This paper explores the fiscal effects of aid in Ethiopia, using national data from 1960 to 2009, which is a longer series than most studies in this literature. This data includes the measure of aid that is flowing through the budget as measured by the recipient. We use the Cointegrated VAR methodology to model complex long run and short run dynamics amongst the following variables: aid grants, aid loans, tax revenue, non-tax revenue, and public expenditure. We also estimate an alternative model, where expenditure is disaggregated into capital and recurrent components (with aggregated domestic revenues to preserve degrees of freedom) in order to explore aid-spending relationships. The CVAR analysis is complemented by an in-depth qualitative understanding of the Ethiopian context, which ensures sound model specification and sensible interpretation of estimated results. Taking into account the major political regime changes, the data suggests three main conclusions regarding long run equilibrium relationships: government long-term spending plans are based on domestic sources, treating aid as an additional source of revenue; aid is positively associated with, and adjusts to, spending, with a particularly strong relation between capital expenditure and grants; and both grants and loans are positively related to tax revenue, both in the long and in the short-run.