Policies are usually associated with objectives. In planning documents, objectives are of two main types: aims and goals. Statements of objectives are also complemented by statements of a mission or purpose, and statements of values. These terms are not fixed, and sometimes they are used interchangeably, but the types of things which are referred to are reasonably distinct.

• A mission (or vision) is a statement of purpose. It is, then, a general statement of aims and values which comes before any specific policy has been determined. Examples of mission statements might be ‘promoting a knowledge economy’, ‘developing sustainable communities’ or ‘helping people to have a healthier old age’. A mission could also be referred to as an aim, but the important issue is that it is generalised and at a higher level than specific service objectives.

• Values are an important dimension in statements of objectives, but they are not always identified explicitly. Values are moral principles or norms. They can be used positively, to lead policy in a particular direction, or negatively, to forestall certain options. Positive examples might be promoting ‘health’ or ‘social inclusion’; negative examples might be ‘confidentiality’ (which is a statement of how information will not be used) or ‘prudence’ (a commitment not to do anything rash).

• Aims are what a policy is supposed to achieve. General purposes have to be ‘operationalised’. That means that they have to be translated into terms which can be realised, or put into practice. If, for example, the mission is to promote a knowledge economy, the aims might be to increase the proportion of people attending higher education and the proportion of those in work who undertake training and knowledge-based staff development.

In an ideal world, what an administration might be expected to do is what the agency is directed to do. Policy goes in at the top, and a service delivery comes out at the bottom. This is the model of ‘perfect administration’. Perfect administration is an ideal type, rather than a realistic expectation. It is commonly associated with a bureaucratic model, where rules are made centrally and the operation of the service follows an undeviating chain of command from top to bottom. (This idea is often attributed to Weber – Haralambos, for example, suggests that bureaucracy, to Weber, is ‘rational action in an institutional form’¹⁹⁹ – although Weber’s work on rational administration is actually rather more sophisticated.²⁰⁰) It is important to realise that there are discrepancies between the aim and the process: if we wanted ‘perfect administration’, bureaucracy would not be the way to do it. Perfect administration calls for a command structure, like the structure of an army. In an ideal army, each person works as part of a unit. There is a surrender of individual decision making and submission to higher authority in all things. In a bureaucracy, by contrast, each person has a defined role and set of tasks, makes decisions in their sphere of influence, and is accountable to the next person above them. In the army, if no-one issues instructions, nothing happens, unless someone assumes command. In a bureaucracy, if no-one issues instructions, every member of the bureaucracy knows exactly what to do – which, incidentally, is why local government departments can survive without anyone apparently being in charge for months or years.

The most basic part of making an assessment of the environment is to find out what is happening. Finding things out is generally treated, in academic writing, as a form of ‘research’. Research is an essential component of policy analysis – although properly speaking, it is not one component, but rather a term that stands for a wide-ranging set of issues. Unfortunately, the idea of research is beset with stereotypes, and many people, in practice as well as in the academic world, will respond to the demand to ‘do some research’ by doing what they have been taught to do. Many researchers are schooled to use disciplinary approaches to research, doing things the way that economists, sociologists or psychologists do them. People who are not very sure about research often imagine that it is some sort of exercise featuring questionnaires, tick-boxes and computers. Policy research is not like that – or at least, it should not be like that, because the object is to find out what can be found. Policy research, Majchrzak argues, deals with problems that are complex and multidimensional. It focuses on the kinds of issue which a policy can affect, rather than those it can’t. It is heavily influenced by the needs of users. And it is also, crucially, value-laden; it is a moral, and political, activity.⁷⁸

The term ‘methodology’ refers to the rationale people give for different kinds of method. There are several widely used distinctions between different types of methodology. The first lies between inductive and deductive approaches. Inductive’ approaches begin by collecting material and seeking to classify and organise it after it is collected. An example is ‘grounded theory’, which sorts material by putting it into categories and stopping when there is no material left to classify.⁷⁹ ‘Deductive’ approaches begin with a hypothesis or model, against which circumstances can be compared.

Extensive research attempts to consider the extent of an issue – how big it is, and what proportions are involved. This is usually a quantitative exercise. Quantitative data are used at many points in policy analysis. They tend to give the impression of accuracy and precision, and it is probably appropriate to begin with a health warning. The kinds of problems which are dealt with in public policy are often fairly illdefined, and the implications are often uncertain. Numbers are used to give shape to issues, and to identify relationships. This is generally interpreted, in policy studies, as work on ‘indicators’. An indicator is a statistic which is taken to mean something else besides the core information it contains.101 It is a signpost.

The term ‘indicator’ is generally used to show that quantitative information about social issues represents not simple ‘facts’ but rather ways of putting together complex and uncertain information. The moment something can be counted, it is likely to be treated as if it was a ‘fact’. Examples are classifications about whether a crime has been committed, or whether someone is unemployed. These are judgments, and they are open to argument. Indicators point out the direction, rather than showing ‘the facts’: they have to be interpreted.

Indicators are not the same thing as measurements. A good measurement is accurate, precise, and reflects the characteristics of the issue it is measuring. A good indicator is associated with the issue, robust, consistent over time and available. For example, low birthweight is usually a good indicator of poverty.

Evaluation has generated a literature, a set of disciplinary approaches and a specialised industry all its own. There is some overlap between evaluation and analysis for policy, and some of the books on evaluation talk about the subject in terms that could be read as books about policy analysis.²⁴⁵ The focus of most guides to evaluation, however, falls on outcomes – they have relatively little to say about strategic planning, targets, indicators, forecasting, efficiency, robustness, the process of implementation, or indeed most of the other material covered up to this point.

