A number outlives the man who knew what it was for. The discipline that governed it seldom does.
The previous essay carried the pattern into the examining room, where the patient cannot see the dashboard that shapes the advice she is given.
This one steps back from all five rooms and asks what they have in common.
* * * * *
In the fourth essay I promised that the last one would be my own accounting — that I had built one of these instruments myself, and that the reckoning was coming. I am going to keep that promise, though the reckoning turned out to be a different kind than I expected. A confession is too small a thing to end six essays on. My instrument is in here, but as evidence, not as the subject. What I owe you is the general case.
● ● ●
I. What the five were trying to do
Take them in order. For each: the job it attempted, what it actually achieved, how it was promoted, and two clocks — how often the number reported, and how long the thing it stood for took to show itself.
The body count. Statistical Control was built in 1942 because the Army Air Forces could not account for its aircraft. Its men found airplanes lost on paper, cut the abort rate, and found that B-29s flying the Hump burned more fuel reaching China than they delivered. That part of the story is usually skipped, and it should not be. The same method then went to a war with no lines on a map. The command was not starved of measurement; it was drowning in it. The damage was one level down, where the officer producing the count was also graded on it. That is promotion by evaluation. The count was reported daily. Whether a hamlet stayed secure showed over seasons.
Systematized ignorance. In 1965 Bulletin 66-3 directed the civilian departments to build program budgets — objectives, outputs, and costs, set side by side. Alice Rivlin’s analysts at Health, Education, and Welfare applied the method and reported that they had systematized their ignorance: they now knew exactly what they did not know. That was a finding, not a failure. A durable product was an analytical staff — and, in time, a profession of analysts — as Rivlin herself concluded in 1971. When the Congressional Budget Office reviewed the whole movement in 1993 it reached a similar and similarly qualified verdict. Meanwhile a cost per enrollee was written into law as a ceiling on the Job Corps — promotion by threshold, a figure that forecloses without grading anyone. The same law ordered follow-up on graduates at six and eighteen months, so the slower evidence was at least asked for. Appropriations came yearly. Whether a Job Corps graduate held a job showed years later.
The number Kuznets warned about. Simon Kuznets helped build the national accounts and wrote into the 1934 report that the welfare of a nation can “scarcely be inferred” from national income. What he practiced was a conditional: use the aggregate for questions the aggregate can answer, and when the question turns to who bears the cost, go below it. The number traveled and the condition was not carried with it. In November 1961 the OECD adopted a target of fifty percent growth in real gross national product over the decade for its twenty members taken together, and national income became something governments could fail at in public — promotion by target. The figures came annually, and the target ran a decade. What that growth cost, and who paid, showed over the decades after. Here delay and omission travel together, and I will not pretend the clock is the whole of it.
Teaching to the instrument. This is where the series corrected itself. American schools were running efficiency surveys by 1915, half a century before McNamara, so what arrived in a Texarkana school district in 1969 — a company paid a base rate of eighty dollars for each grade level a child gained in a subject — was re-import, not infection. That is promotion by transfer: the number became the payment. Donald Campbell had set the achievement test beside the body count by the mid-1970s as one mechanism appearing twice. The warning was available, and at the end of 2001 Congress voted accountability by test score into national law — the No Child Left Behind Act. The scores drove real gains in fourth-grade mathematics; that much the instrument achieved. Scores came yearly. Whether a child’s reading held showed years later.
When the metric becomes the medicine. The resource-based relative value scale was built for payment, to narrow the gap by which Medicare paid far more to cut than to think, and it narrowed it. Then it moved onto employed physicians’ compensation statements and into appointment templates — evaluation again, this time of the clinician. The relative value units posted monthly in many systems. The quality of a physician’s judgment shows over years with a panel of patients, much of it in procedures correctly declined, which generate no billable entry to reward. The fifth essay proposed a repair: comply-or-explain, a way for a physician to document why the standard case did not fit this patient, reviewed by people who can read a chart. It is a proposal, not a proven cure.
So what is the wrong, exactly?
