Conference on Emergent Risk

28 to 29 September 2012
216 Aaron Burr Hall, Princeton University

Miguel Angel Centeno and Alex Tham, Sociology (Princeton University)

Abstract

There is little doubt that globalization has produced an ever more intricate set of interactive relationships between individuals, organizations, and states. To an increasing degree, the business of the world depends on other parts doing their business well. Failures to produce, to loan, to purchase, or to regulate no longer have limited local effects, but can have global and potentially disastrous consequences. This project seeks to explore this danger through the concept of "emergent risk". It will both ground the concept theoretically and produce empirical examples of how it may be studied. Finally, it will consider what regulatory solutions may exist for this danger.

Scholarly Motivation

The global web of trade, finance, travel, and communication can be best understood as a complex adaptive system (Urry, 2003). As such, it has several salient properties. It includes a very large number of actors and multitude of relations between them. The global system is a huge and (in places) dense network of transactions (see http://etc.princeton.edu:8080/maptrade/ and http://qed.princeton.edu/index.php/MG for graphic illustrations). (Consider a small part of this network, air traffic: there are over 500,000 trans-Atlantic passenger flights yearly, transport-ing over 150 million people of practically every nationality, and using over 8 billion gallons of fuel.) Relations between these nodes and links are further complicated by their interactions (continuing with the air-travel example, my willingness to book a flight is related to the price of the ticket which is related to the price of oil). These interactions are in turn networked in a massive causal loop of feedback and adaptation. This project seeks to better understand the structure, nature, and challenges of this complex interactive system.

The number and complexity of transactions and interactions makes any kind of conventional description or analysis arguably impossible to compute or comprehend. These flows are self-organized in that no single authority or set of authorities directs the system. Rather it emerges out of autonomous and individualized decision of millions of agents receiving and sending a broad array of signals (e.g., there is no central office determining the network of flights on any particular day). This organization is hierarchical and self-similar. It is hierarchical in that the system consists of sub-systems of nodes and flows and each sub-system shares similar structural characteristics (e.g., the airline hub and spoke system is reproduced at various geo-graphical levels). The system and its components are adaptive to changes in its inputs and to environmental perturbations (e.g., airline traffic adapts to demand, fuel prices, and weather). The system is non-linear in that adaptation to some changes will be non-proportional to the relative size of the cause (e.g., a blown tire in Detroit may lead to the cancelation of numerous flights in Tokyo). Finally, the system resists any static equilibrium and must remain dynamic in order to exist (e.g., a fixed air schedule that did not respond to environmental change would soon become useless).

The most important characteristic of such systems for present purposes is that they generate emergent properties, "collective phenomena collaboratively created by individuals yet not reducible to explanations in terms of individuals" (Sawyer 2001). Their emergence may reflect both systemic causality and simply the limits to available information and human analytical ability. The bottom line is that emergent properties are inherently unpredictable using deterministic models.

Many of the global system’s emergent properties are benign: e.g., the global price convergence for goods and services makes international trade flows possible. The billions of daily transactions of the global web have created a meta-structure that has indisputably improved the lives of many. These are not the only emergent properties, however. This project is particularly concerned with analyzing a significant consequence of such systems: emergent risk.

Emergent risk is different from "emerging risks". The latter are dangers and challenges that arise from new technologies or interdependencies, which "have not yet occurred but are expected to grow greatly in significance" such as terrorist attacks or natural disasters (Richardson and Gerzon 2004). Emergent risk, on the other hand, is better understood by comparing it with "systemic risk". Whereas systemic risk is the threat that individual failures or accidents represent to a system through the process of contagion, emergent risk is the threat to the individual parts produced by their participation in and interaction with the system itself.

Whether or not the difference between emergent and systemic risk is just a matter of perspective along the causal chain or something qualitatively unique is a matter for discussion. One bad bank (if big enough) or one sick patient (if contagious and social enough) could bankrupt or infect apparently sound and healthy equivalents that are of a significant distance from them. The process by which this becomes a system-wide crisis is the focus of systemic risk, while emergent risk is the process by which system-wide interactions produce new and unpredictable dangers for individual actors that overwhelm formerly adequate defenses. With emergent risks, the line between endogenous and exogenous failures becomes so blurry as to make hazard estimates difficult, if not impossible. What is critical about emergent risks is that they can emerge without mutations or malfeasance, as a result from the "invisible hand" slapping the system dependant on it (Gorton, 2009). If systemic risk is about how one bad apple can spoil the barrel, emergent risk is all about how being in the barrel may spoil the apples.

