by Leah Guthrie
Simon A. Levin, a Professor of Biology at Princeton University, opened up the World Science Fair (WSF) salon on Predicting the Collapse of Complex Systems by relating a story of how ecologists predicted the 2008 financial collapse. He starts at a meeting hosted by the New York Federal Reserve on systemic or undiversifiable risk that involves the collapse of an entire market, as opposed to a specific industry (1). Levin was one of three ecologists in attendance, and they all were all struck by the evident parallels between ecological and financial systems. Their collective thoughts were published in February of 2008 in a Nature paper entitled Ecology for Bankers (Nature 451, 893-895).
In working with ecological data and models, Levin and his colleagues had come to appreciate the concept that robust networks have modularity. In this context, modularity refers to the extent that the nodes of a network can be broken down into smaller, discrete components that can function independently of one another (2). They argued that over-connected networks or ecosystems are more susceptible to collapse because they have less modularity, or localized independence. For example, a forest fire is less dangerous if there are firebreaks, or gaps in vegetation and other combustible material, that act as a way to compartmentalize the fire. When Levin and his colleagues looked at the network topology of financial systems, the over-connected structures of these networks presented a red flag. Firebreaks that could slow down the spread damage between nodes were missing, and in September of 2008, the collapse of the global investment banking firm Lehman Brothers triggered a financial crisis that extended across the financial spectrum. Industries and markets with little or no direct ties to the mechanics of the subprime mortgage crisis found themselves battered by the turbulence of a financial system that had suddenly turned on its head.
What the vertebrate immune system says about preparing for a crisis
Levin and his colleagues recognized financial and ecological systems as complex adaptive systems (CASs), containing numerous interacting parts that are individually able to influence the activities of other parts. These systems are mathematically “messy”, generating emergent properties as a result of nonlinear interactions at different spatial and temporal scales – “the whole is greater than the sum of its parts”. CASs include physiological systems, social systems and economies. While the components and interactions of such systems may vary, there are general principles that can be applied to their study. Levin, who has participated on panels dealing with financial and other types of crises, considers the CAS of the immune system as an instructive paradigm towards trying to understand the financial regulatory system. The immune system is designed to protect us from various threats, without knowing in advance what those threats might be. Key features of our immune system include the ability to recognize self from non-self, innate immunity and adaptive immunity. Our bodies have ways of generating early and generalized immune responses by recognizing commonly expressed molecular patterns shared by many pathogens, as well as more specific and stronger adaptive responses in reaction to very particular molecular signatures (or “epitopes”) of a given pathogenic species. With the framework of the immune system in mind, we can ask: how can we develop systems that can analyze financial markets globally, assess systemic risk and respond accordingly? Do regulations and laws like the Dodd-Frank Wall Street Reform and Consumer Protection Act encourage this goal?
As a response to the 2008 financial crisis, the Dodd-Frank law was passed in June of 2010 (4). The law effectively created new financial regulatory institutions and fortified the powers of others. Controversially, it identifies some financial institutions as systemically important and requires large financial institutions to annually submit a report describing how they could be liquidated without causing another crisis (4). The types of systems and tools that would improve the financial regulatory system, as well as the identification of the underlying assumptions of the Dodd-Frank law, warrant a more extensive review that will not be addressed here. The recognition of certain institutions as having the potential to cause wide-scale problems is an important first step, but developing both broad and narrowly–targeted ways to assess and act on systemic risk are important subsequent actions. It may be useful to view financial regulatory systems as similar to the immune system, in that they both deal with normally beneficial or commensal micro-environments that are occasionally infiltrated by pathogenic species. High-risk investments can have high reward when successful and, so an excess of pro-active regulations can have a stifling effect on economic growth. However, when investors make choices that are most driven by short-term greed and embrace highly fallible endeavors, there can be considerable consequences or emerging pathogenicity – in a financial context, let us define it as “a localized aggregation of financial rewards accrued at the expense of destabilizing the surrounding economic environment.” Regulatory analytics, oversight and laws that take into account the intricate ways in which financial decisions are made present an opportunity as to significantly monitor and influence the financial regulatory system. Much of the problem of financial pathogenicity comes down to the issue of transparency– similar to how a pathogen may try to mask a susceptible molecular epitope from discovery by the immune system, a financial player with an interest in appearing more economically robust than it really is may try to cover up signs of market vulnerability. The challenge lies in increasing the availability of information and communication about market “health”, so that all of the interested regulatory parties are able to see a developing crisis unfold in real-time and act accordingly.
