Summer Reading 2026

Industrial Dynamics

The summer reading series arrives at the wellspring: Jay Wright Forrester's 1961 book, the bedrock beneath circularity's systems thinking.

"Information systems owe their behavior to three characteristics — structure, delays, and amplification." Image generated by Midjourney.

The term “scientific management” is at once widely accepted and vaguely mysterious. The pervasive need for more data, for the creation and following of more metrics, the proliferation of colorful dashboards and experiments all point to the business world borrowing the methods of the physical sciences.

But where does this come from? What is the origin of this perspective? The subject of this Letter, the latest in our summer reading series, is thematically related to the others in two ways.

Industrial Dynamics and its author, Jay Wright Forrester, are cited by Donella Meadows as formative for her work. She studied under Forrester at MIT, and later cited Industrial Dynamics, the book and the discipline, as the basis for her work in systems theory. We read her Thinking in Systems last week.

But there is another connection. I have found Roger Martin’s application of Aristotle’s philosophy to strategy as a useful lens for probing what is missing in the advancement of circularity. He has referred to the single-minded application of scientific management techniques as “the pervasive analytical blunder of strategists.”

So this week our reading serves two purposes. We continue our exploration of the intellectual origins of circularity. We also examine the “blunder of strategists” charge against a capable defendant.

On the latter purpose, it feels fair to spoil the ending up front. It is the same conclusion we have drawn elsewhere in this series: scientific management is necessary. It is not sufficient. We know this because the people who built it told us so:

In his introductory guide to the circular economy, Walter Stahel lamented in his closing pages that the supply-side logic of circularity was not enough to create a longing for it.
Donella Meadows codified the analytical toolkit of circularity, then made the appeal to something beyond for its fulfillment: a future “envisioned and brought lovingly into being.”

In her TED talk, Ellen MacArthur puts the system itself at the center — not the production or consumption per se, but the operating system underneath. Systems thinking as circularity practices it was codified by Meadows. Meadows learned it inside the group Jay Forrester founded at MIT. Four steps back from that TED talk is a book from 1961, written by an engineer who spent the war building weapons.

Consider the mood in America in those days. Two years before Forrester published Industrial Dynamics, the Ford and Sloane Foundations decided America's business schools were not scientific enough. That they could learn from the war effort. Forrester’s work at MIT was funded by the Ford Foundation in support of that mood.

It’s a surprising wellspring for the circularity movement. It deserves our understanding.

An unexpected debt

It turns out, I owe much of my own thinking to Forrester. I named this practice Circudyne (for “circular dynamics”) before knowing there was a book called Industrial Dynamics, or that a discipline by that name had been taught for fifty years. I’m no more immune to the power of his message than anyone else, I guess.

There's more. Forrester opens his book by defining and delineating art from science. "We succeed before we understand why," he writes. Every worthwhile human endeavor begins as an art, practiced on judgment and experience, and the underlying science arrives later, out of a need to understand the foundation the art was already standing on.

The first letter I published here was called the Art and Science of Circular Transformation.

His purpose rhymes too. He says at the outset that industrial dynamics offers a single framework for integrating the separate functions of management, and that the point of the framework is to design better industrial and economic systems. That is what I am after in arguing that circularity should be promoted from a sustainability tactic to a transformation strategy. One frame across the functions, aimed at design.

So: I declare the debt. My nomenclature, much of my aesthetic, and some of my intellectual furniture are from Forrester's generation.

His legacy is prodigious. His central tenet was “the information-feedback system.” This was the engineering principle first developed at Bell Labs that led the way to weapons in WWII that could aim themselves. Information-feedback as a progenitor technology is recognizable in many modern vocabularies. Engineers call it control theory. Management calls it systems thinking. Every cloud platform runs on what its builders call a reconciliation loop. Machine learning calls it the agent and its environment. Executives call it a flywheel. Climate scientists just say feedback loops.

Supply chain gave his discovery a name of its own, the bullwhip effect, and then later named it after him. You can look it up.

