In its latest annual summary published at the end of June, the IIF found that total nominal global debt had risen to a new all time high of $217 trillion, or 327% of global GDP...
... largely as a result of an unprecedented increase in emerging market leverage.
While the continued growth in debt in zero interest rate world is hardly surprising, what was notable is that debt within the developed world appeared to have peaked, if not declined modestly in the latest 5 year period. However, it now appears that contrary to previous speculation of potential deleveraging among EM nations, not only was this conclusion incorrect, but that developed nations had been stealthily piling on just as much debt, only largely hidden from the public eye, in the form of swaps and forwards.
According to a just released analysts by the Bank of International Settlements, "FX swaps and forwards: missing global debt?" non-banks institutions outside the United States owe large sums of dollars off-balance sheet through instruments such as FX swaps and forwards. The BIS then calculates what balance sheets would look like if borrowing through such derivative instruments was recorded on-balance sheet, as functionally equivalent repo debt, and calculates that the total "is of a size similar to, and probably exceeding, the $10.7 trillion of on-balance sheet dollar debt", potentially as much as $13-14 trillion.
* * *
The BIS then provides substantial background data on who, where and how uses FX swaps (as both a lender and borrower), as well as where this "missing debt" can be found when looking away from the balance sheet. Here are the details:
Every day, trillions of dollars are borrowed and lent in various currencies. Many deals take place in the cash market, through loans and securities. But foreign exchange (FX) derivatives, mainly FX swaps, currency swaps and the closely related forwards, also create debt-like obligations. For the US dollar alone, contracts worth tens of trillions of dollars stand open and trillions change hands daily. And yet one cannot find these amounts on balance sheets. This debt is, in effect, missing.
As a result, we know little about it. How much is owed, by whom and for what purpose: trade hedging, asset-liability management, market-making? What does it imply for measures of international credit like the BIS global liquidity indicators (GLIs)? Answers to these questions can inform assessments of global financing conditions and financial stability. For instance, serious strains seized the FX swap market during the Great Financial Crisis (GFC). In response, central banks had to replace lost dollar funding that financed dollar assets.
This special feature frames the issues and suggests some answers. To do so, it breaks ground in combining data on the aggregate amount of outstanding derivatives contracts (from the BIS derivatives statistics) with information from the international banking statistics and from ad hoc surveys to form a view of the size, geography and use of the missing foreign currency debt. The more detailed analysis focuses on the dollar segment, given the currency's outsize role in the foreign exchange and other financial markets.
The debt remains obscured from view. Accounting conventions leave it mostly off-balance sheet, as a derivative, even though it is in effect a secured loan with principal to be repaid in full at maturity. Only footnotes to the accounts report it.
To be sure, the main reason to resort to derivatives such as swaps and forwards is precisely to keep it as obscure and hidden from the public eye as possible. Further complicating matters is the ongoing debate of just what happens when gross derivative notional collapses to net, and just where the liability resides. What is publicly known and disclosed largely remains in the domain of gross notionals; the BIS summarizes it as follows:
The outstanding amounts of FX swaps/forwards and currency swaps stood at $58 trillion at end-December 2016 (Graph 1, left-hand panel). For perspective, this figure approaches that of world GDP ($75 trillion), exceeds that of global portfolio stocks ($44 trillion) or international bank claims ($32 trillion), and is almost triple the value of global trade ($21 trillion). The outstanding amount has quadrupled since the early 2000s but has grown unevenly (Graph 1, left-hand panel). After tripling in the five years to 2007, it fell back sharply during the GFC, even more than international bank credit. This most likely reflected a reduction in hedging needs, as both trade and asset prices collapsed.
Another problem with such debt-masking derivatives is that they are largely short-term in nature, a problem which violently emerged during the Global Financial Crisis when an overnight freeze in shadow banking and other non-bank lending conduits effectively resulted in a liquidity crunch for many of the world's largest financial institutions, forcing central banks around the globe to step in and provide trillions in short-term funding.
