Third time round this particular mountain, Herman. We've done who actually keeps the powerful alive and functioning, we've done the people who work for the ultra-wealthy, and now Daniel wants to go inside the calendar itself. He's asking three things that are really one thing: what does a day off actually look like for a head of state, how do the super-CEOs function, staff and schedule-wise, and what does the data say about executive pay having risen out of all proportion — including, he says, a point Sir Ronald Cohen made about reining in unfettered capitalism.
And that last clause is the one I want to hold separately, because it's going to take some careful handling. But the through-line first: every one of these questions is about how the machinery around a powerful person manages their attention. That's the actual subject. Not the power, the attention.
Which is a scarce resource for a sloth in a way you people will never understand.
You nap fourteen hours a day.
Eleven. Discipline.
My point is that all three of Daniel's questions share one answer shape. Time off, at the top, is not a cessation of work. It's a reallocation of attention — same brief, different venue. And the machinery that makes that possible is staff. Let's start with the person who can never actually leave the office, because the paper trail there is wonderful.
Before you get into it — why does this tier come first? Why not start with the CEOs, since there's more data on them?
Because the presidency is the limiting case. If we can establish what a day off looks like for the one person on earth who structurally cannot have one, then everything the CEOs do reads as a softened version of it. The CEO tier is the presidency with an exit ramp. Start with the hard case and the rest falls into place.
Fine, but I'm holding you to the exit ramp.
The official line first. Jay Carney, standing at the podium in twenty eleven while Obama was on Martha's Vineyard: "There's no such thing as a presidential vacation. The presidency travels with you." And the thing about that line is it isn't spin. It's an operational description.
Walk me through why it isn't spin, because it has the shape of spin. It's exactly what a press secretary would say to make golf look like governing.
Because you can test it against the logistics. Spin collapses when you check the manifest — and the manifest here is extraordinary. Take Hawaii, December twenty ten. Every morning before eight a.m. local time, an intelligence official physically travels to the vacation rental to deliver the presidential daily brief — the same classified briefing he gets at the White House. After that, a chain of staffers feeds him extra material by secure phone, video, and email. So the holiday isn't a break from the brief. The brief commutes.
The brief has a better work ethic than I do.
And it's worth pausing on what that actually means in practice, because it's not one man with a folder. The briefing officer has to be cleared, transported, housed near the vacation site, and available at the same hour every day regardless of time zones. When the president is in Hawaii, that's a ten-hour shift difference from Washington. Somebody's entire job, for two weeks, is to be the White House's morning, relocated.
So the vacation is real for the president's family and fictional for roughly everyone accompanying him. There must be staff who spend every August in Martha's Vineyard and never once see a beach.
The advance teams, the comms teams, the protective detail — none of them are on holiday. The presidential vacation is, functionally, a small branch office that gets assembled in a resort town and then disassembled two weeks later. The only person who experiences it as leisure-adjacent is the principal, and even he gets the brief before eight.
And historians point out the hardware follows him too?
Right — communications gear is always within immediate reach. The "in the loop" requirement is satisfied by equipment as much as by people. But here's the detail I love most, because it's the smallest and it says everything. Jimmy Carter, on vacation aboard the Delta Queen riverboat. The wire story describes his "typical vacation day": he slept until seven a.m., which the report notes was two hours later than normal, and the famously punctual president was late to a picnic.
The man negotiated peace between Israel and Egypt and his measurable downtime was permission to be late to a picnic.
And even then — even in that two-hour lie-in regime — he still got the daily intelligence briefing and news summary aboard the boat, over what the story calls sophisticated gear. Fifty-three-year-old sternwheeler, cutting-edge comms. The relaxation was a scheduling allowance, not a state of being.
It's the vocabulary that gets me. "Two hours later than normal." The wire service is reporting on the president's sleep the way you'd report on a machine's tolerances. Like the deviation itself is the news.
That's because, functionally, it is. The presidency is treated as a continuous system, and the vacation is logged as a variance from baseline. Nobody writes "the president, a human being, rested." They write "he slept until seven, two hours late."
So the question Daniel's really asking — how do aides orchestrate this so the person is technically in the loop but disconnected — has a mechanical answer. It's pre-packaging. Walk me through the White House version, because I read about the Book last night and it's better than my leaf medicine.
The Staff Secretary's office. The White House Transition Project calls it "the last substantive control point before papers reach the Oval Office." Everything that goes to the president passes through it — recommendations, speech drafts, signature forms, guest lists, intelligence, calendars. And six evenings a week, that office produces the Book. It covers tomorrow, three weeks out, and three months out.
Three time horizons in one binder, delivered the night before. Why three? Why not just tomorrow?
Because tomorrow alone would make the president reactive — he'd only see what's urgent. The three-week and three-month views are what let him steer. It's the difference between a driver who can see the bumper of the car ahead and a driver with a map. The Book is the map, and the nightly delivery is what makes the morning calm enough to use it.
And how thick is this thing? Are we talking a pamphlet or a phone directory?
It varies with the news, but the honest answer is that thickness is the whole design problem. Too thin and the president is under-informed; too thick and he skim-reads, which is worse than not reading, because he now believes things. The Staff Secretary's office exists to make that thickness decision hundreds of times a day — what gets in the Book, what gets summarized in a line, what gets held back entirely.
So the day off is manufactured. Not by absence, by pre-processing.
And here's the part that should worry anyone tempted by the job: the staff day is longer than the principal's. People who worked in that office describe six a.m. starts, a "fire drill" between quarter past six and eight — before the president even arrives at ten to seven — and staff still there at half nine, ten at night. The Executive Clerk's office shows up in FDR's three p.m. appointments in the day logs from nineteen thirty-nine. The paper-flow apparatus has been attached to the presidency for nearly a century.
The president arrives at six fifty. The people who prepared his morning have been at desks since six. The man at the top of the org chart has the shortest day in his own building.
Nearly.
I'd take that deal and the Book with it. Where do I sign, and does it come in leaf-green?
You'd be the first president in history to fall asleep during the fire drill.
During the oath, Herman.
There's one more piece of this that's structurally interesting, which is what the press gets told. Trump's seventeen-day working vacation in August twenty seventeen — the aides' approach was to disclose only the work-related calls and briefings, not the golf or the dinners. CNN's Kevin Liptak summed it up: "While we know when he's 'working,' we don't really know when he's on 'vacation.'"
Which means the definition of the day off is being written by the press office in real time. Whether he's resting is a communications decision.
And the categories are controlled by the institution. Hold that thought, because it comes back in the pay segment — it's the same shape.
One more contrast and then I want the CEOs. Obama's first day out of office. He described it as "pretty mundane" — slept in, puttered around the house, and said he "did not miss the trappings of the office." The trappings. Exactly the things that made the vacation not a vacation — the brief, the comms kit, the staff — turn out to be the things you notice the absence of.
The whole apparatus is removable. It just can't be removed while you still need it, and you can't stop needing it, so it never comes off.
The presidency travels with you — until it doesn't, and then you sleep in.
He did seem happy about that.
Carter got a lie-in and picnic lateness on a riverboat. Obama got the whole thing off. Progress, of a kind.
Two hundred and thirty years for the first full morning of sleep. That's the frontier.
Right, the CEO tier — same machinery, different budget. And you've been waiting all segment to tell me about the Harvard study, so go on.
I have, a little. Michael Porter and Nitin Nohria ran a study at Harvard Business School tracking chief executives of large companies around the clock for thirteen weeks, in fifteen-minute increments — where they were, who with, doing what. Twenty-seven CEOs, sixty thousand hours of data, companies averaging over thirteen billion dollars in revenue. It's the most detailed picture we have of how that tier actually lives.
Twenty-seven people. Two women.
Two of twenty-seven, yes. Which is itself a finding, just not the one the study was after. So the hours: nine point seven hours per weekday, which sounds reasonable — that's under fifty a week. But then seventy-nine percent of weekend days, at just under four hours each. And seventy percent of vacation days, at two and a half hours each. Add it up: sixty-two and a half hours a week.
They work on seventy percent of their vacations. That's a hobby with a salary attached.
The study's own framing was that it's over fifty percent more than a traditional forty-hour employee. And the twenty-four-seven split is the part I find quietly grim: thirty-one percent working, a quarter personal time awake, twenty-nine percent sleeping — six point nine hours a night — and five percent on vacation. Total. Five percent of existence.
I sleep more than a Fortune 500 CEO and I have no revenue responsibilities whatsoever.
You have a hammock and a following.
I have listeners, Herman. That's a responsibility.
Here's the number I keep turning over, though. Seventy-two percent of their work time is meetings. Seventy-five percent of all their time was scheduled in advance. So three quarters of the waking life of a person at that level is booked by someone else, in rooms, usually with other people in them.
And the confession?
