Policy Tools: Analysis, Choice and Action

 [This paper was originally written in 2005 but the arguments it makes are still relevant to today’s debates about policymaking?]


This paper seeks to lay out some of the issues surrounding policy formulation and implementation from a ‘tools of policy’ perspective. This is a relatively underdeveloped approach in public policy but one which offers some tangible benefits in expanding the scope of policy choices and developing more realistic analytical, diagnostic and evaluative tools.

The paper is based mainly on a review of the existing literature – official and academic – on policy tools plus the experience of the author with a wide variety of policy and implementation issues within government in the UK and internationally. However it is also situated in a wider understanding of the lack of joined-up thinking in much of the policy arena.

I have argued elsewhere that a lot of research and practice in considering how best to achieve results through state action has fragmented into un-helpful over-specialisation and lack of ‘consilience’, i.e. integration of knowledge (Talbot 2005).

Thus the ‘tools’ approach, which is small but significant (e.g. Hood 1983; Bemelmans-Videc, Rist and Vedung 1998; Blair 2002; Salamon 2002), has been largely ignored by mainstream policy analysis and evaluation (which in any case largely ignore each other (Talbot 2005)).

The ideas outlined are therefore very provisional in nature and would require considerable additional desk and field research to flesh out and justify.

The ideas are presented in fairly ‘telegraphic’ form as a foundation for discussion.


The existing, very extensive, literature on policy analysis – which we have not summarised here – tends to see policies in terms of ‘options’. Policy options invariably include more than one class of policy tool (see below) explicitly and also others implicitly.

Some policy tools – especially those in what Hood calls the ‘nodality’ class – often only appear during the implementation phase of policies, even when they are clearly tools in their own right.

Moreover the fact that particular options are in fact made up of a particular combination of different types of tools is rarely acknowledged, discussed or included in their evaluation. This means that most policy analysis and evaluation is actually only partial – evaluating policies as if they were made up of a much more limited set of interventions than they actually are.

Various categorisations of policy tools have appeared in the academic literature over the past 20 years or more but not, as far as we can tell, anywhere in the official literature which offers advice on policy making and implementation (at least in the UK – there may be some in the USA).

Of the categorisations on offer (for some see Appendix A) the most useful seems to be Hood’s “NATO” model which categorises policy tools into four groups:

  • Nodality – control over information flows
  • Authority – control over formal (legal) authority
  • Treasure – control over money (both gathering and distributing)
  • Organisation – direct or indirect organisation of actions

There are cases in which policies rely almost exclusively on only one of these tool sets:

  • Nodality – cases where governments rely exclusively on control of information – e.g. through public information programmes – as a means of attempting to implement change
  • Authority – where policies are solely couched in regulatory form
  • Treasure – where policies rely solely on transfers – taxes and benefits – as their instrument of change
  • Organisation – where governments ‘do it themselves’ by directly or indirectly organise the production of goods or services

However, even in all of these relatively ‘pure’ cases governments often rely on the other tools to some degree:

  • Public information campaigns have to be funded and organised
  • Regulations have to be enforced, usually through publicly funded agencies (although there are examples of self-funding regulation)
  • Treasure has to be collected and distributed, usually government organisations
  • Organisation has to be authorised, financed and justified

Most policy analysis (pre-implementation) however only focuses a limited sub-set of the tools which are actually to be used. Typically these are the issues of the use of authority (laws, rules, regulations) and treasure (taxes and exemptions, benefits, subsidies). Regulatory Impact Assessment, for example, is now a highly developed field. Likewise, cost-benefit analysis has long been with us. The equivalents for information and organisation simply do not exist.

Informational tools are mostly completely ignored even though they form an intrinsic part of nearly every policy initiative. Governments everywhere have increased their information budgets enormously and in the UK this has been often commented upon although rarely analysed. The part that informational strategies play in success or failure of policies is rarely discussed, although there is clearly evidence (e.g. taxation; the new alcohol licensing arrangements) where it appears to be critical.

Likewise, issues of the direct organisation of activities have often been ignored in policy formulation. A good example is the Child Support Agency, where the organisation of proposed service proved a critical factor but was largely ignored in the policy debates about child support assessments.

Curiously, the issue of the organisation of public activities has become a field of policy in its own right (in the so-called New Public Management debates). But this is seen as separate and apart from policy analysis and is as being somehow ‘policy neutral’.

For example, the idea of autonomous organisations with a quasi-contractual relationship with their sponsoring organisation (e.g. civil service agencies; trust hospitals; academy schools; etc) is treated as a policy issues in its own right, separate from other policies in the same domain (e.g. preventative versus acute health policies).


Given the above analysis, it will come as no surprise that the official and academic literature has little to say about policy tool choice. This neglect has two sources. Firstly, the issue of policy selection is a general area of neglect in the policy analysis literature – it is something of black-box.

