This paper studies the channels through which shocks from one country transmit in the global economy. Using local projections with instrumental variables (LP-IV), we document significant effects of US monetary shocks on output, interest rates, and trade flows in a large panel of countries. We find substantial heterogeneity in the foreign output responses depending on international trade. This association is due to the output share of international trade (extensive margin) and not the elasticity of trade (intensive margin), the exchange rate, or the foreign interest rate responses. To distinguish between the direct and indirect transmission of shocks, we further collect detailed data on bilateral trade linkages and combine the LP-IV approach with spatial econometric models. The indirect spillovers account for nearly half of the total response of foreign output, indicating substantial amplification engendered by the trade network. But the heterogeneity in the total effects is almost entirely due to the direct effects. Moreover, accounting for network effects within this framework leads to estimates of the mean output response up to one-third larger in magnitude than in the nonspatial specifications. Thus, studies that do not explicitly model cross-country linkages may present an incomplete view of international business cycles.

本文研究了来自一个国家的冲击在全球经济中传播的渠道。我们使用具有工具变量的局部预测 (LP-IV)，记录了美国货币冲击对大量国家的产出、利率和贸易流量的重大影响。我们发现取决于国际贸易的外国产出反应存在很大的异质性。这种关联是由于国际贸易的产出份额（外延边际）而不是贸易弹性（密集边际）、汇率或外国利率反应。为了区分冲击的直接和间接传递，我们进一步收集了双边贸易联系的详细数据，并将 LP-IV 方法与空间计量经济学模型相结合。间接溢出占外国产出总反应的近一半，表明贸易网络产生了实质性放大。但总效应的异质性几乎完全是由于直接效应。此外，在此框架内考虑网络效应导致对平均输出响应的估计值比非空间规范大三分之一。因此，没有明确模拟跨国联系的研究可能对国际商业周期提出不完整的看法。在此框架内考虑网络效应导致对平均输出响应的估计值比非空间规范大三分之一。因此，没有明确模拟跨国联系的研究可能对国际商业周期提出不完整的看法。在此框架内考虑网络效应导致对平均输出响应的估计值比非空间规范大三分之一。因此，没有明确模拟跨国联系的研究可能对国际商业周期提出不完整的看法。