Sentencing in New Zealand: a
statistical analysis
Foreword | Acknowledgements | Executive Summary | Introduction | Chapter 2 | Chapter 3 | Chapter 4 | Chapter 5 | Chapter 6 | Chapter 7 | Chapter 8 | Chapter 9 | Chapter 10 | Chapter 11 | References | Appendix I | Appendix II | Tables | Figures
10.1 Background
The monetary penalties included in this analysis are the fines and reparation imposed by the courts where the monetary penalty is the most serious sentence in the case. Reparation to the victim can be imposed whenever an offender has caused any loss or damage to any property or, since 1987, where the victim has suffered emotional harm.
Orders for court costs in the absence of a sentence are not included as a monetary penalty in this analysis. This study is limited to cases involving imprisonable offences. Just under a quarter of the total proved cases in 1995 were for non-imprisonable offences, for which the maximum penalty is a monetary penalty. Also excluded are monetary penalties not imposed by the court, such as fines resulting from unpaid infringement fees for speeding and parking infringements.
This chapter examines which statistical factors most influence the use of monetary penalties, looking first at the variables individually and then at their combined effects using multivariate modelling techniques. Changes in the use of monetary penalties are analysed in the final section of this chapter.
10.2 Single variable analysis of current factors influencing sentencing
The relationships between the probability of receiving a monetary penalty and the various statistical factors (Figure 10.1) are almost exactly the opposite of the relationships for imprisonment (Figure 4.1).
Offenders who have committed an offence of low seriousness have a far higher probability of receiving a monetary penalty. Almost 60% of cases involving an imprisonable offence with a seriousness score of less than 20 result in a monetary penalty, compared to less than 20% for all other offences. The probability of receiving a monetary penalty also drops rapidly as the number of proved charges in the current case increases.
Similarly, monetary penalties are more likely for people who have a limited offending history, especially those with a low rate of conviction or less serious previous convictions. There is a negative, more or less linear, relationship between the probability of a monetary penalty and the number of previous cases. If the person has not had a conviction for some years the probability of receiving a monetary penalty increases.
Figure 10.1: Percentage of offenders receiving a monetary penalty in 1995 for selected variables
Women have a lower probability of receiving a monetary penalty (35%) than men (40%), despite having characteristics that should make them more likely to receive a monetary penalty, such as the lower average seriousness of the offences committed by women. Youth offenders also have a lower probability of receiving a monetary penalty (11%), compared to adult offenders (37-41%). Monetary penalties are used less for Māori (28%) and Pacific peoples (30%) than for Pakeha/Other offenders (42%). Some of these differences are due to the characteristics of the different offender groups. The multivariate analysis presented below takes these factors into account.
As expected, the offence groups with the highest probability of receiving a monetary penalty are those largely comprised of offences of low seriousness. Thus, around half of the proved cases involving traffic offences (54%), drug offences (52%) and disorder offences (48%) result in a monetary penalty, compared to 36% of minor offences against the person, 29% of property offences, 18% of other offences against justice, 13% of domestic violence, 10% of breaches of periodic detention and 5% of serious offences against the person.
The percentage of offenders receiving a monetary penalty is higher if the most recent sentence prior to the current case is also a monetary penalty (54%), than if the previous sentence was a prison sentence (20%), periodic detention (22%), community programme (21%), community service (29%) or supervision (26%).
10.3 Multivariate analysis of current factors influencing sentencing
10.3.1 The fit and accuracy of the 1995 model
The logistic regression models for monetary penalties achieved a significant overall fit to both the full and half 1995 data, as indicated by log likelihood ratios significant at the 0.0001 level of probability and the non-significant residual (unexplained variation) terms.
The results of the test phase (Figure 10.2) indicate that the monetary penalty model gives good general predictive accuracy. That is, when the actual percentage of offenders receiving a monetary penalty is plotted against the probability of receiving a monetary penalty predicted by the model, the results are very close to the ideal line (the ideal line being where the actual probability is equal to the predicted probability). The model also produced predictions over almost the full range of probability (i.e. from 0 to 1), indicating that the model can discriminate between offenders with virtually no probability of a monetary penalty up to those who almost certainly will receive a monetary penalty.
One of the potential problems of developing the regression models is that a good fit to the data may be achieved by testing a large number of variables. Thus the model should be tested on some unseen data. The outcome of this test showed that the results were equally good when the model developed using one half of the 1995 data (the 'fitted data') was used to predict probabilities on the other, unseen half of the 1995 data (the 'test data').
Figure 10.2: Plot of the predicted probability versus the actual proportion of offenders receiving a monetary penalty, 1995 fitted and test data
10.3.2 Results of the 1995 model
As expected, offences of low or moderately low seriousness and offenders convicted on only one or a few charges are much more likely to receive a monetary penalty, all other factors being equal (Table 10.1). Thus, the odds ratio relative to the reference groups (>0-1 for seriousness and one charge) decrease sharply as the seriousness or number of charges increase. Similarly, the more serious offences (e.g. offences against the person and offences against justice) have very low odds ratios relative to property offences.
However, there are other equally significant variables in the monetary penalty model that are less intuitively obvious. In particular, women have a significantly lower probability of receiving a monetary penalty than men, even after taking account of the lower average seriousness of offences committed by women and other factors. Conversely, women are also more likely than men to receive community service or no sentence (e.g. a conviction and discharge; see Appendix I).
Māori and Pacific peoples are also less likely to receive a monetary penalty (and less likely to receive no sentence), but are more likely to receive community service, community programme or periodic detention.
Table 10.1: Logistic regression model for monetary penalties, 1995
Note: An odds ratio of >1.0 indicates a high relative risk (i.e. more likely to receive this sentence than the reference group). The most significant variable (highest Wald Chi-square, lowest probability P), is rank '1'.
