Navigation – Plan du site

AccueilNuméros9-1Revue des livresComptes rendusRichard Bookstaber, The End of Th...

Revue des livres
Comptes rendus

Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction

Connel Fullenkamp
p. 189-192
Référence(s) :

Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction, Princeton: Princeton University Press, 2017, 240 pages, ISBN 978-069116901-9

Texte intégral

Afficher l’image
Crédits : Princeton University Press

1I begin this review with what I consider high praise: Richard Bookstaber’s book The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction reminds me in many ways of Ricardo Rebonato’s Plight of the Fortune Tellers: Why We Need to Manage Financial Risk Differently. Each book points out profound and irreparable shortcomings in one of the workhorse models of modern financial economics, and advocates for a different (though not new) approach to the task these models are currently employed in. In addition, each book liberally uses examples from the history of the profession as well as from literature to drive its points home. And in each case, the result is a highly readable yet sophisticated and convincing argument that is accessible to academics and practitioners, and even undergraduate students with the appropriate guidance.

2The model that Rebonato targets is Value at Risk, and he argues that it should be displaced by both Bayesian and commonsense approaches to measuring and managing financial risks. Bookstaber, on the other hand, attacks the entire class of intertemporally optimizing, rational representative agent models for their failure to anticipate financial crises as well as for their inability to capture any of the essential features of crises. These should be abandoned, according to Bookstaber—hence the book’s main title, The End of Theory—in favor of agent-based models that utilize heuristics based on observed human behavior and rely on computer simulation for their conclusions and predictions.

3The book is roughly evenly divided between an indictment of the representative agent model and an introduction to agent-based modeling for financial economics. Bookstaber’s main argument against the representative agent model at first sounds very much like arguments we’ve heard many times before—theorists want economics to be like physics, and their theories prize mathematical elegance over realism—but quickly and fortunately shifts its focus to what the author calls the “Four Horsemen of the Econopalypse.” These are four characteristics of the real world, and consequences of actual human behavior, that intertemporally optimizing, rational representative agent models are fundamentally incapable of capturing because the standard framework basically assumes them away. In addition, these same characteristics are also either essential features of financial crises, or of any model that has a reasonable hope of representing them.

4Bookstaber’s Four Horsemen include computational irreducibility, emergent phenomena, non-ergodicity, and radical uncertainty. They are distinct from each other, yet each one is a reason why the financial markets and macroeconomy experience outcomes that are beyond the ability of standard economic models to predict. Computationally irreducible systems have outcomes that cannot be summarized by equations of motion. Instead, they must be experienced, or in the case of models, simulated, period by period in order to find out what actually happens. Emergent phenomena are situations in which the system’s actions differ from the actions of the individuals in it. In other words, the individuals’ actions contribute to outcomes at the system level that deviate from the agents’ actions, and in some cases, countermand the agents’ actions or even bring harm to them, as with stampedes. Non-ergodic systems’ outcomes depend critically on history and context, so that every situation is different and can produce outcomes that differ significantly from previous situations that are similar but not the same. Finally, radical uncertainty (also known as Knightian uncertainty) means that not only are the probabilities of various events and disturbances unknown, but the event space itself is not fully known.

5The Four Horsemen are manifestations of an underlying phenomenon: complexity. Bookstaber’s treatment of complexity is interesting, but is not very well connected to the discussion of the Four Horsemen. In fact, he doesn’t actually present a definition of complexity, preferring instead to give examples from a continuum of complexity. One of the most important and relevant examples of complexity for financial markets is Soros’ (inaptly named) reflexivity, in which agents’ actions in the financial markets change the financial environment, which in turn alters their subsequent actions in an endless feedback cycle. Bookstaber also argues that some players in the financial markets even intentionally induce complexity, as combatants do in war, in order to gain an advantage over competitors. Although this discussion is quite useful to Bookstaber’s overall argument, the book would have benefited greatly by drawing more explicit and careful connections between complexity and the Four Horsemen.

