9. The Sovereign–Bank Nexus via OTC Derivatives
- Manmohan Singh
- Published Date:
- October 2016
This chapter focuses on the role of collateral in the over-the-counter (OTC) derivative contracts between sovereigns and large banks. Specifically, due to the sizable volume of business (and associated revenue), most banks do not force sovereigns to post collateral when the sovereigns are “out of the money” on their derivative contracts. However, if banks are out of the money, they generally have to post collateral. The rhetoric about cutting the umbilical cord between banks and sovereigns will not get full traction unless sovereigns post collateral on their derivatives contracts with banks. Estimates of “out-of-collateral” positions are not trivial and thus cannot be ignored when discussing the sovereign–bank nexus. The official sector is presently focused on appropriate haircuts (and risk weights) for the sovereign bonds; presently the flexibility allows for inconsistency (and thus overvaluation of the bonds) relative to other risk metrics (Hannoun 2011). This chapter suggests that there are par amounts (and not only haircuts to par) that need to be on the radar screens also!
As we have already seen in Chapter 6, present market practices for those using OTC derivative contracts result in residual derivative liabilities and derivative assets. By “residual” we mean what is left after all possible netting has been achieved within the OTC derivatives book of a large bank, and after subtracting the (limited) collateral posted on the contracts. Thus, this residual risk captures the shortfall of collateral stemming from clients of large banks not posting their share of collateral to the banks and vice versa – loosely called one-way CSAs (credit support annexes) rather than two-way CSAs, whereby both parties to the contract post collateral to each other. Earlier research finds that the 10–15 largest banks in the OTC derivatives market may have about US$1.5–2.0 trillion in undercollateralisation for derivative assets and a similar amount in derivative liabilities (BIS 2013; Singh 2010). Such residual liabilities and assets exist because clients of large banks such as sovereigns (and related entities), AAA insurers and pension funds, large corporates, multilateral institutions (eg, EBRD), Fannie Freddie, and the “Berkshire Hathaway” types of corporate do not post their full share of collateral. They are viewed by large banks as privileged and (presumably) safe clients.
From a risk-management angle, large banks’ credit-valuation-adjustment (CVA) teams need to hedge their “in the money” positions, or derivative assets, when there is a likelihood that these positions may not be paid in full. CVA teams largely use credit default swaps (CDSs) to hedge their OTC derivatives books. Under the old regulations there was no CVA capital charge. However, Basel III’s recommendations will require a capital charge for undercollateralised exposures. Basel III will recognise capital relief if such exposures are hedged only via CDSs (in other words, other forms of hedge will not be recognised by the accord). Regulations such as Capital Requirement Directive IV in Europe exempt capital charge on undercollateralised exposure to sovereigns etc (ie, those exempted from clearing). However, hedging derivative assets due from a sovereign pushes up the CDS spreads on the sovereign, as seen in peripheral Europe in previous years. This in turn may impact on the sovereign’s debt issuance costs (since CDS spreads impact on the spreads of the underlying bonds).
New Regulations and the Sovereign–Bank Nexus
Due to Basel III, the demand for hedging such growing derivative assets leads to a rise in the CDS spreads of the underlying so-called SSAs, or sovereigns, supranational entities (ECB, EBRD, World Bank and so on), and other agencies such as Fannie Mae and Freddie Mac that is “out of the money”. Market sources indicate that between 10% and 15% of a large dealer’s assets that need to be hedged may stem from SSAs; an equal fraction is from corporate clients.1 Thus, addressing the undercollateralisation issues is important to the understanding of CDS spreads when sovereigns and quasi-sovereigns are in distress – since that is when the CVA teams will hedge positions. There is no rule as to when the CVA teams will hedge; some may act sooner than others. But there is no threshold that CVA teams have to hedge if clients are rated BBB or below. At present, published sources (eg, Risk 2011) indicate that SSAs may be out of the money by US$150 billion to the banks (who are “in the money”). The banks in turn have similar funding costs as they hedge these positions with another bank and thus have to post collateral on being “out of money” on their hedges.
Typically, the “street”, or the large 10–15 dealers active in the OTC derivative market, are often on the same side of a trade (a few of them may hold similar positions when dealing with a sovereign). For example, informal discussions with the official sector indicate that the street is “in-the-money” on interest-rate-swap (IRS) positions that have been written since the early 2000s. Most dealers took the floating-rate leg of an IRS (which is an OTC derivative), and the sovereign typically took the fixed leg. As global interest rates have remained low since the 2008 crisis (and are likely to remain low), many of the IRS positions are now substantial derivative assets on the books of the banks, since sovereigns typically do not post collateral to the large banks and settle at maturity (and IRSs can typically be 30 years in tenor). Figure 9.1 shows a simplified example of a bank’s derivative position where the bank is “in the money” on an IRS (on, say, Italy) but “out of the money” on a CDS on Italy. As SSAs do not post collateral, the bank does not get any cashflows from being “in the money” by US$10 million, but has to fund the hedged position via the CDS that is “out of the money” by US$8 million.
