills and what training arrangementsthey put in place. While some insights intothese are available from household surveys,like HILDA, the full picture only emerges whenthe unit of analysis is the workplace and therespondents include managers.All of these data sources can be utilised todiscern medium-term trends and likelydevelopments during the next five years. Insome cases, these will apply at the nationallevel, but in many cases both state-level andregional-level trends will be apparent.

7.3 Key informants

While a quantitative analysis of the labourmarket is fundamental, much is to be gainedfrom qualitative research as well. However,rather than interview people as‘subjects’—where their experiences are theobject of enquiry—it is preferable to interviewpeople as ‘key informants’.The problem with a focus on subjects isgeneralisability. While the data gathered withthis approach may be much richer than withstructured questionnaires (as in HILDA, forexample), sample selection is highlyproblematic and collection costs for anadequate sample size are enormous. Bycontrast, if key informants are the focus, thesample does not need to be large, and sampleselection is based on deliberation rather thanprobability. This means choosing people whoare positioned at strategic locations in thelabour market and whose insights arelong-term and far-reaching.For example, interviewees might include:members of employer associations and unionswho have been around a long time and havesound knowledge of developments occurringin their sectors; senior teachers andadministrators in the training and educationfield who have observed changing patterns inenrolments and outcomes over the last decadeand can discern emerging trends; senior staffin the recruitment industry who possess agood overview of a number of industry sectors.In-depth interviews with people such as thesecan provide rich material on labour marketdevelopments. While it may not becomprehensive, when augmented with thequantitative material, the overall picture shouldbe adequate for medium-term planning.

7.4 Context and regions

The importance of both thesesources—qualitative and quantitative—is thatthe research outcomes are much richer than isthe case with economic forecasting. The lattersimply produces bald numbers and presentsthe policy maker with the quandary discussedearlier: is this number realistic, or is it too highor too low? The results are largelydecontextualised and only the assumptions


are open for discussion. The overall planningprocess is largely a black-box affair.By contrast, the strategy being discussed hereprovides information which is heavilycontextualised and the process is transparent.The research outcomes contain numbers, ofcourse, but they are provided in the context oftrends, patterns and processes. While some ofthese may be matters for judgment—as withthe assumptions in the forecastingapproach—all of this is open for discussion. Ifsome of the trends appear unrealistic, then thiscan be openly discussed: the material is thereto be scrutinised.Finally, regional planning can become animportant element in this process. Witheconomic forecasting the inputs for themacro-models are overwhelming national-leveldata, or at best, state-level. Even when thepolicy focus is explicitly regional, theunderlying assumptions are largely aboutdevelopments at a national or state level. Insome cases, the local ‘share’—some fixedproportion—is simply applied to aggregateprojections.By contrast, more useful outcomes are likelywith a medium-term planning approach whichmakes uses of regional-level data—such asthe Census or the Workforce Profile—inconjunction with key informants who are basedin the region. This means that the actualresearch outcomes are explicitly regional innature and thus consistent with the policyfocus. The analysis of the Hunter regioncarried out by researchers at the Centre of FullEmployment and Equity at the University ofNewcastle (see, for example Cook et al. 2008)illustrates the useful possibilities which aregional research focus can bring to policyformulation.


References

ABS 2005–06, Australian National Accounts:Input-Output Tables - ElectronicPublication, Cat. 5209.0.55.001, AustralianBureau of Statistics, Canberra.

— 2006, Adult Literacy and Life Skills Survey,Australia 2006 (Reissue), Cat.42280DO001, Australian Bureau ofStatistics, Canberra.

— 2009 a, 2006 Census of Population andHousing - TableBuilder, Cat. 2065.0,Australian Bureau of Statistics, Canberra.

— 2009 b, Population by Age and Sex,Australian States and Territories, Cat.320109, Australian Bureau of Statistics,Canberra.

Access Economics 2006, Future Demand forVocational Education and Training in NewSouth Wales, Report for IndependentPricing and Regulatory Tribunal of NSW(IPART), Access Economics Pty Ltd.

— 2009 b, Economic modelling of skillsdemand - State results, Report for SkillsAustralia, Access Economics Pty Ltd.

— 2009 a, Economic Modelling of SkillsDemand, Report for Skills Australia,Access Economics Pty Ltd.

— 2010, The NSW economy in 2020: Aforesighting study, Report for NSWInnovation Council, Access Economics PtyLtd.

Arrighi, G 2007, Adam Smith in Beijing:Lineages of the Twenty-first Centry, Verso,London.

Australian Academy of Science 2010, TheScience of Climate Change: Questions andAnswers, Australian Academy of Science,Canberra.

Bosworth, D & Kik, G 2009, Ambition 2020:Technical Report, UKCES: UKCommission for Employment and Skills.

Brenner, R 2002, The Boom and the Bubble,Verso, London.

— 2006, The Economics of GlobalTurbulence: The Advanced CapitalistEconomies from Long Boom to LongDownturn, 1945–2005, Verso, London.

Commonwealth Treasury 2002,Intergenerational Report 2002–03,Commonwealth of Australia.

