Reproduce “Residual Learning in Empirical Asset Pricing” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dexin Peng, Xiaoyu Wang
ResNet+ ranks U.S. stocks by next-month excess returns with residual blocks that preserve existing representations and add learned updates. In the 1987–2023 out-of-sample backtest, deep ResNet+ achieved an annualized Sharpe ratio of 2.07 for value-weighted monthly portfolios buying the top forecast decile and shorting the bottom, versus 0.89 for matched NN+; shallow ResNet+ scored 1.92. Findings apply to this stock panel and portfolio design.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.