Concerns about the health of the global economy, originated by adverse economic news and policy fumbles in China but amplified by poor advanced economy data, have dominated financial markets in 2016. Sharp oil prices and equity falls have seen a fresh wave of central bank dovishness: see here for my ECB preview, a March Fed hike now looks touch and go at best and pressure is growing on the BoJ to ease further on Friday (although this still seems unlikely). Recent events have illustrated that China/EME worries can impact the global economy via a wide variety of channels: e.g. China exporting deflationary pressures; trade links (see here); commodity price falls, risk aversion and financial market contagion.
Here I use BIS data to examine G10 banking sector exposures to EMEs, and the possible third wave of the global financial crisis. Loose G10 monetary policy encouraged substantial borrowing by a wide range of EMEs, but they now look to be heading for trouble with their bad debts consequently set to rise. That would represent the third wave of the global financial crisis which started with the initial US subprime crisis and was followed by the European sovereign debt crisis. Those bad debts will then feed back onto the Western banks which were so keen to lend to EMEs in the low-return environment, with consequent adverse impacts. The chickens set loose by G10 Central Bank policy could be about to come home to roost.
The main points are:
• Developing economies’ external banking debt rose from $2tn in 2006 Q1 to $4.8tn in 2015 Q2, driven by a $1.5tn rise in Asian debt (China up $0.7tn) and around $0.6tn rises in Developing Europe and Latam debt (Brazil up $0.3tn).
• EME prospects have deteriorated sharply, with the IMF seeing this as one of the main challenges for the global economy (along China’s hoped-for rebalancing and Fed liftoff).
• Banking sector exposures to China/EMEs are significantly larger than trade exposures but geographically spread very differently: double whammy effects look unlikely.
• UK banks have the largest exposures to developing economies overall (30% of UK GDP) and a disproportionate exposure to China and Asian economies (where this bias has become more apparent in recent years).
• The BoE FSR identified that a China/EME slowdown would cut UK banks CET1 ratio by 4.4pp, and I suspect there could be upsides here, helping account for the BoE’s recent volte face on EME developments.
• The aggregate euro area banking sector exposure to developing economies is substantially smaller (15.5% of GDP), although EA the share of EME claims in the total is in the same ballpark as the UK.
• Several individual EA members’ banks (Austria, Spain and Greece) have larger developing economies exposures to (up to 41% of GDP). But Spanish banks’ Latam exposures (32% of GDP, including a 10.5% of GDP exposure to Brazil), looks most worrisome given the extremely poor prospects there.
• Swiss banks are also relatively-highly exposed to EMEs (21% of GDP), but with a more diversified portfolio than UK banks.
• US banks have the lowest exposure to EMEs (4.3% of GDP) while the relatively low Australian banks’ exposure (11%of GDP) helps ameliorate the high trade exposure.
Developing economy external bank debt up $3tn (150%) since 2006 led by Asia
One of the side-effects of the extraordinary monetary policy measures introduced by G10 Central Banks to counter the effects of the global financial crisis was a surge in foreign borrowing by EMEs. Low borrowing rates encouraged EMEs to borrow while G10 banks were keen to lend given subdued demand and paltry returns available on domestic lending. Specifically, developing economies’ external banking debt rose from $2tn in early 2006 to a peak of $5.1tn in mid-2014 (i.e. a rise of $3.1tn or 150%), before subsequently falling slightly to stand at $4.8tn in 2015 Q2 (latest data). The chart below illustrates that most of this reflected a $1.5tn (peak) debt build up by Asian countries, with Chinese external banking debt rising by just under half of that ($0.7tn increase, $0.8 bn level in 2015 Q2). Latam and Emerging Europe debt both also rose by around $0.6tn, with Brazil accounting for around half of the former.
But EMEs are now in trouble, so bad debts set to rise
Unfortunately, EMEs are now slowing sharply or already in outright recession. The continued weakness in the Chinese manufacturing PMIs, which was an important initial driver of the 2016 market volatility, is shared by other prominent EMEs like Brazil and India. Hard data on industrial production confirm the downbeat picture – with China IP growth of 5.9% y/y being less than half that reported in 2014 and with Brazil IP falling by an eye-watering 12.4% over the past year. Asia industrial production is also generally weak, unsurprisingly given the strong regional links to the Chinese manufacturing powerhouse. Meanwhile Chinese GP growth has continued slowing (and seems likely to overstate the strength) while Brazil and Russia are competing for the sharpest GDP fall: the -4.5% y/y Brazilian GDP figure makes it the worst recession in a century, with the IMF last week downgrading its Brazilian 2016 growth forecast to -3.5% (see here). And today the World Bank joined in the negative commentary on EME prospects. So worrying times indeed for EMEs, with bad debts seeming pretty to likely to rise as a result.
Overall exposures to developing economies: UK largest but watch out for some EA members
So which banking systems are exposed to the likely increase in bad debts from the developing economies? The Chart below illustrates that UK banks’ 30% of GDP exposure to developing countries is the largest of the major banking sectors, compared to 20% of GDP for Swiss banks and 16% of GDP for Euro Area (aggregate) banks. Thats said, several indivdual EA members have larger exposures to developing countries – with both Spanish and Austrian banks claims both over 35% of GDP (and I illustrate below that the location of these exposures is different to the UK). By contrast, all the other major banking sectors exposures are less than 10% of GDP (US banks’ exposure is 4.3% of GDP, but the Fed has nevertheless shown concern about EME stresses).
