Foreign exchange (FX) swaps and their close cousins, FX forwards, are an important and growing segment of global financial markets.1 The latest BIS Triennial Survey indicates that, at $4.3 trillion, they accounted for some 65% of the average daily turnover in global FX trading in April 2019. In fact, they have been its fastest-growing component, accounting for roughly 75% of the increase in aggregate turnover since April 2016, the date of the previous survey (BIS 2019a, Schrimpf and Sushko 2019a). The amount outstanding at end-June 2019 was as high as $72 trillion, with FX swaps an estimated three-quarters of this total.2 For perspective, this figure was equivalent to 84% of global GDP, exceeded that of global cross-border portfolio stocks ($59 trillion) and international bank claims ($35 trillion), and was almost triple the value of global trade ($25 trillion). Not surprisingly, the US dollar reigns supreme. It is almost always one of the two currencies exchanged: 89% and 90% in terms of outstanding and turnover, respectively.
FX swaps and forwards play a key role in funding. Whenever an agent wishes to acquire, say, a US dollar asset on a hedged basis, they can do so in four ways: borrow dollars on an uncollateralised basis, enter a repo, enter an FX swap or, equivalently, combine a spot purchase of dollars with a forward sale, possibly with two different counterparties. The last three transactions are, in effect, forms of collateralised borrowing. It is only when an agent already has revenue streams and wishes to hedge them that they can simply use a forward. In this case, while not a financing vehicle, the forward is a form of debt. Of course, the instruments can also underpin speculative positions.
FX swaps and forwards differ from most other types of derivatives in one crucial respect: the notional amount is exchanged. They generate a debt obligation for the full face value of the contract. As a result, while notional amounts for other derivatives simply act as a reference for the much smaller amounts actually exchanged, the same is not true for FX swaps and forwards. Notional amounts must be paid and thus matter a great deal; in fact, there is nothing notional about them.
The special character of these instruments does not stop here (Borio et al. 2017). Their accounting treatment differs fundamentally from that of other forms of collateralised debt. Compare an FX swap with a repo. When an agent borrows against a security in a repo, the security stays on the balance sheet: the balance sheet is grossed up, as it includes the new borrowing and whatever is acquired through it. By contrast, if the same agent enters into an FX swap, using another currency as collateral, the new currency simply replaces the old one on the asset side of the balance sheet: the size of the balance sheet does not change.3 One transaction shows up as additional debt; the other is not treated as debt at all – the debt remains ‘hidden’. Why are instruments that are economically (roughly) equivalent treated so differently? Because accountants consider cash to be special.
This asymmetric treatment has important implications. One has to do with our understanding of the geography of FX swap and forward debt; the other with their regulatory treatment and financial stability implications more broadly. Consider each in turn.
The geography of FX swap and forward debt: US dollars outside the US
Out of sight may not quite be out of mind, but a lack of transparency does complicate things. When the Great Financial Crisis (GFC) broke out, the FX swap market came under substantial strain (Baba et al. 2009, McGuire and von Peter 2009), as funding in the wholesale unsecured segment froze. The extent of the strains took many by surprise, as did the underlying demand for US dollars, especially as this came from European banks. Had the amount of FX swaps and the banks in need been more broadly known, the surge would have been less unpredictable or at least more easily understood. The funding disruptions were so serious that they prompted major central banks to put in place FX swap arrangements to channel the necessary US dollar funding to those that needed it most.
