The Scottish Government

« Previous | Contents | Next »

Listen

C hapter 5 Costing SIVYC

Introduction

5.1 A costing task was undertaken as part of the pilot assessment. This was in parallel with the main monitoring exercise, with some of the information obtained during routine contacts and some during specific meetings with each of the three main partners. The purposes of the exercise were to:

• identify the type and scale of activities undertaken by each of the main stakeholders in the scheme and the expenditure incurred for each type;
• identify the outputs resulting from the activities and expenditure in order to provide an indication of the cost effectiveness of different support and information provision; and
• allow the scale of work implied by a national rollout (or other extension of the pilot) to be estimated and to provide cost information in a structure that would assist with preparing cost estimates for different rollout options.

5.2 The assistance of the stakeholders in providing financial data is gratefully acknowledged. The financial data were taken as provided, though with some reorganisation. The caseload numbers used in preparing unit costs were derived from the database outputs provided to the evaluation by stakeholders.

5.3 A spreadsheet of unit costs linked to a workload indicator was prepared and provided to the Scottish Executive separately from this report. This is the basis of the unit cost data presented in Tables 5.6-5.8.

Costing Scope and Constraints

5.4 The costing included only those SIVYC tasks and actions that were additional to the existing work of the organisations involved. Tasks already undertaken routinely on which the pilot depended (for example, Reporters making referral decisions) were not costed since they were subsumed in existing costs.

5.5 The stakeholders were not able to provide a firm cost for every component. This was most likely to be the case for tasks undertaken within existing resources. Strictly speaking, these are not additional costs. Such omissions do not seem significant in terms of providing information for estimating rollout costs.

5.6 We estimated missing costs elements occasionally when this was necessary and practicable, eg postage can be estimated without stakeholder data.

5.7 The costing exercise was restricted in scope and purpose, in that costs were not estimated for rollout options. Rollout strategies have not yet been determined by the Scottish Executive. Our objectives were limited to presenting the information derived from the pilot and from the stakeholders in the most accessible way possible, allowing the workload of a national rollout to be identified and drawing out the implications of the pilot for costing rollout strategies that might be considered by the Scottish Executive.

5.8 The costs presented here are of three types. 'Set up' and 'fixed' costs are not significantly affected by the scale of workload and 'unit' costs (excluding the fixed elements) are approximately scaleable to workload. These are presentational headings, rather than strict categories.

5.9 The unit cost elements can be extrapolated from the pilot to the national (or other) scale of casework, though some judgements about the impact of scale and distribution might be required. However, the fixed cost elements of a rollout cannot be derived from the pilot. They will vary, depending on the scale and characteristics of the rollout, including the administrative model adopted, the geographic distribution and the detailed specification of the service.

5.10 The detailed information on case types and task costs provided here is strongly indicative, but it cannot be definitive. In particular, the numbers of cases and workload components have been derived with difficulty from independent stakeholder databases containing inconsistent information. When preparing unit costs our usual approach has been to derive the quantity of each task from the data provided by the stakeholder undertaking the task. This can mean ignoring inconsistencies. Sometimes approximate figures have been used in order to complete the data tables and preserve their function as a set of unit costs suitable for extrapolation to other scales of scheme.

5.11 An issue with unit costs is the number of 'cases' calculated or assumed. There are issues with both definition and enumeration of cases. The unit cost is obviously sensitive to the unit definition and this may be a case where alternative definitions could be used. For example, the total number of referrals to VSS (defined as victims brought to their notice) would be a more useful indicator of activity - even efficiency - than of output production. The number of VSS referrals would include the 351 cases where no contact with the victim was made and all cases where support was not required.

5.12 The unit costs we have calculated use annualised data. The costs are the best annual estimates of ongoing costs, based on the pilot experience. The case numbers are also annualised.

5.13 A point that does not affect the costing process, but which might make the estimates vulnerable to change, is that the only indicator that can be used for expanding workload estimates from the pilot to national totals, Reported Offence Grounds, is an indicator of police activity and practice as well as of underlying demand for the SIVYC service. Changes in practice by individual forces, or more widely, might affect the number of Reported Offence Grounds significantly. These could be changes in policy on the use of warnings or changes in police activity leading to different detection rates. An illustration of the potential volatility of these data is that the Forth Valley number of Reported Offences is nearly twice that implied by its population. This is likely to reflect police detection rates rather than unusually high juvenile crime prevalence.

