Sie sind auf Seite 1von 252

University of Wollongong

Research Online
University of Wollongong Thesis Collection

University of Wollongong Thesis Collections

2014

Monetary transmission mechanism analysis in a


small, open economy: the case of Vietnam
Ha Thanh Nguyen
University of Wollongong

Recommended Citation
Nguyen, Ha Thanh, Monetary transmission mechanism analysis in a small, open economy: the case of Vietnam, Doctor of Philosophy
thesis, School of Accounting, Economics and Finance, University of Wollongong, 2014. http://ro.uow.edu.au/theses/4286

Research Online is the open access institutional repository for the


University of Wollongong. For further information contact the UOW
Library: research-pubs@uow.edu.au

Monetary transmission mechanism analysis


in a small, open economy: The case of Vietnam

This research thesis is submitted as part of the requirements


for the degree
DOCTOR OF PHILOSOPHY
of

By
Ha Thanh Nguyen

B.Econ, National Economic University, Vietnam


MS.Econ, National Economic University, Vietnam

SCHOOL OF ACCOUNTING, ECONOMICS AND FINANCE


2014

CERTIFICATION

I, Ha Thanh Nguyen, declare that this thesis, submitted in fulfilment of the


requirements for the award of Doctor in Philosophy in the School of Accounting,
Economics and Finance, Faculty of Business, University of Wollongong, is wholly
my own work unless otherwise referenced or acknowledged. The document has not
been submitted for qualification at any other academic institution.

Ha Thanh Nguyen
Wollongong, June 2014

ii

ABSTRACT

The transmission mechanism of monetary policy describes the dynamic stages


in which a central banks monetary policies are transmitted to real output and prices.
It plays a crucial, perhaps even central, role in the study of monetary economics.
However, few studies have focussed on developing small, open economies, and even
fewer have covered characteristics such as the aggregate demand components, the
low independence of monetary policy, the weak developing financial markets and
structural changes in the economy. Few quantitative empirical studies have been
conducted on the monetary transmission mechanism in Vietnam, and they do not
include a non-recursive structural vector autoregression model with structural
breaks, and characteristics of a small, open economy. This study attempts to fill such
research gaps.
The aim of this thesis is examining the role of monetary policy in the
Vietnamese economy, in terms of shocks to the economy via different channels, and
the international dimension of the monetary transmission mechanism in the small,
open economy of Vietnam. To answer these questions, the study uses quarterly data
for the period 2000:1 to 2011:4 and the Structural Vector Autoregression (SVAR)
approach. The SVAR is more suited than the Vector Autoregression (VAR)
approach to examining structural shocks as a test of multivariate models. The
findings show that the statistically significant break dates are found in this study, so
a dummy variable is included in the SVAR model of Vietnam.
The studys findings reveal that the impacts of a monetary contraction on
domestic variables are largely consistent with theories, except for its impact on price
level. A monetary contraction causes the output to decrease, but weak. The
monetary channels explain about 20 percent of fluctuations in real output. The
higher output causes the domestic price to increase. A money shock is not a main
reason in increasing the domestic price because of its small and statistically
insignificant effects. Vietnams current price-controlling regimes are effective in
ensuring price stability in spite of shocks in world oil and rice prices.

iii

The outcomes of this research illustrate that the interest rate channel is the
most effective channel for transmission to the price fluctuations and the exchange
rate for output variations. The study found the stock price channel as the least
effective, which is consistent with the modest role of the stock market in the
Vietnamese economy. This limits the statistically significant contemporaneous
effects of monetary policy (the interest rate and credit) on stock prices. The
monetary aggregate could not be a reliable tool to contemporaneously transmit
monetary policy signals to the financial market.
In addition, the study finds that domestic shocks have a more significant
impact on the Vietnamese economy than foreign shocks do. Policymakers need to
focus on solving the internal problems of the economy. However, they should keep a
keen eye on some foreign factors in the short run because of significant effects of
foreign output on domestic output, of foreign output on the monetary aggregate, of
world gold price on the monetary aggregate, and of world rice price on the domestic
price.
In the analysis of aggregate demand components, significant contemporaneous
relationships (the domestic price-exports, the domestic output-imports, and exportsstock prices) are evident. These relationships suggest policymakers cautious
consideration on the trade-off between economic growth and price stability. Imports
and exports are quickly affected by a monetary contraction; however, negative
impacts on import demand contribute to improvements in Vietnams trade balance
in the short run. The credit channel plays an important role in trade activities. The
role of the exchange rate channel is less for prices, imports, and exports. Finally, the
results illustrate that the effects of a monetary contraction on private investment and
private consumption are consistent with economic theory and expectation, but these
responses occur in a short time. The influence of monetary transmission channels on
private consumption is higher than that on private investment.
In this study, there is no evidence of the output and forward discount puzzles.
The puzzles of price, liquidity and exchange rate are addressed but they are
transitory, occurring in the short run and disappearing over time.
Based

on

the

major

findings,

this

study

suggests

some

policy

recommendations: (i) there is a need to focus on solving the internal problems of the
iv

Vietnamese economy to create a strong motivation for the development of Vietnam;


(ii) monitoring the important transmission channels, including the interest rate,
exchange rate and credit channels in formulating monetary policy; coordinating
monetary policy with other policies to effectively control inflation in Vietnam.

ACKNOWLEDGEMENTS

First, I would like to thank my principal supervisor, Dr. Khorshed Chowdhury, for
his encouragement and understanding. His expertise and enthusiasm have helped
broaden my knowledge. Moreover, he has provided many materials and much
advice for addressing econometrics problems and using computer programmes; all
of them have been useful for my research.

Second, I would like to thank the co-supervisor for my thesis, A/P Ed Wilson, for
his critical comments to develop my ideas about constructing structural models
appropriate for the case of Vietnam as well as analysing the model. These made the
arguments and analyses in my thesis more convincing.

I would also like to express my sincere thanks to Dr. Kankesu Jayantakumaran for
discussing with me issues concerning the thesis. I would like to thank the
researchers working at the State Bank of Vietnam and the National Economic
University (Vietnam) for their support in my study. Moreover, I would like to thank
Program 165 (Vietnam), which provided a scholarship to enable me to pursue my
doctoral studies at University of Wollongong. Also, I would like to thank Maree
Horne, Helen Harman and Phil Luskan, who took care of the research students
studying at the Faculty of Commerce (Business). I would like to thank my fellow
PhD students in School of Economics for their friendship and the interesting
discussion with them about studying and living issues. I also acknowledge the
editorial assistance of Laura E. Goodin.

Most importantly, I would like to thank my family, especially my parents, my wife


and my lovely daughters, who are always by my side during my study at University
of Wollongong, Australia.

vi

TABLE OF CONTENTS

CERTIFICATION .............................................................................................. ii
ABSTRACT....................................................................................................... iii
ACKNOWLEDGEMENTS ............................................................................... vi
TABLE OF CONTENTS.................................................................................. vii
LIST OF FIGURES ........................................................................................... xi
LIST OF TABLES ........................................................................................... xiv
CHAPTER 1 ...................................................................................................... 1
INTRODUCTION ............................................................................................ 1
1.1

RESEARCH BACKGROUND AND MOTIVATION .................................... 1

1.2

RESEARCH OBJECTIVES AND QUESTIONS ............................................ 4

1.3

RESEARCH SCOPE ........................................................................................ 6

1.4

STRUCTURE OF THE THESIS ...................................................................... 7

CHAPTER 2 .................................................................................................... 10
LITERATURE REVIEW .............................................................................. 10
2.1

INTRODUCTION .......................................................................................... 10

2.2

CHANNELS OF MONETARY TRANSMISSION MECHANISM.............. 10

2.2.1

Channels of the monetary transmission mechanism and their roles ............... 10

2.2.2

The interest rate channel ................................................................................. 13

2.2.3

The credit channel........................................................................................... 14

2.2.4

The exchange rate channel .............................................................................. 16

2.2.5

The asset price channel ................................................................................... 17

2.3

EMPIRICAL

STUDIES

ON

THE

MONETARY

TRANSMISSION

MECHANISM IN SMALL, OPEN ECONOMIES .................................................... 19


2.3.1

Some selected studies examining small, open economies .............................. 19

2.3.2

Empirical studies on the monetary transmission mechanism of Vietnam ...... 22

2.4

STRUCTURAL

FACTORS

AND

MONETARY

TRANSMISSION

MECHANISM ............................................................................................................. 26
2.5

MODELS IN ANALYSING MONETARY POLICY ................................... 29

2.6

CONCLUDING REMARKS .......................................................................... 32

CHAPTER 3 .................................................................................................... 34
THE VIETNAMESE ECONOMY, INSTITUTION AND POLICY ......... 34
vii

3.1

INTRODUCTION .................................................................................. 34

3.2

VIETNAMS ECONOMIC HISTORY FROM 1975 TO 2011............. 35

3.2.1 Period from 1975 till 1986 ..................................................................... 35


3.2.2 The Economic Renovation in 1986 ........................................................ 37
3.2.3 Post-Doi Moi (after 1986) ...................................................................... 37
3.3

THE ECONOMIC GROWTH AND INFLATION CONCERN ........... 40

3.4

THE OPENNESS OF VIETNAMESE ECONOMY ............................. 44

3.5

ENTERPRISE SECTORS IN VIETNAMS ECONOMY .................... 47

3.5.1 State-owned enterprises.......................................................................... 47


3.5.2 Non-state sector ...................................................................................... 48
3.6

VIETNAMS DOMESTIC FINANCIAL MARKET ............................ 49

3.6.1 Vietnams banking sector ....................................................................... 50


3.6.2 The stock market of Vietnam ................................................................. 52
3.7

INSTITUTIONS AND POLICIES FOR A MARKET ECONOMY..... 53

3.7.1 Legal system and policies for the enterprise environment and the market53
3.7.2 Legal system and policies for the financial market ................................ 54
3.7.3 Changes in the States management approach ....................................... 55
3.8

LEGAL SYSTEM FOR CONDUCTING MONETARY POLICY AND

THE ROLE OF STATE BANK OF VIETNAM .............................................. 55


3.9

THE CONDUCT OF VIETNAMESE MONETARY POLICY AND

VIETNAMESE MONETARY TRANSMISSION MECHANISM .....................


3.9.1 The interest rates liberalisation process ................................................. 58
3.9.2 Vietnams exchange rate regime ............................................................ 59
3.9.3 Vietnamese monetary transmission mechanism .................................... 61
3.10 CONCLUDING REMARKS ................................................................. 62
CHAPTER 4 .................................................................................................... 66
DATA AND STABILITY TESTS ................................................................. 66
4.1

DATA AND SOURCES ........................................................................ 66

4.2

SEASONALITY ANALYSIS ............................................................... 69

4.2.1 Seasonal-factor identification ................................................................. 69


viii

4.2.2 Modelling seasonality with dummy variables ........................................ 75


4.2.3 Seasonal adjustment with complicated computing procedures .............. 81
4.3

DESCRIPTIVE STATISTICS ............................................................... 85

4.4

UNIT ROOT TESTS .............................................................................. 87

4.5

CONCLUDING REMARKS ................................................................. 91

CHAPTER 5 .................................................................................................... 93
THE STRUCTURAL VECTORAUTOREGRESSION MODEL ............. 93
5.1

INTRODUCTION .................................................................................. 93

5.2

THEORETICAL SVAR MODEL ......................................................... 94

5.3

APPROPRIATE DATA FORM FOR SVAR MODELS ....................... 96

5.4

BENCHMARK MODELS ..................................................................... 97

5.4.1 The empirical study of Cushman and Zha (1997) .................................. 98


5.4.2 The empirical study of Kim and Roubini (2000) ................................... 99
5.4.3 The empirical study of Afandi (2005) .................................................. 101
5.5

MODEL DESIGN FOR VIETNAMS ECONOMY ........................... 102

5.6

CONCLUDING REMARKS ............................................................... 111

CHAPTER 6 .................................................................................................. 113


ESTIMATION RESULTS AND ANALYSIS ............................................ 113
6.1

INTRODUCTION ................................................................................ 113

6.2

LAG LENGTH AND VAR STABILITY CHECK ............................. 115

6.3

CONTEMPORANEOUS MATRIX .................................................... 117

6.4

IMPULSE RESPONSE FUNCTION .................................................. 121

6.4.1 Responses of domestic variables to positive foreign interest rate shocks122


6.4.2 Responses of domestic variables to positive foreign output shocks .... 124
6.4.3 Responses of domestic variables to positive world price shocks ......... 126
6.4.4 Responses of domestic variables to monetary policy shocks ............... 131
6.5

VARIANCE DECOMPOSITION........................................................ 138

6.6

THE SVAR ANALYSIS FOR INTERNATIONAL TRANSMISSION144

6.6.1 Contemporaneous matrix ..................................................................... 144

ix

6.6.2 Impulse response functions of exports and imports to a contractionary


monetary policy .............................................................................................. 150
6.6.3 Variance decomposition ....................................................................... 151
6.7

THE SVAR ANALYSIS FOR CONSUMPTION AND INVESTMENT

BEHAVIOUR ................................................................................................. 153


6.7.1 Contemporaneous matrix ..................................................................... 153
6.7.2 Impulse response functions of private investment and consumption to a
contractionary monetary policy ...................................................................... 156
6.7.3 Variance decomposition ....................................................................... 157
6.8

THE ROBUSTNESS OF RESULTS ................................................... 159

6.8.1 The adjustment in the lag length, sample length and standard errors .. 160
6.8.2 The revision in restrictions ................................................................... 160
6.9

CONCLUDING REMARKS ............................................................... 162

CHAPTER 7 .................................................................................................. 167


SUMMARY AND RECOMMENDATIONS.............................................. 167
7.1

SUMMARY ......................................................................................... 167

7.2

POLICY RECOMMENDATIONS ...................................................... 172

7.3

CONTRIBUTION AND SIGNIFICANCE OF THE RESEARCH..... 174

7.4

SUGGESTION FOR FURTHER STUDIES ....................................... 176

APENDIX A - Appendix to Chapter 2 ........................................................ 178


APENDIX B - Appendix to Chapter 4 ........................................................ 189
APENDIX C Appendix to Chapter 6 ....................................................... 208
REFERENCES.............................................................................................. 224

LIST OF FIGURES
Figure 2.1: Effects of the Monetary Transmission Mechanism on the Economy ...... 12
Figure 3.1: Vietnams M2 growth, GDP growth, CPI (%) ........................................ 42
Figure 3.2: Vietnams Trade Volume (millions of USD) .......................................... 45
Figure 3.3: Structure of the Financial Market in Vietnam ......................................... 49
Figure 3.4: Share prices in Vietnam from 2000 to 2011 ............................................ 52
Figure 3.5: Organisational structure of the State Bank of Vietnam ........................... 57
Figure 3.6: Important Milestones about Monetary Institutions and Policies ............. 58
Figure 3.7: The deposit and lending interest rate in VND during the period of 20002011 ............................................................................................................................ 59
Figure 4.1: Foreign variables ..................................................................................... 71
Figure 4.2: Domestic variables................................................................................... 72
Figure 4.3: Gross Domestic Product (Y), Private Investment (PI) and Private
Consumption (PC)...................................................................................................... 74
Figure 6.1: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in the Federal Fund Rate .......................................................... 123
Figure 6.2: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in Foreign Output ..................................................................... 125
Figure 6.3: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in World Oil Prices ................................................................... 127
Figure 6.4: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in World Rice Price .................................................................. 128
Figure 6.5: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in World Gold Price ................................................................. 130

xi

Figure 6.6: Impulse Responses of Domestic Variables to Structural Shocks of One


Standard Deviation in Domestic Interest Rates........................................................ 133
Figure 6.7: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in Money Aggregate ................................................................. 134
Figure 6.8: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in Credit .................................................................................... 135
Figure 6.9: Impulse Responses of Domestic Variables to Structural Shocks of One
Standard Deviation in Real Effective Exchange Rate .............................................. 136
Figure 6.10: Variance Decomposition of Y (output) ................................................ 138
Figure 6.11: Variance Decomposition of CPI.......................................................... 140
Figure 6.12: Variance Decomposition of R (interest rate) and M (money supply) .. 141
Figure 6.13: Variance Decomposition of CR (credit), E (real effective exchange rate)
and VNI (stock price index)...................................................................................... 142
Figure 6.14: Impulse Responses of Exports (VE) and Imports (VI) to Structural
Shocks of One Standard Deviation in the domestic Interest Rate (R). .................... 150
Figure 6.15: Variance Decomposition of Exports .................................................... 151
Figure 6.16: Variance Decomposition of Imports .................................................... 152
Figure 6.17: Impulse Response of Private Investment (PI) and Consumption (PC) to
Structural Shocks of One Standard Deviation in Interest Rate (R) .......................... 156
Figure B1: Seasonal adjusted data for additive model (Data1_SA) and multiplicative
model (Data_SA) in X12-ARIMA of Y, PI and PC ................................................ 189
Figure B2: Gross domestic product (GDP), Private investment (PI), and Private
consumption (PC): seasonally adjusted value with X12 and T/S ............................ 194
Figure B3: Irregular factors of investment (PI) and consumption (PC) estimated by
X12-ARIMA and T/S. .............................................................................................. 195
xii

Figure B4:

Gross domestic product (Y), private investment (PI) and private

consumption (PC): original data, seasonally adjusted data and seasonal factor from
X12-ARIMA. ........................................................................................................... 197
Figure C1: The impulse responses for the Monte Carlo approach ........................... 220

xiii

LIST OF TABLES

Table 2.1: Common Puzzles in Empirical Studies ..................................................... 20


Table 2.2: Influence of an Economys Features on the Monetary Transmission
Mechanism ................................................................................................................. 27
Table 3.1: Vietnams Real GDP Growth and Inflation from 1987 to 1990 ............... 38
Table 3.2: Some Economic Indicators from 1991 to 1995 ........................................ 39
Table 3.3: Real GDP and Sectoral Growth Rates in 2001-2005 ................................ 39
Table 3.4: Some Economic Indicators from 2005 to 2011 ........................................ 40
Table 3.5: Value Added in Agriculture, Industry and Services Sectors of Vietnam
(% of GDP)................................................................................................................. 41
Table 3.6: Gross Capital Formation of Vietnam (% of GDP) .................................... 43
Table 3.7: Influence of Factors from Demand Side on Production ........................... 44
Table 3.8: The Rate of Export, Import and Trade Volume in GDP from 1994 to 2011
(% of GDP)................................................................................................................. 46
Table 3.9: Structure of GDP....................................................................................... 47
Table 3.10: Financial Market in Vietnam .................................................................. 49
Table 3.11: Number of Credit Institutions ................................................................. 50
Table 3.12: Share of Total Asset by Type of Institutions (%) ................................... 51
Table 3.13: Share of Credit by Type of Institutions (%) ............................................ 51
Table 3.14: Stock Market Capitalisation .................................................................... 52
Table 3.15: Vietnams Financial Development Index 2008-2011 ............................. 60
Table 4.1: Data Description ....................................................................................... 67
Table 4.2: Estimated Values for Gross Domestic Product (Y), Private Investment
(PI) and Private Consumption (PC) ........................................................................... 77
xiv

Table 4.3: Estimated Value for Net Exports (VE - VI) .............................................. 80
Table 5.1: The SVAR model of Cushman and Zha (1997) ........................................ 98
Table 5.2: The SVAR model of Kim and Roubini (2000) ....................................... 100
Table 5.3: The SVAR model of Afandi (2005) ........................................................ 101
Table 5.4: The contemporaneous matrix (the A matrix) .......................................... 104
Table 6.1: VAR Lag Order Selection Criteria.......................................................... 115
Table 6.2: VAR Residual Serial Correlation LM Tests ........................................... 116
Table 6.3: Roots of Characteristic Polynomial ........................................................ 116
Table 6.4: Estimated Contemporaneous Coefficients of Model VN1 ..................... 117
Table 6.5: Estimated Contemporaneous Coefficients of Model VN2 ..................... 118
Table 6.6: The Models for International Transmission ............................................ 145
Table 6.7: Results of the Contemporaneous Coefficients for the International
Channel..................................................................................................................... 147
Table 6.8: Extended Model with Investment and Consumption Behaviour ............ 153
Table 6.9: Results of the Contemporaneous Coefficients for the Extended Model
with Investment and Consumption Behaviour ......................................................... 154
Table 6.10: Revised Contemporaneous Matrix ........................................................ 161
Table A1: Milestones of Vietnam economy from 1986 ........................................... 178
Table A2: Milestones of Vietnam financial market and monetary policy from 1986
to 2011 ...................................................................................................................... 183
Table B1: Quality measures and criterion ................................................................ 190
Table B2: Quality indexes under both models (additive and multiplicative) .......... 192
Table B3: X12-ARIMA and T/S used in countries .................................................. 193
Table B4: Moving holidays in Vietnam ................................................................... 196
Table B5: Descriptive statistics for the foreign sector (2000:1-2011:4) .................. 199
xv

Table B6: Descriptive statistics for the domestic sector (2000:1-2011:4) ............... 200
Table B7: Summary of results of univariate unit root tests with break(s) using
Nelson-Plosser data set............................................................................................. 201
Table B8: Unit root tests without structural break ................................................... 202
Table B9: Results from LS tests with one and two structural breaks (seasonally
adjusted and log form).............................................................................................. 205
Table B10: Summary of results from unit root tests ................................................ 207
Table C1: VAR Stability tests with two lags ........................................................... 208
Table C2: Variance Decomposition of Domestic Variables .................................... 210
Table C3: VAR Stability tests with the international channel ................................. 214
Table C4: Variance Decomposition of Exports and Imports ................................... 215
Table C5: VAR Stability tests with the extended model with the investment and
consumption behaviour. ........................................................................................... 217
Table C6: VAR Stability tests with two sub-samples .............................................. 218

xvi

DEDICATED WITH LOVE

To my parents, my wife Hang Thi Thu Nguyen, and my beloved


daughters Linh Ha Nguyen and Ngoc Minh Nguyen.

I am so grateful for your love and support. You suffered a lot during the
period of my PhD study.

xvii

CHAPTER 1
INTRODUCTION

1.1

RESEARCH BACKGROUND AND MOTIVATION


The monetary transmission mechanism (MTM) describes the dynamic

stages in which a central banks monetary policy is transmitted to real output and
prices. This means that changes in the official cash rate or suitably defined
monetary aggregate affect changes in the financial-sector price and quantity
variables, and subsequently real output and prices. According to Mishkin (1995),
the mechanism includes the interest rates, exchange rates, other asset prices and
the credit channels 1 . This transmission mechanism plays a crucial role in
macroeconomic policies. Studying the transmission channels contributes to the
understanding of the effectiveness of channels as well as the way monetary
shocks affect the economy. Dungey and Fry (2009) conclude that understanding
the effects of monetary shocks via MTM channels will contribute to better
management of the economy.
Vietnam is an emerging economy, and its economic reforms beginning in
1986 resulted in significant achievements in economic growth and trade activities.
However, there are some weaknesses in maintaining sustainable growth, such as
high inflationary pressures, trade deficits, and an underdeveloped financial
system, that could make the Vietnamese economy vulnerable to internal and
external shocks. Moreover, Vietnamese monetary policy plays a crucial role in
developing the economy. The goals of monetary policy are to stabilise the
currencys value, ensuring the security of the banking system and speeding up
socio-economic development. Therefore, assessing the role of each channel of
monetary policy is necessary for examining the role of monetary policy in
developing the Vietnamese economy.

Wilson (2011) added that inflation expectations change as a channel in studying MTM and
demonstrated that the wealth effect is expressed in MTM channels as the quantity of bank
lending, the market valuations of bonds, foreign exchange, capital, and the expected changes in
commodity prices.
1

Whilst there are various studies on the MTM, most of them focus on
developed countries such as the U.S., Australia, and other Western countries. A
small open economy is defined by Deardorff (2010) as the economy whose
participation in the international markets is small enough so that its policies will not
affect the world markets prices or incomes. Although there have recently been more
studies of small open economies, only some of them analyse developing countries
such as China, Thailand, Indonesia, and Iran. There have not been many studies of a
small, open economy like that of Vietnam, which is a developing country with an
under-developed financial market and a low independence of monetary policy (these
characteristics are considered in the next sections). Moreover, there are only a few
quantitative studies on the MTM for Vietnam by Le and Pfau (2009) and Tran
(2009). There are some limitations in these studies, so further explorations are
necessary to better assess the effects of the Vietnamese MTM on key
macroeconomic variables, including economic growth, inflation and the trade
balance. Due to the vast literature on the MTM, this study focuses on selected papers
from both developed and developing countries.
This research aims to utilize the knowledge and experience of those countries
cited above to propose policy implications related to the MTM for the small, open
economy of Vietnam.
First, the research will contribute to the existing knowledge on MTM analysis
in small open economies, especially for developing countries. Studying monetary
shocks based on the effects of MTM channels and understanding this mechanism is
not simple. Bernanke and Gertler (1995, p. 27) describe MTM as a black box when
analysing the effects of monetary policy. Mishkin (1995) also illustrates that there
could be unpredicted or unwanted effects from implementing monetary policy.
Therefore, studying the MTM seems to pose quite a challenge. The current study
will examine the international dimension of the Vietnamese MTM, because Vietnam
is a small open economy in which import turnover accounted for 80 percent of GDP
in 2010, and export turnover made up 68.4% of GDP (General Statistics Office of
Vietnam, 2010). Thus, examining the role of the international transmission of the
Vietnamese MTM is worthy of study. This study will focus on many significant
implications of the effects of the shocks on the Vietnamese monetary policys

international dimension; specifically, monetary shocks on the trade balance.


Moreover, examining MTM channels in the case of Vietnam will contribute to
enriching the existing knowledge of the MTM in diverse countries.
Second, examining all MTM channels is of significant importance for all
economies, including the Vietnamese economy. This study will consider the
theoretical and empirical studies on the role of monetary policy in a developing
economy. Mishkin (1995) asserts that monetary policies have been at the centre of
macroeconomic policy making. The MTM is the process that monetary authorities
conduct to control the money supply, with a view to achieving economic growth
targets as well as price stability. According to Bain and Howells (2003), there are
three main ways to change monetary policy: interest rate control, monetary base
control and direct controls. The first preferred mostly by monetary authorities in
developed countries is changing short term interest rate to relieve liquidity shortages
in the banking system. The second plays a crucial role in developing economies, as a
central banks change in the monetary base altering the money supply will result in
expansionary monetary policy (to fight unemployment issues) or contractionary
monetary policy (to curb inflation). The third implies that changes in regulations
directly affect banks lending growth. This study examines the role of monetary
policy in the Vietnamese economy, especially its role in enhancing economic growth
over the past two decades. Determining the transmission channels and lag lengths of
monetary policy are of significant importance in providing policy implications for
Vietnams policymakers. Although modelling the MTM has been popular in the
developed world, it is still new for Vietnam's policymakers, who have so far
focussed more on qualitative rather than quantitative analyses. Therefore, further
studies on the MTM are necessary for an emerging economy like Vietnam to find
evidence on the influence of shocks on macroeconomic activities.
Shocks in economics are defined as exogenous impulses in the economy, and
researchers usually study them using impulse response functions to explain how
the economy reacts over time to exogenous impulses. In addition, through this
research, the issue of whether all channels affect the course of the real economy and
business-cycle dynamics will also be analysed. In addition, the study contributes to
an examination of the role of MTM channels in Vietnam in the context of dynamic

economics, considering the role of each MTM channel in different stages of the
Vietnamese economy. The world economy in general and Vietnam's economy in
particular, have gone through the Asian Financial Crisis in 1997-1998 and the 20082010 Global Financial Crisis. Thus, there have been many significant changes in
economic policies, including monetary policy. Some questions need to be answered,
such as whether financial innovation or institutional changes in recent years have
influenced the MTM, whether the role of MTM channels determined in the past
might be made consistent with the new economic conditions. Therefore, the current
study will re-examine the role of MTM channels in Vietnam to give policy
implications for Vietnams policymakers.
Third, the current research estimates structural vector autoregression (SVAR)
models to investigate monetary policy for Vietnam. Applying SVAR to an analysis
of all four MTM channels (interest rates, credit, exchange rates and stock prices) has
not been done for Vietnam. Previous studies, such as Le and Pfau (2009), used the
vector autoregression (VAR) model to study some MTM channels; Tran (2009) used
SVAR to study the role of the gold-price gap in Vietnams monetary policy. These
are detailed in Section 2.3.2 of the literature review.
1.2

RESEARCH OBJECTIVES AND QUESTIONS


The general objective of this study is to examine the role of monetary policy in

the Vietnamese economy and the role of the MTMs channels (interest rate, credit,
exchange rate and stock prices). Moreover, the study aims to examine the
characteristics of a small open economy, such as the comparison in the effects of
foreign and domestic factors, the role of the exchange rate and the influence of
foreign monetary policy, foreign output, and world prices on the domestic economy.
The study includes some specific objectives and questions as follows.
(1) Are monetary shocks associated with fluctuations of output, price and
other fluctuations in the economy?
This question is consistent with the first objective of measuring the role of
monetary policy in Vietnam. To answer this question, SVAR procedures, including
a contemporaneous matrix, impulse response functions, and variance decomposition,
are conducted. The study considers the degree to which the effects of monetary
policy shocks, determined via impulse response functions, are consistent with
4

economic theories. Variance decomposition provides results to identify whether


fluctuations in the Vietnamese economy result from monetary shocks. The effects of
monetary shocks on the economy are analysed in both the short and medium run.
In light of the analysis, this study aims to determine how monetary shocks
affect output, inflation and other economic indicators, and to uncover evidence on
how to promote economic growth by adjusting monetary policy. Another concern is
how the monetary factor influences inflation fluctuations in Vietnam; thus, this
study examines the causality between money, output and inflation.
(2) Is the interest rate channel the most important transmission channel
of monetary policy?
This question relates to the second objective to ascertain which transmission
channels are the most and least important in the case of the Vietnamese economy.
Because the interest rate channel is defined as a traditional transmission channel, reexamining its role is important. A comparison of the role of channels is mainly
conducted via SVAR techniques. Contemporaneous matrix analysis identifies
significant instantaneous effects, allowing the study to determine whether the
transmission channels result in these effects. Impulse response examines the lag length
of effects via channels to the economy, implying how fast monetary shocks affect
output, price and other economic indicators. Moreover, results of variance
decomposition provide evidence for how much each of the four channels (interest
rates, credit, exchange rates, and stock prices) explains fluctuations in output and price
level. Determining the importance of each channel and comparing the strength of
different channels could suggest a practical policy for the present as well as the future.
(3) Do foreign shocks have a more significant impact on the Vietnam
economy than domestic shocks?
Apart from examining the role of the exchange rate channel mentioned in the
second research objective, this question contributes evidence for the third objective
about interactions among variables under the model of a small open economy.
Specifically, it answers what shocks (foreign-sector or domestic-sector) have the bigger
contribution in explaining fluctuations in the domestic economy. The variance
decomposition technique helps to answer the question. Another aspect related to this
question is examined using unit root tests with structural breaks: evidence about
5

significant break dates found in unit root tests implies a relationship between the
openness of the Vietnamese economy and structural changes in the domestic economy.
(4) Does a monetary contraction positively affect Vietnams aggregate
demand components?
This question is examined to answer another aspect of the third objective: the
SVAR analysis for international transmission. This work is conducted via adding
trade variables to the base model. Theoretically, a monetary contraction (an increase
in the domestic interest rate) results in an appreciation of the home currency, and
thus stimulates the growth of imports. SVAR procedures are applied to give the
answer. Moreover, they help to identify the lag length of monetary policy shocks to
exports/imports and the period length of these effects.
In addition, this question also follows the same objective as the first; however, it
is examined using an extended SVAR model including two variables: private
investment and private consumption. SVAR techniques are repeated to answer whether
and how monetary contraction affects investment and consumption behaviour. This
work could contribute to a clearer picture about the stages of the MTM.
(5) Do puzzles appear in the model? If yes, what are they and what do
they imply?
There are some theoretical puzzles in the empirical studies: the output puzzle,
the price puzzle, the liquidity puzzle, the exchange rate puzzle and the forward
discount bias puzzle (see Section 2.3.1). Puzzles are found in the empirical literature
of both closed and open economies; therefore, identifying the puzzles could provide
more understanding about Vietnams current policy.
These questions will be answered in Chapter 6. The study uses RATS software
for unit root tests with structural breaks, and Eviews software to give results
including the optimal lag length, VAR stability check, an estimation of the
contemporaneous

matrix,

the

impulse

response

functions,

and

variance

decomposition.
1.3

RESEARCH SCOPE
This thesis focuses on identifying and assessing the effectiveness of the

transmission channels of the Vietnamese monetary policy for the period 2000 to
6

2011, which covers many important events in the domestic economy, as well as the
international economic integration of Vietnam. Characteristics of a small open
economy are discussed in designing a suitable model for Vietnam.
The sample period beginning with 2000:1 is considered for a number of
reasons. First, previous studies have focussed on the period of the financial crisis of
1997-1998; in contrast, this research will examine the time between this crisis and
the 2008-2010 crisis. The sample beginning in 2000 includes more than a decades
data, to ensure that there is enough data to run the model.
Second, the year 2000 was recorded as one of the important milestones for
reforms to the banking system. The restructuring of commercial banks was rapidly
enhanced at that time. Especially, there was a marked liberalisation in interest rates
via the adoption of a new interest rate mechanism, in which the State Bank of
Vietnam announced the base interest rate and banks adjusted their offered rates to
follow. Thus, from 2000, policy instruments (interest rates) and intermediate targets
(monetary aggregates) could be defined to help modelling Vietnams MTM.
Third, the Vietnamese stock market went into operation in July 2000, creating
a new channel for accumulating capital for the economy, as well as another MTM
channel apart from the existing interest rate, credit and exchange rate channels.
Lastly, in terms of the whole economy, the year 2000 was the time Law on the
enterprise took effect to create more opportunities to develop Vietnamese
businesses, potentially making the effect of MTM channels more significant.
1.4

STRUCTURE OF THE THESIS


The thesis includes seven chapters. This chapter gives the background,

motivation and research objectives for the study. Chapter 2 provides a literature
review of previous studies on the MTM. The chapter also includes the theoretical
framework for monetary transmission channels: the interest rate, credit, exchange
rate and asset price channels, Monetary policy shocks are transmitted via these
channels to the financial (information) market, the production sector. Responses of
investment and consumption behaviour are also observed from these transmission
channels. This chapter provides a brief review of selected empirical studies
examining small, open economies in both developed and developing countries. Such

review is useful to seek an appropriate approach in studying the Vietnamese


economy as a small open economy. Moreover, the overview of models used to
research the MTM provides suggestions to choose and develop models to study the
transmission mechanism of monetary policy in the case of Vietnam.
Chapter 3 provides a background for examining and discussing various aspects
of the Vietnamese economy, including issues related to Vietnamese institutions and
policies. This chapter briefly describes the development periods of Vietnams
economy, including a discussion of its growth and inflation concerns, its openness,
the development of enterprise sectors, and the Vietnamese domestic financial
market. Other topics in this chapter include aspects of the institution and policy, the
legal system for conducting monetary policy and the role of the Vietnamese
monetary authority (the State Bank of Vietnam), the conduct of monetary policy and
the Vietnamese monetary transmission mechanism.
Chapter 4 reviews the data set for this study and tests its stability. Results
obtained from tests described in this chapter offer an understanding of the datas
seasonality and stationary. Unlike previous studies on Vietnam, this study conducts
tests with structural breaks to identify the break dates that could affect the
Vietnamese economy. Different tests for seasonal adjustment and unit root are
implemented to give a comparison of the tests suitable for the data set of Vietnam.
Chapter 5 proposes a structural vector autoregression (SVAR) model for the
case of Vietnam. Initially, the chapter reviews the theoretical SVAR model and
input data for this model, which suggests the choice of stationary or non-stationary
data, a recursive or non-recursive model, and short run or long run restrictions. Next,
some selected benchmark models are briefly discussed. The main aim of this chapter
is to design a SVAR model based on constructing theoretical assumptions and
appropriateness for the characteristics of the Vietnamese economy. Two versions of
a basic SVAR model, with 12 variables covering foreign and domestic sectors, are
proposed. Next, two extended models are added to the basic SVAR model.
Chapter 6 uses SVAR procedures to examine the interaction among variables
in the model. Using the models proposed in Chapter 5, this chapter gives results of
the

contemporaneous

matrix,

impulse

response

functions,

and

variance

decomposition. The significant coefficients reflect the contemporaneous effects


8

between variables. The chapter examines whether results from the contemporaneous
matrix and impulse response are consistent with economic (monetary) theory.
Results from the variance decomposition illustrate what sources contribute to the
fluctuation of variables in the Vietnamese economy, especially output and inflation.
All results help to identify which channel of monetary policy is the most or least
effective. From the base model, the study extends to examine international
transmission by adding trade variables. Moreover, another extended SVAR model is
used to examine the effects of monetary policy on investment and consumption
behaviour. To ensure the robustness of the study, some adjustments, such as the lag
length, the period of study, and changes in restrictions, are considered.
Finally, Chapter 7 provides a summary of the thesis and discusses possible
policy implications. Moreover, it mentions several limitations of the study and
suggests further research directions.

CHAPTER 2
LITERATURE REVIEW

2.1 INTRODUCTION
In the literature, the monetary transmission mechanism (MTM) refers to how
monetary policy affects real economic activities. Mishkin (1995) states that
monetary policy, the process that monetary authorities conduct to control the money
supply with a view to gaining economic growth targets and price stability, has been
at the centre of macroeconomic policymaking. Loayza and Schmidt-Hebbel (2002)
highlight that central banks set their policy instruments to affect the economy via
various transmission channels; different markets and economic variables are
influenced at various levels of speed and intensity. Thus, identifying MTM channels
is very important for researchers. To understand the MTM, it is necessary to know
what it is and how it acts, its effects, factors that could affect it, and models used to
analyse it in empirical studies.
This chapter reviews the more-recent papers on the MTM and empirical
studies on Vietnams MTM to identify the gaps in the literature. It attempts to
mention the main issues relevant to the research topic. In the next section, MTM
channels and their theoretical impacts are identified. Section 2.3 mentions empirical
studies on the MTM in small, open economies, including some prominent studies on
Vietnams MTM. Section 2.4 concentrates on the important factors in the
effectiveness of MTM channels, including legal and market structure. Section 2.5
briefly describes models for analysing monetary policy. The last section summarises
the issues discussed in this chapter.
2.2 CHANNELS OF MONETARY TRANSMISSION MECHANISM
This section presents the channels of the MTM, their roles, and the
transmission mechanism in each channel.
2.2.1

Channels of the monetary transmission mechanism and their

roles
According to Mishkin (1995; 2010), monetary policy is transmitted through
four main channels: interest rates, credit, exchange rates and asset prices. There are
10

a number of different views of the MTM, of which the financial-market price and
credit are two primary views. Taylor (2002) discussed the differences between
these views 2.
The financial-market price view, identified by Taylor (1995), focuses on the
effects of monetary policy on the prices of financial assets, such as interest rates,
exchange rates, and bond prices. These prices affect firms and households spending
decisions. One of this views branches considers the role of exchange rates in the
degree of openness in an economy. This suggested the inclusion of an exchange rate
variable in existing studies on open economies, such as Cushman and Zha (1997)
and Kim and Roubini (2000).
The second view, known as the credit view, is first mentioned by Bernanke
and Gertler (1995). According to this view, changes in bank lending and other
financial institutions are defined as an alternative to financial-market price. When
analysing the credit view, researchers must pay attention to inputs, including
quantitative data about credit and enterprises cash flows.
As far as the significance of studying the MTM is concerned, it is necessary to
understand and identify an economys transmission channels. Specifically, Loayza
and Schmidt-Hebbel (2002) listed three issues with regard to MTM channels: the
best set of policy instruments, the timing (lag lengths) of policy, and the primary
restrictions in central banks decision-making processes.
Figure 2.1

illustrates

relationships

between

monetary policy rules,

transmission channels, goods markets, sectoral prices, and aggregate output and
prices. Based on monetary policy rules, monetary policy is transmitted via MTM
channels to aggregate demand, domestic and imported goods prices, and ultimately
output and prices. Necessary information from these goals is reflected back to
monetary policy rules. In general, feedback is useful to help policymakers
understand the transmission mechanism and possible reactions.

Taylor (2002) shows that there are five different views: the financial-market price view, the credit
view, staggered price setting, limited participation and rational expectations.
11

Figure 2.1: Effects of the Monetary Transmission Mechanism on the


Economy
Central Bank

Money
and Asset
markets

Goods
markets

Sectoral Prices

Aggregate
Output and
Prices

Domestic
Goods Prices

Aggregate
Output

Imported
Goods Prices

Aggregate
Prices

Monetary
and Credit
Aggregates

Market
Interest Rate
Structure

Monetary
Policy
Actions

Aggregate
Demand

Asset Prices

Exchange Rate
Monetary
Policy Rule

Monetary
Policy
Objectives

Source: Loayza and Schmidt-Hebbel (2002).

In the review of the relevant literature, Loayza and Schmidt-Hebbel (2002)


noted two stages of the transmission mechanism: the first stage includes the impact
of policy instruments on financial-market prices (lending interest rate, exchange rate
and stock price); in the second stage, these prices affect firms and households
spending decision. Afandi (2005) argued that most research has empirically
examined the first stage of the MTM, while relatively scant attention has been paid
to aspects of the second stage. This gap creates a need for further studies on the
effects of monetary policy shocks on investment and consumption behaviour.
The remainder of Section 2.2 introduces existing theoretical and empirical
studies related to each MTM channel.
12

2.2.2

The interest rate channel

In the interest rate channel, a monetary tightening (a decrease in M) results in


an increase in real interest rate (R), which causes consumption (C) and/or
investment (I) to fall. Aggregate demand (Y) and output (y) decrease, and finally
aggregate prices and inflation () decrease. This process is denoted as follows:
M => R => C (I ) => Y => y =>
This channel is known as the traditional textbook Keynesian channel. In this
mechanism, there are some noticeable relationships, including: (1) nominal and real
interest rates, (2) short-term and long-term real interest rates, and (3) aggregate
demand and output-prices (Loayza & Schmidt-Hebbel, 2002). In the first
relationship, the real interest rate, not the nominal rate, has a large influence on
consumption and investment decisions. Due to the assumption of fixed prices in the
short run, higher nominal interest rates generally lead to higher real interest rates.
This explains investment decisions where the real borrowing cost depends on the exante real interest rate, including the known nominal interest rate and the uncertain
inflation rate. However, where the assumption of sticky prices is not accepted, the
interest rate channel is still active when a decrease in the money supply will increase
real interest rates, resulting in less spending and output. Second, the term structure
of interest rates dominates the relationship between short run and long run real
interest rates. Third, the combination of a Phillips Curve with temporary nominal
price rigidities explains for the link between output-prices and aggregate demand
(Loayza & Schmidt-Hebbel, 2002, p. 4).
The role of the interest rate channel is also assessed from different viewpoints.
While Taylor (1995) concurred that the interest rate channel plays a crucial role in
transmitting monetary policy to firms and households behaviour, Bernanke and
Gertler (1995) emphasised another view, known as the credit view.
The interest rate channel is suggested as the primary channel at work in the
MTM, the centre of the IS-LM model and the development process of the New
Keynesian macroeconomic models (Clarida et al., 2000). This channel is of
significance when monetary authorities consider adjusting nominal interest rate
below zero, as in the case of Japan. In an empirical study, Claus (2011) found
evidence in New Zealand that the interest rate channel is more important than other
13

channels because its impacts in transmitting shocks to the New Zealand economy
are larger than those of the credit and exchange rate channels. Specifically, the
interest rate channels impacts on the cost of consumption, rate of return to capital
and exchange rate are highly significant.
Another view is raised by Bernanke and Gertler (1995) when reporting the
empirical evidence for the failure of the interest rate channel in the United States via
the (user) cost of capital which is introduced by Jorgenson (1963). This means that
there could be channels at work other than the interest rate channel. Consequently,
other possible MTM channels have received increasing attention (Mishkin, 2010).
Some studies were related to developing countries, such as Charoenseang and
Manakit (2007) for Thailand and Poon and Wong (2011) for China. Charoenseang
and Manakit (2007) contended that the interest rate channel becomes weaker,
whereas the credit channel is still valid in the Thai MTM. Poon and Wong (2011)
indicated that the major role of the interest rate channel in China before 2007 was
replaced by the asset price channel after the 2008 financial crisis.
2.2.3

The credit channel

The credit channel has two sub-channels: the bank lending and the balance
sheet. The transmission mechanism is stated in the bank-lending sub-channel (or the
narrow credit channel) as shown in the following schematic diagram:
M => bank deposit => bank loans => I => Y
Contractionary monetary policy leads to fewer bank deposits, and in turn to
fewer loans. Fewer companies can borrow, leading to lower investments and output.
This sub-channel arises from market failure where there is a lack of information for
borrowers and lenders in meeting together in the financial market; thus banks act as
financial intermediaries, connecting borrowers and lenders. Thus, this sub-channel
could be vital for small companies, which have limitations in issuing their bonds and
shares on the stock market.
The literature on the bank lending channel has shown that its efficiency
depends on two factors: the number of bank-dependent borrowers, and the supply of
bank loans under the influence of monetary policy (Mishkin, 2010). Such factors
seem to be a useful approach for studying the case of developing economies, where
the structure of financial market is imperfect due to the weaker role of markets other
14

than the money market, such as the stock/ bond market. The domination of an
economys banking system induces this channel to become stronger than other
channels. This is one of the characteristics worth noting in analysing the credit
channel in developing economies.
The balance-sheet sub-channel, known as the broad credit channel, uses the
net worth of companies. This sub-channel is related to problems of adverse selection
in financial markets and moral hazard. Monetary policy can influence companies
balance sheets through the following ways:
(1) M => equity prices (Pe) => adverse selection and moral hazard =>
lending => I => Y
(2) M => R => cash flow => adverse selection and moral hazard =>
lending => I => Y
(3) M => Pe => financial assets => likelihood of financial distress =>
consumer durable and housing expenditure => Y
Bernanke and Gertler (1995) found evidence for the response of the real
economy to monetary policy shocks via the credit channel in the United States. An
explanation is that central banks decisions about interest rates generally depend on
the cost of available credit. The balance-sheet sub-channel describes the effects of
monetary changes on borrowers balance sheets and income statements. The banklending sub-channel shows that changes in monetary policies affect the loan supply
of depository agents. The authors concluded that the existence of the first subchannel is well established while there has been controversy about the second.
Monetary tightening results in sustained decreases in GDP and price level with a
lag; the biggest reduction has been in residential investment and fixed business
investment.
Wilson (2011) reported that the credit channel has been found to be a major
channel in the U.S. and other countries. Mishkin (2010) pointed to some reasons for
the important role of the credit channel. These include this channels effects on
companies employment and spending decisions, small firms credit-constrained
status, and the use of information from credit-market imperfections to explain other
economic phenomena.

