Trade Volatility in the Association of Southeast
Asian Nations Plus Three: Impacts
and Determinants
Thi Nguyet Anh Nguyen, Thi Hong Hanh Pham, Und
Thomas Vallée∗

This paper investigates trade volatility in the Association of Southeast Asian
Nations Plus Three (ASEAN+3) and its links with output volatility, export
diversification, and free trade agreements. To achieve this research objective, Wir
apply several econometric estimators to data from all ASEAN+3 member states
over the period 1990–2016. We first find evidence of a positive relationship
between output volatility and trade volatility. Zweite, we reveal that the way
export diversification is measured can influence its impacts on bilateral export
volatility. Darüber hinaus, the relationship between income volatility, trade volatility,
and export diversification seems to depend on country size and the level of
economic development.

Schlüsselwörter: ASEAN+3, export diversification, FTA, output volatility, trade
volatility
JEL-Codes: F02, F14, F15, F21, F40

ICH. Einführung

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

International trade is considered one of the most volatile components of gross
domestic product (BIP). According to Bennett et al. (2016), trade volatility and
GDP volatility have tended to move together over the past 20 years—from slightly
falling in the mid-1990s until 2008 to sharply increasing after the global crisis of
2008–2009. Obviously, trade has become a transmission mechanism of external
shocks throughout the world economy. Zusätzlich, researchers have found a positive
correlation between trade openness and volatility (Di Giovanni and Levchenko
2009), which suggests that trade openness has also played a role in explaining
the volatility of economic growth (Rodrik 1997). Caselli et al. (2015) find that

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

∗Thi Nguyet Anh Nguyen (Korrespondierender Autor): Faculty of Business Management, National Economics University,
Hanoi, Vietnam. Email: anhnn27@gmail.com; Thi Hong Hanh Pham: LEMNA, Institute of Economics and
Management, University of Nantes, Frankreich. Email: thi-hong-hanh.pham@univ-nantes; Thomas Vallée: LEMNA,
Institute of Economics and Management, University of Nantes, Frankreich. Email: thomas.vallee@univ-nantes.fr. Wir
would like to thank the managing editor and the anonymous referee for helpful comments and suggestions. Der
Asian Development Bank recognizes “China” as the People’s Republic of China and “Korea” as the Republic of
Korea. The usual ADB disclaimer applies.

Asiatischer Entwicklungsbericht, Bd. 37, NEIN. 2, S. 167–200
https://doi.org/10.1162/adev_a_00153

© 2020 Asian Development Bank and
Asian Development Bank Institute.
Veröffentlicht unter Creative Commons
Namensnennung 3.0 International (CC BY 3.0) Lizenz.

168 Asiatischer Entwicklungsbericht

countries that suffer from big country-specific shocks experience a reduction in
volatility as a result of opening up to international trade, because trade is a source
of diversification.

Trade volatility also varies among countries at different levels of economic
Entwicklung. Developing countries are substantially more volatile than developed
ones (Hakura 2009) because developing countries usually concentrate their exports
in a small number of sectors that are particularly sensitive to external shocks.
daher, diversification in trade should diminish volatility effects. Darüber hinaus,
developing countries are characterized by underdeveloped financial markets and
weak monetary and fiscal policies. Daher, diversification of the sectoral composition
can enhance the development of financial markets and alleviate uncertainty in the
economy.

ASEAN+3 is known as a successful model of economic cooperation
between the Association of Southeast Asian Nations (ASEAN) and three East Asian
countries—Japan, Volksrepublik China (VR China), and the Republic of Korea.
Although trade integration and growth prospects from ASEAN enlargement over
the last 2 decades have been investigated in the literature, trade volatility in the
ASEAN+3 has attracted little attention. This paper contributes to empirical trade
studies by analyzing the volatility of trade flows and its links with output volatility,
international institutions, and trade diversification in ASEAN+3. Mit anderen Worten,
our study aims to address the following three questions:

(ich) Does output volatility move together with trade volatility in ASEAN+3?

(ii) What are the links between output volatility and export diversification in

ASEAN+3?

(iii) How do export diversification and free trade agreements (FTAs) impact

ASEAN+3’s trade volatility?

To answer these questions, we conduct multivariate statistical tests using
data from 13 ASEAN+3 countries from 1990 Zu 2016. Erste, we investigate the
impacts of trade variables, notably trade volatility and export diversification, An
ASEAN+3’s output volatility by applying the generalized method of moments
(GMM) estimator. Zweite, using a cross-sectional dataset covering ASEAN+3
bilateral exports, we explore the potential determinants of ASEAN+3’s bilateral
export volatility by applying fixed effect (FE) and instrumental variable (IV)
Techniken.

Der Rest der Arbeit ist wie folgt gegliedert. Section II summarizes the
literature on the link between trade volatility and economic growth. Section III
characterizes the trend in trade growth volatility, openness, and diversification over

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 169

die Vergangenheit 20 years in ASEAN+3. Section IV explains the impact of trade volatility
on output volatility. Section V explores the determinants of trade volatility. Der
concluding remarks are in section VI.

II. Literature Framework

The theoretical and empirical links between international trade and economic
growth have largely been investigated. Helpman and Krugman (1985) and Bhagwati
(1988) find that export growth promotes economic growth and stimulates both
the supply and demand sides of the economy. Easterly (2007) also argues that
exports boost economic efficiency with better allocation of resources and raise
economic growth in the long run. Our study relates to the literature that predicts the
relationship between trade volatility, diversification, and output volatility. In an early
Papier, Rodrik (1997) considers whether trade openness increases macroeconomic
volatility but finds that the effects of openness are still ambiguous. Bejan (2006)
analyzes the relationship between trade openness and output volatility and finds
that higher trade openness is associated with higher output volatility. Over the last
decades, several empirical studies have investigated this trade openness–volatility
relationship (Easterly, Islam, and Stiglitz 2001; Calderón, Loayza, and Schmidt-
Hebbel 2005; Cavallo 2008; Jansen et al. 2009; Balavac and Pugh 2016), Aber
openness remains the most controversial determinant of volatility. Aus diesem Grund,
we investigate the relationship between trade volatility and economic growth
volatility in this study.

The existing literature has also examined the role of export diversification
in controlling risks arising from trade volatility. Einerseits, several empirical
studies find a negative but not always a significant effect of export diversification
on volatility (z.B., Cavallo 2008 and Cavalcanti, Mohaddes, and Raissi 2012).
Calderón and Schmidt-Hebbel (2008) discover a negative relationship between
openness and volatility only when exports are diversified. Ähnlich, Haddad et al.
(2013) find strong evidence that export diversification plays an important role in
reducing the vulnerability of countries to global shocks. They also find that product
diversification clearly moderates the effect of trade openness on growth volatility,
while market concentration measures yield much more mixed results. In der Zwischenzeit,
Bejan (2006) shows that the interaction term between openness and export product
concentration is significant only in advanced economies. Andererseits,
several scholars have found evidence of a positive effect of diversification on
Wachstum. According to Melitz (2003), an increase in export diversification can boost
productivity given that exporters are more productive than nonexporters. Darüber hinaus,
Melitz (2003) suggests that export diversification can reduce exposure to external
shocks, thus reducing macroeconomic volatility and increasing economic growth.

