There is no easy definition for what decentralization is. If you ask ten different people about it, you will likely get ten different answers, according to Paul Smoke, professor of Public Finance and Planning at New York University.

Most tend to agree that decentralization is a broad term which encompasses the many different ways responsibilities are delegated to local authorities. The three common forms of decentralization are: devolution, which transfers power to semi-autonomous local governance represented by locally elected officials; deconcentration, where responsibilities are delegated to subnational administrations that are still part of the national government; and finally, delegation, where national governments give only specific functions to subnational authorities, or even non-governmental actors. 

Smoke notes that devolution is what the international community tends to push for, but most countries have a mix of these types. In Africa, each country has pursued a different decentralization plan according to context and the driving political forces. 

Smoke recently presented his insights on “Decentralization in Africa” at a talk for the American University in Cairo’s Public Policy and Administration program. Smoke has written countless articles on decentralization and is the co-editor of several books on the subject, including Making Decentralization Work: Democracy, Development and Security and Decentralization in Asia and Latin America: Towards A Comparative Interdisciplinary Perspective. 

One Concept, Many Forms

Decentralization can be undertaken in multiple ways under different conditions. Some countries may do it for development reasons and others for political ones. Commenting on why some countries would be motivated to decentralize, Smoke said: “Most countries I know of that have decentralized substantially do it in a time of crisis.” Ethiopia, for example, which was a one-party state with a strong central government, decentralized following Eritrea’s secession, as it wanted to stabilize the situation and didn’t want to face further threats of secession, so it gave more autonomy to regional governments.

Another example is Uganda, which went through many years of civil war. After much instability, it sought to create decentralization based on grassroots democracy, so it went straight to the local level. Smoke notes that this was partly politically motivated, as the Ugandan state sought to circumvent the different regions that were historically tied to ruling kingdoms.

Both Kenya and Uganda redistributed power at the subnational level. In both cases, most revenue still comes from the national government, but in Kenya, there is a strong emphasis on unconditional intergovernmental transfer of funds, while in Uganda it is the opposite. Smoke claims that one can argue that Kenya has too much unconditional transfer, giving local government more liberties than is efficient, while Uganda gives too little, undermining the ability for local governments to adjust to local needs. The Ugandan case aptly demonstrates the idea that subnational spending doesn’t necessarily imply decentralization, as funds can be given under strict instructions on how they are to be spent, thus making the decentralization process void in this regard.

This diversity in context and application, Smoke said, should “raise caution about best practice reform” where models are built based on a specific experience and do not always work with different local conditions. Even the idea of decentralizing for stability, which Smoke gave two relevant examples of, is highly dependent on the situation the country finds itself in; “post-conflict countries have decentralized because they felt it could help stabilize the state, although others centralized for stability.”

Context also matters with regards to different types of regimes and political systems. “The kind of decentralization I’m talking about,” Smoke said, “tends to assume some degree of democratic regime in the country, but it does work in other cases,” adding that some of the best citizen-based monitoring systems he has seen were in one-party states.

As such, there is no single model that fits all. Rather, it is a matter of maintaining balance. Smoke warned that “one can go too far in decentralizing power and authority over resources”. On the other hand, if the central government holds too tight a grip, then it may lose out on efficiency gains, more successful locally adapted policy implementation and service delivery, and vibrant local politics (though local elite dominance is still an issue), not to mention a potential stabilization mechanism in times of crisis.

History and Implementation

Smoke sees decentralization as a public policy tool that must be adapted to the local context, even between different localities within one country. 

Decentralization in developing countries became increasingly common from the 1980s onward, and a lot of aid has gone into supporting intergovernmental systems. In Africa, many of the decentralization processes have been ongoing for the past few decades. However, the context and history of each differs greatly from country to country. Not only are there disparate formal frameworks for decentralization in African countries, but there are also varying degrees  of implementation—“a lot of cases where experts would say this is good, but it’s not being implemented.”

Smoke said that there has been some evidence suggesting a general increase in local services, although the existence of reliable, generalizable evidence is lacking. Despite that, capacity development remains a major problem in Africa. When decentralization takes place without the required capacity-building measures, or without a proper strategy, then failure is likely, and governments would often recentralize power in their hands after such failures. This has in fact become a recent trend in many African countries, such as Uganda. 

In the early days of the move toward decentralization in developing countries, many  adopted a “sink or swim” approach—when national governments create some sort of decentralization scheme and all local governments have to cope with it or fail. More recently, however, a new, more development-based approach is garnering attention, where local governments are allocated a certain degree of responsibility, which they are able to handle and can progress at their own pace. Furthermore, decentralization is beginning to be seen as more than the simple “passive assumption of former central functions,” making way for more general, non-specific empowerment of local governments in different ways.