Not the use of a number in a decision. Numbers belong in decisions, and some instruments are built to decide.
The wrong is authority beyond what the number was validated to carry. Three questions find where to look.
• What conclusion could this number actually support?
• What authority was attached to it anyway?
• Can anyone still enter an exception — this case is different, and here is why — and have it count?
They are questions for instruments that claim to measure a person or a program. A plain limit a legislature sets deliberately is a different act, and
I give it a different account below.
I have used one sentence for this in four essays: the instrument was built to describe; it was promoted to decide.
That is the last of it. The three questions are more exact.
● ● ●
II. The clock
Look at the pair of clocks in each entry above. In every case the number reported faster than the thing it stood for could be judged. The pairs are not all the same kind of mismatch — some are delay, some omission, some aggregation — and I set them side by side for the one feature they share.
That is the whole claim, and I want to hold it to its size. I do not say the mismatch caused any of these outcomes. Career incentives, statutory ceilings, command authority, and plain politics were in every one of those rooms. I say only that the mismatch is present in all five, in varying strength, and that it is worth naming because it lets you see five unrelated institutions as one kind of thing.
Others have seen the mechanism more generally. In 1995 Peter Smith warned that performance indicators across public services can induce managerial myopia, the pursuit of short-term targets at the expense of long-term objectives. In 2010 Andrew Haldane of the Bank of England argued that more frequent reporting of performance risks driving out patient investors in favor of impatient ones. James March showed that activities whose returns are quick, certain, and close at hand tend to win out over those whose returns are slow and distant. What this series adds is not a mechanism. It is a comparison across a war, a budget office, the national accounts, a school, and a clinic.
Reporting before outcomes mature is also simply management; nobody can wait a generation to decide. The mismatch becomes dangerous at the point the three questions mark — when the fast reading is given authority it cannot bear and the door to an exception closes.
No single line of descent runs through all five. Some transfers are documented: the bulletin that went out to the agencies, and the consultant from the Office of the Secretary of Defense who carried the warrant into Texarkana. Others have no carrier I can find. An institution that must account for itself to outsiders on a fixed schedule faces a pressure common enough that the same solution can be reached more than once, by people who never heard of each other.
● ● ●
III. The slow proposal
The sharpest picture I have of a fast number beside a slow proposal involves the same man twice.
In the spring of 1965 the Marines wanted an expeditionary airfield on a stretch of sand in Quang Tin Province. The Air Force’s Pacific command estimated a concrete field at about eleven months. Lieutenant General Victor Krulak proposed aluminum planking, and when Robert McNamara asked how long, Krulak said twenty-five days. He got his airfield. The first jets landed and flew combat sorties on 1 June — the twenty-fifth day. Twenty-five days was a promise a man could be held to, inside a period a Secretary could watch, and he hit it to the day.
In December of that year General Krulak wrote a seventeen-page strategic appraisal arguing the war was being fought wrong. He set enemy manpower against the kill ratio and concluded that the American price of attrition would be intolerable — something like a hundred seventy-five thousand American lives to cut the enemy’s manpower pool by a fifth.
What he proposed instead — protect the population, strike the North’s ports and rail and power, put both governments into pacification, press Saigon on land reform — came with figures, but no promise that would settle on a date a Secretary keeps, because pacification shows its results slowly.
In 1966 he took his case to the President and, as he told it, got as far as the ports — the fast, countable half of his case — before Johnson walked him to the door. There were reasons enough for that rejection — the fear of drawing in the Soviets or the Chinese among them — and I am not claiming the reporting cycle sank him. I am pointing at one difference between his two proposals. The first was a promise that would settle on a schedule a decision-maker keeps. The second, for all its figures, was not.
Call it the Krulak Principle, in two halves. A proposal whose results will not arrive inside a decision-maker’s own horizon starts at a disadvantage against one that will. And a number whose result will not settle inside the reporting period holds a man to less than it appears to, because the thing everyone is waiting to see has not happened yet.
