Last week I attended Stanford University’s Quality of Governance conference, expertly organized by a rising star of the field, Saad Gulzar. I thought I’d follow in the footsteps of Dave Evans and others and summarize the findings of the papers presented. They provide a sketch of the frontier of research on state capacity.
In his plenary, Francis Fukuyama (presenting a paper co-authored with the Bank’s Francesca Recanatini) highlighted how corruption initiatives implemented by the World Bank since Wolfensohn’s ‘Cancer of Corruption’ speech – mainly focused on measurement - have basically failed. That much of corruption is essentially political is consistent with the findings of the World Development Report 2017 and Stuti Khemani’s Policy Research Report. Many of the papers presented below highlight the underlying politics of state capacity.
Accountability
A replication and extension of Bjorkman and Svensson’s famous ‘Power to the People’ (P2P) study across 376 health centers sees increases in the quality of care and patient satisfaction, but fails to replicate impacts on health outcomes and finds no impacts on community mobilization. #RCT #Uganda (Pia Raffler, Daniel Posner and Doug Parkerson)
35% of Ghanaian capital projects do not start (see great complementary video by Martin Williams on this). Central government audits reduce political interference in project implementation significantly, more than civil society monitoring, with corresponding impacts on reported corruption. #RCT #Ghana (Erik Wibbels)
Citizens
Decentralization shifts focus of politician-citizen interactions to local level. However, more centralized politicians act as alternative intermediaries for navigating the state. #RCT #India [paper] (Jennifer Bussell)
Citizens prefer public officials who pursue retribution against law-breakers, as it signals that the officials have capacity and moral character. #ConjointExperiments #China (Lily Tsai)
Allocation of Bureaucrats to Jobs in India
Affirmative action has no impact on the quality of district-level implementation of the Mahatma Gandhi National Rural Employment Guarantee Scheme. #IV #India [paper] (Rikhil Bhavnani and Alexander Lee)
Bureaucrats assigned to their home states are perceived to be more corrupt and less able to withstand illegitimate political pressure. #IV #India [paper] (Marianne Bertrand, Robin Burgess and Guo Xu)
Applying search and matching theory to public sector labor markets highlights the negative impacts of the recent changes in the assignment of senior #Indian civil servants. A particularly important contribution of this paper is that it provides a formal framework for understanding how equity concerns (frequently present in the civil service setting) impact the quality of assignments in the public sector labor market. #MatchingTheory [paper] (Ashutosh Thakur)
Political Communication
Information campaigns are a cost-effective way to increase voter leverage over politicians, but incumbent politicians respond by increasing vote buying, leading to no effect on incumbent vote share. #RCT #Philippines [paper] (Cesi Cruz, Philip Keefer and Julien Labonne)
Besides sharing plans for an exciting field experiment in Pakistan, Miriam Golden also previewed her work showing that incumbency rates rise with GDP per capita. #Pakistan #Global (Miriam Golden; field experiment joint with Saad Gulzar and Luke Sonnet)
Information
Whilst peer effects facilitate adoption of technologies with minimal externalities (like agricultural practices), it can be more difficult for innovations with significant positive externalities (like political accountability) to spread through a network. #RCT #Uganda [paper] (Romain Ferrali, Guy Grossman, Melina Platas and Jonathan Rodden)
Yelp for artificial insemination of cows! Providing feedback on the quality of government vets leads to a 37% higher insemination rate. ‘No bulls’ is a great title. #Pakistan [paper] (Ali Hasanain, Yasir Khanand Arman Rezaee)
Political Candidacy
A striking gender gap in political participation and representation persists in #India, despite several decades of targeted policy interventions. Sol’edad Prillaman presented a very cool study design to understand the determinants of this fact, based on a factorial, yes factorial, natural experiment! #FactorialNaturalExperiment (Sol ́edad Prillaman)
Offering a lawyer to help non-elites file papers for candidacy improves the likelihood that they run. #RCT #Pakistan [paper] (Saad Gulzar and Muhammad Yasir Khan)
Oversight and Delegation
Switching to democracy changes the way bureaucrats are promoted. Democratization amplified the positive effects of educational attainment on career advancement but simultaneously worsened the career prospects of female and religious minority civil servants. #EventStudy #Indonesia [paper] (Jan Pierskalla, Adam Lauretig, Andrew Rosenberg and Audrey Sacks)
Giving politicians formal power over bureaucracy through membership of legislative committees increases project completion rates by 18% in members’ constituencies but reduces the likelihood of being of satisfactory quality of those projects by 15%. #IV #Nigeria [paper] (Dan Rogger; disclaimer: this is me)
Providing procurement offices with autonomy reduces procurement prices and might increase quality. The net effect of performance incentives is zero. #RCT #Pakistan (Adnan Khan, Oriana Bandiera, Michael Best, and Andrea Prat)
Government Jobs
Paying public officials using biometrically-validated mobile money systems reduces the total wage bill by eradicating ghost workers. #RCT #Afghanistan (Michael Callen)
The reward systems for slave-soldiers in #Medieval Egypt undermined the long-run viability of the state by incentivizing them to transfer land assets out of state control. Looks like the rulers got their just deserts. #OLS (Lisa Blaydes)
Patronage
Political party turnover leads to greater replacement of headmasters and teachers, lowering 4thand 8th grade test scores by 0.05 to 0.08 standard deviations. (For comparison, the World Development Report 2018 [http://www.worldbank.org/en/publication/wdr2018] highlights, “across eight countries that have been studied, a 1.00 standard deviation increase in an index of [school] management capacity is associated with a 0.23–0.43 standard deviation increase in student outcomes.”) #RDD #Brazil [paper ] (Mitra Akhtari, Diana Moreira and Laura Trucco)
Road contracts connected to local politicians are more expensive and are less likely to generate an actual road. #RDD #India [paper] (Jonathan Lehne, Jacob Shapiro and Oliver Vanden Eynde)
Institutions
Self-assessment and certification of local state capacity and public goods provision have parallel impacts: real increases in local government performances for the already strong local governments, and manipulation of assessment scores for the weaker local governments. #DiD #Mexico (Soeren Henn, Horacio Larreguy and John Marshall)
Kate Vyborny presented initial results of a pilot experiment studying the impact of different types of information on donor decision-making. Exciting scale-up results coming soon. #RCT #Pakistan (Hamna Ahmed, Asha Gul, Simon Quinn and Kate Vyborny)
Leonard Wantchekon has developed a framework to explain how the intrinsic effect of a political institution (such as voting, deliberation or decentralized governance) can be identified by randomized experiments. #Theory #Everywhere [paper] (joint with Yves Atchade, Pierre Nguimkeu and Leonard Wantchekon)
Join the Conversation