20

Market-Based Credit Intermediation

Shadow Banks and Systemic Risk

20.1 Introduction
So far, much of our account of financial regulation has tracked the traditional distinc-
tion between securities regulation and bank regulation, in Parts B and D respectively.
In broad outline, these regulatory systems pursue different strategies: disclosure is at
the core of securities regulation, whereas prudential oversight is at the core of banking
regulation. (Indeed, we saw the segue from one of these strategies to the other in Part
C.) This regulatory difference maps onto what has been taken to be the central function
of each regime. The canonical financial claim of securities regulation is equity, which
gives the holder a proportionate share of the firm’s business success. The value of
equity depends upon the profitability of the issuing firm and is assessed over time
through trading in secondary markets, which provide the means for value realization.
Disclosure of the firm’s performance is thus central to the determination of value, both
upon the initial public offering of the claim and in its subsequent trading. By contrast,
the canonical claim of banking regulation is the deposit, issued by a bank and redeem-
able by that bank. The value of a deposit is designed to be invariant: a sum certain. The
regulatory task, then, is taken to be the protection of the deposit claim, which requires
prudential oversight of the risk-taking and balance sheet of the bank.1
One of the most salient developments over the past fifty years has been the growth of
‘market-based credit intermediation’. By this we mean the use of securities markets
to provide debt financing for firms and households, which would otherwise have
been supplied by banks. The phenomenon has been particularly strong in the US,2
but a similar trend has also been evident in the EU.3 Indeed, the most recent FSB report
on the largest economies indicates that non-bank financial intermediaries now hold
$137 trillion in assets, approximately 40 per cent of total financial system assets in the

1
See B Holmstrom, ‘Understanding the Role of Debt in the Financial System’, BIS Working Paper No
479 (2015); see also K Judge, ‘Information Gaps and Shadow Banking’, Working Paper Columbia Law
School (2015).
2
F Edwards and F Mishkin, ‘The Decline of Traditional Banking: Implications for Financial Stability
and Regulatory Policy’ (1995) Federal Reserve Board of New York Economic Policy Review, July, 27. In the
US, banks’ share of total financial assets was below that of pension funds, insurers, and other financial
intermediaries over the period 2004–14; the bank share percentage ranged in the low 20s over the period.
FSB, ‘Global Shadow Banking Monitoring Report 2015’ (2015), 59.
3
See J De Haan, S Oosterloo, and D Schoenmaker, European Financial Markets and Institutions
(Cambridge: CUP, 2009), 181. In the UK, for example, banks’ share was consistently above that of insurers,
pension funds, and other financial intermediaries throughout the period 2004–14; the bank share percent-
age ranged mostly in the 50s: FSB, n 2, 59.

Principles of Financial Regulation. First Edition. John Armour, Dan Awrey, Paul Davies, Luca Enriques, Jeffrey N.
Gordon, Colin Mayer, and Jennifer Payne. © John Armour, Dan Awrey, Paul Davies, Luca Enriques, Jeffrey
N. Gordon, Colin Mayer, and Jennifer Payne 2016. Published 2016 by Oxford University Press.
434 Principles of Financial Regulation

surveyed jurisdictions.4 In the US, where non-bank financial intermediation is most

advanced, the bank share of financial assets has been less than 25 per cent for at least
the last ten years.5 As we shall explain in sections 20.2–20.4, this phenomenon has been
driven in part by valuable financial innovation and in part by regulatory arbitrage.
Prior to the financial crisis, regulators generally believed that these developments
had reduced the overall level of systemic risk. This is because, as we explored in
Chapter 13, the banking model entails a certain level of institutional fragility and,
because of its tie to the payment system, represents a locus of systemic risk in the
economy. Thus, it was reasoned, the substitution of credit provision by markets rather
than banks would reduce systemic risk. We know now that this view was mistaken. The
shift in the financial system did not so much reduce systemic risk as change the ways in
which it arose. Thus a regulatory framework that focused its search for systemic risk on
‘banks’ was not apt to spot, and still less to control, emergent varieties of that risk. More
specifically, a regulatory framework that regarded disclosure as the principal strategy
for control of market-based risks did not quickly perceive the need for prudential
oversight of the forms of market-based credit intermediation that performed bank-like
liquidity and maturity transformation.
In terms of the framework for understanding systemic risk articulated in Chapter 19,
the change in the structure of the financial system brought with it a change in the cross-
sectional configuration of systemic risk. Regulators were stuck in a framework that did
not contemplate the migration of systemic risk, and there was consequently no serious
attempt to monitor this from a macroprudential perspective. As a result, the new
sources of systemic risk slipped cleanly under the radar. Understanding how the
changes in the financial system brought with them new sources of systemic risk goes
to the heart of current regulatory endeavours, and of this book’s project. Obviously, it
matters a great deal that we understand what happened in the run-up to the financial
crisis in order to reduce the chances of a recurrence. To preserve financial stability,
regulators need to understand the new cross-section of systemic risk. Of course, this
cross-section surely will continue to evolve. Better still, then, that regulators should
understand what might be called the ‘primitives’ of systemic risk: the basic elements
that combine to create instability as the financial system reconfigures itself over time.
The same is true for anyone who wants to evaluate, or even understand, the new
strategies regulators have deployed in response. In Part E, we consider how the changes
in the structure of the financial system altered the incidence of systemic risk and how
regulators have responded.
This chapter is organized as follows. Section 20.2 explains historically (and analyt-
ically) the rise of market-based credit intermediation. This is in two parts: first, the flow
of retirement and personal savings into institutional investor intermediaries that did
not need the liquidity, maturity, or credit transformation services of banks; and second,
the (US) story of how Glass–Steagall helped foster a set of financial firms—investment
banks—with strong incentives to develop market-based alternatives to bank finance.

