Local Government Finance Inquiry
We continue our local government finance inquiry, for which we have been joined, again, by Rita Hale, whom I welcome. It is also good to see Sylvia Jackson again. Welcome back, Sylvia.
I welcome our first witness, Professor Glen Bramley. The procedure is that you give a short presentation, after which I will open up the discussion to the committee for questions. Over to you.
Professor Glen Bramley (Heriot-Watt University):
I thank the committee for the invitation to speak on a subject that has been of quite long-standing interest in my career. I apologise for the late hour at which my submission was delivered—I hope that members have had a chance to read it—and for its length. Had I had longer, I might have produced something shorter, as the saying goes. At the beginning of my submission, there is a summary of just over a page, which I will speak to. Perhaps we can enlarge on particular paragraphs of that summary at the questioning stage.
Without further ado, I will run through the main points about Scotland's grant system that I want to emphasise. First, there is a choice to be made on how far the system emphasises a formulaic approach to grant distribution. Opposed to that is an approach that relies more on bids and proposals being made by local authorities, to which central Government responds in a discretionary way. Under such a system, central Government makes decisions to allocate resources selectively to some areas in support of some projects and not others. That would be a more competitive model, as it were.
There has been quite a lot of explicit debate about the more competitive approach in England. In Scotland, there has been a similar development towards competitive allocation of grants for particular areas and initiatives. My general advice is that some restraint is required in that direction; I re-emphasise the virtues of a formula-based approach to the allocation of grant for the mainstream services and for the main part of the grant. Although there are particular roles and circumstances in which a more selective and competitive approach may be more appropriate, such allocations should be time-limited and should be rolled back into the main grant and consolidated later on. We should not have an endless proliferation of special allocations.
As we all know, the grant system is designed to equalise. It tries to put different local authorities that are in different circumstances on a common basis. It tries to pay enough grant to recognise differences in needs, in the cost of provision and in the resources that are available for local taxation. The consensus on the need for such equalisation is pretty strong—I note that the committee's other witnesses this afternoon take the same view—and, in a broad sense, that consensus extends across Europe and into other countries as well.
However, the devil is in the detail. There is less consensus—and perhaps less clarity—about exactly what that means in practice when one looks at the formula for a particular service. Grant-aided expenditure aims for the general goal of a similar standard of service, but what is the nature of the standard that we are trying to equalise? That is a big question, which is not always addressed when people look at particular formulae. That question goes to the heart of the matter. We need to ask: what are we trying to achieve through the money that we are allocating? What are the Government's policies? What policies does the Government share with its local government partners? What are we aiming for? We need to try to capture all that in the notion of a standard and we need to capture what it costs to deliver that. Sometimes, there is a missing link.
I usually illustrate that point by referring to education, which is the most important service that local government is involved in. Is the similar standard of service simply to do with the inputs, such as the number of teachers? Is it something to do with the outputs that we are trying to produce, such as delivering a certain curriculum? Or is it to do with the outcomes of the service, such as the proportion of children who attain a certain level in standard grade or other test regimes or who get into further education? According to how the standard is defined, quite different answers are obtained on the nature of the formula and how much it should discriminate between different areas.
The Scottish system exhibits the phenomenon of having a lot of separate formulae for different bits of money—in other words, disaggregation. My view is that the Scottish system has perhaps gone too far in that direction. It seems difficult to believe that 93—or whatever the current number is—separate formulae are required, and I urge a consolidation into fewer service groupings. In particular, I do not think that it is appropriate to split up the inputs to such services as primary schooling—for example teachers' salaries, the school premises and the security services. Decisions on those inputs should be managerial decisions and should be taken in the light of local knowledge. There should not be separate formulae for such things. More noise could be introduced into the process of assessment by making such splits, which would not be helpful.
In some circumstances, particularly involving smaller services, splits could be helpful—for example, if there is a need to consider what the main bits of the service are, what the service is trying to achieve and who the main client groups are. In such circumstances, it could be helpful to break up some of the services, but I would not necessarily go as far as the English have gone. In England, an enormous bundle of "other services" are all thrown together—I think that that is going too far in the other direction.
Another important practical aspect of the system is the basis of the recent move from one-year settlements to the three-year settlement. I am sure that people in local government have generally welcomed the new length of settlement as a move towards stability and a way to facilitate medium-term financial planning. However, we need to consider certain implications of that change. First, at the end of the three-year period, a new set of distribution formulae will be needed. There is plenty of time to work on those formulae, but whatever they are, and even if they do not change very much, there will be quite a big bump, or change, in distribution. That is because the data will have changed—there has been a census and the demographic numbers will have changed. There will probably also be changes in the formulae. As a result, there will be some big redistributional changes at the end of the three-year period. Damping will have to become a fairly permanent feature of the system, as it is in the health service.
I admit that the other part of the settlement—the reciprocal commitment of local authorities to make only moderate council tax changes over the three years—seemed to be taken alongside other continuing features of the system. That is what I might call a form of quasi-capping. In other words, we have got rid of the formal capping of local budgets, but we still have what is, in effect, a similar system. Together with some of the committee's other witnesses today, I wonder whether that gives enough scope for local decision making.
The system in Scotland, as in other parts of the UK, has always tended, and increasingly tends, to treat the actual level of expenditure that local authorities have incurred as the primary criterion for judging the distribution of grant. I do not accept that that should necessarily be regarded as the primary criterion; we should always go back and consider the service standards that we are trying to achieve. If those standards have been achieved, and it is just a matter of fine-tuning within the existing distribution, that is okay. However, if what we are trying to achieve is a bit different from what has been done in the past, just looking at the criteria for those standards and at expenditure will not be enough.
One of the characteristics of capping—actual capping or the quasi-capping that we are getting into now—is that we tend to get a progressive convergence of expenditure and GAE. That means that, in a way, the process becomes circular and self-feeding. We get less new information from looking at expenditure, because we effectively start to get last year's or the previous year's GAE. That is one of the reasons why you might want to consider alternative methods of investigation and research, and alternative criteria for what need might mean for the purposes of grant distribution.
The technical side of analysing expenditure has tended to involve the use of the statistical technique of regression analysis. I do not know whether the committee has discussed that much and I do not want to go into the technical details, as there is a large amount of academic literature on the subject. Basically, the technique is a way of giving expression to the general approach to reflecting past expenditure in the grant system.
There are a lot of technical problems. I was involved in some work on the social work assessments, including the possibility of calibrating them on the basis of expenditure data. That reinforced the lesson that there are a lot of problems in the process and further convinced me that we should consider alternatives. My memorandum outlines a number of alternatives that the committee might wish to consider in the three-year interregnum before the next significant new distribution. There is an opportunity for you to do some slightly more serious research than the quick ad hoc exercises that have sometimes been done in the past. A number of approaches, which differ from past approaches, may be appropriate for particular services—I am not suggesting that all of them will be appropriate across the board. Most of the attention will be on the higher-spending services such as education and parts of social work.
The methods include examining individual survey data, such as those provided by the new Scottish household survey, and the data on service usage and expenditure at small-area level, such as postcode sectors. Examining the variation within local authority areas would give us more evidence of how need is reflected in spending. My submission also talks about methods that are designed to separate out the effects of variations in efficiency and methods that examine the outcomes of services. Perhaps the growing amount of work that is being done into educational outcomes, such as pupil attainment, could be used to inform resource allocation. There is a lot of scope for more research in that regard, but, as an academic, I would say that, would I not?
