from Follow the Money

China September Exports: Not Quite as Bad as They Seem?

October 13, 2016

Blog Post
Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.

More on:


The 5.6 percent fall—in the yuan data—in China’s September exports was a surprise. Exports had been rising in yuan terms, and in volume terms, since March. I expected the rise to continue, largely because the pickup in volumes is consistent with the expected impact of the 8 percent fall in the broad yuan (using the BIS index) since last July.

And I am very conscious of the risk of interpreting data to fit your prior beliefs, and thus missing a new signal.

That said, I do think there are a couple of reasons why the fall in exports may not be indicative of a shift in trend.

The first is straightforward: there was one fewer working days in China this September than last September (22 versus 21; data are here). Nominal exports, in yuan, per working day, fell by 1 percent.

This argument should not be overstated. There were more working days this August than last August, so nominal exports, in yuan, per working day, were down in August.

The more important reason is a bit more complicated. Chinese export prices jumped last September, in the immediate aftermath of the yuan’s August depreciation. Each dollar in exports generated more yuan. Over time, though, export prices have come down. They are now lower than their pre-August devaluation levels.


China hasn’t released its export price index for September. But if you compare the export price index for August (e.g. assume there was no change in September) to the export price index last September, export prices should be down about 5 percent y/y.

After adjusting for the likely change in exports prices—but not adjusting for the change in working days—I get a 1 percent estimated fall in export volumes. Not great, but not as bad as the headline numbers.

Adjusting for the number of working days, I get positive growth in export volumes per working day. But, to be consistent, adjusting for working days also requires a downward adjustment for August. the 7 percent y/y growth China reported for August export volumes looks a bit overstated.


Where does that leave me—my best guess is that export volumes are still creeping up, not down.

And q3 import volumes are all over the map. China reported a 3 percent fall in import volumes in July, a 9-10 percent rise in August, and I estimate a 0.5 to 1 percent rise (noise) in September.

For q3 as a whole I estimate a 3 percent y/y rise in (goods) export volumes, a 2.5 percent y/y rise in (goods) import volumes—both down a bit from the 3-4 percent increases in q2. The q3 numbers are not strong numbers by any measure, but they also aren’t numbers that show a major change in trend.

But as anyone who has worked with the Chinese data knows, there is a lot of noise in the monthly data. Remember China doesn’t report underlying levels for export volumes. Only nominal exports and a nominal price index, and y/y changes in volumes. The actual level of real exports and imports has to be inferred.

This chart, which simply deflates nominal exports by the price index, gives some sense of the amount of “noise” in the underlying data. I use Haver’s seasonal adjustment, but the seasonal adjustment isn’t enough to get a smooth series.


Bottom line. The data looks a bit on the weak side, but not as weak as the headline numbers suggest.

October of course gains added importance after a weak September (which after adjusting for working days turns into a weak August). The comparison with last October is clean: the number of working days is the same (18). And it is likely that there will also be a roughly 5 percent fall in export prices, in yuan terms, as export prices remained on the high side last October.

More on: