7. The Changing Collateral Space
- Manmohan Singh
- Published Date:
- June 2014
This chapter provides a snapshot of the changing collateral space and how it may shape the global demand–supply for collateral. We first identify the key collateral pools (relative to the “old” collateral space). However, post-Lehman and continuing into the European crisis, official sector efforts via quantitative easing (QE) are significantly altering the collateral space. Moreover, regulatory demands stemming from Basel III, Dodd–Frank, EMIR, etc, new debt issuance and collateral connectivity via custodians will also affect collateral movements.
The importance of collateral has been investigated in several strands that relate to each other in the theoretical literature. One strand is that on collateral and default, which has primarily focused on the role of margin and “haircuts” and “fire sales” (Geanakoplos 2003; Krishnamurthy, Nagel and Orlov 2010). Another strand is on securitisation, where collateral serves to support specific asset values (Shleifer and Vishny 2011). However, this chapter is not about haircuts, fire sales, or securitisation but about how collateral that can be reused comes to the market and about the new entrants in this market.
In this sense it is closer to the discussion on the supply and demand of safe assets. Empirical evidence that the safe-asset share has been relatively stable was postulated by Gorton, Lewellen and Metric (2012) using flow-of-funds data only. Recently, concerns have been raised about the supply of safe assets. The IMF’s “Global Financial Stability Report” (2012) estimated a US$74 trillion figure for safe assets, which would appear to be ample. However, a large fraction of such safe assets is held by buy-and-hold investors and is not available for reuse in financial markets. Some market sources conclude that there is little evidence to support the assertion that good collateral will be in short supply (JP Morgan 2012). Others argue that there could be such a shortage and that safe assets should be provided as a public good to avoid financial instability associated with the private supply of safe assets (Gourinchas and Jeanne 2012). This thinking is now being reflected in the recent Fed’s reverse repo programme (discussed in Chapter 4 and Panel 4.1).
The Old and the New Collateral Space
A great deal of short-term financing is generally extended by private agents against financial collateral. In the “old” global financial system, nonbanks were the primary actors that allowed reuse of their collateral in lieu of other considerations. Chapter 2 highlighted that the key providers of pledged collateral to the “street” (or large banks/dealers) are: (a) hedge funds; (b) custodians on behalf of pensions, insurers, official sector accounts, etc (Figure 7.1).1 In this banks–nonbanks nexus, “supply” of pledged collateral is typically received by the central collateral desk of the large banks/dealers that reuse the collateral to meet the “demand” from the financial system.
Figure 7.1.The old collateral space
The rectangle in the centre of Figure 7.1 depicts the volume in the old collateral space (in the orange area) and illustrates the reduction in collateral volumes as of end-20131, relative to end-2007. The financial crisis resulted in elevated counterparty risk leading to incomplete markets and idle – and thus stranded – collateral pools. Also, some central banks’ purchases of good collateral have contributed to shrinkage in the pledged collateral market from US$10 trillion prior to the Lehman crisis (end-2007) to about US$6 trillion (end-2013).
We earlier emphasised that the pledged collateral market (in the old collateral space) is different from some “restricted” collateral markets. For example, securitisation-based structures (SIVs etc) that have lien against specific pieces of collateral are impossible to repledge. Also, the triparty repo (TPR) market is a primary source of funding for banks in the US, standing at US$1.8 trillion at end-2013. It provides banks with cash on a secured basis, with the collateral being posted to lenders – such as money-market funds – through one of two clearing banks, BNY Mellon and JP Morgan. However, such pledged collateral sits with custodians, was left over and not used in the bilateral market, and is not as easily rehypothecable to the street (see Chapter 1). We ignored such restricted markets in the old collateral space, since collateral was not reusable, or with restricted velocity.
The “new” collateral space covers not only the bank–nonbank nexus (where collateral generates a velocity), but other participants who are now significantly impacting on collateral availability. The increasing role of central banks, regulators and collateral custodians is significantly changing the collateral landscape. These new dimensions involve (i) aspects of unconventional monetary policies pursued by some advanced-economy central banks that remove good collateral from markets to their balance sheet, where it is siloed; (ii) regulatory demands stemming from Basel III, Dodd–Frank, EMIR and so forth that will entail building collateral buffers at banks, CCPs etc; (iii) collateral custodians who are striving to connect with the central security depositories (CSDs) to release collateral from silos; and (iv) net debt issuance from AAA /AA-rated issuers.
