Frequently Asked Questions

Bounding the cost analysis work

Any attempt at any type of cost analysis will require the setting of boundaries. For instance, how will crime and disorder be defined, which agencies will be included when looking at the cost of prevention? An obvious example of these problems is ‘disorder’ which will be defined in different ways by different agencies, making it difficult to measure rates of incidence and hence produce figures to base cost analysis on.

Crime and disorder is under-reported and this makes estimating levels of crime difficult. Without accurate figures, cost analysis becomes difficult.

The British Crime Survey (BCS) suggests that approximately three quarters of crime is either unreported or unrecorded. However, levels of under-reporting and under-recording vary for different types of crime and disorder. Some types of crime are particularly prone to under-reporting e.g. racially motivated crime, homophobic crime, drug possession and offences associated with prostitution. Therefore, it is not possible to use BCS figures to apply a single multiplier to recorded crime statistics to ascertain true levels of crime. Other types of crime are not covered by the BCS (e.g. handling stolen goods, some fraud and forgery) and therefore it is not possible to use it to ascertain a multiplier for these types of crime.

For its work on the economic and social costs of crime, the Home Office developed multipliers from the BCS where possible and from other research and expert opinion where offenders were not covered by the BCS. See the British Crime Survey website. www.homeoffice.gov.uk/rds/bcs1.html

Cost-effectiveness / cost-benefit analysis of mainstream provision will require differentiation between the different roles of specific organisations and the disaggregation of funding streams.

Most services will not be entirely devoted to crime reduction in the preventative sense. Crime prevention will either be a secondary aim (e.g. the police) or an unstated aim (e.g. LEA). To map out the costs of a particular service’s contribution to crime prevention, it will therefore be necessary to differentiate between the different types of work that a particular organisation does and then disaggregate funding streams in order to apportion budgets across different functions of an organisation.

Modern financial management systems make this task difficult. Shapland (2000) describes the process researchers used in auditing criminal justice in Milton Keynes. The audit was designed to identify the cost of local agency’s contributions to the criminal justice system in Milton Keynes and recognised that many agencies would make some contribution, even if the support of the criminal justice system was not their primary role. Therefore it was necessary to identify the personnel involved in the criminal justice system and then ascertain the actual costs of those personnel. Shapland and her team found that each functional unit of an organisation was likely to be a cost centre and that this facilitated matching up ‘direct’ spend such as salaries, with personnel and their functions. However, while the identification of direct costs was relatively straightforward, the increasing centralisation of purchasing, fleet management and buildings management, meant that ‘indirect’ costs such as support services and premises, were costed separately at an organisational level, and were difficult to link to those individuals identified as contributing to the criminal justice system.

Undertaking evaluative cost-effectiveness and cost-benefit analysis will involve an element of evaluation whereby it is possible to quantify the impact of the initiative in terms of the number of crimes prevented.. This raises a number of issues including

• The onset, longevity and decay of outcomes will vary according to the type of intervention being assessed. This will have implications for the period of time over which outputs and outcomes will have to be measured in order for cost-effectiveness / cost-benefit ratios to be calculated. For some social crime prevention interventions the period of time to be taken account of will be significant.

• There may be additionality of outcomes. One of the most difficult aspects of an impact evaluation is assessing to what degree outcomes can be attributed to the intervention under scrutiny as opposed to external influences.

• There may be displacement of crime or a diffusion of benefits from an intervention.

For long-term measures, where the impact upon crime or criminality will not be evident in the short-term it will be necessary to identify intermediate outcomes that can be assessed. For instance, in a criminality reduction project in a secondary school the impact upon criminality might not be evident for 3 – 5 years. However, intermediate outcomes such as a reduction in truancy, a reduction in bullying, or increased attainment can be measured in the shorter-term and national research on the links between these risk factors and criminality used to extrapolate to the possible impact upon criminality.

Evaluation design will have to take account of the possibility of displacement of crime and of the diffusion of benefits. Both displacement and diffusion may occur between the intervention area and neighbouring areas (e.g. introduction on CCTV in one car park might make other car parks more vulnerable), but also between one type of crime and another (e.g. a reduction in burglary might be accompanied by a rise in street robbery), or between one time and another (e.g. the introduction of neighbourhood wardens might change the time at which crime occurs).

It will be necessary to demonstrate clear causal links between inputs, outputs and outcomes and this may involve complex qualitative and quantitative analysis involving extensive multi-variate analysis. Where appropriate a sensitivity analysis should be carried out which involves varying key parameters in the analysis to test the robustness of results. One such parameter is the assumption of how much of the observed outcomes can be attributed to the intervention. Varying assumptions about this level of attribution will give a range of estimates of outcomes and benefits. The level of flexibility of results provides a guide to the robustness of findings.

Cost-effectiveness and cost-benefit analysis have complex policy implications.

By undertaking cost effectiveness and cost-benefit analysis it is possible to calculate the relative cost of different interventions and the overall benefits accrued or savings made by a particular intervention. However, this then raises difficult questions for policy makers as to what they should do with this information.

One problem is that although cost analysis might allow us to put a cost on the prevention of each crime, it does not follow that the organisation or project that has brought about that reduction and made the saving will actually accrue a realisable benefit. There are two problems. First the benefit may accrue to a different organisation and this will raise issues that will test the true extent of partnership working in an area. Second, most of an organisation’s costs will be tied to personnel and accommodation, so that any reduction in crime would have to be large and sustained, before it was possible to reduce staff numbers and realise savings. Therefore any cost analysis must pay careful attention to the calculation of marginal costs.

Another problem is that the savings that an organisation achieves through crime reduction might accrue to a different organisation. For instance, if the local Youth Service runs a successful Summer Youth Club the Fire Service might make savings because a reduction in incidents of arson.

Back to Cost Analysis