Research and analysis

Integrating a systems approach into Defra

Published 23 May 2022

“Systems thinking puts world leading evidence at the forefront of Defra’s strategy and decision making. This multidisciplinary and research-led approach to problem solving is already shaping innovative and well-informed solutions to some of the UK’s greatest environmental challenges.

“The use of systems approaches, through the launch of the systems research programme, has enabled Defra to be more strategic in research funding and to work collaboratively with government departments on cross cutting issues. Systems approaches equip us to better protect and understand the natural and socio-economic systems that are influenced by Defra’s policy decisions.

“I welcome this document, which is the first of its kind to introduce a standardised systems approach for application in Defra and across government. It demonstrates the value and potential that systems approaches hold for the future of Defra policy. I hope policy teams will find it of use across the full breadth of the environment sector and beyond.”

Gideon Henderson, Chief Scientific Advisor at Defra

About this document

Defra’s Systems Research Programme, launched in 2019, has been applying systems methods across the organisation to support complex decision making.

The outcomes of the programme have revealed the many benefits of applying systems approaches to Defra’s work. These include:

• facilitating policy join-up
• mitigating negative unintended consequences of policy decisions
• enhancing the robustness of Defra policy

This document is inspired by the findings and insight acquired by the Systems Research Programme to date. It has been carefully designed to:

• highlight the value and ongoing need for systems thinking in Defra
• raise awareness of systems tools and methods that are applicable to Defra
• encourage the use of systems methods which can support and enhance the work of Defra colleagues across the network

Defra network colleagues and policy makers can refer to this document to:

• familiarise themselves with the key principles of systems thinking
• improve their awareness of ‘systems approaches’
• explore how systems approaches may be applied to their own work and to Defra decision making

The primer can be used with or without prior systems experience. Its content and language has been loosely defined to encourage user interpretation, broaden applicability and optimise opportunities.

Further queries, or expressions of interest, are welcome via the following email address: Systems.Research@defra.gov.uk

1. Introduction

This section explores the complexity of Defra’s challenges and the subsequent need for systems approaches. It defines how complex system interactions contribute to Defra’s ‘wicked’ problems.

1.1 Overview

In Defra we work in a complex and ever-changing arena. Our areas of policy responsibility cover natural and human systems that interact in unpredictable and dynamic ways. The issues we work on are highly interconnected and our key stakeholders often have very different perspectives of what is important.

Defra policy strives to meet a multitude of ambitious outcomes. The breadth and variety of these desired outcomes make Defra’s role and overall ability to deliver robust, joined-up policy challenging and complex.

Teams need to think about how their decisions affect goals in other parts of government. For example, decision makers need to define policies that deliver the many interacting goals in the 25 Year Environment Plan without compromising other national and international goals and commitments (for example, National Infrastructure Strategy, Clean Growth Strategy and UN Sustainable Development Goals).

Our thinking needs to encompass social, environmental, economic, technological and political systems in order to consider policy impacts on different stakeholder groups from local to international spatial scales and timeframes spanning months to generations (Figure 1).

Systems thinking helps us to embrace this complexity, providing tools, approaches and ways of thinking that can be used to explore the interconnected nature of policies, to decide which interconnections should be accounted for in decision making and which can be left out. We can use systems approaches to reframe situations, to account for the conflicting needs of different groups, make informed decisions when reframing is not possible and take advantage of synergies between policies across government.

Systems approaches can be used to explore complex, real-world, problems to reveal unknown variables and interrelationships. They can provide a simplified version of reality to reveal hidden links between cause and effect and help to develop solutions to well-defined and data-driven challenges. Systems thinking is more than a practical way of working, it changes mind-sets, how issues and opportunities are approached and the philosophical basis upon which decisions are made.

The systems field is extensive and there are many useful books, courses and web resources that provide detailed guidance on methods. This primer does not attempt to replicate them but gives an overview of how systems working can provide insights to the types of situations faced by Defra’s policy teams.

In particular, this document explores the principles of systemic working, as well as introducing simple techniques for applying systems thinking. It aims to help build skills and awareness of how systems approaches can be used in 3 areas:

• framing policy discussions
• working at the science-policy interface
• structuring dialogue with stakeholders and the public

Figure 1: An example of the complex relationships and dynamics operating within a whole-system approach. Represented by a map of multiple (social, environmental, economic and governing) systems all influencing the outcomes of the United Nations (UN) sustainable development goals.

Figure 1 is a graphic showing the complex dynamics operating within a whole-system approach. The figure represents how social, environmental, economic and governing systems influence the outcomes of the United Nations (UN) Sustainable Development Goals (SDG’s) alongside other desired outcomes.

The graphic demonstrates that a whole systems approach considers:

• multiple stakeholders (i.e. purposes, values and boundary judgements)
• multiple spatial scales (i.e. local, national and international)
• multiple timescales (i.e. months, years, decades and generations)

1.2 Understanding wicked problems and sources of complexity

1.2.1 Complex and wicked problems

The types of problems requiring decision making by government can be classified as ‘simple’, ‘chaotic’, ‘complicated’, or ‘complex’ based on the predictability of cause-and-effect relationships ^{[footnote 1]}.Understanding the type of problem can help identify the most appropriate approach.

• A ‘simple’ problem is where the relationship between cause and effect is established. For example, inspecting a business for compliance with relevant regulations. Solutions are likely to follow a well-known and ordered process
• A ‘complicated’ problem is where the relationship between cause and effect can be predicted by experts. For example, predicting the response of a crop when given nutrients. Solutions require the allocation of sufficient time and resource to conducting a thorough analysis
• A ‘complex’ problem is where the relationship between cause and effect is unpredictable. For example, understanding the many interacting socio-ecological factors affecting land use change. Solutions require systems thinking, engaging with a plurality of stakeholders with decision makers learning and acting as they go
• A ‘chaotic’ problem is where the relationship between cause and effect is unknowable. For example, the emergence of the next global pandemic. The solution is rapid, responsive and decisive action

The environmental systems, food systems and resource systems that influence Defra’s policies often behave in ways that are complex and unpredictable. In areas like land management, waste management, biodiversity conservation, pollution prevention, food security and fisheries, the cause and effect of arising issues are seemingly complex or unattainable.

