Chinese exceptionalism can't last

The International Monetary Fund recently predicted that China would soon overtake the United States as the world's largest economy, ending nearly 150 years of American dominance. If this call is not startling enough, many respected international forecasters envisage a coming age of Chinese dominance.

For example, the OECD believes China will grow at an average rate of 6.6 per cent between 2010 and 2030. The World Bank thinks that from 2011 to 2025, China will maintain an average growth rate of 7.2 per cent. Even the official US National Intelligence Estimates expect China's share of the global economy to grow from 6.4 per cent in 2010 to between 17 and 23 per cent in 2030.

To translate these predictions into dollars terms would mean that the size of the Chinese economy would grow to $36 trillion by the end of 2033 -- more than twice the current size of the US economy.

All these predictions are based on extrapolating China's current performance and projecting it into future decades.

It is not hard to see why economists are making prediction based on China's recent performance. The country has delivered one of the strongest economic growth rates in modern history, growing at close to double digits for the past three decades.

However, two prominent American economists from Harvard University, Lant Pritchett and Lawrence Summers, have challenged this predominant forecasting method in a National Bureau of Economic Research paper called Asiaphoria meets regression to the mean.

They warn that economists are not paying enough attention to a tested and robust rule of cross-national growth rates -- regression to the mean.

The rule is simple and straightforward. A country's growth rates don't persist over medium and long-run horizons and “current growth has very little predictive power for future growth.” The ‘regression to the mean' principle was first developed in 1993 and has since stood the test of time and new data.

Historically speaking, countries eventually converge to average growth rates of 2 per cent, with a standard deviation of 2 per cent. The regression to mean also shows developing countries tend to experience very large accelerations and decelerations of growth, often marked by sharp discontinuities.

Summers, a former US Treasury Secretary, and Pritchett show that episodes of rapid growth, defined as more than 6 per cent, tend to be extremely short-lived. The average duration of a rapid growth period is nine years. The duo identified 28 episodes during which such high growth occurred and only two countries --Taiwan and Korea – have come close to China's current miraculous run.

China holds the distinction of being the only country, quite possibly in the history of mankind, to sustain growth of more than 6 per cent for more than 32 years. Statistically speaking, the chance of China continuing to grow at such a pace is getting slimmer and slimmer or in the parlance of an economist, it's an extraordinary tail event.

Even when they expand their data set to include 70 episodes of growth above 4 per cent, once again, their median duration is only nine years. Only four countries grew by 4 per cent a year for more than nine years, and none of them lasted as long as China's current episode of rapid economic expansion.

More importantly, at the end of these rapid growth periods, growth rates plummeted to 1.85 per cent in the vast majority of countries, which equates to full regression to the mean (2 per cent). Pritchett and Summers can only identify four cases in which a high growth period ended with an acceleration in economic growth, and, once again, China is one of those exceptions.

The Harvard duo's argument is simple. It is wrong, or at least, very imprudent for economic forecasters to rely on the assumption that in the absence of any reason to think otherwise, that current growth rates will persist. They argue this assumption is not tenable in light of the robust and empirically tested ‘regression to the mean' rule.

“Our argument is that the default prediction/projection/forecast should be that a country's growth rate will be subject to regression to the mean. What has to be justified with argumentation is why the growth rate would persist at rates higher or lower than the world mean growth rate,” they say in the paper.

They believe China would not be able to defy the rule of regression to the mean forever. “Our guess is that growth will slow, substantially,” they say, “Why will growth slow? Mainly because that is what rapid growth does.”

They also argue that China's autocratic political system makes it more prone to a boom and bust cycle than industrial democracies.

“The risks of sudden stops are much higher with weak institutions and organisations for policy implementation,” the write. The pair identify only two countries that have managed to grow rich without becoming democratic -- Oman and Singapore.

So what happens if China reverts to the mean, which implies only 2.1 per cent growth after a period of rapid growth? It means a massive deceleration of the world's second largest economy and it would have profound impact on the world's growth rate and more importantly for commodity exporting countries like Australia.

“China's super-rapid growth has already lasted three times longer than a typically episode and is the longest ever recorded. The ends of episodes tend to see full regression to the mean, abruptly,” warns the former US Treasury Secretary and his Harvard colleague.

As an optimist on China's economic future, this paper has severely tested and challenged my understanding. The big question is, how much longer can China continue to defy the law of economics? It's time for our boffins at Treasury and the Reserve Bank to look at their forecasts again.