There is also more attention given by policy analysts to a political-economy analysis of decentralization now, as “decentralization is not just a technical thing” as Smoke puts it. More emphasis is being placed on  performances and incentives as well as data collection, management, and analysis in what appears to be a strengthening of the decentralization process to support a flexible, evidence-based approach.

How to Make It Work

According to Smoke, there is a need for a national framework which clearly outlines the roles and delegation of tasks by the different levels of government, and defines the relationship between them, and the scope of their interaction with other actors through public-private partnerships. Moreover, there are also economic and political dimensions; local governments must be given certain fiscal powers and a share of national revenues, have access to taxes, and enjoy a fiscal responsibility and borrowing framework if they were to borrow. They also have to be held accountable by their citizens; even in democratic states with no real local elections, there are feedback-collection mechanisms regarding service delivery.

Clearly, there are many aspects that have to come together so the benefits of decentralization can be sown, but no country can do all of this at once. Each aspect, whether economic, political, or administrative, is a huge undertaking in and of itself, and each actor has its own interests to pursue—“the ministry of finance cares about finance; the ministry of health cares about health.” It can’t all be done at once, so Smoke advises that it must be done gradually. “Do something significant on the local level which signals the possibility for change” he said, recognizing that such a task is not easy in its own right, as “it’s very difficult to figure out a starting point.” Ultimately, central governments have to think “not just about how to control local governments, but also how to empower them.” This raises the question of autonomy and accountability. 

Autonomy is generally considered desirable as local governments find themselves better able to respond to local concerns and are able operate in a more integrated manner. All decentralized systems have some framework for autonomy. However, autonomy is never absolute, and with autonomy comes the need for accountability. 

Smoke clarified that there is always a need for a strong national government, even in the most decentralized of systems, as the national government plays  a significant role in ensuring the accountability of subnational government. This is called upwards accountability, where local governments are answerable for their performance to the national government, which has in its power defined tools to discipline and leverage local governments according to the framework in place.

Downward accountability, on the other hand, involves citizens holding local governments accountable, possibly through elections, but also through other channels of civic action and monitoring. Smoke noted that “elections are a very blunt instrument of accountability; just because there are elected councils doesn’t mean that there is accountability” and that “people need to develop a comfort for using additional accountability mechanisms” through active civic engagement. 

In addition to these two forms of directional accountabilities, there also needs to be an element of horizontal accountability where local authorities can hold accountable service providers that were previously answerable to the national government. Smoke describes this as “one of the most difficult transitions for any decentralizing system” as it involves a tricky restructuring  of the administrative hierarchy to insert service providers, who previously dealt with the national government, into the local government’s new jurisdiction. 

Smoke said that one of the biggest challenges for achieving decentralization in African countries is getting the right balance between upward and downward accountability. Usually, there is too much of either, and the outcome is that the former doesn’t respond to local citizens, as it’s too restricted to the national government’s instructions, while the latter can’t handle responsibility, because it doesn’t follow the standards set by the national government. 

The Challenges

A certain political will is needed to decentralize, but that alone is not enough. Many challenges present themselves on the bureaucratic front. Some bureaucrats will try to obstruct the decentralization process as they stand to “lose control over powers and functions they had in the past”. In this sense, the separation between those who make the decisions and those who execute them becomes clear, as well as an understanding that both sides of the equation have to be on board for smooth sailing.

Another issue is the assumption that “local governments will behave—but we know that there are power struggles, local elites who may capture the locally elected councils, leaving citizens with little or no voice”. This is why Smoke described elections as a “blunt instrument”. True citizen participation is much harder to capture, and there has to be an awakening where citizens “understand how to hold their local government accountable”.

Further challenges arise in implementation, where it can come too quick, and overwhelm local governments, or too slow, where token, useless responsibilities are given “that serve no purpose except to make it seem that the center is decentralizing”.

Finally, Smoke reemphasized the importance of capacity-building to address the challenges of incompetence—“just because a place is big or small, or because it’s urban or rural, doesn’t mean it’s going to be able to do its job … it has to be based on its capacities.” Capacity-building must be targeted and it needs to be learned on the ground, Smoke said, adding that there is also a lack of awareness with regards to the political challenges of decentralization.

Decentralization is ultimately a balancing act, between the technical and the political, between going slow or fast, and between control and autonomy. The bottom line is that decentralization reform has to be realistic, and it is important to develop the capacity for civic engagement because citizens in a decentralized system have to be better able to hold local governments accountable.

Ultimately, “decentralization is not a one-time action, it’s a learning process” and it must be treated as such.