Human services: The next wave of productivity reform
Commissioner Stephen King delivered the Queensland Productivity Commission (QPC) Productivity Lecture in Brisbane on 8 November 2019.
Download the speech
Read the speech
More than 25 years ago, the report on National Competition Policy, otherwise known as the Hilmer report, kicked off a wave of productivity reform. It was centred on government-owned monopolies in key infrastructure industries, including telecommunications, transportation and energy. The report recognised that:
[i]n the case of many public monopolies … protection from market forces through government regulation or other government policies has often allowed enterprises to develop structures unlikely to be found under normal market conditions.1
This is a polite way of saying that these sclerotic government monopolies had high costs, low innovation and were a drag on Australia’s productivity. Reforms flowing from the Hilmer report transformed these key infrastructure industries and helped underpin almost three decades of economic growth in Australia.
In my opinion, we are at the beginning of a similar wave of productivity reform in human services. Like the public monopolies considered by the Hilmer report, human services operate in markets that are anything but normal. They are dominated by government expenditure, regulation and, for some services, government provision. For many human services, consumers are disempowered, competition—if it exists—is managed, and incentives are distorted. In some areas, markets have been designed for the benefit of providers rather than consumers, and innovation is discouraged. Funding is often based on inputs or block grants rather than outcomes and where markets involve multiple levels of government, services either overlap or are missing.
In this talk, I will briefly outline why productivity reform to human services is important, why it is needed, some lessons for system wide reform and the potential benefits that can be gained from reform.
What are human services?
The problem with talking about human services as the next wave of productivity reform, is that it is far from clear what we are talking about. The definition of ‘human services’ is, at best, vague.
In Australia, the term refers to a group of services.2 But which ones?
Unfortunately, the answer is unclear. The 2015 Competition Policy Review chaired by Professor Ian Harper doesn’t help much.
The human services sector covers a diverse range of services, including health, education, disability care, aged care, job services, public housing and correctional services.3
But is this the right list of human services?
Similarly, when the Commonwealth Government asked the Productivity Commission to look at reforms to human services, it didn’t actually define the term.
The human services sector plays a vital role in the wellbeing of the Australian population. It covers a diverse range of services, including health, education and community services, for example job services, social housing, prisons, aged care and disability services.4
And I am sad to say that the Productivity Commission did not rectify this omission in our report. Rather we simply noted that “[t]he terms of reference for this inquiry do not define ‘human services’, or provide a definitive list of which human services are within scope”.5
So, I will be pragmatic. I am interested in productivity gain and government influence. And I am from the Productivity Commission. And each year the Commission puts out a Report on Government Services that covers six broad areas: Childcare, education and training; Justice; Emergency management; Health; Community services; and Housing and homeless. If we are going to define human services by a list of services to the community that centre around government, then that list is a pretty good starting point. So, at least for this talk, human services will be defined by these six areas.
Why do these services matter from a productivity perspective?
There are at least four reasons why we care about this list of services.
First, these services are a substantial part of Australia’s economic activity. Even if we exclude housing, these services make up around 20% of our national output, with education at around 6% and Health at around 10% of GDP.6 So improving productivity in human services can impact a large part of the economy.
Second, these services matter for the wellbeing of many people. This is rather obvious for services like housing, health and education. Even without government expenditure in these areas, private markets for these services would still exist, and they continue to exist, and in some cases thrive, despite government intervention and expenditure. For example, in education, about one-quarter of all expenditure is non-government.7 In health, the equivalent number is about one-third.8 And in housing, private expenditure and ownership dominates. There are about 20 private dwellings for every social housing dwelling.9
Another way to see the importance of these services is to note how they are accessed over an individual’s lifetime. As the Commission noted in its final report on Human Services:
[E]veryone will access human services in their lifetime, including children, the elderly, people facing hardship or harm, and people who require treatment for acute or chronic health conditions.10
So, improving productivity in human services will benefit all Australians.
Third, these services provide critical underpinnings for our society. As the Commission’s 2018 report into inequality in Australia showed, in-kind transfers—which are basically human services—significantly reduce the level of inequality in Australia. Technically, they reduce the Gini coefficient of inequality from about 0.25 for private consumption to around 0.2 for final consumption. By way of comparison, they have a larger impact on reducing inequality (as measured by the fall in the Gini coefficient) than our progressive income tax system.
So, improving productivity in human services helps to reduce inequality in Australia.
