The Shadow Banking System is an Unstable Funding System for Banks, Not Assets
There are many definitions of shadow banking. A New York Federal Reserve Bank monograph effectively equates shadow banking to securitization, or the process by which individual loans are packaged into bundles, used to issue a wide variety of collateralized assets, and sold to investors. The New York Fed monograph is often used to demonstrate how complicated and virtually incomprehensible the shadow banking system is – it includes a “map” of the shadow banking system that, for legibility, the authors recommend printing as a 36” by 48” poster.
More commonly, however, the term shadow banking refers to the use of money market instruments to provide short-term finance to long-term assets, and thus focuses attention on bank runs and on the fact that shadow banks can face such runs, just as traditional banks do. For this reason securitization should not be equated with shadow banking, because a significant portion of private sector securitized assets were financed on a long-term rather than on a short-term basis. This post will limit its focus – as does most of the literature on shadow banking – to the role played by money markets in longer-term finance.
This post finds that our current money markets play only a very small role in the direct finance of private sector long-term assets and for the most part are used as a financing system for investment banks. In short, the “market-based” credit system that some equate with the shadow banking system, is very small – and relies heavily on commercial bank guarantees. To the degree that a substantial shadow banking system continues to exist, it does not fund long-term assets directly, but instead provides wholesale funding for investment banks, and to a lesser degree commercial banks.
To be clear, the focus here is on finance of private sector banks and assets. Thus, although Fannie Mae and Freddie Mac played a very important historical role in the development of the shadow banking system, by pioneering the practice of financing long-term mortgage debt on money markets through the issue and roll over of short-term debt that was at least nominally a private-sector obligation, they now officially have government support, and, for the purposes of this paper their debt is treated not as part of the shadow banking system, but as a government obligation.
This post provides a simple framework for understanding the shadow banking system that is organized around the two instruments, commercial paper and repurchase agreements, that play an important role in money markets and that are, very roughly, comparable to deposits. Studying how these instruments are used not only allows a distinction to be drawn between the direct finance of assets and the finance of assets that sit on bank balance sheets, but also makes clear why the shadow banking system is unstable.
This analysis finds that the money market instruments have in the past played three roles: they have funded banks and non-financial firms directly, they have funded assets that lie off bank balance sheets, and in order to play these roles, they have created a need for commercial bank guarantees that induce lenders to lend off-balance-sheet or in the case of tri-party repo to investment banks. In practice, the direct funding of assets now takes place only on a very small scale.
Because the two money market instruments, commercial paper and repurchase agreements (repos), are both short-term, it is easy for those who invest in them to “run,” or to decide that they no longer wish to invest their funds with a specific issuer or, indeed, in privately issued money market assets at all. Because these investors can always choose to put their money in Treasury bills or bank deposits, runs in the money markets are associated with unmanageably sudden shifts in investor preferences across short-term assets. In short, a fundamental attribute of the shadow banking system is that the decisions of money market investors can destabilize the money markets.
Money market mutual funds and enhanced cash funds (that promise liquidity, but are less regulated than money market funds) are the most obvious money market investors, but the buy-side of the money market is composed of a huge array of institutional investment funds, corporations, and government bodies that have funds they wish to keep in liquid form. All of these entities can be part of a run in the shadow banking system. In addition, as will be explained in detail below, in the repo market it is possible for the recipients of funds, such as prime brokerage clients and banks in the interdealer market, to run.
Now that the basic instability of the money markets has been established, the next step in understanding the shadow banking system is to understand the different ways in which commercial paper and repo-based instruments are used; this is discussed in sub-part A. The following sub-parts evaluate what shadow banking does, and discuss why it is more unstable than traditional banking.
