Fire Funding Formula Options Paper 21 December 2017

Fire Funding Formula Options Paper 21 December 2017 Background Funding for the fire and rescue service in England is calculated via the Fire and Rescue Relative Needs Formula. This applies to all English Fire and Rescue Authorities (FRAs) including stand-alone fire authorities and FRAs which are a part of a larger county or unitary authority. The fire formula is one of many formulas alongside those covering services including children s services, adult social care and highways. How the result of that formula translates into funding allocations is more complicated and involves an authority s relative levels of need, relative size of tax base per head of population and their business rates as well as a degree of ministerial judgement. Income for the fire and rescue service also comes from council tax as well as other fees and charges. Different areas raise different proportions of their budget from council tax and fees/charges. This paper has been drafted by the Somerset Technical Support Team in response to the Department for Communities and Local Government (DCLG) issuing a Fair Funding Review consultation paper (published here on 19 th December). Hence, this paper only examines the current fire relative need formula (RNF) as this is an element of the Review. This paper examines the current formula s distribution before looking at some alternative options that the NFCC Finance Committee might discuss further at their next meeting on 4 th January. The Fire and Rescue Relative Needs Formula As has tended to be the case with recent funding formulae, the fire need formula was originally created by looking for measures which, when put into a formula, roughly matched past spending. One benefit of using previous spend as an indicator for need in this way is that data is readily available whilst another is that it doesn t produce funding cliff-edges (where an authority loses a significant amount of funding due to the new formula). On the contrary, criticisms of this approach are that it can lock in previous funding disparities (as previous spend is dependent on previous funding) and that the resultant formulas can be complex and lack transparency. The main factor of the fire RNF is projected population which is multiplied by the result of a basic amount added to an array of top-ups (see below). The result of this multiplication is then increased by an Area Cost Adjustment (ACA). The ACA is intended to reflect the different costs associated with service provision in different areas and currently includes a measure of local wage pressure and business rates. The ACA will be dealt with in greater detail later on in this paper. Technical Support Team - September 2017 Page 1

The current relative need values range from 0.0000875412 (Isle of Wight) to 0.005627349 (GLA). As the name suggests they do not relate to cash values but simply indicate, in this case, that the GLA s needs are 64 times the Isle of Wight s needs (64.3 = 0.005627349/0.0000875412). These RNF values currently pass through the Four Block Model mechanism where, in theory at least, allowances are made for resources (council tax) and transitional arrangements. The graph in figure 1, below, illustrates the close relationship between calculated need according to the current formula and population. 20.0% Relative Need v Relative Population 15.0% 10.0% 5.0% 0.0% Avon Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire Cleveland Cornwall Cumbria Derbyshire Devon & Somerset Dorset Durham East Sussex Essex GLA Gloucestershire Greater Manchester Hampshire Hereford & Worcester Hertfordshire Humberside Isle of Wight Kent Lancashire Leicestershire Lincolnshire Merseyside Norfolk North Yorkshire Northamptonshire Northumberland Nottinghamshire Oxfordshire Shropshire South Yorkshire Staffordshire Suffolk Surrey Tyne and Wear Warwickshire West Midlands West Sussex West Yorkshire Wiltshire Proportion of Relative Need Formula Proportion of Population Figure 1: The relationship between need allocations and the shares of population (please refer to Annex A for the corresponding table) Technical Support Team - September 2017 Page 2