The submission comments on the two ends of the geographic spectrum in Scotland. Scotland represents the remote rurality phenomenon more than other parts of the UK, although not necessarily more than other countries, and also demonstrates the extremes of urban conditions. As I have said, the process is about standards of service and the type of provision, but extremely remote and rural areas lead us to question whether the notion of a uniform package of services makes sense in all circumstances, and to conclude that it does not and that things should be done differently in sparsely populated areas. It might make sense for sparsely populated areas not to provide some services, in which case perhaps the residents of those areas should be compensated for the fact that they do not have access to those services, although they might have access to other amenities. My submission contains a graph detailing council tax rates over the past five years for differing classes of local authority in Scotland. It is probably dangerous to suggest this, but that information suggests that the islands in particular and the remote areas generally have not been ungenerously treated by the grant system.
The paper contains some inconclusive discussion of the role of cities. I think that the Executive and the Parliament will be considering the role of cities in Scotland in the coming period. A commission may be set up, which ought to consider funding and other issues, such as the tax base, the funding of major infrastructure and the role of population changes in determining resources.
I want to draw attention to deprivation. The Scottish Executive's policies put a lot of emphasis on social justice. Some evidence in the paper suggests that the GAE system is not as responsive to deprivation as might be expected, given that policy emphasis. That issue would repay closer investigation.
As you know, the committee is inquiring into local government finance. We have with us Professor Peter Smith, from the University of York, whose paper suggests that there should be a development of spending assessments based on small area analysis and multi-level modelling techniques. Have you used those approaches in your work? If so, what were the results? What would you consider to be the strengths and weaknesses of those approaches? In which circumstances—or relative to which services—would you consider it appropriate to use such small area analysis and multi-level modelling techniques?
There are many questions within your question. Peter Smith is more of an authority on this than I am; he has been involved in two substantial research projects that have led to practical, implementable systems of resource allocation that follow from that methodology. I cannot claim to have done that, although I did quite a big project for the Department of the Environment, Transport and the Regions in selected cities in England, to examine how far one would estimate spending at small area level across the range of services.
That approach is one of five possible areas for development of the system. I would not want to overemphasise it at the expense of the others. The greater use of individual household survey data is as important an avenue for development—indeed, it has some similarities. Multi-level modelling techniques can be used at the individual level; it just involves going down a further level to the ultimate unit, which is the individual. To identify the people who are using services and who need to use services—the individual level—is in some ways superior even to the small area approach. One overcomes what statisticians call the ecological fallacy. There are many people with a characteristic in an area and much use of services—or spending—in that area, but they may not be the same people. If we go down to the individual level we have gone a step further.
I would not want to emphasise the small area approach, but I think it is quite a good compromise tool that can be used for certain services. Clearly, some of the social work services that are delivered to people at home would be amenable to that approach, although they would also be amenable to other approaches.
There is another version of the small area approach. If we consider education, the primary focus for analysis would be the school. There are technical issues about how we can relate schools to their catchment characteristics. Again, we can do multi-level modelling in that context, but we may wish to go down to the unit of the pupil and recognise the influences at all levels. It is one of a family of partly-related approaches which depend on the practicalities of data availability. Over 10 years, the health service has established a system of in-patient statistics, which records the postcode of patients. The health service can estimate the cost of in-patient services at the small area level fairly well. We do not have that kind of mandatory system for most local government services. Perhaps we should; and perhaps there has to be more top-down or uniform practice before we can have data that are uniform enough to analyse.
In Peter Smith's research on social work, he was able to consider only a quarter of the social work authorities in England that had what were considered to be the best data. Even then, there were problems with the data. I am, therefore, not sure that enough data are yet available to make the approach fully practicable. We must try to get closer to the relationship between need and the service that must be delivered to meet that need. We must be less confounded and confused by policy differences between local authorities—and possibly efficiency and organisational differences as well.
I would like to ask about the amount of data available, regardless of the form of assessment. You give the impression that local authorities are not yet in a position to deliver the data. I do not know whether that is just my interpretation or whether that is what you actually said. However, regardless of the situation in local authorities, do you believe that the Scottish Executive is in a position to collect the data on which to base assessments? Would there be any problems with that? Would the assessments be service-specific? How could any problems be overcome?
Data availability is improving quite a bit, after having been a bit moribund or even going backwards. Some of the data that we rely on are compiled by local authorities through their research and policy units or through the routine performance of their various statutory responsibilities. It is important to get greater consistency through auditing and checking of that kind of data.
A problem in local government in recent years has been pressure on staffing levels. Less staff have been available for basic research and data functions. Local authority staff with those skills have often been diverted into the kind of bidding exercises that I talked about earlier—bidding for the various challenge funds and for this, that and the other. There is an argument for getting those people back to the basics of providing good data.
Some other developments have been helpful. A number of administrative systems are maintained nationally. The Department of Social Security now routinely makes available data on claimant counts in different categories—in local authority areas and in smaller areas such as postcode sectors. Building on that, there is an initiative—called neighbourhood statistics in England and something similar in Scotland—to try to have a comprehensive set of data that will be available fairly soon for smaller areas as well as local authority areas. The Scottish household survey is another major initiative. That can be adapted over time with additional questions being asked on particular topics; the questions do not have to remain the same every year. That survey can give a lot of coverage.
For some indicators, once the sample is big enough you will be able to see the picture for virtually all the unitary authorities. However, that is perhaps less important than being able to analyse how particular characteristics relate to particular types of household and particular areas and neighbourhoods across Scotland.
The data situation is improving. I do not know whether a joint liaison committee for statistics exists between central and local government, but sometimes that kind of central vehicle is required to agree on things that need to be done jointly. Getting all the local authorities to go along with that would be helpful.
Statistics on input costs are another contentious issue—specifically labour costs, but also other input costs. Apparently, that is an issue in England at the moment. Is an examination required of the different input costs, and their impact on service delivery, in different geographical areas of Scotland?
In Scotland, that has been an issue mainly for the islands. Special investigations into the island allowances have considered data on input costs, among other things. It may emerge as an issue elsewhere, as the Scottish economy develops unevenly and regions such as greater Edinburgh are in danger of overheating, in comparison with other regions. One hopes that, in the broad scheme of things, Scotland might be able to achieve a more even pattern of economic development and avoid the worst excesses of overheating that are now being experienced in London and the south-east.
The scale of the problem in Scotland is not yet of the same order of that in London and the south-east, which is serious. A whole industry of research, consultancy reports and counter-reports has grown up to argue the case to and fro. Overheating has tended to follow the economic cycle: the more the economy booms—as it is at the moment—the worse the problem of cost pressures and the difficulty of recruitment. The experience in London and the south-east is that the official allowance of London weighting is not sufficient to buy labour of equivalent quality or stability. There is now a serious problem in maintaining services in those areas. Let us hope that Scotland can avoid the worst of that.
In 1993, you wrote "Variations on a theme: Needs assessment and grant distribution in six European countries". Would any of the needs assessments that you analysed be suitable for Scotland? If so, which?
Gosh. I am very impressed by your background reading. A more recent, up-to-date study of about 20 countries was conducted by PricewaterhouseCoopers for the Department of the Environment, Transport and the Regions. If you are interested, you might want to have a look at that as well. That study is more up to date than the one that I published, the conclusions of which I am trying to remember. In general, the systems in England and Scotland have tended to go further than the European systems in giving attention to the assessment of expenditure need. In some ways, therefore, the European cousins have tended to look to us, rather than the other way round. However, they might not have liked what they saw when they looked closely.