A. Shadow Banking Instruments
1. Commercial Paper
Commercial paper is traditionally an unsecured obligation to make a payment that has a maturity of one year or less. It is analogous to the commercial bills that were used to finance economic activity in 19th c. Britain, and indeed has existed in one form or another for centuries.
i. Issued by financial institutions
A little over half the commercial paper issued in the United States, or approximately $550 billion, is issued directly by financial institutions. Because this market-based funding source is much less stable as a funding source than retail deposits, it is categorized along with other bank funding sources that are prone to runs as wholesale funding. The case of Lehman Bros. illustrates the instability of this form of funding. When Lehman declared bankruptcy, its commercial paper went into default, and set off a run by investors who feared money market mutual fund losses on money funds that invested in commercial paper; as a result the commercial paper market itself faced a run.
ii. Issued by non-financial corporations
Approximately one quarter of commercial paper is unsecured and issued by non-financial corporations. Because non-financial corporations have less access to liquidity than banks, there is a risk that when their commercial paper is due they will be unable to roll it over into a new issue and will be unable to honor their commercial paper obligations due to this liquidity risk. For this reason, almost all non-financial commercial paper is protected by a liquidity facility provided by a bank, which promises to retire the commercial paper if the issuer is unable to do so. Observe that when Lehman failed, the run on commercial paper was not carefully targeted to financial commercial paper, and as a result non-financial commercial paper was subject to a run as well.
b. Collateralized: Asset Backed Commercial Paper
In recent decades, sponsoring banks have moved assets that they originated into financing vehicles that are “bankruptcy-remote,” or not available to the sponsor’s creditors in the event that the sponsor declares bankruptcy. In addition, in theory any support that would be provided by the sponsor to the vehicle was defined in a contract, so the sponsor had contractually limited exposure to the vehicle’s liabilities. Thus, these vehicles were designed as a means of removing assets from the sponsoring bank’s balance sheet.
The ABCP market was one of the key markets that collapsed in the early days of the financial crisis – from $1.2 trillion outstanding in early August 2007 to $905 billion three months later. Since then the market has continued to decline slowly, and it now hovers around $250 billion.
Because these vehicles finance long-term assets they face the same liquidity risk as non-financial issuers when issuing commercial paper. In addition these vehicles face credit risk in the event that the value of the assets falls below the value of the commercial paper, and the vehicle is no longer fully collateralized. Both liquidity and credit risk must be addressed before the vehicle can receive a credit rating that is high enough for it to issue asset-backed commercial paper (ABCP) that is secured by the assets in the vehicle. The three principal means by which liquidity and credit risk were resolved are discussed below.
i. Bank supported ABCP: Conduits
Prior to the financial crisis most ABCP was issued by ABCP conduits that were sponsored by banks. The banks typically provided both a liquidity facility, which guaranteed that the commercial paper would be retired even if it could not be rolled over, and a credit facility, which promised to honor some fraction of the commercial paper in the event that the value of the collateral fell too low to cover the costs of repaying the commercial paper.
In August 2007 when the crisis started there was a sudden loss in confidence in the ABCP market and many conduits could not roll over their commercial paper. The banks had to step in and honor the liquidity guarantees that had been made – and in order to do so they had to seek regulatory exemptions that are documented by the Federal Reserve.
ii. Liability structure supported ABCP: SIVs, LPFCs, etc.
Some ABCP-issuing vehicles guaranteed the payment of ABCP by funding the assets with a combination of bonds, medium-term notes and ABCP. These vehicles took many forms; the most common were called structured investment vehicles (SIVs).
The concept behind these vehicles was that, in the event that the commercial paper could not be rolled over or the value of the assets fell below a trigger point, assets would have to be sold to pay off the ABCP and any losses would fall to the longer term debt holders. In 2007 most SIVs hit their triggers and were unwound. Because of the losses that were incurred by both longer-term and commercial paper investors (after lawsuits determined the allocation of proceeds), they are no longer a popular investment product.
iii. Repo Conduits – discussed below
2. Repurchase Agreements
A repurchase agreement (repo) is a simultaneous agreement to sell an asset today and to repurchase it a specific date and time in the future. It has the same economic effect as a collateralized loan. Typically the amount lent is less than the value of the collateral; the percentage difference is called a haircut.
