This dissertation comprises three parts. Each part has been set up as a self-contained research project and deals with a specific aspect of foreign exchange option markets. Part I examines the accuracy of option-implied density forecasts in predicting future realizations of the spot rate. To produce density forecasts, a range of interpolation techniques is used, among them a novel method that dynamically updates the information content of currency options. We observe that the risk-neutral density generally provides a biased estimate of the physical return distribution. This finding is robust to different choices of information sets. In Part I, we further establish empirically a link between forecasting accuracy and surrogates for variance and jump risk. Part II examines the variance risk premiums in foreign exchange markets using a model-free approach. When realized variance is computed from intraday data with low frequency, variance risk premiums are significantly negative. In contrast, estimates based on high-frequency data provide a somewhat different picture. This latter finding is likely owed to microstructure effects. Further investigations suggest that variance risk premiums are essentially unexplained by classic risk factors or fear of jump risk. However, we find a significant relationship between the success of a variance swap strategy and the VIX, the TED spread and the shape of the implied volatility function. Overall, we conclude that foreign exchange markets feature a separately priced variance risk factor with time-varying risk premium. In Part III, the structure of currency returns is examined at the case of two unspanned information sets, namely the prices for vanilla and one-touch options. A variety of models with increasing complexity is calibrated to the vanilla market and subsequently applied to quotes for one-touch options. This approach allows to draw conclusions on the coherence between the two markets a.

This paper challenges the conventional view that foreign exchange risk premiums are small, not volatile, and unrelated to macroeconomic variables. For the Italian lira (1987-94), unconditional risk premiums—constructed using survey data to measure exchange rate expectations—are found to be sizable (relative to the dimension of the forward premium), highly volatile (relative to the variability of the forward bias), and predictable. Estimation of structural models of the risk premium suggests that anticipated fiscal contractions in Italy and lower uncertainty about the future path of fiscal policy are associated with a lower risk premium on lira-denominated assets.

First Published in 2001. Routledge is an imprint of Taylor & Francis, an informa company.