Two additional concepts are useful in order to appreciate the implication of emergent risk: "normal accidents" (Perrow 1984) and "black swans" (Taleb 2010). The first refers to the inherent unpredictability of complex and tightly coupled interactions. As we connect more and more parts to each other, the number of possible outcomes increases exponentially, producing statistically "normal" accidents. The second reminds us that even the most unlikely of outcomes is still possible and that there is no such thing as absolute certainty. Emergent risk is the risk that unexpected and unlikely interactions may lead to unpredicted threats to system survival.

The risks associated with such emergent properties are significant and growing. Not only are we increasingly dependent on the continuous flow of complex global transactions for our daily lives, but the increase in these connections and the exponential rise of associated interactions make the system more fragile even as its apparent robustness increases. Some strategic attention has been paid to the susceptibility of the global system to intentional attack on critical nodes; much less has been paid to the fact that interacting and apparently insignificant failures in distant and apparently inconsequential nodes could produce a catastrophic outcome (Bostrom and Cirkoic 2008). While the probability of such causal reactions may be small, the subsequent severity of such outcomes and even the increases in unexpected variance are high enough as to merit significant attention (Christopher and Peck 2004).

The Global Financial Crisis (GFC) of 2007-2008 is a stark example of emergent risk. The prime mover was the US housing bubble and misbehavior by parts of the financial community. Consequent failures produced the possibility of ramifying crisis through contagion and excess le-verage. However the fundamental global threat did not come from bad loans in themselves, but from emergent interdependencies among participants in capital markets that made it almost impossible for actors in the system to estimate the real risk of many transactions and permitted the decimation of relatively healthy and well behaved economic agents. Another example of emergent risks is the global supply chain on which production of even the simplest good depends (www.husdal.com). Critical aspects of contemporary manufacturing management have contributed to this fragility: just-in-time inventories, economies of scale, outsourcing, and consolidation of suppliers. Each and every one of these trends has increased productivity and is economically optimal, but past a point they also increase the systemic danger to the whole. As in the case of finance, we may speak of systemic risk due to the power-law distribution of trade (Kaluza, Kolzsch, Gastner, and Blasius 2010) as well as the concentration of freight along six geographical bottlenecks (Panama, Suez, Gibraltar, Bosporus, Hormuz, and Malacca). More significantly, the interlocking nature of purchasing and supply may create disastrous bottlenecks or shortages even when the strategically central nodes and links are functioning.

In short, globalization carries within it specific risks and challenges over and above those associated with other forms of market or system failure. The very logic of increasing global-ization carries within it the seeds of its own destruction. Even in light of this, few would argue for a return to autarchy (even if it were possible). What is needed is to design governance struct-ures and institutional arrangements that to reduce the probability of such dangers arising and lower the costs when they do. For this, we need to first understand the nature of the danger.

Research

There has been much work on financial systemic risk and, of course, the GFC (Gorton 2008, Haldane 2009). Less attention has been paid to global supply chain risks (Manuj and Mentzer, 2008). There is, of course, a growing literature on complexity theory and its social ap-plications (Mitchell, 2009) and how it points to systemic risks of a new sort (Crutchfield 2009, Helbing 2009). Some scholars are incorporating some of these insights into analyses of global governance in the 21st century (Urry 2003, Goldin 2010, Oborne et al 2009). To our knowledge, however, these scholarly communities --- complexity theorists and modelers, substantive experts in globalization, and policy analysts – have operated isolation.

Policy Implications

Is emergent risk the policy equivalent of the random space asteroid? That is, no matter its salience, is it possible to do anything about it other than to know that it is probable?

