Causal Loop Diagram¶

Causal-structure model — instantiates Migration-Resistant Hazard Control

Draws the pressure behind a hazard, the feedback loops that regenerate it, and the delays between them, so a control can be aimed at the loop rather than the symptom it displaces.

A Causal Loop Diagram maps the variables that drive a hazard and the directed, signed links between them, marking where reinforcing and balancing loops close and where time delays sit. Its distinguishing move in this archetype is that it treats the pressure behind a hazard as endogenously generated — produced and sustained by a loop inside the system — so it can show why a barrier that blocks one outlet simply reroutes the same pressure to another. Where a fault tree traces a single failure downward and a mass balance counts what crosses a boundary, the loop diagram reveals the generating structure: the reinforcing loop that keeps making the pressure, the field of places that pressure can flow to, and the balancing loop that could dissipate it. That is what lets a designer aim at the loop rather than at whichever symptom is visible today.

Example¶

A city introduces a downtown congestion charge to cut gridlock. Before anyone writes the bylaw, a small group sketches a causal loop diagram of why the centre is congested at all. The core is a reinforcing loop: easy driving raises traffic, which raises pressure to add road capacity, which the city has historically supplied — making driving easier again. The charge is a balancing intervention on one node (downtown trips), but the diagram makes visible what it does not touch: the underlying demand to travel at peak. With delays marked, the group can read off the likely reroute — trips shifting to the ring road, to just before the charging window opens, or to the untolled neighbouring boroughs — because the generating pressure (people needing to move at 8 a.m.) is untouched. The same diagram surfaces a balancing loop that would absorb the pressure rather than move it: off-peak transit capacity that drains peak demand. The output is not a decision but a structural hypothesis — charge the cordon and the trips migrate; feed the absorption loop and the pressure actually falls.

How it works¶

Name the pressure as a variable, not an event. The hazard is treated as the current outlet of a persistent driver; the first job is to find and name that driver.
Close the loops. Trace links until they return on themselves. A reinforcing loop that regenerates the pressure is exactly what a one-shot barrier cannot stop.
Mark the delays. Displacement usually travels down a lagged link, so an unmarked delay is where a "successful" control quietly leaks.
Read off the field and the absorption loop. Every node the pressure can reach is a candidate destination; every balancing loop that drains it is a candidate real fix.

Tuning parameters¶

Diagram boundary — how many variables and loops you admit. Draw it too tight and the destination sits just off-page (the very failure this archetype warns of); too wide and it becomes unreadable.
Aggregation level — one "demand" node versus a dozen segmented ones. Finer nodes expose specific reroutes; coarser ones keep the generating loop legible.
Delay marking — whether and how precisely lags are annotated. Delays are where displacement hides, so under-marking them flatters a fast-looking control.
Qualitative vs. quantified — a signed sketch versus a stock-and-flow model with rates. Quantifying tests which loop dominates but costs far more and invites false precision.
Loop-tracing depth — how many feedback loops you follow before stopping. More loops catch indirect reroutes; fewer keep the leverage story sharp.

When it helps, and when it misleads¶

Its strength is that it is the one tool here that explains why a barrier displaces rather than eliminates — because it exposes the loop that keeps regenerating the pressure — and in the same picture it points at the balancing loop, the absorption path, that would actually drain it. That reframes the design question from "how do we block harder" to "which loop do we change."

Its central weakness is that a causal loop diagram is a hypothesis, not evidence. Polarities, and especially which loop dominates, are easy to assert and hard to prove, so a confident, elegant diagram can encode the author's assumptions as if they were findings. Its classic misuse is to be drawn after an intervention is chosen, with the arrows arranged to lead to the pet fix. The discipline that guards against this is to build it with people who hold different mental models, to label which links are evidenced versus assumed, and to treat every loop as a claim to be tested — especially against the Shifting the Burden pattern, where the symptomatic fix is precisely the one the diagram should make suspect.^[1]

How it implements the components¶

generative_hazard_pressure_model — the reinforcing loop the diagram centres on is the model of the pressure that generates the hazard.
migration_field_model — every node reachable from that pressure along a signed link is a place the hazard can re-express itself; the link topology is the migration field.
pressure_absorption_path — the balancing loop the diagram surfaces is the path along which the pressure could be safely dissipated instead of displaced.

It does not count what crosses a boundary — that conservation check is Mass Balance — nor does it measure whether displacement actually happened, which is Before–After–Elsewhere Evaluation.

Instantiates: Migration-Resistant Hazard Control — the diagram supplies the generating-structure hypothesis the other controls are aimed with.
Sibling mechanisms: Fault Tree Analysis · Whole-System Impact Map · Mass Balance · Hazard Analysis · Adaptive Circumvention Red Team · Before–After–Elsewhere Evaluation · Agent-Based Experiment or Simulation · System-Wide Net-Risk Dashboard · Boundary Expansion Review · Cross-Boundary Hazard Ledger · Cross-Jurisdiction Incident Review · Intervention Displacement Stress Test · Migration Sentinel Network · Pressure-Absorption Redesign Workshop · Source-Reduction or Safe-Dissipation Plan

Notes¶

A causal loop diagram earns its place early — it is a framing device that tells the other mechanisms where to look: which destinations to monitor, which boundary to draw, which absorption loop to fund. But it proves nothing on its own. Pair it with a mechanism that carries evidence — a mass balance, an evaluation, a simulation — before betting a control on the loop it proposes.

References¶

[1] Shifting the Burden is one of the classic system-dynamics archetypes: a symptomatic fix relieves pressure in the short run but lets the underlying problem — and the capacity to address it — worsen, so reliance on the fix deepens. A causal loop diagram is the standard way to make that structure visible. ↩