Alf Rattigan Lecture: Bill English on data customers and the limits of universal public services
27 November 2019
● News and media
Sir Bill English KNZM delivered the 2019 Alf Rattigan lecture on 27 November at The Shine Dome in Canberra. Below is the full transcript of the former New Zealand Prime Minister’s speech.
The Rattigan Lecture is the third in an annual series in honour of G.A. Rattigan, the renowned Chairman of the Tariff Board and Industries Assistance Commission (predecessors of today’s Productivity Commission).
The purpose of the Lecture is to elevate the importance of good process, sound governmental institutions and effective leadership to achieving nationally beneficial public policy and reform. It is delivered each year by an Australian or New Zealander who has played a significant role in promoting or implementing these ideals.
It’s a tribute to the free flow of ideas across the Tasman that a New Zealander has been asked to give this lecture in honour of Alf Rattigan; a quintessential Australian economist and reformer. We owe a debt of gratitude to Alf Rattigan, Fred Hilmer, Gary Banks, and a generation of Australian policy makers who helped make New Zealand a successful economy by making Australia a successful economy.
Thanks to our close economic and policy histories, global events and intellectual tides have shaped both our nations in ways that are at once similar and different enough to learn from and share with one another.
For policy minded New Zealanders, Australia is an offshore laboratory running with seven or eight experiments on everything relevant to us. Today is a rare occasion when we can give back a bit of advice. I acknowledge the work of ANZSOG and its leaders over the years encouraging the ongoing Trans-Tasman flow of knowledge, creativity, and learning.
Alf Rattigan thought good process, sound institutions, and leadership mattered for good public policy.
Public policy determines the institutions that define and regulate the market, and the governance of those institutions requires a continuous flow of reliable performance information.
In the marketplace, this information comes largely from prices. Because they show matches and mismatches in supply and demand. Prices provide system participants with rich information about where resources should flow, and the market provides a constant feedback loop for complex systems to learn, adapt, and adjust. Public policy-making for government services by definition lacks meaningful access to this rich information input.
Like the market, government services have to manage incredible complexity, but without the informative input of prices. Instead, institutions and process must fill the information gap and make decisions by selecting from a wide range of information and signals, and therefore the quality of these processes and institutions matters more than it does in a market. Today I will focus on the largest portion of public services – government funded social services, broadly speaking the education, health, justice and welfare sectors. The quality of these institutions matters to the economy because they are large. The government funded social sector makes up around 18% of GDP and as our economies tilt towards services that proportion will grow. Direct skilled employment in the government-funded social sector will account for around 60% of all job growth in Australia in the next five years. In New Zealand the public health system is bigger than the dairy industry and now growing faster.
More importantly these are the services that focus on building and maintaining human capital, and protecting the most vulnerable. Their reach and impact span the entire population, and their governance is at the heart of our democracy. Their impact on the lives of children and the elderly define our sense of fairness and care.
I have had the privilege of political positions where I could make decisions to give effect to this aspiration. I found individual stories of people who experienced the services as illuminating as official advice or statistics. One story I heard was about a sole parent with a disabled child and struggling with debt. She documents her visits to 14 agencies over two exhausting weeks. Only one of these remembered her name, knew her details, made her a cup of coffee, and provided her with an immediate solution to one of her problems. That agency was a payday lender charging 30% interest and outperforming social services in compassion and problem solving in the process.
Can social services do better than a payday lender?
I have great admiration for the thousands of people who care and work hard in the social services. To question the institutions they work in is to honour their work and the greater good that can be achieved.
I believe it’s time to reflect deeply on the core institutions and processes of the post-war welfare state.
The current paradigm assumes that our institutions are driven by what works.
A paper prepared for last year’s review of the Australian public service concluded that this generally isn’t what happens.
It concluded that public service decision makers generally don’t use evidence, and don’t learn from past experience.
So, when politicians and the public see problems they resort to alternatives to public service policy institutions. Royal Commissions, inquiries and reviews are deployed to exert external independent influence. The public believe public institutions either don’t know what happens to the customers or can’t be relied on to tell it how it is.
But customers know how it is.
A homeless man sleeping in the park knows how it is. He can’t get emergency cash for accommodation because he can’t prove his identity – although his identity was certain enough for him to be imprisoned for seven years, and then released without a drivers’ license, passport, or utility bill to show a service provider.
A father of four who has prompted two police callouts for domestic violence to his home in a single year knows how it is. Someone found a relative who was able to help him learn how to deal with the pressures of a low income family household. The previous year there were 85 domestic violence callouts to that household. After 85 instances of violence in that home, someone finally changed something.
These are our customers. They have a very clear understanding of what the service is. Not what it should be, or what well intentioned people designed it to be, but how it is.
I will use the word ‘customer’ through this lecture. The public sector prefers words which define people by the public service they are using at the time – patients, students, prisoners, clients. These terms are useful but conceive of people as partial, and sometimes as problems. The term citizen is more positive but historically positions the person in a civic role as a voter and party to a social contract to support public institutions who support them. I prefer the word ‘customer’ because it recognises that people have an existence independent of the public services they use, and that their personhood gives them the ability to know their own needs and make choices. So customer asserts a different sort of relationship between the institution and the person.
For too many of our customers, social service institutions look unreliable, ineffective, threatening, and cynical – because that’s how these institutions behave.
Policy reformers like Alf Rattigan tell us to look at institutional arrangements as an explanation for the behaviour we see, and the results we get.
The post war welfare state was designed to deliver similar services to large numbers of people, regardless of their ability to pay – a monopoly provision of free services. And this strategy has been relatively successful; we now have a basic income safety net, and universal health care and education.
But there are tradeoffs, which have significant consequences.
The trade-off for monopoly free provision is lack of differentiation.
Although historical patterns of use have some influence on supply, most social services come in commodity standardised packages. So, intense demand can come in the form of repeated use of the same inadequate service – such as 85 police callouts to the same address. Intense demand comes also in the form of people with low resource making haphazard connections between many untimely standardised services. The people with the toughest problems have to negotiate the least coordinated parts of the system. The result is well known problems with well-known solutions stay unresolved, with lifelong impacts for the customer.
Analysis performed in New Zealand ranked everyone in the population by the present value of the public services they used. The work demonstrated the Mason curve – named after the official who thought of it – which shows that 85% of the population use 50% of public services, while the 15% with the highest needs consume the remaining 50%.
The universal model of provision, which delivers roughly uniform services to the entire population, doesn’t account for this distribution of need. Instead it assumes a normal distribution, designing and delivering services for a mythical ‘average’ person. The reality of need is a logarithmic curve, and for the people at the sharp end of it, the services delivered via universal provision are not responsive to their needs. It’s not clear the existing institutions want to or know how to deal with the hardest social problems. Of the 1% of New Zealand children who are at the greatest risk of poor outcomes in their lives, 20% achieve their secondary school leaving qualification. These remarkable children experience a toxic mix of welfare dependency, violence, addiction and abuse in their homes. The generalised statistical and policy discourse frames their prospects as hopeless and in my experience the normal policy process hadn’t bothered to find out that these children succeeded let alone why they succeeded.
One reason is policy culture. Many in the policy and media world are educated in frameworks of hopelessness. Too many believe that broad categories of people are destined for victimisation and failure but they still believe with blind faith in public institutions which persistently fail to change that destiny.
This 15% of the population at the hard end of the spectrum of need are characterised by their diversity not their similarity – their needs are complex, their situations are unique, and the solutions to their compounding challenges will not come from the broad-strokes approach of universal provision. These 15% are thousands of different complex households, needing thousands of unique relationships of sustained trust, and thousands of different pathways to change.
The suppliers of universal services cannot respond adequately to complexity because their vertical service structures are defined by legislation, professional boundaries, and funding flows. They are set up to deal with only one part of a person, not the intersection of factors that shape real lives in real communities. Their annual budget cycles are designed to minimise immediate cost – cost accounted for in terms of dollars spent, not outcomes purchased.
These bureaucratically determined costs are mismatched against the needs of the 15%. These people have needs significantly higher than average. So the 15% tend to be forced to the edge of the system or cost shifted to another system or deliberately underserviced.
Providers can’t respond to greater need by adapting the services neither can they increase resources for the 15% on the scale required to have an impact. Policy makers become addicted to repeating the litany of failure because failure is rewarded with more resources.
This relationship between institution and customer is not one of care – its co-dependency where each is maintained by the dysfunction of the other. Dysfunctional households with few resources continue to interact with delivery systems that sap their energy, their resilience, and their trust. The purpose of the welfare state is to reduce misery, not to feed off it.
The post-war welfare state model of one-size-fits-all services has reached its limit.
Collaboration is not an adequate answer. The 90% failure rate is necessary to find out which 10% of collaborations are sustainable in the real world. Collaboration among large universal providers, or between large monopolies and small NGOs, is an exhausting and high cost process that will only work if every party shares a strong sense of purpose, if the leadership is inspired, if there is enough funding, and if nothing goes wrong. Collaborations often fail at the first hurdle because agencies don’t align governance with their different operational and financial decision processes.
Doing more of the same but all in the room together rarely works.