4
FSB, n 2, 2. This is a ‘broad measure’ of non-bank financial institutions that includes insurance
companies, pension funds, and ‘other financial intermediaries’ (such as bond funds and investment banks).
5
See n 2.
 Market-Based Credit Intermediation 435

Section 20.3 explores the systemic dangers in market-based finance that arose from
the effort to substitute market-based liquidity and maturity transformation for bank-
based transformation. We use one example for each transformation service: first, the
mortgage-backed securitization structures that offered liquidity, backstopped by
private lender of last resort (‘LOLR’) assurances from the sponsoring banks; and
second, the funding of long-term debt securities on dealer bank balance sheets by
short-term liabilities supported by a private form of deposit insurance, repo finance.
The systemic danger is that in financial distress, private LOLR assurances and private
deposit insurance may crumble, producing a massive run on the market-based system
that may spill over into the banking system itself.
Section 20.4 explores the systemic dangers in market-based finance that arise from
the effort to substitute market-based credit transformation for banks. The example here
is mortgage-backed securitization. This purported to provide a superior form of
diversification through rearrangement of cash flows to provide claims that were as
safe as deposits, that were as liquid, but that offered higher yield. When the alchemy
for producing superior credit transformation—tranching overseen by credit-rating
agencies—was shown to be flawed, this form of market-based finance crumbled.
Section 20.5 then explores how regulators around the world have attempted to
distinguish this systemically sensitive activity, ‘shadow-banking’—for which more
stringent regulation is required—from other, benign, forms of market-based credit
intermediation. There is a methodological divide here: is the right approach to try to
identify institutions that perform ‘bank-like’ functions, and—if they achieve sufficient
scale to become ‘dangerous’—then sweep them into the (prudential) regulatory per-
imeter? Or is it better to try to identify certain financial market practices that appear to
increase systemic risk, such as the use of short-term wholesale credit to finance risky
assets, and place economy-wide limits on the practice? A ‘functional’ approach requires
regulators to understand how the various institutions operate within markets in
assessing how to protect financial stability.
This chapter can be seen as an introduction to Part E. Chapter 21 provides a more
institutionally detailed account of the elements of market-based finance. Chapter 22
takes a detailed look at asset managers, who are critical actors in the market-based
credit intermediation system, because as famed (US) bank robber Willie Sutton once
said, ‘that’s where the money is’. The point is to assess the potential for systemic risk in
their diverse business models. Chapter 23 then asks whether seeking to break links
between various components of the large banking organizations that navigate both
banking and securities markets would enhance systemic stability. This poses a con-
temporary version of the Glass–Steagall question.

20.2 The Rise of Market-Based Credit Intermediation

The rise of market-based credit intermediation since the 1960s had two main drivers:
first, the flow of funds into non-bank financial intermediaries (pension funds, life
insurers, mutual funds), which did not need the transformation services that banks
provided; and second, the way that regulation in the US—most notably the Glass–Steagall
436 Principles of Financial Regulation

Act and geographic limits on bank expansion—gave investment banks the incentive
and opportunity to pursue market-based credit finance vigorously.6

20.2.1 The flow of savings into institutional investors not banks

Throughout the post Second World War period, increasing private retirement provision—
boosted by strong tax incentives in the UK and US—has enhanced the mobilization of
capital into the hands of institutional investors who manage these funds on behalf of
their ultimate beneficiaries.7 The most notable such institutions are pension funds and
life insurers.8 These institutions do not need the intermediation services that banks
provide, specifically the transformation services, and thus are candidates for market-
based alternatives.
Why do they not need banks? The core reason is that the beneficiaries/customers of
the pension funds and life insurers, respectively, are not entitled to receive/contract for
liquid claims. First, the beneficiaries of pension funds and life insurers have claims that
are (in general) payable only upon contractually specified future events: retirement,
and death; they are not immediately redeemable. Pension funds and life insurers can
thus plan for highly predictable future needs for cash. This means that pension funds
and insurers do not need the liquidity transformation services provided by bank
deposits. Similarly, because their cash needs are predictable and can be satisfied by
laddering the payout of maturing debt, pension funds and life insurers do not experi-
ence a potential maturity mismatch in holding long-term claims. Thus they do not
need the bank’s maturity transformation services. Indeed, their planning horizon
means that such institutional investors can hold the longest-term debt more readily
than a bank.
Second, pension funds and insurers invest on a scale that means that they can justify
the expense of obtaining their own high-quality investment advice. This means that
like banks, they can minimize risk through well-advised credit-selection and portfolio
diversification. Indeed, the scale of the invested funds far exceeds a bank’s deposit
insurance ceiling, so depositing the funds in a bank would provide no greater margin of
safety than the bank’s own diversified asset portfolio, which such institutional investors
can at least replicate, if not improve upon.9 Moreover, by buying only investment grade
bonds, a typical practice of pension funds and insurers, the institution can substantially

6
An enabling factor was technological advance, particularly the increasing availability of low-cost
computer power and high-speed communications, which greatly facilitated the functioning of securities
markets. These advances lowered the transaction costs of trading, meaning that institutional investors could
rearrange their portfolios at lower cost, and enabled the design of portfolios, indeed the debt securities
themselves, that permitted institutional investors to create diversified debt portfolios and otherwise manage
risk at low cost.
7
See the discussion in Chapter 12, section 12.2.
8
See GL Clark, Pension Fund Capitalism (Oxford: OUP, 2000); EP Davis and B Steil, Institutional
Investors (Cambridge, MA: MIT Press, 2001); RJ Gilson and JN Gordon, ‘The Agency Costs of Agency
Capitalism: Activist Investors and the Revaluation of Governance Rights’ (2013) 113 Columbia Law
Review 863.
9
In the US, historically the most generous provider of deposit insurance, the ceiling was $100,000 per
deposit account. The current limit is $250,000. See generally Chapter 15, section 15.5.
 Market-Based Credit Intermediation 437

