Essays on Currency Risks and Returns

Thesis (Ph.D.)--University of Washington, 2019Chapter 1 proposes using foreign exchange rate currency options with different strike prices and maturities to capture both currency risks and expectations, for helping understand currency return dynamics. We show that currency returns, which are notoriously difficult to model empirically, are well-explained by the term structures of forward premia and options-based measures of FX expectations and risk. Although this finding is to be expected, expectations and risk have been largely ignored in empirical exchange-rate modeling. Using daily options data for six major currency pairs, we first show that currency options-implied standard deviation, skewness, and kurtosis consistently improve the explanatory power of quarterly currency returns than a standardized UIP regression. We then show that adding term structure information of options-implied moments further improves the explanatory power. Our results highlight the importance of expectations and risk in explaining currency returns and suggest that this information may be particularly useful during a crisis period. Chapter 2 studies the term structure of currency risk using FX options data, and finds it able to explain the cross-sectional variation of currency excess returns. With the tool of a new FX risk index, "FCX", I look into currency risk term structure and measure its shape by level and slope. I consistently find that for currencies paired by US dollars, the term structure of currency risk is flat at a low level prior to the 2008 crisis, upward-sloping after the crisis and peaks at a high level with a prominently negative slope during the crisis. This work is believed to be new in the currency research field. I then use this information to build trading strategies, earning a profit by longing currencies with the highest level or slope and shorting ones with the lowest level or slope. The profit by sorting slope is significantly high and robust to the 2008 crisis period, with a low correlation to the Carry Trade return, suggesting extra information in risk than the interest rate. Next, I extract global risk factors by level and slope to help understand the currency excess return, a long-lasting puzzle. The global risk factor by level substantially improves the cross-sectional explanatory power in currency excess returns compared to Lustig et al. (2011). Furthermore, I show that there is certain high risk corresponding to a high level and low slope, and high interest rate currency earns returns co-varying negatively to this risk, implying that it is a risky asset and thus requires a high risk premium, which explains the Carry Trade return well. Chapter 3 explores the possible macroeconomic connection in currency markets through the channel of FX risk term structure. There is a consensus in the literature that exchange rates are empirically “disconnected” from fundamentals, but a possible theoretical insight is that macroeconomic volatility shocks induce time-varying risks in the exchange rates. This chapter empirically investigates the connection between macroeconomic fundamentals and time-varying currency risks captured by the FX risk term structure, following the main findings of chapters 1 and 2. This chapter use both a small dataset of directly observable, country-specific key macroeconomic and international variables implied by exchange rate structural modeling and a small number of macroeconomic factors constructed from a large dataset of 126 U.S. macroeconomic series by principal component analysis. We perform a VAR analysis to examine impulse responses of FX risk term structure to the shocks of macroeconomic events and find that production variables can generate a relatively consistent and systematic impact pattern, which suggests potential macroeconomic connection. We also perform a direct single regression, regressing the 126 macroeconomic series of eight different groups on the FX risk term structure and apply the group LASSO technique for variable selection. Variables among both macroeconomic fundamentals and financial series are commonly selected, which suggests that financial markets’ co-movements also exist besides potential macroeconomic connection