It is tempting to conclude that nobody chose any of this — that the instruments promoted themselves. That excuses too much. A board approved the compensation formula. The project’s auditors in Texarkana found that a company had put test items into the study materials. An officer signed a count he had estimated in the direction his career pointed. A legislature wrote a cost ceiling — a deliberate policy choice, which is owed a different account than a falsified report. These are not equal acts, and an honest reckoning sorts them by what each person knew, could have known, and had the power to change. Dennis Thompson argues that the way to keep individual responsibility alive in large organizations is to attend to responsibility for their design. Applied to these pages, that means asking who built the instrument, who decided what turned on it, and whether anyone was ever assigned to notice when it stopped telling the truth.
Assigning that watch is itself a design duty. The pressure explains why the choice was hard. It does not make the choice.
● ● ●
IV. My own instrument
For a stretch of years I read position specifications for a living — the document a company writes when it has decided to hire and must state what it wants. In the mid-1980s, at the request of MBA students who wanted to know what companies were actually looking for, I collected about nine hundred of them, thirty apiece for thirty positions: Vice President–Marketing, Vice President–Finance, and the like. The goal was a skill profile for each position — a mosaic that could show what the job actually asked for and guide a young MBA’s development toward it. These were positions companies had paid to fill, not answers to a survey. That makes them good evidence of what employers asked for, and I will not stretch them one inch further.
What went on paper was what could be verified: the degree, the years, the prior titles. What companies actually wanted came out on the telephone, in stories — the controller who found the ruinous sentence in the lease nobody else had read, the executive who turned an idea into a working business. The specifications named the credentials. The performance behind the stories had no line. The mosaics were built from both.
The finding that mattered cut across all thirty positions. Whatever the job, certain skills kept appearing. I called them the Critical Skills — five at first, later expanded to eight. Trustworthiness was not on the list I made. I considered it a quality, a different kind of thing — which is a classification, not a finding.
Field Studies was the method for teaching them — students doing real work for real organizations, practicing the skills where the skills live. Washington was moving the same way. The Labor Department’s SCANS commission had defined workplace competencies of its own — cousins to mine, not copies — and in 1993 its publication Teaching the SCANS Competencies profiled Field Studies as one way to teach them. The next year, the School-to-Work Opportunities Act put federal seed money behind skills programs in the schools.
Then I built the instrument proper, and you are owed a disclosure before I describe it. Coop2000 was a commercial product of my own company, and Critical Skills has been my brand ever since. It gave schools a system for work-based learning with local businesses: each placement tied what the student would do on the job to a skill framework — mine where a district chose it, and several did; SCANS or the state’s standards where it did not — and the student’s performance was assessed against the skills he was there to practice, assessed where the work was done, by people who watched him do it. It ran in thousands of schools across the country, rooms I never stood in.
And it did one more thing, and I sold it partly on that one more thing. It gathered the assessments into the reports a district owed for its federal money. My instrument fed distant rooms too. I built it to live in the world the fast number was making.
The School-to-Work money expired on the first of October, 2001, by a sunset written into the Act itself. That December Congress voted accountability by test score into national law, and the President signed it in January. The software itself was aging out of its platform by then. I cannot separate those causes and will not pretend to. What I can tell you is what I watched. The schools I knew answered for test scores now, and hour by hour they turned toward the tested subjects. Coop2000 tapered until it was gone. A skill shown in real work reports slowly, locally, one student at a time. A test score reports every spring, everywhere, in a form a legislature can read. In the schools I watched, the hours went where the reporting was.
Set beside it the other instrument I built. Aboard a nuclear submarine, the USS Ray, I learned what a submarine’s displays could not show an officer with seconds to act, and afterward, as a tactical instructor at the U.S. Naval Submarine School, I built the Geographic Plot from those lessons. It measured something that mattered — where is he, right now — and its reading was used by the officer who made it, in the moment, with the knowledge of its limits standing right there. It also provided a measure of ship safety – plotting a worse-case scenario. And it was useful for later reconstructing what happened in an encounter. As far as I ever saw, it was never asked to be anything else, and it decided things every day.
