Description
This paper aims to investigate whether stock markets can reduce the output costs of
banking crises. The work is motivated by Alan Greenspan’s claim that capital markets serve as a
financial “spare tire” in the event of a banking crisis.
Journal of Financial Economic Policy
Stock markets and the costs of banking crises
Tess DeLean J oseph P. J oyce
Article information:
To cite this document:
Tess DeLean J oseph P. J oyce , (2014),"Stock markets and the costs of banking crises", J ournal of Financial
Economic Policy, Vol. 6 Iss 4 pp. 342 - 361
Permanent link to this document:http://dx.doi.org/10.1108/J FEP-01-2014-0003
Downloaded on: 24 January 2016, At: 21:50 (PT)
References: this document contains references to 37 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 237 times since 2014*
Users who downloaded this article also downloaded:
Tao Zeng, (2014),"Derivative financial instruments, tax aggressiveness and firm market value", J ournal of
Financial Economic Policy, Vol. 6 Iss 4 pp. 376-390http://dx.doi.org/10.1108/J FEP-02-2014-0013
Dogga Satyanarayana Murthy, Suresh Kumar Patra, Amaresh Samantaraya, (2014),"Trade Openness,
Financial Development Index and Economic Growth: Evidence from India (1971-2012)", J ournal of Financial
Economic Policy, Vol. 6 Iss 4 pp. 362-375http://dx.doi.org/10.1108/J FEP-10-2013-0056
Abdul Rashid, Zainab J ehan, (2014),"The response of macroeconomic aggregates to monetary policy
shocks in Pakistan", J ournal of Financial Economic Policy, Vol. 6 Iss 4 pp. 314-330http://dx.doi.org/10.1108/
J FEP-04-2013-0016
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Stock markets and the costs of
banking crises
Tess DeLean and Joseph P. Joyce
Department of Economics, Wellesley College, Wellesley,
Massachusetts, USA
Abstract
Purpose – This paper aims to investigate whether stock markets can reduce the output costs of
banking crises. The work is motivated by Alan Greenspan’s claim that capital markets serve as a
fnancial “spare tire” in the event of a banking crisis.
Design/methodology/approach – We test the impact of stock market capitalization, liquidity and
turnover on the output losses of 76 banking crises in 66 countries over the period of 1975-2008.
Findings – Our results indicate that stock markets can mitigate the effect of banking crises on
economic activity. There is also some evidence that foreign equity holdings lower output costs.
Practical implications – These results suggest that the development of equity markets will
contribute to reducing the costs of banking crises. Such development, however, should be accompanied
by adequate supervisory and regulatory oversight.
Originality/value – Our analysis is the frst direct empirical investigation of the impact of stock
markets on the output costs of banking crises. This paper demonstrates that equity markets can lessen
the severity of such crises.
Keywords Financial markets, Equity, Banks, Stock markets, Banking crises, Output losses,
Foreign equity holdings
Paper type Research paper
1. Introduction
The banking crises that occurred during the Global Financial Crisis of 2007-2009 served
as a reminder that such events can impose signifcant costs upon the economies where
they occur. Banking crises interrupt the fow of credit, forcing otherwise viable frms
into bankruptcy and disrupting investment and other expenditures[1]. The crises also
showed that such events occur in advanced economies as well as in emerging markets
and developing economies.
This paper investigates whether stock markets can reduce the impact of bank crises
on economic activity[2]. The topic is motivated by Greenspan’s (1999) claimthat capital
markets can act as a “spare tire” during a bank crisis, providing another channel for
allocating credit to frms and thus mitigating the costs of the crisis. We specifcally test
whether stock markets have an impact on the output costs associated with banking
crises. This issue is relevant for the study of the contributions of banks and capital
markets to fnancial development.
In addition, we examine whether any such effects depend on the income level of the
economy. Several recent papers have claimed that the impact of the stock market on
economic activity varies across nations with different incomes[3]. We also test whether
JEL classifcation – G01, G10, G21
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JFEP
6,4
342
Journal of Financial Economic Policy
Vol. 6 No. 4, 2014
pp. 342-361
©Emerald Group Publishing Limited
1757-6385
DOI 10.1108/JFEP-01-2014-0003
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the openness of a country’s stock market to foreign investors affects its impact on
banking crises.
Our analysis is the frst direct empirical investigation of the impact of stock markets
on the output costs of bank crises. Our results indicate that stock market activity, as
measured by its depth, liquidity or turnover, does lower the costs of such crises. This
effect does not depend on a country’s income level. The impact of fnancial openness is
more mixed: capital account openness contributes to a fall in output, but foreign
holdings of equities may lower the size of the contraction. Abroad-based and diversifed
fnancial sector, therefore, is likely to recover from a crisis more quickly.
The next section provides a survey of related work previously done on this issue.
Section 3 describes the data and model specifcation, and the following section presents
our results. Section 5 contains the results of tests of robustness, and Section 6 concludes
with some policy implications.
2. Literature review
There is an extensive literature on the impact of banking crises on output (summarized
below), which has found that a number of macroeconomic, fnancial and regulatory
variables contribute to the depth of the contraction. But there are few studies that
examine the linkages of banks and equity markets and their economic impact during
banking crises. Their relationship merits investigation, as a diversifed fnancial system
may lessen the economic impact of a shock to one component of it, such as the banking
sector.
Levine (2005) reviews the theoretical literature on the comparative advantages of
bank-based and market-based fnancial systems, and also the view that they both
promote economic growth. Proponents of bank-based systems claim that banks are
better able to acquire information about frms and monitor their managers. Supporters
of market-based systems maintain that capital markets facilitate risk management and
avoid ineffciencies that can be associated with banks. Critics of this “debate” believe
that banks and markets both provide fnancial services that contribute to growth. The
empirical studies that Levine (2005) reviews show that fnancial intermediaries and
stock markets both enhance growth. Levine and Zervos (1998), for example, found that
stock market liquidity and banking sector development are associated with higher
economic growth. Beck and Levine (2004) reported similar results.
Allen et al. (2012) examined the relationship of the structure of fnancial systems and
crises, drawing upon data from75 banking crises in 65 countries over the period of 1980
to 2008. Their analysis revealed that while bank credit and stock market expansions are
positively linked before a crisis, this relationship is reversed during a crisis and for
several years thereafter. The size and turnover ratio of stock markets rise when bank
credit is still declining. They interpreted their results as evidence that frms can shift
fnancing from banks to equity markets, and that a better-developed fnancial system
allows a quicker recovery from disruption.
Laeven and Valencia (2013a) used frm-level data in a study of the impact of fnancial
frictions during the global fnancial crisis. They found that stock market development,
as measured by stock market capitalization, has a growth-enhancing effect on business
frms. They maintained that this result showed that stock markets can serve as a “spare
wheel” during times of banking distress.
343
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
As stated above, there are also studies that seek to ascertain the determinants of the
output losses that accompany banking crises[4]. This work includes papers by
Demirgüç-Kunt and Detragiache (1998), Domaç and Martinez Peria (2003), Angkinand
(2009), Cecchetti et al. (2009) and Aizenman and Noy (2013). We drew upon these works
in formulating our research strategy.
The primary focus of Demirgüç-Kunt and Detragiache’s (1998) paper was on the
causes of banking crises. Using data from the period of 1980-1994, they included
macroeconomic, fnancial and institutional variables as regressors. They used the same
variables in testing the determinants of the fscal costs of the crises in a sample that
included 19 to 24 countries, depending upon the specifcation. Their results indicate that
low GDP growth, adverse changes in the terms of trade, higher real interest rates and
infation rates, increases in the ratio of M2 to reserves (a sign of vulnerability to a
balance of payments crisis) and the share of private sector credit to GDP, and higher
lagged growth rates of private credit increase the fscal costs. The presence of a deposit
insurance scheme also increases this fscal cost, as does the duration of the crisis, while
an effective legal system has the opposite effect.
Domaç and Martinez Peria (2003) extended this work to examine the determinants of
bank crises, their fscal and output costs and their duration. They utilized data from the
period of 1980-1997 in a sample of 16-39 countries, depending on the specifcation. They
classifedtheir explanatoryvariables into macroeconomic, governmental andexchange rate
measures. They reported that higher infation, real interest rates and credit growth all
increase the output costs of banking crises, while capital infows and fscal surpluses lower
them. Peggedexchange rate regimes whenmeasuredonade jure basis alsoraise these costs.
Angkinand (2009) used data from 47 banking crises in advanced and emerging market
countries during the 1970s to 2003 to investigate the impact of fnancial regulations on their
output costs. She found that increased capital regulations and the existence of deposit
insurance schemes lower output costs, while restrictions on the range of bank activities
increase them. She interpreted the latter result as an indication that banks with a more
diverse range of activities are more robust to shocks. In addition, the coeffcients of the
pre-crisis GDP growth rate and real GDP per capita are negative and signifcant in many of
the output loss regressions, while currency crises have a positive impact.
Cecchetti et al. (2009) examined the length, depth and output costs of 40 banking
crises that have occurred since 1980. Their results showed, as did the previous works,
that declines in output are inversely related to GDP growth rates preceding the crisis.
They also found that the depth of a contraction is larger when a banking crisis is
accompanied by a currency or sovereign debt crisis.
Aizenman and Noy (2013) studied the determinants of the severity of banking crises
in high- and middle-income countries over the period of 1980-2010. They found that GDP
growth in the period prior to the crisis lowers the output loss in the full sample. They
also reported that capital account openness raised the output loss during the global
fnancial crisis.
Our analysis, therefore, builds upon research from several areas. We investigate the
determinants of the size of the economic contractions that accompany banking crises,
using control variables drawn from the previous literature. We add measures of stock
market development that have been utilized in studies of fnancial development and
growth. We leave the impact of bond markets on output costs for future analysis, as their
behavior can be very different from equity markets during crises[5].
JFEP
6,4
344
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
3. Data and methodology
The data on the dates of banking crises and their output losses are the focus of our
analysis. We drew upon the dataset provided by Laeven and Valencia (2013b). Earlier
versions of this database have been used in many papers, including several of the
studies of output costs cited in the previous section.
Laeven and Valencia (2013b) report the dates of the banking crises over the period of
1970-2011 and the resulting output losses, which are calculated as the percentage deviation
of actual real GDPfromits trendvalue[6]. We deletedobservations withmissingdata, which
resulted in a sample of 76 crisis episodes in 66 countries. The crises are listed in Table I. The
size of our sample is larger than those of the previous studies of output losses.
Figure 1 shows the number of crises in our sample by the decade of their start dates,
and indicates that the largest number of crises took place during the decade of the 1990s,
Table I.
Countries and initial crisis
year
Algeria (1990) Latvia (2008)
Argentina (1989, 1995, 2001) Luxembourg (2008)
Austria (2008) Madagascar (1988)
Belgium (2008) Malaysia (1997)
Bolivia (1986, 1994) Mexico (1981, 1994)
Brazil (1994) Nepal (1988)
Burundi (1994) Netherlands (2008)
Cameroon (1987, 1995) Niger (1983)
Cape Verde (1993) Nigeria (1991)
Central African Republic (1995) Norway (1991)
Chad (1992) Panama (1988)
Chile (1981) Philippines (1983, 1997)
Colombia (1982, 1998) Portugal (2008)
Denmark (2008) Russia (2008)
Ecuador (1998) Senegal (1988)
Egypt (1980) Sierra Leone (1990)
El Salvador (1989) Slovak Republic (1998)
Finland (1991) Slovenia (2008)
France (2008) Spain (2008)
Germany (2008) Sri Lanka (1989)
Ghana (1982) Swaziland (1995)
Greece (2008) Sweden (1991, 2008)
Guinea-Bissau (1995) Switzerland (2008)
Haiti (1994) Thailand (1983, 1997)
Hungary (2008) Togo (1993)
India (1993) Tunisia (1991)
Indonesia (1997) Turkey (1982, 2000)
Ireland (2008) Uganda (1994)
Jamaica (1996) United Kingdom (2007)
Japan (1997) United States (2007)
Jordan (1989) Venezuela (1994)
Kenya (1992) Vietnam (1997)
Korea (1997) Zambia (1995)
Source: Laeven and Valencia (2013b)
345
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
a period of fnancial instability in many emerging economies. Almost all of the crises of
the decade of the 2000s occurred during the global fnancial crisis of 2008-2009.
Figure 2 displays the output losses associated with the crises episodes. There were a
large number of crises for which a zero output loss is recorded, and many of these
occurred in low- or middle-income countries in Africa, Asia and Latin America.
Our base regression equation is specifed as:
LOSS
i
? ? ? ?’X
i
? ?
(
Stock Market
)
i
? e
i
(1)
where LOSS is the output loss accompanying a banking crisis in country i, X
i
is a vector
of control variables, Stock Market
i
is an indicator of the stock market and e
i
is the error
term. If stock markets do diminish the impact of a banking crisis, ? should be negative.
We relied on the existing literature in choosing economic, fnancial and regulatory
control variables[7]. The initial specifcation included the growth rate of real GDP
0
5
10
15
20
25
30
35
40
1980s 1990s 2000s
N
u
m
b
e
r
o
f
C
r
i
s
e
s
Source: Laeven and Valencia (2013b)
Figure 1.
Number of crises by
decade
0
2
4
6
8
10
12
14
16
18
N
u
m
b
e
r
o
f
C
r
i
s
i
s
Source: Laeven and Valencia (2013b)
Figure 2.
Output losses of banking
crises
JFEP
6,4
346
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
(YGR), infation as measured by changes in the GDPPrice Defator (INF) and the amount
of private credit provided by banks and other fnancial institutions relative to GDP
(CRED)[8]. We also include a dummy variable to account for the presence for a deposit
insurance scheme (INS). All the variables were lagged by one year before the frst year
of the banking crisis to avoid feedback from the crisis. In addition, there is a dummy
variable to indicate whether a currency crisis occurred before or at the time of the
banking crisis (CUR). The sources for these variables are listed in the Data Appendix.
We use several indicators of stock markets in the analysis. The variables measure
different characteristics of stock markets, as no one proxy can adequately address all the
features of these markets. These data were obtained from the Database on Financial
Development and Structure (Beck et al., 2000, 2010). The variables, which were lagged
by one year before the frst year of the crisis, include:
• stock market capitalization scaled by GDP, a measure of the relative size of a stock
market (CAP);
• the value of the shares traded scaled by GDP, a measure of the liquidity of a stock
market relative to the domestic economy (LIQ); and
• the value of the traded shares scaled by market capitalization, the turnover of the
stock market with respect to its own size (TUR).
The frst variable represents the depth of the stock markets, which has been linked to the
rate of economic development. The value of shares traded relative to GDP is a measure
of liquidity, which contributes to lower costs and more effcient markets. The value of
shares traded scaled by market capitalization is another measure of liquidity as well as
of turnover and transactions costs. These features contribute to the ability of equity
markets to allocate savings to their most productive use[9].
The model is estimated using a Tobit specifcation, as the output losses are truncated
at zero. However, we also estimated the main equations using Ordinary Least Squares
OLS, and those results are reported below as one of the tests of robustness.
4. Results
4.1 Stock market variables
Table II reports our initial results for the control variables and the stock market
variables. Equation 2.1 includes the control variables. Ahigher lagged economic growth
rate lowers the cost of a banking crisis, and the coeffcient is signifcant at the 5 per cent
level. Faster growing economies contract less during a crisis. This result is consistent
with the fndings of the studies reported in Section 2.
While some previous studies of output losses found infation to have a signifcant
impact, the positive coeffcient reported in our results is insignifcant. This result is most
likely driven by crisis episodes with high infation but low output costs. The 1994
banking crisis in Brazil, for example, occurred during a period of hyperinfation (over
2,000 per cent in the year preceding the crisis), but a recorded output loss of 0 per cent.
Arise in the ratio of private credit to output, on the other hand, has a positive impact
that is signifcant at the 1 per cent level. Demirgüç-Kunt and Detragiache (1998),
Angkinand (2009) and Furceri and Zdzienicka (2012) also reported that higher credit/
GDP ratios led to larger output costs. A credit “boom” will be followed by a subsequent
“bust”, as borrowers default on their obligations and credit markets take time to recover.
347
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table II.
Output costs: stock
market variables
V
a
r
i
a
b
l
e
s
(
2
.
1
)
(
2
.
2
)
(
2
.
3
)
(
2
.
4
)
(
2
.
5
)
Y
G
R
?
2
.
9
0
6
*
*
(
1
.
1
4
7
)
?
0
.
3
9
2
(
1
.
0
3
4
)
?
0
.
1
5
7
(
1
.
0
1
7
)
?
0
.
5
2
7
(
1
.
0
3
2
)
?
0
.
5
4
1
(
0
.
9
9
7
)
I
N
F
0
.
0
0
1
(
0
.
0
0
3
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
2
6
5
*
*
*
(
0
.
0
9
3
)
0
.
1
0
2
(
0
.
0
8
3
)
0
.
2
3
3
*
*
(
0
.
1
0
6
)
0
.
2
1
2
*
*
(
0
.
0
9
9
)
0
.
1
3
9
(
0
.
0
9
5
)
I
N
S
?
1
3
.
2
7
4
(
9
.
7
3
8
)
?
1
4
.
2
3
7
(
8
.
6
1
5
)
?
1
3
.
3
4
9
(
8
.
3
9
9
)
?
1
1
.
9
7
3
(
8
.
5
4
4
)
?
8
.
7
3
2
(
8
.
8
0
4
)
C
U
R
1
7
.
5
5
1
(
1
0
.
6
4
0
)
2
.
7
4
7
(
9
.
1
9
2
)
6
.
2
6
7
(
9
.
1
3
6
)
2
.
8
6
8
(
9
.
0
7
2
)
3
.
3
2
6
(
8
.
8
1
9
)
S
M
L
S
?
8
9
.
7
4
3
*
*
*
(
1
6
.
8
9
1
)
?
8
7
.
9
0
1
*
*
*
(
1
6
.
8
4
3
)
?
8
7
.
1
4
9
*
*
*
(
1
6
.
9
5
1
)
?
8
6
.
9
9
7
*
*
*
(
1
6
.
3
8
4
)
C
A
P
?
0
.
1
8
1
*
(
0
.
0
9
5
)
L
I
Q
?
0
.
1
5
0
*
*
(
0
.
0
7
2
)
T
U
R
?
0
.
1
6
6
*
(
0
.
0
9
6
)
C
o
n
s
t
a
n
t
2
5
.
6
9
8
*
*
*
(
7
.
7
8
9
)
4
4
.
0
8
5
*
*
*
(
7
.
0
3
5
)
4
1
.
7
4
4
*
*
*
(
6
.
9
7
0
)
4
2
.
1
7
4
*
*
*
(
7
.
1
4
6
)
4
4
.
9
4
8
*
*
*
(
6
.
9
1
2
)
P
s
e
u
d
o
R
2
0
.
0
2
2
0
.
0
9
4
0
.
0
9
9
0
.
1
0
2
0
.
1
0
2
N
7
6
7
6
7
6
7
4
7
4
N
o
t
e
s
:
*
S
i
g
n
i
f
c
a
n
t
a
t
1
0
%
;
*
*
a
t
5
%
;
*
*
*
a
t
1
%
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
Y
G
R
?
l
a
g
g
e
d
g
r
o
w
t
h
r
a
t
e
o
f
G
D
P
;
I
N
F
?
l
a
g
g
e
d
i
n
f
a
t
i
o
n
r
a
t
e
;
C
R
E
D
?
l
a
g
g
e
d
p
r
i
v
a
t
e
c
r
e
d
i
t
/
G
D
P
;
I
N
S
?
l
a
g
g
e
d
d
e
p
o
s
i
t
i
n
s
u
r
a
n
c
e
d
u
m
m
y
;
C
U
R
?
c
u
r
r
e
n
c
y
c
r
i
s
i
s
d
u
m
m
y
;
S
M
L
S
?
o
u
t
p
u
t
l
o
s
s
l
e
s
s
t
h
a
n
1
%
d
u
m
m
y
;
C
A
P
?
l
a
g
g
e
d
s
t
o
c
k
m
a
r
k
e
t
c
a
p
i
t
a
l
i
z
a
t
i
o
n
/
G
D
P
;
L
I
Q
?
l
a
g
g
e
d
m
a
r
k
e
t
v
a
l
u
e
o
f
s
h
a
r
e
s
t
r
a
d
e
d
/
G
D
P
;
T
U
R
?
l
a
g
g
e
d
m
a
r
k
e
t
v
a
l
u
e
o
f
s
h
a
r
e
s
t
r
a
d
e
d
/
s
t
o
c
k
m
a
r
k
e
t
c
a
p
i
t
a
l
i
z
a
t
i
o
n
JFEP
6,4
348
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The coeffcient of the deposit insurance dummy variable is negative but not signifcant,
while the coeffcient of the currency crisis indicator is positive and not signifcant. The
pseudo R
2
at 0.022 is relatively low, which refects several factors. First, pseudo-R
2
s
from Tobit estimations are often lower than the adjusted R
2
s reported in the OLS
estimations of the same equations. Table VII below with the results of OLS estimations
of these equations reports adjusted R
2
s with higher values. Second, the government’s
policies in reaction to the initial occurrence of a bank crisis are most likely a major factor
in determining the fnal cost of a crisis. These responses would be endogenous to the
crisis. Finally, the results may refect the distribution of output losses that we show in
Figure 2. Many of these are zero or slightly higher. When we use a dummy variable to
account for those crises with output losses below1 per cent (SMLS) as we do in equation
2.2, the variable itself is quite signifcant and the pseudo-R
2
rises to 0.094. However, the
coeffcient of the growth rate of output falls in signifcance.
We add the stock market variables in the following three equations. The coeffcient of
the capitalization of the stock market scaled by GDP in equation 2.2 is negative and
signifcant at the 10 per cent level. We repeat the specifcation with the value of shares
traded scaled by GDP in equation 2.3, and it is also negative and signifcant at the 5 per
cent level. Finally, we utilized the measure of turnover in the stock market in equations
2.4, and its coeffcient is negative and signifcant at the 10 per cent level. When the three
equations are estimated without the dummy for crises with small losses, the signifcance
levels of the stock market variables rise to 5, 1 and 10 per cent, while the GDP growth
variable is signifcant at the 5 or 1 per cent levels.
Could these results result from causality from the output loss to the lagged stock
market? It is possible that investors who foresee that a banking crisis will not be severe
could bid up stock market prices in the year preceding such a crisis. But such foresight
seems to be absent frommost banking crises, including the recent global crisis. Reinhart
and Rogoff (2009) fnd that equity prices usually peak before the occurrence of a banking
crisis and then decline as the crisis approaches before recovering. Overall, therefore, our
results are consistent with the hypothesis that equity markets can lower the output loss
of a banking crisis.
Table III presents the marginal effects of the stock market variables evaluated at
the mean values of the independent variables. For the sake of comparison, we repeat
the results for their impact on the latent variable from Table II with those computed
for actual values of the output loss (also called the “unconditional expected value”)
and those computed using only uncensored observations. If we use the latter values,
we see that the results for the three stock market variables are quite similar, ranging
from ?0.083 to ?0.101.
Table III.
Output costs: stock
market variables:
marginal effects at means
Stock market variable
On latent
variable
On actual
variable
Uncensored
observations
Stock market capitalization/GDP ?0.182 ?0.140 ?0.101
Value of shares traded/GDP ?0.150 ?0.115 ?0.083
Value of shares traded/stock market capitalization ?0.160 ?0.128 ?0.092
Note: Estimates derived from Equations 2.3, 2.4 and 2.5 in Table II
349
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
We extended the results reported in Table II by estimating these equations with
additional variables. We used two lagged macro policy variables, a short-term nominal
interest rate and government expenditures scaled by GDP. The results, available from
the authors, show that neither policy variable is signifcant. Other variables that we
tested, including M2/Reserves and capital fows scaled by GDP, yielded similar results.
4.2 Stock markets and income levels
In this section, we examine whether the impact of stock markets on the output loss
of a banking crisis differs across income levels, as well as test for whether the size of
the output loss is directly infuenced by income. Several recent papers have reported
that the economic effects of stock market activity do vary across income levels.
Demirgüç-Kunt et al. (2013) for example, found that the sensitivity of output to bank
development falls as economies grow, while the sensitivity of output to the
development of securities markets rises. Similarly, Rioja and Valev (2014) claim that
stock markets have positive effects on productivity and capital accumulation in
high-income countries but not in low-income economies. Barajas et al. (2013)
reported that the impact of stock market turnover on growth is smaller in
low-income countries.
We frst used real GDP per capita (YPOP) as our income measure, and these results
appear in Table IV. The private credit variable is highly signifcant in all the
estimations, as is the dummy for crises with small losses. The income per capita variable
has a negative coeffcient that is signifcant at different levels in fve of the six
estimations. Countries with higher income levels have lower output costs, although the
size of the impact is small.
The stock market liquidity measure is signifcant at the 10 and 5 per cent levels in the
two equations that include it, while the interaction term here is positive and signifcant
at the 10 per cent level. Neither of the two other interaction terms is signifcant.
We then used the World Bank classifcations of nations as low-income,
middle-income (aggregated fromlower- and upper-middle income countries to formone
category) and high-income countries to include dummies for the frst two categories
(LOW, MID). We interacted these with each of the stock market measures, and the
results with and without the interaction terms appear in Table V. The stock market
capitalization and liquidity measures are signifcant at the 10 per cent level, while the
private credit variable is signifcant at different levels in all the estimations. But neither
the income category variables nor their interaction terms were signifcant, and omitting
the interaction terms does not change the signifcance of the results for the two income
variables. The results, therefore, are similar to those in the previous table. There is
virtually no evidence to indicate that the impact of the stock market on the output losses
of bank crises varies systemically by income level.
4.3 Stock markets in open economies
We next tested whether the impact of the stock market variables is infuenced by the
openness of the stock markets to foreign investors. Such an effect could be either
positive or negative. On the one hand, foreign equity infows may lead to deeper and
more resilient markets. On the other hand, capital fight by foreign and domestic
investors could depress asset prices and reduce liquidity.
JFEP
6,4
350
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table IV.
Output costs: stock
market variables and real
income per capita
V
a
r
i
a
b
l
e
s
(
4
.
1
)
(
4
.
2
)
(
4
.
3
)
(
4
.
4
)
(
4
.
5
)
(
4
.
6
)
Y
G
R
?
0
.
3
6
6
(
1
.
0
1
4
)
?
0
.
3
4
0
(
1
.
0
0
2
)
?
0
.
7
6
6
(
1
.
0
0
3
)
?
0
.
6
3
7
(
0
.
9
7
6
)
?
0
.
7
0
3
(
0
.
9
6
8
)
?
0
.
6
7
8
(
0
.
9
7
0
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
2
8
6
*
*
(
0
.
1
1
2
)
0
.
3
3
4
*
*
*
(
0
.
1
1
9
)
0
.
3
5
6
*
*
*
(
0
.
1
1
8
)
0
.
4
7
4
*
*
*
(
0
.
1
3
1
)
0
.
2
5
9
*
*
(
0
.
1
1
0
)
0
.
2
7
1
*
*
(
0
.
1
1
7
)
I
N
S
?
8
.
4
6
4
(
9
.
0
4
3
)
?
6
.
4
5
9
(
9
.
1
3
7
)
?
4
.
6
7
3
(
8
.
9
4
0
)
?
3
.
6
7
1
(
8
.
7
2
5
)
?
2
.
5
2
9
(
9
.
0
2
7
)
?
1
.
7
4
2
(
9
.
3
3
8
)
C
U
R
1
.
8
1
9
(
9
.
6
1
7
)
2
.
6
9
8
(
9
.
5
5
7
)
?
1
.
8
9
3
(
9
.
0
4
8
)
1
.
4
5
3
(
8
.
9
5
1
)
?
0
.
9
0
4
(
8
.
7
8
1
)
?
0
.
3
0
4
(
8
.
9
6
3
)
S
M
L
S
?
8
7
.
9
7
4
*
*
*
(
1
6
.
3
9
5
)
?
8
8
.
7
0
8
*
*
*
(
1
6
.
2
2
4
)
?
8
6
.
7
9
5
*
*
*
(
1
6
.
1
7
6
)
?
8
6
.
4
0
4
*
*
*
(
1
5
.
6
4
5
)
?
8
6
.
5
0
9
*
*
*
(
1
5
.
7
5
6
)
?
8
6
.
5
9
2
*
*
*
(
1
5
.
7
6
2
)
Y
P
O
P
?
0
.
0
0
1
(
0
.
0
0
1
)
?
0
.
0
0
1
*
(
0
.
0
0
1
)
?
0
.
0
0
1
*
*
(
0
.
0
0
0
)
?
0
.
0
0
2
*
*
*
(
0
.
0
0
1
)
?
0
.
0
0
1
*
(
0
.
0
0
0
)
?
0
.
0
0
1
*
(
0
.
0
0
1
)
C
A
P
?
0
.
1
1
7
(
0
.
1
0
5
)
?
0
.
2
0
0
(
0
.
1
2
7
)
C
A
P
*
Y
P
O
P
0
.
0
0
0
(
0
.
0
0
0
)
L
I
Q
?
0
.
1
3
4
*
(
0
.
0
7
0
)
?
0
.
4
2
7
*
*
(
0
.
1
7
1
)
L
I
Q
*
Y
P
O
P
0
.
0
0
0
*
(
0
.
0
0
0
)
T
U
R
?
0
.
1
4
7
(
0
.
0
9
3
)
?
0
.
1
9
9
(
0
.
1
8
7
)
T
U
R
*
Y
P
O
P
0
.
0
0
0
(
0
.
0
0
0
)
C
o
n
s
t
a
n
t
4
2
.
1
3
7
*
*
*
(
6
.
8
6
5
)
4
2
.
7
0
0
*
*
*
(
6
.
8
1
6
)
4
1
.
5
5
1
*
*
*
(
6
.
8
9
1
)
3
9
.
7
5
2
*
*
*
(
6
.
7
4
4
)
4
4
.
4
6
1
*
*
*
(
6
.
6
8
7
)
4
4
.
4
9
7
*
*
*
(
6
.
6
8
0
)
P
s
e
u
d
o
R
2
0
.
1
0
3
0
.
1
0
4
0
.
1
0
9
0
.
1
1
4
0
.
1
0
8
0
.
1
0
8
N
7
6
7
6
7
4
7
4
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
Y
P
O
P
?
r
e
a
l
G
D
P
/
p
o
p
u
l
a
t
i
o
n
i
n
U
S
d
o
l
l
a
r
s
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
351
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table V.
Output costs: stock
market variables and
income category variables
V
a
r
i
a
b
l
e
s
(
5
.
1
)
(
5
.
2
)
(
5
.
3
)
(
5
.
4
)
(
5
.
5
)
(
5
.
6
)
Y
G
R
?
0
.
5
5
4
(
1
.
1
0
1
)
?
0
.
4
8
6
(
1
.
1
3
4
)
?
0
.
8
6
6
(
1
.
1
1
3
)
?
0
.
6
9
0
(
1
.
1
5
4
)
?
1
.
0
1
8
(
1
.
0
8
8
)
?
1
.
0
6
0
(
1
.
1
1
3
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
?
0
.
0
0
0
(
0
.
0
0
2
)
?
0
.
0
0
1
(
0
.
0
0
2
)
C
R
E
D
0
.
3
1
5
*
*
*
(
0
.
1
1
7
)
0
.
3
1
5
*
*
*
(
0
.
1
1
7
)
0
.
2
8
5
*
*
(
0
.
1
1
2
)
0
.
3
0
0
*
*
(
0
.
1
1
4
)
0
.
1
8
1
*
(
0
.
1
0
6
)
0
.
1
9
5
*
(
0
.
1
0
6
)
I
N
S
?
6
.
8
2
5
(
9
.
3
8
8
)
?
4
.
5
4
3
(
1
0
.
5
8
2
)
?
6
.
3
9
4
(
9
.
5
3
2
)
?
5
.
8
0
3
(
1
0
.
8
2
5
)
?
6
.
0
9
3
(
9
.
4
3
7
)
?
4
.
7
7
0
(
1
0
.
5
1
8
)
C
U
R
1
.
4
4
6
(
9
.
3
6
6
)
0
.
0
0
6
(
9
.
8
6
2
)
?
1
.
3
0
5
(
9
.
3
0
0
)
0
.
5
9
0
(
9
.
7
9
9
)
?
0
.
1
2
4
(
9
.
1
8
2
)
3
.
2
9
1
(
9
.
8
4
3
)
S
M
L
S
?
8
8
.
0
5
6
*
*
*
(
1
6
.
3
4
9
)
?
8
7
.
6
9
6
*
*
*
(
1
6
.
2
9
0
)
?
8
7
.
7
4
2
*
*
*
(
1
6
.
5
6
1
)
?
8
7
.
3
9
9
*
*
*
(
1
6
.
6
3
1
)
?
8
6
.
9
6
4
*
*
*
(
1
6
.
2
2
8
)
?
8
6
.
4
7
1
*
*
*
(
1
6
.
6
2
2
)
L
O
W
1
7
.
7
5
5
(
1
6
.
9
2
4
)
1
8
.
8
5
9
(
1
9
.
6
0
3
)
1
5
.
8
6
7
(
1
7
.
2
6
5
)
2
0
.
4
1
5
(
1
9
.
7
7
8
)
7
.
8
5
3
(
1
8
.
2
0
6
)
1
2
.
5
7
7
(
2
0
.
1
6
1
)
M
I
D
2
2
.
4
5
6
*
(
1
3
.
2
1
2
)
2
0
.
5
4
5
(
1
5
.
1
7
4
)
2
0
.
6
4
4
(
1
3
.
5
3
0
)
2
3
.
9
3
1
(
1
4
.
3
8
7
)
1
7
.
0
0
0
(
1
4
.
1
8
9
)
2
5
.
5
9
7
(
1
6
.
6
4
9
)
C
A
P
?
0
.
1
6
0
*
(
0
.
0
9
3
)
?
0
.
1
9
3
(
0
.
1
2
1
)
C
A
P
*
L
O
W
?
1
.
1
5
4
(
3
.
3
4
8
)
C
A
P
*
M
I
D
0
.
0
6
9
(
0
.
1
6
4
)
L
I
Q
?
0
.
1
2
7
*
(
0
.
0
7
2
)
?
0
.
1
1
2
(
0
.
0
7
6
)
L
I
Q
*
L
O
W
?
3
6
.
4
7
6
(
1
5
9
.
9
3
4
)
L
I
Q
*
M
I
D
?
0
.
1
2
3
(
0
.
1
9
9
)
T
U
R
?
0
.
1
2
4
(
0
.
1
0
3
)
?
0
.
0
9
5
(
0
.
1
0
8
)
T
U
R
*
L
O
W
?
1
.
3
2
1
(
8
.
6
5
1
)
T
U
R
*
M
I
D
?
0
.
3
1
4
(
0
.
3
2
8
)
C
o
n
s
t
a
n
t
2
1
.
1
6
9
(
1
5
.
9
6
3
)
2
1
.
5
3
7
(
1
7
.
7
4
8
)
2
3
.
6
3
5
(
1
6
.
3
5
5
)
1
9
.
3
9
9
(
1
7
.
8
2
4
)
3
2
.
3
8
8
*
(
1
7
.
0
8
6
)
2
6
.
7
8
7
(
1
8
.
4
5
8
)
P
s
e
u
d
o
R
2
0
.
1
0
5
0
.
1
0
5
0
.
1
0
6
0
.
1
0
7
0
.
1
0
5
0
.
1
0
7
N
7
6
7
6
7
4
7
4
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
L
O
W
?
l
o
w
-
i
n
c
o
m
e
c
o
u
n
t
r
y
d
u
m
m
y
;
M
I
D
?
m
i
d
d
l
e
-
i
n
c
o
m
e
c
o
u
n
t
r
y
d
u
m
m
y
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
352
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
There are several channels through which foreign holdings of equities could
infuence an economy. Mishkin (2006) maintains that there are both direct and indirect
benefts to liberalizing restrictions on equity fows. The direct benefts of stock market
liberalizations are the increase in the availability of funds with lower costs, while the
indirect benefts are the institutional reforms that decrease informational asymmetries
and allow a more effcient allocation of savings. Bekaert et al. (2001, 2005) provide
empirical analyses of these effects.
Jeanne et al. (2012) suggest that any crisis-induced equity outfows are likely to be
short-lived. As evidence, they cite the reemergence of equity fows to emerging markets
in the wake of the Asian fnancial crisis of 1997-1998 and also after the recent global
fnancial crisis. The return of foreign investors can contribute to the reestablishment of
fnancing for frms cut off from other domestic sources.
We used two measures of fnancial openness, one de facto and the second de jure. The
frst is the lagged value of portfolio equity liabilities scaled by GDP (EQL), which are
included in the updated dataset of gross foreign assets and liabilities provided by Lane
and Milesi-Ferretti (2007). This variable demonstrates the relative size of foreign equity
holdings. We also interact this variable with the currency crisis variable to see if its
effect changes in the presence of such a crisis.
Our second measure is a broad-based measure of capital account openness, the
lagged Chinn and Ito (2006, 2013) index (KAO). The index is based on data reported
in the International Monetary Funds (IMF’s) Annual Report on Exchange
Arrangements and Restrictions on the existence of multiple exchange rates,
restrictions on the current and capital accounts and requirements to surrender
export proceeds. The Chinn-Ito index is the frst principal component of these four
indicators, and we rescaled the measure to range from 0 (completely closed) to 100
(completely liberalized). There are other more specifc de jure measures of capital
regulations, such as Schindler’s (2009) measures of restrictions on equity fows, but
these have more limited coverage in terms of both countries and dates.
Table VI presents the results of the estimations of these specifcations. The stock
market capitalization and the value of shares traded scaled by GDP continue to be
negative and signifcant at the 10 and 5 per cent levels[10]. The foreign equity
liabilities variable is not signifcant in these equations, nor is its interaction with the
currency crisis variable. However, the variable has a negative coeffcient which is
signifcant the at 10 per cent level in equations 6.5 and 6.6, where the domestic stock
market measure is the turnover proxy, which itself is not signifcant. The
signifcance level of the variable rises to 5 per cent when the small loss dummy
variable is omitted from the estimations. The signifcance of the foreign equity
variable here may refect the impact of foreign holdings on the amount of trading
activity, with foreign investors who are more diversifed better able to resume
trading after a crisis erupts.
The openness of a country’s stock markets, therefore, may have a different impact on
its economy during a crisis than the negative impact that fnancial openness in general
may have. If foreign investors return after the outbreak of the crisis, their purchases may
allow the domestic market to be better able to provide credit to frms. The policy
implication is that emerging markets should consider removing restrictions on foreign
access to domestic to equity markets relatively early on as part of a phased opening of
the capital account.
353
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table VI.
Output costs: stock
market variables and
equity market openness
V
a
r
i
a
b
l
e
s
(
6
.
1
)
(
6
.
2
)
(
6
.
3
)
(
6
.
4
)
(
6
.
5
)
(
6
.
6
)
Y
G
R
?
0
.
1
7
1
(
1
.
0
1
9
)
?
0
.
3
8
7
(
1
.
0
3
9
)
?
0
.
4
5
2
(
1
.
0
3
8
)
?
0
.
4
4
8
(
1
.
0
7
0
)
?
0
.
6
3
9
(
1
.
0
0
1
)
?
0
.
5
9
0
(
1
.
0
3
0
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
1
8
9
(
0
.
1
2
3
)
0
.
2
0
3
(
0
.
1
2
3
)
0
.
1
9
1
(
0
.
1
2
0
)
0
.
1
9
1
(
0
.
1
2
1
)
0
.
2
7
2
*
*
(
0
.
1
2
1
)
0
.
2
7
6
*
*
(
0
.
1
2
3
)
I
N
S
?
1
3
.
2
4
3
(
8
.
5
2
8
)
?
1
1
.
2
1
9
(
8
.
7
3
7
)
?
1
1
.
6
4
5
(
8
.
5
7
9
)
?
1
1
.
6
7
8
(
8
.
7
7
6
)
?
8
.
6
7
2
(
8
.
7
9
9
)
?
8
.
9
7
5
(
8
.
9
2
8
)
C
U
R
6
.
8
0
8
(
9
.
1
7
0
)
1
.
6
5
5
(
1
0
.
6
9
6
)
2
.
7
7
7
(
8
.
9
5
7
)
2
.
8
7
9
(
1
0
.
6
9
7
)
2
.
6
2
2
(
8
.
6
4
8
)
3
.
7
4
8
(
1
0
.
3
0
4
)
S
M
L
S
?
8
8
.
4
1
1
*
*
*
(
1
7
.
0
3
4
)
?
8
7
.
5
0
7
*
*
*
(
1
6
.
8
7
2
)
?
8
7
.
5
3
2
*
*
*
(
1
7
.
1
3
1
)
?
8
7
.
5
4
1
*
*
*
(
1
7
.
1
4
2
)
?
8
3
.
9
9
0
*
*
*
(
1
6
.
5
6
2
)
?
8
4
.
0
5
2
*
*
*
(
1
6
.
5
9
0
)
K
A
O
?
0
.
0
2
5
(
0
.
1
4
4
)
0
.
0
1
0
(
0
.
1
4
8
)
?
0
.
0
2
6
(
0
.
1
4
6
)
?
0
.
0
2
6
(
0
.
1
4
7
)
0
.
0
0
0
(
0
.
1
4
1
)
?
0
.
0
0
3
(
0
.
1
4
2
)
C
A
P
?
0
.
1
9
4
(
0
.
1
0
8
)
?
0
.
2
8
4
(
0
.
1
4
7
)
L
I
Q
?
0
.
1
6
2
*
(
0
.
0
7
6
)
?
0
.
1
6
2
*
*
(
0
.
0
7
8
)
T
U
R
?
0
.
0
6
0
(
0
.
1
2
5
)
?
0
.
0
6
2
(
0
.
1
2
5
)
E
Q
L
1
0
.
1
2
2
(
7
.
3
6
9
)
1
0
.
2
1
4
(
7
.
3
2
9
)
1
0
.
3
5
2
(
7
.
4
0
8
)
1
0
.
3
4
2
(
7
.
4
2
9
)
?
7
6
.
4
3
8
*
(
4
1
.
8
6
3
)
?
7
5
.
4
1
9
*
(
4
2
.
1
6
3
)
E
Q
L
*
C
U
R
1
3
2
.
9
4
8
(
1
4
5
.
3
2
5
)
?
1
.
9
6
5
(
1
1
1
.
8
1
6
)
?
2
1
.
1
9
4
(
1
0
5
.
4
4
3
)
C
o
n
s
t
a
n
t
4
4
.
0
2
0
*
*
*
(
7
.
7
6
6
)
4
3
.
8
3
6
*
*
*
(
7
.
7
0
3
)
4
3
.
6
4
5
*
*
*
(
8
.
0
8
9
)
4
3
.
6
3
6
*
*
*
(
8
.
1
0
8
)
4
1
.
3
3
7
*
*
*
(
7
.
8
5
2
)
4
1
.
2
0
2
*
*
*
(
7
.
8
8
1
)
P
s
e
u
d
o
R
2
0
.
1
0
4
0
.
1
0
5
0
.
1
0
8
0
.
1
0
8
0
.
1
0
9
0
.
1
0
9
N
7
5
7
5
7
3
7
3
7
3
7
3
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
.
K
O
P
?
C
h
i
n
n
-
I
t
o
I
n
d
e
x
o
f
c
a
p
i
t
a
l
a
c
c
o
u
n
t
o
p
e
n
n
e
s
s
;
E
Q
L
?
F
o
r
e
i
g
n
e
q
u
i
t
y
l
i
a
b
i
l
i
t
i
e
s
/
G
D
P
.
S
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
354
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
5. Robustness
In order to test the robustness of the results, we re-estimated the base equations. First,
the equations were re-estimated with OLS. Second, the one-year lags on the independent
variables were replaced with three-year average lags. Last, the model was estimated
without those crises with reported output losses of zero.
The results using OLS are displayed in Table VII and largely confrm the results
from Table II using the Tobit estimation. The adjusted R
2
s range from 0.271 to 0.299
when the small loss variable is included. The coeffcient on the private credit variable is
signifcant at the 5 per cent level in three of the equations. The three stock market
variables are signifcant at the 10 per cent level.
In Table VIII, we used the average values of GDP growth, infation, credit/GDP and
the three stock market variables over the three years preceding the crisis to determine
whether their longer-term effects are also signifcant. The GDP growth variable is now
positive and signifcant at the 5 or 10 per cent level in all the equations, while the
insurance variable is signifcant at the 10 per cent level in two estimations.
The coeffcients on the stock market variables are negative and similar in magnitude
and levels of signifcance to the coeffcients on the same variables in Table II. In this
case, the stock market capitalization measure is signifcant at the 5 per cent level and the
liquidity measure at the 10 per cent level. The results using stock market data from
several years before the occurrence of a crisis strongly suggest that the line of causality
runs from the stock market to the output loss. It is unlikely that stock markets could
anticipate the size of output losses three years before a crisis took place.
Lastly, those crises that recorded an output loss of zero were dropped from the
analysis to determine if they biased the fndings. The number of observations fell from
a maximum of 76 to 60. Among the explanatory variables, the coeffcient of the ratio of
private credit to GDP remained positive and signifcant at the 5 per cent level in two
estimations. The coeffcients of the stock market capitalization and the values of the
shares traded scaled by GDP are signifcant at the 10 and 5 per cent levels, respectively,
while the last stock market variable is signifcant at the 10.3 per cent level. These results,
available from the authors, indicate that our base results for the stock market variables
are not distorted by the presence of output losses of zero in our dataset.
6. Conclusions
Our results contribute to the existing literature on banking crises by exploring the
impact of stock markets on the output losses of these crises. Greenspan’s suggestion that
capital markets could act as a “spare tire” for channeling credit in the event of a crisis
implies that these markets could reduce the severity of a crisis. The evidence suggests
that greater stock market size and liquidity lower the output costs of banking crises. The
results provide support for Greenspan’s hypothesis, and are consistent with empirical
results presented by Allen et al. (2012) and Laeven and Valencia (2013a). They also
confrm the fnancial services view that both banks and markets provide important
fnancial services, and indicate that a country with a broad-based fnancial system is
likely to recover more quickly from a crisis.
In terms of policy implications, these results suggest that continued development of
equity markets is worthwhile, not only to effectively allocate capital to its most
productive uses, but also to reduce the impact of fnancial crises. Such development,
however, should be accompanied by adequate supervisory and regulatory oversight.
355
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table VII.
Output costs: stock
market variables (OLS)
V
a
r
i
a
b
l
e
s
(
7
.
1
)
(
7
.
2
)
(
7
.
3
)
(
7
.
4
)
(
7
.
5
)
Y
G
R
?
1
.
8
6
9
*
(
0
.
9
5
8
)
?
0
.
2
9
2
(
0
.
9
1
2
)
?
0
.
2
0
2
(
0
.
9
0
2
)
?
0
.
5
2
3
(
0
.
9
2
4
)
?
0
.
4
7
5
(
0
.
8
8
6
)
I
N
F
0
.
0
0
1
(
0
.
0
0
3
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
1
8
9
*
*
(
0
.
0
7
7
)
0
.
0
8
4
(
0
.
0
7
2
)
0
.
1
9
2
*
*
(
0
.
0
9
6
)
0
.
1
8
8
*
*
(
0
.
0
9
1
)
0
.
1
2
1
(
0
.
0
8
5
)
I
N
S
?
1
1
.
0
7
6
(
8
.
1
6
7
)
?
1
0
.
1
8
6
(
7
.
2
0
7
)
?
9
.
0
8
5
(
7
.
1
4
2
)
?
8
.
7
2
6
(
7
.
2
3
6
)
?
5
.
2
4
6
(
7
.
4
9
3
)
C
U
R
1
2
.
0
9
6
(
9
.
1
0
8
)
3
.
1
6
9
(
8
.
2
6
8
)
6
.
3
2
1
(
8
.
3
6
9
)
3
.
6
3
8
(
8
.
2
7
1
)
3
.
7
3
0
(
7
.
9
9
2
)
S
M
L
S
?
3
9
.
6
6
6
*
*
*
(
8
.
6
6
5
)
?
3
7
.
0
9
0
*
*
*
(
8
.
6
8
5
)
?
3
6
.
2
1
7
*
*
*
(
8
.
9
1
7
)
?
3
8
.
3
2
1
*
*
*
(
8
.
4
5
0
)
C
A
P
?
0
.
1
3
0
*
(
0
.
0
7
7
)
L
I
Q
?
0
.
1
0
7
*
(
0
.
0
5
6
)
T
U
R
?
0
.
1
4
9
*
(
0
.
0
8
7
)
C
o
n
s
t
a
n
t
3
1
.
7
0
1
*
*
*
(
6
.
5
8
6
)
4
2
.
7
2
4
*
*
*
(
6
.
2
8
9
)
3
9
.
9
5
6
*
*
*
(
6
.
4
1
8
)
4
0
.
2
0
9
*
*
*
(
6
.
5
2
9
)
4
3
.
1
9
4
*
*
*
(
6
.
2
0
8
)
A
d
j
u
s
t
e
d
R
2
0
.
0
6
4
0
.
2
7
1
0
.
2
9
1
0
.
2
9
9
0
.
2
9
6
N
7
6
7
6
7
6
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
e
q
u
a
t
i
o
n
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
356
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Our results also imply that countries should seek to implement prudent fnancial
openness measures in combination with measures targeted towards equity market
deepening. Moreover, not all frms cut off frombank credit during a crisis will be able to
raise funds in the stock market. Policy measures to reestablish the role of banks in the
fnancial system after a crisis are also needed.
There are several areas for potential future extension of this work. First, while this
research shows that stock markets can lessen the impact of fnancial crises, it does not
directly identify the mechanism through which this effect occurs. Further work could
seek to ascertain the channels through which stock markets reduce output losses.
Additionally, investigation of the timing of foreign equity fows could determine when
infows are reestablished after the outbreak of a banking crisis.
Notes
1. In addition, the magnitude of a government bailout can be signifcant, and combined with
protracted declines in tax revenues infict a fscal cost. Reinhart and Rogoff (2009) report that
central government debt after a banking crisis increases on average by about 86% in real
terms.
2. Bank crises are one type of fnancial crises; other types include currency crises, “sudden
stops” (or capital account crises) and debt crises. A bank crisis can overlap with one of the
other types. Claessens and Kose (2014) reviewthe literature on the different types of fnancial
crises.
3. See literature cited below in Section 4.2.
4. In addition, there are studies that deal with economic contractions due to crises over medium-
and long-run periods. See Cerra and Saxena (2008), Abiad et al. (2009) and Furceri and
Zdzienicka (2012).
5. The difference in behavior during a crisis refects in part the more limited role that bond
markets have in many emerging markets as compared with equity markets. Herring and
Chatusripitak (2006) attribute the absence of bond markets in these countries to weak
Table VIII.
Output losses: Stock
market variables (3-year
lags)
Variables (8.1) (8.2) (8.3) (8.4)
YGR (3) 2.503* (1.267) 2.983** (1.254) 2.397* (1.248) 2.364* (1.254)
INF (3) 0.002 (0.006) 0.003 (0.006) 0.003 (0.006) 0.002 (0.006)
CRED (3) 0.040 (0.080) 0.189* (0.105) 0.142 (0.097) 0.107 (0.089)
INS ?14.058* (8.292) ?14.153* (8.050) ?12.476 (8.225) ?8.541 (8.932)
CUR ?2.341 (9.240) 1.160 (9.120) ?1.322 (9.128) ?0.435 (9.229)
SMLS ?99.000*** (17.492) ?98.899*** (17.558) ?97.034*** (17.371) ?97.613*** (17.367)
CAP (3) ?0.212** (0.102)
LIQ (3) ?0.158* (0.086)
TUR (3) ?0.169 (0.104)
Constant 40.331*** (7.032) 37.621*** (6.939) 38.537*** (7.185) 39.774*** (7.140)
Pseudo R
2
0.099 0.106 0.105 0.104
N 77 77 76 76
Notes: Standard errors in parentheses; see Table II for description of variables; GDP growth, infation
rate, private credit/GDP, stock market capitalization/GDP, value of shares traded/GDP and value of
shares traded/stock market capitalization are measured as three-year averages
357
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
fnancial infrastructure. Spiegel (2009) claims that less developed bond markets exhibit yield
curve anomalies and other distortions. Burger et al. (2012) provide an update on bond market
activity in emerging economies.
6. Trend output is calculated by applying the Hodrick–Prescott flter to the log of real GDP over
(t-20, t-1) before the crisis, where the data are available for the full time range. At least four
observations are required in the pre-crisis period. Real GDP during the crisis period is
extrapolated from the trend growth rate. Output losses equal the sum of the differences
between actual and trend GDP.
7. Demirgüç-Kunt and Detragiache (2005) provide a review of the literature on the causes of
bank crises.
8. Drehmann et al. (2012) report that the ratio of credit to GDP is a reliable indicator of the cycle
in credit and property prices, and peaks in these cycles are linked to banking crisis.
9. Allen et al. (2012) use the same variables in their analysis of the relationship of banks and
stock markets.
10. We also included an exchange rate regime variable in several specifcations, but the variable
was never signifcant.
References
Abiad, A., Balakrishnan, R., Brooks, P.K., Leigh, D. and Tytell, I. (2009), “What’s the damage?
Medium-term output dynamics after banking crises”, IMF Working Paper No. 09/245,
International Monetary Fund, Washington, DC.
Aizenman, J. and Noy, I. (2013), “Macroeconomic adjustment and the history of crises in open
economies”, Journal of International Money and Finance, Vol. 38 No. 8, pp. 41-58.
Allen, F., Gu, X. and Kowalewski, O. (2012), “Financial crisis, structure and reform”, Journal of
Banking & Finance, Vol. 36 No. 11, pp. 2960-2973.
Angkinand, A.P. (2009), “Banking regulation and the output cost of banking crises”, Journal of
International Financial Markets, Institutions & Money, Vol. 19 No. 2, pp. 240-257.
Barajas, A., Chami, R. and Yousef, S.R. (2013), “The fnance and growth nexus re-examined: Do all
countries beneft equally?”, IMF Working Paper No. 13/130, International Monetary Fund,
Washington, DC.
Beck, T. and Levine, R. (2004), “Stock markets, banks, and growth: panel evidence”, Journal of
Banking & Finance, Vol. 28 No. 3, pp. 423-442.
Beck, T., Demirgüç-Kunt, A. and Levine, R. (2000), “Anewdatabase on fnancial development and
structure”, World Bank Economic Review, Vol. 14 No. 3, pp. 597-605.
Beck, T., Demirgüç-Kunt, A. and Levine, R. (2010), “Financial institutions and markets across
countries and over time: the updated fnancial development and structure database”, World
Bank Economic Review, Vol. 24 No. 1, pp. 77-92.
Bekaert, G., Harvey, C.R. and Lundblad, C. (2001), “Emerging equity markets and economic
development”, Journal of Development Economics, Vol. 66 No. 2, pp. 465-504.
Bekaert, G., Harvey, C.R. and Lundblad, C. (2005), “Does fnancial liberalization spur growth?”,
Journal of Financial Economics, Vol. 77 No. 1, pp. 3-55.
Burger, J.D., Warnock, F.E. and Warnock, V.C. (2012), “Emerging local currency bond markets”,
Financial Analysts Journal, Vol. 68 No. 4, pp. 73-93.
Cecchetti, S.G., Kohler, M. and Upper, C. (2009), “Financial crises and economic activity”, NBER
Working Paper No. 15379, National Bureau of Economic Research, Cambridge, MA.
JFEP
6,4
358
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Cerra, V. and Saxena, S.C. (2008), “Growth dynamics: the myth of economic recovery”, American
Economic Review, Vol. 98 No. 1, pp. 439-457.
Chinn, M. and Ito, H. (2006), “What matters for fnancial development? Capital controls,
institutions, and interactions”, Journal of Development Economics, Vol. 81 No. 1,
pp. 163-192.
Chinn, M. and Ito, H. (2013), “Notes on the chinn-ito fnancial openness index 2011 update”,
available at:http://web.pdx.edu/?ito/Readme_kaopen2011.pdf
Claessens, S. and Kose, A. (2014), “Financial crises: explanations, types, and implications”, in
Claessens, S., Kose, A., Laeven, L. and Valencia, F. (Eds), Financial Crises: Causes,
Consequences, and Policy Responses, International Monetary Fund, Washington, DC,
pp. 3-59.
Demirgüç-Kunt, A. and Detragiache, E. (1998), “The determinants of banking crises in developing
and developed countries”, IMF Staff Papers, Vol. 45 No. 1, pp. 81-109.
Demirgüç-Kunt, A. and Detragiache, E. (2005), “Cross country empirical studies of systemic bank
distress: a survey”, National Institute Economic Review, Vol. 192 No. 1, pp. 68-83.
Demirgüç-Kunt, A., Karacaovali, B. and Laeven, L. (2005), “Deposit insurance around the world: a
comprehensive database”, World Bank Policy Research Working Paper No. 3628, World
Bank, Washington, DC.
Demirgüç-Kunt, A., Feyen, E. and Levine, R. (2013), “The evolving importance of banks and
securities markets”, World Bank Economic Review, Vol. 27 No. 3, pp. 476-490.
Domaç, I. and Martinez Peria, M.S. (2003), “Banking crises and exchange rate regimes: is there a
link?”, Journal of International Economics, Vol. 61 No. 1, pp. 41-72.
Drehmann, M., Borio, C. and Tsataronis, K. (2012), “Characterising the fnancial cycle: don’t lose
sight of the mediumterm!”, BIS Working Paper No. 380, Bank for International Settlements,
Basel.
Furceri, D. and Zdzienicka, A. (2012), “Banking crises and short and mediumtermoutput losses in
emerging and developing countries: the role of structural and policy variables”, World
Development, Vol. 40 No. 12, pp. 2369-2378.
Greenspan, A. (1999), “Lessons from the global crises”, paper presented to World Bank Group
and International Monetary Fund Program Seminar, 27 September, available at:
www.federalreserve.gov/Boarddocs/speeches/1999/199909272.htm
Herring, R.J. and Chatusripitak, N. (2006), “The case of the missing market: the bond market and
why it matters for fnancial development”, in Klein, L. and Shabbir, T. (Eds), Recent
Financial Crises: Analysis, Challenges and Implications, Edward Elgar, Cheltenham,
Northampton, MA, pp. 148-194.
Jeanne, O., Subramanian, A. and Williamson, J. (2012), Who Needs To Open the Capital Account?,
Peterson Institute for International Economics, Washington, DC.
Laeven, L. and Valencia, F. (2013a), “The real effects of fnancial sector intervention during crises”,
Journal of Money, Credit and Banking, Vol. 45 No. 1, pp. 147-177.
Laeven, L. and Valencia, F. (2013b), “Systemic banking crises database”, IMF Economic Review,
Vol. 61 No. 2, pp. 225-270.
Lane, P.R. and Milesi-Ferretti, G.M. (2007), “The external wealth of nations mark II: revised and
extended estimates of foreign assets and liabilities, 1970-2004”, Journal of International
Economics, Vol. 73 No. 2, pp. 223-250.
Levine, R. (2005), “Finance and growth: theory and evidence”, in Aghion, P. and Durlauf, S.N.
(Eds), Handbook of Economic Growth Volume 1A, Elsevier, Amsterdam, pp. 865-934.
359
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Levine, R. and Zervos, S. (1998), “Stock markets, banks, and economic growth”, American
Economic Review, Vol. 88 No. 3, pp. 537-558.
Mishkin, F.S. (2006), The Next Great Globalization: How Disadvantaged Nations Can Harness
Their Financial Systems to Get Rich, Princeton University Press, Princeton, NJ, Oxford.
Reinhart, C.M. and Rogoff, K.S. (2009), This Time is Different: Eight Centuries of Financial Folly,
Princeton University Press, Princeton, NJ, Oxford.
Rioja, F. and Valev, N. (2014), “Stock markets, banks and the sources of economic growth in low
income and high income countries”, Journal of Economics and Finance, Vol. 38 No. 2,
pp. 302-320.
Schindler, M. (2009), “Measuring fnancial integration: a new data set”, IMF Staff Papers, Vol. 56
No. 1, pp. 222-238.
Spiegel, M. (2009), “Developing asian local currency bond markets: why and how?”, ADBInstitute
Working Paper No. 182, Asian Development Bank, Tokyo.
Further reading
Joyce, J.P. and Nabar, M. (2009), “Sudden stops, banking crises and investment collapses in
emerging markets”, Journal of Development Economics, Vol. 90 No. 2, pp. 314-322.
About the authors
Tess DeLean was a student at Wellesley College when this research was done, and is currently a
senior analyst at Analysis Group.
Joseph P. Joyce is a Professor of Economics at Wellesley College. His research deals with issues
in fnancial globalization. His book, The IMF and Global Financial Crises: Phoenix Rising?, was
published in 2012 by Cambridge University Press.
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
JFEP
6,4
360
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Appendix
Table AI.
Data defnitions and
sources
Variable Name Defnition and units Source
CAP Capitalization/
GDP
Value of listed shares in stock
market divided by GDP (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
CRED Credit/GDP Claims on private sector by
deposit money banks and
other fnancial institutions
divided by GDP (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
CUR Currency Crisis Dummy ?1 if currency crisis
occurred before or during
banking crisis
Laeven and Valencia (2013b)
EQL Equity
Liabilities/GDP
Value of foreign equity
liabilities divided by GDP (%)
Lane and Milesi-Ferretti (2007)
INF Infation Change of GDP price defator
(%)
World Development Indicators
INS Deposit
Insurance
Dummy ?1 if explicit
deposit insurance plan exists
Deposit Insurance Around the
World Dataset, Demirgüç-
Kunt et al. (2005)
KAO Capital
openness
Index of capital account
openness), normalized: 0
(closed) ?100 (liberalized)
Chinn and Ito (2006, 2013)
LIQ Shares
Traded/GDP
Value of total shares traded in
stock market divided by GDP
(%)
Database on Financial
Development and Structure
Beck et al. (2000, 2010)
LOW,
MID
Low Income,
Middle Income
Income category based on
World Bank classifcation at
time of crisis
World Bank
LOSS Output Loss of
Banking Crisis
Cumulative sum of
differences between actual
and trend real GDP over
periods t to t ?3, expressed
as percentage of trend GDP
Laeven and Valencia (2013b)
SMLS Small Output
Loss
Dummy ?1 if LOSS ?1 %
TUR Shares Traded/
Capitalization
Value of total shares traded
divided by market
capitalization (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
YGR GDP Growth Change in real GDP (%) World Development Indicators
YPOP GDP Per
Capita
GDP per capita (constant
2000 US$)
World Development Indicators
361
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_883223044.pdf
This paper aims to investigate whether stock markets can reduce the output costs of
banking crises. The work is motivated by Alan Greenspan’s claim that capital markets serve as a
financial “spare tire” in the event of a banking crisis.
Journal of Financial Economic Policy
Stock markets and the costs of banking crises
Tess DeLean J oseph P. J oyce
Article information:
To cite this document:
Tess DeLean J oseph P. J oyce , (2014),"Stock markets and the costs of banking crises", J ournal of Financial
Economic Policy, Vol. 6 Iss 4 pp. 342 - 361
Permanent link to this document:http://dx.doi.org/10.1108/J FEP-01-2014-0003
Downloaded on: 24 January 2016, At: 21:50 (PT)
References: this document contains references to 37 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 237 times since 2014*
Users who downloaded this article also downloaded:
Tao Zeng, (2014),"Derivative financial instruments, tax aggressiveness and firm market value", J ournal of
Financial Economic Policy, Vol. 6 Iss 4 pp. 376-390http://dx.doi.org/10.1108/J FEP-02-2014-0013
Dogga Satyanarayana Murthy, Suresh Kumar Patra, Amaresh Samantaraya, (2014),"Trade Openness,
Financial Development Index and Economic Growth: Evidence from India (1971-2012)", J ournal of Financial
Economic Policy, Vol. 6 Iss 4 pp. 362-375http://dx.doi.org/10.1108/J FEP-10-2013-0056
Abdul Rashid, Zainab J ehan, (2014),"The response of macroeconomic aggregates to monetary policy
shocks in Pakistan", J ournal of Financial Economic Policy, Vol. 6 Iss 4 pp. 314-330http://dx.doi.org/10.1108/
J FEP-04-2013-0016
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Stock markets and the costs of
banking crises
Tess DeLean and Joseph P. Joyce
Department of Economics, Wellesley College, Wellesley,
Massachusetts, USA
Abstract
Purpose – This paper aims to investigate whether stock markets can reduce the output costs of
banking crises. The work is motivated by Alan Greenspan’s claim that capital markets serve as a
fnancial “spare tire” in the event of a banking crisis.
Design/methodology/approach – We test the impact of stock market capitalization, liquidity and
turnover on the output losses of 76 banking crises in 66 countries over the period of 1975-2008.
Findings – Our results indicate that stock markets can mitigate the effect of banking crises on
economic activity. There is also some evidence that foreign equity holdings lower output costs.
Practical implications – These results suggest that the development of equity markets will
contribute to reducing the costs of banking crises. Such development, however, should be accompanied
by adequate supervisory and regulatory oversight.
Originality/value – Our analysis is the frst direct empirical investigation of the impact of stock
markets on the output costs of banking crises. This paper demonstrates that equity markets can lessen
the severity of such crises.
Keywords Financial markets, Equity, Banks, Stock markets, Banking crises, Output losses,
Foreign equity holdings
Paper type Research paper
1. Introduction
The banking crises that occurred during the Global Financial Crisis of 2007-2009 served
as a reminder that such events can impose signifcant costs upon the economies where
they occur. Banking crises interrupt the fow of credit, forcing otherwise viable frms
into bankruptcy and disrupting investment and other expenditures[1]. The crises also
showed that such events occur in advanced economies as well as in emerging markets
and developing economies.
This paper investigates whether stock markets can reduce the impact of bank crises
on economic activity[2]. The topic is motivated by Greenspan’s (1999) claimthat capital
markets can act as a “spare tire” during a bank crisis, providing another channel for
allocating credit to frms and thus mitigating the costs of the crisis. We specifcally test
whether stock markets have an impact on the output costs associated with banking
crises. This issue is relevant for the study of the contributions of banks and capital
markets to fnancial development.
In addition, we examine whether any such effects depend on the income level of the
economy. Several recent papers have claimed that the impact of the stock market on
economic activity varies across nations with different incomes[3]. We also test whether
JEL classifcation – G01, G10, G21
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
JFEP
6,4
342
Journal of Financial Economic Policy
Vol. 6 No. 4, 2014
pp. 342-361
©Emerald Group Publishing Limited
1757-6385
DOI 10.1108/JFEP-01-2014-0003
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the openness of a country’s stock market to foreign investors affects its impact on
banking crises.
Our analysis is the frst direct empirical investigation of the impact of stock markets
on the output costs of bank crises. Our results indicate that stock market activity, as
measured by its depth, liquidity or turnover, does lower the costs of such crises. This
effect does not depend on a country’s income level. The impact of fnancial openness is
more mixed: capital account openness contributes to a fall in output, but foreign
holdings of equities may lower the size of the contraction. Abroad-based and diversifed
fnancial sector, therefore, is likely to recover from a crisis more quickly.
The next section provides a survey of related work previously done on this issue.
Section 3 describes the data and model specifcation, and the following section presents
our results. Section 5 contains the results of tests of robustness, and Section 6 concludes
with some policy implications.
2. Literature review
There is an extensive literature on the impact of banking crises on output (summarized
below), which has found that a number of macroeconomic, fnancial and regulatory
variables contribute to the depth of the contraction. But there are few studies that
examine the linkages of banks and equity markets and their economic impact during
banking crises. Their relationship merits investigation, as a diversifed fnancial system
may lessen the economic impact of a shock to one component of it, such as the banking
sector.
Levine (2005) reviews the theoretical literature on the comparative advantages of
bank-based and market-based fnancial systems, and also the view that they both
promote economic growth. Proponents of bank-based systems claim that banks are
better able to acquire information about frms and monitor their managers. Supporters
of market-based systems maintain that capital markets facilitate risk management and
avoid ineffciencies that can be associated with banks. Critics of this “debate” believe
that banks and markets both provide fnancial services that contribute to growth. The
empirical studies that Levine (2005) reviews show that fnancial intermediaries and
stock markets both enhance growth. Levine and Zervos (1998), for example, found that
stock market liquidity and banking sector development are associated with higher
economic growth. Beck and Levine (2004) reported similar results.
Allen et al. (2012) examined the relationship of the structure of fnancial systems and
crises, drawing upon data from75 banking crises in 65 countries over the period of 1980
to 2008. Their analysis revealed that while bank credit and stock market expansions are
positively linked before a crisis, this relationship is reversed during a crisis and for
several years thereafter. The size and turnover ratio of stock markets rise when bank
credit is still declining. They interpreted their results as evidence that frms can shift
fnancing from banks to equity markets, and that a better-developed fnancial system
allows a quicker recovery from disruption.
Laeven and Valencia (2013a) used frm-level data in a study of the impact of fnancial
frictions during the global fnancial crisis. They found that stock market development,
as measured by stock market capitalization, has a growth-enhancing effect on business
frms. They maintained that this result showed that stock markets can serve as a “spare
wheel” during times of banking distress.
343
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
As stated above, there are also studies that seek to ascertain the determinants of the
output losses that accompany banking crises[4]. This work includes papers by
Demirgüç-Kunt and Detragiache (1998), Domaç and Martinez Peria (2003), Angkinand
(2009), Cecchetti et al. (2009) and Aizenman and Noy (2013). We drew upon these works
in formulating our research strategy.
The primary focus of Demirgüç-Kunt and Detragiache’s (1998) paper was on the
causes of banking crises. Using data from the period of 1980-1994, they included
macroeconomic, fnancial and institutional variables as regressors. They used the same
variables in testing the determinants of the fscal costs of the crises in a sample that
included 19 to 24 countries, depending upon the specifcation. Their results indicate that
low GDP growth, adverse changes in the terms of trade, higher real interest rates and
infation rates, increases in the ratio of M2 to reserves (a sign of vulnerability to a
balance of payments crisis) and the share of private sector credit to GDP, and higher
lagged growth rates of private credit increase the fscal costs. The presence of a deposit
insurance scheme also increases this fscal cost, as does the duration of the crisis, while
an effective legal system has the opposite effect.
Domaç and Martinez Peria (2003) extended this work to examine the determinants of
bank crises, their fscal and output costs and their duration. They utilized data from the
period of 1980-1997 in a sample of 16-39 countries, depending on the specifcation. They
classifedtheir explanatoryvariables into macroeconomic, governmental andexchange rate
measures. They reported that higher infation, real interest rates and credit growth all
increase the output costs of banking crises, while capital infows and fscal surpluses lower
them. Peggedexchange rate regimes whenmeasuredonade jure basis alsoraise these costs.
Angkinand (2009) used data from 47 banking crises in advanced and emerging market
countries during the 1970s to 2003 to investigate the impact of fnancial regulations on their
output costs. She found that increased capital regulations and the existence of deposit
insurance schemes lower output costs, while restrictions on the range of bank activities
increase them. She interpreted the latter result as an indication that banks with a more
diverse range of activities are more robust to shocks. In addition, the coeffcients of the
pre-crisis GDP growth rate and real GDP per capita are negative and signifcant in many of
the output loss regressions, while currency crises have a positive impact.
Cecchetti et al. (2009) examined the length, depth and output costs of 40 banking
crises that have occurred since 1980. Their results showed, as did the previous works,
that declines in output are inversely related to GDP growth rates preceding the crisis.
They also found that the depth of a contraction is larger when a banking crisis is
accompanied by a currency or sovereign debt crisis.
Aizenman and Noy (2013) studied the determinants of the severity of banking crises
in high- and middle-income countries over the period of 1980-2010. They found that GDP
growth in the period prior to the crisis lowers the output loss in the full sample. They
also reported that capital account openness raised the output loss during the global
fnancial crisis.
Our analysis, therefore, builds upon research from several areas. We investigate the
determinants of the size of the economic contractions that accompany banking crises,
using control variables drawn from the previous literature. We add measures of stock
market development that have been utilized in studies of fnancial development and
growth. We leave the impact of bond markets on output costs for future analysis, as their
behavior can be very different from equity markets during crises[5].
JFEP
6,4
344
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
3. Data and methodology
The data on the dates of banking crises and their output losses are the focus of our
analysis. We drew upon the dataset provided by Laeven and Valencia (2013b). Earlier
versions of this database have been used in many papers, including several of the
studies of output costs cited in the previous section.
Laeven and Valencia (2013b) report the dates of the banking crises over the period of
1970-2011 and the resulting output losses, which are calculated as the percentage deviation
of actual real GDPfromits trendvalue[6]. We deletedobservations withmissingdata, which
resulted in a sample of 76 crisis episodes in 66 countries. The crises are listed in Table I. The
size of our sample is larger than those of the previous studies of output losses.
Figure 1 shows the number of crises in our sample by the decade of their start dates,
and indicates that the largest number of crises took place during the decade of the 1990s,
Table I.
Countries and initial crisis
year
Algeria (1990) Latvia (2008)
Argentina (1989, 1995, 2001) Luxembourg (2008)
Austria (2008) Madagascar (1988)
Belgium (2008) Malaysia (1997)
Bolivia (1986, 1994) Mexico (1981, 1994)
Brazil (1994) Nepal (1988)
Burundi (1994) Netherlands (2008)
Cameroon (1987, 1995) Niger (1983)
Cape Verde (1993) Nigeria (1991)
Central African Republic (1995) Norway (1991)
Chad (1992) Panama (1988)
Chile (1981) Philippines (1983, 1997)
Colombia (1982, 1998) Portugal (2008)
Denmark (2008) Russia (2008)
Ecuador (1998) Senegal (1988)
Egypt (1980) Sierra Leone (1990)
El Salvador (1989) Slovak Republic (1998)
Finland (1991) Slovenia (2008)
France (2008) Spain (2008)
Germany (2008) Sri Lanka (1989)
Ghana (1982) Swaziland (1995)
Greece (2008) Sweden (1991, 2008)
Guinea-Bissau (1995) Switzerland (2008)
Haiti (1994) Thailand (1983, 1997)
Hungary (2008) Togo (1993)
India (1993) Tunisia (1991)
Indonesia (1997) Turkey (1982, 2000)
Ireland (2008) Uganda (1994)
Jamaica (1996) United Kingdom (2007)
Japan (1997) United States (2007)
Jordan (1989) Venezuela (1994)
Kenya (1992) Vietnam (1997)
Korea (1997) Zambia (1995)
Source: Laeven and Valencia (2013b)
345
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
a period of fnancial instability in many emerging economies. Almost all of the crises of
the decade of the 2000s occurred during the global fnancial crisis of 2008-2009.
Figure 2 displays the output losses associated with the crises episodes. There were a
large number of crises for which a zero output loss is recorded, and many of these
occurred in low- or middle-income countries in Africa, Asia and Latin America.
Our base regression equation is specifed as:
LOSS
i
? ? ? ?’X
i
? ?
(
Stock Market
)
i
? e
i
(1)
where LOSS is the output loss accompanying a banking crisis in country i, X
i
is a vector
of control variables, Stock Market
i
is an indicator of the stock market and e
i
is the error
term. If stock markets do diminish the impact of a banking crisis, ? should be negative.
We relied on the existing literature in choosing economic, fnancial and regulatory
control variables[7]. The initial specifcation included the growth rate of real GDP
0
5
10
15
20
25
30
35
40
1980s 1990s 2000s
N
u
m
b
e
r
o
f
C
r
i
s
e
s
Source: Laeven and Valencia (2013b)
Figure 1.
Number of crises by
decade
0
2
4
6
8
10
12
14
16
18
N
u
m
b
e
r
o
f
C
r
i
s
i
s
Source: Laeven and Valencia (2013b)
Figure 2.
Output losses of banking
crises
JFEP
6,4
346
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
(YGR), infation as measured by changes in the GDPPrice Defator (INF) and the amount
of private credit provided by banks and other fnancial institutions relative to GDP
(CRED)[8]. We also include a dummy variable to account for the presence for a deposit
insurance scheme (INS). All the variables were lagged by one year before the frst year
of the banking crisis to avoid feedback from the crisis. In addition, there is a dummy
variable to indicate whether a currency crisis occurred before or at the time of the
banking crisis (CUR). The sources for these variables are listed in the Data Appendix.
We use several indicators of stock markets in the analysis. The variables measure
different characteristics of stock markets, as no one proxy can adequately address all the
features of these markets. These data were obtained from the Database on Financial
Development and Structure (Beck et al., 2000, 2010). The variables, which were lagged
by one year before the frst year of the crisis, include:
• stock market capitalization scaled by GDP, a measure of the relative size of a stock
market (CAP);
• the value of the shares traded scaled by GDP, a measure of the liquidity of a stock
market relative to the domestic economy (LIQ); and
• the value of the traded shares scaled by market capitalization, the turnover of the
stock market with respect to its own size (TUR).
The frst variable represents the depth of the stock markets, which has been linked to the
rate of economic development. The value of shares traded relative to GDP is a measure
of liquidity, which contributes to lower costs and more effcient markets. The value of
shares traded scaled by market capitalization is another measure of liquidity as well as
of turnover and transactions costs. These features contribute to the ability of equity
markets to allocate savings to their most productive use[9].
The model is estimated using a Tobit specifcation, as the output losses are truncated
at zero. However, we also estimated the main equations using Ordinary Least Squares
OLS, and those results are reported below as one of the tests of robustness.
4. Results
4.1 Stock market variables
Table II reports our initial results for the control variables and the stock market
variables. Equation 2.1 includes the control variables. Ahigher lagged economic growth
rate lowers the cost of a banking crisis, and the coeffcient is signifcant at the 5 per cent
level. Faster growing economies contract less during a crisis. This result is consistent
with the fndings of the studies reported in Section 2.
While some previous studies of output losses found infation to have a signifcant
impact, the positive coeffcient reported in our results is insignifcant. This result is most
likely driven by crisis episodes with high infation but low output costs. The 1994
banking crisis in Brazil, for example, occurred during a period of hyperinfation (over
2,000 per cent in the year preceding the crisis), but a recorded output loss of 0 per cent.
Arise in the ratio of private credit to output, on the other hand, has a positive impact
that is signifcant at the 1 per cent level. Demirgüç-Kunt and Detragiache (1998),
Angkinand (2009) and Furceri and Zdzienicka (2012) also reported that higher credit/
GDP ratios led to larger output costs. A credit “boom” will be followed by a subsequent
“bust”, as borrowers default on their obligations and credit markets take time to recover.
347
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table II.
Output costs: stock
market variables
V
a
r
i
a
b
l
e
s
(
2
.
1
)
(
2
.
2
)
(
2
.
3
)
(
2
.
4
)
(
2
.
5
)
Y
G
R
?
2
.
9
0
6
*
*
(
1
.
1
4
7
)
?
0
.
3
9
2
(
1
.
0
3
4
)
?
0
.
1
5
7
(
1
.
0
1
7
)
?
0
.
5
2
7
(
1
.
0
3
2
)
?
0
.
5
4
1
(
0
.
9
9
7
)
I
N
F
0
.
0
0
1
(
0
.
0
0
3
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
2
6
5
*
*
*
(
0
.
0
9
3
)
0
.
1
0
2
(
0
.
0
8
3
)
0
.
2
3
3
*
*
(
0
.
1
0
6
)
0
.
2
1
2
*
*
(
0
.
0
9
9
)
0
.
1
3
9
(
0
.
0
9
5
)
I
N
S
?
1
3
.
2
7
4
(
9
.
7
3
8
)
?
1
4
.
2
3
7
(
8
.
6
1
5
)
?
1
3
.
3
4
9
(
8
.
3
9
9
)
?
1
1
.
9
7
3
(
8
.
5
4
4
)
?
8
.
7
3
2
(
8
.
8
0
4
)
C
U
R
1
7
.
5
5
1
(
1
0
.
6
4
0
)
2
.
7
4
7
(
9
.
1
9
2
)
6
.
2
6
7
(
9
.
1
3
6
)
2
.
8
6
8
(
9
.
0
7
2
)
3
.
3
2
6
(
8
.
8
1
9
)
S
M
L
S
?
8
9
.
7
4
3
*
*
*
(
1
6
.
8
9
1
)
?
8
7
.
9
0
1
*
*
*
(
1
6
.
8
4
3
)
?
8
7
.
1
4
9
*
*
*
(
1
6
.
9
5
1
)
?
8
6
.
9
9
7
*
*
*
(
1
6
.
3
8
4
)
C
A
P
?
0
.
1
8
1
*
(
0
.
0
9
5
)
L
I
Q
?
0
.
1
5
0
*
*
(
0
.
0
7
2
)
T
U
R
?
0
.
1
6
6
*
(
0
.
0
9
6
)
C
o
n
s
t
a
n
t
2
5
.
6
9
8
*
*
*
(
7
.
7
8
9
)
4
4
.
0
8
5
*
*
*
(
7
.
0
3
5
)
4
1
.
7
4
4
*
*
*
(
6
.
9
7
0
)
4
2
.
1
7
4
*
*
*
(
7
.
1
4
6
)
4
4
.
9
4
8
*
*
*
(
6
.
9
1
2
)
P
s
e
u
d
o
R
2
0
.
0
2
2
0
.
0
9
4
0
.
0
9
9
0
.
1
0
2
0
.
1
0
2
N
7
6
7
6
7
6
7
4
7
4
N
o
t
e
s
:
*
S
i
g
n
i
f
c
a
n
t
a
t
1
0
%
;
*
*
a
t
5
%
;
*
*
*
a
t
1
%
;
s
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
Y
G
R
?
l
a
g
g
e
d
g
r
o
w
t
h
r
a
t
e
o
f
G
D
P
;
I
N
F
?
l
a
g
g
e
d
i
n
f
a
t
i
o
n
r
a
t
e
;
C
R
E
D
?
l
a
g
g
e
d
p
r
i
v
a
t
e
c
r
e
d
i
t
/
G
D
P
;
I
N
S
?
l
a
g
g
e
d
d
e
p
o
s
i
t
i
n
s
u
r
a
n
c
e
d
u
m
m
y
;
C
U
R
?
c
u
r
r
e
n
c
y
c
r
i
s
i
s
d
u
m
m
y
;
S
M
L
S
?
o
u
t
p
u
t
l
o
s
s
l
e
s
s
t
h
a
n
1
%
d
u
m
m
y
;
C
A
P
?
l
a
g
g
e
d
s
t
o
c
k
m
a
r
k
e
t
c
a
p
i
t
a
l
i
z
a
t
i
o
n
/
G
D
P
;
L
I
Q
?
l
a
g
g
e
d
m
a
r
k
e
t
v
a
l
u
e
o
f
s
h
a
r
e
s
t
r
a
d
e
d
/
G
D
P
;
T
U
R
?
l
a
g
g
e
d
m
a
r
k
e
t
v
a
l
u
e
o
f
s
h
a
r
e
s
t
r
a
d
e
d
/
s
t
o
c
k
m
a
r
k
e
t
c
a
p
i
t
a
l
i
z
a
t
i
o
n
JFEP
6,4
348
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
The coeffcient of the deposit insurance dummy variable is negative but not signifcant,
while the coeffcient of the currency crisis indicator is positive and not signifcant. The
pseudo R
2
at 0.022 is relatively low, which refects several factors. First, pseudo-R
2
s
from Tobit estimations are often lower than the adjusted R
2
s reported in the OLS
estimations of the same equations. Table VII below with the results of OLS estimations
of these equations reports adjusted R
2
s with higher values. Second, the government’s
policies in reaction to the initial occurrence of a bank crisis are most likely a major factor
in determining the fnal cost of a crisis. These responses would be endogenous to the
crisis. Finally, the results may refect the distribution of output losses that we show in
Figure 2. Many of these are zero or slightly higher. When we use a dummy variable to
account for those crises with output losses below1 per cent (SMLS) as we do in equation
2.2, the variable itself is quite signifcant and the pseudo-R
2
rises to 0.094. However, the
coeffcient of the growth rate of output falls in signifcance.
We add the stock market variables in the following three equations. The coeffcient of
the capitalization of the stock market scaled by GDP in equation 2.2 is negative and
signifcant at the 10 per cent level. We repeat the specifcation with the value of shares
traded scaled by GDP in equation 2.3, and it is also negative and signifcant at the 5 per
cent level. Finally, we utilized the measure of turnover in the stock market in equations
2.4, and its coeffcient is negative and signifcant at the 10 per cent level. When the three
equations are estimated without the dummy for crises with small losses, the signifcance
levels of the stock market variables rise to 5, 1 and 10 per cent, while the GDP growth
variable is signifcant at the 5 or 1 per cent levels.
Could these results result from causality from the output loss to the lagged stock
market? It is possible that investors who foresee that a banking crisis will not be severe
could bid up stock market prices in the year preceding such a crisis. But such foresight
seems to be absent frommost banking crises, including the recent global crisis. Reinhart
and Rogoff (2009) fnd that equity prices usually peak before the occurrence of a banking
crisis and then decline as the crisis approaches before recovering. Overall, therefore, our
results are consistent with the hypothesis that equity markets can lower the output loss
of a banking crisis.
Table III presents the marginal effects of the stock market variables evaluated at
the mean values of the independent variables. For the sake of comparison, we repeat
the results for their impact on the latent variable from Table II with those computed
for actual values of the output loss (also called the “unconditional expected value”)
and those computed using only uncensored observations. If we use the latter values,
we see that the results for the three stock market variables are quite similar, ranging
from ?0.083 to ?0.101.
Table III.
Output costs: stock
market variables:
marginal effects at means
Stock market variable
On latent
variable
On actual
variable
Uncensored
observations
Stock market capitalization/GDP ?0.182 ?0.140 ?0.101
Value of shares traded/GDP ?0.150 ?0.115 ?0.083
Value of shares traded/stock market capitalization ?0.160 ?0.128 ?0.092
Note: Estimates derived from Equations 2.3, 2.4 and 2.5 in Table II
349
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
We extended the results reported in Table II by estimating these equations with
additional variables. We used two lagged macro policy variables, a short-term nominal
interest rate and government expenditures scaled by GDP. The results, available from
the authors, show that neither policy variable is signifcant. Other variables that we
tested, including M2/Reserves and capital fows scaled by GDP, yielded similar results.
4.2 Stock markets and income levels
In this section, we examine whether the impact of stock markets on the output loss
of a banking crisis differs across income levels, as well as test for whether the size of
the output loss is directly infuenced by income. Several recent papers have reported
that the economic effects of stock market activity do vary across income levels.
Demirgüç-Kunt et al. (2013) for example, found that the sensitivity of output to bank
development falls as economies grow, while the sensitivity of output to the
development of securities markets rises. Similarly, Rioja and Valev (2014) claim that
stock markets have positive effects on productivity and capital accumulation in
high-income countries but not in low-income economies. Barajas et al. (2013)
reported that the impact of stock market turnover on growth is smaller in
low-income countries.
We frst used real GDP per capita (YPOP) as our income measure, and these results
appear in Table IV. The private credit variable is highly signifcant in all the
estimations, as is the dummy for crises with small losses. The income per capita variable
has a negative coeffcient that is signifcant at different levels in fve of the six
estimations. Countries with higher income levels have lower output costs, although the
size of the impact is small.
The stock market liquidity measure is signifcant at the 10 and 5 per cent levels in the
two equations that include it, while the interaction term here is positive and signifcant
at the 10 per cent level. Neither of the two other interaction terms is signifcant.
We then used the World Bank classifcations of nations as low-income,
middle-income (aggregated fromlower- and upper-middle income countries to formone
category) and high-income countries to include dummies for the frst two categories
(LOW, MID). We interacted these with each of the stock market measures, and the
results with and without the interaction terms appear in Table V. The stock market
capitalization and liquidity measures are signifcant at the 10 per cent level, while the
private credit variable is signifcant at different levels in all the estimations. But neither
the income category variables nor their interaction terms were signifcant, and omitting
the interaction terms does not change the signifcance of the results for the two income
variables. The results, therefore, are similar to those in the previous table. There is
virtually no evidence to indicate that the impact of the stock market on the output losses
of bank crises varies systemically by income level.
4.3 Stock markets in open economies
We next tested whether the impact of the stock market variables is infuenced by the
openness of the stock markets to foreign investors. Such an effect could be either
positive or negative. On the one hand, foreign equity infows may lead to deeper and
more resilient markets. On the other hand, capital fight by foreign and domestic
investors could depress asset prices and reduce liquidity.
JFEP
6,4
350
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table IV.
Output costs: stock
market variables and real
income per capita
V
a
r
i
a
b
l
e
s
(
4
.
1
)
(
4
.
2
)
(
4
.
3
)
(
4
.
4
)
(
4
.
5
)
(
4
.
6
)
Y
G
R
?
0
.
3
6
6
(
1
.
0
1
4
)
?
0
.
3
4
0
(
1
.
0
0
2
)
?
0
.
7
6
6
(
1
.
0
0
3
)
?
0
.
6
3
7
(
0
.
9
7
6
)
?
0
.
7
0
3
(
0
.
9
6
8
)
?
0
.
6
7
8
(
0
.
9
7
0
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
2
8
6
*
*
(
0
.
1
1
2
)
0
.
3
3
4
*
*
*
(
0
.
1
1
9
)
0
.
3
5
6
*
*
*
(
0
.
1
1
8
)
0
.
4
7
4
*
*
*
(
0
.
1
3
1
)
0
.
2
5
9
*
*
(
0
.
1
1
0
)
0
.
2
7
1
*
*
(
0
.
1
1
7
)
I
N
S
?
8
.
4
6
4
(
9
.
0
4
3
)
?
6
.
4
5
9
(
9
.
1
3
7
)
?
4
.
6
7
3
(
8
.
9
4
0
)
?
3
.
6
7
1
(
8
.
7
2
5
)
?
2
.
5
2
9
(
9
.
0
2
7
)
?
1
.
7
4
2
(
9
.
3
3
8
)
C
U
R
1
.
8
1
9
(
9
.
6
1
7
)
2
.
6
9
8
(
9
.
5
5
7
)
?
1
.
8
9
3
(
9
.
0
4
8
)
1
.
4
5
3
(
8
.
9
5
1
)
?
0
.
9
0
4
(
8
.
7
8
1
)
?
0
.
3
0
4
(
8
.
9
6
3
)
S
M
L
S
?
8
7
.
9
7
4
*
*
*
(
1
6
.
3
9
5
)
?
8
8
.
7
0
8
*
*
*
(
1
6
.
2
2
4
)
?
8
6
.
7
9
5
*
*
*
(
1
6
.
1
7
6
)
?
8
6
.
4
0
4
*
*
*
(
1
5
.
6
4
5
)
?
8
6
.
5
0
9
*
*
*
(
1
5
.
7
5
6
)
?
8
6
.
5
9
2
*
*
*
(
1
5
.
7
6
2
)
Y
P
O
P
?
0
.
0
0
1
(
0
.
0
0
1
)
?
0
.
0
0
1
*
(
0
.
0
0
1
)
?
0
.
0
0
1
*
*
(
0
.
0
0
0
)
?
0
.
0
0
2
*
*
*
(
0
.
0
0
1
)
?
0
.
0
0
1
*
(
0
.
0
0
0
)
?
0
.
0
0
1
*
(
0
.
0
0
1
)
C
A
P
?
0
.
1
1
7
(
0
.
1
0
5
)
?
0
.
2
0
0
(
0
.
1
2
7
)
C
A
P
*
Y
P
O
P
0
.
0
0
0
(
0
.
0
0
0
)
L
I
Q
?
0
.
1
3
4
*
(
0
.
0
7
0
)
?
0
.
4
2
7
*
*
(
0
.
1
7
1
)
L
I
Q
*
Y
P
O
P
0
.
0
0
0
*
(
0
.
0
0
0
)
T
U
R
?
0
.
1
4
7
(
0
.
0
9
3
)
?
0
.
1
9
9
(
0
.
1
8
7
)
T
U
R
*
Y
P
O
P
0
.
0
0
0
(
0
.
0
0
0
)
C
o
n
s
t
a
n
t
4
2
.
1
3
7
*
*
*
(
6
.
8
6
5
)
4
2
.
7
0
0
*
*
*
(
6
.
8
1
6
)
4
1
.
5
5
1
*
*
*
(
6
.
8
9
1
)
3
9
.
7
5
2
*
*
*
(
6
.
7
4
4
)
4
4
.
4
6
1
*
*
*
(
6
.
6
8
7
)
4
4
.
4
9
7
*
*
*
(
6
.
6
8
0
)
P
s
e
u
d
o
R
2
0
.
1
0
3
0
.
1
0
4
0
.
1
0
9
0
.
1
1
4
0
.
1
0
8
0
.
1
0
8
N
7
6
7
6
7
4
7
4
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
Y
P
O
P
?
r
e
a
l
G
D
P
/
p
o
p
u
l
a
t
i
o
n
i
n
U
S
d
o
l
l
a
r
s
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
351
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table V.
Output costs: stock
market variables and
income category variables
V
a
r
i
a
b
l
e
s
(
5
.
1
)
(
5
.
2
)
(
5
.
3
)
(
5
.
4
)
(
5
.
5
)
(
5
.
6
)
Y
G
R
?
0
.
5
5
4
(
1
.
1
0
1
)
?
0
.
4
8
6
(
1
.
1
3
4
)
?
0
.
8
6
6
(
1
.
1
1
3
)
?
0
.
6
9
0
(
1
.
1
5
4
)
?
1
.
0
1
8
(
1
.
0
8
8
)
?
1
.
0
6
0
(
1
.
1
1
3
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
?
0
.
0
0
0
(
0
.
0
0
2
)
?
0
.
0
0
1
(
0
.
0
0
2
)
C
R
E
D
0
.
3
1
5
*
*
*
(
0
.
1
1
7
)
0
.
3
1
5
*
*
*
(
0
.
1
1
7
)
0
.
2
8
5
*
*
(
0
.
1
1
2
)
0
.
3
0
0
*
*
(
0
.
1
1
4
)
0
.
1
8
1
*
(
0
.
1
0
6
)
0
.
1
9
5
*
(
0
.
1
0
6
)
I
N
S
?
6
.
8
2
5
(
9
.
3
8
8
)
?
4
.
5
4
3
(
1
0
.
5
8
2
)
?
6
.
3
9
4
(
9
.
5
3
2
)
?
5
.
8
0
3
(
1
0
.
8
2
5
)
?
6
.
0
9
3
(
9
.
