We should receive September data over the next couple of days, but it is possible to get some information intra-month via partial indicators from coal-burn data and hydro generation flows.
On the coal side, data on coal consumption from "key power plants" is available every 10 days. This sample is fixed, so it becomes a little less representative as more generation capacity is added. Its also measures tonnes of coal consumed rather than power generated, so improvements in efficiency or changes in the coal mix may also affect estimates of generation.
But as the chart below shows, changes in this sample broadly mirrors changes in thermal generation from the monthly NBS generation series. It makes sense that it trends well below growth rates in the total for the reasons above, but the directional changes largely mirror the headline thermal generation
Looking at this data more closely in the last few weeks, its clear that after the pick up seen in August, coal burn has faded badly in September and early October.
Early October is the weakest point of coal burn given the holiday, and it may be the case that activity is picking up post the break. But there are few signs in other indicators that this is happening.
What about other sources of generation? We think nuclear and wind generation should be growing at a similar strong rate as seen in the last few months given capacity additions. But there is greater uncertainty about hydro generation, which has started to decline year-on-year in the last few months.
One way to get a partial read on this is to look at inflows into generation facilities operated by the China Three Gorges Corp, which are published everyday. While this is only 15pc of total hydro generation, it is representative of hydrological conditions in the main hydro producing region.
This continued to fall away sharply in September, although has not been as weak in October. Nevertheless, it seems the weak rainfall during June and July means that hydro generation will be weak for the rest of the year.
So with coal-fired generation and hydro generation likely falling, it suggest that in the last 6 weeks, total power generation has also been down year-on-year.
This would suggest that so far that growth rates potentially getting worse rather than stabilizing. It also suggests that as of early-October, the efforts of policy makers to bolster growth that have seen variables like money supply and home sales improve for months now is yet to have made a meaningful impact on real activity.
Its hard to be too definitive on October though given the impact of holidays, which this year have had a bigger impact in the days before and after the break than in prior years.
But those holding out for an improvement in Chinese economic growth in Q4 will be disappointed if if the signs from power generation persist in the next few weeks. Using the statistics above, we should be able to get a handle on this before the official stats are released.