McGuire and von Peter (2009) took a first stab at estimating non-US banks’ US dollar borrowing via FX swaps and forward debt more generally. Their key assumption was that banks generally avoided taking significant FX exposures, largely reflecting prudential regulations and risk management practices. In other words, they assumed that banks used FX swaps to ‘fill the gap’ in the net currency exposure shown on their balance sheets. Based on the same methodology, estimates of this gap derived from the BIS international banking statistics suggest that non-US banks’ net dollar borrowing via FX swaps was less than $1 trillion at end-June 2019. However, rough estimates based on some ancillary assumptions detailed in Borio et al. (2017) and derived from the BIS OTC derivatives statistics indicate that the corresponding gross amount exceeded $30 trillion, more than double their gross on-balance-sheet dollar debt ($13 trillion).4
What about the amount of US dollar debt in the form of FX swaps/forwards by non-banks (financial and non-financial entities) outside the US? Here, obtaining an estimate requires a triangulation between the aggregate amount of FX swaps/forwards outstanding and banks’ positions. Using the same ancillary assumptions, we estimate that, at end-June 2019, the amount was some $18 trillion or so. This was actually larger than the equivalent amount of on-balance sheet debt, estimated at $11.9 trillion.5
FX swaps, regulation, and financial stability
As a rule, prudential regulation generally follows accounting, with first-order implications for the treatment of FX swaps/forwards. Compare, again, an FX swap with a repo. While the corresponding cash flows are captured and treated equivalently in liquidity regulation, the picture is quite different for the leverage ratio in particular. As a first approximation,6 FX swaps are exempt; repos included in full. This is surprising, given that the two instruments are roughly equivalent from an economic perspective.
The implications can be substantial. Precisely because the instruments are off-balance sheet, a systematic analysis is not possible. Still, we just saw how large non-US banks’ dollar borrowing (on net) via FX swaps is and how the figures are an order of magnitude larger for gross positions. The net-gross distinction is bound to be especially large for banks acting as market-makers, which have both long and short positions in the instrument.
To be sure, we are not arguing for a specific treatment of repos and swaps. Nor are we saying that the treatment needs to be identical, at least if the uses of the instruments and broader implications for financial stability are considered. Indeed, a key function of FX swaps is to facilitate hedging. We are simply saying that such a large difference is odd. The difference has implications for regulatory arbitrage, for market dynamics around reporting dates and for business models more generally.7 Appropriately, the Basel Committee on Banking Supervision (BCBS) is monitoring banks’ behavioural responses to regulations, including potential regulatory arbitrage transactions and balance sheet optimisation techniques (BCBS 2019c).
Looking forward, regardless of their regulatory treatment, FX swaps are likely to come under strain again should global financial markets face stress at some point. Their role in US dollar short-term funding is so large and important that it could not hardly be otherwise. Almost three quarters (72%) of the average daily turnover in April 2019 was in maturities up to one month, adding to rollover pressure.8 Any financial stress in this market could easily also engulf the shorter-term repo market: both compete for spare dollar funding, which is quite concentrated these days (Dizard 2019). And this at a time when the authorities are grooming secured interest rates to become the new reference (Schrimpf and Sushko 2019b). In fact, the direction of contagion could go either way. The surprising ructions of September 2019 in the US dollar repo market had only a limited effect on the FX swap segment (BIS 2019b). But, in periods of extreme stress, even fully collateralised markets can seize up, as the events of the GFC showed.
FX swaps/forwards are a critical segment of global financial markets. Despite their role, the geography of their utilisation remains opaque. And, largely because of accounting conventions, their regulatory treatment differs markedly from that of instruments that, economically, are also forms of secured debt. Both of these aspects deserve more attention than they have generally received so far.
Our analysis has implications also for academic work on bank funding and lending patterns. That work generally has to rely exclusively on on-balance sheet data, for which the BIS international banking statistics are a key source. Authors should be aware and acknowledge that they are capturing only part of overall activity, often not even the larger one if the focus is on the US dollar. Here as elsewhere, ‘caveat emptor’ applies.
Authors’ note: The views expressed are those of the authors and do not necessarily represent those of the Bank for International Settlements.
Baba, N, F Packer and T Nagano (2008), ”The spillover of money market turbulence to FX swap and cross-currency swap markets”, BIS Quarterly Review, March, pp. 73–86.
Bank for International Settlements (2018), Annual Economic Report, Box III.A, “Banks’ window-dressing: the case of repo markets”, June, pp. 49–50.
Bank for International Settlements (2019a), Triennial Central Bank Survey: OTC interest rate derivatives turnover in April 2019, September.
Bank for International Settlements (2019b), “Easing trade tensions ease sentiment”, BIS Quarterly Review, December.
Basel Committee on Banking Supervision (2018), “Statement on leverage ratio window-dressing behaviour”, Newsletter, 18 October.
Basel Committee on Banking Supervision (2019a), “Basel Framework. Leverage ratio. LEV 30: Exposure measurement”, 15 December.
Basel Committee on Banking Supervision (2019b), “Basel Committee finalises revisions to leverage ratio treatment of client cleared derivatives and disclosure requirements to address window-dressing”, press release, 26 June.
Basel Committee on Banking Supervision (2019c), “The Basel Committee’s work programme”, 14 January.
Behn, M, G Mangiante, L Parisi and M Wedow (2018), “Does the G-SIB framework incentivise window-dressing behaviour? Evidence of G-SIBs and reporting banks”, ECB Macroprudential Bulletin 6, October.
Borio, C, P McGuire and R McCauley (2017), “FX swaps and forwards: missing global debt?”, BIS Quarterly Review, September, pp. 37–54.
Dizard, J (2019), “Derivatives to crash markets again?”, Financial Times, 1 November.
McGuire, P and G von Peter (2009), “The US dollar shortage in global banking”, BIS Quarterly Review, March, pp. 47–63.
Nakaso, H (2017), “Monetary policy divergence and global financial stability: from the perspective of demand and supply of safe assets”, speech to the International Bankers Association of Japan, 20 January.
Schrimpf, A and V Sushko (2019a), “Sizing up global foreign exchange markets”, BIS Quarterly Review, December, pp. 21–38.
Schrimpf, A and V Sushko (2019b), “Beyond LIBOR: a primer on the new benchmark rates”, BIS Quarterly Review, March,
1 The quantitative estimates of in this column are an aggregate of FX swaps, FX forwards and currency swaps, since separate statistics are generally not available for outstanding amounts. Currency swaps are FX swaps with a maturity longer than one year in which coupons are also exchanged. Our aggregate also includes non-deliverable forwards (NDFs), or forwards where the notional amount is not exchanged, making them materially different from swaps and deliverable forwards. It is not possible to exclude NDFs from the outstanding amounts, although turnover data for these instruments suggest that they are a single-digit share of the market. Figures for the average daily turnover and amounts outstanding of FX swaps and forwards are taken from the BIS Triennial Survey and the semi-annual OTC derivatives statistics, respectively. Figures for internationally active banks’ net on-balance sheet open currency positions are derived from the BIS international banking statistics.
2 The estimate of outstanding FX swaps (separate from outright forwards) is derived from the Triennial survey, where turnover figures for these instruments are reported separately. In what follows, whenever we refer to funding via FX swaps, we have in mind combinations of spot and forward transactions.
3 Outright forwards combined with a spot transaction with a different counterparty would be recorded similarly.
4 In addition to market-making activities (see below), the gross figure is boosted by the vehicle currency role of the US dollar. For instance, a European institution seeking to invest in a Thai baht asset may swap euro for dollars and then dollars for baht, i.e. both borrow and lend dollars via FX swaps.
5 See the BIS Global Liquidity Indicators for background.
6 This relates to counterparty risk, in the form of the market value of the instrument (replacement cost) and potential future exposures, which are included in both cases. But these are tiny in comparison to the principal amounts. For details, see BCBS (2019a).
7 For instance, it is well known that banks “window-dress” their balance sheets around reporting dates (BIS 2018, Behn et al. 02018). Indeed, the Basel Committee on Banking Supervision has issued guidance to address this problem (BCBS 2019b, 2018). Not least because of the regulatory treatment, the adjustment takes place largely via repos.
8 Of course, some authorities have been very alert to such risks. For instance, given the hundreds of billions of swaps of yen for dollars by Japanese banks, the Japanese authorities have encouraged their banks to extend the maturities of their swaps (Nakaso 2017).