Together with the link between trade volatility and export diversification,
we also analyze the impact of international trade institutions on trade volatility,

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

170 Asiatischer Entwicklungsbericht

particularly the impact of FTAs. Initial work by Keohane (1983) points out that
various international regimes are explicitly designed to promote “orderly patterns
of behavior among members,” implying that they would enhance stability and
reduce volatility in interactions among members. Martin and Simmons (1998)
state that international institutions “lock in a particular equilibrium, providing
stability.” According to Abbott (2000), trade agreements are adopted to reduce
the risk for private traders and investors. Bagwell and Staiger (2002) maintain that
the purpose of trade agreements is to allow countries with market power to reach
more efficient market access exchanges. Most recently, Mansfield and Reinhardt
(2008) conducted the first large-scale, multivariate statistical tests and strongly
confirmed that preferential trade agreements and the World Trade Organization
regime significantly reduced export volatility and also increased export levels.

In sum, based on the theoretical background discussed above, we first
investigate the impacts of trade volatility and export diversification on ASEAN+3’s
economic growth volatility. Zweite, we focus on trade volatility in ASEAN+3 by
providing an empirical analysis on its determinants.

III. Trade Volatility, Openness, and Diversification in ASEAN+3

Over the last 2 decades, ASEAN+3 has witnessed fluctuating trade growth.
Figur 1 plots the trend of ASEAN+3 trade growth, which is measured by the
average growth rate of intra-ASEAN+3 exports and imports. As shown in Figure
1, ASEAN+3 experienced a significant decrease in intraregional trade during the
1997 Asian financial crisis and the 2008 global financial crisis. Trade flows in the
region increased by around 20% after recovering from the 1997 Asian crisis but
dropped again in 2002, before climbing by more than 30% In 2005. After a difficult
Zeitraum, aufgrund der 2008 global financial crisis, intra-ASEAN+3 trade attained its
highest growth rate in 2010, but has experienced a decreasing trend since 2011.

We now turn our attention to the volatility of ASEAN+3 trade integration
over the period 1990–2016. Figur 2 plots the volatility of total trade growth
and of intra-ASEAN+3 and extra-ASEAN+3 trade growth for each member
country. As seen in Figure 2, ASEAN+3 member states can be divided into two
groups according to trade volatility levels. The first group includes the PRC, Die
Demokratische Volksrepublik Laos (Lao PDR), Kambodscha, Indonesien, and Viet
Nam, which are characterized by a higher level of trade volatility. Darüber hinaus, In
these countries, intra-ASEAN+3 and extra-ASEAN+3 trade volatilities do not
move together. Zum Beispiel, the volatility of trade flows between Viet Nam and
extra-ASEAN+3 countries has an increasing trend, while a decreasing trend in
volatility is observed for Viet Nam’s intra-ASEAN+3 trade. The second group
includes Brunei Darussalam, Japan, the Republic of Korea, the Lao PDR, Myanmar,

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 171

Figur 1. Average Trade Growth Intra-ASEAN+3

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

ASEAN+3 = Association of Southeast Asian Nations plus Japan, Volksrepublik China, and the Republic
von Korea.
Quelle: Authors’ illustration from UN COMTRADE data.

Malaysia, die Phillipinen, Singapur, and Thailand. These countries have a fairly
stable evolution of trade volatility at different levels. Among all countries, Malaysia
and Japan have the lowest trade volatility.

Together with a significant change in trade growth, ASEAN+3 has also
experienced an important evolution in trade openness, which is measured by the
ratio of total exports and imports to GDP. As reported in Table 1, the average of
this ratio over the period 1990–2016 ranges from 7% for Japan to over 100% für
Singapore in terms of intra-ASEAN+3 trade. In terms of total world trade, this ratio
is much higher for all countries, in particular for Viet Nam, Malaysia, and Singapore
(mostly over 100%).

Tisch 1 also shows that the intra-ASEAN+3 openness indicator increased
in most countries in the 2000s, reflecting the growing importance of intraregional
trade. This ratio continued to rise in the 2010s in many countries such as Cambodia,
Japan, the Lao PDR, the Republic of Korea, Thailand, and Viet Nam, but slightly
declined in other countries. Japan and the Republic of Korea,
insbesondere,
experienced a significant increase in this ratio over the study period, indicating
greater trade integration with the region. We also observe a similar trend in
ASEAN+3 trade integration in terms of total world trade.

172 Asiatischer Entwicklungsbericht

Figur 2. Volatility of Trade Growth by Country

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

ASEAN+3 = Association of Southeast Asian Nations plus Japan, Volksrepublik China, and the Republic
von Korea; BRU = Brunei Darussalam; CAM = Cambodia; INO = Indonesia; JPN = Japan; KOR = Republic of
Korea; LAO = Lao People’s Democratic Republic; MAL = Malaysia; MYA = Myanmar; PHI = Philippines; PRC =
Volksrepublik China; SIN = Singapore; THA = Thailand; VIE = Viet Nam.
Quelle: Authors’ illustration.

Trade Volatility in ASEAN+3: Impacts and Determinants 173

Tisch 1. ASEAN+3 Openness Indicator

Intra-ASEAN+3 Trade

Total World Trade

Full
Period

1990–
1999

2000–
2009

2010–
2016

Full
Period

1990–
1999

2000–
2009

2010–
2016

0.67
0.34
0.24
0.36
0.79
0.27
0.29
1.30
0.46
0.57
0.14
0.07
0.24

0.73
0.31
0.23
0.35
0.76
0.21
0.23
1.17
0.32
0.34
0.13
0.04
0.17

0.64
0.28
0.25
0.35
0.86
0.33
0.37
1.48
0.52
0.61
0.17
0.08
0.25

0.61
0.50
0.22
0.38
0.73
0.29
0.26
1.22
0.57
0.85
0.12
0.12
0.34

0.87
0.73
0.45
0.47
1.56
0.39
0.66
2.78
0.97
1.08
0.43
0.21
0.62

0.94
0.44
0.48
0.49
1.56
0.30
0.63
2.69
0.73
0.66
0.36
0.15
0.49

0.84
0.84
0.47
0.47
1.71
0.50
0.82
3.05
1.09
1.15
0.52
0.23
0.63

0.81
1.00
0.38
0.45
1.33
0.35
0.48
2.53
1.12
1.56
0.42
0.28
0.79

Country

Brunei Darussalam
Kambodscha
Indonesien
Lao PDR
Malaysia
Myanmar
Philippinen
Singapur
Thailand
Vietnam
Volksrepublik China
Japan
Republik Korea

ASEAN+3 = Association of Southeast Asian Nations plus Japan, Volksrepublik China, and the Republic
von Korea; Lao PDR = Lao People’s Democratic Republic.
Notiz: The openness indicator refers to the ratio of total exports and imports to gross domestic product.
Quelle: Authors’ calculation.

