Description
The purpose of this paper is to investigate whether the bond market disciplines all banks
equally, in the sense of demanding the same relative risk premium across banks of different risk over
the business cycle.
Journal of Financial Economic Policy
Do markets “discipline” all banks equally?
J oa˜o A.C. Santos
Article information:
To cite this document:
J oa˜o A.C. Santos, (2009),"Do markets “discipline” all banks equally?", J ournal of Financial Economic
Policy, Vol. 1 Iss 1 pp. 107 - 123
Permanent link to this document:http://dx.doi.org/10.1108/17576380910962402
Downloaded on: 24 January 2016, At: 21:35 (PT)
References: this document contains references to 23 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 2388 times since 2009*
Users who downloaded this article also downloaded:
Anand Kumar J aiswal, (2008),"Customer satisfaction and service quality measurement in Indian
call centres", Managing Service Quality: An International J ournal, Vol. 18 Iss 4 pp. 405-416 http://
dx.doi.org/10.1108/09604520810885635
J eannette A. Colyvas, Walter W. Powell, (2009),"Measures, metrics, and myopia: The challenges
and ramifications of sustaining academic entrepreneurship", Advances in the Study of
Entrepreneurship, Innovation & Economic Growth, Vol. 19 pp. 79-111http://dx.doi.org/10.1108/
S1048-4736(2009)0000019004
J osé C. Alves, Kathi J . Lovelace, Charles C. Manz, Dmytro Matsypura, Fuminori Toyasaki, Ke (Grace) Ke,
(2006),"A cross-cultural perspective of self-leadership", J ournal of Managerial Psychology, Vol. 21 Iss 4 pp.
338-359http://dx.doi.org/10.1108/02683940610663123
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Do markets “discipline”
all banks equally?
Joa˜o A.C. Santos
Research Department, Federal Reserve Bank of New York,
New York, New York, USA
Abstract
Purpose – The purpose of this paper is to investigate whether the bond market disciplines all banks
equally, in the sense of demanding the same relative risk premium across banks of different risk over
the business cycle.
Design/methodology/approach – To test this hypothesis, the paper compares the difference
between the credit spreads in the primary market of bank and ?rm bonds with the same credit rating
issued during expansions with that same difference of spreads for bonds issued during recessions.
Findings – The paper ?nds that during recessions investors demand higher risk premiums.
Importantly, the paper ?nds that the impact of recessions is not uniform across banks – it affects
riskier banks more than safer ones. In other words, in recessions investors are relatively more
demanding on riskier banks than on safer ones.
Originality/value – These ?ndings are novel. They also have important policy implications because
they show that a bond-issuance policy aimed at promoting market discipline could affect the relative
funding costs of banks over the business cycle. They also indicate that the information which can
be extracted fromthe credit spreads on bank bonds varies across banks for reasons unrelated to their risk.
Keywords Banks, Risk analysis, Financial markets
Paper type Research paper
1. Introduction
Growing complexity of banks coupled with the development of debt markets has
increased the interest in the potential market discipline of banks. This interest has
motivated many empirical studies, most of them investigating whether the yields on
bank bonds are correlated with the issuers’ risk of default. This paper adds to this
literature by investigating if the market always charges the same relative risk premium
across banks of different riskiness. This is an important issue because otherwise a
bond-issuance policy aimed at promoting market discipline will affect the relative
funding costs of banks. Further, the information that can be extracted from the credit
spreads on bank bonds will vary across banks for reasons unrelated to their risk.
Banks’ simultaneous provision of liquidity insurance and monitoring services leads
to a mismatch between liquid liabilities and illiquid assets[1]. In the event of a liquidity
shock, the same information asymmetries that lead banks to adopt this asset and
liability structure make it dif?cult for them to borrow the necessary funds in the
market. As a result, they may be forced into bankruptcy. The premature liquidation of
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – E44, G32
The author thanks Adam Ashcraft, Scott Frame, and seminar participants at the Federal
Reserve Bank of New York for useful comments and suggestions, and Chris Metli and Kyle Lewis
for outstanding research assistance. The views stated herein are those of the author and are not
necessarily the views of the Federal Reserve Bank of New York or the Federal Reserve System.
Do markets
“discipline” all
banks equally?
107
Journal of Financial Economic Policy
Vol. 1 No. 1, 2009
pp. 107-123
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380910962402
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
bank assets is costly because it ends valuable relationships and it may develop into a
bank panic that culminates in a system failure. This risk of a system failure underpins
the classical argument for government interference in the business of banking.
Government interference materializes in different forms, including the introduction
of regulations covering such things as capital requirements and permissible activities;
the establishment of public institutions to perform supporting functions to banks,
including the provision of deposit insurance and lending of last resort; and the
establishment of public institutions like supervisors to monitor banks’ activities. As
banking organizations have evolved within the increasingly global and complex
?nancial markets, it has become dif?cult for the institutions responsible for supporting
and overseeing banks to meet their mandates. For this reason, they have expressed
growing interest in the use of market-related oversight to supplement their own
methods of monitoring.
Motivated by this interest, researchers have studied whether markets provide any
discipline to banks by investigating whether the credit spreads of bonds issued by
banks varied with their risk[2]. Early studies, including Avery et al. (1988) and Gorton
and Santomero (1990), ?nd that credit spreads on bank subordinated debentures were
virtually unrelated to traditional accounting measures of bank risk. Subsequent
studies, including Flannery and Sorescu (1996), Jagtiani et al., 2002 and Sironi (2003),
however, ?nd a relationship between the credit spreads on bank subordinated
debentures and risk[3].
To date, the literature on the market discipline has investigated extensively whether
credit spreads on uninsured bank debt are correlated with bank risk, but it has paid only
limited attention to such closely related issues as the pricing of bank risk, with the
exception, perhaps, of the “subsidy” that the so-called “too-big-to-fail” banks receive
from investors[4]. Our paper contributes to ?ll this gap in the literature by investigating
if the market charges all banks the same “price” of risk. Speci?cally, we investigate if
investors always demand the same relative risk premium across banks of different risk,
in other words, if the slope of the credit-spreadcurve of bank bonds is constant over time.
To test this hypothesis, we ?rst compare the credit spreads (over treasuries) in the
primary market of bonds issued by banks during expansions with the spreads of
bonds with the same credit rating issued during recessions. We ?nd that, ceteris
paribus, during recessions investors demand higher risk premiums. Importantly, the
impact of recessions is not uniform across banks, instead it affects riskier banks more
than safer ones. Implicit in the ?rst test, and following rating agencies’ claim that their
ratings are comparable over the business cycle, is the assumption that a bond issued
during an expansion and which receives a given credit rating carries the same risk as a
bond issued during a recession that carries the same rating[5]. Also implicit in that test
is the assumption that access conditions to the bond market do not vary over the
business cycle in such a way that make it relatively more dif?cult for some issuers to
access it, for example, during recessions[6].
Since violations of these assumptions could explain the results of our ?rst test, we
perform a second test in which we expand our sample to include bonds issued by ?rms
and compare the difference between the bank and ?rm credit spreads of bonds with the
same rating issued during expansions with that same difference but for bonds issued
during recessions. As we found for banks, we ?nd that recessions increase the cost
to access the bond market for riskier ?rms by more than it does for safer ?rms.
JFEP
1,1
108
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Further, we ?nd that the marginal impact of recessions on the cost to access the bond
market is the same for banks and ?rms that have a Baa credit rating. Importantly, we
?nd that recessions increase the cost of accessing the bond market to safer banks by
less than it does to ?rms with the same risk.
These results con?rm the ?ndings of our ?rst test that the market does not
discipline all banks equally – it is relatively more demanding on riskier banks than on
safer ones. This difference is important. It implies, for example, that a bond-issuance
policy aimed at promoting market discipline will make funding relatively more
expensive for riskier banks than for safer banks, a difference which will be ampli?ed
during recessions[7]. It also implies that the information content in credit spreads of
bank bonds not only depends on the risk of the issuer but also on the state of the
economy. As a result, the implementation of a monitoring program, like a prompt
correction scheme, parameterized on credit spreads on bank bonds, would need to take
into account both the riskiness of the issuer and the state of the economy in order to
promote similar levels of monitoring across banks over time[8].
The remainder of the paper is organized as follows. Section 2 presents our data and
methodology. Section 3 presents the results of our tests. Section 4 concludes the paper.
2. Data and methodology
2.1 Data
The data for this paper came from SDC’s Domestic New Bond Issuances Database. Our
sample of bonds includes only bonds issued in the USA in US dollars by American
commercial banks and non?nancial companies between 1982:2 and 2002:2[9]. However,
we use information on bonds issued since 1970:1 to identify the ?rst time ?rms issued
bonds and to measure the frequency ?rms have issued bonds over time.
Our sample of bonds includes both shelf and non-shelf issues, and bonds issued in
the public market as well as those privately placed. We exclude asset-backed bonds
and convertible bonds. We also exclude bonds for which we did not have the necessary
information to compute their credit spreads over the treasury at issue date, and bonds
with maturities longer than 30 years. Finally, we exclude bonds that do not have
ratings from both Moody’s and Standard & Poor’s (S&P) at issue date. These criteria
left us with a sample of 11,587 bonds, of which 1,537 were issued by commercial banks.
We use NBER’s identi?cation of troughs and peaks to ?nd out if a bondwas issued in a
recession or in an expansion. We de?ne a recession as the time period between a peak and
a trough, and an expansion as the time period between a trough and a peak. We classify a
quarter to be a recession (expansion) if either all months or the majority of months in the
quarter are in a recession (expansion) period. According to this classi?cation, there are 72
quarters of expansion and nine quarters of recession during the period 1982:2-2002:2[10].
Of the 1,537 bonds issued by banks, 12 per cent were issued in recessions and of the 10,050
bonds issued by non?nancial ?rms 10 per cent were issued in recessions.
Table I characterizes our samples of ?rm and bank bonds. There are some important
differences between our two samples of bonds. For example, there are far more risky
?rm bonds than bank bonds. While 31 per cent of the ?rm bonds are below investment
grade, less than 1 per cent of bank bonds have a speculative rating. Further, while
2 per cent of the ?rm bonds have a triple-A rating only 0.5 per cent of the bank bonds
have that rating. These differences suggest that when we compare bank bonds with ?rm
bonds we should focus on investment grade bonds that are not triple-A rated.
Do markets
“discipline” all
banks equally?
109
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Another important difference between the two samples of bonds is the frequency of
split ratings between the two major US rating agencies, Moody’s and S&P, which
Morgan (2000) argues is due to banks being more opaque than ?rms. These agencies
disagreed on 15 per cent of the ratings they assigned to bonds issued by ?rms. In
contrast, they disagreed on 34 per cent of the ratings they assigned on bonds issued by
banks. Still based on Table I, we ?nd that ?rms issue far more bonds with a sinking
fund and a put provision than banks do. They also issue more callable bonds and
senior bonds than banks do. Firms, in addition, do substantially more private
placements than banks do. In contrast, banks do more shelf issues.
Firm bonds Bank bonds
Variables 1982:2-2002:2 Expansions Recessions 1982:2-2002:2 Expansions Recessions
Volume and number of issues
Amount issued
a
11,653.56 10,204.21 1,449.35 1,667.64 1,497.27 170.37
Number of issues 10,050 9,080 970 1,537 1,349 188
Shares of the number of issues by credit rating
b
Aaa 1.930 1.586 5.155 0.520 0.519 0.532
Aa 8.318 8.139 10.000 35.784 34.841 42.553
A 32.179 31.189 41.443 47.885 48.258 45.213
Baa 26.030 26.267 23.814 15.029 15.493 11.702
Ba 8.617 8.733 7.526 0.791 0.890 0.000
B 21.791 22.863 11.753 0.000 0.000 0.000
Below B 1.134 1.289 0.309 0.000 0.000 0.000
Shares of the number of issues with split ratings
Split 15.423 15.529 14.433 34.027 33.062 40.957
SP . Mood 56.387 56.429 56.383 40.344 41.588 40.135
Shares of the number of issues by design of the issue
Callable bonds 36.607 36.905 33.814 23.552 21.201 40.426
Bonds with
a sinking
fund 8.060 8.150 7.217 0.260 0.148 1.064
Senior bonds 84.617 83.921 91.134 67.534 65.678 80.851
Shelf bonds 59.602 59.548 60.103 73.910 73.536 76.596
Private placements 20.985 20.595 24.639 0.065 0.074 0.000
Average maturity
(years) 10.466 10.567 9.519 8.980 8.722 10.837
Bonds with
a putable provision
on maturity
c
1.622 1.762 0.309 0.716 0.815 0.000
Shares of the number of issues by issuer features
First issue 26.886 26.355 13.711 9.109 9.562 5.861
Public company 68.706 68.480 70.825 79.245 78.206 86.702
Notes: It includes all new non-convertible bonds issued by American non?nancial ?rms and
commercial banks in the USA in US dollars over the period 1982:2-2002:2 that had ratings from both
Moody’s and S&P, information on amount issued, maturity and yield to maturity; recessions and
expansions de?ned according to NBER;
a
millions of US dollars, issues de?ated by the core urban
consumer price index (CPI) with the average of 1982-1984 ¼ 100;
b
shares computed based on Moody’s
ratings;
c
bondholders have the option of selling the bond back to the issuing ?rm before the maturing
date
Table I.
Sample characterization
JFEP
1,1
110
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
2.2 Methodology
The methodology we use in this paper has two parts. The ?rst part focuses on
banks and attempts to test the following hypothesis: if the market disciplines all banks
equally, then the impact of a change in the state of the economy on the cost to access
the bond market should not vary across bonds of different creditworthiness. Under
this hypothesis, we would have that the marginal impact of recessions on bond spreads
at issue date would be independent from the issuer’s risk. To test this hypothesis, we
estimate the following model of (bank) bond spreads:
Spread ¼ c þa
i
X
6
i¼1
Rat
i
þb Rec þg
i
Rec
X
6
i¼1
Rat
i
þc
i
X
L
i¼1
X
i
þ1 ð1Þ
where Spread is the bond’s spread over treasury at issue date. We proxy for the
creditworthiness of the bank by the credit rating of its bonds. We de?ne dummy
variables Rat
i
for each credit rating that Moody’s can assign a bond using the
following convention: R
1
equals one for double-A rated bonds; R
2
equals one for
single-A rated bonds; R
3
equals one for Baa-rated bonds, R
4
equals one for Ba-rated
bonds, R
5
equals one for single-B rated bonds and ?nally R
1
equals one for bonds with
a Caa or lower credit rating[11]. Under this criterion, the bonds left out are those with a
triple-A credit rating at issue date[12].
We control for the state of the economy at the time the bond is issued by including a
dummy variable Rec, which takes the value one if the bond is issued during a
recession. Recessions and expansions are denned according to NBER. Since, we want
to ascertain if recessions affect banks of different creditworthiness differently, we
include in our model of bond spreads the interaction of the recession dummy, Rec, with
the bond-rating dummies Rat
i
. If the market were to discipline all banks equally we
would expect our estimates of the coef?cients g
i
to be all statistically insigni?cant. In
this case, if b is positive we would have that recessions make it equally more expensive
for all banks to access the bond market.
We estimate these marginal effects controlling for a set of factors related to the
design of the bond and a set of characteristics of the issuer, which other studies of bond
pricing have shown help explain bond credit spreads[13]. We include dummy variables
to control for private placements, shelf and callable bonds, bonds with a put option,
and bonds with a sinking fund. We also control for the maturity of the bond and for the
amount issued. Finally, following the evidence that rating splits affect credit spreads,
we include a dummy variable, Split, which takes the value one if the bond received a
rating from S&P which differs from the rating it received from Moody’s[14]. In
addition, we include a dummy variable SP . Mood, which takes the value one if the
bond received a more favorable rating from S&P than from Moody’s. With respect to
the issuer, we control for the ?rst bond issued by the bank, the number of times the
bank issued in the past, and the length of time since the bank’s last bond issue.
Finally, we consider a set of other controls, including a time trend, and the total
number of bonds issued in each quarter in the investment and speculative grade
markets to control for a potential market segmentation resulting, for example, from
regulations. In addition, we control for the ten-year constant treasury yield and the
slope of the yield curve (difference between the 30- and ?ve-year yields).
We ?rst estimate our model of bond pricing with the ordinary least squares (OLS)
correcting the standard errors for heteroskedasticity. We then reestimate this model
Do markets
“discipline” all
banks equally?
111
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
with bank ?xed effects to account for potential systematic differences in the pricing of
bonds across banks that may arrive from such things as the so-called “too-big-to-fail”
status. In addition, to account for the disproportional level of bond issuance across the
banks in our sample we reestimate our ?xed-effects model with weighted least squares,
computing the weights based on the number of bonds each bank issued over the
sample period.
As we noted earlier, implicit in this ?rst test is the assumption, that ratings are
com-parable over the business cycle (as claimed by rating agencies). If that were not
the case, then our null hypothesis would be violated for reasons that have nothing to do
with market discipline. To reduce concerns with this limitation of our ?rst test, in
the second part of our methodology we add a control group – the bonds issued by
?rms – that could also be subject to these same problems. In this part of our
methodology, we want to ?nd out if the difference between the spread at issue date of a
bank bond and that of a ?rm bond with the same rating varies over the business cycle.
We also want to ascertain if that difference between bank and bond spreads varies
across bonds of different creditworthiness. To this end, we estimate the following
model of bond pricing:
Spread ¼ c þa
i
X
6
i¼1
Rat
i
þb Rec þg
i
Rec
X
6
i¼1
Rat
i
þhBk þz
i
Bk
X
6
i¼1
Rat
i
þf Rec Bk þd
i
Rec Bk
X
6
i¼1
Rat
i
þc
i
X
L
i¼1
X
i
þ1;
ð2Þ
where Spread, Rat
i
, Rec and X
i
are de?ned as in our ?rst test. The key new variable of
our second test is the dummy variable Bk, which takes the value one for bonds issued
by banks. We also include in our new model of bond spreads, the interaction of this
variable with the credit-rating and recession dummy variables as well as with the
interaction of the latter two variables to allow for different pricing of bank and ?rm
bonds with the same credit rating in good times as well as in recessions.
Under the null hypothesis that the market disciplines all banks equally we would
expect the relative difference between bank and ?rm bond spreads across bonds of
different credit-worthiness not to vary over the business cycle. In terms of the
parameters of our new model of bond pricing, this suggests that the coef?cients d
i
should not be statistically different from zero.
3. Results
3.1 Credit spreads on bank bonds
Table II characterizes the credit spreads at issue on the bonds issued by American
commercial banks in the USA over the period 1982:2-2002:2. We compute these spreads
as the percentage point difference between the yield to maturity of the bond and the
yield on an equivalent maturity US treasury bond. The average credit spread for the
1,537 bank bonds in our sample is 95 basis points. As expected, bonds with lower
credit ratings carry higher credit spreads. While on average a triple-A credit rating
bond carries a credit spread of 54 basis points, a Ba-rated bond, the lowest rated bonds
in our sample of bonds issued by banks, carries a credit spread of 242 basis points.
Comparing the credit spreads at issue date of bonds issued during expansions with
those of bonds issued during recessions, we ?nd that on average recessions increase
JFEP
1,1
112
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the cost of bond issuance by 44 basis points. Note, however, that recessions do not
affect the credit spreads of all bonds equally. On average, they tend to have a larger
impact on riskier bonds. Recessions do not affect the cost of accessing the bond market
for banks that issue triple-A rated bonds and it increases the cost of this funding source
for banks that issue double-A rated bonds by only 27 basis points. In contrast,
recessions increase the cost of bond ?nancing for banks that issue Baa-rated bonds by
112 basis points.
These results suggest that the risk structure of credit spreads on bank bonds
become steeper during recessions, an indication that investors impose higher costs on
riskier banks in recessions when it comes to market funding than they do on safer
banks. We investigate next if this difference continues to hold when we account for
other variables that also explain bond credit spreads.
3.2 The risk structure of credit spreads on bank bonds
Table III presents the results of our ?rst test on the impact of recessions on the risk
structure of credit spreads on bank bonds. Recall that under this test we consider only
bonds issued by banks and try to ascertain if the marginal impact of recessions on the
cost of bond issuance is constant across issuers of different risk. The ?rst three models
were estimated without any controls. These models con?rm the insights of our
univariate analysis of bond credit spreads: the lower the bond credit rating the higher
the credit spread (Model 1); on average, recessions increase spreads of bank bonds by
about 47 basis points (Model 2); the impact of recessions is not uniform across
issuers of different risk, instead, they have a larger impact on riskier issuers than on
safer ones.