Presentation

5.14 The presentation of costs here is under the categories indicated above. 'Set-up' costs are those associated with the launch of the pilot and are unique to that period. They include management costs associated with the launch of the pilot and components such as IT development and equipment purchase. Some elements, such as staff training, might appear as a start up cost and continue at a lower level thereafter. It is usually convenient to discuss the set up costs alongside the 'fixed' costs, which, though continuous, are so described because they are not a direct function of the volume of casework generated by referrals. The fixed heading includes components such as strategic management and accommodation.

5.15 'Unit' costs are a function of casework load, covering components such as postage, contacts with victims and decision taking on individual cases. We have also included under this category some tasks that appeared to be scaleable using casework load as an indicator, ie they are not 'fixed'. An example is volunteer recruitment. Such a task must be a function of workload to a significant extent, though other factors, such as geographic cover, would be relevant.

5.16 An assessment of the casework load in the pilot is presented with the unit costs. This is related to the annual number of Reported Offence Grounds in Forth Valley to provide an annual 'activity rate', ie the number of occasions a task requirement is generated per referral. This can be used to extrapolate the unit cost components of SIVYC to any other area where the number of Reported Offence Grounds is known. However, presentation of costs for other areas does not come within the scope of this report. While the unit costs are essentially scaleable there may be a less direct relationship between caseload and cost than is implied by the extrapolation method. An operation identical to the pilot would have to be assumed if a direct relationship were postulated and the scale of operation will vary at local levels in ways that might cause some variance from the pilot costs.

Set Up and Fixed Costs

Police

5.17 The set up costs for Central Scotland Police were low. Prior to SIVYC, the Force operated a comprehensive scheme for the supply of information to Victim Support and the introduction of SIVYC therefore had a lesser impact than the introduction of an entirely new mechanism. It was estimated by the police that SIVYC had increased the workload in respect of Victim Support referrals by around 10%.

5.18 Table 5.1 shows that the main set up costs were alterations to the search engine utilised on Crimefile, (the Force's crime recording system), and training for Support Staff in how to mark such cases for action.

Table 5.1 Police Set-Up Costs

5.19 While there will have been some management costs for the police during the pilot period, they were not systematically monitored and the police were unable to provide them. Effectively, their contribution to SIVYC was undertaken within existing resources, which accounts for the low set up costs presented here.

Victim Support Scotland

5.20 VSS differs from the other stakeholders in that their total pilot costs can be derived from their Scottish Executive funding, ie they are already known. A note of caution is raised, however, since the outturn costs are likely to be similar to the initial estimates on which the funding was based. It may be that some of the assumptions on which the costs are based could be reviewed based on the pilot experience and on optional rollout strategy specifications.

5.21 Table 5.2 provides a summary of VSS start-up costs for the pilot.

Table 5.2 Victim Support Set-up Costs

5.22 Victim Support played a central role in the early establishment and awareness raising activities undertaken at the start of the pilot. This is reflected in the set up costs presented above.

5.23 Unlike the police, Victim Support Scotland had to undertake considerable review and design of their database for recording SIVYC cases. Despite this, the cost of database development is proportionately low in the set up costs and would also be low in a rollout of the scheme since VSS local offices would, it is assumed, use the same database nationally. Whilst some modifications may be required following further testing of the scheme, it is assumed that these would be minor as the VSS database does, at present, provide a robust source of victim information that can be easily audited and used for monitoring SIVYC referral numbers.

5.24 In addition to the set up costs, VSS provided details of staff time and resources required in running the scheme (Table 5.3). These were presented as per-year costs. They estimated the annualisation from an analysis of data collated over a longer period. Their staff cost presentation was derived from a top-down disaggregation of individual staff members' time to a range of tasks.

5.25 The italicised entries in Table 5.3 also appear in the later unit cost data tables and therefore double count with our unit cost presentation. They are included in the unit costs because, though not attributable to itemised tasks, they do appear to be scaleable to workload to some extent. The term 'direct' and 'indirect' costs are used to indicate costs that can be directly applied to a specific function and costs that can not. For example, the £54,468 of management costs in the data table for ' SIVYC Management and Admin' is an indirect cost. It was judged to be spent on general support of the work on referrals, but cannot be associated directly with individual tasks or functions. The large salary amount was specifically allocated to work on referrals by VSS but the smaller items were not. The unit costs were derived from a total of 1600 referrals taken from VSS data provided.

Table 5.3 VSS Running Costs

5.26 The above costs were considered by VSS to be largely fixed, as the main variable element is free (ie volunteer time). However, these costs are unlikely to be completely fixed for any geographic spread and scale of survey. It seems likely that an important variable in determining the scale of management and support costs linked to referrals would be the extent to which resources (other than volunteers) had to be provided locally throughout a wide geographic distribution of the service.