The DCLG s Fair Funding Consultation refers to common cost drivers this can be considered as a presentational issue but it would appear that, in the current fire formula at least, that population is a common cost driver. With the notable exception of the GLA almost all authorities allocations would be immaterially different if the population were the only driver of need (please see Annex A for the corresponding table). In each formula that is created for a purpose such as this it is often said that there needs to be a balance struck between complexity and fairness. The pervading logic is that a more complex formula is capable of being fairer, whilst a simple formula may end up sacrificing some accuracy or fairness. This is not something that the DCLG necessarily agree with believing that a simiple formula is capable of also being fair. The Fire and Rescue RNF itself is not overly complex, in that it can be calculated with relative ease and it does appear to include intuitive measures given the activities performed by the service. There are 7 top-ups in the current formula, in addition to the basic amount (which can be considered a per head allocation). These top-ups can be summarised as follows: 1. Coastline Top-Up - Put in place to help fund the additional pressures faced by FRSs which assist Port/Harbour Authorities and/or the Maritime and Coastguard Agency. 2. Population Density Top-Up - This top-up reflects the increased threat of fire in densely populated areas. 3. Population Sparsity Top-Up - At the other end of the scale, some FRSs face a significant challenge in reaching sparsely populated parts of its authority. This is reflected through this top-up. 4. Deprivation Top-Up - There is a strong link between deaths and injuries from house fires and social deprivation particularly regarding children. To reflect this increased risk a Risk Index has been calculated using the following four factors: (i) The proportion of working age adults with no qualifications (ii) The proportion of the working population which received Incapacity Benefit, Severe Disability Allowance or Employment support Allowance, or were New Deal Program Starters/Apprenticeship learners. (iii) The proportion of persons who are, or whose partner is, in receipt of Income Support/Income Based Jobseeker s Allowance/the Guarantee Element of Pension Credit (iv) Mortality Ratio Technical Support Team - September 2017 Page 3

5. High Risk Top-Up - Reflecting the increased risk of COMAH (Control of Major Accident Hazards) sites, this top-up distributes additional funding in proportion to the number of COMAH sites per resident. 6. Property and Societal Top-Up Based on the Secretary of State s estimate of property and societal risk, in turn based on Valuation Office Agency and 2006 Fire Services Emergency Cover Toolkit information, on a per person basis. 7. Community Fire Safety Top-Up - This top-up reflects the increased risk associated with children, OAPs and those residents which require fire safety education. As well as the numbers of children and elderly people it also includes a measure of the number of residents with a greater need for fire safety assistance. These top-ups are then summed together with the basic amount and multiplied by the projected population to give a measure of need prior to the application of the Area Cost Adjustment. It is possible to demonstrate, for each FRA in England, how much of their overall need allocation comes from either the basic amount (population) or the top-ups. Figure 2 demonstrates that for all FRAs the primary measure determining allocations is population. The graph below tells us some interesting things some of which are obvious and some of which are less so: Community Fire Safety makes up a very similar proportion of need in all FRAs The proportion attributable to Population Density is also similar in each FRA with the main notable exceptions of the GLA and Cornwall The High Risk Top-Up contributes a very low proportion of need for the vast majority of FRAs, the exceptions being Cheshire, Cleveland and Humberside. Technical Support Team - September 2017 Page 4

100% Weightings within Need Allocations 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Avon Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire Cleveland Cornwall Cumbria Derbyshire Devon & Somerset Dorset Durham East Sussex Essex GLA Gloucestershire Greater Manchester Hamshire Hereford and Worcester Hertfordshire Humberside Isle of Wight Kent Lancashire Leicestershire Lincolnshire Merseyside Norfolk North Yorkshire Northamptonshire Northumberland Nottinghamshire Oxfordshire Shropshire South Yorkshire Staffordshire Suffolk Surrey Tyne and Wear Warwickshire West Midlands West Sussex West Yorkshire Wiltshire Population Coastline Population Density Population Sparsity Deprivation High Risk Property & Societal Risk Community Fire Safety Figure 2: Constituent weightings contained within RNF calculations Technical Support Team - September 2017 Page 5