Those European systems were most interesting for the way in which they dealt with other issues, such as resource equalisation, the degree of equalisation that was employed, the local tax bases that were used and the nature of the compact between central and local government—what was written into the constitution. Those aspects were the most interesting parts of that study. However, I am not sure whether there is much that you could pull off the shelf. Subsequently, some European countries have become more interested in developing their systems further to deal with the kinds of issues with which our system has been grappling, such as deprivation.
Thank you. On page 16 of your submission, you write that local authority policies
"are increasingly motivated towards attracting and retaining population, particularly working households who may be expected to bring more financial resources in than they consume in services. This may not be a bad thing, but it could lead to some distortions in planning decisions, and the competitive stance of authorities might militate against rational strategic planning at the subregional level."
What distortions and irrationalities did you have in mind?
As you will know, the major urban regions of Scotland have moved from a system in which regional authorities were the strategic planning authorities, responsible for structure plans for the regions, to a system in which there is more or less voluntary co-operation between the unitary authorities in the performance of that function. We and colleagues from the University of Glasgow have been examining that in the course of a research project that is part of the city's programme for the Economic and Social Research Council.
We are quite worried that such a mechanism is not adequate, perhaps where harder decisions about priorities for development within the region have to be taken. That is the basis for the second part of the comment that you just quoted. As for the first part, some key environmental assets, such as greenbelt land, might be squandered as a lot of authorities try to release the most attractive greenfield sites to attract buyers.
So your basic concern is about making short-term decisions instead of planning for the long term.
Yes.
In your experience, might such a situation lead to the replacement of large council housing estates by new owner-occupier estates that lack a combination of infrastructure, such as recreation, leisure and education facilities and so on?
The provision of large public housing estates without any infrastructure was very much a characteristic of Scotland in the 1950s, 1960s and 1970s and is a historical example of the sort of thing I am talking about. We are not building public housing of that kind any more and I think—I hope—that those lessons have been learned. Some quite large owner-occupier estates are being developed which might not have an adequate infrastructure, and furthermore are not adequately balanced in social terms and do not have more affordable housing within them. Perhaps we are straying a wee bit from the point.
That said, my evidence is trying to draw attention to the fact that there are connections between planning and urban development and fiscal matters.
The reason I asked that question is that I am concerned that the new, sprawling owner-occupier developments in Glasgow seem to be repeating many of the mistakes of 30 years ago, only this time the housing is owner-occupied not public. Thank you for your response.
Professor Bramley, in your written submission you acknowledge that there is widespread dissatisfaction with grant-aided expenditure. Can you give the committee more detail about the advice that you would give to the Scottish Executive about GAE reform?
As I attempted to highlight the main points in my summary, I will re-emphasise just some of them. Things are too disaggregated, and we should consider consolidating some of the separate formulae. There is an opportunity to research—and a case for researching—some of the main spending services rather more thoroughly, before any decisions are made about a new set of distribution formulae that would be implemented from 2004. I have suggested some avenues for research that could be pursued.
On the whole, such research would move away from the traditional emphasis on trying to mimic the past pattern of spending between the individual local authorities and would try instead to understand more closely what determines need at an individual or small neighbourhood level, with a view to incorporating such data into the formulae. Perhaps we should concentrate only on some of the major services; it is not necessarily worth undertaking much expensive research on small headings of expenditure.
As I have said, the system of resource allocation should particularly reflect the various manifestations of deprivation such as concentration and so on. I suspect that the education service probably requires the most attention in that respect. Quite a bit of attention has been given to social work, although the outcome has not been very satisfactory and it might merit further attention, but education is the area to examine.
I believe that there are currently 93 spending blocks. How many blocks would you envisage—10 or 20?
It is difficult to say. I think that 20 or 30 blocks could be justified. Sometimes there are good reasons for breaking up some of the medium-sized services into their constituent parts and identifying a key client group for each one.
A service such as environmental health, for example, caters to a number of different sectors or client groups. We can talk about one formula with four factors or four sub-formulas with one client group each. That is not a bad way of dealing with some of those medium-sized services, which is why there might be 20 or 30 spending blocks rather than five or 10.
How would you group the services?
Partly for practical reasons, I do not suggest doing anything radically different, for example grouping client groups such as children and the elderly. We might want to investigate that but, because various services operate under different legislative and professional frameworks, and because guidance and the data systems are geared to those services, it is inevitable that the major client groups within the traditional service categories—primary schools, secondary schools and so on—will be used.
Some councils have moved towards using thematic schemes such as on children or the community. Will you comment on that?
As a hard-bitten researcher, I am of the opinion that it takes a long time to change data systems. We want data systems that are consistent throughout Scotland, which means that there must be consistency with traditional accounting categories in relation to the returns for health, social work and so on.
There is merit in taking a cross-client group approach. My deprivation example is relevant in this context. Deprivation has wide-ranging and varying effects on many services, but it tends to get overlooked in a sectoral client-group approach. From that point of view, Mr Harding is right to say that we should consider cross-cutting issues. However, accounting data or client data must be in a form that can be related to the long-established structure.
Your job will become more difficult in that case, because there will not be consistent data from the councils that are moving towards using thematic schemes.
Local government needs to address that issue more seriously. When we conducted our small-area spending study in England, we found that there was a trade-off. In relation to the authority that had the most radically different structure for service provision, which was a London borough that had used contracting and separate service units, it was difficult to produce data that would fit the more traditional framework. Some data were good on some services, but there were gaps in other areas. That illustrates the problem. As I said, data have been in some ways neglected and local government must with the Government's statisticians make a concerted effort to maintain common standards. It is not all bad news, however. More new data are becoming available in some respects.
Should it be possible to produce a set of spending assessments that reflect the needs of all areas and councils, whether they are deprived, rural, urban or whatever?
It is possible and it is being done. There must be safety valves, however. Certain new initiatives must be funded by other mechanisms, such as challenge bidding arrangements and specific grants. Most expenditure must be covered through the kind of mechanism that we are discussing, if such a mechanism is developed sensibly.
Will it still be necessary to have arrangements to address issues such as the islands?
We would have to audit whether the system is dealing with the islands adequately and whether some of that should be done through a separate allowance or through the main formula. I do not have strong views on the subject.
Thank you for your submission, Professor Bramley. I now have a fantastic overview of the matter, although I have not read all the detail. I am particularly interested in the suitability for use in Scotland of alternative approaches to assessment. You mentioned first regression-based assessments and the new approaches to the use of regression analysis. How suitable are they for use in Scotland?
It is becoming more possible to know that now because more quality and standards indicators can be included, whether from the household survey or from the best-value indicators. There have been problems with the expenditure data not being consistent or reliable, post re-organisation. I do not know whether, four or five years on, the problems have settled down or whether they are still plaguing the system. I would not be surprised if some gremlins still exist in the expenditure data.
I still regard such approaches as second best because they emphasise that the ultimate criterion is expenditure, and I am less enamoured of that than I am of other criteria. I prefer method (b)—the analysis of individual household survey data—for services that are delivered to individuals or households in the community that have particular needs. If it can be made to work, I favour that method over method (c), although there are some cases in which method (c) would be practically feasible, whereas method (b) would not.
Method (d) is mainly a way in which to explore efficiency. Therefore, it might be more relevant to issues that are in the context of best value, rather than in the context of needs assessment. It may be that method (d) could yield some insights and give one greater confidence in certain formulae. Indeed, it could be a way in which to develop formulae for needs assessments, but it has not been developed to do that so far. It might reassure us that we are not rewarding inefficiency and that we are taking proper account of matters.
You referred to method (d). Which approach did you mean?
Frontier functions.
Can you explain a little more about frontier functions?
