There’s Enough Math in Finance Already. What’s Missing is Imagination

There’s Enough Math in Finance Already. What’s Missing is Imagination.

For some of us, it was Spock. For others, a humiliating performance as a pilgrim in the kindergarten musical.  For me, it was William Blake’s relentless (and beautiful) attacks on Reason. But everyone at some point encounters – and many of us swallow – the dangerous notion that creativity and calculation are irreconcilable enemies.

This perspective lives at the very heart of our school curricula from first grade through graduate school, as our talents are identified and we, complicit in the scheme, label ourselves ‘artistic’ or ‘sporty’ or ‘scientific.’ No doubt there are real, epigenetic differences in the way people think and see the world, but in epigenesis lies the key: Nature gives us talents, but nurture determines how we use them, and how mono or multidimensional our minds become.

Like many quants – the mathematicians whose equations shape high-stakes decision making on Wall Street – Emanuel Derman arrived on Wall Street with little knowledge of economic theory. Unlike many of his colleagues, though, he had a background in theoretical physics, a field in which imagination and mathematics are happy bedfellows. From 1990-2000, Derman led Goldman Sachs’ Quantitative Strategies group, presiding over the rise of mathematical modeling as the engine driving financial betting on Wall Street.

The nearly insurmountable challenge of mathematical modeling in finance, says Derman, author of the forthcoming Models.Behaving.Badly., is that you are trying to predict the future based the behavior of highly erratic variables – i.e. people. To forecast human behavior with some reasonable degree of accuracy, mathematical models need highly imaginative designers with a profound grasp of human psychology – oracles who can foresee shifts in the global financial landscape and predict people’s reactions to them. Such designers are rare indeed, if they exist at all.

Still, like weather forecasts, the best mathematical models can help us to make better decisions, so long as we keep in mind that their predictive power depends upon the creative vision of the people who designed them, and that it is always vulnerable to refutation by developments in the real world. In other words, so long as we control the tools, not the other way around.

What's the Significance?

Faulty financial models, and Wall Street's overreliance on them, certainly played a role in the collapse of 1998, whose ongoing repercussions are manifest in the demonstrations going on right now in New York City and nationwide. And there are other massive, underlying issues – greed, unequal distribution of opportunity – that bear significant responsibility and intense scrutiny, of course. But math is not to blame – it is a tool, like science, that when guided by imagination and human understanding can expand the scope of our vision. It can help us to peer, however dimly, into the future.

Math’s not to blame, but a religious reverence for mathematical objectivity might be. We are susceptible to a modern-day form of idol worship whereby the sophistication of our technology and data-crunching power sometimes dazzles us into forgetting that these are human creations, designed by people, for people. And when they are applied to human problems, we cannot afford to treat them as perfect, alien entities whose wisdom exceeds our own.

As the early Internet pioneer Jaron Lanier points out in his digital-age cautionary manifesto You Are Not a Gadget, “Software expresses ideas about everything from the nature of a musical note to the nature of personhood. Software is also subject to an exceptionally rigid process of “lock-in.”* Therefore, ideas (in the present era, when human affairs are increasingly software-driven) have become more subject to lock-in than in previous eras.”

If digitized ideas are uniquely vulnerable to lock-in, we need to be exceedingly careful and creative in crafting them. For example, the government-driven “school reform movement” that has swept public education nationwide since 2001’s No Child Left Behind Act, seeks to transform learning based on data-collection from standardized tests, themselves a highly imperfect model of learning. In many cases, school systems rely on data from these tests to make hiring and instructional decisions, to a degree entirely unwarranted by the tests’ current level of sophistication. The clarion call of this movement is the demand for “objective” learning. That’s what the tests are supposedly designed to measure. But learners aren’t really objective. Nor is thinking, which is what, in the end, schools are supposed to teach.

Computers, databases, and other models of reality aren’t going anywhere – the technology will continue to become more sophisticated and its role in our lives will continue to deepen. Now, therefore, is the time for us to recognize where machines can meet our needs and where they cannot – and to design them for human use rather than seeking to adapt ourselves to their shortcomings.

*The process whereby an idea or a model becomes the basis for subsequent development, so that its flaws cannot be remedied without dismantling the entire system that is built upon it.