— 2007, Intergenerational Report 2007,Commonwealth of Australia.

— 2010, Australia to 2050: future challenges,Commonwealth of Australia.

Cook, B & Mitchell, W & Quirk, V & Watts, M2008, Creating effective local labourmarkets: a new framework for regionalemployment policy, Centre of FullEmploymentloyment and Equity, Universityof Newcastle, Australia.

Cully, M 1999, ‘A More or Less SkilledWorkforce? Changes in the OccupationalComposition of Employment, 1993 to1999’, Australian Bulletin of Labour,vol.25, no.2, pp.98–104.

— 2002, ‘The Cleaner, The Waiter, theComputer Operator: Job Change,1986–2001’, Australian Bulletin of Labour,vol.28, no.3, pp.141–162.

Cully, M & Delaney, M & Ong, K & Stanwick, J2006, Matching skill development toemployment opportunities in New SouthWales, Independent Pricing andRegulatory Tribunal of NSW.

Davidson, S 2004, ‘Ethanol: what’s it allabout’, Ecos, no.119, pp.14–18.

Garnaut, R 2008, Climate Change ReviewFinal Report, Cambridge University Press,Port Melbourne.

Hatfield-Dodds, S & Turner, G & Schandl, H2008, Growing the Green Collar Economy:Skills and labour challenges in reducing ourgreenhouse emissions and nationalenvironmental footprint, Report to theDusseldorp Skills Forum June, CSIROSustainable Ecosystems, Canberra.


Henwood, D 2003, After the New Economy,Scribe Publications, Melbourne.

Hirsch, RL & Bezdek, R & Wendling, R 2005,Peaking of World Oil Production: Impacts,Mitigation, & Risk Management, “TheHirsch Report”, A report commissioned bythe US Department of Energy.

Huang, Y 2008, Capitalism with ChineseCharacteristics: Entrepreneurship and theState, Cambridge University Press,Cambridge.

Lee, CK 2007, Against the Law: LaborProtests in China’s Rustbelt and Sunbelt,University of California Press, Berkeley.

Linsley, I 2005, ‘Causes of Overeducation inthe Australian Labour Market’, AustralianJournal of Labour Economics, vol.8, no.2,pp.121–143.

Mason, P 2009, Meltdown: The End of the Ageof Greed, Verso, London.

Mavromaras, KG & McGuinness, S & O’Leary,NC & Sloane, PJ & Fok, YK 2007, TheProblem of Overskilling in Australia andBritain, Discussion Paper No. 3136, IZA.

Mavromaras, K & McGuinness, S & Fok, YK2009, ‘Assessing the Incidence and WageEffects of Overskilling in the AustralianLabour Market’, The Economic Record,vol.85, no.268, pp.60–72.

NCVER 2008, Confidentialised unit recorddata from Student Outcomes 2008,National Centre for Vocational EducationResearch, Adelaide, SA.

NSW 2010, NSW State Plan: Investing in aBetter Future, NSW Government.

Pusey, M 1991, Economic Rationalism inCanberra: A Nation Building StateChanges it Mind, Cambridge UniversityPress, Cambridge.

Shah, C 2010, Demand for qualifications andthe future labour market in Australia,Report for Department of Education,Employment and Workplace Relations,Centre for the Economics of Education andTraining (CEET), Monash University.

Skills Australia 2009, Australia 2025: ThreeGlobal Scenarios, A Framework forAnalysing Australia’s Possible FutureWorkplaces and Workforce DevelopmentRequirements, Canberra.

— 2010, Australian Workforce Futures: ANational Workforce Development Strategy,Canberra.

Stern, N 2006, Stern Review on theEconomics of Climate Change, HMTreasury (UK), London.

Stiglitz, J 2002, Globalization and itsdiscontents, Penguin Books, London.

UKCES 2010, Annual Report 2009–2010, UKCommission for Employment and Skills.

Watson, I 2008, Skills in use: labour marketand workplace trends in skills use inAustralia, Report for NSW Department ofEducation and Training (Skills EcosystemProject).

Watson, I & Buchanan, J & Campbell,I & Briggs, C 2003, Fragmented Futures:New Challenges in Working Life,Federation Press, Sydney.

Wood, EM 2005, Empire of Capital, Verso,London.

Wooden, M 2000, ‘The Changing SkillComposition of Labour Demand’,Australian Bulletin of Labour, vol.26, no.3,pp.191–198.

Wright, M & Hearps, P 2010, Zero CarbonAustralia Stationary Energy Plan, BeyondZero Emissions and Energy ResearchInstitute, University of Melbourne.


For further information see www.bvet.nsw.gov.au

GPO Box 33 Sydney NSW 2001 Australia T (02) 9561 1500 ISBN 978‐1‐921084‐38‐6 February 2011 NSW Department of Education & Training

The report was produced as a research project funded by the NSW Board of Vocational Education and Training. The views and opinions expressed in this document are those of the author and do not necessarily reflect the views of the New South Wales Government.

modelling of future skills demand: the implications for...

Documents