Improving the measurement and analysis of African agricultural productivity : promoting complementarities between micro and macro data

This paper identifies numerous instances where poor data have led to incorrect estimates of African land and labor productivity, and argues that better coordination of macro, meso, and micro data collection, reporting, and analysis efforts can lower costs and improve ability to monitor trends and to quantify determinants of agricultural productivity. The report makes the following key points. (1) Missing or poorly measured variables used in the numerator (output) or denominator (land and labor, for example) are biasing productivity ratios. (2) In most cases, these errors underestimate levels of agricultural productivity in Africa and distort trends. (3) Micro data are an important source of information for identifying the existence and magnitude of these errors in macro and meso data. (4) Information from micro data can improve estimates of productivity ratios when macro data are not available and too costly to collect. (5) Detailed micro data sets are the best source of information on the farm-level determinants of agricultural productivity; this information contributes to the development of productivity-enhancing policies and technologies. (6) Micro data play an important role in identifying the appropriate variables to monitor in macro and meso series. (7) Only consistently high-quality macro data in unbroken time series can provide adequate information about productivity trends and the contribution of policy and technological change to national agricultural productivity over time. The report includes recommendations for improving agricultural productivity data and analyses. (Author abstract, modified)