Adaptive Importance Sampling Estimation of an Open Economy Model With Fiscal Policy

We propose and estimate an open economy general equilibrium model that includes international trade between Canada and the United States. We estimate our model using a new adaptive methodology based on the mixture of student's t$$ t $$ by Importance Sampling weighted Expectation-Maximisation (MitISEM). Our findings for the Canadian economy indicate that an increase in productive spending initially worsens the trade balance, whereas unproductive spending leads to a more persistent improvement through the contraction in private demand. We also show that the domestic real exchange rate appreciates on impact after both productive and unproductive government spending shocks, although the response is sharper and less persistent following a productive expenditure shock. Our structural framework uncovers highly asymmetric spillovers from the United States to Canada, successfully resolving the international transmission puzzle highlighted by Justiniano and Preston [12]. Finally, we conduct counterfactual exercises that consider alternative trade-elasticity assumptions and a policy experiment involving a sizeable increase in US trade barriers.