Local authorities or other organisations that provide services produce service outputs, through combining different inputs. Some authorities do that more efficiently than others. The most efficient come to what the economists call the "frontier". In other words, they are getting more output from their inputs than others.
Such a process has always intrigued me. Would an efficient council with good outputs be disadvantaged under a formula because its outputs were good, as a result of which it needed extra money to increase the outputs?
Such a council might be disadvantaged if the formula were based on traditional expenditure criteria and traditional regression methods, whereas it should be treated correctly and it should take the approach that identified an efficiency issue. The ideal grant system is one that assumes efficient provision or at least relatively and equally efficient provision.
You said that method (b) might work better than method (c)? Will you expand on that?
That takes me back to an earlier point. Need is essentially a characteristic of the individual and the household of that individual, such as an elderly person who needs support with daily tasks. In general, that comes under the heading of home-based care-of-the-elderly services. If that is mainly a function of an individual's circumstances in his or her household, an analysis of the level of the individual's and the household's need—using the data on the 15,000 households in our survey—will pinpoint most reliably the relationship between those needs and the circumstances.
However, if an analysis were carried out of the level of need in postcode sectors, which have an average of 2,000 or 3,000 households in each sector, there is some danger of what statisticians call an "ecological fallacy" occurring. A correlation could be made between, for example, an area having many income support claimants and a high level of take up of a service, but those might not be the same people. Such a situation could arise almost by accident.
In your conclusion, you talked about the need for greater local financial autonomy. Will you explain a little more about what should be happening? In particular, do you consider that the local tax base should be extended and if so, by what methods? What should be the appropriate balance between central and local government support?
I, like others this afternoon, favour greater local autonomy. Why do we have local government if we do not believe that it can make better decisions locally? There are good grounds for arguing that, if we have an appropriate financial base and a link between expenditure and revenue raising, a sounder relationship will be built between accountability, decision makers, their electorate, local communities and local businesses. There are also technical advantages to having a bigger tax base; there is a less-sensitive, more highly geared relationship than the present one that exists because the local tax base is narrow.
Historically, the majority of resources were raised locally and, until 12 years ago, half the resources were raised locally for revenue expenditure. Such a local revenue system operates in many other countries. However, if one level of government has more autonomy than another, it must act responsibly. The system must promote accountability and well-informed decision making. Such a system also means that another level of government has influence, but less direct control. That is where the tension lies. I regret to observe that, throughout the United Kingdom, central Government is reluctant to relinquish so much control. Indeed, it wants to exercise much control over the delivery of local services; that has become a central policy issue.
I am straying into the political arena, so I shall not continue down that track. As for whether the local tax base can be extended, the council tax is reasonable. It could be modified, but not raised much higher than its present level. I favour giving some part of local business-rate revenue to local authorities. I do not support giving it all to local authorities, as happened pre-1989, but local discretion could be brought into play. That would help to build a more direct relationship between local authorities and their business communities.
One thing that is lacking at the moment is the ability for a local authority to act entrepreneurially and to invest for the future of its area. How can it get that investment back? The business rate could offer one way. There are other forms of taxation that relate to the environment, traffic and transport, for example. They can be somewhat controversial, but they are logical and they have a role to play. In some cases, the price of having the revenue sources is that they might be hypothecated, and it might not be realistic to expect some of those revenue sources to be available without their use being directed to the sector in which they were being raised. Overall, the model of hypothecated taxes and charges could play a bigger role.
There is no technical reason why we should not have local income tax, but the politics of taxation is such that I do not realistically expect that that will happen tomorrow.
This point follows on from the idea of the way in which responsible local government can be achieved in relation to the grant distribution system. A reading of your paper indicates that you support the retention of a formula-based distribution system. Will you explain what you consider to be the strengths and weaknesses of each of the types—the bid/plan system and the formula-based system? Are there any circumstances under which you think that the bid/plan system would be appropriate?
The strengths of and case for the bid/plan system are that it encourages local authorities to plan—which might be a good thing—rather than taking traditional one-year, incremental, ad hoc budgeting decisions. It also encourages them to evaluate options, to look more widely and to do the sort of things that they are urged to do as good practice, such as work in partnership with other agencies. It might be that some incentives are needed to encourage local authorities to do more of that.
There are particular circumstances in which bid/plan systems are most appropriate. First, there are new initiatives that are being started on a pilot basis, in which close joint working that is being tried out experimentally in selected areas between central Government and local government can facilitate learning about what works and what does not. Secondly, there are large lumpy capital projects, in which the infrastructure might affect wider areas and in which it is important to go through appropriate project appraisal procedures. Thirdly, there are needs or problems that affect only a small number of authorities, for example flooding or problems that affect the islands.
If we believe in local autonomy and responsibility, we must trust local government to get on with it. We must be a bit hands-off and allow for variation, deviation, diversity and learning at local level. That is part of the overall picture; we must let go. I think that a formula-based system allows that. Such a system says, in effect, "Here's the main bulk of the money. Use it and get on with it." Of course, it must be spent within the legal framework within which an authority must operate.
The bid/plan system leads to ministerial decisions on individual allocations, and tends strongly towards a kind of punishment-and-reward syndrome for following—or not—the particular nostrums that central Government might be promoting at the time. That is why I believe that that system is less compatible with local autonomy or with a division of responsibility than is a formula-based system. Formula-based systems are generally likely to be fairer and more equitable, and to be seen to be so.
I take it that you also want less hypothecation in the formulae.
I mentioned an example of where we might require a bit more hypothecation to make progress—transport is such an example. I am not absolutely agin hypothecation in all areas.
It might be that central Government is more interested in some services than it is in others. Understandably—perhaps rightly—national standards and control are more important for some services than they are for others. In recognition of that, we have long had about a 50 per cent police grant, and grant aid for certain other matters. Housing benefit is largely funded by specific grants. It is not unreasonable to look for a certain amount of variable geometry or different treatment of different groups of services. However, many services are of mainly local interest and involve significant local discretion. In such cases, a less interventionist approach with less specific grant might be appropriate.
From the various projects that you have undertaken, have you reached any conclusions on local government finance in Scotland, other than the many that you have mentioned?
There are probably enough in my submission to be going on with; some of them would repay further investigation.
If the Scottish Executive commissioned work on new needs assessments in, for example, 2002, would you expect the Executive to be able to use those new needs assessments to distribute grant? Would it be sensible to introduce new needs assessments before the 2001 census data had become available?
It would certainly be technically feasible to commission research that might be undertaken during 2002 and into 2003. Such research could produce useful results that might be considered as a basis for formulae that could be used for the 2004 settlement. Census data would, I hope, be available by the end of 2003. So, in developing the research, the Executive would need to work with pre-census data. Formulae would need to be rechecked and recalibrated as census data became available. That might lead to some changes in some of the distribution outcomes, but it would probably not fundamentally affect the basic tenor of the research.
The research would add a further reason for significant changes to the distribution in 2004. However, the distribution will change quite a bit anyway, because of all the data factors. There is bound to be damping over a period of years, as there is in the health service, for example, with the Scottish health authorities revenue equalisation formula—the SHARE formula.
In an answer to Keith Harding, you said that data might become more difficult to produce and that local government and central Government would have to talk to one another. The committee would be absolutely delighted if that were to happen daily.
In an answer to Sylvia Jackson, you said that councils should be treated correctly. Again, the committee would agree with Dr Jackson's worry that councils whose outcomes were positive might lose money in the distribution. The committee would also agree that there should be a greater level of local autonomy. We would pursue that.
Thank you for your time. If we need you back again, we will ask you.
Thank you for the opportunity to speak to the committee.
We now have with us Professor Peter Smith from the University of York.