The maturity of the instruments is largely short-term (Graph 1, centre and right-hand panel). At end-2016, three quarters of positions had a maturity of less than one year and only a few percentage points exceeded five years. Turnover data show that the modal forward (a customer-facing instrument) matures between one week and one year while the modal swap (an inter-dealer instrument) within a week (BIS (2016)). The long-term share has risen since the 2000s, as capital markets have boomed.
In terms of currency distribution, it's all about the US dollar, or as the BIS puts it, "the dollar reigns supreme in FX swaps and forwards. Its share is no less than 90% (Graph 2), and 96% among dealers (Table 1)."
Both exceed its share in denominating global trade (about half) or in holdings of official FX reserves (two thirds). In fact, the dollar is the main currency in swaps/forwards against every currency. For instance, it predominates in forwards in the Norwegian krone, the Swedish krona and the Polish zloty, currencies that trade in the spot market more against the euro
With north of $50 trillion in total notional FX derivatives, it is hardly surprising that the daily trading volumes are massive: while outstanding amounts lump FX swaps with forwards, turnover data show that FX swaps are the instrument of choice. Swaps/forwards and currency swaps amounted to over $3 trillion per day in 2016, over 60% of total FX turnover. Of that, FX swaps accounted for three quarters, forwards for 22% and currency swaps for the rest.
The next question is who are the counterparties reliant on such FX exposure.
According to the BIS, the counterparty breakdown of outstanding amounts provides a big-picture indication of obligors and, indirectly, uses (Table 1). Reporting dealer positions vis-à-vis customers amount to $33 trillion, or over half (57%) of the total; the rest, $25 trillion, are positions among reporting institutions themselves ("inter-dealer"). This large share is a trademark of decentralised, over-the-counter (OTC) markets. Non-dealer financial firms bulk large among customers, with a share of about 80%.
A further breakdown of the client base reveals widespread usage among both non-financial, non-bank financial as well as bank entities. A quick summary of each category, as well as their principal considerations:
- Non-financial customers: international trade and liability hedging. Apart from speculative use, the non-financial sector employs FX forwards and currency swaps to hedge international trade and foreign currency bonds, notably those issued to cheapen funding costs. Since most international trade contracts are short-term, forwards serve as hedges. One can relate non-financial FX swaps/forwards and currency swaps, in an admittedly stylised fashion, to international trade and bond issuance, respectively. If firms use $5.1 trillion of short-term FX forwards to hedge global trade of $21 trillion, then the ratio implies that importers and exporters hedge at most three months' trade. Similarly, if firms and governments use $2.4 trillion of currency swaps to hedge $4.8 trillion of international bonds, then they hedge half or less.
- Non-bank financial customers: hedging assets and liabilities. Other non-bank financial firms come in two groups. The first comprises institutional investors, asset managers and hedge funds that use FX forwards to hedge their holdings and to take positions. The second includes non-bank financial firms that use currency swaps to hedge their FX bonds. A quick look at the two groups: institutional investors and asset managers use outright forwards (90% and 69%, respectively), followed by FX swaps (48% and 33%) and, to a lesser extent, currency swaps (37% and 19%). Hedge ratios vary by asset class. Those for foreign currency bonds range from 50 to 100%; those for foreign equities 20 to 60%. Heavy use of FX forwards widens maturity mismatches. The $6.6 trillion in currency swaps that non-bank financial firms have contracted stand at almost 80% of their outstanding international debt securities.
- Banks: FX swaps were a key part of non-US banks' total US dollar funding, amounting to an estimated $0.6 trillion, roughly 6% of the total in March 2017. The rest, about $9.4 trillion, mostly took the form of deposits from US and non-US non-banks (red and blue areas), and dollar debt securities (yellow area).
Drilling down into the last category, which is of most relevance for the topic of "masking" debt exposure, the BIS writes that banking systems can be sorted into US dollar net borrowers and lenders via FX swaps (Graphs 5 and 6). As noted, net borrowers' balance sheets show more dollar assets than liabilities, offset by missing off-balance sheet dollar debt. Net borrower banks use these dollars, along with those from FX reserve managers (red lines), to fund claims on non-banks (green lines) and other banks (solid and dashed blue lines).
