China Debt Dynamics: Part 1 - Government Debt

This is Part 1 of a two-part series. This piece was originally carried in the Shocks & Surprises Global Weekly Summary for week ending 3rd January 2014. 

Holiday seasons are generally a convenient time to slip out bad news, and so it happened that China published its latest and best effort to tally government debts on New Year's Eve 2013. A serious effort this was as well, with 54.4k auditors deployed to ferret out the liabilities of five levels of government – central, provincial, prefectural, county and township administrations are all counted. The total as of June 2013 came to 30.27tr yuan, equivalent to a not-so worrying 56% of GDP, but also to a distinctly worrying 248% of total govt revenues (tax and non-tax) for the 12m to June 2013.

Even the maximalist 30.27tr yuan tally comes to only 56% of GDP, which by itself is hardly an alarming proportion. Despite the headlines, this number is not directly comparable to the previously announced 10.7tr yuan estimate of local govt debt as of end-2010 – that tally accounted for only for provincial, prefectural and county govt debt. Even though the debt total is sharply higher than previously thought, this is still not debt-meltdown data. But it ought to alert us to the way this larger-than-expected debt total is likely to bear on policy-choices and policy-making.

The key weakness, and therefore potentially a key policy-constraint, is that the debt is large relative to China's ability to raise taxes: 30.27tr is equivalent to 248% of total central and local revenues, including both tax and non-tax revenues.  That means that every one percentage point rise in financing costs has the potential of attracting (over time) a rise in interest payments equivalent to 2.5% of total revenues.   Now, in the 12m to September 2013, China's fiscal deficit of 811mn yuan was equivalent to 6.5 percent of total revenues. In other words, whether China likes it or not, current fiscal policies are now significantly leveraged to financing costs.

The obvious conclusion from this is that PBOC probably faces more constraints on raising interest rates than previously thought.  And from that, we should also expect that 2013's experience of allowing a rising Rmb to exert increasing economic discipline is likely to be extended in 2014 and beyond. The strong Rmb may be painful for exporters, but their interests are unlikely to outweigh those of the government.

The second obvious conclusion is that China is going to need to raise its government revenues relative to GDP. In the 12m to September, revenues were equivalent to 22.6% of GDP. This proportion has been largely unchanged since 2011, and has interrupted the steady progress made in raising this ratio since 1998 at least (when it was just 11.7% of GDP).  Evidently, for the last few years, the tactical need to support the economy over-rode the strategic need to raise the tax-take.  At some point, that strategic issue will have to be addressed.

The renegotiation of tax raising powers and responsibilities between the Centre and the Provincial governments outlined in the 3rd Plenum were always going to be one of the tougher reforms to deliver. These numbers make those negotiations tougher – a lot tougher – as well as more urgent. Quite possibly this is the single biggest and toughest reform upon which the rest of the package will stand or fall. [So far this year (as of early March) it has not featured significantly as a topic of discussion.]

And there is a third thought: if the debt-total highlights the government's unexpected sensitivity to rising financing costs, then it also raises the classic Chinese question:  who pays?  In this case, it's not difficult to imagine that sections of the financial industry – banks in particular – will find themselves picking up some of the costs involved in keeping the apparent financing costs down, even if market rates rise.  The temptation to extend financial repression – ie, by keeping deposit rates negligible  - even in the face of liberalizing lending interest rates is likely to be strong.  Similarly, we can imagine that the Ministry of Finance is rather more attached to those 20% deposit reserve ratios than previously thought – commandeering deposits is, after all, the cheapest source of financing of all.

And there is this final thought: although it may seem that New Year’s Eve is a suspiciously quiet time to slip out such important data,  the fact that is was published at all is important. It is a message from China’s leadership of the seriousness of their intentions, and the limits of their room for manoeuvre.

Northeast Asia Exports: China’s New Year Caveats, Japan’s J-Curve in the Wings

Let’s start with a serious caveat: the way Chinese New Year wanders around the calendar makes it dangerous to read too much into China’s economic data during the first two to three months of every year. Trade data, retail data, inflation, money - all are affected - and even after decades of trying, I’m still uncertain how to which numbers will be affected, and in what way.

This year, it ought to be relatively simple: Chinese New Year fell on January 31st,  whilst in 2013 it fell on February 10th.  So generally, we ought to expect January to have had fewer working days this year relative to 2013 and consequently figures for industrial output and exports to be relatively depressed in January before rebounding in February.