I am delighted that you have taken your jacket off, because it is very warm. I ask you to give us your presentation and I will then open up the meeting for questions. You have been sitting at the back of the room, so you will have got the idea.
Professor Peter Smith (University of York):
Thank you very much, convener, and thank you for the invitation to attend.
Local government finance is an important issue. I have been working in the area for about 20 years and I have seen many of the shifts, changes and disasters that have occurred over that time. I have given the committee a short submission. I believe that members are especially interested in the work that my colleagues and I have done on resource allocation and grant allocation methods.
I have been involved in the much broader area of local government finance for a long time. A lot of our specific work has been in the health care sector. That reflects the fact that the health care sector tends to fund research much more generously than the local government sector does, rather than any preference on my part. The submission concentrates on trying to calculate what I perceive to be the general desire to achieve a fair grant for local authorities. That opens up a lot of issues, which I do not propose to go into now.
The broad thrust of most grant allocation mechanisms in the UK is that one is trying to enable local authorities to provide—given that they adopt standard practices and standard levels of efficiency—a standard level of services; the grant is an enabling device. The traditional way of doing that has been to examine the expenditure of local authorities to tease out the average expenditure across the country, taking account of the needs within local authorities. The approach has been to try to find an average response to needs within local authorities. That approach assumes that local authorities are, on average, meeting the needs of their community and that, in delivering services, local authorities are not systematically skewing expenditure against certain sections of the community. There will be fluctuations between policies, efficiency and perceptions in local authorities, but the use of actual expenditure data assumes that, on average, needs are being met.
It may be felt that certain sections of the community are suffering because their needs are not being met by the local government sector, but one cannot, as a matter of national policy, use existing expenditure patterns to allocate grant to those sections of the community. One must move towards a grant that is based on issues other than historical expenditure. The committee may want to come back to that point; I would be happy to elaborate on it.
What methods should one use? As Professor Bramley said, the UK has, in many ways, led the world—certainly in the intricacy of its methods if not necessarily in how well they meet the requirements. The Scottish example is perhaps the most intricate in the UK. I have a lot of respect for it—the notion of focusing on client groups has a lot to commend it. It is based on saying, "These are the people for whom services are being delivered, so let us try to model the needs of those groups." However, there are some serious technical shortcomings with the method as it is currently applied.
My submission to the committee focuses on the use of aggregate local government expenditure as the basis for estimating expenditure needs. For many years, my colleagues and I have felt that that is unsatisfactory, as aggregate spending is influenced by many things besides needs: history is an example, as expenditure patterns may take a long time to shift; local policies might influence expenditure to move away from the national standard; local levels of efficiency might vary; and even local accounting conventions might vary. The latter becomes especially important when we consider narrow services, which is the case with GAEs.
There are many influences on local spending that one would, in principle, like to wash out of the system of estimating needs assessment. The method that we have used in health care and social work in England—and in health care in Northern Ireland and other countries—is to go below the level of the local authority and to examine expenditure patterns in small areas. Within a local authority, needs will vary between small areas, although efficiency, local authority policies and local authority accounting choices will be fairly constant in them all. It is our belief that the expenditure differences within a local authority more nearly approximate a response to needs than the aggregate data, which are confused by the other issues that I mentioned. That is the rationale for using small-area data.
I agree strongly with Professor Bramley that the ideal would be to go down to the individual level. Sweden has a remarkable system of individual records—a census of individuals—that is kept live and contains information about individuals' marital status, housing tenure, income, illness and many other highly sensitive issues. In this country, that would raise all sorts of issues regarding privacy and personal freedom; nevertheless, the database enables Sweden to develop extraordinarily detailed estimates of citizens' needs. For example, a local authority can refer to the national records and find out how much a single-parent mother in a certain income band, who has two children, is in employment, lives in rented accommodation and has no long-standing illness will require in expenditure. The Swedes are able to do that because of their superb record system, at least for health care.
One could envisage a system in which we all had barcodes across our backs and our use of resources was tracked. From a grant-aided expenditure point of view, that would be the ideal. There would be practical problems implementing such a method, but it is important to bear it in mind. The small-area approach is an intermediate approach that is practical given the data that are available.
Thank you. You advocate the development of spending assessments based on small-area analysis and multilevel modelling techniques. Could that sort of approach be used to produce needs assessments for a range of local government services? You zoned in on health and social services. Would that approach be applicable to leisure services and all the other services?
That model is highly dependent on the availability of data. It could be argued that really one is interested in the total expenditure on local government services that a small area incurs, regardless of which services those are. That is a policy choice within local government. One simply adds up the entire expenditure across the range of services, examines how much goes to a specific small area and tries to correlate that to the needs profile of that area to get an aggregate model of expenditure needs.
In practice, that approach is likely to be unfeasible. One should look at the easy and large services first. Perhaps the easiest would be education, as one has a captive client group—captive is perhaps the wrong word to use, but one has some hope of at least counting the people in the group and considering their characteristics. An individual-based approach in school education would be feasible, provided that there was some ingenuity and some investment in data systems. We have demonstrated that the small-area approach is feasible in social work for children and in the police. One could argue that, because the area incurs the expenditure rather than the individuals in the police service, a small-area approach is more appropriate in analysing expenditure.
I have talked about the big services, in which the biggest immediate gains could be secured, but many other services are amenable to that sort of approach—especially those that are directed at individuals or specific areas. Leisure services are an example.
Have you applied small-area needs assessments in a country with the diversity of Scotland, where there are large, urban areas, such as Glasgow, and remote islands? There is often a huge variation in area types even within councils and the size of council areas can make a difference. Fife has fairly urban parts, but it has fairly rural parts as well. Can the model cope with such diversity?
I had the great pleasure of being involved in the Scottish review of the special islands needs allowance, so I visited some of your islands last year. The extraordinary variation in circumstances was striking, but the issues to which that gives rise affect only a tiny number of authorities. Developing a formula to accommodate issues that are not experienced universally is asking for trouble. Special circumstances—especially those in very remote and very urban, inner-city areas—may have to be treated separately. For example, in England, circumstances in Westminster are treated as special—a one-off.
Would it be possible to construct a sample group of authorities in Scotland for the purpose of analysis?
Yes. There is no reason why one should base a national formula on data from every local authority. For example, some local authorities might be manifestly failing in the service that is being considered; one would not want to include them in a statistical analysis. Work that was carried out on health care for the Arbuthnott review focused on the pattern of expenditure in a region where services were thought to be in line with the desired policy. That seems a perfectly reasonable response.
To design a formula that meets certain requirements, one should consider the local authorities that seem to be delivering services in line with those requirements. That goes back to the standard level of care. The formula is supposed to indicate a standard level of care. Perhaps one should use only data from those authorities that are delivering that standard, or at least a satisfactory level of service or care.
Can you indicate the likely cost of developing a full set of models for local authorities in Scotland?
Are you talking about the research cost?
Yes; the cost for the multilevel modelling techniques.
Whatever the cost, it will be microscopic compared with the sums of money that we are discussing. The major cost is for data capture—setting up suitable data collection systems and ensuring that the data are valid. I point members to the work that we did for social work in England. We won a contract from the Department of Health to undertake the work. The bulk of our expenditure was not on the statistical analysis, but on the footwork that was needed—going round local authorities, finding out what data they had, examining and verifying them and turning them into a form that we could use. The expense of implementing some of the new methods is contingent on the state of the information base. As I am not intimately involved with the Scottish system, I would not like to speculate on its information base.
