A crime "crystal ball" is using big data to estimate the probability of New Zealanders committing or being victims to crime.

Cutting-edge computer data modelling is tapping into a powerful IDI database of Government information, which provides data from the tax, education, benefit and justice systems.

It maps the probability of New Zealanders committing or experiencing crime over their lifetime.

They are then assigned to a group - "vulnerable adults", "career criminals", "petty criminals", "at-risk young people", "vulnerable children" and "not at risk".

Advertisement

The data is anonymised - officials are not working out how likely certain individuals are to commit crime in their lifetime.

Rather, the work is useful because it can give some idea of how many New Zealanders are at risk of committing or being victim to crime - and estimate the total future burden of crime on society.

The actuarial-type model - developed by PricewaterhouseCoopers - can then be used to estimate how that burden would change if more money is put into certain initiatives or policy.

A diagram explaining the work included in a recent four year plan for the criminal justice sector included a depiction of a crystal ball.

Early work has resulted in judges being told that in certain cases a fine could be a better option than community work, after analysts found criminals getting the latter were more likely to reoffend and rely on the dole.

Offenders given community work were found to be 4 to 7 percentage points more likely to be reconvicted within two years, compared with offenders who were fined.

Justice Minister Amy Adams said the investment approach aimed to prevent people from being victimised in the first place.

"The model has used births, census and deaths data on the IDI to estimate the New Zealand population. As more data becomes available in the IDI, more detailed analysis can be carried out in order to better understand the risk and protective factors of offending and victimisation.

"The statistical model is able to predict future offending and victimisation for groups of New Zealanders. The model is also being used to test the return on new investment in high-risk cohorts. These statistics are not yet available but we would expect to be able to release information later this year."

Labour's justice spokeswoman Jacinda Ardern said evidence-based policy-making was something to aspire to.

"For me the most important question is, what are we doing with that information? Are we then looking at adopting practices both in criminal justice and the social sector that we know work and know make a difference.

"Modelling who is at risk is one thing. Doing things differently and then addressing that risk is something completely different."

The prison population has increased at a rate much higher than forecast in the past two years, and soared past 10,000 in November. It has been driven by tougher bail laws, stricter sentencing and repeat offending by drug and alcohol addicts.

The Government plans to spend up to $2.5 billion on new prisons and expanding existing ones over the next few years.

In 2012 "better public service" targets were set to be met by June this year. The Justice sector has made big reductions in total crime (down 14 per cent since 2011) and youth crime (down 32 per cent).

However, violent crime and the re-offending has edged upwards in the past year, and have only reduced by 2 per cent and 4 per cent respectively since 2011.

Treasury has also used the Statistics NZ IDI database and has released research that highlighted indicators that increased the likelihood of children suffering hardship later in life.

Prime Minister Bill English has championed that work and has appointed Adams to the new role of Minister Responsible for Social Investment.

How it works

• A model has been built that taps into a powerful database of data from the tax, education, benefit and justice systems.

• The model estimates the probability of New Zealanders committing or experiencing crime in their lifetime.

• It separates the population into groups with different levels of risk, such as "vulnerable adults", "career criminals", "petty criminals" and "vulnerable children", and estimates the "total future burden of crime" on society.

• The model can simulate how the burden of crime might change if investment is made in certain initiatives or policies.