Causal Loop Diagram (CLD)

A causal loop diagram (CLD) explains the behavior of a system by showing a collection of connected nodes and the feedback loops created by the connections. One or more of the nodes represent the symptoms of the problem. The rest of the nodes are the causal chains causing the problem.

The simplest possible CLD contains two nodes. Below is an example from video 3 in The Dueling Loops Video Series. The R signifies this is a reinforcing loop. The graph shows the loop's general behavior by showing the symptom of interest, population growth.

Population Growth CLD

The insight is that problem symptoms are caused by certain key feedback loops. It follows that to solve a problem you must first model its key feedback loops.

Now take a look at the same causal loop diagram transferred to a system dynamics simulation model.

Introduction to causal loop diagramming

All you need to draw CLDs is pencil and paper or a whiteboard. Below is an actual sketch of a CLD from my notebook on October 20, 2004. On the next notebook page this became The Race to the Bottom loop in the first version of the Dueling Loops model. (A solid arrow is a direct relationship. A dashed arrow is an inverse relationship.)

Notebook sketch of first Race to the Bottom

Let's walk the reinforcing loop, denoted by an “R.” As fallaciousness of arguments goes up, so does false attacks and false promises. As these increase, so does supporters due to falsehood. That in turn causes an increase in political power due to falsehood. Since the strategy of mass deception worked so well, that in turn increases the next round of fallaciousness of arguments and the loop starts all over again. Loop growth creates a social force. This loop increases the force of mass deception in the political powerplace. Note this force does not arise from individual nodes. It comes from the feedback loop. That is the key insight offered by causal loop diagramming. The tool allows you to find where the fundamental forces causing a problem are coming from.

Here's how causal loop diagramming works. A CLD consists of nodes, arrows, and feedback loops. The nodes on this one are truth ratings, detected deception, false attacks, false promises, and so on. A node has a name and a value and represents something in the real world. What matters is not the exact value, but the value relative to what it was earlier or in relation to other nodes.

There are two types of arrows for showing how one node influences another. A solid arrow is a direct relationship, where the value of node A varies directly with the value of node B. A dashed arrow indicates an inverse relationship, where as the value of A goes up the value of B goes down.

There are two types of feedback loops: reinforcing and balancing. In a reinforcing feedback loop a change in a node goes around the loop to cause a change in that same node in the same direction, which causes the loop to grow or decline. In a balancing feedback loop the change is in the opposite direction, which causes the loop to balance its behavior as it seeks a goal of some kind.

Below is probably the most famous CLD in existence. It's the high level diagram of how the World3 model of The Limits to Growth works. Figure 1-3 is from Dynamics of Growth in a Finite World, by Meadows at al, 1974, page 14. This larger book (637 pages) fully describes the World3 model in lots of technical detail.

Causal loop diagram of World3 model

In this diagram style a "+" indicates a direct relationship and a "-" is an indirect relationship.

There are no Rs or Bs denoting the many reinforcing and balancing feedback loops, except for one. The lower left loop has a "(+)" in it. That's a positive or reinforcing loop. The book emphasized this was the key loop pushing the human system beyond the limits of growth, and thus toward eventual collapse unless the sustainability problem was proactively solved.

That loop is the Industrial Growth loop. An increase in industrial investment causes growth in industrial capital. Growth in that causes growth in industrial output. Most of that goes towards industrial output per capita. But some is reinvested, which causes further growth in industrial investment, and the loop goes through another cycle of growth.

The environmental sustainability problem arises from the way growth in industrial output causes more pollution and decreases nonrenewable resources (as well as decreasing renewable resources, not shown). These effects have long delays so there is no reason to solve the problem now, in the minds of most people. It is only advanced minds, those educated by models like this, that can see the future before it happens and decide to commit to solving the problem proactively.

Wouldn't it be nice if a majority of the world's population could grasp the truth of this model and decide to aggressively solve the sustainability problem?

That is currently impossible due to high systemic change resistance.

Browse the Glossary
Previous Next
The Next Step After Causal Loop Diagramming

is simulation modeling. The type of simulation modeling that seems to fit the sustainability problem and our problem solving process the best is system dynamics.

The Thinker's Toolkit: 14 Powerful Techniques for Problem Solving

Thinker's Toolkit coverThis is a classic problem solving book by Morgan Jones, former CIA analyst. The book contains the author's top 14 techniques for solving difficult problems. One technique is causal loop diagramming, which the book calls causal flow diagramming. The chapter on that says:

“All events in life are the results, the outcomes, of previous events. Life itself is a collage of ecological cycles, all intertwining, mingling, ever-evolving chains of cause-and-effect relationships. Likewise, every problem we analyze is the product of a definable cause-and-effect system.

“It stands to reason, then, that identifying that system's components—the major factors—and how they interact to produce the problem is essential to effective problem solving.

The chapter then lists “The Five Steps of Causal Flow Diagramming:

Step 1. Identify the major factors

Step 2. Identify cause-and-effect relationships

Step 3. Characterize the relationships as direct or indirect

Step 4. Diagram the relationships

Step 5. Analyze the behavior of the relationships as an integrated system.”