Finally, despite the significant levels of private expenditure, each of these services involves substantial amounts of government expenditure and in each case the government is intimately involved with the production and the consumption of the services. The nature of a service market fundamentally changes when over half of the expenditure on a service is derived from government. But that is the case for each of these human services.11 Who gets this funding, and how they are allowed to use it, matters.
In summary, if we want to find productivity gains for Australia, looking at a set of services that make up around one-fifth of our output, are consumed by a wide variety of people, are particularly important for those who are struggling, and have a high level of government involvement, are a pretty good place to start. Human services fit these criteria.
Human services markets face a range of distortions that limit productivity and effectiveness. These differ between services. However, they fall under four broad categories—failure to define what the service is meant to achieve; failure of service design to achieve the desired objectives; failure to implement the designed service effectively in the real world; and failure to measure the service outcomes. Put simply: objectives, design, implementation and measurement.
The objective for human services can be unclear. While a ‘high level’ objective may be easy to state—such as ‘user choice of medical service’, ‘educating our children’ or ‘community safety’—these high-level objectives can mask complex and conflicting elements.
For example, in its draft report on Imprisonment and Recidivism , the Queensland Productivity Commission noted the complexity of ‘community safety’.12 They noted that it involves a time frame—presumably over the longer term. It involves a requirement for the justice system to meet “community expectations about justice and fairness”. It has to be cost effective, with the benefits outweighing the costs. And it must operate within resource constraints using the tools of deterrence, detention and rehabilitation.
‘Community safety’ is really a set of multiple objectives.
Having multiple objectives for human services is not a problem by itself, so long as the objectives are clear and it is understood how conflicts between objectives will be resolved. 13 However, when objectives are unclear or unstated, then someone will need to make decisions about which objectives are pursued. Often this will be the service provider. To the degree that these decisions are not aligned with either government objectives or user preferences, service effectiveness is compromised.
For example, as the Commission found for end-of-life care services, the ambiguity in objectives means that clinicians are often placed in a situation to make judgements about the location and nature of care, and that these decisions, while medically appropriate, are often not in line with the preferences of the service users. The result is unwanted treatment and death in hospital rather than home.14
A human service can only achieve the desired outcomes if the delivery system is well designed. But market design is hard. Getting the right incentives for governments, service providers and users can be difficult even if the objectives are clear.
Let’s start with government.
The Commission’s recent draft report into the mental health system notes how design problems can arise between governments:
The division in mental health care responsibilities between the Australian Government and State and Territory Governments has led to service gaps. … These gaps result from two key factors. First the relevant roles and responsibilities are unclear. Second, State and Territory Governments face incentives to direct resources towards acute care instead of providing more care in the community.15
Mental health is not the only example of these problems and having either service gaps and/or overlapping services can reduce service effectiveness, increase costs and lower user outcomes.
Service providers also need to face the correct incentives. For example, under the jobactive program, unemployed Australians are supposed to be assisted by a service provider to find employment. But the service providers face conflicting objectives—both to assist their clients to find work but also to monitor client compliance with the mutual obligation requirements of the program. As a result, some clients find their providers acting as barriers to finding employment.16
Unfortunately, poor incentives for service providers are common in human services.
Finally, service users need to be motivated to use the ‘right service at the right time’. But this is often not the case. For example, in healthcare, people are sometimes driven by price to access healthcare through a hospital rather than a community-based service—despite the community service having a far lower cost to the taxpayer.
The failure in human services to design user-centred programs results in poor outcomes and high costs.
Failure of implementation
A well-designed human service still requires practical implementation. But to quote the Scottish poet Robert Burns, “The best laid schemes o' mice an' men Gang aft a-gley”.
This holds for human services. Unless carefully implemented, well-designed human services programs can fail to meet their objectives.
For example, the National Disability Insurance Scheme (NDIS) is a well-designed program that empowers some of the most disadvantaged people to access the services they need, when and where they need them.
But the roll out of the NDIS has been problematic due to overly optimistic timing for users to receive their plans and for service providers to expand to meet demand. The Productivity Commission concluded that the “focus on participant intake has compromised the quality of plans and participant outcomes” and that “the rollout timetable for participant intake will not be met”.17 The Commission also noted that there are significant workforce shortages in some areas. The National Disability Insurance Agency has responded by using price controls to try and balance the availability of services and value for service users. The results, predictably, satisfy no one.
Failure to measure
Finally, there is the failure to measure.
Measurement, by itself, has no value. It is simply data. The issue is what to do with this data.
Outcome measurement is a key input to two key aspects of human services—evaluating the services to work out what works and informing consumers so they can best match their requirements with the services that are supplied.
