If Hitler Had Died in a 1930 Car Crash, Would World War II Still Have Happened?
Jared Diamond uses natural experiments to answer when leaders matter: CEOs explain roughly 30% of the variation in profits, and a sports coach's biggest effect isn't tactics — it's training players not to commit fouls.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Geographic determinism is not racial determinism
Diamond says he realized after one day in New Guinea that New Guineans are as intelligent as Europeans, or even more curious, so the differences in writing, metal tools, and states could not possibly come from brains. His answer is environment: the Fertile Crescent had the wild ancestors of the five most valuable animals — cattle, sheep, goats, pigs, and horses — plus wheat, peas, beans, and lentils, so its hunter-gatherers got a head start on agriculture 11,000 years ago, and only with food surpluses could they support full-time kings and inventors. Australia is the only continent that had no agriculture in modern times, because not a single Australian plant or animal has ever been successfully domesticated; European farmers tried to domesticate kangaroos and failed because of the kangaroo's social system.
— Jared DiamondEurope won by being divided; China lost by being unified
Diamond says Europe itself had no agricultural advantage — the only thing domesticated there was the poppy — but Europe sits right next to the Fertile Crescent, and its plants and animals reached Turkey by 8000 BC, then Greece, then spread out. What really let Europe win was geography: Europe faces the Atlantic and has the Mediterranean, the largest inland sea, as a practice ground for shipbuilding and navigation, whereas China had a fleet larger than Europe's between 1400 and 1432, then abolished the great fleet in 1432, because China had only one emperor, and the emperor thought these expensive fleets brought back nothing but white giraffes when China already had everything. Europe stayed politically divided because of its peninsulas, islands, radial rivers, and mountains, and division meant competition: if one prince said guns were too expensive and he didn't want them, the prince next door would invade.
— Jared DiamondWinning the Pulitzer got him denied a promotion
Diamond's doctorate was in medical physiology, he made his name studying the gallbladder, and later he studied birds in New Guinea. He says academia holds a belief: you should stick to what you know, and crossing into another discipline means you've run dry. At UCLA he was denied a promotion exactly once — right after Guns, Germs, and Steel was published and won the Pulitzer Prize. His colleagues in the physiology department had previously known only his gallbladder research, not his New Guinea bird research, and the promotion file said: Jared wrote this book about history and geography, but we can't evaluate it, so we won't consider it, and it noted that his New Guinea bird paper had a low citation count — that paper had been cited 49 times, while the people studying the 73rd amino acid of beta-galactosidase had 2,000 citations.
— Jared DiamondThe Danish CEO bereavement experiment: a relative dies, profits fall
Diamond says you can't actually go kill a CEO's relatives, but relatives die anyway, and that's a natural experiment. Denmark has about 78,000 companies, the government keeps public statistics on all companies and CEOs, and hospitalization or death in hospital is a public record. Business school professors entered all this data and checked whether each company's CEO had a child, spouse, sibling, or parent die, and found that after the death of a close relative, company profits suffer for a year or two, because the CEO is distracted. The biggest effects come from the death of an only child or a spouse, then siblings, and the smallest from the death of a parent; the only death that improved company profits was the death of a mother-in-law, but it wasn't statistically significant.
— Jared DiamondCEOs explain about 30% of profit variation
Diamond cites research by a social scientist, Allison Mackie, who looked only at samples where a CEO moved from one company to another, which separates the CEO effect from the company effect. Her conclusion is that CEOs explain about 30% of the variation in profits, industry and the specific company each about 20%, and the year effect averages only about 3%. If you relax the requirement that the CEO must change companies, the database gets bigger and the CEO effect drops to about 15%, with industry and company each at about 20%. So the CEO, industry, and company effects are of comparable magnitude, and the year effect is the smallest — except, of course, for 1930, 2002, and 2008.
— Jared DiamondA coach's biggest job is teaching players not to foul
Diamond says coaches get fired constantly for losing; the Argentine football league has about 400 firings a year, and German coaches last an average of only a year and a half, but losing has a lot of random factors, and improvement the next year after a firing is often just regression to the mean. Once you use analysis of variance to separate the effects of coach, year, industry, and company, the conclusion is that coaches do matter, but more in college sports than in professional sports, because pro players stay longer — Tom Brady spent 18 years under Bill Belichick, and only when Brady went to Tampa did it become clear that the Super Bowl success had a lot to do with Brady. And the biggest effect an American football coach has on outcomes isn't deciding when to pass, it's training players not to fumble and not to commit fouls.
— Jared DiamondReligious leaders need three different skill sets
Diamond says the biggest difference between religion and other domains is that religious leaders must be charismatic, whereas in politics and business charisma sometimes matters and sometimes doesn't — Hitler and Churchill were charismatic, Chile's Pinochet was not, nobody would call him charismatic, and you can be an effective political leader by making people afraid or by being good at guiding group discussion. But for a religion to succeed and spread worldwide, charisma alone isn't enough; it also needs organizational ability, and often a military organization, because religions get attacked as they grow: Mormonism was attacked as it arose, and Joseph Smith built a Mormon army to defend and to attack non-Mormons; early Christianity was a small sect of Judaism, too small to even talk about building an army, and it didn't have one until centuries later. So a new religion needs three roles, possibly filled by the same person: Jesus was the charismatic figure, Saint Paul was the great organizer, and only much later came Christian generals; in Mormonism, Joseph Smith was the charismatic figure and Brigham Young was the great organizer.
— Jared DiamondThree kinds of leaders: those who pick the moment, those the moment picks, and those who stumble into it
Diamond sorts the relationship between leaders and timing into three categories. The first can recognize and choose the moment: de Gaulle was president of France twice, and he stepped down less than two years after World War II because he judged the timing wrong, then in 1958, during the Algerian crisis, he said his moment had come; Bill Gates and Steve Jobs both saw the Intel microchip on the cover of Popular Electronics in 1976 or 1978, and Gates had been telling Paul Allen the time wasn't right until he saw that cover and said now was the time. The second can't manufacture the moment but knows when it lands on them: Churchill wrote in his memoirs that on the night he was made prime minister, he knew this was the moment he had prepared for his whole life. The third happens to be in the right place at the right time and has no idea: Genghis Khan, for instance — dendrochronologists found that the wettest period in Central Asia in the past 2,000 years fell right around his birth, and wet meant plenty of feed for sheep, goats, and horses, and every Mongol warrior needed 10 horses; had he been born 20 years later or 200 miles to the south, we would never have heard of Genghis Khan.
— Jared DiamondIn their own words · checked verbatim
Paradoxically, though, unity has been a disadvantage, and disunity in Europe by stimulating competition has been an advantage.
Jared Diamond16:30
There was only one time when I was denied a promotion at UCLA, and that was immediately after 'Guns, Germs, and Steel' was published and won the Pulitzer Prize.
Jared Diamond24:30
And if the mother-in-law dies, the only death in the family that improves the profits of the business is the death of mother-in-law, but it's not statistically significant, and so it's unsure.
Jared Diamond39:10
The claim is not that the most important part of winning is having fewer penalties, it's that where coaches have the most leverage is in training their teams to have fewer penalties.
Jared Diamond1:03:23
But in the world of religion, you have to be charismatic.
Jared Diamond1:04:59
on that evening when I was offered the prime ministership, I knew that this was the moment for which my entire life was a preparation.
Jared Diamond1:10:12
But in my book, I explain the magic word is discretion. Discretion not in the sense of being discreet and polite, but discretion in the sense of having the power to make decisions.
Jared Diamond1:13:30
Figures
| Number of Danish companies | about 78,000 | 39:10 |
| Share of profit variation explained by the CEO (CEO-changes-company sample) | about 30% | 40:56 |
| Share of profit variation explained by industry | about 20% | 40:56 |
| Share of profit variation explained by the specific company | about 20% | 40:56 |
| Share of profit variation explained by year | about 3% | 40:56 |
| Share of profit variation explained by the CEO (relaxing the CEO-changes-company condition) | about 15% | 40:56 |
| Coaching firings per year in the Argentine football league | about 400 | 57:06 |
| Average tenure of a German football coach | about a year and a half | 57:06 |
| Years Tom Brady played under Bill Belichick | 18 years | 57:06 |
| Voter turnout in the Los Angeles mayoral election | 20% | 1:01:00 |
Glossary
- natural experiment
- Using differences that occur by chance in the real world (such as a leader's accidental death) to infer causation, rather than manipulating variables deliberately.
- ANOVA
- A statistical method for splitting the total variation in an outcome into different sources such as leader, year, industry, and company.
- regression to the mean
- Extreme performance tends to return to average afterward, so improvement after a coaching change isn't necessarily the coach's doing.
- dendrochronology
- Inferring historical climate from the width of tree rings, which Diamond uses to show Central Asia was unusually wet when Genghis Khan was born.
How to listen
Founders and investors interested in historical causation, leadership research, and using data for causal inference — especially anyone who wants to understand how big the CEO effect actually is.
The host's introduction and the Guns, Germs, and Steel retrospective from 0:00-5:24 can be fast-forwarded; the new book's material starts at 28:43.