War was Forrester’s disruptor

Forrester cited four foundations:

The theory of information feedback, worked out in weapons development, where the problem was computing an automatic firing solution against a moving airplane and correcting it as fresh radar returns arrived.
A new account of how decisions actually get made, out of military doctrine. Quickening the cadence of battle meant shifting attention from tactics to strategy, and that required writing the tactics down to make them repeatable.
The experimental approach to systems analysis, which we would now call simulation and scenario planning. Also a war product.
The digital computer, an accelerant for the other three.

Notice what the war is in that account. It is the source of the tools. It is also, to him, the supreme disruption — something so violent and so poorly damped that he made a career and a science out of the study of modulation.

Consider that against our own moment. We are enchanted by disruption. The prevailing attitude in technology is “move fast and break things.” Forrester would find that anathema. To his generation disruption was the pathology, the thing a well-designed system mitigates. The subject of Clayton Christensen's study would have struck them as a category error. (We’ll cover Christensen in a few weeks.)

That might partially explain how Digital Equipment Corporation was disrupted by the PC while Forrester sat on its board.

Forrester also published a scorecard, calling it “the challenge to management.” It laid out the three main practical benefits he predicted would accrue from the careful application of industrial dynamics.

“International economic development” — by which he meant building firms that could serve a global market. Check. Globalization happened, and partly by these methods.
“Middle management and technical effectiveness” — Check. We got Six Sigma, we got OKRs, we got a KPI for everything.
“Long-range planning” — not so much. I think we can say without argument that modern management has failed to fulfill this possibility.

Instead there was a migration toward metrics that drove short-termism. Models run on data, and bigger models need more data. This led to a managerial bias towards activities that produce data, and away from intangible goods; away from the future, from which there is no data.

This failed prediction explains why circular economy strategies, if you can call them that, have remained steadfastly technocratic and immune to broader appeal.

Defending the source of growth

In a 1968 paper, he struck a note that might surprise readers of his protégé Meadows, and the degrowth crowd.

Nearly the entire literature on feedback, he says, concerns what she called “the balancing loop” — the thermostat, the correction, the return to a setpoint. Almost none of it addresses the “reinforcing loop.” And every source of growth in a system is a reinforcing loop. He felt the discipline had lost its sense of balance.

That matters because of the next link in the chain to circularity. The Limits to Growth arrived in 1972 out of his own group at MIT, arguing that growth meets limits: overshoot, ceiling, constraint. A great and necessary book, it set the tone for fifty years of environmental thought.

Then Ellen MacArthur reached back to the precursor of Limits to Growth: Forrester.

In her account, she describes learning about finite materials and responding the way a thoughtful person responds in the face of it — traveling less, doing less, using less. Then she says that made her feel uneasy. It felt like buying time. Even if everybody did it, the problem would remain, because the system would remain.

That is the half everybody quotes. But the essence of her vision is incomplete without the other half. She put the question to system analysts: could a circular economy "decouple growth from resource constraints"? Forrester's descendants said yes.

Restraint is insufficient. Growth is not the enemy. What has to change is the relationship between them. And Forrester was all about the relationship between elements in a system.

MacArthur spent three months alone at sea, where everything she had is what she carried aboard. She says no experience could have taught her the meaning of the word finite so well. A boat is a closed system with a person living inside it, which is a scenario Forrester would have recognized.

She also says she had to understand the system in order to win. Not to survive. Not to balance. Win.

Circularity's leading light is growth-positive. But the community that answered her call is drawn mainly to the balancing loop. Forrester cautioned against that system bias 58 years ago. That is why many in the circular movement can discuss restraint fluently but not the desire for growth.

Forrester is aware of prevalent human and system imperfection. In the beginning of the book he lists the evidence that we are not yet good enough at designing social systems: labor turmoil, bankruptcy, inflation, economic collapse, political unrest, revolution, war. Every item is a disturbance to be damped. The man defending growth also treats change itself as a problem and stability as the goal. Both are true of him.

A propagation error

In his 1968 self-assessment, seven years after the book’s publication, Forrester assessed his own progress and found a bottleneck he was unable to resolve: that the science as he codified it was inaccessible to all but the most trained systems engineers. There was no reliable route to competence except apprenticeship. His book, he warned, reads as though it can be understood without specialized training, which makes it misleading: a reader finishes it feeling instructed, turns to the world, and cannot do the work. The only people likely to succeed were those who had studied feedback dynamics in an engineering curriculum first.