But quite possibly it won’t work out like that. The  nearest similar timing would be in 2009 (when Chinese New Year fell on Jan 26th, vs Feb 07 in the previous year): in January 2009, China’s exports were down 0.4SDs from trend in January before collapsing 1.9SDs in February.   But one can counter that with what happened in 2003 (Feb 1st 2003 vs FEb 12th 2002), when January’s exports were 2.3SDs stronger than expected, before giving back 0.8SDs in February.

So perhaps China’s Ministry of Commerce spokesman was telling nothing but the truth, when he was quoted yesterday as warning the 1Q exports will be 'volatile'  and 'we can't rule out the possibility that February’s trade figures will show an abnormal change from last year.'  And ‘we should be clear that [January’s] export surge doesn't necessarily indicate favourable times for Chinese exporters’, even if the outlook for the year as a whole is ‘cautiously optimistic’.

And whilst we’re reading the runes, the willingness to let the Rmb depreciate over the last week is also compatible with nervousness about February’s export numbers.

Nevetheless, January’s export data was strong not just for China, but for Japan and Northeast Asia as a whole in a way which defies the popular belief that Asian economies are slowing, and which essentially has not been recognized. In dollar terms, NE Asia’s exports rose 4.7% yoy in January against 2013’s toughest base of comparison (exports jumped 15.5% yoy in Jan 2013).

In momentum terms, January’s movement was 1.6SDs above historic seasonal trends.  As the chart below shows, this was merely an acceleration of a trend of rising momentum that had been quietly building throughout 4Q13. In fact, on a 6m basis, the underlying momentum in January reached its highest point for three years.  Whilst consensus seems resigned to another year of single-digit  export growth from NE Asia in 2014 (following 2.5% in 2013, 3.1% in 2012), even if the unnoticed outperformance of the last six months dies right now, Northeast Asian exports are likely to grow in the high 'teens in 2014.  I suspect this is not yet on anyone's radar.

Although China inevitably bulks large over this data, since China accounts for nearly 59% of NE Asia’s exports, it is not the only contributor.  In Japan, too (still about 19% of NE Asia exports), January was strong, with exports up 9.5% yoy in yen terms, which beat historic seasonal trends by 0.4SDs. The statistical impact of yen devaluation is beginning to slide out of the data: the yen was down 14.7% yoy against the dollar in Jan vs 19.2% yoy in December. Consequently, the fact that in dollar terms exports fell 6.6% yoy disguised an outperformance vs seasonal trends of 0.6SDs, allowing the 6m momentum trendline to break into substantially positive territory for the first time since May 2012.


The chart below measures 6m changes in export momentum for China, Japan, S Korea and Taiwan, both in local currency terms, and also in volume terms. By January, each economy was showing positive and accelerating momentum, led by China and Japan. 

Finally, there is one other piece of surprising news - Japan’s trade numbers may finally be beginning to benefit from the J-curve impact of the depreciation in the yen. The J-curve’s arrival has been delayed so long we’ve stopped expecting it. And on the face of it, the record Y2.79tr trade deficit in January is just another demonstration of its absence. However, there are two reasons to see beyond that. First, January’s trade balance was principally the victim of a 37.7% yoy collapse in ship exports - the classic ‘lumpy item’ distortion.

Second, the fall of Japan’s share of NE Asia’s exports is now finally slowing. As the chart below shows, the peak of Japan losing market share came in August 2013, when over the previous 12 months it had slipped by 3.2pps to 19.5%. By January, that market share loss had slowed to 2.5pp, with a market share of 18.7%. It's not dramatic – indeed, it still hasn't broken the surface. But it is there: the J-curve is in the wings.

Foreign Investors and the US Taper

The news that in December foreign investors sold a net US$45.88bn of US securities capped a year in which, unprecedentedly, the world repatriated capital from the US.  Although it may not seem immediately obvious, there are circumstances in which this might matter a lot to the US in 2014 and beyond.

The capital outflow came to US$133bn during 2013, and it contrasts with the US$613bn net capital inflow into the US in 2012 – so it's a turnaround in financial behaviour of approximately US$746bn in a single year. What's more, the fact that the US saw a net capital repatriation by its foreign investors is something that has happened only once before in post-Cold War financial history, as the chart shows. Even at the depth of the financial crisis of 2008/09 capital still flowed into the US. Not now.

This seems like a genuinely historic change in behaviour. Changes in financial behaviour tend to be extremely rare – a once in a generation reconsideration of tactics and strategies. If it proves so, it will influence the underlying flows of capital and trade for years and possibly decades to come.

