Using data to fight crime

  • Amy Adams

Good afternoon and thank you for the invitation to speak at the Government Economics Network today.

For as long as can be remembered, Governments have used data to inform economic policy.

Forecasts and robust modelling sit comfortably alongside fiscal policy. Detailed economic data stretching back into the nineteenth century has helped inform future economic policy and its impacts.

From issues such as managing inflation, regulating labour markets and auctioning off rights to the radio spectrum, Ministers and policy makers have enjoyed the benefit of sophisticated and rigorous analysis.

The same has not always been true of social policy.

Too often, anecdotal advice is more prevalent and the data much more limited.

However, I’m pleased to say that with the Investment Approach to Justice, this is changing. And it’s changing dramatically.

As the Justice Minister, one of my core jobs is to analyse how well the system is working to reduce the burden of crime on our communities, and search for continual improvement

While not naïve enough to think that any iteration of legislation, policies, processes or programmes would end crime, I absolutely believe that we can, as a society, have an enormous impact.

In order to do this, we have to understand the crime that is occurring, the relative harm it causes, the offenders involved and what precipitated their behaviour. 

Then we need methods of understanding the impacts various interventions we can make have on particular crime and particular offenders.

I’m a firm believer in the value of using the vast amount of data that exists to ensure justice policy is as well informed as it can possibly be, and to underpin the business case for making what we term social investments.

This in its most basic form just means investing ahead of demand to reduce future demand. 

All this makes perfect sense but to make that case we have to be able to show with a high degree of likelihood what will happen without intervention and the impact the proposed intervention will have. That can only happen with well-developed, robust, data modelling.

And this is where the Government’s Social Investment Approach comes in.

What is the Investment Approach

The Government’s Social Investment Approach is about improving the lives of New Zealanders by applying evidence-based investment practices to social services.

By using information and technology to better understand those New Zealanders requiring public services and adjust services to fit better, we understand what services or combination of services work best.

I understand you’ve already heard from Finance Minister Bill English today on the topic.

Thanks to the Integrated Data Infrastructure, our social data is now rivalling our economic data for depth and breadth. With the Investment Approach, we are using this data to support more sophisticated and effective policy design.

The purpose of the Investment Approach to Justice is to boost crime prevention and reduce harm from crime in New Zealand communities.

The Investment Approach to Justice is made up of two parts: working out who is experiencing what crimes and where; and then what initiatives works to prevent it.

The first component is a micro-simulation model based on the Integrated Data Infrastructure. This model takes every person in New Zealand and predicts how many times they are likely to commit an offence or be victimised over the rest of their life.

For example:

  • This model shows that in the case of a nine-year old boy who has had CYF intervention and whose family is supported by benefit, we can expect 3.3 offences over the next 15 years.
  • And in the case of a 34-year old man under management by Corrections in the community, we can expect 2.9 offences over the next 15 years.

This helps us understand how much crime we could prevent by focusing on different groups of people.

The second component of the Investment Approach is research to understand what works.

For example:

  • we know that for the nine-year old boy I mentioned, interventions to improve the quality of the parenting will lead to him offending three fewer times
  • And for the 34-year old man, he is more likely to offend again if given a community work sentence rather than a fine.

This helps us understand what we should do to help change the trajectory of these people’s lives. 

This is why the Justice Sector has committed $2 million to build exactly that sort of actuarial model.

It doesn’t just test justice sector interventions, but it will be capable of assessing and informing decisions right across the social sector in terms of the impact they will have on future offending. 

The final step is using this information to take better decisions.

It will help Ministers decide what to invest in. And it will help police officers, social workers, probation officers and others determine who to focus on and what can be done to help.

Officials are working with the Social Investment Unit and others to make these insights broadly available.

New insights about at-risk cohorts

The most powerful use of this data involves putting New Zealanders at the centre of our thinking.

Every single data point in the Integrated Data Infrastructure is about a real person, a real story.