When mapping the changing collateral space in Figure 7.2, we assume that the debt and GDP of developed countries will not increase significantly (otherwise, the topic of collateral shortage is moot). Also, we assume that regulation and collateral standards will not become so lax that junk will be deemed as “good collateral” with only a token haircut. We also acknowledge a new supply source: the recent reverse repo by the Fed that has started to provide collateral to banks and nonbanks (Chapter 4). We focus on collateral “flows”, since, whatever the stock of good collateral, only a fraction flows to markets to seek economic rent.
Figure 7.2.The changing collateral space
Factors Driving the Collateral Dynamics in the Near Term
Although there are many new “entrants” and developments in the collateral space, we discuss the four salient ones, which in our view will have a significant impact on the collateral dynamics.
Central banks (Figure 7.2, blue area on left)
Despite the European Central Bank’s (ECB’s) efforts to keep the ratio of good to bad collateral high in the EU financial markets, actions of the Swiss National Bank (SNB) and other central banks are at odds with this objective. Since the Swiss franc–euro peg in September 2011, the SNB balance sheet grew considerably to about US$500 billion. About half of the assets – consistent with past asset allocation pattern – now comprise short-tenor, “core” euro bonds and equities. This reflects prudent asset-liability management at the SNB.2 However The SNB’s bond purchases withdraw the best and most liquid collateral from the eurozone; this reduces the collateral reuse rate since these bonds are siloed at the SNB and not pledged in the financial markets. Siloed collateral has zero velocity by definition. The ECB has expanded collateral eligibility, including lowering the asset-based securities threshold and relaxing the foreign-exchange collateral requirement (ie, non-euro collateral is eligible).
The Federal Reserve in the United States continues QE3 (its third round of quantitative easing) although at a lowered rate (ie, tapering) as labour markets turn around. Since the Lehman crisis, and due to QE efforts, the Fed is housing over US$3 trillion of “good collateral” – largely US Treasuries and mortgage-backed securities (MBSs). Under Operation Twist (which ended in 2012), the Fed took in long-tenor debt of about US$45 billion per month and released short-term Treasuries. That programme kept the total size of the balance sheet unchanged. Then, QE3 expanded the Fed’s holdings by another US$45 billion per month of long-term US Treasuries (without a parallel sale of short-term debt). QE3 also bought US$40 billion MBSs per month. Thus, the Fed’s balance sheet was expanding at US$85 billion per month. At this rate, the Fed silo-ed over US$1 trillion additional good collateral in 2013 and beyond.
This is likely to have first-order implications for collateral velocity and global demand and supply of collateral. However, as discussed in Chapter 4, Fed’s recent reverse repo could be a game changer on the collateral front.
The Bank of England’s QE efforts have taken about £375 billion in gilts onto its balance sheet; however, looking forward, BoE is attempting to keep good collateral in the market domain with no more envisaged QE. Also, the Bank of Japan is expected to buy an additional ¥50 trillion (US$500 billion) of Japanese Government Bonds (JGBs) annually. However, in the past, JGBs have very low velocity, since they are not used in “upgrade” trades, and are generally held by domestic investors.
New regulations (Figure 7.2, green area on the top)
Regulatory demands stemming from Basel III and Dodd–Frank estimate that about USUS$2–4 trillion of additional collateral will be needed. Higher liquidity ratios at banks, along with collateral needs for CCPs (and non-cleared OTC derivatives), are some of the other key regulatory changes that will impact on collateral markets. These safety buffers will silo the associated collateral and significantly drain collateral in the financial markets (see Figure 7.2).
Custodians (Figure 7.2, pink area on the right side)
In 2011, the ECB said that the eurozone had €14 trillion in collateral, much of it locked in “depositories” and thus not easily accessible for cross-border use. However, Euroclear and Clearstream (the key hubs for eurozone collateral) are actively engaged with the local and national central security depositories (CSDs) to alleviate collateral constraints. The interconnections to the CSDs will be via the Target 2 Securities (T2S) system, which will provide a single pan-European platform for securities settlement in central-bank money – see Chapter 8. In the US, JP Morgan and the Bank of New York may also improve collateral flows from within the US triparty system; however, reforms on the triparty and money market funds will play a role in this effort.
Preliminary estimates suggest that perhaps up to €1–1.5 trillion of AAA /AA-quality collateral may be unlocked in the medium term via efforts of custodians to optimise collateral and build a collateral highway or global liquidity hub. This collateral in unlikely to reach markets but will enhance accounting debt and credits to “break” the silo. However, the internal “plumbing” (ie, operations, workflows, technology, staff and so forth) required to process and manage trillions of collateral balances needs to be smooth.