There are two repo markets: the bilateral repo market and the tri-party repo market. In the bilateral repo market the lender must have the capacity to receive and manage the collateral, whereas in the tri-party repo market the tri-party clearing banks, JP Morgan Chase and Bank of New York Mellon, provide collateral management services for the lenders. Money market investors like mutual funds lend only on the tri-party repo market where the principal borrowers are the dealer banks (although a few hedge funds and private institutions are credit-worthy enough to be accepted as counterparties on this market).
The clearing banks also provide bank guarantees of liquidity to the tri-party repo market. Because it is the broker-dealers that borrow heavily on this market and because every trade in the market is unwound at the start of each trading day giving the borrowers access to their assets during the day, the two tri-party clearing banks extend credit to the borrowers during the day until the trades are rewound in the late afternoon. Thus the tri-party clearing banks provide a guarantee to the market and bear the risk of a broker-dealer failure during the day. While reform of the tri-party repo market has been high on the Federal Reserve’s agenda, five years after the financial crisis 70% of the market is still being financed by the clearing banks on an intraday basis.
On the bilateral market, where the lender must manage the collateral, the dealer banks are the lenders. The borrowers are prime brokerage clients, such as hedge funds, and other dealers.
As a result of this structure, funding generally enters the repo market via tri-party repo and the dealer banks, then, distribute this funding more broadly to their prime brokerage clients on the bilateral repo market. Thus, when a hedge fund buys an asset on margin, it borrows a significant fraction of the purchase price from the dealer bank that is its broker and posts the asset as collateral for the loan in a repo transaction. The dealer bank can then repo the asset on the tri-party repo market so that the dealer bank is effectively intermediating lending from the tri-party market to its client and earning an interest rate spread for the intermediation services. When the asset is of a type that cannot be used as collateral in the tri-party repo market, the dealer may choose to use the asset to raise funds on the inter-dealer segment of the bilateral repo market.
The dealer banks also hold collateral that is posted against derivatives contracts by other dealers and by prime brokerage clients. Whereas the inter-dealer derivatives contracts may have symmetrical collateral posting requirements, prime brokerage clients have typically been required to post collateral without having the right to require that dealer bank follow the same rule when the balance on the derivatives contracts is in the brokerage client’s favor. As a result a dealer bank is almost certain to receive collateral from its prime brokerage services when its client accounts are aggregated. The collateral posted by prime brokerage clients can then be used by the dealer to borrow in the tri-party repo market. As a result of this structure collateral posting by prime brokerage clients on their derivatives liabilities is also a form of financing for the dealer banks.
Thus, dealers often finance their own inventories, their prime brokerage clients’ assets, and any collateral that is posted against derivatives liabilities by other dealers or prime brokerage clients on the tri-party repo market.
The repo market is very different from the ABCP market and from commercial paper markets in general, because a run in one of the latter markets can only be caused by end investors. In the repo market a run can be started either by end investors or by other dealers and/or prime brokerage clients. Darrell Duffie has explained the many channels by which funding can be withdrawn in a repo market. These include: brokerage clients can move their accounts – together with all the collateral they have posted – to another dealer; dealers or brokerage clients who are derivatives counterparties can seek a novation (i.e. transfer) of a derivatives contract in order to post collateral to or expect payment from a more creditworthy dealer; dealers or brokerage clients may seek to reduce new exposures by entering into derivatives contracts that will require a dealer to post collateral; or repo lenders may increase haircuts or stop lending entirely to the dealer. In short, the repo market is subject to inter-dealer and brokerage client runs, as well as to runs by repo investors.
In 2008 it is very clear that both Bear Stearns and Lehman faced a withdrawal of funding from other dealers, from brokerage clients, and from end investors in the repo market.
3. Repo Conduits
A repo conduit is a bankruptcy remote financing vehicle. The vehicle issues commercial paper that is backed by a repo with a maturity that matches the commercial paper. Thus, a repo conduit is backed primarily by the credit of the repo counterparty. Only if the repo counterparty fails to pay, can the repo conduit foreclose on the repo collateral. Because the term of the repo matches the term of the commercial paper, rating agencies do not require that a repo conduit have a backup liquidity facility.