eliminate its need for high-level monitoring and loan management. Such institutions
therefore do not need the credit transformation services of a bank account.
A parallel development augmented the supply of funds available to market-based
finance and extended the range of corporate borrowers that could obtain credit in
markets rather than from commercial banks. This was the shift in pension provisioning
from ‘defined benefit’ plans, for which the company managed a fund towards the goal
of providing an employee with a fixed payout, towards ‘defined contribution’ plans, for
which companies annually paid a fixed amount into accounts managed by the indi-
vidual employee. These defined contribution accounts were typically administered by
an asset manager that provided employees a menu of investment choices, including
‘bond funds’. These bond funds would assemble debt securities of varying credit risk—
not necessarily limited to investment grade—and would be sold on the basis of yield,
reflecting risk-bearing and diversification.10 Shares in the bond funds were redeemable
daily on the basis of the proportionate share of the fund’s current net asset value
(‘NAV’). Because a fund’s obligation to the investor was to redeem at NAV, the fund
was not transforming maturity or credit risk. An investor in a bond fund had what
might be regarded as ‘imperfect liquidity’—the right to receive cash, as with a bank
deposit, but not a sum certain, because the value of the bond portfolio could fluctuate
with interest rate or credit factors. In short, this was not bank-like liquidity transform-
ation. The fund might include a certain level of cash in its portfolio to satisfy
redemption requests without the need to sell credit assets into the market at possibly
depressed prices, but this was principally as a matter of protection for non-redeeming
investors, not to avoid the insolvency of the fund.

20.2.2 The regulatory stimulants to market-based credit intermediation

The rise of life insurers, pension funds, and bond funds thus generated a vast pool of
funds looking for debt-like claims but not needing the transformation services that
banks supplied. This was the precondition for the growth of market-based credit
intermediation. The key institutional catalyst in the creation of this market was the
investment bank, which could replace commercial banks in the ‘origination’ of loans,
except that the form of this origination was the underwriting of debt securities of the
corporate issuer. Unlike a commercial bank, which held credit claims that it originated,
the investment bank would ‘distribute’ debt securities to willing institutional buyers,
which would bear credit risk, albeit in diversified portfolios. Investment-grade com-
panies quickly learned that they could obtain more favourable interest rates and other
credit terms through securities issuance to pension funds and life insurers than through
bank loans, whose interest charge would have to cover the costs of the bank’s trans-
formation services (including the cost of regulatory capital). The investment banks
could earn substantial profits from the underwriting fees, while distributing the risk-bearing
to the institutional buyers. Thus the key fact: market-based credit intermediation undercuts

10
If an issuer fell into financial distress, the bond fund could sell the bond to a hedge fund that
specialized in managing ‘distressed credit’ situations; in effect, the bond fund could outsource this element
of credit transformation.
438 Principles of Financial Regulation

a core lending business of large commercial banks. The rise of institutional investor
intermediaries—pension funds, life insurers, and bond funds—provided an alternative
to banks in the holding of debt claims. And a particular sort of market intermediary,
investment banks, provided an alternative to bank-based loan origination in the form
of underwriting.
In what way did regulation play a role in stimulating this development? First, the
separation of investment banking from commercial banking that followed Glass–Steagall
created a set of market-focused intermediaries: the investment banks. Equity issuances
are infrequent; but firms are continually raising debt—if only to roll-over maturing
indebtedness. Because investment banks were blocked from conventional commercial
banking, they had strong incentives to figure out how to propagate market-based credit
intermediation and then pursued the cost advantages relentlessly. As European finance
demonstrates, a universal bank with a strong commercial lending franchise will be
reluctant to cannibalize its existing business through market-finance innovations.
Second, US banks have historically been subject to geographic restrictions on the
scope of their operation. Until federal legislative change in 1994, interstate bank
operations were tightly circumscribed by a mix of federal and state laws. Indeed, the
most common bank had been the ‘unit’ bank, a single location in a single state. These
restrictions may have facilitated monitoring of local bank activity by in-state regulators
(or protected the local banking monopolies of local elites), but they also impeded the
extent to which commercial banks were able to offer finance at a scale appropriate for
large US public companies. This had the effect of steering corporate finance-raising
away from bank loans and towards bond markets. To raise substantial sums through
bank loans required the formation of a syndicate of dozens, sometimes hundreds, of
banks, typically organized by a lead ‘money centre’ bank. The transaction costs were
quite high, especially if a debt-restructuring proved necessary. By contrast, a small
group of investment banks could readily tap into a nationwide bond market of large
institutional investors to raise funds through a securities issuance. Moreover, innov-
ations in the Bankruptcy Reform Act of 1978 that facilitated bondholder coordination
on a value-conserving reorganization of a failing firm reversed the advantage that banks
had previously enjoyed in restructuring the debt of a distressed corporate debtor.11
The rise of market-based credit intermediation provided the impetus for the steady
erosion of Glass–Steagall through a process of regulatory exceptions over the 1980s and
1990s and that led to its ultimate repeal in 1999.12 US banks were now permitted to
affiliate with securities firms. An immediate consequence was that the balance sheets of
the US bank holding companies that these securities firms joined became much larger
and more complex—and thus more of a supervisory challenge. Nevertheless, in its
origins, market-based credit intermediation looks to have been motivated by efficiency.

11
On the role of reorganization law in facilitating debt restructuring, see S Djankov, O Hart, C McLiesh,
and A Shleifer, ‘Debt Enforcement Around the World’ (2008) 116 Journal of Political Economy 1105; J
Armour, G Hertig, and H Kanda, ‘Transactions with Creditors’, in R Kraakman, J Armour, P Davies, L
Enriques, H Hansmann, G Hertig, K Hopt, H Kanda, M Pargendler, G Ringe, and E Rock, The Anatomy of
Corporate Law, 3rd ed (Oxford: OUP, 2016), Ch 5.
12
Gramm–Leach–Bliley Act of 1999, Pub L 106-102, 113 Stat 138.
 Market-Based Credit Intermediation 439

The debt securities ‘buy side’ simply did not need the transformation services of banks.
These costs, which included the bank’s loan origination and monitoring infrastructure
as well as the costly balance sheet constraints necessary to maintain the bank’s capacity
to provide liquidity transformation, could be avoided without immediately creating
powerful new vectors of systemic risk.