15

To examine the impacts of monetary policy shocks in Thailand via the credit
channel, Kubo (2008) used a SVAR model with monthly data3, and found that this
channel plays an important role in the conduct of monetary policy in Thailand.
Moreover, through credit channels, the shocks negatively affect economic agents
with large credit loans.
Sharifi-Renani (2010) used a SVAR model to examine the role of MTM
channels (exchange rate, credit and asset price) in Iran. The findings highlighted that
the credit channel is dominant, and that monitoring the credit channel contributes to
the success of conducting monetary policy in Iran.
Other studies have downplayed the importance of the credit channel in
transmitting monetary shocks. For example, an empirical study on Australia by
Suzuki (2004) showed that the credit channel cannot be an effective transmission
channel when banks borrow overseas via selling securities to offset tight monetary
policy. Moreover, Black et al. (2010) could not find any evidence that a decrease in
U.S. banks mortgage lending after monetary policy tightening as these banks did
not supply credit for subprime borrowers or did not depend on retail deposits.
2.2.4

The exchange rate channel

The exchange rate channel transmits monetary policy as follows:


M => R => the real exchange rate (E) => net exports (NX) => Y
An increase in the value of domestic currency is denoted by E. This model
illustrates that when monetary policy leads to an increase in the real exchange rate,
the effects are transmitted to net exports and output. Moreover, this channel is
related to interest rate effects. Specifically, a tightening in the money supply from
monetary authorities causes an increase in the domestic real interest rate (R), which
makes domestic-currency-denominated assets more attractive than foreign-currencydenominated assets. For small, open economies with flexible exchange rates, this
channel is particularly important because it influences both aggregate demand and
aggregate supply (Dennis, 2003). In the case of small, open economies with fixed
exchange rates, this channel explains that the domestic interest rate must be adjusted

Variables include the consumer price index, the industrial production index, the producer price
index, the inter-bank overnight lending rate, and private credit aggregates.
16

to equal the world interest rate. If the exchange rate is fixed, policymakers can adjust
the real exchange rate by affecting the price level (Kamin et al., 1998).
The international IS-LM type model known as the Mundell-Fleming model
analyses policies under flexible and fixed exchange rate. This model is after two
studies of Mundell (1963) and Fleming (1962). This model is an extension to the
open economy of the IS-LM model, a Keynesian model of business cycle is
extended with international trade. According to the Mundell-Fleming model, one of
the important relationships is a positive link between the real interest rate and the
real exchange rate. Under a flexible exchange rate regime, monetary contraction
leads to a decrease in output and interest rates are at international levels. Under a
fixed exchange rate regime, effects of domestic monetary shocks are only
temporary, not in the long run. (Blanchard & Sheen, 2013)
2.2.5

The asset price channel

The schematic diagram below represents the asset price transmission channel:
M => Pe => q => I => Y
This channel is based on Tobins q theory (1969) to explain how monetary
policy affects the economy by considering the impacts on the valuation of firms. The
value q denotes the ratio of the market value of firms to the replacement cost of their
capital. The basic idea of this channel is that a contractionary monetary policy
results in lower stock prices (lower market value), and, in turn, a lower q, because
there are more attractive opportunities offered by an increasing market interest rate,
and firms can sell stocks for less money to spend. In other words, a lower q leads to
less investment. Moreover, Jorgenson (1963) introduced the user cost of capital
which is the sum of the real interest rate and the depreciation rate and its relationship
with the optimal size of the fixed capital stock. Changes in the optimal stock lead to
investment adjustments. The user cost of capital is included in estimating the
neoclassical investment function and the higher the real interest rate, the higher the
user cost, the lower the level of investment (Blanchard & Sheen, 2013, p. 378).
According to Mishkin (1995), another effect of asset price channels is wealth
effects on consumption which originates from the life-cycle hypothesis of Modiglina
and Brumberg (1954), Ando and Modigliani (1963). This hypothesis implies that
consumption is a function of lifetime resources of consumers, so a monetary
17

contraction affects stock prices, implying effects on consumers wealth, leading to a


decrease in consumption and finally, a decrease in aggregate demand. The life-cycle
hypothesis of Modiglina and Brumberg (1954), Ando and Modigliani (1963) and the
permanent income hypothesis of Friedman (1957) have the most similar assumption
that individuals utility is maximised by the balance of earnings and consumption.
Consumption is an important component of income, so a decrease in consumption
leads to a decrease in income, resulting in a decrease in aggregate demand.
Changes in asset price could result in financial fluctuations in different agents
in the economy, including banks, businesses, and households. These fluctuations, in
turn, threaten price and economic stability. The role of the stock price channel has
been shown to be more important than other channels, in line with the improvement
of capital markets in developing countries (Poon & Wong, 2011). Li et al. (2010)
found evidence that differences in stock-market responses to a monetary contraction
depend on differences in the openness of the particular financial market.
Specifically, the response of the U.S. stock market is determined to be larger than
that of Canada.
In summary, the literature discussed in Sections 2.2.1 to 2.2.5 reveals some
notable characteristics of the MTM. First, monetary policy plays a crucial role in
developing the economy. The findings of many studies indicate that business-cycle
dynamics are affected by MTM channels. Therefore, studying the MTM is expected
to suggest valuable policy implications.
Second, different channels have been identified as being the most important in
diverse countries. Studies have emphasised that comparing the impulse responses of
each channel helps to determine the role of MTM channels. The credit channel is
highlighted as an important channel in many economies, including developed and
developing, but not all economies. Other channels are regarded as more important in
some developing countries. Wilson (2011) noted the difficulty in disentangling the
MTM channels as a reason for such varied findings.
Third, at different times, the role of each channel in an economy is also
changing, as it happened in China after the 2008 financial crisis (Poon & Wong,
2011) or Thailands MTM under different periods of monetary policy (Kubo, 2008).

18

This is due to structural changes as well as financial innovations (Kubo, 2008; Poon
& Wong, 2011).
2.3

EMPIRICAL STUDIES ON THE MONETARY TRANSMISSION

MECHANISM IN SMALL, OPEN ECONOMIES


2.3.1

Some selected studies examining small, open economies

This section will present some selected studies that help to suggest different
approaches for studying the MTM in small, open economies. These studies are
Cushman and Zha (1997) for Canada, Kim and Roubini (2000) for non-US G-7
countries, Dennis (2003) for Australia, Afandi (2005) for Indonesia, Aslanidi (2007)
for Georgia, Raghavan and Silvapulle (2008) for Malaysia and Kubo (2008) for
Thailand. The selection of studies over several time frames offers a broader
perspective on small, open economies.
To examine the Canadian monetary policy, Cushman and Zha (1997)
constructed a SVAR model that included foreign variables (that is, variables from
the U.S.). This approach was based on the view that a small, open economy could
not greatly affect the world market. Moreover, they included trade related variables
(exports and imports) in their model, so the effects of monetary policy on trade
activities (via the exchange rate channel) were captured. They found a decrease in
exports and an increase in imports due to the appreciation of the home currency after
the monetary contraction (the monetary aggregate decreased and the interest rate
increased); however, the study also recorded a fall in imports in the initial period
before their expected increase.
In their study of non-US countries, Kim and Roubini (2000) used the U.S.
interest rate and the world oil price as foreign variables to control their shocks,
emphasising that world oil prices had a contemporaneous effect on the monetary
policy reactions. They highlighted a number of puzzles occurring in empirical
studies, such as the liquidity puzzle, the price puzzle and the exchange rate puzzle
(Table 2.1). Their results are consistent with the condition of uncovered interest
parity as the domestic currency depreciates a few months after its appreciation due
to monetary contraction.

19

Table 2.1: Common Puzzles in Empirical Studies


Puzzle
Output puzzle

Description
A positive innovation in interest rates leads to an increase in
output, rather than a decrease.

Liquidity puzzle

A positive innovation in monetary aggregates results in an


increase, rather than a decrease, in nominal interest rate.

Price puzzle

A positive innovation in interest rates leads to an increase in


price, rather than a decrease.

Exchange rate

A positive innovation in interest rates leads to a depreciation

puzzle

of home currency, rather than an appreciation.

Forward discount

A positive domestic interest rate innovation relative to the

puzzle

foreign interest rates results in a persistent appreciation of the


home currency, rather than a persistent depreciation over time
after the appreciation under uncovered interest parity holds.

Source: Authors summary from Kim and Roubini (2000).

According to Dennis (2003), an important issue in the literature about small


open economies is whether changes in the exchange rate are considered in
making monetary policy. Dennis (2003) found that micro-founded sticky-price
models applied in studies seem to have little or no evidence that the
policymaking process responds to exchange rate changes. By estimating a small,
open model for the Australian economy with two sticky-price models in which
the terms of trade are added, Dennis (2003) stated that optimal monetary policy
rules should take the changes in the terms of trade into account and, in particular,
the real exchange rate. In an economy where inflation targeting is applied, like
Australia, considering the real value of the exchange rate should be significantly
emphasised (Dennis, 2003).
In the case of Indonesia, Afandi (2005) based his work on Kim and Roubinis
(2000) model of small, open economies to construct a SVAR model for the
Indonesia economy using two external variables the world oil price and the U.S.
Federal Fund rate to isolate exogenous shocks. His study also examined the role of
each MTM channel in Indonesia, with structural breaks like the Asian financial
20

crisis in 1998. The findings showed that the effect of the exchange rate channel is
not strong, but that there is a link between shocks of monetary tightening and the
exchange rate depreciation. Moreover, Afandi (2005) emphasised that exchange rate
depreciation affects an economys foreign debts, thus worsening firms balance
sheets; thus, this financial variable greatly affects businesses.
In a study on Georgian monetary policy, Aslanidi (2007) chose the U.S.
interest rate and the Russian output in the SVAR model for examining external
dimensions, and compared the influence on the economy from the U.S. interest rate
shocks and domestic monetary policy shocks. Using the U.S. interest rate as the
surrogate of foreign monetary policy is similar to the approach of Kim and Roubini
(2000) and Afandi (2005). Findings in Aslanidis study illustrate that an exchange
rate shock had a stronger impact than an interest rate shock on the economy of
Georgia. Furthermore, the author found that the impacts of U.S. interest rate shocks
were as insignificant as that of the domestic interest rate shocks on this economy.
In their examination of the monetary policy framework in Malaysia before and
after the 1997 financial crisis, Raghavan and Silvapulle (2008) set up four variables
for a foreign sector in their SVAR model: the world commodity price index and
three U.S. variables (industrial production, the consumer price index, and the
Federal Fund rate). Also, Raghavan and Silvapulle (2008) emphasised that using
U.S. variables as a proxy of foreign variables is fairly popular in the literature on the
MTMs of small, open economies; for example, in the studies of Cushman and Zha
(1997), Dungey and Pagan (2000). With their model, Raghavan and Silvapulle
(2008) aimed to examine the economys responses not only to domestic shocks but
also to foreign shocks. They came to the conclusion that the impacts of foreign
shocks on monetary policy in the post-crisis period were stronger than those in the
pre-crisis period.
Kubo (2008) analysed the relationship between shocks and the trade related
variables (nominal trade balance, the volume of exports and imports, and the terms
of trade) in his SVAR model for Thailand. The purpose of this approach was to
examine the interaction between monetary shocks and export/import changes. To
examine this, these variables were treated as not being able to affect the monetary
policy contemporaneously. The results of this study points toward the idea that the
21

shocks of monetary tightening caused Thailands import demand to decrease


although there was a decrease in import prices.
In summary, the review of the literature for small open economies shows a
number of research approaches. First, the main difference between open economies
and closed economies in most studies is the appearance of the exchange rate
channel, so examining this monetary transmission channel is necessary for studying
small open economies. Second, trade related variables should be used to examine the
international transmission of monetary policy. This helps to develop the scope of
empirical studies about the MTM. Third, many studies on small, open economies are
of the view that foreign variables, such as foreign interest rate, foreign output, and
world price, are used to examine external shocks to the domestic economy. This
approach results from the assumption that international economic development will
affect small, open economies, so it is useful to study the possible impact of
international economic fluctuations on these economies. Foreign interest rates and
world prices are used to control for changes in foreign monetary policy and
international prices in setting domestic monetary policy. Considering foreign output,
based on the assumption that external shocks from the main trade partners could
affect the domestic economy, is an additional approach for studying small, open
economies. This review of the literature suggests that studying the MTM in small,
open economies should, first, examine the effects of foreign shocks (foreign
monetary policy, world prices, foreign output) on the domestic economy, and
compare the impacts of domestic (monetary) shocks and foreign (monetary) shocks
on the domestic economy. Second, the study can aim to identify the effects of
monetary shocks on trade related variables.
2.3.2

Empirical studies on the monetary transmission mechanism of

Vietnam
There are only a few known quantitative studies on the MTM in the case of
Vietnam, such as Le and Pfau (2009), Tran (2009) and Nguyen (2010). Although
Vietnams MTM is mentioned in some international and domestic papers,
quantitative research is limited.
The VAR study of Le and Pfau (2009) uses quarterly, seasonally adjusted
observations from 1996Q1 to 2005Q4, and nine variables (real industrial output,
22

CPI, broad money, real lending rate, domestic credit, index of the real effective
exchange rate, world oil price, rice price and U.S. Federal Funds rate) to examine
the effect of monetary policy on the economy via three channels of the MTM
(interest rate, credit, and exchange rate). To examine the openness of Vietnams
economy, Le and Pfau (2009) used the world oil price, world rice price, and Federal
Fund rate as exogenous variables. Le and Pfau found that: (1) while monetary policy
in Vietnam affects real output, there was no relationship between money and
inflation; (2) in Vietnam, the credit and exchange rate channels were more
significant than the interest rate channel; and (3) there was a lag effect, with the
strongest influence of monetary policy occurring after four quarters.
Another study by Tran (2009) used SVAR to study Vietnams monetary rules
from 1992 to 2002. Variables used in this model included monthly growth of a
seasonally adjusted monetary aggregate (M1 or M2), monthly growth of seasonally
adjusted CPI, monthly growth of seasonally adjusted official exchange rate, gap
between seasonally adjusted industrial output and industrial output trend derived
from the seasonal adjustment programmes, and percentage gap between domestic
gold price and world gold price. The author focussed on examining the domesticinternational gold price gap, and showed that the State Bank of Vietnam was using a
McCallum-type rule rather than a Taylor interest rate rule in adjusting monetary
growth to respond to fluctuations in price and exchange rate. This study confirmed
that the gold price gap is regarded in Vietnam as a valuable indicator for the
successful implementation of monetary and exchange rate policy.
In a domestic published quantitative study, Nguyen (2010) used a SVAR
model to examine Vietnams MTM from 1998 to 2009. This model consists of nine
monthly variables: the world price index, U.S. industrial output, U.S. CPI, Fed rate,
domestic industrial output, CPI, broad money M2, short run interest rate, and
exchange rate between U.S. dollar (USD) and Vietnams currency (VND). The
SVAR model was established in the recursive form, based on the methodology of
Raghavan and Silvapulle (2008). Nguyen (2010) concluded that broad money (M2)
played a positive role; however, this role was small, while M2 and credit growth in
this period were quite high. The possible reason for this, as illustrated by Nguyen
(2010), is the misapplication of loans for securities and real-estate investment, while

23

these loans should be granted for manufacturing. In addition, exchange rate


fluctuations were affected mainly by interest rates and M2 changes, and a lesser
degree by external shocks. Another finding was that international fluctuations had a
significant effect on the domestic banking system. Finally, Nguyen (2010) showed
that the lag length for an effective impact from adjusting the interest rate was from
one quarter to two quarters, and the greatest impact of monetary policy on CPI was
from three to four quarters.
Although the above studies are valuable, they still contain some limitations.
First, essential information is not sufficient in Vietnams statistical system. The
approach using the industrial output as a proxy for GDP (common in studies of other
countries) could not fully reflect the effects of the MTM on the Vietnamese
economy, which is still largely agricultural. Although the industrial and construction
sector growth have remained in the double digits since 1992, except for the crisis
periods in 1998-1999 and 2008-2010, the weight of these two sectors was between
36 percent and 41 percent of GDP in the period of 2000-2010. Using the industrial
output as a proxy of GDP is explained from the fact that the GDP database before
2000 is not in quarterly form (Le & Pfau, 2009). This result suggests that neither
money nor output Granger causes inflation, which seems to be different from the
assertion in monetary theory that inflation is always a monetary phenomenon.
Vuong and Tran (2009) argue that using the industrial-output-proxy approach
significantly reduces the results trustworthiness.
Second, previous studies focussed on money market, without examining the
link between the money market and other markets, such as the stock market and
real-estate market (the asset price channel). The reason commonly cited is that the
Vietnam stock-market channel was only established in 2000, and is pressured by
domestic investors speculation (Le & Pfau, 2009). The real estate market is in a
similar situation. However, this approach has kept researchers from developing a
clear picture of different MTM channels, which could lead to changes in the role of
each channel. An example is Poon and Wong (2011), who confirmed that the main
transmission channel changed from the interest rate before the 2008 financial
tsunami to the asset price channel thereafter. This thesis considers the asset price
channel, focussing particularly on the stock price channel, because the real-estate

24

markets statistics are available only as annual data and does not meet our research
requirement.
Third, models used in the previous studies do not reflect the structural factors
or do not connect the macro and micro approaches in studying the MTM.
Specifically, the previous empirical studies on Vietnam did not significantly
mention using a SVAR model. Le and Pfau (2009) used a VAR model to examine
the MTM in Vietnam. In the literature, the VAR model is the simultaneous
equations model in which the explanation of each endogenous variable is based on
its lagged values and/or other variables lagged values (Gujarati & Porter, 2009).
Although Tran (2009) applied the SVAR model, the study focussed on shocks to the
gold-price gap. In addition, Trans (2009a) model, based on the SVAR model of
Pagan (1995), was a small SVAR model with only four variables: consumer price
index, exchange rate, gold-price gap and growth of monetary aggregate (M1 or M2).
This model could not fully assess the monetary and external shocks to Vietnams
economy. A better SVAR model with external and domestic variables was used by
Nguyen (2010), but the author applied the recursive structure proposed in Raghavan
and Silvapulle (2008). This is an obvious limitation since using the non-recursive
form is more useful in analysing structural shocks (Kim & Roubini, 2000).
Specifically, non-recursive approaches help to identify restrictions in different
equations with contemporaneous structural relationship of parameters (Sims, 1986).
Therefore, comparing results from both approaches (recursive and non-recursive)
could give more meaningful insights.
Fourth, researching Vietnam as a small, open economy is important and is
mentioned to a degree in the previous studies, but a study covering more aspects of
small, open economies would be more useful. Le and Pfau (2009) used exogenous
variables like world oil price, rice price, and Federal Funds rate as external shocks
due to the openness of Vietnam. The same approach was also used by Nguyen
(2010). However, other aspects are not included in other studies of Vietnam. Kubos
(2008) study on the effects of monetary shocks on trade related variables and the
work of Aslanidi (2007) in controlling foreign output could be examples for further
study on Vietnam. The current research will focus on the effects on the Vietnamese
economy, rather than the effect of this economy on international markets. This

25

corresponds to the explanation of small, open economies mentioned in Section 1.4,


giving any policy implications relevance to a small, open economy like Vietnam. In
addition, Mankiw (2010) emphasised that due to small, open economies minor
international role, the domestic interest rate is determined by the world interest rate.
Because of this characteristic, whether an external or domestic monetary shock has
more significant effects on the economy needs to be an essential research concern.
Finally, previous studies such as Le and Pfau (2009), Tran (2009), Vuong and
Tran (2009), and Nguyen (2010) mentioned Vietnamese monetary policy in the
1997-1999 Asian crisis, but did not necessarily represent the MTM of Vietnam in
the period of the 2008-2010 financial tsunami. Therefore, further studies comparing
Vietnams MTM before and after the recent financial tsunami could be both
interesting and meaningful.
These limitations confirm the necessity for doing further research on the MTM
channels of Vietnam. The study contributes to addressing the weaknesses in
previous studies.
2.4

STRUCTURAL FACTORS AND MONETARY TRANSMISSION

MECHANISM
There is a consensus that changes in monetary policy affect the economy. The
above literature illustrates that different economies react to shocks in different ways.
The question is to determine the factors leading to such differences. Loayza and
Schmidt-Hebbel (2002) confirm that particular features of an economy significantly
affect the transmission of channels and the effectiveness of monetary policy. In the
first stage of the MTM, factors including the financial structure, size, and openness
scale of the economy affect the transmission process from policy instrument to
financial-market prices. In the next stage, factors that affect the MTM are financial
development and firms and households balance-sheet position. The possible effects
of some selected factors are summarised in Table 2.2.

26

Table 2.2: Influence of an Economys Features on the Monetary


Transmission Mechanism
Characteristics

Possible effects

Structure - The diversification of financial Quicker and closer transmission from


of the

institutions.

monetary policy changes to market interest

financial

- The diversification of financial rates and other financial prices.

system

products.
- Monopoly power of a few The financial institutions independence
financial institutions (primarily from the central bank in determining
banks)

financial-market prices.

- The poor supply of financial


alternatives.
- The financial constraints to Firms and households are less responsive to
firms and households.
-

Shallow

and

financial-market prices.
poor - The small importance of the asset price

diversification of the financial channel (the low capitalisation of the


system (dependent on a few securities market).
banks)

- A weak interest rate channel (a few banks


have monopoly power).
- The large influence of the credit channel.

- Financially underdeveloped economies

Foreign

exchange

transactions

are

controlled and the exchange rate channel is


ineffective.

- Deeper and more diversified - The asset price, interest rate and exchange
financial system.

rate channels are more important.

Size and

- The exchange rate channel has an

openness

important role.

of the

- Monetary policy can be determined by

economy

domestic interest rates.

Source: Loayza and Schmidt-Hebbel (2002)


In their 2002 study, Loayza and Schmidt-Hebbel focus only on analysing the
factors related to the first stage of the MTM, including the structure of the financial
system and the economys size and openness while the factors relevant to the second
27

stage (financial development and firms and households balance-sheet position) are
suggested as the issues that need to be studied. This is consistent with a theme in the
literature identified by Afandi (2005) that studying the MTMs first stage (the focus
of most empirical studies) needs to be supplemented with an examination of the
second stage.
Some empirical studies give attention to these factors. Cecchetti (1999)
examined the relationship between legal structure, financial structure and the MTM
in countries in the new Eurosystem. The study provided evidence that different
responses in the MTM were mainly explained by differences in financial structure.
Specifically, Cecchetti (1999) highlighted such structural factors as the banking
system (size, concentration and health) and the financial-market structure, including
the direct- and indirect-financing markets. This view is supported by Kamin et al.
(1998), who mentioned the structure of the economy (such as changes relevant to
financial institutions and their technology development) and firms balance-sheet
positions as the important drivers for monetary policy.
Next, Cecchetti and Krause (2001) conducted a study on the link between
financial structure, macroeconomic stability and monetary policy using data from
the 1980s and 1990s for 23 countries, including developed and emerging economies,
They found evidence that the structure of the banking system and the financial
markets dominate the transmission of the interest rate channel to domestic output
and prices. Moreover, a decrease in the ownership ratio of the state in the banking
system contributes to improved monetary policy efficiency and the macroeconomys stable development. Cecchetti and Krause (2001) explained that if the
government controls almost all the banking system, there is little room for the
central banks monetary policy, and the policies response to market fluctuation is
less effective.
However, Elbourne and de Haan (2006) argue, based on their findings in 10
EU members in the area of Central and Eastern Europe4, that there is little evidence
for a close relationship between financial-structure changes and monetary policy in
these countries. This difference could be explained by using different indicators to
measure the development of the financial system. In particular, Elbourne and de

28

Haan (2006) used three separate indicators the importance of small banks in a
countrys the financial system, the health of the banking system, and the importance
of external finance sources5 whereas Cecchetti (1999) used a combined indicator
of the financial structure. Elbourne and de Haan (2006) ranked their indicators in
increasing order, combining the Kendal and Spearman rank correlations and
recording the p-values for every statistic.
Another interesting aspect about this topic mentioned by Cecchetti (1999) is
the link between the legal system and the financial structure. The difference in each
countrys financial structure results from the separate characteristics of its legal
structure. If the close link between the financial structure and the MTM is accepted,
the relationship between the legal system and the financial structure is significantly
meaningful for explaining many economic issues when studying the MTM,
especially for emerging developing countries.
2.5

MODELS IN ANALYSING MONETARY POLICY


Mishkin (2010) points out that there has been a debate for 70 years on the role

of monetary policy in fluctuations in the economy, resulting in two schools: the


monetarist school, following Milton Friedman, and the Keynesian school, following
John Maynard Keynes. In the former, reduced-form models are used to test the role
of money on economic activities. In the latter, economists attempt to expand the
understanding of the transmission channels by which monetary policy affects
aggregate demand by focussing on structural models.
Many authors use the technique of vector autoregression, a reduced-form
model, known as a VAR model, to examine the MTM. The use of a VAR model was
first proposed by Sim in the 1980s. Subsequently, VAR modelling was extended and
refined by other researchers, such as Johansen (1988), Johansen and Juselius (1990),
and Bernanke and Gertler (1995). To examine the U.S. economy, Sims (1986)
4

The 10 EU members included Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania,
Poland, Romania, the Slovak Republic, and Slovenia.
5
The indicators about the size and concentration of the banking sector included: asset share of
five largest banks (%), loans share of five largest banks (%), deposits share of five largest banks (%),
number of banks, share of foreign banks, and banks per million people. The indicators about bankingsector health included: return on asset (%), non-performing loans (% total assets), non-performing
loans (% loans), EBRD indicator for bank reform, net interest margin (% total assets), and average
capital ratio. The indicators about the importance of external and bank finance included: number of
publicly traded firms, publicly traded firms per million capita, stock market capitalisation (% GDP),
and domestic credit of banks (% GDP).
29

proposed a VAR model with six variables (real GNP, real business fixed investment,
the GNP price deflator, the M1 measure of money, unemployment, and the
Treasury-bill rates). Bernanke and Gertler (1995) used the VAR technique, in which
a set of variables is regressed on lagged values of this variable and the other
variables, to study the U.S. monetary policy. Their work illustrated the convenience
of VAR models in reviewing the dynamic relationship of variables because the
response of a variable reflects both its own disturbance and the disturbance to other
variables. In the case of developing countries, there are studies using VAR models,
such as those of Charoenseang and Manakit (2007) on Thailand, Le and Pfau (2009)
on Vietnam, and Poon and Wong (2011) on China. The ordering in the VAR model
is based on the assumptions about the effects of money-policy variable to
macroeconomic indicators. However, as mentioned above, this method has certain
limitations and it does not help researchers obtain non-recursive orthogonalisation of
the error terms in analysing impulse response, which is conducted when using
SVAR models.
There are many studies using structural VAR (SVAR) models to overcome the
main pitfalls in applying VAR models. Gottschalk (2001) introduced the SVAR
methodology as a useful tool in analysing the dynamics of models to understand the
effects of unexpected shocks. The SVAR model, which is based on the VAR, seems
to be better for examining the structural shocks in testing multivariate time-series
models (Bhattacharyya & Sensarma, 2007; Elbourne & de Haan, 2009). Compared
to VAR models, SVAR models are useful to identify parameters from the model and
recover structural shocks as they require economic theory in analysing simultaneous
interaction of variables (Aslanidi, 2007).
The use of SVAR models is mentioned in some recent studies on small, open
developing economies. These include Afandis (2005) study on MTM channels in
Indonesia after the financial crisis in 1998, Kubos (2008) study on the impacts of
monetary policy shocks in Thailand via the credit channel from 2000 to 2006, Trans
(2009b) on Vietnams monetary rules from 1992 to 2002, and Sharifi-Renanis
(2010) on monetary policy in Iran from 1989 to 2009. However, except for Afandi
(2005), the other studies only consider one of two structural approaches for their
SVAR model: recursive or non-recursive. Moreover, they do not take into account

30

the second stage of the MTM process, the transmission of financial-market prices
into firms and households spending behaviour. Obviously, this affects the scale of
mentioned models. Specifically, this approach results in a smaller number of
variables and equations related to consumption and investment.
In addition to the above models, other methods such as a vector error
correction model (VECM) or a dynamic stochastic general equilibrium (DSGE) are
also proposed in some studies to analyse monetary policy. A VECM method adds
error correction features to a VAR model. This model includes the differenced
equations and an error correction term for the deviation of previous period variables.
VECM is appropriate for variables that may be cointegrated (Gujarati & Porter,
2009). VECM is used to examine the MTM in latest reseach, such as Oliver et al.
(2004) for Germany, Mello and Pisu (2010) for Brazil, Gambacorta (2011) for the
U.S., Hespeler (2013) for Uzbekistan, Waal and Eyden (2014) for South Africa. A
DSGE method is another approach in which economic phenomenon and behaviour
are explained by the models developed from microeconomic foundation. Agents
decisions are derived from their own views and expectations about the future
economy. Del Negro and Schorfheide (2004) demonstrated that in working with
unprocessed factual statistics which has not been affected by removing the trend or
filtering or regression, unrestricted multivariate models including VAR (SVAR)
models seem to outperform DSGE models; however these authors confirm DSGEs
role as a mechanism to support VARs estimation. Sbordone et al. (2010) and Tovar
(2009) emphasise that DSGE models play a crucial role in formulating monetary
policy in many central banks. This approach should not be vulnerable to the Lucas
critique (1976) which asserts that examining the effects of a policy shock based on
observed historical data will be nave. According to Ireland (2006), based in the
context of DSGE models, recent theoretical researches on the MTM are conducted to
examine the traditional Keynesian interest rate channel. The basic New Keynesian
model related to three variables: output, inflation and short term interest rate (Ireland,
2006; Sbordone et al., 2010).
In my thesis, SVAR models are chosen because of their benefit in analysing
the dynamics of a model via subjecting it to an unexpected shock. The difference
between VAR/SVAR and VECM is that while using VAR/SVAR helps to decide if

31

shocks have permanent or temporary effects, using VECM means the effects of
shocks are permanent (Raghavan & Silvapulle, 2008). SVAR models are enough for
examining how domestic and foreign shocks affect domestic macroeconomic
variables of interests, so using DSGE models should be explored for further studies
on Vietnam.
2.6

CONCLUDING REMARKS
This chapter gives a brief understanding about channels of the monetary

transmission mechanism (interest rate, credit, exchange rate and asset price), the
empirical studies of this topic, the effects of structural factors on the transmission
mechanism and the models used to analyse this mechanism. Previous empirical
studies suggest some notable concerns in examining the role of the MTM and the
interaction of variables in the economy.
First, most studies on small, open economies have focussed on developed
countries, or on developing countries with a relatively developed financial market;
however, there is a lack of research on the MTM in small, open economies with a
developing financial market and a low independence of monetary policy. Thus,
studying these cases is necessary for not only theoretical analysis but also practical
policy implications. Under this approach, Vietnam could be a suitable case for
research.
Next, most empirical studies focus on the first stage of the MTM the
transmission of policy instrument to the financial-market prices or the overall
consideration of the response of the production sector (output and price level) to a
monetary policy shockwithout paying attention to the second stage in which shocks
are transmitted to households and firms spending. This preference is explained by
the fact that studying the transmission from the money market to the information
(financial) market (the first stage) or the production sector helps economists and
policymakers identify a set of policy instruments, as well as the lag length of such
instruments on financial-market variables. Thus, investment and consumption
behaviour are usually neglected in most empirical studies. Furthermore, the
international transmission of monetary policy, via examining the response of
international trade activities of a small open economy to a monetary shock, is not
widely studied.
32

Third, although there has been a debate on whether and how structural factors,
specifically the financial structure, affect the MTM, no one side seems to dominate
in the literature. Moreover, this helps to explain many different economic issues in
emerging developing countries, so considering this aspect is useful in studying their
MTM.
Fourth, few quantitative empirical studies have been conducted on the MTM
in Vietnam and those suffer from serious limitations. This further strengthens the
case for studying this topic in Vietnam.
Finally, the SVAR model is better than the VAR model for analysis of
contemporaneous responses between economic variables in the non-recursive multiequation model. The SVAR model could help in studying non-recursive
orthogonalisation for impulse response analysis as well as identifying short run and
long run restrictions.

33

CHAPTER 3
THE VIETNAMESE ECONOMY, INSTITUTION AND POLICY

3.1

INTRODUCTION
Vietnam was unified after ending the war in 1975, and the Vietnamese

economy as a whole began to develop. Since then, Vietnams economy has


witnessed different development periods: pre-1986, the Economic Renovation in
1986 (known as the name Doi Moi), and post-Doi Moi. This third stage was affected
by two financial crises: 1997-1998 and 2008-2010. The reform process in Vietnam,
with the implementation of a series of five-year socio-economic development plans,
helped the economy move from a centrally planned to a market economy. Vietnam
advanced from the list of the worlds low-income countries and to become a middleincome economy in 2011. Vietnams economic integration achieved significant
goals, especially joining in Association of Southeast Asian Nations (ASEAN),
ASEAN Free Trade Area (AFTA) in 1995, normalising the relations between
Vietnam and the United States in 1995, and joining the World Trade Organization
(WTO) in 2007.
The 25 years of Doi Moi has seen remarkable changes in the economic
policies in Vietnam. The international community has expressed approval of these
changes, as documented in the International Monetary Funds announcement on
Vietnam, which stated, Vietnam is one of the fastest growing and most dynamic
economies in Asia (Ishii, 2007), and by (Miyazaki, 2010), who wrote, Vietnams
transition to the market economy in the past 20 years has been truly remarkable.
High economic growth and poverty eradication became the leading targets in
Vietnams socio-economic development plans. However, over time, Vietnams
economy has shown weaknesses, such as unsustainable economic growth, inflation
threats, and the economys dependence on imports, weak financial markets, and low
competitiveness of domestic enterprises.
Institutions, which play a crucial role in the development of society, have been
identified as the foundation of social life (Campbell, 2004). This means that
institutional changes will result in changes in a societys economic and political

34

environment. Tang (2011) uses Norths (1994, p. 360) definition of institutions as


the humanly devised constraints that structure human interaction. They are made up
of formal constraints (rules, laws, constitutions), informal constraints (norms of
behaviour, conventions, and self-imposed codes of conduct), and their enforcement
characteristics. This definition characterises institutions as being shaped by two
kinds of rules: formal, such as constitutions, laws, and regimes, and informal, such
as norms and conventions. Similarly, the World Bank has written Institutions are
not buildings or organizations, they are the rules by which citizens, firms and the
state interact (World Bank, 2010, p. i). Policy is defined as rules that help to guide
decisions, or systems of formal rules such as law. In this study, economic policies
relate to the governments activities in the field of macroeconomics. In the history of
the Vietnamese economy, there have been many changes in institutions and policies.
This chapter outlines the main economic institutions and policies for a market
economy in general, and for conducting monetary policy, which is in line with the
aim of this research.
This chapter provides the background for the model of the Vietnamese
economy developed in the next chapters. Section 3.2 briefly outlines the economic
history of Vietnam. Next, different issues of the economy are discussed in Sections
from 3.3 to 3.6, including economic growth and inflation, the openness of the
Vietnamese economy, enterprise sectors in Vietnams economy, and Vietnams
domestic financial market. Section 3.7 raises the issues about institutions and
policies that support a market economy in Vietnam. Section 3.8 gives an overview
of the legal system for conducting monetary policy. Sections 3.9 discusses the
conduct of monetary policy and its transmission mechanism. The last section is a
summary of the discussion.
3.2

VIETNAMS ECONOMIC HISTORY FROM 1975 TO 2011


This section gives a brief summary about Vietnamese economy from 1975 to

the present. Changes are outlined in detail in Table A1, Appendix A.


3.2.1

Period from 1975 till 1986

The year 1975 was a major landmark in the history of Vietnam. With the
achievement of national unity, the whole country entered a period of building a
socialist state. In 1975, a new currency was issued in the South until another new
35

currency for the whole nation replaced it in 1978. The old banking system was
nationalised, when the National Bank of South of Vietnam was established and took
over the position of the previous government in the South in major international
financial organisations, including the International Monetary Fund (IMF), Asia
Development Bank (ADB), and World Bank (WB). In 1978, the monetary union
was conducted as one of the major objectives of renovation plan in industry and
trade sector. Measures included focussing on developing heavy industry, eliminating
the comprador form of management, renovating nationalist bourgeois under the
socialism, promoting the decisive role of state enterprises, and building collective
cooperatives as pilot enterprises. However, the States renovation policy resulted in
an economy crisis, with stagnated production in agriculture, slow distribution in
domestic trade, high inflation, and a poor standard living. The economic growth was
low (0.4 percent per annum for the period 1977-1980).
The period from 1979 to 1982 witnessed changes in economic planning, with
new policies about pricing agriculture products and subcontracting in production.
Thus, the economy saw greater economic growth rates (2.3 percent in 1981, 8.8
percent in 1982). However, these changes had some negative effects, such as trade
imbalances, which led to increased prices and disruptions to central planning. From
1982, the Vietnamese government decided to focus on developing agriculture,
creating a reasonably strong industry-agriculture nexus. During 1981-1985,
economic growth rate increased to 6.4 percent per year.
In 1985, the Vietnamese government decided to apply a new mechanism,
which they named 'Price-Salary-Money' policy with the following main
components: (1) including all reasonable costs in production; (2) implementing a
one-price mechanism; (3) ensuring an adequate salary for a labourers life; and (4)
setting up self-control for economic units. Items supplied by the State and farmers
were priced under the rice price. The base salary for labourers was planned to
increase by 20 percent. The State expected to print more money, but the cost for this
was too expensive, so changing money was chosen for replacing. Ticket trade with a
fixed amount of goods was used for goods supplied at a low price; price
management was unreasonable, and serious food shortages developed. The targets of
this plan were not attained. The consequence was that prices increased rapidly:

36

inflation in 1986 increased by 774.7 percent, and the country saw three-digit
inflation for the next few years. Leaders seemed to realise that reforms cannot
succeed in a centrally planned economy. These issues created the need for
significant changes.
3.2.2

The Economic Renovation in 1986

A series of economic reforms (known collectively as the Doi Moi Reform)


began in 1986 with three main targets: (1) moving the economy from a centrally
planned to a socialist-oriented market economy under the management of the State;
(2) building a democratic society with the State formed of, by and for the people;
and (3) following an open policy, cooperating with all countries for general benefit.
Based on these targets, three pillars were also defined: rapid and sustainable
economic growth, a stable socio-political climate, and integration into the
international economic community. Following this reform programme, autonomy
for state-owned enterprises (SOEs) was increased, the States monopoly on foreign
trade was eliminated and private business on a small scale was allowed.
The Doi Moi Reform was revised based on previous lessons in developing the
economy and dissatisfaction with the society from both the people and the State.
While people hoped to have a new policy for solving problems in production and
trade distribution, the State agencies realised that patchy plans could not work. Doi
Moi was considered as the beginning of reform measures that spanned the next few
years.
3.2.3

Post-Doi Moi (after 1986)

The economic period after 1986 witnessed different sub-periods with domestic
economic fluctuations as well as world changes including the collapse of
communism in the Soviet Union and Eastern European countries in 1989, and
economic crises in 1997-1998 in Asia and 2008-2010 in the U.S. This period could
be divided into five sub-periods: 1986-1990, 1991-1995, 1996-1999, 2001-2007,
and from 2008 to the present.
In the first sub-period (1986-1990), the Vietnamese economy gradually
abandoned its old system, which had been based on administrative subsidies. This
was known as the period of economic renovation (Phan et al., 2006). Three
economic programs were conducted, focussing on food, consumption goods, and
37

export goods. Restrictions to domestic trade between provinces were eliminated, and
a multi-sector economy was accepted. The Contract 10 system introduced in 1988
gave farmers 15-year usage rights with an agricultural tax, reducing the previously
significant role of the collective cooperatives (Masina, 2006). Thus, this system was
defined as the main factor for dismantling the rural basis of the command economy.
With these changes, Vietnam became an exporting country in items such as rice (for
which it became the third leading exporter after Thailand and the U.S.) and crude
oil. The average economic growth from 1987-1990 was 5.37 percent. The year 1989
was the first year in which three-digit inflation was curbed and inflation fell
consistently thereafter (Table 3.1).
Table 3.1: Vietnams Real GDP Growth and Inflation from 1987 to 1990
1987

1988

1989

1990

GDP growth (%)

3.3

5.1

8.0

5.1

Inflation (%)

317

311

76

67.5

Source: General Statistics Office, www.gso.gov.vn.


In this period, the Vietnamese economy was negatively affected by the
collapse of the Soviet Union and the Council for Mutual Economic Assistance
(COMECON)6 in 1989, and from this time, aids to Vietnam from the COMECON
countries totally ceased. This event urged Vietnam to accelerate reforms to develop
the national economy. Assessing the reforms in this period, Phan et al. (2006, p. 32)
confirmed that Vietnam's success "encouraged the government to take a positive and
decisive step in accelerating the process of transition to a market economy, with a
combination of structural reforms and stabilisation measures.
The next sub-period (1991-1995) was highlighted as the first step towards a
market economy with macroeconomic stabilisation. In this period, the basis for a
market mechanism was established. In June 1991, the political programme for
Vietnams transition period to socialism was approved with the combination of two
tasks: strengthening industrialisation and modernisation and establishing sustainable
development in agriculture. The annual growth rate was strong, with the highest rate
recorded in 1995 (Table 3.2). In 1991, direct subsidies for SOEs were virtually

More details about the COMECON are available at http://www.shsu.edu/~his_ncp/CMEA.html


38

eliminated. The fiscal balance was improved with increases in revenue, mainly from
foreign trade, tax, and crude-oil exports. One noticeable feature was the emphasis
mentioned by Masina (2006) on rapid economic growth combined with poverty
reduction; specifically, GDP per capita increased from 100 USD in 1987 to 300
USD in 1996. In addition, the inflation rate was significantly curbed from 67.6% in
1991 to 17.5% in 1992, 5.2% in 1993, 14.4% in 1994, and 12.7% in 1995. Foreigntrade activities rapidly increased, but imports began to outstrip exports from 1993
and this gap continued to increase during this period.
Table 3.2: Some Economic Indicators from 1991 to 1995
1991

1992

1993

1994

1995

Real GDP growth rate (%)

6.0

8.6

8.1

8.8

9.5

Inflation rate (%)

67.6

17.5

5.2

14.4

12.7

Exports (billions of USD)

2.08

2.58

2.99

4.05

5.45

Imports (billions of USD)

2.34

2.54

3.92

5.83

8.16

Source: General Statistics Office, www.gso.gov.vn.


In the period 1996-1999, the regional financial crisis spread from Thailand.
This crisis caused reduced investment and trade in many countries. Because the
Vietnamese economy was not particularly open at that time, the Vietnamese
economy was not badly affected by the crisis. There were some negative effects,
such as decreasing foreign direct investment (FDI) and slowing down economic
growth. FDI decreased from 2.95 billion USD in 1997 to 1.9 billion USD in 1998.
Importantly, economic growth dropped from 8.2% in 1997 to 4.8% in 1998.
However, exports had been increasing during this period (7225.9 million USD in
1996, 9185 million USD in 1997, 9360.3 million USD in 1998 and 11541.4 million
USD in 1999).
Table 3.3: Real GDP and Sectoral Growth Rates in 2001-2005
Real GDP

Agriculture

Industry

Services

Planned target (%)

7.5

4.8

13.1

7.5

Real value (%)

7.5

5.4

16

7.6

Source: General Statistics Office, Vietnam, www.gso.gov.vn.

39

The period 2000-2007 witnessed the recovery of the Vietnamese economy


after the financial crisis with economic growth increasing at high and quite stable
rates (6.8% in 2000, 8.4% in 2005 and 8.46% in 2007). An assessment of the 20012005 five-year plan shows that GDP growth and the value of production areas
equalled or were greater than the planned target (Table 3.3). However, this period
did see economic deflation in some sub-periods.
Table 3.4: Some Economic Indicators from 2005 to 2011
2005

2006

2007

2008

2009

2010

2011

Real GDP growth rate (%)

7.6

7.0

7.1

5.7

5.4

6.4

6.2

Inflation rate (%)

8.3

7.5

8.3

23

6.9

9.2

18.6

Exports (billions of USD)

32.4

39.8

48.6

62.7

57.1

72.2

96.9

Imports (billions of USD)

36.8

44.9

62.8

80.7

69.9

84.8

106.7

Source: General Statistics Office, Vietnam, www.gso.gov.vn.


In the sub-period 2008 to the present, the Vietnamese economy has been
heavily influenced by the global financial crisis. Economic growth fell from 2008.
Moreover, the Vietnamese economy has had to suffer high inflation; for example,
the CPI was 23.1% in 2008 and 18.6% in 2011). To respond to external shocks,
Vietnam changed its policy from fighting inflation in 2008 to supporting economic
growth in 2009. In 2009, the Government gave stimulus packages, including interest
assistance and tax reduction to stimulate the demand side of the economy. In 2010,
the government pursued three targets namely curbing inflation, stabilising the
economy and ensuring the pace of growth.
3.3

THE ECONOMIC GROWTH AND INFLATION CONCERN

Since the Doi Moi period economic growth rate has been quite high.
According to GSO statistics, the average rates for the periods 1986-1990, 19911995, 1996-2000, 2001-2005 and 2006-2010 were 3.9%, 8.18%, 6.95%, 7.62% and
7%, respectively. For the period 2001-2010, the overall growth rate was 7.26%. Real
GDP in 2011 increased to 3.8 times that of 2000. Vietnam, thus, had one of the
highest growth records worldwide (if the year 2011 is included, there were 25
continuous growth years from 1986). The economic growth contributed to increased
living standards, represented by rising GDP per capita. In 1990, Vietnam was the
40

poorest country in the world, with a GDP per capita of 98 USD, but GDP per capita
in 2011 was 1,411 USD, and Vietnam emerged as a middle-income country (World
Bank, 2011).
Table 3.5: Value Added in Agriculture, Industry and Services Sectors of
Vietnam (% of GDP)
Year

Agriculture

Industry

Services

1985

40

27

32

1986

38

29

33

1987

41

28

31

1988

46

24

30

1989

42

23

35

1990

39

23

38

1991

40

24

36

1992

34

27

39

1993

30

29

41

1994

27

29

44

1995

27

29

44

1996

28

30

43

1997

26

32

42

1998

26

32

42

1999

25

34

40

2000

25

37

38

2001

23

38

39

2002

23

39

38

2003

23

39

38

2004

22

40

38

2005

21

41

38

2006

20

42

38

2007

20

41

39

2008

22

40

38

2009

21

40

39

2010

21

41

38

2011

20

40

40

Source: World Development Index WDI, World Bank.


41

The economic structure changed from focussing on agriculture to the industry


and services sectors. This is reflected via changes in value added in sectors of the
economy. The statistics reflect the increasing role of the industry and services
sectors, especially industry, in improving Vietnams economic growth. From 1986
to 2011, the share of agriculture in GDP decreased from 40% to 20%, the industry
share increased from 27% to 40%, and the services share increased from 32% to
40%.
Despite these achievements, Vietnams sustainable growth needs to be
improved to counter internal and external shocks. Economic growth has decreased
from 2006 and there are increasing inflationary pressures (Figure 3.1). This
instability results from unclear policy messages from the Vietnamese government in
their efforts to maintain high growth and ensure macroeconomic stability.
The inflation rate began to increase rapidly from 2003, and Vietnams
inflation rate has been higher than that of other countries in the Asian region. A
report by the International Monetary Fund (2006) shows that from 2002, there has
been a close relationship between the growth of the money supply (M2) and
inflation with a 12-month lag. Figure 3.1 illustrates that the fluctuation of inflation
was bigger than that of economic growth.
Figure 3.1: Vietnams M2 growth, GDP growth, CPI (%)

Source: General Statistics Offices website, www.gso.gov.vn

42

In examining the sources for economic growth of Vietnam, Ngoc (2008)


measured the contribution level of different factors, including labour, technology,
and capital, from 1975 to 2003. The findings illustrate that capital accumulation is
the most significant source of economic growth. Also, statistics show that gross
capital formation has accounted for a high rate of GDP. This becomes more
significant in studies of Vietnam where the capital accumulation has been based on
the banking system and the economys openness has been high.
Table 3.6: Gross Capital Formation of Vietnam (% of GDP)
Year

Gross capital formation

1986

14

1987

14

1988

18

1989

15

1990

13

1991

15

1992

18

1993

24

1994

25

1995

27

1996

28

1997

28

1998

29

1999

28

2000

30

2001

31

2002

33

2003

35

2004

35

2005

36

2006

37

2007

43

43

Year

Gross capital formation

2008

40

2009

38

2010

39

2011

35

Source: World Development Index WDI, World Bank.