In the debriefs, the CEOs themselves admitted that one-hour meetings could often have been cut to thirty minutes or even fifteen. They knew. The schedule is a garment everyone agrees is too big, and nobody takes it in.
But hold on — if they knew, why didn't they fix it? These are people whose entire reputation is decisiveness. They cut divisions. They cut product lines. Why can't they cut a meeting?
Because the cost of the too-long meeting is invisible and the cost of the cut meeting is visible. If a meeting runs long, nothing happens — no one files a report about the wasted half hour. If a meeting gets cut and something goes wrong six months later, the meeting is remembered. The schedule punishes only the omissions, never the padding. Every calendar in the world works this way, but at the top there's no one above you to force the discipline.
So the asymmetry scales with power. The more powerful you are, the less anyone can tell you the meeting should have been an email.
And there's a compounding effect, because the people who might say it are the ones whose access depends on the meetings existing. The chief of staff who tells the CEO to halve the calendar is arguing for the shrinkage of his own influence. The honest advisors in that structure are the ones cutting their own ration.
The whole building has an incentive to keep the garment too big.
What fills the meetings? Or do I not want to know what a billion-dollar afternoon looks like from inside.
Roughly a quarter on people and relationships, a quarter on functional and business-unit reviews, a fifth on strategy, sixteen percent on organization and culture. And then the sting: six percent of work time with rank-and-file employees, three percent with customers. The two groups the entire enterprise exists for get single digits between them.
Three percent with customers. The people who actually buy the thing.
The study's phrasing was that frontline employees and customers "struggle for executive attention." Face-to-face time — which is sixty-one percent of their communication — is read internally as a signal of who matters. Porter and Nohria put it that people watch this more carefully than most CEOs realize. Who gets on the calendar is a broadcast, whether or not the CEO knows he's transmitting.
So the calendar is the org chart, the newspaper, and the constitution all at once.
That's the deep finding of the study, I think. The schedule isn't a convenience. It's the medium of power.
Now the staffing question, because it's the one Daniel asked straight: how many secretaries and assistants does one of these people actually have? And my strong suspicion is that there's no published number.
There isn't. I looked for a headcount for a Fortune fifty CEO's support team and nobody publishes one — not the companies, not the trade press. What the literature describes instead is function. The executive assistant as gatekeeper: filtering information, protecting confidentiality, keeping non-urgent interruptions off the desk. There was a study of EAs at big global companies that found their influence runs through exactly two channels — gatekeeping and relationships.
Which is a polite way of saying the assistant decides what the CEO knows.
One writer on the EA role, Jan Jones, makes a distinction I like — that it should be "gateway" rather than "gatekeeper." Gatekeeping means handling the routine matters yourself so the boss never sees them; a gateway lets legitimate access through. It's the difference between a wall and a door with a doorman.
A door with a doorman who files the doorbell.
And the newer structural answer, the one that's spreading right now, is the corporate chief of staff. HR trade press this year reported that so many CEOs now want their own. The complication is that nobody can define the job — a psychologist at Claremont McKenna who co-authored one of the only peer-reviewed studies of the role said, "If you say chief financial officer, everyone knows what that means. But you say chief of staff, and it's not clear."
A C-suite position that defies definition, spreading by demand. What could go wrong.
In practice it tends to mean the person who runs the CEO's operating rhythm — the meetings, the pre-reads, the follow-through. The Staff Secretary with stock options, honestly.
There's the parallel then, and it's tighter than I expected. The president's Book and the CEO's chief of staff are the same invention: a person or a document standing between the principal and the flood, deciding what arrives and when. The difference is what tethers them — the president can't disconnect because of national security, and the CEO can't disconnect because of the meeting schedule itself.
The CEO's vacation work — those two and a half hours on seventy percent of vacation days — is mostly not crises. It's the schedule continuing without him having chosen it. Sixty-two and a half hours isn't a workaholic's number, it's a calendar's number.
Carter got two hours extra sleep on a riverboat; the modern CEO gets a laptop on the beach. Nobody's actually solved the problem in ninety years, they've just changed the venue.
Daniel worked for Sir Ronald Cohen, you know — the philanthropist. Granular detail, weekly meetings. So when he mentions Cohen in a question about executive pay, there's history there.
And that's exactly the seam I want to press on, because the pay data first — then the Cohen correction, which I suspect is the most useful thing we'll say today. Give me the trend.
The headline series is the Economic Policy Institute's CEO-to-worker compensation ratio, and it was just updated in September. The realized ratio — what CEOs actually took home after stock compensation was sold — was twenty point six to one in nineteen sixty-five. Thirty-one to one in nineteen seventy-eight. In twenty twenty-four: two hundred and eighty-one to one.
