'One hundred percent in my pocket' Transparency, debt and development

The dissertation is comprised of three main chapters that study the interplay between institutions, public debt and economic development. The first chapter introduces the dissertation.;The second chapter develops a model of transparency in government to study its impact on economic performance, the provision of education, and democratization. The chapter argues that information acquisition costs, which are linked to transparency of public finances, access to the media and overall levels of education are central to the problem of political corruption (diverting public funds for politicians' private benefit). Corruption, which reduces economic growth and voter welfare, is driven by imperfect information about the actions of politicians, and welfare increases when voters can acquire even limited information. Voter welfare is shown to be a decreasing function of information acquisition costs. Two extensions are developed: the first shows that politicians' incentives to invest in growth-enhancing policies that also reduce informational costs (such as education) may be weak, resulting in lower steady-state growth; the second proposes a mechanism through which transparency (or improved education) may impact the process of democratization and finds that the probability of a dictatorship transitioning to democracy increases with lower informational costs.;The third chapter argues that the concept of "odious debts" is closely linked to a principal-agent problem: debt arguably becomes "odious" when the agents contracting debt---the government---do not use it for the benefit of the principals---the population. A number of authors have called for the cancellation of such odious debts under different frameworks. A political agency model is used to analyze three such frameworks. Ex-ante "loan sanctions" (whereby creditors to sanctioned regimes are denied legal protections against default) are found dominate ex-post loan-by-loan audits. Welfare sometimes decreases in both cases when repressive governments also undertake productive investments; in some cases these governments cease investing but continue to use repression to remain in power. A "responsible lending" approach of verifying the use of loan proceeds may be superior to loan sanctions under some circumstances. An extension of the model suggests that an unintended consequence of odious debt frameworks may be to make it more attractive for creditors to bail out troubled odious regimes.;The fourth chapter investigates the channels through which institutions influence the probability that a country experiences a debt crisis. Two channels are postulated: a direct effect on the willingness to repay, and an indirect effect through institutions' impact on promoting sustained economic growth. I find evidence that improvements in institutions lead to sustained economic growth, which in turn reduces the probability of a debt crisis. The methodology used is based on identifying episodes of economic growth that are sustained over long periods. These "growth episodes" correlate well with long-term growth but are uncorrelated with growth in previous years. This allows a test of whether institutional measures (which are correlated with contemporaneous growth) contain information that can inform the path of future growth. The findings appear to be particularly strong for a sub-sample of poorer countries, but the results are robust to a number of alternate specifications, most importantly to changes to the parameters in the definition of a growth episode. The results obtained with the growth episode methodology are confirmed by standard panel data analysis. The chapter also considers whether high debt causes both debt distress and slow growth (which would lead to an overestimate of the observed impact of growth on debt distress), but does not find evidence for the latter hypothesis.