The terms ‘wicked’ ^{[footnote 2]} or ’super-wicked’ are sometimes used to describe problems like climate change, biodiversity loss or soil degradation. Such problems may have characteristics including:

• problems seem intractable with root causes that are hard to identify
• solutions are contested
• different actors have their own perspectives or interests, which influence their understanding of the situation and the solutions they propose
• coordinated action is needed across different sectors
• those seeking to solve the problem may sometimes exacerbate the issues
• there may be no central authority
• time can be pressured
• key processes and interrelationships may be only partly understood and evidence is often fragmented

In these kinds of situations, policy needs to be designed in the absence of perfect knowledge of how human and natural processes interact, and in a way that draws together fragmented evidence and multiple perspectives. Systems approaches can provide structure in such cases.

1.2.2 Understanding sources of complexity

Sources of complexity for Defra include environmental, human and organisational barriers that stand in the way of effective policy making or problem solving. Some examples of how these complexities present themselves through environmental, human and political factors are listed below:

Environmental factors of complexity

Examples of environmental sources of uncertainty which can influence Defra’s decision making include:

• natural time-lags - for example, nitrate levels continuing to rise after mitigation policies have been implemented, due to the slow rate at which nitrates and pollutants percolate down to the groundwater
• multiple drivers affecting a single goal - for example, the conservation of farmland bird populations is dependent on multiple drivers, including habitat availability, food availability, the threat of predation and the exposure to toxins and hunting along migration routes
• multiple outcomes from single interventions - for example, a single agri-environmental intervention may impact productivity, biodiversity, water quality, air quality, greenhouse gas emissions, carbon sequestration and cultural value of landscapes
• interdependencies - such as public engagement with the environment depends on thriving plants and wildlife
• unpredictable emergent problems - for example, reduced air pollution levels resulted in sulphur deficiency in livestock in some areas
• feedback loops and tipping points - for example, raising nutrient levels in water bodies led to an ecological shift in water bodies that is very difficult to reverse
• site specificity - for example, what works to address biodiversity declines in one place might not work in another

Human factors of complexity

There are a multitude of interacting social, cultural and economic factors which drive human behaviour and decision making. When humans play a role within a system they can be described as an ‘actor’ within a system. The complexities associated with human actors within a system stem from:

• culture
• values
• perceptions
• knowledge
• power dynamics

These sources of human complexity can make the outcomes from that system very difficult to predict. An example could be predicting the role humans would play in a shift to novel proteins or the complex role of human actors in addressing plastic pollution.

Political factors of complexity

Examples of political sources of uncertainty which can influence Defra’s decision making include:

• scale issues - for example, where an action is implemented on a localised level (such as in habitat restoration) do not always contribute to national shifts (such as species abundance)
• evidence challenges - for example, where the available evidence is limited, fragmented, outdated or when constrained by the limits of experimental design (for monitoring there can be challenges of attributing environmental impacts to specific pressures)
• plurality of interests - for example, when engaging with different stakeholders, each individual or group will have different values, goals and perceptions, which can hamper local conservation efforts

Within the decision-making process, each decision maker will have their own perspectives and assumptions about a situation. The views and opinions of decision makers can influence the solutions they propose and can sometimes exacerbate the problem further. These differences in perspectives can allow hierarchies of power to undermine systemic thinking.

Systems thinking helps to unpick situations such as these, which are complex or seemingly chaotic. Applying systems thinking ensures that:

• key processes and human and natural interrelationships are understood
• boundaries are explored
• multiple perspectives are accounted for and positive relationships are built amongst stakeholders
• policy is supported by strong evidence

1.3 What is a system?

For the purpose of this document, a system has been defined as “a collection of entities that are seen by someone as interacting together to do something.” ^{[footnote 3]}.

The entities in this definition may be physical elements, human activities, or conceptual ideas. The degree of complexity within a system can be determined by the number of elements, the attributes of these elements and the number of interactions.

From the definition above, it is possible to identify different systems within Defra’s areas of work. For example: a farm, a fishery, the UK food network and even Defra itself may be considered a system.

To understand a system and its complexities, 4 key principles need to be explored. These principles form the building blocks of any practical ‘systems approach’ ^{[footnote 4]}:

1. Acknowledging the purpose of the system according to the key stakeholders.
2. Understanding the interrelationships between entities.
3. Determining whose perspectives should be accounted for.
4. Defining the boundaries of (or being conscious of which elements are included within) the system.

When exploring each of these systems principles several key questions should be considered.

When defining the purpose of a system it is important to ask:

• what are the main outcomes we (the key stakeholders) are interested in
• what does the system do
• how do different people view the purpose

When analysing the interrelationships between entities within a system it is important to ask:

• which processes or relationships are key to deliver the purpose
• where are there feedback loops
• where could we intervene

When determining whose perspectives should be accounted for it is important to ask:

• who are the main actors
• what are their goals
• what are their values
• how do they perceive the system
• who has control

When defining system boundaries, it is important to ask:

• which elements and processes are integral to the system
• what can be left out
• what are the key external drivers

A systems approach can be imagined as working at an intersection of these different principles, combining the known facts about the issue, the motivations and perspectives of the stakeholders involved, the decision about what to include and the interactions between these ^{[footnote 5]} (this concept is represented within Figure 2).

Figure 2. Systems approaches for policy making should be iterative: spiralling through an exploration of purpose, interrelationships, boundaries and perspectives to gain a greater understanding of the main tensions and watch-points.

Figure 2 represents the iterative nature of a systems approach for policy making. Represented by a spiral running through the 4 principles of purpose, interrelationships, boundaries and perspectives.

When thinking in systems each of these four principles (and associated groups of questions) need to be looked at through the lens of others, in order to strengthen understanding of relationships and key elements.

Each of these 4 groups of questions need to be looked at through the lens of the others. For example, different actors may have varied perspectives on a systems purpose. As understanding of the key elements and interrelationships improves, there might be a need to adjust boundaries.

