Complexity, as a science, first emerged in the late 1940s, although it has been studied by humanity for 1000s of years (Castellani & Gerrits, 2024). The term ‘complexity’ is rooted in the Greek word plektös, which means woven together. Historically, the concepts of simplicity and complexity have been linked to perceived ease or difficulty of understanding (Flood & Carson, 2013). Such interpretations, however, are inherently subjective and vary according to individual perception.

‘Complexity’ describes a phenomenon whose interconnected nature surpasses the explanatory scope of mechanistic or linear models. While certain individual elements may be understood in isolation, the behaviour of larger and more intricately related systems requires a focus on emergent patterns, learning, innovation and adaptive processes (The Santa Fe Group, 1996).

The principles governing the emergence of both knower and known, the boundedly rational yet optimally complex biological and social agents, form the foundation of complexity science (Kauffman, 1991). This body of thought contends that the fragile coherence and stability of biological and social systems are precarious and vulnerable to sudden disruptions or phase transitions. Systems positioned at the edge of these thresholds, sometimes referred to as the edge of chaos, serve as paradigmatic expressions of adaptive order (Kauffman, 1991). That said, systems do find positions of stability, the defining characteristic being that complex systems can be disrupted (go through a phase transition of exponential change) with relatively small disrupters.

The study of the behaviour of collectives, like groups of people and organisations, in terms of how they structure themselves and their dynamics is endowed innately with the potential to evolve over time; they remain emergent (Coveney & Highfield, 1996).

A common understanding around the definition of complexity (Gleick, 1989; Parker & Stacey, 1994; Stacey, 1996; Wilding, 1998) highlights that complex behaviours occur in different kinds of systems, such as biological, ecological, economic and social, encompassing chaos because of its nonlinearity. Chaotic dynamics are evident in business environments, where system sensitivity and nonlinear relationships can separate causes and their effects across time and space. Such conditions render future outcomes inherently uncertain, posing substantial challenges for organisational decision-making and strategy, and therefore governance (Mena, 2003).

Complexity theory offers a conceptual foundation and methodological rubric for applying study through disciplines, such as both traditional, agent-based modelling and systems dynamics, to understand complex environments.

Traditional modelling: Traditional modelling often relies on linear, reductionist approaches that emphasise equilibrium and predictability. It typically adopts linear and reductionist assumptions, focusing on aggregated patterns and controllable structures to explain organisational outcomes (Hoogervorst, 2009).

Complexity theory, however, uses traditional modelling approaches, such as formal system models and information theory, to develop understandings of systems by capturing their interconnected components and dynamic behaviours (Varley, 2023). These models provide frameworks to represent complex systems’ emergent properties and interactions. By integrating various data sources and managing uncertainty, traditional modelling helps to bridge metaphorical descriptions and precise representations, supporting better insights into systemic complexity and decision-making (Batty & Torrens, 2001). Traditional models offer a structured way to explore complex systems beyond linear causality.

Agent-based modelling: In agent-based modelling, a system can be understood as an assemblage of autonomous agents (these could be individuals, organisations or systems within systems) that interact with one another and with their environment (Janssen, 2005; Chu et al., 2003). Each agent functions as an independent decision-maker, adapting its behaviour to internal environmental conditions and the actions of other agents within the system in accordance with specified rules. Inevitably, these dynamics can extend beyond the explicit boundaries of the model. Agent-based modelling generally does not incorporate open systems for this reason, as they exchange energy, matter or information with an external environment. Agent-based modelling focuses on the internal dynamics within a system, simulating the behaviours and interactions of individual agents, rather than the system’s interactions with the external world (Deguchi, 2011).

Agent-based models can therefore be used to understand closed complex systems that evolve under conditions far from equilibrium and are shaped by historical trajectories. They consist of numerous interacting agents, whose relationships with each other are dynamic, nonlinear and often localised (Bonabeau, 2002). Individual agents typically lack awareness of the overall system behaviour, while their interactions generate both reinforcing and balancing feedback loops, giving rise to emergent properties and adaptive dynamics within the system (Maguire et al., 2006).

System dynamics modelling: System dynamics modelling is a prominent technique within the complexity theory framework used to understand systems and their interactions by capturing feedback loops, delays and nonlinear relationships (Bala et al., 2017; Sterman, 2002). It enables simulation of system behaviour over time, linking structure to dynamic outcomes, which helps in identifying leverage points and testing scenarios for better decision-making. System dynamics modelling treats systems as a whole rather than isolated parts, emphasising interdependencies and adaptation. This approach supports the exploration of emergent properties and dynamic complexity in social, environmental and organisational contexts (Coyle, 1997; Sterman, 2002).

Much of the complexity research has been in physical and biological systems, although there is a growing body of research in social systems (Kauffman, 2022; Sawyer, 2005). Generally, the difference identified is that of independent free will, which is not generally present in biological and physical systems. It has been suggested that purposeful elements and adaptive responses, such as those used in biological systems, can also be extended to social organisations, with complexity theory providing a framework that accommodates free will (ed. Kane, 2011).

The concept of free will implies that conscious beings possess sufficient autonomy from physical constraints to intentionally shape future outcomes through goal-directed action, a form of intrinsic intentionality that is unique solely to conscious organisms (Shepherd, 2012). Therefore, organisations, which are groups of humans with free will, cannot be reduced to transformation processes or control systems; their survival requires mechanisms for interpreting ambiguity and generating meaning and direction for participants (Kauffman, 1995). Organisations are human and therefore ‘meaning’ systems; this distinguishes them from lower level biological systems (Daft & Weick, 1984). It is perhaps Kauffman’s (2022) definition of complexity which bridges this divide between biological and social. He describes complexity as a systemic condition marked by both stability and adaptability. Such systems exhibit a high degree of internal organisation that ensures coherence while simultaneously maintaining the capacity for flexibility and unexpected outcomes. The relative autonomy of individual components, coupled with their diverse patterns of interaction, gives rise to nonlinearity, unpredictability and emergent properties. In this sense, complexity emerges from the dynamic tension between order and indeterminacy. Perhaps, this is where we find governance.

Complex systems, therefore, are distinguished by their interconnected nature, fundamental unpredictability and controllability, as well as nonlinearity. Within such contexts, scientific inquiry, including modelling, moves beyond the pursuit of prediction and control, instead prioritising an enhanced understanding of systemic dynamics and interrelationships. This shift in focus enables more informed and contextually appropriate forms of engagement within the systems to which we belong and allows us to choose to participate in the system with more strategic intelligence (Wahl, 2019), perhaps precisely what is required in the management of governance.

Governance structures and systems, therefore, can be seen as inherently complex and refer to the intricate and multifaceted nature of organisational and institutional systems through which authority is exercised, decisions are made, resources are mobilised and the way in which these challenges are addressed (Commission on Global Governance, 1995). It is an ongoing process that allows for the reconciliation of diverse interests and the facilitation of cooperative action. Governance includes formal institutions and regulatory frameworks with the authority to enforce compliance, as well as informal arrangements that actors recognise or adopt based on perceived mutual benefit (Carlsson & Ramphal, 1995).

Within global governance, Fuß et al. (2021) describe this complexity in structural terms, highlighting dimensions such as scale, diversity and density. Scale captures the proliferation of governance elements across three dimensions: the growing number of actors (e.g. governments, firms, non-governmental organisations (NGOs) and civil society groups), the expanding range of institutions (formal and informal rule systems and the organisations that embody them) and the interaction of multiple levels (local, national, regional and global). Diversity reflects the heterogeneity of these actors and institutional forms, while density refers to the intensity and overlap of their interconnections. These aspects underscore why governance is irreducible to linear processes of control and therefore must be understood as a dynamic and adaptive system, shaped by interdependence and emergent outcomes (Klijn & Koppenjan, 2016).

