External Research
We move on to agenda item 4. I am pleased to welcome Norman Flynn to the committee. As members know, he is carrying out external research on outcome budgeting. We have all received a copy of his initial note.
If you would like to say a few words first, please do so, otherwise please move on to the presentation that you have prepared for us.
Thank you. I preface my presentation by saying that it is a privilege to work for the committee. It is nice to be involved not in re-inventing or reforming a Government but in inventing one. I find it refreshing to work on a new Government because, for many years, I worked to help change Governments that had been around for a long time.
I will outline my method—my presentation is really a summary of the proposal that I made when I was hired to do the work for the committee.
What do we expect from budgets? Given the committee's request for this piece of work, I think that one of its problems is that although the budget seems to be reasonably good—better than I thought, until this morning—at controlling and planning expenditure purely in financial terms, it is less good at reflecting priorities and resource allocation. At present, it does not seem to be much good at all at evaluating performance—least of all policy performance. The purpose of my research is to try to move the budget process towards a focus on outcomes and to make it better as a way of reflecting priorities—not in terms of spending but in terms of required outcomes—and, ultimately, as a way of evaluating whether policies have worked well, as we will see whether or not the outcomes have been achieved. That is my interpretation of the committee's brief.
That approach leads us to an examination of the other planning and evaluation processes. I want to undertake an examination of best value in local government and of the efforts that have been made to introduce best value into central Government. I also want to examine the residual comprehensive spending review, the process of outcome definition in the health service and so on. That work will allow us to see how many of the planning and evaluation processes are in place already—as I said, we do not need to re-invent them—and to incorporate them in some way into the budget process. Therefore, my focus is not simply on the budget process but on the other work that is taking place to allow those planning and evaluation processes to happen.
Other jurisdictions have experimented with this sort of work for rather a long time. Part of my research will be to draw together work from other jurisdictions, mainly from secondary sources and from personal experience and contacts. For example, the USA passed its Government Performance and Results Act in 1993, following which every federal agency was supposed to produce an outcome-based budget, starting in April 2000. Although that did not happen, sufficient experience in federal programmes exists to be of use to us.
Probably more important, all but three states in America—47 out of 50—have had experiments with outcome-based budgeting over the past five years. Those experiments provide some useful stuff. Although we could not necessarily adopt the individual measures, we could take lessons about the process, the number of measures and how they are incorporated into the legislative and executive processes.
I included New Zealand partly because the Finance Committee seems rather keen on it. I understand that the committee had a videoconference with June Pallot and I thought that we should build on that. Everyone in the world looks at New Zealand's resource accounting and budgeting, which Andrew Likierman introduced to the Westminster Government after visiting New Zealand. It would be foolish not to take account of New Zealand.
Nobody considers Sweden at all. In our research, we discovered that Sweden has had executive agencies for 300 years. We should draw—nobody else seems to—on that country's useful experience of trying to get outcome-based agreements between Government departments and agencies.
I chose Singapore partly because I know it. It is probably one of the tightest-run Governments in the world, whatever members might think of its style of government. The Singapore Government rejected outcome-based budgeting altogether and went for something called performance-based budgeting, on the ground that outcomes should be dealt with elsewhere. I wanted to include Singapore as an interesting counter-example of somewhere that rejected resource accounting and budgeting after examining what New Zealand was doing.
I do not know what to call the Westminster Government. I have called it "England" in the paper, but it could be called Westminster, Great Britain, the United Kingdom or whatever. It is that lot south of the border.
That could spark a debate in the committee.
I hope not. It could spark a debate, but I will not join in.
At Westminster, the incorporation of outcomes in the formal budget process has been given up, but outcomes are being incorporated through the spending reviews and the public service agreements. The details of the public service agreements are good. Whether it is desirable to keep them separate from the budget process in such a way is for the committee to decide, but I will report on that.
The experience at state level in Australia—especially in the more populated states—is probably the most useful to study. New South Wales, Victoria and Queensland have been experimenting with outcome-based budgeting for about five years.
The Queensland budget is an example of that—it is not included in members' papers but I have a slide on economic development to illustrate it. There is a hierarchy, with a big overall outcome of economic prosperity at state level. There are then some supporting outcomes, without numbers attached, and an outcome indicator. The supporting outcomes are: a skilled work force; cost-effective assistance to small business; and regional industries being strengthened and expanded. The simple outcome indicator—rate of employment growth—is broken down into the different departments: education and training, state development and tourism. Each of those has specific output targets that must be met: numbers of apprenticeships and traineeships, numbers of businesses supported and so on. The departments are then held to account for those figures. In that way, rate of employment growth supports the various outcomes, which support the overall outcome of economic prosperity.
It has taken Queensland five years to arrive at what looks rather simple. The process, however, was not simple. That illustrates how overseas comparisons might speed us up.