2. Systems approaches

This section provides an overview of the systems field and the 4 key principles introduced in section 1.

It discusses the benefits and considerations of different approaches and reveals how these can be used to explore purpose, interrelationships, perspectives and boundaries of Defra’s systems.

2.1 The systems field

The systems field is very broad, arising from multiple disciplines such as biology, cybernetics, ecology, engineering, management and social theory. The systems field introduces us to a spectrum of approaches ranging from ‘soft’ (people focused) approaches to ‘harder’ (data driven) approaches. The two ends of the spectrum of systems approaches can be described as follows:

People-focused approaches

These softer systems approaches focus on engagement and communication to explore different perspectives on a situation and resolve conflict.

Such approaches can be useful where evidence is ambiguous and end goals are undefined.

Examples include soft systems methodology (5.1.1) and critical systems heuristics (5.1.7), which can be used to provide structure to discussions and organise ways of thinking.

These approaches are well suited to exploring behaviours and the ways in which people (including Defra policymakers) interact with the environment and with each other.

Model-focused approaches

These harder systems approaches focus on modelling a well-defined problem or situation. They can be data driven or based on qualitative information.

Such approaches can be useful when working with complex systems, where data and evidence is often partial and may need to be complemented by expert judgement. Examples of model focused approaches include system dynamics (5.1.2) and environmental modelling (5.1.3). These approaches are well suited to examining natural processes such as water or nutrient cycles and are often used to explore scenarios in Defra.

Across this spectrum of people or model focused approaches, a range of tools and methodologies exist which can be applied to explore systems of interest. Systems methods and approaches can be used flexibly or in combination. The way they are applied is dependent on the available data, the understanding of key processes and the nature of questions that are being asked during the process.

Different systems methods, such as those outlined in Annex 1 can be adapted from their original purpose and used in combination with more traditional (non-systemic) types of analysis (Figure 3).

Combined approaches are often the most appropriate way to shed light on the complex and chaotic situations faced by Defra. The key thing is to apply the systems principles (of boundaries, perspective, interrelationships and purpose).

Example: A working example of ‘air pollution’ can be used to demonstrate the need for flexibility in applying systems methods and approaches. A hard systems approach may lend itself more to the biophysical parts of the system, characterising pollution sources and modelling atmospheric processes. Soft approaches may focus more towards polluting behaviours and any impacts of exposure. In practice, these two approaches are complementary and should feed into each other to develop an understanding of the whole system. A combined approach enables the key principles (of boundaries, perspective, interrelationships and purpose) to be explored.

Figure 3: Systems analysis can combine activities from across the systems field alongside more conventional activities such as literature review and data mining

Figure 3 is a chart of systems analysis activities (including evidence synthesis, logic mapping, soft systems methodology and expert elicitation). The chart represents the breadth of systems analysis options within the systems field. Demonstrating how quantitative, qualitative, group and individual analysis types can be combined within a systems approach.

2.2 Key principles for a systems approach

The following paragraphs outline some key benefits and considerations for exploring the 4 principles of a systems approach defined in section 2.1.

2.2.1 Defining purpose

The purpose of a system is what it does. And what it does is a matter of perspective. Different stakeholders often have different views on a system’s purpose. A key task in any systems approach is to bring people together to find common ground and agree shared goals. The process of agreeing a system’s purpose supports collaborative discussion to determine which elements should be considered (boundaries), whose perspectives are important, which are the most important outcomes of interest and which are the key interrelationships.

The purpose of a system can be explored using Soft systems methodologies (SSM) (5.1.1). SSM can be used to explore common values which bring actors and stakeholders together and help form agreeable outcomes.

The way in which the purpose is defined can have a profound influence on decision outcomes. For example, reframing recycling from a waste disposal system to a system of resource provision affects the way people think about the situation and possible interventions.

2.2.2. Examining interrelationships

Acknowledging interrelationships (between elements within a system) and understanding the influence of those relationships (on the system as a whole) are key parts of system analysis. Exploring interrelationships should be an iterative process whereby systemic understanding of interactions can be further developed by evidence analysis (Figure 4). This approach unfolds as follows:

Qualitative system mapping

Qualitative system mapping provides insights into the dominant relationships and processes within a system and can help to identify which variables and relationships to focus on for evidence gathering (5.1.2).

Evidence synthesis

Conventional evidence review approaches can be used to explore relationships based on previous research (5.1.4).

Expert elicitation

Where research evidence is missing, new research can be commissioned if time and resources allow. Alternatively, expert elicitation can be used to shed light on key relationships (5.1.5).

Modelling and further mapping

Evidence from syntheses or expert elicitations can be used to iteratively improve qualitative system maps or to build quantitative models (5.1.3).

Figure 4: Using system mapping to target evidence gathering which in turn is used to improve maps in an iterative cycle:

Figure 4 shows how using system mapping to target evidence gathering in turn is used to improve maps in an iterative cycle. Arrows represent how the processes of systems mapping (which provides the framing) and evidence synthesis (which provides the evidence) feed in to each other in an iterative cycle.

In this process systems mapping reveals key interrelationships. These relationships prompt research questions for focused analysis.

Focused analysis (through evidence synthesis and expert elicitation) informs understanding of the wider system, which can be used to expand systems mapping.

System maps developed through this iterative process can be analysed to reveal feedback loops and the key factors affecting outcomes of interest within the system. They can be used to:

1. Identify potential points of intervention in the system. For example, interrelationships to strengthen or block through policy action in order to affect desired goals. In some cases, small changes in one part of the system may lead to a shift in the way the system works, resulting in significant change. Identifying points of intervention makes it possible to explore which actions (including decisions, shifts of perspective and policy interventions) have the potential to change the system in the most effective way.

2. Identify ‘watch points’ within the system. For example, variables that act as indicators for a key process or outcome to show that something in the wider system is changing. This can help design effective monitoring systems.