20.3 Liquidity and Maturity Transformation

through Markets: Enter Systemic Risk
Pension funds and insurers were serving beneficiaries and customers who were not
seeking liquidity; bond funds served customers who were satisfied with imperfect
liquidity (liquidity without a sum certain). The challenge for market-based credit
finance was how to take advantage of funding sources that demanded ‘perfect’ liquidity
(liquidity with sum certain) and that were pursuing yield. Here are two examples that
show how market-based efforts to supply this product massively increased systemic
risk. The first is bank-sponsored securitizations in which banks entered into private
LOLR arrangements to support the liquidity of short-term payment obligations. The
second is the use of repo finance to create private deposit insurance for short-term
financing of long-term assets held by dealer banks. In both cases the goal was to avoid
runs that would force the destabilizing sale of debt securities into unreceptive markets.

20.3.1 Securitizations, asset-backed commercial paper and

bank LOLR provision
We discuss securitization in considerable detail in Chapter 21, section 21.3. For present
purposes these elements are useful: securitization of residential mortgages arose in part
to solve the problems revealed in the US savings-and-loan crisis of the 1970s and
1980s,13 in particular the maturity mismatch between long-term mortgage loans and
short-term deposits, and the risk of correlated defaults because of regional economic
shocks.14 Securitization was a technique to separate the origination of the mortgage
loan from the risk-bearing function. Whole mortgage loans could be purchased by a
‘special investment vehicle’ (‘SIV’) that would in turn issue (or distribute) long-term
‘mortgage-backed securities’ (‘MBS’), the proceeds of which would cover the purchase
of the loans. The SIV needed to issue short-term securities as well, ‘asset-backed
commercial paper’ (‘ABCP’), both to fund initial loan purchases and to smooth out
payments to MBS holders in light of the potential bumpiness of mortgage repayments.
The SIVs were structured so that the ABCP was over-collateralized, meaning pur-
chasers would take no credit risk, and so ABCP could be sold to money market funds

13
We discuss this in section 16.4.2.
14
In the US, the Government-sponsored enterprises (‘GSEs’) of ‘Fannie Mae’ (Federal National Mort-
gage Association) and ‘Freddy Mac’ (Federal Home Loan Mortgage Corporation) securitized most high-
quality mortgage loans below a ceiling amount, absorbing credit risk but not early repayment risk or
interest rate risk. Banks and investment banks did ‘private label’ securitizations of lower-quality loans
(including ‘subprime’ loans) and ‘jumbo’ (very large) loans, and fashioned the complex securitization
products that figured in the financial crisis.
440 Principles of Financial Regulation

and others as a substitute for a bank certificate of deposit. At each rollover of the ABCP
as it matured, the holder could stay invested or receive payment—in effect, redeeming.
What if there came to be uncertainty about the payment streams, such as the
mortgage repayment rates on which the cash flows depended? This could lead to a
run on the SIV as ABCP holders refused to roll-over their investments and potential
new buyers stayed away. As the run gained momentum, the SIV’s liquidity reserves
would be drained, forcing an asset liquidation and leading to losses for the non-
redeeming ABCP holders, the anticipation of which would accelerate the run. This is
a classic story of structural fragility associated with liquidity transformation.
The solution was private provision of LOLR services. The sponsoring bank would
issue a ‘liquidity put’ to the SIV, meaning that if the market froze, the bank would
provide ‘liquidity’ to the SIV, sufficient to pay off the redeeming APCB holders. Under
the Basel rules that then applied, this obligation was regarded as a remote contingent
obligation for the sponsoring bank that carried only a nominal charge. In the event, the
bank’s performing as a LOLR—lending against ‘good’ collateral—would mean taking
onto its balance sheet the credit risk of the SIV’s assets that had supposedly been taken
off its balance sheet by the securitization. In the financial crisis, tens of billions of
dollars of MBS boomeranged back onto bank balance sheets in this way, abruptly
changing capital ratios and in some cases creating destabilizing solvency questions
about the bank itself. Market-based credit intermediation that purported to provide
liquidity transformation outside the official banking system had failed. The key point is
this: the private LOLR mechanism that stabilized the structure would not scale—the
banks could not substitute for the Federal Reserve as securitization came into wide-
spread use. This form of market-based credit intermediation added to systemic risk.

20.3.2 Dealer banks and private deposit insurance

One element in the rise of market-based credit intermediation was the growth in the
balance sheets of ‘dealer banks’, which held long-term debt securities financed with
short-term liabilities.15 Since legally (even after repeal of Glass–Steagall) dealer banks
were not permitted to issue ‘deposits’, these short-term issuances were necessarily funded
through the short-term wholesale credit market. The structural fragility was immediately
apparent, because the dealer bank was providing maturity transformation (and liquidity
transformation) without either a LOLR or deposit insurance. The private-ordering
substitute that would sustain this form of market-based credit intermediation was a
form of secured lending, ‘repo’, which could act as a kind of private deposit insurance and
thus minimize the run risk of highly sophisticated counterparties who were constitu-
tionally unable to take credit risk (money market funds) or simply unwilling to do so.
We explain ‘repo’ and other forms of wholesale finance in Chapter 21, section 21.2.
For present purposes it is sufficient to describe repo as a collateralized loan structured
as a sale and repurchase transaction in which the ‘seller’ (borrower) sells a security to
the ‘purchaser’ (lender) for $x with the understanding that it will repurchase the

