The decentralisation of government and public administration is a huge phenomenon across all of the world’s regions, cultures, income levels, and politico-administrative systems. It began in the aftermath of decolonization and built through the 1980s and 1990s.
Around the early 2000s, many of us working in the field thought the decentralization wave was cresting. Pranab Bardhan and Dilip Mookherjee (2006) produced a seminal book – Decentralization and Local Governance in Developing Countries – which took stock of what we knew about the different types of decentralization in the developing world. But to the surprise of many of us, the decentralization wave continued to build. All around the world, there were more reforms at different levels of devolution with greater creativity about the technologies employed.
At about the same time, a huge amount of new, finer-grained data on not just fiscal flows, but also political, social, and cultural variables became available across many developing countries. This allowed us to ask difficult, nuanced questions that could not be answered before.
Our analysis shows that local development is boosted most only when local leaders are elected by majority voting rather than oligarchy.
The book which we have recently published with LSE Press, Decentralised Governance: Crafting Effective Democracies Around the World, takes advantage of this new evidence to take stock of the state of knowledge on decentralisation in the world today, with a special focus on governance and democracy-building in developing countries. Its chapters include thematic, critical surveys of recent advances in decentralisation, combined with focused, cutting-edge empirical studies.
Empirical chapters include broad cross-country studies as well as detailed explorations of Bangladesh, India, Pakistan, China, Indonesia, Ghana, Kenya and Colombia. We’re proud to include younger, up-and-coming scholars alongside some of the most established and influential thinkers in this field. Many of these authors are themselves from developing countries, and their papers employ the latest evidence and methods to explore complex issues analytically with a blend of qualitative and quantitative data.
Overall, we update Bardhan and Mookherjee (2006) in two novel ways:
- We explore the various impacts of decentralisation on local politics that remain little understood (chapters 4-8);
- We identify three key mechanisms through which decentralised governance can effectively deliver public goods and services: (i) innovative recent reform initiatives; (ii) corruption monitoring; and (iii) enhanced local insights in the past two decades (chapters 9-12).
The remainder of this article reviews some of the biggest themes that emerge across the book’s chapters, and concludes with a few over-arching lessons.
What Decentralisation Is and Is Not
One of the key ideas that underpins all our analysis is the recognition that decentralisation is not ‘a thing’. It is many things. It is not a switch one can flip with results that are predictable and symmetric.
Rather – and by definition – decentralisation generates a heterogeneity of responses that, in any particular dimension (e.g. education, health, transportation), differ from one another as much as the underlying districts, provinces, municipalities, etc. differ from one another.
Democratic accountability and government responsiveness can be greatly improved with effective decentralisation.
Any country that decentralises should expect not one tidy outcome replicated in many localities, but rather a wide variety of outcomes that range from the strongly negative, through the null (“no change”), all the way to highly positive responses. Such heterogeneity is not a problem. It’s the way decentralisation is supposed to work. Indeed, in a deep sense it’s why you do it.
This idea is tied to wealth of conceptual and methodological insights that we’re only beginning to comprehend. For over 50 years, researchers asked questions of the type: Is decentralisation good or bad for X? Here, X represents a significant policy-related outcome, such as primary enrolment rates, access to water and sanitation, or levels of corruption. Such studies approached decentralisation as if it were a technocratic issue. They treated the specifics of reform – the many decisions about how to unpick centralised public services and decentralise which components to what levels – as if these were given, choosing instead to compare ‘the effects of decentralisation’ across countries.
Researchers were not entirely unaware of these issues. Their methodological stance was always more a convenience than an assertion of principle. But it coloured the empirical literature all the same, affecting what questions were asked, and how studies were structured. Only recently have researchers began to internalise that the many decisions about how to decentralise precisely which state functions are not fundamentally technocratic issues. They are political issues everywhere, all the time.
The Complex Relationship Between Politics and Decentralisation
Decentralisation’s political effects are carefully considered in chapter 4 of Decentralised Governance, ‘Realising the promise of partial decentralisation’, where its authors, Shanta Devarajan and Stuti Khemani, point out that decisions to decentralise are often taken by leaders seeking political advantage. They are implemented – or not – by officials whose power and status will be directly affected by reform.
