VoxDev Development Economics

How do courts work when they work well? You would expect them to be impartial, neutral, and consistent. In much of the Global South that is a tall order. 
So when courts fall short of it, are they failing?
Development institutions ask states to build strong courts on the North American and Western European model. Good governance follows, they argue. This model treats poorer, less democratic systems as deviations from a norm rather than as institutions doing different work.
Fiona Shen-Bayh (University of Maryland) joins Tim Phillips to review the evidence on what courts in the Global South actually do, and who they help. Where the state is weak, customary elders, NGOs, even rebel groups step in to adjudicate, and people often trust these forums more than the state's own courts. 
Taliban courts in Afghanistan upheld due process during civil war. Dictators sometimes build genuinely independent courts, because property rights attract investment and citizens' lawsuits tell the centre what local officials are doing.
The research behind this episode:
Rios-Figueroa, Julio, and Fiona Shen-Bayh. 2025. "Courts in the Global South." Annual Review of Political Science 28. 
To cite this episode:
Phillips, Tim, and Fiona Shen-Bayh. 2026. "Courts in the Global South." VoxDev Talks (podcast). 

Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.

About the guest

Fiona Shen-Bayh is Assistant Professor of Government and Politics, with a joint appointment at the College of Information Studies, at the University of Maryland. Her research spans authoritarian regimes, judicial politics, and the use of legal and judicial institutions as instruments of power, often drawing on digitised archives and text-as-data methods. Her book Undue Process: Persecution and Punishment in Autocratic Courts (Cambridge University Press) won the APSA-IPSA Theodore J. Lowi First Book Award, the Giovanni Sartori Book Award, and the Juan Linz Best Book Prize.
The paper is co-authored with Julio Rios-Figueroa, Professor in the Department of Law at the Instituto Tecnologico Autonomo de Mexico (ITAM), whose work spans comparative judicial politics, the rule of law, and empirical legal studies, with a focus on Latin America.
Research cited in this episode

The triad logic of conflict resolution. Drawn from Martin Shapiro's Courts: A Comparative and Political Analysis (1981). A court is effective when two parties appeal to a third to settle their dispute, and three conditions hold: the parties believe the third party is impartial; the third party is neutral, not predisposed to favour either side; and the rationale for the decision is consistent with existing norms, the idea of precedent. The review deliberately relaxes the assumption that courts are effective only when all three conditions are met.
The rule of law revival. The wave of good-governance programmes promoted by the United States and Western European governments and NGOs after the fall of the Soviet Union, presenting strong courts as a remedy for corruption, instability, and tyranny in the Global South.
The fallacy of legalism. The belief that creating law through statute, legislation, or precedent is enough to bring about social change. The phrase, from Sandra Joireman's work on property rights in Africa, names a habit of thought rooted in the Western experience, where the state has historically enforced property rights and contracts. In much of the Global South the absence of the state does not mean the absence of rules and order.
Stateness. The extent to which a state exercises authority across its territory: its monopoly on violence in the classic Weberian sense, but also the creation of law and the administration of public affairs. Where stateness is low, non-state actors fill the judicial vacuum.
Taliban and Islamic State courts. Recent fieldwork-based research finds that Sharia courts run by the Taliban in Afghanistan upheld notable degrees of due process and impartiality, offering predictability during civil war, and that the coercion associated with the Taliban featured in only a minority of the cases their courts heard.
Courts in authoritarian regimes. A growing literature shows what courts do for dictators: establish credible property rights that attract foreign capital; monitor administrative conflict, as in China, where citizens' grievances against the state feed information upward to the centre; and, at other times, repress opponents or legitimise the regime by delivering popular moral outcomes even against the letter of the law.
Political competition and judicial independence. Electoral competition can sustain independent courts in healthy democracies, partly because divided governments struggle to coordinate against unfavourable rulings. Under instability or an expected change of regime, the relationship can reverse: incumbents pack courts to entrench their interests before leaving, and judges may rule strategically to align with whoever they expect to hold power next, a pattern visible across Latin America.
Access to justice and legal mobilisation. Social transformation through courts depends on people developing a "legal conscience", an understanding of the law and how to use it, and on support structures outside the judiciary: civil society organisations, bar associations, prosecutors, lawyers, and human rights groups that help citizens bring and sustain claims.
Courts and democratic backsliding. Courts hold neither the purse nor the sword, which makes them easy targets for hostile rhetoric, legislative threats, pressure to resign, and court-packing. Courts that are neither impartial nor neutral can still stabilise a democracy while rival parties remain uncertain of each other's intentions, provided both still accept competitive elections. Once a party, especially an incumbent, abandons that commitment, there is little a court can do alone.
Digitised judicial data. The digitisation and free publication of court records across the Global South has opened large-scale, fine-grained study of everyday jurisprudence, useful to scholars and to the judges, lawyers, and litigants who can now see how the law works in their own context.

