Essays on Financial Stability

Abstract

This dissertation studies covered interest parity, global dollar funding conditions, and the predictability of future financial crises using recent global data on the foreign exchange derivatives market and a historical macrofinancial database.The first chapter examines how demand for FX swaps impacts persistent devia- tions from CIP. Using a novel dataset on FX swaps from the CLS Group, I analyze the factors driving CIP deviations across multiple interest rates and tenors. Specifically, I focus on non-bank demand for FX swaps and examine heterogeneity across industries and currency pairs. I estimate that on average, a 1% increase in FX swap volume results in a 2% widening of the CIP basis. As a result, a persistent dollar financing premium allows us to better understand why CIP does not hold in the data. The second chapter extends on the first and analyzes the price impact of FX order flow for swaps and outright forwards. I find evidence of the substitution channel: mar- ket participants who draw on swap lines reduce demand for dollars via the FX swap and forward market. In times of financial stress, such as the COVID-19 Pandemic, swap lines can be beneficial by providing cross-border liquidity. During quarter-ends, demand rises and price makers adjust prices to adjust for higher order flow. The third chapter studies the impacts balance sheet ratios on the probability of banking crises. I utilize the Jorda-Schularick-Taylor Macrohistory Database to predict crises over a longer sample period incorporating macroprudential target ratios. I supplement these results using machine learning models in order to model non-linear relationships and decompose crisis probabilities using the Shapley value approach. I find that the Loan-To-Deposit ratio is a robust predictor of financial crises and outperforms other variables proposed in the literature, such as overall credit growth, asset price growth, or the yield curve.