That is my sample of experiences in other jurisdictions, some of which have implemented outcome budgeting, some of which have not—for good reasons.
When it comes to what to do in Scotland, it seems that it is crucial to develop commitment to outcome budgeting among the people who will be asked to do it. My plan is to talk to as many people as possible in the Executive, in executive agencies, possibly in local government and certainly in the health service to get their views and to state my view and, in the end, that of the committee. I intend not just to work technically on what kind of outcome indicators and process we want, but to establish that outcome budgeting is a good idea. I will also talk to people who are not directly involved in the budgeting process. For example, yesterday I spoke to the deputy inspector of prisons about the outcomes that exist in prisons. I will talk to as many people as possible in the short time that is available.
We should not try to do everything all at once next year. It took the USA from 1993 until 2000 to get outcome-based budgeting in place in the federal Government and it took the Australian states at least five years. We should be selective in our approach and concentrate on where the big money is going and where the increments in the budget are. It is sensible to ask what the results of any extra money have been rather than what the results of all of the £20 billion have been. However, the situation becomes a bit obscure when one starts to examine the increments.
The budget document that the committee has been discussing shows big increments in certain areas: central support for local authorities has risen by 13.6 per cent; support for children and central Government education has risen by 14 per cent; support for health has risen by 15.5 per cent, which is nearly £1 billion; justice has risen by 8.4 per cent, which is £56 million; social justice has risen by 16 per cent, which is £118 million; and transport has risen by 14 per cent, which is £188 million. Members will be aware of those numbers, which show where the money is going. The question is, what are we going to get for all those increments between now and 2003-04? I will concentrate on the extras. Of course, once we start asking where all the extra support for local authorities is going, we will be led to think about the McCrone report and so on and we will realise that it will be difficult to pin down which particular programmes the spending has gone towards.
The other point of the sample that I will take is to discern the areas in which outcome-based budgeting is not already well developed. In certain areas, such as economic development, it is rather well developed—for example, Scottish Enterprise and Highlands and Islands Enterprise have targets that are tied to output in a well-organised way. I will report on the areas that are well developed—for example, the health service tells me that it is in the process of producing outcome targets—and will try to do more work on the areas that are less well developed.
My sample will deal with 10 services: prisons; universities; health; schools; police; children and young people, including pre-school children; further education; social inclusion, specifically Scottish Homes; rural development; and environmental protection. I will also deal with existing processes that seem to have some focus on outcomes, such as the comprehensive spending review. I will examine best value in local government to determine whether it is being applied at the Executive level. I will also examine the ministerial working group on best value and budget review, which was referred to this morning—I do not know what it is yet, but I found the person who is in charge of it.
The first five interviews and meetings will take place in the week beginning 5 November and the second five will take place two weeks after that. Members will note that my deadline is soon after the end of my meetings. The report that I have been asked to produce on 23 November will definitely be an interim report as some of the interviews will have been conducted only two days before then. The final report, with implementation plans and so on, will be available in February.
We all understand why New Zealand has become the international reference point in this matter, but you said that Sweden is also a good example. Why does nobody take any notice of it, particularly as its system has been in existence for such a long time?
That is partly because not many people speak Swedish. The Swedes produce a budget summary in English—
Most people in Sweden speak English.
They do and they are helpful. However, the Swedes have never wanted to make themselves famous in the same way that New Zealand politicians—who promoted themselves as world leaders in reinventing government—have. Nonetheless, the Swedes have good experience.
Is that accessible?
It is quite accessible. The whole budget cannot be obtained in English, but a summary of how it works can. The budget process is available in English.
I hope that there are about 35 hours in your day, Mr Flynn, because the programme seems incredibly ambitious for the short time that has been set. I do not know how it will be completed in the allocated time unless you already have a lot of data on international comparators. I wish you luck and hope that the programme is realistic as well as ambitious.
Paragraph 1 on page 2 of the submission contains some of the hard meat of the matter. Our system is all about inputs. There are some targets, but they are not related to outcomes and they inevitably tend to relate to the longer term. There are some long-term health targets, including some that are set for 2010 and, I think, a UK child poverty target that has been set for 2020. However, I can never find out how budgets relate to outcomes. The link is difficult to follow. Paragraph 1 on page 2 refers to the difficulty of attributing outcomes
"to the actions of a specific agency over a short time period".
We hinted that we far too often think of budgets in terms of one year. A programme of expenditure should be designed for an outcome and we should see an incremental improvement over a period.
An example of a target that is not being reached is suicide rates. Since 1992, targets have been set to reduce suicide rates, but rates are rising. New strategies are therefore produced, which is great, but what happens? I do not know whether I am making sense, but realistic outcomes are needed. We have been great at setting targets, but we have not linked things up. We cannot see where money is spent in the longer term. Can you help with that?
Would you like me to reply to the question, convener, or shall I simply make a note of it?
I have an additional question. Paragraph 1 on page 2 of the submission mentions "high level outcome indicators". Does that mean long-term indicators? Morbidity rates, educational attainment and environmental quality are mentioned.
There is a hierarchy in time and identifiable accountabilities. If the suicide rate increases, we cannot point to one person and say that it is their fault. Many people would be at fault. However, if the suicide rate in prisons increases because prisoners are banged up for 23 hours a day, one could say that the increase is the prison governor's fault. The overall rate should be broken down into, for example, suicides in prisons, suicides by drug abusers and suicides by mentally ill people who are not cared for in the community. That is what I mean by high level and low level.
In terms of time frame, many of the policies will be long-term ones. For example, the Singapore targets would be to transform the country from a manufacturing economy to a knowledge-based economy in 10 years, which has been achieved.
It is worth having long-term targets but then it has to be decided whether it is the job of the education system to get 50 per cent of people into higher education within three years, for example. That is the same problem—the target is both long term and high level. It has to be broken down into accountabilities and actions.
We should be holding managers and the people who work for them to account for outputs. However, it is our job to ensure that we are buying—to use the New Zealand jargon—the outputs that are relevant to the outcomes that we want. Therefore, returning to the prison suicide example, we cannot hold prison officials accountable for overall suicide levels, but we can hold the prison governor accountable for a 23-hour in-cell regime in his or her prison. Everyone knows that a 23-hour in-cell regime leads to suicides—that is not controversial—so the governor should be held to account for time spent in the cell.
Your submission mentions trying to get outcomes as a consequence of the additional money or growth in the budget. There is a difficulty with that because almost all of that money will be spent on initiatives and we will end up monitoring those initiatives rather than the delivery of core services. I understand that that might be easier because, while the programme is being set up, the mechanisms for monitoring the outcomes could be put in place. I can see some value in that. However, it works on the assumption that the rest of the money is always going to be used for the same things, which is not the case. Some programmes will come to an end. There is more flexibility in the overall budget than just in the growth part of it. Have you any thoughts on that? Moreover, do you have any thoughts on the point that we will end up considering the plethora of initiatives as opposed to the generality of what government is about?
The shorthand in my presentation might have misled you. What I meant to say was "incremental", which means the bits that change—they could go up or down. For example, what would the consequences be of cutting common agricultural policy spending over the next five years? What would the consequences be of increasing spending on pre-school education? I am not just talking about growth; I am talking about any change in the budget from one year to the next.
It is difficult to find those changes in the draft budget. You can find the bits that are growing and you can apply some sort of deflationary indicator but, as Arthur Midwinter's report shows, the general deflators do not apply across the board. It is hard to see the difference between one year and the next in the draft budget in what an economist would call real terms. However, I take your point.
Some aspects of your approach leapt out at me. How are you going to develop a model such as the one in Singapore, for example, in which Governments decide that they are going to be around to see out the long-term planning and make commitments that would have to be either dismantled or not touched if there was a change of Government? That is an issue about the way in which we operate our democracy. I am not arguing about the principles because, in business, there are long-term objectives and short-term objectives. Obviously, if an outside force comes along, one has to respond in some way and one needs flexibility to do so. However, that cannot all be done from a reserve fund. How do you take account of that in your modelling?
Diplomatically, I would say that the differences between the policies of the parties in Scotland are no greater than the differences within, for example, the People's Action Party in Singapore. There is dissent within the party in Singapore, although that does not get into the newspapers. There is agreement that all Singaporeans want economic prosperity and they achieve that by moving up the value chain. There is also agreement that there should be a compulsory provident fund to look after people in their old age. There is some consensus but there is also some disagreement about how to achieve the Government's big objectives. My guess is that none of the parties in Scotland is arguing for lack of economic growth. There may be some fringe Greens who say that economic growth is a bad thing, but in general we all want prosperity, decent education and a health service that works. I honestly do not think that members of the committee disagree very much about what the high-level, long-term goals are.
Do you believe that, because of your contacts, important civil servants will be honest with you? They would certainly never be honest with us, so it is a big plus if they are prepared to be honest with you.
You mentioned a target of getting X per cent of people into further or higher education, for example. I do not know whether this is true, but there is a widespread view that the experience of the average student in the average university is distinctly worse than it used to be, because of cuts and so on. Presumably you should measure the quality, as well as the quantity—I know that measuring quantity is much easier. The ultimate goal of Government is 5 million—or whatever the figure is—happy Scots wandering about smiling broadly. Has any country tried to measure such a factor? Are we delivering happiness to the people? Perhaps an election measures that.
I would be worried if I saw 5 million Scots going about smiling.
