I develop a theory of the optimal currency choice for invoicing goods for international trade in the presence of imperfect financial hedging of currency risk. I demonstrate that the classic irrelevance result–that the cost of financial hedging does not impact the choice of currency invoicing–rests on the assumption that sellers set prices ex-ante and fulfill any order size ex-post. I show that when quantities are also sticky, in the sense that the order quantity is in part pre-specified, then financial hedging affects the optimal currency of invoicing choice. My theory incorporates the cost of FX financial hedging into the classic theory of optimal currency choice, which relies on real hedging. I show that the optimal currency choice takes into account the relative ability of buyers and sellers to bear exchange rate risk. This financial hedging channel generates feedback between macroprudential policies, such as capital controls, and the optimal currency of invoicing. I highlight a “dollarization dilemma”: capital control policies that aim to reduce dollar borrowing are partially offset by an endogenous substitution into dollar-invoiced trade, which amplifies the local economy’s exposure to the dollar’s movements.

While household portfolios are well know to exhibit strong home currency bias, many firms borrow predominantly in foreign currencies from foreign lenders. Access to these foreign lenders might have large impacts on firm cost of borrowing and borrowing quantities. We propose a quantitative model of firm debt financing across debt markets which are segmented by currency. A key theoretical takeaway is that existing approaches to estimating cost of borrowing savings based on estimates of savings on the marginal unit of debt are only valid if firms face perfectly flat credit supply curves. We then measure key empirical quantities and estimate the parameters in our model using a novel, comprehensive, security level dataset of global corporate debt issuance which covers firms in 41 countries. Our fully estimated model is the first to be able to match key stylized facts about global debt markets and corporate debt issuance decisions. We find that credit supply curves are not perfectly flat and that there are meaningful inframarginal gains. We estimate many firms save up to 2% on their cost of borrowing by accessing foreign currency debt markets. We then perform model counterfactuals in which we change the fixed costs to entry. These counterfactuals highlight the roles of debt market heterogeneity in determining firm outcomes.