The relation between urbanisation and development is a new major paradigm in the global South at the beginning of the 21st century. This new equation raises obviously new and difficult questions. One has tried here to bring some answers on the basis of empirical studies conducted in the last 20 years. The paper reconsiders the practice of the author of analysing and measuring the urban polarity in East Germany Algeria and sub-Saharan Africa from a double point of view: investigating the measurement of the effects of the agglomerations in a context of globalisation on the one hand and, on the other hand, questioning the range of these measurements and the concepts which underlie them, and thus about the transferability in East, in North Africa and sub-Saharan Africa of the tools developed in Western Europe in regional science. The principal conclusion is that if the urban polarity is undeniably one of the main explanatory factors of the regional performances in industrialised countries, it is less clear elsewhere. In Algeria, this variable does not have any effect on regional performances, contrary to other indicators like agglomeration and metropolisation, and in sub-Saharan Africa, the city-effect is manifest but tends to decrease. These results underline the need for revisiting the conceptualisation of the developing effect of cities and their spatial transcription for the developing countries.