We extend the theory of Gabaix and Maggiori (2015a, 2015b) to study currency risk premia in a multi-country world with imperfect financial markets. Currency returns are connected to financiers’ limited commitment, captured by the complexity of their balance sheets in the trade imbalance network. Guided by the theory, we construct a Centrality Based Characteristic (CBC), based on the centrality of the imbalance network and variance-covariance of currency returns. Sorting currencies on CBC generates a high Sharpe ratio, and the resulting excess returns cannot be explained by standard currency factors and intermediary asset pricing factors, suggesting a novel source of currency predictability.

We estimate the impact of banks' cross-currency lending on exchange rates to shed light on the importance of flows as a major force affecting FX market outcomes. When non-US banks extend more loans in US dollars (USD) relative to US banks originating foreign currency-denominated loans, the USD appreciates significantly. When a foreign bank grants a cross-currency USD loan, it needs to obtain USD liquidity which puts pressure on funding markets and leads to an appreciation of USD. This effect -- which we estimate via a granular instrumental variable approach -- has greatly intensified since the global financial crisis and crucially depends on how banks fund the provision of cross-currency loans. In line with this mechanism, we show that cross-currency lending also affects the FX swap market (and deviations from covered interest parity), as well as other segments of the US short-term funding market.

We show that US monetary policy is transmitted internationally through the factor structure of exchange rates. Following an unexpected easing, investment funds sell safe and buy risky currencies. Global US banks, similarly, tilt their distribution of foreign loan origination towards currencies with greater systematic risk. The effects of monetary policy on flows of foreign exchange and international bank lending persist for several months. We argue that the exchange rate factor structure is a lens through which we can understand the international transmission of monetary policy.