15
Dealer banks, which Chapter 21 discusses in detail, can be regarded as the market-making and
proprietary trading part of an investment bank.
 Market-Based Credit Intermediation 441

security (commonly the next day) at $x + y B.P., the y Basis Points reflecting the
interest charge.16 Commonly, the market value of the security exceeds the loan, and
the difference is called the ‘haircut’ (or ‘margin’) and reflects the extent to which the
transaction is over-collateralized. If the borrower defaults on its repurchase obligation
(repayment of the loan), the lender can retain the security or sell it and apply the
proceeds to the loan. Laws of most jurisdictions protect the lender (purchaser) from
insolvency laws that might otherwise stay a secured party’s foreclosing on the borrow-
er’s collateral, a bankruptcy safe harbour. By protecting the value of the lender’s
‘deposit’, repo provides a kind of private deposit insurance.
A key constraint on the size of a dealer bank’s balance sheet was the capacity to
provide such private deposit insurance because unsecured short-term lending is prone
to runs. We know that before the advent of deposit insurance, commercial banks
carried much higher levels of capital and much higher levels of cash and cash
equivalents.17 So in effect, the capacity of the dealer banks to provide transformation
services was increasing in the extent to which they could offer ‘private deposit insur-
ance’ in the way we have described. Two interrelated developments in the run-up to the
financial crisis increased the extent to which they could do so. The first was the use of
the securitization mechanism (discussed in the next section) to engage in a particular
form of credit transformation of risky cash flows from individual mortgages into AAA-
rated MBS. The second was the expansion in 2005 of the bankruptcy safe harbour to
include mortgage-backed securities as collateral eligible for this special treatment.18
The combination of such high credit ratings and access to the bankruptcy safe harbour
meant that these new securities could be financed using repos.
The mortgages underlying the AAA-rated MBS were long-term obligations. These
generated a stream of interest revenue that was higher than the short-term interest
rates payable for overnight financing in the repo market. This meant the dealer banks
had an incentive to expand the size of their holdings of these AAA-rated MBS to
exploit the interest rate spread between long-term and short-term rates.19 They did this
by tapping into the unprecedented worldwide flow of funds looking for safe, liquid
assets.20 The private deposit insurance permitted the dealer banks to provide maturity
and liquidity transformation services that were limited only by the ready supply of

16
A ‘basis point’ is one-hundredth of a percentage point.
17
AG Haldane, ‘Capital Discipline’, Speech to American Economic Association, Denver, 9 January 2011
(2011); CW Calomiris and M Carlson, ‘Corporate Governance and Risk Management at Unprotected
Banks: National Banks in the 1890s’ (forthcoming 2016) 119 Journal of Financial Economics.
18
The 2005 expansion is discussed in detail in ER Morrison and J Riegel, ‘Financial Contracts and the
New Bankruptcy Code: Insulating Markets from Bankrupt Debtors and Bankruptcy Judges’ (2005) 11
American Bankruptcy Institute Journal 641. See 11 USC §101(47). The consequences and a call for reform
are elaborated in ER Morrison, ME Roe, and CS Sontchi, ‘Rolling Back the Repo Safe Harbors’ (2014) 69
Business Lawyer 1015.
19
The dealer banks could also profit from trading bonds if they could accurately forecast changes in
long-term interest rates and from fees on trading on behalf of customers.
20
See Z Pozsar, ‘Institutional Cash Pools and the Triffin Dilemma of the US Banking System’, IMF
Working Paper 11/190 (2011); BS Bernanke, C Bertaut, LP DeMarco, and S Kamin, ‘International Capital
Flows and the Returns of Safe Assets in the United States, 2003–2007’, Federal Reserve Board International
Finance Discussion Paper No 1014 (2011).
442 Principles of Financial Regulation

AAA-rated MBS that could be financed through repo.21 The overnight nature of repo
borrowing meant the dealer banks could adjust interest rates daily to assure that
‘depositors’ did not withdraw funds to pursue higher yields. The availability of interest
rate swaps22 permitted the dealer banks to adjust their exposure to interest rate risk on
their long-term debt holdings without selling them. Thus they could stably transform
daily demandable liquidity into long-term holdings. A large and robust repo market
meant that dealer banks could handle the ‘withdrawal’ of large sums by ‘depositors’
who needed funds for some other purpose, because there were ample substitute short-
term credit providers attracted by the private deposit insurance of the repo system.
Because this form of ‘deposit insurance’ was not capped, it was especially attractive to
wholesale providers.
What happened next to unravel the private deposit insurance system is well
known.23 As the level of mortgage foreclosures began to increase in 2007 and 2008,
the value of the AAA-rated MBS became suspect; in particular, the flaws in the credit
intermediation technology (described in the next section) became manifest. Short-term
credit providers, who remained reliant on private deposit insurance, demanded
increasingly large haircuts on the securities used for repo, or, as valuations became
extremely uncertain, withdrew altogether. This can be described as a ‘run’. The point is
that as the private deposit insurance capacity of the dealer banks dramatically shrank,
so did their ability to draw in and hold the short-term funds necessary to finance their
assets. This led to rounds of fire sales that in some cases rendered dealer banks
insolvent and that had terrible knock-on effects for the commercial banks, which
carried similar assets on their balance sheet. Dealer banks had purported to offer
maturity and liquidity transformation services sustained by a market-based form of
deposit insurance. But the stability the ‘private deposit insurance’ provided was only
relative. It reduced run risk in a repo market on a modest scale, but eventually its
capacity to control run risk was exceeded. When that point was reached, this market-
based credit finance had massively increased systemic risk.24

20.4 Securitization and the ‘Perfection’ of Credit

Transformation
In addition to liquidity and maturity transformation, banks engage in ‘credit trans-
formation’. This means the transformation of a collection of risky assets (and thus
variable payout streams) into a safe asset with a fixed payout stream, the deposit. Banks