The following points should be considered when following the iterative (systems mapping and evidence synthesis) process:

• Trying to map everything can add to the confusion. A simple map is often more useful than a complicated one so a key question to consider as you go along is ‘what can be left out?’ A good ‘system map’ highlights key relationships of interest and provides insights into points of intervention
• Many models and modelling frameworks already exist for Defra’s environmental systems. The challenge is in bringing existing models together into an integrated systemic format which enables trade-offs and synergies to be explored (5.1.3)
• The iterative process for exploring interrelationships, set out above, requires ways of working which are both systemic (systems focused) and systematic (structural or methodical). Understanding the differences between these ways of working and applying them in balance, can help shape a well-rounded systems approach

2.2.3 Understanding perspectives

People can have very different, and sometimes opposing, perspectives of a given situation based on their knowledge, personal values and priorities. The way in which a system is modelled will depend heavily on who has been involved in the modelling process.

It is important to ensure that different perspectives are accounted for and consider which actors are included or left out of a systems process. Participatory approaches can be used to explore multiple perspectives and develop mutual understanding, especially where stakeholders’ goals diverge (5.1.6).

Multiple perspectives can be captured and understood by applying the following systems approaches (in a participatory way):

• Soft systems methodology (5.1.1) includes a range of approaches that can be applied to harness knowledge, beliefs and opinions in a constructive and collaborative manner
• Participatory system mapping (5.1.2) Developing system maps within participatory workshops can help to develop a shared understanding of what is happening in a system or situation of conflict. A map developed in this way encourages debate about which elements are important for decision making and which can be left out
• Conceptual models (5.1.8) Developing shared conceptual models, for example through participatory workshops, can be a useful way to gather perspectives on the various activities that contribute to a system purpose, showing the ideal system in action

Exploring and understanding perspectives, for example through participatory approaches can help to:

• Agree solutions that are acceptable to groups with conflicting agendas. Participants can come out of such an engagement with a better understanding of how others frame the system and better insights into unintended consequences and levers for change
• Bring policy teams together to design coherent policies that deliver multiple goals
• Facilitate constructive conversations with other government departments and external stakeholders

The following points should be considered when planning or applying any participatory process:

• Early engagement with actors is important to identify potential conflicts, power dynamics and contrasting motivations. Paying attention to these dynamics ensures that key groups are given the opportunity to share their views
• A ‘group-think’ mindset can develop between stakeholders inside the process which would cause the perspectives of those outside the development process to be ignored
• Marginalized stakeholders may be difficult to access and could need time and space to develop their thinking separately. When this is not possible, careful consideration needs to be made of how best to capture their perspectives. For example, it may be possible to find a third-party witness who can represent their views and needs. This may also be the case for representing the needs of future generations in decisions with long-term implications
• Individuals applying systems approaches are part of the system. They carry their own perspectives and will influence the way the system is viewed and presented. Within government, policy teams have different objectives which can lead to fundamental differences in the way problems are framed

2.2.4 Exploring and setting boundaries

Drawing a boundary around a system determines which factors should be included when tackling an issue. Exploring boundaries is a process which establishes the feasible and desirable outcomes that can be achieved within a defined set of parameters. Boundary critique begins with the understanding that not every interrelationship or perspective can or should be included in systems analysis.

Exploring and setting boundaries is an iterative process which can help to clarify the purpose and values of different interest groups. Boundary setting involves asking questions such as:

• What should we be looking at?
• What issues should we include?
• Who should we listen to?

Through asking such questions, boundary setting explores purpose, resources, scale, control, whose knowledge is relevant and who ought and ought not to be involved.

Exploring, analysing and setting system boundaries can be approached using:

• Qualitative system mapping (5.1.2) can be used to explore system boundaries and identify main actors and entities within the system. The mapping process can also reveal more about the main resource dependencies, assets and relationships associated with each actor or entity. Throughout the mapping process it is important to be clear about scope and related scale and level issues (for example, space, time, jurisdiction).

Exploring and setting system boundaries helps to:

• Clarify purpose and values of different interest groups by following an iterative process.
• Establish a solution space for developing interventions and defining which actors or stakeholders are responsible for developing solutions.

The following points should be considered when establishing system boundaries:

• The relative influence of stakeholders’ views should be taken into consideration. Each stakeholder will have their own view of what to include and what to exclude from a system boundary. In analysing a system, it is therefore important to be explicit about whose perspectives are being incorporated and to carefully consider the influence of these views within the decision-making process.
• Drawing a system boundary should be an iterative process ^{[footnote 6]} of drawing boundaries and critiquing them. The boundary setting process should repeatedly ask questions of: “what would happen if I left out x?” wherein decisions about who will benefit from the system are constantly articulated.
• If a boundary is drawn too tightly, relevant factors may be omitted. If it is too wide, it may be impossible to map or hard to draw inferences.
• Once a boundary is drawn, it can be tested for robustness by making sure that relevant factors that drive system behaviour are not omitted or that the system in question is not too big to be useful.

Example: In a hypothetical exercise to determine fishing quotas a fishery system map might include fish populations and fishing effort within the boundary but selectively exclude the wider environmental factors of climate change and global commodity prices. These wider environmental factors still impact the system, but currently lie outside the chosen system boundary (as they don’t have an immediate effect on fishery quotas). However, if considering fishery dynamics over a longer time frame (introducing a new perspective), then the boundary can be expanded to include these elements.

3. Using systems in Defra

This section brings together the need for systems thinking and the knowledge of systems approaches gained in the previous sections to focus on how systems can be applied in within 3 Defra scenarios:

1. Framing policy conversations on cross-cutting topics.
2. Making sense of fragmented evidence.
3. Engaging external stakeholders to deliver cross-cutting, evidence-driven work.

The content within this section should be used to inspire ideas and conversations around how systems approaches could be adapted and applied to emerging and ongoing issues across the Defra network.

3.1 Planning a systemic approach to policy making

An individual systems project may employ a range of analytical and systems methods used in combination. These may include quantitative, qualitative, individual and group activities (Figure 3). When planning a systems approach, the systems principles (set out in Section 2) can be built into a process that is tailored to a specific policy question. A generic systemic process could include the activities set out in Figure 5.

A combination of workshops, mapping, evidence gathering and analysis can be built into a systems approach. Each step is adaptable and can be made applicable to any area of policy making.

