The role of financial models in society, according to the author, is largely misunderstood. The author makes the provocative claim that models do not primarily seek to represent reality but rather to facilitate decision making in a context of extreme uncertainty. The supposed tension between models and judgment turns out to be largely resolved in practice. Although often academic in style, this book will be of particular interest to financial professionals who develop, evaluate, or otherwise rely on models in their daily work.
In the second chapter of her provocative new book Financial Models and Society, Ekaterina Svetlova, an associate professor in finance and accounting at the University of Leicester in the United Kingdom, recounts a striking tale of mountaineers lost in the Alps. In the story (originally attributed to the American organizational theorist Karl Weick, whose claims regarding the source of the anecdote are the subject of some controversy), one of the mountaineers eventually produces a map that leads the group to safety. What is noteworthy is that the map turns out to be a map of the Pyrenees, not the Alps. Nevertheless, it “enable[s] the group to become active and alert and, thus, to save itself.”
This anecdote provides a succinct introduction to Svetlova’s view of modeling. Her central claim is that the primary role of financial models is not to offer a representation of financial reality but rather to provide a tool for inciting action in the context of the profound uncertainty of financial markets. This interpretation, which Svetlova supports with extensive empirical study of practitioners, has far-ranging implications for the way academics, regulators, and investment professionals talk about and evaluate models. If not entirely convincing, Svetlova’s arguments nonetheless raise critical questions for market participants who rely on models in their everyday practice.
Like the storied mountaineers, actors in the financial markets must make decisions in an environment of radical uncertainty. Instead of figuring out which way to turn in an unknown wilderness, however, investors must solve what Svetlova calls “the problem of investability.” As she puts it, “if uncertainty cannot be eliminated, we have to understand why investors are not paralyzed by it, why they continue to bring their money to the market—that is, how they solve the problem of investability under circumstances in which no rational person would have any good reason to invest.”
The point of a model, then, is not to establish with precision a security’s intrinsic value, for example, but rather to facilitate decisions in a world in which no amount of evidence can provide a dispositive guide to action. Svetlova uses the expression “action-like decision-making” to describe the way models enable market participants to act despite a dearth of clear signposts to guide their way. “Action-like decision-making,” she argues, “enables an individual to cope with genuine uncertainty and undecidability in an incomplete situation by engaging with the world.” The notion that financial models are a form of this action-like decision making is key to Svetlova’s understanding of models as prods to action rather than pictures of reality.
If a correct understanding of financial models relieves them of a burden of accuracy that is ultimately misplaced, then many of the criticisms we hear in the press lose their edge. Svetlova goes to great lengths to show, for example, that financial practitioners are rarely the slaves of algorithms and models to the extent that a casual observer might fear. Such fears, she writes, are “based on a principal misunderstanding of models’ roles in markets.” Rather, she argues that practitioners are generally well aware of the inadequacies of models and rarely follow them blindly without a qualitative check.
Svetlova explores such examples as the discounted cash flow model (DCF), foreign exchange models (forex), and modern portfolio theory. Relying primarily on interviews with practitioners, she shows that, contrary to common perceptions, a healthy skepticism regarding the reliability of models abounds. Far from being “model dupes,” practitioners constantly bring judgment to bear on model-driven analyses, particularly to fill in gaps that quantitative models inevitably leave open. Speaking of the DCF, Svetlova writes that “judgment reduces the distortion between the model and reality by taking most of the relevant factors and dynamics that were previously excluded from or not specified in the model and ‘bringing them back’ into the decision-making process.” She sees judgment having similar prominence in forex and portfolio theory. It turns out that the famous conflict between quantitative models and judgment is largely resolved in the day-to-day work of analysts and investment advisers.
Given Svetlova’s view that financial models, unlike scientific models, do not aspire to represent any truth but rather serve mainly to facilitate action in the face of uncertainty, the criteria for evaluating them should be revised. For example, the fact that a certain model seems formally inadequate from a scientific perspective should not be fatal for considering it useful in a financial context. As Svetlova puts it, “traders don’t look for mathematical truth, they look for a way to express their truth easily and comfortably.” Thus, efficiency and elegance for communicative purposes are as important, if not more important, than accuracy, particularly in the circumstances of radical uncertainty that characterize financial markets.
Svetlova devotes an entire chapter to the role such models as the CAPM play in “decision-selling.” Here she goes so far as to say that the supposed objectivity of models is mostly an act: “To motivate investing, non-existing ‘objective’ knowledge about the future must be feigned and performed as reliable and objective.” What matters, then, is not the model’s ability to furnish objective knowledge but its ability to appear to do so in the service of bringing participants into the markets. The context of model use is thus crucial for evaluation: We should be wary of condemning any one calculative approach without properly situating it within a larger setting of communicative and cultural practices.
Financial Models and Society is rich with original insights about the way investment practitioners use—and do not use—models. Many of the observations will ring true to those of us who work with financial models daily and are intimately familiar with their inevitable limitations. Svetlova’s basic claim that financial models are distinct from scientific models in that they do not seek to represent reality allows her to analyze model use from a variety of perspectives that are novel and unfailingly thought provoking.
And yet, even if we agree with many of her underlying premises, Svetlova seems to conclude too much from the facts of her argument. After all, the notion that a map of the Pyrenees might prod mountaineers to find their way out of an alpine wilderness surely does not imply that there is no such thing as a map of the Alps. Whether or not one agrees on the nature of financial models, however, Svetlova’s book will encourage greater and more mature reflection on a fraught subject, which can only be welcome.