The complexity of governance also involves the perspectives of simple view, thin complexity and thick complexity (Strand & Fjelland, 2000). They distinguish between three ways of understanding complexity in governance. The first one is a simple view, which observes the system as linear and reducible, where outcomes can be predicted through stable causal relationships. By contrast, a thin conception of complexity acknowledges uncertainty and nonlinearity yet seeks to manage this by simplifying these challenges into models or heuristics. A thick conception of complexity describes the system as irreducible, constitutive of reality and that uncertainty, emergence and dynamic interactions are inherent features of governance systems. They further emphasise the emergent qualities of governance processes, recognising that interactions among actors, institutions, organisations and environments cannot be fully controlled or eliminated. This perspective shows the limits of prediction and highlights the importance of embracing adaptive, reflexive and participatory forms of governance.

Strand and Fjelland further indicate the importance of complementarity and contextuality when understanding the complexity of governance. The principle of complementarity suggests that different perspectives on governance complexity can coexist but are not mutually exclusive. Contextuality stresses that the appropriateness of any governance approach depends on its context. Complex systems cannot be governed with a ‘one-size-fits-all’ model; instead, responses must be sensitive to cultural, institutional and situational factors.

Stacey (2000) also suggests that the worldview of thin complexity must be enriched with the understanding of complementarity and contextuality. This implication is faithful to complex systems theory, where the question is not ‘How can I govern the system into a new attractor’ (a desired state)? but rather ‘What is my role in this system, and how do the actions of me and others affect the system?’. This aligns closely with Wahl’s (2019) perspective, where he suggests that it is the participation with the system that is critical in influencing the performance of the system.

To operationalise effective governance in such contexts, specifically within organisations, involves fostering collaboration across multiple levels and sectors, integrating diverse knowledge sources and understanding the aspects of uncertainty and unpredictability. This perspective aligns with Kreienkamp’s and Pegram (2021) assertion that existing global governance structures are often ill-equipped to manage complex global catastrophic risks, advocating for a design model that accommodates complexity through iterative processes and adaptive mechanisms. Therefore, contemporary governance must evolve to embrace complexity, within and outside of the organisational environment, to ensure resilience and responsiveness in the face of global challenges (Kreienkamp & Pegram, 2021).

Modelling and understanding governance as a complex system necessitates a dynamic, adaptive approach that transcends traditional hierarchical models, suggested Duit and Galaz (2008). They go on to suggest that conventional governance frameworks often struggle to address the nonlinear dynamics, threshold effects and cascading events characteristic of complex adaptive systems. These systems, marked by interdependencies and emergent behaviours, require governance structures, models and measures that are flexible, decentralised and capable of learning and evolving in response to changing conditions.

Modelling governance from the perspective of complexity required the consideration of how the complexity governance model (Model) would be both pragmatic and theoretical. Given the continued challenges facing governance and board evaluations and the expectation from regulators and other stakeholders to approach and measure aspects of governance, such as culture, decision-making and oversight, these factors should be central to the model. The model also needed to account for both the circularity of governance and decision-making, as well as oversight processes commonly found in organisations and their interconnected nature.

Three core foundational domains to the model emerged through the review of literature: Decision-making, Implementation and Oversight, and Organisational Culture. To acknowledge and visualise the intersecting nature of these three, they were drawn as a Venn diagram, assisting with the immediate recognition that these foundational elements are interconnected; see Figure 1. The decisions made (Crivelli & Balconi, 2023; Morrison, 2025) and the actioning of those decisions (Ellinas et al., 2017; Franca & Hollnagel, 2023; Zheng et al., 2024), individually and/or collectively, are influenced by the prevailing culture of any group, and the culture of the group influences the manner in which decisions are made and the nature of those decisions. Organisational governance is more than the sum of its parts; it is more than the aggregate.

Governance Quality (GovQ) is the outcome of the three interconnecting domains, each of which has a reciprocal relationship with the others, the outcome being the interplay between the domains. The domains themselves are also constantly in flux individually, feeding back into the overall system, resulting in a complexity model with observable emergent properties and internal feedback mechanisms. The ‘dashed’ arrows indicate the interconnectedness and that this is multidirectional.