A serious issue in the whole of the UK—this was discussed at a meeting of the Royal Statistical Society, where we presented some of our work last November—is the tiny core of research expertise. One cannot simply conjure up expertise at will. The sad fact is that Governments of all persuasions have neglected to nurture a research community that can undertake such work efficiently and quickly. Inevitably, there will be a problem in ensuring that satisfactory contractors can be found to undertake the research. However, I repeat that I think it will be feasible to carry out such research. Although the one-off amount involved across local government finance might run into millions, that is small when one considers the large sums of money that we are talking about.
Can you tell us the cost of the study that you undertook for social work and the Department of Health?
The cost of both studies was about £150,000 to £200,000.
In your evidence you advocate the extension of the local tax base. Do you have any views on how it should be extended?
There are some difficulties with extending the tax base, although I have no doubt that broadening it would ease much of the pressure on GAE or its equivalent because of the gearing effect. As GAE is so influential on the revenue that local authorities receive, it has come under inordinate pressure and scrutiny.
The move towards a local business tax would undoubtedly be feasible and desirable, but there is a difficulty—and this relates to income tax and to sales tax in particular—in that traders or citizens who do not like a local tax rate are becoming increasingly mobile. There is a difficulty in pinning down a tax base and ensuring that companies, traders and households are not able to live in one area that has low tax rates and use the services in an adjacent area, which has to levy high tax rates. That has been a profound problem in the United States where inner cities have progressively lost their tax base. They have had proportionately higher needs as the wealthy tax base moves out and a low-tax-base, high-need area remains. One must consider the options critically and carefully, but the general principle of a modest enhancement of the local tax base is desirable and should be explored vigorously.
You mentioned a business tax. Is that in addition to rates or are you referring to business rates?
I am sorry. I was referring to business rates.
What is the appropriate balance between central and local funding?
I thought the balance was broadly right in England in the early 1980s when the ratio of local to national sources reached about 50:50. It may be unfeasible to get down to that level now, but I think that that target would result in a much more balanced approach towards local government finance.
Are you advocating a move away from the uniform business rate and a return to the bad old days when councils put business rates up by 122 per cent?
I hope that councils are much more enlightened nowadays about the implications of setting high business rates. However, that is one reason why the matter must be treated cautiously. Immediate and complete freedom to set local business rates at whatever level a locality chooses could result in some very adverse consequences. First, therefore, I recommend the power to be quite modest and circumscribed in certain ways. I have not thought through exactly how to implement that, but I am sure that the Scottish Executive has many ingenious options that it could put to the committee.
Earlier, you said that when you examine the GAEs, you assume that needs are being met. How would you address the present imbalance in many of the GAEs, which are well under what is spent by councils, and funds that are not allocated?
Are you saying that some councils spend much less—
There is probably a 20 per cent imbalance between what peripheral services such as libraries and leisure receive and what they spend.
That is local democracy. The GAEs add up to a total allocation of resources that the centre chooses to make available to localities. The essence of local democracy is that localities choose whether to use that money on services, as the centre assumes they do, use it in some other way or pass it back to council tax payers through lower council taxes. That is a matter of local choice, which is subject—and this is perfectly legitimate—to the centre placing certain requirements on the money that it disburses, in the form of ring-fenced funding, for example. That is equally valid. However, once the centre lets go and says that a matter is for local choice, it should be so.
There is a perfectly legitimate reason for highlighting and publishing variations from the norm; it helps local electors to gain a clearer idea of where their council stands in relation to others. They can then make an informed choice about whether that state of affairs is acceptable to them.
On small-area analysis, if satisfactory contractors could be found, what do you believe would be the time scale for the collection of necessary data and the implementation of such a system? That is in the Scottish context, of course.
The short answer is: longer than you would think. The typical time scale that we have been allowed in the projects that we have undertaken, where we have had to go out looking for data and verifying it, is approximately 12 months. The bulk of that is taken up with assembling the data. We have kept within those time scales. However, the results that one gets at that stage will inevitably come under intense scrutiny. There will be requests for additional analysis. I suggest that the entire process, from deciding to commission the work through to wrapping it up, would take 18 months at least—it is more likely to be two years.
That is assuming that satisfactory researchers could be found and that there were no information technology bottlenecks.
And that the information was readily available. For example, with the English hospital statistics that we use as the basis for some of our analysis, we have got to the stage where we can almost literally flick a switch at IBM and we will get the data that we need within two weeks. That means that we can do sophisticated, multilevel analysis for the Department of Health within three months of its requesting that analysis. However, that is because the data are sitting there for us.
That shows the time relativities. The analysis takes three months, but typically the collection of data—if one has to go out into the field—takes nine months at the very least.
What drawbacks, if any, do you see in the widespread use of multilevel modelling techniques in the development of the needs assessments used to distribute government grants between local authorities?
There is one major drawback, which we come back to time and again. It is the issue that I referred to at the start of this session—we are assuming that local authorities are, on average, meeting need. However, suppose that a group of the population—say, drug abusers—is systematically not getting the services to which Scotland feels that they are entitled, no amount of multilevel analysis will tell us where those services should be directed, because the services are not being provided. The expenditure is not there. Unmet need is the fundamental drawback to the method, as of course it is to virtually all empirical, data-based methods. One has to consider methods that are based on other forms of evidence. There are secondary drawbacks, one of which is the difficulty of undertaking the work, which is specialised—the expertise exists but it is spread thinly.
It can be difficult to convey to policy makers the essence of what is being done. We have gone through many years of trying to do so and we increasingly feel that it is feasible for us to convey the essence of the method used. Members might not believe this after today, but we feel that we are getting somewhere on that. However, the approach is novel and it takes a bit of persuasion.
Policy makers generally seem to be happy with ordinary regression methods, which form the basis for much GAE. My view is that regression methods are complex and subtle—many of us may not understand them as well as we think we do. The transparency of the method can be a problem, but that is the case with almost all technical methods in this arena.
We sympathise with you on that point. Much of our concern about the current system lies in the fact that very few people understand it. That is clearly an issue with the system that you have spent so much of your career working on and perfecting. How difficult would it be to convey it not just to policy makers but to the general public?
To be candid, I think that the general public will never be interested in or prepared to invest the time to get to grips with such issues. The English experience is that the general public do not really want to get involved in such things.
Simplicity is an important objective to bear in mind. There are four broad criteria against which a grant system ought to be judged. The system should look plausible; it should be statistically reliable; it should be simple; and it should not embody any perverse incentives. Simplicity is a desirable characteristic of any grant system. However, we are not dealing with simple things, and my view is that, if we have to sacrifice some simplicity in order to make an allocation formula fairer and more plausible, the price is perhaps worth paying. I do not think that the general public will express much interest, whatever method is used. However, whatever method is used must be transparent to independent audit, to peer review and to challenge, so that it can be debated properly among those with a professional interest.
How transparent and accountable do you believe such a system would be, relative to the current model?
The level of transparency would be a matter for whoever undertakes the work. They would have to ensure that the work is conveyed properly to those with a professional interest in it. When we had drawn up our formula for social services, we took a roadshow around England to help local finance people, local authority accountants and councillors to understand what we were doing. The feedback that we received was broadly positive—the roadshow was thought to have been of some help in increasing the transparency of what was being done. It did not make the system any simpler, but at least the assumptions underlying the analysis, and the strengths and weaknesses of what had been done, were better understood. The debate about the formulas that had been developed could therefore go forward in a reasonably informed environment.
The GAE system suffers from a number of transparency problems. To be honest, I have a problem with exactly what is going on with the secondary indicators. Part of that problem is that there are so many influences on the regression analysis, other than pure needs.