This asymmetric impact of recessions on credit spreads continues to hold when we
add the controls that other studies of bond spreads have found to help explain credit
spreads (Model 4). As with these studies, we ?nd that callable bonds, bonds with a
sinking fund, longer maturity bonds and senior bonds carry higher credit spreads, but
bonds with a put provision and shelf issues pay lower credit spreads. We also ?nd that
larger issues as well as private placements pay higher credit spreads. Issues at the time
the treasury yield is high pay lower credit spreads. We do not ?nd evidence of a secular
decline in spreads of bank bonds. Somewhat surprisingly we ?nd that public banks
pay higher spreads than, private banks. This, however, may be due to the reduced
number of private banks in the sample.
1982-2002 Expansion Recession Rec-Exp t-test
All 0.952 0.899 1.334 0.435 10.340
*
Aaa 0.537 0.485 0.906 0.421 1.417
Aa 0.841 0.802 1.068 0.266 4.969
*
A 0.885 0.829 1.317 0.488 9.758
*
Baa 1.370 1.263 2.385 1.122 7.665
*
Ba 2.421 2.421 – – –
Notes:
*
Signi?cant at 1 per cent level; it includes all new non-convertible bonds issued by American
commercial banks in the USA in US dollars over the period 1982:2-2002:2 that had ratings from both
Moody’s and S&P, information on amount issued, maturity and yield to maturity; recessions and
expansions de?ned according to NBER
Table II.
Spreads of bank bonds by
credit rating arid state
of the economy
Do markets
“discipline” all
banks equally?
113
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
M
o
d
e
l
7
M
o
d
e
l
8
M
o
d
e
l
9
a
M
o
d
e
l
1
0
M
o
d
e
l
1
1
M
o
d
e
l
1
2
a
C
o
n
s
t
a
n
t
0
.
5
3
7
(
5
.
4
7
1
*
*
*
)
0
.
4
7
8
(
5
.
7
3
4
*
*
*
)
0
.
4
8
5
(
5
.
0
6
1
*
*
*
)
0
.
3
5
5
(
1
.
4
4
6
)
0
.
4
1
4
(
1
.
6
6
2
*
)
2
.
0
4
4
(
6
.
8
0
4
*
*
*
)
0
.
3
0
8
(
2
.
3
2
6
*
*
)
0
.
7
8
5
(
2
.
9
9
3
*
*
*
)
1
.
8
1
7
(
5
.
7
8
7
*
*
*
)
0
.
2
0
9
(
1
.
4
1
7
)
1
.
3
1
1
(
2
.
9
5
1
*
*
*
)
2
.
2
2
9
(
4
.
5
5
9
*
*
*
)
A
a
0
.
3
0
4
(
3
.
0
3
4
*
*
*
)
0
.
2
9
4
(
3
.
4
4
3
*
*
*
)
0
.
3
1
8
(
3
.
2
3
8
*
*
*
)
0
.
3
6
6
(
3
.
9
4
8
*
*
*
)
0
.
3
2
3
(
3
.
0
0
2
*
*
*
)
0
.
3
9
8
(
3
.
9
9
9
*
*
*
)
0
.
2
7
8
(
2
.
9
4
6
*
*
*
)
0
.
2
6
1
(
2
.
2
7
0
*
*
)
0
.
4
5
4
(
3
.
1
1
5
*
*
*
)
0
.
2
7
7
(
2
.
7
7
3
*
*
*
)
0
.
2
7
3
(
2
.
2
9
2
*
*
)
0
.
4
3
8
(
2
.
6
8
3
*
)
A
0
.
3
4
8
(
3
.
4
9
2
*
*
*
)
0
.
3
5
3
(
4
.
1
6
0
*
*
*
)
0
.
3
4
4
(
3
.
5
4
1
*
*
*
)
0
.
4
1
6
(
4
,
4
3
6
*
*
*
)
0
.
4
5
1
(
4
.
1
3
5
*
*
*
)
0
.
6
5
3
(
6
.
4
6
7
*
*
*
)
0
.
3
9
1
(
4
.
0
3
7
*
*
*
)
0
.
4
6
8
(
3
.
7
3
0
*
*
*
)
0
.
7
1
2
(
4
.
6
8
4
*
*
*
)
0
.
3
9
7
(
3
.
4
8
7
*
*
)
0
.
3
1
1
(
2
.
2
9
4
*
*
)
0
.
5
7
3
(
3
.
1
7
6
*
*
)
B
a
a
0
.
8
3
2
(
7
.
6
1
3
*
*
*
)
0
.
8
4
6
(
8
.
9
5
5
*
*
*
)
0
.
7
7
8
(
7
.
3
5
4
*
*
*
)
0
.
9
1
4
(
8
.
7
6
0
*
*
*
)
0
.
9
6
1
(
8
.
0
8
9
*
*
*
)
1
.
0
9
4
(
1
0
.
1
0
3
*
*
*
)
0
.
7
8
3
(
7
.
0
8
1
*
*
*
)
1
.
0
2
8
(
7
.
2
6
7
*
*
*
)
1
.
1
5
0
(
7
.
1
0
5
*
*
*
)
0
.
5
0
1
(
3
.
7
6
3
*
*
*
)
0
.
4
3
6
(
2
.
4
5
5
*
*
)
0
.
7
3
9
(
3
.
4
2
5
*
*
)
B
a
1
.
8
8
3
(
6
.
3
0
3
*
*
*
)
1
.
9
4
2
(
6
.
6
0
1
*
*
*
)
1
.
9
3
6
(
6
.
4
9
6
*
*
*
)
1
.
9
3
0
(
7
.
2
4
9
*
*
*
)
1
.
9
5
8
(
7
.
3
6
0
*
*
*
)
2
.
1
0
1
(
8
.
2
0
9
*
*
*
)
1
.
1
6
8
(
3
.
0
6
4
*
*
*
)
1
.
5
3
9
(
3
.
8
9
7
*
*
*
)
1
.
7
2
6
(
5
.
0
5
8
*
*
*
)
0
.
9
6
9
(
2
.
1
5
7
*
*
)
1
.
0
2
7
(
2
.
2
9
6
*
*
)
1
.
4
1
3
(
3
.
6
7
8
*
*
*
)
R
e
c
0
.
4
7
2
(
1
1
.
7
0
5
*
*
*
)
0
.
4
2
1
(
4
.
3
9
1
*
*
*
)
0
.
3
9
0
(
3
.
9
4
9
*
*
*
)
0
.
4
2
0
(
3
.
7
2
6
*
*
*
)
2
0
.
1
3
3
(
2
1
.
0
8
6
)
0
.
1
2
6
(
1
.
0
1
8
)
2
0
.
0
2
7
(
2
0
.
1
8
7
)
2
0
.
2
4
7
(
2
1
.
4
8
8
)
0
.
1
9
3
(
1
.
5
3
2
)
0
.
0
0
3
(
0
.
0
1
7
)
2
0
.
2
2
6
(
2
1
.
0
2
9
)
R
e
c
A
a
2
0
.
1
5
5
(
2
1
.
4
5
6
)
2
0
.
2
0
5
(
2
1
.
8
0
6
*
)
2
0
.
2
5
7
(
2
1
.
9
9
3
*
*
*
)
0
.
2
3
9
(
1
.
7
8
9
*
*
*
)
0
.
0
0
7
(
0
.
0
5
6
)
0
.
0
9
3
(
0
.
5
9
4
)
0
.
3
2
0
(
1
.
8
4
9
*
)
0
.
2
1
5
(
1
.
0
2
3
)
0
.
2
3
4
(
1
.
2
8
2
)
0
.
3
8
3
(
1
.
6
3
4
)
R
e
c
A
0
.
0
6
7
(
0
.
6
0
7
)
0
.
0
7
7
(
0
.
6
8
1
)
0
.
0
6
4
(
0
.
5
1
6
)
0
.
2
2
0
(
1
.
7
6
3
*
*
*
)
0
.
2
1
5
(
1
.
5
2
4
)
0
.
4
0
7
(
2
.
6
3
0
*
*
*
)
0
.
3
7
5
(
2
.
2
5
5
*
*
*
)
0
.
4
1
0
(
2
.
7
7
0
*
*
*
)
0
.
5
3
5
(
3
.
0
6
2
*
*
*
)
0
.
4
8
6
(
2
.
2
2
6
*
*
*
)
R
e
c
B
a
a
0
.
7
0
2
(
3
.
9
0
3
*
*
*
)
0
.
6
8
3
(
3
.
8
9
5
*
*
*
)
0
.
6
4
7
(
3
.
5
2
5
*
*
*
)
0
.
6
9
8
(
4
.
1
5
1
*
*
*
)
0
.
9
3
1
(
4
.
9
7
2
*
*
*
)
0
.
9
4
9
(
4
.
7
8
2
*
*
*
)
0
.
7
1
2
(
3
.
4
3
2
*
*
*
)
1
.
0
9
8
(
4
.
2
4
1
*
*
*
)
1
.
1
2
5
(
3
.
2
4
2
*
*
*
)
0
.
9
6
0
(
2
.
8
7
0
*
*
*
)
O
t
h
e
r
i
s
s
u
e
r
f
e
a
t
u
r
e
s
P
u
b
l
i
c
c
o
m
p
a
n
y
0
.
0
7
7
(
2
.
5
2
9
*
*
)
0
.
0
8
3
(
2
.
3
2
0
*
*
)
0
.
0
3
7
(
1
.
3
5
9
)
0
.
1
2
4
(
3
.
0
6
7
*
*
*
)
0
.
0
7
5
(
2
.
1
9
2
*
*
)
0
.
1
2
0
(
1
.
9
3
7
*
)
0
.
1
1
7
(
2
.
1
5
5
*
)
F
i
r
s
t
i
s
s
u
e
0
.
0
5
8
(
1
.
2
2
4
)
2
0
.
0
4
6
(
2
0
.
9
0
0
)
0
.
0
5
1
(
0
.
0
.
5
2
0
)
O
r
d
e
r
o
f
i
s
s
u
e
0
.
0
3
7
(
3
.
5
6
0
*
*
*
)
2
0
.
0
0
1
(
2
0
.
9
0
0
)
T
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
2
0
.
0
0
4
(
2
0
.
3
7
0
)
2
0
.
0
1
0
(
2
0
.
9
0
2
)
0
.
0
0
3
(
0
.
2
4
7
)
I
s
s
u
e
d
e
s
i
g
n
C
a
l
l
p
r
o
v
i
s
i
o
n
0
.
1
1
4
(
3
.
8
0
7
*
*
*
)
0
.
1
0
9
(
3
.
6
6
7
*
*
*
)
0
.
0
9
4
(
3
.
5
7
2
*
*
*
)
0
.
0
9
1
(
3
.
0
7
8
*
*
*
)
0
.
0
8
3
(
3
.
1
6
4
*
*
*
)
0
.
0
0
5
(
0
.
0
9
7
)
2
0
.
0
1
8
(
2
0
.
3
6
2
)
P
u
t
p
r
o
v
i
s
i
o
n
2
0
.
6
7
0
(
2
7
.
4
1
0
*
*
*
)
2
0
.
6
5
6
(
2
7
.
0
2
9
*
*
*
)
2
0
.
4
0
4
(
2
4
.
3
7
8
*
*
*
)
2
0
.
5
2
9
(
2
5
.
4
2
6
*
*
*
)
2
0
.
3
9
6
(
2
4
.
3
4
5
*
*
*
)
2
0
.
5
6
6
(
2
6
.
0
1
8
*
*
*
)
2
0
.
3
9
9
(
2
4
.
2
2
2
*
*
*
)
S
i
n
k
i
n
g
f
u
n
d
0
.
6
8
9
4
.
8
8
0
*
*
*
)
0
.
6
7
3
(
6
.
2
4
0
*
*
*
)
0
.
9
2
0
(
4
.
8
3
1
*
*
*
)
0
.
7
5
2
(
4
.
9
4
4
*
*
*
)
1
.
0
0
5
(
3
.
3
3
5
*
*
*
)
0
.
7
3
3
(
3
.
5
9
2
*
*
*
)
0
.
7
0
5
(
2
.
9
0
9
*
*
*
)
S
h
e
l
f
2
0
.
2
0
2
(
2
6
.
1
4
7
*
*
*
)
2
0
.
2
1
5
(
2
6
.
5
8
9
*
*
*
)
2
0
.
2
0
6
(
2
7
.
3
9
5
*
*
*
)
2
0
.
2
1
3
(
2
6
.
3
7
0
*
*
*
)
2
0
.
1
9
4
(
2
6
.
6
1
7
*
*
*
)
2
0
.
0
9
8
(
2
1
.
4
8
8
)
2
0
.
1
2
9
(
2
2
.
3
7
4
*
*
)
L
o
g
(
m
a
t
u
r
i
t
y
)
0
.
2
0
9
(
1
2
.
0
7
4
*
*
*
)
0
.
1
9
7
(
1
1
.
1
3
0
*
*
*
)
0
.
1
7
9
(
1
1
.
3
7
9
*
*
*
)
0
.
1
7
5
(
9
.
8
0
9
*
*
*
)
0
.
1
9
0
(
1
2
.
2
5
8
*
*
*
)
0
.
2
1
7
(
6
.
9
4
6
*
*
*
)
0
.
2
4
9
(
8
.
2
4
4
*
*
*
)
S
e
n
i
o
r
0
.
0
7
4
(
2
.
2
0
4
*
*
)
0
.
0
6
9
(
2
.
0
4
2
*
*
)
0
.
0
0
6
(
0
.
2
1
9
)
0
.
0
2
8
(
0
.
7
8
8
)
0
.
0
0
7
(
0
.
2
4
0
)
0
.
0
1
2
(
0
.
1
3
6
)
0
.
0
3
9
(
0
.
5
0
2
)
(
c
o
n
t
i
n
u
e
d
)
Table III.
Bank-bond spreads over
treasury at issue date
JFEP
1,1
114
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
M
o
d
e
l
7
M
o
d
e
l
8
M
o
d
e
l
9
a
M
o
d
e
l
1
0
M
o
d
e
l
1
1
M
o
d
e
l
1
2
a
O
t
h
e
r
c
o
n
t
r
o
l
s
L
o
g
(
a
m
o
u
n
t
i
s
s
u
e
d
)
0
.
0
2
6
(
2
.
6
3
7
*
*
*
)
0
.
0
3
4
(
3
.
3
8
2
*
*
*
)
0
.
0
3
8
(
4
.
3
7
5
*
*
*
)
0
.
0
2
5
(
2
.
0
4
0
*
*
)
0
.
0
2
5
(
2
.
1
8
0
*
*
*
)
0
.
0
2
0
(
0
.
8
1
8
)
0
.
0
1
3
(
0
.
5
1
4
)
P
r
i
v
.
p
l
a
c
e
m
e
n
t
0
.
1
8
7
(
3
.
4
2
1
*
*
)
0
.
1
3
0
(
2
.
2
9
4
*
*
)
2
0
.
0
4
2
(
2
0
.
8
3
2
)
0
.
1
5
5
(
1
.
3
8
0
)
2
0
.
0
7
6
(
2
0
.
7
4
6
)
0
.
1
8
3
(
1
.
4
0
7
)
2
0
.
0
2
8
(
2
0
.
2
2
2
)
L
o
g
(
t
i
m
e
t
r
e
n
d
)
0
.
0
0
8
(
0
.
2
2
9
)
2
0
.
0
2
0
(
2
0
.
5
9
9
)
2
0
.
0
0
2
(
2
0
.
0
5
5
)
2
0
.
0
5
2
(
2
1
.
3
6
3
)
0
.
0
2
1
(
0
.
4
8
7
)
2
0
.
1
4
1
(
2
1
.
8
6
6
*
)
2
0
.
0
6
3
(
2
0
.
8
6
5
)
T
e
n
-
y
e
a
r
T
r
e
a
s
u
r
y
y
i
e
l
d
2
0
.
0
2
7
(
2
1
.
8
3
6
*
)
2
0
.
0
2
3
(
2
1
.
5
2
9
)
2
0
.
0
8
0
(
2
4
.
8
6
4
*
*
*
)
2
0
.
0
4
7
(
2
3
.
1
9
0
*
*
*
)
2
0
.
0
7
8
(
2
4
.
9
6
9
*
*
*
)
2
0
.
0
8
4
(
2
2
.
9
5
2
*
*
*
)
2
0
.
1
1
3
(
2
4
.
2
5
7
*
*
*
)
Y
i
e
l
d
c
.
s
l
o
p
e
2
0
.
2
2
6
(
2
8
.
2
9
5
)
2
0
.
2
2
6
(
2
8
.
3
8
6
*
*
*
)
2
0
.
1
8
7
(
2
6
.
8
2
0
*
*
*
)
2
0
.
2
5
7
(
2
9
.
6
0
4
*
*
*
)
2
0
.
1
8
3
(
2
7
.
0
8
0
*
*
*
)
2
0
.
1
8
2
(
2
1
.
6
8
7
*
)
2
0
.
1
7
9
(
2
1
.
9
2
1
)
S
p
l
i
t
0
.
2
2
3
(
5
.
4
5
8
*
*
*
)
0
.
0
8
7
(
2
.
3
7
7
*
*
*
)
0
.
2
3
7
(
4
.
7
5
9
*
*
*
)
0
.
1
3
1
(
3
.
2
3
0
*
*
*
)
2
0
.
0
0
9
(
2
0
.
0
7
7
)
2
0
.
0
9
3
(
2
0
.
9
1
6
)
S
P
.
M
o
o
d
2
0
.
2
2
2
(
2
3
.
7
4
5
*
*
*
)
2
0
.
2
0
5
(
2
3
.
8
7
1
*
*
*
)
2
0
.
2
5
8
(
2
3
.
6
3
7
*
*
*
)
2
0
.
2
4
4
(
2
4
.
0
7
5
*
*
*
)
2
0
.
1
9
2
(
2
1
.
1
7
5
)
2
0
.
1
8
7
(
2
1
.
3
4
6
)
W
i
t
h
b
a
n
k
?
x
e
d
e
f
f
e
c
t
s
N
o
N
o
N
o
N
o
N
o
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
W
i
t
h
w
e
i
g
h
t
e
d
o
b
s
.
N
o
N
o
N
o
N
o
N
o
N
o
N
o
N
o
N
o
Y
e
s
Y
e
s
Y
e
s
A
d
j
u
s
t
e
d
R
2
1
5
.
9
9
4
2
3
.
6
1
3
2
5
.
7
9
7
3
9
.
8
1
7
4
1
.
1
6
4
5
9
.
9
3
0
4
0
.
5
0
2
5
3
.
2
7
1
6
6
.
7
0
2
7
6
.
5
4
6
8
2
.
2
5
9
8
7
.
1
3
1
L
o
g
l
i
k
e
l
i
h
o
o
d
2
1
,
1
4
8
.
2
6
2
1
,
0
7
4
.
6
9
2
1
,
0
5
0
.
3
9
2
8
8
1
.
3
5
2
8
6
1
.
4
1
2
5
6
4
.
6
8
2
7
6
4
.
0
3
2
5
6
8
.
3
5
2
3
0
3
.
7
7
2
1
,
3
3
3
.
9
0
2
1
,
1
0
9
.
3
3
2
8
5
9
.
0
6
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
a
t
t
h
e
*
1
0
,
*
*
5
,
*
*
*
1
p
e
r
c
e
n
t
l
e
v
e
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
;
m
o
d
e
l
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
a
r
e
h
e
t
e
r
o
s
k
e
d
a
s
t
i
c
-
c
o
n
s
i
s
t
e
n
t
;
t
o
t
a
l
n
u
m
b
e
r
o
f
o
b
s
e
r
v
a
t
i
o
n
s
1
,
5
3
7
.
t
-
s
t
a
t
i
s
t
i
c
s
i
n
p
a
r
e
n
t
h
e
s
i
s
;
t
h
e
d
e
p
e
n
d
e
n
t
v
a
r
i
a
b
l
e
i
s
t
h
e
b
o
n
d
c
r
e
d
i
t
s
p
r
e
a
d
o
v
e
r
t
h
e
t
r
e
a
s
u
r
y
w
i
t
h
t
h
e
s
a
m
e
m
a
t
u
r
i
t
y
;
s
p
r
e
a
d
s
a
r
e
c
o
m
p
u
t
e
d
a
t
i
s
s
u
e
d
a
t
e
;
r
e
c
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
i
s
s
u
e
d
d
u
r
i
n
g
a
r
e
c
e
s
s
i
o
n
a
c
c
o
r
d
i
n
g
t
o
N
B
E
R
;
p
u
b
l
i
c
c
o
m
p
a
n
y
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
i
s
s
u
e
r
i
s
a
p
u
b
l
i
c
c
o
m
p
a
n
y
;
?
r
s
t
i
s
s
u
e
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
t
h
e
b
a
n
k
’
s
?
r
s
t
b
o
n
d
i
s
s
u
e
s
i
n
c
e
1
9
7
0
:
1
;
o
r
d
e
r
o
f
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
t
i
m
e
s
t
h
e
b
a
n
k
i
s
s
u
e
d
b
o
n
d
s
s
i
n
c
e
1
9
7
0
:
1
;
t
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
y
e
a
r
s
s
i
n
c
e
t
h
e
b
a
n
k
m
a
d
e
i
t
s
l
a
t
e
s
t
b
o
n
d
i
s
s
u
e
;
c
a
l
l
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
c
a
l
l
a
b
l
e
;
p
u
t
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
b
o
n
d
h
o
l
d
e
r
s
c
a
n
s
e
l
l
t
h
e
b
o
n
d
b
a
c
k
t
o
t
h
e
c
o
m
p
a
n
y
p
r
i
o
r
t
o
m
a
t
u
r
i
t
y
;
s
i
n
k
i
n
g
f
u
n
d
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
h
a
s
a
s
i
n
k
i
n
g
f
u
n
d
;
s
h
e
l
f
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
a
s
h
e
l
f
i
s
s
u
e
;
m
a
t
u
r
i
t
y
,
m
a
t
u
r
i
t
y
o
f
t
h
e
b
o
n
d
i
n
y
e
a
r
s
;
a
m
o
u
n
t
i
s
s
u
e
d
i
n
m
i
l
l
i
o
n
s
o
f
U
S
d
o
l
l
a
r
s
;
i
s
s
u
e
s
d
e
?
a
t
e
d
b
y
t
h
e
c
o
r
e
u
r
b
a
n
C
P
I
w
i
t
h
t
h
e
a
v
e
r
a
g
e
o
f
1
9
8
2
-
1
9
8
4
¼
1
0
0
;
p
r
i
v
.
p
l
a
c
e
m
e
n
t
d
u
m
m
y
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
p
r
i
v
a
t
e
l
y
p
l
a
c
e
d
;
t
i
m
e
t
r
e
n
d
,
l
i
n
e
a
r
t
i
m
e
t
r
e
n
d
;
a
i
n
c
l
u
d
e
d
i
n
t
h
i
s
r
e
g
r
e
s
s
i
o
n
b
u
t
n
o
t
s
h
o
w
n
i
n
t
h
e
t
a
b
l
e
a
r
e
t
h
e
n
u
m
b
e
r
o
f
b
o
n
d
s
i
s
s
u
e
d
i
n
t
h
e
q
u
a
r
t
e
r
i
n
e
a
c
h
o
f
t
h
e
n
i
n
e
m
a
i
n
c
r
e
d
i
t
r
a
t
i
n
g
c
l
a
s
s
e
s
a
s
d
e
?
n
e
d
b
y
M
o
o
d
y
’
s
w
h
o
l
e
r
a
t
i
n
g
s
,
A
a
a
,
.
.
.
,
C
Table III.
Do markets
“discipline” all
banks equally?
115
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Given that nearly all bonds issued in the USA are rated by both Moody’s and S&P and
given the evidence of other studies showing that split ratings, that is instances where
the two rating agencies rate the bond differently, affect the credit spread, we expanded
our model of bond credit spreads and added two new dummy variables. The ?rst
variable takes the value one when Moody’s and S&P announce a different rating for
the bond and the second variable takes the value one when S&P announces a higher
rating than Moody’s. As the results of Model 5 show, bonds that are rated differently
by Moody’s and S&P pay on average an additional 22 basis points. Interestingly,
among the split-rated bonds only those that S&P assigns a lower rating than Moody’s
pay higher credit spreads (recall that we have used Moody’s ratings to control for the
bond’s creditworthiness). Model 5 continues to indicate that recessions affect
disproportinaly more banks that issue lower rated bonds, even if they are of
investment grade quality, than banks that issue higher rated bonds.
That differential impact of recessions on bond credit spreads continues to hold when
we control for the acquaintance of investors with the bank’s bonds by distinguishing the
?rst bond issue of the bank and by accounting for the time that has elapsed since the
previous issue, and the number of times the bank has issued bonds since 1970 (Model 6).
It also continues to hold when we estimate our model with bank ?xed effects to control
for potential differences across banks on the cost to access the bond market (Models 7-9),
and when with estimate it with weighted least squares, computing the weights off the
number of bonds each bank issued over the sample period, to account for differences in
bond-issuance activity across banks (Models 11-12).
In sum, all of our tests show that the risk structure of the credit spreads on bank
bonds shifts up but becomes steeper during recessions, suggesting that the market
does not discipline all banks equally. Instead it is more demanding with riskier banks
than with safer ones. This ?nding, however, is not the only possible interpretation of
our results since our interpretation relies on some assumptions. For example, if bond
ratings were not comparable over the business cycle then our ?ndings could have an
alternative explanation, namely that bonds issued during recessions are riskier than
bonds issued during expansions with the same credit rating and the risk difference
between these two cohorts of bonds increases with risk. Note that even when ratings
are comparable, if the safer ?rms in each rating bucket drop out of the market in
recessions and this effect increases with risk, this too could explain the marginal
impact of recessions on credit spreads that we identi?ed. To account for these and
other like explanations, we add a control group that is also susceptible to these
problems – the bonds issued by ?rms over the same time period – and investigate the
marginal impact of recessions on the difference between the spreads on bank bonds
and those on ?rm bonds with the same credit rating. Our null hypothesis is that if the
markets were to discipline all banks equally then the marginal impact of recessions on
that difference in spreads should not vary across bond issuers with different risk.
3.3 Credit spreads on bank and ?rm bonds
Table IV presents the results of our second test. Model 1 shows that in good times, on
average, the spreads on bank bonds are 4 basis points higher than the spreads on ?rm
bonds. This model also shows that while recessions increase the spreads of ?rm bonds
by 72 basis points, they increase the spreads of bank bonds by only 49 basis points.
Model 2 shows that this differential effect of recessions continues to hold when we
JFEP
1,1
116
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
C
o
n
s
t
a
n
t
0
.
5
2
8
(
1
4
.
5
2
9
*
*
*
)
0
.
5
3
2
(
1
4
.
1
1
7
*
*
*
)
0
.
6
6
0
(
1
7
.
4
5
0
*
*
*
)
2
.
4
9
7
(
1
5
.
9
3
5
*
*
*
)
2
.
4
9
1
(
1
6
.
1
7
0
*
*
*
)
5
.
9
0
0
(
2
6
.
4
9
3
*
*
*
)
A
a
0
.
1
1
3
(
2
.
9
7
6
*
*
*
)
0
.
0
4
9
(
1
.
2
3
1
)
2
0
.
0
3
8
(
2
0
.
9
5
4
)
0
.
0
6
3
(
1
.
8
3
9
*
)
0
.
0
8
0
(
2
.
1
7
7
*
*
)
0
.
1
9
1
(
5
.
1
7
0
*
*
*
)
A
0
.
3
0
0
(
8
.
1
7
4
*
*
*
)
0
.
3
0
5
(
7
.
9
8
7
*
*
*
)
0
.
1
9
5
(
5
.
0
1
4
*
*
*
)
0
.
3
1
4
(
9
.
4
9
2
*
*
*
)
0
.
3
6
3
(
1
0
.
1
8
2
*
*
*
)
0
.
5
0
8
(
1
4
.
2
1
0
*
*
*
)
B
a
a
0
.
8
3
5
(
2
1
.
6
9
7
*
*
*
)
0
.
8
3
8
(
2
0
.
9
7
5
*
*
*
)
0
.
6
8
6
(
1
6
.
9
6
4
*
*
*
)
0
.
7
6
7
(
2
1
.
9
9
9
*
*
*
)
0
.
8
4
9
(
2
2
.
6
5
5
*
*
*
)
1
.
0
0
0
(
2
6
.
8
6
1
*
*
*
)
B
a
2
.
3
1
5
(
4
2
.
9
3
7
*
*
*
)
2
.
3
1
8
(
4
2
.
0
9
5
*
*
*
)
2
.
1
4
2
(
3
8
.
2
3
1
*
*
*
)
2
.
0
5
7
(
4
0
.
1
6
4
*
*
*
)
2
.
1
2
5
(
4
1
.
3
6
2
*
*
*
)
2
.
3
4
1
(
4
7
.
1
5
4
*
*
*
)
B
4
.
1
3
6
(
8
7
.
1
3
4
*
*
*
)
4
.
1
3
3
(
8
5
.
2
4
1
*
*
*
)
4
.
0
0
4
(
8
1
.
5
7
8
*
*
*
)
3
.
7
2
8
(
6
8
.
8
1
9
*
*
*
)
3
.
8
3
3
(
6
9
.
0
8
5
*
*
*
)
4
.
1
1
0
(
7
4
.
1
6
8
*
*
*
)
C
a
a
5
.
9
5
7
(
2
9
.
1
4
0
*
*
*
)
5
.
9
5
3
(
2
9
.
0
8
9
*
*
*
)
5
.
7
9
7
(
2
7
.
8
3
5
*
*
*
)
5
.
2
9
5
(
2
5
.
7
1
8
*
*
*
)
5
.
5
8
6
(
2
6
.
5
5
3
*
*
*
)
5
.
9
5
0
(
2
9
.
0
9
0
*
*
*
)
B
k
0
.
0
3
7
(
2
.
3
3
1
*
*
)
2
0
.
0
5
4
(
2
0
.
5
8
5
)
2
0
.
1
7
5
(
2
1
.
7
0
3
*
)
2
0
.
2
3
8
(
2
2
.
3
1
7
*
*
)
2
0
.
2
4
7
(
2
2
.
3
1
8
*
*
)
2
0
.
2
8
3
(
2
2
.
5
5
4
*
*
)
B
k
A
a
0
.
2
4
5
(
2
.
5
9
8
*
*
*
)
0
.
3
5
6
(
3
.
3
5
8
*
*
*
)
0
.
4
5
7
(
4
.
3
4
0
*
*
*
)
0
.
3
3
4
(
3
.
0
5
1
*
*
*
)
0
.
3
1
9
(
2
.
8
0
4
*
*
*
)
B
k
A
0
.
0
4
7
(
0
.
5
1
1
)
0
.
1
4
9
(
1
.
4
2
8
)
0
.
3
3
2
(
3
.
1
5
1
*
*
*
)
0
.
3
8
6
(
3
.
5
1
6
*
*
*
)
0
.
5
1
7
(
4
.
5
6
5
*
*
*
)
B
k
B
a
a
0
.
0
0
8
(
0
.
0
7
9
)
0
.
0
9
2
(
0
.
8
1
3
)
0
.
4
2
6
(
3
.
6
7
1
*
*
*
)
0
.
5
3
0
(
4
.
3
7
9
*
*
*
)
0
.
5
8
4
(
4
.
8
0
1
*
*
*
)
B
k
B
a
2
0
.
3
7
6
(
2
1
.
2
5
4
)
2
0
.
2
0
6
(
2
0
.
6
7
8
)
0
.
2
0
2
(
0
.
7
2
8
)
0
.
4
0
2
(
1
.
5
1
3
)
0
.
3
2
7
(
1
.
1
9
9
)
R
e
c
0
.
7
1
8
(
2
2
.
8
0
7
*
*
*
)
0
.
7
1
9
(
2
2
.
8
6
3
*
*
*
)
0
.
2
2
1
(
2
.
7
1
4
*
*
*
)
0
.
2
1
5
(
3
.
2
2
0
*
*
*
)
0
.
2
7
4
(
3
.
8
8
2
*
*
*
)
0
.
1
1
3
(
1
.
4
3
4
)
R
e
c
A
a
0
.
1
4
6
(
1
.
4
7
2
)
0
.
1
4
3
(
1
.
7
6
3
*
)
0
.
0
5
8
(
0
.
6
9
3
)
2
0
.
0
2
2
(
2
0
.
2
2
9
)
R
e
c
A
0
.
3
5
2
(
4
.
0
6
2
*
*
*
)
0
.
3
3
4
(
4
.
6
3
2
*
*
*
)
0
.
2
7
2
(
3
.
6
1
5
*
*
*
)
0
.
1
3
3
(
1
.
5
9
3
)
R
e
c
B
a
a
0
.
7
7
0
(
7
.
6
4
0
*
*
*
)
0
.
6
9
8
(
8
.
0
8
6
*
*
*
)
0
.
6
2
6
(
7
.
0
3
5
*
*
*
)
0
.
4
4
9
(
4
.
6
1
9
*
*
*
)
R
e
c
B
a
1
.
0
6
7
(
5
.
9
9
7
*
*
*
)
0
.
9
8
7
(
6
.
0
9
1
*
*
*
)
0
.
9
6
2
(
6
.
0
0
3
*
*
*
)
0
.
7
8
9
(
4
.
8
4
7
*
*
*
)
R
e
c
B
0
.
5
1
0
(
2
.
9
3
9
*
*
*
)
0
.
4
7
5
(
2
.
8
8
5
*
*
*
)
0
.
4
3
4
(
2
.
6
5
5
*
*
*
)
0
.
2
1
2
(
1
.
2
7
9
)
R
e
c
C
a
a
1
.
5
4
7
(
1
.
4
8
2
)
1
.
6
2
0
(
1
.
5
6
0
)
1
.
5
6
2
(
1
.
5
6
8
)
1
.
2
5
7
(
1
.
2
5
5
)
R
e
c
B
k
2
0
.
2
3
1
(
2
4
.
6
2
8
*
*
*
)
2
0
.
2
4
7
(
2
4
.
8
2
8
*
*
*
)
0
.
2
0
0
(
1
.
5
8
7
)
0
.
5
1
1
(
4
.
2
9
4
*
*
*
)
0
.
5
2
5
(
4
.
2
4
8
*
*
*
)
0
.
5
4
0
(
3
.
9
1
5
*
*
*
)
R
e
c
B
k
A
a
2
0
.
3
0
1
(
2
2
.
0
6
9
*
*
)
2
0
.
7
2
9
(
2
5
.
3
4
1
*
*
*
)
2
0
.
7
3
5
(
2
5
.
1
2
5
*
*
*
)
2
0
.
6
4
6
(
2
4
.
0
3
3
*
*
*
)
R
e
c
B
k
A
2
0
.
2
8
5
(
2
2
.
0
3
1
*
*
*
)
2
0
.
6
1
8
(
2
4
.
4
9
8
*
*
*
)
2
0
.
5
9
4
(
2
4
.
2
1
0
*
*
*
)
2
0
.
7
3
2
(
2
4
.
7
2
3
*
*
*
)
R
e
c
B
k
B
a
a
2
0
.
0
6
9
(
2
0
.
3
3
3
)
2
0
.
3
0
1
(
2
1
.
4
0
4
)
2
0
.
2
4
5
(
2
1
.
1
3
7
)
2
0
.
2
8
8
(
2
1
.
3
0
7
)
O
t
h
e
r
i
s
s
u
e
r
f
e
a
t
u
r
e
s
P
u
b
l
i
c
c
o
m
p
a
n
y
2
0
.
1
3
1
(
2
6
.
9
9
5
*
*
*
)
2
0
.
1
3
6
(
2
7
.
3
5
8
*
*
*
)
2
0
.
1
3
3
(
2
7
.
8
7
3
*
*
*
)
F
i
r
s
t
i
s
s
u
e
0
.
1
9
1
(
6
.
0
2
4
*
*
*
)
O
r
d
e
r
o
f
i
s
s
u
e
0
.
0
0
1
(
0
.
0
7
0
)
T
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
0
.
0
0
7
(
1
.
5
7
8
)
I
s
s
u
e
d
e
s
i
g
n
C
a
l
l
p
r
o
v
i
s
i
o
n
0
.
2
6
2
(
1
1
.
1
5
2
*
*
*
)
0
.
2
3
9
(
1
0
.
3
7
8
*
*
*
)
0
.
1
4
4
(
6
.
3
8
5
*
*
*
)
P
u
t
p
r
o
v
i
s
i
o
n
2
0
.
6
9
9
(
2
1
3
.
5
0
3
*
*
*
)
2
0
.
6
8
0
(
2
1
3
.
4
3
0
*
*
*
)
2
0
.
5
3
2
(
2
1
2
.
2
8
8
*
*
*
)
S
i
n
k
i
n
g
f
u
n
d
0
.
3
0
4
(
8
.
5
4
8
*
*
*
)
0
.
2
6
8
(
7
.
7
4
5
*
*
*
)
0
.
1
3
6
(
3
.
9
3
5
*
*
*
)
(
c
o
n
t
i
n
u
e
d
)
Table IV.
Bank and ?rm-bond
spreads over treasury at
issue date
Do markets
“discipline” all
banks equally?
117
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
F
l
o
a
t
e
r
s
1
.
4
5
5
(
4
.
3
0
0
*
*
*
)
1
.
4
2
0
(
4
.
1
5
5
*
*
*
)
1
.
2
2
1
(
3
.
6
7
9
*
*
*
)
S
h
e
l
f
2
0
.
2
4
3
(
2
1
1
.
0
9
8
*
*
*
)
2
0
.
2
1
5
(
2
1
0
.
1
0
4
*
*
*
)
2
0
.
1
6
0
(
2
7
.
5
3
1
*
*
*
)
L
o
g
(
m
a
t
u
r
i
t
y
)
0
.
1
0
8
(
1
0
.
0
6
4
*
*
*
)
0
.
0
9
5
(
8
.
8
8
0
*
*
*
)
0
.
1
3
7
(
1
3
.
2
6
7
*
*
*
)
S
e
n
i
o
r
0
.
2
3
2
(
6
.
2
3
0
*
*
*
)
0
.
2
7
0
(
7
.
2
8
3
*
*
*
)
0
.
3
1
0
(
8
.
7
9
1
*
*
*
)
O
t
h
e
r
c
o
n
t
r
o
l
s
L
o
g
(
a
m
o
u
n
t
i
s
s
u
e
d
)
0
.
0
3
3
(
5
.
9
3
7
*
*
*
)
0
.
0
3
6
(
6
.
4
7
3
*
*
*
)
0
.
0
0
6
(
1
.
0
1
0
)
P
r
i
v
.
p
l
a
c
e
m
e
n
t
0
.
0
8
2
(
2
.
0
9
9
*
*
)
0
.
1
0
1
(
2
.
6
2
1
*
*
*
)
0
.
1
5
4
(
4
.
3
7
1
*
*
*
)
L
o
g
(
t
i
m
e
t
r
e
n
d
)
2
0
.
2
2
3
(
2
9
.
8
7
7
*
*
*
)
2
0
.
2
4
4
(
2
1
1
.
0
0
8
*
*
*
)
2
0
.
2
3
8
(
2
8
.
6
0
6
*
*
*
)
T
e
n
-
y
e
a
r
t
r
e
a
s
u
r
y
y
i
e
l
d
2
0
.
1
7
6
(
2
1
9
.
2
5
4
*
*
*
)
2
0
.
1
7
5
(
2
1
9
.
5
2
0
*
*
*
)
2
0
.
3
2
6
(
2
2
6
.
7
8
8
*
*
*
)
Y
i
e
l
d
c
.
s
l
o
p
e
2
0
.
2
0
0
(
2
1
2
.
7
2
8
*
*
*
)
2
0
.
1
9
8
(
2
1
2
.
7
8
0
*
*
*
)
2
0
.
2
2
2
(
2
1
1
.
3
1
5
*
*
*
)
S
p
l
i
t
0
.
3
8
5
(
1
3
.
8
6
7
*
*
*
)
0
.
3
2
8
(
1
2
.
4
2
7
*
*
*
)
S
P
.
M
o
o
d
2
0
.
7
1
1
(
2
1
8
.
3
9
3
*
*
*
)
2
0
.
7
0
0
(
2
1
9
.
4
4
6
*
*
*
)
A
d
j
u
s
t
e
d
R
2
7
5
.
5
4
7
7
5
.
5
7
4
7
5
.
7
7
0
7
8
.
3
2
5
7
8
.
9
4
1
8
1
.
7
4
9
L
o
g
l
i
k
e
l
i
h
o
o
d
2
1
4
,
9
2
6
.
6
2
1
4
,
9
1
8
.
1
2
1
4
.
8
6
6
.
4
2
1
4
,
2
1
4
.
4
2
1
4
,
0
4
6
.
2
2
1
3
,
2
1
2
.
2
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
a
t
t
h
e
*
1
0
,
*
*
5
,
*
*
*
1
p
e
r
c
e
n
t
l
e
v
e
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
;
m
o
d
e
l
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
a
r
e
h
e
t
e
r
o
s
k
e
d
a
s
t
i
c
-
c
o
n
s
i
s
t
e
n
t
;
t
o
t
a
l
n
u
m
b
e
r
o
f
o
b
s
e
r
v
a
t
i
o
n
s
1
1
,
5
8
7
;
t
-
S
t
a
t
i
s
t
i
c
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
t
h
e
d
e
p
e
n
d
e
n
t
v
a
r
i
a
b
l
e
i
s
t
h
e
b
o
n
d
c
r
e
d
i
t
s
p
r
e
a
d
o
v
e
r
t
h
e
t
r
e
a
s
u
r
y
w
i
t
h
t
h
e
s
a
m
e
m
a
t
u
r
i
t
y
;
s
p
r
e
a
d
s
a
r
e
c
o
m
p
u
t
e
d
a
t
i
s
s
u
e
d
a
t
e
;
r
e
c
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
i
s
s
u
e
d
d
u
r
i
n
g
a
r
e
c
e
s
s
i
o
n
a
c
c
o
r
d
i
n
g
t
o
N
B
E
R
;
p
u
b
l
i
c
c
o
m
p
a
n
y
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
i
s
s
u
e
r
i
s
a
p
u
b
l
i
c
c
o
m
p
a
n
y
;
?
r
s
t
i
s
s
u
e
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
t
h
e
?
r
m
’
s
?
r
s
t
b
o
n
d
i
s
s
u
e
s
i
n
c
e
1
9
7
0
:
1
;
o
r
d
e
r
o
f
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
t
i
m
e
s
t
h
e
?
r
m
i
s
s
u
e
d
b
o
n
d
s
s
i
n
c
e
1
9
7
0
:
1
;
t
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
y
e
a
r
s
s
i
n
c
e
t
h
e
?
r
m
m
a
d
e
i
t
s
l
a
t
e
s
t
b
o
n
d
i
s
s
u
e
;
c
a
l
l
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
c
a
l
l
a
b
l
e
;
p
u
t
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
b
o
n
d
h
o
l
d
e
r
s
c
a
n
s
e
l
l
t
h
e
b
o
n
d
b
a
c
k
t
o
t
h
e
c
o
m
p
a
n
y
p
r
i
o
r
t
o
m
a
t
u
r
i
t
y
;
s
i
n
k
i
n
g
f
u
n
d
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
h
a
s
a
s
i
n
k
i
n
g
f
u
n
d
;
s
h
e
l
f
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
a
s
h
e
l
f
i
s
s
u
e
;
m
a
t
u
r
i
t
y
,
m
a
t
u
r
i
t
y
o
f
t
h
e
b
o
n
d
i
n
y
e
a
r
s
;
a
m
o
u
n
t
i
s
s
u
e
d
i
n
m
i
l
l
i
o
n
s
o
f
U
S
d
o
l
l
a
r
s
;
i
s
s
u
e
s
d
e
?
a
t
e
d
b
y
t
h
e
c
o
r
e
u
r
b
a
n
c
o
n
s
u
m
e
r
C
P
I
w
i
t
h
t
h
e
a
v
e
r
a
g
e
o
f
1
9
8
2
-
1
9
8
4
¼
1
0
0
;
p
r
i
v
.
p
l
a
c
e
m
e
n
t
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
p
r
i
v
a
t
e
l
y
p
l
a
c
e
d
;
t
i
m
e
t
r
e
n
d
l
i
n
e
a
r
t
i
m
e
t
r
e
n
d
;
a
i
n
c
l
u
d
e
d
i
n
t
h
i
s
r
e
g
r
e
s
s
i
o
n
b
u
t
n
o
t
s
h
o
w
n
i
n
t
h
e
t
a
b
l
e
a
r
e
d
u
m
m
y
v
a
r
i
a
b
l
e
s
f
o
r
t
h
e
i
s
s
u
e
r
’
s
s
e
c
t
o
r
o
f
a
c
t
i
v
i
t
y
a
s
d
e
?
n
e
d
b
y
S
I
C
o
n
e
-
d
i
g
i
t
c
o
d
e
,
a
n
d
t
h
e
n
u
m
b
e
r
o
f
b
o
n
d
s
i
s
s
u
e
d
i
n
t
h
e
q
u
a
r
t
e
r
i
n
e
a
c
h
o
f
t
h
e
n
i
n
e
m
a
i
n
c
r
e
d
i
t
r
a
t
i
n
g
c
l
a
s
s
e
s
a
s
d
e
?
n
e
d
b
y
M
o
o
d
y
’
s
w
h
o
l
e
r
a
t
i
n
g
s
,
A
a
a
,
.
.
.
,
C
Table IV.
JFEP
1,1
118
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
allow for ?rm bonds to be priced differently than bank bonds. According to this model,
recessions increase the cost of bond ?nancing to ?rms and banks by 72 and 47 basis
points, respectively.
Model 3 investigates whether that differential effect of recessions applies to all banks
equally by allowing bank bonds of different risk issued in good and bad times to be priced
differently. As we noted in the methodology section, we are particularly interested in the
variable Rec Bk rating of this model as it tells us whether the slope of the credit spread
curve on bank bonds changed differently than the slope of the same curve for ?rm bonds
during recessions. According to the results of that, effectivelythe slope of the credit spread
curve on bank bonds becomes relatively steeper than that on ?rm bonds during
recessions. This difference arises because recessions have the same impact on ?rms and
banks that issue Baa-rated bonds (the coef?cient on Rec Bk Baa is not statistically
different fromzero), but theyhave a signi?cantlysmaller impact onbanks that issue either
double or single-A rated bonds than ?rms that issue bonds with these same ratings (the
coef?cients on Rec Bk Aa and Rec Bk Aare negative and statistically different fromzero).
This differential impact of recessions continues to hold after we control for the
standard factors that researchers have found help explain bond spreads (Model 4). One
noticeable difference is that the variable Rec Bk, which measures the differential
impact of recessions on triple-A rated bank bonds and ?rm bonds with the same rating,
is now statistically signi?cant. Given that there are only nine bank bonds with a
triple-A rating in the sample, this ?nding does not appear to be meaningful and so we
chose to continue to focus our analysis on investment-grade bonds other than triple-A
rated bonds[15]. As for the new controls we added to the regression, they produced
similar results when we included them in our regressions on bank bonds. There are,
however, three differences that are worth noting. First, we now ?nd a secular decline in
the spreads of bonds. Second, consistent with the literature, we ?nd too that public
companies pay lower credit spreads than private companies. Finally, we ?nd that the
slope of the treasury yield curve affects negatively the spreads on ?rm bonds. In our
previous analysis of bank bonds none of these variables were statistically signi?cant.
Models 5 and 6 present the results of two additional robustness tests. Model 5
accounts for the potential impact of disagreements between rating agencies on the
rating of the bond, and Model 6 investigates the impact of controlling for the
acquaintance of investors with bond issuers. The results of the former model con?rm
our earlier ?ndings that disagreements between rating agencies increase the cost of
accessing the bond market, but only when the issuer gets a rating from Moody’s which
is more favorable than the rating it gets from S&P. The results of the latter model, in
turn, show that the ?rst time an issuer comes to the market it pays a premium, but
issuing frequently does not bring the costs of bond ?nancing down. Importantly,
though the results of both models continue to show that recessions steepen the slope
of the credit spread curve on banks bonds, that is, they have a relatively large impact
on the cost of accessing the bond market for Baa-rated banks than for safer banks, say
banks that issue double or single-A rated bonds (when compared to the similar effect
on ?rms). This difference is shown in Figure 1, which shows the estimated marginal
impact of recessions on the cost of accessing the bond market for ?rms and banks with
the same credit rating as well as the difference between these marginal costs[16].
Since these results hold when we consider bank bonds alone and when we also
account for bonds issued by ?rms, this suggests that they are driven by factors other
Do markets
“discipline” all
banks equally?
119
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
than differences in the risk associated with credit ratings over the business cycle or
possible patterns of bank-bond issuance over the business cycle. Instead, a possible
explanation for our results is that in recessions there is a ?ight toward bank bonds, and
in particular toward safer banks. This suggests that the market does not discipline all
banks equally – it appears to be relatively less demanding on safer banks. Thus, far,
we have shown that these differences in the funding costs matter in a statistical sense.
Next, we investigate if they also matter from an economic point of view.
3.4 Economic signi?cance
Based on the estimates of Model 6 of Table IV, we have that in good times the estimated
difference between the spreads of Baa-rated bank bonds and the spreads of double-A
bank-rated bonds is 70 basis points. Still according to this model, recessions increase the
spreads of the former bonds by 51 basis points and those of the latter bonds by 10 basis
points. As a result, during recessions the difference between the spreads of Baa bank
bonds and those of double-A bank bonds increases by 41 basis points.
Under these conditions, a policy requiring banks to issue, say, $100 million, at a
certain frequency can have an important impact on the relative funding costs of banks.
Consider the case of the two banks, one rated double-A and other one rated Baa, and
assume that the credit rating of these banks does not change with the state of the
economy. In good times, that policy costs the Baa-rated bank $700,000 more than it
does the double-A rated bank. In recessions, however, this difference in funding costs
does not stay the same, even though the two banks continue to have their same ratings.
Instead, it increases to $1,110,000. As a result, in recessions the funding costs of the
Baa-rated bank increase by $410,000 relative to the funding costs of the double-A rated
bank. This suggests that the difference in the way the market disciplines banks of
different risk, which we identi?ed above, is not only statistically signi?cant but is also
important from an economic point of view.
4. Final remarks
In this paper, we showed that credit spreads on bank bonds vary not only across banks
of different risk but also with the state of the economy at the time of the issue. We also
showed that the impact of the state of the economy is not uniform across issuers of
Figure 1.
Cost of recessions to banks
and ?rms by credit rating
–0.4
–0.2
0
0.2
0.4
0.6
0.8
1
Aa A Baa
Firms Banks Banks minus firms
JFEP
1,1
120
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
different risk. These ?ndings imply that the information content of credit spreads on
bonds of banks varies both with the risk if the bank and the state of the economy.
These results are relevant for bank regulators and policymakers. They suggest, for
instance, that the proposals which have been put forth to introduce prompt corrective
action programs triggered by the credit spreads on bank bonds need to be
parameterized not only on the risk of the issuer, but also on the state of the economy if
they are to produce similar levels of monitoring across banks[17]. Similarly, proposals
to make bond issuance by banks mandatory need to account for this difference in the
market discipline across banks, otherwise they will alter the relative funding costs of
banks and thus promote different levels of monitoring across banks.
Notes
1. Diamond and Dybvig (1983) explain the role of demand deposits for the liquidity insurance
provision, and Diamond (1984) explains the role of loans for the monitoring services
provision. Calomiris and Kahn (1991), Flannery (1994) and Diamond and Rajan (1998)
explain the advantages of combining these two functions.
2. See Flannery (1998) and Kwast et al. (1999) for a review of the literature on the market
discipline of banks.
3. Bliss and Flannery (2002) do not ?nd that market discipline in?uences bank management
policies.
4. See Morgan and Stiroh (2005) for evidence of this subsidy.
5. Moody’s states “As a rule of thumb, we are looking through the next economic cycle or
longer. Because of this, our ratings are not intended to ratchet up and down with business or
supply-demand cycles [. . .]”.
6. Santos (2006) shows that it is costlier to access the bond market in recessions, and the
additional cost that comes with recessions does not affect issuers uniformly.
7. See Calomiris (1999) and Evanoff and Wall (2000) for proposals to make it mandatory for
banks to issue subordinated debt.
8. Evanoff and Wall (2000, 2001) propose the use of credit spreads on bank bonds over
treasuries instead of the capital to assets ratio as the trigger variable in the prompt
correction action scheme in place in the USA on the grounds that credit spreads are less
prone to manipulation by bank managers and are better predictors of bank risk.
9. We do not consider bonds issued prior to the second quarter of 1982 because Moody’s started
to use alpha-numeric ratings only in April of 1982 and use this information in our
methodology.
10. Between 1982:2 and 2002:2, the peaks occurred in (months in brackets): 1990(July)
and 2001(March), and the troughs occurred in: 1982(November), 1991(March) and
2001(November).
11. As we saw in the previous subsection, the lowest rated bonds issued by banks have a
Ba Moody’s rating. We have, however, bonds issued by ?rms with a Moody’s rating equal to
Caa as well as bonds with an even lower credit rating. Since the number of bonds with a
rating at issue date lower than Caa is rather small we chose to pool all of these bonds
together under the credit-rating dummy R
6
.
12. An important advantage of this approach to control for the creditworthiness of bonds over
the alternative approach sometimes adopted by bond pricing models, which uses a linear
variable to capture the rating of bonds, is that it does not assume that each unit change in
ratings has the same effect on credit spreads.
Do markets
“discipline” all
banks equally?
121
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
13. See, for example, Blackwell and Kidwell (1988), Fenn (2000), Collin-Dufresne et al. (2001),
Harrison (2001) and Santos (2006).
14. See Santos (2006) for evidence on the impact of rating splits on the cost of ?rms’ access to the
bond market.
15. The number of bank bonds in our sample with the highest below grade rating, Ba, is not
meaningful.
16. The estimates shown in Figure 1 were computed based on Model 6 of Table IV. We have
excluded from this ?gure triple-A rated bonds because there are only nine bank bonds in the
sample with this rating, and Ba-rated bonds because there are no bank bonds with this
rating that were issued during recessions.
17. Note that parameterizing these programs on spreads of bank bond computed over, say,
triple-A rated bonds, as is sometimes advocated, will not solve this problem because as we
saw in recessions the credit spread (over treasuries) curve on bank bonds not only shifts up
but it also becomes steeper.
References
Avery, R.B., Belton, T.M. and Goldberg, M.A. (1988), “Market discipline in regulating bank risk:
new evidence from the capital markets”, Journal of Money and Credit and Banking, Vol. 20
No. 4, pp. 597-610.
Blackwell, D.W. and Kidwell, D.S. (1988), “An investigation of cost differences between public
sales and private placements of debt”, Journal of Financial Economics, Vol. 22, pp. 253-78.
Bliss, R.R. and Flannery, M.J. (2002), “Market discipline in the governance of US bank holding
companies: monitoring vs. in?uencing”, European Finance Review, Vol. 6 No. 3, pp. 361-95.
Calomiris, C.W. (1999), “Building and incentive compatible safety net”, Journal of Banking &
Finance, Vol. 23, pp. 1499-519.
Calomiris, C.W. and Kahn, C.M. (1991), “The role of demandable debt in structuring optimal
banking arrangements”, American Economic Review, Vol. 81 No. 3, pp. 497-513.
Collin-Dufresne, P., Goldstein, R.S. and Martin, J.S. (2001), “The determinants of credit spreads
changes”, Journal of Finance, Vol. 56 No. 6, pp. 2177-207.
Diamond, D.W. (1984), “Financial intermediation and delegated monitoring”, Review of Economic
Studies, Vol. 51, pp. 393-414.
Diamond, D.W. and Dybvig, P.H. (1983), “Bank runs, deposit insurance and liquidity”, Journal
of Political Economy, Vol. 91, pp. 401-19.
Diamond, D.W. and Rajan, R.G. (1998), “Liquidity risk, liquidity creation and ?nancial fragility:
a theory of banking”, Journal of Political Economy, Vol. 109 No. 2, pp. 287-327.
Evanoff, D.E. and Wall, L.D. (2000), “Subordinated debt and bank capital reform”, in Kaufman,
G.G. (Ed.), Bank Fragility and Regulation: Evidence from Different Countries and Different
Times, JAI Press, Greenwich, CT, pp. 53-120.
Evanoff, D.E. and Wall, L.D. (2001), “Sub-debt yield spreads as bank risk measures”, Journal of
Financial Services Research, Vol. 20 No. 2, pp. 121-45.
Fenn, G.W. (2000), “Speed of issuance and the adequacy of disclosure in the 144A high-yield debt
market”, Journal of Financial Economics, Vol. 56, pp. 383-405.
Flannery, M.J. (1994), “Debt maturity and the deadweight cost of leverage: optimally ?nancing
banking ?rms”, American Economic Review, Vol. 84, pp. 320-31.
JFEP
1,1
122
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Flannery, M.J. (1998), “Using market information in prudential bank supervision: a review of the
US empirical evidence”, Journal of Money and Credit and Banking, Vol. 30 No. 3,
pp. 273-305.
Flannery, M.J. and Sorescu, S.M. (1996), “Evidence of bank market discipline in subordinated
debenture yields: 1983-1991”, Journal of Finance, Vol. 51 No. 4, pp. 1347-77.
Gorton, G. and Santomero, A.M. (1990), “Market discipline and bank subordinated debt: note”,
Journal of Money and Credit and Banking, Vol. 22 No. 1, pp. 119-28.
Harrison, P. (2001), “The impact of market liquidity in times of stress on the corporate bond
market: pricing, trading, and the availability of funds during heightened illiquidity”,
Mimeo, Board of Governors of the Federal Reserve System, Washington, DC.
Jagtiani, J., Kaufman, G. and Lemieux, C. (2002), “Do markets discipline banks and bank holding
companies? Evidence from the post-FDICIA period”, Journal of Financial Research, Vol. 25
No. 4, pp. 559-75.
Kwast, M.L., Covitz, D.M., Hancock, D., Houpt, J.V., Adkins, D.P., Barger, N., Bouchard, B.,
Connolly, J.F., Brady, T.F., English, W.B., Evanoff, D.D. and Wall, L.D. (1999), “Using
sub-debt as an instrument of market discipline”, Staff Study 172, Board of Governors of
the Federal Reserve System, Washington, DC.
Morgan, D.P. (2000), “Rating banks: risk and uncertainty in an opaque industry”, American
Economic Review, Vol. 92 No. 4, pp. 874-88.
Morgan, D.P. and Stiroh, K.J. (2005), “‘Too big to fail’ and market discipline of banks:
a cross-sector study of investor perceptions”, Staff Reports 220, Federal Reserve Bank of
New York, New York, NY.
Santos, J.A.C. (2006), “Why ?rm access to the bond market varies over the business cycle:
a theory and some evidence”, Journal of Banking & Finance, Vol. 30 No. 10, pp. 2715-36.
Sironi, A. (2003), “Testing for market discipline in the European banking industry: evidence from
subordinated debt issues”, Journal of Money Credit and Banking, Vol. 35 No. 3, pp. 443-72.
Corresponding author
Joa˜o A.C. Santos can be contacted at: [email protected]
Do markets
“discipline” all
banks equally?
123
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
This article has been cited by:
1. Martin Byford, Sinclair Davidson. 2013. Informing Depositors: A Proposed Reform Of The
Australian Banking Sector. Economic Papers: A journal of applied economics and policy 32:10.1111/
ecpa.2013.32.issue-4, 417-425. [CrossRef]
2. Luc Laeven. 2011. Banking Crises: A Review. Annual Review of Financial Economics 3, 17-40. [CrossRef]
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_767855241.pdf
The purpose of this paper is to investigate whether the bond market disciplines all banks
equally, in the sense of demanding the same relative risk premium across banks of different risk over
the business cycle.
Journal of Financial Economic Policy
Do markets “discipline” all banks equally?
J oa˜o A.C. Santos
Article information:
To cite this document:
J oa˜o A.C. Santos, (2009),"Do markets “discipline” all banks equally?", J ournal of Financial Economic
Policy, Vol. 1 Iss 1 pp. 107 - 123
Permanent link to this document:http://dx.doi.org/10.1108/17576380910962402
Downloaded on: 24 January 2016, At: 21:35 (PT)
References: this document contains references to 23 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 2388 times since 2009*
Users who downloaded this article also downloaded:
Anand Kumar J aiswal, (2008),"Customer satisfaction and service quality measurement in Indian
call centres", Managing Service Quality: An International J ournal, Vol. 18 Iss 4 pp. 405-416 http://
dx.doi.org/10.1108/09604520810885635
J eannette A. Colyvas, Walter W. Powell, (2009),"Measures, metrics, and myopia: The challenges
and ramifications of sustaining academic entrepreneurship", Advances in the Study of
Entrepreneurship, Innovation & Economic Growth, Vol. 19 pp. 79-111http://dx.doi.org/10.1108/
S1048-4736(2009)0000019004
J osé C. Alves, Kathi J . Lovelace, Charles C. Manz, Dmytro Matsypura, Fuminori Toyasaki, Ke (Grace) Ke,
(2006),"A cross-cultural perspective of self-leadership", J ournal of Managerial Psychology, Vol. 21 Iss 4 pp.
338-359http://dx.doi.org/10.1108/02683940610663123
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Do markets “discipline”
all banks equally?
Joa˜o A.C. Santos
Research Department, Federal Reserve Bank of New York,
New York, New York, USA
Abstract
Purpose – The purpose of this paper is to investigate whether the bond market disciplines all banks
equally, in the sense of demanding the same relative risk premium across banks of different risk over
the business cycle.
Design/methodology/approach – To test this hypothesis, the paper compares the difference
between the credit spreads in the primary market of bank and ?rm bonds with the same credit rating
issued during expansions with that same difference of spreads for bonds issued during recessions.
Findings – The paper ?nds that during recessions investors demand higher risk premiums.
Importantly, the paper ?nds that the impact of recessions is not uniform across banks – it affects
riskier banks more than safer ones. In other words, in recessions investors are relatively more
demanding on riskier banks than on safer ones.
Originality/value – These ?ndings are novel. They also have important policy implications because
they show that a bond-issuance policy aimed at promoting market discipline could affect the relative
funding costs of banks over the business cycle. They also indicate that the information which can
be extracted fromthe credit spreads on bank bonds varies across banks for reasons unrelated to their risk.
Keywords Banks, Risk analysis, Financial markets
Paper type Research paper
1. Introduction
Growing complexity of banks coupled with the development of debt markets has
increased the interest in the potential market discipline of banks. This interest has
motivated many empirical studies, most of them investigating whether the yields on
bank bonds are correlated with the issuers’ risk of default. This paper adds to this
literature by investigating if the market always charges the same relative risk premium
across banks of different riskiness. This is an important issue because otherwise a
bond-issuance policy aimed at promoting market discipline will affect the relative
funding costs of banks. Further, the information that can be extracted from the credit
spreads on bank bonds will vary across banks for reasons unrelated to their risk.
Banks’ simultaneous provision of liquidity insurance and monitoring services leads
to a mismatch between liquid liabilities and illiquid assets[1]. In the event of a liquidity
shock, the same information asymmetries that lead banks to adopt this asset and
liability structure make it dif?cult for them to borrow the necessary funds in the
market. As a result, they may be forced into bankruptcy. The premature liquidation of
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JEL classi?cation – E44, G32
The author thanks Adam Ashcraft, Scott Frame, and seminar participants at the Federal
Reserve Bank of New York for useful comments and suggestions, and Chris Metli and Kyle Lewis
for outstanding research assistance. The views stated herein are those of the author and are not
necessarily the views of the Federal Reserve Bank of New York or the Federal Reserve System.
Do markets
“discipline” all
banks equally?
107
Journal of Financial Economic Policy
Vol. 1 No. 1, 2009
pp. 107-123
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380910962402
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
bank assets is costly because it ends valuable relationships and it may develop into a
bank panic that culminates in a system failure. This risk of a system failure underpins
the classical argument for government interference in the business of banking.
Government interference materializes in different forms, including the introduction
of regulations covering such things as capital requirements and permissible activities;
the establishment of public institutions to perform supporting functions to banks,
including the provision of deposit insurance and lending of last resort; and the
establishment of public institutions like supervisors to monitor banks’ activities. As
banking organizations have evolved within the increasingly global and complex
?nancial markets, it has become dif?cult for the institutions responsible for supporting
and overseeing banks to meet their mandates. For this reason, they have expressed
growing interest in the use of market-related oversight to supplement their own
methods of monitoring.
Motivated by this interest, researchers have studied whether markets provide any
discipline to banks by investigating whether the credit spreads of bonds issued by
banks varied with their risk[2]. Early studies, including Avery et al. (1988) and Gorton
and Santomero (1990), ?nd that credit spreads on bank subordinated debentures were
virtually unrelated to traditional accounting measures of bank risk. Subsequent
studies, including Flannery and Sorescu (1996), Jagtiani et al., 2002 and Sironi (2003),
however, ?nd a relationship between the credit spreads on bank subordinated
debentures and risk[3].
To date, the literature on the market discipline has investigated extensively whether
credit spreads on uninsured bank debt are correlated with bank risk, but it has paid only
limited attention to such closely related issues as the pricing of bank risk, with the
exception, perhaps, of the “subsidy” that the so-called “too-big-to-fail” banks receive
from investors[4]. Our paper contributes to ?ll this gap in the literature by investigating
if the market charges all banks the same “price” of risk. Speci?cally, we investigate if
investors always demand the same relative risk premium across banks of different risk,
in other words, if the slope of the credit-spreadcurve of bank bonds is constant over time.
To test this hypothesis, we ?rst compare the credit spreads (over treasuries) in the
primary market of bonds issued by banks during expansions with the spreads of
bonds with the same credit rating issued during recessions. We ?nd that, ceteris
paribus, during recessions investors demand higher risk premiums. Importantly, the
impact of recessions is not uniform across banks, instead it affects riskier banks more
than safer ones. Implicit in the ?rst test, and following rating agencies’ claim that their
ratings are comparable over the business cycle, is the assumption that a bond issued
during an expansion and which receives a given credit rating carries the same risk as a
bond issued during a recession that carries the same rating[5]. Also implicit in that test
is the assumption that access conditions to the bond market do not vary over the
business cycle in such a way that make it relatively more dif?cult for some issuers to
access it, for example, during recessions[6].
Since violations of these assumptions could explain the results of our ?rst test, we
perform a second test in which we expand our sample to include bonds issued by ?rms
and compare the difference between the bank and ?rm credit spreads of bonds with the
same rating issued during expansions with that same difference but for bonds issued
during recessions. As we found for banks, we ?nd that recessions increase the cost
to access the bond market for riskier ?rms by more than it does for safer ?rms.
JFEP
1,1
108
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Further, we ?nd that the marginal impact of recessions on the cost to access the bond
market is the same for banks and ?rms that have a Baa credit rating. Importantly, we
?nd that recessions increase the cost of accessing the bond market to safer banks by
less than it does to ?rms with the same risk.
These results con?rm the ?ndings of our ?rst test that the market does not
discipline all banks equally – it is relatively more demanding on riskier banks than on
safer ones. This difference is important. It implies, for example, that a bond-issuance
policy aimed at promoting market discipline will make funding relatively more
expensive for riskier banks than for safer banks, a difference which will be ampli?ed
during recessions[7]. It also implies that the information content in credit spreads of
bank bonds not only depends on the risk of the issuer but also on the state of the
economy. As a result, the implementation of a monitoring program, like a prompt
correction scheme, parameterized on credit spreads on bank bonds, would need to take
into account both the riskiness of the issuer and the state of the economy in order to
promote similar levels of monitoring across banks over time[8].
The remainder of the paper is organized as follows. Section 2 presents our data and
methodology. Section 3 presents the results of our tests. Section 4 concludes the paper.
2. Data and methodology
2.1 Data
The data for this paper came from SDC’s Domestic New Bond Issuances Database. Our
sample of bonds includes only bonds issued in the USA in US dollars by American
commercial banks and non?nancial companies between 1982:2 and 2002:2[9]. However,
we use information on bonds issued since 1970:1 to identify the ?rst time ?rms issued
bonds and to measure the frequency ?rms have issued bonds over time.
Our sample of bonds includes both shelf and non-shelf issues, and bonds issued in
the public market as well as those privately placed. We exclude asset-backed bonds
and convertible bonds. We also exclude bonds for which we did not have the necessary
information to compute their credit spreads over the treasury at issue date, and bonds
with maturities longer than 30 years. Finally, we exclude bonds that do not have
ratings from both Moody’s and Standard & Poor’s (S&P) at issue date. These criteria
left us with a sample of 11,587 bonds, of which 1,537 were issued by commercial banks.
We use NBER’s identi?cation of troughs and peaks to ?nd out if a bondwas issued in a
recession or in an expansion. We de?ne a recession as the time period between a peak and
a trough, and an expansion as the time period between a trough and a peak. We classify a
quarter to be a recession (expansion) if either all months or the majority of months in the
quarter are in a recession (expansion) period. According to this classi?cation, there are 72
quarters of expansion and nine quarters of recession during the period 1982:2-2002:2[10].
Of the 1,537 bonds issued by banks, 12 per cent were issued in recessions and of the 10,050
bonds issued by non?nancial ?rms 10 per cent were issued in recessions.
Table I characterizes our samples of ?rm and bank bonds. There are some important
differences between our two samples of bonds. For example, there are far more risky
?rm bonds than bank bonds. While 31 per cent of the ?rm bonds are below investment
grade, less than 1 per cent of bank bonds have a speculative rating. Further, while
2 per cent of the ?rm bonds have a triple-A rating only 0.5 per cent of the bank bonds
have that rating. These differences suggest that when we compare bank bonds with ?rm
bonds we should focus on investment grade bonds that are not triple-A rated.
Do markets
“discipline” all
banks equally?
109
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Another important difference between the two samples of bonds is the frequency of
split ratings between the two major US rating agencies, Moody’s and S&P, which
Morgan (2000) argues is due to banks being more opaque than ?rms. These agencies
disagreed on 15 per cent of the ratings they assigned to bonds issued by ?rms. In
contrast, they disagreed on 34 per cent of the ratings they assigned on bonds issued by
banks. Still based on Table I, we ?nd that ?rms issue far more bonds with a sinking
fund and a put provision than banks do. They also issue more callable bonds and
senior bonds than banks do. Firms, in addition, do substantially more private
placements than banks do. In contrast, banks do more shelf issues.
Firm bonds Bank bonds
Variables 1982:2-2002:2 Expansions Recessions 1982:2-2002:2 Expansions Recessions
Volume and number of issues
Amount issued
a
11,653.56 10,204.21 1,449.35 1,667.64 1,497.27 170.37
Number of issues 10,050 9,080 970 1,537 1,349 188
Shares of the number of issues by credit rating
b
Aaa 1.930 1.586 5.155 0.520 0.519 0.532
Aa 8.318 8.139 10.000 35.784 34.841 42.553
A 32.179 31.189 41.443 47.885 48.258 45.213
Baa 26.030 26.267 23.814 15.029 15.493 11.702
Ba 8.617 8.733 7.526 0.791 0.890 0.000
B 21.791 22.863 11.753 0.000 0.000 0.000
Below B 1.134 1.289 0.309 0.000 0.000 0.000
Shares of the number of issues with split ratings
Split 15.423 15.529 14.433 34.027 33.062 40.957
SP . Mood 56.387 56.429 56.383 40.344 41.588 40.135
Shares of the number of issues by design of the issue
Callable bonds 36.607 36.905 33.814 23.552 21.201 40.426
Bonds with
a sinking
fund 8.060 8.150 7.217 0.260 0.148 1.064
Senior bonds 84.617 83.921 91.134 67.534 65.678 80.851
Shelf bonds 59.602 59.548 60.103 73.910 73.536 76.596
Private placements 20.985 20.595 24.639 0.065 0.074 0.000
Average maturity
(years) 10.466 10.567 9.519 8.980 8.722 10.837
Bonds with
a putable provision
on maturity
c
1.622 1.762 0.309 0.716 0.815 0.000
Shares of the number of issues by issuer features
First issue 26.886 26.355 13.711 9.109 9.562 5.861
Public company 68.706 68.480 70.825 79.245 78.206 86.702
Notes: It includes all new non-convertible bonds issued by American non?nancial ?rms and
commercial banks in the USA in US dollars over the period 1982:2-2002:2 that had ratings from both
Moody’s and S&P, information on amount issued, maturity and yield to maturity; recessions and
expansions de?ned according to NBER;
a
millions of US dollars, issues de?ated by the core urban
consumer price index (CPI) with the average of 1982-1984 ¼ 100;
b
shares computed based on Moody’s
ratings;
c
bondholders have the option of selling the bond back to the issuing ?rm before the maturing
date
Table I.
Sample characterization
JFEP
1,1
110
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
2.2 Methodology
The methodology we use in this paper has two parts. The ?rst part focuses on
banks and attempts to test the following hypothesis: if the market disciplines all banks
equally, then the impact of a change in the state of the economy on the cost to access
the bond market should not vary across bonds of different creditworthiness. Under
this hypothesis, we would have that the marginal impact of recessions on bond spreads
at issue date would be independent from the issuer’s risk. To test this hypothesis, we
estimate the following model of (bank) bond spreads:
Spread ¼ c þa
i
X
6
i¼1
Rat
i
þb Rec þg
i
Rec
X
6
i¼1
Rat
i
þc
i
X
L
i¼1
X
i
þ1 ð1Þ
where Spread is the bond’s spread over treasury at issue date. We proxy for the
creditworthiness of the bank by the credit rating of its bonds. We de?ne dummy
variables Rat
i
for each credit rating that Moody’s can assign a bond using the
following convention: R
1
equals one for double-A rated bonds; R
2
equals one for
single-A rated bonds; R
3
equals one for Baa-rated bonds, R
4
equals one for Ba-rated
bonds, R
5
equals one for single-B rated bonds and ?nally R
1
equals one for bonds with
a Caa or lower credit rating[11]. Under this criterion, the bonds left out are those with a
triple-A credit rating at issue date[12].
We control for the state of the economy at the time the bond is issued by including a
dummy variable Rec, which takes the value one if the bond is issued during a
recession. Recessions and expansions are denned according to NBER. Since, we want
to ascertain if recessions affect banks of different creditworthiness differently, we
include in our model of bond spreads the interaction of the recession dummy, Rec, with
the bond-rating dummies Rat
i
. If the market were to discipline all banks equally we
would expect our estimates of the coef?cients g
i
to be all statistically insigni?cant. In
this case, if b is positive we would have that recessions make it equally more expensive
for all banks to access the bond market.
We estimate these marginal effects controlling for a set of factors related to the
design of the bond and a set of characteristics of the issuer, which other studies of bond
pricing have shown help explain bond credit spreads[13]. We include dummy variables
to control for private placements, shelf and callable bonds, bonds with a put option,
and bonds with a sinking fund. We also control for the maturity of the bond and for the
amount issued. Finally, following the evidence that rating splits affect credit spreads,
we include a dummy variable, Split, which takes the value one if the bond received a
rating from S&P which differs from the rating it received from Moody’s[14]. In
addition, we include a dummy variable SP . Mood, which takes the value one if the
bond received a more favorable rating from S&P than from Moody’s. With respect to
the issuer, we control for the ?rst bond issued by the bank, the number of times the
bank issued in the past, and the length of time since the bank’s last bond issue.
Finally, we consider a set of other controls, including a time trend, and the total
number of bonds issued in each quarter in the investment and speculative grade
markets to control for a potential market segmentation resulting, for example, from
regulations. In addition, we control for the ten-year constant treasury yield and the
slope of the yield curve (difference between the 30- and ?ve-year yields).
We ?rst estimate our model of bond pricing with the ordinary least squares (OLS)
correcting the standard errors for heteroskedasticity. We then reestimate this model
Do markets
“discipline” all
banks equally?
111
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
with bank ?xed effects to account for potential systematic differences in the pricing of
bonds across banks that may arrive from such things as the so-called “too-big-to-fail”
status. In addition, to account for the disproportional level of bond issuance across the
banks in our sample we reestimate our ?xed-effects model with weighted least squares,
computing the weights based on the number of bonds each bank issued over the
sample period.
As we noted earlier, implicit in this ?rst test is the assumption, that ratings are
com-parable over the business cycle (as claimed by rating agencies). If that were not
the case, then our null hypothesis would be violated for reasons that have nothing to do
with market discipline. To reduce concerns with this limitation of our ?rst test, in
the second part of our methodology we add a control group – the bonds issued by
?rms – that could also be subject to these same problems. In this part of our
methodology, we want to ?nd out if the difference between the spread at issue date of a
bank bond and that of a ?rm bond with the same rating varies over the business cycle.
We also want to ascertain if that difference between bank and bond spreads varies
across bonds of different creditworthiness. To this end, we estimate the following
model of bond pricing:
Spread ¼ c þa
i
X
6
i¼1
Rat
i
þb Rec þg
i
Rec
X
6
i¼1
Rat
i
þhBk þz
i
Bk
X
6
i¼1
Rat
i
þf Rec Bk þd
i
Rec Bk
X
6
i¼1
Rat
i
þc
i
X
L
i¼1
X
i
þ1;
ð2Þ
where Spread, Rat
i
, Rec and X
i
are de?ned as in our ?rst test. The key new variable of
our second test is the dummy variable Bk, which takes the value one for bonds issued
by banks. We also include in our new model of bond spreads, the interaction of this
variable with the credit-rating and recession dummy variables as well as with the
interaction of the latter two variables to allow for different pricing of bank and ?rm
bonds with the same credit rating in good times as well as in recessions.
Under the null hypothesis that the market disciplines all banks equally we would
expect the relative difference between bank and ?rm bond spreads across bonds of
different credit-worthiness not to vary over the business cycle. In terms of the
parameters of our new model of bond pricing, this suggests that the coef?cients d
i
should not be statistically different from zero.
3. Results
3.1 Credit spreads on bank bonds
Table II characterizes the credit spreads at issue on the bonds issued by American
commercial banks in the USA over the period 1982:2-2002:2. We compute these spreads
as the percentage point difference between the yield to maturity of the bond and the
yield on an equivalent maturity US treasury bond. The average credit spread for the
1,537 bank bonds in our sample is 95 basis points. As expected, bonds with lower
credit ratings carry higher credit spreads. While on average a triple-A credit rating
bond carries a credit spread of 54 basis points, a Ba-rated bond, the lowest rated bonds
in our sample of bonds issued by banks, carries a credit spread of 242 basis points.
Comparing the credit spreads at issue date of bonds issued during expansions with
those of bonds issued during recessions, we ?nd that on average recessions increase
JFEP
1,1
112
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the cost of bond issuance by 44 basis points. Note, however, that recessions do not
affect the credit spreads of all bonds equally. On average, they tend to have a larger
impact on riskier bonds. Recessions do not affect the cost of accessing the bond market
for banks that issue triple-A rated bonds and it increases the cost of this funding source
for banks that issue double-A rated bonds by only 27 basis points. In contrast,
recessions increase the cost of bond ?nancing for banks that issue Baa-rated bonds by
112 basis points.
These results suggest that the risk structure of credit spreads on bank bonds
become steeper during recessions, an indication that investors impose higher costs on
riskier banks in recessions when it comes to market funding than they do on safer
banks. We investigate next if this difference continues to hold when we account for
other variables that also explain bond credit spreads.
3.2 The risk structure of credit spreads on bank bonds
Table III presents the results of our ?rst test on the impact of recessions on the risk
structure of credit spreads on bank bonds. Recall that under this test we consider only
bonds issued by banks and try to ascertain if the marginal impact of recessions on the
cost of bond issuance is constant across issuers of different risk. The ?rst three models
were estimated without any controls. These models con?rm the insights of our
univariate analysis of bond credit spreads: the lower the bond credit rating the higher
the credit spread (Model 1); on average, recessions increase spreads of bank bonds by
about 47 basis points (Model 2); the impact of recessions is not uniform across
issuers of different risk, instead, they have a larger impact on riskier issuers than on
safer ones.
This asymmetric impact of recessions on credit spreads continues to hold when we
add the controls that other studies of bond spreads have found to help explain credit
spreads (Model 4). As with these studies, we ?nd that callable bonds, bonds with a
sinking fund, longer maturity bonds and senior bonds carry higher credit spreads, but
bonds with a put provision and shelf issues pay lower credit spreads. We also ?nd that
larger issues as well as private placements pay higher credit spreads. Issues at the time
the treasury yield is high pay lower credit spreads. We do not ?nd evidence of a secular
decline in spreads of bank bonds. Somewhat surprisingly we ?nd that public banks
pay higher spreads than, private banks. This, however, may be due to the reduced
number of private banks in the sample.
1982-2002 Expansion Recession Rec-Exp t-test
All 0.952 0.899 1.334 0.435 10.340
*
Aaa 0.537 0.485 0.906 0.421 1.417
Aa 0.841 0.802 1.068 0.266 4.969
*
A 0.885 0.829 1.317 0.488 9.758
*
Baa 1.370 1.263 2.385 1.122 7.665
*
Ba 2.421 2.421 – – –
Notes:
*
Signi?cant at 1 per cent level; it includes all new non-convertible bonds issued by American
commercial banks in the USA in US dollars over the period 1982:2-2002:2 that had ratings from both
Moody’s and S&P, information on amount issued, maturity and yield to maturity; recessions and
expansions de?ned according to NBER
Table II.
Spreads of bank bonds by
credit rating arid state
of the economy
Do markets
“discipline” all
banks equally?
113
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
M
o
d
e
l
7
M
o
d
e
l
8
M
o
d
e
l
9
a
M
o
d
e
l
1
0
M
o
d
e
l
1
1
M
o
d
e
l
1
2
a
C
o
n
s
t
a
n
t
0
.
5
3
7
(
5
.
4
7
1
*
*
*
)
0
.
4
7
8
(
5
.
7
3
4
*
*
*
)
0
.
4
8
5
(
5
.
0
6
1
*
*
*
)
0
.
3
5
5
(
1
.
4
4
6
)
0
.
4
1
4
(
1
.
6
6
2
*
)
2
.
0
4
4
(
6
.
8
0
4
*
*
*
)
0
.
3
0
8
(
2
.
3
2
6
*
*
)
0
.
7
8
5
(
2
.
9
9
3
*
*
*
)
1
.
8
1
7
(
5
.
7
8
7
*
*
*
)
0
.
2
0
9
(
1
.
4
1
7
)
1
.
3
1
1
(
2
.
9
5
1
*
*
*
)
2
.
2
2
9
(
4
.
5
5
9
*
*
*
)
A
a
0
.
3
0
4
(
3
.
0
3
4
*
*
*
)
0
.
2
9
4
(
3
.
4
4
3
*
*
*
)
0
.
3
1
8
(
3
.
2
3
8
*
*
*
)
0
.
3
6
6
(
3
.
9
4
8
*
*
*
)
0
.
3
2
3
(
3
.
0
0
2
*
*
*
)
0
.
3
9
8
(
3
.
9
9
9
*
*
*
)
0
.
2
7
8
(
2
.
9
4
6
*
*
*
)
0
.
2
6
1
(
2
.
2
7
0
*
*
)
0
.
4
5
4
(
3
.
1
1
5
*
*
*
)
0
.
2
7
7
(
2
.
7
7
3
*
*
*
)
0
.
2
7
3
(
2
.
2
9
2
*
*
)
0
.
4
3
8
(
2
.
6
8
3
*
)
A
0
.
3
4
8
(
3
.
4
9
2
*
*
*
)
0
.
3
5
3
(
4
.
1
6
0
*
*
*
)
0
.
3
4
4
(
3
.
5
4
1
*
*
*
)
0
.
4
1
6
(
4
,
4
3
6
*
*
*
)
0
.
4
5
1
(
4
.
1
3
5
*
*
*
)
0
.
6
5
3
(
6
.
4
6
7
*
*
*
)
0
.
3
9
1
(
4
.
0
3
7
*
*
*
)
0
.
4
6
8
(
3
.
7
3
0
*
*
*
)
0
.
7
1
2
(
4
.
6
8
4
*
*
*
)
0
.
3
9
7
(
3
.
4
8
7
*
*
)
0
.
3
1
1
(
2
.
2
9
4
*
*
)
0
.
5
7
3
(
3
.
1
7
6
*
*
)
B
a
a
0
.
8
3
2
(
7
.
6
1
3
*
*
*
)
0
.
8
4
6
(
8
.
9
5
5
*
*
*
)
0
.
7
7
8
(
7
.
3
5
4
*
*
*
)
0
.
9
1
4
(
8
.
7
6
0
*
*
*
)
0
.
9
6
1
(
8
.
0
8
9
*
*
*
)
1
.
0
9
4
(
1
0
.
1
0
3
*
*
*
)
0
.
7
8
3
(
7
.
0
8
1
*
*
*
)
1
.
0
2
8
(
7
.
2
6
7
*
*
*
)
1
.
1
5
0
(
7
.
1
0
5
*
*
*
)
0
.
5
0
1
(
3
.
7
6
3
*
*
*
)
0
.
4
3
6
(
2
.
4
5
5
*
*
)
0
.
7
3
9
(
3
.
4
2
5
*
*
)
B
a
1
.
8
8
3
(
6
.
3
0
3
*
*
*
)
1
.
9
4
2
(
6
.
6
0
1
*
*
*
)
1
.
9
3
6
(
6
.
4
9
6
*
*
*
)
1
.
9
3
0
(
7
.
2
4
9
*
*
*
)
1
.
9
5
8
(
7
.
3
6
0
*
*
*
)
2
.
1
0
1
(
8
.
2
0
9
*
*
*
)
1
.
1
6
8
(
3
.
0
6
4
*
*
*
)
1
.
5
3
9
(
3
.
8
9
7
*
*
*
)
1
.
7
2
6
(
5
.
0
5
8
*
*
*
)
0
.
9
6
9
(
2
.
1
5
7
*
*
)
1
.
0
2
7
(
2
.
2
9
6
*
*
)
1
.
4
1
3
(
3
.
6
7
8
*
*
*
)
R
e
c
0
.
4
7
2
(
1
1
.
7
0
5
*
*
*
)
0
.
4
2
1
(
4
.
3
9
1
*
*
*
)
0
.
3
9
0
(
3
.
9
4
9
*
*
*
)
0
.
4
2
0
(
3
.
7
2
6
*
*
*
)
2
0
.
1
3
3
(
2
1
.
0
8
6
)
0
.
1
2
6
(
1
.
0
1
8
)
2
0
.
0
2
7
(
2
0
.
1
8
7
)
2
0
.
2
4
7
(
2
1
.
4
8
8
)
0
.
1
9
3
(
1
.
5
3
2
)
0
.
0
0
3
(
0
.
0
1
7
)
2
0
.
2
2
6
(
2
1
.
0
2
9
)
R
e
c
A
a
2
0
.
1
5
5
(
2
1
.
4
5
6
)
2
0
.
2
0
5
(
2
1
.
8
0
6
*
)
2
0
.
2
5
7
(
2
1
.
9
9
3
*
*
*
)
0
.
2
3
9
(
1
.
7
8
9
*
*
*
)
0
.
0
0
7
(
0
.
0
5
6
)
0
.
0
9
3
(
0
.
5
9
4
)
0
.
3
2
0
(
1
.
8
4
9
*
)
0
.
2
1
5
(
1
.
0
2
3
)
0
.
2
3
4
(
1
.
2
8
2
)
0
.
3
8
3
(
1
.
6
3
4
)
R
e
c
A
0
.
0
6
7
(
0
.
6
0
7
)
0
.
0
7
7
(
0
.
6
8
1
)
0
.
0
6
4
(
0
.
5
1
6
)
0
.
2
2
0
(
1
.
7
6
3
*
*
*
)
0
.
2
1
5
(
1
.
5
2
4
)
0
.
4
0
7
(
2
.
6
3
0
*
*
*
)
0
.
3
7
5
(
2
.
2
5
5
*
*
*
)
0
.
4
1
0
(
2
.
7
7
0
*
*
*
)
0
.
5
3
5
(
3
.
0
6
2
*
*
*
)
0
.
4
8
6
(
2
.
2
2
6
*
*
*
)
R
e
c
B
a
a
0
.
7
0
2
(
3
.
9
0
3
*
*
*
)
0
.
6
8
3
(
3
.
8
9
5
*
*
*
)
0
.
6
4
7
(
3
.
5
2
5
*
*
*
)
0
.
6
9
8
(
4
.
1
5
1
*
*
*
)
0
.
9
3
1
(
4
.
9
7
2
*
*
*
)
0
.
9
4
9
(
4
.
7
8
2
*
*
*
)
0
.
7
1
2
(
3
.
4
3
2
*
*
*
)
1
.
0
9
8
(
4
.
2
4
1
*
*
*
)
1
.
1
2
5
(
3
.
2
4
2
*
*
*
)
0
.
9
6
0
(
2
.
8
7
0
*
*
*
)
O
t
h
e
r
i
s
s
u
e
r
f
e
a
t
u
r
e
s
P
u
b
l
i
c
c
o
m
p
a
n
y
0
.
0
7
7
(
2
.
5
2
9
*
*
)
0
.
0
8
3
(
2
.
3
2
0
*
*
)
0
.
0
3
7
(
1
.
3
5
9
)
0
.
1
2
4
(
3
.
0
6
7
*
*
*
)
0
.
0
7
5
(
2
.
1
9
2
*
*
)
0
.
1
2
0
(
1
.
9
3
7
*
)
0
.
1
1
7
(
2
.
1
5
5
*
)
F
i
r
s
t
i
s
s
u
e
0
.
0
5
8
(
1
.
2
2
4
)
2
0
.
0
4
6
(
2
0
.
9
0
0
)
0
.
0
5
1
(
0
.
0
.
5
2
0
)
O
r
d
e
r
o
f
i
s
s
u
e
0
.
0
3
7
(
3
.
5
6
0
*
*
*
)
2
0
.
0
0
1
(
2
0
.
9
0
0
)
T
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
2
0
.
0
0
4
(
2
0
.
3
7
0
)
2
0
.
0
1
0
(
2
0
.
9
0
2
)
0
.
0
0
3
(
0
.
2
4
7
)
I
s
s
u
e
d
e
s
i
g
n
C
a
l
l
p
r
o
v
i
s
i
o
n
0
.
1
1
4
(
3
.
8
0
7
*
*
*
)
0
.
1
0
9
(
3
.
6
6
7
*
*
*
)
0
.
0
9
4
(
3
.
5
7
2
*
*
*
)
0
.
0
9
1
(
3
.
0
7
8
*
*
*
)
0
.
0
8
3
(
3
.
1
6
4
*
*
*
)
0
.
0
0
5
(
0
.
0
9
7
)
2
0
.
0
1
8
(
2
0
.
3
6
2
)
P
u
t
p
r
o
v
i
s
i
o
n
2
0
.
6
7
0
(
2
7
.
4
1
0
*
*
*
)
2
0
.
6
5
6
(
2
7
.
0
2
9
*
*
*
)
2
0
.
4
0
4
(
2
4
.
3
7
8
*
*
*
)
2
0
.
5
2
9
(
2
5
.
4
2
6
*
*
*
)
2
0
.
3
9
6
(
2
4
.
3
4
5
*
*
*
)
2
0
.
5
6
6
(
2
6
.
0
1
8
*
*
*
)
2
0
.
3
9
9
(
2
4
.
2
2
2
*
*
*
)
S
i
n
k
i
n
g
f
u
n
d
0
.
6
8
9
4
.
8
8
0
*
*
*
)
0
.
6
7
3
(
6
.
2
4
0
*
*
*
)
0
.
9
2
0
(
4
.
8
3
1
*
*
*
)
0
.
7
5
2
(
4
.
9
4
4
*
*
*
)
1
.
0
0
5
(
3
.
3
3
5
*
*
*
)
0
.
7
3
3
(
3
.
5
9
2
*
*
*
)
0
.
7
0
5
(
2
.
9
0
9
*
*
*
)
S
h
e
l
f
2
0
.
2
0
2
(
2
6
.
1
4
7
*
*
*
)
2
0
.
2
1
5
(
2
6
.
5
8
9
*
*
*
)
2
0
.
2
0
6
(
2
7
.
3
9
5
*
*
*
)
2
0
.
2
1
3
(
2
6
.
3
7
0
*
*
*
)
2
0
.
1
9
4
(
2
6
.
6
1
7
*
*
*
)
2
0
.
0
9
8
(
2
1
.
4
8
8
)
2
0
.
1
2
9
(
2
2
.
3
7
4
*
*
)
L
o
g
(
m
a
t
u
r
i
t
y
)
0
.
2
0
9
(
1
2
.
0
7
4
*
*
*
)
0
.
1
9
7
(
1
1
.
1
3
0
*
*
*
)
0
.
1
7
9
(
1
1
.
3
7
9
*
*
*
)
0
.
1
7
5
(
9
.
8
0
9
*
*
*
)
0
.
1
9
0
(
1
2
.
2
5
8
*
*
*
)
0
.
2
1
7
(
6
.
9
4
6
*
*
*
)
0
.
2
4
9
(
8
.
2
4
4
*
*
*
)
S
e
n
i
o
r
0
.
0
7
4
(
2
.
2
0
4
*
*
)
0
.
0
6
9
(
2
.
0
4
2
*
*
)
0
.
0
0
6
(
0
.
2
1
9
)
0
.
0
2
8
(
0
.
7
8
8
)
0
.
0
0
7
(
0
.
2
4
0
)
0
.
0
1
2
(
0
.
1
3
6
)
0
.
0
3
9
(
0
.
5
0
2
)
(
c
o
n
t
i
n
u
e
d
)
Table III.
Bank-bond spreads over
treasury at issue date
JFEP
1,1
114
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
M
o
d
e
l
7
M
o
d
e
l
8
M
o
d
e
l
9
a
M
o
d
e
l
1
0
M
o
d
e
l
1
1
M
o
d
e
l
1
2
a
O
t
h
e
r
c
o
n
t
r
o
l
s
L
o
g
(
a
m
o
u
n
t
i
s
s
u
e
d
)
0
.
0
2
6
(
2
.
6
3
7
*
*
*
)
0
.
0
3
4
(
3
.
3
8
2
*
*
*
)
0
.
0
3
8
(
4
.
3
7
5
*
*
*
)
0
.
0
2
5
(
2
.
0
4
0
*
*
)
0
.
0
2
5
(
2
.
1
8
0
*
*
*
)
0
.
0
2
0
(
0
.
8
1
8
)
0
.
0
1
3
(
0
.
5
1
4
)
P
r
i
v
.
p
l
a
c
e
m
e
n
t
0
.
1
8
7
(
3
.
4
2
1
*
*
)
0
.
1
3
0
(
2
.
2
9
4
*
*
)
2
0
.
0
4
2
(
2
0
.
8
3
2
)
0
.
1
5
5
(
1
.
3
8
0
)
2
0
.
0
7
6
(
2
0
.
7
4
6
)
0
.
1
8
3
(
1
.
4
0
7
)
2
0
.
0
2
8
(
2
0
.
2
2
2
)
L
o
g
(
t
i
m
e
t
r
e
n
d
)
0
.
0
0
8
(
0
.
2
2
9
)
2
0
.
0
2
0
(
2
0
.
5
9
9
)
2
0
.
0
0
2
(
2
0
.
0
5
5
)
2
0
.
0
5
2
(
2
1
.
3
6
3
)
0
.
0
2
1
(
0
.
4
8
7
)
2
0
.
1
4
1
(
2
1
.
8
6
6
*
)
2
0
.
0
6
3
(
2
0
.
8
6
5
)
T
e
n
-
y
e
a
r
T
r
e
a
s
u
r
y
y
i
e
l
d
2
0
.
0
2
7
(
2
1
.
8
3
6
*
)
2
0
.
0
2
3
(
2
1
.
5
2
9
)
2
0
.
0
8
0
(
2
4
.
8
6
4
*
*
*
)
2
0
.
0
4
7
(
2
3
.
1
9
0
*
*
*
)
2
0
.
0
7
8
(
2
4
.
9
6
9
*
*
*
)
2
0
.
0
8
4
(
2
2
.
9
5
2
*
*
*
)
2
0
.
1
1
3
(
2
4
.
2
5
7
*
*
*
)
Y
i
e
l
d
c
.
s
l
o
p
e
2
0
.
2
2
6
(
2
8
.
2
9
5
)
2
0
.
2
2
6
(
2
8
.
3
8
6
*
*
*
)
2
0
.
1
8
7
(
2
6
.
8
2
0
*
*
*
)
2
0
.
2
5
7
(
2
9
.
6
0
4
*
*
*
)
2
0
.
1
8
3
(
2
7
.
0
8
0
*
*
*
)
2
0
.
1
8
2
(
2
1
.
6
8
7
*
)
2
0
.
1
7
9
(
2
1
.
9
2
1
)
S
p
l
i
t
0
.
2
2
3
(
5
.
4
5
8
*
*
*
)
0
.
0
8
7
(
2
.
3
7
7
*
*
*
)
0
.
2
3
7
(
4
.
7
5
9
*
*
*
)
0
.
1
3
1
(
3
.
2
3
0
*
*
*
)
2
0
.
0
0
9
(
2
0
.
0
7
7
)
2
0
.
0
9
3
(
2
0
.
9
1
6
)
S
P
.
M
o
o
d
2
0
.
2
2
2
(
2
3
.
7
4
5
*
*
*
)
2
0
.
2
0
5
(
2
3
.
8
7
1
*
*
*
)
2
0
.
2
5
8
(
2
3
.
6
3
7
*
*
*
)
2
0
.
2
4
4
(
2
4
.
0
7
5
*
*
*
)
2
0
.
1
9
2
(
2
1
.
1
7
5
)
2
0
.
1
8
7
(
2
1
.
3
4
6
)
W
i
t
h
b
a
n
k
?
x
e
d
e
f
f
e
c
t
s
N
o
N
o
N
o
N
o
N
o
N
o
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
Y
e
s
W
i
t
h
w
e
i
g
h
t
e
d
o
b
s
.
N
o
N
o
N
o
N
o
N
o
N
o
N
o
N
o
N
o
Y
e
s
Y
e
s
Y
e
s
A
d
j
u
s
t
e
d
R
2
1
5
.
9
9
4
2
3
.
6
1
3
2
5
.
7
9
7
3
9
.
8
1
7
4
1
.
1
6
4
5
9
.
9
3
0
4
0
.
5
0
2
5
3
.
2
7
1
6
6
.
7
0
2
7
6
.
5
4
6
8
2
.
2
5
9
8
7
.
1
3
1
L
o
g
l
i
k
e
l
i
h
o
o
d
2
1
,
1
4
8
.
2
6
2
1
,
0
7
4
.
6
9
2
1
,
0
5
0
.
3
9
2
8
8
1
.
3
5
2
8
6
1
.
4
1
2
5
6
4
.
6
8
2
7
6
4
.
0
3
2
5
6
8
.
3
5
2
3
0
3
.
7
7
2
1
,
3
3
3
.
9
0
2
1
,
1
0
9
.
3
3
2
8
5
9
.
0
6
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
a
t
t
h
e
*
1
0
,
*
*
5
,
*
*
*
1
p
e
r
c
e
n
t
l
e
v
e
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
;
m
o
d
e
l
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
a
r
e
h
e
t
e
r
o
s
k
e
d
a
s
t
i
c
-
c
o
n
s
i
s
t
e
n
t
;
t
o
t
a
l
n
u
m
b
e
r
o
f
o
b
s
e
r
v
a
t
i
o
n
s
1
,
5
3
7
.
t
-
s
t
a
t
i
s
t
i
c
s
i
n
p
a
r
e
n
t
h
e
s
i
s
;
t
h
e
d
e
p
e
n
d
e
n
t
v
a
r
i
a
b
l
e
i
s
t
h
e
b
o
n
d
c
r
e
d
i
t
s
p
r
e
a
d
o
v
e
r
t
h
e
t
r
e
a
s
u
r
y
w
i
t
h
t
h
e
s
a
m
e
m
a
t
u
r
i
t
y
;
s
p
r
e
a
d
s
a
r
e
c
o
m
p
u
t
e
d
a
t
i
s
s
u
e
d
a
t
e
;
r
e
c
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
i
s
s
u
e
d
d
u
r
i
n
g
a
r
e
c
e
s
s
i
o
n
a
c
c
o
r
d
i
n
g
t
o
N
B
E
R
;
p
u
b
l
i
c
c
o
m
p
a
n
y
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
i
s
s
u
e
r
i
s
a
p
u
b
l
i
c
c
o
m
p
a
n
y
;
?
r
s
t
i
s
s
u
e
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
t
h
e
b
a
n
k
’
s
?
r
s
t
b
o
n
d
i
s
s
u
e
s
i
n
c
e
1
9
7
0
:
1
;
o
r
d
e
r
o
f
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
t
i
m
e
s
t
h
e
b
a
n
k
i
s
s
u
e
d
b
o
n
d
s
s
i
n
c
e
1
9
7
0
:
1
;
t
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
y
e
a
r
s
s
i
n
c
e
t
h
e
b
a
n
k
m
a
d
e
i
t
s
l
a
t
e
s
t
b
o
n
d
i
s
s
u
e
;
c
a
l
l
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
c
a
l
l
a
b
l
e
;
p
u
t
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
b
o
n
d
h
o
l
d
e
r
s
c
a
n
s
e
l
l
t
h
e
b
o
n
d
b
a
c
k
t
o
t
h
e
c
o
m
p
a
n
y
p
r
i
o
r
t
o
m
a
t
u
r
i
t
y
;
s
i
n
k
i
n
g
f
u
n
d
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
h
a
s
a
s
i
n
k
i
n
g
f
u
n
d
;
s
h
e
l
f
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
a
s
h
e
l
f
i
s
s
u
e
;
m
a
t
u
r
i
t
y
,
m
a
t
u
r
i
t
y
o
f
t
h
e
b
o
n
d
i
n
y
e
a
r
s
;
a
m
o
u
n
t
i
s
s
u
e
d
i
n
m
i
l
l
i
o
n
s
o
f
U
S
d
o
l
l
a
r
s
;
i
s
s
u
e
s
d
e
?
a
t
e
d
b
y
t
h
e
c
o
r
e
u
r
b
a
n
C
P
I
w
i
t
h
t
h
e
a
v
e
r
a
g
e
o
f
1
9
8
2
-
1
9
8
4
¼
1
0
0
;
p
r
i
v
.
p
l
a
c
e
m
e
n
t
d
u
m
m
y
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
p
r
i
v
a
t
e
l
y
p
l
a
c
e
d
;
t
i
m
e
t
r
e
n
d
,
l
i
n
e
a
r
t
i
m
e
t
r
e
n
d
;
a
i
n
c
l
u
d
e
d
i
n
t
h
i
s
r
e
g
r
e
s
s
i
o
n
b
u
t
n
o
t
s
h
o
w
n
i
n
t
h
e
t
a
b
l
e
a
r
e
t
h
e
n
u
m
b
e
r
o
f
b
o
n
d
s
i
s
s
u
e
d
i
n
t
h
e
q
u
a
r
t
e
r
i
n
e
a
c
h
o
f
t
h
e
n
i
n
e
m
a
i
n
c
r
e
d
i
t
r
a
t
i
n
g
c
l
a
s
s
e
s
a
s
d
e
?
n
e
d
b
y
M
o
o
d
y
’
s
w
h
o
l
e
r
a
t
i
n
g
s
,
A
a
a
,
.
.
.
,
C
Table III.
Do markets
“discipline” all
banks equally?
115
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Given that nearly all bonds issued in the USA are rated by both Moody’s and S&P and
given the evidence of other studies showing that split ratings, that is instances where
the two rating agencies rate the bond differently, affect the credit spread, we expanded
our model of bond credit spreads and added two new dummy variables. The ?rst
variable takes the value one when Moody’s and S&P announce a different rating for
the bond and the second variable takes the value one when S&P announces a higher
rating than Moody’s. As the results of Model 5 show, bonds that are rated differently
by Moody’s and S&P pay on average an additional 22 basis points. Interestingly,
among the split-rated bonds only those that S&P assigns a lower rating than Moody’s
pay higher credit spreads (recall that we have used Moody’s ratings to control for the
bond’s creditworthiness). Model 5 continues to indicate that recessions affect
disproportinaly more banks that issue lower rated bonds, even if they are of
investment grade quality, than banks that issue higher rated bonds.
That differential impact of recessions on bond credit spreads continues to hold when
we control for the acquaintance of investors with the bank’s bonds by distinguishing the
?rst bond issue of the bank and by accounting for the time that has elapsed since the
previous issue, and the number of times the bank has issued bonds since 1970 (Model 6).
It also continues to hold when we estimate our model with bank ?xed effects to control
for potential differences across banks on the cost to access the bond market (Models 7-9),
and when with estimate it with weighted least squares, computing the weights off the
number of bonds each bank issued over the sample period, to account for differences in
bond-issuance activity across banks (Models 11-12).
In sum, all of our tests show that the risk structure of the credit spreads on bank
bonds shifts up but becomes steeper during recessions, suggesting that the market
does not discipline all banks equally. Instead it is more demanding with riskier banks
than with safer ones. This ?nding, however, is not the only possible interpretation of
our results since our interpretation relies on some assumptions. For example, if bond
ratings were not comparable over the business cycle then our ?ndings could have an
alternative explanation, namely that bonds issued during recessions are riskier than
bonds issued during expansions with the same credit rating and the risk difference
between these two cohorts of bonds increases with risk. Note that even when ratings
are comparable, if the safer ?rms in each rating bucket drop out of the market in
recessions and this effect increases with risk, this too could explain the marginal
impact of recessions on credit spreads that we identi?ed. To account for these and
other like explanations, we add a control group that is also susceptible to these
problems – the bonds issued by ?rms over the same time period – and investigate the
marginal impact of recessions on the difference between the spreads on bank bonds
and those on ?rm bonds with the same credit rating. Our null hypothesis is that if the
markets were to discipline all banks equally then the marginal impact of recessions on
that difference in spreads should not vary across bond issuers with different risk.
3.3 Credit spreads on bank and ?rm bonds
Table IV presents the results of our second test. Model 1 shows that in good times, on
average, the spreads on bank bonds are 4 basis points higher than the spreads on ?rm
bonds. This model also shows that while recessions increase the spreads of ?rm bonds
by 72 basis points, they increase the spreads of bank bonds by only 49 basis points.
Model 2 shows that this differential effect of recessions continues to hold when we
JFEP
1,1
116
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
C
o
n
s
t
a
n
t
0
.
5
2
8
(
1
4
.
5
2
9
*
*
*
)
0
.
5
3
2
(
1
4
.
1
1
7
*
*
*
)
0
.
6
6
0
(
1
7
.
4
5
0
*
*
*
)
2
.
4
9
7
(
1
5
.
9
3
5
*
*
*
)
2
.
4
9
1
(
1
6
.
1
7
0
*
*
*
)
5
.
9
0
0
(
2
6
.
4
9
3
*
*
*
)
A
a
0
.
1
1
3
(
2
.
9
7
6
*
*
*
)
0
.
0
4
9
(
1
.
2
3
1
)
2
0
.
0
3
8
(
2
0
.
9
5
4
)
0
.
0
6
3
(
1
.
8
3
9
*
)
0
.
0
8
0
(
2
.
1
7
7
*
*
)
0
.
1
9
1
(
5
.
1
7
0
*
*
*
)
A
0
.
3
0
0
(
8
.
1
7
4
*
*
*
)
0
.
3
0
5
(
7
.
9
8
7
*
*
*
)
0
.
1
9
5
(
5
.
0
1
4
*
*
*
)
0
.
3
1
4
(
9
.
4
9
2
*
*
*
)
0
.
3
6
3
(
1
0
.
1
8
2
*
*
*
)
0
.
5
0
8
(
1
4
.
2
1
0
*
*
*
)
B
a
a
0
.
8
3
5
(
2
1
.
6
9
7
*
*
*
)
0
.
8
3
8
(
2
0
.
9
7
5
*
*
*
)
0
.
6
8
6
(
1
6
.
9
6
4
*
*
*
)
0
.
7
6
7
(
2
1
.
9
9
9
*
*
*
)
0
.
8
4
9
(
2
2
.
6
5
5
*
*
*
)
1
.
0
0
0
(
2
6
.
8
6
1
*
*
*
)
B
a
2
.
3
1
5
(
4
2
.
9
3
7
*
*
*
)
2
.
3
1
8
(
4
2
.
0
9
5
*
*
*
)
2
.
1
4
2
(
3
8
.
2
3
1
*
*
*
)
2
.
0
5
7
(
4
0
.
1
6
4
*
*
*
)
2
.
1
2
5
(
4
1
.
3
6
2
*
*
*
)
2
.
3
4
1
(
4
7
.
1
5
4
*
*
*
)
B
4
.
1
3
6
(
8
7
.
1
3
4
*
*
*
)
4
.
1
3
3
(
8
5
.
2
4
1
*
*
*
)
4
.
0
0
4
(
8
1
.
5
7
8
*
*
*
)
3
.
7
2
8
(
6
8
.
8
1
9
*
*
*
)
3
.
8
3
3
(
6
9
.
0
8
5
*
*
*
)
4
.
1
1
0
(
7
4
.
1
6
8
*
*
*
)
C
a
a
5
.
9
5
7
(
2
9
.
1
4
0
*
*
*
)
5
.
9
5
3
(
2
9
.
0
8
9
*
*
*
)
5
.
7
9
7
(
2
7
.
8
3
5
*
*
*
)
5
.
2
9
5
(
2
5
.
7
1
8
*
*
*
)
5
.
5
8
6
(
2
6
.
5
5
3
*
*
*
)
5
.
9
5
0
(
2
9
.
0
9
0
*
*
*
)
B
k
0
.
0
3
7
(
2
.
3
3
1
*
*
)
2
0
.
0
5
4
(
2
0
.
5
8
5
)
2
0
.
1
7
5
(
2
1
.
7
0
3
*
)
2
0
.
2
3
8
(
2
2
.
3
1
7
*
*
)
2
0
.
2
4
7
(
2
2
.
3
1
8
*
*
)
2
0
.
2
8
3
(
2
2
.
5
5
4
*
*
)
B
k
A
a
0
.
2
4
5
(
2
.
5
9
8
*
*
*
)
0
.
3
5
6
(
3
.
3
5
8
*
*
*
)
0
.
4
5
7
(
4
.
3
4
0
*
*
*
)
0
.
3
3
4
(
3
.
0
5
1
*
*
*
)
0
.
3
1
9
(
2
.
8
0
4
*
*
*
)
B
k
A
0
.
0
4
7
(
0
.
5
1
1
)
0
.
1
4
9
(
1
.
4
2
8
)
0
.
3
3
2
(
3
.
1
5
1
*
*
*
)
0
.
3
8
6
(
3
.
5
1
6
*
*
*
)
0
.
5
1
7
(
4
.
5
6
5
*
*
*
)
B
k
B
a
a
0
.
0
0
8
(
0
.
0
7
9
)
0
.
0
9
2
(
0
.
8
1
3
)
0
.
4
2
6
(
3
.
6
7
1
*
*
*
)
0
.
5
3
0
(
4
.
3
7
9
*
*
*
)
0
.
5
8
4
(
4
.
8
0
1
*
*
*
)
B
k
B
a
2
0
.
3
7
6
(
2
1
.
2
5
4
)
2
0
.
2
0
6
(
2
0
.
6
7
8
)
0
.
2
0
2
(
0
.
7
2
8
)
0
.
4
0
2
(
1
.
5
1
3
)
0
.
3
2
7
(
1
.
1
9
9
)
R
e
c
0
.
7
1
8
(
2
2
.
8
0
7
*
*
*
)
0
.
7
1
9
(
2
2
.
8
6
3
*
*
*
)
0
.
2
2
1
(
2
.
7
1
4
*
*
*
)
0
.
2
1
5
(
3
.
2
2
0
*
*
*
)
0
.
2
7
4
(
3
.
8
8
2
*
*
*
)
0
.
1
1
3
(
1
.
4
3
4
)
R
e
c
A
a
0
.
1
4
6
(
1
.
4
7
2
)
0
.
1
4
3
(
1
.
7
6
3
*
)
0
.
0
5
8
(
0
.
6
9
3
)
2
0
.
0
2
2
(
2
0
.
2
2
9
)
R
e
c
A
0
.
3
5
2
(
4
.
0
6
2
*
*
*
)
0
.
3
3
4
(
4
.
6
3
2
*
*
*
)
0
.
2
7
2
(
3
.
6
1
5
*
*
*
)
0
.
1
3
3
(
1
.
5
9
3
)
R
e
c
B
a
a
0
.
7
7
0
(
7
.
6
4
0
*
*
*
)
0
.
6
9
8
(
8
.
0
8
6
*
*
*
)
0
.
6
2
6
(
7
.
0
3
5
*
*
*
)
0
.
4
4
9
(
4
.
6
1
9
*
*
*
)
R
e
c
B
a
1
.
0
6
7
(
5
.
9
9
7
*
*
*
)
0
.
9
8
7
(
6
.
0
9
1
*
*
*
)
0
.
9
6
2
(
6
.
0
0
3
*
*
*
)
0
.
7
8
9
(
4
.
8
4
7
*
*
*
)
R
e
c
B
0
.
5
1
0
(
2
.
9
3
9
*
*
*
)
0
.
4
7
5
(
2
.
8
8
5
*
*
*
)
0
.
4
3
4
(
2
.
6
5
5
*
*
*
)
0
.
2
1
2
(
1
.
2
7
9
)
R
e
c
C
a
a
1
.
5
4
7
(
1
.
4
8
2
)
1
.
6
2
0
(
1
.
5
6
0
)
1
.
5
6
2
(
1
.
5
6
8
)
1
.
2
5
7
(
1
.
2
5
5
)
R
e
c
B
k
2
0
.
2
3
1
(
2
4
.
6
2
8
*
*
*
)
2
0
.
2
4
7
(
2
4
.
8
2
8
*
*
*
)
0
.
2
0
0
(
1
.
5
8
7
)
0
.
5
1
1
(
4
.
2
9
4
*
*
*
)
0
.
5
2
5
(
4
.
2
4
8
*
*
*
)
0
.
5
4
0
(
3
.
9
1
5
*
*
*
)
R
e
c
B
k
A
a
2
0
.
3
0
1
(
2
2
.
0
6
9
*
*
)
2
0
.
7
2
9
(
2
5
.
3
4
1
*
*
*
)
2
0
.
7
3
5
(
2
5
.
1
2
5
*
*
*
)
2
0
.
6
4
6
(
2
4
.
0
3
3
*
*
*
)
R
e
c
B
k
A
2
0
.
2
8
5
(
2
2
.
0
3
1
*
*
*
)
2
0
.
6
1
8
(
2
4
.
4
9
8
*
*
*
)
2
0
.
5
9
4
(
2
4
.
2
1
0
*
*
*
)
2
0
.
7
3
2
(
2
4
.
7
2
3
*
*
*
)
R
e
c
B
k
B
a
a
2
0
.
0
6
9
(
2
0
.
3
3
3
)
2
0
.
3
0
1
(
2
1
.
4
0
4
)
2
0
.
2
4
5
(
2
1
.
1
3
7
)
2
0
.
2
8
8
(
2
1
.
3
0
7
)
O
t
h
e
r
i
s
s
u
e
r
f
e
a
t
u
r
e
s
P
u
b
l
i
c
c
o
m
p
a
n
y
2
0
.
1
3
1
(
2
6
.
9
9
5
*
*
*
)
2
0
.
1
3
6
(
2
7
.
3
5
8
*
*
*
)
2
0
.
1
3
3
(
2
7
.
8
7
3
*
*
*
)
F
i
r
s
t
i
s
s
u
e
0
.
1
9
1
(
6
.
0
2
4
*
*
*
)
O
r
d
e
r
o
f
i
s
s
u
e
0
.
0
0
1
(
0
.
0
7
0
)
T
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
0
.
0
0
7
(
1
.
5
7
8
)
I
s
s
u
e
d
e
s
i
g
n
C
a
l
l
p
r
o
v
i
s
i
o
n
0
.
2
6
2
(
1
1
.
1
5
2
*
*
*
)
0
.
2
3
9
(
1
0
.
3
7
8
*
*
*
)
0
.
1
4
4
(
6
.
3
8
5
*
*
*
)
P
u
t
p
r
o
v
i
s
i
o
n
2
0
.
6
9
9
(
2
1
3
.
5
0
3
*
*
*
)
2
0
.
6
8
0
(
2
1
3
.
4
3
0
*
*
*
)
2
0
.
5
3
2
(
2
1
2
.
2
8
8
*
*
*
)
S
i
n
k
i
n
g
f
u
n
d
0
.
3
0
4
(
8
.
5
4
8
*
*
*
)
0
.
2
6
8
(
7
.
7
4
5
*
*
*
)
0
.
1
3
6
(
3
.
9
3
5
*
*
*
)
(
c
o
n
t
i
n
u
e
d
)
Table IV.
Bank and ?rm-bond
spreads over treasury at
issue date
Do markets
“discipline” all
banks equally?
117
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
D
e
p
.
v
a
r
i
a
b
l
e
s
M
o
d
e
l
1
M
o
d
e
l
2
M
o
d
e
l
3
M
o
d
e
l
4
M
o
d
e
l
5
M
o
d
e
l
6
a
F
l
o
a
t
e
r
s
1
.
4
5
5
(
4
.
3
0
0
*
*
*
)
1
.
4
2
0
(
4
.
1
5
5
*
*
*
)
1
.
2
2
1
(
3
.
6
7
9
*
*
*
)
S
h
e
l
f
2
0
.
2
4
3
(
2
1
1
.
0
9
8
*
*
*
)
2
0
.
2
1
5
(
2
1
0
.
1
0
4
*
*
*
)
2
0
.
1
6
0
(
2
7
.
5
3
1
*
*
*
)
L
o
g
(
m
a
t
u
r
i
t
y
)
0
.
1
0
8
(
1
0
.
0
6
4
*
*
*
)
0
.
0
9
5
(
8
.
8
8
0
*
*
*
)
0
.
1
3
7
(
1
3
.
2
6
7
*
*
*
)
S
e
n
i
o
r
0
.
2
3
2
(
6
.
2
3
0
*
*
*
)
0
.
2
7
0
(
7
.
2
8
3
*
*
*
)
0
.
3
1
0
(
8
.
7
9
1
*
*
*
)
O
t
h
e
r
c
o
n
t
r
o
l
s
L
o
g
(
a
m
o
u
n
t
i
s
s
u
e
d
)
0
.
0
3
3
(
5
.
9
3
7
*
*
*
)
0
.
0
3
6
(
6
.
4
7
3
*
*
*
)
0
.
0
0
6
(
1
.
0
1
0
)
P
r
i
v
.
p
l
a
c
e
m
e
n
t
0
.
0
8
2
(
2
.
0
9
9
*
*
)
0
.
1
0
1
(
2
.
6
2
1
*
*
*
)
0
.
1
5
4
(
4
.
3
7
1
*
*
*
)
L
o
g
(
t
i
m
e
t
r
e
n
d
)
2
0
.
2
2
3
(
2
9
.
8
7
7
*
*
*
)
2
0
.
2
4
4
(
2
1
1
.
0
0
8
*
*
*
)
2
0
.
2
3
8
(
2
8
.
6
0
6
*
*
*
)
T
e
n
-
y
e
a
r
t
r
e
a
s
u
r
y
y
i
e
l
d
2
0
.
1
7
6
(
2
1
9
.
2
5
4
*
*
*
)
2
0
.
1
7
5
(
2
1
9
.
5
2
0
*
*
*
)
2
0
.
3
2
6
(
2
2
6
.
7
8
8
*
*
*
)
Y
i
e
l
d
c
.
s
l
o
p
e
2
0
.
2
0
0
(
2
1
2
.
7
2
8
*
*
*
)
2
0
.
1
9
8
(
2
1
2
.
7
8
0
*
*
*
)
2
0
.
2
2
2
(
2
1
1
.
3
1
5
*
*
*
)
S
p
l
i
t
0
.
3
8
5
(
1
3
.
8
6
7
*
*
*
)
0
.
3
2
8
(
1
2
.
4
2
7
*
*
*
)
S
P
.
M
o
o
d
2
0
.
7
1
1
(
2
1
8
.
