Although the week's dataflow delivered more positive surprises than negative shocks, we'll start with the shocks, because potentially the most worrying almost managed to slip by undetected, a triumph of camouflage.
It's not as if China's monthly monetary data isn't widely watched, but most commentary on November's data dwelt on the relative strength of new bank lending (up 562.2bn yuan on the month) and the modesty of the slowdown in M2 (12.7% in November, vs 12.9% in October). In my opinion, they should have spent more time thinking about the 'shock' fall in growth of M1 7.8%, slowing still further from the 8.4% recorded in October. There is a ready explanation for both the weakness of M1 and the relative resilience of M2, which notices that inflation has retreated, and urges that inflation-related liquidity preference (M1/M2) has similarly retreated. The only problem with this explanation is that it's wrong (as one could also tell from the surprising strength of the previous week's retail sales numbers – up 17.3% YoY). Rather, what's happening is that China's M2 numbers since October have been newly bolstered by the inclusion of deposits from China's Housing Provident Fund – a mandatory forced savings scheme in which the size of new deposits are tied to the size of wages. M2 numbers are staying within range of expectations only because of these forced savings.
In fact, the M1 numbers almost certainly tell the right story – a story of deteriorating private liquidity and cashflows. Even with these forced savings included in the story, for the second month in a row, China's banks gave out substantially more new loans than they took in new deposits. Since the end of September, China's banks have extended 1.148tr yuan in new loans, but have taken in only 100bn yuan of new deposits (including the forced savings). To be clear, this is not just some seasonal effect showing up in the data – it's a genuine deterioration in private sector liquidity. Lucky, then, that the central bank can offset it by cutting reserve ratios.
The week's positive surprises continue to be discovered in the relative health of the world's industrial economy, which so far is almost managing to shrug off the appalling dangers European politicians seem keen to subject their citizens to in order to 'save the Euro'.
It no longer completely surprises that the more timely surveys of industrial conditions in the US show sharp improvements. This week, the Empire State Manufacturing index delivered its best reading since May, and the Philadelphia Fed Survey gave the strongest reading since April. In both cases, the improvements were driven by a sharp uptick in new orders. Nor is it beyond comprehension that Japan's machine tool orders managed a sequential jump which was 1.1 SDs above seasonal historic trends – as we've pointed out elsewhere, the minutiae of recent trade and output data has already showed the resilience of capex spending globally.
But it is a surprise that Europe's manufacturers are surviving better than expected: this week the Eurozone Composite PMI, the Eurozone Manufacturing PMI, the Eurozone Services PMI, and (separately) manufacturing and services PMIs for both Germany and France all arrived stronger than consensus had expected. True, only German and French services PMI readings managed to crawl over the expansion/contraction reading of 50 – but outside those readings, the pace of contraction was less than expected. Similarly, the 3.5% YoY fall in new car registrations in the EU25 actually hid a sequential rise of 2.4% MoM – which was 1.1SDs above seasonal historic trends.
It would, of course, be misleading to ignore the fact that the hardest of data – which also arrives systematically later than most survey data – was generally worse than expected. In the US, industrial production fell 0.4% MoM and retail sales rose only 0.2% MoM in November. In the Eurozone, industrial production grew only 1.3% YoY during October, China's Leading Index declined by 0.1% MoM – the first sequential fall since Dec 2010. This latter indicator is particularly badly-named – how can a short-term 'leading indicator' be doing its job properly if October's 'leading indicator' is delivered two weeks after the hard release of economic data for November?