Regarding export diversification, Figur 3 shows how export diversification
has evolved in the last 2 decades in the ASEAN+3 market.1 Figure 3 also illustrates
the difference in export concentration between ASEAN and the Plus Three countries
(Japan, die VR China, and the Republic of Korea). As displayed in Figure 3, the evolution
of both indicators, the Herfindahl–Hirschman Index (HHI) and the Theil index, Ist
similar in each country group.2 ASEAN has a higher concentration in exports with
an average HHI of up to 0.3 and an average Theil index of up to 3.3, verglichen
mit 0.1 Und 1.9, jeweils, for the Plus Three countries. ASEAN experienced
two significant upward trends in export concentration after 1997 Und 2009. Erste,
zwischen 1997 Und 2000, the HHI climbed from 0.17 Zu 0.25 and the Theil index
increased from 2.1 Zu 3. This trend could be explained by an important change in
trade policy of ASEAN countries due to the financial crisis. Jedoch, seit 2010,
these indicators have dramatically decreased, corresponding with the trend toward
export diversification in ASEAN. For the Plus Three countries, both the HHI and
the Theil index varied significantly over the study period. Compared to ASEAN
Länder, the Plus Three countries experienced a significant increase in both the
HHI and the Theil index. During 1990–2016, the HHI ranged from 0.08 Zu 0.11,
and the Theil index ranged from 1.62 Zu 1.83.

Figur 4 illustrates the evolution of export concentration in each ASEAN+3
country. A high value of HHI indicates a high level of export concentration.

1In line with our research objective, Figur 3 only displays trends in export concentration. For import

concentration, see Figure A1.1 in Appendix 1.

2See section IV.A for details on computing the HHI and the Theil index.

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

174 Asiatischer Entwicklungsbericht

Figur 3. Export Concentration: ASEAN versus Plus Three

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

ASEAN = Association of Southeast Asian Nations, HHI = Herfindahl–Hirschman Index.
Notiz: Plus Three refers to Japan, Volksrepublik China, and the Republic of Korea.
Quelle: Authors’ illustration.

Entsprechend, Brunei Darussalam, which mainly specializes in oil and natural
resources, shows the highest concentration in exports with a value of 0.6. Low-
and middle-income economies such as Cambodia, the Lao PDR, Myanmar, Und
Viet Nam maintained a high level of export concentration until 2010. Jedoch,
seit 2010, the HHIs in these countries have fallen, suggesting the start of a
trade diversification policy. Compared to other countries, the PRC and Japan have
maintained a low level of HHI throughout the period 1990–2016, clearly affirming
their trade diversification policy.

Zusamenfassend,

these stylized facts on trade volatility, openness, and export
diversification provide more details about trade patterns in ASEAN+3, which partly
support the research objectives of this paper.

IV. Impacts of Trade Volatility

This section addresses the question of whether trade volatility contributes
to economic growth volatility in ASEAN+3. Darüber hinaus, we also investigate the
potential role of export diversification in explaining output volatility.

Trade Volatility in ASEAN+3: Impacts and Determinants 175

Figur 4. Export Diversification Trends by Country

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

BRU = Brunei Darussalam, CAM = Cambodia, HHI = Herfindahl–Hirschman Index, INO = Indonesia, JPN =
Japan, KOR = Republic of Korea, LAO = Lao People’s Democratic Republic, MAL = Malaysia, MYA = Myanmar,
PHI = Philippines, VR China = Volksrepublik China, SIN = Singapore, THA = Thailand, VIE = Viet Nam.
Quelle: Authors’ illustration.

A. Methodik

This paper aims to explain the pace in output volatility and its variation
across ASEAN+3 by employing an empirical model, which will allow us to test the
main hypothesis of interest. Given this objective, we try to make maximum use of
both time and cross-country dimensions of our panel dataset, including data at an
annual frequency. Based on an annual dataset, our empirical model will consider
the possible slow adjustment of output volatility to changes in other variables
in any given year. daher, we specify a dynamic equation, which includes a
lagged dependent variable. The potential impacts of trade volatility and export
diversification on output volatility are formulated as follows:

GDPvoli,t = α0 + β1Tvoli,t−1 + β2DIVi,t−1 + β3CONi,t−1 + β4FTAi,T + β5Zi,T + ui,T

(1)
where the dependent variable, GDPvoli,T, is the volatility of the growth of real
GDP per capita for country i in period t. As argued in many studies (Bejan 2006,
Cavallo 2008, Jansen et al. 2009, and Haddad et al. 2013), we use the growth rate
instead of output level, since it is of greater interest to policy makers to measure
economic growth stability. Tvoli,t−1 is the volatility of total trade (sum of exports

176 Asiatischer Entwicklungsbericht

and imports) of country i in period t − 1. In diesem Papier, we use the 5-year rolling
standard deviation to measure the volatility of output and trade.

In equation (1), uit is an error term that contains country and time-specific

fixed effects:

uit = μi + εt + ϑit

(2)

where ϑit is independent and identically distributed with mean 0 and variance σ 2
ϑ .
DIVi,t−1 is the level of export diversification in country i in period t − 1, welche
is proxied by different alternative concentration indicators including the HHI, Die
Theil index, and the similarity index (SI). Cadot, Carrère, and Strauss-Kahn (2013)
decompose the Theil index into two indicators for the intensive and extensive
product margins. The extensive Theil index captures concentration in the number
of products (extensive product margin), whereas the intensive Theil index measures
concentration in the sales volume of products (intensive product margin).

The Theil index is thus T = T ext + T int, where the intensive Theil index is

given by
T int = 1
Nx

(cid:2)

k∈Nx

Rk
¯Rx

ln

(cid:4)

(cid:3)

Rk
¯Rx

and the extensive Theil index is
(cid:4)

(cid:3)

T ext = ln

N
Nx

(3)

(4)

where Nx denotes the number of exported products, Rk is the value of exports of
product k, and ¯Rx represents the mean value of exported products. Like the HHI, A
higher value means higher export concentration.

The HHI, by construction, measures the changes in the distribution of
export shares. Mit anderen Worten, this indicator is a measure of the degree of product
concentration. The following normalized HHI is used in order to obtain values
zwischen 0 Und 1:
(cid:5)

(cid:6)

(cid:7)

(cid:8)
2

N
i=1

xi j
X j
√
1 −

1/N

√

−

1/N

H j =

X j =

N(cid:2)

i=1

xi j

(5)

where xi j is the value of exports for country j and product i, and n is the number
of products at the level of classification. An index value closer to 1 indicates that
a country’s exports or imports are highly concentrated on a few products. Auf der
contrary, values closer to 0 mean that exports or imports are more homogeneously
distributed among a series of products.

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 177

Different from the HHI and Theil index, the similarity index SI is considered
the simplest measure for comparing export content across countries or across time.
To calculate export SI, we use the original export similarity index first introduced by
Finger and Kreinin (1979) and largely applied to trade data. The Finger and Kreinin
similarity index between two countries c and d is given as follows:

SI F K
C,D

=

(cid:4)

(cid:3)

min

N(cid:2)

i=1

xc,ich
Xc

, xd,ich
Xd

(6)

and xd,ich
Xd

where xc,ich
are the shares of good i in the total exports of country c and
Xc
D, jeweils. This index ranges from 0 to 1—from completely different export
shares to identical export shares. According to Finger and Kreinin (1979), Das
measure should not be affected by the relative sizes or scales of total exports
as it is intended to compare only patterns of exports across product categories.
All three concentration indexes are calculated at the 4-digit level of the Standard
Industry Trade Classification Revision 3 from the United Nations Commodity Trade
Statistics Database.