What is the ACA? The cost of providing the same public services can vary between local authorities for a number of reasons. One reason is because of particular area characteristics, such as numbers of children, or cumulative length of road these are cost drivers which are (at least theoretically) accounted for in the Relative Needs Formulas (RNFs). Another reason is because of differences in the costs of inputs which local authorities need to buy; for example buildings and wages. Area Cost Adjustments (ACA) form part of a needs assessment and adjust local authority allocations for the latter differences. Currently the ACA reflects two sources of differences in costs between areas: i) Differences in labour costs; and ii) Differences in business rates paid on premises. There are currently different ACAs for different service areas reflecting the shares of budgets considered to be affected by rates and local labour costs. The table in Annex B lists the Fire & Rescue Area Cost Adjustments for each FRA in England. The ACA s are multiplicative in other words an ACA of 1.1 means that a FRA s need allocation is then multiplied by 1.1 (increased by 10%) to take account of other costs. An ACA of 1.0 indicates no additional costs in that FRA area. The Fair Funding Consultation asks whether the ACA should be extended to include costs associated with widely dispersed populations and also whether deprivation should be included in the ACA. Do NFCC members have any evidence for increased costs (not demand/risk) as a result of deprivation or sparsely populated areas? What will the fair funding review cover? The Technical Support Team is relatively confident that the Four Block Model mechanism will no longer be used in determining allocations although so far in the review process the DCLG has not ruled anything in or out. To carry out a full review of the current funding formula, the DCLG will need to consider: A new formula to measure need for each authority which takes account of the services they each deliver A way to account for varying costs (e.g. wages, rents) in each area the area cost adjustment, mentioned above Technical Support Team - September 2017 Page 6

A way to account for different areas differing abilities to raise income from council tax Possibly a way to account for differing incomes from sales, fees and charges Some kind of transitional arrangement The Fire Finance group of the NFCC will need to eventually discuss all these points and decide on a consensus but for now the DCLG s questions are concerning the need formula. What are the formula options? The DCLG have presented a number of loose scenarios ranging from doing nothing (keeping old data and old formulae), through updating the data but keeping the old formula to creating a whole new formula. Most technical experts will acknowledge that the formula and data are now old and that the financial climate alone will have changed the way in which services are delivered. Therefore, in most cases, the technicians will be supportive of a new formula. The Fair Funding Review Technical Working Group (TWG) - formerly the Needs and Redistribution Technical Working Group - meets regularly to discuss the formula review. The TWG has heard from the DCLG that it is attracted to a new formula based on Multi-Level Model regressions to past spend and/or activity - the recent Children s Services tender document and Fair Funding Review consultation appear to confirm this. Presumably, this is because reliable spend data is available over a number of years and past formulae have used similar methodologies that both civil servants and, perhaps more importantly, the ministers are familiar with the technique. The local government representatives within the TWG are of a slightly different opinion in that they can see that basing a formula on past levels of spend/activity is only going to replicate current funding patterns and would be unlikely to address any funding inequalities that might exist. Instead they favour a model built upon cost drivers identified by the service rather than ones that fit regression formulae. To date there has been very little steer from ministers about their thoughts on the need formula. Working groups have, so far, been primarily concerned with determining cost drivers. There has been a strong argument for a simplified funding formula based on just a handful of measures. An alternative method, that is similar to that being suggested by the local government representatives and that NFCC might want to consider is outlined in Annex C. This draft paper has been prepared on behalf of the Society of County Treasurers which plans to present the paper at the January Technical Working Group. Technical Support Team - September 2017 Page 7

Are the NFCC supportive of an alternative to regression? Although the current top-ups do seem to be intuitively correct, the NFCC will probably want to give thought to what elements of a fire authority s activity is not currently captured. One might be attending road traffic accidents (RTAs). A possible measure of RTAs which crucially need a fire engine to attend is unlikely to include simply the length of roads but also the average speed travelled on the roads. What do members consider might be an appropriate measure for RTAs to ask the DCLG/HO to investigate further for possible inclusion in a fire formula? Are there any other omissions from the current formula? The chart on page 2 demonstrates the importance of population on formula allocations. Would the NFCC be satisfied if population was the only cost driver for the fire service? It could be argued that, for quite a significant amount of additional work to create a formula, allocations are not that different from what could have been achieved using a simple per-head allocation. What other data sets does the fire service have access to? The Government currently collects and publishes information by fire authority (including those within counties/unitaries) on the type of incident being responded to broken down into the following categories: False alarms as a result of faulty apparatus, with good intent or maliciously Chimney fires Dwelling fires Fires in other buildings Outdoor fires Road Vehicle fires Secondary Fires; and Non-fire Technical Support Team - September 2017 Page 8

