Essays in small open macroeconometrics

This thesis is composed by three papers connected to the study of the business cycle of small open economies under a double prospective: the consequences related to the transmission mechanism of global shocks and the employment of forecasting methods to date future economic developments at national level. Specifically, the first chapter analyses the role of international shocks on a small non-oil commodity-exporting economy, using Australia as representative case. The Australian natural resource sector is independent from oil and this is a single feature among developed resource-rich countries. Identifying global shocks through a combination of sign and bounds restrictions, the results reproduce the main stylized features commonly documented in this type of economies, e.g. concerning the dynamics of commodity currency and terms of trade. Particularly interesting is the response of trade-related sectors after the rise in commodity price; indeed no Ducth disease is found even if the reaction of nontradable sectors testifies its function is essential for the transmission of beneficial effects of a global shocks on real domestic variables. Furthermore it is shown that for some Australian variables the source behind the movement of resource price is not relevant while for other indicators the origin of the shock must be distinguished. The second chapter develops an early warning system for predicting incoming recessions for a group of small open economies. Combining models and data from previous contributions on early warning systems and business cycle analysis in small open economies, the results turn out to highlight the presence of some commonalities among this type of economies in terms of signaling for recession outbreaks. Out-of-sample performance also confirms the ability of the model to signal for recession outbreaks. Despite this, the baseline model is tested with different business cycle identification and augmented with global factors outside national policymakers control, suggesting the dependence of the results to specific specification and in particular to recession duration. In the third paper the predictability of monthly inflation for three small open economies is assessed. Through the Bayesian estimation of a set of linear models, it is worth claiming that including a measure of global inflation to a standard AR(1) model for national inflation is not beneficial either in terms of point and density forecasts. This is then faced considering the dynamic of regional inflation together with the global one in the benchmark specification. This strategy suggests the importance of sub-global developments in driving domestic indicators; the results are consistent at point and density level and for two different prior specifications.