5.27 VSS made the point that they would expect to see the unit cost drop substantially (to about half) in the context of a larger operation of the scheme. However, the amount of reduction would inevitably depend on the detailed definition of the rollout and its implications for management.

5.28 It is worth noting that the use of volunteers brings added value to the VSS project, though there are no costs to report.

SCRA

5.29 Cost data for the SCRA set-up is presented in Table 5.4 below. Of each of the main stakeholders, the SCRA database ( RAD) required the most modification to accommodate the SIVYC scheme. This is reflected in the table below with more than 90% of the set-up costs being allocated to this task.

Table 5.4 SCRA Set-Up Costs

5.30 Training in the use of the SIVYC specific fields of the RAD would be a potentially large cost for a full national rollout. It was reported during consultation that around 370 staff would require such training. Extrapolating from the pilot training (which covered four staff) would give a total of nearly £140,000. This would be in-house training of groups of staff and it is possible that some economies of scale might be found. Despite the high cost of training, however, this would not continue beyond the rollout period. The skills would become embedded and form part of the induction of new staff along with other standard capabilities.

5.31 The large cost of database development would be marginal in a national rollout since, unlike the police, the RAD provides a nationally accessible and nationally shared standard. Although some changes are required to the database as it currently stands, these costs are fixed and could be shared by any new schemes being set-up elsewhere. SCRA have already commissioned database adjustments to reflect the ongoing needs of SIVYC administration. The costs of this have been reported to the Scottish Executive separately.

5.32 As with VSS, SCRA provided additional costs which are difficult to place within a unit framework and which we have, therefore, positioned within the set-up and fixed costs element. Table 5.5 below shows these costs.

Table 5.5 Additional SCRA Pilot Costs

5.33 As with the police, strategic management resources came from individuals who have other responsibilities and so the SYVIC costs have not been additional or, at least, have not been measurable. They may not be fully represented in the figures shown above, unless they come through in the evaluation costs. However, we doubt if they would be meaningful in terms of rollout implications.

5.34 Authority reporters currently attend a monthly liaison meeting and that is a small cost not included in the above.

5.35 Other costs would be co-ordination and IT support from Logica. These, and other management and administrative costs, would vary depending on the scale and model of operation.

Unit Costs

5.36 These are defined by type of case and by task. The case types for which it was suitable and possible to prepare data are:

• Case Type A: Victim receives letter about SIVYC from the police but opts out of the scheme (both information and support);
• Case Type B: Victim opts in to information direct from SCRA but opts out of support or information from Victim Support; and
• Case Type C: Victim opts in to information and support via Victim Support.

5.37 The tasks within each case type have been defined in as disaggregate a way as is practicable, given that cost estimates for each task are required. Ideally, the cost categories should be consistent with the main components of the service and generally they are.

5.38 A mechanism is needed to scale the unit costs to national or other levels. As discussed earlier, the absence of a shared unique identifier for victims means that it was not possible to carry out analysis of costs at a shared unit level ie costs per victim. The stakeholder data provided by the police and by Victim Support Scotland contained counts of 'victim' referrals while the SCRA data provided counts of 'alleged offence ground' referrals. It proved impossible to create a relational database for victims referred to the pilot.

5.39 For this reason, the offence grounds reported to the SCRA by the police (Reported Offence Grounds) is the only indicator of scale that can be used to predict SIVYC activity levels in a national (or other) context, ie if the number of reports in the pilot area gave rise to a number of actions recorded in the pilot it is assumed that this proportion would be replicated at national level. The information is collated nationally and by area whereas number of recorded youth crimes (police) and numbers of victims of youth crimes ( VSS) are not.

5.40 The unit cost data are presented here as annual task numbers and costs. When necessary, the numbers of actions reported in the databases have been adjusted to create annual estimates. As the pilot period was a year, this incorporated all cases initiated in the year. For some cases, the processes involved would not complete within the pilot period. Without a relational database of cases it was not possible to select cases deemed to be complete to discover, for example, the proportion that went on to receive different types of support. Annual numbers of some tasks were therefore estimated by deciding on the valid period from the time of offence reporting. If it appeared from the databases that an action took three months (from the time of the offence report) to complete for most cases the activity rate could be developed only from cases reported within the first nine months of the pilot. The rate could then be applied to the number of reports in the full 12-month pilot period to give the estimated annualised total. Alternatively, the data from the part year could be expanded proportionately to the full year. It should be acknowledged that the data were not always clear about which would be the valid period and elements of approximation were involved.

5.41 An illustration of annualisation is the SCRA provision of decision information. It was (somewhat approximately) determined that the valid period was 10 months, during which the following quantities were found in the SCRA data.