In the case of a primary fire; there is a lot of information collected and published, including: The type of incident dwelling, vehicle, appliance, other building The type of dwelling, vehicle, appliance or other building The day of the week How long the fire had been burning before fire crews were called Whether it was deliberate or accidental The response time Number of fire fighters in attendance Number of vehicles in attendance Length of time at the scene Fatalities or casualties Whether rescues were required Is there any other data that the NFCC consider is needed in order to capture what FRAs do? Do we know whether this data is accurate? Is it suitable for inclusion in a fire formula? Could this incident data be skewed by past over/under funding patterns? There is also data on the number of fire fighters by wholetime, retained duty, fire control and support staff. The number of wholetime and retained duty officers are also given in terms of the following roles: Brigade Manager Area Manager Group Manager Station Manager Watch Manager Crew Manager Fire Fighter Technical Support Team - September 2017 Page 9

Clearly, FRAs do not only respond to fires and other incidents, they also carry out fire safety and fire prevention work. There is also data published on a number of prevention/fire safety-type activities: The number of fire risk checks and the hours spent on them: in total, where resident are aged 65+ or disabled and also checks carried out by partners The number of fire safety audits In total The number that were satisfactory The number that were unsatisfactory The number of enforcement notices served under Articles 254, 255, 256 and 257 The number of serviced notices that resulted in a satisfactory outcome and The number of premises known to the FRA Campaigns and Initiatives o In total o Firesetter, anti-social behaviour schemes and other youth diversions o Other youth fire safety programmes Do the NFCC consider that this captures all of the prevention work being carried out? How do we know if this data is accurate? Is there a danger of FRAs doing more prevention work where there are the resources available and less when/where resources are scarce? Do the NFCC wish to request that the Home Office/DCLG further investigate whether the current prevention activity is effective? Every FRA must produce an Integrated Risk Management Plan (IRMP), which is available to the public. The DCLG s guidance includes business continuity, community safety, environmental protection, equality and diversity, heritage buildings, road traffic collisions and wildlife. Clearly, if plans are required for each FRA area, they will all be different. Does the NFCC consider whether there is any merit in analysing these IRMPs? Does the NFCC have any better data? Technical Support Team - September 2017 Page 10

Annex A Proportions of RNF versus proportions of population Fire Authority Fire and Rescue RNF Proportion of Fire and Rescue RNF Population Proportion of Population Avon 0.00054982 1.9% 1,096,388 2.0% Bedfordshire 0.00030946 1.1% 634,550 1.2% Berkshire 0.00043127 1.5% 885,829 1.6% Buckinghamshire 0.00034093 1.2% 772,388 1.4% Cambridgeshire 0.00036856 1.3% 825,381 1.5% Cheshire 0.00052220 1.8% 1,038,892 1.9% Cleveland 0.00036844 1.3% 562,422 1.0% Cornwall 0.00035493 1.3% 544,216 1.0% Cumbria 0.00028585 1.0% 501,217 0.9% Derbyshire 0.00047732 1.7% 1,033,567 1.9% Devon & Somerset 0.00084934 3.0% 1,689,881 3.1% Dorset 0.00035477 1.3% 757,647 1.4% Durham 0.00031628 1.1% 626,153 1.2% East Sussex 0.00040642 1.4% 811,820 1.5% Essex 0.00087246 3.1% 1,763,285 3.3% GLA 0.00562735 19.9% 8,459,566 15.6% Gloucestershire 0.00026939 1.0% 607,509 1.1% Greater Manchester 0.00145385 5.1% 2,722,679 5.0% Hampshire 0.00084252 3.0% 1,791,362 3.3% Hereford & Worcester 0.00034257 1.2% 758,333 1.4% Hertfordshire 0.00052678 1.9% 1,141,136 2.1% Humberside 0.00055202 1.9% 927,659 1.7% Isle of Wight 0.00008754 0.3% 139,895 0.3% Kent 0.00084177 3.0% 1,766,154 3.3% Technical Support Team - September 2017 Page 11