The reasons for the lower use of monetary penalties for women, Māori and Pacific peoples cannot be distinguished using the statistical data available in this analysis. Some possible explanations are that there may be gender or ethnic differences in the circumstances of the average case or possibly differences in the ability to pay a fine or a perhaps a perception that community service is more appropriate in other ways for these groups.
Youth offenders (aged less than 17) are less likely to get a monetary penalty than other offenders, although the effect is not as strong as it is for other sentences, for which youth offenders have very low odds ratios. The few youth offenders who are formally prosecuted, as opposed to being dealt with by Family Group Conferences, are more likely to receive no sentence or one of the other sentences specifically for youth offenders (e.g. youth supervision and community work orders).
As with all other sentence types, the previous sentences served have a significant impact on the current sentence type. The probability of a monetary penalty is reduced if the offender has previously served a prison sentence, periodic detention, community service or supervision, especially as the most recent previous sentence.
Other aspects of the offender's previous criminal history are also relevant, although these are not amongst the most significant of the variables in the model. Offenders with an offending history of moderate to high seriousness or who have a high rate of conviction or a recent conviction, are less likely to receive a monetary penalty than the reference groups. A guilty plea also increases the probability of a monetary penalty.
10.4 Changes in the use of monetary penalties
The total number of cases resulting in a monetary penalty as the primary sentence decreased by a third (from about 83,000 to 56,000) over the 1983 to 1995 period. Much of this decrease is accounted for by the non-imprisonable offences excluded from this study. The number of monetary penalties imposed for non-imprisonable offences almost halved (from about 42,000 to 23,000) over the 1983 to 1995 period. Most of this decrease is due to the decriminalisation of minor offences (especially some minor traffic offences), although the greater use of alternatives to prosecution for minor offences has also contributed (Triggs 1998).
For the imprisonable offences included in the present study, there has been a 20% decrease in the number of monetary penalties imposed over the 1983 to 1995 period and a 33% decrease in the use of monetary penalties as a percentage of the sentences imposed for proved cases (Figure 10.3). Most of the change occurred between 1985 and 1992, with a small increase in use in the last few years.
Figure 10.3: The percentage of proved cases resulting in a monetary penalty for imprisonable offences, 1982-1997
To test whether this change is due to changes in sentencing practice or due to changes in the type of offence and offender being sentenced, the logistic model developed using 1995 data was used to predict the probability of receiving a monetary penalty for offenders in earlier years. This means that an offender in 1983 who has the same criminal history and current case characteristics as a person in 1995 would have the same predicted probability of a monetary penalty. The predicted probabilities are then compared to the actual proportion of people receiving a monetary penalty in each year. If no change in sentencing practice with respect to the statistical factors has occurred, then the predicted probability should be the same as the actual proportion receiving a monetary penalty.
The results indicate that there has been a significant change in sentencing practice, but also that some of the decrease in the use of monetary penalties can be explained by changes in statistical factors.
Overall, 59% of offenders in 1983 received a monetary penalty, compared to 55% in 1987, 41% in 1991 and 39% in 1995 (Table 10.2). Had there been no change in sentencing practice, the predicted probability of receiving a monetary penalty would still have been somewhat higher in 1983 (43%) than in 1995 (39%), due to statistical factors such as the lower average seriousness of cases proved in 1983 and the lower percentage of persistent offenders. The estimated decrease in the use of monetary penalties due to changes in sentencing practice is the difference between the actual percentage for in 1983 (59%) and the predicted percentage (43%).
Thus, the percentage of proved cases receiving a monetary penalty decreased by 20% in absolute terms (i.e. 59% minus 39%), most of which (16%) was estimated to be due to changes in sentencing practice, with an estimated 4% due to changes in statistical factors.
Changes in sentencing practice over the range of probabilities are shown in Figure 10.4. For all levels of probability, the 1983 and 1987 lines lie considerably above the 1991 and 1995 lines, indicating that the actual probability of receiving a monetary penalty was higher in the 1980s than the predicted probability based on 1995 sentencing practice. Thus, any offender sentenced in the 1980s had a higher probability of receiving a monetary penalty than an offender with the same statistical characteristics in the 1990s.
Some of this change in sentencing practice may be due to changes in factors that could not be quantified in this analysis. For example, any change in the average circumstances of offenders (such as the ability to pay a fine) would have had an influence on changes in sentencing practice, as the ability to pay a fine must be taken into account by the court.
Figure 10.4: Comparison among years of the actual and predicted probabilities of receiving a monetary penalty, with predictions made using the 1995 model
Very significant changes in sentencing practice have occurred over the 1983 to 1995 period for all levels of seriousness, all types of offence, all categories of criminal history, and all demographic groups of offenders (Table 10.2, Figure 10.5). In all categories, only a small proportion of the total change can be explained by changes in statistical factors.
Many of the offenders who would previously have got a monetary penalty are now receiving other community-based sentences, especially community service, or no sentence (see section 11.2).
Table 10.2: Actual percentage receiving a monetary penalty for each variable and year and predicted percentage based on 1995 sentencing practice
Note: The predicted percentage is the percentage of offenders who would have received the sentence had the sentencing practices of 1995 been applied, as predicted by the 1995 logistic model. The difference between the predicted percentage for 1995 and other years indicates the proportion of the total change due to changes in statistical factors (e.g. the increase in average seriousness) while the difference between the actual and predicted percentage for each year indicates the proportion of total change due to changes in sentencing practice.
Figure 10.5: Actual percentage receiving a monetary penalty and predicted percentage based on 1995 sentencing practice, by offence seriousness and previous conviction history, 1983-1995
