Electronic copy available at: http://ssrn.

com/abstract=1962590
Securitisation
and
Bank Intermediation Function
+
Maxim Zagonov
,
September 2011
Abstract
The move from the originate-to-hold to originate-to-distribute model of lend-
ing profoundly transformed the functioning of credit markets and weakened
the natural asset transformation function performed by nancial intermedi-
aries for centuries. This shift also compromised the role of banks in channeling
monetary policy initiatives, and undermined the importance of traditional
asset-liability practices of interest rate risk management. The question is,
therefore, whether securitisation is conducive to the optimal hedging of bank
interest rate risk. The empirical results reported in this work suggest that
banks resorting to securitisation do not, on average, achieve an unambiguous
reduction in their exposure to the term structure uctuations. Against this
background, banks with very high involvement in the originate-to-distribute
market enjoy lower interest rate risk. This however by no means implies su-
perior risk management practices in these institutions but is merely a result
of disintermediation.
JEL classication: G21; G28; E52; C23
Keywords: Financial institutions; Interest rate risk; Securitisation

The author would like to thank Hippokrates A. Hadjiilias, Aneel Keswani, Ian W. Marsh, David
Stolin, Elyas Elyasiani, Santiago Carb Valverde and Ning Gong for their valuable suggestions that
helped to greatly improve this work. Further, the author is greatful to the seminar participants at
the annual meetings of the Australasian Banking and Finance 2010, Midwest Finance Association
2011, Southwestern Finance Association 2011 and Financial Management Association 2011. The
ndings and conclusions expressed in the paper are entirely those of the author.
y
Toulouse Business School, Universit de Toulouse, 20 Boulevard Lascrosses, 31068 Toulouse,
France. Email: [email protected]; Tel: +33 561 294 974; Fax: +33 561 294 994.
Electronic copy available at: http://ssrn.com/abstract=1962590
1 Introduction
The recent nancial crisis prompted by the US subprime mortgage meltdown has
demonstrated the detrimental impact a troubled banking sector has on the wider
economy both domestically and internationally. The nancial markets worldwide
suered disastrous losses, with massive declines in portfolio values of various, in-
cluding highly rated, securities. The crisis also led to a severe liquidity shortfall
that adversely aected all economic agents. As credit tightened, the myriad of for-
mally prosperous businesses were forced to le for bankruptcy, resulting in soaring
unemployment and unprecedented decline in international trade.
Securitisation is generally regarded as the key culprit in the subprime debacle,
thus provoking copious discussions on possible remedies for the market for securi-
tised assets. Recently, a plethora of contributions addressed these issues both em-
pirically and analytically
1
. Together these works suggest that the root causes of the
crisis are by no means exogenous, and reside in managers opportunistic behaviour,
propensity to short-termism, and concomitant regulatory policies that abetted these
trends. Beyond this point of agreement, the issue remains an ongoing debate among
academics, practitioners, and policymakers with many of the underlying causes yet
to be fully understood.
Interestingly, none of the aforementioned causes is new, and they have all been
acknowledged as the primary determinants of the major nancial crises in the past.
Three common causes are particularly emphasised: moral hazard and information
asymmetries; global imbalances; and a poorly designed multi-layered regulatory
framework which further aggravated an already present misalignment of incentives.
However, what makes the current crisis dierent is a contagion which was man-
ifested due to highly developed inter-linkages between international nancial cor-
porations, their complexity, multi-sector involvement, and a speedy transmission
of news and investment ows. What started as a relatively isolated US subprime
mortgage episode was then propagated to the rest of the nancial sector worldwide,
aecting all major asset classes. In response, a great deal of research has focused
on examining the market mechanism by which the nancial contagion is prolifer-
ated, proposing even more solutions to contain the shock spill-overs in the future
(Brunnermeier, 2009; Longsta, 2010).
Further contributions have also addressed the role of rating agencies, condemn-
ing their inability to properly rate the securitised products (Skreta and Veldkamp,
2009). Agencies incentives, and conict of interest are also emphasised (Bolton,
1
A detailed discussion on the mechanisms of the subprime mortgage crisis is oered by Brun-
nermeier (2009).
2
Freixas, and Shapiro, 2011). The design of the compensation structure, with man-
agers rewards being tied to short-term mark-to-market prots rather than the long-
term protability and solvency of created positions, has also been acknowledged for
contributing to the crisis (Erkens, Hung, and Matos, 2009). In a similar vein, the
regulatory architecture which allowed, and in some instances abetted, such short-
termist behaviour has also been denounced (Acharya and Richardson, 2009).
While much has been learnt from these contributions, they have predominantly
concentrated on the underlying causes of the current events, not the risks facing
the nancial system in the aftermath of the crisis. For instance, none has explicitly
addressed the issue of bank interest rate exposure, the importance of which was
reasserted by recent developments in the monetary environment.
Following an unprecedented reduction in the nominal interest rates, today the
concern exists that banks have relaxed their asset-liability management practices and
are less protected than ever against rising interest rates
2
. This concern is reected
in the speech of then Vice Chairman of the Board of Governors of the Federal
Reserve System, Donald L. Kohn at the Federal Deposit Insurance Corporations
Symposium on Interest Rate Risk Management in January 2010. In his speech, Dr.
Kohn stressed that "... interest rate risk is inherent in the business of banking..." and
"... it is especially important now for institutions to have in place sound practices to
measure, monitor, and control this risk". He further cautioned that as the economy
recovers, it is reasonable to expect a tightening in monetary policy, with associated
developments in the entire shape of the term structure being hard to predict. In this
respect, the unprecedentedly high issuance of government debt worldwide, coupled
with increasing inationary pressure, may trigger sharp changes in the interest rate
environment. As suggested by Dr. Kohn, it is highly unlikely that the interest
rate volatilities will "...return to their previous quiescent state", thereby posing
further concerns for the stability of the nancial sector. The shape of the term
structure is also likely to undergo signicant changes. As the investors return to
higher risk leveraged positions, the yields oered on sovereign instruments will have
to be revisited in order to successfully nance the scal decit. Furthermore, due to
the crisis-induced liquidity constraints, many institutions were forced to shorten the
maturity of their liabilities and are accordingly exposed to greater renancing risk
3
.
2
Over the last two years, the US yield curve has experienced a considerable steepening, with
the interest rate spread widening to a multi-decade level high. This steepening poses a signicant
challenge to the asset-liability managers, particularly in addressing possible non-parallel shifts in
the term structure. The empirical evidence on the adverse impacts of low interest rates on bank
risk is provided in Altunbas, Gambacorta, and Marqus-Ibez (2010).
3
Further to this, according to the Oce of Thrift Supervision Quarterly Review of Interest Rate
Risk, in the rst quarters of 2010 the median percentage ratio of xed-rate mortgage loans held by
3
And while the prudently managed companies will presumably access the required
funds, the increased competition for credit may escalate its cost. On the asset side,
as many households nd the value of their debt exceeding the value of the underlying
equity, the rate of defaults is likely to peak with interest rates.
Such economic conditions raise the fundamental question of what are the most
eective and suitable ways to hedge against unanticipated developments in the yield
curve. In this respect, the theoretical benets of securitisation for ecient manage-
ment of bank interest rate risk are unambiguous. On the one hand, securitisation
serves as a channel to transfer interest rate risk from the nancial intermediary to
parties better equipped to bear and manage this exposure. On the other hand, it
provides an opportunity to align the duration of interest rate sensitive assets and li-
abilities, thereby reducing the balance sheet duration gap and concomitant exposure
to interest rate movements. Further, securitisation income oers the potential to
improve revenue diversication, thus reducing bank reliance on interest-generating
activities
4
. Despite these sound theoretical grounds, no empirical account of the
impact of securitisation on bank interest rate risk has hitherto been conducted.
Accordingly, the objective of the work reported here is to circumvent the afore-
mentioned issues in addressing the impact of securitisation on bank interest rate
risk. In particular, the paper oers three major contributions to the literature.
First, utilising an extensive sample of publically traded US bank holding compa-
nies, this work empirically veries the importance of interest rate exposure for the
majority of analysed institutions over the 2001 to 2009 period. Nearly 95 percent
of analysed nancial intermediaries are adversely aected by yield curve shocks at
one time or another, with the yield curve slope being the most signicant source of
risk. The banks resorting to asset securitisation are aected to a higher degree by
term-structure movements than their non-securitising counterparts.
Second, this is the rst study which explicitly relates the level of bank securiti-
sation activities to its interest rate exposure. While the empirical evidence to date
suggests that securitisation aects the level of bank credit risk, its solvency, and
eciency, no empirical test to assert its impact on bank interest rate risk has been
conducted. The results reported here oer a valuable insight to both managers and
the US thrifts to their total assets was at the level of 40.6%, while the corresponding proportion
of all adjustable-rate mortgage loans to total assets was at only 22.3%. The eective duration gap
in the thrift industry also remained positive, highlighting the rms susceptibility to rising interest
rates.
4
As argued by Keswani, Marsh, and Zagonov (2011), since activities that generate non-interest
income are imperfectly correlated with those generating interest revenues, with raising interest
rates, the diversication of revenue sources should help stabilizing operating income and give rise
to a more stable stream of prots. This view is supported by the empirical ndings of Smith et al.
(2003) and Chiorazzo, Milani and Salvini (2008).
4
regulators looking into securitisation to curb bank interest rate risk. This is partic-
ularly important in the aftermath of the global nancial crisis, with the monetary
policy decisions creating a unique environment for interest rate exposure.
Third, the current research also studies whether the securitisation of assets with
dierent maturities and risk characteristics impacts dierently on bank interest rate
exposure. The empirical tests suggest that interest rate risk generally increases
with the maturity of assets securitised. To decouple the eect of securitisation
from other factors, I consider further channels that may have aected bank risk.
These include numerous bank-specic characteristics and the macroeconomic envi-
ronment in which banks operate. Further, the research covers both pre-crisis and
crisis episodes, thereby oering an opportunity to compare the eectiveness of se-
curitisation in curbing bank interest rate risk between the two periods. I nd that
banks resorting to asset securitisation are subject to greater interest rate exposure
in the second, crisis sub-period.
The remainder of the paper is organised as follows: Section 2 provides a brief
review of the literature and outlines a set of testable hypotheses. Section 3 presents
a theoretical model of nancial intermediary risk-taking behaviour, while Section 4
continues by outlining the supporting empirical framework. The description of the
data sample follows in Section 5. Empirical results are discussed in Section 6, while
Section 7 concludes the paper.
2 Literature review and hypotheses formulation
Securitisation is a relatively straightforward process of transforming a pool of
illiquid assets into marketable securities via cash ow repackaging; yet it has sub-
stantially reshaped the credit markets in recent decades. While originally conned
to the US residential mortgages, today, securitisation is applied to a wide range
of asset classes, including credit card, commercial and industrial, automobile, and
home equity loans, among others. Since its inception in the late 1960s, the issuance
of securitised assets in the US has been growing steadily to amount to nearly US
$2.11 trillion as of the year end 2009
5
.
5
Aggregate of the US mortgage-related (MBS) and asset-backed securities (ABS) issuance,
based on the data compiled by the Securities Industry and Financial Markets Association,
http://www.sifma.org. The fastest growth was enjoyed by the MBS sector, with a nearly 11.2%
(15.8%) compound annual growth rate between 1996 and 2009 (1996 and 2006). The correspond-
ing growth rates for the US ABS issuance are -0.8% and 16.3% respectively. The declining trend
in MBS is likely to persist in the foreseeable future, owed to weak house sales, mortgage loan
origination, and new housing start-ups following the crisis. The number of house sales in the US
has reached its peak of 1.28 million in 2005, and declined since to 0.38 million in 2009. The same
is true for new housing start-ups, declining at a compound rate of 28.1% per year between 2005
5
On the theoretical front, access to the market for securitised products may sub-
stantially benet the originator by (a) allowing to eciently diversify its credit
portfolio; (b) improving asset-liability management; (c) reducing the cost of nan-
cial intermediation; and (d) providing an opportunity to prot by specialising in
operations in which it enjoys a comparative advantage
6
. As suggested by Loutskina
and Strahan (2009), securitisation eases the inuence of bank nancial conditions
and local funding shocks on credit supply. As a result, it increases liquidity and
facilitates the reduction of funding and therefore banks intermediation costs. Fur-
ther, securitisation provides a means to eciently transfer the risk from the banks
balance sheet to other economic players better equipped to bear it, thereby removing
the impediment to further growth implied by capital and balance sheet constraints.
In terms of bank interest rate risk, securitisation oers an opportunity eectively
to tailor the balance sheet duration gap induced by the banks asset transformation
function. Thanks to heterogeneity in the maturity of assets admissible for secu-
ritisation, the duration of rate sensitive assets can be perfectly matched to that
of corresponding liabilities. Further, by securitising assets with embedded prepay-
ment provisions, the lender, in eect, resells the position held in these options and
therefore hedges its exposure to unanticipated increases in interest rate volatility.
Despite the unambiguous theoretical benets oered by securitisation, the em-
pirical evidence and the state of market predicament to date suggest that nancial
institutions may have been unable fully to enjoy such advantages. With many rms
moving from an "originate-to-hold" to "originate-to-distribute" business model, the
agency problems become ever more apparent and a vast literature analyses this issue
in depth (Berndt and Gupta, 2009; Drucker and Puri, 2009). In particular, due to
the separation of asset ownership and control functions, the loan originator lacks
the incentive to exert enough eort in monitoring the credit quality of any pursued
projects. Provided with a channel to alleviate its credit exposure, the intermediary
is more concerned with the fees it extracts from the new loan origination rather
than the underlying quality of these loans. As demonstrated by Keys et al. (2010),
the likelihood of originating sub-quality loans increases with the probability of the
loans being sold. Furthermore, the funds released from asset shifts are commonly
used to nance more protable, yet riskier avenues (Cebenoyan and Strahan, 2004;
Purnanandam, 2011). And while various mechanisms were introduced to minimise
moral hazard and to better align the interests of bankers and investors (Gorton
and Pennacchi, 1995), inecient contractual environment and misplaced regulatory
and 2009 (source: US Census Bureau, http://www.census.gov).
6
For more insightful discussion on the benets of securitisation, see Greenbaum and Thakor
(1987).
6
eorts precluded a complete resolution of these problems.
Besides, under poorly designed regulatory capital charges, banks have an incen-
tive to securitise safer, low-yield assets while retaining riskier and more protable
ones. As demonstrated by Ambrose, Lacour-Little and Sanders (2005), intermedi-
aries commonly securitise safer mortgages and retain the more risky ones on the
balance sheet. An extensive scope of works provides further empirical evidence to
support this "regulatory arbitrage hypothesis" for asset securitisation. Many also
agree that even with no capital distortion, the banks are likely to shift safer as-
sets, owed to excessive costs involved in distribution of riskier instruments due to
the "lemons" problem (Akerlof, 1970). Additionally, despite the fact that under
the FASB140 rule (Financial Accounting Standards Board) securitisation is clas-
sied as an asset sale, in practice, this transaction resembles a typical nancing
arrangement with securitisers commonly retaining their credit exposure by provid-
ing various credit enhancements and guarantees. For this reason, the o-balance
sheet treatment of such transactions has been greatly criticised in the literature.
Moreover, with the increased popularity of securitised products, a myriad of non-
depository market players entered the lending business directly to compete with tra-
ditional intermediaries. This translated into increased market competition, forcing
many nancial institutions to accept higher risks to remain competitive. From this
perspective, securitisation is unlikely to be utilised as a risk-transfer mechanism, but
is rather motivated by the desire for greater protability.
On the basis of the discussion so far, and following the recent events in global
nancial markets, the possibility of banks utilising securitisation to curb interest
rate risk seems rather elusive. This view is reected in the rst testable hypothesis:
Hypothesis
1
: Banks resorting to asset securitisation face greater interest rate ex-
posure. The extent of this exposure varies with the duration of assets securitised.
Against this background, there is evidence to suggest that in the run up to
the subprime crisis banks successfully shifted a great deal of riskier assets owing
to favourable monetary and regulatory conditions. This trend was majorly fuelled
by a low interest rate environment, the increased market demand for securitised
products, and investors excessive reliance on credit ratings reinforced by copious
regulatory provisions. This view is empirically supported by Mian and Su (2009)
and DellAriccia, Igan, and Laeven (2008). However, both papers report a pro-
nounced decline in the lending standards associated with higher securitisation rates.
The former contribution also reports a signicant upturn in bank "disintermedia-
tion" over the 2001-2005 period, with a substantial increase in loans sold shortly
7
after origination.
In the same vein, many have argued that in the last decade banks have moved
from a traditional spread generating strategy to a new equity-maximisation fees-
generating strategy. By assertively strengthening its involvement in the "originate-
to-distribute" market, many intermediaries, in eect, function as brokers who ex-
tract the fees for joining borrowers and lenders. And while the asset repackaging
and sale is costly to the originator, the costs associated with joining the complemen-
tary transactions between borrowers and securitised-debt investors are considerably
reduced through the standardisation of securitised products. Besides, the company
achieves economies of scale by specialising in structured nance transactions. It also
enjoys increasing returns to scale in evaluating the borrowers credit quality due to
lax monitoring. Furthermore, the active players in the securitisation market enjoy
better access to derivative instruments which, as demonstrated by Purnanandam
(2007), enable these companies to preserve the extent of loan origination even as
monetary conditions tighten.
With this business model, the importance of interest generating revenues declines,
and so does the eective duration of assets held on the balance sheet. Accordingly,
the duration gap remains at minimal levels, and the intermediary is less exposed
to the risk of changing interest rates. On the basis of this argument, the following
hypothesis is added into the analysis:
Hypothesis
2
: The relationship between bank interest rate risk and asset securitisa-
tion is non-linear. The risk initially increases with the value of assets securitised,
but declines with bank "disintermediation".
3 Theoretical background
The interest rate exposure represents a natural risk faced by all nancial inter-
mediaries due to the nature of their maturity transformation business model. In
particular, this type of risk may arise from three key sources. First, by transforming
the short-term savings to long-term investments, banks unavoidably mismatch the
duration of the interest sensitive assets and liabilities. The "Duration Theorem"
independently proposed by Samuelson (1945) and Hicks (1946) states that if the
weighted duration of the asset stream is greater (less) than the weighted duration
of the liability stream, the interest rate increase (decrease) will reduce the individ-
uals net worth. With therefore a positive duration gap, measured as the dierence
between the durations of assets and liabilities, rising interest rates reduce the value
of assets more than the value of corresponding liabilities. The earlier attempt to
8
formalise the practical applications of the proposed theory can be traced to the work
of Redington (1952) who introduces the so-called "immunisation rule". Under this
simplied rule, the agent chooses to always hedge against interest rate shocks by
matching the durations of rate sensitive assets and liabilities.
Second, when the rates earned on the underlying assets are not perfectly corre-
lated with the rates paid on the liabilities, the banks earnings are exposed to interest
rate uctuation. This is referred to as the interest rate margin risk. Following the
Federal Reserves decision to reduce the interest rates to unprecedentedly low levels,
the bankers have enjoyed a substantial increase in the interest rate margins. These
conditions may substantially change as the monetary policy tightens, with many
banks nding it dicult to renance some of their xed rate assets with variable
rate liabilities. Finally, the third source of interest rate risk arises from optionality
embedded in some assets and liabilities (e.g. prepayment options). This asymmetric
source of interest rate risk gained its prominence in recent decades.
To theoretically formalise the aforementioned sources of interest rate risk, and
to see how securitisation may be used in curtailing these exposures, this section
presents the model of bank intermediation and describes its key attributes. For sim-
plicity, the model concentrates on the banks duration transformation function and
discounts any other claim attributes and risks. Formally, I assume that the interests
of shareholders and managers are aligned in their combined utility maximisation
(A.1). Accordingly, the bank pursues the strategy of maximising its after-tax prof-
its. The credit market is perfectly competitive la Besanko and Thakor (1987),
with the credit contracts designed to maximise the expected utility of borrowers.
At each planning date t the manager can choose the amount to be invested in
assets and liabilities of dierent maturities, conditional on her choices in preced-
ing periods. The maturity of available projects is limited by 1, which represents
the managers investment horizon. Some divergences from the target asset mix
are inevitable in the short-run, though the banks choice of principal specialisa-
tion determines the market condition it faces and its ability to promptly adjust
the composition of the asset portfolio. Bank liabilities are subject to similar con-
straints, with relatively stable, manager controlled federal funds, though volatile
deposit base. The latter contracts represent a relatively stable funding source in the
presence of a deposit insurance guarantee. Assuming further that t is continuously
dened on the closed interval [0. 1], the banks asset and liability streams over the
investment horizon are (t) and 1(t) respectively. The interest rates are stochastic
and independent of the banks choice of balance sheet structure, with the function
1(t) characterising the market term structure over the interval [0. 1]. The interme-
9
diary can nonetheless negotiate favourable rate conditions on its assets and liability
contracts (e.g., spreads over index rates such as LIBOR) owing to its market power.
The BHCs equity value Q is therefore simply the dierence between the present
values of its asset and liability streams:
Q =
Z
T
0
(t)c
R(t)t
dt
Z
T
0
1(t)c
R(t)t
dt = 1 (1)
where the present values of asset and liability streams are denoted by and 1
respectively.
In a similar manner, the BHCs net income \t 0 is dened as:
1 = 1
a
(t)(t) 1
l
(t)1(t) (2)
where 1
a
(t) and 1
l
(t) are interest rates charged on assets and liabilities respectively.
For convenience, the regulatory capital charges, as well as the operational costs of
servicing the asset and liability portfolios are assumed away in this specication.
Accordingly, following Assumption 1 (A.1) above, the bank shareholders are
concerned with maximising the value of bank prots:
:(t) = 1
a
(t)(t) 1
l
(t)1(t) +Q (3)
Note that the equity value Q is unaected if the yield curve remains unchanged over
the period; and the bank prots are driven by the net interest margin.
As, however, the term structure evolves, both the bank interest margin and its
equity value would be aected in a number of ways. The exact nature of such
response is convoluted due to the direction of rate movements, the occurrence of
non-parallel shifts in the term structure, and the relationships between the bank
assets and liabilities rates. These considerations unnecessary complicate the model,
and a number of simplifying assumptions are introduced as follows:
A.2 The shifts in the interest rate yield curve are parallel in nature: given a contin-
uous random variable with a probability density function ,() _ 0 and c _ _ /,
the future yield curve can be described by 1(t) +, \t [0. 1].
Accordingly, assuming 1(t) = 1 in (3), the bank interest income remains unaected
as long as the adjustment speed of the rates charged on assets and the rates paid
on liabilities is the same:
J1
a
(t)
J1
=
J1
l
(t)
J1
(4)
10
Under this condition, the prots are determined by the term-structure driven changes
in the market values of the intermediarys assets () and liabilities (1):
J:
J1
=
JQ
J1
=
Z
T
0
(t)c
R(t)t
dt
R
T
0
t(t)c
R(t)t
dt
R
T
0
(t)c
R(t)t
dt
+
+
Z
T
0
1(t)c
R(t)t
dt
R
T
0
t1(t)c
R(t)t
dt
R
T
0
1(t)c
R(t)t
dt
(5)
It is easy to see that
R
T
0
t(t)c
R(t)t
dt
R
T
0
(t)c
R(t)t
dt
and
R
T
0
t1(t)c
R(t)t
dt
R
T
0
1(t)c
R(t)t
dt
are the weighted average time to maturity, or durations, of assets and liability
streams respectively. Denoting the duration of assets with `1
A
and the duration
of liabilities with `1
L
, we get:
J:
J1
=
JQ
J1
= 1 `1
L
`1
A
(6)
It therefore follows that the managers decision problem is to choose the `1
L
and
`1
A
that maximise the value of bank equity Q. Assuming, however, the stochastic
nature of the interest rate movements [1()=
R
b
a
,()d], adjusting the durations
is barely an improvement over the immunisation strategy.
Since banks commonly assume a positive asset-liability duration mismatch, to
reduce the sensitivity of a companys value to interest rate uctuation, the risk
manager must either reduce the duration of assets `1
A
or increase the duration
of liabilities `1
L
. In this respect, securitisation oers an elegant solution to the
rst problem, owed to heterogeneity in the assets admissible for securitisation. In
particular, the lender with a positive duration mismatch can use securitisation in
at least two ways to curtail its interest rate exposure: (a) it can securitise the long
term-assets, such as mortgages, o the balance sheet, thereby reducing the eective
duration gap; (b) it can securitise assets with embedded prepayment provisions and
thus hedge its exposure to unanticipated increases in interest rate volatility.
11
4 Methodological framework
4.1 Yield curve modelling
The standard research methodology of assessing the interest rate exposure pro-
poses to use a single interest rate factor (Stone, 1974). Therefore, it fails to recognise
the time-varying nature of the yield curve shape.
In this study, I account for the sensitivity of BHCs stock returns to the changes
in the entire shape of the term structure by employing simultaneously the level,
slope and curvature of the interest rate yield curve. These measures are calculated
via the Diebold and Lee (2006) factorisation of the Nelson and Siegel (1987) model:

