Back to nature
In the wake of the 2008 crisis, some economists have started to base their financial models on an approach that mirrors biological ecosystems, focusing on how companies interact and markets evolve. Vanessa Drucker reports from New York.
A model is like a little toy. Social and natural sciences use theories, based on assumptions, in an imperfect attempt to make sense of the world outside. While the history of science traces refinements in our understanding, reality may, over time, reveal how painfully inaccurate the models were.
Just as new theories shook the foundation of physics in the early 20th century, financial theory faces its own upheaval. Before the crisis struck in 2008, regulators tended to focus on individual financial firms and banks, rather than on addressing systemic risk across the entire global network. Traditionally, a micro-style financial theory looked individually at how to evaluate a company, a derivative, a swap or even a collateralised debt obligation (CDO), or how to optimise a portfolio or annuity. More recently, a movement has been gaining momentum to model how firms relate and interact dynamically.
To tackle the problem, economists consider an “evolutionary” paradigm, as an adjunct to, or replacement for, the rational expectations framework that has dominated the field for more than 60 years. Both physical and financial activities can be described by parabolic partial differential equations, a tool that could illustrate either heat diffusion or options pricing. (article continues below)
An emphasis on quantitative mathematical modelling increased over the decades, as the economics profession struggled with “physics envy”. Biologists, on the other hand, often rely on carefully documented empirical studies, like Charles Darwin’s On the Origin of Species, which employed no equations or mathematical formulas. Both biological and classical rational expectations models share common threads: evolution is a great optimiser, always making improvements with an invisible hand.
”Economies evolve, dictated by the behaviours of individuals, who learn from experience and positive feedback”
Some finance scholars and a few practitioners are moving away from a mechanical approach, derived from Newtonian physics, to a more organic one, based on biological observations. Attention is shifting, in recognition that economic systems, based on human interactions, can become more complex and unpredictable than planets or toasters. Financial theorists are turning to organic biological models, to help monitor the global economy and regulate key institutions that may be too big or interconnected to fail. Vivid metaphors serve to describe the phenomena. Let us examine some of the main elements, and the biological imagery that can be applied: complexity, modularity, feedback, innovation and market cycles.
The first feature of complexity is intuitively easy to grasp when applied to microcosms such as tropical rainforests or marine ecosystems. A threat to any one of the multiple species, flora or fauna, which inhabit those spheres can set off a series of repercussions that affect the stability of the entire structure. Thus, complexity can pose inherent dangers.
Lord (Robert) May, a former chief scientific adviser to the British government and professor of zoology at Oxford University, describes his own “accidental” breakthrough realisation in about 1970: “Before that time, ecology texts had taken up mystical notions and implicit assumptions about a balance of nature. Framed by G Evelyn Hutchinson, a Yale professor, these suggested that the richer the web of interconnections, the more stable the system.” May, however, then proved a surprising theorem. In ecosystems with multiple interactions, each species might have a mechanism to stabilise the population; yet once it interacted in a critical transition, if the number of links connecting the species and the strength of the interactions exceeded a certain value, the nature of the system would “tip” and become unstable.
May’s insight has profound implications for banking and other human industries. Consider by how many orders of magnitude complexity has grown since just 1900. At that time the earth’s population was estimated at 1.5 billion, compared with 6.9 billion in 2008, more than a four-fold increase. Couple that move with advances in technology, including the internet.
In a simple system, most threats are likely to come from outside. But a highly interconnected arrangement “can hatch its own black swan”, explains Ren Cheng, referring to the symbol for outlier events, popularised by Nassim Nicholas Taleb. Cheng, who is chief investment officer for life-cycle funds at Fidelity in Boston, has been focusing on abnormal distributions for about 15 years, and notes how any complex system has a tendency to go to extremes with so-called fat tails. To explain how such a system can self-destruct, He uses the metaphor of a brain malfunction. Normally, individual neurons in the brain will be constantly firing, to produce consciousness. But when neurons fire in large numbers they might bring about a seizure. “You don’t need to put a bullet in somebody’s head! The brain alone can do it.”
In living organisms, a large gene pool promotes genetic diversity, leading to biological fitness and survival. Homogeneity and inbreeding can result in fragility or even extinction. Today’s baroque financial system, operating in a tangle of complexity, exhibited its own symptoms of homogeneity. In principle, securitisation was slicing and dicing risk, but it turned out to be concentrating rather than diffusing it. At the same time, risk management strategies all grew to look more alike, and turned into “near-cloning”, as expressed by Andrew Haldane, executive director at the Bank of England, in an April 2009 speech in Amsterdam.
”Evo-devo [evolution and development] can rely on recombinations, as building blocks, to make bigger leaps forward”
In nature, as in banking, modularity may serve to limit damage. Fire breaks for forest fires are a clear example. Think of a net, composed of a series of linkages and nodes, which can be partitioned into modules. The American Fedwire interbank payment network, representing $1.2 trillion (£800 billion) of transactions a day, offered an image for the concept, as presented at a 2006 conference (see box, above). Just 66 banks, the nodes in this case, performed three-quarters of the daily value of transactions. It seems the linkages among financial firms may be multiplying, according to Haldane. Over the past 20 years, nodes have multiplied 14-fold, and links expanded six-fold.