Apart from trade, Easterly, Islam, and Stiglitz (2000) find evidence of
bivariate correlation between per capita growth volatility and financial sector
Entwicklung, and price policies. Following this work, we also introduce in equation
(1) a set of control variables CONit, which includes inflation volatility and the
level of financial development, which is measured by the ratio of private sector
banking credit to GDP and a financial openness indicator. The quality of domestic
institutions has also been recognized as a cause of macroeconomic instability in the
Literatur. Zum Beispiel, research by Rodrik (1999) and Acemoglu et al. (2003) hat
found that corruption and conflict can lead to misleading trade policy. Ähnlich,
Mobarak (2005) and Klomp and de Haan (2009) have found that a democracy or
a stable political regime supports macroeconomic stability. Daher, we introduce in
equation (1) an institution variable which is proxied by the index of political rights
from Freedom House.3 The index ranges from 1 Zu 7, mit 1 indicating the most
“free” country. We also include the variable FTAi,t to see the impact of a free trade
agreement in ASEAN on the volatility of output. Zusätzlich, time dummies are
included to control common output fluctuations during financial crises, notably the
Asian economic crisis (1997–1999) and the global financial crisis (2007–2009).

The data are summarized in Table 2, which provides definitions and sources
for all variables and their units of measurement, mean, standard deviations,
and minimum and maximum values. This table also reports the correlation
coefficients between output volatility, trade volatility, and independent variables.
The signs, magnitudes, and significance of correlation coefficients help in modeling
and confirming the choice of variables. Jedoch, the values of the correlation
coefficients are diverse, ranging from negative to positive and small to large. Das

3See Balavac and Pugh (2016).

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

178 Asiatischer Entwicklungsbericht

0
0
0
1

.

8
1
1
0

.

1
9
0
0

.

8
1
1
0

.

*
5
0
4
0

.

*
1
1
2
0

.

*
9
4
2
0

.

6
3
0
0
−

.

1
0
3
4

.

5
3
0
.
0

6
3
5
.
0

7
0
3
.
0

4
7
2

E
D
A
R
T
M
Ö
C
N
U

7
0
7
0

.

6
6
5
5

.

5
5
0
.
0

3
0
2
.
1

5
4
1
.
0

8
3
8
.
0

2
1
2
.
0

0
0
4
.
2

4
7
2

4
7
2

E
D
A
R
T
M
Ö
C
N
U

E
D
A
R
T
M
Ö
C
N
U

H
T
ich

w
N
Ö
ich
T
A
l
e
R
R
Ö
C

e
D
A
R
T

e
M
Ö
C
N
ICH

S
C
ich
T
S
ich
T
A
T
S
e
v
ich
T
P
ich
R
C
S
e
D

S
C
ich
T
S
ich
T
A
T
S
j
R
A
M
M
u
S
:
l
e
D
Ö
M
j
T
ich
l
ich
T
A
l
Ö
V

T
u
P
T
u
Ö

.
2

e
l
B
A
T

j
T
ich
l
ich
T
A
l
Ö
V

j
T
ich
l
ich
T
A
l
Ö
V

X
A
M

N
ich
M

D
S

N
A
e
M

S
B
Ö

e
C
R
u
Ö
S

T
N
e
M
e
R
u
S
A
e
M

e
l
B
A
ich
R
A
V

0
0
0
1

.

9
5
4
0

.

0
0
0
.
0

7
7
0
.
0

5
0
1
.
0

5
2
3

ICH

D
W

F
Ö

N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S
G
N
ich
l
l
Ö
R

R
A
e
j
–
5

j
T
ich
l
ich
T
A
l
Ö
v

e
M
Ö
C
N
ICH

6
6
7
1

.

3
2
0
.
0

6
7
2
.
0

5
3
2
.
0

5
2
3

E
D
A
R
T
M
Ö
C
N
U

F
Ö

N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S
G
N
ich
l
l
Ö
R

R
A
e
j
–
5

j
T
ich
l
ich
T
A
l
Ö
v

e
D
A
R
T

H
T
w
Ö
R
G

A
T
ich
P
A
C

R
e
P
P
D
G

S
T
R
Ö
P
M

ich

D
N
A

S
T
R
Ö
P
X
e

l
A
T
Ö
T

N
Ö
ich
T
A
T
u
P
M
Ö
C

'
S
R
Ö
H
T
u
A

N
Ö
ich
T
A
T
u
P
M
Ö
C

'
S
R
Ö
H
T
u
A

l
ich
e
H
T
e
v
ich
S
N
e
T
N
ICH

X
e
D
N
ich

ICH

H
H

N
Ö
ich
T
A
T
u
P
M
Ö
C

'
S
R
Ö
H
T
u
A

l
ich
e
H
T
e
v
ich
S
N
e
T
X
E

X
e
D
N
ich

*
5
1
2
0
−

.

*
5
3
2
0
−

.

4
7
3
2

.

4
0
9
.
1
−

4
3
4
.
1

2
3
0
.
0

9
1
3

)
8
0
0
2
(

Ö
T
ICH

D
N
A

N
N
ich
H
C

P
D
G
Ö
T

N
E
P
Ö
A
K

*
4
2
1
0

.

3
9
0
0
−

.

*
8
6
1
0

.

4
0
0
0
−

.

*
6
5
7
0

.

*
3
1
5
0

.

2
3
1
6

.

6
1
6
4
1

.

9
2
6
3
5

.

6
7
2
.
1

0
7
7
.
0

9
1
0
.
0

0
2
1
.
1

7
8
7
.
2

9
4
8
.
7

7
0
7
.
2

8
4
3
.
7

1
2
5
.
5

4
7
2

4
7
2

4
3
3

2
9
0
0
−

.

1
1
0
0

.

1
8
0
1
3

.

0
6
0
.
0

5
1
6
.
5

9
6
4
.
6

1
2
3

ICH

D
W

F
Ö

N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S
G
N
ich
l
l
Ö
R

R
A
e
j
–
5

e
T
A
R
N
Ö
ich
T
A
fl
N
ich

ICH

D
W

T
ich
D
e
R
C

G
N
ich
k
N
A
B

R
Ö
T
C
e
S

e
T
A
v
ich
R
P

F
Ö
Ö
ich
T
A
R

E
D
A
R
T
M
Ö
C
N
U

E
D
A
R
T
M
Ö
C
N
U

N
Ö
ich
T
A
T
u
P
M
Ö
C

'
S
R
Ö
H
T
u
A

N
Ö
ich
T
A
T
u
P
M
Ö
C

'
S
R
Ö
H
T
u
A

X
e
D
N
ich
j
T
ich
R
A
l
ich

M
ich
S

X
e
D
N
ich

l
ich
e
H
T

T
N
e
M
P
Ö
l
e
v
e
D

S
S
e
N
N
e
P
Ö

l
A
ich
C
N
A
N
ich
F

j
T
ich
l
ich
T
A
l
Ö
v

l
A
ich
C
N
A
N
ich
F

N
Ö
ich
T
A
fl
N
ICH

*
7
0
2
0

.

1
8
0
0

.

0
0
0
7

.