4
3
7
)
?
4
.
7
7
0
(
1
0
.
5
1
8
)
C
U
R
1
.
4
4
6
(
9
.
3
6
6
)
0
.
0
0
6
(
9
.
8
6
2
)
?
1
.
3
0
5
(
9
.
3
0
0
)
0
.
5
9
0
(
9
.
7
9
9
)
?
0
.
1
2
4
(
9
.
1
8
2
)
3
.
2
9
1
(
9
.
8
4
3
)
S
M
L
S
?
8
8
.
0
5
6
*
*
*
(
1
6
.
3
4
9
)
?
8
7
.
6
9
6
*
*
*
(
1
6
.
2
9
0
)
?
8
7
.
7
4
2
*
*
*
(
1
6
.
5
6
1
)
?
8
7
.
3
9
9
*
*
*
(
1
6
.
6
3
1
)
?
8
6
.
9
6
4
*
*
*
(
1
6
.
2
2
8
)
?
8
6
.
4
7
1
*
*
*
(
1
6
.
6
2
2
)
L
O
W
1
7
.
7
5
5
(
1
6
.
9
2
4
)
1
8
.
8
5
9
(
1
9
.
6
0
3
)
1
5
.
8
6
7
(
1
7
.
2
6
5
)
2
0
.
4
1
5
(
1
9
.
7
7
8
)
7
.
8
5
3
(
1
8
.
2
0
6
)
1
2
.
5
7
7
(
2
0
.
1
6
1
)
M
I
D
2
2
.
4
5
6
*
(
1
3
.
2
1
2
)
2
0
.
5
4
5
(
1
5
.
1
7
4
)
2
0
.
6
4
4
(
1
3
.
5
3
0
)
2
3
.
9
3
1
(
1
4
.
3
8
7
)
1
7
.
0
0
0
(
1
4
.
1
8
9
)
2
5
.
5
9
7
(
1
6
.
6
4
9
)
C
A
P
?
0
.
1
6
0
*
(
0
.
0
9
3
)
?
0
.
1
9
3
(
0
.
1
2
1
)
C
A
P
*
L
O
W
?
1
.
1
5
4
(
3
.
3
4
8
)
C
A
P
*
M
I
D
0
.
0
6
9
(
0
.
1
6
4
)
L
I
Q
?
0
.
1
2
7
*
(
0
.
0
7
2
)
?
0
.
1
1
2
(
0
.
0
7
6
)
L
I
Q
*
L
O
W
?
3
6
.
4
7
6
(
1
5
9
.
9
3
4
)
L
I
Q
*
M
I
D
?
0
.
1
2
3
(
0
.
1
9
9
)
T
U
R
?
0
.
1
2
4
(
0
.
1
0
3
)
?
0
.
0
9
5
(
0
.
1
0
8
)
T
U
R
*
L
O
W
?
1
.
3
2
1
(
8
.
6
5
1
)
T
U
R
*
M
I
D
?
0
.
3
1
4
(
0
.
3
2
8
)
C
o
n
s
t
a
n
t
2
1
.
1
6
9
(
1
5
.
9
6
3
)
2
1
.
5
3
7
(
1
7
.
7
4
8
)
2
3
.
6
3
5
(
1
6
.
3
5
5
)
1
9
.
3
9
9
(
1
7
.
8
2
4
)
3
2
.
3
8
8
*
(
1
7
.
0
8
6
)
2
6
.
7
8
7
(
1
8
.
4
5
8
)
P
s
e
u
d
o
R
2
0
.
1
0
5
0
.
1
0
5
0
.
1
0
6
0
.
1
0
7
0
.
1
0
5
0
.
1
0
7
N
7
6
7
6
7
4
7
4
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
L
O
W
?
l
o
w
-
i
n
c
o
m
e
c
o
u
n
t
r
y
d
u
m
m
y
;
M
I
D
?
m
i
d
d
l
e
-
i
n
c
o
m
e
c
o
u
n
t
r
y
d
u
m
m
y
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
352
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
There are several channels through which foreign holdings of equities could
infuence an economy. Mishkin (2006) maintains that there are both direct and indirect
benefts to liberalizing restrictions on equity fows. The direct benefts of stock market
liberalizations are the increase in the availability of funds with lower costs, while the
indirect benefts are the institutional reforms that decrease informational asymmetries
and allow a more effcient allocation of savings. Bekaert et al. (2001, 2005) provide
empirical analyses of these effects.
Jeanne et al. (2012) suggest that any crisis-induced equity outfows are likely to be
short-lived. As evidence, they cite the reemergence of equity fows to emerging markets
in the wake of the Asian fnancial crisis of 1997-1998 and also after the recent global
fnancial crisis. The return of foreign investors can contribute to the reestablishment of
fnancing for frms cut off from other domestic sources.
We used two measures of fnancial openness, one de facto and the second de jure. The
frst is the lagged value of portfolio equity liabilities scaled by GDP (EQL), which are
included in the updated dataset of gross foreign assets and liabilities provided by Lane
and Milesi-Ferretti (2007). This variable demonstrates the relative size of foreign equity
holdings. We also interact this variable with the currency crisis variable to see if its
effect changes in the presence of such a crisis.
Our second measure is a broad-based measure of capital account openness, the
lagged Chinn and Ito (2006, 2013) index (KAO). The index is based on data reported
in the International Monetary Funds (IMF’s) Annual Report on Exchange
Arrangements and Restrictions on the existence of multiple exchange rates,
restrictions on the current and capital accounts and requirements to surrender
export proceeds. The Chinn-Ito index is the frst principal component of these four
indicators, and we rescaled the measure to range from 0 (completely closed) to 100
(completely liberalized). There are other more specifc de jure measures of capital
regulations, such as Schindler’s (2009) measures of restrictions on equity fows, but
these have more limited coverage in terms of both countries and dates.
Table VI presents the results of the estimations of these specifcations. The stock
market capitalization and the value of shares traded scaled by GDP continue to be
negative and signifcant at the 10 and 5 per cent levels[10]. The foreign equity
liabilities variable is not signifcant in these equations, nor is its interaction with the
currency crisis variable. However, the variable has a negative coeffcient which is
signifcant the at 10 per cent level in equations 6.5 and 6.6, where the domestic stock
market measure is the turnover proxy, which itself is not signifcant. The
signifcance level of the variable rises to 5 per cent when the small loss dummy
variable is omitted from the estimations. The signifcance of the foreign equity
variable here may refect the impact of foreign holdings on the amount of trading
activity, with foreign investors who are more diversifed better able to resume
trading after a crisis erupts.
The openness of a country’s stock markets, therefore, may have a different impact on
its economy during a crisis than the negative impact that fnancial openness in general
may have. If foreign investors return after the outbreak of the crisis, their purchases may
allow the domestic market to be better able to provide credit to frms. The policy
implication is that emerging markets should consider removing restrictions on foreign
access to domestic to equity markets relatively early on as part of a phased opening of
the capital account.
353
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table VI.
Output costs: stock
market variables and
equity market openness
V
a
r
i
a
b
l
e
s
(
6
.
1
)
(
6
.
2
)
(
6
.
3
)
(
6
.
4
)
(
6
.
5
)
(
6
.
6
)
Y
G
R
?
0
.
1
7
1
(
1
.
0
1
9
)
?
0
.
3
8
7
(
1
.
0
3
9
)
?
0
.
4
5
2
(
1
.
0
3
8
)
?
0
.
4
4
8
(
1
.
0
7
0
)
?
0
.
6
3
9
(
1
.
0
0
1
)
?
0
.
5
9
0
(
1
.
0
3
0
)
I
N
F
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
1
8
9
(
0
.
1
2
3
)
0
.
2
0
3
(
0
.
1
2
3
)
0
.
1
9
1
(
0
.
1
2
0
)
0
.
1
9
1
(
0
.
1
2
1
)
0
.
2
7
2
*
*
(
0
.
1
2
1
)
0
.
2
7
6
*
*
(
0
.
1
2
3
)
I
N
S
?
1
3
.
2
4
3
(
8
.
5
2
8
)
?
1
1
.
2
1
9
(
8
.
7
3
7
)
?
1
1
.
6
4
5
(
8
.
5
7
9
)
?
1
1
.
6
7
8
(
8
.
7
7
6
)
?
8
.
6
7
2
(
8
.
7
9
9
)
?
8
.
9
7
5
(
8
.
9
2
8
)
C
U
R
6
.
8
0
8
(
9
.
1
7
0
)
1
.
6
5
5
(
1
0
.
6
9
6
)
2
.
7
7
7
(
8
.
9
5
7
)
2
.
8
7
9
(
1
0
.
6
9
7
)
2
.
6
2
2
(
8
.
6
4
8
)
3
.
7
4
8
(
1
0
.
3
0
4
)
S
M
L
S
?
8
8
.
4
1
1
*
*
*
(
1
7
.
0
3
4
)
?
8
7
.
5
0
7
*
*
*
(
1
6
.
8
7
2
)
?
8
7
.
5
3
2
*
*
*
(
1
7
.
1
3
1
)
?
8
7
.
5
4
1
*
*
*
(
1
7
.
1
4
2
)
?
8
3
.
9
9
0
*
*
*
(
1
6
.
5
6
2
)
?
8
4
.
0
5
2
*
*
*
(
1
6
.
5
9
0
)
K
A
O
?
0
.
0
2
5
(
0
.
1
4
4
)
0
.
0
1
0
(
0
.
1
4
8
)
?
0
.
0
2
6
(
0
.
1
4
6
)
?
0
.
0
2
6
(
0
.
1
4
7
)
0
.
0
0
0
(
0
.
1
4
1
)
?
0
.
0
0
3
(
0
.
1
4
2
)
C
A
P
?
0
.
1
9
4
(
0
.
1
0
8
)
?
0
.
2
8
4
(
0
.
1
4
7
)
L
I
Q
?
0
.
1
6
2
*
(
0
.
0
7
6
)
?
0
.
1
6
2
*
*
(
0
.
0
7
8
)
T
U
R
?
0
.
0
6
0
(
0
.
1
2
5
)
?
0
.
0
6
2
(
0
.
1
2
5
)
E
Q
L
1
0
.
1
2
2
(
7
.
3
6
9
)
1
0
.
2
1
4
(
7
.
3
2
9
)
1
0
.
3
5
2
(
7
.
4
0
8
)
1
0
.
3
4
2
(
7
.
4
2
9
)
?
7
6
.
4
3
8
*
(
4
1
.
8
6
3
)
?
7
5
.
4
1
9
*
(
4
2
.
1
6
3
)
E
Q
L
*
C
U
R
1
3
2
.
9
4
8
(
1
4
5
.
3
2
5
)
?
1
.
9
6
5
(
1
1
1
.
8
1
6
)
?
2
1
.
1
9
4
(
1
0
5
.
4
4
3
)
C
o
n
s
t
a
n
t
4
4
.
0
2
0
*
*
*
(
7
.
7
6
6
)
4
3
.
8
3
6
*
*
*
(
7
.
7
0
3
)
4
3
.
6
4
5
*
*
*
(
8
.
0
8
9
)
4
3
.
6
3
6
*
*
*
(
8
.
1
0
8
)
4
1
.
3
3
7
*
*
*
(
7
.
8
5
2
)
4
1
.
2
0
2
*
*
*
(
7
.
8
8
1
)
P
s
e
u
d
o
R
2
0
.
1
0
4
0
.
1
0
5
0
.
1
0
8
0
.
1
0
8
0
.
1
0
9
0
.
1
0
9
N
7
5
7
5
7
3
7
3
7
3
7
3
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
.
K
O
P
?
C
h
i
n
n
-
I
t
o
I
n
d
e
x
o
f
c
a
p
i
t
a
l
a
c
c
o
u
n
t
o
p
e
n
n
e
s
s
;
E
Q
L
?
F
o
r
e
i
g
n
e
q
u
i
t
y
l
i
a
b
i
l
i
t
i
e
s
/
G
D
P
.
S
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
354
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
5. Robustness
In order to test the robustness of the results, we re-estimated the base equations. First,
the equations were re-estimated with OLS. Second, the one-year lags on the independent
variables were replaced with three-year average lags. Last, the model was estimated
without those crises with reported output losses of zero.
The results using OLS are displayed in Table VII and largely confrm the results
from Table II using the Tobit estimation. The adjusted R
2
s range from 0.271 to 0.299
when the small loss variable is included. The coeffcient on the private credit variable is
signifcant at the 5 per cent level in three of the equations. The three stock market
variables are signifcant at the 10 per cent level.
In Table VIII, we used the average values of GDP growth, infation, credit/GDP and
the three stock market variables over the three years preceding the crisis to determine
whether their longer-term effects are also signifcant. The GDP growth variable is now
positive and signifcant at the 5 or 10 per cent level in all the equations, while the
insurance variable is signifcant at the 10 per cent level in two estimations.
The coeffcients on the stock market variables are negative and similar in magnitude
and levels of signifcance to the coeffcients on the same variables in Table II. In this
case, the stock market capitalization measure is signifcant at the 5 per cent level and the
liquidity measure at the 10 per cent level. The results using stock market data from
several years before the occurrence of a crisis strongly suggest that the line of causality
runs from the stock market to the output loss. It is unlikely that stock markets could
anticipate the size of output losses three years before a crisis took place.
Lastly, those crises that recorded an output loss of zero were dropped from the
analysis to determine if they biased the fndings. The number of observations fell from
a maximum of 76 to 60. Among the explanatory variables, the coeffcient of the ratio of
private credit to GDP remained positive and signifcant at the 5 per cent level in two
estimations. The coeffcients of the stock market capitalization and the values of the
shares traded scaled by GDP are signifcant at the 10 and 5 per cent levels, respectively,
while the last stock market variable is signifcant at the 10.3 per cent level. These results,
available from the authors, indicate that our base results for the stock market variables
are not distorted by the presence of output losses of zero in our dataset.
6. Conclusions
Our results contribute to the existing literature on banking crises by exploring the
impact of stock markets on the output losses of these crises. Greenspan’s suggestion that
capital markets could act as a “spare tire” for channeling credit in the event of a crisis
implies that these markets could reduce the severity of a crisis. The evidence suggests
that greater stock market size and liquidity lower the output costs of banking crises. The
results provide support for Greenspan’s hypothesis, and are consistent with empirical
results presented by Allen et al. (2012) and Laeven and Valencia (2013a). They also
confrm the fnancial services view that both banks and markets provide important
fnancial services, and indicate that a country with a broad-based fnancial system is
likely to recover more quickly from a crisis.
In terms of policy implications, these results suggest that continued development of
equity markets is worthwhile, not only to effectively allocate capital to its most
productive uses, but also to reduce the impact of fnancial crises. Such development,
however, should be accompanied by adequate supervisory and regulatory oversight.
355
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Table VII.
Output costs: stock
market variables (OLS)
V
a
r
i
a
b
l
e
s
(
7
.
1
)
(
7
.
2
)
(
7
.
3
)
(
7
.
4
)
(
7
.
5
)
Y
G
R
?
1
.
8
6
9
*
(
0
.
9
5
8
)
?
0
.
2
9
2
(
0
.
9
1
2
)
?
0
.
2
0
2
(
0
.
9
0
2
)
?
0
.
5
2
3
(
0
.
9
2
4
)
?
0
.
4
7
5
(
0
.
8
8
6
)
I
N
F
0
.
0
0
1
(
0
.
0
0
3
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
0
.
0
0
0
(
0
.
0
0
2
)
C
R
E
D
0
.
1
8
9
*
*
(
0
.
0
7
7
)
0
.
0
8
4
(
0
.
0
7
2
)
0
.
1
9
2
*
*
(
0
.
0
9
6
)
0
.
1
8
8
*
*
(
0
.
0
9
1
)
0
.
1
2
1
(
0
.
0
8
5
)
I
N
S
?
1
1
.
0
7
6
(
8
.
1
6
7
)
?
1
0
.
1
8
6
(
7
.
2
0
7
)
?
9
.
0
8
5
(
7
.
