Cross-Sectional Targets and Neutralisation
"You are building a stock-selection model for a market-neutral book. What should the target be?" Raw future returns are the obvious answer, and the wrong one for this book. A complete answer removes from the target whatever the book will not hold, says how, and does the same to the features.
What a raw return contains
On any date, a stock's return has two parts: a move shared with the market or its sector, and a move relative to its peers. A market-neutral book holds equal amounts long and short, so the shared move cancels in its profit. It earns only the relative part.
A model trained on raw returns has to learn both parts. On days when the market rises, every stock's target rises with it. The model then spends capacity on the market's direction, which the book cannot trade. Worse, a feature that is correlated with a stock's beta can look predictive simply because the training period had more up days than down days. The fix is to train on the part of the return the book will earn.
The rest of this lesson is for subscribers
Unlock every lesson in Machine Learning for Quantitative Research, and every other premium course.
Subscribe to continueTest your knowledge
Keep reading Machine Learning for Quantitative Research
27 lessons in this course, and every other premium course, on one subscription.
- Every lesson in every course, with the worked examples and interactive simulators
- Graded questions on every lesson, with explanations for the wrong answers as well as the right one
- The trainers, timed assessments and brainteaser library that go with them