The following paragraphs outline the benefits and considerations of applying a systems approach to 3 decision-making scenarios. The remainder of this section represents how a systems approach, reflective of the format in Figure 5, could be adapted to provide different outcomes.

Figure 5: System purpose, perspectives, boundaries and interrelationships can be explored in an evolutionary ‘stepwise’ manner through workshops, system mapping, evidence review and policy analysis.

There should be considerable overlap between these activities. Each step should loop back to refine insights gained through previous steps. The process should allow conclusions to emerge iteratively.

Stakeholder engagement is important throughout the process to foster co-design and co-ownership of the systems analysis.

Figure 5 outlines a linear order for conducting a systems approach:

1. workshops (to explore issues, boundaries, purpose, stakeholders and values)
2. system mapping (for example: material flows or socio-ecological links)
3. evidence review (to explore key interrelationships and feedback loops)
4. Policy analysis (for example using logic mapping to find intervention points)

Each stage links back to the previous steps to represent how each stage of the process can be used to strengthen and inform the others.

3.2 Systemic approaches for framing conversations within government

3.2.1 The need for systems thinking in policy

Many of the issues facing the Defra network cut across organisational structures and disciplines. Resolving cross-cutting issues requires coordination between multiple policy teams who may have synergistic or opposing goals and perspectives. Systems approaches can provide insights on key interactions between policy areas.

In order to develop well integrated and complementary policies, teams across government need to come together to identify cross-cutting challenges and explore the interactions between their priorities. Critical systems heuristics (5.1.7) and soft systems methodology (5.1.1) both provide useful tools to help bring teams together in exploring the boundaries around a problem and in exposing diverging interests.

Example: Land use is an example of an area that cuts across established organisational structures and scientific disciplines. This issue requires coordination between food, agriculture, biodiversity, water supply, flooding, water quality and air quality policy areas. The issue also requires Defra to coordinate with other government departments working on transport, infrastructure, housing development and bioenergy.

Systems approaches can be used to frame and develop conversations between Defra policy teams and different government departments in order to identify cross-cutting challenges and find synergistic solutions.

3.2.2 Potential roles for systems approaches in policy framing

Systems approaches can be used to frame cross-team conversations within government to:

• Unpack complex situations. providing new insights.
• Identify potential unintended consequences of policies. Applying systems thinking moves participants away from managing issues in isolation to considering the connectedness between them.
• Identify interdependencies between policies. A systems approach can help Defra and cross-government teams to understand the interrelationships between policy areas and develop strategic plans for intervention.
• Manage conflicts. Systems approaches can expose how different groups and stakeholders view a certain policy. By revealing the different goals, perspectives and boundaries of policy teams it is possible to develop a shared understanding of the problem and reveal common values to work from.
• Identify areas for policy intervention. A systems approach can help decision makers to understand the dominant processes and linkages in a system and possible leverage points for policy intervention.

3.2.3 Considerations for systemic policy framing

Systemic approaches to cross cutting policy issues should:

• Consider the perspectives of multiple actors. This needs to go beyond the usual suspects and consider actors who may be indirectly affected by a policy decision. Actor mapping can help to expand understanding of who is indirectly affected by a policy intervention, including marginalised or fragmented groups who may lack a voice through lobbying.
• Build networks around complex areas. Cross-team working groups can be formed on a formal basis, but more flexible, informal interactions can be supported through workshops. Systems analysts can play an important role by acting as honest and mutual brokers to facilitate discussion between policy teams detached from government silos, departments and disciplines.
• Collaboratively agree on boundaries and map the system in question. Workshops involving multiple policy teams can be an effective way of exploring and mapping issues. They are useful for developing a shared conceptual understanding and framing of a situation (5.1.6).

Systems thinking is not solely the domain of analysts. It can be embedded in mainstream policymaking in Defra and across Whitehall.

3.3 Systems at the science-policy interface

3.3.1 Bridging world views

A key challenge at the science-policy interface arises from the differences in scope and timeframe between policy questions and research questions. Policymakers often need rapid answers to questions that are too broad to be easily answered through research. Research projects often address questions that are too narrow over time frames that are too long (for example, 1 to 5 years) to respond to rapidly evolving, complex policy needs. The contrast in language, framing and approaches of highly detailed scientific research and fast evolving policy can generate barriers between researchers and policy makers.

Despite an increased focus on applied/translational research in recent years, with incentives for academics to produce science with ‘impact’ (with clear economic or societal benefits), most research funding does not directly inform policy. This is especially problematic in the face of urgent environmental issues, such as climate change, that require rapid policy responses often with sparsely available evidence ^{[footnote 7]}.

Systems approaches such as participatory mapping can be applied to help bridge this gap, bringing researchers and policy makers together in a process of co-learning. For example, bringing disparate strands of research evidence together into coherent and robust policy narratives. Systems mapping can also be used to target new research, for example identifying key variables for data collection and interrelationships to explore.

Over recent years, ways of working at the science-policy interface have evolved significantly. A predominant unidirectional concept of ‘knowledge transfer’, whereby scientific advisers present evidence to policy makers, has evolved into multi-directional concepts of ‘knowledge exchange’ or ‘knowledge brokering’, whereby knowledge is jointly produced by researchers and stakeholders. Knowledge brokering to inform policy can include approaches from passive one-way communication to a full participatory joint-learning process ^{[footnote 8]}.

3.3.2 The role of systems approaches at the science policy interface

Applying systems methods at the science-policy interface can be used to:

• Prioritise new research and evidence synthesis. The use of systems mapping within a systems approach can draw out the key variables and relationships in a given policy situation. The mapping process can be used to ask the right questions of analysts and researchers, to develop understanding where knowledge is lacking and to help focus and prioritise research efforts.
• Bring together fragmented evidence into a coherent narrative. Systems maps and conceptual models can provide a framework to pull together and contextualise evidence from multiple disciplines. This can help to piece together a holistic picture of a policy situation and facilitate interdisciplinary working. A more structured way of filling gaps in a fragmented evidence-base is through expert elicitation as part of a systems approach.
• Interpret evidence from models of partial systems. Highlighting which elements of a system are included and excluded from models (by applying boundary setting principles) can help analysts to contextualise or challenge outputs and identify potential risks.