21
This of course stimulated the product of such assets, both through mortgage origination but also
through financially engineered credit transformation.
22
See Chapter 21.
23
See G Gorton and A Metrick, ‘Securitized Banking and the Run on Repo’ (2012) 10 Journal of
Financial Economics 425; G Gorton and A Metrick, ‘Who Ran on Repo?’, NBER Working Paper No 18455
(2012).
24
Note that this account of this form of market-based liquidity and maturity transformation also suggests
its own corrective: limiting the bankruptcy safe harbour to risk-free assets like US Treasury issuances or
other sovereign debt issuances that will hold their value even in a financial crisis. In other words, if the same
balance sheet that is behind public deposit insurance is reflected in the instrument used for repo finance,
then private deposit insurance will offer equivalent protection. See Chapter 21, section 21.2.3.
 Market-Based Credit Intermediation 443

do this by credit screening in the origination of loans, diversification to minimize

concentrated borrower, industry, and, to the extent possible, geographic risks, and then
follow-on monitoring of loan performance, including intervention to minimize losses
on any particular loan. An important element to this credit transformation was the
bank’s loss-absorbency capacity, particularly shareholder equity and, on occasion,
subordinated debt. Securitization as promoted by its underwriters purported to provide
a superior version of credit transformation and at lower cost. The canonical residential
mortgage-backed security (‘RMBS’) aggregated mortgages that were geographically
diverse; these mortgages were also issued to buyers of lesser creditworthiness, ‘sub-
prime’ or just above.25 The novel feature was to rearrange the incoming cash flows like
a ‘waterfall’, so that the securitization vehicle could issue securities of different credit
quality.26 In the simplest version, the senior ‘tranche’ would have priority over all
incoming cash flows until its claims were paid and thus could receive an AAA-rating.
As a result of this prioritization scheme, ‘many of the manufactured tranches [were] far
safer than the average asset in the underlying pool’.27 This credit transformation was
thought to be superior to a bank’s, because it produced AAA-rated securities from
loans that would weigh down the bank’s balance sheet, and without the cost of
regulatory capital.28 Moreover, the ‘mezzanine’ tranches could be pooled and on the
same prioritization principle, the resulting Collateralized Debt Obligation (‘CDO’)
could also issue a certain fraction of AAA-rated securities. In the hands of a financial
institution, an AAA-rated MBS could also produce immediate liquidity, because as we
discussed in section 20.3.2, it would be taken as collateral for a repo loan close to the
face amount.
We now know that the credit transformation of securitization was in many respects
inferior to a bank’s. First, while a prioritization waterfall could indeed produce a layer
of AAA-rated credit, the financial engineering models—used by the issuer banks and
by the rating agencies—that produced a large percentage of the AAA-rated tranches
were based on inadequate historical data about housing prices, including an assump-
tion about continual price appreciation, and failed to understand that the national
diversification strategy designed for risk reduction had created a national real estate
market that undermined local diversity and added correlation to default risks. More-
over, the relatively small number of underwriters were able to ‘ratings shop’ among the
small number of credit rating agencies (CRAs) to squeeze more AAA ratings out of the
same risk pool.29
Second, it was hard to value an MBS tranche. The complexity of the prescribed
movement of cash within a particular securitization vehicle made their payoff

25
Other multi-sector securitizations might include, for example, automobile loans, credit card receiv-
ables, and even commercial real-estate loans.
26
In practice there were essentially three tranches: AAA, BBB (the ‘mezzanine’ tranche), and the ‘equity’
tranche, which picked up the residuals left by the higher-rated tranches.
27
J Coval, J Jurek, and E Stafford, ‘The Economics of Structured Finance’ (2009) 23 Journal of Economic
Perspectives 3.
28
The securitizer could further enhance the credit quality of the tranche by arrange for third-party
insurance in the form of a credit default swap.
29
See LJ White, ‘Markets: The Credit Rating Agencies’ (2010) 24 Journal of Economic Perspectives 211.
See also Chapter 6, section 6.3.3.
444 Principles of Financial Regulation

structures immensely complex.30 To some extent this facilitated the process of liquidity
and maturity transformation, since short-term credit suppliers to a dealer bank could
feel that the opacity of the MBS structure protected them against potential adverse
selection problems in the proffer and acceptance of collateral.31 On the other hand, as
mortgage defaults increased, real estate values went into decline, and credit rating
agencies’ ratings—and conflicts—came under withering fire,32 the very complexity of
the MBS—and a fortiori, their more exotic brethren—made value determinations
difficult if not impossible. Short-term credit suppliers knew that valuation was import-
ant, knew that they didn’t know it, and thus pulled back from repo finance; they ran.
When the method for producing superior credit transformation—tranching overseen
by credit rating agencies—was shown to be flawed, this form of market-based finance
crumbled.
That is the negative result on one side of the credit transformation process. On the
other side, securitization seems to have exacerbated losses from increasing mortgage
default rates.33 Securitization unbundled the housing finance function through con-
tracts with various agents, including servicers, who collected payments from the
homeowners and forwarded them. The servicers were also supposed to manage any
necessary loan renegotiation and modification with a financially constrained mortga-
gee. But servicers were not paid enough to cooperate; their incentives were to foreclose,
irrespective of ultimate recovery rates on the defaulted mortgage or the impact on
surrounding property values. The complexity of many securitization vehicles made it
very difficult to avoid irritating some fraction of loss-bearing holders. By contrast,
banks, which internalize the cost of foreclosure, achieved more modifications and
ultimately preserved more value. In short, securitization provided a market substitute
for the bank’s credit transformation, but it presented its own set of institutional frailties
and showed a propensity for adding, not subtracting, systemic risk.
In sum, the commercial banks through their off-balance sheet SIVs, the dealer banks
in the financing of their balance sheets, and the securitizers in their reshaping of
financial claims, had each engaged in forms of transformation activity traditionally
performed by banks. Because of this, and the fact that this activity was not regulated as
‘banking’, it has come to be known as ‘shadow banking’.