Large scale change to the system of universal provision is technically difficult and politically hazardous, which is why the public policy world doesn’t believe it’s possible to significantly improve the system.
I believe it is possible. And furthermore, I believe that the patterns I have described are so negative for the lives of our most vulnerable that we are obliged by their suffering to take risks to save lives that can be saved, treat untreated illness, reduce violence and abuse, rebuild families and realise human potential. We can better use public resource to change the lives of the most disadvantaged.
Social investment is intended not to replace universal provision but to overcome its limits by reframing old problems to generate new solutions. In New Zealand, we found social investment to be a sustainable and pragmatic way forward.
Social investment tools enable the public sector to do functions it does now but much better – to differentiate needs, to understand impact of its interventions, to prioritise where it spends resource and to build feedback loops for continuous adaptation. This approach encourages the social sector to do better what it says it does – change lives by investing now for long term benefit.
The key features of an investment approach to social services are:
Focus on the customer, not the institution
Integrate data to understand the customer
Use actuarial life course projection to understand future impact
Create an Investment framework
Supervise the investment process independent of service providers
Focus on outcomes and accountability
Unlike the traditional universal approach which seeks to shift aggregate indicators, these tools seek to identify where change is needed most. It seeks out the most amenable opportunities for change and ranks them by return on investment.
I will comment briefly on each of the key features of the investment approach with some lessons from nine years of implementing social investment.
Focus on the customer and not the institution
Focusing on the customer is harder than it sounds. The universal welfare state resists the idea that the people receiving its services are more important than the existence of the service. This is one reason that ‘customer’ is such a powerfully uncomfortable word in public policy. It’s argued that the word ‘customer’ trivialises the relationship between the citizen and the state. However, from the perspective of the customer, the outcome is all that matters. Why should they go on receiving chaotic and inefficient services in order to maintain institutions they don’t trust?
Customer focus for policy makers means making decisions as if the person we are talking about is sitting in the room. How would the young woman suffering from episodic schizophrenia think of how we discuss the policies designed for her? Or the angry 14-year-old boy in state care? Or the mother struggling to obtain the right support for her disabled child? Status quo policy discussion is dominated by the needs of politics, funding, and existing institutions. This mother should hear a conversation where she is sufficiently important that the social agency must change to align with her needs.
I have found three simple questions prompt the right sort of customer focused thinking:
Which people will benefit from this policy?
Who has a relationship with them to deliver the solution effectively?
How will that person know the impact of their intervention?
Senior managers in social service agencies will know that the budget was spent and that the rules weren’t broken. But they should also be able to answer all three of these questions, and explain how their organisation is organised to align with the customer.
Integrate data to understand the customer
Linking data about people is a powerful way to understand what they need in order to live their full potential. Mountains of data exist about people using public services, but public agencies are where useful information goes to die and turn into statistics.
Agency culture is to control all flows of information in the name of reputational risk management, but that control comes at a cost to the customer and to the community.
No public agency offers to share data. They all resist and it takes determined political leadership to succeed. Senior managers don’t know about data because they don’t use it. When someone comes asking about their data they find out their agency has lots of expensively collected low quality data that no one uses, that it’s chaotic and siloed because it’s organised around collection not customers. That’s when their risk management skill clicks in and they lie low hoping it will all go away.
Eventually agencies can be embarrassed about resisting more knowledge and putting their interest above the collective. The hardest cases have to be reminded of their constitutional responsibilities. Once they concede there might value for them and their customers in linking data, they will try to control all aspects of the process to stop bad news that might affect their credibility or expose their minister. All this is called “privacy concerns”.
The best way to improve data is to use it. Expose it to scrutiny and invite questions about its relevance to the decisions a government can make. Then the agency will find the will and the resource to treat information about people with respect and change their methods of collection, cleaning, warehousing, and sharing.
By law in New Zealand the data belong to the person described, not to the agency. The agency has a statutory role as a steward who should use the data in the best interests of the customer, not their own best interest.
The early and easy gains comes from using data within a sector or agency. Within welfare, education, and corrections, data driven investment made gains regardless of other influences. The demonstration of this value feeds a growing appetite for data integration.
More significant gains come from linking data across two agencies who share the same customers. An example is to link data from the health and welfare systems about sickness and disability recipients. This group is often neglected under the status quo because they don’t appear close to employment, and in the short run doing nothing for them doesn’t cost much. However positive change for even a small number of these customers has a large impact because of their high probability of long term dependency.
Health systems don’t face the cost of under-treatment, and the welfare system doesn’t know much about health needs. But bringing the data together uncovers untreated and undertreated illness that traps people in welfare and creates the opportunity for health and welfare to see the common benefits of new and better services.