Nine times wider in a working lifetime.
And the direction is completely uncontested. Every source, whatever its politics, shows the same slope. But the magnitude is a mess, and this is the part worth slowing down on. EPI's own twenty twenty-four ratio, depending on methodology, ranges from a hundred and sixty-nine to one to three hundred and eighty to one.
A factor of two inside one organization's own number.
The main split is realized versus granted. Realized counts what the stock was actually worth when it was sold; granted counts what it was valued at on the day it was awarded. In twenty twenty-four that's two hundred and eighty-one versus about two hundred and twelve. And if you chart the two series over time, the granted line crosses above the realized line in the mid-nineties — you can see the stock-option boom in the data the way you see a flood line on a wall.
So when someone at a dinner party says CEOs earn three hundred times a worker, the straight answer is somewhere between a hundred and seventy and three hundred and eighty, depending on when you count the stock.
The peak on the realized measure was four hundred and eight to one, in twenty twenty-one. Absolute numbers: average realized CEO compensation at the three hundred and fifty largest US firms was nearly twenty-three million dollars in twenty twenty-four, up about six percent on the year. EPI calculates CEO pay has grown forty-two times faster than the typical worker's since nineteen seventy-eight.
Forty-two times faster. My leaf medicine appreciates slower than that and I still insist it's a growth asset.
The trajectory is worth hearing as a slope, because it isn't gradual. It's roughly thirty in the late seventies, fifty by the mid-eighties, seventy-five by nineteen ninety, then it more than triples across the nineties — a hundred and seventeen in ninety-five, two hundred and twenty-three by ninety-seven, past three hundred by ninety-eight.
The nineties did that.
The stock-option culture did it, mostly. Compensation committees started paying in options because accounting rules made them look free, and the ratio detached from anything it had historically tracked. This is where Stiglitz comes in — this is the inequality critique I can actually source. Joseph Stiglitz, in a Chatham House lecture, attributed the CEO pay explosion to "rent-seeking from monopolies to corporate governance." His argument is that the pay isn't the market clearing at the value of talent — it's insiders setting the terms of their own evaluation, with monopoly profits and compliant boards doing the rest.
Which, whatever you think of his remedies, does explain why a number can grow nine-fold without a nine-fold change in the difficulty of the job. A job doesn't get nine times harder in a generation. The work of running a company in nineteen ninety-eight was not nine times the work of running one in nineteen seventy-eight.
Now the transparency question, because there's a parallel here I don't think gets drawn enough. The Atlantic made the point in twenty fifteen: the US has required companies to disclose CEO pay since the nineteen thirties. Pay soared anyway. Disclosure exists — it has existed for ninety years — and it changed nothing about the slope.
Because the categories are controlled by the institution being disclosed.
Same shape as the vacation press strategy. The Trump aides disclosed the work and withheld the golf; the companies disclose the pay in categories their own compensation consultants designed. In both cases there is transparency, and in both cases the disclosed picture is curated by the party it reflects on.
We did an episode on this exact gap, actually — the Human Development Index one, where Mahbub ul Haq wanted a number to dethrone GDP and got a compromise metric instead. Same lesson: once the measurement becomes the famous thing, the fight moves inside the measurement.
The EPI ratio has had exactly that happen to it. It's quoted as if it were one hard number, and it's a family of numbers with a direction in common.
Which takes us to Cohen, and I want to be careful here, because I haven't seen — anywhere — Ronald Cohen making a specific call about executive pay framed as reining in unfettered capitalism. Not in his G20 piece, not in the summit speeches, not in the coverage of either.
And I'd be surprised if we did, because it doesn't sound like him. What Cohen actually argues is what he calls the impact revolution. To the G20, in twenty eighteen, he wrote: "The existing social contract has expired — we need to draw up a new one. It's time for a new financial model that values impact alongside risk and return." And at the GSG Impact Summit in Buenos Aires the following year he called for "overthrowing the dictatorship of profit and putting impact by its side."
Overthrowing the dictatorship of profit — that's a strong sentence for a man whose reform program is about accounting.
That's the misreading, though, and it's an easy one to make. "Overthrowing" sounds like capping, and capping is what people hear. But his actual mechanism is that every investment decision should carry three numbers instead of two — risk, return, and impact — so that capital flows change because the measurement changed. He's not saying pay executives less. He's saying measure companies on something besides profit, and let the pay follow from that if it follows.