All of which makes it remarkable that it has had no noticeable impact on the US: money supply, bond yields, dollar value, net foreign liability position of the US banking system – none seem to have been challenged by this abrupt reversal of international capital flows. If it is a historic change, it is one which has arrived seemingly without consequence.

The reason, of course, is the link between the Fed's buying (and guidance) and net foreign investment. On the one hand, the Fed has simply been prepared to buy all the bonds foreign investors wanted to sell.  Whilst foreign investors sold US$133bn of securities in 2013, the Fed bought no fewer than US$1.08tr. The chart below shows the way in which Fed buying has tended to offset net foreign buying over the last few years. 

But of course the relationship between the Fed and foreign investors has another strand: like other investors, foreign investors spent much of 2013 anticipating the onset of the Fed’s tapering. Net selling started in February, but then peaked in the 3m to June with US$115bn of net selling, as markets absorbed April’s FOMC news that tapering was probably on its way.  So the Fed’s forward guidance might be said to have both encouraged foreign capital out of US treasury markets at the same time as buying the securities they wanted to sell. 

Even so, such net foreign selling should be expected to erode the positive monetary stimulus of the Fed’s treasuries buying.  And indeed it has: the Fed initiated QE3 in September 2012, and the Fed’s balance sheet shows its holdings of securities rising $51.1bn only in 2012, but $1.0844tr in 2013.  But once you adjust this for the reversal of foreign capital flows, the net buying changes only from US$664bn in 2012 to US$951.2bn in 2013.  As the chart shows, the net effect has been . . . . moderate. 

What matters for now is whether the foreign net outflow of 2013 was purely a trading response to the realization that the taper was on its way, or whether it represents a fundamental shift in global financial behaviour.

If the former, then we can conclude that the exit of foreign capital essentially initiated a de facto tapering some nine months earlier than officially executed. More, it raises the possibility that the visible negative impact of the Fed’s actual tapering this year may itself now be positively offset by a return of foreign capital inflows, or at least a cessation of the outflow.   

Since December, the Fed has announced a cut in its monthly bond buying from the original US$85bn per month to US$65bn. This tightening of US$20bn a month would be halved simply by a cessation of net capital outflows (which averaged US$11bn a month in 2013).

If, by contrast, the willingness to repatriate capital from US securities markets represents a once-in-a-generation re-ordering of financial risks and rewards, then the Fed’s taper could be a gamble which turns out to be unexpectedly painful.

In short, the monthly net long-term capital flows announcements from the US now demand watching closely.

Finally, we need to look at the other side of the transaction:  US$133bn is a sizeable amount of capital repatriation. It would be enough to form a powerful offset to the view that news of Fed tapering has stripped easy capital from emerging markets and helped precipitate a range of currency crises. But alas, the main selling has come not from Asia, but from Europe.  In December, for example, when total net selling came to US$45.9bn, European investors sold US$48.8bn, whilst Asian investors bought a net US$7.1bn, with China’s net US$5.36bn selling offset by net buying from Japan of US$6.73bn.

Turkey - The J-Curve Can't Come Soon Enough

Turkey’s current account deficit to US$8.32bn, the worst since March 2011, tells us that Turkey’s cashflows continued to deteriorate sharply in 4Q, with no sign yet of the sort of turnaround which would justify any relaxation of the pressures on the Lira.  December’s current account deficit featured not only a $2.67bn widening of the trade deficit, but also a $767mn rise in gross international income payments and a $1bn narrowing of the services surplus.

Eventually, of course, we should see the devaluation of the Lira produce a J-curve upswing in trade flows, but for December, exports rose only 4.5% yoy, on a monthly movement which was 1SD below seasonal historic trends, whilst imports rose 16.7% yoy, which was 0.1SDs above historic trends.

In fact, private sector liquidity trends are probably slightly worse than the current account deficit trends suggest, since the total savings/investment balance is actually being improved by 4Q fiscal restraint in Turkey.  In every month in 4Q, the government’s total domestic debt actually fell slightly, in all by L4.939bn, which in theory should have moderated the decline in the current account deficit.  So the Turkey’s private sector savings position was in deficit to the tune of L36.9bn in 4Q, equivalent to an estimated 9.4% of 4Q GDP, and taking the 2013 private sector savings deficit to 6.9% of GDP.  This deficit expanded throughout 2013, and compares to 4.9% in 2012.