0
0
0
.
1

4
1
2
.
2

9
4
7
.
4

1
5
3

e
S
u
Ö
H
M
Ö
D
e
e
R
F

e
H
T

G
N
ich
T
N
e
S
e
R
P
e
R
1
H
T
ich

w

,
7
–
1

F
Ö

e
R
Ö
C
S

S
T
H
G
ich
R

l
A
C
ich
T
ich
l
Ö
P

*
1
1
1
0

.

*
7
3
1
0

.

0
0
0
1

.

0
0
0
.
0

6
6
4
.
0

4
8
6
.
0

1
5
3

T
A
ich
R
A
T
e
R
C
e
S
N
A
E
S
A

7

D
N
A
M
Ö
D
e
e
R
F

F
Ö

e
e
R
G
e
D
T
S
e
T
A
e
R
G

M
Ö
D
e
e
R
F

F
Ö

e
e
R
G
e
D
T
S
e
l
l
A
M
S

e
H
T

j
M
M
u
D

T
N
e
M
e
e
R
G
A

e
D
A
R
T

e
e
R
F

,
N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S

=
D
S

,
X
e
D
N
ich

Ö
T
ICH
–
N
N
ich
H
C
=
N
E
P
Ö
A
K

,
X
e
D
N
ICH

N
A
M
H
C
S
R
ich

H
–
l
H
A
D
N
fi
R
e
H
=

ICH

H
H

,
T
C
u
D
Ö
R
P

C
ich
T
S
e
M
Ö
D

S
S
Ö
R
G
=
P
D
G

,
S
N
Ö
ich
T
A
N
N
A
ich
S
A

T
S
A
e
H
T
u
Ö
S
F
Ö

N
Ö
ich
T
A
ich
C
Ö
S
S
A
=
N
A
E
S
A

.
S
R
Ö
T
A
C
ich
D
N
ICH

T
N
e
M
P
Ö
l
e
v
e
D
D
l
R
Ö
W
=

ICH

D
W

,
e
S
A
B
A
T
A
D
S
C
ich
T
S
ich
T
A
T
S
e
D
A
R
T
j
T
ich
D
Ö
M
M
Ö
C
S
N
Ö
ich
T
A
N
D
e
T
ich
N
U
=
E
D
A
R
T
M
Ö
C
N
U

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

.
S
e
C
R
u
Ö
S

S
u
Ö
ich
R
A
v
N
Ö
D
e
S
A
B
N
Ö
ich
T
A
l
ich
P
M
Ö
C

'
S
R
Ö
H
T
u
A

:
e
C
R
u
Ö
S

.
l
e
v
e
l

%
5
e
H
T

T
A

e
C
N
A
C
fi

ich
N
G
ich
S

S
e
T
A
C
ich
D
N
ich

*

:
e
T
Ö
N

Trade Volatility in ASEAN+3: Impacts and Determinants 179

means we should expect different potential impacts of the independent variables on
output volatility.

Due to the inclusion of lagged dependent variables in equation (1), Die
combination of fixed effects and lagged dependent variables introduces serious
econometric bias. According to Nerlove (1967) and Nickell (1981), the ordinary
least squares estimates of the lagged dependent variable’s coefficient in a dynamic
panel model are biased due to the correlation between the fixed effects and the
lagged dependent variable.4 Judson and Owen (1999) also suggest that the bias is
inversely related to panel length (“T”), but potentially severe biases remain even at
T = 30. In this case, the preferred estimator is the GMM suggested by Arellano and
Bond (1991), which was then developed and extended by Arellano and Bover (1995)
and Blundell and Bond (1998) for two reasons. Erste, the GMM differences the
model of interest to remove country-specific effects or any time-invariant country-
specific variable. Zweite, this method eliminates possible correlation between
the country-specific effects and the regressors. In the GMM estimator, moment
conditions utilize the orthogonality conditions between differenced errors and
the dependent variable’s lagged values. This means that the disturbances ϑit are
serially uncorrelated and that the differenced error is a moving average with 1
Verzögerung (MA[1]). daher, two diagnostics are computed to test for first-order and
second-order serial correlation in the disturbances. We expect to reject the null of
the absence of first-order serial correlation and not reject the absence of second-
order serial correlation. In the GMM procedure, the number of moment conditions
increases with time span t. Infolge, the Sargan test is performed to test for
overidentification restrictions.

B.

Results and Discussion

Our empirical estimations are organized in two steps. Erste, the benchmark
dynamic GMM estimation treats all right-hand side variables other than the lagged
dependent variable as if they were exogenous. In this benchmark estimation, Wir
lag all independent variables by one period. Zweite, we run the dynamic GMM
estimator in which the export volatility is treated as endogenous by using an
additional instrument suggested by related literature. We follow Frankel and Romer
(1999) who developed “natural openness” instruments for openness. According to
Wei (2000), natural openness is found by estimating the level of trade openness
depending on a country’s size and geographic and linguistic characteristics. In
besondere, we estimate the following equation:

log (Openness) = β21Remoteness + β22 log (Bevölkerung) + β23Lang

+ β24Geo + ε21

4See further Baltagi (2008).

(7)

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

180 Asiatischer Entwicklungsbericht

(cid:6)

J

distancei j
where Remoteness =
GDPj/GDPw , a formula that measures a country’s average
weighted distance from its trading partners (Kopf 2003), where the weights are
the partner countries’ shares of total GDP (denoted by GDPw). “Lang” is proxied
by two dummies of language, “English” and “Chinese,” each of which takes
the value of 1 if the country speaks the respective language and 0 ansonsten.
“Geo” is constructed from the land area and a dummy called “landlock” if the
country is landlocked. All data are adopted from Centre d’Etudes Prospectives
et d’Informations Internationales (CEPII) Daten. The method to construct an IV is
based on the same concept as the gravity model. Appendix 2 reports a succession
of regressions on ASEAN+3 openness over the period 1990–2016. The predicted
value of the log of openness (trade-to-GDP ratio) from equation (7) is used as an IV
for trade volatility to deal with the biases.

All empirical results are summarized in Table 3. We estimate equation (1)
separately for each of our five export diversification indexes. It is also worth noting
that in all the output volatility regressions reported in Table 3, all diagnostics
are satisfactory, irrespective of whether the trade volatility terms are treated as
exogenous or endogenous. Erste, the presence of first-order serial correlation is not
rejected, while the presence of second-order serial correlation is rejected. Zweite,
the Sargan test does not reject the overidentification restrictions. Zusätzlich, Die
lagged dependent variable in all regressions is positive and statistically significant.
These results allow us to conclude that the dynamic GMM is an appropriate
estimator.

Going straight to the hypothesis of interest, we note that the estimated
coefficients of trade volatility have the expected sign with a high level of
significance. This result suggests that an increase in intra-ASEAN+3 trade
volatility leads to an increase in income volatility of member countries. Zusätzlich,
when the trade volatility terms are treated as endogenous, the sign and significance
of their estimates are qualitatively similar to those obtained in the benchmark GMM
Modelle. Jedoch, due to the endogeneity treatment, the empirically estimated
coefficients are somewhat smaller. The positive link between trade and income
volatility is also illustrated in Figure 5.

In sum, our empirical results confirm that trade has a quantitatively large
and robust positive effect on income (Frankel and Romer 1999). This result is also
consistent with that of Cavallo (2008) who argues that output volatility naturally
relates to the frequency and size of the shocks affecting an economy and to the
manner in which the economy handles the shocks. Entsprechend, trade openness is
associated with greater output volatility. Mit anderen Worten, the more exposed a country
is to trade, the more vulnerable it is to shocks coming from abroad.

