Speech to the Third Data Hui
E ngā mana, e ngā reo
o ngā iwi o te motu,
o ngā tangata Pasifika
me ngā roopu hei āwhina i te Pani me te hunga Rawakore
tēnā tatau katoa
Aku hoa o te Whare Miere
me ngā Kaiwhakahaere, ngā Kai-takatu o ngā Tari o te Kawanatanga
Tēnā koutou katoa.
He huinga tangata he whakaaro kia rātau kua wehi atu ki te pō,
Te hunga mate ki te hunga mate
Te hunga ora kua tau mai nei,
ki te hui tuatoru a nga Whakaaturanga,
Tēnā tatau katoa
Can I acknowledge Dame Diane Robertson, chair of the Data Futures Partnership, John Whitehead, from the Partnership Government Statistician Liz MacPherson, Brendan Boyle, Chief Executive of the Ministry for Social Development who we just heard from, and my Parliamentary colleagues joining us here today.
It’s fantastic to see such a great turnout here for a topic which a year ago none of us would have thought that much about.
But let us first remind ourselves why we’re here today.
We’re here because we need to change more lives.
I’ve just been through a Budget process, energised yet puzzled by the process which gives us the privilege of allocating billions of other people’s money.
Money that is hard-earned week-by-week. The GST that is paid to us on the groceries.
We owe it, at least to the tax payer, but absolutely to the people needing our help, to use every tool available to change lives.
Lives which are described by the data.
The data is just a story.
Brendan referred to the 10,000 most challenging of our community who will cost us $6.5 billion over their lifetime.
That’s 10,000 stories. 10,000 people who have over the last 30 years been the focus of endless discussion, multiple measurement, and beautifully-launched strategies – but their stories tell us that hasn’t changed much.
They find themselves in lives that, for 20 or 30 years, are spent on the lowest income you can have in New Zealand trying to deal with the most complex family and personal challenges we can imagine.
This is the context for our discussions today on data.
While we’re talking about some of the things that we find challenging, we should just keep in mind that by comparison our challenges are easy.
Our problems are in talking to another government department, whether we’ll share some data. Others would love to have something that easy to deal with.
This is the third Data Hui following events in September and December, and the Government has been a consistent advocate of increasing transparency through information sharing.
And we’ve made good progress over the last few years.
You’ve heard some of that this morning from the Chief Executive of MSD, insights largely unknown until only recently.
Isn’t it amazing how much we didn’t know about the people we told everyone we cared for?
We now have a much better understanding of our most challenging customers and a partial view of their journey through our system.
That understanding is starting to translate into results.
Government actions have reduced the welfare system’s future lifetime cost by $12 billion in just four years.
That equates to 60,000 people each spending 15 fewer years on benefits.
These are real achievements, but we’re only just beginning.
We’ve found that greater openness delivers a range of benefits.
It increases trust between government and the public.
And it rewards governments for successfully getting on with business.
For example, the Treasury now publishes a Major Projects Report which rates the performance of projects covering $18.5 billion of spending across government.
This level of transparency, just on our projects, not people, has helped to sharpen the focus of public servants and Cabinet.
The uncomfortable fact is that, despite billions of dollars of spending on social services, we don’t have that degree of transparency in social spending.
We need lift our effectiveness in helping people get back on their feet after setbacks.
As a Government, we’ve found that being as honest and straightforward as we can with the public, and with ourselves, about that uncomfortable fact means the conversation can focus on real solutions for real people.
It’s always tempting for any government to throw more money into a system that isn’t working.
What’s difficult is to sit down and attempt to understand the drivers of poverty, or to unpick the downwards spiral, which we too often assume is inevitable, or to adjust systems so that we can actually change the lives of those people in the greatest need.
So today is about taking the next step on the path to improved sharing and use of data.
A reason there is such a broad range of people here is because the problems are complex and the solutions also sit outside the public service.
Today is also about recognising the opportunity to build on the progress that has already been made by Iwi, NGOs and Pasifika in the use of data in their daily operations.
As a Government, we are guided by a simple and obvious principle.
We’ll do what works.
We welcome the accountability that goes with being measured publically on results, for example through the Better Public Services targets – still unique in the developed world.
And we welcome the public’s rising expectation that we will use their money effectively.
Data sharing is an essential part of this.
If we are going to understand what works, we need to know what services went to whom – something we don’t yet know.
We need to know whether those interventions changed outcomes and by how much – something we’re only now starting to see.
And we want to understand the long term fiscal and social consequences of those changes.