In terms of aggregate demand, the theory shows that gross domestic product
(GDP) is defined by consumption, including spending by households and the
government; investment by government, businesses and households; and net
exports. Nguyen and Bui (2011) illustrates that the influence of factors from the
demand side in Vietnam in two periods (2000-2005 and 2007-present) on production
has been increasing. Table 3.7 shows that the final household and government
consumption significantly affected production, with increases of 21% and 27%,
respectively. Meanwhile, investment and exports made the same contribution at 5%.
Table 3.7: Influence of Factors from Demand Side on Production
Factors

2007-present

2000-2005

1.8

1.49

1.44

1.13

Investment

1.69

1.61

Exports

1.53

1.46

Final consumption by
household
Final consumption by
government

Source: Nguyen and Bui (2011).


3.4

THE OPENNESS OF VIETNAMESE ECONOMY

In line with the development of the country, trade activities were centrally
planned in the period 1975-1986. The State operated all functions (state
management and business), trade was run under administrative management in a
one-sector economy, trade between provinces was restricted, and international trade
activities were limited to socialist countries (the Soviet Union and the countries in
Eastern Europe).
44

The post-Doi Moi period from 1986 to the present, brought significant
changes in the field of trade. External trade focussed on supporting exports;
international relations were expanded, thus contributing to improved foreign trade.
Vietnams economic integration involved joining regional associations in 1995,
particularly signing the Bilateral Trade Agreement (BTA) with the United States in
2001 and joining the World Trade Organization (WTO) in 2007.
Vietnam is characterised as a small, open economy with total imports and
exports accounting for 150 percent of GDP. Total turnover in export and import both
tended to increase from 2000 to 2008, but decreased in 2009 because Vietnams
trade was significantly affected by international economic changes (Figure 3.2). The
rate of trade volume in GDP gradually increased from 60.7 percent in 1994 to 143
percent in 2011 (Table 3.8).
Figure 3.2: Vietnams Trade Volume (millions of USD)

Source: General Statistics Office of Vietnams website, www.gso.gov.vn


Another important characteristic of Vietnams trade is the high excess of
imports over exports, and the trade-balance deficit has been increasing over time.
Figure 3.2 illustrates that the balance of export-import in the period 2000-2001
45

began to increase in 2003, with the increase accelerating from 2007 to the present. In
addition, most machinery for domestic production is from imports with its weight in
import volume at 27 percent in 2006. This is because domestic manufacturing
industries are weak while the demand for the items it produces is necessary for
economic development. Thus, the domestic economy could be affected by world
price fluctuations. These problems negatively affect Vietnams macroeconomic
stability in the long run.
Table 3.8: The Rate of Export, Import and Trade Volume in GDP from
1994 to 2011 (% of GDP)
Year

Export

Import

Trade

1994

24.9

35.8

60.7

1995

26.3

39.3

65.6

1996

24.9

45.2

74.6

1997

34.3

43.3

77.6

1998

34.5

42.4

76.8

1999

40.2

40.9

81.2

2000

46.5

50.2

96.6

2001

46.2

49.9

96.1

2002

47.5

56.1

103.6

2003

50.1

62.6

112.7

2004

58.2

70.3

128.6

2005

61.1

69.2

130.3

2006

65.3

73.6

138.9

2007

68.2

88.1

156.3

2008

69.2

89.0

158.2

2009

57.2

63.2

120.4

2010

65.3

68.2

133.5

2011

72.7

71.0

143.7

Source: General Statistics Office of Vietnams website, www.gso.gov.vn

46

3.5

ENTERPRISE SECTORS IN VIETNAMS ECONOMY

Under the Law of Enterprise, Vietnams enterprise system includes stateowned enterprises (SOEs), and domestic non-state and foreign-owned enterprises.
3.5.1 State-owned enterprises
The development of SOEs links to the development strategy of Vietnam on the
transition to socialism. In any case, this enterprise type still plays the most important
role in the economy. However, the State also realised that it is necessary to help
these enterprises become effective economic forces of the State in the process of the
countrys industrialisation and modernisation. Currently, SOEs still hold the major
share of national capital assets, land and resources. However, the SOEs growth and
investment effectiveness have been lower than expected. According to the Ministry
of Finance statistical data in 2008, state enterprises and corporations accounted for
75 percent of the countrys fixed assets and about 60 percent of total domestic credit
and foreign loans, but their production was about 35 percent of GDP. This problem
is still a question as the state sectors sharing in fixed assets and domestic credit and
foreign loans largely unchanged while the contribution of this sector has been
decreasing (Table 3.9).
Table 3.9: Structure of GDP
2005
Total

100.00

2006

2007

2008

2009

2010

2011

100.00 100.00 100.00 100.00

100.00 100.00

State sector

37.62

36.69

35.35 35.07

34.72

33.46

32.68

Non-state sector

47.22

47.24

47.69 47.50

47.97

48.85

49.27

15.16

16.07

16.96 17.43

17.31

17.69

18.05

Foreign-invested
sector

Source: General Statistics Offices website, www.gso.gov.vn.


The present reform of SOEs is another major concern for Vietnam. This
reform is being conducted through the equitisation process, initial public offerings
(IPOs) and listings on the stock market, which are expected to improve these
enterprises efficiency. However, SOEs equitisation and reform have been quite
slow. The General Statistic Office database shows in the 2001-2010 period, 3,390
47

SOEs were equitised. In October 2011, there were 1309 SOEs. Moreover, some
large state-owned corporations, such as VINASHIN, VINALINES, have huge debts
(80,000 billion of VND and 43,000 billion of VND 7 ); these and other poor
performance measures point to problems in state management and ownership.
3.5.2 Non-state sector
Domestic non-state enterprises have been rapidly developing. These
enterprises were set up under the 1990 Law on Private Enterprises and provisions in
the 1992 Constitution. Originally the number of registered companies was small,
and most of them focussed on trade. This may have been because of initial concerns
that the private sector would challenge the countrys political system (Masina,
2006). With the increasing investment of foreign-invested enterprises and deeper
international integration of the Vietnamese economy, regulations for business
became more open, and the community of domestic companies grew bigger and
bigger. The Law on Enterprises taking effect in 2000 created a breakthrough for the
developing private sector. In 2004, there were 35,000 registered private enterprises;
in 2009, according to Ministry of Planning and Investment (MPI)s statistics, 83,000
enterprises were newly registered, creating 460,000 enterprises in total a 1,500
percent increase. GDP share of domestic private enterprises has increased every
year, accounting for 48 percent of GDP in 2010. Furthermore, this sector accounts
for 50 percent of total employment.8
However, this enterprise community has its limitations. They exhibit the weak
competitiveness and lending difficulties of the domestic non-state sector, which
could limit its development. Moreover, due to the open policy of the Law on
Enterprises, management after registration is not given much importance, so there is
a gap between the number of registered enterprises and operating businesses. This
makes the business environment more risky for enterprises, investors, and society.
The development of foreign-invested enterprises began in 1987, when the Law
on Foreign Investment was approved. These enterprises contributed significantly to
the development of the Vietnamese economy, increased the competitive strength of
7

http://dantri.com.vn/c20/s20-406434/Tong-no-cua-Vinashin-la-hon-80000-ty-dong.htm and
http://tuoitre.vn/Kinh-te/496954/Vinalines-no-hon-43000-ti-dong.html (VINASHI: Vietnam Shipping

Industry Corporation; VINALINES: Vietnam National Shipping Lines)


8

http://www.infotv.vn/doanh-nghiep/tin-tuc/52415-doanh-nghiep-tu-nhan-dong-gop-48-vao-gdp-nam-2010

48

the domestic private sector, and helped modernise the countrys economy. In the
GDP structure, this sector contributed about 19%. The World Banks statistics show
that foreign direct investment increased 769 times, from 10.4 million USD in 1987
to 8,000 million USD in 2010, but this figure decreased to 7,430 million USD in
2011. Lessons from the financial crises show that the activities of these enterprises
depend on the world economic picture.
3.6

VIETNAMS DOMESTIC FINANCIAL MARKET

After 1986, Vietnams financial sector was established with a single state
bank. In 1998, a new banking system was set up that included the central bank
system and the credit institutions. The stock market was formed in July 2000.
Changes in Vietnams financial market and monetary policy are summarised in
Table A2, Appendix A.
Figure 3.3: Structure of the Financial Market in Vietnam
The financial market

The money market

The stock market

Credit institutions

Stock companies, funds

Enterprises, households

Table 3.10: Financial Market in Vietnam


Unit: % of GDP
2008 2010 2011

2004

2005

2006

2007

Loans

61

70

75

93

93

n/a

n/a

Deposits

60

67

78

99

92

n/a

n/a

Share market (total capitalisation)

3.50

5.55

22.61

43.38

15

33

20

Outstanding bonds

8.4

8.2

8.1

13.7

15.1

n/a

n/a

Insurance premiums (life and non-life)

2.00

1.63

1.54

1.54

1.44

1.56

1.44

Pension funds

4.12

4.04

3.7

n/a

n/a

n/a

n/a

Source: The State Bank of Vietnam and Ministry of Finance, Vietnam.


49

3.6.1 Vietnams banking sector


Before 1988, the mono-bank system, in which the State Bank of Vietnam
covered both central- and commercial-bank functions, was considered a tool to
support the government's policies, to meet fiscal needs as well as the financial needs
of SOEs. In 1988, state-owned commercial banks were established, allowing the
State Bank of Vietnam to focus on its central-bank functions. Joint-stock
commercial banks were set up in 1991, and foreign banks were licensed to open
their branches in 1992. In 2010, 100% foreign-owned banks were allowed to be
established in Vietnam. Vietnams financial system has depended significantly on
the banking system, especially before the establishment of the stock market. Total
banking assets make up nearly 140% of GDP. Loans and deposits accounted for
93% and 92% of GDP at the end of 2008, respectively. Statistics show that there
have been significant changes in the number of credit institutions and share of total
assets (Tables 3.11 and 3.12).
Table 3.11: Number of Credit Institutions
Type of Credit Institution

1997

2006

2007

2008

2010

2011

State-owned commercial banks

Joint-stock commercial banks

51

36

34

40

37

35

Foreign bank branches

23

34

41

45

48

50

100% foreign-owned bank subsidiary

Joint-venture banks

Policy banks

Financial companies

17

17

18

Financial leasing companies

11

12

13

13

12

939

938

986

1019

1057

1095

Peoples credit funds

Source: The State Bank of Vietnam, annual reports.

50

Table 3.12: Share of Total Asset by Type of Institutions (%)


Type of Credit Institution

1997

2006

2007

2008

State-owned commercial banks

66.57

62.30

53.30

51.48

Joint-stock commercial banks

11.86

22.80

31.50

32.45

Foreign bank branches

16.6

9.80

9.60

10.26

Joint-venture banks

3.48

1.10

1.20

1.25

Finance companies

0.21

2.10

3.40

3.10

Leasing companies

0.12

0.80

0.70

0.97

Peoples credit funds

0.14

1.10

0.64

0.86

Source: The State Bank of Vietnam, www.sbv.gov.vn.


In brief, the banking sector has been the major source of domestic financing
for growth for many years. There are some concerns about changes in the Vietnam
banking system. First, although there is an increasing number of credit institutions,
state-owned commercial banks still maintain a crucial role in the banking system
(Table 3.13). Second, payment services have been developing for 10 years. The
Interbank Electronic Payment System came into operation in 2002 and was
expanded to all provinces and cities. Up to April 2009, 16 million bank cards were
issued, and 8,000 ATMs were installed. This contributed to a reduction of the ratio
of cash to total liquidity, from 20.3% in 2004 to 14.6% in 2008.
Table 3.13: Share of Credit by Type of Institutions (%)
Type of credit institution

1997

2006

2007

2008

2011

State-owned commercial banks

64.00

67.10

59.70

58.20

56.03

Joint-stock commercial banks

11.76

19.60

27.50

26.54

Foreign bank branches

19.85

8.30

8.56

10.27

Joint-venture banks

2.39

1.39

1.20

1.30

Finance companies

0.22

1.00

2.30

1.92

Leasing companies

0.03

1.30

1.10

1.19

Peoples credit funds

1.69

1.50

1.06

1.10

Source: The State Bank of Vietnam, www.sbv.gov.vn.

51

43,97

3.6.2 The stock market of Vietnam


Vietnams stock market was established in 2000. There are two stock
exchanges in Hanoi and in Ho Chi Minh City, respectively. The managing state
authority is the Securities State Commission (SSC) which is under the Ministry of
Finance (MOF). In the initial period from 2000 to 2006, it was an independent
agency; however, to combine SOEs equitisation process and the development of the
stock market, SSC was merged with the MOF in 2006.
Table 3.14: Stock Market Capitalisation

% of GDP

2003

2004

2005

2006

2007

2008

2009

2010

2011

0.4

0.5

0.9

14.9

27.5

10.5

21.8

19.2

14.8

Source: World Development Index, World Bank.


The Vietnamese government has made efforts to improve market regulation
and developing market instruments. The current regulation for the stock market is
the Securities Law, which was introduced in 2006 and revised in 2010. The
development of the Vietnamese stock market has included many achievements, such
as expanding the scale and liquidity of the market, helping enterprises and the
government mobilise capital, and strengthening SOEs disclosure. Trading volume
in 2009 increased 96 times over that of 2005. Foreign investment in the stock market
increased from 4.5 billion USD in 2007 to 9 billion USD in 2009. Table 3.14 shows
that the market capitalisation increased rapidly in 2006-2007, despite a decrease in
2008 (10.5% of GDP) when the world economy fell into recession.
Figure 3.4: Share prices in Vietnam from 2000 to 2011

Source: International Financial Statistics (IFS), IMF.


52

The stock market has developed through a number of periods. Although the
period 2006-2007 witnessed heat and uninterrupted growth (VN-Index at 1170.67
points), the stock market fell into a recession due to the negative effects of the 19971998 crisis and domestic measures for fighting inflation. To improve the market
structure, the SSC applied measures to expand organised stock markets and restrict
the free market, such as building the UpCom trading system in 2009 for public
companies, which was not listed on the formal stock exchanges. This contributed to
reduced risk for investors and increases in the markets transparency. However,
results from market supervision of the SSC illustrate that the stock market has been
driven by portfolio-investment inflows, some of which are speculative; this could
limit its role as an important transmission channel.
3.7

INSTITUTIONS AND POLICIES FOR A MARKET ECONOMY

Vietnam began to move from a planned economy to a market economy


beginning in 1986. To support a market economy, Vietnam developed many
significant institutions and policies. The basis for making these rules has been the
direction of the Communist Party, and its 10-year strategies and five-year and
annual plans for socio-economic development. Policies are set up to support
sectors/fields in the economy.
There are many different aspects of the Vietnamese governments large-scale
economic institutions and policies, such as the legal system for managing the
economy and the management method of the State. This section focuses on three of
these aspects: the legal system and policies for the enterprise environment and the
market; the legal system and policies for the financial market; and changes in the
States management approach.
3.7.1 Legal system and policies for the enterprise environment and the
market
The biggest change in institutions and policies was Doi Moi Reform in 1986.
From this time, Vietnam developed a multi-sector economy, with legal and policy
frameworks to increase enterprises autonomy and responsibility. A new
Constitution in 1992 institutionalised this policy. Other laws, such as the 1987 Law
on Foreign Investment, the Law on Enterprises and the Law on Private Enterprises
in 1990 confirmed the existence of many different sectors, both state and non-state,
53

in the economy. Subsequent laws, such as the 1995 Law on State Enterprises and the
1996 Law on Cooperatives limited the States intervention in business activities.
Noticeably, the 1999 and 2005 Laws on Enterprises created a breakthrough in
developing Vietnams enterprise community, as they allowed enterprises to do
business in any field not specifically prohibited by law. The States monopoly has
been gradually eliminated. Unreasonable sub-licences have been removed and
procedures are simplified. All kinds of enterprises are equal in business under the
law. Moreover, the 2005 Law on Investment improved the investment environment,
making all domestic and foreign investors equal and providing simpler investment
procedures. Vietnams policymakers have also established a legal system covering
aspects such as possession, contracts, competitiveness, taxation, customs, exportimport and bankruptcy.
To develop the market, Vietnam approved the Ordinance on Economic
Contracts, Civil Code (1995) and the Law on Commerce (1997), which established a
framework for free trading. The one-price policy was introduced in the 1980s, and
price management improved with the Ordinance on Price in 2002.
3.7.2 Legal system and policies for the financial market
A financial market is established with different branches: money market, the
State Bank of Vietnam bill and Treasury bond market, foreign exchange market, and
stock market. The legal system in Vietnam was set up to support the development of
these markets.
Ordinances on banks in 1989 and 1990 divided the management and business
functions, setting up the system of the State Bank and commercial banks. Next, the
Law on the State Bank and the Law on Credit Institutions were approved in 1997
and 1998, respectively (and revised in 2003, 2004 and 2010). These laws
significantly contributed to a legal framework for reforming the organisation and
operation of the State Bank of Vietnam and credit institutions. The State Bank of
Vietnam had a legal base to reform its operations including conducting monetary
policy and banking supervision. The types of credit institutions were expanded
(state-owned banks, joint-stock banks and foreign-invested banks). Bankinggovernance regulations were set up to enhance the effectiveness of commercial
banks. The Law on Securities was approved in 2006 and revised in 2010, creating a
54

legal system for ensuring that the stock market was operated publicly, transparently,
safely, and effectively. Along with these laws, many legal documents were issued to
specify regulations for operations in the financial market, such as open-market
operations and bond bids.
3.7.3 Changes in the States management approach
The State has also made reforms in its economic management functions. These
policies include dividing the functions of the States management, ownership of
SOEs and business autonomy, as reflected in Vietnams current equitisation process.
Moreover, the primary management approach has been in the degree of intervention
from the Government. This means that the State no longer directly interfere with
SOEs activities, which is the most important characteristics in the period of central
planning. Policies have been stabilised over a longer term, which has enterprises
develop more sustainably. The States management measures have changed from
administrative regulations to indirect policy tools to ensure that all sectors and
enterprises operate under the legal system.
However, as Vietnam is a transition country, many institutions and policies
lack practically and do not link different sources of growth because of conflict
between policies, and thus limit the development of the economy. In some fields
such as oil, petrol, and electricity, the State is still maintaining its monopoly role.
3.8

THE STATE BANK OF VIETNAM AND MONETARY POLICY

3.8.1 The State Bank of Vietnam


Currently, the regulations for the banking system include the Law on the State
Bank of Vietnam and the Law on Credit Institutions which were introduced in 19971998, then revised in 2003-2004 and re-introduced in 2010. These are an important
legal base for conducting the State Bank of Vietnams function in implementing
monetary policy.
One of the concerns in examining Vietnams monetary system is assessing
whether the State Bank of Vietnam has independence in making and conducting
monetary policy. To do this, it is necessary to discuss three aspects: the goals of
monetary policy; the mechanism for appointing the Governor of the State Bank and
the authority of the State Bank of Vietnam in using monetary policy tools.

55

Regarding the first aspect, the Vietnam National Assembly decides on and
supervises the implementation of monetary policy; the State Bank of Vietnam is a
governmental body that acts as the central bank of Vietnam, implementing state
management of currency and banking operations, and as a bank for credit
organisations. This ensures that monetary policy goals support the targets of the
governments socio-economic development plans. Monetary policy has the ultimate
goals of stabilising the currencys value and speeding up socio-economic
development. Monetary policy in Vietnam is no exception, although, like other
countries where central banks are not independent due to having a charter or
constitution. Vietnam faces particular challenges in implementing monetary policy.
In regard to the second aspect, the Governor of the State Bank of Vietnam is
nominated by the Prime Minister and elected by the National Assembly. Regarding
the third aspect, the Governor of the State Bank of Vietnam has the authority to
decide on the use of tools for implementing national monetary policy, in accordance
with the policies of the government. In practice, after the government approves the
annual money supply indicator, the State Bank of Vietnam adjusts its monetary
policy tools to achieve the goals.
According to the Law, the State Bank of Vietnam utilises interest rates,
exchange rates, reserve requirement, open market operations and other instruments
in order to implement the national monetary policies; monitors and supervises
banking activities; to control credit activities; and handles all violations in the
monetary and banking activities in accordance with law. It manages the current
transactions, capital transactions and foreign exchange spending in the Vietnamese
territory; manages the State foreign exchange reserves; and controls international
reserves; determines the exchange rates of Vietnam dong versus foreign currencies;
develops foreign currency market; and develops foreign exchange mechanism to be
submitted to the Prime Minister for approval.
In terms of the implementation of the functions of the Central Bank, the State
Bank of Vietnam carries out refinancing in order to provide short-term credit and
payment instruments for the economy; regulates the money market; and carries out
the open-market operations; organises the payment system via banks; conducts state
management of payment activities; provides payment services; and pursues the
56

policy of encouraging and strengthening non-cash payment under the approval of


the relevant authorities; develops the banking information system and provide
banking information services; manage credit information organizations; and
conducts credit rating for Vietnamese enterprises and other functions.
The State Bank of Vietnam is structured with 27 entities including
departments and organisations at its headquarters and State Bank branches of
provinces and the cities (Figure 3.5).
Figure 3.5: Organisational structure of the State Bank of Vietnam

Source: The State Bank of Vietnam.


57

In brief, the State Bank of Vietnam has a low degree of independence in


making and conducting monetary policy. In future, the State Bank of Vietnam is
likely to have greater independence, so assessing the current MTM channels and the
policy implications of an effective MTM will help the State Bank of Vietnam fulfil
its central bank functions. This is a new organisational direction in Vietnam, and has
already begun with the State Audit Offices change in legal position from a
governmental body to an agency of the Vietnam National Assembly in 2006.
3.8.2 The interest rates liberalisation process
The negotiable interest rate mechanism in VND commercial lending was
applied in 2002; since then, the State Bank of Vietnam announced the monthly base
rate, and credit institutions have used this rate as a direction for deciding their
commercial rates. This was one of the milestones in conducting monetary policy
based on market rules. Under this mechanism, credit institutions lend at interest rates
in line with market demand and supply. The State Bank of Vietnam also enhanced
credit institutions business autonomy and competitiveness.
Figure 3.6: Important Milestones about Monetary Institutions and Policies
1997-1998: Banking laws were introduced

2003-2004: Banking laws were revised


2010: New banking laws

2008: the interest rate cap (150% of the base rate)


2002: the negotiable interest rate
2000: the base interest rate announced by the State Bank of Vietnam

1999: the exchange rate management changed from administrative method to


market rules-based management: the average interbank rate the band (0.1% in
2000 to 0.25% in 2002 and 5% in 2009)
Source: Authors summary
58

However, due to the instability of the money market with interest rate
volatilities in the banking system, the State Bank of Vietnam returned in 2008 to an
interest rate cap mechanism, in which raising and lending interest rates could not
exceed 150% of the base rate. The State Bank of Vietnam understood that the
interest rate cap was instrumental in stabilising money market conditions, and that it
can be removed when there is a more stable market and a decline in the general level
of the interest rate.
Figure 3.6 illustrates the changes in the deposit and lending interest rate in
VND from 2000 to 2011. In 2008, both interest rates increased dramatically,
resulting in contractionary monetary policy to curb the high inflation at this time
(23%). After that, they decreased sharply in 2009 to be appropriate with the demand
stimulation policy of the Vietnamese government on that time.
Figure 3.7: The deposit and lending interest rate in VND during the period
of 2000-2011

Source: International Financial Statistics (IFS), IMF.

3.8.3 Vietnams exchange rate regime


Major reforms in managing the exchange rate were taken in 1999 when the
State Bank of Vietnam changed from administrative management to market rulesbased management. From this time, the State Bank of Vietnam no longer announced
the official exchange rate; instead, it announced the average interbank rate so the
59

commercial banks could set their rates around the State Bank of Vietnams rate. The
band was also adjusted from 0.1% in 2000 to 0.25% in 2002 and 5% in 2009.
Moreover, the inter-bank foreign exchange market continued to be developed with a
larger number of trading sessions. The State Bank of Vietnams statistics show that
the inter-bank transaction volume in 2008 increased by about 25%, and 79 banks
participated in the market. Foreign exchange management policy was reformed to
ensure the liberalisation process. The foreign exchange surrender ratio was reduced
from 80-100% in 1998 to 0% in 2003. Regulations and management mechanisms on
foreign currency became more flexible. The legal system for managing foreign
exchange, the Ordinance on Foreign Exchange Management, was made less
restrictive, with current account transactions liberalised, capital transaction control
loosened and the obligation of Article VIII of the International Monetary Fund's
Charter on lifting restrictions on foreign currency transactions fully accepted.
However, there remain some limitations in foreign exchange management, such as
weak management and regulation inefficiency, which have resulted in a dollarisation
problem in the economy (Table 3.15).
Table 3.15: Vietnams Financial Development Index 2008-2011
Rank

Score

49 (out of 52)

3.0

- Change in real effective exchange rate

14

-0.8

- Dollarisation vulnerability indicator

34

70.0

45 (out of 55)

3.0

- Change in real effective exchange rate

44

-0.7

- Dollarisation vulnerability indicator

36

47.9

46 (out of 57)

3.0

- Change in real effective exchange rate

n/a

n/a

- Dollarisation vulnerability indicator

42

104.5

50 (out of 60)

2.98

- Change in real effective exchange rate

n/a

n/a

- Dollarisation vulnerability indicator

43

93.7

Index 2008

Index 2009

Index 2010

Index 2011

Source: World Economic Forum, Financial Development Report 2008-2011.


60

3.8.4 Vietnamese monetary transmission mechanism


Monetary policy in Vietnam is a tool to implement the broad economic
objectives laid out in the national socio-economic development plan. Determining
targets and policy instruments is crucial for modelling and analysing MTM.
However there are some difficulties in the case of Vietnam due to the State Bank
of Vietnams unclear policies between 1986 and 2011. While the final goals of
monetary policy can be easily determined by legislation, the intermediate target
and policy instruments have not been clear over time. The final targets of
monetary policy, including stabilising inflation and supporting economic growth,
are regulated by the Law on the State Bank. Under the Law on the State Bank,
monetary policy tools include refinancing, interest rate, exchange rate, required
reserve and open-market operations (Vietnam National Assembly, 2011a).
In conducting monetary policy from 1986 to 2000, the State Bank of
Vietnam has not formed a clear intermediate target as along with the main policy
instrument. The State Bank of Vietnam controlled broad money (M2), credit
growth, key interest rates and official exchange rates at different times.
Therefore, in some cases, the exchange rate was considered as an intermediate
target, and some monetary policy tools such as M2, interest rates, and required
reserves were defined as policy instruments. This is actually a major limitation in
the State Bank of Vietnams operating mechanism of monetary policy. However,
from 2000, M2 has been considered an intermediate target, and the base interest
rate as announced by the State Bank of Vietnam has been a policy instrument.
This is because, in 2000, a new monetary policy mechanism was established
using this base interest rate as the State Bank of Vietnams main instrument in
implementing monetary policy. Meanwhile, the State Bank of Vietnam regulated
money indicators based on the annual government-approved money supply, so
M2 is the suitable intermediate target to achieve the final monetary policy
targets. M2 is also a main monetary indicator regularly reported to the National
Assembly (Vietnam National Assembly, 2011b). Thus, the variable M2 was
employed as the policy variable in the studies of Le and Pfau (2009) and Tran
(2009) on Vietnams MTM.

61

3.9

CONCLUDING REMARKS

The Vietnamese economy has changed significantly since the Doi Moi Reform
in 1986, moving from a centrally planned economy to a market economy. These
changes affect different economic indicators, such as economic growth, investment,
and trade activities. The history of the Vietnamese economy shows that reforms
started from the field of agriculture and spread to other fields. Moreover, the reforms
were refined based on lessons learned from mistakes in previous years, so studying
and learning economic trends is necessary for determining new policies in the
future.
One of the most important characteristics after Doi Moi is a dramatic increase
in the openness of Vietnams international economy. Vietnam joined international
organisations such as ASEAN and WTO, and this has helped Vietnam integrate
deeply into the world economy. Export has become one of the significant sources for
economic growth which has been positive and quite high, even in the period of the
financial crises in 1998-1999 and 2008-2010. Before these financial crises,
economic growth was high, with 8.7% in 1992-1997 and 7.8% in 2002-2007.
Vietnam has applied a step by step approach in its reforms, which has progressed
from a low level to a higher level as programs are gradually adjusted (World Bank,
2011). However, Vietnams sustainable growth is facing difficulties from internal
weaknesses as well as external shocks, so it is important to study the effects of
shocks on the Vietnamese economy.
A recent trend in the Vietnam economy has been a significant increase in the
inflation rate along with a reduction in economic growth. The growth of the money
supply has been suggested to have close ties with inflation. Thus, the fluctuations in
these variables should be examined in the framework of monetary policy.
Specifically, the study will identify the internal monetary shocks to output and
inflation via four MTM channels (interest rate, exchange rate, credit, asset price).
The outcomes will help identify monetary policy adjustments required to balance the
Vietnams output and inflation objectives.
Second, since Vietnam is a small, open economy, it will also be affected by
external shocks, such as changes in foreign output and incomes, foreign interest rate,
and world prices. For example, crude oil is one of Vietnams leading export
62

products, but its price is affected by fluctuations in the world economy. However,
the domestic prices of oil and petrol are managed administratively by the State.
Vietnams main export products are agricultural products, such as rice, and price
fluctuations of these products are unpredictable. Therefore, these foreign variables
are included in the model to examine effects of external shocks on the Vietnamese
economy.
Moreover, due to increasing Vietnams trade deficit issues, it is necessary to
examine further the international dimension of monetary policy. Examining the
international dimension of the MTM has not been done in Vietnam before, so this
study will help to identify the relationship between monetary shocks and the selected
trade variables, as well as comparing the effects of an external shock to those of
domestic monetary shocks. Thus, the trade variables need to be included in the
model of the study. In addition, the openness of the Vietnam economy to
international transactions will illustrate the role of the exchange rate channel, as well
as explore how the domestic interest rate is affected by external monetary shocks.
Third, the share of SOEs and state banks in the leading sector ensures a large
number of loans for state enterprises. However, inefficiencies in this sector will
influence the role of the MTM channels in Vietnam, especially the credit channel.
The significance of the MTM channels could be reduced because of the negative
impacts of non-performing loans to state-owned enterprises.
Fourth, the structure and depth of the financial sector will decide which
channels are important. The fact that Vietnams financial system depends on banks
could make the asset price channel less important. In addition, the power being held
by relatively few banks (specifically, state commercial banks) might cause the
interest channel to be weaker. Innovations in Vietnams banking system, such as
internet banking and electronic payments, will change the effects of the MTM
channels. Specifically, improvements in payment technology reduce the use of cash
in market transactions. This could, in turn, reduce the role of the interest rate
channel and increase the role of the asset price channel.
In addition, examining the stock price channel is necessary because this
channel has not been considered in previous studies of the MTM in Vietnam.
Furthermore, the stock market has developed rapidly since 2000 and it is potentially
63

an important emerging channel in the development of Vietnam. This has been found
in developing countries where their capital market is being strengthened. For
example, a shift in the most important role from the interest rate channel to the stock
price channel has been recorded in China before and after 2008 (Poon & Wong,
2011). This study aims to confirm a more positive role for this channel in Vietnam.
However, some factors affecting the Vietnam stock market, such as portfolioinvestment inflow or speculation, could limit the effectiveness of this channel.
In constructing a model, it is necessary to take several above features of the
Vietnamese economy into account to examine the operation of MTM channels.
Vietnam is a transition economy, and building and implementing institutions and
policies cannot be accomplished overnight. To develop a market economy,
Vietnam policymakers approved important laws and policies. These institutions
and policies created a legal framework for establishing and managing the market
economy in Vietnam. One of the most important changes has been a move from
management via administrative decision to the use of policy and legal tools.
However, institutional changes in the field of money and banking activity are
determined at a low level in comparison with reforms of the business environment
as a whole. However, these factors cannot be ignored in studying the current
situation of the Vietnamese economy. A number of institutional features could
affect the operation of the MTM channels.
First, the State Bank of Vietnams small degree of independence will affect
the speed and effectiveness of the transmission mechanism. Specifically,
monetary policy making is controlled by many agencies (such as the National
Assembly and the government), meaning that effects from activity in MTM
channels could have long lags. According to Sharifi-Renai (2010), the operating
procedures of central banks is the most important factor in examining the MTM.
This procedure is generally similar in central banks in many countries, although
institutional details may differ. All central banks use policy tools such as changes
in interest rates, reserve requirements, or management regulations to gain final
monetary policy targets.
Second, the positive changes in liberalising interest rates and exchange rates
along with institutional changes in money markets, will contribute to alterations of
64

the MTM. For instance, the existence or removal of an interest rate ceiling could
change the role of the interest rate channel. Sellon (2002) argued that in the case of
the United States, applying interest rate ceilings gave a larger role to the credit
channel than to the interest rate channel. In addition, the dollarisation issue in
Vietnam could change the significance of internal and external shocks. Specifically,
high dollarisation could result in the domestic interest rate reflecting both changes in
the domestic economy and adjustments in U.S. interest rates. This could lead to a
more important role for domestic interest rates than foreign in the domestic economy
(Aslanidi, 2007).
Third, changes in institutions and policies have created an equal business
environment for all type of enterprises (state-owned and non-state), so the private
sectors approach to bank credit is better. This could affect the analysis of the credit
channel in different periods.

65

CHAPTER 4
DATA AND STABILITY TESTS

This chapter presents the data and stability tests that were applied in this study
to understand the nature of the data. These tests include seasonality analysis,
statistical analysis, and unit root tests. The structure of this chapter is as follows.
Section 4.1 reviews the data for the study and its sources. Section 4.2 analyses the
seasonality of the data. Section 4.3 reports on the statistical analysis. The last section
summarises the issues discussed in this chapter.
4.1

DATA AND SOURCES


In this study, 48 quarterly time-series observations were collected from 2000:1

to 2011:4. The system for a small open economy like that of Vietnam is as follows:

yt ( y1,t , y2,t ) '


where:

y1,t

represents the foreign sector, including the variables related to the world

economy or foreign countries: WP (World oil price), WRP (World rice price), WGP
(World gold price), FY (Gross domestic product of China) and FFR (U.S. Federal
fund rate).
y2,t

represents the domestic sector, including the variables of the Vietnamese

economy: CPI (Consumer price index), Y (Gross domestic product), PI (Private


investment), PC (Private consumption), R (Short-term interest rate), M (Money
supply, M1), CR (Domestic credit aggregates), VNI (Stock index), E (Real effective
exchange rate) and trade variables (exports and imports).
The data and its sources are shown in Table 4.1.
The data were chosen bearing a number considerations in mind.
The choice of Gross Domestic Product for the variable Y overcomes the
weakness in Le and Pfau (2009) who selected industrial output as a proxy for GDP.

66

Table 4.1: Data Description


Variable

Data description

Data source

FOREIGN SECTOR
WP

World oil price, USD/barrel, log

IFS-IMF

WRP

World rice price, USD/ton, log

IFS-IMF

WGP

World gold price, USD/troy ounce, log

IFS-IMF

FY

Gross domestic product of China, millions of IFS-IMF


RMB, log

FFR

U.S. Federal fund rate, % per annum

IFS-IMF

DOMESTIC SECTOR (the Vietnamese economy)


CPI

Consumer price index (2005=100), log

GSO

Gross domestic product, billions of VND, log

GSO

Short-term interest rate, % per annum

IFS-IMF

Money supply, M1, billions of VND, log

IFS-IMF

CR

Domestic credit aggregates, billions of VND, IFS-IMF


log

VNI

Vietnam stock index

MOF

Real effective exchange rate (Q4/2007=100)

Calculated from the


database

of

Datastream
Aggregate demand components
PI

Private investment, billions of VND, log

MPI

PC

Private consumption, billions of VND, log

MPI

VE

Volume of exports, millions of USD, log

MPI

VI

Volume of imports, millions of USD, log

MPI

Note: IFS = International Financial Statistics; IMF = International Monetary Fund; MOF =
Ministry of Finance (Vietnam); MPI = Ministry of Planning and Investment (Vietnam), GSO =
General Statistics Office (Vietnam).

67

As mentioned in previous chapters, the variables R and M are regarded as


important targets in formulating the Vietnamese monetary policy. The variable R
also represents the interest rate channel in the MTM channels.
The variables CR, E and VNI are considered to examine the credit, exchange
rate and asset channels, respectively.
In this study, E the CPI-based real effective exchange rate (REER) is
measured as the change in the real value of the Vietnamese currency (VND) against
the basket of the trading partners of Vietnam9, so an increase in the REER indicates
a real appreciation of VND. This indicator is chosen to give a better reflection of
changes in the competitiveness of Vietnam with other countries.
Using exogenous variables aims to isolate exogenous policy changes.
+ Using the variable WP (world oil price) is similar to the approach of Kim
and Roubini (2000) and Afandi (2005); using the variable WRP (world rice price)
originated from the study of Le and Pfau (2009), and aims to control for external
shocks because this is an important export product in Vietnams economy. The idea
to include the variable WGP (world gold price) is from the conclusion of Tran
(2009) about the crucial role of gold price for the successful implementation of
monetary policy in Vietnam.
+ The aim in using the variable FFR (the U.S. Federal Fund rate) is to
control exogenous changes from foreign monetary policy. Obviously, using the
Federal Fund rate is the most suitable option because of the unchangeable role of the
U.S. dollar in the international market. Moreover, due to the dollarisation in the
Vietnam economy, this variable needs to be included in the model.
+ Considering the foreign output (FY) variable is necessary to understand
external shocks (Aslanidi, 2007). In the case of Vietnam, the study suggests
choosing the Chinese Gross domestic product as China is one of Vietnams biggest
trade partners. In 2010, the import turnover to China was 17.9 billion USD,

The real effective exchange rate of Vietnam (REER) = (The nominal effective exchange rate of
Vietnam * the consumer price index of Vietnam)/the geometrically weighted average of CPI indices
of trading partners of Vietnam. The base is 2007:4. The basket includes 20 trading partners: Japan,
Singapore, China, South Korea, the United States, Thailand, Australia, Hong Kong, Germany,
Malaysia, France, Indonesia, the United Kingdom, the Netherlands, Russia, the Philippines,
Switzerland, Italy, Belgium, and India. The database is from Datastream (Reuter). The quarterly
approach is based on the REER calculation suggested by Bruegel a non-profit international
68

accounting for 24.9 percent of the total of import turnover of Vietnam (General
Statistics Office of Vietnam, 2010). Thus, shocks in the Chinese economy can
potentially affect the Vietnamese economy.
Using trade related variables, including VE (exports) and VI (imports), aims to
examine the role of the international channel on the MTM.
4.2

SEASONALITY ANALYSIS
This section tests the existence of seasonal effects in the data. First, variables

are identified for the existence of seasonal patterns. For those variables with
seasonal patterns, procedures are applied to analyse the seasonality. Two methods
are used for seasonal adjustment (1) modelling and estimating dummy variables
with the Eviews package; and (2) using Eviews computing procedures, such as
X12-ARIMA and TRAMO/SEATS (Time Series Regression with ARIMA Noise,
Missing Observations and Outliers/Signal Extraction in ARIMA Time Series) for
adjusting each time series. These approaches allow the comparison of results when
analysing seasonal adjustments. Seasonality analysis is crucial for studies of the
Vietnamese economy because all statistical figures are published without seasonal
adjustment. Moreover, seasonal adjustment has not been considered in many studies
of the Vietnamese economy (see Goujon (2006), Phan et al. (2006), Ngoc (2008),
Hoang (2009), Hoang and Dung (2009), Le and Pfau (2009), Tran (2009), Phuc and
Duc-Tho (2009), Nguyen (2010), and Nguyen et al. (2012)). Therefore, seasonal
adjustmentcan provide additional insights in our study.
4.2.1

Seasonal-factor identification
Seasonal factors play a significant role in time-series data. Diebold (2007,

p.99) stated that the seasonal pattern is one that repeats annually. Moreover, such
patterns usually appear in many monthly or quarterly economic variables. A time
series comprises of seasonal, trend, cyclical, and random (irregular) components
(Gujarati & Porter, 2009, p.290).

TSt f (TCt , St , It )
where: TSt is the original time series at time t,
association with the members from EU Member State governments and international corporations
and institutions. The website of Bruegel explains this approach in details.
69

TCt is the trend-cycle


St is the seasonality
It is the irregularity.
A time series is defined as a multiplicative or additive function of these
factors. According to Diebold (2007), factors such as weather or holidays could
affect seasonal patterns, and need to be considered in analysing seasonality. Vietnam
is an agricultural country, so the weather is a very important seasonal factor which
could affect agricultural production as well as some aspects of industrial production,
construction, and transport. In agriculture, there are three main cropping seasons,
including winter-spring (the first four to five months of a year), summer-autumn
(April to July), and the main crop season (June to November). Harvest activities
usually occur in mid-June (for the first season), August and September (for the
second season) and December (for the third season). Furthermore, there are also the
dry and wet seasons in the southern areas of the country. The wet season, which
normally happens from late April to October, affects many sectors including
industrial and construction activities. Moreover, Vietnam is known to be very
vulnerable to natural disasters, such as floods and storms.
Holidays are a surprisingly complex factor in this analysis. The country uses
two types of calendars the Gregorian and the Lunar, giving rise to four fixed
holidays based on the Gregorian calendar (New Year on 1/1, Unification Day on
30/4, Labour Day on 1/5 and National Day on 2/9) and two moving holiday periods
based on the Lunar calendar (four days for Lunar New Year and one day for the
birthday of the Vietnamese ancestor King Hung). To ensure appropriate conditions
for modelling and testing, the seasonal components must be removed from the time
series. This process is known as seasonal adjustment or deseasonalisation.
This study uses 16 variables, of which two (Chinas GDP and Vietnams
money suppy) are available in seasonally adjusted form, while the others have not
been seasonally adjusted. Therefore, the seasonal-adjustment procedure is applied to
estimate and remove seasonal effects from all time-series variables used in the
study.

70

To identify the seasonal patterns, the time-series data in the study is plotted
with a time period from quarter 1 in 2000 to quarter 4 in 2011 (see Figure 4.1 for
foreign variables and Figure 4.2 for domestic variables).
Figure 4.1: Foreign variables
(Federal Fund Rate (FFR), %/annum; Chinas Gross Domestic Product (YC), Real, constant
2005, millions of RMB, seasonally adjusted; World oil (crude oil) price (WOP), US$ per Barrel;
World rice price (WRP), US Dollars per Metric Ton; World gold price (WGP), USD/troy ounce.)

WP

WRP

140

1,000

120
800
100
80

600

60

400

40
200
20
0

0
2000

2002

2004

2006

2008

2010

2000

2002

2004

WGP

2006

2008

2010

2008

2010

FFR

2,000

7
6

1,600
5
1,200

800

3
2

400
1
0

0
2000

2002

2004

2006

2008

2010

2008

2010

2000

YC
8,000,000
7,000,000
6,000,000
5,000,000
4,000,000
3,000,000
2,000,000
2000

2002

2004

2006

Source: Authors calculation.

71

2002

2004

2006

Figure 4.2: Domestic variables


(Gross domestic product (Y), Real, constant 1994, millions of VND; Consumer Price Index
(CPI), (2005=100); Private investment (PI), billions of VND, Private consumption (PC), billions of
VND; Money (M), billions of VND, Seasonally Adjusted; Short-term interest rate (R), %/annum;
Domestic credit aggregates (CR), billions of VND; Vietnam stock index (VNI); The real effective
exchange rate (E), 2007M12=100; The volume of exports (VE), millions of USD; The volume of
imports (VI), millions of USD).
Y
200,000

160,000

CPI

PI

240

200,000

200

160,000

160

120,000

120

80,000

80

40,000

120,000

80,000

40,000

40
00 01 02 03 04 05 06 07 08 09 10 11

0
00 01 02 03 04 05 06 07 08 09 10 11

PC

00 01 02 03 04 05 06 07 08 09 10 11

400,000

800,000

24

300,000

600,000

20

200,000

400,000

16

100,000

200,000

12

0
00 01 02 03 04 05 06 07 08 09 10 11

8
00 01 02 03 04 05 06 07 08 09 10 11

CR
4,000,000

00 01 02 03 04 05 06 07 08 09 10 11

VNI
1,200

REER
130

1,000

120

3,000,000
800

110
2,000,000

600
100
400

1,000,000

90

200
0

0
00 01 02 03 04 05 06 07 08 09 10 11

80
00 01 02 03 04 05 06 07 08 09 10 11

VE

VI

30,000

30,000

25,000

25,000

20,000

20,000

15,000

15,000

10,000

10,000

5,000

5,000

0
00 01 02 03 04 05 06 07 08 09 10 11

00 01 02 03 04 05 06 07 08 09 10 11

Source: Authors calculation.

72

00 01 02 03 04 05 06 07 08 09 10 11

In the foreign sector, Chinas seasonally adjusted GDP has consistently


increased since 2000. It is notable that there was a slight decrease in 2008 when the
2008-2010 recession happened. However, this crisis seems to have little effect on
the upward trend of GDP. Other foreign variables had many fluctuations in the
period 2000-2011. As can be seen from Figure 4.1, the Federal Fund rate reached a
peak in the third quarter of 2000 before decreasing until the fourth quarter of 2003.
After that, there was a recovery until 2007:1, afterwards this rate has been
decreasing. A decrease in both the world oil price and world rice price in 2008:2
reversed their previous increasing trend. An interruption in the increasing trend of
the world gold price appeared from 2008:2 to 2008:4.
In the domestic sector, the money supply variable (M) is the only domestic
variable available in seasonally adjusted form. As shown in Figure 4.2, there are
different patterns of fluctuations in Vietnams domestic variables. During the 20002011 period, Vietnams important macroeconomic variables, such as GDP (Y),
private consumption (PC), private investment (PC), exports (VE), and imports (VI),
increased, but in different patterns. Also, some variables affecting growth quality,
consumption and investment behaviour, such as CPI, credit (CR), and the money
supply (M), rose in the given period. All time-series data, except M, are also
adjusted seasonally, but only some variables are discussed to highlight the seasonal
factors in Vietnam.
Figure 4.3 illustrates that a seasonal component exists in Y from 2000:1 to
2011:4. Specifically, the lowest value of GDP occurs in the first quarter and the
highest in the fourth quarter every year. There are some plausible explanations for
this pattern. In Vietnam, there are many holidays in quarter 1, including the New
Year (1/1), the Lunar New Year (four days in January or February) and other
agriculture-related holidays. Therefore, people usually do not work on these
occasions. After these holidays, cultivation activities are conducted, so it takes more
time for them to affect the GDP value. Besides, other holidays in quarter 2, such as
Victory Day (30/4), Labour Day (1/5), and King Hungs Birthday Anniversary
(perhaps in April or May), could affect the production. In quarter 4, industrial
production, business cycles and agricultural harvests are expected to complete.
Moreover, many industrial and construction activities are believed to increase after
October when the wet season finishes. Regarding the PC and PI variables, seasonal
73

factors appear in a given period. Figure 4.1 confirms that these variables should be
treated as multiplicative because there is a proportional relation between the
magnitude of the seasonal changes and the time series (Central Bureau of Statistics,
2013).
Figure 4.3: Gross Domestic Product (Y), Private Investment (PI) and
Private Consumption (PC)
400,000

Billions of VND

350,000
300,000
250,000
200,000
150,000
100,000
50,000
0
00

01

02

03

04

05

06
PI

07

08

09

10

11

Year

PC

Source: Authors calculation.


An array of methods may be used for the deseasonalisation process, including
the simple moving-average method, the dummy-variable method, and some
sophisticated hybrid methods like the X11/X12 procedure (Diebold, 2007). In this
study, the latter two approaches are considered with the Eviews software, as they
give more advanced analyses than the first approach. X11 style methods are defined
as filter-based methods in which the ratio to moving-average process is used. This
procedure, introduced by the National Bureau of Economic Research includes three
stages: (1) use a moving average to estimate the trend, (2) leave the seasonal and
irregular factors to remove the trend, and (3) use moving averages to smooth
irregularities, and thus estimate the seasonal component. X12 ARIMA is the new
version of the X11 family and was released by the U.S. Census Bureau in the late
1990s (Australian Bureau of Statistics, 2013). This method differs from other
methods in that changes in the seasonal factors appear annually when using X11,
74

and no change appears for quarters in the moving-average method (Microsoft,


2009a).
In the next section, seasonal adjustment procedures are applied to some
domestic variables exhibiting seasonal patterns for making seasonal adjustments.
4.2.2

Modelling seasonality with dummy variables


In this approach, determining the number of dummy variables to use is an

important step. Diebold (2007) and Gujarati and Porter (2009) confirm that the
number depends on the given data (for example, the number of observations in a
year). Specifically, in this study, data are collected quarterly so the number of
seasons equals 4, indicating four seasonal dummy variables Di (i=1, 2, 3, 4). For
example, D1 indicates whether economic activities have been observed in the first
quarter (the value is 1 if such activities have been observed and 0 otherwise); D2, D3
and D4 are used similarly.
Gujariti and Porter (2009) emphasise that omitting the intercept is necessary
for avoiding the dummy-variable trap, in which an inclusion of an intercept and all
dummy variables are present. Such a trap produces the econometric problem called
perfect multicollinearity (Diebold, 2007). In this study, this would mean that four
dummies are determined for four quarters of the year. This approach is shown in the
following model:
(Model 1)
4

TSt ai Dit ut

(4.1)

where: TSt is the value of the time series at time t (t = 2000,, 2011),
ai is the dummy coefficient and is the mean value of quarter i (i = 1,
2, 3, 4), indicating the seasonal factors,
Dits are the seasonal dummies, and
ut is the error term.
Using Model 1 equation (4.1) to perform a regression for each variable gives
the t-value results of the given dummy coefficient; this helps to determine if there
are seasonal factors in given quarters. In this case, the t-value is statistically
significant (see Table 4.2).

75

Another econometric procedure that is popularly applied in the seasonaladjustment studies is treating a season as the reference (Diebold, 2007; Gujarati &
Porter, 2009). While the first approach is used to test for seasonal effects, this
approach is applied to test the existence of effects from a given time period.
Identifying the benchmark time should be based on the seasonal characteristics.
Regression results indicate the seasonal changes in relation to the omitted season.
Such an approach is used by Asteriou and Kavetsos (2006) to test for the existence
of the January effect in the stock-market return of eight transition economies
(Slovenia, Slovakia, Russia, Romania, Poland, Lithuania, Hungary, and the Czech
Republic). These authors chose January as the time for testing due to their
hypothesis on tax-loss selling: in December, the end of the tax year, the pricedeclined stocks are sold, and in the first month of a new year firms maintain
business with higher risk (higher return), resulting in the January effect. Because of
this, the January effect is commonly studied in relation to the seasonality in stock
markets (see Fountas and Segredakis (2002), Bohl and Salm (2010), Agnani and
Aray (2011)). In this study, the same approach is used by considering the first
quarter as the reference quarter for the following reasons. First, this season coincides
with holiday periods, both fixed and moving. These holidays could affect economic
variables like GDP, investment, and consumption. Second, at the end of this quarter,
production and business activities begin to increase with the hope of a bright future
in a new year. Third, data on GDP, investment and consumption show that the value
in quarter 1 is lower than that of other quarters, after reaching its highest level in
quarter 4 of each year. Such reasoning produces the following model, with an
intercept and three dummy variables indicating the second quarter to the fourth
quarter.
4

(Model 2):

TSt a1 ai Dit t

(4.2)

where: TSt is the value of the time series at time t (t = from 2000 to 2011),
a1 is the dummy coefficient and is the mean value of quarter 1,
ai is the dummy coefficient (i = 2, 3, 4),
Dits are the seasonal dummies (i = 2, 3, 4), and

t is the error term.


76

The difference between the two models is illustrated in the following aspects.
First, the null hypotheses differ between the models. In Model 1, testing for seasonal
effects is based on the null hypothesis that the dummy coefficients (ai) are equal: a1
= a2 = a3 = a4. This means that if the null hypothesis is not rejected, there is no
seasonal effect for domestic variables in Vietnam; otherwise, seasonal effects exist.
The null hypothesis in Model 2 is that all dummy coefficients (ai, i = 2, 3, 4) are
zeros. If this hypothesis is rejected, the appearance of seasonal effects is confirmed.
Second, the dummy coefficients have different expressions. If they express the mean
value corresponding to a given quarter in Model 1, they are the differences between
the value of the first quarter and the ith quarter. Third, while regressing Model 1
gives only one value for all four standard errors, there are two values of standard
errors in Model 2. This is because the value of all dummy variables is 1 or zero, so
the estimated coefficients have the same standard errors (see Gujariti and Porter
(2009) and Asteriou and Kavetsos (2006)).
Both models are estimated by the OLS. The regression results for all three
variables in each model are shown in Table 4.2.
Table 4.2: Estimated Values for Gross Domestic Product (Y), Private
Investment (PI) and Private Consumption (PC)
- For gross domestic product (Y)
Model 1
Variable

Coefficient

Model 2

t-Statistic

Prob.

Variable

(Std. Error)
D1

80410.330

108997.100

10.440

0.000

103925.800

14.152

0.000

D2

123337.900

13.493

0.000

D3

10.441

0.000

28586.750

2.624

0.012

23515.420

2.159

0.036

3.941

0.000

(10892.370)
16.014

0.000

D4

(7702.071)
R2

80410.330

(10892.370)

(7702.071)
D4

Prob.

(7702.071)

(7702.071)
D3

t-Statistic

(Std. Error)

(7702.071)
D2

Coefficient

42927.580
(10892.370)

R2

0.268

- For private investment (PI)


77

0.268

Model 1
Variable

Coefficient

Model 2

t-Statistic

Prob.

Variable

(Std. Error)
D1

41213.380

69376.850

3.497

0.001

64549.210

5.887

0.000

D2

71956.450

5.477

0.000

D3

3.497

0.001

28163.470

1.689

0.099

23335.830

1.400

0.169

1.845

0.073

t-Statistic

Prob.

4.918

0.000

0.763

0.449

1.038

0.306

1.595

0.119

(16666.030)
6.106

0.000

D4

(11784.670)
R2

41213.380

(16666.030)

(11784.670)
D4

Prob.

(11784.670)

(11784.670)
D3

t-Statistic

(Std. Error)

(11784.670)
D2

Coefficient

30743.060
(16666.030)

R2

0.096

0.096

- For private consumption (PC)


Model 1
Variable

Model 2

Coefficient t-Statistic

Prob.

Variable

(Std. Error)
D1

130201.600

(Std. Error)
4.918

0.000

(26475.350)
D2

158786.700

169048.0

5.998

0.000

D2

189907.6

6.385

0.000

D3

38846.380
(37441.810)

7.173

0.000

D4

(26475.350)
R2

28585.110
(37441.810)

(26475.350)
D4

130201.600
(26475.350)

(26475.350)
D3

Coefficient

59706.010
(37441.810)

R2)

0.062

0.062

Source: Authors calculation.

As shown in Table 4.2, the estimated coefficients produced in Model 1 show


that the average values in quarter 1 are the smallest (80,410 in Y, 41,213 in PI and
130,202 in PC), and those in quarter 4 are the highest values (123,338 in Y, 71,956
in PI, and 189,908 in PC). In Model 2 when the first quarter is treated as the
78

benchmark, the results show that there is no negative value of dummies, so no


quarters have lower GDP, private investment, or private consumption than the mean
value in quarter 1. Dummy variables account for a small part of the variation of the
variables (R2 = 0.27, 0.1 and 0.062 for Y, PI and PC, respectively).
For the Y variable, regression results in both models illustrate that p-values are
less than 5 percent and all t-statistic values are statistically significant, so the null
hypothesis is rejected, supporting the existence of seasonal effects and the impact of
the quarter 1s effects on GDP. This finding confirms that the average GDP values
for the second, third and fourth quarters are statistically different to the mean GDP
in the first quarter.
For the variables PC and PI, the findings from Model 1 are similar to those for
the variable GDP: seasonal effects appear in all quarters. However, there is a distinct
result in running Model 2 for these two variables. P-values for dummy coefficients
are greater than 5 percent, so the null hypothesis is not rejected and there is no
evidence for the existence of the quarter 1s effects on investment and consumption.
This implies that these coefficients for the next three quarters are not statistically
significant, and that the investment and consumption values are not statistically
different to the mean value for the first quarter.
Treating Quarter 1 as the benchmark quarter leads to the conclusion that
regression results for Y contrast with those for PC and PI. While the value of Y in
the following three quarters is statistically different to the average value in the first
quarter, there is no statistical difference for investment and consumption between
quarter 1 and the three remaining quarters of the year. As consumption (PC) and
investment (PI) are the main components in measuring GDP, this finding suggests
that the changes in Vietnams GDP could not derive from changes in private
consumption and private investment.
In economics, GDP is calculated under the expenditure approach, adding final
expenditures on goods and services with net exports equalling exports minus
imports as follows:

GDP PC PI G (VE VI )
where: PC, PI, VE and VI are consumption, investment, exports and imports,
respectively. G is the government expenditure on goods and services.
79

Because investment and consumption in the last three quarters are not
statistically different to the values in quarter 1, quarterly changes in GDP can be
explained by two other factors namely, government expenditure (G) and net exports
(VE VI).
Continuing the analysis, this studys data on exports and imports are used for
regression under two dummy variable models. By repeating the steps above, the
regression results for the variable net exports are generated (Table 4.3).
Table 4.3: Estimated Value for Net Exports (VE - VI)
Model 1
Variable

Model 2

Coefficient t-Statistic

Prob.

Variable

(Std. Error)
D1

D2

D3

0.005

Prob.

-2.974

0.004

D2
-4.286

-0.927

0.358

0.000

0.213

0.832

0.011

-0.888

0.379

160.148
(751.263)

D4
-4.231

-696.524
(751.263)

D3
-2.673

-1580.093
(531.223)

-2247.580
(531.223)

R2

-2.974

-1419.944
(531.223)

D4

-2276.617
(531.223)

t-Statistic

(Std. Error)

-1580.093
(531.223)

Coefficient

0.000

-667.487
(751.263)

R2

0.046

0.046

Source: Authors calculation.

Table 4.3 shows that results for net exports, PI and PC are similar in that
seasonal effects appear in all quarters, and quarter 1s effects do not occur in other
quarters. The net exports in the last three quarters are not statistically different to the
mean net exports in quarter 1. Seasonal dummies represent 4.5 percent of the
variation in net exports. Given these results, the impact of investment, consumption,
exports, and imports on GDP seems to be statistically insignificant. Therefore, the
only factor affecting on quarterly changes in GDP is government expenditure (G).
This implies that in a developing economy like that of Vietnam, changes in GDP
seem to originate from fiscal policy, instead of monetary policy.

80

4.2.3

Seasonal adjustment with complicated computing procedures


As mentioned above, two popular seasonal-adjustment procedures used to

extract time-series components are X12-ARIMA and TRAMO/SEATS (T/S). There


is an ongoing debate on what procedure is suitable for the data of different countries
(Eo, 2010). While the former is based on non-parametric moving averages, the latter
is conducted with parametric ARIMA. Moreover, comparing these two methods
shows that the X12 method is significantly different to the T/S method in that the
latter allows missing values (Microsoft, 2009a). In this study we have compared the
results from the application of both procedures, after selecting the appropriate model
(additive or multiplicative) in X12 to produce suitable adjusted data. A spectrum
including better smoothness and reduced seasonality of adjusted data ensures the
goodness of adjustment (Cleveland & Devlin, 1980). Applying X12 procedures
allows a clear comparison between additive and multiplicative models for each
variable.
[Insert Figure B1 here]
Figure B1 illustrates that the multiplicative model is more appropriate than the
additive one due to the goodness of adjustment in all three variables. This is
confirmed when plotting the variables, which reveals a proportional relationship
between the magnitude of seasonal changes and the data; in other words, a signal of
a multiplicative form (Central Bureau of Statistics, 2013). Another approach is using
diagnostic indices to measure the quality of adjustment. In the X12 procedure, there
are some important quality measures (Norway, 2008)10:
- M2 (the relative contribution of the irregular component to the variance of
the stationary portion of the series).
- M7 (the amount of stable seasonality relative to the amount of moving
seasonality).
- M10 (the size of seasonal-component fluctuations in recent years)
- M11 (the size of linear movement in the seasonal component in recent years)
- Q (the collective measure of quality in X12-ARIMA)
These measures, used to identify several criteria for the ideal model, are
specified in Table B1 (Appendix B).

81

[Insert Table B1 here]


Based on the criteria above, the results from applying both the multiplicative
and additive models are compared in Table B2 (Appendix B).
[Insert Table B2 here]
Table B2 illustrates that both models are acceptable for Vietnams database,
but all indices under the multiplicative model are smaller than those under the
additive model. This proves that the multiplicative model is better than the additive
model in determining seasonal adjustments. However, there is a problem in the
multiplicative model, in that both M10 and M11 of the variable PI are higher than 1.
This means that fluctuations in private investment are too large, its seasonal
adjustment is no longer stable (as M10>1), and its fluctuation is not random (as
M11>1). With better adjustment in applying the multiplicative model in the X12
procedure, the seasonally adjusted data under this model were used for the next
analysis.
The next work in this section is the comparison of the different computing
procedures in deriving the adjusted series. As mentioned, two popular procedures
are X12 and T/S, which have been variously applied in different countries (Table
B3, Appendix B).
[Insert Table B3 here]
With such varied applications, it is significant to compare both procedures for
their suitability for application to the case of Vietnam. According to Eo (2010), in
terms of theoretical comparison, T/S has advantages over X12-ARIMA. However,
when comparing these two procedures in the case of Korea, the author recommends
that replacing X12-ARIMA by T/S would not be a good option (Eo, 2010). Figure
B4 (Appendix B) shows the comparison of these procedures for each variable.
[Insert Figure B2 here]
Figure B2 shows the seasonally adjusted GDP. It appears that both procedures
produce very similar results, except for a slight difference between 2010 and 2011.
Therefore, it may be difficult to identify whether applying X12 or T/S makes a
significant distinction in the case of the GDP variable.

10

Details about the quality measures are contained in Table B1.


82

As can be seen from the graph on investment (PI) and consumption (PC), the
same results are observed from 2000 to 2007 for both procedures. Movements
recorded from 2008 to 2010 are unfit, but main trends seem to be unchanged. Due to
missing data for the PI and PC variables in 2011, only the T/S procedure is
appropriate, as, unlike the X12 procedure, it can interpolate missing values
(Microsoft, 2009a). With T/S, values of investment and consumption are appended
till the end of 2011. A noticeable aspect is that the difference in seasonally adjusted
data using both procedures occurs in the period of the financial crisis from 2008.
This consideration is appropriate given the difference in identifying irregular
components of these two procedures in the period 2008-2010. (Figure B5, Appendix
B)
[Insert Figure B3 here]
As Figure B3 shows, if, in the 2000-2007 period, there is a similar shape in
irregular factors between the two approaches, the different movement is obviously
denoted for both variables in the next period from 2008. The irregularity factor
includes three relevant parts, including calendar changes, extreme one-time events,
and residual irregularity (Central Bureau of Statistics, 2013, p9). However, in this
period there is no change to the calendar in Vietnam; thus, excluding measurement
errors, the main event comes from economic instability. Therefore, unpredicted
factors in the time of financial crisis seem to result in the difference in applying X12
and T/S in the seasonal adjustment process for Vietnams economic indicators. Due
to the unavailability of data and the scope of this study, this finding could not be
examined in the case of other economies to obtain a more clear comparison for the
crisis time. Currently, another procedure has been developed with the combination
of X12-ARIMA and T/S, known as DEMETRA, which automates seasonal
adjustment and trend estimation 11 .In the study of Eo (2010), a measure of
idempotency, known as the relative mean absolute difference (RMAD), is used to
identify whether X12-ARIMA or T/S is more appropriate for seasonal adjustment.
This measure is based on the assumption that the seasonal-adjustment procedure is
considered a good choice if the adjusted data is unchanged after repeating the same

11

See information from http:/forum.europa.eu.int/irc/dsis/eurosam/info/data/demetra.htm.


83

adjustment steps (Eo, 2010). Specifically, the formula for this measurement is
proposed as follows:

I (%)

1 T Yt , SA1 Yt , SA2
x100
Y
T t 1
t , SA1

where: T: the number of observations,


Yt,SA1: the first seasonally adjusted data, and
Yt,SA2: the second seasonally adjusted data.
Applying the same calculation does not work correctly for this study because
the T/S procedure does not produce the second seasonally adjusted data that is
produced by the X12-ARIMA procedure. Thus, to prevent computational errors in
estimating Vietnams variables, the X12-ARIMA procedure is preferred to the T/S
procedure.
Returning to the application of the X12-ARIMA procedure, one important
issue is moving holidays. Seasonal adjustment is more complicated when applying a
popular seasonal adjustment method like X12-ARIMA for countries in which
moving holidays are based on different calendar systems (Shuja et al., 2007). In
Vietnam, the Lunar year is normally calculated by subtracting approximately one
month from the solar year; for example, an event that takes place in March of the
Lunar year takes place in April of the solar year. Table B4 (Appendix B) presents
moving holidays in Vietnam.
[Insert Table B4 here]
Table B4 shows that there is variation in the month in which a holiday occurs
every year, yet there is no change in the quarter in which the holiday occurs. That
means in the observed period (2000-2011), Vietnamese New Year falls in quarter 1
and King Hungs birthday anniversary is in quarter 2. Therefore, the effect of
moving holidays does not affect quarterly data in this study, and applying seasonal
adjustment for moving holidays is not necessary. If the study was based on monthly
data, however, determining the effects of moving holidays would be essential due to
the changes shown in Table B4.
Apart from the issues mentioned above, applying X12-ARIMA helps to
explain clearly discrepancies. For example, the unadjusted values of GDP at quarter
84

2 are greater than the values at quarter 3, but the adjusted data increase from quarter
2 to quarter 3 is because the seasonal factors in quarter 2 are higher than the quarter
3 values. A similar explanation is appropriate for the PI and PC variables (see Figure
B6, Appendix B).
4.3

DESCRIPTIVE STATISTICS
In this section, seasonally adjusted data is used for some preliminary analysis

for the descriptive statistics


The descriptive statistics of variables during the period 2000 to 2011 are
presented in Tables B5 and B6 (Appendix B) for the foreign and domestic sectors,
respectively.
[Insert Table B5 here]
To examine the effects of external shocks, it is important to identify the role of
fluctuations in variables in the foreign sector. Table B5 shows that the average value
of the Federal Fund rate (FFR), used to consider the foreign monetary shock,
reached around 2.48 percent in the 12-year period from 2000 to 2011. In light of
this, the previous section could be reviewed. This mean value is supported by high
levels over the periods 2000:1 to 2001:3, with the peak at 6.52 percent in 2000:3,
and 2005:2 to 2008:1, with the peaks at 5.25 percent in 2006:3, 2007:1, and 2007:2.
These high levels contrast with the low levels in the remaining periods.

The

reduction in FFR from 2008:1 is consistent with the fact that the Fed and other
monetary authorities injected a considerable amount of money to support liquidity in
the money market as they sought to prevent the credit crisis (Sultan, 2012). The
volatility of foreign output (the YC variable) and international prices (the WRP,
WOP and WGP variables) is less than that of the foreign monetary policy. This is
highlighted by comparing the mean value and the standard deviation of these
variables. While the standard deviation of the first three variables account for less
than two-thirds of the average value, these two values for the FFR variable are
similar (2.12 and 2.48, respectively). Compared to the mean value, the highest
volatility belongs to the Federal Fund rate, followed by the gold price, rice price, oil
price and Chinas GDP.
With regard to the distribution, it is necessary to consider the skewness,
kurtosis and Jarque-Bera statistics. When a variable is normally distributed, the
85

skewness and kurtosis are valued 0 and 1, respectively (Vogelvang, 2005). Positive
skewness suggests a long right tail in the distribution. The kurtosis values of all
foreign variables, except rice price (WRP), are less than the kurtosis of the normal
distribution (3), so they are flat relative to the normal. Meanwhile, the kurtosis of the
WRP variable is peaked relative to the normal. The Jarque-Bera test shows that the
null hypothesis of a normal distribution is not rejected for three series (FFR, YC and
WOP), but is rejected for the WRP variable at all levels. The WGP variable has a
normal distribution at the 1 percent significance level, but not at the 5 percent level
(Microsoft, 2009a)12.
[Insert Table B6 here]
In the domestic sector, comparing the mean value and the standard deviation
in Table B6 suggests that domestic credit (CR) had the highest level of volatility,
followed by M (money supply), VE (exports), VI (imports), VNI (stock price index),
PI (private investment), PC (private consumption), CPI (consumer price index), Y
(domestic output), R (short run interest rate) and E (real effective exchange rate).
The results also reveal that all domestic variables distribution has a right tail due to
positive skewness. Only two kurtosis values (the interest rate and the stock index)
are more than 3; this implies that the distribution is peaked (leptokurtic) relative to
the normal distribution. In econometrics, the leptokurtic is explained as periods of
the high volatility followed by periods of relative stability. Specifically, in this
study, the leptokurtic appearance of the figure relates to volatility in the financial
market (financial variables are interest rate and securities index). Regarding testing
the normal distribution, Jarque-Bera statistics suggest different results. Specifically,
the distribution for Y, E and TM is normal at all levels, and the null hypothesis for
any of the four variables (PI, PC, M and VE) is rejected at 10 percent significance
level, and not rejected at the 5 percent level. Meanwhile, two CPI and CR variables
do not reject the hypothesis of normal distribution at the 5 percent level and reject
this hypothesis at the 1 percent significance level. For the two remaining variables
(R and VNI), the abnormal distribution is evident (Microsoft, 2009a).

12

Under the null hypothesis of a normal distribution, the Jarque-Bera statistic is distributed as 2 with
two degrees of freedom. The critical value is 4.61 at the 10% level, 5.99 at the 5% level, and 9.21 at
the 1% level. The Jarque-Bera statistic exceeds the critical value, so the null hypothesis is rejected.
86

4.4

UNIT ROOT TESTS


In this section, stationary of data is analysed. The main characteristic of a

stationary variable is that its mean and variance do not change over time13. This
characteristic is essential to prevent spurious analyses. When data is not stationary
standard regressions like OLS could result in incorrect results. Therefore, time series
properties of the variables should be considered before conducting estimations. The
first task before conducting regressions is running unit root tests which are dominant
in most studies.
The existence of the unit root received great attention in research beginning
with Nelson and Plosser (1982). A review of unit root tests as well as models
applied in previous studies is summarised in the studies of Glynn et al. (2007),
Chowdhury (2011) and Chowdhury (2012a). The difference between tests could be
from the null hypothesis and structural breaks mentioned in tests (Chowdhury,
2012b). There have been three generations of unit root tests: (i) with no break and
break with a known date (or an exogenous break), (ii) with a single endogenous
structural break (or an unknown break date), and (iii) with multiple endogenous
structural breaks.
Augmented Dickey and Fuller (ADF) is the traditional test for checking unit
root. Tests without structural breaks with higher power include the Elliot,
Rothenberg, and Stock test (ERS), the Ng and Perron test (NgP), and the
Kwiatkowski-Phillips-Schmidt-Shin test (KPSS). Among them, the ADF, ERS, and
NgP tests have the same null hypothesis (that variables are non-stationary), and the
KPSS tests null hypothesis is that the variables are stationary. The ERS test is
known as the generalised least squares (GLS) detrended Dickey-Fuller. In all four
tests, trend and intercept are included in the test equation. This is implicit regarding
the presence of the deterministic components (ao and a2t) in the equation of the test,
as below:
p

yt a0 yt 1 a2t t yt i ut
i 1

where:

the difference of the given variable,

yt:

the series being tested at the time t,

87

p:

the number of lags, and

ut:

the residual is white noise.

Based on this approach, a generalisation of the ADF test procedure is


developed with other tests. With the ADF, ERS and NgP tests, the Schwarz Info
Criterion (SIC) is used to determine the optimal lag length (p).
For tests with one structural break and multiple breaks, the approaches of Lee
and Strazicich (2003; 2004) with minimum Lagrange multiplier (LM) are applied in
this study (LS test).
According to Lee and Strazicich (2004, p2), by combining the two-break unit
root test of Lee and Strazicich (2003) with the one-break test developed here,
researchers can more accurately determine the correct number of breaks. They also
warn that applying previous tests with endogenously determined structural breaks
could create spurious rejections (see Lee and Strazicich (2003) and Glynn et al.
(2007)).
Unit root testing is mentioned in some studies on Vietnam (see Goujon (2006),
Le and Pfau (2009), Nguyen (2010), Nguyen et al. (2012)). All studies, except
Nguyen et al. (2012), use the ADF test as the only approach for testing the unit root.
Nguyen et al. (2012) use the KPSS test along with the ADF test for their study. This
aspect, along with the different time periods covered by the studies, could result in
different results. For gross domestic product (Y), industrial output is used as a proxy,
and it is stationary at I(1) in Goujon (2006), Le and Pfau (2009), and Nguyen
(2010); and I(0) for both the ADF and KPSS tests in Nguyen et al. (2012). For
consumer price index (CPI), the result I(1) is obtained in all studies. For money
supply (M), Le and Pfau (2009), Nguyen (2010) and Nguyen et al. (2012) use M2
for the study and the results are I(1) in Le and Pfau (2009) and Nguyen (2010), and
I(0) in Nguyen et al. (2012). For the interest rate (R), Le and Pfau (2009) and
Nguyen (2010) get the result I(1), while I(0) is recorded for Nguyen et al. (2012).
Similar results are obtained for exchange rate, world rice price, and world old price
in these studies. The last variable, credit (CR), is integrated of order 1 in the study of
Le and Pfau (2009).

13

In case that a variable is nonstationary, it is defined as having a unit root.


88

One of the recommendations by Shrestha and Chowdhury (2005) is to use a


mix of different tests rather than selecting one specific test to obtain better results. In
this study, all eight tests for unit roots are conducted, namely ADF, ERS, NgP,
KPSS, and LS with one break and two breaks in Crash and Break models (see
Tables B5, B6, and B7, Appendix B). However, the LS test with two structural
breaks is preferred in this study. First, tests treating break dates as endogenous are
more popular than tests with exogenous break dates (Narayan & Popp, 2010), and
LS tests are better than previous tests with endogenous breaks (Lee & Strazicich,
2003; Glynn et al., 2007). Second, applying Lee and Strazicichs approaches implies
that the structural-break results should be considered along with important
phenomena in the Vietnamese economy during the period 2000-2011, including
external and internal shocks. Applying LS tests is important in the case of Vietnam
for several reasons. First, there are no prior studies on Vietnam concerning structural
breaks, so spurious results could appear. Second, Vietnam is an emerging economy,
so breaks for the economy are unavoidable. Pahlavani et al. (2006) argue that
determining time for breaks supports analysis on the level of effect (immediate or
gradual) versus breaks on the variables. Third, LS test procedures (known as a thirdgeneration unit root test) are newly applied to empirical studies (Chowdhury,
2012b), so it is interesting to use them for this study on Vietnam.
In applying this method, there are some important differences in estimation.
First, TB is defined by considering all possible structural break periods with the
minimum t-statistics. Another difference with the above traditional tests is using the
trimming region (0.1T, 0.9T) with the T sample size for determining the appropriate
results.
Second, the critical values for assessing results are derived from Lee and
Strazicich (2003; 2004).
Third, applying different software programs requires the use of codes with
different estimation ranges. In this study, the models are tested with the WinRATS
8.10 program, so a common estimation range is defined by maxlag plus 1, which is
different to GAUSSs code equalling the lag plus 114. As the quarterly data is used
for this study, the maximum lag length (kmax) is set to 4. Also, the trimming value
14

See www.estima.com (the RATS web site)


89

equals 0.10 for this study to prevent the information loss of variables. The RATS
procedure (lsunit.src) is used in this study for computing the LS test15. With this
procedure, the number of lags is identified under the effect of choosing the
breakpoints.
Fourth, the LS test could be appropriate for testing multiple breaks, implying
more than two structural breaks. Also, the RATS procedure fits tests with two or
more breaks. However, no studies have followed this approach, as the trend of a
given variable could be comprehensively broken with three and more breaks, which
easily results in inaccurate signals of unit root results. Moreover, in LS tests the
sample size is 100, while this study has 48 observations, so the time is too short to
do more breaks; this could result in spurious findings with variables broken trend.
With the assumption that structural changes could occur frequently in an emerging
economy like that of Vietnam, considering Break model is enough to analyse
these changes. Therefore, testing with two breaks as in Lee and Strazicich (2003;
2004) is considered reasonable for this study.
As shown in Table B9 (Appendix B), there are nine rejections of the null
hypothesis of unit root, implying that nine variables are trend stationary: Y, CPI, PI,
PC, M, R, VNI, VE and VI. Non-stationary variables are five foreign variables and
two domestic variables (CR and E). Table B8 (Appendix B) compares these results
with those from other unit root tests. This work is similar to that mentioned by
Narayan and Popp (2010). This study supports the argument that tests with breaks
have more power than tests without breaks. Perron (1989) supposes that the standard
tests are biased toward non-rejection of the unit root null hypothesis due to not
applying structural-break tests. Moreover, Ben-David et al. (2003) argue that if not
considering breaks, researchers could accept the null hypothesis of unit root by the
ADF test or tests integrating to one break. Obviously, this argument supports for the
opinion of Perron (1989). Based on the comparison of Narayan and Popp (2010)
about the number of rejections of unit root, this study supports this view.
Specifically, the rejections of unit root are 4 for ADF, 7 for ERS, 7 for NgP, 10 for
KPSS, 4 for LS1 (model A), 2 for LS1 (model C), 6 for LS2 (model A), and 9 for
LS2 (model C) (Table B10, Appendix B). Following this understanding, applying

15

Download from www.estima.com/procs_perl/lsunit.src.


90

the LS test with two structural breaks (model C), as supported by the argument of
Perron (1989) and Ben-David et al. (2003) is convincing for the case of Vietnam.
By conducting LS tests, the study obtains the break dates for the variables, as
shown in Table B9. The endogenous break points vary across variables but no break
occurs in the years 2000, 2001, or 2011 for domestic indicators. It is noticeable that
many break dates focus on the 2007-2010 period of the financial crisis (15 out of the
32 structural break dates). However, the study fails to find statistically significant
structural break dates for all variables, and it finds a statistically significant break
only for one variable, Gross Domestic Product at the second quarter of 2008 in
model C and the fourth quarter of 2010 in model A. These dates could coincide with
some important events occurring in the economy: (1) increased trade openness after
Vietnam joined the World Trade Organisation (WTO) in 2007; (2) the financial
crisis from 2007; and (3) the governments policies to stimulate the economy in
2009. If this relevance exists, the structural breaks in the period 2007-2009 are
crucial to changes in the Vietnamese economy. The statistically significant date
suggests that our model needs to include a dummy variable to cover this break date.
This study uses the dummy variable to reflect the important break dates, especially
covering structural changes from the financial crisis.
4.5

CONCLUDING REMARKS
In this chapter, data is analysed via seasonal adjustment. This work is essential

to produce adjusted data appropriate for the estimates in the studies on Vietnams
economy. This has not been done before in studies on Vietnam. Two different
approaches are used in this study, including estimating dummy variables and
applying X12-ARIMA and TRAMO/SEATS procedures for adjusting each time
series. Results from the use of dummy variables show strong evidence for the
existence of seasonal components as well as the quarter 1 effects for GDP. For
investment and consumption, while the existence of seasonal factors is evident, there
is no evidence for the quarter 1s effects. Analysis demonstrated that changes in
mean GDP are not caused by factors such as investment, consumption, or net export
under the measurement for aggregate-demand components. Therefore, government
expenditure could be shown to be the only factor causing variation in average GDP.
However, dummy variables were shown to account for a small proportion of the
91

variation in the variables. With the second approach, the following results were
obtained. First, the multiplicative model is better than the additive in testing
Vietnams GDP, investment and consumption after considering spectrum analysis
and assessment about quality indices of adjustment. However, quality measurement
shows that the fluctuation in private investment is too large, its seasonal adjustment
no longer stable, and its fluctuation not random.
Second, as a whole, applying popular seasonal adjustment procedures like
X12-ARIMA and TRAMO/SEATS produced relatively similar results, especially in
the period 2000-2007. Differences in adjusted data of investment and consumption
are consistent with different results of irregular factors from 2008 for Vietnam. The
study confirms that by applying the relative mean absolute difference (RMAD), the
X12-ARIMA option is preferred for Vietnams data. Third, due to the distinct
characteristics of moving holidays in Vietnam, together with choosing quarterly
data, adjustments for moving holidays in the study period do not affect this study.
Finally, applying X12-ARIMA gives a clearer explanation of the discrepancy
problems than using dummy variables.
Results from unit root tests illustrate that there are nine stationary variables
and seven non-stationary variables in the data set of this study. Stationary variables
include gross domestic product, consumer price index, private investment, private
consumption, monetary aggregate, domestic interest rate, stock price index, exports,
and imports. Two remaining domestic variables in the non-stationary form are
domestic credit and real effective exchange rate. All foreign variables are nonstationary. After differencing non-stationary data, we find that six first-differenced
variables are stationary, in other words, I(1). The only variable in the second
integration, I(2), is FFR. The unit root test with two structural breaks (model C) is
suggested as the most appreciate test for the data set of Vietnam due to its greater
power as well as the usefulness of the test to check break dates in the case of
Vietnam. The statistically significant break dates (2008:2 and 2010:4) found in this
study seems to relate to some important events in the Vietnamese economy such as
joining the WTO, the financial crisis from 2007 and Vietnams stimulation policies
in 2009. Thus, this suggests using a dummy variable in the SVAR model for the
case of Vietnam in the next chapter.

92

CHAPTER 5
THE STRUCTURAL VECTORAUTOREGRESSION MODEL

5.1

INTRODUCTION
The literature review in Chapter 2 revealed that most studies use two main

methods to identify and examine the significance of MTM channels, namely the
vector autoregression (VAR) and structural vector autoregression (SVAR)
procedures. The analyses in Chapter 2 also showed that SVAR models are superior
to VAR models, and that they are useful in identifying parameters from the model
and recover structural shocks in analysing simultaneous interaction of variables.
Therefore, this chapter presents the theoretical framework about modelling SVAR.
The specification of SVAR models present important issues: should the
structural model be recursive or non-recursive, and should it apply short run or long
run restrictions? Each of these specifications requires different procedures, which
will be discussed in this chapter.
The analysis of the MTM for Vietnam was conducted using a two-stage
approach. First, a basic SVAR model of the Vietnamese economy was constructed
to examine the role of monetary policy and process of the MTM including the
interest rate, credit, exchange rate and asset price channels. In this stage, the study
examined the interaction of variables; specifically, the effects of foreign variables on
domestic variables and the effects of monetary policy on domestic variables. Next,
the base models were extended by adding trade variables to examine further the
international dimension of the MTM.
The rest of this chapter is structured as follows. Section 5.2 briefly reviews the
theoretical SVAR modelling procedure. Section 5.3 gives a brief overview of the
data used in the estimation of the models. Section 5.4 discusses the benchmark
models, whilst Section 5.5 proposes the preferred models for the case of Vietnam.
The last section summarises the discussion.

93

5.2

THEORETICAL SVAR MODEL


The proposed equations used in the model are:
A0 yt = c0 + A1 yt-1 + A2 yt-2 +...+ Ap yt- p + D+vt

(5.1)

This system is a basic SVAR model, where: yt is an (n x 1) vector of


endogenous variables, c0 is a (n x 1) vector of constants, Ai is a (n x n) matrix (i =
0,, p) of structural parameters, D is a (n x 1) vector of exogenous variables and

vt is a (n x 1) structural innovation, assumed to be orthogonal, and uncorrelated.


Constructing the structural matrix is very important and depends on two
issues: (i) whether the model is represented as recursive or non-recursive, and (ii)
whether short run or long run restrictions are imposed.
For the first issue, the causation relationship of variables helps to identify
whether the SVAR model is recursive or non-recursive. In a recursive model,
different disturbance terms are uncorrelated, and there is a unidirectional causation
among variables. The structural matrix for a recursive model is defined as a lowertriangular matrix. Meanwhile, in a non-recursive model, causation is reciprocal (n
directional) and the residuals are correlated. According to Berry (1984), nonrecursive models should be used rather than recursive models because it is not
realistic to assume that there is no reciprocal causal relationships between variables
in the model. Kim and Roubini (2000) supported this view and further argued that a
non-recursive approach is useful in identifying structural shocks, including monetary
policy shocks. This approach also helps to avoid the limitations associated with the
assumptions of the structural ordering of the variables in a recursive approach16.
The second important issue relating to SVAR identification is the imposition
of short run or long run restrictions. In the short run approach, depending on which
matrix is defined as an identity matrix I, there are three different possible models: (i)
the AB restriction type if A and B are different to I, (ii) the A restriction type if B = I
and (iii) the B restriction type if A = I. Amisano and Gianini (1997) use the AB form
for imposing short run restrictions. The residuals of this model are assumed to have
a linear relationship to structural shocks ( t ), so vt B t , where B is a (n x n)
diagonal matrix. A SVAR model could be changed to a reduced VAR, with the

94

restrictions identified on the A and B matrices different to zero. A second restriction


is proposed by Blanchard and Quah (1989) when estimating potential output. In this
approach, the accumulated long run response C to structural innovations is in the

form: C A1B, where: ( I A1 .... A p )1 is the estimated accumulated


response to the observed reduced-form shocks. In the case where Cij = 0, the
accumulated response of the ith variable to the jth structural shocks is zero in the long
run (Microsoft, 2009b, p474).
The traditional VAR procedure estimates the parameters of the reduced form
of the model using OLS, and these values are then used to calculate the structural
form parameters. The coefficients can be estimated via the ordinary least squares
(OLS) method. The structural coefficients can then be calculated from the reducedform coefficients (Gujarati & Porter, 2009).
Alternatively the SVAR procedure restricts the A0 matrix according to
economic theory. Indirect least squares (ILS) is used to estimate the parameters in
two stages:
Stage 1: A SVAR model is put into a reduced VAR form by multiplying both sides
of (5.1) by the inverse matrix A0-1:

A-10 A0 yt A0-1c0 A0-1A1 yt 1 A0-1A2 yt 2 ... A0-1Ap yt p A0-1B t

(5.2a)

yt A-10 c0 A0-1A1 yt 1 A0-1A2 yt 2 ... A0-1Ap yt p A0-1B t

(5.2b)

The reduced VAR form is:


yt C0 C1 yt 1 C2 yt 2 ... Cp yt p ut

where:

(5.3)

C0 A01c0 , Ci A01 Ai (i = 1, , p) and ut A01B t

This VAR can be estimated by maximum-likelihood estimation (MLE) to obtain the

VAR residuals (u t ) .
Stage 2: Identifying the contemporaneous matrix A0 .
The reduced form residuals ut = A0-1 Bt , can be represented in the structural
form A0ut = Bt or A0ut = vt and all that remains isto identify the system. The total
16

Kim and Roubini (2000) give further details about limitations of the recursive approach.
95

number of elements of the matrices A0 and B are 2n2; and the number of parameters
in the reduced form is n(n+1)/2. Thus, the system requires at least [2n2 n(n+1)/2] =

[n(3n-1)/2] additional restrictions to identify A0 and B .


The specification of matrices A0 and B for the case of Vietnam are discussed in
Section 5.5. The next section considers complications of the specification of the
variables, whilst Section 5.4 reviews the benchmark models.
5.3 APPROPRIATE DATA FORM FOR SVAR MODELS
In the literature, there are two different approaches in estimating SVAR
models with non-stationary data. The first, it is necessary to transform variables that
are non-stationary to a stationary form by taking first differences in their level
values (see Du et al. (2010), Nguyen (2010), Narayan et al. (2012) and Seema
(2013)). This is necessary to avoid the risk of finding spurious relationships between
non-stationary I(1) variables treated as stationary. Moreover, transforming nonstationary variables to stationary variables is necessary to correctly analyse the
impulse response functions (IRF) and the forecast error variance decompositions
(FEVD). Nguyen (2010) and Seema (2013) and other studies characterise all
variables as non-stationary, whereas Du et al. (2010) and Narayan et al. (2012) and
others 17 transform any non-stationary variables to first (or second) difference
stationary variables.
The second approach contends that it is more useful to use variables in levels
in SVAR estimation rather than using the differencing approach. This is argued and
practised by Sims (1992), Cushman and Zha (1997), Bernanke and Mihov (1998),
Kim and Roubini (2000), Afandi (2005), Berkelmans (2005), Aslanidi (2007),
Raghavan and Silvapulle (2008), Dungey and Pagan (2000; 2007), Nakahira (2009),
Bicchal (2010), Zaidi and Fisher (2010) and Sharifi-Renani (2010). The main reason
is that using first differences leads to loss of information about the long run
relationships. Another reason is the purpose of SVAR (and VAR) is to analyse the
mutual relationships between variables, rather than estimating parameters. This
implies the presence of non-stationary variables do not affect statistical inference
(Kim & Roubini, 2000). Moreover, including lagged values in the VAR model could
17

These two cases are understood as: all pure I(0) or a mix of I(0) and I(1).
96

remove the non-stationarity effects of the variables and help the residuals become
stationary.
However, the use of variables in levels raises two concerns, the exclusion of a
possible long run cointegrating relationship and the risk of finding spurious results
(Bernanke & Mihov, 1998; Berkelmans, 2005; Aslanidi, 2007). Cointegration tests
help to examine the first issue before conducting VAR estimation, but considering
the second issue seems to be more difficult as it depends on the correlation of data
and appropriateness of assumptions about the variables relationships. Aslanidi
(2007) and Berkelmans (2005) have noted that spurious relationships between the
I(1) variables is still a problem in the second approach. Moreover, Zaidi and Fisher
(2010) argue another difficulty in applying this approach is that many variables have
unit roots, which results in a very large forecast horizons, in turn requiring cautious
interpretation of variance decomposition results.
The unit root tests reported in the previous chapter with and without breaks
using seasonally-adjusted data produced mixed I(0) and I(1) results. With the second
approach, some doubts about non-stationary data appear in estimating SVAR.
Therefore, the first approach is applied because this study only focuses on
examining the temporary effects, so the concerns about the loss of long run
information by differencing non-stationary data are solved. Moreover, our study
uses short run restrictions to examine temporary influences in the short and medium
term.
5.4 BENCHMARK MODELS
This section provides an overview of selected benchmark models including the
models of Cushman and Zha (1997) for Canada, Kim and Roubini (2000) for nonUS G-7 countries, and Afandi (2005) for Indonesia. These models have been
selected as benchmarks for the design of a specific model for the Vietnamese
economy in the next section, for the following reasons. First, these studies examined
small, open economies which include two sectors (foreign and domestic), relevant to
modelling a small, open economy like Vietnam, subject to exogenous shocks in
foreign interest rates, foreign output and world prices. Secondly, these studies
assume flexible exchange rate regime, which Vietnam has operated under since
1999. Finally, the studies specify the SVAR model as non-recursive include three
97

segments of the domestic economy: the production, the money market, and the
financial market 18 . Moreover, all models are over-identified and non-recursive.
Therefore, these benchmark models create useful references in modelling and
constructing the interaction among variables.
The contemporaneous matrix ( A0 ) is specified with three possible values:
cells with zero indicate no effects exist between the two variables (or the lagged
effect), cells including unity and cells containing the contemporaneous parameters
(for example a2,1 implies the immediate influence of the first variable on the second
variable).
5.4.1

The empirical study of Cushman and Zha (1997)

Cushman and Zha (1997) constructed an 11-variable model to identify


monetary policy in Canada as an example of a small open economy with a flexible
exchange rate regime.
Table 5.1: The SVAR model of Cushman and Zha (1997)
WP

FCPI

WP

FCPI

a2,1

FY

a3,1

a3,2

FFR

a4,1

FY

FFR

CPI

VE

a9,10

a9,11

a10,9

a11,9

a11,10

a5,5

VE

a6,5

a6,6

VI

a7,5

a7,6

a7,7

a9,1

a9,4

M
a11,1

a3,4

a2,4

CPI

VI

a11,2

a11,3

a11,4

a10,5

a10,6

a11,5

a11,6

a11,7

a11,8

Notes: WP = World total exports commodity price index, FCPI = U.S. consumer price index,
FY = Foreign output (the U.S. industrial production), FFR = Federal fund rate, CPI = Consumer price
index, Y = Gross domestic product, VE = total exports, VI = total imports, R = Short-term interest
rate, M = Monetary aggregate, and E = the exchange rate (US dollar price of the Canadian dollar).
Source: Cushman and Zha (1997).

18

Cushman and Zha (1997) include the production sector, the money market, the information market.
98

The four foreign sector variables are world total exports commodity price
index, U.S. consumer price index, U.S. gross domestic product, and the Federal
Fund rate.
The domestic sector is divided into three segments: the production market,
with four variables (the home industrial production, consumer price index, total
imports, and total exports), the money market with two variables (short-term interest
rate and monetary aggregate), and the financial market with one variable (exchange
rate).
There are some meaning assumptions in the study of Cushman and Zha
(1997). First, in the production sector, while the output is assumed to be affected
contemporaneously by price level, the price variable only influences the output
variable with the lagged effects. Moreover, other variables have no effect, or affect
through lag, on the output and price variables. Both total imports and total exports
have simultaneous effects only on other variables in the production market, implying
the exclusion of effects from variables of other markets on exports and imports.
Cushman and Zha (1997, p438) explained that these lagged effects reflect trade
contracts and advance production planning. Second, in the money market, the form
of the money demand equation is determined by four variables: M (the money
supply), CPI (price), Y (the output) and R (the interest rates), which is common in
monetary studies, such as that of Cushman and Zha (1997). Meanwhile, the money
supply equation is specified based on the availability of information to the monetary
authority in the time period of the study. In the case of Canada, the information
available for the Central Bank in a month includes WP (the world oil price), FFR
(the U.S. Federal Funds Rate), R (the interest rate), M (the money supply) and E (the
real effective exchange rate). Third, the exchange rate equation contains all
variables, reflecting other information of indirect effects on the identification of
monetary policy (Cushman & Zha, 1997).
5.4.2

The empirical study of Kim and Roubini (2000)

To study the effects of monetary policy in open economies, Kim and Roubini
(2000) constructed a non-recursive SVAR model for non-U.S. G-7 countries.

99

Table 5.2: The SVAR model of Kim and Roubini (2000)


WP
WP

FFR

a2,1

CPI

a3,1

a4,1

a51

FFR

a71

a34

a56

a57

M
E

CPI

a72

a63

a64

a65

a73

a74

a75

a76

Note: WP = World oil price, FFR = Federal fund rate, CPI = Consumer price index, Y =
Industrial production, R = Short-term interest rate, M = monetary aggregate, E = the exchange rate

(units of home currency for one unit of U.S. dollars).