Each of the three (n = 3) foundational domains of GOVIndicia^® has nine (n = 9) categories that contribute towards and influence them. These categories are also intimately connected to each other, and so the model recognised this by interconnecting each of these groups of three categories and by interconnecting them to their respective core domain. The categories are noted in Table 1 – The Domains, Categories and Indices of the Governance Complexity Model, GOVIndicia^®.

The nine categories themselves also have three indices that influence them. The model, therefore, has three (n = 3) foundational domains, nine (n = 9) categories and 27 indices, as shown in Figure 2. The indices are noted in Online Appendix 1 – The Domains, Categories and Indices of the Governance Complexity Model, GOVIndicia^®.

All models are a distillation of the real world, and not all functionally operate. However, if they do, they should inform the model user of something about the world that it represents. Therefore, observing and visualising the way or ways a model operates helps in developing an understanding of how the different facets within a model interact with each other. Models rarely operate in isolation, as they tend to be the representation of a part of a larger system. It was a requirement to acknowledge that this model interacts with its environment (everything outside the system), noting that this interaction was not part of this model. This issue was acknowledged earlier when system dynamics and agent-based modelling were discussed, and while this is not an agent-based model, agents (individuals within an organisation) are, in fact, those acting out many of the categories and indices within the model. This model is not the whole system; it is a part of a bigger, even more complex system. This model sought to model in a functional and pragmatic way the complex nature of interactions between aspects of governance in an internal setting (organisational governance) where social interactions, behavioural dynamics and institutional norms converge to influence collective outcomes.

The main essence of ‘modelling’ any system is to try and draw out key themes and present them in a way that is meaningful and will support better understanding. In this case, the model seeks to further understanding around intra-organisational governance. Governance practice, in fact, most practices within organisations are cyclical, to provide for decision, action and then review; this model accommodates this. It also accommodates the innate complexity and interconnectedness of governance as highlighted in the literature review.

The model is cyclical. It models the process of decision-making from an internal governance perspective rotating in a clockwise direction, and although there may be no specific start or end point, for want of illustration, by way of example, it starts with the category Purpose & Values.

Organisations (ISO 37000) to practise good governance should congregate around a sense of purpose and values (this might not always be an aspirational purpose, but a purpose, nonetheless). Leaders and leadership emerge, and along with that certain behaviours and an organisational ethic. At this point in the cycle the true nature of governing occurs; making decisions towards the purpose. This process includes the gathering of information, the flow of information and the quality of the decision-makers (experiences, knowledge and understandings of the world and its affect on decision-making). Once any decision has been made the decision may need to be cascaded or communicated. The methods of communication, their clarity and how emotional intelligence is used to develop and nurture communication feeds into the overall communication of any decision. This will also impact how well the decision is implemented; its operational efficacy (the efficiency and effectiveness of an organisation’s processes, those that are formal and informal). Finally, to establish the outcome of our decision and implementation, most organisations have some mechanisms to detect this measure of success in the cycle through feedback and learning, which the model calls ‘reasoned persistence’ (do we persist with the current decision and implementation, or do we need to adjust, reconsider or take an alternative course – we reason our persistence with a particular decision). Then, of course, it starts all over again. As with all models, this cycle is by no means fixed, as sometimes we might miss a stage or two in the process.

The proposed model moves towards the conceptual idea of complexity by visualising the interconnected nature of the three core domains of the organisational governance system. To illustrate how this manifests itself in the model, it can be seen that there are clear relationships that cross the cycle. The purpose and values of an organisation are likely to have some level of connection with the nature of the decisions made and the nature of the implementation. The behaviours and ethics that emerge through the nature of leadership are likely to have an impact on the communication.

The model indicates these two operational aspects through the dotted and solid arrows. The solid arrow indicates the cyclical nature, and the dotted arrow indicates the interconnected nature.