Forecasts of China's growth always attract interest, even when they are a year old. Larry Summers and Lant Prichett are getting another good run with the paper they published last year (see my earlier post), which analyses emerging-economy growth in general, but of China and India in particular.

There is sophisticated econometrics here, but the key argument is a simple but powerful rule-of-thumb: 'reversion to the mean'. One of the great insights (not just in economics) is that natural phenomena vary around a mean, and when there is an observation well away from it, chances are the following observation will be closer to the average. You might flip a coin and get three 'heads' in a row, but the best forecast for the next toss is still 50:50. Applying this rule-of-thumb to China tells Summers and Pritchett that it's growth rate during the three decades before 2008 is an outlier in the history of global economic growth, and so in the future there is likely to be a lot less, and somewhere around the mean.

Economics has other rules-of-thumb which would support the idea that China's growth will slow. Herb Stein's famous 'law' is that 'unsustainable events don't go on forever'. And of course 'trees don't grow to the sky'.

But a powerful case can be made that reversion to the mean of global economic growth is not the most likely outcome for China, at least any time soon.

Even the most powerful rules-of-thumb must be used in the right context. Let's start with another one, the rule of convergence. In the right circumstances, poor countries will converge towards the levels of per capita GDP achieved by the mature economies, because the technology to do so already exists. Accumulating the necessary capital and technology has been done before by quite a few countries. If they can do it, why not China and India? In this context, the more relevant mean is the average per capita GDP in mature economies. China and India have a long path of potential convergence –adding capital and technology – before they will run into the technological frontier where the mature economies currently are.

What about another favourite rule-of-thumb: the story of the statistician who drowned while crossing a river with an average depth of only a metre? The moral here is that there is a range of experiences – often widely different – hidden within the average. True, Brazil had more than two decades with no growth at all in per capita income. On the other hand Taiwan, Singapore, South Korea and Hong Kong (and earlier, Japan) had quite long periods of fairly sustained economic growth which have taken them to high levels of per-capital income. What relevance does Brazil's failed growth experience have for China? It's a warning that you can mess things up for a sustained period, but China already knows this from its own experience.

Where does this leave us? It's stale news to say that China can't sustain the double-digit economic growth that occurred in the three decades before 2008. Since then, China's underlying growth has been 7-8%, with an exceptional year in 2010 when there was a gigantic temporary policy stimulus. Proper analysis shouldn't rely too much on the average experience of all emerging economies (many of them clearly quite different from China), but look at the variety of experience in the convergence process, and ask if China's actual circumstances will allow it to mimic the success stories rather than the failures. 

The serious debate is whether China can sustain an underlying rate of economic growth around the current pace or whether there are specific factors, such as financial problems, environment, demographics and rebalancing difficulties, which could take this down to 3%, as Michael Pettis argues. The history of emerging economies tells us that it's easy to mess up and fall off the convergence path. But China has done pretty well for the past three decades, and that experience is relevant to the forecast.

To look at these specifics is more useful than thinking in terms of 'reversion to the global mean'. If China achieves even 5% growth until 2050, it will reach OECD average per capita GDP. That's amazing, but not unrealistic.

Photo courtesy of Flickr user Richard Atkinson.