Examining the estimated coefficients associated with export diversification
Variablen, we first note that their signs depend on the way export diversification
is measured. Zum Beispiel, estimates of the extensive Theil index have a negative
value, while those of the intensive Theil index have a positive value. Zweite, In

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 181

ICH
S

l
ich
e
H
T

e
v
ich
S
N
e
T
N
ICH

e
v
ich
S
N
e
T
X
E

ICH
H
H

N
Ö
ich
T
A
C
fi
ich
S
R
e
v
ich
D

T
R
Ö
P
X
E

F
Ö
e
R
u
S
A
e
M

l
e
D
Ö
M
j
T
ich
l
ich
T
A
l
Ö
V

T
u
P
T
u
Ö
e
H
T

F
Ö
S
T
l
u
S
e
R
R
Ö
T
A
M

ich
T
S
E
S
T
N
e
M
Ö
M

F
Ö
D
Ö
H
T
e

M
D
e
z
ich
l
A
R
e
N
e
G

.
3

e
l
B
A
T

*
*
4
6
4
0

.

*
*
1
8
0
0

.

)
6
6
1
0
(

.

)
7
3
0
0
(

.

4
0
0
0

.

*
*
1
1
0
0

.

)
6
0
0
0
(

.

)
5
0
0
0
(

.

0
0
0
0

.

)
0
0
0
0
(

.

5
0
0
0

.

*
6
1
0
0

.

)
3
1
0
0
(

.

*
1
8
7
0

.

)
9
6
3
0
(

.

*
6
2
1
0

.

)
3
6
0
0
(

.

*
1
1
0
0

.

)
5
0
0
0
(

.

8
0
0
0

.

)
0
1
0
0
(

.

1
0
0
0
−

.

)
1
0
0
0
(

.

9
1
0
0

.

)
3
2
0
0
(

.

6
0
0
0

.

)
8
0
0
0
(

.

8
5
0
0
−

.

)
8
0
0
0
(

.

6
4
0
0
−

.

*
*
1
2
0
0

.

)
3
4
0
0
(

.

)
9
0
0
0
(

.

)
9
2
0
0
(

.

7
1
0
0

.

)
1
1
0
0
(

.

