Feb's Headline-Grabbing Manufacturing Purchasing Manager Index

·         UK Feb Markit Manufacturing PMI Disappoints Down 2.1pts to 50.8

o   Suggests Most Marginal Monthly Expansion Since April 2013

o   New Orders Slow for Consumer and Capital Goods, Rise for Intermediates; Export Orders Fall for Second Month

o   Output Growth Slows to 7m Low, Slowed by Consumer and Capital Goods Sectors

o   Employment Declines Slightly for Second Month

o   Margins: Input Price Deflation Eases to 7m Low, Selling Prices Fall for Sixth Month

Comment: The crucial thing to know about this purchasing managers' index is that historically there is no statistical link between this and the monthly official surveys of industrial output - realistically, therefore, it contains no useful information. However, Markit's surveys provide regular fodder for newspaper headlines. Technically, an index reading above 50 is said to indicate an improvement in business conditions, below 50 a deterioration. So Feb's 50.8 result suggests a marginal improvement.  But it's not clear how such precise results are produced. 

January Consumer Credit

• UK Jan Net New Consumer Credit Surprises +£1.56bn
• Almost Double Jan 2015 Total, and 1.5SDs Above Monthly Average for 2015
• Credit Card Lending £509mn,  Others £1,044mn
• Banks Lent £796mn, Others Lent £661m
• Gross Outstanding Consumer Debt is £179.2bn, up 4.6% yoy
• Gross Consumer Debt is Equivalent to c9.6% of GDP, up from 9.4% in 2014

Comment: Bank of England's data shows steady rise in net additions to consumer credit - but still sharply below what was seen between 2002 to 2005. Meanwhile, data released the same day shows no growth at all in Jan's M4 - no surprise, since it has been essentially unchanged since the beginning of 2010 !

No Mystery in Britain's Job Growth

The rise in British employment is one of the statistical mysteries of our time: in the 3m to August, the number of jobs rose by 212k qoq, with 162k new employees and a further 35k newly self-employed.  Moreover, the number of those economically inactive also fell by 138k during the same period. As the chart shows, this means that the number of British jobs is now 84k now than at the pre-crisis peak of early 2008.

Quite often, surprise improvements in unemployment rates can be traced to strange movements in the denominator. Not in this case: the improvement looks quite genuine. But the improvement is in stark contrast to two things: first, GDP data which tells us the British economy is currently 3.8% smaller than it was at its pre-crisis peak. Second, it is also a great contrast to what is happening in the Eurozone, where in 2Q employment levels had fallen by 4m, or 2.7% from early 1Q2008 levels, with no sign of recovery.

In fact, the divergence between GDP and employment may be less surprising than it immediately seems. Certainly, if all the data is correct, it implies an erosion in labour productivity. However, once real output per worker is discounted by changes in capital per worker, it becomes clear that declining real labour productivity  has been the norm since at least 1998, and that, after steep cyclical productivity declines in 2008/09, current productivity levels have simply recovered to the (falling) 1998-2008 trendline.  The sharp rise in employment since 2010 ran alongside the recovery back to that (falling) trendline. From here on in, we should expect 'normal' employment patterns to be resumed. 

Britain's Volatile Export Data - ONS Replies

The ONS has replied to my query about the extraordinary volatility of Britain's export data this year. The brief reply is worth retailing in full, in part because it discloses their own (muted) concern about the series.  Actually,  times of significant structural economic change inevitably do tend to throw seasonally-adjusted data out of whack for a while -  Japan's statisticians struggled with these problems for years after the Bubble imploded. . . .  

The nature of a monthly dataset does tend to highlight the volatility of the statistics - but it is true that the Goods series has become particularly volatile as of late.

Although we don't quantify the impact of special events, the anecdotal evidence indicates that a lot, but not all, of the impact comes from specific economic events over the past year (additional factory closures etc).

The issue is that because we have a series of genuine erratic movements, there is an element of negative/positive autocorrelation in the data (as expected). Although we don't doubt the quality of our dataset we're currently looking at our seasonal adjustment and time series compilation techniques to ensure that our treatment continues to be correct.

Strange Data from ONS

What's going on in Britain's trade balance? Given that the 2Q current account deficit was a record  £20.8bn blowout largely owing to the £28.1bn deficit in traded goods, one would imagine it is easy to tell. But it isn't:

• in June  the £4.33bn trade deficit reported was far larger than the range of analysts' expectations;  
• in July, a dramatic recovery in the deficit, to just £1.71bn similarly defied the range of analysts' expectations, and
• in August, the deficit ballooned again, to £4.17bn - once again, completely unexpected.

Now this is seasonally adjusted data, which usually suppresses monthly volatility (there are exceptions -  Japan's 1Q GDP, for example). But right now, the volatility of Britain's export performance is extremely unusual.  Consider this chart of Britain's export of goods, data from the Office for National Statistics: 

This year, the data is showing a monthly volatility quite unlike anything we've previously seen. This can be measured: during 2000 to the end of 2011, the monthly change in Britain's exports had a standard deviation of 4% a month. This year, that volatility has jumped to 7.2%. It is as if, come 2012, we are dealing with a completely different population of data. 

To make matters more puzzling, this spike in volatility shows up nowhere else in Britain's trade data: exports of services, imports of goods, imports of services -  all these have retained their historic  volatility pattern. 

 I have emailed the ONS seeking an explanation. If they reply, I will post it here. 

In the meantime, if one wants to track Britain's trading trends, it makes more sense to smooth - in this case to a 6m average. It's bad. . . . but probably not as bad as the monthly data has announced.