5.42 The total number of cases was factored up from 397 to 476 to reflect the part year. However, annualisation could disregard the 'awaiting decision' element because in any long term scheme this would have to match the number of decisions required. The 69 undecided cases were therefore attributed proportionally to the 'letter' categories giving the following annualised totals in the data table.

5.43 The strategy of grossing up proportionately to the full year was adopted because we had more confidence in the activity data than in the total number of cases.

5.44 Unit costs are presented in Tables 5.6 to 5.8 below, being one cost table for each of the three main case types identified in the pilot.

5.45 The elements in the data tables are defined below:

• Activities are the elements that are costed, mainly specific actions or cost items that can be enumerated.
• Grounds reported to the SCRA or Offence Grounds is the indicator of scale in the data that is used to predict SIVYC activity levels in a national (or other) context. An area generating twice the number of reported cases would generate twice the SIVYC activity.
• An activity rate is the number of times an action or cost is recorded in the pilot as a proportion of the number of Reported Offence Grounds. Thus, 1000 offences and 50 instances of an activity yield an activity rate of .05. These rates can then be applied to any scale of operation as altering the number of Reported Offence Grounds in a spreadsheet data table creates a new estimate of the costs.
• In the case of re-estimates to national or other scales the number of actions is the product of the activity rate and the number of reported offences.
• The unit cost is the cost per action, normally obtained from stakeholders.
• The total cost is the product of the unit cost and the number of actions.

Cost Summaries

5.46 Table 5.6 provides a summary of costs for case type A. During the pilot, only a small number of cases were recorded as 'opting out' of both the information and support elements of the SIVYC scheme (n=170). For all such victims, the main activity rests with the police who are responsible for issuing the initial SIVYC offer letter and for notifying SCRA of all opt outs of victim information and support. It is important that the police notify SCRA of opt outs to ensure that information is not sent directly to the victim in the absence of VSS contact.

Table 5.6 Summary of Annual Costs for Case Type A: Victim receives letter about SIVYC from the police but opts out of the scheme (both information and support)

5.47 Although there is no requirement in the SIVYC protocol for VSS to be notified of victims opting out of the SIVYC scheme entirely, an analysis of VSS data suggested a total of 81 victims having declined VSS support or information and they are noted here only to indicate the marginal costs that might be associated with communicating opt out decisions between the police and VSS.

5.48 Staff time of £2 per case (as advised by the police) covers notification to VSS/ SCRA. An allowance of 40 pence per letter for post and stationary has been made for the letters. An allowance of 20 pence per item for communication (mainly fax) has been made for the notifications, here and elsewhere.

5.49 For Case Type A it is possible to generate a unit cost of £3.80 per victim opting out of the SIVYC scheme. It is worth noting, however, that time required for logging of victim decisions on the SCRA database may vary depending on the number of young people and alleged offences associated with each opt out victim. Multiple offenders and alleged offence grounds per victim would increase the time required for recording victim opt outs on the RAD. Since the data provided by SCRA is grounds based and not victim based, it is not possible to calculate a variable cost for this task relative to grounds referred.

Table 5.7 Summary of Annual Costs for Case Type B: Victim opts in to information direct from SCRA but opts out of support or information from Victim Support

Table 5.8 Summary of Costs for Case Type C: Victim opts in to information and support via Victim Support

5.50 It is important to note that the resources for case type C include volunteer time that is not covered here. VSS reported that much of the increased resources of a larger operation would be taken up by volunteers at little cost so there may not be an increase in costs proportionate to the increase in Reported Offence Grounds (or associated victims referred).

5.51 The figures for individual contacts made with victims were reported to us by VSS and are presented here to give an impression of the type of communications made. We have assumed that they come under the 'other' referral costs as they are not covered elsewhere in the figures provided to us by VSS.

5.52 The tasks and costs that are italicised were identified as referral-related by VSS but not attributed to specific outputs. Conversely, telephone and postage, printing and stationary, volunteer recruitment, expenses and training were not identified as variable by VSS but we have assumed a significant variable element.

5.53 An issue that affects the VSS presentation of unit costs is how they should best be calculated, particularly with respect to staff time. Specific activities such as contacting victims could be costed in two ways. The 'bottom-up' approach would involve estimating the average time per case spent by staff at different levels and then valuing the time. The 'top-down' approach would involve identifying the proportion of staff time spent on the activity as a whole, valuing this as a proportion of the staff cost, then deriving the unit cost from the number of actions. The costs presented here are derived from top-down disaggregations.

