China Auto Sales: What’s the Deal?

The China Automotive Technology & Research Center (CATRC) reported that passenger car sales jumped a whopping 59.3% y/y in August to 977,300 units, up from a 15% y/y and 822,300 units in July. The CATRC data supposedly tracks retail deliveries of autos, so in theory it should be a better metric of demand than the more prevalent China Association of Automobile Manufacturers (CAAM) data that tracks wholesale delivers.

Still, these two series should run more or less in line with each other. Unless there is a gigantic amount of inventory rotation, wholesale figures should net out to retail sales within a few months. Now, I don’t have the CATRC data series, and cannot find any English or Chinese-language news articles from last year reporting the figure, but I do have the CAAM data and I can use the implied CATRC data from the y/y growth rates reported. CAAM reported that passenger car sales were 858,278 in August 2009, which itself was a massive 70% y/y jump. If the CATRC sales increase is correct, then these 858,278 wholesale deliveries only resulted in 613,497 retail sales in August last year. The implied CATRC figure for July was 712,565 and June would have seen 756,743 retail deliveries. At first it seems absurd that retail sales would have fallen by about 100,000 units m/m in August 2009 when wholesale deliveries remained robust, but I think it’s plausible (if just barely).

First, there has been a wide divergence between the CAAM and CATRC data in the past few months. For example, in June 2010, the CAAM wholesale figure was 1,042,818, but CATRC reported only 839,228 in retail sales. Second, the demand in H1 2009 for autos was well beyond what anybody could have expected, and the demand shifted more than expected to the smaller engine cars with the largest government subsidy. This could have resulted in a shortage of smaller autos on dealers’ lots in August 2009, delaying deliveries for those cars. (Still, in June-July 2009, wholesale deliveries outpaced retail sales by a wide margin, so even this theory requires some big assumptions to remain plausible.)

The recent data do seem to indicate that dealership inventories likely explain part of the discrepancy. CATRC reported that inventories fell to 57 days from 58 in August 2010, so wholesale deliveries must have lagged retail sales for the month. In the previous two months, there were 519,710 more wholesale deliveries than retail sales, and inventories were reported to be building. Dealers started cutting prices in July, and deliveries started to dip as well as demand looked softer.

If this is the case, the 60% pace is not sustainable (if it’s even accurate), and it signals future weakness in demand and producer margins, not resilience. The markets didn’t take it that way though: SAIC’s stock was up 10%, and rubber prices jumped too. I wouldn’t be shocked if when the CAAM figures come out next week auto stocks get beat up on the disappointing news of weaker wholesale deliveries.

As ever, China’s data keeps things interesting for analysts.

All rights reserved, Roubini GlobalEconomics, LLC