Create pictures of people’s lives with actuarial life course projection
Linking data across agencies allows us to see the complexity of people’s lives – linking data longitudinally allows us to understand and measure the impact on their lives of policy decisions over time.
With a life course view of the customer, it’s possible to make the transition from cost-based thinking to putting a meaningful price on outcomes, and linking investments to future impact. It enables the development of predictive analytical models, so that we can see how lives are changed by good investments in services that work, and calculate the reduction in future costs or the increase in potential as a result. The Federal government has built such a long term model. The same technique has been used by the New South Wales government for children in the TFM project.
The logic of prevention is easy to grasp but the tools to make it more than a policy cliche are few and far between. Prevention should be focused where it’s most likely to prevent something. With linked, longitudinal data constructing a life course view, we can transform social services from a damage control mechanism to proactive investments in a better future.
These predictive models are really just stories about people’s lives but those stories about real people in actual communities are in my experience profoundly motivating for policy makers. We found the Million dollar kids You can predict from the age of seven, the life course of children likely to experience a million dollars’ worth of services by the age of 35, in benefits, emergency care, hospitalisation, court costs, and imprisonment. It’s less than 1% of a cohort, in New Zealand 600 in a year.
Think of the life we know these children will lead and the courage and resource they will need to survive and hopefully to thrive. They will deal daily with poverty, violence, illness abuse and broken relationships. All we have to do to help them better is wrestle with funding, contracts, responsibilities and more money than we know how to use effectively.
The numbers tell us a story we didn’t know. Now we know, how could anyone stand by when these children face endless misery? There aren’t many of them and one by one their lives can be changed. This approach has driven a complete overhaul of the system for children in care in New Zealand.
Create an Investment framework to allocate funds according to return on investment
An investment framework is a useful tool to show where and how to spend new invest for maximum impact later. It links what we know about the value of outcomes and the cost of the status quo to what we know about customers and their different needs. We know from insights generated by longitudinal integrated data that long term impacts for youth at risk of going on welfare by the age of 20 are large enough that in their relatively small numbers they warrant individual attention. With an investment framework in place, we have a consistent structure around which to analyse and organise the activities of multiple system factors in the service of this specific group of young people.
Putting aside funding issues an investment framework creates space for abundant thinking. Social service culture is conditioned by the pressures of living within current year budgets, a discipline empowered by the Finance Minister because there is no better tool. Where there are commodity services for average need customers this sort of simple fiscal control isn’t a constraint. But where need is complex and intensive it severely constrains the solution set.
Rheumatic fever is a disease NZ children shouldn’t get but it persists and we needed some new answers. The data tells us some children are much more likely to get the disease, and where those children are. An investment framework tells us the benefit for a child of avoiding rheumatic fever is large, in fact so large it was worth spending a lot more to prevent the disease even if it worked only for some. All this we know. But the argument was persuasive enough that officials were set the task of determining a way to reduce rheumatic fever incidence with no budget limit. At first they didn’t believe that there was no budget limit even though it was the Finance Minister telling them.
I was surprised how long they took to come back with a proposition. Their thinking had been constrained by randomly limited current year operational funding. Instead of managing the issue by endlessly restating the problem and handwringing about social determinants, they had to think deeply about what mechanisms they could use to prevent one more child from contracting the disease. They came back with an expensive thoughtful solution we fully funded. Rheumatic fever rates dropped by 40% over five years, when they had previously been considered unshiftable.
It was a good investment.
Supervise the investment process independent of agency service providers
I recall a policy discussion about a group of 2000 highest risk young offenders uncovered by the simple measure that they had five police contacts by the age of 17. They have a 70% probability of incarceration. One way to empty an officials meeting room is to ask which agency wants to be responsible for reducing that probability. No agency has a direct interest in identifying and funding a coherent sustainable service to cope with this challenging group.
Ideally a cabinet mandated process identifies priority groups in the population independent of the service agencies. The same independent process is necessary to create new service models for high priority populations free of the preconceptions of existing agencies. Agencies sometimes don’t want more insights because the insight creates more demand for change in their funding and services. An independent entity can also apply measurement that concentrates on collective returns, rather than fiscal savings to a single agency.
Finally, results matter
Results are those incontestably good things to do, like reducing the number of cases of child abuse, increasing qualifications for disadvantaged students, or shorter wait times in hospital emergency departments.