Based on the chart above:
- Japanese banks had by far the largest on-balance sheet mismatch in dollar positions at end-Q1 2017 (top right-hand panel). They were net borrowers of an estimated $1 trillion via FX swaps.Borio et al (2016) argue that constraints on global banks' balance sheets, combined with increasing demand for cross-currency funding by Japanese banks, pension funds and other non-bank financial entities, led the yen/dollar basis to widen after 2014.
- Several large European banking systems also draw dollars from the FX swap market to fund their international dollar positions (top centre panel). Pre-GFC, German, Dutch, UK and Swiss banks, in particular, had funded their growing dollar books via interbank loans (blue lines) and FX swaps (shaded area). The meltdown in dollar-denominated structured products during the crisis caused funding markets to seize up and banks to scramble for dollars. Markets calmed only after coordinated central bank swap lines to supply dollars to non-US banks became unlimited in October 2008. Post-GFC, these European banks' aggregate dollar borrowing via FX swaps declined, along with the size of their dollar assets. In particular, German, Swiss and UK banks reduced their combined reliance on FX swaps from $580 billion in 2007 to less than $130 billion by end-Q1 2017.
- Australian banks (bottom left-hand panel), have relied on direct dollar interbank and bond borrowing (blue line) to fund mostly Australian dollar investments at home, and with an estimated $200 billion in FX swaps hedging the currency risk. Selected European and non-European banking systems add almost $300 billion to this total.
To summarize, non-US banks' net dollar lending through the FX swap market falls short of their net borrowing, with the gap widening over time (Graph 6, right-hand panel); dollar borrowing against other major currencies, like the yen and the euro (left-hand panel), exceeds dollar lending against secondary and emerging market currencies (centre panel).
* * *
Now that we know the borrowers, who are the lenders to non-US banks vis FX swaps? Four candidates are: US banks, central banks, European agencies and supranational organisations, and private non-banks. All of these appear to provide some funding, with US banks and central banks together closing about half the gap.
- As regards the first candidate, US banks naturally lend dollars via FX swaps: $150 billion in the latest data. This figure combines positions from offices outside and inside the United States.
- Second, central banks lend dollars via FX swaps against either their own currency or third currencies. Against their own currency, some Asian central banks provide about $200 billion in swaps as they manage the domestic liquidity consequences of FX reserve accumulation (Graph 8, left-hand panel). They first buy dollars spot (increase their FX reserves) and then drain domestic liquidity by swapping (lending) the dollars for (against) domestic currency. On net, however, central banks' dollar supply against their own currencies is close to zero, since other central banks are actually borrowing dollars via FX swaps. They do so in order to finance their accumulation of FX reserves without incurring currency risk ("borrowed reserves"). Against foreign currencies, some central banks lend dollars via swaps in the management of their FX reserve portfolio. For instance, the Reserve Bank of Australia swaps US dollars for yen (Debelle (2017)). We estimate that such operations by reserve managers sum to at least $300 billion
- Third, European supranationals and agencies have opportunistically borrowed dollars to swap into euros to lower their funding costs. While their operations mostly require euros, they have done so to take advantage of the breakdown in covered interest parity. Five European supranationals and agencies together had over $400 billion in dollar debt in June 2017. These alone have provided $300 billion in swaps against the euro.
- Fourth, non-bank private sector entities can provide hundreds of billions of dollars. Like US banks, US-based asset managers are obvious candidates. In June 2014, the then largest US bond fund, PIMCO's Total Return Fund, reported $101 billion in currency forwards, no less than 45% of its net assets. Since the overall US holdings of foreign currency bonds were $600 billion at end-2015, a 50% hedge ratio would extrapolate to $300 billion. On the equity side, US investors' hedge ratios are thought to be 40-50% for advanced economy equities. In addition, US firms that hold cash in offshore affiliates to avoid US corporate tax on repatriated earnings could be sizeable lenders as well.
* * *
Putting the above together, the BIS calculates that according to its estimates of "on-balance sheet" dollar debt equivalents of non-banks outside the United State, such cash market obligations, both bank loans and bonds, totalled $10.7 trillion at end-March 2017. The BIS then calculates the corresponding additional debt borrowed through the FX derivatives markets, i.e. "off balance sheet": it finds the answer to be in a similar magnitude: "the missing debt amounts to some $13-14 trillion."