3.4 Systemic approaches for public engagement

The way in which stakeholders and the public are engaged in the process of policy making can greatly impact the outcome. Systems approaches can be applied to integrate public and stakeholder perceptions into policy development and the decision-making process at an early stage. For example, by incorporating participatory systems methods into citizen assemblies and public dialogues. Citizen assemblies are achieving growing recognition as a means to engage civil society in a participatory approach to policymaking (McGonigle et al, 2020 A knowledge brokering framework for integrated landscape management). Similarly, public dialogues have been recognised as a valuable approach for sharing different perspectives and finding workable solutions.

A citizen assembly is a group of people who are brought together to discuss a public issue, or issues and reach a conclusion about what they think should happen. The people who take part are chosen to reflect the diversity of the wider population. This is a form of collective decision making sometimes called deliberative democracy. Examples include: Oxford City Council’s citizen assemblies on their stated climate emergency and the National Citizens’ Assemblies on Climate Change.

A public dialogue is a collaborative process whereby participants work together towards a shared understanding of issues and perspectives. The goal of a public dialogue is to explore common ground and reveal differences between public perspectives in a safe, respectful exchange. Participants are invited to listen to understand and gain insight into the different perspectives, derived from different lived experiences. A dialogue assumes that many people have pieces of the answer to complex problem-solving and that together they can make them into a workable solution. Examples include Sciencewise and Structured Democratic Dialogue.

Applying systems approaches in stakeholder engagement and public dialogue When designing and facilitating citizens assemblies, public dialogues and stakeholder engagement, participatory systems methods can be incorporated to:

• Shape communication so as to uncover what matters to different people. Soft systems methodology (5.1.1), participatory system mapping (5.1.2) and conceptual modelling (5.1.8) can be applied to capture the values and motivations behind stakeholders. The insights acquired may enable better-informed judgements about what to include or exclude from policy.
• Capture and account for multiple perspectives. Soft systems methodologies (5.1.1) can be applied to facilitate the articulation of different voices. This can lead to the emergence of new solutions through synergistic, multi-value analyses.
• Uncover conflicts and structure negotiations. Critical systems methodologies (5.1.7) can help to bridge different perspectives and purposes. For example, by identifying lock-ins, barriers and areas of poor governance.

Potential benefits

Potential benefits of evidence-informed participatory approaches include:

• Highly visible public engagement to act on issues such as climate change.
• Lending legitimacy and public ownership to major decisions, for example, to achieve the net zero target.
• Promoting personal agency and pro-environmental behaviour changes through a co-creation process.
• Evidence gathering on public attitudes and on the public acceptability of different types of policy intervention, gaining better intelligence on what is politically possible.
• Crucial knowledge on multiple outcomes of policy interventions provided by the scientific evidence gathering process. This offers a new way of doing science to provide timely and credible evidence which is tailored to informing specific policy decisions.
• Assumptions on social values in scientific and economic models can be parameterised through public engagement.
• Engaging stakeholders and establishing collaborations around complex environmental issues, especially at local levels.

What to consider

When planning participatory systems methods for public engagement it is essential to consider:

• How to introduce scientific evidence to facilitate relevant discussions: Using scientific experts to explain key issues to participants at the start of citizen assemblies/public dialogues can be effective in some instances but problematic for extensive topics, such as UK land use transformation. Alternatively, iterative interactions with a scientific network enables policy options and the subsequent outcomes to be considered using the best evidence, together with modelling (Figure 6) and Section 5.1.6
• How to make best use of facilitators: the facilitator’s role is key in shaping emerging patterns of communication so that multiple voices and perspectives are recognised and tensions are brought into the open

Figure 6. Interacting with researchers, civil society and multiple levels of government enables an evidence-informed participatory approach to policy-making.

A key principle of any systems approach is to engage the main stakeholders within the design and the implementation of the analysis. This increases the chances that the systemic interventions identified will be practical, acceptable and co-owned.

This collaborative approach to policy making, reconciles national to local scales, (vertical policy coherence), and allows navigation of trade-offs and synergies (horizontal policy coherence).

4. Conclusions: embedding systems thinking at Defra

The systems field is broad and can seem daunting to engage with. While we think in systems in many aspects of our work in Defra, the systems field can provide formal approaches to help us navigate complexity. More importantly, it provides basic principles to help us ask good questions about our policies, focused on purposes, perspectives, boundaries and relationships.

Systems research is more than an analytical process, it provides a way of thinking and a language to unpick cross-cutting problems. Systems approaches do not always provide hard answers but can be invaluable in exploring issues and asking “what if?”. In systems, the journey is as important as the destination. Approaches should be people-centred, co-designed and co-owned with participants.

All this means is that it is important to consider the motivation of stakeholders (including policy teams) when deciding who to engage in policy decisions. Those developing systems approaches need to invest in relationship-building and consider how to align their work with participants’ personal objectives. Getting senior stakeholder champions on board early can help secure engagement in systems projects. Engaging people in small groups can lead to better conversations. Throughout, it is important to be open to different ways of thinking about the situation of interest.

Systems approaches are at their best when they are flexible, mixing methods from across the systems field and beyond. They should be adaptive to the changing policy agenda and to unexpected insights that emerge. A process of continuous self-reflection can help avoid groupthink.

Truly embedding systemic ways of working in Defra has broader implications for the way that policy is developed and implemented, particularly in the way government engages stakeholders and the public in decision making. Engaging interested parties in a participatory, systemic process can help balance the conflicting goals of different stakeholder groups particularly when the most desirable mix of outcomes is ambiguous or contested. Systemic policy design might involve a more iterative, ‘experimental’ approach where policies are piloted with on-going monitoring and evaluation to underpin learning.

Finally, systems thinking should not solely be the domain of experts. Basic systems literacy should be encouraged across the department at all levels. Learning by doing is the best way to build understanding of systems principles.