Fire Authority Fire and Rescue RNF Proportion of Fire and Rescue RNF Population Proportion of Population Lancashire 0.00076557 2.7% 1,474,707 2.7% Leicestershire 0.00049016 1.7% 1,036,842 1.9% Lincolnshire 0.00036272 1.3% 731,723 1.4% Merseyside 0.00078794 2.8% 1,382,551 2.6% Norfolk 0.00043762 1.5% 874,729 1.6% North Yorkshire 0.00037737 1.3% 807,062 1.5% Northamptonshire 0.00032318 1.1% 710,407 1.3% Northumberland 0.00018147 0.6% 318,152 0.6% Nottinghamshire 0.00053260 1.9% 1,108,111 2.0% Oxfordshire 0.00029939 1.1% 663,236 1.2% Shropshire 0.00021224 0.7% 478,907 0.9% South Yorkshire 0.00068107 2.4% 1,360,525 2.5% Staffordshire 0.00051537 1.8% 1,110,374 2.1% Suffolk 0.00033897 1.2% 740,134 1.4% Surrey 0.00050950 1.8% 1,159,941 2.1% Tyne and Wear 0.00059747 2.1% 1,117,163 2.1% Warwickshire 0.00025507 0.9% 554,620 1.0% West Midlands 0.00150043 5.3% 2,784,932 5.2% West Sussex 0.00036359 1.3% 824,719 1.5% West Yorkshire 0.00114890 4.1% 2,273,284 4.2% Wiltshire 0.00030709 1.1% 695,993 1.3% Technical Support Team - September 2017 Page 12

Annex B Area Cost Adjustments in the Fire & Rescue Relative Need Formula FRA ACA Increase Avon 1.0479 4.8% Bedfordshire 1.0507 5.1% Berkshire 1.1205 12.1% Buckinghamshire 1.0930 9.3% Cambridgeshire 1.0424 4.2% Cheshire 1.0137 1.4% Cleveland 1.0000 0.0% Cornwall 1.0000 0.0% Cumbria 1.0000 0.0% Derbyshire 1.0000 0.0% Devon & Somerset 1.0000 0.0% Dorset 1.0000 0.0% Durham 1.0000 0.0% East Sussex 1.0078 0.8% Essex 1.0350 3.5% GLA 1.1773 17.7% Gloucestershire 1.0223 2.2% Greater Manchester 1.0194 1.9% Hampshire 1.0461 4.6% Hereford and Worcester 1.0000 0.0% Hertfordshire 1.0924 9.2% Humberside 1.0000 0.0% Isle of Wight 1.0461 4.6% FRA ACA Increase Kent 1.0133 1.3% Lancashire 1.0000 0.0% Leicestershire 1.0000 0.0% Lincolnshire 1.0000 0.0% Merseyside 1.0060 0.6% Norfolk 1.0000 0.0% North Yorkshire 1.0000 0.0% Northamptonshire 1.0131 1.3% Northumberland 1.0000 0.0% Nottinghamshire 1.0115 1.2% Oxfordshire 1.0707 7.1% Shropshire 1.0000 0.0% South Yorkshire 1.0000 0.0% Staffordshire 1.0000 0.0% Suffolk 1.0027 0.3% Surrey 1.1336 13.4% Tyne and Wear 1.0000 0.0% Warwickshire 1.0245 2.5% West Midlands 1.0134 1.3% West Sussex 1.0177 1.8% West Yorkshire 1.0031 0.3% Wiltshire 1.0250 2.5% Technical Support Team - September 2017 Page 13

