This thesis provides an empirical assessment of the impact of migrants' remittances on economic development of their origin country of origin. It shows that this impact, assumed in the literature as generally positive, is likely to be reduced due to the existence of a "dependency effect". This effect is particularly important in countries where remittances are an important share of GDP. This demonstration takes place in three stages: The first chapter provides a literature survey on the main determinants and impacts of remittances. Remittances are determined by individual behaviours, whose effects are observable at the macro level. The low volatility and the cyclical nature of remittances give them a stabilizing effect on receiving economies. This positive effect has nevertheless a setback: the remittances receiving countries can be subject to a dependency effect, resulting in a decrease in economic activity. The second chapter presents two results: First, it proposes the establishment of an empirical threshold of remittances' dependency. Thus, a country heavily dependent is characterized by a remittances to GDP ratio above the average of developing countries. The second result is the confirmation of a negative impact of this dependency according to the results of a panel data analysis on 32 countries. The results refute the hypothesis of a consistently positive impact of remittances on GDP growth and on gross capital formation (GFCF). The third chapter examines the impact and determinants of TFM in the case of a country heavily dependent: Tajikistan. This case study provides further analysis of the economics of remittances. Different macroeconomic determinants are tested to compare the role of Russian and Tajik economic activities on remittances. The results, robust to different estimation methods, confirm the effect of dependency. Moreover, remittances are more determined by Russian economic activity than by the Tajik one.
