Technical Report The effect of locally hired teachers on school outcomes (the Dose response function estimation evidence from Kenya)

Wakano, Ayako

16-15pp.1 - 30 , 2016-05 , Graduate School of Economics and Osaka School of International Public Policy (OSIPP) Osaka University
Do locally hired teachers benefit pupils’ school achievements more than governmental employed teachers? In Republic of Kenya (below referred as Kenya), there are two types of teachers in public primary schools. One is those employed by the government and the other is those hired by the local school community, named “PTA teacher”. Though locally hired teachers are in general less qualified in terms of educational background and paid substantially lower than that of governmentally employed teachers, past randomized experiment results show that the marginal product in terms of test score is positive and significant when pupils are taught by PTA teachers (Duflo et al. 2012, and Bold et al., 2013). By using a nationally representative rich data set, with the Generalized propensity score matching method, the present study examines the effect of PTA teacher ratio (ratio of PTA teachers out of total number of teachers) on education outcome. The question of this study is “if PTA teachers have superior performance, proved by the Randomized Controlled trial in Kenya, should higher PTA teacher ratio in one school bring better educational outputs?”. With the nationally representative dataset containing rich educational school inputs as well as individual pupils’ background and household information, this paper estimates the dose response function of school average outcomes. Provided that government teachers’ allocation and school selection by the parents can be an endogenous to pupils’ school outcomes, this paper utilizes the generalized propensity score method by Hirano and Imbens (2004) which enable us to estimate the function of the continuous treatment effect, PTA teacher ratio. The result consistently shows that the PTA teacher ratio affects school outcomes nonlinearly.

Number of accesses :  

Other information