Annex C - A possible alternative to expenditure based regression Introduction For at least the last decade, and possibly considerably longer, funding formulae for local government have relied upon statistical regressions. Even those with an element of judgment, ministerial or otherwise, have been based on regression. Regression is a perfectly good statistical technique. In essence it determines the relationship between a dependent variable of any number of noncorrelated, independent measures. In the case of local government funding then the dependent variable we are modelling must be need to spend/demand for services. The independent measures should be drivers of demand or indicators of need. For example an intuitive indicator of ASC need is the number of elderly people in the population. Regression analysis will not only identify the indicators that have the greatest influence but also how much weight they should carry in a formula (i.e. how important they are). One of the key questions using any type of regression, including multi-level modelling, is to determine what our dependent variable is. In other words how do we measure need or service demand in the first place in order to establish indicators which model it well? This current formula review timeline indicates that the DCLG has the luxury of more time than in recent formula reviews to fully examine the current formula and suggest a replacement. Academics have already been recruited to carry out multi-level modelling in the Adults and Children s service areas, something which can be very time consuming. It would be remiss of the Technical Working Group to not use the time available to consider other alternatives which we believe to still be evidence-based but might also offer the chance to improve the effectiveness of local government funding as well as better understand the demand facing the sector. What is wrong with using historic spend or activity levels? In the past there have tended to be two primary datasets used as a proxy for need or demand ; past spending and past activity levels. To use these datasets presumes that systematic patterns of unmet need and/or unjustified supply can be isolated and controlled for. They cannot. Consequently, we have to question whether either spend or activity is a suitable measure to predict appropriate future service demand. It is argued that expenditure based regression methodologies have inherent circularity, i.e. they perpetuate existing patterns of service provision precisely because the allocation of resources to different types of populations will reflect the use they make of services that are already differentially available. Technical Support Team - September 2017 Page 14

All local authorities are different they all have political leaders with slightly different priorities, they serve different demographics and they often deliver different services. This will frequently be manifested in local authorities spending differing proportions of their budget on different services. For example, in some areas, schools will take priority over the roads, in other areas pressures in social care means libraries have to close. The same is true for many activities. People will only be able to use a bus route if there is one available. Children will only be attending a children s centre if there is one open near their home. In addition, each local authority has a different level of precept, different levels of grant funding and different levels of reserves. All of these things will mean that they spend differing amounts. However, spending decisions will always be constrained by the total funding available. Local authorities are required to produce balanced budgets. Reserves are finite and council tax is constrained. The scenario is simple; where services are better funded (relative to need) they will tend to be more available (relative to need). This will be reflected in activity data and result in models and allocations which overestimate the actual level of need. Services remain well-funded, activity remains high and a positive feedback loop is created. Conversely, of course, these models risk underestimating the needs of populations which have a lower level of activity precisely because of their lower levels of funding. There is an intuitive and inextricable link between past funding allocations and the levels of spending and/or activity in these years. Creating a regression formula which models future allocations by basing them on past patterns of spending or activity will only serve to lock in the current distribution patterns. If an area is under or over funded; this cycle will be perpetuated and fairness cannot be achieved. Promoting policy objectives This kind of regression- based approach not only tends to reflect and reinforce patterns of historic activity. It is also poorly equipped for promoting progressive policy objectives. This can undermine attempts to drive forward improvements in efficiency and service quality. For example, there is recognition that the strengthening of preventative services is the right thing to do. It will reduce demand for service further down the line and thus make monetary savings in the longer term and positively impact on quality of life. This policy is not reflected in funding allocation formulae. Financial austerity (as seen in recent years) strengthens the likelihood that resources will be focused on crisis intervention rather than prevention. Technical Support Team - September 2017 Page 15