The other barrier to transparency in the GAE system is the large number of expenditure groupings. Whatever system is adopted in the future, I strongly urge that the number of those groupings be reduced. As I said, one possibility is to reduce the system to a single formula for all local government expenditure, but that is not feasible. However, there is a strong argument for significantly reducing the number of groupings.
The next question is from Michael McMahon.
It is okay, convener. Professor Smith has answered my question in response to other questions.
He did the same with Sylvia Jackson's question.
Professor Smith, you said that simplicity in the system was the most important objective. I think that we would all agree with that. The more that we go into this independent—as independent as we can be—review of local government finance, the more we realise how complicated it is. Simplifying the system would certainly help us and I am sure that it would help those who have to deliver the services.
I was interested in what you said about Sweden. The problem in Scotland would be getting people over the idea that we know too much about them, even if we were delivering much better services.
This has been an interesting debate and the members' questions have been good—at least the members had read your paper, which is always good from a convener's point of view. Thank you very much for coming. If we need to see you again, we shall invite you back.
My pleasure.
I welcome Jim Twomey from Pion Economics, who has been sitting patiently at the back of the room. I apologise for the fact that we are overruning, but the committee wanted to ask as many questions as possible. Thank you for coming along. You know the drill; you may make some introductory remarks before we ask you questions.
Jim Twomey (Pion Economics):
I agree with the witnesses who have already spoken, but I welcome the opportunity to provide an input to the review exercise.
Pion Economics has worked in local authority resource allocation for a number of years and now has reasonable experience of grant systems in Scotland and across Wales and England. We have undertaken a number of reviews in Scotland in the past two or three years, some of which have focused on specific services and some of which have contributed to attempts to improve measures of population distribution. Some of that work has subsequently been extended to assist standard spending assessment distribution in Wales.
We have made a written submission to the review, and I do not propose to indulge myself by making a verbatim statement. I would like to point to a few key elements in our submission before answering questions.
Having worked within the system for the past few years, we feel that, despite its complexity, a structure that involves a broad primary indicator per client group and a secondary indicator to reflect differential need or cost of delivery is, on the whole, rational. Clearly, that depends on the quality of the indicators and the extent to which they are able to detect some of the underlying complications within service delivery.
There are a number of problems with definitions, the first of which is rather obvious but no less important: not every service has an identifiable client. We focused on some of the major services, such as education and social services, and the nature of their client groups is clear. However, the client groups of other services are not so clear—not only are the beneficiaries of certain services less well defined, but it is clear that residents from outside an authority may participate in and benefit from some of the services that are provided by that authority.
In that context, an area in which further work is required is the important role that is played by many urban and regional centres in the provision of services of national or international standing. It is clear that in certain areas, such as Edinburgh, large numbers of visitors are attracted to certain events and imposed cost pressures may require additional consideration. Our experience under the GAE system is that those pressures are recognised, but if they are not captured to the fullest extent, as a result of data difficulties, they might go unrecognised.
A far more contentious debate resides in the arena of differential need and service delivery costs. The major problem, which has been mentioned several times, is the quality of level of service. The concept of a standard service underpins both GAE and the system in England and Wales. However, that concept is imprecise, rarely defined and open to significantly different interpretations. It is not surprising that authorities have substantial disagreements about the nature of the standards and the extent to which a pattern of service delivery reflects needs rather than choice. It is clear that, in many circumstances, authority A's need is authority B's choice, and vice versa.
Along a similar line, some of the financial information comes down to a single figure on a piece of paper in the current system. One has only to consider some of the services that are under discussion, such as leisure services, which range from bowling greens, where costs depend on how many times the grass is cut, to golf courses, ski facilities and so on. The range of services in any given GAE is huge, yet here we are with a single figure—say £X million—on a piece of paper, and that figure determines the profile of future GAE.
Some of the previous witnesses' comments also indicated that those issues create major difficulties for regression analysis, and there are substantial risks that the results that are being used contain statistical bias.
I want to raise another issue, although I am not sure about the jargon. There is substantial risk of what one might call a service reduction spiral. If an authority provides a relatively low level of service, perhaps from choice—we have discussed how important choice is—one might expect expenditure on that service to be rather lower than average. That clearly feeds into the next GAE process, given that there is an authority or a set of authorities with lower levels of expenditure. As a consequence, and all other things being equal, lower levels of grant will be given. What will we do with such an authority if it decides that it wants to improve its service? Because of the starting position, it is locked into a system in which it provides a low level of service and receives a low level of grant. That takes us into a difficult situation.
When I was sitting in the public gallery, I heard the previous discussion about other areas of improvement. Hierarchical regression models were discussed before I arrived and I do not want to make much comment about them, apart from saying that they represent a feasible way forward. However, there are difficulties. Such models place extensive demands upon management information services in authorities. Evidence has demonstrated that. Professor Peter Smith's estimate of 18 months to two years is probably about right. The National Assembly for Wales is having similar difficulties to do with personal social services. The techniques represent an improvement, but not necessarily a solution to all the difficulties that have been mentioned.
To follow up on something that was said a few minutes ago, one really has to sit back and evaluate whether the scale of any improvement is sufficient to offset the risk of the increased alienation that inevitably follows as technical procedures continually progress and those who are required to explain them fall further and further behind, in both their capacity to demonstrate their understanding of the process and their capacity to demonstrate the way in which they are accountable and can respond to pressures from their electorate. There are difficulties there.
A few years ago, I hoped that best value might begin to help in the taking account of non-expenditure service characteristics. I am not sure whether that potential has been overlooked, or whether it has been decided that the prospect of bringing together two huge structures—best value and resource allocation—is just too complicated to countenance.
Another area of difficulty—and one with which we have been very involved—concerns issues such as economies of scale, population dispersion and settlement size. In the past few years, we have had at least two reviews of population distribution and GAE. We have played a role in determining some of the developments and the indicators. Members will be pleased to know that I do not intend to discuss any of the detail of that. However, it is clear that population distribution is likely to be a key secondary indicator for some services, especially where the potential for economies of scale is limited and where duplication of facilities is necessary to provide adequate levels of services. Different services are probably unlikely to exhibit the same relationship to population density or settlement. It would be simplistic to expect that. Detailed studies are probably needed, at the level of individual services, to assess the patterns.
Our first experience of issues related to sparsity came with the English SSA system. We left that exercise with quite a deep sense of dissatisfaction about the extent to which our sparsity indicator was used. We left with deep doom on our foreheads. Coming to Scotland provided us with some relief, because we found serious attempts to investigate such underlying pressures. The solution may not be in evidence, and there is some way to go towards improvement, but there have been serious attempts to investigate underlying secondary pressures. That is welcome.
Is there room for further improvement? Clearly, the answer is yes. I suspect that the other witnesses would agree with that. However, sometimes we have to stand back and tell ourselves that no mechanical system of grant distribution can ever hope to replicate the complexity of reality in delivering services. We have to stand back and say, "Okay, we can develop the technology as far as we want, but the reality on the ground will never be replicated." We must recognise that.
What we can hope to do is make progress and attempt to accommodate key needs and key costs of delivery in a balanced way, minimising potential error in the distribution process. My appeal would be to maintain at least one foot in the practical, rather than walking too far towards the ideal.
Your company has done a lot of work in investigating the effect of population concentration and dispersal—sparsity—on the cost of providing local government services. Have you devised a robust method for taking those costs into account?
We were asked to conduct a review of the current set of indicators and we did so. During the exercise, one of the indicators was substantially revised. As members may or may not know, within the current settlement indicator there is a threshold that is used to determine the value of the indicator. When we asked why the threshold was set where it was, the answer was not particularly clear. We therefore set about trying to investigate how one might come up with a particular figure.