It doesn’t take much time looking at the Integrated Data Infrastructure to see the truth of that.

For example, we have been looking closely at two groups of New Zealanders where the model shows us they are highly likely to offend again.

The data clearly notes that the offences committed by these people are just a subset of much deeper social problems.

What the Investment Approach does is bring hard numbers to the long-held “truths” people working in the Justice sector have intuitively known.  

14-16 year old serious offenders

Firstly, I want to touch on the 1800 New Zealanders between ages 14 and 16 who have committed a serious offence such as assault with intent in the past two years.

On average, these 1800 young people have already been dealt with by police for six offences each.

Well over 10,000 victims owe their experiences to this small group of young people.

If we continue to do what we’ve always done with these young people, they are predicted to commit 10 further offences each over the next 30 years.

This would include 2800 serious property offences and 1600 serious violent offences.

We can break this down further.

We know that 65 per cent of this cohort is Māori and 79 per cent is male.

We also know that 62 per cent have a parent with a history of Corrections management and eight out of ten have previously been notified to CYF for a care and protection matter.

58 per cent have been stood down from school twice or more or suspended and 45 per cent have accessed mental health services.

Half of this cohort is the child of a beneficiary, and half are in a sole parent family.

Never before have we had such insights available to us.

In the past, our justice data would have shown us the young people solely as offenders. The information we had about them was limited largely to what courts they attended and what orders they were being given.

Now we can use data to look behind their offending into the root causes, and what we can do to help them from continuing down the path of crime.

We are creating models that are capable of assessing and informing decisions right across the social sector in terms of the impact they will have on future offending. 

And this is allowing us to know where to focus for maximum impact.

Some of the learning we have already is instinctive, other parts are counterintuitive. 

For example, treating young offenders as adults, or simply trying to scare them into submission, actually increases reoffending. This creates more victims and a higher cost to the Crown. 

Our data now challenges us to look at the broader social environment of these young people, and other opportunities where we can intervene.

We know from international research that if schools can manage the behaviour of young people and keep them in school, that this reduces offending, increases educational achievement, and leads to better fiscal outcomes.

We can use this information to better target efforts to support these young people, as they are concentrated in certain schools, in certain neighbourhoods – and indeed many of them will be friends with each other.

For example, many of the serious young offenders in Auckland go to just 10 schools, and are concentrated in areas of the city such as the eastern and southern parts of Auckland.

It might not occur to New Zealanders from the outset that these schools play an important role in supporting crime reduction.

20-24 year old serious property offenders

The other cohort I want to talk to you about are the 2000 New Zealanders between ages 20 and 24 who have committed a serious property offence such as burglary in the past year.

From the data we can now say that this cohort has already committed 12 offences on average.

Over the next 30 years, this cohort is predicted to commit 15,300 further offences – an average of eight per person.

67 per cent of this cohort is receiving a main benefit and that together they are costing the Government $191 million in welfare benefits. 57 per cent of this cohort is Māori and 82 per cent of the cohort is male

73 per cent has been under Corrections management in the past year.

And 39 per cent of this cohort has used mental health services. Much of this was for substance abuse problems.

But more than just knowing factors that have led to their offending, we know from research what works to address it.

Among options to reduce reoffending by this cohort, the strongest evidence is for cognitive behavioural therapy, alcohol and drug treatment, restorative justice, reintegration services, and prisoner education and employment.

Big picture

As I’ve outlined, the Investment Approach is helping us to understand high-risk cohorts in detail.

It can also help us to better understand the big picture of crime.

We have completed a segmentation analysis that places everyone in New Zealand into one of either the high, medium or low risk groups.  

The segmentation analysis has shown us that the two per cent of people that are at highest risk of offending are predicted to commit 27 per cent of offences next year.

However, nine per cent of people in the medium risk group are predicted to commit 40 per cent of offences next year.

This challenges us to think differently about crime.