I shall tackle the second question first. People have drawn up happiness indicators. One can establish satisfaction-with-quality-of-life indicators and monitor whether they go up or down. I am not suggesting that we do that for Scotland as a whole, but it is worth doing in local areas, and I have done such work. We base the indicators on people's own criteria, asking what makes them smile and whether those things are getting better or worse. That is quite fun, but it requires large-scale social research.
The first point, about honesty, is a great philosophical question. "If you ask civil servants a question, how do you know whether their answer is true? Discuss." I have experience of talking to civil servants in many jurisdictions and I think that I am reasonably skilled at finding out when they are pulling the wool over my eyes. In fact, I had good fun for the first two hours of this morning's meeting listening to the committee's discussion with civil servants from the Executive, wondering which answers I would follow up. Other than that, I cannot say whether civil servants are honest with me. I take the point that they may attempt to mislead, but I am not easily misled.
I also asked about measuring quality as well as quantity.
Singapore is an interesting case in point. The Singapore Government said that it wanted a knowledge-based economy by whatever year it was—it has virtually got it now—and that it would need a certain proportion of people to go through higher education to achieve that. It did not stop there, of course. It decided which courses should be done at the technical universities and which should be done at the National University of Singapore, and what those courses should be in. The Singapore Government has a level of control that the Scottish Executive frankly does not have.
I am going to meet the bloke in charge of higher and further education—the same person is in charge of both, which is a wonderful appointment—to find out how priorities are made about which courses to run and whether that fits in with national priorities about what should be done. In Singapore, if 10,000 people who can design circuit boards are required, we can be pretty damn sure that in three years the Government there will have them.
As a supplementary answer, I think that the levers of control or modes of governance are many and varied. It gives me a headache thinking about the fact that we whack a huge amount of money to local government with virtually no plan and a similar amount to health, but that we have an explicit agreement with the Scottish Prison Service, for example. It must give committee members a headache wondering how they can ensure that any indicators—never mind those of quality and quantity—will hit the citizens in the street. It is easier to control the executive agencies, which are a relatively minor part of the budget, than it is to control the 32 local authorities and the local health boards. Different answers on quantification and quality will be necessary depending on the mode of governance of the services.
I want to follow on from Donald Gorrie's point, although I do not have such a jaundiced view of civil servants as he does. Your paper says that
"the process requires the involvement and commitment of those providing services to make the process effective and workable."
That could mean civil servants. Elsewhere, you say more precisely that one of the
"features of the Scottish governance system is that a large part of the budget is distributed in grants in aid"
and that managers and workers need to be
"involved in the process of outcome definition."
Do you have reason to believe that either civil servants or professionals, for example in the health service, are not likely to provide that sort of co-operation?
That depends on how one goes about it. In my experience, if one tries to hold people accountable for outcomes that are a result of variables over which they have no control, they get terribly nervous. My research will not organise the process, but the process that I propose in the implementation plan will have to involve people who know what works on the ground so that they can say that they are accountable for one thing and not for another. One person cannot be accountable for the morbidity in a health board area, but he can be accountable for the death rates from a certain procedure in hospital.
The problem is that if people are hit—without discussion—with an outcome target for which they will be held to account, they give up; they start cheating and lying, they have bad faith and they will come to the committee and say, "Yes, boss, we are on target for 2020; we are all going to be fit and healthy." As it happens, in 2003 we might all be dying of suicide and lung cancer, but we are told, "Don't worry, the outcome will be fine." That is not helpful to anyone.
I did some work on outcome measures with the fire brigade in Hong Kong, which, before I became involved, had an outcome target of no deaths from fire. That might be achieved through prevention and intervening in the building process, but neither of those is a fire brigade activity. The fire brigade is happy for there to be no deaths from fire, but its target is to put the bloody things out once they have started. The fire brigade's strategy had to be to get fire engines with tall enough ladders to put out fires in the new buildings. A lot of people further up the system should be held accountable for fires. Before I got involved, the fire brigade was told what its target was and that, if it did not reach it, it would be in trouble. The result was cynicism.
If there is cynicism in the system, we will not achieve what we want. We must ensure that such cynicism is rooted out.
We should design the process of arriving at outcomes so that it generates enthusiasm rather than cynicism.
We note your point about the interim report on 23 November and we look forward to receiving that. It occurs to me that it might be useful for you to have a committee member with whom to liaise.
That would be great.
You can liaise with the clerks when necessary. As Richard Simpson suggested the subject, does he have the time to take on that responsibility? I do not think that it would be terribly onerous.
He has bags of time.
I will make time.
Does the committee agree that Norman Flynn should liaise with Dr Simpson?
Members indicated agreement.
We will see Norman Flynn again when he has produced his interim report.
I will be a lot thinner and I will have less hair by then.
The time scale is only a month, so you have a lot of shedding of hair to do. Thank you for your presentation.
As agreed, we will now take items 5 and 6 in private.
Meeting continued in private until 12:31.