The private sector savings deficit is, of course, a key measure of the cashflow position between Turkey’s private sector and its financial system: in this case, we can see that somehow the financial system must either generate enough cashflow to keep the private sector’s activities unchanged – by selling government bonds or by taking on foreign liabilities – or alternatively the private sector’s activities need to change in order to curb its net demand for cash. 

The balance sheet of Turkey's banks shows clearly enough how the private savings deficit has been financed. In the year to Dec, banks' loan books grew 32.9% yoy, or by and amount equivalent to approximately 14.4% of GDP,  but also equivalent to about 170% of estimated nominal GDP growth in 2013. The stock of bank debt at end-2013 was approximately 58.2% of GDP, up from 47.9% at end-2012. December's figures, incidentally, show no sign of any slowdown in lending momentum.   Meanwhile, deposits grew only 21.7% yoy in 2013, so the loan/deposit ratio was pushed up to 103.1% by end-2013.

Turkey's banks financed this in the way one would expect: net foreign liabilities of Turkey's banks grew by US$22bn in 2013, almost doubling to US$5.58bn.   At the same time, the amount of securities held for sale fell 1.4% yoy, and their total securities book grew only 3.4% yoy. 

With Turkey’s central bank raising its overnight lending rate to 12% from 7.75% at the end of January, it is now clear that the private sector’s economic and financial behaviour is expected to make the adjustment. The fact that it has a net cashflow deficit of approximately 6.9% of GDP tells us this will be painful. So that J-curve improvement in trade and current account balances cannot come soon enough. 

But there is a kicker which the rest of Europe is likely to feel. With Turkey's banks willing to fund private sector savings & cashflow deficits, it is no surprise that investment spending continued to surge throughout 2013, with national accounts showing capital formation spending rising 9.6% yoy during the first nine months of  2013. By my estimate (formed by depreciating all gross fixed capital formation over 10yrs), Turkey's capital stock is currently growing around 11.7%. Topline growth has not kept up with this pace, so asset turns, and probably return on capital,  have declined uninterruptedly since mid-2011. 

But what happens now is precisely that this capital spending will stop, and Turkey's industries will raise asset turns and cashflow precisely by satisfying export demand, at no matter what price. The rest of Europe must expect that as Turkey seeks to deploy that capital stock more intensely, its industries will be eating someone else's lunch – almost certainly those of its near neighbours in the Eurozone. 

'Reaping the Benefits'

"In Ireland and Portugal export performance has also been strong, because programme countries are now reaping the benefits from their significant improvement in cost competitiveness." 

Thus Jörg Asmussen,  ECB board member, speaking at Bank of America/Merill Lynch Investor conference in Washington, 20 April 2013.

And here's what he was talking about : 

Ireland

In the 12 months to Feb 2013, Ireland's exports were 4% higher than during the same period of 2009/08 - ie, the nadir.  Currently, momentum is being lost quicker than at any time during that period. 

Portugal

Yes, during 2010 and 2011, Portugal did indeed manage to sustain positive export growth. But there's been no yoy growth during the last six months, and for the last eight months, the 6m momentum trend has been negative. 

I wonder what his speaker's fee was? 

Who's Got Economic Momentum?

• For industrial momentum, the ordering is US,  China, Eurozone, Japan
• For domestic demand, the ordering is US, China & Japan (tied), Eurozone
• Outliers & Likely Corrections: Industrial and domestic  demand momentum tend eventually to balance. On that basis, expect: i) vulnerability in Eurozone industrial momentum, and ii) recovery of Japanese industrial  momentum.

So which of the world’s major economies has the most positive momentum, and in what way?  I compile momentum indicators for the US, the Eurozone, China and Japan on a monthly basis, making separate indicators for the industrial economy and domestic demand. Where possible, the industrial economy indicator tracks production, exports (both local currency value and volume),  inventory/shipment ratios and capacity utilization.  Where possible, the domestic demand indicator takes in retail sales, auto sale, employment and wages. In both cases, the composition will alter slightly according to the availability of monthly data. For each item of data, I measure the deflection of the month’s data from seasonal trends, and express the result as a number of standard deviations from the average error. Expressing the result as a number of standard deviations allows me to take a simple average of the data I’m measuring.  Finally, I take the 6ma as defining the underlying trend momentum.

Industrial Momentum

Taking the industrial economy first, we can compare the 6m momentum trends. However, at this point, it’s worth emphasising that what’s being measured is changes in momentum relative to each country’s individual experience of the past decade, not absolute performance. Thus US industrial output may be growing by 3.5% yoy whilst China’s is growing by 8.9%, but the underlying momentum change may be (is) more positive in the US.