The Justice sector can work with high-risk people, but these results emphasise that crime is everyone’s problem.

Next year, only 22 per cent of crime is predicted to be committed by people under the management of Corrections.

Most medium-risk people are not in contact with the Justice sector, they are out in our communities.

If we are to stop these people from creating victims, the solutions will need to lie in schools, in the mental health system, in local councils, the welfare system and elsewhere.

National Youth Dataset

On the subject of better data, we have a new National Youth Dataset that will allow in-depth measurement of key trends across the whole youth justice system.

After 14 years in the pipeline, we now are able to assess young people and children entering, moving through, and leaving the youth justice system.

The Dataset collects and compares information from across agencies about youth justice and allows us to dissect in new ways.

While agencies have always collected information, prior to the Dataset this information was not always comparable. This made determining trends difficult – if not impossible.

Now a wealth of this data and information is at the fingertips of decision-makers.

For example, we can now use data about serious and persistent young offenders to get a better idea of the numbers who have complex needs, and who are at risk of becoming adult offenders.

This is intuitive for those of you working in the sector, but we now can prove it.

We can also increasingly drill down into data to see what types of crimes those kids are committing or if there are differences between ethnicities and in different areas around the country.

It’s telling us that not only have the number and rate of young people in court decreased, but we’ve also been successful in reducing the number of youth with complex needs or persistent offending.

This type of information helps provide a clear, shared picture of how well we’re achieving our goals, and contributes to our understanding of how we should focus our efforts.

The Dataset is a work in progress and my officials are still determining how we can best use the data to enhance our work across the youth justice sector, but I expect it will have a significant impact.

The future of the Investment Approach

The work we have completed to date has only just scratched the surface of what is possible with these powerful new tools.

There are four areas where we want to focus on, which are evaluating investments, managing the pipeline, supporting the frontline and supporting non-governmental organisations.

Managing investments

The Government invests about $500 million a year across Police, Justice and Corrections to prevent crime, and much more in MSD and elsewhere.

In many cases we don’t know if this spending is effective so we need to evaluate our investments accordingly. We need to substantially increase the number of evaluations we undertake so that we can identify programmes that work and expand them, and stop or redesign programmes that don’t work

Managing the pipeline

As the prison population continues to grow, we need to do more to stop people’s offending careers before imprisonment becomes the only option. We have looked at two cohorts this year, but need to look at many more to help reduce future pressure on the pipeline.

Supporting the frontline

Police officers, judges, the parole board and others take important decisions every day about who to warn, who to sentence, who to release, and what services to offer. The data analytics we can produce hold the potential to support better decisions at the frontline that reduce crime and reduce cost. For example, with an automated tool that provides judges at the time of a bail decision with a prediction about the risk level of the offender.

Supporting non-governmental organisations

Finally, many of the services that can reduce crime are delivered or funded by iwi organisations, philanthropic trusts and non-government social service providers. 

There is an appetite in these organisations for the kind of data and evidence we are producing. We aim to work closely with these providers to help them make their services more effective.


The Investment Approach is about addressing the core issues to stop crime happening in the first place.

With the Investment Approach, we can see clearly in the data exactly who we need to work with in order to make the biggest difference.

We can also see clearly in the data that crime is everyone’s problem. The Justice sector can’t do it alone.

We can see in the data that when an employer is considering a job application from an ex-con – that is an opportunity to reduce crime.

When a council is considering how much to invest in lighting within the CBD – that is an opportunity to reduce crime.

When a GP is talking to a patient with an offending history and a substance abuse problem – that is an opportunity to reduce crime.

Just like economic policy, the Justice Sector is being equipped with the tools and information it needs to reduce the burden of crime on New Zealanders. Big data is making this possible.

We are helping identify the areas of greatest need where the Government can best direct our efforts, and do more of what works.

We are working to make sure as many of these opportunities are taken as possible, to reduce the number of victims.

Thank you.