cd ..

Advanced Portfolio Management by Giuseppe Paleologo (2021)

1. For Whom? Why? And How?

• This book is for people at the intersection of (equity) fundamental analysis & quantitative research.

• Good trading ideas are useless without a portfolio that survives adversity.

• Readers who don’t have access to risk models can still get a lot out of this book.

  • Managers with effective heuristics can be successful without checking their factor risk decomposition every minute.

2. The Problem: From Ideas to Profit

• There is no master theory (yet) of portfolio management.

• Simulations & historical tests should be looked at for limit-testing, but not for confirmation.

• The problem (or steps) dealt in each chapter:

  • All valuations are relative, but to what? (ex. a stock vs its environment) (ch. 3,4,5)

  • Sizing position for your investment: is conviction the only factor? how do other assets in the portfolio affect this size? (ch. 6)

  • A strategy’s performance: stock selection, position sizing, and timing skills. How to quantify and improve upon them? (ch. 8)

  • Transaction costs when picking trade opportunities (sec. 8.2.1, 8.3)

  • Determine rules for risk management that allow various trading ideas while controlling risk (ch.7)

  • Survival is based on avoiding your capital’s critical limit (often via stop-loss policies) (ch.9)

  • How to decide leverage (for managers who are also fund owners) (ch. 10)

  • Analyze alternative data (sec. 8.4)

3. A Tour of Risk and Performance

• A stock index provides a benchmark to compare one’s investment, being a summary of the market (or specific sector).

  • Factor models make them rigorous and extend them to allow decomposition.

• A stock’s return can be written with the linear regression $r = \alpha + \beta \times m + \varepsilon$

  • $r$ is stock return, $\alpha$ is alpha (intercept), $\beta$ is beta (slope), $m$ is market return, $\varepsilon$ is noise (idiosyncratic or specific or residual return)

  • The term $\alpha + \varepsilon$ is only dependent on the company being examined.

  • For expected returns, the equation becomes (expected return) = (alpha) + (beta) × (expected market return)

• Estimates of $\alpha$ tend to be highly variable (and often the error is bigger than the estimate itself), unlike estimates of $\beta$ which are relatively stable.