We did a very simple exercise, examining expenditure on education according to settlement rather than according to school. We plotted expenditure per pupil against settlement size. That revealed very clearly that as the size of settlements declines—towards settlements of 1,000 or less—the costs of delivering that statutory service increase hugely. Below a certain level of settlement, the increases can be of the order of 50, 70, 90 or 100 per cent. We have not replicated that investigation elsewhere, but in my view there is significant evidence that delivering the education service through small settlements imposes significant costs.
Are the indicators that are currently used for GAE assessments capable of dealing with that, or are more changes needed in how GAE is calculated to take account of different population distributions?
I have no doubt that the forthcoming census results will reveal significant changes in population distribution in Scotland. There is probably scope for improving the indicators. We have not yet done that detailed work and I do not want to pre-empt it. Like everyone, when we work to tight deadlines our focus is on delivering the results rather than on considering the options. The current indicators go some way towards recognising the differential pressures that exist. They are far better than the indicators that are used in the English SSA system. That is why a similar approach is being adopted in Wales.
Are there any lessons to be learned from the work that you have done for the DETR or the Welsh Assembly that are applicable to Scotland, or have the lessons gone the other way?
To be frank, they have gone the other way. The exercise in Cardiff was slightly more complicated because we ended up with all sorts of physical configurations to take account of the effects of the valleys. Here, there are equally complicated issues to do with the islands. There is scope for improvement, but that will need further consideration.
Like Iain Smith, I want to ask about the scope for improvement in statistical analysis. Have you developed the ideas that you outlined in your written evidence of using service reviews undertaken by the best-value regime to provide a framework for your analysis? If so, what were your conclusions?
No, we have not. There is a huge opportunity in this area. We are currently involved in a review of local performance indicators in the best-value regime in England. It is clear that best-value structures provide an eminently rational basis on which to make judgments about differential service quality. I am not aware of any effort to relate that information base to any resource allocation information base. However, it does exist.
One of the exercises that we undertook for the Scottish Executive was an examination of some aspects of leisure and recreation services. During that review, it became quite clear that other organisations such as the Association of Direct Labour Organisations were collecting huge amounts of very specific information, which the Scottish Executive could not hope to replicate. We made some attempts to harness that information, but the time scales meant that that was not possible. However, simply bringing some of that information alongside some of the expenditure-based information within the GAE process provided a straightforward opportunity to attempt to assess expenditure relative to level of service.
In developing new forms of needs assessment, are we creating a stricter regime—even a straitjacket—in which local authorities would have to work? Do you have any evidence that needs assessment restricts local authorities' decision making as far as service provision is concerned?
Evidence is a difficult word, in the sense that anecdotal evidence is clearly available. For example, some rural local authorities might feel that they have great difficulty in supplying what they perceive as the same level of service as more urban authorities. In some circumstances, they simply hold up their hands and say that they cannot supply the same quality or level of service, which causes the straitjacket that I talked about earlier.
The process simply revolves; the authorities report their expenditure levels, and the next year they receive a lower level of resource because they are providing a lower level of service. How can services be improved if an authority cannot switch resources into an area to do so? There is a danger that, by providing a lower level of service, the authority risks remaining at that level unless it is willing to switch resources at a cost to some other area.
You mentioned in your submission that a hierarchical model might be used for needs assessment in education and personal and social services. How applicable would such a system be in Scotland, which ranges from diverse island communities to urban Glasgow?
Apart from being incredibly interesting, Scotland is also incredibly diverse, and the problem is that it might contain a greater number of hierarchies than elsewhere in the UK. As you have pointed out, community structures in certain rural parts of Scotland exist alongside the traditional ward-based and regional constituencies. We need to think more carefully about some of the wider dimensions when defining hierarchies. Perhaps the inclusion of some sociological content or perspective would be appropriate, instead of simply focusing on a straightforward geographical or spatial basis.
Can you expand on that comment about including a sociological perspective?
Having worked in a number of services over the years and having had the opportunity to talk to a variety of authorities across Scotland, I think that it is clear that, within certain types of community, there is a different social perspective about the provision of some services such as being reported for minor criminal offences and probation-based exercises that are carried out by social services and their recipients. Those perspectives can change radically over very short distances, but I do not know the extent to which they feed up through the structure into any expenditure-based implications.
All that is very much anecdotal. My suspicion is that there are a larger number of hierarchies and more complicated interrelationships than may be the case elsewhere.
Are there any obvious disadvantages, apart from the fact that such a model would take quite a bit of time?
It would take a long time. It is clear that there are technical advantages, in that the model provides a much better focus than the standing levels of expenditure need. Like you, I am concerned about the timetable. I am also concerned about the process of divorcing local authority members from the resource distribution process. There is evidence that many authorities have, in effect, developed their own corporate elite of GAE people, who are experienced in understanding how the process works and how they can tweak the process next time round to ensure that their pot gets a little bigger.
As Professor Smith said, it is quite evident that many local authority members are struggling. Technical advancement would mean that that gap would extend even further. There may come a time when we say that the price of better quality distribution has to be recognised and that there needs to be some separation of accountability on the ground.
Can I ask specifically about the time scale?
As I said earlier, my impression accords with that of Professor Smith. I was part of a team that undertook a review of the SSA process in Wales in 1998-99. Twelve months were spent on the education multilevel modelling exercise. That work is more or less up to date. Another 12-month contract has recently been put out to make progress on personal and social service. That means that the exercise will have taken two years. I suspect that the 18 months that Professor Smith expounded is a minimum and that we are probably talking about at least two years.
In your submission, you mention the tensions arising because of the gearing ratios. You say that the two possible solutions are to improve the revenue base of local authorities or to introduce safety nets into the system. I would appreciate your views on those options.
The previous witnesses have probably covered those points. We are of one mind on the opportunities available for expanding the tax base through business rates.
Safety nets come back to the process of damping—ensuring that annual change is restricted to a specific banding structure, so that no radical adjustments in service quality result.
Is there not a danger that a safety net system would reward people who are irresponsible in their expenditure?
Yes, I think there is.
We must recognise that an immediate and radical shift of service level from one level to another is not necessarily in the interests of the authority or the recipients of the service. If we develop our thinking to such a point that we can identify poor performers, it would be better, in the interests of the community, to have a measured adjustment process rather than to jump over the side of the ship at the end of one financial year and the start of another.
Do you have any experience of the way in which needs assessments are produced in other European countries? If so, do you believe that Scotland could learn any lessons from other countries?
My experience is limited. I am aware that, as has been suggested, the UK has tended to lead in this area. The Scottish system is quite complicated, but complexity is not necessarily a negative attribute in some circumstances, especially if we are moving to the recognition that, to some extent, we cannot keep members with us and they must follow behind. Professor Bramley probably talked that point through earlier. We are ahead of the game now. Certainly, the detail of the Scottish system in the areas of settlement and population distribution has much to commend it.
I think that there are no more questions.
You said briefly at one point that you thought that best value might have been forgotten somewhere. The subject may be complicated, but I assure you that best value is not forgotten and the committee will look at best value later. You also said that there was room for improvement. We all agree on that. Improvement does not necessarily mean massive change.
You said that there are a number of things to consider, one of which is the reality out there, which we must recognise is complex. As I said earlier, the system is complicated. We have not yet come to any conclusion on whether changing the system would improve it. It may be that the system needs tweaking and adjusting rather than a massive change.
I thank you for your contribution. As I have said to others, if we need you again, we will certainly call you.