The credit rating of a repo conduit typically is based entirely on the credit of the repo counterparty. For this reason, repo conduits can be used – by institutions with high credit ratings – to finance assets that would not be eligible for tri-party repo financing.
B. What Does Shadow Banking Do?
1. Shadow Banking is a Funding Mechanism for Banks
The most important role of the shadow banking system is to provide wholesale funding for banks. Unsecured wholesale funding is provided when a bank issues commercial paper. Secured wholesale funding is provided when a investment bank uses the tri-party repo market to finance inventories, the assets of brokerage clients, and any collateral posted by counterparties in derivatives transactions.
As of Dec. 31, 2013, financial institutions raised $550 billion unsecured on financial commercial paper markets and the dealer banks used the tri-party repo market to borrow on a secured basis close to $1.6 trillion. 80% of the collateral posted is Treasuries and Agencies. Only $330 billion of private sector assets are financed on this market.
2. Shadow Banking is a Funding Mechanism for Assets
Before the crisis, the shadow banking system played an important role in funding assets with liabilities that were secured by assets that were held off of bank balance sheets in bankruptcy remote vehicles. When this secured asset funding relied on bank support, it was usually provided by ABCP conduits. When this secured asset funding was made possible by a tiered liability structure, it was provided by SIVs and similar vehicles. When this secured asset funding relied on a maturity-matched repo, it was provided by a repo conduit.
Before the crisis the ABCP market was the most important source of shadow bank funding of private sector assets. (Not only did the tri-party repo market fund private sector assets that were for the most part on dealer bank balance sheets, but it was dominated by Treasuries and Agencies and thus played a relatively small role in financing private sector assets even indirectly.) In post-crisis markets vehicles like ABCP and repo conduits are financing far fewer assets than they did before the crisis. The ABCP market is continuing its slow but steady decline over time and now hovers in volume around $250 billion.
3. Shadow Banking Allows Money Market Issuers to Rent Bank Credit and Allows Banks to Avoid Capital Requirements
When assets were directly financed by the shadow banking system, it was usually because financing vehicles paid a small fee to “rent” a commercial bank’s credit rating by purchasing a guarantee of the vehicle’s liabilities. Because these guarantees were off-balance sheet, the bank was able to avoid the capital requirements that would have been imposed if the bank had done the lending itself. The role played by the clearing banks in the tri-party repo market is similar: they provide intraday credit in order to give dealer banks access to their assets during the day, but face no capital charge for the credit. Thus, a key function played by shadow banking is the arbitrage of capital regulations.
The liquidity and credit facilities provided by banks to ABCP conduits are examples of unsecured bank guarantees. By contrast, the tri-party clearing banks provide secured guarantees. The intra-day credit that the clearing banks provide to the dealer banks is secured by the collateral that has been posted on the tri-party repo market. Banks may also issue guarantees in the form of swaps that offset the market risk of collateral; these guarantees may be secured or unsecured depending on the derivative contract.
The collapse of the ABCP market since regulators have become attuned to the problem of regulatory arbitrage of capital requirements is just another piece of evidence that the vast majority of financing on the ABCP market at its peak was not driven by economic efficiencies, but by regulatory arbitrage as banks used liquidity and credit facilities to take on credit risk, while avoiding capital requirements. Indeed, the industry reaction to the 2004 Final Regulation governing such liquidity facilities – which resulted in a “reinterpretation” of the regulation that effectively gutted it – is also evidence of the importance of regulatory arbitrage to this market.
C. Collateralized Money Markets Are More Unstable Than Traditional Banks
The use of collateral in repo markets makes them particularly unstable for two reasons: leverage and the fact that not just lenders, but borrowers, can start a run.
When the price of the collateral in a repo contract falls, the borrower is typically required to post more collateral within a day, and, in the event that the collateral call is not met, the collateral that was posted can be liquidated immediately. While this description shows how quickly market price changes can be reflected in the sale of collateral on repo markets, it does not take the leverage that is ubiquitous on repo markets into account. Because of leverage small changes in the market price of an assets can force the borrower to sell off a large fraction of the borrower’s holding of that asset.
