The author studies the macroeconomic consequences of discretionary changes in the fiscal policy instruments for Canada. He adopts a semi-structural vector autoregression framework. Restrictions are based on institutional interactions between some policy and non-policy instruments that mimic a government's decision process. The author characterizes the actual economy's response to fiscal shocks, and proposes a theoretical model for a small open economy with nominal and real rigidities to test for the endogenous transmission mechanisms following shocks to government spending. He pursues a limited-information econometric strategy by comparing the theoretical impluse-response functions with the empirical ones, capturing the effects of a disturbance in government spending. Generally, the results of the model are very close to the observed reactions, especially for consumption, investment, exports, imports, and inflation; however, the model fails to predict the real exchange rate reaction.