2004年-世界发展银行全球_When_Can_School_Inputs_Improve____________Test_Scores__54页_1mb

Summary: When Can School Inputs Improve Test Scores?

Core Content

This paper explores the relationship between school inputs and cognitive achievement, with a focus on how household responses to these inputs affect educational outcomes. The authors argue that traditional production function approaches may not fully capture the impact of school inputs due to unaccounted household adjustments. They propose a household optimization model to better understand the dynamics between school and household inputs and their effects on learning outcomes.

Main Points

• Household Responses Matter: Households respond optimally to changes in school inputs, and these responses can alter the observed relationship between school inputs and cognitive achievement.
• Anticipated vs. Unanticipated Inputs: The paper distinguishes between the effects of anticipated and unanticipated school inputs. It argues that unanticipated inputs have a greater impact on cognitive achievement because households cannot adjust their own expenditures in response.
• Substitutability of Inputs: The substitutability between household and school inputs in the production function for cognitive achievement determines the impact of these inputs. If they are substitutes, unanticipated school inputs lead to greater learning improvements than anticipated ones.
• Empirical Context: The study uses data from Zambia to test these hypotheses, where the government provided fixed cash grants to schools, and alternative funding was unreliable.
• Methodology: The authors develop a theoretical framework based on inter-temporal optimization and durable goods theory, leading to an Euler equation that characterizes the optimal path of educational achievement.

Key Findings

• Substitution Elasticity: The elasticity of substitution between household educational expenditures and school cash grants ranges from -0.35 to -0.52, indicating strong substitutability.
• Impact of Anticipated vs. Unanticipated Grants: Anticipated grants have a small and insignificant effect on cognitive achievement, while unanticipated grants significantly improve learning outcomes.
• Robustness: These results are robust to potential mis-specification and omitted variable bias.
• Policy Implications: The findings suggest that the impact of school inputs on achievement is not solely determined by the production function but also by household responses. This has important consequences for educational policy and research methods.

Structure of the Paper

1. Introduction: Introduces the problem of understanding the relationship between school inputs and cognitive achievement in the presence of household responses.
2. Review of the Literature: Reviews three strands of literature:
   • Educational production functions
   • Household characteristics and educational outcomes
   • Effects of government subsidies on private outcomes
3. Theoretical Framework:
   • Presents a conceptual experiment on the impact of anticipated and unanticipated school inputs.
   • Defines the household optimization model with inter-temporal budget constraints and production functions.
   • Derives the Euler equation for the optimal path of cognitive achievement.
4. Empirical Specification:
   • Imposes conditions on preferences and production functions for estimation.
   • Derives the growth equation for test scores based on user-costs and expectational surprises.
5. Anticipated and Unanticipated Inputs:
   • Analyzes how the nature of the input (anticipated or unanticipated) affects the growth of cognitive achievement.
   • Highlights the counter-cyclical behavior of household spending relative to government spending when inputs are substitutes.
6. Data and Methodology:
   • Uses a unique data set from Zambia covering school inputs and student test scores.
   • Incorporates household data matched with schools using GIS tools to eliminate school choice effects.
7. Results:
   • Finds strong evidence that household educational expenditures and school grants are substitutes.
   • Demonstrates that unanticipated grants significantly boost cognitive achievement, while anticipated grants do not.
8. Policy Relevance:
   • Highlights the importance of considering household responses when evaluating school input effectiveness.
   • Suggests implications for both estimation techniques and educational policy design.

Conclusion

The paper provides a novel framework for understanding the relationship between school inputs and cognitive achievement by incorporating household responses into the model. It shows that unanticipated school inputs have a larger impact on learning outcomes, which has significant implications for educational research and policy. The methodology and data contribute to a more accurate estimation of the true effect of school inputs, moving beyond traditional production function approaches.