Source: Kim and Roubini (2000).

The foreign sector comprises two variables, the world price of oil and the U.S.
Federal Funds rate. Similar to Cushman and Zha (1997), the price of oil is assumed
to contemporaneously affect the U.S. Federal Funds rate because of the role of oil
prices in implementing the U.S. monetary policy. Kim and Roubini (2000) argued
that the U.S. monetary contraction is an effort to respond to oil price related
inflationary shocks. The domestic sector is characterised using five equations
relating to three markets: the production market, with two variables (home industrial
production and consumer price index), the money market, with two variables (shortterm interest rate and monetary aggregate), and the financial market, with one
variable (exchange rate). This model is smaller than the Cushman and Zha (1997)
model as it excludes foreign price level and foreign output from the foreign sector,
as well as total imports and total exports from the domestic-production sector.
The study of Kim and Roubini (2000) includes important assumptions. In the
production sector, similar to Cushman and Zha (1997), variables such as FFR (the
U.S. Federal fund rate), R (the short-term interest rate), M (the money supply) and E
(the exchange rate) are assumed to affect Y (the output) and CPI (price) with a lag.
However, the two models are specified differently: both output and price level are
contemporaneously affected by world oil price because of the importance of oil
price in the production sector. Moreover, the output is assumed to have immediate
100

effects on the price level, but, in contrast to Cushman and Zha (1997), it is only
affected by lagged prices. In the money market, the assumptions for the moneydemand equation follow the common approach that includes M, CPI, Y and R. In
terms of the money supply equation, Kim and Roubini (2000) also apply an
approach based on the information set available to monetary authorities. However,
Kim and Roubini (2000) confirm the difference in the availability of monthly data
and quarterly data that the monetary authority can observe and react. Although FFR
is available within a month, this variable is still excluded from the equation because
it can give no additional information for non-U.S. monetary authorities. Last, Kim
and Roubini (2000) follow the popular assumption for the open economies, that all
variables affect contemporaneously the exchange rate.
5.4.3

The empirical study of Afandi (2005)

Afandi (2005) used the benchmark model of Kim and Roubini (2000) to
construct a nine-variable SVAR model for Indonesia. The contemporaneous matrix
for this economy is reproduced below.
Table 5.3: The SVAR model of Afandi (2005)
WP
WP

FFR

a2,1

FFR

a3,4

R1

a4,1

a7,3

a5,7

1
a6,5

a7,4

a9,2

a9,3

a7,6

a9,4

a9,5

a5,9

a8,4

CR
a9,1

CR

R2

a3,9

a5,2

R1

R2

CPI
Y

CPI

a9,6

1
a8,7

a9,7

a9,8

Note: WP = World oil price, FFR = Federal fund rate, CPI = Consumer price index, Y =
Industrial production, R1 = Interbank call rate, R2 = Interest rate on capital loan, M = Monetary
aggregate, CR = Domestic credit, and E = Exchange rate (units of Indonesian currency for one unit

of U.S. dollars).
Source: Afandi (2005).

101

Whilst most assumptions in Afandi (2005) are the same as those in Kim and
Roubini (2000), there are some notable differences. Specifically, in the production
market, Afandi (2005) argues that as the Indonesian government controls domestic
oil prices, the world oil price could not contemporaneously affect domestic price
level. Moreover, the exchange rate is assumed to have an instantaneous influence on
consumer prices because it is one of the crucial sources of inflation (Afandi, 2005).
In the money market, again because of government control of domestic oil prices,
the world oil price is not included in the money demand equation. However, unlike
the study of Kim and Roubini (2000), FFR (the U.S. Federal Funds Rate) appears in
this equation. This is necessary because the data set in Afandi (2005) includes
monthly and quarterly data. In the financial market, Afandi (2005) considers the
credit equation to examine the effects through the credit channel with the
contemporaneous effects of Y (output) and M (money supply). This study constructs
a SVAR model for the case of the Vietnamese economy in the next section, based on
the three benchmark models.
5.5 MODEL DESIGN FOR VIETNAMS ECONOMY
This study will use a base model (with two versions) and two extended
models.
(1) Two versions of the basic SVAR models (models VN1 and VN2) are used
to examine foreign shocks and domestic monetary policy shocks to the Vietnamese
economy. The short run restrictions are applied and non-stationary variables are
differenced to obtain stationarity. A difference between these SVAR models is
Vietnams administered oil pricing regime, reflecting the strong or weak
administration; therefore, world oil price shocks could have their lagged effects or
the contemporaneous effects.
(2) The SVAR model is extended to include trade related variables to examine
further the international transmission of a monetary contraction to exports and
imports.
(3) Another extension of the SVAR model is to include expenditure behaviour
variables, to examine the effects of a monetary contraction on private investment
and consumption.

102

The basic SVAR model with short run restrictions for the case of Vietnam is
based the AB restriction type of Amisano and Giannini (1997). Links between
innovations ( ut ) and structural shocks ( t ) are denoted in the following system with
the zero coefficients (blank cells) illustrating possible lags in the relationship
between variables, and the non-zero coefficient (aij) representing that the variable i
contemporaneously affects the variable j.
The structural matrix is specified as a non-recursive model of the Vietnamese
economy. This approach makes assumptions about the reciprocal relation of
variables and nonzero correlation of error terms when considering this economy.
Compared to the benchmark models, this study develops a larger SVAR model,
where additional variables include: WRP (world rice price) and WGP (world gold
price) in the foreign sector, and VNI (Vietnamese stock price index) in the financial
market. Chinese gross domestic product and the real effective exchange rate (REER)
are proxies for FY (the foreign output) and E (the exchange rate), respectively.

uWP
WP

uWRP
WRP
uWGP
WGP

u FY
FY
u

FFR
FFR

uCPI
CPI
A.
B.

Y
Y

uR
R

M
M
u

CR
CR
uVNI
VNI
u

E
E

(5.4)

The matrix B is a diagonal matrix with the values bij on the diagonal line. The
contemporaneous matrix A0 in (5.4) is illustrated below.
103

Table 5.4: The contemporaneous matrix (the A matrix)


WP
WOP

WRP WGP

FFR

CPI

a6,7

a7,6

CR

VNI

WRP

WGP

FY

a4,1

FFR

a5,1

CPI
Y

FY

1
1
a6,2

a7,1

a7,2

a7,4

a7,10

a8,3

a8,5

a8,6

a8,7

a8,9

a9,3

a9,5

a9,6

a9,7

a9,8

a10,5

a10,6

a10,7

a10,8

a11,5

a11,6

a11,7

a11,8

a11,9

a11,10

a12,5

a12,6

a12,7

a12,8

a12,9

a12,10

a12,11

CR
VNI

a11,1

a11,2

a11,3

a12,1

a12,2

a12,3

a12,4

a8,10

a8,12
a9,12

Note: WP = World oil price, WRP = World rice price, WGP = World gold price, FY =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Short run interest rate, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real
effective exchange rate.

The time period of this study includes the 2008-2010 financial crisis and the
significant break dates identified in the previous chapter, so it is necessary to use the
dummy-variable approach. However, unlike Afandis research, only one dummy
variable should be used in estimating the SVAR model. This is because the time
period for studying from 2008 to 2011 could be quite short for setting two dummy
variables. The assumptions are described as follows.
The first five equations relate to the foreign sector, including the equations for
the world oil price, world rice price, world gold price, Chinese gross domestic
product, and the U.S. Federal Fund rate. As Vietnam is considered as a small, open
economy, the study applies the assumptions of Cushman and Zha (1997) and Kim
and Roubini (2000). Specifically, the foreign variables are not affected by domestic

104

shocks, while the foreign shocks may contemporaneously influence domestic


economic fluctuations. However, as the study uses Chinese gross domestic product
as foreign output, the WOP (world oil price) variable appears in the FY (foreign
output) equation due to the importance of this product for all economies, as in Kim
and Roubini (2000). The four remaining foreign variables are expression by the
following equations.

uWOP = b1,1WOP

(5.5)

uWRP = b2,2 WRP

(5.6)

uWGP = b3,3 WGP

(5.7)

a4,1WP +uFY = b4,4 FY

(5.8)

Following Cushman and Zha (1997) and Kim and Roubini (2000), the study
assumes the world oil price (WOP) is included in the U.S. interest rate (FFR)
equation as U.S. monetary policy is tightened to respond to oil inflationary shocks.
Hence, the U.S. interest rate equation appears as:

a5,1uWP +uFFR = b5,5 FFR

(5.9)

The next two equations represent the production sector with the two most
important variables: CPI (the consumer price index) and Y (the domestic output).
Like the benchmark studies, this study uses the assumption about the information
lag in the output and prices equations, since this is appropriate for the case of a
developing economy like Vietnam. In particular, the output (Y) and price (CPI) are
not changed contemporaneously when domestic and foreign monetary policy
variables change. This is because of information delays and an undeveloped
financial structure, which together result in the slow reaction speed of monetary
policy in Vietnam. Thus, the variables FFR (the U.S. Federal Funds Rate), WGP
(the world gold price), R (the interest rate), M (the money supply), CR (the bank
credit), VNI (the stock price index), and E (the real effective exchange rate) are
excluded in the equations of output and price, except that the CR variable appears in
the output equation and the E variable in the price level equation. The first exception
is explained by the fact that Vietnamese economic growth is mainly supported by
bank credit, as explained in Chapter 3, so output responds to bank credit. Such an
approach is similar to the study of Safaei and Cameron (2003) and Berkelmans
105

(2005). The second exception is the assumption of Afandi (2005) that the exchange
rate contemporaneously affects the price level, which is different to the studies of
Cushman and Zha (1997) and Kim and Roubini (2000). Afandi (2005) argued that
because tradable goods are important components in calculating the consumer price
index, the price level is assumed to respond instantaneously to the exchange rate
shocks. Moreover, such an approach is advisable since the impact of exchange rate
shocks on inflation is one of the leading concerns for small, open economies (Kim &
Roubini, 2000). As rice is one of the main export products of Vietnam, the variable
WRP appears in the equations for both Y and CPI.
Unlike their behaviour in the benchmark models, the output and consumer
price index in this study are assumed to affect each other contemporaneously. First,
the instantaneous effect of output on price is from the concern of economic growth
and inflationary pressure, so policymakers take into account the output variable in
the price-reaction equation (Kim & Roubini, 2000; Afandi, 2005). Berkelmans
(2005) argues this specification is common in empirical studies. Second, the price
variable has an immediate effect on output. This assumption is based on the LucasPhelps imperfection information model, which assumes that producers tend to
increase their production when observing price increases. Such increases could be
from changes in the relative price or the aggregate price level, and the effects on
production are different, depending on each reason. However, the fact producers
cannot know exactly the reason for the increase in price leads to a rise in output. A
change in the relative price alters the optimal amount to produce. A change in the
aggregate price level, on the other hand, leaves optimal production unchanged
(Romer, 2012, p. 292).
In contrast to Kim and Roubini (2000), this study considers the oil price
controlling regime in reflecting the contemporaneous or lagged effects of world oil
price shocks on the domestic economy. Afandi (2005) argues that the Indonesian
government has adjusted the domestic oil price under its separate oil price
controlling regime; thus, the world oil price does not appear in the price equation as
a contemporaneous effect. Due to the important role of this input on the economic
sectors, the oil price still instantaneously affects output (Afandi, 2005). This is
similar to the regime of controlling domestic oil price in Vietnam. This is because

106

oil is one of the strategic goods for developing the Vietnam economy, so the
government has implemented policies to stabilise retail prices for oil products.
Though it is common to include the world oil price (WP) variable in the price (CPI)
equation, as in Kim and Roubini (2000), and Raghavan and Silvapulle (2008), this
assumption is not appropriate for Vietnam, where the domestic oil price is controlled
by the government. An increase in WP (world oil price) can have lagged effects on
CPI (price). However, in contrast to Afandi (2005), this study tests two assumptions
about Vietnams administered oil pricing regime. The two versions are a strongly
administered oil pricing (SAOP) or a weakly administered oil pricing (WAOP). A
WAOP regime indicates low efficiency in the administered oil-pricing regime. With
each regime, appropriate specifications are used. Specifically, for SAOP, the study
applies the approach of Afandi (2005) with the exclusion of WP (world oil price) in
the CPI (price) equation. However, Kim and Roubinis (2000) inclusion of WP in
the CPI equation is acceptable for a WAOP regime. These cases are denoted in
different versions. The contemporaneous impacts of WP on Y (the output) are
explained by the fact that an increase in WP directly affects Vietnamese output, as
Vietnams oil imports and exports are important. During the period 2000-2011,
exports of oil were 178,582 thousands of tons and imports were 131,172 thousands
of tons19. Therefore, shocks in world oil prices immediately affected these export
and import activities and real output.
In addition, only the benchmark model of Cushman and Zha (1997)
considered the shocks of foreign output. However, Cushman and Zha (1997)
assumed that the U.S. output has no contemporaneous effect on Canadian economic
variables, except the exchange rate. Unlike Cushman and Zha (1997), this study
accepts the opposite assumption to examine the significant role of China as one of
Vietnams most important trade partners. Berkelmans (2005) confirmed that
transmitting outside inflationary pressure is normally reflected via domestic
fluctuations, so there is no immediate effects of foreign output on domestic price
level. Therefore, the FY variable does not appear in the CPI equation.
With the above assumptions, the variables WRP (world rice price) and Y
(output) contemporaneously affect CPI when considering an SAOP regime, denoted
19

See http://www.gso.gov.vn/default.aspx?tabid=433&idmid=3
107

in equation 5.10. However, for a WAOP regime, the study adds the WP variable, as
in equation 5.11.

a6,2uWRP +uCPI + a6,7 uY = b6,6 CPI

(SAOP regime)

(5.10)

a6,1uWP +a6,2uWRP +uCPI + a6,7 uY = b6,6 CPI

(WAOP regime)

(5.11)

The variables WP (world oil price), WRP (world rice price), FY (foreign
output), CPI (consumer price index), and CR (credit) immediately influence Y
(output) in equation 5.12.

a7,1uWP +a7,2uWRP +a7,4uFY +a7,6 uCPI +uY +a7,10uCR = Y

(5.12)

The two money market equations are for the money supply and money
demand. The money supply equation is a Taylor style reaction function of the
monetary authority. As this study accepts the benchmark studies assumption that
this equation reflects the information available to the monetary authority, it includes
the variables CPI (the consumer price index), R (the interest rate), M (the money
supply), CR (the credit) and E (the real effective exchange rate) (Cushman & Zha,
1997; Kim & Roubini, 2000; Afandi, 2005). The exclusion of output from the
money reaction equation is explained by the assumption about the information lag
for the monetary authority. Such an approach is applied not only to monthly data
(for example, the benchmark models) but also to quarterly data (Safaei & Cameron,
2003; Berkelmans, 2005). Although the current study uses quarterly time series, it
does not apply this assumption to Vietnam for several reasons. First, this approach is
more appropriate for monthly data than quarterly (Kim & Roubini, 2000). Second,
as mentioned in Chapter 4, the State Bank of Vietnam (Vietnams monetary
authority) is a part of the government and follows the growth target in implementing
its monetary policy. Therefore, the output variable is available for the State Bank of
Vietnam. The specifications for the money-demand equation based on the theory of
money demand according to the LM equation are the common approach in
numerous studies, such as Cushman and Zha (1997), Kim and Roubini (2000),
Afandi (2005), and Raghavan and Silvapulle (2008).
Along with the above assumptions, some additional specifications are applied
in two money market equations. Tran (2009) concluded that the Vietnamese
monetary authority should take into account the gold price in conducting monetary
108

policy. Therefore, to examine the role of the gold price, this variable is included to
examine contemporaneous relationship to the interest rate and the monetary
aggregate. Moreover, as mentioned in Chapter 3 that the dollarisation in Vietnam is
quite high, so it is necessary to consider the opportunity cost of holding the domestic
currency (the interest rate - R) and the foreign currency (the U.S. Federal Funds Rate
FFR). Thus, the effects of foreign interest rate and the exchange rate are included
in both money market equations.
In addition, the world oil price is assumed to be excluded in the money supply
equation as in Afandi (2005) due to the oil price control of Vietnamese government.
Equation 5.13 expresses the above assumptions, with the contemporaneous
effects of WGP (the world gold price), FFR (the U.S. Federal Funds Rate), CPI (the
consumer price index), Y (the output), R (the interest rate), M (the money supply),
CR (the bank credit) and E (the real effective exchange rate):

a8,3uWGP +a8,5uFFR +a8,6 uCPI +a8,7 uY +uR +a8,9uM +a8,10uCR +a8,12uE = b8,8 R
(5.13)
Equation 5.14 denotes the money demand equation, with the contemporaneous
influences of WGP (the world gold price), FFR (the U.S. Federal Funds Rate), CPI
(the consumer price index), Y (the output), R (the interest rate), M (the money
supply) and E (the real effective exchange rate):

a9,3uWGP +a9,5uFFR +a9,6 uCPI +a9,7 uY +a9,8uR +uM +a9,12uE = b9,9 M (5.14)
Finally, the financial market includes three equations relating to credit, stock
(asset) price, and the exchange rate. These are important financial variables, which
move the market towards equilibrium. The inclusion of these variables helps to
examine different channels of the MTM for Vietnam.
Equation 5.15 expresses the responses of credit to shocks from FFR (the U.S.
Federal Funds Rate), CPI (the consumer price index), Y (the output) and R (the
interest rate).

a10,5uFFR +a10,6 uCPI +a10,7 uY +a10,8uR +uCR = b10,10 CR

(5.15)

Afandi (2005) argued that the output and the interest rate instantaneously
affect credit demand. This assumption is correct, as one of the main factors of
demand credit is expectations about the economy, so borrowers always observe the
109

current activity (output - Y) and loans cost (interest rate - R) to develop their
expectations about future economic conditions when making their decision about
bank loans (Berkelmans, 2005)). Moreover, Berkelmans (2005) argued that
borrowers always consider the real cost of loans, which incorporates the interest rate
and the price level, before deciding to borrow from banks, so credit
contemporaneously responds to price shocks. Moreover, due to the highdollarisation issue of Vietnam as mentioned in Chapter 3, the foreign interest rate
(FFR) is assumed to instantaneously affect credit. Furthermore, as in Afandi (2005),
the credit equation excludes the presence of a monetary aggregate; this differs from
the assumption of Safaei and Cameron (2003). In a study on the credit channel in
Canada, Safaei and Cameron (2003) concurred that money shocks affect banks
deposits, and afterwards resulting in effects on credit; therefore, the influence of
monetary aggregate innovations on credit is lagged.
Equation 5.16 presents the stock (asset) price channel and this channel has
been mentioned in the benchmark models.

a11,1uWP + a11,2uWRP + a11,3uWGP + a11,5uFFR + a11,6 uCPI + a11.7uY + a11,8u R + a11,9uM


+a11,10uCR +uVNI + a11,12uE = b11,11VNI

(5.16)

Trading activities in the stock market affect decisions of institutional and


private investors. In Vietnam, the demand for equities is affected by real income, so
stock price contemporaneously respond to shocks in output and price level.
Moreover, because fluctuations in the stock market relate to the publics
expectations about economic conditions, changes in domestic and foreign variables
are transmitted quickly to investment behaviour in the stock market. Therefore,
stock prices are assumed to respond contemporaneously to all variables except
foreign output and real effective exchange rate. The first exception is because
Chinese GDP is the final target of Chinas economy, so it could be unavailable for
investors decision process in time to have an immediate effect on the stock price.
Similarly, the real effective exchange rate is calculated from the basket of Vietnams
many trading partners, so the E variable (real effective exchange rate) affects the
stock price with the lag. A similar approach to excluding the exchange rate in the
stock price equation was taken by Zaidi (2011).

110

Although Vietnamese authorities have applied the regime to administer the oil
price, the stock market reflects investors behaviour about future expectations
including possible impacts of shocks in world oil price. Moreover, Narayan and
Narayan (2010) found evidence for the significant effect of world oil price (WP) on
share prices (VNI). Therefore, the WP variable is included in equation 5.16 as shown
by coefficient a11,1:
Last, the exchange rate is defined as a forward-looking asset price, as in the
benchmark models, it is assumed to respond contemporaneously to all variables in
the study, as shown in the following equation:

a12,1uWP + a12,2uWRP + a12,3uWGP + a12,4uFY + a12,5uFFR + a12,6 uCPI + a12,7uY


+a12,8uR + a12,9uM + a12,10uCR + a12,11uVNI ++u E = b12,12 E

(5.17)

5.6 CONCLUDING REMARKS


This chapter has presented the major issues on the SVAR model. Moreover,
issues relating to the data to be used to estimate the SVAR model were discussed
with two different approaches. The first is using all data in stationary form
(transforming non-stationary data to stationary data before estimation via
differencing), and the second is using variables in level form (a mixture of stationary
and non-stationary data).the second is. This study applies the first approach with the
focus on examining the temporary effects through using short run restrictions to
study the short and medium term.
Next, the chapter reviewed the benchmark models of Cushman and Zha
(1997), Kim and Roubini (2000), and Afandi (2005). All three models consider
aspects of small, open economies with two sectors (foreign and domestic), three
markets in the domestic sector (the production market, the money market and the
information, or financial market.
Based on the benchmark models, the SVAR models were specified for the
case of Vietnam with two sectors: the foreign sector (the world oil price, world rice
price, world gold price, Chinas GDP, Federal funds rate) and the domestic sector
(the consumer price index, Vietnams gross domestic production, exports, imports,
private investment, private consumption, short term interest rate, money supply,
bank credit, stock price index, real effective exchange rate). Characteristics of the
111

Vietnamese economy were discussed and included in the restrictions of the proposed
models, such as the dependence of Vietnamese monetary authority, the administered
oil-pricing regime, the national food-security policy, the relationship between the
gold price and the monetary policy.
The next chapter presents the estimation results based on the above
assumptions and specifications. It also includes analysis of the contemporaneous
parameters, the impulse responses, and variance decomposition. These analyses help
to understand relationships between variables, including the contemporaneous
effects, the reaction to a shock and sources for fluctuations of a variable.

112

CHAPTER 6
ESTIMATION RESULTS AND ANALYSIS

6.1

INTRODUCTION
Chapter 5 proposes two base non-recursive SVAR models and two extensions

that are appropriate to the characteristics of the Vietnamese economy. Based on


these models, this chapter presents the estimation results and analyses about the
relationship between variables in the case of the Vietnamese economy. The research
methodology is conducted via the following steps: (i) identifying lag length and
ensuring conditions about VAR/SVAR stability; (ii) estimating coefficients in the
contemporaneous matrix; (iii) discussing results of impulse response functions; and
(iv) discussing results of variance decomposition.
The number of lags in the VAR/SVAR could change the effects shown in the
model, so identifying the lag length is necessary before conducting the next steps in
the SVAR procedure. Several selection criteria are used to identify lag length,
including LR - sequential modified LR test statistic (each test at 5 percent level);
FPE - Final prediction error; AIC - Akaike information criterion; SC - Schwarz
information criterion; and HQ - Hannan-Quinn information criterion. Results using
such criteria usually show different lag lengths, but the optimal lag length should be
chosen appropriately (Berkelmans, 2005). In order to check the stability condition of
the VAR/SVAR model, the AR roots test and the Lagrange multiplier (LM) test are
conducted. While the AR roots test is used to identify whether eigenvalues lie in the
unit circle, implying whether the VAR/SVAR satisfies the stability condition, the
LM test determines whether there is serial correlation in disturbances in the model.
Analysing and interpreting the results of the contemporaneous matrix, impulse
response functions (IRF), and forecast error variance decomposition (FEVD)confirm
the interaction of various shocks and variables (Kilian, 2011). The contemporaneous
matrix gives the results of the instantaneous effects in the model; it helps identify the
statistically significant coefficients denoting meaningful relationships between the
variables in the model. Next, the impulse response functions help to analyse how
shocks or innovations of one variable affect other variables with or without the
given lag. Because the aim of the study is applying SVAR to obtain a better
113

understanding of the structural relationship between the innovations and variables,


the structural impulse response functions are conducted. Results from both the
contemporaneous matrix and the impulse response functions are also evaluated for
consistency with economic theory. Lastly, the FEVD (forecast error variance
decomposition) method helps obtain a better understanding of the dynamic
relationships among variables, implying what sources contribute to more or less of a
given variables fluctuations.
The first version of the base model (model VN1) reflects the strongly
administered oil pricing regime while the second version (model VN2) reflects the
weakly administered oil pricing regime. With the two extensions, the current
research aims to analyse components of aggregate demand. In the first extension of
the base model, the study examine whether trade activities play any role in the
Vietnamese monetary transmission. This is done via extending the base model with
the inclusion of trade related variables (total exports and total imports) as in the
study of Cushman and Zha (1997). A similar approach is conducted in the recursive
SVAR model of Kubo (2008) when studying the international transmission of Thai
monetary policy. The results in the current study focus on the relationship between
the trade related variables and other variables in the model. Moreover, the second
extension helps to determine possible effects of a monetary contraction on
investment and consumption behavior. The expression in investment and
consumption equations is mainly based on the Keynesian assumption that aggregate
consumption is primarily a positive function of income and investment is an inverse
function of the interest rate.
This chapter is organised as follows. Section 6.2 determines the appropriate
lag length and checks the VAR stability. Section 6.3 estimates the contemporaneous
coefficients. Section 6.4 illustrates and discusses the results of the impulse
responses. Section 6.5 reveals the results of variance decomposition. Section 6.6 and
Section 6.7 present the analysis of aggregate demand, where Section 6.6 examines
the international transmission of monetary policy while Section 6.7 analyses the
consumption and investment behaviour. Section 6.8 applies some adjustments to
check the robustness of the study. Finally, Section 6.9 concludes the discussion of
this chapter.

114

6.2

LAG LENGTH AND VAR STABILITY CHECK


To determine the lag length of the reduced form (VAR), the study uses

different lag-length selection criteria, including LR, FPE, AIC, SC, and HQ. Table
6.1 shows lag-length results based on these criteria.
As shown in Table 6.1, while the LR, FPE, AIC and HQ criteria choose two
lags as optimal, SC suggests a shorter lag (one lag). The study chooses one lag (one
quarter). The chosen lag length (one quarter) is longer than the one-month lag length
used in the SVAR model of Tran (2009), but shorter the the 4-quarter lag length in
the VAR model of Le and Pfau (2009) study of Vietnams monetary policy. The 2lag result is not chosen as this lag length does not meet the requirement of the VAR
stability check, which is mentioned below.
Table 6.1: VAR Lag Order Selection Criteria
Endogenous variables: WOP WRP WGP YC FFR CPI Y R M CR VNI E
Exogenous variables: C DUM
Sample: 2000Q1 2011Q4

Included observations: 44

Lag

LogL

LR

FPE

AIC

SC

HQ

554.124

NA

5.57e-26

-24.096

-23.123

-23.736

1064.016

695.307

4.38e-33

-40.728

-33.916*

-38.202

1295.003

188.989*

5.30e-34*

-44.682*

-32.030

-39.991*

* indicates lag order selected by the criterion


Notes: WOP = World oil price, WRP = World rice price, WGP = World gold price, YC =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Interest rates, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real effective
exchange rate, C = constant term, DUM = the dummy variable.

Source: Authors calculation.

Before using the optimal lag length to estimate the parameters of the SVAR, it
is necessary to check the conditions of VAR stability using the AR roots and
autocorrelation LM tests. The results, presented in Table 6.2 and Table 6.3, show
that all the eigenvalues in the proposed model lie in the unit circle, so the
VAR/SVAR model satisfies the stability condition. Specifically, all root lie inside
the unit circle, and there is no serial correlation since the LM p-value = 5.35%, so
the null hypothesis of no serial correlation cannot be rejected. When applying the
same stability tests, the LM result for two lags do not obtain the required VAR
115

stability as p-value = 0%, so the null hypothesis is rejected (Appendix C, Table C1).
Therefore, the SC criterion (one lag) is selected because it is the most convenient lag
length for this study.
Table 6.2: VAR Residual Serial Correlation LM Tests
Null Hypothesis: no serial correlation at lag order h
Sample: 2000Q1 2011Q4

Included observations: 45

Lags

LM-Stat

Prob

172.3788

0.0535

Source: Authors calculation.

Table 6.3: Roots of Characteristic Polynomial


Endogenous variables: WOP WRP WGP YC FFR CPI Y R M CR VNI E
Exogenous variables: C DUM
Lag specification: 1 1
Root

Modulus

0.991

0.991

0.952

0.952

0.771

0.771

0.609 - 0.472i

0.771

0.609 + 0.472i

0.771

-0.026 - 0.648i

0.648

-0.026 + 0.648i

0.648

-0.227 - 0.348i

0.415

-0.227 + 0.348i

0.415

-0.289

0.289

0.104 - 0.255i

0.276

0.104483 + 0.255i

0.276

No root lies outside the unit circle.


Notes: WOP = World oil price, WRP = World rice price, WGP = World gold price, YC =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Interest rates, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real effective
exchange rate, C = constant term, DUM = the dummy variable.

Source: Authors calculation.


116

6.3

CONTEMPORANEOUS MATRIX
The estimated coefficients in the contemporaneous matrix are reported in

Table 6.4 with model VN1 for the SAOP (strongly administered oil-pricing) regime
and Table 6.5 with model VN2 for the WAOP (weakly administered oil-pricing)
regime. The difference between these two models is the appearance of world oil
price in the price level equation. The estimation method in the study is scoring with
analytic derivatives 20 . The highlighted values denote that this coefficient is
significant at 5 percent.
Table 6.4: Estimated Contemporaneous Coefficients of Model VN1
WP
WP

WRP

WGP

1.069

-1.842

-0.028

CR

VNI

1
-0.079

-0.049 0.136

2.047
8.090

-2.806 38.831 -130.111

0.175

-0.070 3.815

CR

VNI

FFR -0.553

CPI

WGP

CPI

FFR

WRP

FY

FY

-0.616 0.361 -0.556

1.204
1

0.264

-0.033

0.012 -1.977 -0.661

0.172

78.772 -24.832

54.449
-1.642

1
1

0.068 5.903 -5.276 -0.135 -4.557 3.376

-15952 -28271 10.647 940903 -2853 369622 -78526 -3566 -149663 247212 -7965
Note: WP = World oil price, WRP = World rice price, WGP = World gold price, FY =

Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Short run interest rate, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real
effective exchange rate.

Source: Authors calculation.

20

The method of scoring is the method where the gradient and expected information matrix are
evaluated analytically and it helps to maximize the log likelihood (Microsoft, 2009b). Another
standard-error option (Monte Carlo) is applied to check the robustness of the study in Section 6.8.
117

Estimation shows that the models in Tables 6.4 and 6.5 show similar results.
Most values are the same in sign (45 coefficients) although they differ somewhat in
levels because of the different number of restrictions in the two models.
Both models are over-identified but restrictions are valid under the null
hypothesis of the LR test for over-identification. Specifically, in model VN1, the
structural VAR is over-identified with the log likelihood of 964.292. The identifying
restrictions are not rejected at the 5 percent significance level (LR test chi-square
(20) = 31.107, p-value = 0.053). In model VN2, the null hypothesis that restrictions
are valid, is not rejected (log likelihood: 965.972, chi-square (19) = 27.747, p-value
= 0.088). Therefore, all restrictions in both models VN1 and VN2 are valid.
Table 6.5: Estimated Contemporaneous Coefficients of Model VN2
WP
WP

WRP

FY

FFR

0.841

-0.903

CR

VNI

1
-0.028

FFR -0.553

CPI -0.024 -0.073


-0.056 0.071

1.799

10.816

-3.544 50.840 -181.999

0.191

-0.074 4.071

CR
VNI -0.601 0.384
E

WGP

CPI

WRP

FY

WGP

-0.563

0.980
1

96.037 -31.986

0.110

-0.039

0.009 -1.952 -1.392

0.172

-5.619 -0.137 -4.608

3.321

0.071 5.561

64.434
-1.816

-15153 -27046 -454.558 953662 -2647 351186 -102826 -3794 -155048 246695 -8096
Note: WP = World oil price, WRP = World rice price, WGP = World gold price, FY =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Short run interest rate, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real
effective exchange rate.

Source: Authors calculation.

118

Estimation results of the contemporaneous matrix show the same number of


statistically significant coefficients (seven coefficients), including the instantaneous
impacts of WP (world oil price) on FY (foreign output) and VNI (stock price index),
WRP (world rice price) on CPI (consumer price index), Y (domestic output) on CPI
(consumer price index), and monetary variables (R interest rate, M money
supply, and CR bank credit) on VNI (stock price index). These coefficients are
statistically significant at the 5 percent significance level.
The estimation results show the negative effects of world oil price (WP) on
both foreign output (FY) and domestic output (Y), which is in line with most
empirical findings on the impacts of world oil price on output (Kim & Roubini,
2000). However, this relationship is only statistically significant for the case of
China (fourth row) and not for Vietnam (seventh row). This could be explained by
the oil price control in Vietnam, which lessens the negative effects of the world oil
price on Vietnamese economic growth. The next result is that the world oil price
elasticity of stock prices (eleventh row) is negative and significant. This finding is
consistent with theoretical expectations that an increase in WP leads the public to
have negative attitudes about higher production and living costs. The result becomes
more important when looking at the unexpected result of Narayan and Narayan
(2010) error-correction model finding of a positive and statistically significant link
between the world oil price and stock prices.
The results also show that the coefficients of world rice price and output are
significant in the price level equation (sixth row). The positive output elasticity of
the consumer price index is consistent with economic theory, implying that
economic growth exerts inflationary pressure on the economy. This is crucially
significant for a developing economy like Vietnam. As explained in Chapter 3,
growth rates and price level have been increasing, so policymakers should take into
account this statistically significant relationship between growth and inflation in
Vietnams development. Another significant link between the world rice price and
the domestic price level is negative, which is inconsistent with the theoretical
expectations. These results (-0.079 in model VN1 and -0.073 in model VN2)
indicate that the domestic price level has a slight and contemporaneous decrease
after an increase in the world rice price. A recent paper by the Oxfam organisation

119

offers a possible explanation, stating that changes in the domestic rice selling price
are not always in the same direction as changes in world rice price (Tran et al.,
2013). Specifically, while selling price falls along with a decrease in world rice
price, farm-gate selling prices only minimally increase after an increase in world
rice price. Moreover, to follow national food-security policy and to cope with
inflationary pressure, the Vietnamese government has been applying the rice-export
ban policies to stabilise the domestic price of food. Such policies have contributed to
limiting the increase in the consumer price index.
Except for the impact of world oil price as mentioned above, results in the
Vietnams stock price index equation reveal that only three domestic variables the
interest rate, monetary aggregate and credit, have statistically significant effects on
changes in this index (eleventh row). These findings give support to Narayan and
Narayans (2010) suggestion that domestic factors seem to have a more important
role than world oil price in affecting the stock market. While the impacts of R
(interest rate) and CR (credit) on VNI (stock price index) with negative and positive
signs respectively, are in line with the theory, the contemporaneous effect of M
(money supply) on VNI (with a negative sign) is inconsistent with expectations.
Specifically, an increase in the interest rate, implying a contractionary monetary
policy, results in a decrease in stock prices. Moreover, when bank loans increase, the
stock market booms. Obviously, trading on Vietnams equity market is greatly
influenced by monetary policy. However, the negative effect of monetary aggregates
on stock prices indicates that investors expect the real response to shocks in interest
rates (negative) and credit (positive) to decide their trading behaviour in the stock
market. Therefore, the growth of the Vietnamese stock market should be affected
more by the interest rate and credit, and less by the money supply.
Despite the similarity between the two models, there are some noticeable
differences. Specifically, the first is the elasticity of the world oil price with respect
to the price level in model VN2 (for a WAOP regime). The coefficient of world oil
price in the price level equation is -0.024 (p-value=0.07). The result shows that this
coefficient is statistically significant at the 10 percent significance level, implying
that in a weak administered oil price regime, an increase in the world oil price could
not cause any contemporaneous rise in the domestic price level, as expected.

120

Therefore, the current price controlling regime of the Vietnamese government seems
to be effective in ensuring price stability. Moreover, the elasticities of the domestic
interest rate and the monetary aggregate with respect to the stock prices in model
VN2 are 0.137 and 4.608, respectively. These are higher than those in model VN1
(0.135 and 4.557, respectively). Except for this, other elasticities (world rice price
and output with respect to price level, and world oil price and credit with respect to
stock prices) in model VN1 are higher than those in model VN2. This finding
implies that, when considering external shocks like oil price shocks in the price level
equation, most contemporaneous effects between the variables in the model are
weaker, so the shocks in model VN2 could become less sensitive than in model
VN1. However, this difference could be relatively small due to the values of
elasticities in both models.
6.4

IMPULSE RESPONSE FUNCTION


Estimation results of the contemporaneous coefficient matrix in the previous

section help to compute impulse response functions. Analytic (asymptotic) response


standard errors are used to display responses of domestic variables over 20 quarters.
This period is equivalent to the length of Vietnams five-year socio-economic plans,
which provide the fundamental direction for the development of its economy.
Moreover, to avoid the criticism of using differences to lose the long run
information, as mentioned at Section 5.3, the 20-quarter period is appropriate when
considering responses to shocks in the short and medium run, up to the 20 quarters
(five years) which becomes the long run. Specifically, for this study, four quarters
(one year) represent the short run and one year ahead to five years represents the
medium run.
Responses to structural shocks of one standard deviation are denoted in figures
6.1 to 6.9. In these figures, the responses are confirmed to be significant if the
confidence intervals (between two dashed lines) do not include zero. The results of
impulse responses from two versions of a base model (VN1 and VN2) have similar
patterns.

121

6.4.1 Responses of domestic variables to positive foreign interest rate


shocks
One of the most important foreign shocks are adjustments in the U.S. interest
rate because of the crucial global role of the U.S. economy. Figure 6.1 illustrates the
impulse responses of domestic variables to a positive Federal Fund rate shock.
After a positive U.S. monetary shock, domestic output tends to decrease while
price level increases over the 20-period horizon. The decrease in output is consistent
with the awareness that an increase in FFR induces the aggregate demand and output
to response negatively in a small, open economy (Kim & Roubini, 2000). However,
this decrease seems to be relatively small. This occurs because the U.S. is
considered one of Vietnams important trade partners, for exports, with the export
volume increasing from 5 percent of its total exports in 2000 to 20 percent in 201121;
at the same time, Vietnams imports from the U.S. are quite small, representing 2-4
percent of total imports during the period 2000-201122. Thus, an increase in the U.S.
interest rate benefits Vietnams export turnover, improving the trade balance
between Vietnam and the U.S.
After an increase in FFR, the results show a rise in the domestic interest rate
after the first three quarters. This increase is reasonable because a higher interest rate
in a big economy like the U.S. causes interest rates of other economies to increase,
as the domestic interest rate is increased to counter the possible domestic interest
rate devaluation. Although appearing after the three-quarter lag, this result is similar
to those found in other studies, such as Kim and Roubini (2000) and Afandi (2005).
However, this response is relatively short: it occurs from quarter 3 to quarter 6, and
after that there is a minor decrease between the quarters 6 and 8 before stability in
the remaining quarters. Due to the three-quarter lag, the increase in both the
monetary aggregate and credit are in line with the decrease in the domestic interest
rate. Due to the lag in domestic monetary policy, the exchange rate increases four
quarters after the shocks to FFR.

21
22

See http://www.gso.gov.vn/default.aspx?tabid=393&idmid=3&ItemID=13172
See http://www.gso.gov.vn/default.aspx?tabid=393&idmid=3&ItemID=13168
122

Figure 6.1: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in the Federal Fund Rate
Domestic variables

CPI

Model 1

Model 2

.12

.12

.08

.08

.04

.04

.00

.00

-.04

-.04

-.08

-.08
2

10

12

14

16

18

20

.08

.08

.04

.04

.00

.00

-.04

-.04

-.08
4

10

12

14

16

18

20

1.5

1.5

1.0

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

-1.0

-1.0

-1.5

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-1.5
2

-.08
2

10

12

14

16

18

20

.2

.2

.1

.1

.0

.0

-.1

-.1

-.2

-.2
2

10

12

14

16

18

20

.024
.020

.02

.016

CR

.012
.01
.008
.004
.00

.000
-.004

-.01

-.008
2

10

12

14

16

18

20

0.25

VNI

0.25

0.00

0.00

-0.25

-0.25

-0.50

-0.50

-0.75

-0.75

-1.00

-1.00
2

10

12

14

16

18

20

.02

.01

.01

.00

.00
-.01
-.01
-.02

-.02

-.03

-.03
2

10

12

14

Source: Authors calculation.

123

16

18

20

Over the whole period, the response of the stock prices to an increase in the
U.S. interest rate is negative. This is reasonable because the increases in money
supply and credit are insignificant, and only occur in the three quarters before an
increase in the domestic interest rate. Therefore, the stock market is negatively
affected by a contractionary monetary policy.
Overall, positive innovations to the FFR variable produce the important effects
on macroeconomic indicators for Vietnam. However, in any case, FFR shocks do
not generate statistically significant responses as zero is inside the confidence
intervals in Figure 6.1.
6.4.2 Responses of domestic variables to positive foreign output shocks
Considering the impulse responses to shocks of Chinese output (as a proxy for
foreign output) gives a clear understanding about the interaction between foreign
output and Vietnams economic variables (Figure 6.2).
As shown in Figure 6.2, the response of domestic output to foreign output
shocks is positive and gradually increasing for the whole period. This suggests that
the effect of Chinas output on Vietnams output is relatively significant because of
the role of the Chinese economy as one of Vietnams most important trade partners.
Such a response is consistent with a positively contemporaneous impact of the
coefficient of FY (foreign output) in the Y (domestic output) equation, as denoted in
Tables 6.4 and 6.5. Moreover, this response is statistically significant for the first 12
quarters, so in the short and medium term this indicator could have meaningful
effects on Vietnamese output. A similar response appears for the monetary
aggregate. Price level decreases until quarter 8, but they recover to be asymptotic to
the base line at the end of the period. However, except for the responses of domestic
output and monetary aggregate, other responses are not statistically significant.

124

Figure 6.2: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in Foreign Output
Domestic variables

CPI

Model 1

Model 2

.10

.10

.05

.05

.00

.00

-.05

-.05

-.10

-.10

-.15

-.15
2

10

12

14

16

18

20

.16

.16

.12

.12

.08

.08

.04

.04

.00

.00

-.04

-.04
2

10

12

14

16

18

20

-1

-1

-2
4

10

12

14

16

18

20

.4

.4

.3

.3

.2

.2

.1

.1

.0

.0

-.1

-.1
2

10

12

14

16

18

20

.01

CR

.00

-.01

-.01

-.02

-.02

-.03

-.03

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-.04
2

10

12

14

16

18

20

1.2

1.2

0.8

0.8

0.4

0.4

0.0

0.0

-0.4

-0.4
2

.01

.00

-.04

VNI

-2
2

10

12

14

16

18

20

.03

.03

.02

.02

.01

.01

.00

.00

-.01

-.01

-.02

-.02

-.03

-.03
2

10

12

14

16

Source: Authors calculation.

125

18

20

6.4.3 Responses of domestic variables to positive world price shocks


Figure 6.3 shows the responses of domestic variables to world oil price
shocks. First, there is a decrease in domestic output and an increase in the price
level, which is consistent with expectations. There are some reasons for this result.
First, although crude oil is one of the main exports of Vietnam, this economy has to
import oil for its growth. According to Vietnams General Statistics Office, from
2005 to 2011, the export volume was about 93 million tons of oil, but in the same
time, 82.3 million tons of oil was imported23. Therefore, the benefits of increasing
oil prices to Vietnams economy are limited, because of the difference between oil
imports and exports. Second, the average annual oil consumption growth of Vietnam
from 2000 to 2010 was 6-8 percent (Hoa Binh Securities, 2011). Positive shocks in
oil prices increase production costs, which results in decreasing output. However,
due to the Vietnamese governments control over oil prices, this response occurs
gradually and the effects seem to be negligible. This indicates that the world oil
price could play an important role in production in Vietnam, but price management
for this product under the strict control of Ministry of Vietnam seems to be effective
in lessening the effects of oil price shocks. Third, the impulse response functions
recorded the positive responses of stock prices to the world oil price shocks after a
statistically negative contemporaneous effect of world oil prices on stock prices, as
mentioned in Section 6.3. This result is similar to the findings of Narayan and
Narayan (2010), especially in the period of four to five quarters (short run) after the
shock. For the whole period (middle run), the response stays at the same level, and it
is not statistically significant. Moreover, shocks to oil prices create negative
responses in the monetary aggregate, in credit with some lag, and in the positive
responses of interest and exchange rates, but none of these responses is statistically
significant. Overall, positive shocks in oil prices do not produce great effects on
domestic variables.

23

See http://www.gso.gov.vn/default.aspx?tabid=393&idmid=3&ItemID=13167
126

Figure 6.3: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in World Oil Prices

Domestic variables

CPI

Model 1

Model 2

.16

.20

.12

.15

.08

.10

.04

.05

.00

.00

- .04

- .05
2

10

12

14

16

18

20

.08
.04

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

.05

.00
- .04
- .05

- .12

- .10
2

10

12

14

16

18

20

-1

-1
2

.00

- .08

.10

10

12

14

16

18

20

.2

.3

.1

.2

.0

.1

- .1

.0

- .2

- .1

- .3

- .2
2

10

12

14

16

18

20

.01

.02
.01

.00

.00
- .01

CR

- .01
- .02
- .02
- .03

- .03

- .04

- .04
2

VNI

10

12

14

16

18

20

1.2

1.6

0.8

1.2

0.4

0.8

0.0

0.4

- 0.4

0.0

- 0.8

- 0.4
2

10

12

14

16

18

20

.04

.04

.03

.03

.02

.02

.01

.01

.00

.00

- .01

- .01

- .02

- .02
2

10

12

14

Source: Authors calculation.

127

16

18

20

Figure 6.4: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in World Rice Price
Domestic variables

CPI

Model 1
.16

.16

.12

.12

.08

.08

.04

.04

.00

.00

-.04

-.04
2

Model 2

10

12

14

16

18

20

.100

.100

.075

.075

.050

.050

.025

.025

.000

.000

-.025

-.025

-.050
4

10

12

14

16

18

20

2.0

2.0

1.5

1.5

1.0

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

-1.0

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-1.0
2

10

12

14

16

18

20

.3

.2

.2

-.050
2

.1

.1
.0
.0
-.1

-.1

-.2

-.2
2

CR

10

12

14

16

18

20

.01

.01

.00

.00

-.01

-.01

-.02

-.02

-.03

-.03
2

10

12

14

16

18

20

0.25

VNI

0.25

0.00

0.00

-0.25

-0.25

-0.50

-0.50

-0.75

-0.75

-1.00

-1.00
2

10

12

14

16

18

20

.04

.03

.03

.02

.02
.01
.01
.00
.00
-.01

-.01
-.02

-.02
2

10

12

14

Source: Authors calculation.


