Why we keep “fixing” things and nothing changes, a systems view from Donella Meadows

I keep noticing this pattern where public debates get incredibly intense about the number. The budget amount, the tax rate, the cap, the threshold, the percentage change. It feels like the most real part of the argument because it is concrete, measurable, and easy to turn into a headline. But even when the number changes, the lived outcome often barely shifts, or it shifts briefly and then returns to the same familiar shape. That does not mean nothing matters. It means we keep grabbing the weakest levers in the system and then expecting a structural outcome.

Donella Meadows describes leverage points as the places you can intervene in a system, ranked from low leverage to high leverage. The low leverage stuff is the surface layer. Parameters and numbers. How much funding, what the cap is, how high the rate goes. Buffers and slack also live down here, like reserve capacity, inventory, savings, extra staffing, and housing supply that prevents a system from becoming fragile and panicky. Delays matter here too, because if feedback arrives late and conditions change quickly, the system can oscillate and overcorrect, even if everyone involved is acting rationally with the information they have. These levers can matter, but they usually do not change the system’s personality. They change intensity, timing, or distribution, while the underlying pattern stays.

The system’s personality tends to come from higher leverage layers. One layer is structure, meaning the layout and pathways that determine what can flow where, what bottlenecks exist, and what options are realistic. Another layer is feedback loops, meaning whether the system can notice harm and correct itself. That includes whether oversight has real power, whether enforcement is consistent, and whether corrective mechanisms have teeth instead of being decorative. Then there is information flow, which is quietly one of the biggest levers because you can change outcomes dramatically without changing resources, simply by changing who can see what, when they can see it, and whether the signal is trustworthy. When consequences are hidden or delayed, systems do not self correct. When consequences are visible and timely, behavior often shifts faster than people expect.

Even higher leverage than information are rules and incentives. Rules decide what behavior is rewarded, what behavior is punished, what is easy, what is costly, and what is ignored. Over time, rules do not just shape behavior. They select which kinds of actors thrive inside the system. If the system rewards short term extraction, you will get more extractors. If it rewards long term trust and outcomes, you will get more builders. This is why changing the rules can reshape the entire character of a sector or institution without needing constant moral lectures. People adapt to the selection pressure.

Above rules sits the part people avoid talking about, goals. What the system is optimized to achieve in practice, not what it claims in slogans. If the real goal is maximize profit, maximize attention, avoid blame, or win elections, the system will sacrifice truth, long term health, and stability whenever those things interfere with the goal. That is the brutal part. You cannot patch over a bad goal with good intentions, because the system keeps trading away anything that is not the goal in order to hit the goal. When you change the goal, you change what success means. That is why it is such a high leverage move compared to tweaking a rate or increasing a budget line.

All of this sits inside paradigms, which are the background beliefs people do not realize they are using. Beliefs like growth is always good, markets are neutral, people are mostly selfish, regulation is always oppression, or the only thing that matters is winning. Paradigms matter because they create the goals and they decide which policies even feel legitimate. When a paradigm shifts, the system can transform quickly, because what used to feel impossible suddenly becomes obvious. And the highest leverage is being able to step outside any single paradigm and treat worldviews like tools rather than identities, because that keeps you from becoming trapped in the system’s script.

A practical test I keep using is this. When a proposed fix is mostly a number change, I try to translate it upward. I ask what incentives stay the same, what feedback loops remain weak, what information is still hidden, and what goal is still driving behavior. If you cannot answer those, then you are not redesigning the system, you are adjusting settings on a machine that will keep pursuing the same objective. That is why politics can feel like we change the numbers and nothing changes. Systems do not mainly run on numbers. They run on incentives, feedback, information, and goals. If we want different outcomes, we have to get comfortable arguing about the layers that actually produce outcomes.