The situation has a very strong parallel in the strategic management literature where Henry Mintzberg has pointed out that it has a great deal to say about strategic analysis and even comparison of options (on often complex criteria) but very little to say about actual choice, except in an anecdotal fashion (Mintzberg 1994).

In policy the situation is almost precisely the same. The second factor however, which reinforces this problem, is the absence of analysis of policy tools and their combination in specific policies.

Despite the absence of advice about how to actually choose, the espoused view in most official and some academic literature is that this is a rational process and objectively evaluating alternatives and choosing between them on the basis of a set of (unspecified) criteria.

This espoused position has already been subject to numerous critiques which include:

  • ‘Bounded rationality’ – the idea that there is never sufficient information about all the available options and even if there were human information processing capacity is so limited (in a rational sense) that the best that can be hoped for is approximate judgements founded on limited information
  • ‘Incrementalism’ – the idea that policy legacies, ‘path-dependence’ and pluralistic politics means that policy choices are in reality restricted to a few variations on existing themes
  • ‘Public choice’ – the idea that individuals act as ‘rational utility maximisers’, including voters, politicians and bureaucrats, so that actual policies emerge from a complex system of bargains and negotiations rather than rational analysis

In terms of policy tools, it does seem that there is some evidence that policy makers in particular fields are ‘habituated’ to specific configurations of tools – e.g. they predominantly rely on regulatory or incentive-based approaches and rarely even consider alternatives. This seems to apply not just to the actual decision-makers in government but to the wider “policy networks” within which they are embedded which seem to have predominant policy tool ‘mindsets’.

Having said that, there are also clearly instances of dramatic shifts in pre-dominant policy tool use. Much of the public sector reform movement of the past two-decades has included an attempt to move from provision (organisation) based approaches to more regulatory (authority) and incentive (treasure) based ones. This was neatly summed up in one of the public reform best-sellers as “more steering and less rowing” (Osborne and Gaebler 1992). How far governments have really shifted from “rowing” to “steering” is open to debate, but the attempt has certainly been made.

Some interesting questions which arises from the above – under what conditions do:

  • do certain policy areas and networks seem to have a predilection for certain combinations of policy tools and if so why? Does it relate to the nature of the policies themselves (i.e. some tools are more appropriate to say road building than preventative health)? Or is it more a case of habit and the inertia of past policies?
  • if abrupt shifts in policy tool choice occur, why and how does this happen?


As already indicated, implementation of public policies rarely follows the formal course set out in policy statements or legislation. The academic Nils Brunsson, writing in an organisational context, uses the categories of “talk”, “decisions” and “actions” (Brunsson 1989).

These apply to policies too. Policy ‘talk’ – the rhetoric by which policies are justified – is often far ‘purer’ and uncomplicated than the actual policies which are “decided”. In turn, policy “action” can often turn out to be not merely different but sometimes even the direct opposite of the policy rhetoric.

This is especially relevant to the discussion of policy tools because policy “talk” often revolves around a very limited story about what tools are being used in a specific policy – e.g. just talking about legislation or funding and neglecting other important aspects which undoubtedly will form part of the decisions and actions to implement policies such as organisation or information. Communications campaigns are, for example, almost ubiquitous in most new policy initiatives.

In some cases they may just be informational but in others they may be a critical component. Where behaviour change is a key component of policy objectives this is the case. Likewise, in many ‘human services’ areas – education, health, etc – there is a strong element of what is called ‘co-production’ – i.e. they require active involvement of people to be successful. So ignoring informational and organisational aspects can be extremely detrimental to policy implementation.

Action usually involves some feedback-based adjustment of implementation to take account of actual impact. These usually involve so-called single-loop learning or formative evaluations. This is where the role of policy tools as ‘detectors’ as well as ‘effectors’ becomes important. Application of any tool invariably produces useful data about the object of the policy, which can in turn be used to modify the policy.

For example, the application of a regulation to licensed premises for alcohol consumption will reveal how many there are and how long they stay open and lots of other details (e.g. even the sex or age profile of pub landlords). Such information is potentially useful for both single-loop adjustment and double-loop reassessment of policy goals.

Another critical action issue is the context within which policy tools are deployed. There are several types of contextual factors including importantly: complexity; culture and trust.

Policy tools may be implemented in areas of relatively stable environments, well-known cause-and-effect linearities and easily observable outputs and outcomes. On the other hand they may have to operate in unstable environments, with complex non-linear systems and hard to observe outputs and outcomes. Increasingly many ‘wicked issue’ policy problems seem to be more at the latter end of the spectrum than the former.

In ‘far from equilibrium’ and ‘near chaotic’ complex, non-linear, systems it may be that experimentally based, iterative, interventions using differing mixes of policy tools may be more appropriate. With hard to observe outputs and outcomes, more focus on processes and lower level outputs may be appropriate.

