The Law and Order Referendum

and other related issues in New Zealand


Lock Ďem up and throw away the key....

A Cost/Benefit Analysis of long term imprisonment for repeat violent offenders based on Ministry of Justice Reports except where otherwise specified, plus the outcome of several overseas studies, taking into account the costs incurred by the victim and by society as a whole.

The cost of imprisonment is high - $51,672 per year ($73,882 per year for maximum security) per inmate. However, the contention here is that the cost of releasing certain types of inmates into the community is even higher, and so the public is actually getting good value for money by keeping those inmates in preventative detention until they either die or otherwise cease to be a threat. All the Ministry of Justice Reports which are referred to with numbers in brackets are listed in the Footnotes and can be found here

The first thing that needs to be done here is to arrive at a rigorous definition of the type of inmate to which this strategy should be applied. For the purposes of this analysis, we propose that it be limited to prisoners that have committed three or more offences defined by the Ministry of Justice as violent offences, i.e. murder, manslaughter, attempted murder, kidnapping/abduction, rape, unlawful sexual connection, attempted sexual violation, indecent assault, aggravated burglary and/or robbery, grievious/serious assault, male assault on female/child and threatening to kill or do GBH.

There may be other "offences against the person" (as defined by the Ministry of Justice) which could be drawn into this definition at a later stage.

"61.6% of offenders in prison for a violent offence have six or more previous convictions"

The next thing is to investigate the current outcomes of the justice system as it stands with such offenders. The overall pattern seems to be that we imprison them for a limited period, they are released into the community where they offend again within months, go back through the system and back into prison, are released after a somewhat longer period, offend again and so on. This is borne out by the fact that 61.6% of offenders in prison for a violent offence have six or more previous convictions (1).

Another interesting fact is that 81.8% of all offenders who are imprisoned reoffend within two years of release (2), and the rate for other forms of punishment are not much lower; 76.9% for periodic detention and 69.7% for Community programmes. Another 1998 report (3) tells us that 61% will be reimprisoned (for a recidivism rate of 77% in that year). These figures are for inmates overall, the rates for violent offenders were not specified but given the trends apparent in other data are probably even higher.

And 25.3% of violent offenders donít even make it through their bail period before reoffending. The rate for those whose previous major offence was rape, unlawful sexual connection or attempted sexual violation was even higher; 27.4%, 30.4% and 31.2% respectively.

Worst of all were those whose previous major offence was robbery or attempted robbery. They had a rate of reoffending on bail of 36.4%. And the more previous violent offences committed, the greater the likelihood of reoffending on bail; 46.8% for those with 20 or more such previous offences. All this is from Those on Bail and their Offending 1994 Chapter 4. One has to wonder what the hell we are doing giving bail to such people in the first place.

It seems obvious that if the offender cannot even make it through their bail period without reoffending, then there is little chance that they will not reoffend once released from prison. And the statistics bear this out not only here but overseas. The likelihood of recidivism in the USA was 62% across all types of offenders after three years (from US Department of Justice) and 55% after two years in Britain (from Hansard ). Bear in mind that these two statistics are not broken down for offence type, and that violent offenders have a higher recidivism rate than the overall rate.

The picture that builds up is one of a core group of violent offenders who repeatedly offend and do time as a result. Some actually end up spending more time in prison than out (such as Taffy Hotene). Having established that we have this core group, we now need to assess the cost of the current approach, by working out the likely cost to the public while they are not imprisoned and comparing that cost to that of imprisonment. To do this we need to estimate how many offences they are likely to commit while out, and more importantly their cost to society and the victim.

Overseas research indicates that at least 50% of such offenders commit 12 offences in the year immediately prior to their imprisonment on average, excluding drug crime (4). Local data gives a similar result. In 1997 we had 2528 violent offenders in prison (i.e. offenders whose primary offence was a violent one) (5). The average sentence length was 24.7 months (6), and the number of violent offences in the previous year was 16,576 (it has been around this number give or take a few hundred since 1995) (7). Therefore, given the above, it looks like each offender committed around 6.5 violent crimes on average in the year prior to their imprisonment.

The cost of the crimes committed (8) is estimated, as given in the following table (in US dollars). The first column, Direct Losses, is derived from direct loss data from a number of victim surveys. Given that in some cases victims are forced to relocate, as well as replace goods lost as a result of crime, as well as the expense and hassle of replacing ID, credit cards etc, the figures look to be an accurate assessment.

The second column, Pain and Suffering, attempts to measure the indirect costs by using data from of surveys of jury awards in civil cases where victims have sued their attackers. These indirect costs are generally by far the highest of the three, and also the hardest to measure. When you consider the cost of such things as counselling that many victims will need, plus medication, plus the intangibles that are hard to assess, these figures appear on the low side if anything.

Violent crime is a significant contributor to mental health costs. Between 10 and 20 percent of American mental health care expenditure is on victims being treated as a consequence of crime (This from here, in the Introduction)

The third column, Risk of Death, is derived by multiplying the risk of death by the valuation of a life at $2 million, a similar procedure to that used to assess traffic accident risks etc. All the figures are in US dollars.

