Thinking in Systems

The Bill Arrives Later

Lesson 3.4 handed you a structure for consequential decisions: frame the real question, generate real options rather than one option and two decoys, weigh the evidence honestly, ask whether the door swings both ways, and run the pre-mortem before you sign anything. It closed by asking which decision currently on your desk would fail its own pre-mortem. Hold that decision in your mind, because this lesson is about to complicate it — not to undo the framework, but to show you what the framework quietly assumes.

Here is the assumption. A decision framework takes for granted that the world will grade your paper. You choose, results come back, you learn, you adjust. That loop is how every competent leader you know became competent. It is also the loop that stops working in exactly the situations where the stakes are highest, and it stops for a reason nobody announces: the grading is late, the grade gets mailed to somebody else, and by the time it arrives no one remembers which paper it was for.

Picture a leader who inherits a budget gap and imposes a hiring freeze. The savings appear in ninety days. The applause appears at the next board meeting. The consequence appears in four years, when the people who would have been trained to run the place were never hired, and the institution discovers it has no bench. Nobody will connect those two facts. The freeze will be remembered as a success and the bench problem will be remembered as a talent-market issue. Same decision. Different decade. No feedback loop joining them.

That gap — between the moment you act and the moment the system answers — is where an enormous amount of well-intentioned leadership goes wrong. Notice that the failure is not incompetence. The leader in that example framed the problem, weighed options, and chose the reversible one. He did everything Lesson 3.4 asked. He simply looked at the event and not at the structure it was sitting on.

That is the distinction this whole lesson turns on. Events are what you see: the number, the resignation, the vote, the headline. Patterns are what you see if you keep records and look across time. Structure is what produces the patterns, and it is almost never visible in the room where the decision gets made. Most leaders are excellent at events, adequate at patterns, and have never in their lives been asked to describe a structure.

So let me state the claim plainly, and let me also state its limit, because manufacturing certainty is not how we do things here. The claim: a leader is responsible not only for the decision but for the structure the decision runs inside, and that responsibility extends past the end of his term. The limit: nobody can compute a complex system. Systems thinking is not a crystal ball and any teacher who sells it to you as one is doing you harm. It is a way of seeing that reliably surfaces questions you would otherwise never ask. That is worth a great deal. It is not the same as knowing the future.

Before we go further, do something with this. Name one result in your organization — a number, a turnover figure, a recurring complaint — that everybody currently treats as an event. Write it down. We come back to it with better tools.

Complicated Is Not Complex

The word system gets thrown around until it means almost nothing. So define it usefully: a system is a set of elements connected in a way that produces its own pattern of behavior over time. That phrase — its own — is the whole point. A pile of sand is not a system. A city budget is, because it will do things you did not order it to do.

The most useful distinction a leader can learn here is between complicated and complex. A jet engine is complicated. It has an enormous number of parts, the rules governing them are knowable, expertise transfers from one engine to the next, and the same input produces the same output. Raising a child is complex. The parts respond to you. Your intervention changes what the intervention is intervening on. This complicated-versus-complex distinction is developed in Frances Westley, Brenda Zimmerman, and Michael Quinn Patton's Getting to Maybe, and it sits alongside the Cynefin framework that David Snowden and Mary Boone laid out for executives in the Harvard Business Review in 2007.

Now look at what that means for you. Almost all executive training is complicated-domain training. Procurement, logistics, accounting, project scheduling, legal compliance — these are complicated, and analysis genuinely solves them. But congregations, markets, campaigns, neighborhoods, and city agencies are complex, because the parts read the memo. When you announce a new performance standard, people do not passively receive it. They read it, infer what you actually want, and adapt. The system you measured yesterday is not the system you are operating today, because measuring it changed it.

Underneath the confusion sits an old philosophical habit worth naming. Aristotle taught us to think teleologically — to explain things by the end toward which they aim. It is a rich and defensible way of understanding purposeful action, and Scripture assumes something like it, since the world is going somewhere and was made on purpose. But a leader can quietly smuggle in a false version of it: the assumption that because my intention is good, the outcome will bear the shape of my intention. Complex systems do not honor intentions. They honor structure.

Robert K. Merton put an analytical frame around this in 1936 in a paper called The Unanticipated Consequences of Purposive Social Action. He catalogued why purposeful action produces results nobody planned — plain ignorance, error, and a category he named the imperious immediacy of interest, which is his phrase for wanting the near result so badly that you never examine the further ones. That last one is not a research finding about strangers. That is a description of the last time you needed a win before a board meeting.

Let me steelman the objection to all of this, because it is a serious one and it comes from serious people. The critic says: systems language is how accountability evaporates. Nobody did it, the system did it. The budget failed, the culture failed, the incentives failed — and no human being ever has to stand up and say I decided that. The critic is right that systems talk gets used this way constantly, in corporations and in government, and usually by the person most responsible. Here is the answer, and it has to be more than a slogan. Systems thinking properly used expands accountability rather than dissolving it. It does not excuse you from the decision. It adds to your account the second-order effects you were too busy or too flattered to examine. If your systems analysis ends with nobody responsible, you have not done systems analysis. You have written a defense brief.