30
See R Bartlett, ‘Inefficiencies in the Information Thicket: A Case Study of Derivatives Disclosure
During the Financial Crisis’ (2010) 36 Journal of Corporation Law 1; K Judge, ‘Fragmentation Nodes’ (2012)
64 Stanford Law Review 657.
31
That is to say, the dealer banks offering the securities as collateral could credibly claim they did not
understand the expected cash flows much better than the institutional investors taking them as such in the
repo market. See sources cited in n 1. Such adverse selection as there was in securitizations was at the stage
when the loans were transferred to the SIV from the originator, not in the transfer of the MBS that the SIV
issued: see BJ Keys, T Mukherjee, A Seru, and V Vig, ‘Did Securitization Lead to Lax Screening? Evidence
from Subprime Loans’ (2010) 125 Quarterly Journal of Economics 307.
32
These conflicts are discussed in Chapter 6, section 6.3.3.
33
See T Piskorski, A Seru, and V Vig, ‘Securitization and Distressed Loan Renegotiation: Evidence from
the Subprime Mortgage Crisis’ (2010) 97 Journal of Financial Economics 369; S Agarwal, G Amromin, I
Ben-David, S Chomsisengphet, and D Evanoff, ‘The Role of Securitization in Mortgage Renegotiation’
(2011) 102 Journal of Financial Economics 559.
 Market-Based Credit Intermediation 445

20.5 Shadow Banking and Market-Based

Financial Intermediation
Since the financial crisis, considerable effort has been devoted to reining in the
propensity of the shadow banking sector to create systemic risk. Perhaps inevitably,
what should count as ‘shadow banking’ has become hotly contested in this regulatory
debate.34 A functional approach calls for understanding how the institutions of
market-based credit intermediation function, both separately and collectively, in
assessing the best regulatory techniques to strengthen the market-based system a
whole. This means that the most important regulatory initiatives may be those that
target market-wide practices and behaviour, rather than particular institutions.
Initially the Financial Stability Board (FSB), the agenda-setter for post-crisis regula-
tory reform, pushed for the broadest conceivable definition of ‘shadow banking’,
namely ‘credit intermediation involving entities and activities (fully or partially)
outside the regular banking system’.35 This would encompass virtually all market-
based credit intermediation. It would imply that asset managers are shadow bankers
and that even a pass-through vehicle like a bond fund is a shadow bank, despite the
limited transformation services that we have described. This definition is apt to capture
the full amount of credit intermediation that may be occurring in the financial system
outside the banking sector. This may be relevant for macroprudential authorities
thinking about the application of time-varying macroprudential tools, for example.36
It may be harder to make these work if a larger proportion of credit intermediation occurs
outside the regulated banking sector, beyond the reach of time-varying capital controls,
for example.
But simply because a firm engages in credit intermediation does not mean that it
adds to systemic risk in the cross-sectional sense.37 If we are concerned to identify those
firms, or sectors, which do this, then a more restrictive definition is needed. We can
think about market-based credit intermediation as potentially adding cross-sectional
systemic risk in two distinct ways: (i) through linkages that render the regular banking
sector less stable; and (ii) through systemic risk within the ‘shadow banking’ sector. If
there are no links to the regular banking sector, then the contribution to systemic risk
of any market-based financial intermediation channel depends on its size relative to the
economy as a whole, and its fragility. A large but stable sector has a small risk of failure,

34
A useful article is Z Pozsar, T Adrian, A Ashcraft, and H Boesky, ‘Shadow Banking’ (2013) FRBNY
Economic Policy Review, December, 1.
35
See eg FSB, ‘Strengthening Oversight and Regulation of Shadow Banking: An Overview of Policy
Recommendations’ (2013), iv; see also M Carney, ‘The Need to Focus a Light on Shadow Banking is Nigh’,
Financial Times, 15 June 2014 (‘the extension of credit from entities and activities outside the regular
banking system’). These sentiments were echoed in a report of the US Treasury’s Office of Financial
Research (the research arm of FSOC), ‘Asset Management and Financial Stability’ (September 2013) that
proved to be highly controversial, though it did provoke the SEC to move towards greater prudential
oversight of large asset managers.
36
See Chapter 19, section 19.3.3.
37
Indeed, if activity were to move away from the banking sector in favour of non-bank financial
intermediaries that engaged only in credit transformation and not liquidity or maturity transformation,
the overall cross-sectional systemic risk would likely be reduced.
446 Principles of Financial Regulation