In keeping with the general objectives of this book, this chapter is concerned with evaluation only in a limited sense. Policies are set in a particular context. They have aims. They select methods to implement those aims. They have effects. A basic policy analysis tests at each point whether each of the parts is consistent with the others. The most common pattern of evaluation in policy analysis is probably to check whether or not practice has been consistent with accepted guidelines, but a more fundamental, and often a more appropriate, approach is to see whether the outcomes of policy are consistent with the aims. In other words: has it done what it was supposed to do?

Evaluation is commonly categorised in two main classes, summative and formative. Summative evaluation is the evaluation of a whole policy or process, focusing on the impact of policy. Most policy can be treated in terms of a series of categories – aims, methods, implementation and outcomes.

Selecting methods is, properly speaking, part of policy formation, but understanding the process is also important for policy review. In practice, Scriven suggests, policy analysis is often concerned with things that have not actually happened – alternative approaches which might have been taken, the things that might go wrong, or the ways that policies might be adapted in the future.¹²⁴ The role of the policy analyst in the process of policy development is not always clear – it has to be discussed and negotiated. By the time an analysis gets under way, both the methods and the institutional structure are likely to be decided on, and analysis has to respond to the consequences of the arrangement, not to offer ‘blue sky’ solutions. There may be scope to negotiate some changes, but it is likely to be limited. Understanding approaches to method is important partly because it helps to explain where policies might have gone wrong, and where they might still go wrong, but also partly because making recommendations about developing or improving a policy calls for an understanding of how else things might be done.

The kinds of action which can be taken depend on the powers and competence of the organisation which does them. Central government in the UK is legally unrestricted in most of the actions it can take (although there are treaty obligations, particularly relating to the European Union, which limit the potential course of action). Central government in many other countries, like the US or the Federal Republic of Germany, is limited by a constitution or basic law: governments can only do what they are permitted to do.

There are three main areas of applied policy work: policy formation, public sector management, and policy analysis and review.

• Policy formation. The formation and development of policy depends on knowledge of the specific subject area and ideas about options and approaches. Studies in this field focus on what policies are, what they do, and how else they might be done. Because this is often done in an attempt to bring about change in policy, or to defend particular approaches, it is sometimes referred to as ‘policy advocacy’.²

• Public sector management is mainly concerned with the process of administering policy, implementation and managing organisations. The skills required include project management, resource management and working with people.

• Policy analysis and review. This is about examining policy – finding out and assessing what is happening; monitoring implementation; and evaluation, or finding out whether policies do what they are supposed to do.

Although there are references in this book to all these activities, Policy analysis for practice is mainly concerned with the third area, policy analysis and review. Policy analysis works by trying to establish the criteria by which policies can be judged, to find out how they operate, and to see whether they produce the effects they are supposed to produce. This book introduces the concepts, approaches and techniques which are used to examine policy in practice. It has a direct application to an expanding set of job opportunities, concerned with social administration, service planning and delivery, and strategic public service management.

Policy is a very ambiguous term. In its simplest sense, a policy is a decision about a course of action, but it is also supposed to represent a set of decisions, interrelated and consistent with others. Hogwood and Gunn identify 10 meanings of the term:

• a label for a field of government activity and involvement;

• an expression of the desired state of affairs or general purpose;

• specific proposals;

• the decisions of government;

• formal authorisation;

• a programme of activity – that is, a defined sphere of activity, involving particular measures;

• the output of an agency (what is done);

• outcomes;

• a theory or model – assuming certain results from actions; and

• a process of decision making.¹¹

Much of the literature in political science in this field is concerned with the understanding of how policy is made. It addresses questions like, ‘who makes the policy?’, ‘who is the policy for?’ and ‘who benefits?’. There are several schools of thought for the analysis of policy. Stone argues that policy formation is a process of negotiation or bargaining within the ‘polis’ or political community. This is essentially an irrational process: negotiation and bargaining depend on the reconciliation of conflicting interests, and people are influenced by many irrational factors, such as concealment, bluff, bargaining, log rolling (trading to get the results one wants), influence, loyalty and so forth.¹² The leading views currently are:

• Policy networks. Policy is made through a process of interaction between people in networks, where people negotiate, exchange and share positions. Examples of networks include professional groups, inter-governmental groups and economic producers.

The central issue in selecting methods is effectiveness – the extent to which a method achieves the aims of the policy. However, there are other important issues to consider, and when policy makers are considering what the best options are, new criteria tend to be applied – issues like cost, practicality and how the policy will be interpreted publicly. Just which method is adopted has to depend on the specific circumstances.

The main test of policy is to ask whether it meets its objectives. This is ‘effectiveness’. In real life, however, effectiveness is rarely pursued at the expense of every other consideration. The selection of methods depends not just on whether something works, but how it works when the costs are taken into account, and whether the return is proportionate to the effort. This is often summed up in terms of ‘value for money’. ‘Best value’, for example, placed a duty on local authorities and some other agencies

to make arrangements to secure continuous improvement in the ways its functions are exercised having regard to a combination of economy, efficiency and effectiveness.¹⁷⁴

The issues of economy, effectiveness and efficiency are the main subject of the first part of this chapter. The issues are financial ones, but ‘finance’ goes rather beyond the narrow sense of accountancy which it is usually associated with – and accountants are increasingly being asked to do the sort of work which is part of the training of a policy analyst. The kinds of question which are asked about ‘value for money’ are not just about money; they are also about value.