Transaction Costs and the Stochastic Discount Factor
The DEMS Economics Seminar series is proud to host
Daniele Bianchi (Queen Mary University of London)
joint work with Teng Jiao and Hao Ma
ABSTRACT: Transaction costs determine which characteristic exposures are worth maintaining in equilibrium, yet standard stochastic discount factor (SDF) estimates often ignore them. We embed stock-specific trading costs into the no-arbitrage condition to identify a nonlinear SDF, estimated via adversarial neural networks across a large cross-section of U.S. equities. The resulting transaction-cost-aware pricing kernel endogenously reallocates away from high-turnover signals toward stable fundamentals. This improves cross-sectional pricing, mean-variance efficiency, and anomaly absorption, revealing a taxonomy of implementation-dependent versus cost-invariant risk premia. The results hold across architectures, cost definitions, and market conditions, and extend to a more conventional linear SDF specification.
The seminar will be in presence, Room: 4096 - Building U7