Mörke, Mathis Rudolf WernerMathis Rudolf WernerMörke2023-04-132023-04-132023-02-20https://www.alexandria.unisg.ch/handle/20.500.14171/107736Rebalancing of leveraged ETFs and delta-hedging of equity options are two distinct and economically significant sources of order flow and liquidity demands. Liquidity Provision to Leveraged ETFs and Equity Options Rebalancing Flows finds that delta-hedging effects are persistent, those stemming from leveraged ETFs are decreasing significantly over time. These dynamics arise from different intermediation structures, generating heterogeneous levels of information asymmetry. While leveraged ETF providers generate perfectly predictable flows, option delta-hedgers have flexibility in deciding the strategic timing of their rebalancing, resulting in less predictable flows. Credit Variance Risk Premiums studies the pricing of variance risk in credit markets by employing a unique data set of credit swaptions. Returns of credit variance swaps are negative and economically large, irrespective of the credit rating class. Shorting credit variance swaps yields annualized Sharpe ratios well above their counterparts in other asset classes. The returns remain highly statistically significant when accounting for transaction costs and cannot be explained by established risk-factors and structural model variables. Commodity Tail Risks investigates the cross-section of tail risks in commodity markets. Left and right tail risks play an equally important role in commodity markets. For commodity producers, negative price jumps (left tail risk) might have devastating consequences. For commodity consumers though, e.g., companies that process commodities, right tail risks matter more. Both left and right tail risk implied by option markets are large. Moreover, both risks are priced in the cross-section of commodity futures returns. The option market has witnessed significant growth in recent years. Whereas risk factors in equity markets have been studied in detail, the same does not apply for option returns. Option Return Predictability with Machine Learning and Big Data finds that option returns are highly predictable. Allowing for nonlinearities significantly increases the out-of-sample performance of option and stock characteristics in predicting future option returns. Although option-based characteristics are the most important standalone predictors, stock-based measures offer substantial incremental predictive power. Most of the factor models assume a linear structure. There are, nonetheless, no obvious theoretical or intuitive justifications for this assumption. An Autoencoder Based Factor Model for Option Returns adopts a novel latent factor model, which allows for non-linearities, and applies it to index options. The model excels in explaining variation in risk compensation and out-of-sample trading strategies.enOptionsmarktOptionspreistheorieRohstoffmarktAktienmarktEDIS-5301RohstoffeVorhersagbarkeit von OptionsrenditenMachine LearningKünstliche IntelligenzOption Return PredictabilityTail RiskGamma SqueezeCommoditiesdoctoral thesis