On the former, in the Commission’s recent draft mental health report, we note that the success of our proposed reforms will:
depend on the creation of a strong, evidence-based feedback loop so that program effectiveness can be evaluated with the results being used to help determine which activities are funded in the future.18
On the latter, in our report on human services, we note that publishing outcome information on hospitals and specialists, as already occurs in some countries, can improve productivity by helping consumers locate the best services and, most importantly, by encouraging service providers to benchmark themselves and work to improve their performance.19
Unfortunately, at present, measurement, publication and evaluation of outcome data is rare.
Fixing the problems to raise productivity
So, we have four problems—objectives, design, implementation and measurement.
It would be nice if I could now offer a simple formula to address these problems and raise productivity in one-fifth of our economy. But there is no simple formula. Human services differ on many dimensions and solutions must be service specific. However, we can identify some broad approaches to reform.
A good starting point to address the problems and reform human services is the separation of functions.
The Harper report, in its Recommendation 2, suggested a three-way split of functions in human services, “separating the interests of policy (including funding), regulation and service delivery.”
I would add a fourth separate function—evaluation.
This four-way split lines up with the core problems that beset human services. Policy means defining objectives and sources of funding. What are governments trying to achieve and who is responsible for funding what? Regulation is about service design. How are the correct incentives put in place for all participants and how are the rules enforced? The regulatory body interacts with the service providers in implementing the relevant service delivery models. And the loop is closed through independent evaluation based on outcomes to determine what works and what doesn’t.
Having a clear separation of the four functions—policy, regulation, service delivery and evaluation—creates clear lines of accountability. It also reduces the potential for undesirable interactions between these roles.
To see this, let’s return to the earlier wave of infrastructure reforms in Australia.
The Hilmer inquiry, when looking at government infrastructure monopolies, noted the inherent conflict between having commercial and regulatory functions in the same organisation.20 Indeed, those of you who have grown up post-Hilmer would find absurd the idea that a telecommunications company could also be the telecoms regulator, or an electricity supplier could be the energy regulator. But in the 1980s, that was the norm. The gamekeeper was the poacher.
In human services today, similar conflicts exist, for example where block-funded service providers decide what services will be delivered, where and when. Or government policy makers determine not just what services should be provided but specify who will deliver them. Or where service providers get to evaluate the services they deliver. Or where a government department both designs and delivers a service, regulating itself.
How can separation work?
In the Commission’s draft report into Mental Health, we recommend clear separations between the four functions.
Policy is the role of government. In Australia, this means both the Commonwealth and the State and Territory governments, and we recommend that the Council of Australian Governments (COAG) develop a National Mental Health and Suicide Agreement to make funding and responsibilities clear. There also needs to be a National Mental Health Strategy to establish nationally consistent policy.
Implementation and regulation would remain divided across the health system. For primary care, the existing Medicare system provides incentives that, while imperfect, appear to involve less bureaucracy and more effective service delivery than the alternatives. For high intensity and complex clinical care, and non-health supports, our preferred position is that new Regional Commissioning Authorities are established to determine what is needed at a local level, then to commission evidence-based services to meet these local needs. Together, this provides an implementation approach for stepped-care services across Australia.
For service delivery, service providers would be diverse. They would include private practitioners such as for GPs and psychologists; for-profit companies such as some hospitals and clinics; not-for-profit organisations, such as providers in social housing and disability services; and government bodies, including schools and correctional facilities.
And finally, the Commission recommends that the National Mental Health Commission is expanded to not just monitor the mental health system, but also to evaluate services to determine what works—and what doesn’t—and report these results back to government.
A second broad approach to human services reform is timing.
Different reforms need to proceed at their own pace. There are many changes which can be implemented relatively quickly to improve the outcomes for human services. For example, in the Productivity Commission’s Human Services Report, we note that some human services, such as public dental services, exist in much larger private markets. Shortages in these services can be addressed relatively quickly by policy makers moving away from models that falsely assume that government funded services must be separated from these private markets. Similarly, in primary care, simple changes to GP referrals for specialist care can empower consumer choice (Recommendations 10.1, 10.2 and 10.3).
However, for other reforms, time is needed to test alternative models and allow for adjustments. For example, consumer outcomes can be improved by increasing public medical data and releasing risk‑adjusted information on the clinical outcomes achieved by individual specialists. But this reform needs to be implemented carefully over time to avoid incentives for practitioners to alter their mix of consumers in order to artificially raise their measured performance. Nonetheless, models exist in both the US and the UK for such reporting and in both countries public reporting has been found to save lives.21
Unfortunately, without appropriate leadership from government, neither the short-term nor the long-term gains will be realised.