Table 1. Average Cost of Crime to Victims

Crime Direct Losses Pain and Suffering Risk of Death Total
Kidnapping $1,872 $15,797 $92,800 $110,469
Rape 4,617 43,561 2,880 51,058
Bank robbery 4,422 10,688 3,700 18,810
Assault 422 4,921 6,685 12,028
Car theft 3,069 - 58 3,127
Burglary 939 317 116 1,372

(Source: Cohen, M. (1988) p.546)

Another more recent report, National Institute of Justice Research Report 1996; The Extent and Costs of Crime Victimization: A New Look gives the figures in the table below, as extracted from that report. These figures were derived from an in depth study carried out by T. Miller, B. Wiersema and M. Cohen, Victim Costs and Consequences, a new look, another National Institute of Justice Research Study. The NIJ is a division of the US Department of Justice. All the figures are again in US dollars.

Crime Tangible Costs Intangible Costs TOTAL Costs
Murder $1,030,000 $1,910,000 $2,940,000
Rape/Sexual Assault 5,100 81,400 86,500
Robbery/Attempt with Injury 5,200 13,800 19,000
Assault or Attempt 1,550 7,800 9,350
Burglary or Attempt 1,100 300 1,400

We can now put together a formula that describes this situation,

Cr = v+j+b+p

Cr is the cost of release per year

v is the cost to the victim(s) of offences committed while released from the table above

j is the costs incurred of running any resulting court cases

b is any benefits paid out to the offender while released

p is the cost of probation or supervision

For most repeat violent offenders, Cr>Ci where Ci is the cost of imprisonment, a constant having a value of 51,672 or 73,882 for a maximum security inmate. The situation can also be described graphically as below. The vertical axis is obviously enough the cost to society of whichever course of action is chosen to deal with an offender.

The horizontal axis, Seriousness of Offending, is somewhat more complex, being the number of offences multiplied by the average severity. This scale is one devised by the NZ Ministry of Justice, and can be found scattered across their reports. To give an idea of how it works, such offences as burglary are assigned a seriousness usually less than 100. Lesser traffic infringments are usually less than 50. On the other hand, assaults on the person rate at 300 and up.

There comes a point where itís cheaper to simply put an offender away and not release him, and as demonstrated in the following example, that point is surprisingly low. Let us envisage a hypothetical violent offender and assume that they commit six offences, two being assaults, and err on the conservative side and say that the other four are burglaries.

We will also use the generally more conservative figures from the earlier Cohen study. The total cost of releasing such an offender upon the public will be $29,488 approximately in damages to the victims, the components of which are $12,028 each for the two assault cases, and $1372 for each of the four burglaries. Here we make a further conservative assumption, that the US dollar is equivalent to the NZ dollar - if only!

We now have a cost to the victims of $29,488. Now assume that only 50% of these actually result in a conviction and prosecution (9), and that these comprise the two assaults and one of the burglaries. This is not an unreasonable assumption as assault victims are usually fairly motivated to see something done about prosecuting the offender and are often able to identify the assailant, whereas burglars are much less often seen and therefore apprehended.

We now need to add to this the cost of the court cases which we can derive from data found in the Department for Courts Annual Report 1999 (Requires Adobe Acrobat or equivalent PDF reader) on pages 59-61. In the year 1998/1999, there were 2340 District Court jury trials, at a cost of $18,118,000, giving a cost per case of $7307.

We have left out the cost of High Court cases, of which there were 394 jury cases and 972 appeals. These tend to cost even more. Now, given that assault cases generally result in jury trials (an assumption supported by the fact that there were 2317 cases involving violence resulting in prison sentences in 1998), we now have court costs of $14,600 for the two assaults.

If we were to assume that the burglary merely results in what is called in the Report a "criminal summary jurisdiction - police information", of which there were 269,330 in the period 1998/1999, at a cost of $33,048,000, this gives a court cost of $122 (probably a lot more than that in reality too). Add this to the court cost figure of $14,600 above, and we arrive at say a round figure of $14,700. Add to this the cost to the victims of $29,488, and suddenly this offender has already cost us over $44,000 to release per year.

Add now to this the cost of a benefit. Realistically, it is extremely unlikely that such an offender will be in paid employment, as overall only 35.2% of offenders cited paid employment as their source of income prior to entering prison (12). It would be reasonable to assume that the odds would be even lower for the average repeat violent offender. Given that the average benefit will be roughly $200 per week, this makes another $10,000 per year our hypothetical scumbag is costing the taxpayer.

We are now up to $54,000 per year..

We are now up to $54,000 per year give or take. Add now the cost of probation for which we could not find a cost figure but assume that it will be the same as that for supervision, $2000 per year (13) and we are at $56,000. Hmmm.... where did our "savings" from not putting this lowlife in prison get to I wonder...?