The UK’s relatively large exposure seems to reflect UK banks’ large balance sheets (total external claims ammount to 120% of UK GDP) rather than a particularly large bias within that to EME lending. Specifically, lending to developing economies represents 25% of UK banks’ foreign claims, which is in the same ball-park as the EA and US banking sectors. Indeed, several EA member country banks’ have loan portfolios more skewed to developing economies (Austria, Spain, Portugal, Greece). That said, the Swiss banking sector looks perhaps to have been more canny in avoiding exposures to developing economies – only 9% of its 190% of GDP foreign claims are on developing economies.
Double whammies from banking and trade exposures look rare, banking exposures larger
An important point is that the banking sector exposures are distributed very differently to trade exposures: making it unlikely that countries suffer a double-whammy hit from the prospective EME problems. For example, the UK’s relatively-large banking sector exposure contrasts with the UK having the third-lowest trade trade exposure to developing economies (which MPC member Kirsten Forbes last year quoted as a reason to be sanguine about EME developments). Conversely, the Netherland, Germany and Australia have the largest trade exposures but relatively-low banking sector links. That said, several EA members (Spain, Greece) have both medium trade exposures and high banking sector exposures.
But it’s also notable that the banking sector exposures to developing economies are several times larger than their trade exposures: the highest trade exposure of (11% of GDP for Netherlands) is dwarfed by the near-40% of GDP banking sector exposures of Austria and Spain. So the banking channels seems like a more potent contagion channel (via higher bad debts).
But it’s important to dig into the regional distribution of different banking sector EME exposures – EMEs in different regions may end up faring differently and hence generating different levels of bad debts.
UK banks exposed to Asia
And doing so makes apparent that the UK’s relatively-large banking sector exposures are driven by lending to China (6.5% of UK GDP) and other developing Asian countries (11% of GDP). Moreover, adding in the more developed Asian economies such as Hong Kong (11.8% of GDP) and Singapore (3.4% of GDP) brings the UK bank’s total Asian exposure to around 31% of UK GDP. So UK banks should be taking seriously the potential write-offs generated by the combination of high debt (likely exacerbated by currency depreciations, see BIS work here), weakening growth, overcapacity and capital outflows affecting Asia. And that could feed back onto lending to the UK economy.
And it’s worth stressing that any Asian bad debt issues will affect UK banks signifiantly more than their international compatriots: the closest comparators to UK banks’ large 31% of GDP China/Asia exposure are a long way back with Swiss banks at 12% of GDP exposure, Japanese banks at 10% and Australian banks at 8% of GDP. Gulp.
Of course, much of this exposure arises for historic reasons and is particularly pronounced in two institutions (HSBC and Standard Chartered). But it’s notable that UK banks haven’t really reduced their exposures to developing economies and (broad) Asia in recent years. And that contrasts with the relatively-sharp cut in overall foreign exposures, from around 170% of GDP in 2010 to 119% of GDP in 2015. So UK banks have actually become more biased to Asian/developing economy exposures in recent years: the share of developed economies in total foreign exposures nearly doubled from 12.9% in 2006 to 25.4% in 2015. Double gulp.
And the December BoE Financial Stability Report (see here) illustrated that in a stressed scanario, involving a sharp China/EME slowdown, UK banks’ CET1 ratio would fall by 4.4pp (from a baseline of 12.0% to 7.6%). That reflected a 1.8pp hit from loan impairments, 1.6pp of traded risk losses an 0.3pp lower net interest income. So prretty chunky numbers, although the FSR stressed that the UK banking system would be able to mantain its core functions. But the FSR also noted that the scenario isn’t a worst case. For example, while the scenario included impairments on Chinese/Hong Kong loans more than tripling to almost 5% it’s unclear whether the same assumption as applied to the remaining 40% of UK banks’ Asian exposures. If it hasn’t then the impacts could potentially be 40% larger. And as ever the results are only as good as the (imperfect) models used to generate them.
EA members exposed to Emerging Europe and Latam (Brazil)
By contrast, the aggregate EA banking exposure to China/Asian countries is only 3% of GDP. Even Dutch and French banks’ larger exposures (9.1% and 5.7% of GDP) are chicken feed in comparisn to the UK.
Rather, EA banks foreign exposures are more to latam and developing Europe: 4.4% and 6.9% of GDP respectively at the Euro Area aggregate level. But those relatively-small aggregate numbers conceal substantially larger exposures for individual EA members. The largest and most widespread exposures are to developing Europe. Here Austria leading the way (40% of GDP claim), Greece isn’t far behind (28%) while Belgium and the Netherlands both come in with over 9% exposures. While something to watch going forward, developing Europe isn’t the centre of current concerns.
Rather it’s Spain’s large Latam exposure (32% of GDP) which stands out, given the previously-discussed concerns about Brazil. Indeed, Spanish banks’ claims over Brazil make up a chunky 10.5% of GDP, reinforcing the nascent worries, although such claims have been falling (second chart, although the impact is offset by falling GDP during the period).
Swiss banks have relatively high exposures, but more diversified than UK banks
I’ve already dropped in a couple of nuggets about Swiss banks above. But to re-iterate:
• They’re second to UK banks in terms of their overall EMEs exposures (21% of GDP), despite their overall balance sheet being significantly larger (as they’re less biased to EME lending, only 9% of the total)
• Their exposure to China (2.6% of GDP) and Asia in total (around 12% of GDP) is significantly smaller than the UK – although the latter is nevertheless the second-highest overall.
• Ovearall their EME exposure is more diversified than UK banks – they also have medium-sized exposures to Latam (5% of GDP) – although this can also be spun as them being exposed to more shocks rather than less.