A third broad approach to human services reform focuses on the policy objective. As an economist, the objective is relatively straight forward. The overriding objective should be the outcomes for the consumers who receive the services. And, in general, the consumers themselves are best placed to judge whether desirable outcomes are being achieved.
While focusing on user outcomes might appear obvious to an economist, it is foreign in many areas of human services.
For example, in health, success is often judged by clinical outcomes rather than the quality of the consumer’s life. While this is slowly changing and patient reported outcome and experience measures—PROMs and PREMs—are becoming more accepted, there is still a culture that the clinician knows best.
And when talking with service providers in a range of areas, including aged care, indigenous services, and disability services, I have been surprised how some providers consider that the service system needs to be designed for their benefit, rather than addressing what the consumers want. When challenged, the providers either appeal to costs—that it is too expensive to gear services to consumers’ preferences—or paternalism—that the providers know what is best for the consumers.
Moving from a ‘provider knows best’ to a consumer-centred approach is critical to improving productivity in human services. A system that produces lots of what isn’t valued is unproductive.
The Harper Review took one approach by recommending that “[u]ser choice should be placed at the heart of service delivery” (recommendation 2). User choice drives innovation and productivity in competitive markets. It leads producers to supply what consumers want. But as that Review noted, user choice and competition are not always possible in human services markets. For example, in health markets, consumers often rely on the judgement of professionals. In corrections, I am not quite sure what ‘user choice’ means. And in rural and regional areas, user choice is often limited by markets that are too thin to support multiple suppliers.
However, there are two other tools to improve consumer focus. The first is to ask consumers what they want, by including consumer input in service design, and codesigning the objectives of the relevant services. The second is measuring and evaluating outcomes. The only way to ensure human service outcomes are consumer focussed is to check.
What are the potential gains?
What are the potential gains from reforming human services?
They make up a large part of our economy, and the current approaches to delivering these services appear, at least to an economist, to fall well short of a desirable, far less an optimal, system. But how big are the gains from reform?
We need to place some caveats on this question.
First, there are differences between gains that are measured in national accounts and those accruing to society. This is stark in human services, where the measured inputs and outputs may be modest, while the unmeasured transformation to the lives of individuals and families can be far greater.
Second, even for measured data, the gains may be indirect. A poor education system may cost the same as a great system. The productivity gains of education reform may not be obvious in the directly measured numbers. Rather they are reflected in the improved productivity of the workforce into the future.
With these caveats in mind, what are the benefits of human services reform?
There is little independent evidence on the economic gains from broad-based human services reforms. This is not surprising. Most analysis is based on specific interventions for particular services.
Where a broader reform is examined, care must be taken when bringing it into the Australian context. For example, in 2009, when Chair of President Obama’s Council of Economic Advisers, Christina Romer estimated that slowing the growth of health care costs in the US by 1.5% would increase US real output by about 2.5% by 2020 and by 8% by 2030.22 These are huge numbers but have to viewed with caution. They may say more about the state of healthcare and politics in the US than provide any lessons for Australia.
The Productivity Commission suggested some modest reforms in health and estimated the potential gains in our five-year productivity review.
The net present value of the future stream of economic impacts over twenty years is estimated at about $140 billion (in 2016 prices).23
Much of the benefit would be personal gains to those who access healthcare. But even direct economic gains would be significant with potential gains of “over $4 billion a year” to GDP and potential reductions in long term government health expenditure of over 6%.24
However, with broader reforms, the gains are likely to be higher. For example, the Productivity Commission’s draft Mental health report, estimated that mental illness costs Australia around $180b per year, with around $50b of this being a direct economic cost of reduced productivity. Even reducing this economic cost in one part of the health system by 10% leads to a $5 billion per year gain to the economy While we have not measured the net gain of our full reform program at this draft report stage, the inefficiencies are stark and the benefits of economy wide reform in mental health will be in the tens of billions of dollars over the longer term.
When the Industry Commission examined the potential gains from the National Competition Policy reforms, flowing from the Hilmer report, it estimated that “in the long run, once all adjustments have taken place, there would be an annual gain in real GDP of 5.5 per cent …”.25 Broad reforms to improve Human Services could lead to similar economic gains, and potentially much greater gains to the lives of many Australians.
