On the accuracy of foreign trade statistics (1909-1935): Morgenstern revisited

Abstract

This work questions Morgenstern's pessimistic results on the reliability of aggregate international foreign trade statistics: His comparisons using pairs of countries can only test the misclassification of a country's trade flow. Aggregation, by contrast, eliminates this problem. Therefore, testing the total value of imports and exports with the sum of the same trade flows as registered by their partner countries' statistics, leads to more encouraging conclusions on the aggregate data. Our results strengthen considerably one's trust in the reliability of pre-World War II foreign trade statistics. Diversity in individual countries' accuracy indexes can be partially explained by differences in freight factors and also by minor differences in compilation. © 1991.