*
*
2
1
5
.
0

*
*
2
2
9
.
0

)
5
9
1
.
0
(

*
5
7
0
.
0

)
5
3
0
.
0
(

4
0
0
.
0
−

*
*
0
1
0
.
0

)
2
1
0
.
0
(

)
4
0
0
.
0
(

0
0
0
.
0

)
0
0
0
.
0
(

7
0
0
.
0

)
7
1
0
.
0
(

9
1
0
.
0

*
*
5
2
0
.
0

)
3
4
0
.
0
(

)
0
1
0
.
0
(

)
3
1
0
.
0
(

5
4
0
.
0
−

)
2
7
3
.
0
(

*
0
1
1
.
0

)
9
5
0
.
0
(

7
0
0
.
0

)
4
1
0
.
0
(

6
0
0
.
0

)
9
0
0
.
0
(

1
0
0
.
0
−

)
0
0
0
.
0
(

8
2
0
.
0

*
*
4
1
0
.
0

)
3
2
0
.
0
(

)
7
0
0
.
0
(

1
4
0
.
0
−

)
5
3
0
.
0
(

*
*
*
7
2
0
.
0

)
8
0
0
.
0
(

*
*
2
1
5
.
0

*
*
8
7
0
.
0

)
8
8
1
.
0
(

)
6
3
0
.
0
(

1
0
0
.
0

*
*
1
1
0
.
0

)
2
1
0
.
0
(

)
4
0
0
.
0
(

0
0
0
.
0

)
0
0
0
.
0
(

7
0
0
.
0

)
5
1
0
.
0
(

8
1
0
.
0

*
*
4
2
0
.
0

)
6
4
0
.
0
(

)
0
1
0
.
0
(

)
1
1
0
.
0
(

3
5
0
.
0
−

*
*
6
8
8
.
0

)
5
9
3
.
0
(

*
8
0
1
.
0

)
0
6
0
.
0
(

6
1
0
.
0

)
9
1
0
.
0
(

7
0
0
.
0

)
0
1
0
.
0
(

1
0
0
.
0
−

)
0
0
0
.
0
(

9
2
0
.
0

)
4
2
0
.
0
(

1
1
0
.
0

)
7
0
0
.
0
(

2
5
0
.
0
−

*
*
6
2
0
.
0

)
7
3
0
.
0
(

)
9
0
0
.
0
(

*
1
6
4
.
0

)
7
1
2
.
0
(

*
9
6
0
.
0

)
8
3
0
.
0
(

0
5
0
.
0
−

*
*
1
1
0
.
0

)
5
3
0
.
0
(

)
5
0
0
.
0
(

0
0
0
.
0

)
1
0
0
.
0
(

5
0
0
.
0

)
8
1
0
.
0
(

*
1
2
0
.
0

*
*
5
2
0
.
0

)
4
3
0
.
0
(

)
1
1
0
.
0
(

)
1
1
0
.
0
(

6
4
0
.
0
−

*
3
7
8
.
0

)
4
0
4
.
0
(

*
1
0
1
.
0

)
7
5
0
.
0
(

5
3
0
.
0
−

)
7
2
0
.
0
(

7
0
0
.
0

)
0
1
0
.
0
(

1
0
0
.
0
−

)
1
0
0
.
0
(

8
2
0
.
0

)
6
2
0
.
0
(

7
1
0
.
0

)
0
1
0
.
0
(

3
2
0
.
0
−

*
*
9
2
0
.
0

)
9
2
0
.
0
(

)
1
1
0
.
0
(

*
*
9
1
5
.
0

*
*
2
4
9
.
0

e
l
B
A
ich
R
A
v

T
N
e
D
N
e
P
e
D

D
e
G
G
A
L

)
8
0
2
.
0
(

*
3
7
0
.
0

)
5
3
0
.
0
(

1
7
0
.
0
−

*
*
2
1
0
.
0

)
0
3
1
.
0
(

)
4
0
0
.
0
(

0
0
0
.
0

)
1
0
0
.
0
(

9
0
0
.
0

)
1
2
0
.
0
(

*
0
2
0
.
0

*
*
6
2
0
.
0

)
7
3
0
.
0
(

)
1
1
0
.
0
(

)
1
1
0
.
0
(

0
4
0
.
0
−

)
4
9
3
.
0
(

*
2
1
1
.
0

)
8
5
0
.
0
(

1
3
0
.
0
−

)
4
2
1
.
0
(

5
0
0
.
0

)
9
0
0
.
0
(

1
0
0
.
0
−

)
0
0
0
.
0
(

)
D
e
G
G
A
l
(
j
T
ich
l
ich
T
A
l
Ö
v

N
Ö
ich
T
A
fl
N
ICH

N
Ö
ich
T
A
C
fi
ich
S
R
e
v
ich
D

T
R
Ö
P
X
E

)
D
e
G
G
A
l
(

)
D
e
G
G
A
l
(
j
T
ich
l
ich
T
A
l
Ö
v

e
D
A
R
T

T
N
e
M
P
Ö
l
e
v
e
D

l
A
ich
C
N
A
N
ich
F

)
D
e
G
G
A
l
(

0
3
0
.
0

)
D
e
G
G
A
l
(

S
S
e
N
N
e
P
Ö
l
A
ich
C
N
A
N
ich
F

*
*
4
1
0
.
0

)
5
2
0
.
0
(

)
7
0
0
.
0
(

0
2
0
.
0
−

)
0
4
0
.
0
(

*
*
*
9
2
0
.
0

)
8
0
0
.
0
(

)
D
e
G
G
A
l
(

S
T
H
G
ich
R

l
A
C
ich
T
ich
l
Ö
P

T
N
e
M
e
e
R
G
A

e
D
A
R
T

e
e
R
F

S
ich
S
ich
R

C

S
u
Ö
N
e
G
Ö
D
N
E

S
u
Ö
e
N
e
G
Ö
X
E

S
u
Ö
N
e
G
Ö
D
N
E

S
u
Ö
e
N
e
G
Ö
X
E

S
u
Ö
N
e
G
Ö
D
N
E

S
u
Ö
e
N
e
G
Ö
X
E

S
u
Ö
N
e
G
Ö
D
N
E

S
u
Ö
e
N
e
G
Ö
X
E

S
u
Ö
N
e
G
Ö
D
N
E

S
u
Ö
e
N
e
G
Ö
X
E

e
D
A
R
T

F
Ö

T
N
e
M
T
A
e
R
T

)
9
9
0
0
(

.

.

9
0
1
−

)
9
4
0
0
(

.

.

4
2
1
−

)
7
7
2
0
(

.

3
0
8
3

.

)
4
2
7
0
(

.

)
5
1
2
0
(

.

5
0
4
1

.

)
1
8
7
0
(

.

.

5
6
1
−

.

7
9
1
−

)
9
8
0
.
0
(

2
.
1
−

)
1
3
2
.
0
(

5
1
.
0
4

)
7
3
6
.
0
(

7
.
1
−

)
8
2
0
.
0
(

7
4
.
1
−

)
2
4
1
.
0
(

)
7
9
6
.
0
(

0
4
.
5
1

2
.
2
−

)
5
8
0
.
0
(

9
1
.
1
−

)
6
3
2
.
0
(

)
2
6
6
.
0
(

6
5
.
9
3

2
7
.
1
−

)
9
2
0
.
0
(

7
5
.
1
−

)
7
1
1
.
0
(

)
9
3
7
.
0
(

4
7
.
4
1

8
1
.
2
−

)
5
0
1
.
0
(

7
2
.
1
−

)
4
0
2
.
0
(

)
4
2
7
.
0
(

3
0
.
8
3

2
6
.
1
−

)
3
2
0
.
0
(

3
5
.
1
−

)
6
2
1
.
0
(

7
2
.
2
−

9
3
7
.
0
(

5
7
.
4
1

)
9
9
0
.
0
(

5
2
.
1
−

)
1
1
2
.
0
(

)
5
1
6
.
0
(

8
6
.
0
4

5
6
.
1
−

)
8
2
0
.
0
(

6
4
.
1
−

)
3
4
1
.
0
(

)
0
0
7
.
0
(

5
3
.
5
1

2
.
2
−

)
e
u
l
A
v
–
P
(

T
S
e
T

N
Ö
ich
T
A
l
e
R
R
Ö
C

)
e
u
l
A
v
–
P
(

T
S
e
T

N
Ö
ich
T
A
l
e
R
R
Ö
C

l
A
ich
R
e
S

R
e
D
R
Ö
–
D
N
Ö
C
e
S

F
Ö
T
S
e
T

N
A
G
R
A
S

)
e
u
l
A
v
–
P
(

S
N
Ö
ich
T
C
ich
R
T
S
e
R

N
Ö
ich
T
A
C
fi
ich
T
N
e
D
ich
R
e
v
Ö

l
A
ich
R
e
S

R
e
D
R
Ö
–
T
S
R
ich
F

j
T
ich
l
ich
T
A
l
Ö
v

.
j
l
e
v
ich
T
C
e
P
S
e
R

,
S
l
e
v
e
l

%
0
1

D
N
A

,

%
5
,

%
1
e
H
T

T
A

e
C
N
A
C
fi
ich
N
G
ich
S

l
A
C
ich
T
S
ich
T
A
T
S

e
T
A
C
ich
D
N
ich

*

D
N
A

,
*
*

,
*
*
*

.
S
R
Ö
R
R
e
D
R
A
D
N
A
T
S

e
R
A

S
e
S
e
H
T
N
e
R
A
P
N
ich

S
e
R
u
G
ich
F
:
S
e
T
Ö
N

.
X
e
D
N
ich
j
T
ich
R
A
l
ich

M
ich
S

=

ICH
S

,
X
e
D
N
ICH
N
A
M
H
C
S
R
ich

H
–
l
H
A
D
N
fi
R
e
H
=

ICH

H
H

.
S
e
T
A
M

ich
T
S
e

'
S
R
Ö
H
T
u
A

:
e
C
R
u
Ö
S

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

182 Asiatischer Entwicklungsbericht

Figur 5. Output Volatility versus Trade Volatility

ASEAN+3 = Association of Southeast Asian Nations plus Japan, Volksrepublik China, and the Republic
von Korea; BIP = Bruttoinlandsprodukt.
Notiz: Out-of-sample predictions are from the generalized method of moments estimator with the Herfindahl–
Hirschman Index.
Quelle: Authors’ illustration.

most regressions, the impact of export diversification on output volatility is not
statistically significant. Exceptionally, we find a significant and positive estimated
coefficient on the similarity index in the GMM model that considers trade volatility
as an exogenous regressor. Jedoch, this coefficient loses its significance when
trade volatility is treated as an endogenous variable. Dritte, the empirical results
relating to export diversification are not sensitive to whether the trade volatility
terms are treated as endogenous. Gesamt, we do not find evidence of a significant
and direct link between export diversification and output volatility. Despite the
absence of a direct impact, export diversification could still affect output volatility
through its potential link with export volatility. This mechanism will be studied in
the next section.

We now turn our attention to the impact of institutional variables, represented
by political rights and FTAs. Erste, as expected, we find that
the estimated
coefficients for political rights all have a positive value. These coefficients are
generally significant, except in the intensive Theil index model. This finding means
that a “freer” country experiences less volatile output growth. Mit anderen Worten,
ASEAN+3 countries with a higher level of political rights and democracy have
a lower level of output volatility. Zweite, the estimates associated with FTAs are
negative but statistically insignificant. Tatsächlich, it is noteworthy that there are two
economic aspects of FTA. Einerseits, an FTA could divert trade away from
more efficient suppliers outside the area toward less efficient ones within the area.
Andererseits, an FTA could create trade that may not have otherwise existed
and thus raise a country’s national welfare. Reducing the output volatility of a
country should also be considered a trade creation effect of an FTA. Nach

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 183

to our empirical results, this effect potentially exists, but ASEAN+3 countries have
not yet realized it. Mit anderen Worten, the potential trade creation effects of FTAs are
not working in ASEAN+3 in precisely the way that is described by the relevant
Literatur, suggesting that more nuanced political economy explanations may be
erforderlich. This result can motivate policy makers in ASEAN+3 to conduct further
analysis to find out exactly what makes FTAs beneficial, whether by enhancing
growth or lowering income volatility. The resulting inferences could then be applied
to future FTA negotiations in order to effectively achieve an expected long-run
outcome from FTAs.

