Image: David Stanley from Nanaimo, Canada, CC BY 2.0The rule of 70 and why China's forty years were different
A country whose income per head grows by 2 per cent a year doubles it in 35 years. A country growing at 7 per cent doubles it in 10. Over a lifetime of 70 years the first country becomes twice as rich; the second becomes 128 times as rich. The arithmetic behind those sentences is the rule of 70, and it is the single most useful piece of mental mathematics in economics.
Where the 70 comes from
Something that grows at a constant rate g per year has multiplied by (1 + g)t after t years. It has doubled when that factor equals 2, so the doubling time is ln 2 divided by ln(1 + g). The natural logarithm of 2 is 0.693, and for small rates ln(1 + g) is very close to g itself. The doubling time is therefore about 0.693 divided by g, and if g is written in per cent rather than as a fraction, that is 69.3 divided by the percentage rate. Rounding 69.3 up to 70 makes the division easier and introduces an error of about 1 per cent, which is smaller than the error in most growth rates. Some people use 72 instead, because 72 divides neatly by 2, 3, 4, 6, 8, 9, and 12; the choice is a matter of taste, and Unit 1 of the masterclass uses 70 in the text and 69.3 in its derivation box.
The rule works in both directions. If you know how long something took to double, 70 divided by the years gives the growth rate. World population doubled from 4 billion in 1974 to 8 billion in 2022, 48 years, so it grew at about 1.5 per cent a year on average over that stretch. It also works for anything that compounds: a savings account at 3 per cent doubles in about 23 years, and a price level rising at 2 per cent a year, which is the inflation target of the Fed, the ECB, and the Bank of England, doubles in 35 years, so a person who retires at 65 with a fixed pension and lives to 100 sees its purchasing power halve.
Why the log scale is the honest picture
Compound growth is why economists draw long series on a logarithmic scale. On an ordinary scale, a quantity that grows at a constant rate curves upwards ever more steeply, and the eye reads acceleration where there is none. On a log scale, in which each step up the vertical axis is a multiplication by the same factor, constant growth is a straight line, and the slope is the growth rate. Two countries growing at the same rate are parallel lines; a country growing faster is a steeper line; and a doubling is the same vertical distance whether it happens at an income of 1,000 dollars or 40,000 dollars. The long charts in Our World in Data, built from the Maddison Project Database, let you switch between the two scales with one click, and it is worth doing once to see how much the picture changes.
China's forty years
In 1978 China began the reforms that dismantled the collective farms, opened the coast to foreign investment, and let prices do work that planners had done. According to the World Bank's national accounts data, China's real GDP grew at an average rate of about 9.5 per cent a year between 1978 and 2018, and GDP per head at about 8.5 per cent a year, since the population was also growing. Put 8.5 into the rule of 70 and the doubling time is a little over 8 years. Forty years contain almost five such doublings, and 2 to the power of 4.9 is about 30. The World Bank's series for GDP per head at constant prices rises by a factor of roughly 25 over those forty years, which agrees with the arithmetic once one allows for the years in which growth was below the average.
Here is the same number from the other side. In 1978 China's income per head was about one fortieth of that of the United States at purchasing power parity according to the Maddison Project Database, and about a quarter by 2018. The United States grew at about 1.7 per cent a year per head over the same stretch, a doubling time of about 41 years. The gap closes at the difference in the two rates, roughly 7 percentage points a year, which gives a halving time for the ratio of about 10 years. Four decades, four halvings, and a ratio of 1 to 40 becomes a ratio of about 1 to 2.5, which is the order of magnitude observed. No other large country has done this in the span of a single career; Japan between 1950 and 1990 and South Korea between 1960 and 2000 came closest, at per head growth rates of 6 to 7 per cent.
Why it was different
Three things made the Chinese numbers possible, and all three are the subject of Unit 8 of the masterclass. The first is the starting point. A country that is very poor relative to the frontier can grow fast by adopting technology and methods that already exist, which is the convergence the Solow model predicts for an economy far below its steady state; a country at the frontier has to invent. The second is capital. China's saving and investment rates were among the highest ever recorded, above 40 per cent of GDP for decades, which raised capital per worker at an extraordinary pace. The third is the movement of several hundred million people from farms to factories and cities, which raised output per worker simply by moving workers from low-productivity to high-productivity work. Institutions matter too, as the research spotlight on Daron Acemoglu's work in Unit 8 explains, and the Chinese case is the one that every theory of institutions has to accommodate, since the rules that delivered the growth were far from the textbook list.
The rule of 70 also says why the fast decades had to end. Convergence slows as the gap closes, the supply of farm workers to move runs out, and the return to building yet more capital falls. China's growth rate of GDP per head in the 2020s has been around half of what it was in the 2000s, so the doubling time has roughly doubled, to about 15 years. That is still fast by the standards of rich countries, where 1.5 to 2 per cent a year, a doubling every 35 to 45 years, has been the norm since the war.
Try it yourself
Take any growth rate you read this week and turn it into a doubling time. A pension fund that promises 5 per cent a year is promising to double your money in 14 years. A city growing at 3 per cent a year will need twice the housing in 23 years. An economy forecast to grow at 1 per cent will take two generations to double its income. The rule converts rates, which nobody can feel, into spans of time, which everybody can, and that is why it is the first number worth knowing on the masterclass's key models page.


