What Kinds of Educational Inputs Foster the Completion of Primary Education?: The Case of the Kagera Region of the United Republic of Tanzania

Abstract

  The United Republic of Tanzania (hereafter, Tanzania) is located in central East Africa, having a per capita GDP of 326 US dollars with a population of 40 million people. Poverty is a serious concern for Tanzania with income levels still among the lowest across sub-Saharan African countries. On the other hand, thanks largely to strong assistance from international communities, Tanzania has been successful in raising primary education enrollment rates in recent years. However, the primary completion rate is still very low and there is a substantial gap between enrollments and completions, which is one of the most highly prioritized policy issues in Tanzania.

  Much previous literature investigates the input and output relationship in schools by using the education-production function approach. Much of the research for developed countries shows that educational expenditures on school inputs are not systematically related to student performance (e.g., Hanushek, 1989). However, it is not yet clear whether or not this insight applies to developing countries. Cross-country evidence presented by Heyneman and Loxley (1982) revealed diverse influences of school inputs on student performance in low- and middle-income countries. Previous literature relies heavily on test scores to proxy the educational output in the education-production function, however the present research uses completion rates instead. Completion rates seem to capture more factors than test scores, such as the increase in school age children, the lack of resources to accommodate these extra children, and a rise in schooling costs paid by families, etc., that may more closely reflect the current situations of developing countries such as Tanzania.

  Using the 1994-2004 Kagera Health and Development Survey, fixed- and random-effects models are estimated controlling for unobserved heterogeneity across schools that is constant over time. According to the result of specification tests, the fixed effects would be the model of choice, implying that unobserved heterogeneity across schools is likely to account for the observed effect. The empirical results of the fixed effects model show the coefficients of blackboards and textbooks on Kiswahili, and textbooks on math; class size and student-teacher ratio are statistically significant at a 5% level with expected sign. The results indicate that school resources have a large impact on improving the completion rate in Tanzania.

  However, also very important for decision-makers of school budgets is not only the extent that certain school inputs can improve completion rates but also how much they cost. Decision makers seek the kinds of school inputs which are most cost-effective to meet the goals of policy objectives. Thus, the cost-effectiveness ratio comparing teacher input (teacher salary) and non-teacher inputs (blackboards and textbooks) are estimated. Results show that the cost-effectiveness indicators of non-teacher inputs are 1.8 to 13.4 higher than those for teacher input. This indicates that school resources are not allocated to maximize educational output, and that teachers and principals who have the power to allocate school budgets vastly overuse teacher resources relative to non-teacher resources. To overcome this problem, the government should work to change the incentives of decision makers who have power to implement school budgets. More specifically, decentralization, community involvement, and school choice would shift resource allocations more according to parents’ rather than teachers’ interests to attract new enrollees under competition with neighboring schools.