The power of electricity to distort the growth picture
Most people who follow these things found the raft of economic data released yesterday by the mainland to be reassuring.
The impression was of an economy successfully pulling itself out of the recent downturn and embarking on a steady, sustainable recovery.
Indeed, some people found the numbers rather too reassuring. In a research note, analysts at investment bank BNP Paribas called the data 'almost too perfectly soothing to be credible'.
Given all the recent attention paid to inconsistencies in the mainland's economic statistics, some scepticism was understandable. But economists took comfort from data showing that electricity generation was up 4.7 per cent last month compared with a year earlier.
Analysts doubtful about the accuracy of official economic data have long looked towards electricity generation figures as a more reliable indicator of output.
They reason that electricity use is closely correlated with business activity and argue that because the amount generated is relatively easy to verify, changes in power production are a good proxy for overall economic performance.
Economists made much of this relationship earlier in the year, maintaining that the mainland could not have grown at a 6.1 per cent rate in the first quarter as official data showed, because electricity production was down 4.5 per cent.
Yet electricity generation is at best an imperfect gauge of overall economic activity. On examination, it turns out that both the slump in production at the end of last year and the subsequent rebound are easy enough to account for. Unfortunately, the explanation does not offer quite as much comfort as the optimists would like.
Industrial users are by far the biggest consumers of electricity on the mainland, eating up 70 per cent of all the power generated last year. And by far the biggest industrial users are metallurgical plants, especially aluminium smelters and steel mills.
As a result, as the first chart below shows, when aluminium smelters cut back their production last year, electricity generation dropped steeply. And as aluminium production picked up again over recent months, topping one million tonnes in June, so did the amount of electricity generated.
(The chart for steel production follows a similar trajectory, with monthly output bottoming out in November last year, then rebounding 44 per cent up to last month.)
The problem is that the rebound in metal production, which has driven the increase in power generation, does not necessarily indicate a recovery in overall economic activity.
To see why, take a look at the mainland's aluminium imports, which rocketed from just 14,000 tonnes a month in September last year to 394,000 tonnes in April as domestic production dropped, before falling back as domestic production picked up again.
The reason is simple enough. As the second chart below shows, towards the end of last year, international aluminium prices fell steeply relative to mainland domestic prices, encouraging buyers to purchase imported aluminium. More recently, however, international prices have staged a recovery, closing the gap and making domestically produced aluminium more attractive again.
In other words, the decline and subsequent rebound in domestic aluminium output - and therefore much of the drop and recovery in electricity generation - may be due less to changes in domestic demand than to international pricing anomalies in the metals markets.
That means mainland electricity production figures may not be such a good proxy for economic growth as many analysts like to believe.
This cuts both ways, of course.
It means that although the slowdown at the end of last year may not have been as brutal as electricity data indicated, it also implies that the current recovery may not be as robust as analysts hope.