The linkages are highly enmeshed. Remember the six degrees of separation - demonstrated by Stanley Milgram in 1967 in a chain letter - which allegedly link every human? For types of natural linkages, a marine biologist studies who is eating whom, and in what proportions. On a more delicate note, consider the patterns of pollinators and the plants they pollinate.
Epidemiology displays cases of modularity and natural self-regulation. “A healthy person’s immune system incorporates hundreds of thousands of antibodies, each designed to counteract families of bacteria and viruses,” says Cheng. If the immune system became homogenised, with each antibody performing the same task, it could only repel one type of pathogen. Likewise, if every financial institution follows the same path, then all become subject to the same risks.
In an epidemic, healthcare workers or prostitutes might constitute critical nodes. “In diseases like HIV, we often see a rapid spread and then a slowdown, due to heterogeneous transmission,” says Simon Levin, a professor of biology at Princeton. “If everything were equally connected, the conditions that started the spread would allow it to continue.” That firebreak approach was the Canadian government’s rationale, during the Sars virus outbreak in 2003, for shutting down transport in Toronto and taking measures to isolate the infection. By a similar token, when foot and mouth disease affects cattle, livestock must be culled on neighbouring farms.
In the banking system, uptick rules against shorting stocks or trading circuit breakers are manifestations of modularity. Currency unions, such as the euro, illustrate potential perils of sacrificing sovereign modularity for homogeneity. “But you don’t want to be so modular as to restrict the capital flows the banking system depends on,” Levin warns.
Feedback reactions, another property of economic activity, suggest a farming metaphor: a flock of sheep. (It is a myth that proverbial lemmings commit mass suicide, though.) “Economies evolve, dictated by the behaviours of individuals, who learn from experience and positive feedback,” explains Andrew Lo, professor of finance at the MIT Sloan School of Management. While people shy away from negative feedback, when none intrudes, they carry on with a course of action. “During years of prosperity, with no negative feedback about credit default swaps [CDS] they continued to trade them until it became too late, like an addictive drug,” says Lo.
”Those maths were hugely elaborate but not necessarily brilliant”
On one level, individuals are engaged in strategic behaviour - but that may collide with others’ behaviour. Changing incentives changes behaviour, as CDS contracts have demonstrated. Those who invest in CDS insurance positions that benefit from the defaults of third parties have ulterior incentives to see those third parties flounder.
Rational self-interest models describe what will maximise payoffs. In contrast, the evolutionary model says that people use imitation or social learning, by looking around to see who is doing well, as opposed to rational calculation. “Actually, both models often lead to the same result,” says David Rand, a researcher at Harvard’s Program for Evolutionary Dynamics. “While there are often lots of equilibria for any particular game or setting, an evolutionary approach helps with the equilibrium selection.”
Haldane, in his Amsterdam speech, pointed to examples of bankers’ reactions to the crisis, which precipitated further feedback loops. Banks sought to protect themselves against infection from other institutions by hoarding liquidity rather than on-lending, which soon caused more stress in the money markets. Unable to fund their asset portfolios, some hastened to offload assets and thereby created a vicious downward spiral of prices. Each of these actions, which made sense individually, added to the calamity of the whole.
An important principle of biology is natural selection, whereby the fittest types replace frailer specimens. Other important mechanisms, such as mutation or recombination, generate variety. “Evo-devo [evolution and develop- ment] can rely on recombinations, as building blocks, to make bigger leaps forward,” Levin explains. “So sometimes one organism absorbs another and builds it into the genome.”
For a man-made banking system, think of innovation as a bridge to survival and adaptation to market environments. Investment managers who can come up with a host of capabilities that work suitably in a range of environments are less likely to get wiped out as business conditions shift, Lo explains. He extends the metaphor to compare natural species with distinct groups of market participants, such as pension funds or hedge funds. A kind of natural selection may propel certain groups of investors to exit the market altogether, just as the dotcom fiasco prompted several technology managers to withdraw. Likewise, the popularity of certain hedge fund strategies, say long-short or convertible arbitrage, tends to wax and wane in response to the market climate.
Innovation can still be a curse as well as a blessing. Lo notes that “the rapidity of financial innovation outstripped the ability of the banking industry to cope with it, challenging what banks are and do by the pace at which it evolved.” Organisations must simultaneously innovate and select for the best. The same dilemma can be seen in most industries and firms. How much should any company spend on a particular solution, Levin asks, when it knows that the landscape is bound to keep changing? “How much should it invest in tried and true technology, or how much on the ’crazy stuff’ designed for a changing environment?”
