1.4 Comparison with existing books
The financial literature on data modeling and portfolio design is extensive and diverse. This book aims to provide a unique perspective on these topics, and it is instructive to compare it with some of the existing textbooks:
Financial data modeling: Many excellent textbooks cover financial data modeling, such as (Campbell et al., 1997; Fabozzi, 2007; Fabozzi et al., 2010; Feng and Palomar, 2016; Lütkepohl, 2007; Meucci, 2005; Ruppert and Matteson, 2015; Tsay, 2010, 2013). In this book, Chapters 3 and 4 provide a succinct overview of i.i.d. models and models with temporal structure, respectively. Particular emphasis is placed on heavy-tailed models and estimators (as opposed to the more traditional methods based on the Gaussian assumption), stochastic volatility models (usually not receiving the deserved attention), and the use of state-space models with Kalman filtering as a unified approach with efficient algorithms.
Modern portfolio theoy: Traditional books that focus primarily on portfolio foundations and mean–variance portfolios include (Bacon, 2008; Cornuejols and Tütüncü, 2006; Fabozzi, 2007; Fabozzi et al., 2010; Grinold and Kahn, 2000; Meucci, 2005; Michaud and Michaud, 2008; Prigent, 2007). In this book, Chapters 6 and 7 cover this material with an optimization perspective, including utility-based portfolios, a recent derivation of the otherwise heuristic quintile portfolio as a robust solution, and particularly delving in detail into the nonconvex formulation of the maximum Sharpe ratio portfolio. It also provides a recently proposed universal algorithm for all these portfolios based on different trade-offs of the mean and variance.
Risk parity portfolios: Roncalli’s book (Roncalli, 2013b) provides a detailed mathematical treatment (see also (Feng and Palomar, 2016)), while Qian’s book (Qian, 2016) covers the fundamentals. In this book, Chapter 11 covers risk parity portfolios from an optimization perspective, progressively covering the naive solution, the vanilla convex formulations, and the more practical and general nonconvex formulations, with emphasis on the numerical algorithms.
Backtesting: López de Prado’s book (López de Prado, 2018a) covers backtesting and its dangers in great detail from the perspective of machine learning, while Pardo’s book (Pardo, 2008) focuses on the walk-forward backtest. In this book, Chapter 8 explores the many dangers of backtesting and the different forms of executing backtesting based on market data, as well as synthetic data, with abundant figures.
Index tracking: The topic of index tracking is treated in detail in (Benidis et al., 2018a; Prigent, 2007), with shorter treatments in (Cornuejols and Tütüncü, 2006; Feng and Palomar, 2016). In this book, Chapter 13 provides a concise yet broad state-of-the-art exposure, offering new formulations and a cutting-edge algorithm that automatically selects the right level of sparsity.
Robust portfolios: Robust optimization is widely explored within the context of portfolio design, with standard references including (Fabozzi, 2007) and (Cornuejols and Tütüncü, 2006) (see also (Feng and Palomar, 2016)). In this book, Chapter 14 gives a concise presentation of these techniques for obtaining robust portfolios with illustrative numerical experiments.
Pairs trading: The standard reference to this topic is (Vidyamurthy, 2004); see also (Feng and Palomar, 2016). In this book, Chapter 15 provides a full coverage of the basics and presents a more sophisticated use of Kalman filtering for better adaptability over time.
Machine learning in finance: Recent textbooks that give a broad account of the use of machine learning in financial systems include (López de Prado, 2018a) and (Dixon et al., 2020). In this book, Chapter 16 briefly discusses machine learning and deep learning techniques in the context of portfolio design.