Evaluation of Nigerian agricultural production data

Abstract

The absence of an agricultural census in Nigeria means that samples for surveys to estimate
agricultural production are obtained from a non-comprehensive, non-representative set of farms.
Therefore, aggregate production data quality is questioned. The methods employed herein provide a
new way to empirically evaluate the quality of agricultural production estimates. Two objective types
of data, namely the normalised difference vegetation index (NDVI), a satellite remote-sensing
measure of intertemporal vegetation changes, and prices, which reveal supply-use dynamics, are
used to analyse the degree to which agricultural production estimates reflect adjustments in growing/
market conditions. Broadly weak relationships were found between the production estimates and
these objective measures, but with variations in degree across states. In addition, these two objective
measures are more strongly related to each other than either is to production data. The results imply
that the inclusion of NDVI and prices in agricultural production estimation models would improve
the quality of the Nigerian production estimate.