Martin Williams draws on over a decade of research across six African countries to explain why civil service reforms so often fall short, and what leaders should do instead.
Editor's note: You can listen to this podcast on Spotify, Apple Podcasts, or wherever else you get your podcasts. Reform as Process: Implementing Change in Public Bureaucracies is open access and can be found here.
In development economics, attention often falls on policy design rather than how those policies actually get implemented. Martin Williams, author of Reform as Process: Implementing Change in Public Bureaucracies, spent more than a decade studying civil service reform from the inside, partnering with the Office of the Head of Civil Service in Ghana and examining 131 reform efforts across six countries. In this episode of VoxDevTalks, he explains why even well-intentioned reforms so frequently fall short, and what civil service leaders can do about it.
What makes civil service reform so difficult
Williams traces his research agenda to a single meeting in 2014, when Ghana's newly appointed head of civil service asked him a deceptively simple question: what does the evidence say about how to improve a civil service? The honest answer, Williams admits, was not much.
"It's the type of question that I think most academics don't ask, because it's too big, it's too hard, it's too messy to answer."
Responding to that challenge, Williams set out to study the universe of performance-oriented, system-level civil service reforms he could find across Ghana, Kenya, Nigeria, Senegal, South Africa, and Zambia over the past three decades. The result was a dataset of 131 reform efforts, drawing on government documents, donor reports, academic literature, and interviews with senior civil servants, politicians, consultants, and rank-and-file staff.
Why reform efforts keep identifying the same problems
One of the most striking findings from Williams's research is how consistent civil service reform documents are across countries and decades. Whether in West Africa or Southern Africa, in the 1990s or the 2020s, the same diagnosis appears: civil servants lack motivation, accountability, and performance culture.
"You could almost copy and paste these descriptions from one country to another, from one reform effort to another, over time."
The persistence of the same diagnoses, however, has not produced consistent solutions. Not one of the 131 reforms Williams examined fully achieved its stated goals. Most managed to change something on paper while leaving day-to-day practice largely untouched.
The limits of formal rules and performance incentives
The most common approach to civil service reform – revising formal rules, processes, and structures – faces a fundamental problem. Many of the things that determine whether a civil service performs well cannot be written into a job description or a set of KPIs. Innovation, coordination across teams, and going beyond the minimum required are all what Williams calls "non-verifiable" actions: things a third party cannot objectively confirm were done well.
Formal accountability systems cannot govern these behaviours. Worse, they can actively discourage them. Williams found that performance-linked incentive schemes – the most common category of reform across his six countries, with 34 separate efforts – failed in striking fashion.
"In zero out of these 34 cases were differentiated incentives actually delivered on a sustained basis."
Managers consistently responded by setting soft targets, awarding identical scores to all staff, or treating the process as a box-ticking exercise. The systems did not work because the underlying problem – how to motivate non-verifiable behaviour – was not addressed.
The problem with treating reform as a project
Beyond the content of reform, Williams argues, the way reform is approached is itself a major obstacle. Most reform efforts are "projectised": they have a predetermined end date, a dedicated budget, their own acronym and implementing team, and a focus on measurable outputs that can be reported to donors and politicians.
This structure distorts reform in multiple ways. It encourages a focus on formal, verifiable changes – a new policy exists, a law has been passed – at the expense of the harder-to-measure shifts in culture and practice that actually matter. It also undermines credibility. Civil servants who have lived through multiple failed reform cycles are quick to recognise a project that will end in a few years with little lasting impact.
"As soon as you introduce a predetermined end date, you're just reinforcing this existing perception that civil servants often have, which is that this is just a song and dance that's going to happen for a couple of years, and then people are going to move on and things will go back to normal."
Three rules of thumb for effective reform
Against this backdrop, Williams sets out three principles drawn from the reforms that did produce meaningful, sustained change. The first is to improve performance within existing rules before creating new ones. Working within the current formal framework removes the need for lengthy legislative processes and allows change to begin immediately, building credibility through small, tangible actions.
The second rule of thumb is to treat change as a series of small steps rather than a single transformative intervention. Williams draws an analogy with physical health: getting fit requires ongoing effort, not a single burst of exercise followed by rest. The goal is not to be unambitious, but to recognise that credibility is built incrementally.
The third principle is to decentralise reform leadership. Rather than driving change top-down through a visionary plan, effective reform leaders act as catalysts, creating conditions in which teams across the organisation can identify and address their own performance problems.
"What you need to do is find what things that you can do as a central actor or a leader that help all of those organisations and teams solve their own problems."
Counterintuitively, Williams notes that high-profile, personality-driven reform efforts tend to guarantee their own reversal: once the champion leaves, the change goes with them.
Reasons for cautious optimism
Despite the sobering track record of reform, Williams says he came away from his research more optimistic than when he began. In Ghana, micro-level data collected alongside the Office of the Head of Civil Service demonstrates meaningful improvements in management quality and organisational performance in recent years. The evidence shows that reform is not impossible – only that it is frequently approached in ways that make success unlikely.
"There's so much frustration with the status quo that giving civil servants small but meaningful opportunities to improve can be incredibly powerful."