Specifying results in numbers gives politicians and public servants clear direction. With data driven feedback loops, public servants have a consistent system of measurement and accountability around which to organise. Results cuts through the fog of verbiage, good intentions, and useless statistics. Results subvert the deep urge of monopoly providers to restructure as a substitute for better performance. Many governments have tried targets and results with varying degrees of success. The successful recipe is to pick a dozen, not 50. Put clear, functional systems of measurement in place which focus on results, not outputs. Make an individual Minister and senior official responsible, and publish updates.
The discussion about what to measure and who can feasibly be held responsible is half the value of targets. Results like reducing substantiated cases of child abuse test agencies severely in the right way. Agencies reveal their deepest fears as they face the prospects of unambiguous accountability for what actually happens and shifting power and funding among themselves Ministers and agencies who want to appear all powerful when they are bidding for new money are often the best at explaining why they can’t be held responsible for a challenging result. If their influence is low and failure is likely then the budget should be cut so failure is cheap. All this is flushed out by establishing personal accountability for each result.
So: to remind you of the components of social investment.
Customers, linked data, life course, investment, independent oversight, and results.
There are two further critical dimensions to enacting change – they are timeliness, and trust.
Often, even when the system is performing well, the pace at which it progresses renders its best efforts useless. The months that children in care wait for health appointments or family court decisions are a lifetime for the child. For children, timeliness is everything. A referral may create the illusion that something is being done, but no relevant, meaningful change is occurring in the life of that child while they wait. If a universal service can’t be timely because of the queues, then the service is the problem, not the people in the queues. Lack of timeliness is too often regarded as a small price to pay for solidarity. For some in the queue that’s correct but for the most needy it can be crushing. And for the budget the long run cost of waiting today can be large.
The other critical dimension is trust. Many people do not trust the system because they rightly believe it’s not organised to work for them. Unreliable inconsistent and long delayed service feeds cynicism. Worn out or restructured frontline staff struggle to maintain sustainable relationships with families and communities. Trust is hard to establish and maintain when the state agencies have enforcement roles collecting fines, and debts, arresting adults and removing children. In turn, the system doesn’t trust customers to understand their own needs, and does not trust people to make good decisions for themselves. In this world we can blame the customer or rethink how public agencies can change their own behaviour to rebuild sustainable trusted relationships with customers.
A model which places trust in people to understand their own needs is self-direction for those requiring chronic care. A customer is assessed for their care needs, and this assessment is turned into a flexible cash budget. The customer, or their agent, is trusted to make the best decisions about what to do with that budget. NDIS is a fantastic move in this direction and globally distinctive for its ambition, if somewhat over specified in its detail.
Trust is also the foundation of the navigator model. The navigator model is one in which a single trusted person works alongside individuals or families with complex needs, helping them decide on their aspirations and the steps they must take to achieve them, supported by a pool of discretionary funding. Whanau Ora in New Zealand is a version of this model grounded in the Māori and Pacifica cultural context with its own governance and higher levels of transparency than mainstream programmes. Whanau Ora operates on guidance and trust, not coercion and entitlement. It is respectful of the integrity and resilience of the customers. Part of the recipe is to keep away the many agencies that might turn up, uncoordinated and unreliable, and to help the family to successfully negotiate public services. Because the approach is holistic it’s relatively simple to add a focus on economic opportunities alongside.
Whanau Ora now services thousands of families who increasingly drive the programme with their own plans and aspirations. It was fiscally and politically risky to set up, but less risky than persisting with legacy programmes and persistent failure.
There is a chance the families with Whanau Ora will do better than the families relying on the compassion of the payday lender.
For the people in our society who face the most challenges realising their potential, the monopoly welfare state has become part of the pathology of their deprivation. The current system is not a fact of the cosmos. It’s a choice with costs and benefits and the evidence is clear that it’s the most vulnerable who get the worst deal.
Social investment is a pathway to change and improvement. It applies well understood and widely used methods in a fresh way to the problem of persistent deprivation. Social investment asserts respect for the capacities and aspirations of every person and their ability to change their life for the better . That’s a purpose everyone can share.
ABOUT THE ALF RATTIGAN LECTURE
The 2019 Alf Rattigan Lecture is the fourth in an annual series in honour of G.A. Rattigan, the renowned Chairman of the Tariff Board and Industries Assistance Commission (predecessors of today’s Productivity Commission).
The purpose of the Lecture is to elevate the importance of good process, sound governmental institutions and effective leadership to achieving nationally beneficial public policy and reform.
It is delivered each year by an Australian or New Zealander who has played a significant role in promoting or implementing these ideals.
The Lecture is held by ANZSOG, with the benefit of an initial endowment by the Trans-Tasman Transparency Group and contributions from The Treasury, Productivity Commission, Infrastructure Partnerships Australia and the Minerals Council of Australia.
Previous Alf Rattigan lectures: