Take-Off in Convergence Terms

Evidence essay · Production badge

Economists have argued for seventy years about whether poor countries “catch up” to rich ones. Ask a sharper question — did a country close the distance to the richest economy in the world? — and the picture is stark.

In 1950 there was effectively one high-income economy: the United States. Only eight countries stood even halfway to it, and three of those were tiny oil states riding fixed exchange rates. Bretton Woods — the IMF and the World Bank — was built in that vacuum, and built to fix it: discipline the payments system, steer capital to the poorest, close the gap.

It didn’t close. Of the 46 countries that reached their high-water mark against the US under the Bretton-Woods fixed-rate system (1950–71), 42 have since fallen back below 30% of the US frontier — a 91% non-convergence rate. The countries that did catch up — post-Soviet Europe, then export-led Asia — did it after the fixed-rate system collapsed in the 1970s, under floating rates, and they were mostly countries the 1944 conference never had in mind.

The convergence the founders wanted was real. It just happened after their machine broke, and to strangers.

Horizontal bars of people living in 2025 in nations grouped by the monetary regime in which their real Net Worth per capita was closest to the United States: par value 1950 to 1971, 1.4 billion people in 46 nations; Smithsonian 1972 to 1973, 3 million in 2; float and oil shocks 1974 to 1984, 0.7 billion in 30; Great Moderation 1985 to 2007, 0.4 billion in 16; supercycle and crisis 2008 to 2014, 1.6 billion in 57; convergence 2015 to 2030, 3.8 billion in 55, most still rising
A bimodal century. Each nation sits in the monetary regime in which its real Net Worth per capita was closest to the United States, and each bar counts the people who live there today. The par-value years hold the periphery's high-water mark: 46 nations, 1.4 billion people. Few nations peaked in the floating-rate Great Moderation. The convergence bar is hatched because most of its nations, China and India among them, are still rising.

How the axis works, and the 1950 Club

Two different questions hide inside “take-off”: PPP take-off (Rostow, Maddison) asks whether the economy escaped Malthusian stagnation; US=100 take-off asks whether it closed the distance to the global frontier. PPP take-off tells you whether the economy is functioning; US=100 take-off tells you whether the world is converging. Bretton Woods’ framers conflated the two.

Lines of real income per capita for Korea, Hungary, Turkey, China, Brazil, Indonesia, India, Bangladesh and Pakistan as a percent of US income per capita in 1950, from 1950 to 2023, capped at 100 when reached, with dashed projections at 6 percent a year to 2050, and a grey line for the share of world population living above the US level of 1950
PPP take-off, the other question: real income per capita against the United States of 1950, capped at 100 when reached. Korea, Hungary, Turkey, Brazil and China have crossed; Indonesia, India, Bangladesh and Pakistan are on the way (dashed: 6 percent a year). Grey, right scale: the share of world population living above the US of 1950. Clearing that bar says an economy works; it says nothing about closing the gap with today's frontier, the question in the figure above. Source: Penn World Table 11.0.
Country1950Peak2030Verdict
Qatar440.31971159.1rentier — relative decline
Kuwait166.0195856.0rentier — relative decline
Switzerland81.42013100.3held the frontier
Canada71.7201362.0mild divergence
Sweden66.2198064.7plateau
Australia61.0201368.2plateau
Bahrain56.8197743.2rentier — relative decline
Denmark53.8200983.4late catch-up

Of the eight countries already at the top in 1950, only Denmark used the next eighty years to climb materially closer. The high-income set of 1950 was already, in convergence terms, a closed group.

Opposing reading, in fairness: a defender of the Bretton-Woods record would argue the par-value system’s job was macro-stability and reconstruction, not frontier-convergence, and that the post-1990 take-offs were only possible because of the trade and payments plumbing the earlier system laid down. The essay’s claim is narrower than “the institutions failed”: on the specific convergence metric the founders themselves invoked, the record is 91% non-convergence, and that fact is only visible once every country sits on one normalized balance-sheet axis.

Source: FAND national Net Worth panel, real per capita, US = 100 each year. Check: 46 countries peak in 1950–71; 42 of them sit below 30% of the US at their latest value (91%); robust to re-binning. Production badge.

← back to The Evidence