No Good Deed Goes Unpunished: Political Incentives for Education Reform in Tanzania James Habyarimana, Ken Opalo, and Youdi Schipper September 13, 2018
Research Question Do competitive elections create incentives for improvements in the provision of public goods and services?
Do competitive elections help improve service delivery? Large literature elections and child mortality (Blaydes and Kayser, 2011); health (Fujiwara, 2015) Elections and agricultural policy (Bates and Block, 2013) Elections and road construction (Burgess, 2015; Harding, 2015) Elections and UPE (Stasavage, 2005; D Archy, 2013; Stasavage and Harding, 2014)
How do politicians signal credibility? Unbundling education provision Easily verifiable and attributable inputs (buildings, fees) Hard-to-measure outcomes (learning outcomes) Voters demand easily attributable and verifiable effort This is especially true in contexts of low state capacity
Politics and Education Reform in Tanzania We examine education reform through multiple electoral cycles In 2000 Tanzania had one of the lowest secondary school enrollment rates in the world (7.5%) Ahead of the 2005 election the incumbent party (CCM) promised to improve access (one school per ward) Result: (i) Number of O-level schools ballooned from 1628 (2005) to 4016 (2010); (ii) Performance plummeted, with pass rates decreasing from over 70% to under 40%
Findings There was an initial electoral boost associated with school construction of about 2 percentage points in 2005 (promise) But in 2010 there was a penalty of about 1.4 percentage points We argue that this was driven by the manner in which the policy was implemented (tax increase & learning outcomes)
Access vs Learning Outcomes
Changes in Secondary (O-Level) Pass Rates
Global Shift Towards Increasing Access MDGs and UPE Increased demand for secondary education
Changes in Secondary Enrollment (Bashir et al., 2018
The Political Context CMM s 2005 campaign promise explicitly tied school construction to electoral outcomes Wards are the lowest electoral units in Tanzania (simple plurality SMD) Elected Ward representatives sit in district councils Each district has a District Education Officer (DEO) Tanzania is a hegemonic party-state dominated by CCM (Av. 65.4% vote share in last 5 elections)
Rapid Increase in Number of Secondary Schools
Rapid Increase in Enrollment
An Extensive Construction Program
Drivers of Schools Construction 4000 3500 total number of wards 3000 2500 2000 1995 2000 2005 2010 2015 year
Drivers of Schools Construction No Schools in 2005 At Least One School in 2005 Ward Splits Ward Splits CCM Vote 05 CCM Vote 05 Pop (logged) Pop (logged) Rural Rural Dist. to Road Dist. to Road.1.05 0.05.1.15.5 0.5 1
How do you keep a promise with no cash? Secondary school construction delegated to communities Households contributed money, labor, and materials Government committed to provide roofing and operational assistance This provided strong incentives for politicians to signal effort by delivering on a Ward Secondary School
How do you keep a promise with no cash? They can do it [contribute] but only if they are forced. The ones who refused to attend the meetings are taken to the ward executive officer for penalties; their properties taken to be sold and to fund the works?? (Languille, 2014, p. 122).
Empirical Strategy We estimate whether exposure of local political units (wards) to school construction affects CCM s vote share Our unit of analysis is the ward (2000 boundaries) Data sources: Electoral Commission, MoE school registration
Changes in Ward Boundaries
Distribution of CMM Vote Share vs Poverty Rates
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)
Key Takeaways Elections create incentives for programmatic policies Promises of public goods provision can win votes But the devil is in the details (implementation does reveal information that voters may not like) Current Addition: Examining the electoral impacts of attempts to improve learning outcomes (BRN-Education)
Extension: Politics and Learning Outcomes Examining the electoral impacts of attempts to improve learning outcomes (BRN-Education)
The Politics of Improving Learning Outcomes
Summary Statistics
Orthogonality Tables
Balance (School Registration)
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)
School Construction and CMM Vote Share (Ward-Level)