In the early 1980s, a great number of developing countries were facing severe balance of payments difficulties, and called upon the international financial institutions (IFIs) (the International Monetary Fund and the World Bank) for financial relief. In exchange for this relief, the IFIs devised a set of economic reforms, typically targeting fiscal, financial and monetary issues. This set of reforms, which were the conditions for IFI lending (‘conditionalities’), was later summarised as the ‘Washington Consensus’. These reform programmes during the 1980s and 1990s were, however, not associated with better economic performance. This mixed success led the IFIs to devise in the 1990s a different set of reforms, which this time targeted the functioning of the government and its ‘governance’. Several studies had insisted that the genuine problems of developing countries stemmed from the characteristics of governments described as, e.g., rent-seekers, plagued by corruption and whose policies only expressed the interests of particular groups. The paper analyses and compares these two sets of conditionalities, that of the ‘Washington Consensus’ and that of ‘good governance’ conditionalities, demonstrating their respective limitations. These limitations stem from: i) the concept of conditionality, the mechanism of exchanging finance for reform, per se; ii) the contents of the reforms summarised as the ‘Washington Consensus’ given the economic context of the countries under programme (typically an export structure based on primary commodities) and the weakness of the concept of ‘governance’ in view of these countries’ political economies; and iii) the intrinsic linkages between economic and political conditionalities, whose limitations thus retroact on each other, in particular regarding effectiveness and credibility. These limitations are examined in the light of the current theoretical debates on aid ineffectiveness and on the political economy of developing countries.< Réduire