Human services make up around 20% of Australia’s economy. And they are ripe for reform. It will take effort—including coordination and cooperation across different levels of government. It will require careful planning including making sure that systems are robustly designed and implemented carefully. It will require increased consumer involvement—either through expanding user choice or through co-design. And it will require clear outcome objectives with data being collected and assessed to make sure that these objectives are being met. In other words, human services need reform across four dimensions: objectives, design, implementation and measurement.
These reforms will not be easy. There are vested interests who will fight reform. But this was the case with the reforms to infrastructure following the Hilmer report. And just like in that earlier wave of productivity reform, there will be mistakes. But that is not a problem where the reforms are robust, and mistakes are recognised and rectified. And, as with the earlier wave of reforms, human services reforms are worth it. Not only will they raise our measured productivity, they will significantly improve the lives of most Australians.
1 Hilmer, F., M. Rayner and G. Taperell (1993) National Competition Policy, Commonwealth of Australia, p.216 Return to text
2 In contrast in the US, according to www.humanservciesedu.org the term human services is used to refer to a set of occupations “designed to help people navigate through crisis or chronic situations”. Return to text
3 Commonwealth of Australia (2015) Competition Policy Review, Final Report, p.218. Return to text
4 Terms of reference: Productivity Commission Inquiry into introducing competition and informed user choice into human services (29 April 2016). Return to text
5 Productivity Commission (2016) Introducing competition and informed user choice into human services: Identifying sectors for reform, Study Report, Commonwealth of Australia, p.34. Return to text
6 These figures ignore activities, such as unpaid care provided by voluntary carers including friends and family. This care can be substantial. For example, in its submission to the Productivity Commission’s 2016 Study Report, Carers Australia (submission 259) stated that “[i]nformal carers are major contributors to the human services sector—an estimated 2.7 million family and friend carers provide almost 2 billion hours of care each year.” Return to text
7 Rice, J.M., D. Edwards and J. McMillan (2019) Education expenditure in Australia, Australian Council for Education Research, published July 24. Return to text
8 Australian Institute of Health and Welfare (2019) Health expenditure Australia 2017–18, Health and welfare expenditure series No. 65, Cat No. HWE 77, Canberra. Return to text
9 The number of social housing dwellings is around 436,213 from the Australian Institute of Health and Welfare (2019) Housing assistance in Australia 2019, Cat No. HOU 315. The number of private dwellings is extrapolated to approximately 9m from the ABS numbers from the 2011 census which gives about 7.76m private dwellings in 2011. Return to text
10 Productivity Commission (2017) Introducing competition and informed user choice into human services: reforms to human services, Inquiry report, No. 85, October, p.3. Return to text
11 For housing, I am restricting attention to public housing (including social housing) and homeless services. Return to text
12 Queensland Productivity Commission (2019) Inquiry into imprisonment and recidivism, Draft Report, February, p.9-10. Return to text
13 See for example, Productivity Commission (2016) Introducing competition and informed user choice into human services: Identifying sectors for reform, Study Report. Commonwealth of Australia, p.38. Return to text
14 Productivity Commission (2017) Introducing competition and informed user choice into human services: reforms to human services Inquiry report, No. 85, October, chapter 3. Return to text
15 Productivity Commission, (2019) Mental health, Draft Report, November at p.928. Return to text
16 See: The Senate Education, Employment and references Committee (2019) Jobactive: failing those it is intended to serve, Canberra, February at chapter 7. Return to text
17 Productivity Commission (2017) National disability insurance scheme (NDIS) Costs, p.12. Return to text
18 Productivity Commission, (2019) Mental health, Draft Report, November, p.42. Return to text
19 Productivity Commission (2017) Introducing competition and informed user choice into human services: reforms to human services, Inquiry report, No. 85, October, chapter 11. Return to text
20 Hilmer, F., M. Rayner and G. Taperell (1993) National Competition Policy, Commonwealth of Australia, p.217. Return to text
21 Productivity Commission (2017) Introducing competition and informed user choice into human services: reforms to human services, Inquiry report, No. 85, October, chapter 11. Return to text
22 Romer, C. (2009) The economic case for healthcare reform, Commonwealth Club, June 8. Return to text
23 Productivity Commission (2017), Impacts of health recommendations, Shifting the dial: 5 year Productivity Review, Supporting Paper No. 6, Canberra, p.6. Return to text
24 Productivity Commission 2017, Impacts of health recommendations, Shifting the dial: 5 year Productivity Review, Supporting Paper No. 6, Canberra, p.7. Return to text
25 Industry Commission (1995) The growth and revenue implications of Hilmer and related reforms, A report by the Industry Commission to the Council of Australian Governments, Final Report, Canberra, March at p.53. Return to text