Two instruments, one design principle: the person acting on the reading was the person next to the thing being read. Whether that principle lost on the merits I cannot prove either way, because nobody measured — including me. That is the accounting, since I promised one. I never gathered the evidence I have demanded of the five in these essays: no follow-up on the students, no record of what the assessments predicted, no independent evaluation of the program. Principals and superintendents shared what we had among themselves, and it never entered the rooms where the money was decided. Some of that was geography. Some of it was mine. I would have carried it into those rooms gladly. I still would.
● ● ●
V. What a republic runs on
Let me refuse the likeliest misreading of this series flatly.
This is not an argument against measurement.
Professional judgment is not a substitute for evidence, and an institution that asks to be trusted on judgment alone should be refused. Trust us is what the patronage machine said, and the hospital that buried its mortality figures, and the school that graduated children who could not read. The parent who cannot evaluate a school, the patient who will be unconscious when it matters, the citizen paying for a war he cannot assess — each is owed something better than trust, and measurement is often what gives them a place to stand. The people who built these five instruments were answering that demand, and it was legitimate.
That is the knot at the center, and I cannot untie it. Accountability to outsiders and the overpromotion of a number are close cousins. Outsiders can check a great deal without a number — records, testimony, a documented exception. But a number is usually the cheapest and most portable form the check can take, and once it exists it tends to be pressed past what it can hold. I have not found an arrangement that removes that pressure, which is not proof that none exists. The repairs I know of work by pushing back on it.
And I owe a correction to this essay’s own title, because the fourth essay already made it.
It is not true that the important things cannot be counted. Parts of civic discernment can be assessed; someone built an instrument that measures some of it.
The real question is which measurements have power — which one the day is organized around and the money follows. When schools moved time toward the tested subjects, discernment was not what kept its hours.
I have kept this series largely to episodes the record has settled, because the past holds still long enough to be judged. My own is the exception, and you may weigh it as what it is — testimony. I will not score this week’s controversy for you; that would be the very thing these essays warn against.
But you have the test.
When a number is put in front of you, ask what it was built to answer and whether that is the question it is now settling. Ask how often it reports and how long the thing it stands for takes to show itself. Ask whether anyone can still say this case is the exception and be heard.
And if you built the instrument, write down what it cannot carry, and put it in the procedure rather than the preface. A written rule is not a wall; procedures can be gamed like everything else in these pages. It only gives the next person a chance to meet the limit before the number does his thinking for him. If you can, go and find out what became of it. If you also control what turns on the number — the formula, the ceiling, the report — then the note is the least of what you owe. That is a small repair and not equal to the problem. It is what one person can do.
What a republic actually runs on is seldom what it answers to.
The willingness of a citizen to sit through an argument that cuts against his own interest. The discernment to tell a serious man from a plausible one — a judgment that can go wrong, and is still the one a self-governing people cannot delegate. The capacity to accept, through legitimate process, an outcome you voted against, and go on taking part.
Little of this appears on any form that governs a school day or a legislative calendar. The founders understood that some things should not move at the speed of a passing mood; they gave the Senate six-year terms so that one part of the government would outlast the mood. A republic does not lose its capacity for deliberation because anyone votes to end it.
One way it loses it is quieter.
Deliberation seldom appears on the instruments that drive the calendar; the things that compete with it for the same hours do, and the contest is settled before most people notice it is being held.
* * * * *
Charles Cranston Jett is an author, civic educator, and Professional Certified Coach based in Chicago. A graduate of the U.S. Naval Academy (Class of 1964) and Harvard Business School, he served during the Cold War aboard the nuclear submarine USS Ray (SSN 653); afterward, as a tactical instructor at the U.S. Naval Submarine School, he developed the Geographic Plot, a contact-tracking method for submarine operations. He is the author of six books, including Super Nuke!, hosts four podcasts, and writes across his Critical Skills Blog platform on history, leadership, and the health of the American republic. In his writing he employs AI tools in a limited, supporting role for research, occasional image creation, and editing, while the prose and judgment remain entirely his own. He and his wife, Dr. Nancy Church, live and co-host the Chicago Salons at Water Tower Residences.
Leave a Reply