This thesis studies the predictability of stock and commodity returns. It also examines the sources of return anomalies in financial markets. Chapter 1 introduces the main concepts and delivers an overview of the subsequent chapters. Chapter 2 begins with the analysis of stock return predictability around the globe. By studying more than 80 countries for a sample period of up to 144 years, we conduct the most comprehensive analysis of equity premium predictability that thus far exists. We find substantial evidence of in-sample and out-of-sample predictability for aggregate excess returns of countries in all regions. We detect predictability by examining price ratios, interest-related variables, and economic indicators, as well as forecast combination approaches. Investors in international markets can realize sizable Utility gains. Analyzing the cross-section of countries, we find that markets that are more efficient are generally more predictable, while more variability in business cycles does not lead to better predictability. Building on the analysis of return predictability based on historical measures in Chapter 2, Chapter 3 examines another strand in the literature and comprehensively analyzes the predictive power of several option-implied variables for monthly S&P 500 excess returns and realized variance. The correlation risk premium (CRP) and the variance risk premium (VRP) emerge as strong predictors of both excess returns and realized variance. This is true both in- and out-of-sample. A timing strategy based on the CRP leads to utility gains of more than 5.03 % p.a. Forecast combinations provide stable forecasts for both excess returns and realized variance, and add economic value. Inspired by the remarkable degree of predictability in stock markets, Chapter 4 extends the analysis to commodity spot markets. Using more than 140 years of data, we comprehensively analyze the predictive power of a broad set of macroeconomic variables for commodity prices and volatilities. We find some evidence for short-term predictability, while the predictability is much stronger in the long-term, at least in the case of predicting returns. The level of volatility and the degree of predictability are affected by the introduction of derivatives trading. A business cycle analysis shows that the degree of return predictability is independent of being in a recession or expansion. Volatility predictability is more pronounced in recessions. Motivated by the findings in commodity spot markets in Chapter 4, Chapter 5 translates the analysis to futures markets and studies the predictability of metal futures returns. Additionally, it identifies years of high predictability. Generally, we find a substantial degree of predictability both in- and out-of-sample. Gold returns seem to be best predictable out-of-sample. A timing strategy leads to utility gains of 2.18 % p.a. In particular, the Aruoba-Diebold-Scotti (ADS) business conditions index incorporates relevant information for metal returns, and strongly predicts gold returns. In the previous chapters, stock and commodity markets have been analyzed in isolation. Chapter 6 examines commodity futures markets to draw inferences about stock markets. The analysis is based on the insight that financial markets are populated by a large number of return anomalies. Our main objective is to provide evidence as to which of these are likely behaviorally-based and which have a risk-based explanation. To do so, we examine return anomalies in commodity futures markets. These markets provide an ideal ground for such research since (i) they are populated mostly by institutional investors rather than retail investors and (ii) there are only small limits to arbitrage. We find that downside beta, idiosyncratic volatility, and MAX are likely due to behavioral reasons, while jump risk, momentum, and volatility-of-volatility have a risk-based origin. Finally, Chapter 7 concludes and outlines possible future directions for research questions.