This failure has not yet been fully resolved. The progress of circularity has suffered because of it.

In Industrial Dynamics he acknowledged the problem, and gave management science twenty-five years to deliver a new dimension of management effectiveness. From 1961, that clock ran out in 1986.

In 1989 he addressed a banquet in Stuttgart and described what had happened instead. The early work ran in what he called consultant mode: study a company, go away, build the model, return with recommendations. The recommendations were accepted. People agreed with the logic. And then, under the pressure of daily operations, they went back to doing what they had done before. Changing how decisions get made, he said, is far harder than they had realized. Mental models do not yield to logical argument alone.

Industrial dynamics was not a failure. Two of his three challenges were met, and the apparatus he built remains foundational to our economy’s operating system. It is a partial victory, and the missing part is specific: technocratic quicksand.

Forrester critiqued the perceived shortcomings of the Harvard case method, for its dependence on intuition over formula. He attacked liberal arts education for the same reason. He was right about the question of transferability. But the thing he needed in 1989 — the ability to transcend logic to persuade — is precisely what the liberal arts are for.

This is precisely the reason Martin prescribes Aristotle's Rhetoric.

Something that does not yet exist emits no signal

You might suppose the trouble is a boundary problem; that circularity's error is drawing its system too small, leaving the customer on the outside.

Forrester's own rule permits that possibility. A variable belongs outside your boundary only if nothing inside affects it. And he applies the rule to demand himself. In his model of a new product there is a loop where orders lengthen the backlog, the backlog lengthens the delivery delay, and the delay makes the product less attractive to the customer, which lowers the order rate. Customer desire sits inside the system as a quantity, moved by the factory's own performance.

But look at what kind of desire that is. It responds. It is modulated by delivery times. He labels the loop negative, meaning self-correcting, aimed at a market that already exists. We have a phrase for it now: product-market fit.

For something new to the world, the limit is not where the boundary sits. A feedback system requires a signal, and a thing that does not yet exist emits none.

It’s a logical error. Draw the boundary as wide as you like. There is nothing on the far side to observe. The method needs a history, and novelty by definition has none.

Which is where Aristotle turns up. He divided the world in two: part of it cannot be other than it is; drop a ball and it falls, and a sample of the past predicts the future exactly. The rest can be otherwise, because it involves people. Analysis for the first part, invention and persuasion for the second. Business education ignored that line, which is Martin's lament. The prevailing wisdom of circularity ignores it too.

And circularity lives entirely on the other side of the boundary. It asks people to own differently, buy differently, want differently. No amount of past consumer data will tell you whether they will.

Here is the sharpest way I can put it in systems terms. Product-market fit is a goal-seeking loop with the existing market as its goal. On Meadows' ranking of places to intervene in a system, that sits near the bottom, among parameters and buffers, the weakest places to push. Moving the goal is her third-strongest lever, because a system's behavior is acutely sensitive to what its loops are aiming at.

The disruptive innovators of today move the goal. The systems tradition provides the logic, but not the argument. Perhaps that’s why they are so keen to break things.

None of which makes the systems apparatus wrong. It is sound but incomplete. It has no organ for generating a premise. Ask what would have to be true and you produce conditions; a model consumes conditions. Imagination supplies the premise, simulation stress-tests it, the prototype makes the data. Only one of them can go first.

Art, then science, then art

I mentioned at the top that Forrester had something to say about art and science: that art develops through experience and then plateaus, because its knowledge is disorganized. That science arises to explain and compress that experience. That the science then becomes a new foundation from which the art further evolves.

Think about where the circular economy transition sits on that curve: plateaued on disorganized knowledge — special cases, poor transfer, every firm convinced its situation is unique. That is Forrester's description of a stalled art, and it is the state description of our movement today.

But his analysis offers hope. Art, then science, then art again: a ratchet that turns. Industrial Dynamics was an historic advance in Aristotle's first kind of knowledge. But Aristotle showed us the other kind, too. Two sorts of knowledge, permanently distinct, neither reducible to the other.

The first letter I published here was called the Art and Science of Circular Transformation. The conjunction in that title is the argument.