Figure 9.1Illustrative Example of Funding Costs for Banks “in-the-money” with SSAs
The Experiences of some Sovereigns
In recent years, market sources indicate that the treasuries of some peripheral sovereigns have occasionally been active in the OTC derivatives market to reduce their CDS spreads (Bilal and Singh 2012). In one case, the treasury requested that large banks active in OTC derivatives market should not hedge their exposure to the sovereign since it would lead to an increase in the CDS spreads of the sovereign (Figure 9.2). By the assignment of OTC derivatives contracts from large global banks to the sovereign’s local banks, the original derivatives contract is reassigned from the large global bank to the periphery’s local bank (or another domestic entity) at the request of the treasury, but the out-of-the-money positions of the treasury are not due till maturity of the contract.2 On the other hand, novation would entail a “tear-up” of the original contract, and the treasury would have to “settle” the out-of the-money positions and pay the accrued balance up to the date of the novation at the time of the novation.
Figure 9.2How Assignments Lowered CDS Spreads
CDS volatility in peripheral Europe was on the rise after Greece came to the forefront in 2009. This has had a material impact on the books of large banks active in the OTC derivatives market. For example, Morgan Stanley’s 2011 financial statements states (square-bracketed material is from the original):
On December 22, 2011, the Company executed certain derivative restructuring amendments which settled on January 3, 2012. Upon settlement of the amendments, the exposure before hedges and net exposure for Italy decreased to $2.887bn [from $6.268bn] and $1.522bn [from $4.901bn], respectively, and the exposure before hedges and net exposure for Peripherals decreased to $5.044bn [from $8.425bn] and $3.056bn [from $6.435bn], respectively.
Note that, before the trade, Morgan Stanley’s net counterparty exposure to the Italian sovereign (ie, repurchase transactions, securities lending and OTC derivatives), taking into consideration legally enforceable master netting agreements and collateral, was as much as US$4.2 billion, in addition to US$689 million in exposure to Italian non-sovereigns.3 This example may be the tip of the iceberg, as all SSAs are in a similar position vis-à-vis the banks. However, forth-coming regulations such as Basel III and Dodd–Frank will not address this directly and the likelihood of large global banks having the sovereign umbrella over them will continue.
Most derivatives with large banks active in the OTC derivatives market are via IRSs. When rates fall, the swaps go “in-the-money” for the banks, and they have to hedge by buying CDSs. Large money-centre banks – which have positive exposure to sovereigns and corporates who do not post collateral – are likely to hedge with CDSs. The causality is from IRSs to CDSs. Also, when SSAs are in trouble, there is general risk aversion.
Take another example in peripheral Europe. Risk aversion led to a lower euro swap curve, as bund yields fall (dragging swaps lower) and it may indicate growth issues and/or that perhaps the ECB would keep rates lower for longer. Causality is from the CDS curve to the euro swap curve. Figure 9.3 shows the correlation of five-year US dollar CDSs of a peripheral sovereign and the euro-generic interest-rate swap curve (inverted, 2010–2012). Ignoring such correlation in regressions will lead to misspecified analysis of CDS spreads in the European periphery.
Figure 9.3Five-year CDS and Euro Interest-Rate Swap Curve
A Proposal to Reduce the Sovereign–Bank Nexus
Present market practices result in residual derivative liabilities and residual derivative assets, because (as we saw earlier), sovereigns, AAA insurers, large corporate, multilateral institutions, Fannie, Freddie and the “Berkshire Hathaway” firms are seen as safe and do not post their full share of collateral. We thus suggest that a levy on residual derivative liabilities would be a more transparent approach than moving OTC derivatives to CCPs, especially if the costs to bail out CCPs are to be funded by taxpayers (Singh 2011). If a levy is punitive enough, then large banks will strive to make derivative liabilities reach zero; as a result, there will be minimal systemic risk via the OTC derivatives markets if a large bank fails.
Furthermore, as a by-product of the above levy, we will also address the residual derivative assets (which have also averaged US$100 billion per large bank in the past). This will happen because the large banks typically have matched books (ie, on average, the size of the derivative liability and asset positions at each bank is roughly the same). Since at the time of inception of the OTC derivative contract, it is not known whether the contract will be in-the-money (asset) or out-of-the money (liability), the levy on liabilities will force receiving and paying collateral with every client (ie, no free ride for anyone). Thus, derivative assets will also be minimised (and not get siloed under bankruptcy when a large bank fails). The undercollateralisation issues stemming from all the users symmetrically and related synergies were not being addressed in the CCP discussions.
Some Policy Issues
Understanding the overall OTC derivative portfolio at banks is paramount to explaining the CDS spreads of SSAs (when distressed). As long as all clients of large banks do not post their fair share of collateral (ie, initial margin and variation margin), banks will continue to hedge their exposure on OTC derivatives where they are “in the money”. In fact, the regulatory proposals that will move OTC derivatives to centralised counterparties (CCPs) exempt most SSAs from posting collateral when they are “out of the money”. Recently, some sovereigns have acknowledged that having two-way CSAs and posting and receiving collateral when due may be economically more advantageous (eg, Bank of England). On the other hand, some may be forced to post collateral, as banks may be reluctant to fund positions that may remain profitable but with no collateral coming in for many years (and thus a funding cost if such positions are hedged).