We then need to use that information to inform where to invest next for better measurement, evaluation and feedback.
We’re all becoming familiar with the language.
Services that take the time and efforts of vulnerable individuals and families for no improvement in outcomes reduce resilience and perpetuate the hardship cycle.
One reason to stop things that don’t work is it saves Government effort and tax payers’ money.
A better reason to stop things that don’t work is because it wastes the time, effort and resilience of those for whom we are trying to help.
All too often, officials tell me there isn’t the capability outside the public service to make better use of data to show impact.
I disagree, and the presentations later this morning will show you why. That perception is certainly changing within the public service. And that capability is going to lift.
We can harness and build on the progress already being made outside the public sector to measure, evaluate, and feed back into decisionmaking.
My hope is that officials here today listening to the presentations from service providers later this morning will continue, as they are now, to ask themselves how we can expand the use of these technologies.
Government departments hold enormous stores of information that are directly relevant to understanding what services work.
But too often that information sits in silos within departments. It is often difficult and sometimes impossible to access.
Those limits on access are used to protect the state’s monopoly on the control and use of data.
Today is about further breaking down that monopoly.
By analogy, what we have at the moment is a warehouse stocked full of food.
The door is small, it’s a bit hard to find, and when get to it you’ll only get the food you’re given.
But what we really want is a supermarket.
The kind of supermarket with a door that is easy to find, and that has rules for entry that are strict but clear.
Once you’re inside this supermarket, you can rely on the quality of food that’s on the shelf, and you get to choose what you want.
And on the way out, you’re not going to be stopped by security second-guessing what you’re going to cook when you get home.
Because that’s what we currently do. We say “you can’t have that data because you might misuse it”.
Access to data shouldn’t be the exclusive reserve of government – but that’s what it largely is because in many cases access is being decided in an ad hoc fashion.
So today we’re committing to improving access.
Iwi, NGOs, and Pasifika – many of you here today – have told us getting information out of departments is not easy.
It’s a negotiation. Agency by agency. Official by official.
You’ve told us contracts are entered in to as if each negotiation was the first, with the each negotiation’s success depending on who you are talking to.
But the officials you’re working with are acting consistent with their environment. They are operating in many cases without clear rules on appropriate data sharing.
Sharing brings costs and risks for agencies, including protecting privacy, but few rewards.
It can be a low priority for agencies who feel they have limited resources and have to assertively prioritise.
So officials are reacting in quite reasonable ways to the circumstances they are confronted with.
The government has earned the right to collect and hold data because it has the trust of the majority of citizens. This is our social license.
Up until now, we have maintained that social license by saying “no” to many, perhaps most, requests for external use of that data.
But “just say no” as an approach to data security has meant we have made only limited use of all the data we have gone to the trouble of collecting.
Data has no value if it is not used.
So let’s fix the system.
Let’s reverse the presumption and make data sharing the norm rather than the exception by clarifying the rules.
We need to reverse that assumption, because actually the data belongs to the people it is about.
It belongs to the people and the whanau that you are working for and with.
Some of you here today have spoken of data sovereignty.
This is the idea that you, as citizens, and collectively as Iwi, Pasifika and NGOs acting on behalf of citizens, own the data held by public agencies.
We agree with you. This is not a novel idea.
In fact, that principle is embedded in statute. Principle 6 of the Privacy Act says:
Where an agency holds personal information in such a way that it can readily be retrieved, the individual concerned shall be entitled… to have access to that information.
I’ll say that again – “the individual shall be entitled”. That’s the law.
So I look forward to the day when we will find a very long queue of citizens outside the warehouse who have figured out that the cost of retrieval is about zero and you shall provide them with that information.
It’s just a matter of time.
So agencies are subject to a statutory obligation to share the information it holds about and are obliged to share it when you – or somebody authorised by you to act on your behalf – asks for it.
We’re making changes to the system to make good on that statutory and policy commitment. Ongoing workstreams include:
- The Data Futures Partnership, which is developing guidelines for data sharing through conversations around the country to discover the boundaries of social license, as well as launching catalyst projects for data use.
- We’ve invested in the Integrated Data Infrastructure in Statistics New Zealand, eliminated most fees for use, and established the Analytics and Insights unit in the Treasury to publish insights from the IDI including its recent insightful work on at-risk 0-24 year olds.
- We’ve developed the Cost-Benefit Analysis tool, or “CBAX”, through Treasury, a template for lifetime cost-benefit analysis. This was fully incorporated for the first time in the Budget process this year.
- In February we launched a Social Investment Insights tool, a platform providing a point-and-click access to selected IDI data.
