The thesis addresses the question of how poverty can be overcome by education. The starting point is the following type of poverty trap: poverty forces parents to send their children to work to secure the survival of the household. Due to child labor, children do not attend school and do not form human capital. When these children are adult, the lack of education causes them being as poor as their parents have been. This produces a vicious circle that has to be broken. In an OLG growth model such a type of poverty trap is established as a locally stable steady state. Imperfect capital markets prevent the overcoming of poverty. Moreover, a positive externality of the formation of human capital on future generations is identified, which ought to be internalised. The thesis analyses two types of instruments that could generate economic growth via human capital accumulation: subsidies and land reforms. The paid subsidy or the transferred land must generate an income effect sufficiently big in size to enable the household to escape the poverty trap sustainably. The first part deals with subsidy policies. Subsidies contingent to school attendance are more cost-effective, i.e. require less resources per household to lift a household out of poverty, than the simple, uncontingent lump-sum subsidy, proposed in the literature. Conditional subsidies therefore allow a society to be educated and to be free of poverty in a shorter span of time. However, one has to distinguish between different types of conditional subsidies, since they differ with respect to cost-effectivity. A lot of aspects, like fighting corruption or inventments in the quality of schooling, enter into the optimal, multidimensional state expenditures plan. Successful tax-financed subsidy policies will only arise in democracy if constitutional rules prevent particular political failures. A major reason of failure is, among others, excessive taxation, so that households slip back into poverty. To heal these political failures a flexible majority rule and restricted rights of reelection for certain households or groups are proposed; a constitutional right to a tax exemption for protecting human capital can be justified. The second part of the thesis investigates land reforms as an alternative policy instrument. There is a dynamic connection between land reforms and the formation of human capital that has been ignored as well in the political as in the academical discussion: land reforms not only produce a more equitable distribution of land, they also, beyond this aspect, can generate a dynamic growth impulse which can leads a society out of backwardness and starts a transition towards an educated, developed industrial society. Even in a world without uncertainty, open access to land markets should temporarily not be allowed to land reform beneficiaries, as otherwise early land sales and migration can cause the failure of the land reform. A general equilibrium effect, caused by the land reform, may a wage effect in the rural labor market, so that also unsupported, landless and formerly poor households escape the poverty trap, even without receiving a plot of land. However, it is demonstrated that also the opposite effect may occur, in which case the situation of the agricultural workers get worse by a land reform. The focus in discussions on land reforms ought to be shifted: the aim of a land reform not only has to be a more equitable distribution of land, but it is especially important to target generating growth and education, as otherwise attaining a sustainable overcoming of poverty remains doubtful. Summarising, the thesis emphasises the necessity to analyse problems of poverty in a dynamic framework. Within a restricted static framework, one cannot identify and understand all the essential factors for a sustainable success of development efforts -- and the danger of deducing only temporarily successful policy implications remains.
