An important feature of aid to developing countries is that it is given to the government. As a result aid has the potential to affect budgetary behaviour. Although the (albeit limited) aid-growth literature has addressed the effect of aid on policy, it has tended to neglect the effect of aid on the fiscal behaviour of governments. While fiscal response models have been developed to examine the effects of aid on fiscal aggregates - taxation, expenditure and borrowing - the underlying theory is ad hoc and empirical methods used are subject to severe limitations. This paper applies techniques developed in the "macroeconometrics" literature to estimate the dynamic structural relationship between aid and fiscal aggregates. Using vector autoregressive methods, an impulse response function is estimated to model the effect of aid on fiscal behaviour in Ghana. Results suggest that aid does not have a direct effect on the volume of government spending in Ghana but is treated as a substitute for domestic borrowing. Government spending does rise significantly following aid but this is principally due to an indirect effect arising from higher tax revenue associated with aid inflows. This, aid to Ghana has tended to be associated with reduced domestic borrowing and increased tax effort, combining to increase public spending.