Two boardrooms, one hundred and thirty-two years apart, running the same arithmetic.
May 14, 2026
Chicago, August 27, 1894. George Pullman sits before the federal Strike Commission convened by President Grover Cleveland. He is months past sending his workforce home, weeks past the federal troops, days past the last funeral. The commissioners ask him about the wages. They ask him about the dividend. He answers both questions calmly, because the answers are not, from his seat, in conflict. They are the same answer.
The Pullman Palace Car Company was capitalized at thirty million dollars. It built sleeping cars — the long, plush railroad cars that turned cross-continent travel from punishment into a marketable luxury. By the early 1890s it was the dominant supplier in North America. It also owned a town. Pullman, Illinois, fifteen miles south of Chicago's central business district, was built on company-owned land. The houses, the streets, the gas works, the water mains, the public school, the library, the church, the hotel, the savings bank — all sat on Pullman's balance sheet as real-estate assets earning rent.
In the spring of 1893, the American economy stopped working. The Panic of 1893 closed banks, collapsed railroad bond prices, and within months reduced new orders for Pullman sleeping cars by more than half. The company's directors faced a contraction of the kind that breaks businesses with thinner margins. They responded with a sequence of decisions that, taken together, ended in federal troops on the streets of Chicago a year later.
Between September 1893 and May 1894, the company cut wages in stages. The cuts ran from twenty-five to forty percent depending on the department, applied unevenly across skill grades and trades — heavier on the piece-rate workers in the freight-car shops, lighter on the salaried draftsmen. The aggregate wage envelope contracted by an estimated one and a half to two million dollars over those eight months.
During the same eight months, the company continued to pay an eight percent annual dividend on its thirty-million-dollar capital stock — some two and four-tenths million dollars distributed to shareholders, in quarterly installments, drawn against the same accounts the wage cuts were filling.
The dividend was not an oversight. It was a board decision, recorded in the company minutes and subsequently entered into evidence before the Strike Commission. The wage cut and the dividend were not two separate stories about Pullman in 1893 and 1894. They were the same story, viewed from two sides of the income statement.
The rents told the third side. Pullman's workforce lived, in substantial part, in Pullman houses. Rents had been set at the level the directors considered necessary to produce a six percent return on the company's real-estate investment. When workers asked the company to reduce rents in line with the wage cuts, the company refused. A reduction in rent would have compressed the revenue line of the real-estate segment on the same income statement whose manufacturing segment was already under pressure. The refusal was not a moral position. It was a ledger position. The workers paid the unchanged rent out of the reduced wages and watched the difference disappear into a column on a page they could not see.
Then they walked off.
The strike began at Pullman in May 1894 and spread through the American Railway Union — a young industrial federation led by Eugene Debs — to the railroads themselves. Within six weeks, a quarter million workers across twenty-seven states refused to handle trains carrying Pullman cars. Cleveland mobilized federal troops, deputized marshals, and reinforcing state militia — some twelve thousand armed men in all — citing a presidential responsibility to keep the mail moving. Thirteen people died in the Chicago clashes alone, with further deaths reported along the western rail lines. Debs went to federal prison on an injunction-violation charge. He read Marx, Kautsky, and Bellamy in his cell. His socialism, already forming, hardened there.
This is the part of the story that is usually told. It is in every American labor-history textbook. It is in Almont Lindsey's 1942 book on the strike, still the standard scholarly account. It is in Nick Salvatore's biography of Debs. It is the dramatic version.
What it usually leaves out is the math.
The math is the testimony in front of the Strike Commission, in August 1894, in Chicago. The three commissioners appointed by President Cleveland pulled the company books into evidence. They asked Pullman whether the dividend had continued during the wage cuts. He confirmed that it had. They asked whether the rents had been reduced when the wages were cut. He confirmed that they had not. They asked him whether he considered the decisions defensible.
He said he did.
He was telling the truth. That is the part of the story that matters.
Sit in the boardroom in late November 1893, four months into the panic, six months before the strike.
The orders for new sleeping cars have collapsed. The company's bond covenants presume a continuing capital return at a level the dividend reflects. The capital stock — thirty million dollars on the books — represents the savings of several thousand investors, many of them banks and insurance companies that have placed Pullman securities in their reserve portfolios. A decision to cut the dividend would not stay inside the company. It would propagate outward, repricing Pullman paper, triggering covenant clauses, and likely raising the company's cost of capital in any future borrowing. From the directors' seat, cutting the dividend is not the same act as cutting wages. The two are different in legal form, different in financial consequence, different in audience.
The options the structure actually offers are four. Cut the dividend and break faith with the shareholders. Cut the wages and break faith with the workforce. Cut both. Liquidate.
The unwritten rule of late nineteenth-century industrial finance — internalized so deeply the directors never needed to vote on it — was that capital return came first. Dodge v. Ford, the Michigan Supreme Court ruling sometimes read as formalizing shareholder primacy as a fiduciary duty, was still twenty-five years off. The directors were not breaking a law. They were applying a norm no one in the room thought of as a choice.
From inside that norm, cutting wages while maintaining the dividend was not an extreme answer. It was the boring one — the answer a director who had read the bond covenants and the company minutes was expected to give.
The town made it cleaner. Pullman, Illinois had been built as a financial product as much as a community. Stanley Buder's history documents the explicit calculation: a workforce housed on company land would be more stable; the rental income would generate a separate return; the moral environment of the planned town would dampen the trade unionism the founder considered a threat. Under expansion, the town's arithmetic added to the factory's. Under contraction, the same architecture turned brittle — the housing line was not a sentimental ledger.
This is what is meant by saying the wage cut was defensible. The word does not mean "morally acceptable" or "kind" or "what the workers deserved." It means: a director who voted for it could explain the vote from the seat he was sitting in, against the standards by which his fiduciary performance was being judged, in the language of the system he was operating inside. That is what a defensible decision is.
Replace the directors of the Pullman Palace Car Company in November 1893 with twelve different people. Choose the twelve kindest, most thoughtful, most worker-conscious figures available in 1893 Chicago. Seat them at the same table. Give them the same order book, the same bond covenants, the same balance sheet, the same shareholders, the same set of options. The wage cut still arrives in May 1894. The dividend is still paid through fiscal 1894. The strike still breaks out. Cleveland still sends the troops. Debs still ends up in prison reading Marx.
The choice the directors made was not, in this sense, their choice. It was the choice the structure produced. They were the surface on which the structure expressed itself.
Carnegie's silence after Homestead, two years earlier, runs the same equation in a different industry. He was traveling in Scotland; he could have intervened; he did not. Frick handled it. When the choice is between capital return and labor return under an earnings contraction, the capital return wins. That was not Carnegie's character. It was the equation.
In February 2024, the Swedish financial-technology company Klarna announced that an Artificial Intelligence (AI) assistant integrated into its customer-service operation was performing the work previously done by roughly seven hundred customer-service representatives. The company's chief executive, Sebastian Siemiatkowski, framed the result as both a productivity story and a strategic story: the assistant was resolving customer inquiries faster than human agents, at a fraction of the cost, and the savings would flow through to the company's financial performance ahead of its planned public listing.
Take Klarna's number with the obvious grain of salt. A company that sells financial technology and is preparing an initial public offering has an interest in making its automation story sound transformative. The seven-hundred figure was sourced to internal Klarna analysis, not an independent audit. In the months that followed, Financial Times reporting noted that the company had partially walked back the framing — some of the work had migrated to outsourced agents, some had been reabsorbed, and the headcount narrative was more complicated than the original announcement suggested.
The walk-back does not change the structure. It changes the timeline. The work that Klarna's AI assistant did not handle in 2024 will be handled by a more capable system in 2025 or 2026. The replacement is, in the language of the labor economists, a delayed substitution rather than a cancelled one. The arithmetic — labor cost contracts, capital return holds — is the same arithmetic Pullman's board ran in 1893, set against a different production function and a different century.
Klarna is one boardroom. The S&P 500 is five hundred boardrooms.
For the twelve months ending September 2025, the companies in the S&P 500 returned a record sum nearing one and seven-tenths trillion dollars to shareholders — approximately one trillion in share repurchases, six hundred and sixty-five billion in dividends — according to S&P Dow Jones Indices' quarterly buyback report. The figure is the largest annual capital return to shareholders in American corporate history. It is the dividend column of Pullman's 1894 income statement, scaled to the contemporary economy.
During the same year, the technology and finance sectors continued the layoff cycle that began in late 2022 — by industry trackers' counts, over half a million white-collar positions eliminated since the start of the cycle, with a steadily rising share attributed by the laying-off companies to AI-driven productivity gains. The numbers do not match up in any simple way. Many of the layoffs are unrelated to AI. Many AI-related savings have not yet shown up as headcount reductions. But the direction is what matters. Anthropic's own labor-market data, published in 2026, found that business and finance professionals face a theoretical automation exposure of 94.3 percent against a current actual rate of 28 percent. The gap between what AI can do and what AI is doing is roughly two-thirds of those professionals' work. That gap is the inventory of decisions still to be made by boards over the next five years.