Democratic accountability and government responsiveness can be greatly improved with effective decentralisation. This is achieved through the simple but profound expedient of changing the incentives that local officials face. In a centralise regime, local officials’ incentives are upward-pointing, towards national authorities. By devolving power and resources to locally-elected administrations, decentralisation re-orients officials’ incentives downwards, towards local voters. This is one of the simplest but most powerful lessons of 50 years of reform. In so doing, decentralisation ultimately affects not only local but also national public goods, as well as the responsiveness and accountability of the entire state.
Decentralisation often leads to more, better people getting involved in local politics.
Done correctly, decentralisation should make local governments better attuned to local economic conditions. This should, in turn, help boost economic growth. And the sum of many localities boosting economic growth is stronger national-level growth.
But this concept of decentralisation being ‘done correctly’ hides more than it reveals. Countries decentralise in very different ways, devolving different sets of powers over different public services to different levels of subnational government, with different revenue-raising powers and different degrees of subnational democracy. The reason for this is deeply political.
Any decentralisation programme entails winners and losers, and how the programme is designed and implemented will largely determine who those winners and losers are, as well as the size of their gains and losses. Countries have different kinds of political parties and movements. Those movements throw up different kinds of leaders. And such actors inhabit different political equilibria. This is why the types of decentralisations that get implemented can vary hugely from one country to another.
Analysing the political nature of decentralisation in chapter 4 of Decentralised Governance leads its authors, Devarajan and Khemani, to the striking conclusion that most decentralisations are partial. Not only are they not full expressions of some similar underlying blueprint, they are not even full expressions of the diverse blueprints that national reformers publicly declare or write into law. Politics systematically gets in the way of a full, clean implementation.
By definition, decentralisation requires centralised administrations to hand over power and resources to lower-level politicians with independent mandates. Put simply, central officials often don’t want to do this. They’d rather keep power and resources in their own hands. So they fight a rear-guard action to limit, or even undo, reform. The extent and nature of the reform that is finally implemented is the equilibrium outcome of many battles between those who want to make reform and those who want to stop it.
The upshot of this conflict between central officials and lower-level politicians is that most decentralisations are incomplete in the sense that economic theory would predict, and often feature large mismatches between devolved responsibilities and accountability. These partial decentralisations nonetheless represent political equilibria that balance competing forces, making it difficult to change them at the margins.
Although decentralisation can be undertaken by both democratic and autocratic regimes, our analysis shows that local development is boosted most only when local leaders are elected by majority voting rather than oligarchy.
Devarajan and Khemani also show that politics provides strong grounds for hope. In many countries that have decentralised, greater contestation in local elections has led to improved service delivery. This is because decentralisation often leads to more, better people getting involved in local politics. Reform shines a brighter light on local governance and brings more information into the public realm. As a result, citizens are better able and willing to judge officials’ performance. This leads to greater accountability, better governance, and faster development.
But in decentralisation, politics is not everything. In recent decades, a great deal of experimentation has occurred in explicitly non-political realms, known as mechanism design.
Designing Decentralised Mechanisms
Successful governance is not solely a question of politics. In addition to devolving power and resources, effective decentralisation also implies adapting structures, rules, norms, and behaviours to new actors and dynamics that a centralised system may not have contemplated. Decentralised Governance refers to this broad set of issues as mechanism design.
In chapter 2, we present detailed efficiency criteria for deciding which powers to decentralise to which level of government, and how to design tax and transfer systems that can adequately fund the public expenditure systems that result. These are the tools of what is often called intergovernmental relations; we might call them ‘broad mechanisms’. In theoretical terms, those insights largely summarised the state of play as the 1990s ended.
Dilip Mookherjee brings that story forward to the present day in chapter 3, focusing on the surprising number of policy experiments and technocratic innovations that have been trialled across a wide array of countries in recent years in a bid to tackle problems of elite capture and clientelism. The worry is that these rule-based interventions may, in effect, recentralize public expenditures, something that must be taken into account before a reform is embarked upon.
Recent innovations include the issuing of biometric identification cards in India, where they have been used to verify beneficiaries of employment programs. This has reduced programme leakages from ‘ghost beneficiaries’ by 41%. Beneficiaries were paid more quickly, and reported earnings rose 24%. Most impressively of all, these results were achieved at no additional cost.
