Diagnosing the Credibility of National Agricultural Panel Surveys: Land Rental Markets and Allocative Efficiency in Sub-Saharan Africa

Abstract of the Paper

Land rental markets are widely viewed as an important mechanism for improving the allocation of land in smallholder agriculture, and nationally representative household panel surveys have become an increasingly important source of evidence on their performance. However, credible inference from such surveys depends on the reliability of reported land ownership and rental information. This paper develops a diagnostic framework that exploits repeated household observations to evaluate reporting quality, identify where conventional panel analyzes become vulnerable to reporting errors, and improve subsequent empirical assessment of land rental markets. Using balanced panel data from Ethiopia, Malawi, and Uganda, we document substantial instability in reported ownership holdings, persistent inconsistencies between reported tenant and landlord activity, and systematic underestimation of contemporaneously reported land ownership. We incorporate these diagnostics into benchmark and dynamic analyzes of land rental markets to examine their implications for allocative efficiency and entry barriers. The diagnostics substantially affect estimated farm-size distributions, landlessness assessments, and landlord-side rental statistics, while tenant-side analyzes prove considerably more robust. Although reporting instability is widespread, the principal relationships between land endowments, complementary productive assets, entry barriers, and tenant behavior remain remarkably stable after the identified reporting problems are explicitly taken into account. Our findings demonstrate that nationally representative household panel surveys can provide credible evidence on land rental markets when accompanied by systematic diagnostic analyzes that identify reporting limitations, guide empirical specification, and clarify where reliable inference is and is not possible.