This is a purely hypothetical case, but demonstrates a point. Imprisonment is expensive, but it is still cheaper than all the alternatives, at least for repeat violent offenders. Our hypothetical scumbag is moreover in the lower range of offending and will be close to the crossover point on the graph. Now take a look at that table of costs to victims above, and work out the outcomes where just one rape is involved. The cost to just one victim outweighs the cost of a yearsí imprisonment, without factoring in any of the other costs.

Of course many of the above costs are estimates, but if anything they appear to be somewhat on the low side, at least those incurred by victims, going on some of the comments from The Economic Analysis of Criminal Justice Policy Options Report 1997 surrounding Mark Cohenís and other analyses. Costs such as those of higher insurance premiums, and those of such security measures such as alarms, monitoring, guard patrols, etc, that are borne both by individual victims and by society overall, have not been factored into this analysis either.

And they do not fully take into account the intangible losses incurred by families and friends of victims, something which seems to have completely escaped the likes of Matt Robson also. In fact, he appears to have completely missed out something altogether when calculating the cost of releasing inmates - factor v, the cost to victims.

There are of course a number of situations where it does not pay to imprison offenders. Drug offenders are a case in point; it is usually cheaper to treat the addiction and the causative factors behind it, than to imprison them. Given that in and of itself drug use is victimless, there is of course no cost v to victims, unless it is as a result of other crime committed in order to obtain drugs. Therefore the cost of release of such offenders to society tends to be less than that of imprisonment.

This was recognised by American voters in some states, where referendums proposing treatment rather than imprisonment were held along with the recent elections. Overwhelmingly they voted for treatment rather than imprisonment for drug offenders.

Another example is the situation where someone commits a murder or other serious crime under extremely unusual circumstances that clearly mark it as a one-off type offence, or confesses to a crime after many years of law abiding behaviour. In both situations the liklihood of further offending is very low, and therefore the cost of release will be low, as not only will there be no further cost to victims, but there will also be no further cost within the justice system.

Such offenders are also far more likely to be employed and paying tax, and so it makes far more sense for them to be released where they can continue to contribute to society and perhaps compensate the victim of their one-off crime where appropriate.

We also suspect that the model above has not fully taken into account the administrative savings to be made by keeping inmates in, rather than shuttling them in and out on a regular basis. Potential also exists for cost savings on the current cost of imprisonment, something not examined here. Unfortunately some of the costs, such as those for probation/parole, appear not to be available anywhere, so had to estimated as best as we could. If anyone is able to provide more accurate costs in these or other areas, or good data on the costs of various crimes to NZ victims, we would be most grateful.

As it happens the fact that it is more cost effective to lock serious violent offenders up permanently has been borne out by research by the US National Institute of Justice in the mid 1980's. The following is a report from the Otago Daily Times on this on July 5th 1988

Cheaper to gaol repeat offenders.

Washington (AFP). - It's cheaper in the long run to American society to build new prisons than to relieve overcrowding in gaols by releasing repeat offenders, a US Justice Department study says. The cost of building a new cell and maintaining a prisoner in it averages US$25,000 (NZ$36,207) a year, says the study, which was conducted by the departments National Institute of Justice.

On the other hand each released repeat offender costs society an average US$430,000 (NZ$612,823) a year if one adds up the damages suffered by the victims, the costs of police and court work and security expenses. The Study of 2,190 inmates in California, Texas and Michigan said each released repeat offender committed an average of 187 crimes a year. Each drug purchase or sale was considered as a seperate crime. The confinement of 1,000 repeat offenders a year would thus cost about US$25 million (NZ$36.21 million) a year but would avoid 187,000 felonies costing society a total of US$430 million (NZ$622.77 million) .

"Confinement is not too expensive when weighed against the price of crimes that would otherwise be prevented by incarceration" said institute director, Mr James Stewart. The study concluded that the US$8.6 Billion (NZ$12.45 Billion) cost of operating United States prisons in 1983 was one-tenth the costs of crime to society.

Otago Daily Times - July 5, 1988.


  1. from The Use of Imprisonment in NZ 1998 Chapter 3: The Criminal Histories of Sentenced Inmates
  2. from Appendix II of Sentencing in NZ: a statistical analysis 1999
  3. from the 1998 Responses to Crime Annual Survey
  4. from The Economic Analysis of Criminal Justice Policy Options Report 1997- Chapter 6.3 Piehl and DiIulio
  5. Census of Prison Inmates Chapter 4 Offending History
  6. Conviction and Sentencing of Offenders in NZ 1989-1998 Chapter 3.5
  7. Conviction and Sentencing of Offenders in NZ 1989-1998 Chapter 2.5
  8. from The Economic Analysis of Criminal Justice Policy Options Report 1997 Chapter 6.2 Mark Cohen
  9. in line with the outcomes in the Annex to Responses to Crime: Annual Review 1999
  10. Annex to Responses to Crime: Annual Review 1999
  11. Responses to Crime: Annual Review 1998 Chapter 2.4
  12. Census of Prison Inmates 1997 Chapter 11 Source of Income Prior to Entering Prison
  13. from 1999 Review of Community based sentences

Back to Sentencing
Home Back to Top