Each of those decisions, in the boardroom where it is made, will look like the decision Pullman's directors made in November 1893.
A retail bank in Frankfurt or Charlotte will discover that its trade-finance operation can be run with one-third of its current headcount once a foundation-model-assisted workflow is integrated into the document-processing pipeline. The savings — some tens of millions of euros or dollars annually, depending on the bank — will be calculated and presented to a board. The board will not be asked, in plain English, "should we eliminate these jobs?" The board will be asked: "given that the productivity gain is real, what is the responsible deployment of the savings — wage retention, capital return, or reinvestment?"
The framing produces the answer.
Under quarterly earnings-cycle pressure, under buyback expectations baked into the share price, under activist-investor letters that arrive on the chief executive's desk if return on equity slips below the sector median, the responsible deployment looks like a buyback. It looks like a special dividend. It looks like the same answer Pullman gave the Strike Commission in 1894, expressed in a different vocabulary.
The 2026 boards are not stupider or crueler than the 1894 boards. In most measurable ways, they are more constrained. The shareholder-primacy norm that was unwritten in 1893 is now codified in jurisprudence, in compensation structures, in the corporate-governance manuals that sit on the directors' desks. The fiduciary frame Pullman's directors applied without thinking about it is now applied with thinking. The result is tighter, not looser.
Replace the boards. Take the kindest, most thoughtful, most worker-conscious figures available in 2026. Seat them at the same tables. Give them the same earnings cycles, the same buyback expectations, the same activist letters, the same compensation structures, the same set of options. The same AI-attributed cuts arrive. The same buybacks get authorized. The same column on the page holds the same number.
There is a moral verdict available here. It says: Pullman was greedy. His board was greedy. The S&P 500 boards are greedy. The 2026 directors who authorize the buyback over the wage retention are greedy. The verdict names the greed and moves on, leaving the reader with the satisfaction of correct judgment and no requirement to act on it.
That verdict is not wrong on the facts. George Pullman was, by most accounts of his contemporaries, a man whose private generosity did not extend to the people who built his railcars. Some 2026 directors are no doubt greedy in any sense the word can carry. The moral framing has the additional benefit of being emotionally available — most readers can produce the verdict without consulting any data.
The moral framing does not change the math.
Run the counterfactual once more, in both directions. Replace Pullman in 1893 with a person you would trust with your own savings. Replace every director of every S&P 500 firm in 2026 with people you would trust with the welfare of a town. Hold the bond covenants constant. Hold the quarterly earnings cycle constant. Hold the activist-investor leverage constant. Hold the buyback expectations constant. Hold the rule that says capital return precedes labor return in the order of operations.
The wage cut still arrives. The dividend still gets paid. The buyback still gets authorized.
The math has not noticed that the directors changed.
The structural rules under which boards make these decisions do not permit personal character to be load-bearing on the outcomes the essay is concerned with. A director who chose, against the bond covenants and the fiduciary norm and the activist letters, to retain the wage envelope at the cost of the dividend would not be saluted as a moral hero. He would be sued.
The political problem is therefore not the personnel. The political problem is the rule.
Specifically: the rule that says, when earnings contract, capital return is the protected line and labor return is the variable line. That rule was a default assumption inside the equity-finance norm of 1893. It is, in 2026, written into the compensation contracts of every Fortune 500 executive, the ratings methodology of every major equity analyst, the regulatory framework of every public stock exchange, and the implicit duty of every director who would prefer not to be removed.
It is a rule. It is not a law of nature. It is not the only possible rule. It is the rule the system is currently running.
If you want the math to come out differently, you change the rule. You do not change the people running it.
Pullman did not lose the argument in 1894. He lost the strike. The Strike Commission — known to historians as the Wright Commission, after its chair Carroll D. Wright — found that the company's labor policy had been "unwise and arbitrary." The dividend continued. The town continued. The federal injunction that broke the strike was upheld by the Supreme Court the following year, formalizing in law a tool the federal government had not previously held. The structural rule — capital return ahead of labor return in the order of operations — was not amended by the events of that summer. It was reinforced.
Which is why the 2026 board, looking at a thirty percent customer-service contraction routed through a language model, sees what Pullman saw in December 1893. Not the same workers. Not the same product. The same column on the income statement, holding the same number, defended by the same logic. The argument that changes that outcome is not the argument that Pullman or any of his analogs were bad people. It is the argument that changes the order of operations.