An example (drawn from a Fitch Ratings report) will make the instability inherent in repo market finance more clear. Consider a borrower with a $5 million equity stake, which uses repo markets to finance the purchase of a $105 million portfolio of corporate bonds on which the lender imposes a 5% haircut, so that $1 can be borrowed for every $1.05 in collateral repo’d. The borrower will therefore have a leverage ratio of 21 to 1. A 2% decline in the value of the portfolio would reduce the total portfolio value to $102.9 million, reducing the equity in the portfolio to $2.9 million. If we assume that the borrower has no additional equity to contribute, the borrower can now only finance a $60.9 million portfolio at a 5% haircut. In short, because of the leverage inherent in using repo markets to finance assets, a 2% drop in portfolio value can force a sale of 42% of the assets held. Note that this example doesn’t take into account the possibility that the lender increases the haircut on the repo, which would mean that even more of the assets had to be sold. In short, once a borrower has maximized the use of leverage on repo markets – whether the borrower does this intentionally in order to “maximize” returns or simply ends up in this situation after the collateral has declined in price – very small declines in price can force the borrower to sell a significant fraction of the assets. If the borrower is a large market participant, such as an investment bank, this is likely to be the first step in a liquidity spiral, where asset sales further reduce the value of the collateral and trigger additional assets sales.
Not only does leverage make repo markets inherently unstable, but, in addition, a key characteristic distinguishing the repo market from unsecured credit markets generally is that not only the lenders, but also the borrowers, can start a run. The use of collateral in bilateral repo markets makes a borrower run possible, because the collateral can be rehypothecated, or posted as collateral in a subsequent loan by the recipient of the collateral. In short, the collateral posted by borrowers in the bilateral repo market is a source of liquidity for the lender.
When borrowers decide that they don’t want to be exposed to a troubled lender that may not be able to return the borrowers’ collateral in the event that it fails, the borrowers may seek to transfer their accounts to a lender who is not troubled. When the borrowers’ accounts are transferred, the collateral they have posted it transferred with the accounts, and the troubled lender loses the liquidity that was provided by that collateral.
As a result of this property of the repo market, the dealer bank failures of 2008 were characterized by “runs” by both prime brokerage clients and other dealers, none of whom wanted to be exposed to a failing bank. In fact, Krishnamurthy, Nagel, & Orlov conclude that the evidence supports the view that the 2008 crisis looks more like an inter-dealer credit crunch than a run by end investors on the two firms. For these authors one factor distinguishing the two types of runs is the fact that the dealers are well-informed market participants, whereas end investors typically must decide whether to pull out of the market based on very limited information. In short, it is possible that, far from being comparable to bank runs, the runs that took place in 2008 were runs that started with the most informed participants in financial markets.
Thus, there are two very important differences that make the repo market more unstable than unsecured funding markets. Not only does leverage mean that a small decline in price can easily force a large sale of asste, but in the bilateral repo market a run can be started not only by lenders, but also by borrowers.
In conclusion, it is misleading to describe the shadow banking system that exists today as “money market funding of capital market lending” and to focus on it as a means of financing assets, because at present by far the most important use of shadow banking instruments is to provide wholesale funding for dealer banks and through them indirect financing of assets that sit on their balance sheets. Although the view that shadow banking finances assets directly may have held some truth prior to the crisis when $1.2 trillion of ABCP financed bankruptcy remote vehicles, today, to the degree that shadow banking disintermediates commercial banks, it does so by reintermediating investment banks – using a form of funding that is even more unstable than deposits.
The key question that regulators have yet to answer is whether this collateralized wholesale funding market is a valuable addition to the financial system or whether the risk of instability that accompanies it is so great that lending on this wholesale market should be curtailed.
 Zoltan Pozsar, Tobias Adrian, Adam Ashcraft, & Hayley Boesky, Author’s Note in Shadow Banking, NYFRB Staff Rep. No. 458 (July 2010).