128

16

18

20

Figure 6.4 shows the responses of domestic variables to world rice-price


shocks. Positive rice price shocks generate a positive impact on most domestic
variables, including output, price level, interest rate and exchange rate over 20
quarters. Specifically, output and price increase consistently after an increase in
world rice price. This is because of the important role of rice in the development of
the Vietnamese economy. As mentioned in earlier chapters, rice is one of Vietnams
main exports, with the export volume at 57.5 million tons of rice from 2000 to
201124, higher than that of other products. Vietnam is defined as one of the worlds
top three rice exporters 25 . Along with the increase in interest rates, there are
decreases in monetary aggregate, credit and stock prices, until the interest rate
begins to decrease from quarter 3. The responses of monetary variables, including
the interest rate (positive) and the monetary aggregate, and credit (negative), denote
that a contractionary monetary policy exists after the shocks in rice price. As rice is
a strategic product in Vietnams food-security policy, the government take many
measures to limit inflationary pressure when the world rice price increases, as
mentioned in Section 6.3. Therefore, monetary policy responses are consistent with
this direction. However, such policies could be defined as an example of the tradeoff between fighting inflation and ensuring economic growth. Specifically, the
impulse responses along with the results of contemporaneous effects show that the
Vietnamese economy do not receive greater benefits from an increase in rice price,
although it is one of the worlds top rice exporters. From the social viewpoint, this
also implies that the farmers income is difficult to improve even with increases in
the world rice price; this, in turn affects the success of poverty-alleviation measures.

24
25

See http://www.gso.gov.vn/default.aspx?tabid=393&idmid=3&ItemID=13171
See http://www.foodsecurityportal.org/vietnam?print
129

Figure 6.5: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in World Gold Price
Domestic variables

CPI

Model 1

Model 2

.050

.050

.025

.025

.000

.000

-.025

-.025

-.050

-.050

-.075

-.075

-.100

-.100
2

10

12

14

16

18

20

.04

.04

.02

.02

.00

.00

-.02

-.02

-.04

-.04

-.06

10

12

14

16

18

20

1.5

1.5

1.0

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

-1.0
4

10

12

14

16

18

20

.05

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

.05

.00

.00

-.05

-.05

-.10

-.10

-.15

-.15

-.20

-.20

-.25

-.25
2

10

12

14

16

18

20

.01

.01

.00

.00

-.01

-.01

-.02

-.02

VNI

-1.0
2

CR

-.08
2

-.06

-.08

10

12

14

16

18

20

.8

.8

.6

.6

.4

.4

.2

.2

.0

.0

-.2

-.2

-.4

-.4
2

10

12

14

16

18

20

.02

.02

.01

.01

.00

.00

-.01

-.01

E
-.02

-.02
2

10

12

14

Source: Authors calculation.

130

16

18

20

Figure 6.5 shows the estimated results of the effect of gold-price shocks on the
Vietnamese economy. A positive shock results in a decreasing negative response of
output but it occurs with a two-quarter lag. Similarly, two quarters after the shock,
the price level increases in quarters 3 and 4 before their decrease beginning in
quarter 5. This could be explained by Vietnams policy of controlling the importing
gold. Positive shocks in gold price also lead to responses of monetary and financial
indicators with the same lag. Specifically, the interest rate increases two quarters
after the shock, so the monetary aggregate and credit decrease in response to goldprice shocks. However, the increase in the domestic interest rate only occurs in two
quarters until quarter 4, before falling for the remainder of the period. Notably, the
influence of gold price (WGP) shocks on the monetary aggregate is significant in the
short run. The changes in stock prices (VNI) and the real effective exchange rate (E)
are consistent with the fluctuation of the domestic interest rate. Specifically, after the
first period, stock prices increase slightly when the interest rate decreases, and credit
increases again. In the first two quarters and the period from quarter 4, the E
decreases, implying a depreciation in the real value of the VND, while the increase
in the E from quarter 2 to quarter 4 represents the appreciation of the VND. This
could reflect insufficient responses of Vietnams monetary authority to positive
shocks in gold price. To narrow the gap between the world and domestic gold price,
the State Bank of Vietnam needs to increase the gold supply or depreciate the VND
(Tran, 2009a). The depreciation of the VND in the first two quarters helps the State
Bank of Vietnam follow its policy in narrowing the gold price gap; however, the
appreciation of the VND in the next two quarters implies that the State Bank of
Vietnam does not take full advantage of the increase in the world gold price to
narrow the gap between the domestic and world prices. Moreover, the decrease of E
from quarter 4 does not change the fact that the response of E to the gold price
shocks is positive until the end of the period. Overall, it appears the State Bank of
Vietnam does not operate monetary policy tools effectively to obtain the planned
policy targets in narrowing the gold price gap.
6.4.4 Responses of domestic variables to monetary policy shocks
Figure 6.6 shows the responses of domestic variables. An increase in the
domestic interest rate implies that the State Bank of Vietnam implements a
contractionary monetary policy. The interest rate shocks result in decreases in
131

output, which is consistent with theory, as positive R (interest rate) shocks induce Y
(output) to decrease. Thus, this result illustrates no evidence of an output puzzle26.
However, the price level increases after the R shocks, implying a price puzzle. This
puzzle only occurs in the first six quarters; after that time, the interest rate has a
tendency to decrease. In previous studies, the inclusion of the world oil price in the
system helps to surmount the price puzzle (Kim & Roubini, 2000; Vinayagathasan,
2013), but the price puzzle still appears in model VN2 which includes the world rice
price in the price level equation. Following the R shocks, variables M (money
supply), CR (credit) and VNI (stock price index) decrease, while E increases, as
expected; however, these changes denote the policy direction of the State Bank of
Vietnam and the lag of financial variables. Specifically, it takes two quarters for the
interest rate to increase. This seems to reflect the monetary authoritys caution when
applying contractionary monetary policy to the economy. With this shock, the
monetary aggregate and stock prices start to decrease from quarter 2, while the
credit only begins to fall in quarter 3. The change in monetary aggregate confirms
that there is no evidence of a liquidity puzzle one quarter after the shock. The
appreciation of domestic currency appearing with the lag (from quarter 3)27 is in line
with the existing theory about the response of home currency after an increase in the
domestic interest rate. Thus, a liquidity puzzle and an exchange rate appear in the
short period after the interest rate shock (one quarter and two quarters, respectively).
Figure 6.7 represents the estimated impulse responses of domestic variables to
positive money shocks. When M (money supply) increases and R (interest rate)
decreases, the real value of the VND depreciates. Unexpectedly, output (Y) declines
after a positive money shock, rather than increases, as predicted by monetary theory.
However, this decrease is not far from the base line. This is because the money
shock is relatively small, and only occurs in the first three quarters before decreasing
over the period. Due to such changes, other variables, including price level (CPI),
interest rate (R), credit (CR), stock price (VNI), and E, change in a manner consistent
with expectations. The response of the domestic interest rate is negative in the first
quarter, but begins to increase slightly after that time. Next, the decrease in M
Kim and Roubinis (2000) summary of puzzles is mentioned in Chapter 2.
With measuring the REER in Section 5.2, its increase reflects the national currency
appreciates in real terms relative to the currencies of the countrys main trade partners.
26
27

132

induces CR and VNI to decrease over the period. Because there is a positive
response in price level after the positive money shock, the result implies the
possibility of money shocks increasing inflation in Vietnam; however this response
is relatively small and is statistically insignificant. The decrease in M after its initial
increase cause the output (Y) to decrease.
Figure 6.6: Impulse Responses of Domestic Variables to Structural Shocks
of One Standard Deviation in Domestic Interest Rates
Domestic variables

CPI

Model 1

Model 2

.25

.25

.20

.20

.15

.15

.10

.10

.05

.05

.00

.00

-.05

-.05
2

10

12

14

16

18

20

.08

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

.04

.04
.00
.00

-.04

-.04

-.08

-.08
-.12

-.12

-.16
2

10

12

14

16

18

20

-1

-1

-2

-2
2

10

12

14

16

18

20

.2

.2

.1

.1

.0

.0

-.1

-.1

-.2

-.2

-.3

-.3
2

CR

10

12

14

16

18

20

.03

.03

.02

.02

.01

.01

.00

.00

-.01

-.01

-.02

-.02

-.03

-.03
2

10

12

14

16

18

20

0.8

0.5

0.4

VNI

0.0

0.0
-0.5
-0.4
-1.0

-0.8

-1.2

-1.5
2

10

12

14

16

18

20

.06

.06

.04

.04

.02

.02

.00

.00

-.02

-.02

-.04

-.04
2

10

12

14

Source: Authors calculation.

133

16

18

20

Figure 6.7: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in Money Aggregate
Domestic variables

Model 1

Model 2

.3

.2

.2

.1

.1

CPI

.0
.0
-.1

-.1

-.2

-.2
2

10

12

14

16

18

20

.08

.10

.04

.05

.00

.00

-.04

-.05

-.08

-.10

-.12

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-.15
2

10

12

14

16

18

20

1
0
0
-1

-1

-2

-2
-3

-3
2

10

12

14

16

18

20

.2

.2
.1

.1

.0

.0
-.1
-.1

-.2

-.2

-.3
2

CR

10

12

14

16

18

20

.02

.02

.01

.01

.00

.00

-.01

-.01

-.02

-.02
2

VNI

10

12

14

16

18

20

0.8

0.8

0.4

0.4

0.0

0.0

-0.4

-0.4

-0.8

-0.8

-1.2

-1.2
2

10

12

14

16

18

20

.04

.04

.02

.02

.00

.00

-.02

-.02

-.04

-.04

-.06

-.06
2

10

12

14

Source: Authors calculation.


134

16

18

20

Figure 6.8: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in Credit
Domestic variables

CPI

Model 1

Model 2

.12

.12

.08

.08

.04

.04

.00

.00

-.04

-.04

-.08

-.08
2

10

12

14

16

18

20

.04

.04

.00

.00

-.04

-.04

-.08

-.08

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-.12

-.12
2

10

12

14

16

18

20

3.2
2.8

2.4
2.0

1.6
1.2

0.8
0.4

0.0
2

10

12

14

16

18

20

.1

.1

.0

.0

-.1

-.1

-.2

-.2

-.3

-.3
2

CR

10

12

14

16

18

20

.01

.01

.00

.00

-.01

-.01

-.02

-.02

-.03

-.03
2

10

12

14

16

18

20

.4
0.25
.2

VNI

0.00

.0

-0.25

-.2
-.4

-0.50

-.6

-0.75

-.8

-1.00
2

10

12

14

16

18

20

.03

.03
.02

.02

.01
.01
.00
.00
-.01
-.01

-.02

-.02

-.03
2

10

12

14

Source: Authors calculation.

135

16

18

20

Figure 6.9: Impulse Responses of Domestic Variables to Structural Shocks


of One Standard Deviation in Real Effective Exchange Rate
Domestic variables

Model 1

Model 2

.08

.08
.04

.04

.00

CPI

.00
-.04
-.04

-.08

-.08

-.12
2

10

12

14

16

18

20

.00

.00

-.02

-.02

-.04

-.04

-.06

-.06

-.08

-.08

-.10

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

10

12

14

16

18

20

-.10

-.12

-.12
2

10

12

14

16

18

20

1.5

1.0

1.0

0.5

0.5

0.0
0.0
-0.5
-0.5
-1.0

-1.0
-1.5

-1.5
2

10

12

14

16

18

20

.0

.0

-.1

-.1

-.2

-.2

-.3

-.3
2

CR

10

12

14

16

18

20

.020

.020

.015

.015

.010

.010

.005

.005

.000

.000

-.005

-.005

-.010

-.010
2

10

12

14

16

18

20

.2

0.2
0.0

.0

-0.2

VNI

-.2
-0.4
-.4
-0.6
-.6

-0.8

-.8

-1.0
2

10

12

14

16

18

20

.02

.02
.01

.01

.00
.00
-.01
-.01

-.02

-.02

-.03
2

10

12

14

Source: Authors calculation.

136

16

18

20

Figure 6.8 shows the interaction between credit shocks and other variables in
the economy. As with to money shocks, credit shocks do not produce the expected
response in output, as it decreases rather than increases. This may be explained by
several reasons. First, bank loans may be used in ways, preventing them from
creating positive effects in ensuring economic growth. This causes risks for the
banking system (for example, bad debts). According to the State Bank of Vietnam,
the bad-debt ratio accounted for 3.34% of the credit total at the beginning of 2012,
and it reached a peak of 4.93% in September 2012 28 . This domination of stateowned banks (or state-originated banks via equitisation) in the Vietnamese banking
system, along with the large credit allocation for state-owned enterprises, worsens
the bad-debt problem in Vietnam. Moreover, the increase in price level could be
from the inappropriate allocation of credit in the economy, and this response is
significant in the short run. Second, positive credit only occurs in the first two
quarters; after that, the impacts of its decrease on other monetary variables seem to
outweigh the effects of the increase in credit. As a result, the interest rate increases,
the monetary aggregate decreases, stock prices fall, and the real value of the VND
appreciates. The response of the monetary aggregate is significant in the short run.
Along with considering the responses of money shocks above, this implies that the
impacts of monetary policy could disappear due to short run policy adjustments and
the lag in transmission between the monetary instruments. However, only the
response of interest rates after the decrease in credit is statistically significant in both
the short and the long runs.
Figure 6.9 shows the responses of key macroeconomic indicators to shocks in
the real effective exchange rate (E). Positive E shocks, representing the appreciation
of the domestic currency, generate some statistically significant results. Specifically,
when the VND appreciates, it stimulates imports and reduces the competitiveness of
Vietnamese goods, negatively affecting Vietnams export turnover, so the trade
deficit increases. This induces the output to decrease significantly over the period.
Such a result could reflect the high openness of the Vietnamese economy. The
appreciation in the VND value is also supported by the increase in the domestic
interest rate, leading to a decrease in monetary aggregate, output and stock prices
28

See http://www.sbv.gov.vn/portal/faces/vi/vim/vipages_trangchu/tkttnh/hdhttctd/tlnx?_adf.ctrlstate=16m038634v_129&_afrLoop=3851357282138600
137

over the period. In particular, the response of the monetary aggregate is statistically
significant in both the short and long run terms. Moreover, the response of stock
prices is also significant in the first six quarters after the shock. A decrease in R
(interest rate) from quarter 7 is consistent with the depreciation of the VND at the
same time, but these do not affect M (money supply), VNI (stock price) or Y
(output); this implies that the influence of the previous appreciation of the VND on
many macroeconomic variables is relatively strong. Notably, the appreciation of the
domestic currency causes imports to increase, resulting in an increasing tendency in
price level because import goods account for a large share in calculating CPI;
however, this increase disappears when the VND depreciates.
6.5

VARIANCE DECOMPOSITION
Table C2 (Appendix C) reports the forecast error variance decomposition

(FEVD) of the domestic variables in four forecast horizons: one, four, 10 and 20
quarters. To facilitate comparison, the variance decompositions for real output, Y
are shown in Figure 6.10. The shocks are to world oil price (Shock 1), world rice
price (Shock 2), world gold price (Shock 3), foreign output (Shock 4), foreign
interest rate (Shock 5), domestic price level (Shock 6), domestic output (Shock 7),
interest rate (Shock 8), monetary aggregate (Shock 9), credit (Shock 10), stock price
(Shock 11) and real effective exchange rate (Shock 12).
Figure 6.10: Variance Decomposition of Y (output)

Source: Authors calculation (Appendix C)


138

FEVD is a useful analysis to understand the sources of fluctuations of a


variable, implying that how many of its fluctuations are explained by its own shocks
versus the innovations of other variables in the model (Enders, 2004). Estimation
results for variance decomposition in both models gave similar patterns.
First, as shown in Figure 6.10, the fluctuations of Y (output) are mainly
explained by its own shocks (Shock 7) and the price level (CPI) shocks (Shock 6)
over the short and medium run. The role of price shocks to output is consistent with
the Lucas-Phelps imperfect information models findings about the interaction of
price changes to output fluctuations (Romer, 2012). In terms of transmission
channels, exchange rate shocks (Shock 12) explain more output fluctuations than
interest rate shocks (Shock 8) and credit shocks (Shock 10). The impacts of three
channels (credit, exchange rate and asset price) are bigger in the medium run than in
the short run, while the influences of the interest rate channel decrease. As the
horizon lengthens, especially in the medium run, the contribution of foreign output
shocks in explaining domestic output fluctuations is higher, implying that the
Chinese economy has a significant effect on the Vietnamese economy. Less of the
output variation comes from other shocks. The contribution of world oil prices
(Shock 1) and foreign interest rates (Shock 5) almost disappear in the medium run.
The impact of Shock 1 reflects the Vietnamese governments price-control policy
limiting the negative effects of world oil price shocks to the domestic economy.
Obviously, in both the short and medium run, internal factors play a crucial role in
the fluctuations of domestic output.
Figure 6.11 shows that the explanation of the CPI fluctuations varies in the
short and medium run. In short run, shocks to the world rice price (Shock 2) explain
more variations in CPI than other shocks; the next factors are the interest rate
(Shock 8) and credit (Shock 10). The impact of world rice price shocks seems to be
consistent with the fact that rice is one of the main exports of Vietnam, but it is also
the daily food of the Vietnamese people, so it is very important for the countrys
food security target. The impact of the world oil price on price level (Shock 1)
increases at longer horizons in the short run and remains a significant contribution in
the medium run, implying that the world oil price affects domestic price level with
some lags due to Vietnams administered oil price regime. Because the world oil
price shocks are more sensitive to domestic price level (from the publics behaviour
139

or the flexible adjustments of the authorities), the influence of world oil price in
explaining price level fluctuations in model VN2 is greater than those in model VN1
(see Appendix C). In the later period (medium run), the influence of the rice price is
less, while the interest rate has the greatest contribution in explaining the CPI
variation. The explanation for CPI fluctuations from output shocks and its own
shocks becomes more important as the horizons expand to the medium run. An
examination of the transmission channels shows that the credit channel is more
effective than the other channels in explaining price fluctuations. Compared to the
short run, the impacts of the interest rate, credit, and asset price channels are greater
in the medium run, while the influences of the exchange rate channel are smaller.
Similar to the variance decomposition of output, a substantial proportion of the
consumer price variation is mainly explained by domestic factors, rather than
foreign, especially in the medium run.
Figure 6.11: Variance Decomposition of CPI

Source: Authors calculation.

Domestic shocks also play a bigger role than the foreign shocks in explaining
the fluctuations of monetary policy in Vietnam (Figure 6.12). Much of the interest
140

rate fluctuations are due to credit shocks at all horizons, especially at the first quarter
when the credit (CR) shocks explain about 90 percent of the variations in the interest
rate (R). The next contributions come from world rice price shocks and the interest
rates own shocks. The impacts of output and gold price become higher from quarter
2 and remain stable in the medium run (more than 8 and 10 percent, respectively). A
similar trend occurs to the world oil price shocks. The monetary aggregate explains
less of the interest rate fluctuations. Shocks to credit, interest rate, price level, output
and the world rice price significantly contribute to the money variations in both the
short run and medium run. The contribution of E shocks has been increasing in the
medium run terms reaching about 10 percent at the peak, while the stock price
shocks explain less than 1 percent of the M (money supply) fluctuations. The results
show that the impact of world gold price (WGP) on M is more significant from
quarter 2. The influence of world rice price is significant in the short run (more than
30 percent in quarter 2) but it has been decreasing in the medium run. In the
fluctuations of both interest rate and monetary aggregate, the influence of world oil
price is too small.
Figure 6.12: Variance Decomposition of R (interest rate) and M (money
supply)
Variance Decomposition of R

Source: Authors calculation.


141

- Variance Decomposition of M

Source: Authors calculation.

Figure 6.13: Variance Decomposition of CR (credit), E (real effective


exchange rate) and VNI (stock price index)
- Variance Decomposition of CR

Source: Authors calculation.


142

- Variance Decomposition of E

Source: Authors calculation.

- Variance Decomposition of VNI

Source: Authors calculation.

143

In the financial market (represented by CR the credit, E the real effective


exchange rate and VNI the stock price), the effects of the money market
(represented by R the interest rate and M the money supply) on the E variable is
bigger than its effects on the CR and VNI variables (Figure 6.13). Specifically, much
of the E fluctuation is explained by the R and M innovations (more than 50 percent
in the short run and about 50 percent in the medium run). If the CR shocks are
included, the impacts of the R, M and CR shocks on the E variations are about 80
percent in the short run and about 70 percent in the medium run. Shocks to the world
rice price and world gold price explain more changes in real effective exchange rate
(about 20 percent from quarter 3). Only around 10 percent of the fluctuations in CR
are explained by other financial shocks (R, M, E and VNI). Much of the CR variation
is explained by the Y (domestic output), FY (foreign output) and WRP (world rice
price) shocks with about 40, 20 and 10 percent respectively. Stock prices own
shocks contribute the most (about 30 percent). Other financial variables, including
R, M, CR and E, explain nearly 20 percent in the medium run, but the impact of M is
too small. Notably, the contributions of monetary policy shocks (the R shocks) are
higher than those of output and price level shocks.
6.6

THE SVAR ANALYSIS FOR INTERNATIONAL TRANSMISSION


6.6.1

Contemporaneous matrix

This section presents the first extension of the basic SVAR model to analyse
two aggregate demand components: exports and imports under the effects of the
monetary policy. Also, it examines the international transmission of monetary policy
by analysing the relationship between the trade related variables total exports (VE)
and total imports (VI) - and other variables in the model. Based on the model of
Cushman and Zha (1997), the study constructs a new model, which is one the
extensions to the base model.
Unlike the study of Cushman and Zha (1997), this study restricts the
contemporaneous effect of only the output and price level on two trade variables.
Apart from these restrictions, Cushman and Zha (1997) included the instantaneous
impact of total exports on total imports in order to obtain the triangularised order of
variables in the production segment (Y, CPI, VE and VI). This could be inappropriate
144

in the SVAR model, as the opposite relationship where total imports affect total
exports is not considered. Thus, to avoid this confusion, this study only accepts
Cushman

and

Zhas

(1997)

assumption

that

exports

and

imports

are

contemporaneously affected by the output and price level. Other variables do not
instantaneously affect these two trade variables due to prior setup practices relating
to trade contracts as well as production plans (Sims & Zha, 1998).
Table 6.6: The Models for International Transmission
- Model VN1 (revised)
WP WRP WGP FY FFR CPI
WP

VE

CR

1
a4,1

FFR a5,1
CPI

1
1

a6,7

a7,6

VI

a8,6

a8,7

VE

a9,6

a9,7

a6,2
a7,1

a7,2

a7,4

a10,3

a10,5 a10,6 a10,7

a11,3

a9,5

1
1
1

a10,11 a10,12

a11,10

CR

a12,5 a12,6 a12,7

a12,10

VNI a13,1 a13,2 a13,3

a13,5 a13,6 a13,7 a13,8 a13,9 a13,10 a13,11 a13,12

a14,1 a14,2 a14,3

a9,6

a7,12

a9,7

VNI

WGP

VI

WRP

FY

a10,14
a11,14

1
1
1

a14,4 a14,5 a14,6 a14,7 a14,8 a14,9 a14,10 a14,11 a14,12 a14,13

145

- Model VN2 (revised)


WP WRP WGP FY
WP

VI

VE

CR

WGP

1
a4,1

FFR a5,1

CPI

a6,1

a6,2

a6,7

a7,1

a7,2

a7,6

VI

a8,6

a8,7

VE

a9,6

a9,7

a7,4

a10,3

a10,5 a10,6 a10,7

a11,3

a9,5

a9,6

a7,12
1
1
1

a10,11 a10,12

a9,7

a11,10

CR

a12,5 a12,6 a12,7

a12,10

VNI a13,1 a13,2 a13,3

a13,5 a13,6 a13,7 a13,8 a13,9 a13,10 a13,11 a13,12

VNI

WRP

FY

FFR CPI

a10,14
a11,14

1
1
1

a14,1 a14,2 a14,3 a14,4 a14,5 a14,6 a14,7 a14,8 a14,9 a14,10 a14,11 a14,12 a14,13

Note: WP = World oil price, WRP = World rice price, WGP = World gold price, FY =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, VI =
Imports, VE = Exports, R = Short run interest rate, M = Money supply, CR = Bank credit, VNI =
Stock price index, E = Real effective exchange rate.

Results in Table C3 (Appendix C) shows that all the eigenvalues of the


model lie inside the unit circle confirming the SVAR satisfies stability conditions.
Based on the above restrictions, this study shows the contemporaneous matrix, in
Table 6.7.
146

Table 6.7: Results of the Contemporaneous Coefficients for the


International Channel
- Model VN1
WP
WP

WRP WGP

FY

0.915

-0.896

VI

VE

CR

FY -0.025

FFR -0.289

1
-0.094

-0.021 0.048

1.035

VI

-1.682 -2.562

VE

-1.913 -1.414

0.593
1
1

-55

5.056 -159.73 68.155

-284

-4487

154.74 -66747 -62583

1289

CR

-0.026 -0.448 -0.587

0.094

VNI -0.260 0.218 0.808

0.040 6.620 -2.629 -0.184 -0.936 -0.136 -2.17 2.814

VNI

WGP

CPI

WRP

CPI

FFR

125

37.95
62266

1
1

-4521 -10375 5094 281012 -1846 157140 -38347 -13864 12992 -1590 -31700 78441 -3445
Source: Authors calculation.

147

- Model VN2
- WP
WP

WRP

FY

FFR

CPI

0.836

-0.708

VI

VE

VNI

1
1

-0.025

FFR -0.289
CPI

-0.013 -0.088

-0.023 0.034

1.020

VI

-1.703 -1.361

VE

-1.896 -2.659

0.563
1
1

-53.626

4.895 -153.65 72.193

-4210

145.74 -62952 -59266

1213

-0.026 -0.482 -0.745

0.095

CR

CR

WGP

VNI

WRP

FY

WGP

-0.257 0.227 0.809


-4159 -9542

-276 122.43

36.885

58651
1

0.040 6.542 -2.762 -0.190 -0.933 -0.137 -2.180 2.806

4785 263319 -1728 145574 -38545 -13121 12224 -1502 -29787 73283 -3246

Notes:
* WP = World oil price, WRP = World rice price, WGP = World gold price, FY = Foreign output,
FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, VI = Imports, VE =
Exports, R = Short run interest rate, M = Money supply, CR = Bank credit, VNI = Stock price index,
E = Real effective exchange rate.
* The highlighted cells express the same significant contemporaneous effects with the base model.
The underlined cells express the additional significant results.
Source: Authors calculation.

As Table 6.7 shows, there are eight statistically significant coefficients, with
five results similar to the results in Section 6.3 (the contemporaneous effects of WP
148

on FY, WRP on CPI, Y on CPI, and R and CR on VNI in the highlighted cells). These
results strongly confirm the analysis of the Vietnamese economy given above, in
several ways. First, during the period of increasing world rice prices, the Vietnamese
government applied policies to control inflation and thus ensure the food-security
policy. Second, the economic growth of Vietnam puts presssure on inflation. Third,
monetary policy implementation plays an important role in the boom of the stock
market.
The new model also shows three more significant results related to Vietnams
trade activities (the underlined cells). First, the negative and significant coefficient
of CPI in the VE equation reflects negative effects of an increase in price level on
Vietnams exports. Production costs increase, so the competitiveness of Vietnamese
exports decreases. As a result, exports are negatively affected. Second, the negative
and significant coefficient of Y in the VI equation is consistent with economic
theory. This implies that the increase in domestic output causes imports to decrease,
and thus helps to limit Vietnams trade surplus. Third, the negative coefficient of VE
in the VNI equation illustrates that although an increase in exports contributes to an
improved trade surplus, a slight decrease in stock prices reflects investors negative
reaction to this news. Economic theory would classify this as an inappropriate result,
but it seems to result from investors worry about the increase in exports. This is
explained by the limitation of the export growth of Vietnam. According to
Vietnams Ministry of Industry and Trade, exports mainly come from exploiting
natural resources and cheap labor rental costs. Until 2009, high-technology
exporting goods only accounted for 9 percent of export turnovers. Therefore, the
expansion of exports seems to increase the risk of natural resources exhaustion and
environmental pollution29. The finding implies that in the case of Vietnam, it is not
very likely that the growth in exports is a positive signal for the public and the
investor community, due to the nature of this growth. The relationship between Y,
CPI, VE and VI gives a clear picture about the development cycle of Vietnam;
specifically, higher output causes higher price level to rise, which in turn leads to a
decrease in exports and a negative effect on gross domestic product. Economic
growth helps to limit imports when domestic production is improved to satisfy

29

See http://www.moit.gov.vn/vn/Pages/Trangchu.aspx.
149

domestic demand. The stability of prices contributes to enhancing the


competitiveness of exports, leading to stable economic growth, rather than a bubble
growth and its associated increasing prices and decreasing exports.
6.6.2

Impulse response functions of exports and imports to a

contractionary monetary policy


This study considers the results of the impulse response functions of trade
variables to a contractionary monetary policy shock, as shown in Figure 6.14. This
shock is a positive innovation to the interest rate. The results illustrate that exports
decrease from quarter 1 and reach the lowest level in quarter 6 while imports obtain
the positive response at quarter 1 after the positive interest rate shock. This is
consistent with the monetary theory that an increase in interest rate implies the
appreciation of the home currency, which attracts imports. However, from quarter 2,
imports decrease for two quarters before increasing again in quarter 4 after the
shock. These changes imply that in the short run (one year), a contractionary
monetary policy has negative effects on import demand, which helps to improve the
trade balance; however, in the medium run, the increase in imports could cause the
trade deficit to be more serious, especially in the period from the quarter 4 to quarter
6, when exports continue to decrease and imports increase again.
Figure 6.14: Impulse Responses of Exports (VE) and Imports (VI) to
Structural Shocks of One Standard Deviation in the domestic Interest Rate (R).

Accumulated Response to Structural One S.D. Innovations 2 S.E.


- Response of VE:

Accumulated Response of LNVE_SA to Shock10


.2

.1

.0

-.1

-.2
2

10

12

14

16

18

20

Accumulated Response of LNVI_SA to Shock10


.2

.1

150

.0

-.1

-.2
2

10

12

14

16

18

20

- Response of VI:
Accumulated Response of LNVI_SA to Shock10
.2

.1

.0

-.1

-.2
2

10

12

14

16

18

20

Source: Authors calculation.

6.6.3

Variance decomposition

This section presents the results of the variance decomposition of the two trade
related variables, exports and imports, to understand the contribution of these
variables to the fluctuations in the trade balance. The specific results are shown in
Table C4, Appendix C and are graphed here to compare the sources of trade related
fluctuations. The shocks are to world oil price (Shock 1), world rice price (Shock 2),
world gold price (Shock 3), foreign output (Shock 4), foreign interest rate (Shock 5),
domestic price level (Shock 6), domestic output (Shock 7), imports (Shock 8),
exports (Shock 9), interest rate (Shock 10), monetary aggregate (Shock 11), credit
(Shock 12), stock price (Shock 13) and real effective exchange rate (Shock 14).
Figure 6.15: Variance Decomposition of Exports

Source: Authors calculation.

151

Figure 6.16: Variance Decomposition of Imports

Source: Authors calculation.

Figure 6.15 shows that the fluctuations in exports are mainly explained by
their own shocks (Shock 8) in both the short and the medium run; however, this
contribution decreases when the horizon lengthens. The impacts of price level
(Shock 6) and output (Shock 7) gradually increase at longer horizons. The secondhighest contribution in the medium run belongs to credit shocks (Shock 12). Other
shocks explain less of the variation in exports.
Figure 6.16 illustrates that imports own shocks (Shock 9) explain much of
their fluctuation over the period, although their contribution significantly decreases
in the medium run. Except for imports own shocks, the credit shocks (Shock 12)
explain much of the variations in imports in the medium run. The contributions of
exports (Shock 8), output (Shock 7), and price level (Shock 6) increase over longer
horizons, but not as much as that of credit shocks (Shock 12).
In short, the variance decomposition results show that domestic shocks explain
much of the fluctuation in the trade related variables (about 80 percent). Of this,
credit shocks play an important role in explaining the fluctuations of exports and
imports. Other monetary and financial variables explain less of these variations.

152

6.7

THE SVAR ANALYSIS FOR CONSUMPTION AND INVESTMENT


BEHAVIOUR
6.7.1

Contemporaneous matrix

This section presents the second extension of the basic SVAR model to
examine two other aggregate demand components (consumption and investment
behaviour) under the effects of the monetary policy. As with international
transmission, previous empirical studies, such as Afandi (2005), have shown interest
in examining such interactions of the monetary transmission mechanism. The
extended model puts two new variables - private investment (PI) and private
consumption (PC) - in the place of the two trade variables in Section 6.6 and the
specification is shown in Table 6.8.
Table 6.8: Extended Model with Investment and Consumption Behaviour
WP WRP WGP FY
WP

PI

PC

CR

1
a4,1

FFR a5,1

CPI

a6,2
a7,1

a7,2

a7,4

a6,7

a7,6

PI

a8,6

PC

a9,6

a9,7

a10,3

a10,5 a10,6 a10,7

a11,3

a9,5

a9,6

a7,12
a8,10

1
1

a10,11 a10,12

a9,7

a11,10

CR

a12,5 a12,6 a12,7

a12,10

VNI a13,1 a13,2 a13,3

a13,5 a13,6 a13,7 a13,8 a13,9 a13,10 a13,11 a13,12

VNI

WGP

WRP

FY

FFR CPI

a10,14
a11,14

1
1
1

a14,1 a14,2 a14,3 a14,4 a14,5 a14,6 a14,7 a14,8 a14,9 a14,10 a14,11 a14,12 a14,13

Note: WP = World oil price, WRP = World rice price, WGP = World gold price, FY =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, PI =
Private investment, PC = Private consumption, R = Short run interest rate, M = Money supply, CR =
Bank credit, VNI = Stock price index, E = Real effective exchange rate.

Source: Author's calculation.


153

The equations of PI and PC are based on the Keynesian assumption that


aggregate consumption is primarily a positive function of income and investment is
an inverse function of the interest rate. However, consumption is affected by other
variables including the price level (Sharifi-Renani, 2010). Thus, the equation for
private consumption includes both income and price. Moreover, the price variable is
also added to the equation for private investment, implying that investors in the
private sector observe both loans costs (interest rate) and inflation expectation
before making an investment decision. The inclusion of CPI in the equations for
both PI and PC helps to solve the difference between nominal and real values in this
study. The results in Table C5 (Appendix C) show that all the eigenvalues of the
model lie inside the unit circle, satisfying the stability conditions. Based on the
above restrictions, the results of estimation of the contemporaneous matrix are
shown in Table 6.9.
Table 6.9: Results of the Contemporaneous Coefficients for the Extended
Model with Investment and Consumption Behaviour
WP
WP

WRP

FY

CPI

0.317

23318

-0.025

CPI

CR

VNI

1
1
-0.072

-5682 874.94

195377

PI

-0.056

PC

-16253 395735
-0.065

-0.725

6.475 -32.988

0.436

-0.049

9.928

CR

61902
-0.042

PC

FFR -0.629

VNI

PI

WGP

FFR

WRP

FY

WGP

1
1

-3.293

-0.029

786.89 -141700 -213999


-1.877 -0.332 -1.108
0.598

0.509

0.147

-0.218 18.383
-7.525 0.170

-9.444

-6.16

7616

1
7.446

0.658

1.155

-5.223 -0.200 -7.167

-1.274

-0.092

2.054

154

0.966

48.602 -39.688

-0.053 -0.196 -1.958

1
0.334

Notes:
* WP = World oil price, WRP = World rice price, WGP = World gold price, FY = Foreign output,
FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, PI = Private
investment, PC = Private consumption, R = Short run interest rate, M = Money supply, CR = Bank
credit, VNI = Stock price index, E = Real effective exchange rate.
* The highlighted cells express the same significant contemporaneous effects with the base model.
The underlined cells express the additional significant results.

Source: Authors calculation.

As shown in Table 6.9, the extended model has 11 statistically significant


coefficients in which six results are similar to the basic results (the contemporaneous
effects of WP on FY; WP on VNI; WRP on CPI; and R, M and CR on VNI in the
highlighted cells). The coefficient of Y in the CPI equation is positive but
insignificant. This could be explained by the inclusion of the private sector
expenditure in the extended model of the Vietnamese economy. The new model
shows five more significant results (the underlined cells). First, the positive and
significant coefficient of FFR in the E equation (0.170) is reasonable, because an
increase in FFR leads to an appreciation in the home currency, and thus in the
domestic interest rate. This reflects the reaction of a small open economy like
Vietnam to a shock in U.S. monetary policy. Second, the positive coefficient of PI in
the VNI equation (1.155) illustrates that the stock market positively reacts
contemporaneously with an increase in private investment, reflecting the
contribution of private investment to stock market booms. Third, the negative and
significant coefficient of CR in the R equation is consistent with monetary theory.
Fourth, the two unexpected results are from the relationship between M and R, and
CPI and E. These results could be explained by the problems in the Vietnamese
banking system. The positively contemporaneous effect between M and R is
consistent with economic theory, but it implies that the interest rate channel seems to
be weak in transmitting a change in monetary policy (money-supply expansion or
contraction). This could be from the domination of a group of state-originated banks
(that is, state owned and equitized) banks in adjusting the interest rate. Thus, when
the price level increases, the monetary authority desires to reduce the monetary
aggregate to control inflation. However, the right sign for a change in the interest

155

rate that is, a decrease rather than an increase - does not occur, so the home
currency is depreciated.
6.7.2

Impulse

response

functions

of

private

investment

and

consumption to a contractionary monetary policy


This study discusses the responses of the impulse response functions of
investment and consumption behaviour to a contractionary monetary policy shock,
as shown in Figure 6.17. This shock is a positive innovation to the interest rate
(Shock 10).
The results illustrate that private investment responds with a three-quarter lag
as there is an increase in investment in the first three quarters before a decrease
beginning in the fourth quarter. In the medium run, private investment tends to
fluctuate around the base line. Thus, a money contraction negatively affects private
investment after the 3-quarter lag. Conversely, a negative response in private
consumption is recorded in the first quarter after a contraction in monetary policy.
However, after that time, consumption increases in the short run before decreasing
in the medium run. Figure 6.17 shows that the impulse response functions are not
significant over the period.
Figure 6.17: Impulse Response of Private Investment (PI) and
Consumption (PC) to Structural Shocks of One Standard Deviation in Interest
Rate (R)
- Response of PI:

Response of LNPI_SA to Shock10

.06

.04

.02

.00

-.02

-.04
2

10

12

14

16

18

Response of LNPC_SA to Shock10


.03

.02

156

20

.00

-.02

-.04
2

10

12

14

16

18

20

- Response of PC: Response of LNPC_SA to Shock10


.03

.02

.01

.00

-.01

-.02
2

10

12

14

16

18

20

Source: Authors calculation.

6.7.3

Variance decomposition

This section presents the results of the variance decomposition of investment


and consumption behavior, to understand the contribution of shocks to fluctuations
in these two variables. Figures 6.18 and 6.19 compare the results for private
investment and private consumption.
The shocks are to world oil price (Shock 1), world rice price (Shock 2), world
gold price (Shock 3), foreign output (Shock 4), foreign interest rate (Shock 5),
domestic price level (Shock 6), domestic output (Shock 7), private investment
(Shock 8), private consumption (Shock 9), interest rate (Shock 10), monetary
aggregate (Shock 11), credit (Shock 12), stock price (Shock 13) and real effective
exchange rate (Shock 14).

157

Figure 6.18: Variance Decomposition of Private Investment

Source: Authors calculation.

Figure 6.19: Variance Decomposition of Private Consumption

Source: Authors calculation.

158

Figure 6.18 shows that the fluctuations in private investment are mainly
explained by its own shocks (Shock 8) in both the short and the medium run. In the
medium run, shocks of private consumption (Shock 9) increasingly contribute to the
fluctuations in investment. The impacts of monetary and financial variables on
investment are relatively small, implying the weak influence of transmission
channels on private investment. The credit channel (Shock 12) has more effect on
private investment than do other channels.
Figure 6.19 illustrates that shocks of investment (Shock 8), price level shocks
(Shock 6), and consumptions own shocks (Shock 9) explain much of the fluctuation
in private consumption. The contribution of monetary transmission channels in
explaining the consumption variation is about 20 percent, implying that the role of
monetary transmission channels in consumption is bigger than their role in private
investment. In the short run, the exchange rate channel is the most effective (Shock
14), being replaced by the interest rate channel in the medium run (Shock 10).
In short, the variance decomposition results show that domestic shocks explain
much of the fluctuation in private investment and private consumption (more than
80 percent and 60 percent, respectively). The impact of monetary policy on private
consumption is higher than its impacts on private investment. The credit channel is
the most effective channel in transmitting monetary policy to private investment,
although the exchange rate and interest rate channels are the most effective in the
transmission to private consumption, in the short and medium run respectively.
6.8

THE ROBUSTNESS OF RESULTS


There are many different approaches to check the robustness of their results,

such as changing the number of lags and the sample length in the SVAR estimation
(Berkelmans, 2005), applying other standard errors (Aslanidi, 2007), and adjusting
restrictions in the contemporaneous matrix (Afandi, 2005; Bhuiyan, 2012). Specific
changes to the base model are described below to check the robustness of the current
research.

159

6.8.1

The adjustment in the lag length, sample length and standard

errors
This study could not obtain stability with different numbers of lags. The lag
length selection criteria results reported in Section 6.2 show that the lag length could
be one lag or two lags. However, when applying two lags, the LM test results for the
VAR stability cannot be obtained as the null hypothesis of no serial correlation is
rejected (Appendix C, Table C1). With more than two lags, serial correlation
between variables, appears to make the model unstable due to longer lags indicating
strong (longer) memory and the model facing the losses of degrees of freedom.
Therefore, one lag chosen in Section 6.2 is selected.
Moreover, a similar result appears when the study uses the sub-samples
(2001Q1-2011Q4) or (2000Q1-2010Q4), as the p-value is less than 5 percent, so the
null hypothesis of no serial correlation is rejected (Appendix C, Table C6).
Therefore, this study could not apply this approach to test for robustness. This result
is clarified that the smaller number of samples results in the serial correlation
between variables, leading to the instability of the model.
When computing the response standard errors with the Monte Carlo method
(with 1,000 repetitions), the impulse responses show quantitatively similar results
with the analytical (asymptotic) standard errors (Appendix C, Figure C1). With this
method, the standard deviation of the simulated impulse responses after 1,000
replications is the standard errors result. Computing Monte Carlo standard errors
confirms impulse response results are delivered reasonably because this is a better
measure of the variability of estimates than asymptotic standard errors.
6.8.2

The revision in restrictions

The study subsequently revises the model by adjusting restrictions, as shown


in Table 6.10. Specifically, the Y variable is excluded in the reaction function of the
monetary authority. As mentioned in Section 5.5, our analysis is conducted with the
inclusion of the output variable in the interest rate equation to reflect the targets of
the monetary authority, such as economic growth and price stability. This revision
implies that the assumption of an information lag used in previous studies (Cushman
& Zha, 1997; Kim & Roubini, 2000; Safaei & Cameron, 2003; Afandi, 2005;
Berkelmans, 2005) is included here to test the robustness of this study. The revision

160

also suggests a change in the State Bank of Vietnams contemporaneous reaction in


monetary policy. Specifically, rather than including contemporaneous output (Y) and
price (CPI) as the final target, the State Bank of Vietnam exhibits higher
independence by choosing the inflation target as its only contemporaneous target.
This revision, reflected in the R equation, which includes the CPI variable and
excludes the Y variable, is important, as previous results reported above in this study
also show that higher output leads to a higher pressure of price. Moreover, this
change is different to the Taylor rule, where both output and price are included in
the interest rate equation. With the above revision, this study obtained some
significant results which are reported in Table 6.10.
Table 6.10: Revised Contemporaneous Matrix
WP
WP

WRP

WGP

FY

FFR

CPI

0.842

-0.985

VNI

WGP

FY

-0.028

FFR

-0.553

CPI

-0.024 -0.073

-0.023 0.066

1
1

0.853

0.653

-5.716

2.074 -106.47

0.163

-0.045 0.506 -2.146 -0.012

CR

CR

WRP

VNI

0.011 -2.010 -0.724 0.152


-0.596 0.392 -0.578

0.077

-27.568 -3.738

45.727
0.861

1
1

5.201 -5.266 -0.136 -4.719 3.443

-16323 -29185 -1658 1047118 -2449 359061 -86948 -4067 -178997 280298 -8802
Notes: * WP = World oil price, WRP = World rice price, WGP = World gold price, FY =
Foreign output, FFR = Federal funds rate, CPI = Consumer price index, Y = Domestic output, R =
Short run interest rate, M = Money supply, CR = Bank credit, VNI = Stock price index, E = Real
effective exchange rate. The highlighted cells express the same significant contemporaneous

effects with the base model. The underlined cells express the additional significant results.
Source: Authors calculation.

161

0
1

First, the revised model still obtains all seven statistically significant
contemporaneous coefficients (the highlighted cells with the same sign) as obtained
in the base models. These results support the proposed approach in this study.
Second, the revised model also records some other statistically significant
results at the 5 percent significance level (the underlined cells). Because the revised
model implies that the monetary authority follows the price stability target, rather
than the economic growth target, three more statistically significant coefficients are
revealed: the contemporaneous impacts of FFR on M, Y on M, and E on R. The
negative coefficient of FFR on M implies that an increase in the U.S. interest rate
leads to a decrease in the monetary aggregate; in turn causing the domestic interest
rate to increase. Obviously, this is consistent with the common trend in home
interest rates to avoid the inflationary impact of the devaluation in these economies
currency when the Federal Fund rate increases. Next, the negative significant
coefficient of Y on M shows that the monetary aggregate decreases after an increase
in output, consistent with reactions in monetary policy. Specifically, the increase in
Y results in a significant rise in CPI, implying a higher price level. Due to the
assumption that the monetary authority chooses the domestic price as its optimal
objective, a decrease in M contributes to controlling the increase in the price level.
Last, the positive coefficient of E on R reflects the theoretical consistency that the
appreciation of the home currency (real effective exchange rate increases) induces
the domestic interest rate to increase. This increase supports the suggesting that the
monetary authority acts according to the inflation target. Overall, the revised model
suggests changes in the monetary policy (the monetary authoritys target) can
contribute to control inflation.
6.9

CONCLUDING REMARKS
This chapter has examined possible relationships between foreign and

domestic macroeconomic variables for the case of the Vietnamese economy.