Every institution or market is different; there is a lot of friction in the pipes. Even though, legally, collateral is allowed to be reused, if a counterparty along the collateral chain hasn’t built the system to do anything with it the collateral gets “stuck” in the plumbing. The frictions in aggregate can be quite sizable and may be another reason why the theoretical balances may not add up mathematically.
Even if this collateral does not reach large banks and markets, it allows the collateral to leave CSD silos, improve efficiency, enhance accounting debt and credits, and reduce the burden on markets to provide collateral for liquidity-coverage-ratio-or CCP-related regulatory buffers. The triparty elements in Europe (ie, Euroclear Bank and Clearstream Banking SA) also hold client collateral but, unlike in the US, there is generally no intraday credit to clients.
In the US, JPMorgan and the Bank of New York may also improve collateral flows from within the US TPR system; however, regulatory reforms on the triparty and money market funds may limit the size of the collateral market. Money market mutual funds (MMMFs) are an important money artery to the US financial plumbing system and support about one-third of the TPR market. If US regulations move this industry towards variable NAV, then the money artery may shrink. Lately, US MMMFs have had increasing difficulty finding balance sheets willing to provide investments. That implies that custodial banks such as State Street and BNY will likely grow because of their position as “balance sheet of last resort” for the money market mutual fund industry (unless the Fed’s reverse repo leads MMMFs to shift en masse from the TPR to the Fed directly).
In general, central banks, sovereign wealth funds (SWFs) and long-term asset managers (life-insurance and pension funds) desire collateral that is of low volatility, but not necessarily highly liquid. These entities should be net providers of liquidity, in the form of either cash or liquid collateral. But, critically, their “need” for collateral is relatively static (or, as providers of liquidity, they can dictate that counterparties take a fixed amount). On the other side the hedge funds, money market funds (and, with the new regulations, the dealer banks, too) have a dramatically shifting need for collateral and a large number of counterparties. Their needs are for liquid collateral. So a market for collateral upgrades – in theory – could work.
New (net) debt issuance (Figure 7.2, grey area at the bottom)
Let us assume AAA/AA countries have GDP of around US$25 trillion and, with a deficit of around 4–5%, they have supplied (on average) about US$1 trillion of new (net) debt – sovereign and corporate – every year, with latest data on the lower side.3 Database and market contacts suggest that, on average, about 30–40% of AAA/AA collateral inventory reaches markets via custodians for reuse (on behalf of reserve managers, SWFs, pensions, insurers and so on); however, much of the inventory stays with buy-and-hold investors. So, if debt/GDP remains on trend in developed countries (ie, the ratio does not increase significantly), new debt stemming from the “numerator” may provide up to US$300–400 billion per year to the markets, assuming counterparty risk, especially with European banks, does not elevate. Another 5–10% of new inventory (including equities) may come via hedge funds. With a collateral reuse rate of about 2.5 in recent years (and, as of end-2013, even lower at 2.0 due to the various silos in the “new” collateral space), this may alleviate collateral shortage by about US$800 billion to US$1.2 trillion per year.
Policy Issues and The New Collateral Space
The dwindling number of AAA/AA entities and the potential correlations between borrowers and the collateral they’re pledging create quite sharp mismatches between what looks like plenty (eg, eurozone government bonds) and the extent to which anyone wants to actually take them as collateral from a bank in the same country. Regulations remain in flux. For example, sub-AAA/AA issuance may likely be considered satisfactory collateral. Also, if there is demand, collateral transformation may increase the required supply. On the other hand, debt ceiling issues in the US may entail reduced collateral supply in the form of US Treasuries or Bills than past trend.
The ECB still holds good collateral (for instance, bunds, Dutch and French bonds and other AAA /AA-rated securities). Although the fraction of good collateral has dropped since end-2011, the ECB’s €3 trillion balance sheet still holds about 20% in good collateral (or €600 billion). The ECB may want to “rent” the good collateral it holds, especially if its goal is to keep the good/bad collateral ratio high “in the markets”. So far the ECB has accepted “not so good collateral”, and thus improved the good–bad collateral ratio in the market by decreasing the denominator. Renting of good collateral does not lower the numerator – the collateral is on loan temporarily. Other EU central banks also hold good collateral. Other central banks (eg, the SNB, the Bank of England) do not have the same vested interest as the ECB to prop up collateral markets in the EU. Interestingly, the Fed has begun a reverse repo programme that will supply collateral to both banks and nonbanks.