 Perry Mehrling, Zoltan Pozsar, James Sweeney, and Daniel Neilson, Bagehot was a Shadow Banker (Nov. 2013).
 For example, although only $35 billion of private label residential mortgage-backed securities have been issued since 2008, at the end of 2013 more than $1 trillion of such securities remained outstanding. Data from SIFMA: http://www.sifma.org/uploadedFiles/Research/Statistics/StatisticsFiles/SF-US-Mortgage-Related-SIFMA.xls?n=47986.
 Perry Mehrling, Zoltan Pozsar, James Sweeney, and Daniel Neilson, Bagehot was a Shadow Banker 2 (Nov. 2013).
 See Frank Fabozzi & Michael Fleming, U.S. Treasury and Agency Securities 11 (April 2004), available at http://www.newyorkfed.org/cfcbsweb/Treasuries_and_agencies.pdf.
 Federal Reserve Commercial Paper Release, Outstanding
 In practice, banks sometimes supported these vehicles even in the absence of a contractual obligation to do so, and sometimes did not.
 See the letters granting JPMorgan Chase & Co., Citigroup Inc., and Bank of America Corp. Regulation W exemptions that are dated August 20, 2007, available at the Federal Reserve website: http://www.federalreserve.gov/boarddocs/legalint/FederalReserveAct/2007/.
 Note that in securities lending, where institutional investors provide high-quality, high-demand collateral like Treasuries to the market, haircuts frequently go in the reverse direction. That is, more money must be lent than the value of the collateral in order to induce the securities lenders to lend.
 Tobias Adrian, Brian Begalle , Adam Copeland , Antoine Martin, Repo and Securities Lending, Federal Res. Bank of NY Staff Report No. 529, Feb. 2013 at 5-6.
 Adam Copeland, Darrell Duffie, Antoine Martin, and Susan McLaughlin, Key Mechanics of The U.S. Tri-Party Repo Market, 18 FRBNY Economic Policy Review 17, 22, 24 (2012).
 William C. Dudley, speech, Introductory Remarks at Workshop on “Fire Sales” as a Driver of Systemic Risk in Tri-Party Repo and Other Secured Funding Markets, Oct. 4, 2013.
 Darrell Duffie, How Big Banks Fail 23 – 42 (2011). See also William Dudley, More Lessons From the Crisis, Remarks at the Ctr. for Econ. Policy Studies Symposium, (Nov. 13, 2009), available at http://www.newyorkfed.org/newsevents/speeches/2009/dud091113.html; Adam Copeland, Antoine Martin & Michael Walker, The Tri-Party Repo Market before the 2010 Reforms 56-58 (Fed. Res. Bank of N.Y. Staff Rep. No. 477, 2010).
Duffie observes that when there is a repo market run, the coup de grace is almost always given by a clearing bank when it responds to concerns about a firm’s financial position by exercising its right to offset aggressively, by for example demanding collateral for intraday exposures or refusing to give access to deposits. Duffie, supra note 9, at 41¬-42. See also Tobias Adrian & Adam Ashcraft, Shadow Banking Regulation 17 (Fed. Res. Bank of N.Y. Staff Report No. 559, 2012).
 Duffie, at 23-42.
 Moody’s Revises Approach To Counterparty Rating Actions In Repo ABCP Conduits, Oct. 21, 2009, available at http://www.cranedata.com/archives/all-articles/2541/
 Arvind Krishnamurthy, Stefan Nagel & Dmitry Orlov, Sizing Up Repo 22 (NBER Working Paper No. w17768, 2012).
 Carolyn Sissoko, Note, Is financial regulation structurally biased to favor deregulation, 86 Southern California Law Review 365 (2013). Sissoko also has a discussion of the broader literature on the role of regulatory arbitrage in the ABCP market.
 See id. for details.
 See Sissoko, Deregulatory Bias at.
 Fitch Ratings, at 8.
 Arvind Krishnamurthy, Stefan Nagel & Dmitry Orlov, Sizing Up Repo 19,22 (NBER Working Paper No. w17768, 2012).