Steering without Circe: attending to reinforcing loops in social systems

Abstract

Relating system dynamics to the broad systems movement, the key notion is that reinforcing loops deserve no less attention than balancing loops. Three specific propositions follow. First, since reinforcing loops arise in surprising places, investigations of complex systems must consider their possible existence and potential impact. Second, because the strength of reinforcing loops can be misinferred-we include an example from the field of servomechanisms-computer simulation can be essential. Be it project management, corporate growth or inventory oscillation, simulation helps to assess consequences of reinforcing loops and options for interventions. Third, in social systems the consequences of reinforcing loops are not inevitable. Examples concerning "globalization" illustrate how difficult it might be to challenge such assumptions. However, system dynamics and ideas from contemporary social theory help to show that even the most complex social systems are, in principle, subject to human influence. In conclusion, by employing these ideas, by attending to reinforcing as well as balancing loops, system dynamics work can improve the understanding of social systems and illuminate our choices when attempting to steer them. Copyright © 2008 John Wiley & Sons, Ltd.Syst. Dyn. Rev. 24, 37-61, (2008) We are very honoured and happy to receive the Jay Wright Forrester Award. We are grateful to the members of the award committee, and to Peter Milling for his generous introduction. We are also in debt to many other people who have helped us and we would like to thank at least some of them here.We should like to mention our teachers at the universities of Bristol, Bielefeld and Oxford. We should also like to thank colleagues at LSE and in the fields of problem structuring methods and system science. The very first contact either of us ever had with system dynamics was when David sat in a room by himself for a week with a Mac 512 k, a copy of STELLA and the manual. It was Barry Richmond's fine work that set him off on a new course. He was also lucky enough early on to meet people who were great examples and who encouraged his growing interest in system dynamics. John Sterman, together with David Kreutzer, ran the very first system dynamics course that David ever went to. John's enthusiasm was infectious then and he has been immensely supportive ever since. One of the first people David got to work with on actual projects was John Morecroft and from that experience he saw directly how system dynamics could be used to improve organizations. Some of the earliest advice about work on problem structuring methods came from George Richardson,

Advertising effectiveness and spillover: simulating strategic interaction using advertising

“…When the industry is responsive to advertising, advertising pays off handsomely; when the industry is not responsive to advertising then advertising yields little competitive advantage. Such variation in behaviour is not uncommon in complex systems where loop dominance can change and where dynamic behaviour is not easily induced from infl uence diagrams: hence the usefulness of a simulation model (Lane and Huseman, 2008). The results also suggest that predatory advertising yields more benefi t for a fi rm than does cooperative advertising.…”

‘Till the Muddle in my Mind Have Cleared Awa’: Can We Help Shape Policy Using Systems Modelling?

“…There is a range of possible unintended consequences of this policy, and some of these can also be mapped (Lane and Husemann, ). These ideas are represented in Figure , RHS.…”

The web as an adaptive network

“…to support homophily [23,42]. Both processes will reduce the number of social links between actors of differing opinions (either by changing opinions or by changing links) creating self-reinforcing loops in social structure and behavior [43][44][45]. Different separations of timescales between the two processes (e.g.…”

“…Specifically, in the modern 'structuration' [43] perspective, "social phenomena are treated neither in terms purely of social structure nor in terms purely of human agency, rather a view is adopted which treats structure and agency as influencing each other" [45]. Similar ideas are also current in the field of evolutionary biology where it is recognized that we cannot understand the evolution of an organism independently of its environmental niche, nor can we assume that its environmental niche is fixed, but we must recognize that organisms construct their own niches as a response to selection but thereby also alter their selection [71].…”

“…One of the general organizing principles that appear to be relevant to understanding structure/agency interaction in networks is the notion of self-reinforcing loops [45] -i.e. actors tend promote structural changes in a network that support their current behavior, and structures enable or constrain agent behaviors to those that support (or do not disrupt) the current structure -creating a positive feedback loop.…”