In summary, the decrease in the “churning” of collateral may be significant since there is demand from some SIFIs and/or their clients (asset managers, hedge funds and so forth) for legally segregated/operationally commingled accounts (LSOCs), for the margin that they will post to CCPs (see Chapter 6). After the MF Global and Peregrine sagas there will be a decrease in the “reuse rate” of collateral, as there is increasing demand from several clients (asset managers, hedge funds and so on) for “legally segregated” accounts. An excellent example that is market-based is from the Reserve Bank of Australia (RBA). Its proposal manages to cope with the upcoming regulatory changes that will warrant significant additional high-quality liquid assets (or good collateral), without issuing more debt securities, unlike discussions in some policy circles (Gourinchas and Jeanne 2012). This committed liquidity facility (CLF) is akin to paying a fee to get the guarantee of good collateral from the RBA at a penalty rate. Its suggested route is akin to collateral transformation but this would keep the collateral reuse rate from declining. In other words,
Demandcollateral = Supplycollateral × collateral velocity
This annex highlights recent moves by hedge funds to segregate collateral from their prime brokers. We show that custody via a third party is akin to siloing the excess collateral, and thus reduces the churning and velocity. However, keeping the collateral with a prime broker’s holding company in a separate legal entity allows hedge funds to bargain a more attractive PB fee in lieu of the reuse of the collateral by the holding company (and not the PB unit).
The difference in the PB fees between secured funding (where the PB retains the HF collateral) and unsecured funding can be large. The PB fee could be zero or as low as Libor + 50bp if excess collateral is allowed to stay within the dealer’s umbrella organisation such as Goldman Sachs (but in a segregated entity under the Goldman Sachs Group Holding Company). This compares favourably with Libor + 250 basis points if the HF opts to keep the excess collateral with a third-party custodian such as the BoNY (see Figure 7.1.A for an illustration).4 A dealer can reuse/churn the collateral that stays in its segregated account for securities lending or repo activities that provide an attractive yield; if the collateral moves to a third-party custodian such as BoNY (where it is in a silo), it will not churn (ditto for the Fed’s reverse repo since September 2013, whereby nonbanks cannot rehypothecate but collateral remains in custody within the Fed). Many small and medium-sized HFs would rather show returns, net of fee, to be 2 percentage points higher (ie, Libor + 250bp versus Libor + 50bp) rather than take the extra step of placing excess collateral with a third-party custodian, which will cost them the higher PB fee.5
Figure 7.1A.Large banks’ use of hedge funds’ collateral
GeanakopolosJohn2003 “Liquidity, Default and Crashes” Cowles Foundation Paper No. 1074.
GortonGaryStefanLewellen and AndrewMetric2012 “The Safe Asset Share” American Economic Review: Papers and Proceedings102(3) May.
GourinchasPierre-Olivier and OlivierJeanne2012 “Global Safe Assets” BIS Working Paper No. 399.
International Monetary Fund2012 “Global Financial Stability Report” Annex 2.1 April.
ISDA2012 “Initial Margin For Non-Centrally Cleared Swaps” November.
JP Morgan2012 “Flows and Liquidity-No Shortage of Collateral” October5.
KrishnamurthyArvindStefanNagel and DmitryOrlov2010 “Sizing up Repo” working paperNorthwestern and Stanford Universities.
ShleiferAndrei and RobertVishny2011 “Fire Sales in Finance and Macroeconomics” Journal of Economic PerspectivesWinter.
Standard and Poor’s2012 “Collateral Optimisation is Playing a Transformative Role for Depositories, Custodians, and Clearinghouses” Ratings DirectDecember18.
Hedge funds (HFs) via their prime brokers (PBs) allow for collateral reuse as a quid pro quo for the leverage/funding they receive from dealers. The other nonbank providers of collateral generally lend collateral for various tenors to optimise their asset-management mandates. Commercial banks are not active in this bank (hence de minimis)
Alternately, the SNB could use foreign exchange or other derivatives to hedge the peg.
The Risk Management Association’s (RMA) database summarised inventory on loan to the market. See the Barclays AAA/AA index and www.rmahq.org, which that provides data on securities lending with title transfer.
Note that the “segregated” structure (ie, excess collateral with the GS Group) has not been legally tested.
To illustrate, assume two hedge funds, both showing annual returns on assets they manage for their clients of 8% (before the PB fee). One of the hedge funds, which allow full rehypothecation of its collateral, will pay a PB fee of, say, 0.75%, and give its clients annual returns (after PB fee) of 7.25%. The second HF keeps all excess collateral with a third party (ie, does not give any rehypothecation rights on excess collateral to its PB), and so will give its clients annual returns of 8% minus 2.75% PB fee, or only 5.25%.