Overview
Donella Meadows was an environmental scientist and a lead author of The Limits to Growth. This primer was published in 2008, after her death, from a draft she had circulated for years. It explains how to see a problem as a system of stocks, flows and feedback loops. She argues that hunger, pollution and economic cycles keep recurring because of how systems are built, and that blaming individuals misses this.
In brief
Stocks and flows
A stock is an amount that builds up, such as water in a bath or money in an account. Flows fill and drain it. Stocks change slowly, which is why systems respond with delays.
Feedback loops
A balancing loop pushes a stock back towards a target, as a thermostat does. A reinforcing loop feeds on itself, like interest on savings. Meadows explains system behaviour by which loop dominates.
Where to intervene
Meadows ranks the places where a change can shift a whole system. Adjusting numbers such as tax rates ranks low. Changing the system's goals, or the beliefs behind them, ranks highest.
Unintended results
Fixes often fail because the system pushes back. Meadows describes common traps, such as policy resistance and the overuse of shared resources, and explains how each can be escaped.
Who should read this
People who keep fixing the same problem and watching it come back, whether in a business, a household budget or a public policy.
Who might skip it
Skip if you want the hard-maths version of systems dynamics — Meadows is deliberately qualitative, with diagrams rather than differential equations. Skip also if you're allergic to ecological and sustainability examples; she draws heavily on them, reflecting her own research background.
Stocks, flows and the bathtub
A system, in Meadows's definition, is a set of elements connected in a way that produces a pattern of behaviour over time. A football team, a school and your digestive system all count. Sand scattered on a road does not, because nothing connects the grains.
Her basic tool is the stock and the flow. A stock is anything that accumulates, such as water in a bathtub, money in a bank account or trees in a forest. Flows are what fill or drain it. You can raise the level of the tub either by opening the tap or by partly closing the drain. People tend to look only at inflows. A shopkeeper worried about cash may think only of increasing sales when reducing waste would do as much.
Stocks change slowly, even when flows change suddenly. This slowness gives systems stability, and it also means that problems build up quietly before anyone notices.
Two kinds of feedback
Systems behave as they do because of feedback loops. A loop exists when the level of a stock affects the flows into or out of that same stock.
A balancing loop pushes a system towards a goal and resists change. A thermostat is the usual example. When the room gets cold the heater switches on, and when it is warm enough the heater switches off. Balancing loops are why many systems are stable, and also why they resist reform.
A reinforcing loop amplifies whatever is happening. Money in a savings account earns a return, which increases the balance, which earns a larger return. Such loops never run for ever, because some balancing loop eventually limits them.
Meadows asks readers to stop looking for a single cause and to ask which loops are dominant at the moment.
Why delays cause swings
Information in a system usually arrives late, and responses take time to work. Meadows shows that these delays are a main reason systems swing back and forth instead of settling down.
Her example is a car dealer managing the stock of cars on her lot. She orders more when sales rise. But she needs some time to be sure the rise is real, and the factory takes more time to deliver. By the time the extra cars arrive she has ordered too many, so she cuts back too far, and the stock swings up and down. Oddly, reacting faster makes the swings worse. Reacting more slowly and steadily calms them.
Anyone who has adjusted a shower in an old building knows the feeling. You turn the tap, nothing happens, you turn it more, and then the water comes out scalding.
Traps that catch well-meaning people
Some system structures produce trouble so often that Meadows gives them names and calls them traps. They are caused by the structure, so blaming the individuals caught in them changes nothing.
In the tragedy of the commons, each user of a shared resource gains from taking a little more, while the cost is spread across everyone. Groundwater is pumped until wells run dry. The way out is to regulate access, or to make sure users feel the consequences of their own use.
In shifting the burden, a quick fix relieves a symptom and weakens the system's own ability to cope, so that more of the fix is needed each time. A student who copies solved answers passes the homework and understands less each week.
Drift to low performance occurs when standards are set by recent results. Each disappointing result lowers the expectation, which allows the next result to be worse.
Where to intervene
The best known part of the book is a ranked list of places to intervene in a system, running from the least effective to the most.
At the bottom are numbers, such as tax rates and subsidies. Most public argument is about these, and Meadows thinks they rarely change behaviour much, because the structure stays the same. Higher up is the flow of information. She tells of a group of similar houses in the Netherlands where some electricity meters had been installed in the front hall and others in the basement. The households that could see their meters used noticeably less electricity.
Near the top are the rules of the system and its goals. At the very top is the paradigm, the set of shared assumptions from which the goals arise. Meadows admits that the higher you go, the more the system resists. She presents the list as tentative.
Working with systems you cannot control
Complex systems cannot be fully predicted or controlled, and Meadows ends by asking for a different attitude. Our models of the world are always simplified. People inside a system act sensibly on the limited information they have, and the combined result can still be something nobody wanted.
She describes resilience as the ability of a system to recover from a shock. Managers often sacrifice it for efficiency, for example by keeping no spare stock, and find out the cost when something goes wrong.
Her practical advice is modest. Watch how a system behaves over time before trying to change it. Make your assumptions explicit and let others challenge them. Pay attention to what matters even when it cannot be measured. Expect to be surprised, and treat errors as information about where your model was wrong.
Try this week
- Pick one recurring problem and draw it as a stock with its inflows and outflows on a single sheet of paper.
- For that problem, write down how long it takes between an action and its visible result, and wait that long before adjusting again.
- Track one stock in your own life for a week, such as your savings balance or hours of sleep, noting what adds to it and what drains it.
- Put one piece of useful information where you will see it daily, like your monthly spending total on the fridge door.
Our honest take
The book was put together after the author's death in 2001 from a draft written in the early 1990s, and many examples show their age. It teaches you to describe systems in words and diagrams but not to build or test a real model. The ranked list of intervention points is the author's judgement and has not been tested in any strict way.
"Remember, always, that everything you know, and everything everyone knows, is only a model."
— Donella H. Meadows, Thinking in Systems
If you liked this
The Limits to Growth for the seminal modelling work. For a harder-science follow-up, Non-Linear Dynamics and Chaos by Strogatz.
Questions readers ask
Is Thinking in Systems worth reading?
Yes, if your problems keep coming back and you suspect the cause is the setup rather than the people in it; the chapters on stocks, flows and feedback give you a way of seeing that is hard to unlearn. It is a short primer, not a modelling course, and some of the examples are from the early 1990s.
Who should read Thinking in Systems?
People who keep fixing the same problem and watching it come back, whether in a business, a household budget or a public policy.
How long does this summary take to read?
About 6 minutes. It covers the book's 6 main ideas, who it is for, what to try this week, and an honest view of its weak points.
Read it offline, or listen to it. This summary is in the free MyByteNest app for Android, with 267 others. No sign-up.
Get the app
Share your thoughts
Found a mistake, disagree with the verdict, or have a book to recommend? Email us at support@mybytenest.com — we read everything.