Leverage points are places within a complex system where a small shift in one thing can produce big changes in everything else. Meadows's 1999 paper leverage-points-paper-1999 — adapted from a 1997 publication places-to-intervene-in-a-system-1997 — is her most cited single contribution, presenting a hierarchy of twelve places to intervene in a system, ordered from least to most powerful.
The list emerged from frustration. Meadows had spent decades watching policy debates focus relentlessly on the least effective leverage points — numbers, parameters, subsidies and taxes — while ignoring the structural and paradigm-level interventions that actually change system behavior. Her hierarchy was an attempt to redirect attention.
The twelve leverage points, from least to most powerful:
12. Numbers (constants and parameters) — changing a tax rate, a subsidy level, a flow rate. Almost never changes behavior in a significant way unless the system is near a threshold.
11. Sizes of stocks and buffers — buffers stabilize systems; large buffers relative to flows produce stable behavior. Changing buffer sizes is often physically constrained and expensive.
10. Structure of material flows — the physical arrangement of roads, pipes, factories. Slow and expensive to change, but important. Subject to path dependency.
9. Lengths of delays — relative to the rate of system change. Shorter delays in feedback loops generally improve system performance, but delays in fast loops can cause oscillation.
8. Strength of negative feedback loops — how strongly a balancing-feedback-loops corrects deviations. Regulation, oversight, enforcement.
7. Gain around positive feedback loops — how fast a reinforcing-feedback-loops drives growth. Taxing the gains from reinforcing loops (e.g., compound interest, monopoly) or removing the perverse incentives that drive them.
6. Structure of information flows — who has access to what information when. Meadows considered this highly underappreciated: many problematic behaviors persist because the people in the loop causing the behavior do not experience its consequences. Fixing this does not require new physical infrastructure, just new information connections.
5. Rules of the system — incentives, constraints, and laws. Constitutions, regulations, property rights. Higher leverage than parameters; lower than the following items.
4. Power to add, change, or self-organize the system structure — self-organization, the capacity of a system to create new feedback loops, new structures, new entities. Highly leveraged because it generates all the system-level behaviors above it in the hierarchy.
3. Goals of the system — the purpose or function of the system, the reference level of key balancing-feedback-loops. If a system's goal changes, all the lower-level parameters and rules change accordingly.
2. Mindset or paradigm out of which the system arises — the shared ideas, assumptions, values, and world views from which the system's goals, rules, and information flows arise. Meadows argued that paradigms are the source from which systems grow; change the paradigm and the system changes.
1. Power to transcend paradigms — staying flexible, recognizing that no paradigm is the final truth, maintaining the capacity to question all paradigms including the one you currently inhabit. The highest leverage of all, and the rarest.
Meadows was clear that higher leverage does not mean easier. Changing paradigms is the most powerful intervention and the most resisted. People will fight to the death to defend their paradigms. Systems thinkers are frequently defeated at the institutional and political level precisely because their interventions threaten the paradigm-level assumptions that the system's defenders take as given.
The paper ends with a characteristic Meadows move: having climbed to the top of the hierarchy, she recommends a kind of intellectual humility — the ability to use paradigms as tools while knowing they are tools, not truths. This connects leverage-points to her broader teaching about dancing-with-systems.
The hierarchy has been widely adopted and widely misapplied — often reproduced as a list without the relational logic that gives the ordering its meaning. Meadows herself noted that the list was not perfect and that the items interact; it was a teaching device, not a formula.