Designing State Capacity

Can better information delivery help low-income countries escape the vicious circle of meager resources, leaky systems and poor public services?

Rich nations are rich in information. The government receives dependable data on the income of its citizens and taxes them accordingly. Ideally, this capital, along with more quality information, allows the government to effectively target programs that improve the lives of poor citizens and help them break out of poverty. Poor countries stay poor in part because the government has insufficient information to enable it to gather resources or efficiently allocate what resources it has. In Pakistan, for example, collected taxes only represent around 9-10% of the GDP (compared to 20-30% for many European countries), and in Punjab, the province that contains over half of Pakistan’s population, property tax collection is roughly a fifth to tenth of the level of comparable countries. Insufficient taxes rise to the government, and when government allocates funds to programs, inefficiency and corruption cause great waste. A 2004 study in Uganda compared the amount of a special education block grant sent down from the central government with the amount actually received by schools, and determined that 87% had disappeared. These are dramatic, but not uncommon figures in the developing world; indeed, the research paints a picture of government as a wasteful machine with insufficient fuel going in and leaky pipes leading out. But a well-designed engine can go much farther on less fuel than a badly designed one. Can insightful policy design allow developing countries to increase their capacity at low cost, improving outcomes for the poor?

Policy Engagements

Incentivizing property tax inspectors to increase state revenues

We are collaborating with the government of Punjab, Pakistan on a randomized controlled trial involving over 480 tax inspectors to determine which program has the greatest impact: receiving increased pay regardless of performance, or one of a variety combinations of incentives based on amounts collected, accuracy, and tax-payer satisfaction.

Policy Partner: Department of Excise and Taxation, Punjab, Pakistan
: Asim Khwaja, Benjamin Olken (MIT) and Adnan Qadir (IGC/LSE)
Research Partners
: Center for Economic Research in Pakistan, National Bureau of Economic Research, Innovations for Poverty Action
: National Science Foundation, the International Initiative for Impact Evaluation (3ie), International Growth Centre

Targeting welfare payments to reach the ultra-poor

With the development ministry in Indonesia, we explored cost-effective screening methods that discourage welfare cheats without unduly burdening worthy recipients. Further, we asked whether the truly poor are better identified by the government or their neighbors and found that, based on poverty as defined by the poor, community-based targeting achieved better satisfaction.

Policy Partner: Ministry of National Development and Planning, Indonesia
: Rema Hanna and Benjamin Olken (MIT)
Research Partner
: J-PAL Southeast Asia
: World Bank, the International Initiative for Impact Evaluation (3ie), AUSAID


Making big data useable through data-visualization interfaces

Vast amounts of open data produced by India’s workfare program – probably one of the world’s largest development datasets – were going unused. To change this, we are collaborating with government partners to build an innovative data portal that puts information about the program’s effectiveness into a form policymakers and the public can use to identify successes and failures and weed out waste and corruption.

Policy Partner: Ministry of Rural Development, India
: Rohini Pande, Charity Moore and Eric Dodge


Evaluating the advantages and drawbacks of slum relocation programs

We studied the long-run effects of a lottery that gave 110 female slum dwellers the opportunity to own new houses in a more remote residential area of an Indian city. We found it had no effect on winners’ family income, and that 14 years later most had left the development. Our results demonstrate the importance accounting for behavioral elements and social networks when designing housing policy.

Policy Partner: Self-Employed Women's Association (SEWA), Ahmedabad, India
: Rohini Pande, Erica Field (Duke), and Sharon Barnhardt (IFMR)
Research Partner: Centre for Microfinance, IFMR, Chennai, India
: US Department of Labor

Ensuring child safety in areas with weak governance and transportation

We will design, implement, and evaluate a set of information-gathering protocols that use text message technology to provide a system of support to village-level child protection volunteers that is both sustainable and scalable in Sierra Leone's context of weak transportation infrastructure and state capacity.

Policy Partner: UNICEF, Sierra Leone Ministry of Social Welfare, Gender and Children's Affairs, and Child Fund Sierra Leone
Researchers: Ryan Sheely
Research Partner: Innovations for Poverty Action
Funding: Oak Foundation, UNICEF