On the set of control variables, we find that inflation volatility plays a
significant role in explaining output volatility when trade volatility is treated
as endogenous. This means that a stable macroeconomic environment, welche
is captured by the level of inflation volatility, could promote output stability.
Im Gegensatz, financial variables in all regressions have insignificant estimated
coefficients. Zum Beispiel, as expected, the sign of the coefficients on the level of
financial development is negative, meaning a more developed financial system is
associated with more stable output growth, although the results are not statistically
significant. Looking at the financial openness variable, the estimates of this variable
are positive in all regressions. This positive sign supports the theoretical relationship
between financial openness and income volatility—the more integrated a financial
system the more sensitive a country is to external shocks and the more volatile its
income growth. Notwithstanding, this potential effect is not statistically significant.

V. Determinants of Trade Volatility

In diesem Abschnitt, we use a cross-sectional dataset of intra-ASEAN+3 bilateral
exports over the period 1990–2016 to investigate the potential determinants of
ASEAN+3 trade volatility.

A. Methodik

The potential determinants of ASEAN+3 trade volatility are modeled in the

following equation:

log (Xvol)ich, J,T

= α2 + θ0 log (INCOMEvol)ich,T

+ θ1log(INCOMEvol) J,T + θ2DIVXi, J,T

+ θ3TARIFFi, J,T + θ4FTAi, J,T + θ5Wi, J,T + μi, J + δt + ε3i,T

(8)

where we use the 5-year rolling standard deviation of bilateral exports Xvoli, J,T
between countries i and j as a proxy of trade volatility. Export diversification,
DIVXi, J,T, is proxied by five alternative indicators: the HHI, the SI, the Theil index,
and the intensive and extensive Theil indexes. INCOMEvoli,t and INCOMEvol j,t are
the income volatilities of the reporter country and its trading partner, jeweils.

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

.

/

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

184 Asiatischer Entwicklungsbericht

FTAi, J,t is a dummy variable indicating whether countries i and j are members of
the same FTA in year t. TARIFFi, J,t is the tariff applied on imported products from
country i to country j in year t, which is measured by the most favored nation (MFN)
and the effectively applied (AHS) weighted averages. Wi, J,t is a vector of control
Variablen, which are proxied by the bilateral exchange rate and political rights.5 μi, J
is a country-specific fixed effect, and δt is a year-specific fixed effect. The statistical
summary of all variables in equation (8) is reported in Table 4.

According to the Hausman test results, the fixed-effects estimator is the
appropriate technique to estimate equation (8).6 Zusätzlich, we also reestimate
equation (6) by using the IV technique, which allows us to tackle the potential
endogeneity of the income volatility variable. We use country’s infant mortality
rate and the corruption perception index as two external instrumental variables of
the income volatility variable.7 According to Kalemli-Ozcan (2002), birth rate is
indeed a determinant of economic growth. Lower child mortality results in higher
educational investment and lower fecundity by parents, which in turn causes lower
population growth and higher economic growth. The other variable, corruption
Kontrolle, is an essential instrument of governance, which has a sizable long-run effect
on economic growth (Kaufmann, Kraay, and Mastruzzi 2007).

Außerdem, we distinguish the impact of each ASEAN+1 FTA (zwischen
ASEAN and the PRC, Japan, or the Republic of Korea) on trade volatility by
estimating equation (8) with the presence of three dummies—ACFTA, AJFTA, Und
AKFTA—which represent ASEAN’s FTA with the PRC, Japan, and the Republic
von Korea, jeweils.

B.

Results and Discussion

We summarize the empirical results of FE and IV estimators in Table 5. In
besondere, the IV results that treat income volatility as endogenous indicate the
validity of the instruments, which is shown by several diagnostic tests. Erste, Die
p-values for Anderson’s canonical correlation test confirm the adequate explanatory
power of our instruments. Zweite, the Cragg–Donald Wald F-statistics allow us
to reject the null hypothesis of a weak instrument. Dritte, the overidentification
restriction test (the Sargan statistics) accepts the null hypothesis of the instruments’
validity.

The estimated results in Table 5 allow us to confirm a positive and significant
impact of income volatility of the reporter country on its bilateral export volatility.

5The bilateral real exchange rate is calculated as the product of the nominal exchange rate and the relative
GDP deflator in each country: RERi j= ei jt *(p jt /pit ), where ei jt is the nominal exchange rate (IMF, International
Financial Statistics), p jt is the GDP deflator of the exporter, and pit is the GDP deflator of the importer.

6The Hausman test results are not reported here to save space.
7Infant mortality rate is extracted from World Bank data. Corruption perception index is adopted from

Transparency International.

l

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

P

:
/
/

D
ich
R
e
C
T
.

M

ich
T
.

/

e
D
u
A
D
e
v
/
A
R
T
ich
C
e
–
P
D

l

F
/

/

/

/

3
7
2
1
6
7
1
8
4
6
8
2
8
A
D
e
v
_
A
_
0
0
1
5
3
P
D

/

.

F

B
j
G
u
e
S
T

T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

Trade Volatility in ASEAN+3: Impacts and Determinants 185

*
1
0
3
0

.

7
3
0
0
−

.

*
0
6
4
0

.

*
1
7
3
0

.

*
6
0
4
0
−

.

*
1
9
0
0

.

*
8
6
0
0

.

*
7
8
1
0

.

*
8
9
0
0

.

8
0
0
0

.

*
7
7
0
0
−

.

8
1
0
0
−

.

*
7
9
1
0

.

*
3
8
1
0

.

=
3
+
N
A
E
S
A

N
Ö
ich
T
A
l
e
R
R
Ö
C

T
R
Ö
P
X
E
H
T
ich

w

j
T
ich
l
ich
T
A
l
Ö
V

X
A
M

N
ich
M

D
S

N
A
e
M

S
B
Ö

e
C
R
u
Ö
S

T
N
e
M
e
R
u
S
A
e
M

S
C
ich
T
S
ich
T
A
T
S
j
R
A
M
M
u
S
:
l
e
D
Ö
M
j
T
ich
l
ich
T
A
l
Ö
V

T
R
Ö
P
X
E

.
4

e
l
B
A
T

0
0
0
1

.

6
3
2
5

.

1
0
0
0

.

2
6
3
.
0

2
5
3
.
0

4
4
5
,
2

E
D
A
R
T
M
Ö
C
N
U

F
Ö

N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S

G
N
ich
l
l
Ö
R

R
A
e
j
–
5

j
T
ich
l
ich
T
A
l
Ö
v

T
R
Ö
P
X
e

l
A
R
e
T
A
l
ich
B

S
'
3
+
N
A
E
S
A

*
2
1
1
0

.

*
6
6
0
0

.

1
7
9
0

.

0
0
0
0

.

7
1
1
.
0

0
6
1
.
0

0
6
8
,
2

ICH

D
W

P
D
G

F
Ö

N
Ö
ich
T
A
ich
v
e
D

D
R
A
D
N
A
T
S