The usual way to measure women's power in politics is to count the seats they hold in parliament. But most women who take part in politics never stand for office. They vote, attend meetings, petition, protest, or try to get the water supply fixed. 
In this week's VoxDev Talk, Soledad Artiz Prillaman of Stanford talks to Tim Phillips about her new review of the research into non-elite women's participation in politics, written with Peace Medie (University of Bristol).
They are not elite women with less money, she argues. They want different things and face different constraints. Social norms can prevent them from achieving the change they want. But in the Global South there is evidence that non-elite women are using collective action to gain access to politics, and using that access to renegotiate the norms that hold them back, rather than waiting for those norms to shift first.
The research behind this episode:
Medie, Peace A., and Soledad Artiz Prillaman. 2026. "Nonelite Women's Participation in Politics." Annual Review of Political Science, vol. 29.
To cite this episode:
Phillips, Tim, and Soledad Artiz Prillaman. 2026. "Nonelite Women's Participation in Politics." VoxDev Talks (podcast). 

Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.

About the guest

Soledad Artiz Prillaman is Assistant Professor of Political Science at Stanford University and faculty director of the Inclusive Democracy and Development Lab. Her research spans comparative political economy, development, and gender, with a focus on South Asia and on how and when women gain access to politics, both as citizens and as representatives. She is the author of The Patriarchal Political Order: The Making and Unraveling of the Gendered Participation Gap in India (Cambridge University Press, 2023).
The paper is co-authored with Peace A. Medie, Associate Professor in the School of Sociology, Politics and International Studies at the University of Bristol. Her work covers gender, security, and politics in Africa, including the campaigns to end violence against women.
Research cited in this episode

Elite and nonelite women. The paper defines eliteness by access to political power, not by office held or income alone. Elites include elected representatives, but also academics and business executives whose position gives them access to power. Nonelites are those who lack that access. The distinction matters because policy aimed at getting more women into elite positions only helps everyone else if elite and nonelite women want the same things, and the evidence that they do is thin.
The income puzzle. At the individual level, income is generally uncorrelated with women's turnout; at the national level, GDP predicts nonelite women's participation only in some places. Women in paid work do participate more, but the driver appears to be the networks and information that come with a job, not the wage.
Vote agency. Showing up to vote is not the same as voting freely. Asked whether they would vote for their own preferred party or the one a male gatekeeper preferred, at least half of women in some South Asian settings say they would defer. Work by Sara Khan shows that the women with the least agency are those whose preferences differ most from the men who hold power over them.
Varieties of patriarchy. All societies are patriarchal, but patriarchy operates differently across them. In parts of South Asia it takes the form of explicit, socially sanctioned control over where women go and how they vote. In the United States and Europe it shows up earlier, as socialisation, producing large gender gaps in stated political interest. Same underlying force, different mechanics, different policy conclusions.
Quotas. More than 100 countries have adopted some form of electoral gender quota, making it the most widespread women's empowerment policy in the world. The evidence on whether quotas help nonelite women is mixed; they raise some women's participation in some places, but in others the effect is null or negative. In India, Prillaman notes campaign material for quota seats that pairs the woman candidate's name with a man's photograph.
Collective action. Networks outside the home, through women's groups, microcredit groups, churches, unions or friendship circles, raise women's participation by widening their information and giving them cover against backlash. Prillaman argues that in the Global South women are increasingly using collective action to gain access to politics, and using that access to renegotiate norms, rather than waiting for norms to change first.
More from VoxDev

Where are the Indian female politicians?, an interview with Lakshmi Iyer on why a woman winning office in India does not lead to more women standing next time.
Related reading on VoxDev

Grassroots party activism by women promotes equal political participation, in which Tanushree Goyal finds that women politicians in Delhi recruit women activists, narrowing gender gaps in political knowledge and participation.
Women's microcredit groups empower women politically, in which Prillaman shows that microcredit groups raise women's political participation in India by building their networks, not their bank balances.