t
(t) = ,
1;t
+,
2;t

1 c

`t

+,
1;t

1 c

`t
c

(7)
where t represents the maturity of the underlying xed-income security and ` is a
decay parameter discussed below.
The Nelson-Siegel model uses just a few parameters (compared for example to
spline methods) and provides enough exibility to capture a range of monotonic,
o-type and humped shapes typically observed in the yield curve data. It ts the
term structure using a exible, smooth parametric function based on a Laguerre
function. Notably, due to its ability to provide a good t of the interest rate yield
curves the model is advocated by Diebold and Lee (2006), and Czaja, Scholz and
Wilkens (2010), and is widely used by central banks and practitioners. The central
banks in nine out of thirteen countries members of the Basel Committee of Banking
Supervision construct a sovereign zero-coupon yield curve using the Nelson-Siegel
class of models.
To estimate the yield curve level, slope and curvature, the series of the sovereign
zero-coupon yields of 12 dierent maturities (t = 3, 6 and 12 months, and 2, 3,
. . . , 10 years) are sourced from the U.S. Federal Reserve Board statistical releases.
These series are used as the initial estimates on the left hand side.
Based on the model parameterisation above, the loading on the level (,
1;t
) pa-
rameter is 1 and is independent of time-to-maturity. Taking the limit, it is easy
to see that lim
!1

t
(t) = ,
1;t
and hence the yield curve level can be seen as a
long-term interest rate variable. It also worth noting that an increase in ,
1;t
would
identically aect all yields, thereby shifting the level of yield curve. Similarly, the
loading on the slope parameter ,
2;t
is driven by the exponential function starting at 1
and decreasing monotonically to zero with increasing maturity. Therefore, the slope
parameter might be seen as short-term interest rate variable. An increase in this
12
parameter would amplify the short-rates more than the long ones. In mathematical
terms, given lim
!0

t
(t) = ,
1;t
+ ,
2;t
, it is easy to see that
t
()
t
(0) = ,
2;t
.
The loading on the last parameter ,
3;t
(curvature) is also driven by the exponential
function, now starting at zero (with the t = 0), increasing for the medium maturi-
ties and decaying back to zero as maturity increases. Accordingly, the yield curve
curvature (,
3;t
) can be seen as the medium term interest rate variable.
Following Diebold and Lee (2006), and Czaja, Scholz and Wilkens (2010), to
obtain the estimates of the level, slope and curvature, the identied series of zero-
coupon yields are regressed on the factor loadings and a constant using the cross-
sectional ordinary least squares technique. With this model factorisation the para-
meters on the right hand side are calculated assuming the prexed value of decay
parameter `. Consistent with Diebold and Lee, the value of the decay parameter `
is xed and is chosen to maximise the loading on the curvature parameter. For com-
parison, the time-varying decay parameter ` is also employed. In this specication `
is chosen to maximise the goodness-of-t statistics of the underlying model at each
time t. Both specications yield statistically identical results. To avoid introducing
an additional time-varying component in the yield-curve model, I resort to the xed
` specication.
Figure 1 plots the estimated level, slope and curvature factors, with the pertinent
statistics outlined in the corresponding table.
Compared to the yield curve slope and curvature, the level factor is less volatile.
This observation is not surprising since the yield curve level serves as a proxy for
the long-term interest rate, with the yields at the long end of the term structure
being generally less volatile.
4.2 Interest rate exposure
To address the underlying empirical hypotheses, I follow a two-stage estimation
procedure in line with previous literature in the area. In the rst step, the interest
rate exposure of BHCs stock returns is modelled via a four-factor GARCH(:. :)
parameterisation
7
of the market model formalised as:
7
The GARCH based econometric framework is used to account for a time-varying element in
the distribution of BHCs stock returns. See for instance Elyasiani and Mansur (1998), Flannery,
Hameed and Harjes (1997).
13
1
it
= c +A
0
it
, +
it
(8)
/
it
= .
0
+
n
X
i=1

2
i;t1
+
m
X
i=1

2
/
i;t1
(9)

it
[
t1
~ `(0. /
it
) (10)
where 1
it
represent the weekly logarithmic returns
8
on BHC i (i = 1 to 304) for the
rms scal year t; c is a scalar, , is a 1 1 vector of coecients and A
it
is the
it-th observation on 1 explanatory variables: A/ = (1
M
. 1
Level
. 1
Slope
. 1
Curvature
).
1
M
is return on the S&P500 market index. 1
Level
. 1
Slope
. and 1
Curvature
represent
unanticipated changes in the level, slope, and curvature of the domestic sovereign
zero-coupon yield curve at time t respectively. The use of unanticipated changes is
advocated by previous research suggesting that asset values should already incor-
porate all the anticipated changes in interest rates. I estimate these unanticipated
changes as the dierence between the actual changes in the respective factor at time
t and ones forecasted via the appropriate specication of the ARMA (/. |) model
9
.