So the reform that gets attributed to him — rein in executive pay — and the reform he actually proposes — change what companies are measured on — are different programs that happen to rhyme. One caps the output. The other rewires the scoreboard.
And conflating them matters, because the scoreboard version is arguably the more radical one. A pay cap is a ceiling on the old game. Impact-weighted accounts change the game. Cohen has spent decades on the second thing — he basically built the social investment banking field out of the ashes of his first career.
The pay-ratio critique belongs to EPI and Stiglitz. The measurement critique belongs to Cohen. Two arguments, constantly shipped in the same box.
It's the friendly version of a very old error — attributing to the famous name in the neighborhood whatever the neighborhood is saying.
Well, given who Daniel worked for, I'm glad we checked before we credited it. That's the whole correction, and it's worth making: the strongest documented voice on inequality in impact investing is arguing something more ambitious than the thing he's quoted as saying.
Which of you said the assistant decides what the boss knows?
I did. I stand by it.
Close, but you're one rung off. I did that job for a fabric executive in Manchester for four years — scheduling, not gates. The person who runs the calendar's colors is the one who decides what the boss knows, because the boss stops reading anything, he just trusts the colors. I ran his entire week out of seven colors and he never once asked what a category meant. Cornelius, you'd have liked him — he explained everything in halves too.
I explain things at the correct speed.
The green blocks on his calendar were labeled strategic thinking. They were naps. He had a chaise in the corner office and a blanket from the car. Forty minutes, twice a week, green as grass on the sheet. Cough on track three, Cornelius, I'll take it out in the edit. The blanket stayed in the car so the office never smelled of it — that was my system. I still keep the color legend in my wallet. Laminated. The man from the stationery shop on Deansgate did it for one pound fifty and never asked what it was for.
Seven colors for a whole week. What were the other six, or is that classified?
Calls, mill visits, finance, personnel, travel, and one I called "miscellaneous" that was mostly the owner's brother, who could not be filed under any honest heading. The genius of the system was that nobody ever queried a color. A man who would interrogate a supplier over a half-penny a yard on cotton would look at a green block marked strategic thinking and nod like it was scripture.
You're saying the nap was the schedule working as designed. You built deliberate slack and labeled it in the one color no one questions.
Strategy. Nobody audits strategy. The rest of the sheet was true — his calls, the mill visits, the lot. Only the green lied, and it lied in his favor. He came back from every nap with an opinion about the Portuguese order. Some of them were good opinions.
One pound fifty for the lamination. The most cost-effective management tool of the twentieth century.
It's held up better than the company, that's certain. My claws have cracked the corners and I can still read every category. Try doing that with a touchscreen.
So the one honest number in his whole setup was the naps.
The green was the only thing on that calendar that happened exactly when it said it would.
Which is more than the White House press office can claim.
It does reframe the study, doesn't it — seventy-five percent of a CEO's time pre-scheduled, and the reliability of the label depends entirely on whoever holds the legend.
If you take one thing from this episode, let it be that the day off at the top is never an absence — it's a pre-packaged continuation. The Book for the president, the chief of staff and the colored calendar for the CEO. The brief follows to Hawaii because it can't be batched; everything else was batched the night before. Rest, at that altitude, is a reallocation of attention with a press strategy attached.
The sharpening on it: whoever controls the categories — what counts as work, what counts as strategy, what counts as pay — controls the picture, no matter how transparent everything looks. The Trump aides, the compensation consultants, and Hilbert's green blocks are all the same trick at different salaries.
And that's before you get to the question I'd actually want answered next: whether the chief-of-staff role spreading through the C-suite eventually gets a definition, or whether it stays, in Steven Zhou's words, the one C-suite title nobody can explain. My money is on nobody.
The realized-versus-granted fight isn't going anywhere either — every time a stock swings, the two series diverge again and the headline number gets re-litigated. The direction's settled. The magnitude is permanent theater.
Two brothers and an anteater, all diagnosed with the same condition: we can't stop watching the machinery. Thanks to Hilbert Flumingtop at the desk — the legend really is laminated, I've seen it. If you enjoyed this one, try episode twenty-four eighteen, The Lossy Compression of Human Development; and episode thirty-five sixty-one, Working for the Ultra-Wealthy. This has been My Weird Prompts. If today's episode earned its twenty-five minutes, leave us a review wherever you get your podcasts, and the whole archive lives at my weird prompts dot com. See you tomorrow, Herman.
Tomorrow. I'll be scheduling it.