We study the evolution of the public debt to GDP ratio during 40 fiscal consolidation episodes in 21 OECD countries in 1980-2008. We test within a multivariate regression framework seven hypotheses put forward in the literature on the success or failure of consolidation programmes. These hypotheses concern (i) the composition of the consolidation programme, (ii) its size and persistence, (iii) the gravity of the debt situation, (iv) the influence of the international macroeconomic environment, (v) the role of labour and product market institutions and institutional reform, (vi) the ideological orientation of the government, and (vii) the role of strict fiscal rules. We add a new hypothesis emphasizing the influence of public sector efficiency. We also improve on the literature methodologically by controlling for one-off budgetary measures. Consolidation programmes imply a stronger reduction of the public debt ratio when they mainly rely on spending cuts (except public investment), are large but of short duration, take place when growth in the international economy is high and interest rates are low, are accompanied by product market deregulation, are adopted by left-wing governments, are embedded in a regime of strict and wide fiscal rules, and are executed by highly efficient administrations. Public sector efficiency is important also for the composition hypothesis. Government wage bill cuts do not contribute to lower public debt ratios when public sector efficiency is high. On the hypothesis that consolidation is more likely to succeed in a situation of fiscal emergency, our evidence is mixed. Finally, we find no evidence that labour market deregulation contributes to a reduction of the public debt ratio during consolidation periods.