On that basis, the indicators are pretty unambiguous: the US has the most positive underlying industrial momentum, followed by China, then the Eurozone and lastly Japan.  In absolute terms, both the US and China are gaining momentum, whilst Eurozone and Japan are losing momentum. The leadership of the US is likely to be extended in March’s data, with today’s data showing industrial production up 0.4% mom sa (0.5SDs above trend) and capacity utilization also rising further to 78.5% (1SD above trend). 

Domestic Demand Momentum

For domestic demand, the picture is slightly different, and mostly tells a far more encouraging story. What’s not different is that the US plainly enjoys the most positive domestic demand momentum of these economies, and has does almost continuously since  1Q2011. But since August last year, the improvement in the US momentum trend has found close echoes in both China and (surprisingly) Japan.  

The improvement (relative to their recent experience)  is almost identical for both China and Japan – it is only fractional, but has been sustained now for the past four to five months. 

Finally, the charts confirm the Eurozone as a serious outlier, with domestic demand losing momentum for the past two years with no sign at all of any recovery, hampered by an unemployment rate which has risen almost uninterruptedly from 7.4% at the start of 2008 to 12% now.   Whilst the other major economies  can be seen to have their own cyclical patterns, and can also be seen to respond to other economies’ cyclical fluctuations, there  is no similar pattern in Eurozone domestic demand – rather, we have a continuous erosion of demand momentum. Unlike the rest of  the world, the Eurozone chart suggests Depression not cyclical recession.  

Imbalance and Likely Corrections

Finally, it is worth considering the difference between industrial momentum and domestic demand in each economy, on the basis that over time one would not expect industrial momentum to differ much from domestic demand momentum (and vice versa), in much the same way as one would not expect supply to differ that much, over time, from demand.  Where large deviations between the two occur, we might expect a tendency for them to reconnect – for example, if industrial momentum was sharply more positive than domestic demand, we might expect either industrial momentum to slow, or domestic demand momentum to accelerate.

The following chart, then, simply looks at 6m industrial momentum minus 6m domestic demand momentum. Where the line is positive,  industrial momentum is greater than domestic demand; where negative industrial momentum is not keeping up with domestic demand momentum.  The first thing to notice is that both the US and China are roughly balanced.  Second, for the last three years Eurozone industrial demand has run persistently stronger than domestic demand momentum (even though since 2008, the difference has averaged zero).  One would continue to expect either industrial momentum or domestic demand  momentum to change trend  in order to resolve this disequilibrium.  Personally, I think this suggests latent vulnerability in Eurozone industrial momentum.  Third, in Japan domestic demand momentum has survived better than industrial momentum over the last year. Japanese consumer confidence indicators tells that no abrupt collapse in domestic demand is anticipated: if so, these are grounds for expecting an upturn in Japanes industrial momentum in the short to medium term.  

Chinese Economic Momentum

Given the commentary on China's GDP and industrial data released today (Bloomberg: 'China Growth Loses Momentum in Blow to Global Expansion') I thought it might be useful to show what's happening to my momentum indicators for China's economy.
I compile three indicators: one for monetary conditions (tracks changes in money supply, fx, real interest rates, yield curve structure); one for industrial momentum (exports, both in Rmb value and in volume, industrial production, electricity production); one for domestic demand (retail sales, urban investment, auto sales, real estate climate index).
For each dataline I track, I express the previous month's movement as a number of standard deviations away from seasonalized trends.  Overall movements in underlying momentum are best captured using a 6m average.  The calendrical instability of Chinese New Year offers difficulties which I strive to deal with using as much information as I can muster, but the ramifications of Chinese New Year only really work their way out of the data by April.
I now have most of the data I need to construct these indicators, and where I do not yet have the data (auto sales, electricity generation, real estate climate) I have chosen to simply apply seasonal trends (ie, they are effectively neutralised).
The summary graph first:

Comment: Contrary to popular belief, monetary conditions are tightening, primarily as a result of the strength of the Rmb and positive real interest rates, though it is also true that monetary aggregates M1 and M2 are accelerating far less quickly than the growth of aggregate financing would suggest. Historically, when monetary conditions tighten, domestic demand is vulnerable, but so far momentum has only flattened out.  Industrial momentum is extremely volatile over the holiday period, but has exited with modestly positive momentum.