They have trawled through the data and drawn two powerful conclusions. One is that there is almost no statistical basis for predicting growth from one decade to another. Extrapolation is a mug’s game – or, as they put it, “current growth has very little predictive power”. From 1967 to 1980, Brazil grew at an average annual rate of 5.2 per cent. Few would have predicted, then, that for the next 22 years per capita income would grow at precisely zero.

Their second finding is that episodes of super-rapid growth last a median of nine years. China is the big exception. The only countries with fast-growth episodes that come close are Taiwan and South Korea, which managed 32 and 29 years respectively. According to the authors, once such episodes end, the median drop in growth is 4.65 points. That would cut China’s growth to 4 per cent and India’s to 1.6 per cent – far lower than almost anyone is predicting.

The thesis has huge potential ramifications, both economic and geopolitical. If China and India continue their current growth trajectories, their combined gross domestic product will rise to $66tn by 2033, against $11tn today. If they regressed fully to mean, they would reach a combined GDP of just $24tn. The authors have no need to explain why China’s, or indeed India’s, growth should fall so precipitously. This is just what happens. Their reasoning puts the onus on optimists to explain otherwise.

There are several conceivable rejoinders. The first applies to emerging markets generally. The idea of “convergence” holds that poor countries can grow faster than rich ones. That is partly because there is low-hanging fruit; for example, moving people from unproductive rural jobs to more productive urban ones. Poor countries can also copy rich ones; they do not have to reinvent the wheel. Niall Ferguson, professor of history at Harvard University, refers to “six killer apps of prosperity”: competition, scientific revolution, property rights, medicine, consumer society and the work ethic. Since we already know what they are, they can be “downloaded”. Several Asian economies, including Japan, Taiwan, South Korea and Singapore have more or less caught up with western living standards. If they can do it, why not others? Regression to mean, however, implies that such speedy catch-up is impossible, or at least very difficult.

Is there anything about China specifically that suggests it could buck the trend? First, it has already done so, growing rapidly for more than 30 years. The two economists say this makes a rapid slowdown more likely. But perhaps it is the reverse. Chinese leaders may have learnt how to beat the odds. Second, as Jim O’Neill, who coined the term Brics, says, the authors’ data may be skewed by the mostly disappointing economies of Latin America, the Middle East and Africa. Perhaps Asian economies have discovered a secret sauce. Third, China’s long-lived expansion follows many decades of chaos and suboptimal growth. What we are seeing now could plausibly be a long-pent-up recovery – the country’s own regression to mean. Fourth, China’s size could confer growth-sustaining advantages in terms of economies of scale and size of domestic market. If true, that would also apply to India.

The authors do offer one reason for believing China’s growth rate is unsustainable: “institutional inadequacies”, particularly the failure to control corruption. To some, the argument is an old saw dressed up in new data: the country cannot keep growing because it is not democratic. Still, the regression to mean theory provides a powerful corrective to mechanical extrapolation. The authors are right. Any time you hear the words, “based on current growth trends”, you should pause for thought.