1
4
2
)
?
8
.
7
2
6
(
7
.
2
3
6
)
?
5
.
2
4
6
(
7
.
4
9
3
)
C
U
R
1
2
.
0
9
6
(
9
.
1
0
8
)
3
.
1
6
9
(
8
.
2
6
8
)
6
.
3
2
1
(
8
.
3
6
9
)
3
.
6
3
8
(
8
.
2
7
1
)
3
.
7
3
0
(
7
.
9
9
2
)
S
M
L
S
?
3
9
.
6
6
6
*
*
*
(
8
.
6
6
5
)
?
3
7
.
0
9
0
*
*
*
(
8
.
6
8
5
)
?
3
6
.
2
1
7
*
*
*
(
8
.
9
1
7
)
?
3
8
.
3
2
1
*
*
*
(
8
.
4
5
0
)
C
A
P
?
0
.
1
3
0
*
(
0
.
0
7
7
)
L
I
Q
?
0
.
1
0
7
*
(
0
.
0
5
6
)
T
U
R
?
0
.
1
4
9
*
(
0
.
0
8
7
)
C
o
n
s
t
a
n
t
3
1
.
7
0
1
*
*
*
(
6
.
5
8
6
)
4
2
.
7
2
4
*
*
*
(
6
.
2
8
9
)
3
9
.
9
5
6
*
*
*
(
6
.
4
1
8
)
4
0
.
2
0
9
*
*
*
(
6
.
5
2
9
)
4
3
.
1
9
4
*
*
*
(
6
.
2
0
8
)
A
d
j
u
s
t
e
d
R
2
0
.
0
6
4
0
.
2
7
1
0
.
2
9
1
0
.
2
9
9
0
.
2
9
6
N
7
6
7
6
7
6
7
4
7
4
N
o
t
e
s
:
S
t
a
n
d
a
r
d
e
r
r
o
r
s
i
n
p
a
r
e
n
t
h
e
s
e
s
;
e
q
u
a
t
i
o
n
s
e
s
t
i
m
a
t
e
d
w
i
t
h
O
L
S
;
s
e
e
T
a
b
l
e
I
I
f
o
r
d
e
s
c
r
i
p
t
i
o
n
o
f
v
a
r
i
a
b
l
e
s
JFEP
6,4
356
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Our results also imply that countries should seek to implement prudent fnancial
openness measures in combination with measures targeted towards equity market
deepening. Moreover, not all frms cut off frombank credit during a crisis will be able to
raise funds in the stock market. Policy measures to reestablish the role of banks in the
fnancial system after a crisis are also needed.
There are several areas for potential future extension of this work. First, while this
research shows that stock markets can lessen the impact of fnancial crises, it does not
directly identify the mechanism through which this effect occurs. Further work could
seek to ascertain the channels through which stock markets reduce output losses.
Additionally, investigation of the timing of foreign equity fows could determine when
infows are reestablished after the outbreak of a banking crisis.
Notes
1. In addition, the magnitude of a government bailout can be signifcant, and combined with
protracted declines in tax revenues infict a fscal cost. Reinhart and Rogoff (2009) report that
central government debt after a banking crisis increases on average by about 86% in real
terms.
2. Bank crises are one type of fnancial crises; other types include currency crises, “sudden
stops” (or capital account crises) and debt crises. A bank crisis can overlap with one of the
other types. Claessens and Kose (2014) reviewthe literature on the different types of fnancial
crises.
3. See literature cited below in Section 4.2.
4. In addition, there are studies that deal with economic contractions due to crises over medium-
and long-run periods. See Cerra and Saxena (2008), Abiad et al. (2009) and Furceri and
Zdzienicka (2012).
5. The difference in behavior during a crisis refects in part the more limited role that bond
markets have in many emerging markets as compared with equity markets. Herring and
Chatusripitak (2006) attribute the absence of bond markets in these countries to weak
Table VIII.
Output losses: Stock
market variables (3-year
lags)
Variables (8.1) (8.2) (8.3) (8.4)
YGR (3) 2.503* (1.267) 2.983** (1.254) 2.397* (1.248) 2.364* (1.254)
INF (3) 0.002 (0.006) 0.003 (0.006) 0.003 (0.006) 0.002 (0.006)
CRED (3) 0.040 (0.080) 0.189* (0.105) 0.142 (0.097) 0.107 (0.089)
INS ?14.058* (8.292) ?14.153* (8.050) ?12.476 (8.225) ?8.541 (8.932)
CUR ?2.341 (9.240) 1.160 (9.120) ?1.322 (9.128) ?0.435 (9.229)
SMLS ?99.000*** (17.492) ?98.899*** (17.558) ?97.034*** (17.371) ?97.613*** (17.367)
CAP (3) ?0.212** (0.102)
LIQ (3) ?0.158* (0.086)
TUR (3) ?0.169 (0.104)
Constant 40.331*** (7.032) 37.621*** (6.939) 38.537*** (7.185) 39.774*** (7.140)
Pseudo R
2
0.099 0.106 0.105 0.104
N 77 77 76 76
Notes: Standard errors in parentheses; see Table II for description of variables; GDP growth, infation
rate, private credit/GDP, stock market capitalization/GDP, value of shares traded/GDP and value of
shares traded/stock market capitalization are measured as three-year averages
357
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
fnancial infrastructure. Spiegel (2009) claims that less developed bond markets exhibit yield
curve anomalies and other distortions. Burger et al. (2012) provide an update on bond market
activity in emerging economies.
6. Trend output is calculated by applying the Hodrick–Prescott flter to the log of real GDP over
(t-20, t-1) before the crisis, where the data are available for the full time range. At least four
observations are required in the pre-crisis period. Real GDP during the crisis period is
extrapolated from the trend growth rate. Output losses equal the sum of the differences
between actual and trend GDP.
7. Demirgüç-Kunt and Detragiache (2005) provide a review of the literature on the causes of
bank crises.
8. Drehmann et al. (2012) report that the ratio of credit to GDP is a reliable indicator of the cycle
in credit and property prices, and peaks in these cycles are linked to banking crisis.
9. Allen et al. (2012) use the same variables in their analysis of the relationship of banks and
stock markets.
10. We also included an exchange rate regime variable in several specifcations, but the variable
was never signifcant.
References
Abiad, A., Balakrishnan, R., Brooks, P.K., Leigh, D. and Tytell, I. (2009), “What’s the damage?
Medium-term output dynamics after banking crises”, IMF Working Paper No. 09/245,
International Monetary Fund, Washington, DC.
Aizenman, J. and Noy, I. (2013), “Macroeconomic adjustment and the history of crises in open
economies”, Journal of International Money and Finance, Vol. 38 No. 8, pp. 41-58.
Allen, F., Gu, X. and Kowalewski, O. (2012), “Financial crisis, structure and reform”, Journal of
Banking & Finance, Vol. 36 No. 11, pp. 2960-2973.
Angkinand, A.P. (2009), “Banking regulation and the output cost of banking crises”, Journal of
International Financial Markets, Institutions & Money, Vol. 19 No. 2, pp. 240-257.
Barajas, A., Chami, R. and Yousef, S.R. (2013), “The fnance and growth nexus re-examined: Do all
countries beneft equally?”, IMF Working Paper No. 13/130, International Monetary Fund,
Washington, DC.
Beck, T. and Levine, R. (2004), “Stock markets, banks, and growth: panel evidence”, Journal of
Banking & Finance, Vol. 28 No. 3, pp. 423-442.
Beck, T., Demirgüç-Kunt, A. and Levine, R. (2000), “Anewdatabase on fnancial development and
structure”, World Bank Economic Review, Vol. 14 No. 3, pp. 597-605.
Beck, T., Demirgüç-Kunt, A. and Levine, R. (2010), “Financial institutions and markets across
countries and over time: the updated fnancial development and structure database”, World
Bank Economic Review, Vol. 24 No. 1, pp. 77-92.
Bekaert, G., Harvey, C.R. and Lundblad, C. (2001), “Emerging equity markets and economic
development”, Journal of Development Economics, Vol. 66 No. 2, pp. 465-504.
Bekaert, G., Harvey, C.R. and Lundblad, C. (2005), “Does fnancial liberalization spur growth?”,
Journal of Financial Economics, Vol. 77 No. 1, pp. 3-55.
Burger, J.D., Warnock, F.E. and Warnock, V.C. (2012), “Emerging local currency bond markets”,
Financial Analysts Journal, Vol. 68 No. 4, pp. 73-93.
Cecchetti, S.G., Kohler, M. and Upper, C. (2009), “Financial crises and economic activity”, NBER
Working Paper No. 15379, National Bureau of Economic Research, Cambridge, MA.
JFEP
6,4
358
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Cerra, V. and Saxena, S.C. (2008), “Growth dynamics: the myth of economic recovery”, American
Economic Review, Vol. 98 No. 1, pp. 439-457.
Chinn, M. and Ito, H. (2006), “What matters for fnancial development? Capital controls,
institutions, and interactions”, Journal of Development Economics, Vol. 81 No. 1,
pp. 163-192.
Chinn, M. and Ito, H. (2013), “Notes on the chinn-ito fnancial openness index 2011 update”,
available at:http://web.pdx.edu/?ito/Readme_kaopen2011.pdf
Claessens, S. and Kose, A. (2014), “Financial crises: explanations, types, and implications”, in
Claessens, S., Kose, A., Laeven, L. and Valencia, F. (Eds), Financial Crises: Causes,
Consequences, and Policy Responses, International Monetary Fund, Washington, DC,
pp. 3-59.
Demirgüç-Kunt, A. and Detragiache, E. (1998), “The determinants of banking crises in developing
and developed countries”, IMF Staff Papers, Vol. 45 No. 1, pp. 81-109.
Demirgüç-Kunt, A. and Detragiache, E. (2005), “Cross country empirical studies of systemic bank
distress: a survey”, National Institute Economic Review, Vol. 192 No. 1, pp. 68-83.
Demirgüç-Kunt, A., Karacaovali, B. and Laeven, L. (2005), “Deposit insurance around the world: a
comprehensive database”, World Bank Policy Research Working Paper No. 3628, World
Bank, Washington, DC.
Demirgüç-Kunt, A., Feyen, E. and Levine, R. (2013), “The evolving importance of banks and
securities markets”, World Bank Economic Review, Vol. 27 No. 3, pp. 476-490.
Domaç, I. and Martinez Peria, M.S. (2003), “Banking crises and exchange rate regimes: is there a
link?”, Journal of International Economics, Vol. 61 No. 1, pp. 41-72.
Drehmann, M., Borio, C. and Tsataronis, K. (2012), “Characterising the fnancial cycle: don’t lose
sight of the mediumterm!”, BIS Working Paper No. 380, Bank for International Settlements,
Basel.
Furceri, D. and Zdzienicka, A. (2012), “Banking crises and short and mediumtermoutput losses in
emerging and developing countries: the role of structural and policy variables”, World
Development, Vol. 40 No. 12, pp. 2369-2378.
Greenspan, A. (1999), “Lessons from the global crises”, paper presented to World Bank Group
and International Monetary Fund Program Seminar, 27 September, available at:
www.federalreserve.gov/Boarddocs/speeches/1999/199909272.htm
Herring, R.J. and Chatusripitak, N. (2006), “The case of the missing market: the bond market and
why it matters for fnancial development”, in Klein, L. and Shabbir, T. (Eds), Recent
Financial Crises: Analysis, Challenges and Implications, Edward Elgar, Cheltenham,
Northampton, MA, pp. 148-194.
Jeanne, O., Subramanian, A. and Williamson, J. (2012), Who Needs To Open the Capital Account?,
Peterson Institute for International Economics, Washington, DC.
Laeven, L. and Valencia, F. (2013a), “The real effects of fnancial sector intervention during crises”,
Journal of Money, Credit and Banking, Vol. 45 No. 1, pp. 147-177.
Laeven, L. and Valencia, F. (2013b), “Systemic banking crises database”, IMF Economic Review,
Vol. 61 No. 2, pp. 225-270.
Lane, P.R. and Milesi-Ferretti, G.M. (2007), “The external wealth of nations mark II: revised and
extended estimates of foreign assets and liabilities, 1970-2004”, Journal of International
Economics, Vol. 73 No. 2, pp. 223-250.
Levine, R. (2005), “Finance and growth: theory and evidence”, in Aghion, P. and Durlauf, S.N.
(Eds), Handbook of Economic Growth Volume 1A, Elsevier, Amsterdam, pp. 865-934.
359
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Levine, R. and Zervos, S. (1998), “Stock markets, banks, and economic growth”, American
Economic Review, Vol. 88 No. 3, pp. 537-558.
Mishkin, F.S. (2006), The Next Great Globalization: How Disadvantaged Nations Can Harness
Their Financial Systems to Get Rich, Princeton University Press, Princeton, NJ, Oxford.
Reinhart, C.M. and Rogoff, K.S. (2009), This Time is Different: Eight Centuries of Financial Folly,
Princeton University Press, Princeton, NJ, Oxford.
Rioja, F. and Valev, N. (2014), “Stock markets, banks and the sources of economic growth in low
income and high income countries”, Journal of Economics and Finance, Vol. 38 No. 2,
pp. 302-320.
Schindler, M. (2009), “Measuring fnancial integration: a new data set”, IMF Staff Papers, Vol. 56
No. 1, pp. 222-238.
Spiegel, M. (2009), “Developing asian local currency bond markets: why and how?”, ADBInstitute
Working Paper No. 182, Asian Development Bank, Tokyo.
Further reading
Joyce, J.P. and Nabar, M. (2009), “Sudden stops, banking crises and investment collapses in
emerging markets”, Journal of Development Economics, Vol. 90 No. 2, pp. 314-322.
About the authors
Tess DeLean was a student at Wellesley College when this research was done, and is currently a
senior analyst at Analysis Group.
Joseph P. Joyce is a Professor of Economics at Wellesley College. His research deals with issues
in fnancial globalization. His book, The IMF and Global Financial Crises: Phoenix Rising?, was
published in 2012 by Cambridge University Press.
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
JFEP
6,4
360
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Appendix
Table AI.
Data defnitions and
sources
Variable Name Defnition and units Source
CAP Capitalization/
GDP
Value of listed shares in stock
market divided by GDP (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
CRED Credit/GDP Claims on private sector by
deposit money banks and
other fnancial institutions
divided by GDP (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
CUR Currency Crisis Dummy ?1 if currency crisis
occurred before or during
banking crisis
Laeven and Valencia (2013b)
EQL Equity
Liabilities/GDP
Value of foreign equity
liabilities divided by GDP (%)
Lane and Milesi-Ferretti (2007)
INF Infation Change of GDP price defator
(%)
World Development Indicators
INS Deposit
Insurance
Dummy ?1 if explicit
deposit insurance plan exists
Deposit Insurance Around the
World Dataset, Demirgüç-
Kunt et al. (2005)
KAO Capital
openness
Index of capital account
openness), normalized: 0
(closed) ?100 (liberalized)
Chinn and Ito (2006, 2013)
LIQ Shares
Traded/GDP
Value of total shares traded in
stock market divided by GDP
(%)
Database on Financial
Development and Structure
Beck et al. (2000, 2010)
LOW,
MID
Low Income,
Middle Income
Income category based on
World Bank classifcation at
time of crisis
World Bank
LOSS Output Loss of
Banking Crisis
Cumulative sum of
differences between actual
and trend real GDP over
periods t to t ?3, expressed
as percentage of trend GDP
Laeven and Valencia (2013b)
SMLS Small Output
Loss
Dummy ?1 if LOSS ?1 %
TUR Shares Traded/
Capitalization
Value of total shares traded
divided by market
capitalization (%)
Database on Financial
Development and Structure,
Beck et al. (2000, 2010)
YGR GDP Growth Change in real GDP (%) World Development Indicators
YPOP GDP Per
Capita
GDP per capita (constant
2000 US$)
World Development Indicators
361
Stock markets
and the costs of
banking crises
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
:
5
0
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_883223044.pdf