The first task is more precisely defining and delineating the concept. Is this simply an interaction effect by another name or does the complexity and number of interactions require a new form of nomenclature and analysis? In what areas might we best study this phenomenon? The GFC may be a perfect example of one emergent risk that came to fruition. What are others? How delicate are supply systems or energy deliveries (two obvious candidates). What about other fields? No matter the field, how do we measure emergent risk? At its simplest level, is this a linear function of underlying complexity or are there threshold and feedback effects? Even if these questions have answers, we still have the issue of what to do about it. In this way, the analysis of emergent risk may represent a critical new format for policymaking in that what we are doing is not so much estimating the cost of a disaster or even the likelihood of it occurring, but rather the systemic characteristics of global networks and the inherent uncertainties within them. To borrow from Secretary Rumsfeld, this is about estimating the dangers of the unknown unknowns. One obvious question in this regard is whether policy might best focus on remediating failures (such as facilitating and speeding up responses to crises) or on constructing governance and regulatory regimes that make dangerous intra-system interaction less likely.

Project

This project seeks to create the foundation for a more integrated focus on emergent risk through:

  • Convening an international workshop on "Global Complexities and Emergent Risk" at Princeton on September 28-29, 2012.
    • There will be three panels. The intention is to provide multiple viewpoints of emergent risk as a complex adaptive process. Each panel deals with one part of the overall process of emergent risk:

Section I: Macro perspectives and the longue durée

The first section will focus on broad macro-historical perspectives. A critical question is whether how history is conceptualized, e.g. in terms of a linear trajectory or as cycles, matters for our understanding of emergent risk. Issues to be discussed include: the differences and overlaps between emergent and systemic risk; how today’s networked economy, linked by global trade and the Internet, has changed the nature of these risks.

Section II: Boundaries of Uncertainty – the case of finance

Why did so many experts neglect the warning signs that preceded the 2007-08 global financial crisis? One theory is that the models used to estimate risk gave rise to a situation where any event outside these models’ scope was unpredictable because it could not be measured. This raises the problem of whether we can ever know "unknown unknowns". Is it meaningful to speak of "risk" in situations of uncertainty?

Section III: Complexity as a Paradigm for Emergent Risk

This section examines if and how approaches to harness complexity might be better able to tackle emergent risk. Can a threat be turned into an opportunity? Are there similarities across cases that can provide for a "unified" complexity model for identifying and dealing with risk? Perspectives from different fields will be up for discussion.

The conference will be followed by:

  • Publication of a volume analyzing the realities and policy challenges of emergent risk.
  • Preparation of a proposal for research support to encourage analyses of applications of emergent risk.


References

Bostrom, N., M. Cirkovic. 2008. Global Catastrophic Risks. Oxford: Oxford Univ. Press.
Christopher, M., H. Peck. 2004. Building the Resilient Supply Chain. Inter. J. Logistics Mgmt.: 15: 1-13.
Crutchfield, JP. 2009. "The Hidden Fragility of Complex Systems". Complexity Sciences Center, UC Davis.
Goldin, I., T. Vogel. 2010. Global Governance and Systemic Risk in the 21st Century: Lessons from the Financial Crisis. Global Policy 1: 4-15.
Gorton, G. 2009. "Slapped in the face by the Invisible Hand: Banking and the Panic of 2007". Federal Reserve Bank of Atlanta 2009 Financial Markets Conference, May 11-13.
Gorton, G. 2008. "The Panic of 2007". Federal Reserve Bank of Kansas City, Jackson Hole Con-ference, August.
Haldane, A.G. 2009. "Why Banks Failed the Stress Test". Marcus Evans Conference on Stress-Testing, February 9-10, London.
Helbing, D. 2009. "Systemic Risks in Society and Economics". Santa Fe Institute.
Manuj, I., J.Mentzer. 2008. Global Supply Risk Management. J. Bus. Logistics 29:133-55.
Mitchell, M. 2009. Complexity: A guided Tour. Oxford: Oxford University Press.
Oborne, M. et al. 2009. Horizons: OECD Work on the International futures Program. Brussels.
Perrow, C. 1999. Normal Accidents. Princeton, NJ: Princeton University Press.
Richardson, B., P. Gerzon. 2004. "Emergent Risks", Institute of Risk Management, London.
Sawyer, K. 2001. Emergence in Sociology: Contemporary Philosophy of Mind and Some Implications for Sociological Theory. Am. J. Soc. 107: 551-85.
Taleb, Nassim Nicholas. 2010. The Black Swan (2nd. Edition). New York: Random House.
Urry, John. 2003. Global Complexity. Cambridge: Polity.