Specifically, with the proposed non-recursive SVAR models, the study has
presented results of interaction between five foreign variables (world oil price, world
rice price, world gold price, Chinese output as foreign output, and Federal Fund rate
as foreign interest rate) and seven domestic variables (domestic output, price level,
interest rate, monetary aggregate, credit, stock price as asset price, and real effective
162

exchange rate). To do this, two versions of base model were estimated and analysed
using SVAR techniques including specifying and estimating the contemporaneous
matrix, impulse response functions, and variance decompositions. As mentioned in
Chapter 5, the first version (model VN1) reflects the strongly administered oilpricing regime while the second version (model VN2) reflects the weakly
administered oil-pricing regime.
The study has found that there are seven statistically significant relationships
between variables in terms of their contemporaneous effects. They are the
instantaneous impacts of the world rice price and domestic output on domestic price,
and the effects of the world oil price, the domestic interest rate, and credit on the
domestic stock prices. The negative effect of world oil price shocks on output is
insignificant, as expected due to the controlling regime for oil prices applied in
Vietnam. However, a significantly negative impact of these shocks on Vietnams
stock market is recorded. Moreover, the rapid economic growth in Vietnam creates a
greater hazard of inflation pressure, so this relationship should be one of the leading
concerns of policymakers in terms of the development of the Vietnamese economy.
Changes in monetary policy play an important role in the development of the stock
market. The study also confirms that the Vietnamese governments policies to
ensure national food security, as well as its administered oil price regime, make
important contributions to controlling price stability. Both of the proposed base
models in Chapter 5 are useful because their restrictions are valid according to the
over-identification test. Moreover, although model VN1 (with a strongly
administered regime for oil price) is slightly more sensitive, the difference is
negligible with model VN2. This finding is consistent with the similar patterns
detected for both models in terms of the impulse response functions.
Results from the impulse response functions reveal that foreign shocks do not
generate significant effects to domestic variables over the short and long run.
However, in the short run, foreign shocks, such as those of foreign output on
domestic output, foreign output on the monetary aggregate, world gold price on the
monetary aggregate, and world rice price on the domestic price, are found to have
significant influences. These factors need to be taken into account in policy making,
as they relate to the openness of the Vietnamese economy. Impacts of foreign
interest rates on domestic monetary policy occur with some lags. Moreover, the
163

results also confirm some current policies that the Vietnamese government has been
applying to control oil prices and ensure the effectiveness of the national foodsecurity policy, contribute to stabilising the price level. The Vietnamese monetary
authority seems to be following a policy to narrow the gold price gap, but this has
only been the case for a short time. The impulse responses of domestic variables to
monetary policy shocks suggests the output puzzle does not appear over the 20
quarter period. However, the price puzzle, liquidity puzzle, and exchange rate puzzle
do appear over short periods (six quarters, one quarter, and two quarters,
respectively) and disappear afterwards. Over the medium term, the home currency
depreciates after its appreciations in the short term, implying that this response
complies with the principles of uncovered interest-parity bias, and so there is no
evidence of the forward-discount puzzle. The impulses responses of variables in the
money market and the financial market indicate that short run policy adjustments
could limit the effects of previous policies due to the transmission lag between these
variables. The real effective exchange rate plays an important role in the
development of the Vietnamese economy. Specifically, its effects on the output, the
monetary aggregate is significant over the period, while its impact on the stock
market is significant in the short run. The response of the stock market to a monetary
contraction is not large; this, along with the analysis of variance decomposition of
stock prices, indicates the moderate openness of the Vietnamese financial market (Li
et al., 2010).
In terms of variance decomposition, domestic factors are found to play a
greater role than foreign factors in explaining fluctuations of domestic variables in
both the short run and the medium run. This implies that foreign activity is not a
substantial contributor to domestic activity, so negative effects from events in the
world economy, such as the 2008 financial crisis, do not greatly affect the
Vietnamese economy. These findings are different to those in Dungey and Fry
(2000) for a small, open economy like Australia, in which the overseas sector and
asset prices play an important role in the growth cycle. The difference could be
explained by the fact that Australia is a developed small, open economy, while
Vietnam is a developing small, open economy. The Vietnamese governments
administered policies seem to limit the influence of the foreign sector on the
domestic economy, which is obviously an advantage for this economy in coping
164

with outside recessions. The study finds that foreign output has a bigger role in
explaining fluctuations in domestic output in the medium run, and that the world rice
price has a bigger role in explaining fluctuations in price level in the short run.
Moreover, due to the Vietnamese governments oil price controlling regime, the
impact of world oil price on output is relatively small but its influence on price level
tends to increase in the medium run.
In terms of monetary policy channels, the exchange rate and interest rate
channels have the most effect on real output and price level, respectively; the
interest rate becomes the biggest contributor to the variation of price level in the
medium run. The asset price is the least effective channel in interaction to the
production sector. The weak role of the asset price channel in the case of Vietnam (a
developing economy) is different to its important role in a developed economy like
Australia, as found by Dungey and Fry (2000). Moreover, the small contribution of
the monetary aggregate in explaining the fluctuation of price level seems to imply
that there is no strong link between money and inflation in Vietnam. Le and Pfau
(2009) had surprisingly similar results. In the money market, the fluctuations in
interest rate and monetary aggregate are largely explained by the shocks to monetary
policy variables: credit affects the former, and credit and the interest rate affect the
latter. The gold price is found to have a bigger effect on the interest rate and the
monetary aggregate one quarter after the shock. The role of the gold price in
conducting Vietnams monetary policy had previously been found by Tran (2009),
but confirmed in this study. Changes in the money market affect the variables in the
financial market, but the influences of the interest rate and monetary aggregate on
the real effective exchange rate are stronger. Monetary and financial variables do not
explain much of the fluctuation in credit and asset price; only about 10 and 20
percent, respectively. The output shocks are found to be the biggest contributing
source to credit variation, implying a strong relationship between credit and
economic growth in the Vietnamese economy. Compared to shocks in the
production sector (output and price level), the effects of monetary policy on the
stock market are stronger.
When considering the international transmission in the revised model with
trade related variables, significant contemporaneous relationships between domestic
output and total imports, and between total exports and the stock market are found,
165

but the latter is inconsistent with economic theory. The former result implies the
growth in domestic production helps improve Vietnams trade balance, while the
latter suggests that Vietnam should guard against overly exploiting natural resources
for export growth. The results of the impulse response analysis in this model show
that both imports and exports are quickly affected by a contractionary monetary
policy, and that this induces the trade deficit to increase. However, monetary
contraction also has negative impacts on import demand in the short run (about two
quarters after an increase in the first quarter), which contributes to improving the
trade balance of Vietnam in the short run. A similar point is made by Kubo (2008)
for Thailand. Moreover, the results of variance decompositions indicate that
domestic variables are important in explaining much of the fluctuation in exports
and imports. Credit is the transmission channel with the most effect on the trade
balance. Simultaneously, it is also the second-highest contributor to variations in
exports and imports. Both output and price level do play a more significant role in
affecting the trade balance in the medium term.
This study records a positive and significant relationship between private
investment and stock market booms. The variance decomposition results illustrate
that domestic shocks have a larger role than foreign variables in explaining
fluctuations in private investment and private consumption. Consumption is affected
by monetary policy to a greater degree than private investment. The credit channel is
the most effective channel in transmitting monetary policy to private investment. In
the short run, the exchange rate plays an important role in affecting private
consumption, but the interest rate plays this role in the medium run.
In testing the robustness of the study, different approaches, such as computing
other response standard errors and adjusting restrictions, confirm the previous
results and analyses. Moreover, the results from revised restrictions suggest that the
Vietnamese monetary authority is becoming more independent in choosing inflation
as its monetary policy target, rather than both economic growth and price stability.
However the tradeoff between inflation and economic growth is a challenge for the
monetary authority as poverty is always the biggest concern of an emerging
economy like Vietnam.

166

CHAPTER 7
SUMMARY AND RECOMMENDATIONS

7.1

SUMMARY
This thesis uses a non-recursive SVAR approach to the monetary transmission

mechanism of Vietnam, using quarterly data for the period 2000 to 2011. The
sample period covers important events in the world and Vietnamese economies
including Vietnams continuing international economic integration, the global
economic recession, in the changing practice of Vietnams monetary policy and the
development of the Vietnamese stock market. The study presents two base models
with twelves variables from both the foreign and domestic sectors. SVAR analytic
techniques are used to assess the role of monetary transmission channels and their
effects on the financial markets aggregate demand, real output and prices in
Vietnam.
The literature in Chapter 2 illustrates that few studies have focussed on
developing small, open economies, and even fewer have covered characteristics
such as the low independence of monetary policy (the role of the central bank), the
developing financial markets (few participants, lack of institutions, narrow markets
with thin trading) and the structural changes in the economy. These characteristics
are important for economies, especially with the developing economy of Vietnam,
where policymakers always pay attention to the trade-off between economic growth
and price stability, and there are structural influences on the economy.
Moreover, previous studies have predominantly addressed the transmission of
policy instruments settings to the domestic financial sector (credit, stock prices) or
the production sector (output and prices). They have neglected aggregate
expenditure in the form of investment and consumption behaviour in Vietnam and
the comparison of different monetary transmission mechanisms (including the asset
price channel). The literature review also showed the SVAR approach has not been
widely applied to the case of Vietnam. This thesis fills this identified research gap.
To the best of the researchers knowledge, this study is the first attempt to compare

167

four transmission channels of Vietnamese monetary policy (the interest rate, credit,
exchange rate and stock prices channels).
Chapter 3 provides a general picture of the development of the Vietnamese
economy, institutions and monetary policy. The time series data, relevant to these
characteristics, was collected to study the monetary transmission mechanism in
Vietnam. Chapter 4 describes and analyses the detected complex seasonality
problems in the raw data, associated with moving holidays and adopts, for the first
time, a suitable procedure to account for this. Analyses of the unit root properties of
the data found a mixture of stationary and non-stationary time series. Evidence was
also found of a structural break in the 2007-09 period, reflecting these structural
change in the Vietnamese economy due to the join in WTO, effects of the Global
Financial Crisis.
Adopting a theoretical economic and econometric framework, Chapter 5
proposes two versions of a base SVAR model, one reflecting a strongly
administered oil-pricing regime (SAOP), the other a weakly administered oil-pricing
regime (WAOP). This consideration is meaningful for Vietnam where the
government applies the controlling policy for the domestic oil price. The SVAR
models in this study follow two versions of the SVAR model, as described in
Chapter 5, with two sectors (foreign and domestic) and three dimensions in the
domestic economy (the production market, the money market, and the financial
market). This model, plus two extended models are used to examine in Chapter 6 the
effects of a monetary contraction on international trade activities, and the aggregate
demand components. With this research design, this study addresses and answers the
following important questions raised in the introduction chapter of this thesis:
(1) Are domestic monetary shocks associated with fluctuations of output,
price and other fluctuations in the economy?
The empirical analysis reveals that the impacts of a monetary contraction on
domestic economic variables in Vietnam are largely consistent with theory and
expectations. Whilst this is true for real output it does not hold for price level.
Specifically, in response to a contractionary monetary policy shock, monetary
aggregates fall, the interest rate increases, credit and the stock prices decrease, and
the domestic currency appreciates. However, the price level rises, rather than fall as
168

expected. This result implies that monetary policies seem to be ineffective in


controlling inflation in Vietnam during the period under examination. Results from
the impulse response analysis also show that the appreciation of the home currency
is not strong and persistent, so the monetary authority could intervene to restrain the
appreciation of the domestic currency with a small expansion in the money supply
(about four quarters, from quarters 6 to 10), which in turn create more pressures on
the price level. Moreover, the variance-decomposition results illustrate that the
contribution of monetary policy in fluctuations in price level is relatively large; thus,
the monetary policy has failed to control price level.
As for real effects, contractionary monetary policy causes a reduction in
output, although the response is weak. Furthermore, the variance decomposition
finds that the monetary policy channels explain more than 20 percent of fluctuations
in real output. Economic growth causes inflationary pressures due to the
significantly contemporaneous effect of output on price level, along with the weak
influence of monetary policy on output, so controlling inflation is still one of the
great challenges faced by the Vietnamese economy. Moreover, the effect of
monetary policy shocks and output shocks in economic growth seems to be
consistent with the findings in the seasonality analysis that the dominant determinant
of output could be factors other than monetary policy.
The findings show that structural breaks occurred over the period of 2000-11,
especially the statistically significant break dates (2008:2 and 2010:4) are found in
this study seems to relate some important events in the Vietnamese economy such as
joining the WTO, the financial crisis. Based on this result, a dummy variable is
included in the SVAR model of Vietnam.
(2) Is the interest rate channel the most important transmission channel of
monetary policy?
Results from variance decomposition illustrate that the interest rate channel is
the most effective mechanism for transmission to the price level, but not to output.
Specifically, compared to other channels, interest rate shocks explain a great deal of
fluctuation in price level from the second quarter onwards. In the short run, the
contribution of the credit channel to fluctuations in the price level is larger than that
of the interest rate channel. The credit channel is the second most effective channel
169

in the medium term, while the asset price is the least important channel in explaining
variations in price level. Moreover, results from the impulse response illustrate that
there is no lag in transmitting shocks in the interest rate and credit to price level, and
these responses are significant at particular stages (11 quarters and four quarters,
respectively). As for output, the order of the effective channels is different. The
exchange rate is found to be the most successful channel, followed by the interest
rate channel and the credit channel. The response of output to shocks in the
exchange rate is statistically significant and without lag. The stock price channel is
the least effective channel, which is consistent with the modest role of the stock
market in the development of the Vietnamese economy. Although the study finds
statistically significant contemporaneous effects of monetary policy (especially from
the interest rate and credit) on stock prices, the small impact of the stock prices
channel in the fluctuation of output and price level results in the fact that the
transmission of monetary policy via the stock market was not as effective as
expected.
(3) Do foreign shocks have a more significant impact on the Vietnam
economy than domestic shocks?
The study finds that foreign shocks have less significant impact on the
Vietnam economy than domestic shocks.
This result is different to the assumption that a small, open economy could be
greatly affected by shocks in the world economy. Results from the impulse response
functions and variance decomposition illustrate that domestic shocks are the main
factors in the fluctuations of domestic variables over the short and medium run. The
foreign shocks explain less of the fluctuations in key variables of the Vietnamese
economy, such as about 20 percent of Y (output), VE (exports), VI (imports) and PI
(private investment); about 40 percent of CPI (price) and PC (private consumption).
This finding helps to explain the fact that the Vietnamese economy is not under
significant influence of negative foreign effects, such as the financial crisis in the
periods 2007-2010.
In the short run, the current study finds significant effects of foreign output on
domestic output and the monetary aggregate; world gold price on monetary
aggregate; and world rice price on domestic price level. Also, there are
170

contemporaneous impacts of world prices including world rice price on the price
level and world oil price on stock prices. Results from variance decomposition
illustrate that the role of foreign output in explaining fluctuations in output is bigger
in the medium run. In addition, results from the impulse response suggest that in
examining the effects of foreign monetary policy on domestic monetary policy, there
are some lags in domestic interest rate after the shock of foreign interest rate.
(4) Does a monetary contraction positively affect components of aggregate
demand/ expenditure of Vietnam?
The empirical results reveal that the effect of a monetary contraction on
exports is consistent with both economic theory and expectations, but the expected
positive response of imports only occurs in the first quarter. However, the monetary
contraction has negative effect on imports over a short period (from the second to
the sixth quarter). At the same time, exports still decrease after the monetary
contraction as expected, so the contraction does not produce a deterioration of
Vietnams trade balance.
Results from the variance decomposition reveal that credit is the most
effective channel in transmitting monetary policy to trade activities in Vietnam. This
finding provides more evidence about the necessity of increasing the monitoring role
of the State Bank of Vietnam with respect to the credit channel. The effects of
monetary contraction can be transmitted to credit, which in turn affects exports and
imports. The finding in which domestic output has significantly and negatively
contemporaneous effects on imports suggests that policymakers could use the credit
channel as a means to promote domestic production, in turn limiting the volume of
imports, to improve the trade gap.
In addition, the simulation reveals that the effect of a monetary contraction on
private investment and private consumption is consistent with economic theory and
with expectations but with the 3-4 quarter lag and these expected responses also
appear in a short time. These findings suggest the following implications. First,
Vietnamese private enterprises have difficulty in borrowing from banks, so the
effects of monetary policy credit shocks on investment are not too strong. These
difficulties could be from stringent borrowing requirements or the priority given to
enterprises from the state sector in obtaining loans from banks. Second, households
171

in Vietnam are relatively unfamiliar with borrowing from banks for consumption
purposes. In addition, results from the variance decomposition illustrate that the role
of monetary policy is not dominant for private investment and private consumption,
but the effects on private consumption are bigger than those on private investment.
In terms of transmission channels of monetary policy, credit is the most effective
channel in affecting private investment, and private consumption is most affected by
the exchange rate and the interest rate in the short and medium runs, respectively.
(5) Do puzzles appear in the model? If yes, what are they and what do they
imply?
This study finds no evidence of the output puzzle and the forward discount
puzzle. Specifically, over the period examined, a monetary contraction in the form
of an increase in the domestic interest rate does not cause any increase in output.
In the short run, there is evidence of the price puzzle, the liquidity puzzle, and
the exchange rate puzzle (within six quarters, one quarter and two quarters,
respectively). A depreciation of the home currency appears after its initial
appreciation. Kim and Roubini (2000) assert that the liquidity puzzle implies that
shocks in the monetary aggregate might not be the correct representative for
adjustments in monetary policy. The price puzzle implies that inflationary pressures
are partly reflected in shocks in the interest rate, and this explanation is used to
understand the exchange rate puzzle. It supports the appearance of all three puzzles
in the short run in Vietnam. The price puzzle has the longest effect while the
liquidity puzzle has the shortest effect. The appearance of these puzzles exist in the
short run and they are resolved afterwards.
In short, there is no evidence of the output and forward discount puzzles. The
puzzles of price, liquidity and exchange rate are addressed but they are transitory
and they disappear over time.
7.2

POLICY RECOMMENDATIONS
The examination of the monetary policy transmission channels for the small

open economy of Vietnam provides the following policy recommendations:


(1) There is a need to focus on solving the internal problems of the Vietnamese
economy to create a strong motivation for the development of Vietnam.

172

This recommendation results from the finding that domestic shocks play a
crucial role in changes in the domestic economy. Therefore, to ensure the stable
development of the economy, policies should be taken into consideration that will
explore the sources for economic growth; stabilise price level; improve the
effectiveness of policy instruments; restructure the enterprise sector, especially the
state-owned enterprises; and grow the stock market soundly. Any failure to improve
domestic factors negatively affects the economy.
At the same time, policymakers should know the effects of foreign shocks on
domestic variables, especially output and price level, because they have significant
effects in the short run. Policymakers need to give suitable attention to foreign
factors in considering specific targets. Specifically, to control inflation,
policymakers should pay more attention to world prices, especially in the short run.
Policies should not focus on following overheated economic growth which results in
an increase in price level. Next, increased price level causes a decrease in Vietnams
exports. Policymakers should address foreign output because of its increasing
contribution in explaining fluctuations of domestic output over the longer horizon.
(2) Monitoring the important transmission channels, including the interest
rate, exchange rate and credit channels in formulating monetary policy.
Coordinating monetary policy with other policies to effectively control
inflation in Vietnam.
The findings reveal that these channels should not be ignored as they are the
most effective transmission channels. The exchange rate and the interest rate
particularly affect output; the interest rate and credit particularly affect the price
level; credit particularly affects imports, exports, and investment behaviour; and the
exchange rate and the interest rate particularly affect consumption behaviour in the
short run and the long run, respectively. The stock price channel is found to be the
least significant channel for both the output and price level. The significance of the
exchange rate reflects the unique features of a small, open economy, so this finding
is useful for the case of Vietnam. To best control these transmission channels, it is
necessary to improve legal and institutional factors in the money market as well as
enhance banks financial-intermediation function. To ensure effective monitoring,
the independence of the Vietnamese monetary authority should be improved,
173

especially through a better combination of market-based and non-market-based


solutions. The interest rate liberalisation should be further liberalised. To ensure the
effectiveness of the credit channel, policy makers need to consider restructuring the
state-owned enterprises and banks and adopting effective solutions to reduce bad
debts. The improved effectiveness of the credit channel largely affects the trade
activities via the international transmission. Due to the small contribution of
monetary policy in variations of private investment and private consumption, the
effect of the credit channel is limited. The significant effects of the interest rate and
credit via the asset price (stock price) channel imply that the State Bank of Vietnam
has enough market-based instruments to control bubbles in the stock market.
However, the significant and negative effects of monetary aggregates on stock prices
imply that banks should be soundly supervised to reduce the unexpected interaction
of monetary policy.
In addition, it is necessary to have a comprehensive combination between the
monetary policy (especially the interest rate and credit channels) and other policies
to curb inflation. This recommendation is based on the finding that monetary
contraction has failed to control inflationary pressure from 2000 to 2011 in Vietnam.
An increase in output makes inflationary pressure while monetary policy is not a
dominant source of fluctuations of output. Therefore, it will be better to consider
other policies in efforts to control inflation. One of the solutions is that policy
makers should give attention to fiscal policy with the strict and effective supervision
on government spending.
7.3

CONTRIBUTION AND SIGNIFICANCE OF THE RESEARCH


This research attempts to fill research gaps to provide a comprehensive

analysis by the consideration of the aggregate demand components, characteristics


of the independence of monetary policy and structural changes for a developing
economy. It has makes significant contributions to the study of the transmission
mechanism of Vietnamese monetary policy:
(1) The present study is the first empirical study examining four transmission
channels of the Vietnamese monetary policy - the interest rate, credit, exchange rate
and stock price channels - in the period 2000 to 2011. Le and Pfau (2009) had
previously examined three channels including the interest rate, credit, and exchange
174

rate channels from 1996 to 2005. The Vietnamese stock market was established in
2000, creating another channel to accumulate capital for the development of the
Vietnamese economy. Therefore, this study attempts to compare the strength of
these four channels. Results illustrate the significantly contemporaneous effects of
monetary policy tools on stock prices in the period of study.
(2) This work is the first attempt to apply the large, non-recursive SVAR
model to the Vietnamese economy. Tran (2009) also used a non-recursive SVAR
model, but it only includes four variables that do not cover two sectors (foreign and
domestic) and three dimensions of the domestic economy (the production, the
money market and the financial market). This study uses 12 variables in its base
models and 14 variables in its extended models. Using a non-recursive SVAR model
helps to reflect the structure and characteristics of the Vietnamese economy.
(3) This thesis is the first empirical study to apply different unit root tests and
consider structural breaks in the case of Vietnam. Results from unit root tests
without and with structural breaks are compared. To the best of the researchers
knowledge, this is the first time that tests with endogenous determined structural
breaks proposed by Lee and Strazicich (2003; 2004) are applied in the study on
Vietnam. Based on the significant break dates, this study uses a dummy variable to
cover the 2007-2009 financial crisis.
(4) This research is the first empirical study to examine many different
aspects of a small open economy, including the role of the exchange rate in the
economy; the effects of foreign output shocks, world price shocks and foreign
monetary shocks on the Vietnamese economy; and the impacts of monetary shocks
on trade activities. Moreover, extending the study of Tran (2009), this study is the
first effort to re-examine the relationship between the gold price and Vietnamese
monetary policy.
(5) This thesis is the first empirical study to examine the effects of shocks in
monetary contraction on aggregate demand components of Vietnam: private
investment, private consumption, exports, and imports (although not government
expenditure). The inclusion of exports and imports helps in the examination of the
international transmission of monetary policy, while the inclusion of private
investment and private consumption examines the behaviour of the private sector
175

and households. In the literature not only for Vietnam but also for other countries,
little research has been conducted in these areas.
(6) The research findings provide recommendations for Vietnamese
policymakers in considering the relationship between economic growth and
inflation, the interaction between the monetary policy and economic indicators, and
the effects of foreign and domestic shocks to the development of the Vietnamese
economy.
7.4

SUGGESTION FOR FURTHER STUDIES


The current research is exploratory work and there is much to be done.

Although findings from this study is significant in explaining how different channels
of monetary policy can affect the intermediate and final targets of the Vietnamese
economy, the comparison of effects of foreign and domestic shocks on the domestic
economy, further research can be undertaken with this area:
(1) The study only focuses on the stock prices for the asset price channel,
while this channel includes other important asset prices, such as housing and land
prices. Due to the unavailability of the data, this approach is not attempted in this
study. Therefore, if future studies cover more asset prices to examine the asset price
channel, it will be useful for policymakers to know and compare the effectiveness of
the different asset price sub-channels.
(2) Puzzles appear in this study, although they occur only in the short run. If
they are expected to disappear, it is necessary to consider an expansion to the SVAR
model. This should be employed in future studies on Vietnam.
(3) The study does not estimate two sub-sample periods before and after the
financial crisis of 2007-2009. This research direction is useful but it has not been
included in the scope of this study. To limit the disadvantage of using industrial
output as a proxy of GDP in an agricultural economy like that of Vietnam, this study
uses quarterly GDP from 2000 to 2011, so the sample size is not be enough to divide
into two sub-samples (before and after 2009). Future studies with a longer sample
size could strengthen this approach.
(4) The study only applies one type of restriction (short run) as the SVAR
framework does not allow the consideration of two restrictions (short run and long
run) within a model at the same time. Because the study focuses on temporary
176

effects on the Vietnamese economy (in the short and middle run) the short run
restriction is sufficient here. Studying both restrictions simultaneously could be
explored in future research.
(5) The study uses the Chinese GDP as a proxy for foreign output to examine
the effects of foreign output on domestic output, which is a relationship
characteristic of small, open economies. This approach is useful, as foreign output is
found to affect domestic output over longer horizons, and the effects in the medium
run are not small. However, further research could examine different foreign output
variables to obtain a comparison between the effects of different economies;
information about what foreign output affects domestic output is useful to the
formulation of Vietnamese policies. However, such an approach will require an
expansion in the scope and model of the research.
(6) The study uses the SVAR approach which is useful to analyse the
dynamics of a model via subjecting it to an unexpected shock. However, another
competing methodology such as the VECM and DSGE models could be explored as
further examination on MTM of Vietnam.
The above limitations are not explored in this study due to its scope and the
time limitation; therefore, they should be considered in future studies.

177

APENDIX A - APPENDIX TO CHAPTER 2


Table A1: Milestones of Vietnam economy from 1986
Period

Year

Overview
- The first period of the Reform (known as Doi Moi) with average economic growth 3.9%.

1986-1990
1986

- The year of the beginning of the Reform with three main objectives: (1) Reforming to the market economy under the States
management; (2) Constructing a democratic and jurisdictional society; (3) Implementing open policies, strengthening the friendly
relationship with all countries. The inflation rate at this time was 774% (hyperinflation).
- Diminishing system based on administrative subsidies in all fields of the life and business. Handing over autonomy to stateowned enterprises (SOEs) and solving the state of false profit, true losses in SOEs sector.
- Diminishing the isolated state of the domestic market and initially integrating to the international market.

1987

- Law of foreign investment and Law of land were approved.


- Inflation reduced to 400%.

1988

- Abandoning States distribution role and authorizing SOEs, foreign-invested firms, private firms to conduct directly exportimport activities. Confirming households right of land use.
- First foreign joint venture enterprise was set up.
- Non-tax policies such as quota, export-import monopoly were gradually decreased. Export turnover was 1 billion USD.

1989

- Diminishing the system of administrative targets in state management. Handing all autonomy in business to SOEs and
beginning SOEs restructuring (12000 SOEs in 1989). Diminishing States subsidization to government officials.
- Liberalizing price, terminating the two-price system to follow the system of market price. Terminating compulsory trade
mechanism to farmers. Vietnam became the 3rd largest exporter. The Ordinance on economic contract was approved.
178

Period

Year

Overview

1990

- The State expanded the number of enterprises doing business in the field of international trade, from 40 businesses in 1985 to
270 businesses in 1990. Taking inventory SOEs assets were implemented. Law on foreign investment was revised; Law on
companies and Law on private enterprises were approved.
- GDP growth was 8.3%, 20 million of food was produced, and 2 million of crucial oil was exploited. Registered FDI was over
1 billion of USD.
Super-inflation had been controlled (the inflation in 1986 was about 747.7%, but 17.43% in 1990).
- Making initial changes from planning economy to market economy. Economic growth was 8.18% and the highest rate was

1991-1995

9.54% in 1995.
1991

- No economic aid from the USSR and socialist countries.


- Industrial growth in 1991 was 9%.

1992

- New Constitution with the official confirmation on a multi-background economy. SOEs equitization process was started in
1992. Gradually reducing financing from state budget for weak SOEs.
- Signing a trade agreement with EU. Registered FDI was 5 billion of USD. Inflation was 17.5%.

1993

- The US embargo to Vietnam was removed. Vietnam established the relationship with international donors.

1994

- Terminating export licenses for goods, except rice, timber and oil. Registered FDI was 10 billion of USD.

1995

- Law on state enterprises and Civil Law were approved. Reducing the number of quota to 7 types of good.
- Vietnam joined to ASEAN.

1996-1999

- Economic growth was decreasing in trend because of the Asian crisis. The rate in this period was 7% lower than the 5-year
plan objective.
179

Period

Year
1996

Overview
- Vietnam joined to the ASEAN Free Trade Area (AFTA).
- GDP growth was 9.34%. FDI reached the highest with the registered 27 billion of USD.

1997

- Commercial Law was approved to support for free trading in the market. Terminating obstacles in rice trading in the domestic
market. Three million tons of rice was exported and 10 million tons of crucial oil was exploited. Economic growth was 8.2%.

1998

- Vietnam economy was affected by the 1997 financial crisis. (The economic growth reduced to 5.76% and 4.77% in 1998 and
1999, respectively).

1999

- The economy target changed from fighting high inflation to stimulating inflation due to deflation.
- Law on enterprise (new) was approved. Enterprises are allowed to do business in all economic fields which are not banned by
legal system. Enterprises are freely allowed to export and/or import. 4.5 million of rice was exported. Equitization result in 1999
recorded the highest number from 1992 (249 SOEs). FDI was rapidly decreased. Vietnam joined to APEC.
- Reform focussed on restructuring the economy and SOEs equitization. The economy was restored with the growth 6.8% in

2000-2007

2000, 8.4% in 2005 and 8.46% in 2007. The average rate in this period was 7.7%.
2000-2001

- There was deflation in the economy in the period of 2000-2001. This state was solved from 2002. Vietnam economy was
gradually recovered with an increase in the economic growth. Vietnam signed the Bilateral Trade Agreement (BTA) with the
United State in 2001.

2002

- The Ordinance on price was approved, so most goods price was not subsidized.

2004

- Increasing world price negatively affected domestic production. Avian influenza in most provinces, drought and prolonged
cold caused price to increase. Inflation began to increase rapidly (from 3.0% in 2003 to 9.5% in 2004).
- Law on bankruptcy was amended.
180

Period

Year

Overview

2005

- Law on enterprises was amended. Law on investment was issued to replace the Law on Foreign Investment and Law on
Domestic Investment Promotion. These new laws created an equal environment for all economic sectors and removed the
distinction between domestic and foreign investment.

2006

- The Securities Law was introduced.

2007

- Vietnam became the 150th member of the World Trade Organization (WTO) in 2007. Foreign investment rapidly increased.
- Vietnam economy faced to the effects of subprime crisis of the U.S. and the high oil price at 100 USD/barrel.
- Trade deficit rapidly increased to 14.2 billion of USD. The total market capitalization of the stock market reached 43% GDP.
- Vietnam economy was affected by world economy recession.

2008-2011
2008

- In March, Vietnam government decided to implement solutions for controlling inflation, stabilizing macro-economy,
ensuring social security and sustainable growth and one of them was tightening monetary policy; however, there was a change in
monetary policy from November.
- Fluctuations in world market, especially in oil, gold and rice prices significantly affected Vietnam economy. In addition,
natural disasters, especially strong cold periods, negatively influenced on production.
- Trade deficit recorded to increase to 18 billion of USD, accounting for 29% of export turnover. The total market
capitalization of the stock market quickly decreased to 15% GDP. Vietnam was out of the list of poor countries on the world.

2009

- Vietnam was affected by the 2007-2008 financial tsunamis negative impacts: the high inflation pressure, the decreasing
economic growth, and the recession of the stock market. Trade deficit in 2009 decreased, but remaining at a high level (12.8
billion of USD or 22.4% GDP). However, economic growth remained positive.
181

Period

Year

Overview
- Enterprises business faced with difficulties, such as high lending rate, stagnant consumption market. Vietnam government
implemented stimulus package in 2009.

2010

- Vietnam ranked a higher position in national competitiveness by the World Economic Forum (WEF) from 4.0 (2009) to 4.3
points (2010). Production capacity of the industry increased because of the operation of some main thermal power, hydropower
and oil refineries. The export market was expanded, the trade deficit continued to decline (12.4 billion of USD).
- Vietnam economy worsen: 2-digit inflation, higher input costs for businesses, prolonged floods, declined stock market.

2011

- Vietnam continued to be affected by the negative effects from world depression, debt crisis in Europe, high oil price from
instability in North Africa and the Middle East.
- Vietnam government focussed on controlling inflation, stabilizing macroeconomic, and maintaining economic growth at a
suitable level. Vietnam decided to implement plans to restructure the economy, focussing on three fields: public investment, stateowned corporations, and the banking system.

Source: Pham and Vuong (2009), and authors summary.

182

Table A2: Milestones of Vietnam financial market and monetary policy from 1986 to 2011
Period

Year

Overview

1986-1990
1986

- Gold and foreign currency was allowed to transfer to Vietnam after proving legal origin.

1988

- Loosening regulations on foreign exchange to support enterprises payment.

1989

- A new legal framework for a two-level banking system replaced for a mono-bank system of the State Bank of
Vietnam before 1989.
- Changing from negative real interest rate and different rates for different economic sectors in the period before March
1989 to positive real interest rate contributed to solve difficulties in mobilizing deposits from residents. Deposits increased
to 6.7% of GDP at the end of 1989 (0.8% of GDP in 1988). Inflation reduced to 34.7% at the end of 1989.
- Multi-exchange rate mechanism was replaced by the exchange rate between Vietnam currency (VND) and USD. This
official exchange rate was frequently adjusted to be close with the market rate. Based on the official rate published by the
State Bank of Vietnam, commercial banks determined their rate in the band 5%.

1990

- The Ordinances on the State Bank of Vietnam and commercial banks, credit co-operatives and financial companies
took effects from May. The system of commercial banks was setup and join-stock banks were formed. Foreign banks were
allowed to open their branches or to form a joint-venture.
The Ordinance on the State Bank of Vietnam regulated the ratio of required reserve which was 10% at minimum and
35% at maximum.

1991-1995
1991

- First joint-stock commercial banks were set up.


183

Period

Year

Overview

1992

- The State Bank of Vietnam followed a cautious monetary policy with reducing the growth of money supply M2 (from
78.7% in 1991 to 33.7% in 1992).
- Applying a new positive interest rate policy with the following characteristics: (1) Regulations on the minimum
deposit rate and the maximum lending rate were applied for commercial banks; (2) Average lending rate was bigger than
average deposit rate to cease States finance via credit channel; (3) The State Bank of Vietnam regulated the lending rate
cap in foreign currency for banks.
- The Governor regulated the 10% required reserve ratio for deposits at credit institutions. Terminating financing for
state budget via issuing money. First foreign banks were licensed to open their branches.

19931994
1995

- Two centres of foreign exchange trading in Hanoi and Ho Chi Minh city were set up. Announced official exchange
rate was based on trading results at the centres. In 1994, the required reserve ratio was regulated for different deposits.
- Implementing treasury bond bids and these bonds were not included in the State Bank of Vietnams required reserve
structure from 1995.

1996-1999
1996

- Terminating the minimum deposit rate, maintaining the lending rate cap in the band 0.35%/month.
- The National Assembly terminated the tax on banks turnover.
- The fixed exchange rate caused the value of Vietnams currency to increase relative to other currencies in Asia.

1997-

- The Law on the State Bank of Vietnam and the Law on credit institutions were newly introduced in 1997 and 1998,

1998

respectively. In 1997, the State Bank of Vietnam adjusted four times the bilateral exchange rate of VND and USD with the
higher announced official rate and larger band. In 1998, the lending rate cap for rural area was terminated.
184

Period

Year

Overview

1999

- Reforming the exchange rate management from administrative management in which to market rules-based
management. Announced official rate was replaced by inter-bank market foreign exchange rate.
- The State Bank of Vietnam began to conduct an expansionary monetary policy.
- Overseas currency exchange was not imposed tax, so 1 billion of USD was transferred in.
- The monetary policy in this period continued to support stimulating consumption demand. The State Bank of

2000-2003

Vietnams measures included reducing interest rate, required reverse ratio and refinance ratio. However, many adjustment
times of the State Bank of Vietnam resulted in banks passive business.
2000-

- The stock market was formed in July 2000 to create another channel to accumulate capital.

2001

- In 2000, the State Bank of Vietnam began to manage interest rate under a new mechanism base interest rate in which
the domestic currency lending rates adjusted by banks based a base rate announced by the State Bank of Vietnam plus a
correlative ratio for short-term loans or medium and long-term loans.
- In 2000-2001, the growth of credit was lower than the deposit growth and foreign currency loans growth was lower
than domestic currency loans growth. From 2001, the State Bank of Vietnam began to increase trading times in the open
market. (from one time in ten days in 2000 to 1 time/week in 2001)

2002

- Vietnam applied the negotiable interest rate mechanism in VND commercial lending. The Interbank Electronic
Payment System came to operation. Trading times in the open market increased to 2 times/week.
- The State Bank of Vietnam increased the bands of the exchange rate from 0.1 to 0.25%. Moreover, the State Bank of
Vietnam actively priced VND lower to stimulate exports. The growth of credit was higher than the deposit growth due to the
effects of expansionary monetary policy.

185

Period

Year

Overview

2003

- Revising Law on the State Bank of Vietnam where the State Bank of Vietnam would trade in the short term for
treasury bonds, the State Bank of Vietnams bills and other valuable papers with credit institutions to conduct the monetary
policy. Foreign exchange surrender ratio was 0% in 2003.
- The monetary policy in this period was cautiously conducted due to the fluctuations of the world economy and

2004-2009

domestic economy.
2004

- Law on credit institutions were revised in 2004.


- The State Bank of Vietnam adjusted a twofold increase in the reserve ratio in July 2004. Also, interest rates were
increased to conduct contractionary monetary policy.

2005

- The Ordinance of Foreign Exchange was approved in 2005. Vietnams current account transactions were liberalized.
Trading times in the open market increased to 3 times/week.
- Monetary policy was adjusted in flexibility to gain the objectives of economic growth and price control. Key interest
rates are increased twice. The total of payment and credit were lower than that of 2004.

2006

- Banks enhanced precautions for risks in banking activities. Credit for the economy was decreased.
- On December 24th, the State Bank of Vietnam increased the band between banks trading ratio and the State Bank of
Vietnams announced ratio from 0.5% to 0.75%.

2007

- The State Bank of Vietnam adjusted a twofold increase in the reserve ratio in June 2007.
- The open market operations were daily traded.
- The State Bank of Vietnam bought a large amount of USD for the national reserve. Money supply and credit increased
to 46% and 54%, respectively. The State Bank of Vietnam regulated the ratio of securities lending per total loans at 3%.
186

Period

Year

Overview

2008

- The State Bank of Vietnam conducted monetary policy with different period: tightening before November, expansion

2010-2011

from November. Many adjustments for base rate (8.25% -> 8.75% -> 12% -> 14% -> 13% -> 12% -> 11% -> 10% ->
8.5%), decreasing the required reserve ratio, adjusting average exchange rate 2.35% and expanding the band of the
exchange rate between USD and VND (3%), expanding types of deposits for reserve (over 24 months deposits), tightening
regulations on loans for securities investment, narrowing consumers for foreign exchange loans. Banks had to buy 20300
billion of VND in the State Bank of Vietnams bills.
- Difficulties in liquidity of commercial banks resulted in racing about interest rate (the highest rate is 21%/year). From
2008, interest rate cap mechanism was applied due to the fluctuation in the economy and money market.
2009

- The State Bank of Vietnam decreased the base interest rate from 14% to 7% and required reserve ratio from 11% to
5%. From November 2009, the base interest rate increased 1%.
- The band for exchange rate was adjusted from 3% to 5%, but from November, it was decreased to 3% while the
announced rate was increased 5% compared to the previous rate.

2010

- In 2010, the new laws on the State Bank of Vietnam and credit institutions were approved. 100% foreign owned banks
were allowed to be established in Vietnam. Also, the Securities Law was revised.
- The State Bank of Vietnam remained key interest rates in the first ten months, but from 5/11, these rates were
increased 1%. The negotiable interest rate mechanism was applied from April.
- The State Bank of Vietnam strictly controlled SOEs exchange rate trading and gold market. There were many
unexpected fluctuations in gold market.

187

Period

Year

Overview
- Required reserve ratio was decreasingly adjusted.

2011

- The State Bank of Vietnam conducted tightening monetary policy. Credit for production sectors increased 18% while
credit for non-production sectors decreased 20% in which credit for securities investment and business fell 43%.
- Interest rate cap was regulated at 14% to reduce the lending rate.
- The State Bank of Vietnam began to restructure the banking system with the first merging of three joint-stock
commercial banks.

Source: Pham and Vuong (2009), and authors summary.

188

APENDIX B - APPENDIX TO CHAPTER 4


Figure B1: Seasonal adjusted data for additive model (Data1_SA) and
multiplicative model (Data_SA) in X12-ARIMA of Y, PI and PC
180,000
160,000
140,000
120,000
100,000
80,000
60,000
40,000
00

01

02

03

04
Y

05

06

07

Y1_SA

08

09

10

11

Y_SA

180,000
160,000
140,000
120,000
100,000
80,000
60,000
40,000
20,000
0
00

01

02

03

04
PI

05

06

PI1_SA

07

08

09

10

11

PI_SA

400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000
00

01

Source: Authors calculation.

02

03

04
PC

05

06

PC1_SA

07

08
PC_SA

09

10

11

Table B1: Quality measures and criterion

Quality measures

Criterion

M2

It should ideally be close to zero.

M7

It is close to 0, the seasonal pattern becomes more stable.

M10

Fluctuation is too large, and seasonal adjustment no longer


stable, if M10 >1.

M11
Q value

Fluctuation is not random if M11 >1.


The closer Q is to 0, the higher the quality of seasonal
adjustment becomes.

Source: Norway (2008).

The quality measures selected from the 11 M-measures are M2, M7, M10 and
M11, and are produced automatically by X-12-ARIMA. Q is a weighted average of
all 11 M-measures.
Quality measures
The relative

Description
M2 measures whether the amount of random

contribution of the

variation in the data is small enough for an

irregular component to the

estimation of trend and seasonal components. The

variance of the stationary

values can vary from 0.0 to 3.0, and should ideally

portion of the series (M2)

be close to zero.

The amount of stable


seasonality present relative
to the amount of moving
seasonality (M7)

The formula for M7 is as follows:

M7

1 7 3FM
(

)
2 FM
FS

FS: the relative contribution of the stable part


of the season.
FM: the contribution of the moving part of the
season.
The M7 value is a common quantity in the
evaluation of the setup and routines in use. The value
of M7 for series with a stable seasonal pattern is
190

Quality measures

Description
usually well below 1 and the closer this value is to
0, the more stable the seasonal pattern becomes. The
formula for estimating this quantity is rather
complicated and difficult to interpret intuitively. The
M7 test is often more robust than the total quality
measure (Q, see below) as well as the other M tests
produced by X-12-ARIMA.

The size of seasonal

M10 measures the amount of fluctuation in the

component fluctuations in seasonal component in recent years. Fluctuation is


too large, and seasonal adjustment no longer stable,

recent years (M10)

if M10 >1.
The

size

of

linear

M11 measures the degree of linear movement

movement in the seasonal in the seasonal

component in recent

years.

component in recent years Fluctuation is not random if M11 >1.


(M11)
Q

value

(collective

The Q value is a weighted average of the

measure of quality in X-12- eleven M tests in X-12-ARIMA. The weights reflect


ARIMA Q)

the importance assigned to the various tests by the


developers of X-12-ARIMA.
The closer Q is to 0, the higher the quality of
seasonal adjustment becomes. The M-measures
should be reassessed when Q is greater than 1, in
order to determine whether the variation in the
irregular or the seasonal component is too large.

191

Table B2: Quality indexes under both models (additive and multiplicative)

Variables Indexes
Y

PI

X12 for multiplicative model

X12 for additive model

M2

0.003

0.008

M7

0.069

0.222

M10

0.387

0.715

M11

0.387

0.715

0.15 ACCEPTED

0.24 ACCEPTED

M2

0.160

0.183

M7

0.497

0.892

M10

2.048

2.920

M11

1.844

2.576

0.69 ACCEPTED

0.99 CONDITIONALLY
ACCEPTED

PC

M2

0.041

0.038

M7

0.278

0.574

M10

0.517

1.627

M11

0.443

1.621

0.29 ACCEPTED

0.59 ACCEPTED

Source: Authors calculation.

192

Table B3: X12-ARIMA and T/S used in countries

Method

Countries

X12-

Australia, Canada, New Zealand, Israel, Japan, Korea, Singapore, USA,

ARIMA
T/S

Ireland, Netherlands, Norway, Switzerland, UK.


Bulgaria, Greece, Hungary, Italy, Latvia, Luxembourg, Malta, Poland,
Romania, Slovak Rep., Slovenia, Spain, Turkey.

Both

Austria, Croatia, Cyprus, Denmark, Finland, France, Iceland, Lithuania,


Portugal, Sweden.

Source: Eo (2010)

193

Figure B2: Gross domestic product (GDP), Private investment (PI), and
Private consumption (PC): seasonally adjusted value with X12 and T/S
160,000

140,000

120,000

100,000

80,000

60,000
00

01

02

03

04

05

06

07

08

09

10

11

09

10

11

Final seasonally adjusted series


Y_SA

200,000

160,000

120,000

80,000

40,000

0
00

01

02

03

04

05

06

07

08

PI_SA
Final seasonally adjusted series
450,000
400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000
00

01

02

03

04

05

06

07

08

PC_SA
Final seasonally adjusted series

Source: Authors calculation.

194

09

10

11

Figure B3: Irregular factors of investment (PI) and consumption (PC)


estimated by X12-ARIMA and T/S.

Final irregular component/factor

PI_IR
1.3

110

1.2

100

1.1

90

1.0

80

0.9

70

0.8

60

0.7

50
40

0.6
00 01 02 03 04 05 06 07 08 09 10 11

00 01 02 03 04 05 06 07 08 09 10 11

PC_IR

Final irregular component/factor


125

1.10

120

1.05

115
1.00
110
0.95

105

0.90

100

0.85

95
00 01 02 03 04 05 06 07 08 09 10 11

00 01 02 03 04 05 06 07 08 09 10 11

Source: Authors calculation.

195

Table B4: Moving holidays in Vietnam

Event

Lunar year

Solar year

Lunar New Year

The last day in the 2000: 4-7/2

2006: 28-31/1

as known Tet

old year

2007: 16-19/2

holiday or

1-3/1 in the new 2002: 11-14/2

2008: 6-9/2

Vietnamese New

year

2003: 31/1 3/2

2009: 25-28/1

2004: 21-24/1

2010: 13-16/2

2005: 8-11/2

2011: 2-5/2

2000: 14/4

2006: 7/4

birthday

2001: 3/4

2007: 26/4

anniversary (1

2002: 22/4

2008: 15/4

2003: 11/4

2009: 4/4

2004: 28/4

2010: 23/4

2005: 18/4

2011: 12/4

Year (4 days)

King Hungs

10/3

days)

2001: 23-26//1

Source: Authors summary.

196

Figure B4:

Gross domestic product (Y), private investment (PI) and

private consumption (PC): original data, seasonally adjusted data and seasonal
factor from X12-ARIMA.
Y

Y_SA

200,000

160,000
140,000

160,000

120,000
120,000
100,000
80,000

80,000

40,000

60,000
00

01

02

03

04

05

06

07

08

09

10

11

00

01

02

03

04

05

06

07

08

09

10

11

07

08

09

10

11

Y_SF
1.2
1.1
1.0
0.9
0.8
0.7
00

01

02

03

04

05

06

07

08

09

10

11

PI

PI_SA

200,000

160,000

160,000

120,000

120,000
80,000
80,000
40,000

40,000
0

0
00

01

02

03

04

05

06

07

08

09

10

11

00

PI_SF
1.4

1.2

1.0

0.8

0.6
00

01

02

03

04

05

06

07

08

09

10

11

197

01

02

03

04

05

06

PC

PC_SA

400,000

350,000
300,000

300,000
250,000
200,000

200,000
150,000

100,000
100,000
0

50,000
00

01

02

03

04

05

06

07

08

09

10

11

07

08

09

10

11

00

PC_SF
1.2

1.1

1.0

0.9

0.8
00

01

02

03

04

05

06

Source: Authors calculation.

198

01

02

03

04

05

06

07

08

09

10

11

Table B5: Descriptive statistics for the foreign sector (2000:1-2011:4)

FFR

FY

WRP

WP

WGP

Mean

2.485563

4494729.

358.2527

56.23094

668.0225

Median

1.845000

4182370.

