The world is experiencing an increasing number of free-trade areas between developed and developing countries (think of NAFTA, the Free Trade Area of the Americas, APEC, the Euromediterranean Partnership, or the free trade agreements between the EU and Turkey, Mexico, South Africa or Mercosur and between the United States and Jordan, most of them still in the transitory period before full implementation). However, most of these schemes are being implemented out of a blind confidence in the positive effects of trade liberalisation regardless of the framework conditions, without much true empirical evidence on their impact on developing countries, let alone a serious reconsideration of the standard trade integration (custom unions) theory to adapt it to the special circumstances of developing countries. Due to the high social transformation potential of international trade, this could ultimately prove a dangerous economic engineering experiment for the development prospects of less developed countries, particularly the smaller ones. After a brief summary of the conventional trade integration theory, the paper refers to five aspects of free-trade areas between developed and developing countries which are not tackled by the this theory and which could question its assumptions predictions: the (high) import and (low) export elasticities which might prove wrong the positive impact on the current account balance of trade liberalisation; the possible working of economies of agglomeration favouring concentration of economic activities in more developed areas of a free-trade area; the eventual preference for industry in low-competitiveness countries which without a certain level of protection might lose any chance of industrialisation; the impact of those free-trade areas on inward foreign direct investment into developing countries, which might be actually negative; and the macroeconomic and political sustainability (far from granted) of those free trade areas. The conclusion is that there is a need to adapt conventional trade integration theory to the particular case of free-trade areas between developed and developing countries, and this cannot be done from a purely trade theory perspective, but must also take into account development theory and political economy considerations. In any case, a preliminary analysis seems to indicate that this new framework could favour "deep" integration schemes (where trade liberalisation is supplemented by certain legislative harmonisation, monetary stabilisation schemes and even a sizeable resource transfer from developed to developing countries to support the transition process and compensate the losers) instead of the more frequent "hollow" integration processes (consisting of mere trade barriers removal).