- And we’re reviewing the Statistics Act and the Privacy Act in this context.
I want to talk about another recent change, which is central to the presentations you’ll be hearing later today.
In the middle of 2015, we established the Social Sector Investment Change Programme.
It’s now the Social Investment Unit, and it is currently building a data highway that will be integral to the social investment system.
This data highway will securely connect all participants, including departments – Health, Education, Social Development, Justice, Inland Revenue and ultimately all parts of government – allowing information to be shared and joined under a federated data model.
Each department will be a node on this highway, and their data will be accessible through it.
Our vision is that service delivery providers, including many in this room, will be able to connect to a prototype of this data highway in the near future.
It will be a one-stop shop for access to identified data held by departments, subject to the informed consent of clients and to standards set to achieve compliance with all relevant legislation.
We will call it a Data Access Service.
Those standards will be applied consistently, so departments and service providers can be confident of access provided they meet those standards and hold authorisation for sharing.
Importantly, access will not be subject to veto by departments who may not agree with the findings drawn from the data.
Data will be interoperable, that is, joined up across agencies. Iwi, Pasifika groups and NGOs holding suitable authorisation from their customers will also be able to match data they hold with the data held by agencies.
For social service providers, this will make data access professional, routine, reliable and responsible – rather than an unpredictable negotiation.
And by levelling the playing field on access to government data, subject to appropriate controls consistently applied, new models for social service delivery will emerge.
Models that are more effective in changing lives.
Access rules will be informed by the work of the Data Futures Partnership, as well as the Privacy Commissioner, and the Government Statistician.
Improving access to data amounts to the democratisation of data in an orderly and safe way.
Government holds no unique place in the use of data, even if it is its largest repository. There is no government monopoly on good ideas.
Increasingly, the currency in the delivery of social services will be new and better ideas.
Open access to data will in-effect create an information logistics platform that will be used in ways we can’t yet envisage.
I must emphasise that increasing access to data will not be at the expense of security or risk to privacy.
Quite the opposite – we are improving access by clarifying the rules around sharing, and applying those rules consistently.
Let me also say to agencies here today – we are doing this.
There will be sharing of the data under your control which has not been previously shared.
It is neither an agency’s right nor its responsibility as a data steward to decline to share that data for fear of misinterpretation or of findings you might disagree with.
If you’re worried about misinterpretation, then clarify – but don’t withhold.
We’ll be working with you as we move into this new and, at times, uncomfortable territory.
Improving access to data beyond this new service will involve considered judgment about two issues in particular on which the Government does not yet have a position.
The first is around secondary identification, which occurs when data that has been anonymised by removing names and addresses can nevertheless be used to identify individuals from combinations of data points.
The problem is that it is hard to anticipate all the ways secondary identification can occur.
And so, without legislative protections post-release, agencies understandably respond by being aggressive in sanitising data before putting it into the public domain, for example by requiring high levels of aggregation that will in many cases remove much of the data’s value.
Statistics New Zealand has told me that using their normal procedures for making anonymised data safe from secondary identification will frequently produce empty data where the populations of interest are small.
And, from a provider’s point of view, that is almost always the case.
Legislation setting penalties for secondary identification could enable improved access while protecting the public’s confidence in appropriate use of data.
This proposition is yet to be thoroughly tested by the Government, and we have arrived at no position on it at this stage.
A second way to support social license for improving data access is by establishing citizen-level control of how data is used.
One version of this is called Personal Information Management Services, a technology that gives citizens the ability to opt-in or opt-out of sharing some or all of their personal records.
In effect, it shifts decisionmaking on the use of data from the steward to its owner.
This technology exists now, and it promises a future in which the problem of discovering the boundaries of collective social license is replaced with a system that responds to each of your individual preferences for sharing.
It will also enable the creation of data intermediaries who, working within the legislative boundaries for data use and the permissions granted by individual citizens, will be able to find and distribute information for like-minded or like-behaving people for social services providers.
Again, this proposition has to be thoroughly tested.
Finally, let me close by restating what today’s hui is about.
Today we learn about the technologies supporting measurement, evaluation and feedback that are already in operation among Iwi, NGOs and Pasifika providers of social services.
As Government, we want to build on these achievements rather than replicate them.
Today is also about committing to further improvements in the sharing and use of data on a level playing field.
We want to do this by clarifying rules protecting privacy, applying those rules consistently, and sharing data except where those rules prevent it.
So welcome to the Beehive, thank you all for taking the time to be here today as we take the next step in improving the sharing of data.
That is – your data.