it
is the estimated error term from the mean equation of portfolio i, and /
it
is a
conditional variance of portfolio i over week t. The order of lags (:. :) ensures the
adequate treatment of serial correlation in squared returns, with the formal Engle
ARCH Lagrange multiplier and Ljung-Box Q-statistics determining the correct lag
structure.
The estimated coecients measure the sensitivity of bank is stock returns to
changes in the considered interest rate factor. They are treated as independent
variables in the empirical framework to follow.
4.3 Securitization and interest rate risk
In the second step, the estimated measures of interest rate risk are related to
proxies of bank securitisation and asset sales activities. I use panel data techniques
to fully exploit the potential of the data sample, and to control for unobserved cross-
8
To avoid the bias introduced by the Monday or Friday market eects (French, 1980; Pettengill,
Wingender and Kohli, 2003), the calculation of returns is based on the Wednesday to Wednesday
stock prices. The choice of the weekly sampling interval instead of daily or monthly is determined
by two reasons. First, the ndings of Trzcinka (1986) indicate that the returns calculated at a
daily frequency are not well explained by the normal distribution. By using, however, monthly
sampling frequency the non-normality of daily observation would be avoided just at the expense
of information loss. Second, the use of weekly intervals reduces distortions due to non-trading
holidays and noise trading.
9
For most interest rate factors, the ARMA (k; l) model is specied with autoregressive (k) and
moving average (l) parameters ranging from 1 to 3.
14
sectional and time heterogeneity. The workhorse model specication accounts for
both company specic nancial characteristics and the overall economic and business
conditions in which these rms operate:

,
k
it

= , +o1C
0
i;t1
` +1
0
i;t1
+G
0
t1
+1
0
t
+j
i
+
it
(11)
where, ,
k
it
represents the interest rate risk measure / in year t for bank i. As
discussed above, these measures represent the BHCs equity return sensitivity to
unanticipated changes in the yield curve level, slope, and curvature. ` is an o 1
vector of coecients and o1C
it
is the it-th observation on o securitisation proxies.
Similarly, is an ` 1 and 1
it
is the it-th observation on ` company specic
nancial characteristics; while is an 1 1 and G
t
is the t-th observation on 1
macroeconomic characteristics. 1
t
is a vector of year - dummies of dimension 1 1,
and the company specic eect is measured by j
i
. The model is estimated by either
treating j
i
as xed (xed eect model), thus assuming (` + ` + 1) unknown
coecients, with j = (j
1
. . . . . j
N
)/ being company specic intercepts; or random
(random eect model). In the random eect specication j
i
~ 111(0. o
2

) and
is independent of
i
~ 111(0. o
2
"
). Further, both j
i
and the disturbance term
i
are independent of (o1C
it
. 1
it
. G
t
) for all i and t. For both model specications
the robust standard errors adjusted for serial correlation and heteroskedasticity are
calculated.
In line with Keswani, Marsh, and Zagonov (2011) and Au Yong, Fa and Chalmers
(2009), the absolute values of interest rate betas are used as the dependent variable
in the second step regressions. This aids an economic interpretation of the estimated
results and can be reconciled with the notion that both positive and negative expo-
sures to yield curve shocks represent the risk to bank economic value and should be
treated accordingly. Further, to facilitate the validation of the proposed hypotheses,
various parameterisations of the baseline model are introduced through empirical
investigation.
5 Sample selection
The dataset spans the 2001 to 2009 period and consists of the US publicly traded
bank holding companies (BHC). The choice of sample period is driven by the avail-
ability of required data on BHCs securitisation activities. I identied publicly traded
BHCs by cross-referencing the institutions appearing both in the Federal Reserve
Bank of Chicago Bank Holding Company database and in the dataset supplied by
the University of Chicagos Centre for Research in Security Prices (CRSP). The
15
requisite dataset is accordingly constructed by merging the income statement and
balance sheet data from the Consolidated Financial Statement for Bank Holding
Companies (FR Y-9C form) with the equity market data from CRSP on the basis
of company name and its geographical location. The equity returns are of weekly
frequency, all adjusted for dividend reinvestment and stock splits by CRSP. I further
check for the dataset consistency with Compustat using the CUSIP identier.
The focus on BHCs instead of their commercial bank subsidiaries is determined
by two factors. First, the share price data is commonly available for only the BHC
and not individual banks. Second, the decisions concerning the companys capital
and risk management strategies are ordinarily undertaken at the highest level, and
are not necessarily directed at a single subsidiary.
The banks with missing data on securitisation and asset sales activities, deriv-
ative transactions, total loans and assets, and equity capital are excluded from the
sample. The same applies for the acquired entities. Every eort is taken to detect
and address any outliers arising as a result of measurement or reporting errors in the
underlying datasets. Other non-technical representative outliers, depicting genuine
variability in the considered variables, are dealt with accordingly as per the discus-
sion to follow. This yields a total of 304 bank holding companies with the required
information being continuously available across the entire sample period. The list
of analysed banks is in Appendix A, while the considered variables alongside their
detailed denitions can be found in Appendix B. For each BHC, the annual aggre-
gates of the underlying data are used. The average value of total assets for these
institutions ranges between $16,524 million in 2001 and $35,682 million in 2009,
with the median for two years being $1,017 billion and $2,023 billion respectively.
Bank attributes related to securitisation and loan sales activities are from Sched-
ule HC-S of FR Y-9C lings. For each BHC, I measure the aggregate value of assets,
by category, securitised and sold, or sold but not securitised, within a given scal
year. Additionally, the value of the outstanding principle balance of assets securi-
tised or sold for each bank-year is also considered. The pertinent statistics on these
measures, by year, are reported in Table 1, with a detailed denition for each vari-
able available in Appendix B. Evidently, the loans secured by 1-4 family residential
real estate dominate securitisations and loan sales. This is followed by commercial
and industrial, and credit cards receivable loans.
[Insert Table 1 here]
To account for further bank characteristics and the macroeconomic environment
in which these institutions operate, I introduce two sets of control variables accord-
ingly.
16
5.1 Bank specic control variables
There are six rm level controls, all constructed using FR Y-9C lings. First,
given the evidence of signicant U-shaped relationships between bank capital and
interest rate risk (Keswani, Marsh, and Zagonov, 2011), the ratio of equity capital
to BHCs total assets (CAP) is deployed. It should be noted that by facilitating
the diminution in regulatory capital requirements, securitisation may render the
capital ratios an unreliable approximation of the true bank capital constraints. This,
however, should not signicantly alter the importance of this factor in explaining
the banks interest rate sensitivity because the equity capital itself represents not-
interest rate sensitive liability. Accordingly, rms with higher capital levels are
expected to be less sensitive to interest rate shocks.
Second, following the rationale outlined in previous works, the measure of bank
liquidity (LATA) is also considered. In line with empirical literature, a positive
relationship between banks liquidity and risk is expected. Care should be taken
in interpreting this variable, since securitisation may aect the short-term fund in-
ows and hence inate the bank liquidity ratios. Third, the ratio of non-performing
loans
10
(NPL) is used to measure the quality of the bank asset portfolio. Fourth,
based on the theoretical underpinning outlined in the previous section and in line
with Flannery and James (1984), the measure of balance sheet asset-liability mis-
match (GAP) is calculated as the dierence between interest-earning assets and
interest-bearing liabilities maturing or being repriced within one year, scaled by the
banks total assets. As per the outlined theory, a positive sign on this variable is
expected. Fifth, since the originator commonly retains an equity-like interest in the
transaction, thus maintaining its exposure to credit and prepayment risks, the bank
purchase of credit protection (e.g. credit default swaps) can be seen as an attempt
to hedge this exposure. To this end, I calculate the banks net credit protection
purchase (NECP) as the dierence between the credit protection it buys and sells
in a given scal year.
Finally, to control for the eect of bank activity diversication, a set of asset
and revenue diversication measures is constructed. In line with Laeven and Levine
(2007), the diversication of net operating revenue (ROID) is proxied via a modied
specication of a Herndahl-Hirschman Index (HHI) as follows:
1C11 = 1

Interest income - Non-interest income

(12)
10
A loan is considered delinquent if it fails to acquire interest, or when a payment is 90 days or
more overdue but interest is still acquired.
17
This measure assumes values between 0 and 1, with a higher value suggesting greater
degree of income diversication.
In addition, the income concentration in both interest and non-interest revenue
streams is also captured via a Herndahl-Hirschman Index. In particular, I consider
a broad eight part breakdown for non-interest revenues (H_NOIR), and a twelve
part breakdown for the interest income (H_NITR). In a similar manner, the loan
concentration HHI (H_LOAN) is computed considering ve major categories of
loans. These include agricultural, commercial and industrial, consumer, real estate,
and other loans. More information on the construction of all variables is given in
Appendix B.
To improve the t of the empirical model, I control for further bank character-
istics that may explain the variation in the risk exposures. Namely, the return on
assets (ROA) is utilised to proxy the bank operational performance and eciency,
while the return on equity (ROE) is discounted in the analysis due to its decep-
tiveness for rms with highly leveraged balance sheet. It may also be argued that
the level of bank securitisation, as well as its risk exposure, is determined by the
growth rate of its assets base. Accordingly, the asset growth rate (AGR) is added
to account for this supposition. Finally, as securitisation alters the value of banks
on-balance sheet assets, the size indicator becomes less relevant (DeYoung and Rice,
2004) and it is omitted from the analysis.
To this end, Panel A of Table 2 provides key comparative statistics for the
outlined measures between securitisers and non-securitisers, while Table 3 presents
pairwise correlations for these variables.
[Insert Tables 2 and 3 here]
BHCs resorting to asset securitisation are larger, retain higher capital buers,
and have better diversied non-interest revenues, while their non-securitising coun-
terparts excel in diversifying the interest income. Generally, securitisers seem to
better balance the shares of interest and fee-generating revenues in their total oper-
ating income (ROID). Securitisers also maintain a better diversied loan portfolio,
which, however, seems to be of a lower credit quality as suggested by loan-loss pro-
vision and non-performing loan ratios. Further, these rms purchase more credit
protection than their non-securitising peers. This provides evidence to support the
"regulatory arbitrage hypothesis" for asset securitisation discussed above. Finally,
BHCs not involved in the originate-to-distribute market maintain a lower asset-
liability mismatch on the balance sheet, suggesting that these rms resort to stricter
asset-liability management practices.
18
5.2 Economic environment
In the second group of controls, the overall economic and business conditions
are captured by the annual growth rate in the gross domestic product (GDPG),
and the Aruoba-Diebold-Scotti Business Conditions Index (ADSI) sourced from the
Federal Reserve Bank of Philadelphia database (Aruoba, Diebold, and Scotti, 2009),
respectively. The latter measure accounts for the real economic activity at high
frequency, on the basis of both high- and low-frequency information on six major
economic indicators (i.e. weekly initial jobless claims, monthly payroll employment,
industrial production, personal income less transfer payments, manufacturing and
trade sales, and quarterly real GDP). This index has an average value of zero, with
progressively greater values indicating better than average business conditions and
vice versa. The descriptive statistics for both gures are outlined in Table 2: Panel
B.
To get more stable estimates in the empirical model, all considered explanatory
variables ( = 1. G) are treated for outliers via type I winsorization
11
, with xed
cut-o points of 4^ o. Alternatively, the variables are winsorised at the 1 and 99
percentiles, with the results being robust to the variable winsorisation..
6 Empirical Results
The discussion begins with the results obtained in the rst stage estimation in
Section 6.1. The multivariate regression analysis is discussed in section 6.2.
6.1 Bank interest rate sensitivities
The interest rate exposure of the analysed BHCs is assessed via a four-factor
G1CH (:. :) model formalised in Eq. (8). This model is estimated for each
bank-year, with Table 4 presenting comparative statistics of estimated interest rate
factors for securitisers and their non-securitising peers.
[Insert Table 4 here]
At least 10% of the examined BHCs are signicantly aected by the adverse
movements in dierent components of the interest rate yield curve, thereby indi-
cating the inability of risk managers to timely adopt adequate hedging strategies.
Notably, while the eect of interest rate shocks on the values of both securitisers
11
Type 1 winsorisation commonly refers to the procedure of replacing outliers with the exact
value of the interval limit, while with Type 2 outliers are transformed to predestined weighted
average between their original and the cut-o values.
19
and non-securitisers is similar in its magnitude, the proportion of securitisers signif-
icantly aected by these shocks is appreciably higher. This, in a way, supports the
rst empirical hypothesis which argues that securitisation is unlikely to be employed
as a risk-transfer mechanism.
The majority of the signicant interest rate factors are negative, suggesting that
BHCs maintain a positive duration mismatch between their interest sensitive assets
and liabilities.
6.2 Securitization and interest rate risk
For the main research hypotheses, the panel model in Eq. (11) is rst estimated
with time- and state-xed eects applied to the entire sample of BHCs. The sensi-
tivities of equity values to unanticipated changes in the yield curve level, slope, and
curvature estimated from Eq. (8) are interchangeably used as the endogenous vari-
able in this model. The explanatory variables that control for the company nancial
characteristics, and the country economic conditions, are as discussed in previous
section. All right-hand side measures are lagged to avoid simultaneity bias. When
the economic environment proxies are added into the model, the time-xed eect is
relaxed.
Considering rst the intermediaries exposure to changes in the long end of the
yield curve, Table 5: Panel A outlines the empirical results for Hypothesis 1. The
proxy for bank securitisation activities (TSEC) enters the table positively and signif-
icantly at the one percent level. This implies that BHCs with a greater outstanding
value of securitised assets tend to increase interest rate exposure, with this evidence
providing additional support for the proposed hypothesis. This is also consistent
with the view that securitisation is unlikely to serve as a risk-transfer mechanism,
and is instead motivated by the desire for greater protability.
[Insert Table 5 here]
To attest the second part of the hypothesis, concerning the duration of assets
securitised, I aggregate securitisations by the maturity of the underlying assets into
three categories: long-term (1-4 family residential mortgages), medium term (home
equity lines of credit and commercial and industrial loans), and short-term (auto
loans, credit card receivables, and other consumer and commercial loan and leases).
Given that commercial and industrial loans commonly include short- and medium-
term lending to businesses, they enter both short- and medium-term categories in-
terchangeably. The results, also reported in Table 5: Panel A, are robust to either
specication.
20
It appears that increases in interest rate exposure are mainly driven by securi-
tisation of long-term assets, which are mainly represented by residential mortgages.
This is not surprising given that these type of loans dominate securitisations and
asset sales, and the funds released from these transactions are likely to be reutilised
to extend the loans of similar long-term maturity, yet lower quality. This is in line
with the "regulatory arbitrage hypothesis", which suggests that banks commonly
securitise safer, low-yield, assets and retain more protable, though riskier, ones on
the balance sheet. This also is consistent with the empirical ndings of Ambrose,
Lacour-Little and Sanders (2005), and is further supported by the observation of
higher proportion of non-performing loans and the asset-liability maturity gap mea-
sure for securitising rms. Besides, the distribution of riskier, opaque, assets would
incur a heavy discount due to the "lemons" problem suggested by Akerlof (1970),
and would introduce an impediment to the banks external funding channel once
the market participants learn about the underlying quality of securitised products.
Accordingly, the retained mortgages are subjected to greater interest rate risk,
with their credit quality likely to further deteriorate as the interest rate shocks are
passed on to customers (Drehmann, Sorensen and Stringa, 2010).
Against this background, it can be argued that banks with high involvement in
the originate-to-distribute market function more as brokers, who generate fees by
matching the complementary transactions between borrowers and securitised-debt
investors, than nancial intermediaries. Under this "disintermediation" business
model, the bank shifts the majority of originated loans, and, therefore, has a com-
parative advantage in selecting the projects most suitable for securitisation. Further,
given that loans exit the balance sheet soon after origination, the eective duration
of assets and liabilities held on the balance sheet is short-term and can be closely
matched. Given this background, the active players in securitisation markets are
expected to be less exposed to the risk of changing interest rates, with this view
being reected in the second research hypothesis.
To test this supposition empirically, I reformulate the model in Eq. (11) in a
non-linear form as follows:

,
k
it

= , +`
1
1o1C
i;t1
+`
2
1o1C
2
i;t1
+1
0
i;t1
+G
0
t1
+1
0
t
o +j
i
+
it
(13)
where, ,
k
it
represents the stock return sensitivity of bank i to unanticipated changes
in the yield curve level, slope, and curvature at year t. 1o1C
it
is the it-th obser-
vation on the company securitisation proxy, and 1
it
is the it-th observation on `
company specic nancial characteristics. 1
t
and j
i
are vectors of year- and state-
dummies respectively.
21
Given the model parameterisation, I predict a negative sign on the coecient
estimate for the squared securitisation proxy (1o1C
2
), and a positive sign on 1o1C
variable: `
1
0 and `
2
< 0.
[Insert Table 6 here]
The estimation results in Table 6 support the hypothesised relationship, implying
that interest rate risk initially increases with the value of assets securitised, but
declines with bank "disintermediation". Once again, the results are driven by the
securitisation of long-term assets, with non-linearity being only conrmed for the
long-term interest rates represented by the yield curve level.
In a similar manner, the remaining interest rate factors (yield curve slope and
curvature) are evaluated in Table 5: Panels B and C. For all three measures of
interest rate risk the results are consistent with the theoretical prediction that banks
do not necessarily resort to securitisation to curb their risk exposure. As discussed
above the parameter estimate for the securitisation proxy (TSEC) enters all Tables
signicantly positive. In this respect, the magnitudes of JIRR/JTSEC suggest a
great economic signicance. Thus, a one percent increase in the proportion of total
assets securitised translates into about 0.053 percent increase in BHCs exposure
to shocks in the yield curve level. This, in turn, would imply that a typical US
securitiser will incur an additional $1.79 million decline in its market value following
a typical shock in the yield curve level. The corresponding values for interest rate
slope and curvature are $4.01 million and $1.17 million respectively.
Turning to the remaining bank characteristics in Eq. (11), the majority of co-
ecients estimates are statistically signicant and bear the expected sign. Con-
sistent with prior empirical research, the relationship between equity capital and
bank risk taking is U-shaped. That is, both undercapitalised and well capitalised
intermediaries are generally riskier than banks with intermediate, optimal capital
levels. Further, the institutions with higher degree of revenue heterogeneity also
enjoy lower risk exposures, and so are the companies with higher asset base growth
rate. Not surprisingly, the coecient on the ratio of non-performing loans enters
the table negative, owed to the intrinsic link between credit and interest rate risks
(Drehmann, Sorensen and Stringa, 2010).
6.3 Robustness checks
To corroborate the ndings from the basic model in Eq. (11), I perform a com-
prehensive set of robustness checks. These include the use of dierent time horizons
22
and subsamples; the assumption of alternative model specication and distributional
properties; and an extensive treatment of endogeneity and simultaneity biases.
In the context of this study, endogeneity may arise when the BHCs decision to
participate in the market for securitised products does not only inuence, but is
inuenced by its interest rate exposure. In this scenario, the exogenous treatment
of securitisation activities would introduce simultaneity bias in the regression esti-
mates. Furthermore, additional factors may jointly inuence the variability in both
measures, biasing the ordinary least squares estimation and making it dicult to
infer causal relationship. To address these concerns, I detect potential endogeneity
via a Hausman test and resort to a two-stage least squares (2SLS) panel estimation
procedure by introducing a set of instruments for the BHCs securitisation activi-
ties as appropriate. To identify suitable instruments, I address the banks decision
to securitise by analysing its nancial characteristics in the probit framework (not
reported). The results remain robust to the choice of estimation technique. Col-
umn 1 of Table 7 details the empirical output for the 2SLS regression assuming the
BHCs exposure to the shocks in the yield curve level as an endogenous variable. Al-
though not reported, the results for the remaining interest rate proxies also remain
statistically unchanged.
[Insert Table 7 here]
Furthermore, caution should also be taken in isolating the risk management mo-
tives of asset securitisation from auxiliary inducements. In particular, the incentive
to securitise may be circumscribed by the level of loan demand and current economic
conditions. Faced with unusually high demand for loans, banks would resort to asset
sales to extract higher loan origination rents, and to satisfy the existing customer
demand for funds. On the other hand, weaker loan demand conditions following the
economic downturn make it dicult for an intermediary to successfully perform the
securitisation transaction. This is due to low liquidity and demand for ABS, and
higher credit risk of the underlying asset mix resulting in market mispricing. Such
economic conditions would also aect the level of bank interest rate exposure.
In this respect, the analysed sample period provides a unique opportunity to
explicitly test this supposition by separating the time horizon into pre-crisis and
crisis episodes. This also provides a comparison of the eectiveness of securitisation
in curbing interest rate risk between the two periods. In addition, the sample of
companies is separated into a number of sub-samples on the basis of ranking by
the banks (1) size, (2) liquidity, and (3) net derivative usage (hedging - trading).
Selected are the top 25% and the bottom 75% of values in each category, with a
total of six portfolios constructed.
23
The pertinent results for these tests are also reported in Table 7. The coe-
cient estimates on the bank securitisation proxy remain robust to the considered
time horizon, thus reconrming the ndings in the previous section. Not surpris-
ingly, it appears that BHCs are subjected to greater risk exposure in the second
crisis-episode. Turning to the measure of bank size, the estimate on the securi-
tisation proxy remains signicant only for the smaller companies. This might be
explained by the fact that larger BHCs are better equipped to weather the yield
curve shocks owed to better diversied portfolios and unrestricted access to the
markets for derivative products. On the other hand, these rms might also pursue
the "disintermediation" business model, therefore reducing the balance sheet du-
ration gap and concomitant exposure to interest rate movements. Once the bank
liquidity and derivative activities are considered the estimation suggests that the
risk exposure is greater for the companies retaining higher liquidity buers and for
BHCs which are the net traders of derivative instruments. The intermediary is clas-
sied as the net-trader if the notional amount of all derivative instruments held for
trading exceeds that of instruments held for hedging.
7 Concluding remarks
The recent turmoil in global nancial markets, prompted by the US subprime
mortgage meltdown, has once again accentuated the importance of banking sector
prudency for overall economic stability worldwide. Securitisation is consensually
regarded as the key culprit in the subprime debacle, with a plethora of works ad-
dressing possible remedies for the market for securitised assets. These contributions,
however, are largely concerned with the underlying causes of the current events, not
the risks facing the nancial system in the aftermath of the crisis. None has explic-
itly addressed the issue of bank interest rate risk, the importance of which becomes
increasingly apparent in the current monetary environment. This concern has been
recently agged by regulatory authorities both in the US and in Europe, with super-
visors emphasising the necessity of establishing robust practices to measure, monitor,
and control bank interest rate exposures.
In this context, the move from the originate-to-hold to originate-to-distribute
model of lending profoundly transformed the natural asset intermediation function
performed by banks for centuries and compromised the importance of traditional
asset-liability practices of interest rate risk management. Against this background,
this work empirically examines the impact of securitisation on bank interest rate
risk. In particular, the research questions whether securitisation is conducive to the
24
optimal hedging of bank interest rate risk, or is merely a funding source enabling
these companies to pursue more protable, yet riskier, projects
The empirical results reported in this work suggest that banks resorting to asset
securitisation do not, on average, achieve an unambiguous reduction in their expo-
sure to the term structure developments. It appears that interest rate risk generally
increases with the maturity of assets securitised, with securitisation of long-term
assets driving the results.
In addition, banks with very high involvement in the originate-to-distribute mar-
ket enjoy lower interest rate risk, thereby suggesting an asymmetric U-shape rela-
tionship between bank risk and securitisation. This observation, however, does not
imply superior risk management practices in these institutions but is merely a re-
sult of disintermediation. In particular, I argue that BHCs with high involvement to
the market for securitised products function more as brokers, who generate fees by
matching the complementary transactions between borrowers and securitised-debt
investors, than nancial intermediaries. Under this "disintermediation" business
model, the importance of interest generating revenues declines, and so is the eec-
tive duration of assets held on the balance sheet. Accordingly, the duration gap
remains at minimal levels, and the intermediary is better protected against term
structure developments.
25
References
[1] Acharya,V.V., Richardson, M., 2009. Restoring nancial stability: How to re-
pair a failed system. Wiley.
[2] Akerlof, G.A., 1970. The market for lemons: Quality uncertainty and the
market mechanism. Quarterly Journal of Economics 84, 488-500.
[3] Altunbas, Y., Gambacorta, L., Marqus-Ibez, D., 2010. Does monetary policy
aect bank risk-taking? European Central Bank, Working paper series No 1166.
[4] Ambrose, B. W., Lacour-Little, M., Sanders, A.B., 2005. Does regulatory capi-
tal arbitrage, reputation, or asymmetric information drive securitization? Jour-
nal of Financial Services Research 28, 113-133.
[5] Aruoba, S.B., Diebold, F.X., Scotti, C., 2009. Real-time measurement of busi-
ness conditions. Journal of Business and Economic Statistics 27, 417-427.
[6] Au Yong, H.H., Fa, R., Chalmers, K., 2009. Derivative activities and Asia-
Pacic banks interest rate and exchange rate exposures. Journal of Interna-
tional Financial Markets, Institutions and Money 19, 16-32.
[7] Berndt, A., Gupta, A., 2009. Moral hazard and adverse selection in the
originate-to-distribute model of bank credit. Journal of Monetary Economics
56, 725-743.
[8] Besanko, D., Thakor, A.V., 1987. Competitive equilibrium in the credit market
under asymmetric information. Journal of Economic Theory 42, 167-182.
[9] Bolton, P. Freixas, X., Shapiro, J., 2011. The credit ratings game. Journal of
Finance, forthcoming.
[10] Brunnermeier, M.K., 2009. Deciphering the liquidity and credit crunch 2007-
2008. Journal of Economic Perspectives 23, 77100.
[11] Cebenoyan, S.A., Strahan, P.E., 2004. Risk management, capital structure and
lending at banks. Journal of Banking and Finance 28, 19-43.
[12] Chiorazzo, V., Milani, C., Salvini, F., 2008. Income diversication and bank
performance: Evidence from Italian banks. Journal of Financial Services Re-
search 33, 181-203.
[13] Czaja, M., Scholz, H., Wilkens, M., 2010. Interest rate risk rewards in stock
returns of nancial corporations: evidence from Germany. European Financial
Management 16, 124-154.
[14] DellAriccia, G., Igan, D., Laeven, L.A., 2008. Credit booms and lending stan-
dards: Evidence fromthe subprime mortgage market. IMF Working paper series
No. WP/08/106.
26
[15] DeYoung, R., Rice, T., 2004. How do banks make money? The fallacies of
fee income. Economic Perspectives, Federal Reserve Bank of Chicago fourth
quarter, 34-51.
[16] Diebold, F.X., Lee, C., 2006. Forecasting the term structure of government
bond yields. Journal of Econometrics 130, 337-264.
[17] Drehmann, M., Sorensen, S., Stringa, M., 2010. The integrated impact of credit
and interest rate risk on banks: A dynamic framework and stress testing appli-
cation. Journal of Banking and Finance 34, 713-729.
[18] Drucker, S., Puri, M., 2009. On loan sales, loan contracting and lending rela-
tionships. Review of Financial Studies 22, 2835-2872.
[19] Elyasiani, E., Mansur, I., 1998. Sensitivity of the bank stock returns distribution
to changes in the level and volatility of interest rate: a GARCH-M model.
Journal of Banking and Finance 22, 535-563.
[20] Erkens, D., Hung, M., Matos, P., 2009. Corporate governance in the recent
nancial crisis: evidence from nancial institutions worldwide. University of
Southern California, Marshall School of Business, Working paper series.
[21] Flannery, M.J., James, C.M., 1984. Market evidence on the eective maturity of
bank assets and liabilities. Journal of Money, Credit and Banking 16, 435-445.
[22] Flannery, M.J., Hameed, A.S.,Harjes, R.H., 1997. Asset pricing, time-varying
risk premia and interest rate risk. Journal of Banking and Finance 21, 315335.
[23] French, K.R., 1980. Stock returns and the weekend eect. Journal of Financial
Economics 8, 5569.
[24] Gorton, G.B., Pennacchi, G.G., 1995. Banks and loan sales: Marketing non-
marketable assets. Journal of Monetary Economics 35, 389-411.
[25] Greenbaum, S.I., Thakor, A.V., 1987. Bank funding modes: Securitization ver-
sus deposits. Journal of Banking and Finance 11, 379-401.
[26] Hicks, J.R., 1946. Value and capital. Oxford: Clarendon Press.
[27] Keswani, A., Marsh, I.W., Zagonov, M., 2011. Bank regulations and inter-
est rate risk: an international perspective. Cass Business School and Toulouse
Business School, Working paper series.
[28] Keys, B.J., Mukherjee, T.K., Seru, A., Vig, V., 2010. Did securitization lead to
lax screening? Evidence from subprime loans. Quarterly Journal of Economics
125, 307-362.
[29] Laeven, L., Levine, R., 2007. Is there a diversication discount in nancial
conglomerates? Journal of Financial Economics 85, 331-367.
[30] Longsta, F.A., 2010. The subprime credit crisis and contagion in nancial
markets. Journal of Financial Economics 97, 436-450.
27
[31] Loutskina, E., Strahan, P.E., 2009. Securitization and the declining impact of
bank nance on loan supply: evidence from mortgage originations. Journal of
Finance 64, 861-889.
[32] Mian, A.R., Su, A., 2009. The consequences of mortgage credit expansion:
Evidence from the U.S. mortgage default crisis. Quarterly Journal of Economics
124, 14491496.
[33] Nelson, C.R., Siegel, A.F., 1987. Parsimonious modeling of yield curve. Journal
of Business 60, 473-489
[34] Pettengill, G.N., Wingender, J.R. , Kohli, R., 2003. Arbitrage, institutional
investors, and the Monday eect. Quarterly Journal of Business and Economics
42, 49-63.
[35] Purnanandam, A.K., 2007. Interest rate derivatives at commercial banks: An
empirical investigation. Journal of Monetary Economics 54, 1769-1808
[36] Purnanandam, A.K., 2011. Originate-to-distribute model and the subprime
mortgage crisis. Review of Financial Studies 24, 1881-1915.
[37] Redington, F.M., 1952. Review of the principles of life oce valuations. Journal
of the Institute of Actuaries 78, 286-340.
[38] Samuelson, P., 1945. The eects of interest rate increases on the banking system.
American Economic Review 35, 16-27.
[39] Skreta, V., Veldkamp, L., 2009. Ratings shopping and asset complexity: A
theory of ratings ination. Journal of Monetary Economics 56, 678-695.
[40] Smith, R., Staikouras, C., Wood, G., 2003. Non-interest income and total in-
come stability. Bank of England, Working paper series No 198.
[41] Stone, B.K., 1974. Systematic interest rate risk in a two-index model of returns.
Journal of Financial and Quantitative Analysis 9, 709-721.
[42] Trzcinka, C., 1986. On the number of factors in the arbitrage pricing model.
Journal of Finance 41, 347-368.
28
29

Figure 1. US zero-coupon yield curve level, slope and
curvature

This figure depicts time-series plots of the Nelson and Siegel
(1978) zero-coupon yield curve factors for the US over the
2001 to 2009 period. Shown are the estimates of the interest
rate yield curve level (
1
), slope (
2
) and curvature (
3
).

-10
0
10
2001 2002 2003 2004 2005 2006 2007 2008 2009
Level
Slope
Curvature
Mean St. Dev Min. Max.
5.178 0.647 2.915 6.436
2.678 1.943 -0.990 5.577
-2.665 2.141 -7.544 0.843

Table 1: Bank loan sales and securitization activities by year
This table presents the summary statistics of the US publically traded bank holding companies (BHCs)
securitization and assets sales activities by year. Reported are the average values of assets by category,
expressed as a proportion of BHCs total assets, securitized or sold within a given year, and the percentage
of BHCs (in italics) involved in issuance of new securitization and loan sales transactions in the same year.
The respective data are compiled from Schedule HC-S of the Federal Reserve Systems FY-9C lings for a
sample of 304 nancial intermediaries analysed in this study.
2001 2002 2003 2004 2005 2006 2007 2008
Loan sales
1-4 family residential 0.0424 0.0226 0.0129 0.0164 0.0146 0.0075 0.0062 0.0087
13.36% 11.30% 11.82% 12.58% 12.58% 12.58% 13.25% 13.71%
Home equity lines 0.0000 0.0000 0.0000 0.0000 0.0001 0.0040 0.0024 0.0002
0.00% 0.00% 0.00% 0.00% 0.33% 0.99% 0.66% 0.67%
Credit card receivables 0.0005 0.0000 0.0003 0.0004 0.0000 0.0001 0.0000 0.0006
2.53% 1.37% 2.36% 2.65% 1.99% 1.99% 2.32% 2.34%
Auto loans 0.0000 0.0001 0.0000 0.0000 0.0000 0.0014 0.0000 0.0152
0.00% 0.34% 0.00% 0.33% 0.00% 0.33% 0.00% 0.33%
Other consumer loans 0.0044 0.0091 0.0003 0.0000 0.0000 0.0000 0.0000 0.0003
1.08% 0.34% 0.34% 0.00% 0.00% 0.00% 0.00% 0.33%
C&I loans 0.0126 0.0020 0.0028 0.0018 0.0051 0.0017 0.0107 0.0001
2.53% 2.40% 2.36% 2.65% 2.32% 1.99% 0.66% 0.33%
Other loans 0.0143 0.0019 0.0033 0.0089 0.0040 0.0100 0.0312 0.0168
0.72% 1.03% 1.69% 2.65% 1.66% 1.32% 1.99% 3.34%
Loan securitization
1-4 family residential 0.2218 0.0708 0.0641 0.0342 0.0412 0.0562 0.0325 0.0497
11.55% 8.22% 6.42% 6.29% 4.97% 4.30% 4.30% 5.35%
Home equity lines 0.0086 0.0021 0.0033 0.0093 0.0120 0.0037 0.0000 0.0033
1.81% 0.68% 1.01% 1.66% 1.32% 0.99% 0.33% 1.67%
Credit card receivables 0.0274 0.0049 0.0055 0.0147 0.0101 0.0160 0.0108 0.0060
1.44% 1.03% 0.68% 0.99% 0.99% 1.99% 1.32% 1.67%
Auto loans 0.0147 0.0133 0.0286 0.0119 0.0086 0.0126 0.0203 0.0118
5.42% 1.71% 2.03% 1.99% 0.99% 1.32% 0.33% 1.67%
Other consumer loans 0.0110 0.0031 0.0036 0.0041 0.0040 0.0074 0.0007 0.0011
2.89% 1.03% 0.68% 0.66% 1.32% 1.66% 0.99% 1.00%
C&I loans 0.0264 0.0111 0.0046 0.0036 0.0048 0.0043 0.0029 0.0006
3.61% 3.08% 2.36% 1.66% 1.32% 1.99% 1.32% 1.67%
Other loans 0.0128 0.0055 0.0138 0.0054 0.0094 0.0096 0.0069 0.0246
2.53% 2.05% 2.70% 2.98% 2.98% 3.97% 3.97% 2.01%
30
Table 2: Selected characteristics of bank holding companies
This table provides a comparison of selected nancial characteristics for securitisers and non-
securitisers over the 2001 to 2009 period. A bank holding company (BHC) is dened as securitiser
if it reports at least one securitisation transaction over the analysed period in Schedule HC-S
of the Federal Reserve Systems FY-9C lings. Reported are the mean [median] values of the
considered accounting variables. This includes an institutions asset growth rate (AGR); equity
capital (CAP) calculated as the ratio of BHCs book value of equity capital to its total assets; the
Herndahl-Hirschman (non)interest revenue concentration index H_NITR(H_NOIR) calculated
on the basis of twelve (eight) part breakdown of the (non)interest income; the proportion of total
assets that are liquid (LATA); the Herndahl-Hirschman loan concentration index (H_LOAN)
computed considering ve loan categories; the banks provision for loan and lease losses scaled by
total loans (LLP); maturity gap (GAP) calculated as the dierence between interest-earning assets
and interest-bearing liabilities maturing or being repriced within one year, scaled by the banks
total assets; the net credit protection (protection bought minus sold) NECP purchased by a bank;
the ratio of non-performing loans to total loans is NPL; return on assets (ROA); the measure of
bank revenue diversication (ROID); and the ratio of the institutions risk-weighted to total assets
(TRA). The economic environment is proxied by the annual growth rate in the gross domestic
product (GDPG), and the Aruoba-Diebold-Scotti Business Conditions Index (ADSI). ***, **, and
* denote statistical signicance at the 1%, 5%, and 10% level respectively for an appropriate mean
[median] equality test.
Variable Securitizers Non-securitizers All BHCs Equality test
mean/[median] mean/[median] mean/[median] mean/[median]
Panel A: BHC nancial characteristics
Asset growth rate 0.101 0.126 0.121 1.04
AGR [0.077] [0.091] [0.088] [3.03***]
Capitalisation 0.098 0.091 0.093 -3.20***
CAP [0.088] [0.088] [0.088] [0.34]
Interest income HHI 0.076 0.064 0.067 -1.96*
H_NITR [0.019] [0.016] [0.017] [4.54***]
Liquidity 0.264 0.261 0.262 -0.45
LATA [0.242] [0.238] [0.239] [0.37]
Loan HHI 0.530 0.608 0.590 10.51***
H_LOAN [0.530] [0.601] [0.582] [10.34***]
Loan loss provision 0.006 0.004 0.005 -5.66***
LLP [0.004] [0.003] [0.003] [6.34***]
Maturity gap 0.177 0.160 0.164 -2.56**
GAP [0.141] [0.130] [0.132] [2.09**]
Net credit protection 6.54E-04 1.38E-05 1.61E-04 -3.66***
NECP [0.000] [0.000] [0.000] [0.48]
Non-interest income HHI 0.177 0.213 0.205 5.02***
H_NOIR [0.142] [0.191] [0.177] [6.72***]
Non-performing loans 0.012 0.010 0.010 -4.16***
NPL [0.008] [0.006] [0.006] [7.77***]
Return on assets 0.012 0.009 0.009 -4.78***
ROA [0.011] [0.010] [0.010] [3.59***]
Revenue Diversication 0.427 0.330 0.352 -10.56***
ROID [0.409] [0.300] [0.315] [9.94***]
Total risk adjusted assets 0.749 0.740 0.742 -1.37
TRA [0.758] [0.748] [0.750] [1.58]
Panel B: Economic environment characteristics
GDP growth 0.017 0.023 0.021 13.93***
GDPG [0.020] [0.025] [0.025] [14.21***]
Business conditions index -0.952 -0.095 -0.422 57.01***
ADSI [-1.077] [-0.130] [-0.155] [35.79***]
31
T
a
b
l
e
3
:
C
o
r
r
e
l
a
t
i
o
n
m
a
t
r
i
x
f
o
r
s
e
l
e
c
t
e
d
v
a
r
i
a
b
l
e
s
T
h
i
s
t
a
b
l
e
p
r
e
s
e
n
t
s
t
h
e
b
i
v
a
r
i
a
t
e
c
o
r
r
e
l
a
t
i
o
n
s
b
e
t
w
e
e
n
t
h
e
c
o
n
s
i
d
e
r
e
d
e
x
p
l
a
n
a
t
o
r
y
v
a
r
i
a
b
l
e
s
.
S
p
e
a
r
m
a
n
(
P
e
a
r
s
o
n
)
c
o
r
r
e
l
a
t
i
o
n
c
o
e

c
i
e
n
t
s
a
r
e
a
b
o
v
e
(
b
e
l
o
w
)
t
h
e
d
i
a
g
o
n
a
l
.
V
a
r
i
a
b
l
e
d
e

n
i
t
i
o
n
s
a
n
d
s
o
u
r
c
e
s
a
r
e
p
r
o
v
i
d
e
d
i
n
A
p
p
e
n
d
i
x
B
.
p

v
a
l
u
e
s
a
r
e
i
n
p
a
r
e
n
t
h
e
s
e
s
.
A
G
R
C
A
P
H
_
N
I
T
R
L
A
T
A
H
_
L
O
A
N
L
L
P
G
A
P
N
E
C
P
H
_
N
O
I
R
N
P
L
R
O
A
R
O
I
D
T
R
A
G
D
P
G
A
D
S
I
T
S
E
C
A
G
R
-
0
.
1
3
4
-
0
.
1
3
8
-
0
.
0
8
7
0
.
1
3
3
-
0
.
0
6
4
0
.
0
5
8
0
.
0
1
3
-
0
.
0
5
7
-
0
.
2
2
8
-
0
.
0
0
7
-
0
.
1
1
1
0
.
0
8
6
0
.
0
8
4
0
.
1
4
1
-
0
.
0
3
0
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
2
)
(
0
.
0
0
7
)
(
0
.
5
4
5
)
(
0
.
0
0
7
)
(
0
.
0
0
0
)
(
0
.
7
4
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
1
6
3
)
C
A
P
0
.
0
2
2
-
0
.
0
2
8
-
0
.
0
5
9
0
.
0
0
6
-
0
.
0
7
9
0
.
0
7
3
-
0
.
0
4
7
-
0
.
0
2
2
0
.
0
5
0
0
.
2
5
7
0
.
0
7
9
0
.
0
8
5
-
0
.
0
1
9
-
0
.
0
6
8
0
.
0
3
7
(
0
.
3
0
9
)
(
0
.
1
9
0
)
(
0
.
0
0
6
)
(
0
.
7
7
1
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
0
2
7
)
(
0
.
3
0
8
)
(
0
.
0
1
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
3
6
6
)
(
0
.
0
0
1
)
(
0
.
0
8
3
)
H
_
N
I
T
R
0
.
1
0
0
0
.
0
4
6
0
.
5
9
3
-
0
.
0
9
2
0
.
1
5
0
-
0
.
1
2
0
0
.
0
4
5
0
.
0
6
9
0
.
1
7
0
-
0
.
1
9
4
0
.
0
7
0
-
0
.
5
0
5
-
0
.
3
4
5
-
0
.
1
3
5
0
.
1
2
1
(
0
.
0
0
0
)
(
0
.
0
3
2
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
3
3
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
L
A
T
A
0
.
0
9
1
0
.
0
3
7
0
.
1
3
5
-
0
.
1
5
6
-
0
.
1
9
8
-
0
.
1
3
5
0
.
0
4
4
0
.
0
7
4
-
0
.
1
2
6
0
.
0
8
7
0
.
0
9
3
-
0
.
8
0
0
0
.
0
5
2
0
.
2
2
0
0
.
0
3
7
(
0
.
0
0
0
)
(
0
.
0
7
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
4
0
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
1
4
)
(
0
.
0
0
0
)
(
0
.
0
8
3
)
H
_
L
O
A
N
0
.
0
2
9
-
0
.
0
3
8
0
.
1
0
6
-
0
.
0
4
6
-
0
.
2
4
4
-
0
.
1
4
1
-
0
.
0
2
2
0
.
0
8
4
-
0
.
1
5
4
-
0
.
1
5
2
-
0
.
3
4
6
-
0
.
0
0
3
0
.
0
4
7
-
0
.
0
5
1
-
0
.
2
5
8
(
0
.
1
6
7
)
(
0
.
0
7
1
)
(
0
.
0
0
0
)
(
0
.
0
2
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
3
0
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
8
9
0
)
(
0
.
0
2
7
)
(
0
.
0
1
5
)
(
0
.
0
0
0
)
L
L
P
-
0
.
0
5
3
0
.
0
5
5
0
.
4
1
6
-
0
.
1
6
0
-
0
.
1
4
6
0
.
0
2
3
-
0
.
0
4
4
-
0
.
0
3
5
0
.
5
5
2
-
0
.
2
7
6
0
.
0
3
6
0
.
1
9
5
-
0
.
4
2
0
-
0
.
1
9
4
0
.
1
7
0
(
0
.
0
1
2
)
(
0
.
0
0
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
2
7
6
)
(
0
.
0
3
9
)
(
0
.
0
9
8
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
9
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
G
A
P
-
0
.
0
0
4
0
.
1
6
1
-
0
.
0
5
9
-
0
.
0
9
6
-
0
.
1
3
1
-
0
.
0
0
4
-
0
.
0
1
4
-
0
.
1
1
5
-
0
.
0
7
0
0
.
1
2
1
0
.
1
4
9
0
.
2
3
8
0
.
0
6
4
0
.
0
7
8
0
.
0
8
4
(
0
.
8
4
0
)
(
0
.
0
0
0
)
(
0
.
0
0
6
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
8
3
8
)
(
0
.
5
1
4
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
3
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
N
E
C
P
-
0
.
0
0
1
-
0
.
0
2
1
0
.
0
2
6
0
.
0
1
8
-
0
.
0
8
6
0
.
0
8
7
-
0
.
0
3
1
0
.
0
4
5
-
0
.
0
3
7
-
0
.
0
0
2
0
.
0
6
7
-
0
.
0
4
8
-
0
.
0
0
3
-
0
.
0
2
9
0
.
0
5
2
(
0
.
9
6
4
)
(
0
.
3
2
0
)
(
0
.
2
1
6
)
(
0
.
3
9
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
1
4
7
)
(
0
.
0
3
2
)
(
0
.
0
8
1
)
(
0
.
9
1
1
)
(
0
.
0
0
2
)
(
0
.
0
2
5
)
(
0
.
8
7
4
)
(
0
.
1
7
7
)
(
0
.
0
1
5
)
H
_
N
O
I
R
-
0
.
0
1
2
-
0
.
0
1
0
0
.
0
2
1
0
.
0
1
1
-
0
.
0
4
1
0
.
1
3
3
-
0
.
0
2
6
0
.
4
5
1
-
0
.
0
0
6
-
0
.
0
8
7
-
0
.
1
8
6
-
0
.
1
6
0
-
0
.
0
1
3
-
0
.
0
6
9
-
0
.
2
1
0
(
0
.
5
7
5
)
(
0
.
6
3
6
)
(
0
.
3
1
3
)
(
0
.
6
1
6
)
(
0
.
0
5
2
)
(
0
.
0
0
0
)
(
0
.
2
2
9
)
(
0
.
0
0
0
)
(
0
.
7
6
7
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
5
5
0
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
N
P
L
-
0
.
0
7
9
-
0
.
0
4
3
0
.
4
4
5
-
0
.
0
8
2
-
0
.
0
0
8
0
.
6
0
3
-
0
.
0
7
8
0
.
0
3
1
0
.
0
5
7
-
0
.
2
7
4
0
.
0
0
7
0
.
1
0
0
-
0
.
3
4
1
-
0
.
2
4
2
0
.
1
6
1
(
0
.
0
0
0
)
(
0
.
0
4
3
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
7
1
8
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
1
4
0
)
(
0
.
0
0
7
)
(
0
.
0
0
0
)
(
0
.
7
5
1
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
R
O
A
0
.
0
0
6
0
.
6
0
9
-
0
.
2
5
3
0
.
1
3
2
-
0
.
1
0
2
-
0
.
3
5
5
0
.
1
4
7
-
0
.
0
2
2
-
0
.
0
4
8
-
0
.
3
3
4
0
.
2
4
1
-
0
.
0
0
4
0
.
2
2
9
0
.
2
5
3
0
.
0
8
9
(
0
.
7
6
4
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
3
0
7
)
(
0
.
0
2
2
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
8
5
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
R
O
I
D
0
.
0
0
3
-
0
.
0
5
0
-
0
.
0
5
2
0
.
0
8
6
-
0
.
3
4
1
0
.
0
1
0
0
.
1
5
0
0
.
0
5
5
-
0
.
0
4
0
-
0
.
0
8
3
0
.
1
0
5
-
0
.
0
6
1
0
.
0
8
8
0
.
1
8
3
0
.
2
1
9
(
0
.
8
7
1
)
(
0
.
0
1
7
)
(
0
.
0
1
4
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
6
5
3
)
(
0
.
0
0
0
)
(
0
.
0
0
9
)
(
0
.
0
5
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
4
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
T
R
A
-
0
.
0
7
4
-
0
.
0
4
7
-
0
.
1
1
8
-
0
.
8
1
3
-
0
.
0
9
2
0
.
1
3
6
0
.
2
0
5
-
0
.
0
4
2
-
0
.
0
4
2
0
.
0
7
5
-
0
.
0
9
8
-
0
.
0
1
7
0
.
0
0
5
-
0
.
1
5
1
0
.
0
2
8
(
0
.
0
0
1
)
(
0
.
0
2
5
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
4
9
)
(
0
.
0
4
9
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
4
2
1
)
(
0
.
8
0
4
)
(
0
.
0
0
0
)
(
0
.
1
8
7
)
G
D
P
G
0
.
0
2
2
-
0
.
0
0
8
-
0
.
5
0
6
0
.
0
6
9
0
.
0
3
7
-
0
.
3
7
9
0
.
0
6
7
-
0
.
0
1
6
-
0
.
0
3
6
-
0
.
3
3
6
0
.
2
0
3
0
.
0
9
3
-
0
.
0
0
5
0
.
6
4
4
-
0
.
0
2
3
(
0
.
2
9
5
)
(
0
.
7
1
4
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
0
7
8
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
4
4
4
)
(
0
.
0
9
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
8
1
8
)
(
0
.
0
0
0
)
(
0
.
2
7
4
)
A
D
S
I
0
.
0
5
2
-
0
.
0
4
2
-
0
.
3
3
7
0
.
2
1
1
-
0
.
0
8
0
-
0
.
2
4
4
0
.
0
6
0
-
0
.
0
6
5
-
0
.
0
2
7
-
0
.
2
7
3
0
.
1
8
2
0
.
1
1
7
-
0
.
1
7
7
0
.
4
7
9
0
.
0
0
9
(
0
.
0
1
4
)
(
0
.
0
4
6
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
4
)
(
0
.
0
0
2
)
(
0
.
1
9
5
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
0
0
0
)
(
0
.
6
6
4
)
T
S
E
C
-
0
.
0
0
4
0
.
0
3
9
0
.
0
1
3
-
0
.
0
1
3
-
0
.
0
7
3
0
.
1
0
6
0
.
0
5
7
0
.
0
8
6
0
.
0
6
9
0
.
1
0
3
0
.
0
5
1
0
.
2
1
2
0
.
0
0
1
-
0
.
0
2
9
0
.
0
1
1
(
0
.
8
6
7
)
(
0
.
0
6
7
)
(
0
.
5
2
5
)
(
0
.
5
3
6
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
0
8
)
(
0
.
0
0
0
)
(
0
.
0
0
1
)
(
0
.
0
0
0
)
(
0
.
0
1
6
)
(
0
.
0
0
0
)
(
0
.
9
5
9
)
(
0
.
1
7
0
)
(
0
.
6
0
6
)
32
Table 4: Selected BHCs market measures of risk
This table provides a comparison of selected measures of market risk for securitisers and non-
securitisers over the 2001 to 2009 period. A bank holding company (BHC) is dened as securitiser
if it reports at least one securitisation transaction over the analysed period in Schedule HC-S of
the Federal Reserve Systems FY-9C lings. Reported are the mean [median] values of pertinent
risk measures. The market measures of risk are represented by the coecient estimates from a
four factor GARCH market model. Specically, for each bank-year, I run a four-factor time series
regression of BHC weekly returns on the market returns (MRK), and unanticipated changes in
zero-coupon yield curve level (LEV), slope (SLO), and curvature (CUR). The estimation requires
at least 30 weekly return observations for each bank-year. The corresponding US zero-coupon
yield curve level, slope, and curvature are estimated using Diebold and Lee (2006) parameterisa-
tion of the Nelson and Siegel (1987) model. The unanticipated changes in the yield curve factors
at time tare calculated as the dierence between the actual changes in these factors and ones
forecasted via an appropriate specication of the autoregressive moving average (ARMA) model.
The percentage of coecients signicant at the 5% level (% of which is negative) is in italics.
***, **, and * denote statistical signicance at the 1%, 5%, and 10% level respectively for an
appropriate mean [median] equality test.
Variable Securitizers Non-securitizers All BHCs Equality test
mean/[median] mean/[median] mean/[median] mean/[median]
Systematic risk 9.72E-01 6.62E-01 7.32E-01 -8.78***
[9.04E-01] [5.78E-01] [6.63E-01] [8.99***]
Signicance at 5% level 75.00% 49.62% 55.35%
IR Level 7.14E-04 1.16E-03 1.06E-03 0.15
[2.66E-03] [2.24E-04] [7.29E-04] [0.58]
Signicance at 5% level 9.51% 7.30% 7.80%
% negative -47.06% -46.27% -46.49%
IR Slope -9.94E-03 -8.57E-03 -8.88E-03 0.35
[-9.88E-03] [-9.44E-03] [-9.61E-03] [0.02]
Signicance at 5% level 11.94% 9.64% 10.16%
% negative -68.75% -72.88% -71.78%
IR Curvature 6.27E-05 -8.60E-04 -6.52E-04 -0.86
[3.84E-04] [-4.95E-07] [-6.78E-08] [0.77]
Signicance at 5% level 13.99% 7.24% 8.77%
% negative -61.33% -67.67% -65.39%
33
34

Table 5: Interest rate risk and securitisation by maturity category

This table presents the panel estimation results for the regression which evaluates bank holding
companies (BHC) interest rate risk with respect to the maturity of securitised assets over the 2001 to
2009 period. The dependent variable is the absolute value of the coefficient measuring the sensitivity of
BHCs i equity returns to unanticipated changes in the level (Panel A), slope (Panel B), and curvature
(Panel C) of the US sovereign zero-coupon yield curve at year t. The explanatory variables are as
follows: TSEC is the outstanding principle balance of assets securitised or sold measured as the
proportion of total assets; the outstanding balance of securitised long-, medium-, and short-term loans
are LT_SEC, MT_SEC, and ST_SEC respectively; the asset growth rate (AGR); equity capital (CAP)
calculated as the ratio of BHCs book value of equity capital to its total assets; H_NITR(H_NOIR) is the
Herfindahl-Hirschman (non)interest revenue concentration index calculated on the basis of twelve (eight)
part breakdown of the (non)interest income; the proportion of total assets that are liquid (LATA);
H_LOAN is the Herfindahl-Hirschman loan concentration index computed considering five loan
categories; GAP is the balance sheet maturity gap calculated as the difference between interest-earning
assets and interest-bearing liabilities maturing or being repriced within one year, scaled by the banks
total assets; NECP is the net credit protection (protection bought minus sold) purchased by a bank; the
ratio of non-performing loans to total loans is NPL; ROID is the measure of bank revenue diversification;
return on assets is represented by ROA. The regression also includes year- and state-dummies (not
reported). Heteroskedasticity and autocorrelation consistent t-values based on Whites robust standard
error are in italics. ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively.
Panel A: Yield curve level exposure

(1) (2) (3) (4) (5) (6) (7) (8)

Securitisers All bank holding companies
TSEC 0.053

0.025


3.61***

2.59***

LT_SEC

0.053

0.028



3.52***

2.76***

MT_SEC

-0.015

-0.091



-0.04

-0.33

ST_SEC

0.196

-0.026


1.44

-0.49


AGR -0.028 -0.028 -0.026 -0.030 0.004 0.004 0.004 0.004

-2.06** -2.00** -1.79* -2.04** 1.95* 1.96** 1.98** 1.97**
CAP -0.197 -0.199 -0.100 -0.081 -0.180 -0.179 -0.177 -0.176

-1.02 -1.02 -0.51 -0.41 -2.89*** -2.88*** -2.84*** -2.82***
CAP
2
0.564 0.570 0.440 0.375 0.292 0.295 0.292 0.296

2.14** 2.14** 1.65 1.38 3.14*** 3.17*** 3.14*** 3.17***
H_NITR
-0.060 -0.059 -0.072 -0.074 -0.039 -0.038 -0.041 -0.040

-1.75* -1.70* -2.06** -2.14** -2.29** -2.25** -2.38** -2.35**
LATA
0.063 0.061 0.062 0.068 0.010 0.010 0.010 0.009

2.19** 2.12** 2.13** 2.31** 1.08 1.05 1.02 0.99
H_LOAN
0.050 0.048 0.064 0.071 0.030 0.030 0.031 0.031

2.30** 2.18** 2.93*** 3.18*** 4.09*** 4.00*** 4.16*** 4.08***
GAP 0.004 0.004 0.003 0.002 0.015 0.014 0.015 0.015

0.20 0.19 0.14 0.11 1.97** 1.95* 2.04** 2.02**
NECP 0.476 0.484 0.552 0.530 0.819 0.825 0.875 0.883

1.42 1.44 1.62 1.56 3.19*** 3.21*** 3.41*** 3.43***
H_NOIR 0.000 0.000 0.000 0.000 0.005 0.006 0.005 0.005

1.37 1.41 1.67* 1.50 0.73 0.78 0.67 0.69
NPL
0.702 0.718 0.810 0.751 0.796 0.795 0.817 0.819

3.21*** 3.28*** 3.68*** 3.36*** 8.56*** 8.55*** 8.81*** 8.82***
ROA
-1.574 -1.551 -1.506 -1.472 -0.411 -0.410 -0.401 -0.403

-4.55*** -4.46*** -4.27*** -4.18*** -3.23*** -3.22*** -3.15*** -3.16***
ROID
-0.008 -0.008 0.005 -0.001 -0.010 -0.010 -0.007 -0.006

-0.54 -0.49 0.32 -0.03 -1.68* -1.66* -1.13 -1.03
Constant 0.027 0.028 0.007 0.003 0.006 0.006 0.005 0.004

0.93 0.95 0.24 0.10 0.33 0.34 0.24 0.24


Observations 516 516 516 516 2225 2225 2225 2225
BHCs
68 68 68 68 304 304 304 304
Period fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
State fixed effect
Yes Yes Yes Yes Yes Yes Yes Yes
Adj. R
2

0.23 0.23 0.21 0.22 0.14 0.14 0.14 0.14
35


Panel B: Yield curve slope exposure

(1) (2) (3) (4) (5) (6) (7) (8)

Securitisers All bank holding companies
TSEC
0.083

0.053


3.98***

4.00***

LT_SEC

0.085

0.056



3.98***

4.12***

MT_SEC

0.057

-0.422



0.11

-1.12

ST_SEC

0.223

0.004


1.14

0.05


AGR -0.037 -0.036 -0.033 -0.037 0.000 0.000 0.000 0.000

-1.90* -1.81* -1.59 -1.76* -0.16 -0.14 -0.09 -0.12
CAP
-0.151 -0.148 0.023 0.044 -0.011 -0.009 -0.003 -0.005

-0.55 -0.54 0.08 0.16 -0.13 -0.10 -0.04 -0.06
CAP
2

0.376 0.374 0.150 0.077 0.057 0.062 0.056 0.057

1.00 0.99 0.39 0.20 0.45 0.50 0.44 0.45
H_NITR
0.022 0.026 0.005 0.001 0.000 0.001 -0.004 -0.004

0.46 0.53 0.09 0.03 -0.01 0.04 -0.16 -0.15
LATA
0.087 0.085 0.088 0.094 0.017 0.016 0.015 0.016

2.14** 2.09** 2.11** 2.25** 1.32 1.27 1.20 1.24
H_LOAN 0.034 0.029 0.054 0.061 0.017 0.015 0.018 0.018

1.10 0.92 1.73* 1.93* 1.67* 1.54 1.74* 1.79*
GAP -0.010 -0.009 -0.011 -0.012 -0.001 -0.001 0.000 0.000

-0.35 -0.33 -0.40 -0.43 -0.12 -0.12 0.01 0.00
NECP
0.324 0.334 0.443 0.419 0.061 0.078 0.184 0.175

0.68 0.70 0.91 0.86 0.18 0.23 0.53 0.50
H_NOIR
0.000 0.000 0.000 0.000 -0.007 -0.006 -0.008 -0.007

-1.81* -1.79* -1.45 -1.57 -0.67 -0.60 -0.79 -0.75
NPL
0.345 0.362 0.513 0.445 0.490 0.489 0.535 0.535

1.11 1.16 1.63 1.39 3.89*** 3.88*** 4.25*** 4.24***
ROA
-1.516 -1.480 -1.402 -1.365 -0.472 -0.473 -0.457 -0.454

-3.07*** -2.99*** -2.79*** -2.71*** -2.74*** -2.75*** -2.64*** -2.62***
ROID -0.031 -0.033 -0.013 -0.019 -0.014 -0.014 -0.007 -0.007

-1.42 -1.51 -0.61 -0.85 -1.69* -1.67* -0.84 -0.88
Constant 0.037 0.038 0.005 0.000 0.047 0.047 0.044 0.043

0.89 0.93 0.11 0.01 1.84* 1.85* 1.71* 1.69*


Observations
516 516 516 516 2225 2225 2225 2225
BHCs 68 68 68 68 304 304 304 304
Period fixed effect
Yes Yes Yes Yes Yes Yes Yes Yes
State fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Adj. R
2

0.18 0.19 0.16 0.16 0.10 0.10 0.10 0.10

36


Panel C: Yield curve curvature exposure

(1) (2) (3) (4) (5) (6) (7) (8)