The Four Sight Lines

Here is the framework, and I want it to be simple enough that you can run it in a meeting without a whiteboard. Four sight lines: stocks, loops, delays, incentives. Look down each one before you commit.

The first sight line is stocks. A stock is an accumulation — something that builds or drains over time rather than existing in a moment. Cash is a stock. So is deferred maintenance, institutional trust, pension liability, housing supply, staff experience, and your own credibility. Flows are what change them. Here is the executive insight: stocks change slowly relative to conversation, which is why leaders consistently misjudge them. John Sterman's research at MIT has documented that people are systematically poor at reasoning about accumulation — shown a picture of inflows and outflows, educated adults routinely get the resulting stock wrong. If you cannot reason about a bathtub, you cannot reason about a pension fund. The question this sight line asks: what is accumulating or draining while I am busy?

The second sight line is loops. Systems run on feedback, and feedback comes in two flavors. Reinforcing loops amplify: a well-regarded school district attracts families, which raises the tax base, which funds the school, which improves its reputation. Balancing loops correct: crime falls, the political pressure to fund enforcement falls with it, funding erodes, and conditions drift back. Notice that a reinforcing loop is morally neutral — it will compound your virtues and your rot at the same rate. The question this sight line asks: what does this result do to its own cause?

The third sight line is delays, and it is the one that ruins the most careers. Delay is why intelligent people over-correct. You have felt it in a hotel shower: the water is cold, you turn the handle, nothing happens, so you turn it further, and thirty seconds later you are scalded. Now scale that to a hiring plan, a rate cut, or a police staffing decision. The Beer Distribution Game, developed at MIT to teach system dynamics, puts capable people in a simple four-stage supply chain and watches them generate wild oscillations — not from stupidity, but from acting on information that arrives late. The question this sight line asks: how long until the truth about this shows up, and who will be sitting in this chair when it does?

The fourth sight line is incentives, which are the system's real instructions regardless of what the mission statement says. People respond to what is measured, funded, promoted, and forgiven. If your stated value is candor and your promotion pattern rewards people who never deliver bad news, your organization runs on the promotion pattern. The question this sight line asks: if I were paid, promoted, or re-elected under these rules, what would I actually do? Answer that honestly about your own people and you will predict their behavior better than any survey will.

Run the sight lines in that order and you will notice something. Most leadership energy goes to parameters — the budget number, the staffing level, the tax rate — because parameters are what a leader can visibly change this quarter. Donella Meadows argued these are the least effective places to intervene, and that higher leverage lies in the goals a system pursues and the paradigm those goals arise from. She also made a claim that should keep you humble: people intuit where the leverage points are, and then push them in the wrong direction. Being right about where to push is not the same as being right about which way.

Now take the result you wrote down at the end of the first section — the one everybody treats as an event. Run all four sight lines on it. Most of the time, the thing your organization has been treating as a recurring surprise turns out to be a structure doing exactly what it was built to do.

The Man Who Was Never Elected

New York City supplies the canonical case, and it is not a mayor. Robert Moses never won an election. Over roughly four decades he held appointed positions simultaneously — city parks commissioner, chairman of the Triborough Bridge and Tunnel Authority, and a series of coordinating and clearance posts — and from them built parkways, bridges, expressways, beaches, playgrounds, and housing on a scale no elected New York official has matched. That is established record.

The structural genius, and it was genius, was not in the concrete. It was in the financing. Toll revenue from his bridges and tunnels was pledged to bondholders under covenants that ran for decades. That did two things at once. It gave him a stock — a reliable, growing pool of money that no mayor controlled and no voter appropriated. And it created a reinforcing loop: tolls funded construction, construction moved more cars, more cars generated more tolls, which funded more construction. Elected officials arrived and departed on four-year cycles. The bond covenants did not. A mayor who wanted to redirect that money had to contend with contract obligations that outlived his administration.

Sit with that, because it is the lesson. Moses understood what most elected leaders never learn: whoever writes the structure outranks whoever wins the election. His power was not charisma and it was not a mandate. It was delay and incentive built into legal architecture. He was removed from that base in 1968, when the Triborough authority was folded into the newly created state Metropolitan Transportation Authority — which is to say, it took a structural change at the state level to unseat him, because a structural position cannot be argued out of office.

