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...

We propose a conditional model of asset returns in the presence of common factors and downside risk. Specifically, we generalize existing latent factor models in three ways: we show how to estimate the threshold which identifies the 'disappointment' event triggering the bad state of the world; we...

We develop a new test for threshold-type regime changes in the risk exposures in portfolios with a large number of financial assets whose returns exhibit an approximate factor structure. Unlike existing procedures to detect discrete shifts in factor models, our test is robust to regime-specific...

We study the problem of detecting structural instability of factor strength in asset pricing models for financial returns. We allow for strong and weaker factors, in which the sum of squared betas grows at a rate equal to and slower than the number of test assets, respectively: this growth rate...