This episode follows a wide-ranging panel convened at Stanford's King Center on Global Development, featuring Gyude Moore, as well as Gates Foundation CEO Mark Suzman, former USAID Administrator and Ambassador Mark Green, and Chair and Founder of the Liquidity and Sustainability Facility Vera Songwe - The future of global development: Approaches and partnerships for a new reality.
Bilateral aid to sub-Saharan Africa will fall by between 16% and 28% this year, according to the IMF. In past downturns, multilateral and humanitarian funding tended to fill the gap when bilateral aid dropped. This time those channels are shrinking too.
Gyude Moore, who ran the Liberian President's Delivery Unit under Ellen Johnson Sirleaf, thinks the contraction is structural rather than a passing effect of the Trump administration, and that recipient countries should stop expecting the old arrangement to return. He wants economic growth put at the centre of development rather than treated as one programme among several. Instead of letting donors decide which programmes are run, he says, countries should run a growth diagnostic: a way of identifying the two or three constraints doing most to hold an economy back. Governments can then reorganise their budgets around removing those constraints, and use the diagnostic to decide which offers of aid to take and which to turn down. Moore calls this “sovereignty through analytics”. Aid was meant to be temporary, he argues, and the job now is to quickly reach the point of not needing it.
To cite this episode:
Phillips, Tim, and W. Gyude Moore. 2026. "The end of aid dependency.” VoxDev Talks (podcast). 

Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.

About the guest

W. Gyude Moore is a distinguished fellow at the Energy for Growth Hub and a non-resident fellow at the Center for Global Development. He was Liberia's minister of public works from December 2014 to January 2018, and before that deputy chief of staff to President Ellen Johnson Sirleaf and head of the President's Delivery Unit, which oversaw more than $1 billion of road, power and port projects in a country rebuilding after civil war. He also lectures at the University of Chicago's Harris School of Public Policy. His work covers African infrastructure, energy, industrial policy and development finance.
Cited in this episode

The scale of the cuts. The IMF's October 2025 Regional Economic Outlook for sub-Saharan Africa, using OECD figures, projects bilateral aid to the region falling by 16% to 28% in 2025, with more cuts likely. Moore says the cuts to multilateral and humanitarian funding run higher again, and that the most aid-dependent countries have been hit hardest, through weaker health, education and nutrition systems.
Growth diagnostics. A way of finding the constraints that matter most: the one or two that, once removed, allow others to ease. Moore likens it to a doctor running tests before prescribing. The method is associated with the Growth Lab at Harvard. He suggests governments hire an independent party to run the analysis, so the findings cannot be dismissed as political.
The Millennium Challenge Corporation. A US agency that runs what it calls a constraints analysis, then funds the removal of the constraint it finds. Moore offers it as an existing model for diagnostic-led aid, while noting that it has critics.
Sovereignty through analytics. Moore's phrase for using a credible diagnostic to set the terms with donors. A government can say what it is trying to do, ask for help where it needs it, and decline what does not fit. He points to Ghana, Zambia and Zimbabwe rejecting or walking away from US health agreements under the America First Global Health Strategy as evidence that recipient governments now have that leverage and are willing to use it.
The Development Alliance. Liberia's attempt, around 2014 and 2015, to bring every donor and NGO into one room to map who was doing what, spot duplication and find the sectors nobody was covering. Moore's assessment: useful, but voluntary, not written into law, and not built around a single diagnostic. His conclusion is that such a framework should be put on a legal footing.
Five-year plans. Moore, who teaches in China each autumn, points to the discipline that fixed planning periods impose, and argues that legislation can do a similar job of holding a development strategy steady across changes of government.
Delivery units. Small teams set up to push complex projects through where the wider bureaucracy cannot. Moore ran one in the Liberian presidency and calls them islands of competence; he offers them as a way around weak implementation.
The European politics of aid. Moore's reason for thinking the window may close. Nativist parties are gaining ground across Europe, from the AfD to Reform UK to the PVV in the Netherlands, and an ageing population will pull more public money homeward. Countries that do not adjust, he warns, may find the external funding gone.

