Doubts about China's economic data are nothing new, of course. One good reason for questioning their accuracy is the sheer speed with which they are published. In an economy as large and diverse as China, the National Bureau of Statistics manages to collect, collate and release GDP data within two weeks of the end of the quarter.

In Hong Kong, where the economy is smaller and data collection easier, it takes a full six weeks.

What's more, local officials in China have powerful political incentives to exaggerate the output figures for their districts. This produces big distortions. In 2009, for example, all but one of China's 31 provinces announced a growth rate higher than the national average. In 2010, China's GDP was 40 trillion yuan (HK$49.1 trillion). But if you add up the numbers declared by each province, you get a total of 44 trillion; a discrepancy of 10 per cent.

As a result, investors have long looked to hard, real world indicators like electricity production as a more accurate guide to levels of activity. When he was party boss of Liaoning, Li Keqiang, strongly tipped as China's next premier, famously said he regarded official output figures as fabrications, and that he relied instead on hard data for power consumption, rail shipments and bank loans as a better guide to the health of the provincial economy.

So investors were badly shaken last month when the New York Times accused government officials of ordering power company managers to exaggerate the amount of electricity they are generating, in an attempt to hide the true depth of China's economic downturn.

Professional economists were quick to question the accuracy of the story itself, arguing that it would be difficult to falsify power production figures, which are automatically transmitted to the central government.

But their analysis missed the point. Although electricity production might have been a useful indicator back in the 1990s, these days it is no longer an accurate guide to China's GDP growth.

The reason is simple. Over the last decade, local governments all over the country have built brand-new aluminium smelters. As a result, China's aluminium production has rocketed from less than 3 million tonnes in 2000, to more than 18 million last year.

And because of aluminium's chemical structure, refining it into pure metal takes a lot of energy. Turning two tonnes of the intermediate material alumina into one tonne of aluminium typically uses 14,000 kilowatt-hours of power.

So making 18 million tonnes of the stuff takes a truly enormous amount of electricity. Last year, 5 per cent of all the electricity generated in China was used to power the country's aluminium smelters.

But with the aluminium industry suffering from massive overcapacity, the price of the metal has been falling. As a result, in 2011 China's production of refined aluminium constituted only around 0.6 per cent of its GDP.

Because power is expensive and aluminium cheap, smelters operate on razor-thin margins. That means whenever there is a drop in demand for aluminium, as there has been over recent months with the decline in new property investment, producers quickly idle their smelters. And when they do, electricity consumption - and production - falls rapidly.

But that fall is not necessarily reflected by a corresponding fall in GDP. Because China uses 5 per cent of its electricity to produce just 0.6 per cent of its GDP, turning off aluminium smelters results in a big fall in electricity production, but only a tiny fall in GDP.

So faking electricity production figures in an attempt to persuade people that GDP growth is healthy would be singularly pointless. The relationship is sketchy at best.