Annex 1: Systems tools and approaches

5.1.1 Soft Systems Methodology (SSM)

SSM is an umbrella term for a range of tools that were developed by Peter Checkland and Brian Wilson to facilitate understanding between stakeholders with different perspectives of what constitutes the system in question and therefore what the appropriate solution/intervention should be. This is a common issue in systems thinking as each stakeholder has their own perception on the same situation/system. In some cases this can become a problem for finding a solution because different stakeholders may not be able to agree on problem in question, or the best course of action to remedy it. SSM can therefore be used to aid participants to reflect on their own perceptions and how these relate to others.

SSM is contrasted to harder systems approaches by focussing more on the ‘human’ aspects of the system, such as understanding cultural or political factors that influence the outcome of a system enquiry, rather than using data or models to simplify a complex process or set of relationships. The outcomes of SSM will likely give you a common vision or objective, or a project plan, in which the stakeholders involved are working cooperatively to achieve the best possible outcome for all parties involved. It is therefore often most effectively used as part of a mixed methods approach. The 4 most common tools used are:

• rich pictures to capture people’s understanding of core outcomes and interrelationships.
• conceptual models
• CATWOE (Customers, Actors, Transformation process, Worldview, Owners and Environmental constraints)
• formal system models

It is not essential to use all 4 of these steps. However, in combination, they help to move a group of stakeholders with divergent objectives towards a shared understanding of a system and collaboratively identify changes that need to be made. These tools are intended to counteract the effect of power dynamics which might be at play in a multi-stakeholder environment and might influence decision making.

5.1.2 System mapping

System mapping can be useful to explore interactions or interdependencies between the different elements of a system. If done well, maps can provide powerful tools for communicating and exploring complex or non-linear relationships. Mapping can be done individually or as part of a participatory process, to synthesise partial knowledge from many people into a wider picture, or as a joint learning process.

A system map can be drawn with ease by hand or using various open source and commercially licensed software packages. Depending on the data input, software used and purpose of enquiry, system maps range from simple visualisations to complex and dynamic models used to test assumptions and interventions. A system map represents the perspectives that inform its development. As such, each map of a given system will be different depending on whose views are included or excluded.

Maps of systems can range in complexity. In its most basic form, a system map, rich picture, or mind map can be used to show key elements of a system and how they group together in sub-systems. This can be very useful in exploring the boundaries as conceived by the stakeholders in the mapping process. Maps can be used to show different perspectives and values and, crucially, to identify and illustrate interlinkages within a system. Influence diagrams show basic interrelationships (where one component influences another), whereas causal loop diagrams draw out the direction of influence, usually including feedback loops. Evidence on the strength and confidence of relationships can be collated in associated databases. To understand the effects of these relationships on key outcome metrics, system dynamics models can be developed (view examples from the German Environment Agency). These require numeric parameterisation of the direction and strength of each link. The quality of the overall prediction depends on how well the system diagram captures key processes, as well as the accuracy by which each process (each link) is parameterised. In practice, system dynamics models often produce complex emergent properties, but are difficult to validate and so are more useful to explore different ways a system can respond, rather than anticipate how a system will respond.

5.1.3 Integrated spatial modelling

Various modelling platforms have been developed to explore trade-offs between different environmental outcomes. Each of these integrates models capable of quantifying various elements of natural capital and ecosystem services (for example, water quality, recreational value, biodiversity and pollination services) while some also include valuation of derived-benefits. A review of some of these modelling platforms can be found on the Ecosystems Knowledge Network website and in a recent JNCC report.

The main roles of these models in Defra include:

• Target setting: to help to set targets or normative scenarios for policy. For example, establishing how much conversion of farmland to woodland is needed to meet the net zero C target without adversely affecting other 25 YEP outcomes, food production or the economic implications for the farming sector.
• Exploring management interventions: to predict the likely effect of changes in the way in which land is used and managed on a wide range of different environmental, economic and social outcomes.
• Policy levers: Socio-economic modelling can help predict the effect of government interventions (for example, regulation, incentives, guidance or advice) and contextual factors (for example, market forces) on stakeholder behaviour (for example, land-use and farming practice).
• Monitoring and evaluation: In addition to the ex-ante evaluation of policy levers described above, modelling is needed to support the interpretation of data in ex-post evaluation. For example, modelling can help to attribute the causality of observed environmental changes or to quantify cumulative impacts of multiple local interventions.

Integrated spatial modelling to understand the multiple outcomes of land use decisions is likely to be enhanced through a well-functioning science-policy interface, allowing government to access the newest scientific developments in land use science. This includes: i) state-of-the-art quantitative modelling of natural capital and ecosystem services, including statistical, process-based and agent-based models; ii) coupling of models with big data (environmental, social and economic) to represent interacting processes within landscape systems, iii) computer science innovation to design digital infrastructure for decision support, including standardised open-access model coding protocols and dynamic peer review, environmental dataset integration and accessibility, visualisation of trade-offs and synergies, and uncertainty mapping.

More integration across modelling initiatives is required to combat the issues of fragmentation across research groups, paradigm differences between many natural and social science disciplines (as well as within some disciplines), varying degrees of reliability between different models and the separation from real-world decision making processes. This is important because models are influenced by a broad network of actors, who will not always making decisions based on the best available evidence or effectively balancing a wide range of societal interests.

Such integrated spatial modelling approaches (‘hard-systems’ science) complement ‘soft-systems’ approaches, which aim to develop a functional community of practice between key stakeholders. It is particularly relevant to explore place-based approaches to decision making that reconcile public engagement with local and national governance (See Section 3.3).

5.1.4 Evidence synthesis

Evidence Reviews, in their various forms, are used to systematically search for, summarise and assess evidence to answer specific questions. While not a systems approach, they can be used alongside systems approaches to test the logic and evidence between particular interrelationships.

There exists a spectrum of Evidence Reviews that range in detail and rigour from Literature Reviews to Systematic Reviews. Rapid evidence synthesis (RES) lies in between these two ends of the spectrum and provides a good compromise between speed and rigour. They can most readily be used to understand the impact either of a ‘pressure’ or a policy intervention.

