The Australian experience with employment service purchasing suggests that a hybrid model would work best. This would include a public provider and online service for those requiring limited employment assistance.
The 2010s: A system under pressure
By the late 2000s, governments observed that the Job Network offered little more than low-level job search and compliance and concluded they could reduce its cost with little or no reduction in effectiveness. Funding per unemployed person was cut in successive tenders. By 2018, Australia was spending less than half the average of OECD nations (as a share of GDP) on employment services.
Caught between shrinking public funding, tighter contract management, and a Star Rating system that required them to achieve better results than their competitors, some providers (who in the past had made easy profits from the Job Network) gamed the system, for example by diverting funds to associated businesses or making false claims for outcome payments. This in turn led to tighter controls and gnawed away at public confidence in the program.
One of the major weaknesses of employment services at this time was that the supply-side focus on supervised job search meant that too little attention was paid to relationships with employers.
As unemployment declined and ‘Welfare to Work’ policies directed increasing numbers of people with disabilities, sole parents, and older women on to unemployment payments, the profile of people in the Job Network system became more disadvantaged, yet it was under-powered to help them.
The Job Network placed many people who were already close to employment in part-time and casual jobs in booming sectors such as hospitality and retail. However, its overall share of job vacancies declined as employers associated it with applicants they considered ‘unsuitable’. Instead of being adequately screened for and connected to job vacancies (and offered the training and other support required) too many people with labour market disadvantage were simply sent out into the labour market to apply for the required number of jobs every month.
From Rudd to Abbott: variations on a common tune
Realising that prolonged unemployment had become entrenched, the Rudd and Gillard governments attempted to shift the focus from compliance to assistance for those most disadvantaged in the labour market.
‘The Job Network has failed disadvantaged job seekers who have needed assistance to overcome barriers to employment and to gain the skills that employers need. This is starkly illustrated by the proportion of people on unemployment benefits for more than five years which has increased from one in 10 in 1999, to almost one in four today.’ (O’Connor B (2008), Reform of Employment Services in Australia, Speech to House of Representatives, 2 September 2008).
People excluded from the Job Network due to their deeper labour market disadvantage – including many people with mental illness – were brought into the replacement program (‘Job Services Australia’ or JSA) as ‘Stream 4’ participants attracting the highest level of funding. However overall funding for mainstream employment services was not increased, so the resources devoted to Stream 4 came at the expense of others unemployed long-term.
The government experimented with alternatives to the Job Network model on a smaller scale. Consistent with its emphasis on ‘social exclusion’, pilot schemes were run in regions with high unemployment to ‘join together’ employment and other services for people with ‘complex needs’.
Unlike JSA, which mostly retained the performance management regime of the Job Network, these new schemes were traditional grants-based programs where providers specialised in assisting groups assumed to face major barriers to employment including young single parents and early school-leavers. Providers were encouraged to work collaboratively with JSA providers and other local community agencies to offer ‘wrap-around support’ to people believed to face multiple barriers to employment.
The new schemes were too small-scale and temporary to make a major difference. They were stopped by the Abbott government, along with an evaluation which might otherwise have drawn lessons for future program design.
The JSA program was large enough to make a difference, but it was a poor fit with the new ‘place-based’ approach. Providers were required (in theory) to forge local partnerships with employers and other community services but tight performance management left little room for this in practice. Local employers faced competing providers who found it difficult to cooperate locally.
Notwithstanding the higher ‘Stream 4’ funding, the incentives in the purchasing model reinforced a laser-like focus on quick employment outcomes rather than patient investment and collaboration. The Employment Pathway Fund remained underspent. JSA was an incremental reform that had a limited impact on prolonged unemployment.
Towards the end of the Gillard administration and the JSA contract, there was an attempt to revive the ‘black box’ approach to loosen central controls over services and encourage innovation. To that end, the Abbott government’s jobactive program directed 70% of funding to outcome payments. However, the compliance system for unemployed people was ramped up via a major expansion of Work for the Dole and central controls over employment services remained tight.
The jobactive contract saw the ‘market’ consolidate, from around 100 providers in JSA to 44, and for-profit providers now service the majority of participants. A major contributing factor was the shift under jobactive from contracting to service smaller Employment Service Areas towards servicing the much larger employment regions. Many long-standing local community-based providers left the system. This will make it harder to implement place-based initiatives in future.
As budget savings were progressively squeezed from the system, service quality declined along with the share of vacancies filled, and governments became more reluctant to invest. By 2018, consultant caseloads in jobactive averaged 140 (around 6-10 interviews a day) and staff turnover was high.
In each of its iterations, the Job Network model adapted to new challenges and demands, but a model that was once a source of pride for Australian governments, and was feted by international bodies like the OECD, seemed to be locked in a downward spiral.
By 2019, there were more than 600,000 people on Jobseeker and Youth Allowance payments for more than a year, almost twice as many as the number that inspired the Working Nation package 25 years before.
2018: Employment Services Expert Panel
In 2018, the Turnbull government initiated a major review of employment services led by an Expert Panel (of which I was a member). This was the first major public review of the system since the Productivity Commission inquiry 20 years earlier. The review consulted widely and commissioned research on the experience of jobactive by unemployed people, employers and providers.
Its key findings of the Panel’s report ‘I want to work’ included the following:
- The system was too compliance-focused and unemployed people should instead be supported to exercise more choice and control in their search for employment.