Here is one more parallel, that of market cycles. In ecological systems, every species plays its own role, but all are interconnected. One of the most basic relationships is the cyclical interaction of predator/prey. When a rabbit population increases, foxes can eat more. That encourages them to breed more, so a younger generation of foxes will feast on even more rabbits - causing the rabbit population to contract. With less food available, the foxes begin to die off, so the next round of rabbits has a better chance to survive.
”Imagine assessing the robustness of the electricity grid with data on power stations but not on the power lines connecting them”
The financial system offers its own predator/prey scenarios. Consider two firms that wish to sell bonds into a market with a fixed level of available funds. One of the firms has superior credit, and is accordingly more highly rated. Thanks to its superior rating, it can raise more money. As a result, it becomes more leveraged, causing its rating to sink, relative to that of its competitor. Now the less leveraged company has gained an advantage. Cheng points to other cases, such as the situation described by technical analysis, when a market is overbought or oversold, in other words primed for an imminent reversal. And do not forget the old adage, that when a major magazine touts an investment on its cover, then watch out below.
Having drawn on a medley of relevant ecological metaphors, an obvious question follows: how can regulators or investors apply the inferences from a biological model? Haldane, in addressing the global financial breakdown, divides his policy prescriptions into three compartments: data/communications (or mapping), regulation and restructuring.
Risk needs to be measured, or mapped, before it can be mitigated. That means understanding the linkages, exposures and correlations across the system among various institutions. Haldane laments that, to date, samplings of bank “nodes” has been deficient. “Little was known about the activities of off-balance-sheet vehicles - SIVs [structured investment vehicles] and conduits - ahead of the crisis … imagine assessing the robustness of the electricity grid with data on power stations but not on the power lines connecting them.”
Lord May looks at the financial products themselves. “I would forbid trading in any instrument so complicated that it couldn’t have an intuitively accessible value put on it,” he proposes. By “intuition”, he is not referring to an innate feeling, but rather a “tentative insight into meaning, as opposed to manipulating mathematics.” With hindsight, it appears that some of the intellectual underpinnings of the arbitrage pricing theory that described financial theory were plain wrong, along with the maths that went into them. “Those maths were hugely elaborate but not necessarily brilliant,” says May. “They were computationally so com- plex, no one knew what they were doing.”
Nicholas Beale, who runs the London consultancy Sciteb, describes the distances between any two banks’ risk exposures as a mathematical vector, a simple calculation. Think of yacht racing, he advises. The sport is governed by rules designed to prevent collisions, with an optimal course to follow, given the wind and so on. “Over the years they have worked out sensible regulations for ensuring that all the yachts won’t be in the same place. They don’t tell them how to behave, but how to keep clear if overtaking another boat.”
”As we move from the philosophical to the practical, we need participation from the industry”
Yacht clubs, air traffic controllers and banking regulators need their manuals. Haldane’s second recommendation should “ensure appropriate control of the damaging network consequences of the failure of large, interconnected institutions”. A tension will inevitably result, between robustness and flexibility. The rainforest, as Lo notes, does not delineate zones for geckos or frogs, where “each is into each other’s business”. Unlike in the Amazon, artificial constraints and rigidities imposed on financial regulation risk causing further crises. Regulatory reformers can learn from the environment to develop more fluid systems. Economists have begun to do so already, with proposals for more counter-cyclical measures. Capital adequacy levels need not be static either, but could change with market conditions and animal spirits. Lo recalls how, in 1999, Brooksley Born, chair of America’s Commodities Futures Trading Commission, lobbied to restrict the size of derivatives and CDS markets. Heavyweights Robert Rubin, Alan Greenspan and Larry Summers squashed her warnings. Over the next five years the CDS market exploded, ballooning by a factor of 20. “Born was not wrong, just early,” says Lo. The lesson is a need for financial regulation attuned to managing magnitudes of exposure over time, rather than control by an on/off switch.
Restructuring, Haldane’s third pillar, involves reordering system, players and networks, through new arrangements like central counterparties and intra-system netting. “If you don’t have explicit measures built into the system to discourage herding you will get excessive concentration of banks with similar sets of risk measures,” says Beale. He suggests regulators may need to adjust incentives, like risk weighting, to encourage banks to diversify.
Global systemic risk still looms large. These new approaches remain embryonic. “We have the framework for thinking about it, but need to measure real world quantities that map into the variables in the model,” says Rand. “And as we move from the philosophical to the practical, we need participation from the industry.”
We should note that it is not easy to manage a complex ecosystem, like restoring a wetland. Some humility may be in order. Yet we are in a race against time to implement useful reforms before the next crisis, without the luxury to proceed “on an academic time-scale”, as Beale puts it.
We already have the computational power, so it is the will that is needed. Lo is optimistic that regulators will have access to developed tools over the next five or 10 years. It is urgent that the early steps be taken as soon as possible, while the impressions of last year’s crisis are still palpable. There may be only a narrow window ahead to effect real change before it is back to business as usual.
Have you looked at investment trusts more since RDR?