but would cause a lot of harm to the real economy in the rare circumstances in which it
did fail. Conversely, a smaller but more fragile sector might engender the same overall
level of systemic risk. If the market-based financial intermediation channel is linked to
the regular banking sector, then there may be potential for contagion from one to the
other. In such a case, the consequences of instability in the sector outside the banking
perimeter are greatly magnified, because they potentially encompass the regular bank-
ing system as well.
(i) Linkages to the banking sector. Focusing on the linkages, we can understand that
bank-sponsored SIVs were particularly dangerous because their usage directly threat-
ened the stability of the official banking system through the private LOLR call on the
sponsoring banks. Bank-sponsored securitization was pure regulatory arbitrage—an
expansion of bank leverage permitted by a gap (now closed) in the risk-weighting rules.
This was perhaps not so much ‘shadow’ banking as ‘disguised’ banking designed to fool
the regulator. A specific policy approach—structural regulation of banking—aims to
address the problem of linkages by separating banking from other activities. We
consider this in detail in Chapter 23. Regulators have also responded to the problem
indirectly by seeking to shore up the banking sector through the safeguards discussed
in Chapters 14–18.
(ii) Systemic risk outside the banking sector but within the ‘shadow banking’
sector. The initial question is what sort of ‘shadow banking’ actors exist that are not
linked to the regular banking sector—what we might call a ‘parallel’ banking system?
One example might be dealer banks, and other market actors who engage in bank-like
liquidity and maturity transformation but contract for liquidity protection through
back-up credit lines now provided by hedge funds rather than banks. Another example
might be the finance companies that originate loans but which are not ‘banks’ because
they are not deposit-funded. Instead, they may fund their activities through wholesale
finance (issuing commercial paper, typically to money market funds) and securitiza-
tion (selling their loan assets to minimize capital requirements). So-called ‘fintech’
firms—finance companies that make heavy use of the Internet to reach high-quality
borrowers—fall into this category.
How might we think about systemic risk, then, in a ‘parallel’ banking system? It is
easy to see how the failure of a very large dealer bank, or finance company, could have
systemic effects. For example, the liquidation of the dealer bank is likely to affect the
value of similar assets held by official banks. Moreover, a large dealer bank’s disappear-
ance could well damage the financing of important activity for the real economy.
Similarly, the collapse of a large fintech company, as through the sudden ‘run’ of
commercial paper purchasers, could destroy an important credit channel. Size, there-
fore, is relevant for systemic risk—a point we return to in Chapter 22.
But surely it is not simply a matter of scale. In thinking about the sources of
instability within the market-based credit intermediation sector, it is helpful to look
to the ‘primitives’ we described in the foregoing sections. Certain elements add risk:
short-term finance, maturity mismatches, liquidity mismatches, reliance on a disas-
sembled credit intermediation chain (as with some securitizations), interconnection
with other financial firms that create channels of contagion, high correlation of asset
 Market-Based Credit Intermediation 447

types among financial firms, and size as a proxy of importance to the real economy.
‘Shadow banks’ are unlikely to bundle these elements in the same way as official banks.
They are not monolithic and the levels of systemic risk and fragility will not present in
the same way. A functional approach calls for assessment of the mechanisms or
patterns of bundling and then to regulate accordingly. Regulators have so far responded
to the sources of instability in the shadow banking sector through a number of
measures we consider in Chapter 21. Rather than designate a growing list of institu-
tions as systemically important (which raises the challenging question of tailoring
multiple regimes suitable for diverse institutions), another approach is to designate
specific practices or financing strategies as systemically sensitive and constrain them
accordingly.38 Such a functional approach is aimed at strengthening the resiliency of
market-based credit intermediation as a general system.
The most recent FSB Global Shadow Banking Monitoring Report distinguishes
between the ‘broad measure’ of non-bank financial intermediation and, as a subset of
this, a ‘narrow measure’ of ‘global shadow banking that may pose financial stability
risks’. As of 2014, the narrow measure encompassed assets of $36 trillion globally, as
opposed to $137 million under the broad measure.39 This narrowing should usefully be
understood as not simply an exercise in taxonomy, but rather as motivated by a
functional assessment of the sources of systemic risk. These include the various
primitives we have discussed, such as liquidity and maturity transformation, which
can be performed by the interaction of firms through markets, as well as internalized
within firms.40

20.6 Conclusion
One of the most distinctive practices that has recently emerged in developed financial
systems is the rise of market-based credit intermediation. At first it was thought that
market-based substitutes for bank finance would help to reduce systemic risk, because
they did not seem to rely on the bank balance sheet and did not put at risk the bank-
based payment system. The financial crisis showed that the market-based alternatives
nevertheless created systemic risk. In part, this was because of linkages between market-
based finance and the commercial banking system. One linkage came through contin-
gent liquidity obligations—private LOLR facilities—that banks entered into with market-
finance entities and firms. These facilities burdened—and at times overwhelmed—bank
balance sheets, and constrained their lending capacity. Another linkage came through
damage to bank balance sheets triggered by losses on assets and products originated by
market-finance entities and firms.

38
This is the approach taken by the EU’s recent ‘Securitization initiative’, identifying criteria for ‘simple,
transparent and standardized’ securitizations, which are to be accorded lower capital charges than other
forms of securitization: see European Commission, Proposal for a Regulation laying down common rules
on securitisation and creating a European framework for simple, transparent, and standardized securitiza-
tion, COM/2015/0472 final, Brussels 30.9.2015; European Commission, Proposal for a Regulation amend-
ing Regulation (EU) No 575/2013 on prudential requirements for credit institutions and investment firms,
COM/2015/0473 final, Brussels 30.9.2015.
39 40
FSB, n 2, 2. Ibid, 7–8.
448 Principles of Financial Regulation

Quite apart from the direct and indirect effects on the banking system, the sheer
magnitude of market-based credit intermediation relative to the real economy may
itself threaten financial stability. It is quite possible to imagine circumstances in which a
peer-to-peer lender,41 for example, faces a run on investor confidence, which spreads to
other such firms, and thus brings to a sudden stop activity that is becoming an
important credit channel.
These observations have a number of implications for regulators. First, as a matter of
bank supervision, it will be important for regulators to monitor a bank’s exposure to
market-based credit intermediation, which may attempt to free-ride on the bank’s
balance sheet. Second, non-bank credit institutions may themselves reach systemic size
and scope quite independently of the commercial banking system and, if so, their
importance to the real economy will require their designation and supervision as
systemically important institutions. Third, although institutional size matters, it may
also be that particular funding mechanisms, especially short-term wholesale credit
practices, are especially prone to fragility and so may be most appropriate for regula-
tory intervention. The increasing importance of ‘shadow banking’ calls attention to the
fact that a functional approach requires regulatory attention to the resiliency of market
structures and practices as well as to institutions. We address these issues in more detail
in Chapter 21.

41
Outlined in Chapter 12, section 12.2.5.