Securitisers All bank holding companies
TSEC
0.013

0.008


2.61***

2.50**

LT_SEC

0.013

0.008



2.58**

2.19**

MT_SEC

0.028

0.006



0.23

0.06

ST_SEC

0.076

0.034


1.68*

1.84*


AGR -0.002 -0.002 -0.002 -0.003 0.003 0.003 0.003 0.003

-0.55 -0.51 -0.41 -0.68 4.52*** 4.51*** 4.51*** 4.51***
CAP
-0.089 -0.086 -0.056 -0.049 -0.021 -0.021 -0.020 -0.022

-1.38 -1.33 -0.87 -0.76 -0.97 -0.98 -0.95 -1.05
CAP
2

0.174 0.168 0.129 0.104 0.040 0.041 0.040 0.036

1.97** 1.90* 1.46 1.17 1.25 1.28 1.25 1.13
H_NITR
-0.023 -0.022 -0.026 -0.027 -0.013 -0.013 -0.014 -0.014

-2.00** -1.93* -2.22** -2.33** -2.29** -2.27** -2.37** -2.45**
LATA
0.013 0.013 0.014 0.016 -0.003 -0.003 -0.003 -0.002

1.39 1.37 1.41 1.63 -0.85 -0.88 -0.88 -0.74
H_LOAN 0.002 0.001 0.004 0.007 0.000 0.000 0.001 0.001

0.24 0.11 0.60 0.94 0.15 0.10 0.24 0.47
GAP -0.002 -0.002 -0.003 -0.003 0.002 0.002 0.002 0.002

-0.38 -0.33 -0.41 -0.44 0.63 0.70 0.75 0.81
NECP
0.079 0.081 0.097 0.090 -0.021 -0.016 -0.003 -0.016

0.70 0.72 0.87 0.80 -0.24 -0.18 -0.03 -0.18
H_NOIR
0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000

-2.07** -2.07** -1.87* -2.04** 0.12 0.20 0.12 0.09
NPL
-0.021 -0.020 0.004 -0.019 0.128 0.128 0.134 0.132

-0.29 -0.27 0.05 -0.26 4.01*** 4.02*** 4.23*** 4.14***
ROA
-0.344 -0.332 -0.318 -0.306 -0.102 -0.100 -0.098 -0.094

-2.98*** -2.88*** -2.74*** -2.64*** -2.33** -2.30** -2.23** -2.15**
ROID -0.011 -0.012 -0.009 -0.011 -0.007 -0.007 -0.006 -0.006

-2.13** -2.30** -1.75* -2.09** -3.18*** -3.15*** -2.78*** -3.09***
Constant 0.031 0.031 0.026 0.025 0.018 0.018 0.018 0.018

3.25*** 3.26*** 2.75*** 2.60*** 2.80*** 2.79*** 2.71*** 2.72***


Observations
516 516 516 516 2225 2225 2225 2225
BHCs 68 68 68 68 304 304 304 304
Period fixed effect
Yes Yes Yes Yes Yes Yes Yes Yes
State fixed effect Yes Yes Yes Yes Yes Yes Yes Yes
Adj. R
2

0.15 0.16 0.14 0.15 0.11 0.11 0.11 0.11

37

Table 6: Nonlinearity between interest rate risk and securitization

This table presents the panel estimation results for the regression which evaluates the bank holding companies (BHC) interest rate risk with respect to the maturity of
securitized assets over the 2001 to 2009 period. The dependent variable is the absolute value of the coefficient measuring the sensitivity of BHCs i equity returns to
unanticipated changes in the level (columns 1-5), slope (columns 6-10), and curvature (columns 11-15) of the US sovereign zero-coupon yield curve at year t. Only
BHCs reporting at least one securitization transaction over the analysed period in Schedule HC-S of the Federal Reserve Systems FY-9C filings are considered. The
explanatory variables on the right-hand side are as follows: TSEC is the outstanding principle balance of assets securitized or sold measured as the proportion of total
assets; the outstanding balance of securitized long-, medium-, and short-term loans are LT_SEC, MT_SEC, and ST_SEC respectively; the ratio (and the squared ratio)
of book value of equity capital to banks total assets CAP. Each regression also includes year- and state- dummies, and the following firm-specific variables which are
not reported: the asset growth rate (AGR); the proportion of total assets that are liquid (LATA); H_LOAN is the Herfindahl-Hirschman loan concentration index
computed considering five loan categories; NECP is the net credit protection (protection bought minus sold) purchased by a bank; the ratio of non-performing loans to
total loans is NPL; ROID is the measure of bank revenue diversification; and return on assets is represented by ROA. The regressions in columns 2, 7, and 12 also
incorporate the economic environment proxies (not reported) as follows: annual growth rate in the gross domestic product (GDPG), and the Aruoba-Diebold-Scotti
Business Conditions Index (ADSI). When the economic environment proxies are added, the time-fixed effect is relaxed. Heteroskedasticity and autocorrelation
consistent t-values based on Whites robust standard error are in italics. ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)

Level Slope Curvature
TSEC 0.102 0.093

0.112 0.098

0.014 0.013


3.95*** 3.47***

3.06*** 2.66***

1.65* 1.43

TSEC
2
-0.028 -0.027

-0.026 -0.026

-0.003 -0.003


-1.80* -1.65*

-1.15 -1.12

-0.52 -0.60

LT_SEC

0.103

0.115

0.014



3.79***

3.02***

1.61

LT_SEC
2

-0.028

-0.027

-0.003



-1.74*

-1.16

-0.51

MT_SEC

0.610

0.748

0.365



0.71

0.62

1.30

MT_SEC
2

-11.013

-14.521

-6.705



-0.71

-0.67

-1.34

ST_SEC

0.342

0.352

0.026


2.08**

1.53

0.48
ST_SEC
2

-0.269

-0.713

0.070


-0.55

-1.04

0.44
CAP -0.271 -0.197 -0.274 -0.108 -0.094 -0.187 -0.008 -0.190 0.013 0.030 -0.122 -0.089 -0.121 -0.094 -0.089

-1.40 -0.99 -1.40 -0.55 -0.48 -0.68 -0.03 -0.69 0.05 0.11 -1.89* -1.36 -1.87* -1.47 -1.40
CAP
2
0.726 0.678 0.741 0.502 0.429 0.490 0.466 0.506 0.213 0.140 0.232 0.222 0.231 0.190 0.180

2.79*** 2.52** 2.82*** 1.91* 1.65* 1.33 1.25 1.37 0.58 0.38 2.68*** 2.52** 2.66*** 2.23** 2.09**


Observations 516 516 516 516 516 516 516 516 516 516 516 516 516 516 516
BHCs 68 68 68 68 68 68 68 68 68 68 68 68 68 68 68
Period fixed effect Yes No Yes Yes Yes Yes No Yes Yes Yes Yes No Yes Yes Yes
State fixed effect Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Adj. R
2
0.22 0.16 0.22 0.18 0.19 0.18 0.15 0.18 0.16 0.16 0.14 0.10 0.14 0.13 0.13

38

Table 7: Robustness test (yield curve level)

This table presents the panel estimation results for the regressions which evaluate the bank holding
companies' (BHC) interest rate risk with respect to securitisation, using different time horizons (column
Crisis); subsamples (columns SIZE, LATA, NDUS); and the model econometric specifications
(column 2SLS). The dependent variable is the absolute value of the coefficient measuring the
sensitivity of BHC is equity returns to unanticipated changes in the level of the US sovereign zero-
coupon yield curve at year t. These coefficients are estimated from a four factor GARCH market model.
Specifically, for each bank-year, I run a four-factor time series regression of BHC weekly returns on the
market returns (MRK), and unanticipated changes in yield curve level (LEV), slope (SLO), and
curvature (CUR). The estimation requires at least 30 weekly return observations for each bank-year.
The corresponding US zero-coupon yield curve level, slope, and curvature are estimated using Diebold
and Lee (2006) parameterisation of Nelson and Siegel (1987) model. The unanticipated changes in the
yield curve factors at time t are calculated as the difference between the actual changes in these factors
and ones forecasted via an appropriate specification of the autoregressive moving average (ARMA)
model. Only BHCs reporting at least one securitisation transaction over the analysed period in Schedule
HC-S of the Federal Reserve Systems FY-9C filings are considered. Reported are the coefficient
estimates for the TSEC explanatory variable, which represents the outstanding principle balance of
assets securitised or sold as the proportion of total assets. Each regression also includes year- and state-
dummies, and the following firm-specific variables which are not reported: the ratio (and the squared
ratio) of book value of equity capital to banks total assets (CAP); the asset growth rate (AGR); the
proportion of total assets that are liquid (LATA); the Herfindahl-Hirschman loan concentration index
computed considering five loan categories (H_LOAN); the net credit protection (protection bought minus
sold) purchased by a bank (NECP); the ratio of non-performing loans to total loans (NPL); the measure
of bank revenue diversification (ROID); and return on assets (ROA). All BHCs are split into a number
of sub-samples on the basis of ranking by the banks size (column SIZE); liquidity (column LATA);
and net derivative usage (column NDUS). Selected are the top 25% and the bottom 75% of values in
each category with a total of six portfolios. Coefficients on TSEC are reported for each portfolio. The
test statistics (F-statistics) for the Wald coefficient restriction test with the null hypothesis testing the
equality of the coefficient estimates for the Top 25% and the Bottom 75% portfolios in each category
is reported in column entitled WALD, with the associated p-value reported in brackets below.
Heteroskedasticity and autocorrelation consistent t-values based on Whites robust standard error are
reported in italics. ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively.

2SLS Crisis SIZE LATA NDUS WALD
TSEC
0.192


3.77***

Pre-crisis (2001-2006)

0.035

67.50


2.59***

(0.000)
Crisis (2007-2009)

0.244



9.50***

Top 25%

0.023

3.73


0.94

(0.054)
Bottom 75%

0.084




4.74***

Top 25%


0.153

20.93



6.37***

(0.000)
Bottom 75%

0.018




1.04

Top 25%


0.007 36.79



0.40 (0.000)
Bottom 75%

0.177




7.63***


Observations
516 516 516 516 516

BHCs 68 68 68 68 68

Period fixed effect Yes No Yes Yes Yes

State fixed effect Yes Yes Yes Yes Yes

Adj. R
2
0.08 0.31 0.22 0.25 0.27


A
p
p
e
n
d
i
x
A
P
a
n
e
l
A
:
S
e
c
u
r
i
t
i
s
e
r
s
B
H
C
n
a
m
e
R
S
S
D
I
D
T
i
c
k
e
r
B
H
C
n
a
m
e
R
S
S
D
I
D
T
i
c
k
e
r
1
s
t
S
o
u
r
c
e
C
o
r
p
o
r
a
t
i
o
n
1
1
9
9
6
0
2
S
R
C
E
G
u
a
r
a
n
t
y
F
e
d
e
r
a
l
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
2
6
1
8
9
4
0
G
F
E
D
A
M
C
O
R
E
F
i
n
a
n
c
i
a
l
,
I
n
c
.
1
2
0
8
6
6
1
A
M
F
I
H
u
n
t
i
n
g
t
o
n
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
1
0
6
8
1
9
1
H
B
A
N
A
r
r
o
w
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
0
4
8
8
1
2
A
R
O
W
I
n
d
e
p
e
n
d
e
n
t
B
a
n
k
C
o
r
p
o
r
a
t
i
o
n
1
2
0
1
9
2
5
I
B
C
P
A
s
s
o
c
i
a
t
e
d
B
a
n
c
-
C
o
r
p
1
1
9
9
5
6
3
A
S
B
C
I
n
d
i
a
n
a
U
n
i
t
e
d
B
a
n
c
o
r
p
.
1
2
0
9
1
0
9
M
S
F
G
A
u
b
u
r
n
N
a
t
i
o
n
a
l
B
a
n
c
o
r
p
o
r
a
t
i
o
n
,
I
n
c
.
1
1
2
9
5
3
3
A
U
B
N
I
n
t
e
r
n
a
t
i
o
n
a
l
B
a
n
c
s
h
a
r
e
s
C
o
r
p
o
r
a
t
i
o
n
1
1
0
4
2
3
1
I
B
O
C
B
a
n
k
o
f
A
m
e
r
i
c
a
C
o
r
p
o
r
a
t
i
o
n
1
0
7
3
7
5
7
B
A
C
K
e
y
C
o
r
p
.
1
0
6
8
0
2
5
K
E
Y
B
a
n
k
o
f
N
e
w
Y
o
r
k
C
o
m
p
a
n
y
,
I
n
c
.
1
0
3
3
4
7
0
B
K
L
N
B
B
a
n
c
o
r
p
,
I
n
c
.
1
0
7
1
6
6
9
L
N
B
B
B
O
K
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
8
8
3
6
9
3
B
O
K
F
M
a
r
s
h
a
l
l
&
I
l
s
l
e
y
C
o
r
p
o
r
a
t
i
o
n
1
1
9
9
4
9
7
M
I
C
a
m
d
e
n
N
a
t
i
o
n
a
l
C
o
r
p
o
r
a
t
i
o
n
1
1
3
0
2
4
9
C
A
C
M
B
F
i
n
a
n
c
i
a
l
,
I
n
c
.
1
0
9
0
9
8
7
M
B
F
I
C
a
p
i
t
a
l
O
n
e
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
2
2
7
7
8
6
0
C
O
F
M
i
d
P
e
n
n
B
a
n
c
o
r
p
,
I
n
c
.
1
9
4
4
2
0
4
M
P
B
C
a
p
i
t
o
l
B
a
n
c
o
r
p
L
t
d
.
1
2
4
7
3
3
4
C
B
C
M
o
n
r
o
e
B
a
n
c
o
r
p
1
2
1
0
2
9
9
M
R
O
E
C
a
r
o
l
i
n
a
F
i
r
s
t
C
o
r
p
o
r
a
t
i
o
n
1
1
4
1
5
9
9
T
S
F
G
N
o
r
t
h
e
r
n
S
t
a
t
e
s
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
2
1
0
5
8
9
N
S
F
C
C
h
a
r
l
e
s
S
c
h
w
a
b
C
o
r
p
o
r
a
t
i
o
n
1
0
2
6
6
3
2
S
C
H
W
N
o
r
t
h
r
i
m
B
a
n
c
o
r
p
I
n
c
.
3
0
2
5
3
8
5
N
R
I
M
C
i
t
i
g
r
o
u
p
,
I
n
c
.
1
9
5
1
3
5
0
C
N
o
r
w
e
s
t
C
o
r
p
o
r
a
t
i
o
n
1
1
2
0
7
5
4
W
F
C
C
i
t
y
H
o
l
d
i
n
g
C
o
m
p
a
n
y
1
0
7
6
2
6
2
C
H
C
O
P
a
c
i