Robert Caro's The Power Broker, published in 1974 and awarded the Pulitzer Prize, is the standard source and the reason most people know any of this. It is also a work of interpretation as well as reporting, and this academy asks you to hold those apart. Caro's account of Moses' motives and of the human cost of the Cross Bronx Expressway is powerfully argued and widely accepted, and it is an argument. The frequently repeated claim that Long Island parkway overpasses were built deliberately low to keep buses — and therefore poorer and Black New Yorkers — away from the beaches was popularized in Langdon Winner's 1980 essay Do Artifacts Have Politics? and has been contested since, notably by Bernward Joerges, who challenged both the evidence and the retelling. The 2007 collection Robert Moses and the Modern City, edited by Hilary Ballon and Kenneth Jackson, argues that Caro's portrait undervalues what Moses actually built and the constraints he worked under. You do not have to adjudicate that dispute to lead well. You do have to know which claims are established, which are contested, and never to present the second as the first. That discipline is Lesson 3.6 arriving early.

Then there is the other New York case, the one where the bill arrived all at once. Through the 1960s and early 1970s the city expanded services, payroll, and obligations while its manufacturing base and population were eroding. Mayor Wagner's 1958 executive order recognizing collective bargaining for city employees was, at the time, an act of decent intent toward workers; the compounding cost of the contracts and pensions that followed landed on successors who did not negotiate them. Short-term notes were rolled over. Operating expenses were moved into the capital budget. Each of those was defensible in isolation and each one moved a stock in the same direction.

By 1975 the city could not sell its paper. The state created the Municipal Assistance Corporation and then the Emergency Financial Control Board, effectively suspending the city's fiscal autonomy. The Daily News ran its famous FORD TO CITY: DROP DEAD headline that October; in December the president signed federal seasonal financing legislation. Charles Morris titled his book on the period The Cost of Good Intentions, and the title is the argument. Be honest about the causal dispute: historians still disagree about how much weight to assign labor costs, deindustrialization, federal policy, and accounting practices. What is not disputed is that no single decision caused it. A structure did, over fifteen years, while capable people made individually reasonable choices.

The Metric You Choose Is the Behavior You Get

Let me give you the single most useful operating rule in this lesson. Charles Goodhart, writing about British monetary policy in 1975, observed that a statistical regularity tends to collapse once it is used as a target for policy. The crisp version most people quote — when a measure becomes a target, it ceases to be a good measure — is Marilyn Strathern's later restatement of his point. Donald Campbell made a parallel argument about social indicators. Say it in leadership terms: the metric you choose becomes the behavior you get, and the behavior you get is rarely the behavior you meant.

New York provides an instructive and genuinely contested example. CompStat made precinct commanders accountable for crime numbers in their commands, and that accountability was real and in many ways salutary. It also created exactly the incentive Goodhart described. Criminologists John Eterno and Eli Silverman, surveying retired NYPD commanders, argued that pressure to produce numbers distorted crime reporting more than the department acknowledged. An outside panel appointed by the police commissioner reviewed reporting integrity and found instances of misclassification while concluding the overall downward trend was real. Both can be true, and a leader who cannot hold them together will either dismiss a real problem or reject a real achievement. The structural principle is not in dispute: attach a career to a number and the number will be managed.

Which brings us to the office this curriculum is aimed at. A mayor inherits systems built by predecessors and hands his own to successors, and almost nothing important resolves inside one term. New York's City Water Tunnel No. 3 began construction in 1970 and is still not finished. Pensions negotiated decades ago constrain budgets today. Zoning determines what gets built fifteen years from now. Early-childhood programs produce outcomes researchers are still arguing about a decade later, with some studies showing effects that fade in later grades. If you need to see your own results, you will systematically choose the wrong projects.

The credit will also be misassigned, and you should decide now how you will handle that. Safe Streets Safe City, enacted under Mayor Dinkins, funded a substantial expansion of police hiring; many of those officers were on the street during the following administration, which received most of the public credit for the crime decline. And the decline's causes remain genuinely unsettled among criminologists — policing tactics, the crack market's collapse, demographic shifts, incarceration, and the economy all have serious advocates. A systems thinker says that out loud. A campaign consultant would rather you did not.

One more executive discipline, and it is a matter of integrity rather than technique. Knowing the system includes knowing which levers are actually in your hand. The MTA is state-run. School governance depends on Albany. District attorneys and the courts are independent. A candidate who promises to fix the subways as though the office controls them is not being visionary, he is misdescribing the structure to people who will believe him — and the disappointment that follows corrodes the public trust that every future leader will need. Say what you can do, say who else holds the rest, and say how you will work with them.

So bring this back to yourself, because a framework you do not apply is entertainment. Look at the loop you are currently inside — not the one your predecessor built, the one you are building. What are you measuring, what behavior will that measurement purchase, what is accumulating on your watch that will not be visible for years, and who will pay for it?

And here is the honest difficulty that carries us into 3.6. The hardest part of systems thinking is never the diagram. It is that the diagram usually implicates you. Once you can see structure, you will eventually see a loop that you built, defended publicly, and staked reputation on. Seeing it is worth nothing unless you can say so out loud. That capacity — holding your conclusions accountable to evidence when the evidence indicts you — is the subject of the next lesson and the capstone of this module.