In order to explain CDS spreads in the sovereign context, regressions will remain misspecified unless they include the collateral (or undercollateral) elements that drive the CDS spreads.4 Policymakers need to be wary of interpreting academic studies that may have omitted variables in the regressions that try to explain CDS spreads. Such variables are fragmented operationally across the CVA desks of all the key banks and, even if they are available, are unlikely to be available for regulatory oversight.
Panel 9.1:Lisbon Move Points to End of Risk-Free Sovereigns
Financial Times column by Gillian Tett, January 20, 2011
Another week, another bout of angst about sovereign and municipal risk. But as investors fret about Spain and Belgium – or Illinois and California – they should take a close look at a fascinating little development in Lisbon. On Wednesday [January 19], the Portuguese debt management agency formally announced in an e-mail that it would start posting collateral (such as cash or government bonds) on derivatives trades that it cuts with banks. It is intended to have a “positive effect” on its financing costs, and reduce “credit exposures”, it said. To onlookers, this might all sound dull and technical. And this e-mail has garnered little attention (partly because the Portuguese government first floated this idea last year). But in reality it carries considerable symbolic and practical significance. This week’s is just one sign of a much bigger paradigm shift about how investors and risk managers are now re-evaluating their assumptions about “safe” public sector debt. And this shift could create some fascinating practical challenges in the coming months, not just in the eurozone, but in America too.
The issue at stake revolves round how governments and banks construct derivatives deals in the over-the-counter market. During the past three decades, when banks have cut OTC interest and foreign exchange swaps deals with each other (or other private entities), they have often posted collateral to back those deals. This gives market participants protection against the failure of a counterparty.
However, until now, most governments have generally not provided such collateral, since they were considered “privileged”. This was partly due to logistical challenges (it is tough for bureaucrats to raid budgets to find collateral). However, Western public sector entities were deemed to be (almost) risk free. Thus, while banks were expected to provide collateral, public entities (and some AAA insurance groups and banks) were not.
But the financial crisis has forced investors and risk managers to rethink their assumptions about what is “risk free”, let alone “privileged”. And, unsurprisingly, many banks are now worrying about the swaps deals they previously struck with public entities. That is not because banks are necessarily losing money right now; on the contrary, eurozone swaps deals are typically moving in the banks’ favour due to swings in currency and interest rates. But since swaps are long-term, risk managers are nervous about the future. And though banks have tried to hedge this risk in the credit default swap market, this market is thin – and all this hedging has pushed CDS spreads wider (thus fuelling alarm further). The net result is that banks are furtively pushing for change. So are some regulators who are worried about wild swings in the CDS market. Lisbon’s announcement clearly shows some public entities are listening. After all, the Portuguese government seems to hope that posting collateral for swaps deals will now reduce the need for banks to hedge, thus potentially reducing CDS spreads and cutting sovereign funding costs. Or so the argument goes. Whether it works remains to be seen.
But investors would do well to watch what happens next. For one thing, this saga highlights something banks have long preferred to conceal: namely the wider level of under-collateralisation in the OTC derivatives market. Last year, Manmohan Singh, an economist at the International Monetary Fund, calculated, for example, that if market participants posted sufficient collateral to cover all OTC deals properly, they would need an extra $2,000bn (or about$100bn per big dealer). The TABB consultancy has reached similar conclusions. And while banks dispute this data, these numbers are sobering; particularly since OTC business is now moving on to clearing houses – where collateral will be mandatory.
But the second fascinating question is how many other public entities will follow Portugal’s lead? Or try to use clearing houses to lower the banks’ need to hedge. Some are certainly preparing to move in that direction; however, for many public entities there are huge challenges. Many do not have cash to spare; but it is far from clear that they will be allowed to use their own bonds as collateral instead. And in the US, there are also legal constraints to what local government can do.
Nevertheless, the one thing that is clear is that this debate – and trend – is long overdue. After all, one factor behind the recent bond and derivatives bubble was that the financial system has often failed to price properly all the associated credit, processing and execution risks attached to deals, particularly when entities were labelled AAA, or risk-free. If the financial system is now rectifying that for swaps, then that is a good thing; the only pity is that it has come 10 years too late.
BilalM. and ManmohanSingh2012 “CDS Spreads in European Periphery – some technical issues to consider” IMF Working Paper 12/77.
BIS2013 (and earlier issues) OTC Derivatives Semiannual Reports.
HannounHerve2011 “Sovereign risk in bank regulation and supervision: Where do we stand?” BIS speech in Abu DhabiOctober26.
Risk2011 “One-way CSA pile-up funding risk for banks” February available at http://www.risk.net/risk-magazine/feature/1949147/-csas-pile-funding-risk-banks
SinghManmohan2010 “Collateral, Netting and Systemic Risk in OTC Derivatives MarketIMF Working Paper 10/99.
The sovereigns’ related entities include quasi-sovereign, debt-management office and municipalities.
It is difficult to turn down a request from one of your biggest derivative clients.
Incorrectly leaving out one or more important independent variables leads to omitted-variable bias.