1 Introduction
This paper identifies factors to be considered when thinking about the extent to which functions within government should be centralised or decentralised. The intention of the paper is to assist those considering questions about the extent to which functions such as human resource management, purchasing, or specific output delivery should be centralised, both across the public sector and within agencies. The paper does not suggest specific solutions, which will depend on the particular decision and circumstances in each individual case. Often the solution lies in a “tight/loose” fit between centralisation and decentralisation[1]. By this “tight/loose” fit I mean a pattern whereby there is “tight” or centralised direction and coordination over the major objectives and “loose” or decentralised discretion and autonomy over many of the ways to achieve those overall objectives.
This is a high level paper that attempts to highlight the relevant issues to consider when thinking about centralising or decentralising decision rights rather than anything more prescriptive. Centralisation (or decentralisation) is a complex and multidimensional issue. It is partly for this reason that the paper does not suggest any specific solutions. The solution in any particular case will involve tradeoffs between the factors identified in this paper and is likely to involve some value judgement regarding the ranking of the various factors. This paper aims to identify the relevant issues to consider when faced with a (de)centralisation question and in doing so, should provide a pathway through some of the complexities involved.
The paper seeks to answer the question: “According to what factors should authority be allocated?” Generalisations in this area can be dangerous and the solution is almost always likely to involve a balance between centralised and decentralised decision-making. Nevertheless there are a number of common factors that are generally applicable to (de)centralisation questions. This paper identifies those factors in order to provide some guidance for decisions regarding the level of centralisation or decentralisation.
In determining the degree of centralisation, various values are in tension. Advocates of centralisation often contend that it ensures uniform and consistent standards, minimises inequalities, avoids the duplication of services, allows for the achievement of economies of scale, and increases coherence and coordination. Advocates of decentralisation, on the other hand, point to the enhancement of local autonomy and empowerment, greater customisation and innovation, and increased participation as benefits of decentralisation (Boston, Martin, Pallot and Walsh 1996).
Economic, social, managerial and constitutional perspectives offer a number of theoretical frameworks that are useful in considering issues of (de)centralisation. This paper draws together arguments from across these perspectives and identifies common factors applicable to (de)centralisation questions.
Each of these factors points either to centralisation or decentralisation of decision-making, all other things being equal as shown in Tables 1 and 2. Resolving (de)centralisation questions requires tradeoffs to be made (such as that between the agency costs of inconsistent goals that increase as decentralisation increases, and the costs of gathering and processing information, that decrease as decentralisation increases). Different factors may dominate depending on the particular decision and circumstances.
| Factor | Examples |
|---|---|
| Decisions requiring knowledge of central information, that it is not practical to transfer, an understanding of the needs of the system overall, or knowledge of a number of different organizations and non-localised tradeoffs | Whole of government resource allocation |
| The need for accountability for key judgements to remain clearly with the democratically elected government | Key judgements about public policy values, trade-offs between organisations, broad resourcing and taxation issues |
| The need for the government to manage processes closely in order to manage risk more directly | Management of an international relations crisis |
| Decisions where it is possible to achieve economies of scale | New Zealand Debt Management Office’s overnight cash “sweep”[2] |
| Decisions where it is possible to achieve economies of scope | “Brokerage” roles, budget-setting functions, spreading “best practice” |
| The need for coherence and coordination across the public sector | Leadership on core public policy values |
| The need for uniform and consistent national standards and guidelines | Benefit entitlement and eligibility rules, tax administration and rulings |
| Issues where it is difficult to specify the contract or provide suitable incentives for a subordinate to carry out government objectives | Defence |
Often, the solution to questions of (de)centralisation is likely to involve a combination of centralised and decentralised functions. This solution is likely to be characterised by a “tight/loose” pattern whereby there is “tight” or centralised control over the major objectives that is then joined by “loose” or decentralised discretion over the ways in which those objectives are achieved to varying degrees.
| Factor | Examples |
|---|---|
| Localised non-transferable (or transferable only at a prohibitively high cost) information, knowledge and relationships | Judgements about supplementary benefit payments |
| Issues requiring customisation, innovation and a flexibility to respond to localised conditions | Business development initiatives |
| Situations where there is a span of control problem | Specific functions are decentralised from doctors to nurses (and from nurses to nurse aides) as doctors are unable to make all the decisions relating to the care of their patients |
| Issues where community involvement, local empowerment and local participation are important | Community employment schemes |
| The need to avoid an unnecessary concentration of power and the risk of abuse of power by distancing a decision from political or central bureaucratic control | Specific functions may be decentralised to limit the power of Ministers to other entities such as the Police Complaints Authority and the Commerce Commission (decisions under the Commerce and Fair Trading Acts) |
This paper discusses questions of (de)centralisation in more detail, providing some background on the various perspectives, (economic, social, managerial and constitutional), the theoretical frameworks within each of these perspectives, and the arguments that can be drawn from them. From these arguments common factors relating to questions of (de)centralisation are identified.
Notes
- [1]The concept of a “tight/loose” fit between centralisation and decentralisation originated with Peters and Waterman, 1984.
- [2]The New Zealand Debt Management Office (DMO) is the debt manager of the central government of New Zealand. DMO manage the risk associated with the Government’s fixed income portfolio. The account balances of all New Zealand dollar Crown and departmental bank accounts are consolidated at the end of each banking day. The net balance is then moved to the Crown settlement account at the Reserve Bank of New Zealand. This practice (called the overnight cash ‘sweep’) makes use of economies of scale to help achieve efficient cash management.