This is the conclusion of some ‘complexity theorists’ (Axelrod and Cohen 1999) and has also recently been recognised by the OECD in their recent 20-year survey of the results of modernising government initiatives (OECD 2005). Given the well-known “butterfly effect” in such systems, it may well be that apparently weak levers can sometimes have disproportionate effects compared to rather more heavy-weight interventions. An example might be the highly successful AIDS awareness campaign of the 1980s, which worked far better than many regulatory and incentive based such policies.

The second contextual factor worth mentioning is culture. It is pretty clear that socio-cultural factors can have major effects of the likely impact of different policies. Studies of national cultural difference (Hampden-Turner and Trompenaars 1993; Hofstede 2003) imply strongly that certain types of interventions – e.g. persuasion-based ones – may be better received in some cultures than in others.

Even within national cultures for complex historical reasons some issues are more amenable to some tools rather than others. Whilst one regulatory reform may be widely welcomed, others might be condemned as ‘nanny statism’ purely based on the nature of the issue.

Thirdly, a in closely linked to the issue of culture, is the issue of the level of trust and social capital in society or in a specific policy domain. Doctors can get away with highly intrusive interventions into people’s lives which would be deemed highly unacceptable coming from a police office or a bureaucrat, for example. The crucial differentiator is trust. This also affects the choice of policy tools and there likely reception in action.


This quick canter through the issues of applying a “policy tools” perspective to policy analysis, choice and action has highlighted the utility of such a perspective. It clearly could add a substantial amount to the practice of policy making and implementation. It could:

  • Encourage more creative thinking about alternative policy approaches, something which is widely acknowledged to be a current practice
  • Provide a useful framework for comparing and contrasting policy options which moves beyond current thinking
  • High-light neglected aspects of policy action – especially the issues of nodality and organisation – both of which have, in different ways, been poorly integrated into policy analysis and implementation

However it is also clear that insufficient is currently known about policy tools and the configurations of tools actually deployed in real policy areas. The failure to use a “tools” approach means that many policy analyses and evaluations simply miss highly significant aspects of the actuality of policies in practice. Issues which may have been, or may be, critical to the success or failure of policies (such as nodality and organisation) have often been largely ignored.

An agenda is needed to explore these issues further. This might include:

  • Comparative analysis of actual policies from a ‘tools’ perspective, over time, between policy areas and perhaps between countries on the same policy areas (with due allowance for national cultural differences).
  • Studies of actual policy-making processes to see to what degree they reflect and use a tools perspective in practice and to what extent such a perspective might enhance and extend creative thinking about solving policy problems, especially in ‘wicked issue’ areas.
  • Establishing case examples of good practice, in which a more ‘tools’ oriented approach has been used to good effect.
  • Exploring the complex issue of configurations of, and inter-action between, different policy tools, probably through in-depth case studies.

These are just a few ideas for taking this debate forward. There is clearly something of a gap here in the policy analysis, implementation and evaluation practice and knowledge which could be a useful one to plug.

Prof. Colin Talbot

(Original written November 2005)


Axelrod, R. and M. D. Cohen (1999). Harnessing Complexity – Organizatinal Implications of a Scientific Frontier. New York, The Free Press.

Bemelmans-Videc, M.-L., R. C. Rist and E. Vedung, Eds. (1998). Carrots, Sticks and Sermons – Policy Instruments and Their Evaluation. New Brunswick, Transaction Publishers.

Blair, R. (2002). “Policy Tools Theory and Implementation Networks: Understanding State Enterprise Zone Partnerships.” Journal of Public Administration Research and Theory 12(2): 161-191.

Brunsson, N. (1989). The Organization of Hypocrisy – Talk, Decisions and Actions in Organisations. New York, John Wiley and Sons.

Hampden-Turner, C. and F. Trompenaars (1993). The Seven Cultures of Capitalism. London, Piatkus.

Hofstede, G. (2003). Culture’s Consequence’s: Comparing Values, Behaviours, Institutions and Organizations Across Nations (2/e), Sage Publications Inc (USA).

Hood, C. (1983). The Tools of Government, Macmillan.

Mintzberg, H. (1994). The Rise and Fall of Strategic Planning, Prentice Hall.

OECD (2005). Modernising Government – The Way Forward. Paris, OECD.

Osborne, D. and T. Gaebler (1992). Reinventing Government – How the Entrepreneurial Spirit is Transforming the Public Sector, Addison-Wesley.

Salamon, L., Ed. (2002). The Tools of Government – A Guide to the New Governance. Oxford, Oxford University Press.

Talbot, C. (2005). “Consilience and Performance in the ‘Art of State’.” Public Policy and Administration 20(1).




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.