6Given that standard economic models are built precisely to avoid both complexity and the Four Horsemen, it is impossible for them to be effective at predicting or representing a financial crisis. Thus, it is up to a different type of model—according to Bookstaber: the agent-based model—to take on this responsibility. The author describes the main components of agent-based models, and devotes several chapters to giving readers a taste of how agent-based models could be used to represent financial crises. Bookstaber emphasizes the importance of modeling the actual players in the financial markets, such as bank-dealers, hedge funds, and cash providers, and the rules or heuristics they use to make decisions. Just as important is to model as accurately as possible the environment within which these players interact—that is, the actual financial plumbing that connects them in multiple ways. Using several combinations of players and environment, the author presents examples of how agent-based models can show how credit shocks or asset shocks would propagate in unexpected ways through the system and lead to panic or widespread distress.

7In the last part of the book, Bookstaber turns from general agent-based models of financial crisis to a discussion of how agent-based models would have given insight into recent events such as the Crash of 1987 and the 2010 Flash Crash, as well as developments in 2007 and early 2008 that drove the Global Financial Crisis. The narrative is an engaging mixture of snapshots of possible agent-based models of these events and the author’s own interpretation of them—most of which he had a close-up view of, as a professional risk manager and later an adviser to the Financial Stability Oversight Council. The switch to more of a storytelling approach foreshadows Bookstaber’s concluding chapters, in which he argues that making theories should give way to understanding the markets’ story as it unfolds and evolves in real time. The author claims that this approach, supported by agent-based models, can give us both improved warning of financial storms and insight into how to prevent them or mitigate their fury.

8Although Bookstaber makes a compelling argument that agent-based models could be a very useful tool in finance and macroeconomics, and that economists should start exploring and developing this tool, I think it’s premature to declare the end of theory. Although Bookstaber is correct that theory can only take us so far—and in the case of financial crises, not very far at all—theory has still contributed much to our understanding of finance and macroeconomics. The problem occurs, as Bookstaber points out, when we take theory far more seriously than is merited by its limitations, and we insist that it is the best (or perhaps only) method for advancing our understanding of financial markets and the economy. In addition, theory is a rich source of the stories that Bookstaber believes are important to the narrative-based approach he advocates. As a long-time instructor of principles of economics courses, I am a strong believer in the power of simple theories to help people organize their thinking into coherent, satisfying and surprisingly insightful stories about what takes place in markets and the overall economy.

9To the extent that this book uses agent-based models to open economists’ minds to alternative methods for gaining insights into how markets really work—and fail—it is quite successful. And I am confident that agent-based models will become a workhorse of both finance and macroeconomics in the near future, given their advantages highlighted in this book. But Bookstaber overpromises a scientific revolution to end economic theory as a dominant paradigm. Rather, the emergence of agent-based models introduces another discussant into what has, up to now, been a dialog between theory and data (empirical analysis). This should make for more lively conversations, and I believe that each participant is capable of holding its own in them. Thus, although I understand the attraction of the book’s provocative title, I think a more accurate way to view The End of Theory is the beginning of a new and productive era for learning about financial and economic crises, thanks to the addition of a powerful tool to our kit.

Haut de page


Rebonato, Ricardo. 2007. Plight of the Fortune Tellers: Why We Need to Manage Financial Risk Differently. Princeton: Princeton University Press.

Haut de page

Pour citer cet article

Référence papier

Connel Fullenkamp, « Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction »Œconomia, 9-1 | 2019, 189-192.

Référence électronique

Connel Fullenkamp, « Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction »Œconomia [En ligne], 9-1 | 2019, mis en ligne le 01 mars 2019, consulté le 08 mars 2021. URL : ; DOI :

Haut de page


Connel Fullenkamp

Duke University,

Articles du même auteur

Haut de page

Droits d’auteur

Licence Creative Commons
Les contenus d’Œconomia sont mis à disposition selon les termes de la Licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Haut de page
  • Logo Association Œconomia
  • Logo CNRS
  • Logo DOAJ - Directory of Open Access Journals
  • OpenEdition Journals
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search