5.54 This approach means that the VSS unit cost contain significant general elements, which were attributed to the variable part of their budget under the heading ' SIVYC Management and Admin'. At £54,468 this is a significant part of the total, so it is useful to distinguish it from absolutely fixed costs. We considered attributing this cost to the other more specific items on a pro rata basis but rejected that as too arbitrary. This number has therefore been included amongst the unit costs, but it should be noted that it would not be right to scale it up to national casework volume because economies of scale would be ignored. VSS could advise on these economies in the context of specified rollout plan.

5.55 The costs provided by VSS contain unit elements (that is to say variable elements) that are service activity rather than specific tasks that can be enumerated. Recruitment of volunteers is an example. It seems appropriate to link this to caseload. The rationale for considering this as a variable element linked to activity seems clear and incorporating it into the unit costs gives a transparent process for expansion. In such cases we have used the 12-month estimate of costs as a lump sum that would vary in accordance with the Reported Offence Grounds indicator.

5.56 While the data presented here are based on that provided by the stakeholders it is not always identical. The annualised data provided by VSS includes unit costs calculated using a number of 'referrals'. Rather than use this figure, which we take to be their estimate of likely workload, we have used the annualised number of actions derived from the pilot as the unit cost denominator. If the number of actions could not be identified or if it was an activity that could not be defined in these terms, then the cost was tied directly to the number of Reported Offence Grounds.

5.57 Table 5.9 below provides a summary of costs for each of the three main case types and the overall costs of operation for the pilot period. As the data table shows, Type C cases incur much the highest unit costs. It is not possible to determine the exact number of cases each represents, because of the data and definitional problems, but the total of Types B and C will not differ greatly, ie the Type C unit costs is several times that of Type B. This is unsurprising, given the concentration of activity on this type and the large amount of SIVYC management and administration. Removing this from the unit costs would not effect a meaningful change because the management costs would come through as fixed and, because of VSS's role, would eventually come through as Type C. However, this particular element would reduce as a unit cost in a larger operation.

Table 5.9 Summary of Costs by Case Type

Costs of Further Development of SIVYC

5.58 In addition to the main set up and operational costs detailed above, discussions with stakeholders at the conclusion of the assessment identified a number of operational changes and adjustments that would be required in a rollout of the scheme. Many of these tasks would incur costs that have not been included in the presentation above.

Adjustments to Police Database and VSS notification

5.59 Rolling out the scheme more widely would involve other forces, and we understand that the situation for some would be very different, especially in so far as start up and IT development were concerned. We know of no way to cost this aspect other than to undertake consultation with the forces, probably starting with an approach to ACPOS if that has not already been done. A functional specification is required that can be distributed to forces in order that they might provide estimates of costs associated with each task in the context of current crime recording practices SCRA and VSS referral practices. At a minimum, costs would be required for:

• management
• IT specification and development
• staff training
• added workload for victim support referrals
• ongoing staff training and support
• ongoing cost of postage etc

Victim Information Leaflets

5.60 Whilst the costs of victim information leaflets have been presented above, the monitoring and evaluation exercise has revealed a need for some re-drafting of the information leaflets to make them more accessible for victims.

5.61 In the course of the pilot, VSS developed leaflets for both young people and adult victims. Different media work well with different audiences and separate information leaflets may need to be developed for each. That said, however, it is anticipated that any changes in this regard would be modification, rather than complete re-drafting, thus cutting down significantly on the costs.

RAD Re-development

5.63 The Scottish Children's Reporter Administration are currently commissioning changes to the administration of the SIVYC scheme which will have cost implications.

Monitoring and Evaluation

5.64 Finally, a cost which must be considered as additional to the operational costs highlighted above is that of ongoing monitoring and evaluation of the scheme.

Summary

5.65 The costing exercise shows the following:

• That police unit data are low cost in Forth Valley, however, there can be no presumption of national applicability. The police made the point that their costs were low because they already had a system in place for provision of information. Consultation with the other forces is needed to establish the practicability and costs of achieving a rollout where the scheme would operate consistently throughout Scotland. Because VSS and SCRA databases can be expected to operate consistently, difficulties would probably arise if the police data were inconsistent.
• VSS unit data are much the most complex unit costs, reflecting the range of VSS activities. A significant cost is incurred by operating an opt out system, though this may be weighted against the benefits of offering information and support in a pro-active way. The costs of rollout would depend on scale, the rollout strategy and the detailed specification of the national scheme.
• Reporter decision making is much the most expensive unit cost for SCRA. Rollout costs would again vary, depending on the details of the rollout strategy. The training costs of integrating the scheme and database administration into SCRA working practice appear to be very high and, therefore, worth a review of approach. These could change as a result of review of SIVYC protocol.

« Previous | Contents | Next »