3
9
3
*
*
*
)
2
0
.
7
0
0
(
2
1
9
.
4
4
6
*
*
*
)
A
d
j
u
s
t
e
d
R
2
7
5
.
5
4
7
7
5
.
5
7
4
7
5
.
7
7
0
7
8
.
3
2
5
7
8
.
9
4
1
8
1
.
7
4
9
L
o
g
l
i
k
e
l
i
h
o
o
d
2
1
4
,
9
2
6
.
6
2
1
4
,
9
1
8
.
1
2
1
4
.
8
6
6
.
4
2
1
4
,
2
1
4
.
4
2
1
4
,
0
4
6
.
2
2
1
3
,
2
1
2
.
2
N
o
t
e
s
:
S
i
g
n
i
?
c
a
n
c
e
a
t
t
h
e
*
1
0
,
*
*
5
,
*
*
*
1
p
e
r
c
e
n
t
l
e
v
e
l
s
,
r
e
s
p
e
c
t
i
v
e
l
y
;
m
o
d
e
l
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
a
r
e
h
e
t
e
r
o
s
k
e
d
a
s
t
i
c
-
c
o
n
s
i
s
t
e
n
t
;
t
o
t
a
l
n
u
m
b
e
r
o
f
o
b
s
e
r
v
a
t
i
o
n
s
1
1
,
5
8
7
;
t
-
S
t
a
t
i
s
t
i
c
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
t
h
e
d
e
p
e
n
d
e
n
t
v
a
r
i
a
b
l
e
i
s
t
h
e
b
o
n
d
c
r
e
d
i
t
s
p
r
e
a
d
o
v
e
r
t
h
e
t
r
e
a
s
u
r
y
w
i
t
h
t
h
e
s
a
m
e
m
a
t
u
r
i
t
y
;
s
p
r
e
a
d
s
a
r
e
c
o
m
p
u
t
e
d
a
t
i
s
s
u
e
d
a
t
e
;
r
e
c
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
i
s
s
u
e
d
d
u
r
i
n
g
a
r
e
c
e
s
s
i
o
n
a
c
c
o
r
d
i
n
g
t
o
N
B
E
R
;
p
u
b
l
i
c
c
o
m
p
a
n
y
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
i
s
s
u
e
r
i
s
a
p
u
b
l
i
c
c
o
m
p
a
n
y
;
?
r
s
t
i
s
s
u
e
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
t
h
e
?
r
m
’
s
?
r
s
t
b
o
n
d
i
s
s
u
e
s
i
n
c
e
1
9
7
0
:
1
;
o
r
d
e
r
o
f
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
t
i
m
e
s
t
h
e
?
r
m
i
s
s
u
e
d
b
o
n
d
s
s
i
n
c
e
1
9
7
0
:
1
;
t
i
m
e
s
i
n
c
e
p
r
e
v
.
i
s
s
u
e
i
s
t
h
e
n
u
m
b
e
r
o
f
y
e
a
r
s
s
i
n
c
e
t
h
e
?
r
m
m
a
d
e
i
t
s
l
a
t
e
s
t
b
o
n
d
i
s
s
u
e
;
c
a
l
l
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
c
a
l
l
a
b
l
e
;
p
u
t
p
r
o
v
i
s
i
o
n
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
b
o
n
d
h
o
l
d
e
r
s
c
a
n
s
e
l
l
t
h
e
b
o
n
d
b
a
c
k
t
o
t
h
e
c
o
m
p
a
n
y
p
r
i
o
r
t
o
m
a
t
u
r
i
t
y
;
s
i
n
k
i
n
g
f
u
n
d
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
h
a
s
a
s
i
n
k
i
n
g
f
u
n
d
;
s
h
e
l
f
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
i
s
a
s
h
e
l
f
i
s
s
u
e
;
m
a
t
u
r
i
t
y
,
m
a
t
u
r
i
t
y
o
f
t
h
e
b
o
n
d
i
n
y
e
a
r
s
;
a
m
o
u
n
t
i
s
s
u
e
d
i
n
m
i
l
l
i
o
n
s
o
f
U
S
d
o
l
l
a
r
s
;
i
s
s
u
e
s
d
e
?
a
t
e
d
b
y
t
h
e
c
o
r
e
u
r
b
a
n
c
o
n
s
u
m
e
r
C
P
I
w
i
t
h
t
h
e
a
v
e
r
a
g
e
o
f
1
9
8
2
-
1
9
8
4
¼
1
0
0
;
p
r
i
v
.
p
l
a
c
e
m
e
n
t
d
u
m
m
y
v
a
r
i
a
b
l
e
t
h
a
t
e
q
u
a
l
s
o
n
e
i
f
t
h
e
b
o
n
d
w
a
s
p
r
i
v
a
t
e
l
y
p
l
a
c
e
d
;
t
i
m
e
t
r
e
n
d
l
i
n
e
a
r
t
i
m
e
t
r
e
n
d
;
a
i
n
c
l
u
d
e
d
i
n
t
h
i
s
r
e
g
r
e
s
s
i
o
n
b
u
t
n
o
t
s
h
o
w
n
i
n
t
h
e
t
a
b
l
e
a
r
e
d
u
m
m
y
v
a
r
i
a
b
l
e
s
f
o
r
t
h
e
i
s
s
u
e
r
’
s
s
e
c
t
o
r
o
f
a
c
t
i
v
i
t
y
a
s
d
e
?
n
e
d
b
y
S
I
C
o
n
e
-
d
i
g
i
t
c
o
d
e
,
a
n
d
t
h
e
n
u
m
b
e
r
o
f
b
o
n
d
s
i
s
s
u
e
d
i
n
t
h
e
q
u
a
r
t
e
r
i
n
e
a
c
h
o
f
t
h
e
n
i
n
e
m
a
i
n
c
r
e
d
i
t
r
a
t
i
n
g
c
l
a
s
s
e
s
a
s
d
e
?
n
e
d
b
y
M
o
o
d
y
’
s
w
h
o
l
e
r
a
t
i
n
g
s
,
A
a
a
,
.
.
.
,
C
Table IV.
JFEP
1,1
118
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
allow for ?rm bonds to be priced differently than bank bonds. According to this model,
recessions increase the cost of bond ?nancing to ?rms and banks by 72 and 47 basis
points, respectively.
Model 3 investigates whether that differential effect of recessions applies to all banks
equally by allowing bank bonds of different risk issued in good and bad times to be priced
differently. As we noted in the methodology section, we are particularly interested in the
variable Rec Bk rating of this model as it tells us whether the slope of the credit spread
curve on bank bonds changed differently than the slope of the same curve for ?rm bonds
during recessions. According to the results of that, effectivelythe slope of the credit spread
curve on bank bonds becomes relatively steeper than that on ?rm bonds during
recessions. This difference arises because recessions have the same impact on ?rms and
banks that issue Baa-rated bonds (the coef?cient on Rec Bk Baa is not statistically
different fromzero), but theyhave a signi?cantlysmaller impact onbanks that issue either
double or single-A rated bonds than ?rms that issue bonds with these same ratings (the
coef?cients on Rec Bk Aa and Rec Bk Aare negative and statistically different fromzero).
This differential impact of recessions continues to hold after we control for the
standard factors that researchers have found help explain bond spreads (Model 4). One
noticeable difference is that the variable Rec Bk, which measures the differential
impact of recessions on triple-A rated bank bonds and ?rm bonds with the same rating,
is now statistically signi?cant. Given that there are only nine bank bonds with a
triple-A rating in the sample, this ?nding does not appear to be meaningful and so we
chose to continue to focus our analysis on investment-grade bonds other than triple-A
rated bonds[15]. As for the new controls we added to the regression, they produced
similar results when we included them in our regressions on bank bonds. There are,
however, three differences that are worth noting. First, we now ?nd a secular decline in
the spreads of bonds. Second, consistent with the literature, we ?nd too that public
companies pay lower credit spreads than private companies. Finally, we ?nd that the
slope of the treasury yield curve affects negatively the spreads on ?rm bonds. In our
previous analysis of bank bonds none of these variables were statistically signi?cant.
Models 5 and 6 present the results of two additional robustness tests. Model 5
accounts for the potential impact of disagreements between rating agencies on the
rating of the bond, and Model 6 investigates the impact of controlling for the
acquaintance of investors with bond issuers. The results of the former model con?rm
our earlier ?ndings that disagreements between rating agencies increase the cost of
accessing the bond market, but only when the issuer gets a rating from Moody’s which
is more favorable than the rating it gets from S&P. The results of the latter model, in
turn, show that the ?rst time an issuer comes to the market it pays a premium, but
issuing frequently does not bring the costs of bond ?nancing down. Importantly,
though the results of both models continue to show that recessions steepen the slope
of the credit spread curve on banks bonds, that is, they have a relatively large impact
on the cost of accessing the bond market for Baa-rated banks than for safer banks, say
banks that issue double or single-A rated bonds (when compared to the similar effect
on ?rms). This difference is shown in Figure 1, which shows the estimated marginal
impact of recessions on the cost of accessing the bond market for ?rms and banks with
the same credit rating as well as the difference between these marginal costs[16].
Since these results hold when we consider bank bonds alone and when we also
account for bonds issued by ?rms, this suggests that they are driven by factors other
Do markets
“discipline” all
banks equally?
119
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
than differences in the risk associated with credit ratings over the business cycle or
possible patterns of bank-bond issuance over the business cycle. Instead, a possible
explanation for our results is that in recessions there is a ?ight toward bank bonds, and
in particular toward safer banks. This suggests that the market does not discipline all
banks equally – it appears to be relatively less demanding on safer banks. Thus, far,
we have shown that these differences in the funding costs matter in a statistical sense.
Next, we investigate if they also matter from an economic point of view.
3.4 Economic signi?cance
Based on the estimates of Model 6 of Table IV, we have that in good times the estimated
difference between the spreads of Baa-rated bank bonds and the spreads of double-A
bank-rated bonds is 70 basis points. Still according to this model, recessions increase the
spreads of the former bonds by 51 basis points and those of the latter bonds by 10 basis
points. As a result, during recessions the difference between the spreads of Baa bank
bonds and those of double-A bank bonds increases by 41 basis points.
Under these conditions, a policy requiring banks to issue, say, $100 million, at a
certain frequency can have an important impact on the relative funding costs of banks.
Consider the case of the two banks, one rated double-A and other one rated Baa, and
assume that the credit rating of these banks does not change with the state of the
economy. In good times, that policy costs the Baa-rated bank $700,000 more than it
does the double-A rated bank. In recessions, however, this difference in funding costs
does not stay the same, even though the two banks continue to have their same ratings.
Instead, it increases to $1,110,000. As a result, in recessions the funding costs of the
Baa-rated bank increase by $410,000 relative to the funding costs of the double-A rated
bank. This suggests that the difference in the way the market disciplines banks of
different risk, which we identi?ed above, is not only statistically signi?cant but is also
important from an economic point of view.
4. Final remarks
In this paper, we showed that credit spreads on bank bonds vary not only across banks
of different risk but also with the state of the economy at the time of the issue. We also
showed that the impact of the state of the economy is not uniform across issuers of
Figure 1.
Cost of recessions to banks
and ?rms by credit rating
–0.4
–0.2
0
0.2
0.4
0.6
0.8
1
Aa A Baa
Firms Banks Banks minus firms
JFEP
1,1
120
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
different risk. These ?ndings imply that the information content of credit spreads on
bonds of banks varies both with the risk if the bank and the state of the economy.
These results are relevant for bank regulators and policymakers. They suggest, for
instance, that the proposals which have been put forth to introduce prompt corrective
action programs triggered by the credit spreads on bank bonds need to be
parameterized not only on the risk of the issuer, but also on the state of the economy if
they are to produce similar levels of monitoring across banks[17]. Similarly, proposals
to make bond issuance by banks mandatory need to account for this difference in the
market discipline across banks, otherwise they will alter the relative funding costs of
banks and thus promote different levels of monitoring across banks.
Notes
1. Diamond and Dybvig (1983) explain the role of demand deposits for the liquidity insurance
provision, and Diamond (1984) explains the role of loans for the monitoring services
provision. Calomiris and Kahn (1991), Flannery (1994) and Diamond and Rajan (1998)
explain the advantages of combining these two functions.
2. See Flannery (1998) and Kwast et al. (1999) for a review of the literature on the market
discipline of banks.
3. Bliss and Flannery (2002) do not ?nd that market discipline in?uences bank management
policies.
4. See Morgan and Stiroh (2005) for evidence of this subsidy.
5. Moody’s states “As a rule of thumb, we are looking through the next economic cycle or
longer. Because of this, our ratings are not intended to ratchet up and down with business or
supply-demand cycles [. . .]”.
6. Santos (2006) shows that it is costlier to access the bond market in recessions, and the
additional cost that comes with recessions does not affect issuers uniformly.
7. See Calomiris (1999) and Evanoff and Wall (2000) for proposals to make it mandatory for
banks to issue subordinated debt.
8. Evanoff and Wall (2000, 2001) propose the use of credit spreads on bank bonds over
treasuries instead of the capital to assets ratio as the trigger variable in the prompt
correction action scheme in place in the USA on the grounds that credit spreads are less
prone to manipulation by bank managers and are better predictors of bank risk.
9. We do not consider bonds issued prior to the second quarter of 1982 because Moody’s started
to use alpha-numeric ratings only in April of 1982 and use this information in our
methodology.
10. Between 1982:2 and 2002:2, the peaks occurred in (months in brackets): 1990(July)
and 2001(March), and the troughs occurred in: 1982(November), 1991(March) and
2001(November).
11. As we saw in the previous subsection, the lowest rated bonds issued by banks have a
Ba Moody’s rating. We have, however, bonds issued by ?rms with a Moody’s rating equal to
Caa as well as bonds with an even lower credit rating. Since the number of bonds with a
rating at issue date lower than Caa is rather small we chose to pool all of these bonds
together under the credit-rating dummy R
6
.
12. An important advantage of this approach to control for the creditworthiness of bonds over
the alternative approach sometimes adopted by bond pricing models, which uses a linear
variable to capture the rating of bonds, is that it does not assume that each unit change in
ratings has the same effect on credit spreads.
Do markets
“discipline” all
banks equally?
121
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
13. See, for example, Blackwell and Kidwell (1988), Fenn (2000), Collin-Dufresne et al. (2001),
Harrison (2001) and Santos (2006).
14. See Santos (2006) for evidence on the impact of rating splits on the cost of ?rms’ access to the
bond market.
15. The number of bank bonds in our sample with the highest below grade rating, Ba, is not
meaningful.
16. The estimates shown in Figure 1 were computed based on Model 6 of Table IV. We have
excluded from this ?gure triple-A rated bonds because there are only nine bank bonds in the
sample with this rating, and Ba-rated bonds because there are no bank bonds with this
rating that were issued during recessions.
17. Note that parameterizing these programs on spreads of bank bond computed over, say,
triple-A rated bonds, as is sometimes advocated, will not solve this problem because as we
saw in recessions the credit spread (over treasuries) curve on bank bonds not only shifts up
but it also becomes steeper.
References
Avery, R.B., Belton, T.M. and Goldberg, M.A. (1988), “Market discipline in regulating bank risk:
new evidence from the capital markets”, Journal of Money and Credit and Banking, Vol. 20
No. 4, pp. 597-610.
Blackwell, D.W. and Kidwell, D.S. (1988), “An investigation of cost differences between public
sales and private placements of debt”, Journal of Financial Economics, Vol. 22, pp. 253-78.
Bliss, R.R. and Flannery, M.J. (2002), “Market discipline in the governance of US bank holding
companies: monitoring vs. in?uencing”, European Finance Review, Vol. 6 No. 3, pp. 361-95.
Calomiris, C.W. (1999), “Building and incentive compatible safety net”, Journal of Banking &
Finance, Vol. 23, pp. 1499-519.
Calomiris, C.W. and Kahn, C.M. (1991), “The role of demandable debt in structuring optimal
banking arrangements”, American Economic Review, Vol. 81 No. 3, pp. 497-513.
Collin-Dufresne, P., Goldstein, R.S. and Martin, J.S. (2001), “The determinants of credit spreads
changes”, Journal of Finance, Vol. 56 No. 6, pp. 2177-207.
Diamond, D.W. (1984), “Financial intermediation and delegated monitoring”, Review of Economic
Studies, Vol. 51, pp. 393-414.
Diamond, D.W. and Dybvig, P.H. (1983), “Bank runs, deposit insurance and liquidity”, Journal
of Political Economy, Vol. 91, pp. 401-19.
Diamond, D.W. and Rajan, R.G. (1998), “Liquidity risk, liquidity creation and ?nancial fragility:
a theory of banking”, Journal of Political Economy, Vol. 109 No. 2, pp. 287-327.
Evanoff, D.E. and Wall, L.D. (2000), “Subordinated debt and bank capital reform”, in Kaufman,
G.G. (Ed.), Bank Fragility and Regulation: Evidence from Different Countries and Different
Times, JAI Press, Greenwich, CT, pp. 53-120.
Evanoff, D.E. and Wall, L.D. (2001), “Sub-debt yield spreads as bank risk measures”, Journal of
Financial Services Research, Vol. 20 No. 2, pp. 121-45.
Fenn, G.W. (2000), “Speed of issuance and the adequacy of disclosure in the 144A high-yield debt
market”, Journal of Financial Economics, Vol. 56, pp. 383-405.
Flannery, M.J. (1994), “Debt maturity and the deadweight cost of leverage: optimally ?nancing
banking ?rms”, American Economic Review, Vol. 84, pp. 320-31.
JFEP
1,1
122
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Flannery, M.J. (1998), “Using market information in prudential bank supervision: a review of the
US empirical evidence”, Journal of Money and Credit and Banking, Vol. 30 No. 3,
pp. 273-305.
Flannery, M.J. and Sorescu, S.M. (1996), “Evidence of bank market discipline in subordinated
debenture yields: 1983-1991”, Journal of Finance, Vol. 51 No. 4, pp. 1347-77.
Gorton, G. and Santomero, A.M. (1990), “Market discipline and bank subordinated debt: note”,
Journal of Money and Credit and Banking, Vol. 22 No. 1, pp. 119-28.
Harrison, P. (2001), “The impact of market liquidity in times of stress on the corporate bond
market: pricing, trading, and the availability of funds during heightened illiquidity”,
Mimeo, Board of Governors of the Federal Reserve System, Washington, DC.
Jagtiani, J., Kaufman, G. and Lemieux, C. (2002), “Do markets discipline banks and bank holding
companies? Evidence from the post-FDICIA period”, Journal of Financial Research, Vol. 25
No. 4, pp. 559-75.
Kwast, M.L., Covitz, D.M., Hancock, D., Houpt, J.V., Adkins, D.P., Barger, N., Bouchard, B.,
Connolly, J.F., Brady, T.F., English, W.B., Evanoff, D.D. and Wall, L.D. (1999), “Using
sub-debt as an instrument of market discipline”, Staff Study 172, Board of Governors of
the Federal Reserve System, Washington, DC.
Morgan, D.P. (2000), “Rating banks: risk and uncertainty in an opaque industry”, American
Economic Review, Vol. 92 No. 4, pp. 874-88.
Morgan, D.P. and Stiroh, K.J. (2005), “‘Too big to fail’ and market discipline of banks:
a cross-sector study of investor perceptions”, Staff Reports 220, Federal Reserve Bank of
New York, New York, NY.
Santos, J.A.C. (2006), “Why ?rm access to the bond market varies over the business cycle:
a theory and some evidence”, Journal of Banking & Finance, Vol. 30 No. 10, pp. 2715-36.
Sironi, A. (2003), “Testing for market discipline in the European banking industry: evidence from
subordinated debt issues”, Journal of Money Credit and Banking, Vol. 35 No. 3, pp. 443-72.
Corresponding author
Joa˜o A.C. Santos can be contacted at: [email protected]
Do markets
“discipline” all
banks equally?
123
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
This article has been cited by:
1. Martin Byford, Sinclair Davidson. 2013. Informing Depositors: A Proposed Reform Of The
Australian Banking Sector. Economic Papers: A journal of applied economics and policy 32:10.1111/
ecpa.2013.32.issue-4, 417-425. [CrossRef]
2. Luc Laeven. 2011. Banking Crises: A Review. Annual Review of Financial Economics 3, 17-40. [CrossRef]
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_767855241.pdf