Rapid evidence synthesis can also be applied to more open-ended questions. It provides an understanding of the volume and characteristics of evidence available on a certain topic and makes it more accessible for further scrutiny if required. Hence, RES allows questions to be answered by maximising use of the existing evidence base, whilst also providing a clear picture of the adequacy of that evidence.

5.1.5 Expert elicitation

Expert elicitation is a process that seeks to gather and combine expert judgement on a topic in which there is a high degree of uncertainty, or where information is missing. It may be used to support or enable decision making where there are significant evidence gaps or insufficient statistical or modelling tools. It is particularly pertinent to environmental policy-making due to the high level of uncertainty that exists around future environmental, technological, social and political changes and helps to highlight trade-offs and synergies in policy designed to achieve multiple outcomes. It is useful for honing priorities in a project and refining thinking about the future, especially when operating under time pressure.

It is typically carried out with a group of experts with a diverse range of knowledge and opinions (maximising ‘cognitive diversity’). It can be carried out in person or remotely, meaning that experts from all over the world may be engaged in the process. An example is the Delphi process, whereby initial independent ideas from experts can be refined after group discussion. For example, an expert elicitation approach was recently used to assess potential trade-offs between greenhouse gas emissions and air quality to inform policy interventions recommended by the Committee on Climate Change.

5.1.6 Convening approaches

Systems approaches can be applied in a participatory way. This can be helpful for:

• Capturing multiple perspectives and opening up discussions.
• Bringing in knowledge or views from groups that are often marginalised or left out.
• Convening actors to bridge divergent interests or views.
• Piecing together fragmented knowledge to develop a more integrated picture (for example, by bringing together experts from different disciplines).
• Getting buy-in for decisions.

Participatory approaches can involve small groups coming together to draw system maps (for example, participatory system mapping and participatory modelling. In some cases, groups with divergent views or interests can be deliberately brought together to find common ground (for example, using negotiation games). Exercises can look backwards to develop a learning history (mapping timelines of past events to reflect on system dynamics) or forwards. Back-casting exercises bring system actors together to articulate long term goals and work back to the present, establishing what scenarios can lead to the goals articulated, identifying potential interventions, synergies and trade-offs. Participatory scenario development (for example, Three Horizons workshops) are used to explore alternative futures and encourage out-of-the-box thinking. Larger-scale public engagement and participatory decision-making processes include citizen assemblies and public dialogue, which can be framed around systems concepts (see Section 3.3).

5.1.7 Critical systems boundary questions

The following questions adapted from critical systems heuristics (from Ulrich 2000) can be useful in exploring system boundaries:

Purpose (motivation)

• Who are (should be) the main stakeholders benefited or impacted?
• What are (should be) their purposes?
• How do (should) we measure improvement in these purposes?

Control (power)

• Who are (should be) the decision makers?
• What resources do (should) they control?
• What resources are (should be) outside the core system (part of the broader environment)?

Expertise (knowledge)

• Who are (should be) the experts?
• What kind of expertise is needed (should be drawn on)?
• How can (should) improvement be guaranteed?

Legitimacy

• Who acts (could act) as witnesses for stakeholders who can’t represent themselves?
• What protects (should protect) the interests of those affected?
• Which worldviews are (should be) determining?

5.1.8 Conceptual models

A particularly useful approach can be the co-development of conceptual models, bringing together multiple policy teams to agree on the purpose of enquiry, boundaries and key interrelationships within a policy area. Figure 7 provides an example of a food system conceptual model. Development of a conceptual model through conversations with policy teams and academics can help frame policy discussions and communicate key areas of conflict or trade-off. A conceptual model can help focus questions, identify who needs to be involved in discussions, the scope or boundaries of policy discussions and what the overall purpose of the policy area is. It gives a point of reference against which policy teams can check whether they have considered everything of relevance to a given decision.

Figure 7. An example of a food system conceptual model. This helps set system boundaries showing key drivers, outcomes and system components for consideration in policy decision making

Contributors and acknowledgements

This document has been produced by the Defra Systems Research Programme in the Chief Scientific Adviser’s Office.

Contributors in alphabetical order:

Pam Berry, Frank Boons, Bob Doherty, Bethany Green, Amy Hill, Nick MacInnes, Dan McGonigle, Abigail McQuatters-Gollop, Sarah Moller, Maria Munoz, Tom Oliver, Gary Preece, Victoria Robinson, Yaad Sidhu, Adam Vaughan, Rose Willoughby.

Acknowledgements:

The authors gratefully acknowledge the extensive comments and inputs from Bob Williams, Gerald Midgley (University of Hull), Lorenzo Benini (European Environment Agency), Nigel Gilbert (CECAN) and Ian Boyd (University of St. Andrews).

Suggested citation:

McGonigle, D.F., Berry, P., Boons, F.,… (2021). A Primer for Integrating Systems Approaches into Defra. Report from the Defra Systems Research Programme.

For further information, please contact: systems.research@defra.gov.uk

1. A conceptual classification of different types of problems (adapted from the ‘cynefin’ framework, Snowden, 1999) ↩

2. Rittel H. W.J., & Webber, M.M. (1973) Dilemmas in a general theory of planning. Policy Sciences, 4(2), 155-169. ↩

3. Open University ↩

4. Open University; Cabrera D and Cabrera L (2015). Systems Thinking Made Simple: New Hope for Addressing Wicked Problems. Ithaca NY: Odyssean; Cabrera D, Cabrera L, Powers E (2015). A unifying theory of systems thinking with psychosocial applications. Systems Research and Behavioral Science 32(5): 534–545. ↩

5. Open University Systems Thinking in Practice; Reynolds, Martin (2011). Critical thinking and systems thinking: towards a critical literacy for systems thinking in practice. In: Horvath, Christopher P. and Forte, James M. eds. Critical Thinking. New York, USA: Nova Science Publishers, pp. 37–68. ↩

6. Progressive contextualization (Vayda, 1983) is one useful approach to guide successive iterations. ↩

7. McGonigle et al, 2020 A knowledge brokering framework for integrated landscape management ↩

8. Gastil, J. and R. Richards. 2013. Making Direct Democracy Deliberative through Random Assemblies. Politics & Society 41:253-281. ↩