- The system was transactional and administrative-heavy, and should be redesigned to reward quality and collaboration.
- People who were closer to employment benefited less from employment services and many would be better served by an online service, with a back-up support as needed.
- Budget savings from online services should be redirected to ‘enhanced services’ for people unemployed long-term or at risk, to reduce caseloads and improve service quality.
- Improved outreach to employers that responds to their preferred methods of recruitment.
- Encourage and support local collaboration with employers and other service providers.
- A licensing system to ensure service quality and provide a flexible and transparent point of entry into the system for providers.
- More funding upfront to facilitate entry to the system for new providers and investment in quality services.
If this history of performance-based contracting of employment services seems overly pessimistic, that is because I’m conscious of the potential of employment services to improve the lives of people worst affected by unemployment. A major part of the problem has been the lack of fiscal commitment from governments lulled into complacency about prolonged unemployment by 30 years of non-recessionary growth. That may change now.
Large-scale job search assistance backed by the latest technology can improve the speed and effectiveness of job matches, benefiting employers and unemployed people and reducing the economic waste associated with short-term unemployment.
More importantly, robust evaluations find that substantial investments in paid work experience, training and employer engagement (‘demand-led approaches’) can make a real difference to the employment prospects of people who have been out of paid work for a long time. This has profound social and economic consequences since prolonged unemployment is a major cause of poverty, declining health and social breakdown, and a major hurdle to achieving full employment.
Of course, employment programs will only work if employment is growing strongly and consistently, but growth of itself is no guarantee of full employment. The distribution of job matches also matters.
Over a 30-year period in which Australia succeeded in reducing unemployment but failed to curb long-term unemployment, the main challenge for employment services has been the same: to devise a system that connects individuals at greatest risk of prolonged unemployment to the package of supports that each person – and their prospective employer – needs. This is no easy task. It means strengthening the skills, capabilities and resilience of people left behind in the labour market and getting employers to consider applicants that would normally set aside.
At its heart the solution is simple: employment service consultants must have skills and resources at their disposable (wage subsidies, training and connections with local employers and community services) to guide each unemployed person towards the most suitable job available. Their service must have the right ‘intrinsic motivation’ and financial incentives to use those resources cost-effectively.
Large-scale national programs such as wage subsidies and training are needed, but they won’t be effective unless placed in the hand of skilled, well-connected labour market intermediaries.
Our employment services system operates on the assumption that performance-based contracting is the best way to combine national oversight and cost control with local flexibility and innovation. We have seen how this assumption has been called into question. While the cost of employment services has steadily declined, so has the quality and personalisation of services on the ground to both unemployed people and employers. The system shifts risk to providers so they can innovate and invest, but services have converged towards low-level job search assistance. Successive governments have taken greater control of service inputs as they found that the financial incentives didn’t work as expected.
It is telling that when governments wanted to invest in more substantial help for unemployed people, they did so through old-fashioned grants programs where providers are paid according to service inputs (such as wage subsidies or training) for particular target groups (such as young or older people). We now have five different wage subsidy schemes and three career guidance and training programs for different groups. If we go too far down this track, transaction costs will be high, inequities will emerge in the treatment of different groups (such as the restriction of JobMaker wage subsidies to people under 35 years) and the system will be less responsive to the diversity of individual needs.
These results are consistent with findings of economic research into ‘incomplete contracts’, that when governments contract out human services (where it’s hard to accurately measure outcomes) service quality is often sacrificed to reduce costs (see Hart O 2017, Incomplete Contracts and Control. American Economic Review Vol 107 No 7 pp 1731–1752). Employment services are a monopsony market – with one or a small handful of dominant purchasers. In monopsony markets, as shown by the treatment of farmers by the big retailers, there is a tendency for prices to be squeezed. On the other hand, where there are significant barriers to entry, the largest providers (who have the access to capital and scale to ride out these financial pressures) tend to dominate as the market matures. We may reach a point where the largest providers, including international conglomerates, set the terms by freezing out local competitors – and where others are too big for governments to allow the fail.
Either way, the main benefits of provision of employment services by non-government organisations – diversity and local connections – are diminished.
It doesn’t follow that employment services for those most disadvantaged in the labour market are best provided by a public agency. While it would make sense for the online service for people closer to employment to be managed in this way (for example, using Centrelink as the back-up service), not-for-profit community services that are well connected with their communities and have the right intrinsic motivation or ‘mission’ can probably do a better job for those who need more personalised help. Depending on the priority given to local cooperation, cost minimisation and choice of provider, there could be number of competing providers in each local area, or a single provider (or a set of providers each of whom specialises in assisting a particular group or industry).
The one clear advantage of funding to employment outcomes is that it sharpens a provider’s focus on the end goal. For this reason alone, it is worth preserving, at least in the form of bonus payments above and beyond fees for services provided.
The Australian experience with employment service purchasing suggests that a hybrid model would work best. This would include a public provider and online service for those requiring limited employment assistance. For those needing more personalised and intensive support, a network of not-for-profit services funded for broadly defined service inputs and sharply defined outcomes, underpinned by a robust licensing system to assure quality and share best practice, is probably the best option. There’s no need to bundle up the entire service offer into a single program. Providers should have at their disposal a suite of national programs (not too many, and no more pilots) offering wage subsidies, training, career guidance and other specialised supports. That seems to be the best way to guarantee that everyone unemployed long-term or at risk receives the help they need.
All that’s old is new again, brought back in a different form.