Ecology at work
The similarities between complex systems, particularly between ecological and financial, extend beyond metaphors. Andrew Lo, a Professor of Finance at the MIT Sloan School of Management, used ecological principles to develop a framework for understanding financial interactions, termed the Adaptive Market Hypothesis. Lo explains investor and market behavior by interpreting the basic tenet of the Efficient Market Hypothesis, the idea that stock prices are based on commonly available information that makes it impossible to manipulate or underprice a stock, with an understanding of the role of irrational behavior in making financial decisions. Incorporating the concepts of competition, adaption and natural selection, he argues that if a risky strategy works well the first time, an investor is likely to use it again regardless of other information and continue to do so until the strategy fails. Key to his hypothesis is the idea of shifting the framework for thinking about the interactions within financial systems from the perspective of the physical sciences to the biological sciences, wherein particular nodes act with self-agency and interests instead of merely as the subject of external forces (5).
In ecological and financial systems, a shared key concept is the idea of public goods, common resources such as clean air and water that are owned by none but used by all. One of the hurdles that environmental movements often face is the difficultly of mobilizing people to care about clean air or other natural resources when its endangerment or scarcity does not affect them locally or eminently – this is known as “the tragedy of the commons”. In California, amidst the current water shortage of historically unprecedented impact, some are still adamant about watering their lawns despite the limiting resource that water has become for the whole state. This is a quandary in which people value their local access and use of a public good, in this case towards maintaining the appearance of their yard, but value less the consequences that their actions may have on the system as a whole. This is largely a problem of perception, where people are able to understand the impact they have on a local scale, but cannot extrapolate that understanding towards an appreciation of their systemic impact. This model of a local overconsumption of a systemic resource can also be used to describe a tumor. The cells of a tumor are producing public “goods” (nutrition and growth) for the benefit of the tumor and are acting cooperatively, but at a systemic level the tumor causes damage to the organism by using up common nutrients and releasing toxic products into the environment. Applying evolutionary principles, some scientists are looking for social cheaters that will stop producing for the tumor in what has been termed autologous cell defection (6).
Understanding a Nonlinear World
Mathematical approaches to predicting systemic collapse have developed over the past few decades. Catastrophe theory, during the 1960s-70s, dealt with how a static system could collapse if one parameter was changed. This was followed by Chaos theory in the 1970s, which involved oscillating or rhythmic systems such as the growth and contraction of economic markets. Subsequently, in the 1980s, Complexity theory involved systems with millions of interacting parts. Mathematical models based on ecological theory are already in use to improve financial systems. Predicting the behavior of complex systems requires indulging in messy and nonlinear math as well as an understanding the limits of what these models can predict and explain. Still, the parallels between ecological and financial systems as well as the application of ecological theory to a wide variety of systems present many opportunities for interdisciplinary collaboration and thought. The ability to extrapolate common principles of inter-modal complexity from particular systems, and apply those principles to problems in other systems, can prove to be a useful “thinking-outside-the-box” way to approach problems.
Leah Guthrie is a second year PhD student in the Department of Systems and Computational Biology at Albert Einstein College of Medicine. She is inspired by the intersections between science and technology, and looks forward to a career in problem solving.
- TheDodd–Frank Act: a cheat sheet by http://media.mofo.com/files/uploads/images/summarydoddfrankact.pdf
- Lo, A., 2004, “The Adaptive Markets Hypothesis: Market Eﬃciency from an Evolutionary Perspective”, Journal of Portfolio Management 30, 15–29. http://web.mit.edu/alo/www/Papers/JPM2004_Pub.pdf
- Archetti, M.2013b. Evolutionarily stable anti-cancer therapies by autologous cell defection. Evolution, Medicine and Public Health. doi: 1093/emph/eot014 (in press).