How to measure risk in the real world

Economists are forever running forensics on past financial crises to discover clues as to how the next one might occur. But in the latest work by former Securities and Exchange Commission regulator and Treasury department adviser Richard Bookstaber, the best illustration of how crises play out comes not from the market but from a logistical meltdown during a cultural event in Rome - the White Night, or Notte Bianca, of September 27 2003.

Shops, clubs, bars, restaurants, theatres and galleries had stayed open all night as part of an arts celebration. But around 2:30am, heavy winds knocked out a power line between Switzerland and Italy, which, in a system designed to automatically shift power to other lines during such a failure, eventually overloaded the grid that supplied Rome, shutting down the city. As Bookstaber writes in The End of Theory, the shift was to be expected, since it had been engineered into the system. What wasn't expected was the particular cascade effect - the first power failure disabled a separate communications system that connected the various grids, so that the blackout came in a very unexpected place. Two networks that seemed completely separate interacted in ways that engineers couldn't predict.
Of course, this will sound familiar to anyone who remembers how seemingly separate players - from banks to hedge funds to institutional investors - all interacted in unexpected ways to bring down the financial system in 2008. Add in the wild emotions of the individuals involved, who were as much driven by fear and panic as any stranded Roman reveller, and you get an even more complex stew of interactions.
Bookstaber - who ran risk management for Morgan Stanley, Salomon Brothers and the hedge fund Bridgewater before he decamped to Washington and later the University of California - believes those emotions and systemic interactions are still not properly accounted for in the neoclassical economic theory that underpins most risk models today. The supposedly "dynamic" models that allow for back-and-forth feedback between market players cannot capture the endless complexity of multidimensional networks. Likewise, behavioural economic attempts to model a less "rational" man are merely incremental expansions of the definition of self-interest: evolutionary, not revolutionary.

Tweaking the model

According to Bookstaber, it's time to stop tweaking a 150-year-old model that seems to be getting worse, not better, at predicting crises, and embrace something totally new. In the same way that the massive shifts of the Industrial Revolution took economics from the political theories of Adam Smith, David Ricardo and John Stuart Mill to mathematical neoclassicism, today's complex stew of technology, globalisation and financialisation requires a revolution in the economics profession itself.
His answer? Agent-based modelling, which is a jargon-y but apt description of an economic worldview that starts with people rather than maths. Using everything from the Heisenberg uncertainty principle to Milan Kundera's novel The Unbearable Lightness of Being as points of reference, he walks the reader through just how misplaced most of the assumptions in neoclassical economic models are when you start applying them to the real world. Forget about rationality, fixed preferences and market equilibrium. Like subatomic particles and fickle lovers, we are forever in motion, unknown not only to each other but to ourselves, touching and moving apart in endlessly complex ways.
None of this is news to the smartest market players - hedge fund billionaire George Soros's theory of "reflexivity", for example, posits that the emotions of players within the market can create feedback loops that change the market itself. Bookstaber, whose work was influenced by the Soros-funded Institute for New Economic Thinking, takes that idea and layers on others from chaos theory, network theory, physics, biology and even literature to challenge the conventional wisdom that we can model anything definitively in the markets.

New theories

The deductive science of maths, the very thing that modern economics is based on, is, in Bookstaber's mind, its weakness and fundamental limitation. "We are not robots with fixed, mechanistic responses to inputs," he writes. "We face a changing world that, in turn, changes the context with which we view the world, and that changes us, again all the more so during periods of crisis. The critical implication is that we cannot plug numbers into a model and solve for the future. We cannot know where we will end up until we take the journey. And we cannot retake that journey once completed." Bookstaber argues that we should give up rigid neoclassical models and move to agent-based models that will churn out not a number but an ever-evolving "narrative", something less precise but possibly more correct; think models of weather, or traffic patterns that show where pile-ups might occur (agent-based modelling is already used in urban planning).
It is too early to say whether agent-based modelling represents any sort of single unifying theory that could replace neoclassical economics (it is only just being trialled by academics, edgier financial risk managers, and forward-thinking regulators at places such as the Bank of England). But The End of Theory does hold some important lessons for financial markets today. First, it provides powerful ammunition for those who call for a modernised version of the Glass-Steagall regulation, which would separate commercial lending from investment banking not because such institutions are "too big to fail" but rather because their complexity can transmit too much financial contagion, across too many layers of the system, too quickly. Second, Bookstaber argues persuasively that neoclassical modelling of the last financial crisis has put too much focus on leverage; liquidity is more likely a cause of market troubles.
Finally, and perhaps most usefully, he challenges the economics profession itself, where too many experts still have way too much faith in their own mathematical infallibility. "The point isn't to crank out and act on a number," he writes. "It is to set up a model to see what light can be shed on a real-world problem, and to see if it can fit a larger, intuitive narrative about what is going on . . . is the plot line a reasonable one?" The analogy reveals the frustrated novelist that Bookstaber admits he is. It also shows him to be someone who adds a fair bit of humility into his models - which may be the best reason to read his book.

The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction, by Richard Bookstaber, Princeton University Press, RRP£24.95/$29.95, 240 pages

Copyright The Financial Times Limited 2017

ANDRA LÄSER