It's 1990. A young staff economist walks into a director's office at the World Bank and says the number he's about to publish is "crazy". The director tells him not to worry about it. 
The number was the dollar-a-day poverty line. Lant Pritchett, now of LSE, was that economist. More than three decades later, he's still worrying about it. In this week’s episode he argues that the dollar-a-day line warped how the world thinks about poverty, by setting the bar so low that we can count billions of deprived people as not poor.
In a new paper, co-authored with Martina Viarengo (Graduate Institute, Geneva), their fix isn't to scrap the low line. It's to add a high one as well. They propose a global upper-bound poverty line of $21.50 a day, ten times the extreme-poverty standard, derived from four separate measures of material wellbeing.
Above it, you're no longer poor by any reasonable global standard. Below it, you're poor in a sense worth measuring. By that standard, 99% of Pakistan is poor, and almost no one in Denmark is. Should that affect how we think about anti-poverty policy? 
The research behind this episode:
Pritchett, Lant, and Martina Viarengo. Forthcoming. "Raising the Bar: An Inclusive Global Poverty Line." Journal of Development Economics. Available now as a working paper.
To cite this episode:
Phillips, Tim, and Lant Pritchett. 2026. "What the $1-a-day global poverty line gets wrong." VoxDev Talks (podcast). 

Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.

About the guest

Lant Pritchett is a development economist and Visiting Professor at the School of Public Policy at the London School of Economics. He worked at the World Bank from 1988 to 2007 and taught at the Harvard Kennedy School for nearly two decades. His work spans economic growth, state capability, education systems, and labour mobility.
The paper is co-authored with Martina Viarengo, Professor of International Economics at the Geneva Graduate Institute. Her research spans public policy, labour markets, comparative education, and international migration.
Research cited in this episode

The dollar-a-day poverty line. Created for the World Bank's 1990 World Development Report on poverty and based on the observation that national poverty lines in the poorest countries clustered at a low floor (Ravallion, Datt and van de Walle 1991). Updated for inflation, it now sits at P$2.15 a day in 2017 purchasing power parity. It was only ever meant to mark the lowest a global poverty line could plausibly be, not the line.
The focus axiom. A standard property of poverty measures, originating with Amartya Sen (1976), under which changes in the income of anyone above the poverty line do not register in the measure. Pritchett's objection is that this assigns mathematically zero weight to the near-poor; a household just above the line counts the same as a Danish millionaire, namely zero. He calls it an economic bug that became a political feature, because it takes global redistribution off the table.
Gresham's law applied to poverty. Pritchett's framing for how the simple headcount displaced richer, distribution-sensitive approaches; bad economics drove out better economics because it was easier to understand. He notes the World Bank of the 1970s was preoccupied with distribution, citing Hollis Chenery and Montek Ahluwalia's Redistribution with Growth (1974), so the idea that economists ignored distribution until poverty measurement arrived is a myth.
The two criteria for an upper bound. The proposed line rests on two ideas drawn from the tension between the focus axiom and standard welfare economics. One, material wellbeing achievement; the line sits where a household reaches a standard of living a rich-country citizen would recognise as adequate. Two, near enough satiation; the line sits where the extra wellbeing from another dollar has fallen so low that treating further gains as zero does little violence to reality. At twenty-one and a half dollars the marginal utility of income is roughly three percent of its value at the dollar-a-day line; at the World Bank's current high line of P$6.85 it is still around thirty percent.
Four measures of wellbeing. The number is triangulated across an iso-elastic utility function, food shares in consumption (Engel's Law), a household index of six basic conditions drawn from Demographic and Health Survey data, and a cross-national index of basics. The estimates cluster between twenty and forty dollars a day; twenty-one and a half was chosen because it is exactly ten times the dollar-a-day line, a focal point in the same way one dollar was.
The six minimal conditions of prosperity. Electricity, improved sanitation, safe water, primary schooling completed by older children, no child dying under five, and no young child malnourished. The test Pritchett applies is whether it would be absurd to call a household prosperous while it lacks one of them.
The rich of the poor and the poor of the rich. The tenth percentile in Denmark has higher consumption than the ninetieth percentile in Pakistan or Indonesia. This is why any global line that produces meaningful poverty in rich countries implies poverty rates near one hundred percent across most of the developing world; a point Dani Rodrik (2007) showed is widely misunderstood.
The prosperity gap. A distribution-sensitive welfare measure adopted by the World Bank (Kraay et al. 2025) that weights the whole income distribution rather than counting everyone above a threshold as zero. Pritchett offers it, alongside poverty-gap and squared-poverty-gap measures at a higher line, as the practical route to acting on a global upper bound without reducing everything to a single headcount.
More VoxDev Talks episodes

Rethinking evidence and refocusing on growth in development economics, Lant Pritchett on what the problem might be if we rely exclusively on rigorous evidence in development economics as a guide for policy.
Rethinking how we measure extreme poverty, Charles Kenny asks: is it time for a new measure of extreme poverty?

