On the accuracy of economic observations : do Sub-Saharan trade statistics mean anything?

Matching exporters'freight on board (fob) trade statistics with the corresponding importers'cost, inventory and freight (cif) data, this paper studies the quality of official information on trade between 36 African countries. The author found the disparities in data so great as to make the official trade statistics almost useless for most empirical and policy studies. He concluded that: (a) statistics on matching exports and imports vary so much that the data cannot be used to assess the level of trade between African countries; (b) the data are probably equally useless for assessing the direction of trade since the countries that exporters report as the destination of trade often fail to report corresponding imports; (c) the data are inadequate for determining the composition of trade because countries report on different levels of detail; and (d) large, persistent, apparent differences in trends in intra-African trade may simply reflect different degrees of accuracy in country trade data. Finally, the fact that reported fob exports often exceed the corresponding reported imports suggests either that smuggling is widespread or that importers are intentionally underinvoicing because of high African tariffs or quotas.