Here's the calculation:
First, the BIS OTC derivatives statistics report total dollar-denominated forwards and swaps outstanding vis-à-vis customers (our proxy for non-banks) of $28 trillion (in parentheses in Table 1). Second, we need to make an assumption about the direction of non-banks' positions. We could assume that the amount of dollars lent and borrowed through derivatives by customers is matched. This would require reporting dealers, as a sector, to be also balanced: data slippage aside, customers and dealers make up the whole market. If so, to obtain the total amount of non-banks' gross dollar borrowing through FX forwards, one would divide by two. This gives $14 trillion. The previous section, however, suggested that banks as a whole use the market for net dollar borrowing. If so, one could subtract that amount from the total before dividing. Even then, an upper estimate of the banks' net position would be, say, $2 trillion. This would imply (after dividing the remaining amount by two) a lower amount of non-bank dollar borrowing of $13 trillion. Finally, we need to estimate the fraction of that debt held by non-US residents. If US residents use the bulk of their derivatives for hedging purposes, the amount would be small.
Why is this number important? The BIS' Claudio Borio writes that "regardless of whether the off-balance sheet debt is currency-matched or not, it has to be repaid when due and this can raise risk. To be sure, such risk is mitigated by the other currency received at maturity. Most maturing dollar forwards are probably repaid by a new swap of the currency received for the needed dollars. This new swap rolls the forward over, borrowing dollars to repay dollars."
In other words, the bucket can be kicked indefinitely... unless there is a sharp discontinuity in the value of the dollar, such as what happened during the 2008 financial crisis, forcing the Fed to become a US dollar provider, both on and off balance sheet, of last resort, to avoid a collapse of the financial system which effectively has a massive, multi-trillion dollar bet on its books at any given moment.
Borio confirms as much in the following paragraph:
Even so, strains can arise. In particular, the short maturity of most FX swaps and forwards can create big maturity mismatches and hence generate large liquidity demands, especially during times of stress. Most spectacular was the funding squeeze suffered by many European banks during the GFC. Indeed, in response, the Swedish bank supervisor has applied liquidity requirements separately to banks' dollar and euro positions (Jönsson (2014)). But non-banks may also face similar problems when they run such mismatches, both on- and off-balance sheet. During the GFC, central banks extraordinarily extended dollar credit to non-banks in Brazil ($10 billion programme) and Russia ($50 billion). Moreover, even sound institutional investors may face difficulties. If they have trouble rolling over their hedges because of problems among dealers, they could be forced into fire sales,
Going back to the core point of the article, namely the ability of financial and non-bank entities to transform trillions in dollar obligations from on to off-balance sheet debt, in the process obfuscate funding needs and risk exposure, the BIS concludes as follows:
Obligations to pay dollars incurred through FX swaps/forwards and currency swaps are functionally equivalent to secured debt. In contrast to other derivatives, agents must repay the principal at maturity, not just the replacement value of the position. Moreover, they could replicate those positions through transactions in the cash and securities markets that would show up on-balance sheet. But because of accounting conventions, this debt does not appear on the balance sheet: it has gone missing.
This is a major problem for the regulator of global central banks which admits that this attempt to mask exposure, "greatly complicates any assessment of the missing debt's total amount and distribution, and hence of its implications for financial stability." While the BIS concedes that much more analysis will be necessary before making definitive policy recommendations, the implicit message is that the next time a USD funding crisis strikes, in addition to the trillions in dollar exposure already known, the world will be faced with a USD shortfall potentially as great as $14 trillion. All that would be needed is a sharp repricing in the US dollar, which in turn would likely be catalyzed by another financial crisis. The result would be another central bank firedrill, in which the Fed will have to double down on its emergency global bailout last observed in the days after the Lehman failure, and which we descibed in October 2009 in "How The Federal Reserve Bailed Out The World".
The only difference is that next time, the total dollar injection required will be far, far greater...