c
C
a
p
i
t
a
l
B
a
n
c
o
r
p
.
1
0
2
9
8
8
4
P
C
B
C
C
o
m
m
B
a
n
c
o
r
p
,
I
n
c
o
r
p
o
r
a
t
e
d
1
1
1
8
2
2
9
C
C
B
P
P
o
p
u
l
a
r
,
I
n
c
.
1
1
2
9
3
8
2
B
P
O
P
D
o
r
a
l
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
2
1
8
4
1
6
4
D
R
L
R
e
g
i
o
n
s
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
0
7
8
3
3
2
R
F
E
a
g
l
e
B
a
n
c
o
r
p
,
I
n
c
.
2
6
5
2
1
0
4
E
G
B
N
R
e
p
u
b
l
i
c
B
a
n
c
o
r
p
,
I
n
c
.
1
0
9
7
0
2
5
R
B
C
A
A
E
a
s
t
W
e
s
t
B
a
n
c
o
r
p
,
I
n
c
.
2
7
3
4
2
3
3
E
W
B
C
S
.
Y
.
B
a
n
c
o
r
p
,
I
n
c
.
1
2
4
9
7
3
0
S
Y
B
T
E
x
c
h
a
n
g
e
N
a
t
i
o
n
a
l
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
2
0
3
8
4
0
9
H
W
B
K
S
a
n
t
a
n
d
e
r
B
a
n
c
o
r
p
.
2
8
4
7
1
1
5
S
B
P
F
i
f
t
h
T
h
i
r
d
B
a
n
c
o
r
p
1
0
7
0
3
4
5
F
I
T
B
S
t
a
t
e
S
t
r
e
e
t
B
o
s
t
o
n
C
o
r
p
o
r
a
t
i
o
n
1
1
1
1
4
3
5
S
T
T
F
i
r
s
t
B
a
n
C
o
r
p
2
7
4
4
8
9
4
F
B
P
S
u
n
T
r
u
s
t
B
a
n
k
s
,
I
n
c
.
1
1
3
1
7
8
7
S
T
I
F
i
r
s
t
B
u
s
e
y
C
o
r
p
o
r
a
t
i
o
n
1
2
0
3
6
0
2
B
U
S
E
S
u
s
q
u
e
h
a
n
n
a
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
1
1
1
7
1
5
6
S
U
S
Q
F
i
r
s
t
C
i
t
i
z
e
n
s
B
a
n
c
S
h
a
r
e
s
,
I
n
c
.
1
0
7
5
6
1
2
F
C
N
C
A
T
h
e
C
h
a
s
e
M
a
n
h
a
t
t
a
n
B
a
n
k
1
0
3
9
5
0
2
J
P
M
F
i
r
s
t
E
m
p
i
r
e
S
t
a
t
e
C
o
r
p
o
r
a
t
i
o
n
1
0
3
7
0
0
3
M
T
B
T
h
e
P
N
C
F
i
n
a
n
c
i
a
l
S
e
r
v
i
c
e
s
G
r
o
u
p
,
I
n
c
.
1
0
6
9
7
7
8
P
N
C
F
i
r
s
t
F
i
n
a
n
c
i
a
l
B
a
n
c
o
r
p
.
1
0
7
1
2
7
6
F
F
B
C
T
I
B
F
i
n
a
n
c
i
a
l
C
o
r
p
.
2
4
5
7
9
4
3
T
I
B
B
F
i
r
s
t
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
2
0
8
5
9
5
T
H
F
F
U
.
S
.
B
a
n
c
o
r
p
.
1
1
1
9
7
9
4
U
S
B
F
i
r
s
t
H
o
r
i
z
o
n
N
a
t
i
o
n
a
l
C
o
r
p
o
r
a
t
i
o
n
1
0
9
4
6
4
0
F
H
N
U
n
i
o
n
B
a
n
c
o
r
p
,
I
n
c
.
1
2
0
6
5
9
1
T
R
U
E
F
i
r
s
t
M
i
d
w
e
s
t
B
a
n
c
o
r
p
,
I
n
c
.
1
2
0
8
1
8
4
F
M
B
I
U
n
i
t
e
d
B
a
n
c
o
r
p
,
I
n
c
.
1
0
7
1
5
0
2
U
B
C
P
F
i
r
s
t
M
e
r
i
t
C
o
r
p
o
r
a
t
i
o
n
1
0
7
0
8
0
4
F
M
E
R
U
n
i
t
e
d
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
1
1
3
6
0
0
9
U
B
O
H
F
N
B
C
o
r
p
.
1
1
3
3
4
7
3
F
N
B
N
U
n
i
t
e
d
B
a
n
k
s
h
a
r
e
s
,
I
n
c
.
1
0
7
6
2
1
7
U
B
S
I
F
r
a
n
k
l
i
n
R
e
s
o
u
r
c
e
s
,
I
n
c
.
1
2
4
6
2
1
6
B
E
N
W
H
o
l
d
i
n
g
C
o
m
p
a
n
y
I
n
c
o
r
p
o
r
a
t
e
d
2
8
0
1
5
4
6
W
H
I
F
u
l
t
o
n
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
1
1
1
7
1
2
9
F
U
L
T
W
i
n
t
r
u
s
t
F
i
n
a
n
c
i
a
l
C
o
r
p
o
r
a
t
i
o
n
2
2
6
0
4
0
6
W
T
F
C
G
e
r
m
a
n
A
m
e
r
i
c
a
n
B
a
n
c
o
r
p
1
0
9
8
6
2
0
G
A
B
C
Z
i
o
n
s
B
a
n
c
o
r
p
o
r
a
t
i
o
n
1
0
2
7
0
0
4
Z
I
O
N
39
A
p
p
e
n
d
i
x
A
P
a
n
e
l
B
:
N
o
n
-
S
e
c
u
r
i
t
i
s
e
r
s
B
H
C
n
a
m
e
R
S
S
D
I
D
T
i
c
k
e
r
B
H
C
n
a
m
e
R
S
S
D
I
D
T
i
c
k
e
r
1
s
t
C
o
n
s
t
i
t
u
t
i
o
n
B
a
n
c
o
r
p
.
2
7
8
4
9
2
0
F
C
C
Y
C
a
r
r
o
l
l
t
o
n
B
a
n
c
o
r
p
.
1
4
6
9
8
0
0
C
R
R
B
A
B
C
B
a
n
c
o
r
p
1
0
8
2
0
6
7
A
B
C
B
C
a
s
c
a
d
e
B
a
n
c
o
r
p
.
1
8
4
8
0
0
3
C
A
C
B
A
c
c
e
s
s
N
a
t
i
o
n
a
l
C
o
r
p
.
3
1
0
9
9
0
4
A
N
C
X
C
a
s
c
a
d
e
F
i
n
a
n
c
i
a
l
C
o
r
p
.
2
5
6
8
3
6
2
C
A
S
B
A
l
l
i
a
n
c
e
B
a
n
k
s
h
a
r
e
s
C
o
r
p
.
3
1
2
3
6
3
8
A
B
V
A
C
a
s
s
I
n
f
o
r
m
a
t
i
o
n
S
y
s
t
e
m
s
,
I
n
c
.
1
0
9
8
6
4
8
C
A
S
S
A
l
l
i
a
n
c
e
F
i
n
a
n
c
i
a
l
C
o
r
p
.
1
1
4
0
5
1
0
A
L
N
C
C
a
t
h
a
y
G
e
n
e
r
a
l
B
a
n
c
o
r
p
,
I
n
c
.
1
8
4
3
0
8
0
C
A
T
Y
A
m
e
r
i
c
a
n
N
a
t
i
o
n
a
l
B
a
n
k
s
h
a
r
e
s
I
n
c
.
1
0
7
6
6
9
1
A
M
N
B
C
e
n
t
e
r
B
a
n
c
o
r
p
,
I
n
c
.
1
0
4
8
7
6
4
C
N
B
C
A
m
e
r
i
c
a
n
R
i
v
e
r
B
a
n
k
s
h
a
r
e
s
2
3
1
2
8
3
7
A
M
R
B
C
e
n
t
e
r
F
i
n
a
n
c
i
a
l
C
o
r
p
.
3
0
0
3
1
7
8
C
L
F
C
A
m
e
r
i
S
e
r
v
F
i
n
a
n
c
i
a
l
,
I
n
c
.
1
1
1
7
3
1
6
A
S
R
V
C
e
n
t
e
r
s
t
a
t
e
B
a
n
k
s
o
f
F
l
o
r
i
d
a
,
I
n
c
.
2
8
6
8
1
2
9
C
S
F
L
A
n
n
a
p
o
l
i
s
B
a
n
c
o
r
p
,
I
n
c
.
1
4
7
2
2
5
7
A
N
N
B
C
e
n
t
r
a
l
V
i
r
g
i
n
i
a
B
a
n
k
s
h
a
r
e
s
,
I
n
c
.
1
1
4
0
6
7
7
C
V
B
K
B
a
n
c
C
o
r
p
.
2
7
3
1
8
5
8
S
U
P
R
C
e
n
t
u
r
y
B
a
n
c
o
r
p
,
I
n
c
.
1
1
1
1
0
8
8
C
N
B
K
A
B
a
n
c
F
i
r
s
t
C
o
r
p
.
1
1
3
3
2
8
6
B
A
N
F
C
h
e
m
i
c
a
l
F
i
n
a
n
c
i
a
l
C
o
r
p
.
1
2
0
1
9
3
4
C
H
F
C
B
a
n
c
o
r
p
R
h
o
d
e
I
s
l
a
n
d
,
I
n
c
.
2
8
9
6
4
5
8
B
A
R
I
C
i
t
i
z
e
n
s
B
a
n
k
i
n
g
C
o
r
p
.
1
2
0
5
6
8
8
C
R
B
C
B
a
n
c
o
r
p
S
o
u
t
h
,
I
n
c
.
1
0
9
7
6
1
4
B
X
S
C
i
t
i
z
e
n
s
H
o
l
d
i
n
g
C
o
m
p
a
n
y
1
0
8
3
4
7
5
C
I
Z
N
B
a
n
c
s
h
a
r
e
s
o
f
F
l
o
r
i
d
a
,
I
n
c
.
2
7
9
6
6
2
4
B
O
F
L
C
i
t
y
N
a
t
i
o
n
a
l
C
o
r
p
.
1
0
2
7
5
1
8
C
Y
N
B
a
n
k
o
f
C
o
m
m
e
r
c
e
H
o
l
d
i
n
g
s
1
0
3
0
0
4
0
B
O
C
H
C
N
B
F
i
n
a
n
c
i
a
l
C
o
r
p
.
1
1
1
8
3
4
0
C
C
N
E
B
a
n
k
o
f
G
r
a
n
i
t
e
C
o
r
p
.
1
1
4
3
4
8
1
G
R
A
N
C
o
B
i
z
I
n
c
.
1
0
6
0
3
2
8
C
O
B
Z
B
a
n
k
o
f
H
a
w
a
i
i
C
o
r
p
.
1
0
2
5
3
0
9
B
O
H
C
o
d
o
r
u
s
V
a
l
l
e
y
B
a
n
c
o
r
p
,
I
n
c
.
1
1
4
2
4
7
5
C
V
L
Y
B
a
n
k
o
f
S
o
u
t
h
C
a
r
o
l
i
n
a
C
o
r
p
.
2
2
9
7
7
0
1
B
K
S
C
C
o
l
o
n
y
B
a
n
k
c
o
r
p
,
I
n
c
.
1
0
8
5
1
7
0
C
B
A
N
B
a
n
k
o
f
t
h
e
O
z
a
r
k
s
,
I
n
c
.
1
0
9
7
0
8
9
O
Z
R
K
C
o
l
u
m
b
i
a
B
a
n
c
o
r
p
.
2
3
7
8
4
4
0
C
B
B
O
B
a
n
n
e
r
C
o
r
p
.
2
1
2
6
9
7
7
B
A
N
R
C
o
l
u
m
b
i
a
B
a
n
k
i
n
g
S
y
s
t
e
m
,
I
n
c
.
2
0
7
8
8
1
6
C
O
L
B
B
a
r
H
a
r
b
o
r
B
a
n
k
s
h
a
r
e
s
1
1
1
5
3
8
5
B
H
B
C
o
m
e
r
i
c
a
I
n
c
o
r
p
o
r
a
t
e
d
1
1
9
9
8
4
4
C
M
A
B
B
&
T
C
o
r
p
.
1
0
7
4
1
5
6
B
B
T
C
o
m
m
e
r
c
e
B
a
n
c
s
h
a
r
e
s
,
I
n
c
.
1
0
4
9
3
4
1
C
B
S
H
B
e
r
k
s
h
i
r
e
B
a
n
c
o
r
p
,
I
n
c
.
2
7
2
8
1
5
7
B
E
R
K
C
o
m
m
e
r
c
i
a
l
B
a
n
c
o
r
p
.
1
0
2
9
8
9
3
W
C
B
O
B
N
C
B
a
n
c
o
r
p
.
3
1
4
1
6
5
0
B
N
C
N
C
o
m
m
e
r
c
i
a
l
N
a
t
i
o
n
a
l
F
i
n
a
n
c
i
a
l
C
o
r
p
.
1
8
2
3
7
3
8
C
N
A
F
B
o
s
t
o
n
P
r
i
v
a
t
e
B
a
n
c
o
r
p
,
I
n
c
.
1
2
4
8
0
7
8
B
P
F
H
C
o
m
m
o
n
w
e
a
l
t
h
B
a
n
k
s
h
a
r
e
s
,
I
n
c
.
1
2
5
0
6
0
6
C
W
B
S
B
r
i
d
g
e
C
a
p
i
t
a
l
H
o
l
d
i
n
g
s
3
2
8
0
9
8
8
B
B
N
K
C
o
m
m
u
n
i
t
y
B
a
n
k
S
h
a
r
e
s
o
f
I
n
d
i
a
n
a
,
I
n
c
.
2
3
5
6
0
7
3
C
B
I
N
B
r
i
t
t
o
n
&
K
o
o
n
t
z
C
a
p
i
t
a
l
C
o
r
p
.
1
0
8
4
2
1
2
B
K
B
K
C
o
m
m
u
n
i
t
y
B
a
n
k
S
y
s
t
e
m
,
I
n
c
.
1
0
4
8
8
6
7
C
B
U
B
r
y
n
M
a
w
r
B
a
n
k
C
o
r
p
.
1
1
4
0
9
9
4
B
M
T
C
C
o
m
m
u
n
i
t
y
C
a
p
i
t
a
l
C
o
r
p
.
1
3
9
8
9
3
7
C
P
B
K
C
&
F
F
i
n
a
n
c
i
a
l
C
o
r
p
.
2
1
8
3
4
9
3
C
F
F
I
C
o
m
m
u
n
i
t
y
C
e
n
t
r
a
l
B
a
n
k
C
o
r
p
.
2
4
4
3
5
2
6
C
C
B
D
C
a
m
c
o
F
i
n
a
n
c
i
a
l
G
r
o
u
p
1
2
5
1
2
5
6
C
A
F
I
C
o
m
m
u
n
i
t
y
T
r
u
s
t
B
a
n
c
o
r
p
,
I
n
c
.
1
0
7
0
6
4
4
C
T
B
I
C
a
p
i
t
a
l
B
a
n
k
C
o
r
p
.
2
7
4
1
1
5
6
C
B
K
N
C
o
m
m
u
n
i
t
y
W
e
s
t
B
a
n
c
s
h
a
r
e
s
2
6
2
6
2
9
9
C
W
B
C
C
a
p
i
t
a
l
C
i
t
y
B
a
n
k
G
r
o
u
p
,
I
n
c
.
1
0
8
5
5
0
9
C
C
B
G
C
o
w
l
i
t
z
B
a
n
c
o
r
p
.
1
9
8
4
0
4
0
C
W
L
Z
C
a
r
d
i
n
a
l
F
i
n
a
n
c
i
a
l
C
o
r
p
.
2
6
8
2
9
9
6
C
F
N
L
C
P
B
I
n
c
.
1
0
2
2
7
6
4
C
P
F
C
a
r
o
l
i
n
a
B
a
n
k
H
o
l
d
i
n
g
s
,
I
n
c
.
2
9
4
3
4
7
3
C
L
B
H
C
r
e
s
c
e
n
t
B
a
n
k
i
n
g
C
o
m
p
a
n
y
1
9
5
8
8
2
7
C
S
N
T
40
A
p
p
e
n
d
i
x
A
P
a
n
e
l
B
:
N
o
n
-
S
e
c
u
r
i
t
i
s
e
r
s
(
C
O
N
T