There are strong grounds for developing an approach to formula funding that can provide policy levers to explicitly shift the balance between prevention, low level demand and crisis intervention. What is the alternative method? In simple terms, this alternative method aims to shift the approach to creating a funding formula from what is to what ought to be. On the surface it shares many characteristics with the cost-driver suggestion coming out of the ALATS sub-group of the TWG. Something which local government members of the Technical Working Group are attracted to, primarily for its simplicity but also because it appears to breaks the link with past spending/activity. The first step would be to break local authority services into groupings, probably similarly to those already suggested by the TWG. Then in each area it needs to be established what the services should look like. For example, in the case of Adult Social Care services, the universal, preventative service would be linked to public health spending and might include subsidised Gym membership, smoking cessation support and access to green spaces. The next tier of service provision might be more targeted; perhaps to those in poor housing, living alone or on low incomes. Then the next level might be for those already in poor health, a considerable distance from a hospital or those who are very old. In the fire service the universal services might include visits to all schools and carrying out safety audits, the next tier of more targeted prevention activity might be to carry out safety inspections where there are vulnerable adults identified at being at high risk of fire, or working with road safety campaigners to reduce collisions. It might also involve prevention work with youth groups in certain areas. This evidence, once collected, can be used to construct a formula that distributes funding universally for preventative work and then to the identified populations for the more targeted activities. Clearly local authorities would still be free to spend their allocations as they see fit but this information about what services should look like will be freely available to those that wish to use it. How to find out what ought to be Originally we considered that establishing what services should look like would be a significant hurdle to overcome. It would certainly be a lengthy process and unlike anything seen in recent years for local government. Technical Support Team - September 2017 Page 16

It would likely utilise surveys and focus groups talking to experts from representative sectors for each identified service. These experts might be service directors, involved in service procurement or members of performance teams. The work would involve discussions on prevention, what has already been proven to work (and not work), where investment is needed to stem the demand for expensive services and who/what the more targeted services would be directed at. In order to gain the support of the whole local government sector it would not be appropriate for one special interest group to lead on this work. We would suggest that the work be outsourced to experts in the field of focus groups and/or survey work. However, there may be an even simpler option, namely the What Works Network. A network set up to improve the way government and other organisations create, share and use high quality evidence for decision-making, the WWN aims to support more effective and efficient services across the public sector at national and local levels. The What Works Network The What Works Network (WWN) was set up in 2013 and sits within the Cabinet Office. One can find out more about its work on its website: https://www.gov.uk/guidance/what-works-network. The network is made up of seven independent What Works Centres. The centres are different from standard research centres. They enable policy makers, commissioners and practitioners to make decisions based upon strong evidence of what works and to provide cost-efficient, useful services. The current centres cover the following policy areas: Health and Social Care Educational Achievement Crime reduction Early Intervention Local economic growth Improved quality of life for older people Wellbeing Some of the network s early work included conclusions such as: Technical Support Team - September 2017 Page 17

Early Intervention - The Family Nurse Partnership programme has been shown to be effective in the US for improving children s health and development, with the benefits outweighing the costs by around four to one. Local Growth - Whilst they have intrinsic social value, the local economic impacts of major sporting and cultural projects tend not to be large and are more often zero. More investigation is required to establish whether the WWN findings can be used in constructing a new formula. The Survey Option If it emerges that the WWN is not going to be able to provide answers in the format needed, then, as described above, it may still be possible to obtain the answers through the use of surveys and or focus groups of experts in each service. Not only could this work be used to establish the what ought to be but also to ascertain the current cost/demand drivers and their relative importance in driving service demand/need. Clearly it is the Technical Support Team s suggestion that this information be used to derive a new evidence-driven future-proofed formula based on best practice but it could also be used to determine and weight the cost drivers of current activity. This would help to progress the current ALATS model past its present hurdle where cost drivers have been arrived at but there is no evidence in order to set weightings. Other considerations in a formula This alternative method would not impact on the considerations currently ongoing with regard to the treatment of resources, transitional arrangements and the area cost adjustment. This is simply a way to predict the need to spend, based on evidence around what works, how services should look, including a shift of focus from intervention to prevention. Technical Support Team - September 2017 Page 18