Every civil service reform plan opens with the same list of complaints: poor performance, low motivation, weak accountability. Across six African countries and three decades, governments launched 131 separate reform efforts; not one fully achieved what it set out to do.
Martin Williams spent more than a decade working alongside Ghana's civil service before writing a book called Reform as Process that analyses the lessons from his experience, and the rest of the 131 reforms. For example, 34 programmes across six countries tried to link civil service pay to performance; none delivered. One lesson is that formal rules and accountability systems cannot govern what matters in a civil service: innovation, adaptation, co-ordination, the willingness to act on the spirit of a rule rather than its letter. Meaningful reforms often require no money at all. They require changing expectations from inside, starting small and building credibility, decentralising the leadership of change, and treating new formal rules as a last resort rather than a first step.
The book behind this episode:
Williams, Martin J. 2026. Reform as Process: Implementing Change in Public Bureaucracies. New York: Columbia University Press. Open-access PDF available at uplopen.com.
To cite this episode:
Phillips, Tim, and Martin J. Williams. 2026. "Why civil service reform fails (and what actually works)." VoxDev Talks (podcast).

Assign this as extra listening. The citation above is formatted and ready for a reading list or VLE.

About the guest

Martin J. Williams is Associate Professor of Organizational Studies and Associate Professor (by courtesy) of Political Science and Public Policy at the University of Michigan, and Associate Faculty at the Blavatnik School of Government, University of Oxford. His research spans the politics and management of policy implementation, public service delivery, and bureaucratic reform, with a sustained focus on sub-Saharan Africa. He previously worked as an economist in Ghana's Ministry of Trade and Industry as an Overseas Development Institute Fellow, and as a Senior Researcher at the Economic Policy Research Institute in Cape Town. Reform as Process has been shortlisted for the Douglass North Award for best book in institutional and organizational economics.
Research cited in this episode

Non-verifiable tasks. In organizational economics, a verifiable action is one where a third party (an auditor, a judge, an administrative tribunal) can determine objectively whether it was performed correctly. Non-verifiable tasks are those where no such determination can be made; they include innovation, adaptation, co-ordination across teams, and acting on the spirit of a rule rather than its letter. Williams draws on this framework, which originates in contract theory, to explain why formal accountability systems consistently fall short: they can only govern verifiable outputs, leaving the full range of non-verifiable tasks unaddressed and, in many cases, actively crowded out.
Performance-linked incentive systems. Williams's dataset covers 34 separate reform efforts across Ghana, Kenya, Nigeria, Senegal, South Africa, and Zambia that attempted to tie civil service pay or progression to measured performance. Not one delivered sustained differentiated incentives on an ongoing basis; only two achieved even partial delivery of rewards, and none delivered sanctions based on measured performance. Williams argues this is not isolated implementation failure but reflects a structural incompatibility between formalised performance metrics and the non-verifiable nature of much civil service work. Managers respond rationally: they set soft targets, award uniform scores, and the process becomes a tick-box exercise.
Projectization of reform. Williams uses this term to describe the dominant approach: treating change as a time-bound, discrete intervention with its own budget, acronym, and implementing team, conceived separately from the organisation's core work. This approach systematically distorts reform goals towards formally measurable outputs (new policies, new laws) rather than sustained behavioural change, undermines credibility by signalling a predetermined end date, and reinforces the perception among civil servants that reform is a temporary performance before things return to normal.
Continuous improvement. Williams draws an analogy with physical fitness: achieving a target and then stopping does not sustain the gain. High-performing organisations, in the public and private sectors alike, treat improvement as an ongoing process embedded in daily work, not a periodic project handed to a specialist unit. Starting small is not an absence of ambition; it is how credibility is built and larger changes become possible. Williams argues civil service reform should be reconceived on these terms, with performance improvement treated as the job of everyone in the organisation.
Decentralised reform leadership. The dominant model of reform leadership, Williams argues, is a visionary leader driving a top-down plan. This model is counterproductive. It personalises reform in ways that guarantee reversal when the leader moves on, and it cannot reach the day-to-day interactions among the thousands of individuals and hundreds of teams that determine how a civil service actually works. A more effective model is catalysing rather than forcing: creating conditions in which teams can identify and solve their own problems, escalate issues, co-ordinate with each other, and act on ideas for improvement without fear of being ignored or penalised.
More VoxDev Talks episodes

How government analytics can improve public sector implementation, in which Daniel Rogger and Christian Schuster discuss their efforts to use the data that already exists in governments to better understand how they function.