Tap right to skip · left to go back · hold to pause
I
The fixed-pie intuition
Start with the idea we want to test. It is simple, and within its own assumptions the math is airtight.
Premise
Picture the economy as a pie of fixed size, shared by five people. Each gets 20.
Now the gold slice grows to 40. With the pie fixed, those 20 extra have to come from somewhere, so everyone else drops to 15.
Σ Δ = +20 − 5 − 5 − 5 − 5 = 0
Add up every change and you get exactly zero.
Zero-sum
This is a zero-sum game. One person getting richer requires others getting poorer, and inequality becomes a record of who took from whom.
"The rich get richer and the poor get poorer" follows directly from the premise. Much socialist analysis starts here, with wealth as a fixed stock to be divided or extracted.
The question
Everything hinges on the premise. Let the pie grow to 250 while the shares stay unequal.
Gold now holds 100. Every other slice is 37.5, almost double what each person had under perfect equality. The sum of changes is +150.
So where does growth like this come from, and what stops it? Start with the smallest possible economy.
II
One trade
Two people and one swap. Nothing gets produced, and total value still goes up.
Ana has a fish. Ben has a loaf of bread.
Value is personal, so write down what each item is worth to each of them, in any unit you like. Ana's fish is worth 2 to her. Ben's bread is worth 3 to him. Total value in this tiny world: 5.
Ana would rather have bread, which is worth 5 to her. Ben would rather have fish, worth 6 to him.
Each values the other's item more than their own. That gap is the whole opportunity.
They swap. No fish was caught and no bread was baked. The world contains exactly the same objects as before.
Yet Ana now holds 5, Ben holds 6, and total value went from 5 to 11.
Σ Δ = +3 + 3 = +6
The core idea
Nobody agrees to a trade they expect to lose on. So every voluntary trade is, by construction, expected to leave both sides better off.
Wealth is not only stuff. It is stuff in the hands of the people who value it most, and trade is how it gets there.
Try it
Who gets the surplus?
Ben won't sell his bread for less than 3 coins. Ana won't pay more than 5. Drag along the price line or use the slider.
Anywhere between 3 and 5 the deal happens and the total gain is always 2. The price only decides how that gain is split. Outside the range one side walks away and the gain never exists.
This is what a price ceiling does. Cap bread at 2.50 by law and the trade stops: Ben keeps the loaf, the value is never created, and neither person is better off.
III
Specialization
Swapping existing things is half the story. Trade also changes what gets made.
In a day Ana can catch 6 fish or bake 2 loaves, or any mix along this line. The triangle is everything she can produce on her own.
Splitting her day in half, she ends up with 3 fish and 1 loaf.
Ben is the opposite: 2 fish or 4 loaves. Alone he settles at 1 fish and 2 loaves.
Between them they produce 4 fish and 3 loaves a day.
Now each does what they are relatively good at. Ana only fishes. Ben only bakes.
Combined output jumps to 6 fish and 4 loaves. Same people, same hours, more of both goods.
Beyond the frontier
Ana trades 2.5 fish for 1.5 loaves. She ends at (3.5, 1.5), Ben at (2.5, 2.5).
Both points lie outside their own triangles. Each now has more than they could ever produce alone. Now multiply this by eight billion people and millions of goods.
Try it
Comparative advantage
Change what each person can make in a day. Try making Ana better at both.
1 loaf costs Ana3.0 fish
1 loaf costs Ben0.5 fish
Gain for both+38%
Each gets 38% more of both goods than they could make alone.
Gains disappear only when both people face exactly the same trade-off between fish and bread. Being better at everything doesn't matter. What matters is what each person gives up. David Ricardo worked this out in 1817, and it remains one of the least intuitive and most robust results in economics.
IV
The planner's dilemma
Suppose a well-meaning planner divides everything perfectly equally. Can anyone improve on that without producing a single extra thing?
Experiment
Equal shares, then free trade
48 people, four goods. Everyone starts with 3 of each. Each person has private tastes that nobody else can see. Press Allow trade: people swap one item at a time, and only when both sides gain.
Equal shares. No trades yet.
Total wellbeing100.0
Swaps made0
Better off than equal0 / 48
Worse off than equal0 / 48
New goods produced0
Starting from equal shares, trade can't make anyone worse off, because nobody accepts a swap that hurts them. Wellbeing still climbs, and the final holdings are unequal because people wanted different things.
The planner's best guess gives everyone one ideal bundle based on average tastes. It barely helps, and it hurts people with unusual tastes. The information it needs sits in 48 separate heads. In a real economy it sits in millions of heads and changes every day. Chapter VIII comes back to this.
Model: wellbeing = Σ taste × √quantity, tastes drawn at random per person. Swaps are one-for-one and are made only if both people gain.
V
Two worlds, many rounds
Now repeat the game thousands of times. Two toy economies of 300 people each start from the same modest differences. Every round, people pair up at random.
Simulation
Zero-sum versus positive-sum
Top: each encounter is a coin flip that moves wealth from one person to the other. Nothing is created. Bottom: the same coin flips, plus each encounter creates a small surplus that both sides keep. Dots are colored by starting wealth, poorest to richest.
Round 0 / 800
In the zero-sum world the rich really do get richer and the poor poorer. Wealth drains upward by pure chance, because a string of losses is hard to recover from. This is the world the fixed-pie picture describes, and inside it the picture is correct.
Add a 1.5% surplus per encounter and the story turns around. Luck still makes winners and losers, and the richest pull far ahead. But the poorest tenth ends up many times richer than where it started. Try lowering the slider: below about 1%, bad luck outpaces the gains for the unluckiest. At 0 you are back in the fixed pie.
Toy model, not a forecast. Stakes are 10% of the poorer party's wealth; the surplus is a share of the poorer party's wealth, added to both sides. The zero-sum version is the "yard-sale model" from econophysics (Hayes 2002; Boghosian 2019).
VI
Levels, not shares
Debates about inequality usually compare shares of the pie. People live on the size of their slice.
World A is perfectly equal: a pie of 100 split five ways. World B is ten times bigger and very unequal.
B's poorest person gets only 8% of the pie. That is 80, four times what anyone gets in A. Measured by shares, B looks worse. Measured by what people have, B is better for every single person.
The real pie
In 1820 about three in four people lived in extreme poverty. Today it is roughly one in ten, out of a population more than seven times larger.
The number of people living above extreme poverty went from about a quarter of a billion to over seven billion.
Our World in Data (Moatsos 2021 for 1820; World Bank for today). Rounded.
In 1800, earning an hour of reading light by tallow candle took about six hours of work. With an LED bulb today it takes about half a second.
A king in 1700 couldn't buy what an ordinary worker now doesn't notice paying for.
Nordhaus (1996); Ridley (2010). Today's figure: 10 W LED at average European wages and power prices.
In 1800 about 43 of every 100 children died before their fifth birthday. Today about 4 do.
This is what the pie growing means in human terms. None of it shows up in a chart of shares.
Our World in Data, child mortality (Gapminder, UN IGME).
VII
History's controlled experiments
Economists can't run experiments on whole countries. History ran a few anyway: same people, same culture, split into different systems.
Korea
Divided after 1945. In the early years the North was the more industrialized half. The South turned to export-oriented markets, the North to central planning.
Today the South produces roughly 30 times as much per person.
Bank of Korea estimates of North Korean GNI; ratios vary by method.
Germany
One people, split in 1949. By 1989 East German productivity was around a third of West Germany's.
An East German who ordered a Trabant typically waited 10 to 15 years. In the West, a new car took weeks.
China
In 1981, 88 of every 100 people in China lived in extreme poverty. Market reforms had begun in 1978. By 2019 the figure was below 1 in 100.
Close to 800 million people left extreme poverty, the largest reduction in history.
World Bank (2022), Four Decades of Poverty Reduction in China.
Venezuela
After years of expropriations, price controls and currency controls, the economy shrank by roughly three quarters between 2013 and 2021. About 7.7 million people, around one in four, left the country.
Oil prices fell in 2014 for every oil exporter. None of the others lost most of its economy.
IMF; UNHCR / R4V platform.
Friedman's second claim
Command economies were never equal in practice. Party elites had special shops, dachas and closed compounds like Wandlitz near Berlin while ordinary people queued.
When the state owns everything, access to the state becomes the only real wealth, and it concentrates at the top.
VIII
Why planning fails
The pattern repeats for a reason no planner can fix with better intentions. Friedrich Hayek called it the knowledge problem (1945).
An economy is a web of people who each know things nobody else knows: a local shortage, a better technique, what a customer actually wants.
No single mind holds more than a sliver of it.
A tin mine floods. Tin is suddenly scarce.
Hayek's own example. Who should use less tin, how much less, and what should replace it?
Markets
The price of tin rises. That single number travels through the web. Everyone who uses tin economizes, and some switch to alternatives.
None of them needs to know why. The price carries the information and the incentive to act on it in one signal.
Planning
A planner must first collect every report, process them centrally, and send orders back out.
By the time the orders arrive, conditions have changed. Reports pile up in a queue while shortages spread.
Feedback
Markets also correct their mistakes. Losses tell a business it is wasting resources. Profits reward whoever does better.
A planned economy has no such signal, so errors can persist for decades. That is the mechanism behind the numbers in the previous chapter.
IX
Compounding
Small differences in growth, sustained, become enormous. The gap between Seoul and Pyongyang is a growth rate held for sixty years.
Try it
The rule of 70
Income per person grows by g percent a year. After t years it has multiplied by (1 + g)t, and it doubles about every 70 / g years.
Doubles every35 years
After one lifetime (80 y)×4.9
After a century×7.2
Long-run growth per person in the US and Western Europe has been roughly 2% a year. South Korea managed about 6 to 7% for three decades. A country stuck near 0% stays where it is while its neighbors multiply.