Emerging stock markets are dominated with overconfident individual investors and in those markets short-sale constraints are frequently used to promote market stability. The unique market participants and regulations may influence the formation of stock prices and potentially lead to stock...

We propose a linearization of rule-based algorithms that reveals the most important interactions between characteristics and macroeconomic variables when explaining future stock returns. Our results suggest that the two types of predictors are intertwined, which implies that the relationships...

In March 2020, safe asset markets experienced surprising and unprecedented price crashes. We explain how strategic investor behavior can create such market fragility in a model with investors valuing safety, investors valuing liquidity, and constrained dealers. While safety investors and...

The price-bubble and crash formation process is theoretically investigated in a two-asset equilibrium model. Sufficient and necessary conditions are derived for the existence of average equilibrium price dynamics of different agent-based models, where agents are distinguished in terms of factor...

These lecture notes are part of a course on Big Data Asset Pricing. The topics covered in this Primer on Empirical Asset Pricing include the basic questions in empirical asset pricing, how to make factors, how to use factors in factor models, regressions (time series, cross-sectional,...

These lecture notes are part of a course on Big Data Asset Pricing. The topics covered include the factor zoo, replication in science and in finance, frequentist multiple testing adjustments, a Bayesian model to interpret factor evidence and preserve power in its multiple testing adjustment, and...

These lecture notes are part of a course on Big Data Asset Pricing. The topics covered include an overview of machine learning, supervised and unsupervised learning, and general ML methods such as bias-variance trade-off, hyper-parameter tuning, train-validation-test methods, and feature...

We use deep neural networks to estimate time-varying equity risk premia. The key innovations are the nonlinear and non-parametric generalisation of Fama-Macbeth regressions through partial derivatives of an arbitrary estimator function with respect to its input and the introduction of Jacobian...