Answers are grounded in this paper and cite their sources.
This paper studies how to compute trading strategies that trade off expected returns, risk, and the trading costs generated by price impact in an efficient manner. For a general model where impact is nonlinear and decays gradually, we show that the performance of linear feedback policies can still be computed in closed form. These are optimal for linear impact and linear decay dynamics only. However, when the linear impact parameters are optimized for the nonlinear impact model at hand, then the performance of the resulting policy is competitive with state‐of‐the‐art machine‐learning methods. The gap in performance becomes significant only when the market is so illiquid that neither of the strategies is attractive net of trading costs.
No takes yet. Share an insight, caveat, or question.
Brokmann et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: