The domestic policy response by the Federal Reserve to the current health crisis has thankfully been swift. While monetary policy makers across the world understandably focus on dislocations in their domestic markets first, in a globalized and, more crucially, dollarized world, it falls to the Fed to attenuate global USD funding pressures in order to guarantee the medical response to unfold as unconstrained as possible.
US policy makers have to date largely followed the 2008/09 playbook by reestablishing FX swap lines with foreign central banks. While no doubt necessary, these highly unusual conditions, further strained by understaffed desks, may impede transmission mechanisms and require streamlining or expansion of the Federal Reserve’s international response.
Two ideas will be discussed in this post:
- Expanding the role foreign central banks play in the transmission of USD liquidity by installing them as temporary, localized ‘primary dealers’.
- Instead of reliance on only passive, auction-style provision of USD via currency swap lines, which require pledging of preexisting collateral by local banking systems, central banks might consider also offering USD via regular FX swaps, allowing much more seamless transmission by intermediaries to the buyside. This format would open the possibility of intraday backstopping and/or targeting of X-CCY markets by the recipient central bank alone, or in conjunction with the Federal Reserve via US regulated intermediaries.
Implementation of both ideas would seem relatively straight forward, leave the Federal Reserve in a at minimum as secured position as under current programs, move the USD liquidity provision offshore to an intraday format (as onshore) and eliminate unnecessary strains on US dealers, which currently have to absorb international spillovers.
Foreign central banks as localized primary dealers
1. Consider the creation of a repo facility for foreign official institutions (FOIs). For fastest rollout, any Treasury or Agency security held by such actors in custodial accounts provided by the Fed should serve as collateral. Of the roughly USD 4.5tn Treasuries and Agencies owned by FX reserve managers and sovereign wealth funds with US custodians, about 70% are held in such accounts already. In the event of temporary liquidity shortages, such actors should not have to exert avoidable pressure on private sector intermediaries. If need be, and with slightly more effort, securities held with global custodians (Bank of NY Mellon for the most part historically) could be included, opening the space for extending the underlying collateral basket to mirror, for instance, the PDCF basket.
2. Like primary dealers in quantitative easing auctions, foreign central banks should be able to directly sell Treasury and Agency securities to the Desk at recently quoted bid prices. There is no need to burden dealers warehousing such securities before selling these on to the Federal Reserve under regular QE auctions.
3. If both of the above conditions were met, foreign central banks could distribute dollar liquidity globally without dependence on US dealers. More concretely, encourage foreign central banks to
- lend USD against Treasury/Agency securities held by local private sector institutions by
- having these shift (by way of another repurchase agreement) securities from private custodians into the local central bank’s account with the Federal Reserve, so that utilization of the aforementioned repo facility is possible.
Collateral choice could again be expanded to include securities issued by the private sector. Also, as above, sales by overseas private institutions could be intermediated by respective local central banks, receiving intraday credit to facilitate these purchases and be settled by the subsequent on sale to the NY Fed’s Desk.
Continuous, direct intervention in FX swap markets
Probably even more important at this time is the backstopping of FX-hedged fixed income portfolios of overseas investors. The current design of central bank currency swap lines is largely a result of the GFC, when overseas banks themselves were hit by a drought in synthetic private sector USD funding. This time, the institutions in need are a broader set of non-bank institutions, possibly requiring tweaking in the design of swap lines to ensure system-wide transmission.
By way of brief review:
Over the past decade, USD-denominated fixed income markets have seen unprecedented amounts of inflows from overseas non-bank private sector institutions. Of the USD ~3tn acquisitions of such securities this cycle, about 65% were FX-hedged coming into the year.
The largest counterparties to these FX hedges are FX reserve managers, allocating portions of their liquidity tranches to lending USD in the FX swap market, as well as a plethora of individually smaller private sector actors doing so as well for a yield pick-up resulting from deviations from Covered interest rate parity.
Current stress in FX basis markets is due to three reasons:
- Private sector arbitrageurs (mostly US based) pull back USD lent via swaps as balance sheet exposures generally are reduced and dislocations in US FI markets allow higher profits on an equally fully collateralized basis.
- Central banks may step back from providing USD via FX swaps due to actual (or anticipated) declines in FX reserves. FX reserve managers in normal times act as natural backstop for G10 FX basis markets: they are, unlike private sector actors, unconstrained by regulations and can evaluate sovereign STIR markets on an unlevered basis. However, central banks are slow actors and increases in USD lent via FX swaps are typically not done at quick pace, much less now, with the domestic situation the primary concern in each country.
- Perhaps most importantly, in a crisis, FX hedge ratios frequently increase sharply as henceforth unhedged exposures aren’t viable from a risk management perspective, with volatility in currency markets increasing markedly. Most foreign acquisitions of USD debt this cycle came from debt or debt-like funded real money accounts (lifers, pensions, FI funds, specialist banks) for which the FX risk can quickly exceed even volatile FI exposures.
Why do US policy makers care about these dynamics?
Similarly as when repo markets dry up and require liquidation of collateral, unavailability of reasonably priced FX hedges may equally lead to sales of collateral, creating spillovers back into US markets, if bonds were issued in the United States. Actions by the Federal Reserve are thus, and as the Fed itself notes, not only out of benevolence for the rest of the world, but also out of self-interest, avoiding overseas spillover effects further straining the functioning of US financial markets.
The current design
Under the current system, when a foreign central bank (here, and without loss of generality, the Bank of Japan) calls funds from the Federal Reserve by way of a currency swap line,
- the central banks will exchange USD for JPY respectively at current FX spot rates, and agree to a reexchange these at the same exchange rate at contract maturity.
- Before usage by one of the parties, the funds received are placed in segregated accounts at the other CB, only to be used as part of the currency swap transaction.
- Interest rate payments are handled differently than in cross-currency swaps among private sector parties, with only the initiator (here the BoJ) paying USD interest (at the now lowered rate of OIS +25bps) on the received funds and the Fed agreeing to leave JPY funds received unremunerated in its account with the BoJ.
In terms of optics, the lending of USD by the Fed via swap lines comes, as a result of this design, as close as possible to its regular collateralized lending transactions with domestic counterparties, only that now the collateral is foreign currency held with another central bank.
The amount the BoJ will request is determined by the results of an (in size constrained or full allotment) auction held previously with the local, Japanese banking system, which will subsequently receive the drawn USD funds from the BoJ on a collateralized basis. The collateral basket and haircuts central banks apply is normally in line with the specifications for domestic currency operations, plus frequently a risk factor accounting for currency risk the central bank takes on when advancing USD against non-USD collateral. Delivery of USD-denominated collateral is frequently also possible.
This system is optimal when the bank the BoJ lends USD to is the ultimate recipient in need of USD funding. Such a bank typically has a preexisting holding of USD assets which were funded (or hedged) by an FX swap (or longing the domestic currency in the forward market) with a private sector counterparty. Given the unwillingness of the latter to roll over contracts at reasonable rates, the bank simply pledges the preexisting asset (or, alternatively, any other previously unencumbered asset) to the BoJ and repays the private sector swap counterparty with USD received from the BoJ.
In this instance, the bank is not acting as a dealer for USD liquidity with a net zero position, but as a liquidity taker. The balance sheet size of the bank remains unaffected if it so wishes: it continues to hold a USD-denominated asset which it pledges to the central bank in return for USD funds, which it in turn delivers into the FX swap, in return for its initial yens. Then, should it wish, it can use these to repay the initial yen funding on the liability side of its balance sheet.
It might be self-evident, but one fact of this transmission mechanism will become very important in a moment, so it is worth reiterating. The method by which the central bank drawing USD funds from the Fed currently passes these on to its private sector is mechanistically indifferentiable from when local currency funds are lent in that banks have to pledge securities in return for currency.
Today’s challenge of insufficient intermediation
The present demand for synthetic USD funding differs greatly from the 2008/09 picture, with non-bank institutions in Asia and Europe the primary takers. In such a situation, central banks’ task of ensuring system-wide distribution of USD funds drawn via swap lines is much harder.
The current design of currency swap lines relies implicitly on the fact that by providing USD funds on a collateralized basis cheaper than where FX markets trade will invite sufficiently large arbitrage to close this gap. The central banks’ behavior in this is quite passive, counting merely on banks’ profit motive, perhaps paired with moral suasion. In situations of acute stress, this might not be enough, especially as the current design does not allow efficient matched-book dealing in FX derivatives by intermediaries.
When a dealer considers intermediating relatively cheaper USD funding received from the BoJ to a non-bank, it will first have to pledge securities it presently holds to the BoJ in return for USD. Then, in a second step, it will enter (or renew) the FX swap contract with the non-bank, the ultimate source of demand. Aside from haircuts and margin requirements imposed by central banks (already increasing the costs above the OIS +25bps level offered by the Fed), the major issue for any singular dealer is that such an activity will increase its balance sheet. A desk will receive USD by repoing securities to the central bank, lend USD via FX swaps to a non-bank counterparty and as a result end up with, in its view, superfluous yen deposits.
The resulting increase in balance sheet leverage for the intermediating dealer could hardly come at a less opportune time, with all segments of the associated bank likely under heighted stress and already trying to delever non-essential operations. This unwillingness to accept additional balance sheet pressure is only reasonable and seems to meaningfully hinder transmission, keeping cross-currency basis markets far deeper in negative territory than the Fed’s inexpensive pricing might suggests.
From repos to FX swaps, from discrete to continuous USD liquidity provision
At a later date, a broader discussion might be necessary in order to assess whether highly regulated financial institutions (e.g. life insurers), which under normal conditions do not directly interact with the central bank, should be given more direct access to central bank funds in a crisis situation.
At present, and with much lower operational effort, streamlining how intermediaries receive USD swap-drawn funds could have equally sized effects. More specifically, central banks should not only on-lend USD drawn from the Federal Reserve via transactions collateralized by securities, but should in addition also directly enter into FX swaps with the banking system. Said differently, instead of only accepting securities as collateral, central banks should also allow ‘delivery’ of local currency as collateral.
FX swaps, unlike repos, do not increase balance sheet size when done on a matched-book basis as they receive derivative-accounting-status, despite an exchange of principal at trade initiation and termination. This much more balance sheet light treatment would thus, even should central banks insist on margin requirements exceeding private sector standards, allow much improved transmission of USD funds from the central bank to non-bank sources of demand.
In its simplest form, the provision of USD via FX swaps could simply be added as an alternative delivery option for banks participating in the current auction system at pre-specified times.
It is unclear how beneficial exactly this additional option would be on its own, it would however open the door to further, more aggressive measures central banks could take to more effectively alleviate stress in FX funding markets.
1. Instead of passive reliance on arbitrage considerations of intermediaries, central banks, after calling USD from the Federal Reserve, could initiate FX swap transactions directly as aggressor, pushing USD actively into markets and likely eliciting a much stronger price response.
2. If central banks feel comfortable in this more involved role, it would be straight-forward to supplement the auction format with an intraday provision of USD liquidity. For instance, central banks could target the very front end (< 1w) aggressively, ensuring sufficient intraday liquidity in case unforeseeable and hard to bridge pressures emerge ahead of the next formal auction. If wished, the cross-currency basis curve could also be targeted directly, for instance by establishing a hard emergency floor at a more-costly-than-auction but still relatively contained level. Of course, continuous provisions of USD via FX swap markets would necessitate the Federal Reserve to allow overseas central banks to draw on swap lines not only around scheduled auctions, but continuously as well.
3. Finally, direct interventions need not only be enacted by the foreign central bank but could also be initiated by the Federal Reserve directly, should it consider dislocations severe enough. Lacking the connections a relevant local central bank might have and assuaging issues of oversight, the Federal Reserve could limit its transactions to US resident financial institutions under its supervision. The global nature of this group should suffice in all but the most exotic cases to transmit USD funding to the ultimate non-bank institutions abroad.
The foreign currency the Fed will attain as a result of the FX swap will be placed (as in the more conventional case) with the respective central bank. The only difference is its precise location, no longer in the designated currency swap line account, but rather in the regular, interest-bearing accounts central banks provide another on a reciprocal basis. As long as the remuneration these accounts provide to the Fed are in line with market rates, the Fed will end up with a PnL at least great as the official pricing of the swap line.
The only minor difference from the Fed’s perspective is the counterparty risk it assumes when current FX spot markets deviate from those at initiation, which now lie with a US-based private sector institution instead of a foreign central bank. Daily exchanges of variation margin and the insistence to transact with well capitalized banks only should serve to sooth concerns of this nature.
image credit: Bank of England, Financial Stability Paper No. 36 – February 2016: Stitching together the global financial safety net by Edd Denbee, Carsten Jung and Francesco Paternò