This paper shows that a puzzling yet strong correlation exists between imports and exports. Cross-section and time-series analysis for the broadest possible country sample (159 countries over 1960-2006) attests to the robustness of this finding, even though causality is not explored in depth. After ruling out common factors driving simultaneously imports and exports, as well as systematic government interventions, we favor two weighty explanations. The first one is the presence of international financial constraints, in the spirit of the Feldstein-Horioka literature. However, we contend that this is an incomplete rationale, and thus we put forward a classic argument relating exports, imports, and economic growth in the short and the long-run. In the short run, a simple Mundell-Fleming logic can be invoked, whilst a longer-run perspective can be rooted in export-led growth models. Our results open room to rich and provocative policy implications as well as to the reinterpretation of commonly accepted notions in the analysis of open economies. The main policy outcome is that active, merchantilist strategies are in practice ineffectual to bring countries into large and persistent trade surpluses. In this light, trade, fiscal and exchange rate policies are deemed to be of little impact, letting alone their distortive side effects, in the medium and long term. At most they can aim to correct short-term imbalances. As a result, export-oriented policies should not be primarily judged in terms of growing trade balances but in terms of the indisputable benefits of increased export volumes. Imports, as an endogenous variable, would simply accompany exports. In the research arena, the paper’s results suggest that academic work should place some effort in reaching a better articulation between international trade, international finance, and saving and investment theory and empirics. These all are normally treated as separate fields, when in truth they are closely connected by the elemental fact that the trade balance is part of the current account balance, and the latter is by definition equal to saving minus investment. To a great extent, this is a consequence of the intellectual wall in macroeconomics between the Mundell-Fleming and the intertemporal approach to the balance of payments.