Mookherjee also identifies the mechanism of hiring non-elite citizens as monitors to help reduce elite capture of public benefits. As part of these efforts, programme management is often contracted out to NGOs or private firms. In some instances, formula-bound programmes reduce the authority of locally-elected officials to allocate public funds, in effect replacing politics with technocracy in the name of greater targeting precision.
Integrating ideas from across these experiences, Mookherjee suggests a big data approach to the decentralised provision of services and benefits that makes use of survey data to predict the level of poverty of each individual in a country. Benefits could then be distributed via a nationwide ID system with biometric identification, combined with electronic transfers to low-cost bank accounts or mobile phones. Such a system could improve pro-poor targeting significantly at very low cost, while reducing losses and distortions due to capture, corruption, and bureaucratic inefficiency.
All of these measures are examples of technocratic alternatives to decentralised politics that achieve some of the same outcomes, but in very different, non-political ways.
Concluding Thoughts
Decentralisation is not in itself a good or a bad thing. Designed strategically to take advantage of the insights offered in our book on politics and mechanism design, it can promote democracy, efficiency, and accountability. We can say that with confidence because we document it there.
First-order theorising is of little use in this field. Decentralisation is complicated. Reformers face a daunting set of choices as they design and execute real decentralisation programmes. The low-hanging fruit available to first-order decentralisation theory was largely plucked by previous generations of thinkers. The issues that remain are complicated and nuanced; potential solutions are highly contingent. These ‘details’ are not only not trivial, they are crucial if decentralisation is to fit a country’s challenges.
Getting decentralisation right is difficult, but also immensely valuable because it can improve the quality of a country’s governance. But how do we ‘get decentralisation right’? The short answer is via a combination of political and technocratic measures that work with the grain of national and subnational incentives to target resources and hold public officials to account, e.g. through various monitoring processes that ensure transparency and accountability.
The chapters in this book shine a light on how to do this. They answer questions about when and why decentralisation works across the globe under very different conditions, what it can achieve when it works, and when and why it often fails. In so doing, it points to important gaps in our knowledge where further research is needed.
Decentralisation works best when local governments are democratically elected. Many readers in the West may find surprising that it is not so much the level of income as its distribution that is important for securing the efficacy of decentralization. High income inequality opens the door to elite capture and clientelism that disrupt the decentralisation process.
We must think harder about the power and the potential of decentralisation. A casual glance around the world shows that citizens are unhappy as inequality soars. Democracy is stumbling everywhere. If we are serious about responding to political unrest, we need to consider alternatives that generate better, more effective governance. It’s a big challenge that has never been more important.
References and Further Reading
- Besley, T., J.P. Faguet and M. Tommasi. 2003. A Synoptic Guide to Decentralization and Intergovernmental Relations. Initiative for Policy Dialogue, Columbia University.
- Eaton, K., J.P. Faguet, I. Harbers, A.H. Schakel, L. Hooghe, G. Marks, S. Niedzwiecki, S. Chapman Osterkatz and S. Shair-Rosenfield. 2018. “Measuring and theorizing regional governance.” Territory, Politics, Governance, 7 (2): 265-283.
- Faguet, J.P. 2004. “Building Democracy in Quicksand: Altruism, Empire and the United States.” Challenge: The Magazine of Economic Affairs, 47: 73-93.
- Faguet, J.P., Q. Khan and D.P. Kanth. 2021. “Decentralization’s Effects on Education and Health: Evidence from Ethiopia.” Publius: The Journal of Federalism, 51 (1): 79-103.
- Dilip Mokherjee. 2023. Decentralised targeting of transfer programmes: A reassessment. InternationalDevelopment@LSE blog.
- Anirban Mitra and Sarmistha Pal. 2023. Impact of fiscal decentralisation and ethnic heterogeneity on choice of local polity. Ideas for India blog.
- Zaki Wahhaj and Abu S. Shonchoy. 2023. Birth Registration in the Developing World: Bridging the gap between policy and practice. InternationalDevelopment@LSE blog.
Profs. Jean-Paul Faguet and Sarmistha Pal are co-editors and contributing authors to Decentralised Governance: Crafting Effective Democracies Around The World. This book is free to read and download from the LSE Press website via Open Access publishing.