293.3400

56.31700

519.7700

Maximum

6.520000

7568636.

953.0400

121.1130

1700.120

Minimum

0.073000

2405840.

164.7100

19.31300

263.4600

Std. Dev.

2.122980

1591584.

184.5224

28.70970

406.8972

Skewness

0.465871

0.389996

1.046785

0.569095

0.997022

Kurtosis

1.804010

1.856272

3.543602

2.251820

2.998358

Jarque-Bera

4.597067

3.833004

9.357072

3.710501

7.952433

Probability

0.100406

0.147121

0.009293

0.156414

0.018756

Sum

119.3070

2.16E+08

17196.13

2699.085

32065.08

Sum Sq. Dev.

211.8311

1.19E+14

1600281.

38739.60

7781570.

Observations

48

48

48

48

48

Source: Authors calculation.

199

Table B6: Descriptive statistics for the domestic sector (2000:1-2011:4)


Y

CPI

PI

PC

CR

VNI

VE

VI

Mean

103818.4

117.203

60996.02

161102.1

289654.5

11.463

966958.5

101.536

398.978

10373.05

12254.10

Median

101411.5

103.975

48605.74

131397.5

230931.7

10.850

602447.8

100.925

364.250

8941.350

9587.000

Max

151017.0

207.510

154414.0

342820.7

634314.5

20.100

3063869.

120.460

1035.000

26514.00

28165.00

Min

66607.41

79.420

22087.91

70217.14

72364.15

8.520

119729.9

88.220

113.500

3111.000

3358.000

Std. Dev.

25259.96

38.125

35631.38

83373.39

183501.6

2.658

897646.0

7.695

237.793

6376.132

7458.181

Skewness

0.185

0.849

0.860

0.721970

0.551

1.577

1.018778

0.414

1.338

0.796

0.546

Kurtosis

1.762

2.519

2.717

2.132137

1.861

4.868

2.720

2.743

4.357

2.775

2.012

Jarque-

3.340

6.235

5.580

5.203272

5.027

26.880

8.459

1.505

17.267

5.173

4.338

Prob

0.188

0.044

0.061

0.074152

0.080

0.000

0.014

0.471

0.000

0.075

0.114

Sum

4983285.

5625.760

2683825.

7088493.

13903418

550.270

46414006

4873.740

18353.00

497906.2

588197.0

Sum Sq.

3.00E+1

68317.48

5.46E+10

2.99E+11

1.58E+12

332.070

3.79E+13

2783.572

2544557.

1.91E+09

2.61E+09

48

44

44

48

48

48

48

46

48

48

Bera

Dev.

Obs

48

Source: Authors calculation.

Table B7: Summary of results of univariate unit root tests with break(s)
using Nelson-Plosser data set.

Number of rejections of unit


Authors

Break(s)

root (with possible breaks)


(rejection frequency)
1%

5%

10%

Total

ADF with no break (0)

Perron (1989)

1 exogenous break

10

Zivot and Andrews (1992)

1 endogenous break

Sen 2003 (2003)

1 endogenous break

Popp (2008)

1 endogenous break

Lumsdaine and Papell

2 endogenous breaks

Lee and Strazicich (2003)

2 endogenous breaks

10

Narayan and Popp (2010)

2 endogenous breaks

Nelson and Plosser (1982)

(1997)

Source: Narayan and Popp (2010)

Table B8: Unit root tests without structural break*


Variable: LnY
Test

Ta=1

Decision

ADF

-1.341

NS

ERS

4.816

Ng-Perron

2.757

KPSS

0.159

NS

ADF

-1.619

NS

ERS

94.433

NS

Ng-Perron

26.364

NS

KPSS

0.227

NS

ADF

-6.016

ERS

6.397

NS

Ng-Perron

4.935

KPSS

0.075

ADF

-2.836

NS

ERS

11.763

NS

Ng-Perron

10.033

NS

KPSS

0.195

NS

ADF

-4.366

ERS

2.122

Ng-Perron

2.391

KPSS

0.099

ADF

-5.159

ERS

2.313

Ng-Perron

2.137

KPSS

0.099

Variable: LnCPI

Variable: LnPI

Variable: LnPC

Variable: LnM

Variable: R

202

Variable: LnCR
ADF

-2.944

NS

ERS

5.499

Ng-Perron

5.635

NS

KPSS

0.147

NS

ADF

-1.520

NS

ERS

19.861

NS

Ng-Perron

18.347

NS

KPSS

0.177

NS

ADF

-2.431

NS

ERS

8.140

NS

Ng-Perron

7.350

NS

KPSS

0.103

ADF

-2.590

NS

ERS

9.979

NS

Ng-Perron

9.532

NS

KPSS

0.083

ADF

-3.304

NS

ERS

4.188

Ng-Perron

4.279

KPSS

0.105

ADF

-3.717

ERS

2.841

Ng-Perron

2.580

KPSS

0.100

ADF

-3.444

NS

ERS

4.282

Ng-Perron

4.179

KPSS

0.105

Variable: LnE

Variable: LnVNI

Variable: LnVE

Variable: LnVI

Variable: FFR

Variable: LnWP

203

Variable: LnWRP
ADF

-2.800

NS

ERS

16.808

NS

Ng-Perron

13.235

NS

KPSS

0.099

ADF

-3.200

NS

ERS

34.285

NS

Ng-Perron

20.235

NS

KPSS

0.218

NS

ADF

-2.064

NS

ERS

45.574

NS

Ng-Perron

36.244

NS

KPSS

0.123

Variable: LnWGP

Variable: LnFY

ADF Test critical values at 5 percent level are -3.51; ERS Test critical values: 5.72;
Ng-Perron Test critical values: 5.48; KPSS Test critical values: 0.146. The null
hypothesis for ADF, ERS and Ng-Perron tests is of a unit root (nonstationarity), for
KPSS is stationarity.
* implies that the result is significant at 5 percent and testing includes a constant and
a trend.
Source: Authors calculation.

204

Table B9: Results from LS tests with one and two structural breaks
(seasonally adjusted and log form)*
Ho Unit root

Unit root

Ha Levels Stationary, Crash Model A

Trend Stationary, Break Model C

t-statistic Break 1

Break 2

Decision t-statistic

NC

NS

Break 1

Break 2

Decision

-2.904

2007:4

NC

NS

LnY
LS1

-3.765

2010:4**

LS2

-4.282

2009:4 2010:4**

-5.784

2004:3 2008:2**

LS1

-3.188

2007:4

NC

NS

-4.399

2007:1

NC

NS

LS2

-3.720

2006:3

2007:4

NS

-6.256

2007:1

2008:3

LS1

-5.383

2007:4

NC

-3.891

2007:3

NC

NS

LS2

-5.939

2007:1

2007:4

-7.755

2007:2

2009:4

LS1

-0.906

2007:4

NC

NS

-4.634

2004:1

NC

NS

LS2

-1.546

2004: 1

2008:2

NS

-5.646

2002:2

2005:3

LS1

-5.029

2009:4

NC

-6.343

2006:3

NC

LS2

-5.947

2006:4

2009:4

-7.257

2006:3

2010:1

LS1

-5.297

2009:4

NC

-5.475

2008:4

NC

LS2

-5.440

2004:4

2009:4

-6.649

2007:3

2009:2

LS1

-2.368

2009:4

NC

NS

-3.797

2008:2

NC

NS

LS2

-2.784

2008:3

2010:1

NS

-5.025

2002:2

2008:4

NS

LS1

-2.746

2004:3

NC

NS

-3.738

2003:1

NC

NS

LS2

-2.939

2004:3

2007:3

NS

-4.629

2004:3

2010:1

NS

LS1

-3.031

2003:4

NC

NS

-4.400

2006:2

NC

NS

LS2

-3.255

2003:4

2008:2

NS

-5.567

2002:4

2006:2

LnPC

LnPI

LnCPI

LnM

LnCR

LnE

LnVNI

205

Ho Unit root

Unit root

Ha Levels Stationary, Crash Model A

Trend Stationary, Break Model C

t-statistic Break 1

Break 2

Decision t-statistic

Break 1

Break 2

Decision

LnVE
LS1

-3.926

2005:2

NC

-3.947

2008:2

NC

NS

LS2

-4.354

2003:4

2005:2

-5.779

2003:3

2009:1

LS1

-3.138

2009:4

NC

NS

-4.001

2007:3

NC

NS

LS2

-3.916

2006:4

2008:2

-6.178

2007:3

2009:1

LS1

-3.337

2008:4

NC

NS

-3.341

2008:4

NC

NS

LS2

-3.571

2007:3

2008:4

NS

-4.379

2004:2

2009:3

NS

LS1

-1.874

2004:3

NC

NS

-3.025

2006:2

NC

NS

LS2

-2.345

2004:3

2009:1

NS

-5.216

2003:3

2006:4

NS

LS1

-3.299

2005:2

NC

NS

-4.161

2005:2

NC

NS

LS2

-3.523

2003:3

2005:2

NS

-4.728

2003:2

2006:1

NS

LS1

-2.114

2007:4

NC

NS

-2.831

2003:4

NC

NS

LS2

-2.179

2004:1

2007:3

NS

-3.751

2002:2

2008:1

NS

LS1

-1.239

2007:4

NC

NS

-2.907

2004:2

NC

NS

LS2

-1.733

2008:4

2009:4

NS

-4.388

2002:3

2004:3

NS

LnVI

FFR

LnFY

LnWP

LnWRP

LnWGP

LS1, LS2 are LS tests for one break and two breaks, respectively.
S = Stationary, NS = Nonstationary, NC = Not calculated.
The LM test critical values are -3.937 for Model A, and -5.620 for Model C at the
5% significance level.
* The statistical significance at the 5% level.
** The statistically significant break dates.
Source: Authors calculation.

206

Table B10: Summary of results from unit root tests


ADF

ERS

NgP

KPSS

LS1-A

LS1-C

LS2-A

LS2-C

LnY

NS

NS

NS

NS

LnCPI

NS

NS

NS

NS

NS

NS

NS

LnPI

NS

NS

LnPC

NS

NS

NS

NS

NS

NS

NS

LnM

LnCR

NS

NS

NS

NS

NS

NS

NS

LnE

NS

NS

NS

NS

NS

NS

NS

NS

LnVNI

NS

NS

NS

NS

NS

NS

LnVE

NS

NS

NS

NS

LnVI

NS

NS

NS

FFR

NS

NS

NS

NS

LnWP

NS

NS

NS

NS

NS

LnWRP

NS

NS

NS

NS

NS

NS

NS

LnWGP

NS

NS

NS

NS

NS

NS

NS

NS

LnFY

NS

NS

NS

NS

NS

NS

NS

4S-12NS

7S-9NS

7S-9NS

10S-6NS

4S-12NS

2S-14NS

6S-10NS

9S-7NS

Source: Authors calculation.

APENDIX C APPENDIX TO CHAPTER 6


Table C1: VAR Stability tests with two lags
Roots of Characteristic Polynomial
Endogenous variables: WOP WRP WGP YC FFR CPI Y R M CR VNI E
Exogenous variables : C DUM, Lag specification : 1 2
Root

Modulus

0.996938

0.996938

0.340928 + 0.841930i

0.908338

0.340928 0.841930i

0.908338

0.667592 0.598473i

0.896576

0.667592 + 0.598473i

0.896576

0.830245

0.830245

0.735683 0.378523i

0.827351

0.735683 + 0.378523i

0.827351

0.184186 0.780926i

0.802352

0.184186 + 0.780926i

0.802352

-0.671589 0.403182i

0.783319

-0.671589 + 0.403182i

0.783319

-0.329629 0.704672i

0.777958

-0.329629 + 0.704672i

0.777958

0.483416 0.406506i

0.631615

0.483416 + 0.406506i

0.631615

-0.604856

0.604856

-0.362796 0.413334i

0.549968

-0.362796 + 0.413334i

0.549968

-0.225984 + 0.398052i

0.457727

-0.225984 0.398052i

0.457727

0.281636 + 0.292936i

0.406363

0.281636 0.292936i

0.406363

0.143891

0.143891

No root lies outside the unit circle.


VAR satisfies the stability condition.
Source: Authors calculation.

208

VAR Residual Serial Correlation LM Test


Null Hypothesis: no serial correlation at lag order h
Sample: 2000Q1 2011Q4
Included observations: 44
Lags

LM-Stat

Prob

1224.072

0.0000

NA

NA

Probs from chi-square with 144 df.

Source: Authors calculation.

209

Table C2: Variance Decomposition of Domestic Variables

Model VN1
Innovations Quarters

WP

WRP

WGP

FY

FFR

Forecast error variance for variable


Y

CPI

CR

VNI

0.02

0.02

0.00

0.51

0.02

10.84

0.07

0.66

8.72

8.89

1.68

1.14

13.42

2.71

10

1.61

12.45

9.53

2.52

3.05

11.40

3.51

20

1.82

10.14

9.43

1.86

3.13

11.31

3.53

3.62

42.02

2.82

8.18

2.02

16.17

7.86

2.45

19.95

23.67

11.81

13.77

10.38

7.53

10

2.86

11.22

21.52

11.81

13.70

9.75

8.56

20

2.57

9.86

21.37

8.58

13.66

9.66

8.57

0.32

0.30

0.00

2.42

0.28

0.56

0.08

2.45

0.82

8.44

12.34

5.94

1.44

10.73

10

2.86

1.44

8.15

8.41

5.60

1.61

9.97

20

2.71

1.99

8.09

6.55

5.61

1.60

9.95

2.28

2.12

0.79

0.17

22.27

10.92

1.44

11.47

6.38

4.36

8.26

19.99

8.68

3.51

10

15.46

5.14

3.93

9.48

18.40

9.83

3.47

20

17.18

4.73

4.02

12.01

18.33

9.91

3.52

5.22

4.85

0.34

13.09

4.66

0.96

0.49

1.97

1.68

1.01

4.13

3.29

5.42

1.74

10

0.76

2.59

1.65

3.08

3.25

6.05

1.95

20

0.40

2.21

1.64

1.98

3.24

6.03

1.96

23.94

22.25

5.29

27.93

53.73

5.01

0.45

23.25

4.80

9.42

7.66

34.71

6.50

4.01

10

17.95

8.02

10.35

10.44

32.85

5.43

5.48

20

16.40

12.06

10.48

12.99

32.73

5.51

5.49

210

Innovations Quarters

CPI

CR

VNI

Forecast error variance for variable


Y

CPI

CR

VNI

45.57

10.76

1.05

0.38

0.00

2.17

8.36

39.69

1.73

0.57

14.21

0.53

7.83

6.74

10

37.43

6.55

1.53

21.10

0.78

8.48

6.33

20

36.32

17.73

1.87

28.34

0.79

8.69

6.32

3.19

6.21

0.05

29.66

5.97

1.47

27.72

1.25

28.78

11.31

14.57

4.93

5.92

18.16

10

1.08

32.66

14.44

12.49

5.84

8.63

18.34

20

1.05

25.52

14.35

9.01

5.91

8.63

18.37

3.19

2.97

0.03

14.17

2.85

0.70

42.27

1.25

7.27

1.98

4.14

3.16

0.78

28.38

10

1.08

6.03

2.08

2.91

3.07

0.85

25.65

20

1.05

4.70

2.07

2.03

3.08

0.84

25.54

3.44

3.20

89.59

1.04

3.08

2.19

10.56

2.44

16.81

29.17

17.62

7.21

3.30

14.33

10

3.71

10.75

25.38

12.48

7.42

3.78

14.43

20

3.97

7.73

25.16

9.33

7.45

3.78

14.40

0.60

0.56

00.0

0.26

0.54

39.36

0.08

2.35

0.66

0.72

0.09

2.56

30.92

1.69

10

1.98

1.44

0.84

0.37

3.48

28.66

1.74

20

1.70

1.13

0.84

0.49

3.52

28.46

1.74

5.11

4.75

0.04

2.19

4.57

9.63

0.64

8.32

2.40

0.46

3.50

2.77

5.42

0.48

10

9.84

1.71

0.59

4.93

2.56

5.54

0.57

20

10.55

2.19

0.68

6.84

2.55

5.58

0.60

Source: Authors calculation.


211

Model VN2

Innovations

WP

WRP

WGP

FY

FFR

Quarters

Forecast error variance decomposition for variables


Y

CPI

CR

VNI

2.98

1.13

0.14

1.18

0.00

13.77

0.27

1.41

10.82

9.38

0.51

1.23

17.87

3.80

10

0.58

17.22

9.85

1.94

3.32

15.60

4.47

20

0.32

16.67

9.77

1.66

3.42

15.53

4.48

2.59

42.42

2.98

7.70

1.86

15.53

7.34

2.47

20.31

23.55

12.13

13.86

10.30

7.23

10

2.18

11.17

21.43

11.65

13.76

9.79

8.27

20

1.93

9.47

21.28

8.22

13.72

9.71

8.28

0.51

0.31

0.01

2.53

0.25

0.58

0.10

2.25

0.78

8.34

12.19

6.01

1.45

10.79

10

2.57

1.34

8.07

8.28

5.66

1.64

10.03

20

2.40

1.76

8.01

6.31

5.67

1.63

10.01

2.76

1.66

1.08

0.24

22.37

11.00

1.30

12.79

5.60

4.15

8.79

19.93

8.40

3.33

10

17.25

4.38

3.78

10.37

18.31

9.51

3.30

20

19.13

4.39

3.88

13.23

18.23

9.58

3.35

7.21

4.33

0.42

11.54

3.52

1.26

0.77

2.60

1.22

1.25

3.60

2.67

5.88

2.08

10

1.12

1.70

1.71

2.78

2.66

6.45

2.20

20

0.71

1.30

1.70

1.82

2.66

6.44

2.21

26.92

16.16

3.4

31.57

57.89

5.93

0.88

27.23

5.23

10.85

8.71

37.26

6.72

4.79

10

21.27

10.60

11.77

12.53

35.28

5.63

6.36

20

19.38

15.77

11.93

15.62

35.15

5.71

6.37

212

Innovations

CPI

CR

VNI

Quarters

Forecast error variance decomposition for variables


Y

CPI

CR

VNI

28.71

16.99

2.40

0.03

0.27

1.39

7.08

25.62

3.25

0.72

9.62

0.40

4.46

5.26

10

24.53

6.59

1.23

14.79

0.55

4.94

4.87

20

23.75

14.16

1.45

19.39

0.57

5.07

4.86

10.31

6.19

0.19

28.38

5.04

0.98

24.62

5.35

27.31

10.56

15.50

4.34

6.42

16.19

10

7.40

29.48

13.81

13.53

5.26

9.16

16.55

20

8.53

22.19

13.74

10.56

5.33

9.18

16.60

4.36

2.61

0.08

11.99

2.13

0.41

45.71

2.03

5.87

1.46

3.99

2.69

0.76

29.99

10

1.81

4.45

1.63

2.79

2.60

0.82

27.06

20

1.77

3.29

1.62

2.17

2.60

0.82

26.95

5.96

3.58

89.16

1.19

2.91

1.75

11.44

4.55

17.47

28.91

20.68

7.21

3.78

14.64

10

6.36

10.88

25.61

14.73

7.48

4.37

14.85

20

6.74

7.84

25.40

11.72

7.51

4.38

14.83

0.82

0.49

0.02

0.39

0.40

36.90

0.05

2.80

0.50

0.60

0.13

2.37

28.17

1.55

10

2.41

1.12

0.72

0.48

3.27

26.24

1.60

20

2.08

0.84

0.72

0.65

3.30

26.05

1.60

6.88

4.13

0.13

3.26

3.36

10.50

0.44

10.89

1.61

0.23

4.13

2.03

5.80

0.35

10

12.52

1.08

0.38

6.14

1.87

5.84

0.44

20

13.26

2.33

0.50

8.63

1.86

5.89

0.47

Source: Authors calculation.

213

Table C3: VAR Stability tests with the international channel

Roots of Characteristic Polynomial


Endogenous variables: DLNWOP DLNWRP DLNWGP DLNYC DDFFR LNCPI_SA
LNY_SA LNVE_SA LNVI_SA I_SA LNM_SA DLNCR LNVNI_SA DLNREER
Exogenous variables: C DUM
Lag specification: 1 1
Root
Modulus
0.979842 - 0.016408i
0.979980
0.979842 + 0.016408i
0.979980
0.598390 - 0.553638i
0.815221
0.598390 + 0.553638i
0.815221
0.717616 - 0.118563i
0.727345
0.717616 + 0.118563i
0.727345
-0.400901 - 0.544661i
0.676297
-0.400901 + 0.544661i
0.676297
0.002612 - 0.610552i
0.610558
0.002612 + 0.610552i
0.610558
0.096172 - 0.391119i
0.402769
0.096172 + 0.391119i
0.402769
-0.388714
0.388714
0.095831
0.095831
No root lies outside the unit circle.
VAR satisfies the stability condition.
Source: Authors calculation.

214

Table C4: Variance Decomposition of Exports and Imports


Variance Decomposition of Exports
Shock1

Shock2

Shock3

Shock4

Shock5

Shock6

Shock7

Shock8

Shock9

Shock12

Shock13

Shock14

0.130

6.356

0.002

0.001

0.0045

5.351

0.036

88.095

0.003

0.002

0.000

0.008

0.001

0.005

0.920

10.797

0.043

1.179

8.439

5.869

0.792

69.731

0.008

0.333

1.347

0.149

0.341

0.045

2.348

9.318

0.180

1.087

6.943

5.968

5.494

59.290

3.3768

1.306

1.601

2.727

0.281

0.076

3.740

8.761

0.744

1.296

5.926

5.344

7.309

55.728

2.920

3.090

2.087

2.649

0.330

0.068

3.526

7.814

1.209

1.294

5.568

5.160

9.747

52.791

2.795

3.874

2.990

2.662

0.500

0.063

3.176

7.495

2.117

1.182

5.013

5.286

11.506

50.283

2.999

3.660

2.945

3.466

0.695

0.168

2.858

6.659

2.792

1.339

4.510

5.787

11.639

47.186

3.704

3.302

2.632

5.985

1.042

0.558

2.951

6.022

2.675

1.883

4.167

6.634

11.523

43.971

3.632

3.349

2.403

8.326

1.385

1.072

3.206

5.652

2.476

2.542

3.865

7.364

11.251

41.599

3.394

3.414

2.350

9.581

1.712

1.588

10

3.394

5.538

2.332

3.145

3.650

7.854

10.996

39.970

3.206

3.324

2.287

10.228

2.047

2.020

11

3.521

5.457

2.231

3.628

3.489

8.195

10.989

38.836

3.064

3.192

2.213

10.533

2.309

2.334

12

3.570

5.319

2.173

3.970

3.351

8.433

11.112

38.021

2.951

3.066

2.134

10.802

2.509

2.583

13

3.581

5.128

2.154

4.238

3.213

8.617

11.245

37.275

2.880

2.939

2.047

11.205

2.669

2.803

14

3.594

4.907

2.147

4.493

3.073

8.791

11.351

36.478

2.838

2.820

1.959

11.733

2.795

3.014

15

3.619

4.682

2.126

4.758

2.936

8.971

11.368

35.637

2.812

2.731

1.879

12.339

2.903

3.231

16

3.668

4.477

2.077

5.042

2.810

9.157

11.316

34.790

2.774

2.676

1.814

12.934

3.005

3.452

17

3.734

4.304

2.015

5.331

2.698

9.338

11.245

34.008

2.716

2.637

1.766

13.431

3.103

3.665

18

3.799

4.161

1.955

5.606

2.599

9.497

11.183

33.331

2.653

2.597

1.727

13.826

3.199

3.862

19

3.854

4.037

1.901

5.853

2.511

9.631

11.146

32.753

2.593

2.549

1.688

14.150

3.289

4.038

20

3.898

3.921

1.857

6.069

2.431

9.740

11.132

32.250

2.542

2.500

1.650

14.438

3.370

4.194

215

Shock10 Shock11

Variance Decomposition of Imports


Shock1

Shock2

Shock3

Shock4

Shock5

Shock6

Shock7

Shock8

Shock9

Shock10

Shock11

Shock12

Shock13

Shock14

0.244

6.938

0.059

0.037

0.128

9.180

1.024

0.045

81.841

0.065

0.002

0.239

0.027

0.166

4.562

10.424

1.395

0.355

0.309

7.040

3.203

2.852

65.246

2.909

0.373

1.075

0.097

0.153

4.085

8.448

1.217

1.217

0.430

5.718

5.699

5.029

52.480

10.213

3.433

1.065

0.669

0.290

3.343

10.180

3.540

1.104

0.523

4.657

9.301

6.412

43.073

11.033

4.547

1.152

0.890

0.238

2.663

8.925

6.683

1.058

0.417

4.464

9.660

8.635

37.568

8.787

3.887

5.469

1.243

0.534

2.815

7.401

6.124

1.751

0.547

5.577

9.354

8.807

32.165

8.447

3.264

10.970

1.564

1.207

3.546

6.534

5.373

2.756

0.526

6.644

8.690

8.556

28.260

8.700

3.312

13.278

1.886

1.932

4.021

6.383

5.022

3.681

0.504

7.231

8.146

8.425

26.075

8.483

3.353

13.867

2.295

2.507

4.278

6.484

4.845

4.349

0.482

7.549

7.995

8.344

24.974

8.151

3.299

13.792

2.606

2.844

10

4.354

6.463

4.702

4.707

0.486

7.725

8.142

8.481

24.260

7.916

3.217

13.682

2.807

3.051

11

4.340

6.309

4.611

4.898

0.475

7.832

8.404

8.838

23.550

7.702

3.122

13.749

2.956

3.205

12

4.301

6.080

4.561

5.051

0.458

7.930

8.698

9.272

22.773

7.422

3.009

14.038

3.056

3.343

13

4.266

5.815

4.491

5.220

0.443

8.059

8.884

9.711

21.925

7.123

2.880

14.550

3.125

3.501

14

4.275

5.543

4.355

5.439

0.438

8.233

8.929

10.059

21.038

6.884

2.766

15.157

3.192

3.686

15

4.331

5.308

4.189

5.706

0.434

8.430

8.906

10.300

20.189

6.708

2.686

15.657

3.265

3.884

16

4.400

5.130

4.037

5.980

0.427

8.611

8.873

10.492

19.449

6.552

2.629

15.994

3.346

4.074

17

4.457

4.993

3.908

6.227

0.419

8.760

8.870

10.671

18.826

6.397

2.577

16.216

3.429

4.241

18

4.497

4.870

3.802

6.434

0.411

8.877

8.908

10.860

18.292

6.243

2.523

16.386

3.506

4.383

19

4.520

4.748

3.712

6.604

0.402

8.969

8.969

11.065

17.809

6.094

2.467

16.557

3.572

4.506

20

4.534

4.622

3.632

6.751

0.395

9.046

9.032

11.270

17.352

5.952

2.410

16.751

3.629

4.616

Source: Authors calculation.

216

Table C5: VAR Stability tests with the extended model with the investment
and consumption behaviour.

Roots of Characteristic Polynomial


Endogenous variables: DLNWOP DLNWRP DLNWGP DLNYC DDFFR
LNCPI_SA LNY_SA LNPI_SA LNPC_SA I_SA LNM_SA DLNCR LNVNI_SA
DLNREER
Exogenous variables: C DUM
Lag specification: 1 1
Root
0.996253
0.931987
0.654876 - 0.555196i
0.654876 + 0.555196i
0.683213 - 0.178714i
0.683213 + 0.178714i
0.068671 - 0.693952i
0.068671 + 0.693952i
-0.272462 - 0.405718i
-0.272462 + 0.405718i
-0.454772 - 0.165614i
-0.454772 + 0.165614i
0.054999 - 0.451126i
0.054999 + 0.451126i

Modulus
0.996253
0.931987
0.858548
0.858548
0.706200
0.706200
0.697342
0.697342
0.488716
0.488716
0.483989
0.483989
0.454466
0.454466

No root lies outside the unit circle.


VAR satisfies the stability condition.
Source: Authors calculation.

217

Table C6: VAR Stability tests with two sub-samples


Sub-sample 1 (2001Q1-2011Q4):
Roots of Characteristic Polynomial
Endogenous variables: DLNWOP DLNWRP DLNWGP DLNYC DDFFR
LNCPI_SA LNY_SA I_SA LNM_SA DLNCR LNVNI_SA DLNREER
Exogenous variables: C DUM
Lag specification: 1 1
Root

Modulus

0.992133

0.992133

0.922570

0.922570

0.780013

0.780013

0.604520 - 0.472821i

0.767466

0.604520 + 0.472821i

0.767466

-0.035488 - 0.687559i

0.688474

-0.035488 + 0.687559i

0.688474

-0.177893 - 0.307171i

0.354964

-0.177893 + 0.307171i

0.354964

-0.302594

0.302594

0.099439 - 0.248002i

0.267195

0.099439 + 0.248002i

0.267195

No root lies outside the unit circle.


VAR satisfies the stability condition.
VAR Residual Serial Correlation LM Tests
Null Hypothesis: no serial correlation at lag order h
Sample: 2001Q1 2011Q4
Included observations: 44
Lags

LM-Stat

Prob

176.8462

0.0326

Probs from chi-square with 144 df.

Source: Authors calculation.

218

Sub-sample 2 (2000Q1-2010Q4):
Roots of Characteristic Polynomial
Endogenous variables: DLNWOP DLNWRP DLNWGP DLNYC DDFFR
LNCPI_SA LNY_SA I_SA LNM_SA DLNCR LNVNI_SA DLNREER
Exogenous variables: C DUM
Lag specification: 1 1
Root

Modulus

0.996133

0.996133

0.934987

0.934987

0.610352 - 0.500988i

0.789632

0.610352 + 0.500988i

0.789632

0.719399

0.719399

-0.060813 - 0.657593i

0.660399

-0.060813 + 0.657593i

0.660399

-0.214948 - 0.396333i

0.450869

-0.214948 + 0.396333i

0.450869

-0.448554

0.448554

0.200519 - 0.204033i

0.286073

0.200519 + 0.204033i

0.286073

No root lies outside the unit circle.


VAR satisfies the stability condition.

VAR Residual Serial Correlation LM Tests


Null Hypothesis: no serial correlation at lag order h
Sample: 2000Q1 2010Q4
Included observations: 41
Lags

LM-Stat

Prob

179.1711

0.0248

Probs from chi-square with 144 df.


Source: Authors calculation.

219

Figure C1: The impulse responses for the Monte Carlo approach
Accumulated Response to Structural One S.D. Innovations
Accumulated Response of LNCPI_SA to Shock1

Accumulated Response of LNCPI_SA to Shock2

Accumulated Response of LNCPI_SA to Shock3

Accumulated Response of LNCPI_SA to Shock4

Accumulated Response of LNCPI_SA to Shock5

.1 2

.1 2

.1 2

.1 2

.1 2

.0 8

.0 8

.0 8

.0 8

.0 8

.0 4

.0 4

.0 4

.0 4

.0 4

.0 0

.0 0

.0 0

.0 0

.0 0

- .0 4

- .0 4
5

10

15

Accumulated Response of LNY_SA to Shock1

- .0 4
5

20

10

15

Accumulated Response of LNY_SA to Shock2

- .0 4
5

20

10

15

Accumulated Response of LNY_SA to Shock3

- .0 4
5

20

10

15

20

Accumulated Response of LNY_SA to Shock4

.0 8

.0 8

.0 8

.0 8

.0 4

.0 4

.0 4

.0 4

.0 4

.0 0

.0 0

.0 0

.0 0

.0 0

- .0 4
5

10

15

- .0 4
5

20

10

15

- .0 4
5

20

10

15

20

- .0 4
5

20

15

Accumulated Response of LNY_SA to Shock5

.0 8

- .0 4

10

10

15

20

10

15

20

Accumulated Response of I_SA to Shock1

Accumulated Response of I_SA to Shock2

Accumulated Response of I_SA to Shock3

Accumulated Response of I_SA to Shock4

Accumulated Response of I_SA to Shock5

1.5

1.5

1.5

1.5

1.5

1.0

1.0

1.0

1.0

1.0

0.5

0.5

0.5

0.5

0.5

0.0

0.0

0.0

0.0

0.0

- 0.5

- 0.5
5

10

15

Accumulated Response of LNM_SA to Shock1

- 0.5
5

20

10

15

Accumulated Response of LNM_SA to Shock2

- 0.5
5

20

10

15

Accumulated Response of LNM_SA to Shock3

- 0.5
5

20

10

15

20

Accumulated Response of LNM_SA to Shock4

.2

.2

.2

.2

.1

.1

.1

.1

.1

.0

.0

.0

.0

.0

- .1
5

10

15

Accumulated Response of DLNCR to Shock1


.0 1

- .1
5

20

10

15

Accumulated Response of DLNCR to Shock2


.0 1

- .1
5

20

10

15

Accumulated Response of DLNCR to Shock3


.0 1

20

- .1
5

20

15

Accumulated Response of LNM_SA to Shock5

.2

- .1

10

10

15

20

Accumulated Response of DLNCR to Shock4

10

15

20

Accumulated Response of DLNCR to Shock5

.0 1

.0 1

220
.0 0

.0 0

.0 0

.0 0

.0 0

- .0 1

- .0 1

- .0 1

- .0 1

- .0 1

.2

.2

.2

.2

.2

.1

.1

.1

.1

.1

.0

.0

.0

.0

.0

Figure C1 (to
be continued)
-.1
5

10

15

-.1

Accumulated Response of DLNCR to Shock1

-.1
5

20

10

15

Accumulated Response of DLNCR to Shock2

-.1
5

20

10

15

Accumulated Response of DLNCR to Shock3

-.1
5

20

10

15

20

Accumulated Response of DLNCR to Shock4

.01

.01

.01

.01

.00

.00

.00

.00

.00

-.01

-.01

-.01

-.01

-.01

-.02
5

10

15

Accumulated Response of LNVNI_SA to Shock1

-.02
5

20

10

15

Accumulated Response of LNVNI_SA to Shock2

-.02
5

20

10

15

Accumulated Response of LNVNI_SA to Shock3

10

15

20

Accumulated Response of LNVNI_SA to Shock4

.6

.6

.6

.6

.4

.4

.4

.4

.4

.2

.2

.2

.2

.2

.0

.0

.0

.0

.0

-.2

-.2

-.2

-.2

-.2

-.4
5

10

15

Accumulated Response of DLNREER to Shock1

-.4
5

20

10

15

Accumulated Response of DLNREER to Shock2

-.4
5

20

10

15

Accumulated Response of DLNREER to Shock3

10

15

20

Accumulated Response of DLNREER to Shock4

.02

.02

.02

.02

.01

.01

.01

.01

.01

.00

.00

.00

.00

.00

-.01

-.01

-.01

-.01

-.01

-.02
5

10

15

20

-.02
5

10

15

20

-.02
5

221

10

15

20

20

10

15

20

Accumulated Response of DLNREER to Shock5

.02

-.02

15

-.4
5

20

10

Accumulated Response of LNVNI_SA to Shock5

.6

-.4

20

-.02
5

20

15

Accumulated Response of DLNCR to Shock5

.01

-.02

10

-.02
5

10

15

20

10

15

20

Figure C1 (to be continued)


Accumulated Response to Structural One S.D. Innovations
Accumulated Response of LNCPI_SA to Shock6

Accumulated Response of LNCPI_SA to Shock7

Accumulated Response of LNCPI_SA to Shock8

Accumulated Response of LNCPI_SA to Shock9

Accumulated Response of LNCPI_SA to Shock10

Accumulated Response of LNCPI_SA to Shock11

Accumulated Response of LNCPI_SA to Shock12

.12

.12

.12

.12

.12

.12

.12

.08

.08

.08

.08

.08

.08

.08

.04

.04

.04

.04

.04

.04

.04

.00

.00

.00

.00

.00

.00

.00

-.04

-.04
5

10

15

Accumulated Response of LNY_SA to Shock6

-.04
5

20

10

15

Accumulated Response of LNY_SA to Shock7

-.04
5

20

10

15

Accumulated Response of LNY_SA to Shock8

-.04
5

20

10

15

Accumulated Response of LNY_SA to Shock9

-.04
5

20

10

15

Accumulated Response of LNY_SA to Shock10

-.04
5

20

10

15

20

Accumulated Response of LNY_SA to Shock11

.08

.08

.08

.08

.08

.08

.08

.04

.04

.04

.04

.04

.04

.04

.00

.00

.00

.00

.00

.00

.00

-.04

-.04

-.04

-.04

-.04

-.04

-.04

-.08

-.08
5

10

15

Accumulated Response of I_SA to Shock6

-.08
5

20

10

15

Accumulated Response of I_SA to Shock7

-.08
5

20

10

15

Accumulated Response of I_SA to Shock8

-.08
5

20

10

15

Accumulated Response of I_SA to Shock9

-.08
5

20

10

15

Accumulated Response of I_SA to Shock10

10

15

20

Accumulated Response of I_SA to Shock11

-1
5

10

15

Accumulated Response of LNM_SA to Shock6

-1
5

20

10

15

Accumulated Response of LNM_SA to Shock7

-1
5

20

10

15

Accumulated Response of LNM_SA to Shock8

-1
5

20

10

15

Accumulated Response of LNM_SA to Shock9

-1
5

20

10

15

Accumulated Response of LNM_SA to Shock10

10

15

20

Accumulated Response of LNM_SA to Shock11

.4

.4

.4

.4

.4

.4

.2

.2

.2

.2

.2

.2

.2

.0

.0

.0

.0

.0

.0

.0

-.2
5

10

15

Accumulated Response of DLNCR to Shock6

-.2
5

20

10

15

Accumulated Response of DLNCR to Shock7

-.2
5

20

10

15

Accumulated Response of DLNCR to Shock8

-.2
5

20

10

15

Accumulated Response of DLNCR to Shock9

.04

.04

.04

.04

.02

.02

.02

.00

.00

.00

-.2
5

20

10

15

Accumulated Response of DLNCR to Shock10

10

15

20

Accumulated Response of DLNCR to Shock11

10

15

20

10

15

20

Accumulated Response of DLNCR to Shock12

.04

.04

.04

.02

.02

.02

.02

.00

.00

.00

.00

222

20

-.2
5

20

15

Accumulated Response of LNM_SA to Shock12

.4

-.2

10

-1
5

20

20

Accumulated Response of I_SA to Shock12

-1

15

-.08
5

20

10

Accumulated Response of LNY_SA to Shock12

Figure C1 (to be continued)


Accumulated Response t o Struct ural One S.D. Innovations
A ccumulated R esponse of LN C P I_S A to S hock6

A ccumulated R esponse of LN C P I_S A to S hock7

A ccumulated R esponse of LN C P I_S A to S hock8

A ccumulated R esponse of LN C P I_S A to S hock9

A ccumulated Response of LNC P I_S A to S hock10

A ccumulated Response of LNC P I_S A to S hock11

A ccumulated Response of LNC P I_S A to S hock12

.12

.12

.12

.12

.12

.12

.12

.08

.08

.08

.08

.08

.08

.08

.04

.04

.04

.04

.04

.04

.04

.00

.00

.00

.00

.00

.00

-.04

-.04
5

10

15

A ccumulated R esponse of LNY _S A to S hock6

-.04
5

20

10

15

A ccumulated R esponse of LN Y _S A to S hock7

-.04
5

20

10

15

A ccumulated R esponse of LN Y _S A to S hock8

-.04
5

20

10

15

A ccumulated R esponse of LNY _S A to S hock9

.00

-.04
5

20

10

15

A ccumulated R esponse of LNY _S A to S hock10

-.04
5

20

10

15

20

A ccumulated R esponse of LN Y _S A to S hock11

.08

.08

.08

.08

.08

.08

.04

.04

.04

.04

.04

.04

.04

.00

.00

.00

.00

.00

.00

.00

-.04

-.04

-.04

-.04

-.04

-.04

-.04

-.08

-.08
5

10

15

-.08
5

20

A ccumulated R esponse of I_S A to S hock6

10

15

-.08
5

20

A ccumulated R esponse of I_S A to S hock7

10

15

-.08
5

20

A ccumulated R esponse of I_S A to S hock8

10

15

-.08
5

20

A ccumulated R esponse of I_S A to S hock9

10

15

10

15

20

A ccumulated Response of I_S A to S hock11

-1
5

10

15

A ccumulated R esponse of LNM_S A to S hock6

-1
5

20

10

15

A ccumulated R esponse of LN M_S A to S hock7

-1
5

20

10

15

A ccumulated R esponse of LN M_S A to S hock8

-1
5

20

10

15

A ccumulated R esponse of LNM_S A to S hock9

10

15

A ccumulated R esponse of LNM_S A to S hock10

10

15

20

A ccumulated R esponse of LN M_S A to S hock11

.4

.4

.4

.4

.4

.4

.2

.2

.2

.2

.2

.2

.2

.0

.0

.0

.0

.0

.0

-.2
5

10

15

A ccumulated R esponse of D LN C R to S hock 6

-.2
5

20

10

15

A ccumulated R esponse of DLN C R to S hock 7

-.2
5

20

10

15

A ccumulated R esponse of D LN C R to S hock 8

-.2
5

20

10

15

A ccumulated R esponse of D LN C R to S hock 9

10

15

A ccumul ated R es ponse of D LN CR to S hock10

10

15

20

A ccumul ated R es ponse of D LN C R to S hock11

.04

.04

.04

.04

.04

.04

.02

.02

.02

.02

.02

.02

.02

.00

.00

.00

.00

.00

.00

-.02
5

10

15

A ccumulated R esponse of LN V N I_S A to S hock6


.8

-.02
5

20

10

15

A ccumulated R esponse of LN V N I_S A to S hock7


.8

.4

-.02
5

20

10

15

A ccumulated R esponse of LN V N I_S A to S hock8


.8

.4

-.02
5

20

10

15

A ccumulated R esponse of LN V N I_S A to S hock9


.8

.4

10

15

A ccumulated Response of LNV N I_S A to S hock10


.8

.4

10

15

20

A ccumulated Response of LNV N I_S A to S hock11


.8

.4

.4

.0

.0

.0

.0

.0

-.4

-.4

-.4

-.4

-.4

-.8
10

15

-.8
5

20

10

15

-.8
5

20

A ccumulated R esponse of D LN R E E R to S hock7

10

15

-.8
5

20

A ccumulated R esponse of DLN R E E R to S hock8

10

15

-.8
5

20

A ccumulated R esponse of D LN R E E R to S hock9

10

15

10

15

20

A ccumulated R esponse of D LN RE E R to S hock11

.04

.04

.04

.04

.04

.04

.02

.02

.02

.02

.02

.02

.02

.00

.00

.00

.00

.00

.00

.00

-.02

-.02

-.02

-.02

-.02

-.02

-.02

10

15

20

10

15

20

10

15

20

Source: Authors calculation.

223

10

15

20

10

15

20

10

15

20

10

15

20

A ccumulated R esponse of D LN R E E R to S hock12

.04

20

-.8
5

20

A ccumulated R esponse of D LN R E E R to S hock10

15

.4

.0
-.4

10

.8

.0

-.8

20

A ccumulated Response of LNV N I_S A to S hock12

-.4

A ccumulated R esponse of D LN R E E R to S hock6

15

-.02
5

20

10

.00

-.02
5

20

20

A ccumul ated R es ponse of D LN C R to S hock12

.04

-.02

15

-.2
5

20

10

.0

-.2
5

20

20

A ccumulated R esponse of LN M_S A to S hock12

.4

-.2

15

-1
5

20

10

-1
5

20

20

A ccumulated R esponse of I_S A to S hock12

-1

15

-.08
5

20

A ccumulated R esponse of I_S A to S hock10

10

A ccumulated Response of LN Y _S A to S hock12

.08

10

15

20

REFERENCES

Afandi, A. (2005). 'Monetary policy transmission mechanism and structural breaks in


Indonesia', PhD thesis, University of Wollongong, New South Wales.
Agnani, B. and Aray, H. (2011). 'The January effect across volatility regimes',
Quantitative Finance, 11(6): 947-953.
Amisano, G. and Giannini, C. (1997). Topics in structural VAR econometrics,
Springer, Berlin.
Ando, A. and Modigliani, F. (1963). 'The "Life Cycle" Hypothesis of Saving:
Aggregate Implications and Tests', The American Economic Review, 53(1): 5584.
Aslanidi, O. (2007). The optimal monetary policy and the channels of monetary
transmission mechanism in CIS-7 countries: The case of Georgia, Discussion
Paper Series of Center for Economic Research and Graduate Education, No.
2007-171, Charles Universtiy, Czech Republic.
Asteriou, D. and Kavetsos, G. (2006). 'Testing for the existence of the 'January effect'
in transition economies', Applied Financial Economics Letters, 2(6): 375-381.
Australian Bureau of Statistics. (2013). Time series analysis: seasonal adjustment
methods, Australian Bureau of Statistics, viewed 19/2/2013, www.abs.gov.au.
Bain, K. and Howells, P. G. A. (2003). Monetary economics: policy and its theoretical
basis, Palgrave Macmillan Ltd, Basingstoke.
Ben-David, D., Lumsdaine, R. L. and Papell, D. H. (2003). 'Unit roots, postwar
slowdowns and long-run growth: evidence from two structural breaks', Empirical
Economics, 28(2): 303-319.
Berkelmans, L. (2005). Credit and monetary policy: an Australian SVAR, Reseach
Discussion Paper, No. 2005-06, Reserve Bank of Australia.
Bernanke, B. S. and Gertler, M. (1995). 'Inside the black box: the credit channel of
monetary policy transmission', Journal of Economic Perspectives, 9(4): 27.
Bernanke, B. S. and Mihov, I. (1998). 'Measuring monetary policy', The Quarterly
Journal of Economics, 113(3): 869-902.
Berry, W. D. (1984). Nonrecursive causal models, Sage Publications, Beverly Hills,
CA.
Bhattacharyya, I. and Sensarma, R. (2007). 'How effective are monetary policy signals
in India?', Journal of Policy Modeling, 30(1): 169-183.
Bhuiyan, R. (2012). 'Monetary transmission mechanisms in a small open economy: a
Bayesian structural VAR approach', Canadian Journal of Economics/Revue
canadienne d'conomique, 45(3): 1037-1061.
224

Bicchal, M. (2010). Monetary policy and inflation in India: a structural VAR analysis,
viewed 15/4/2012, http://ssrn.com/abstract=1813886.
Black, L., Hancock, D. and Passmore, S. (2010). The bank lending channel of
monetary policy and its effect on mortgage lending, Finance and Economics
Discussion Series, No. 2010.39, Divisions of Research and Statistics and
Monetary Affairs-Federal Reserve Board, Washington, D.C.
Blanchard, O. and Sheen, J. R. (2013). Macroeconomics: Australasian edition,
Pearson Australia Group, Frenchs Forest, N.S.W.
Blanchard, O. J. and Quah, D. (1989). 'The dynamic effects of aggregate demand and
supply disturbances', The American Economic Review, 79(4): 655-673.
Bohl, M. T. and Salm, C. A. (2010). 'The other January effect: international evidence',
The European Journal of Finance, 16(2): 173-182.
Boivin, J., Kiley, M. T. and Mishkin, F. S. (2010). 'Chapter 8 - How has the monetary
transmission mechanism evolved over time?',in Handbook of Monetary
Economics, edited by M. F. Benjamin and W. Michael, Elsevier, 3: 369-422.
Campbell, J. L. (2004). Institutional change and globalization, Princeton University
Press, Princeton, N.J.
Cecchetti, S. G. (1999). 'Legal structure, financial structure, and the monetary policy
transmission mechanism', Economic Policy Review - Federal Reserve Bank of
New York, 5(2): 9.
Cecchetti, S. G. and Krause, S. (2001). 'Financial structure, macroeconomic stability
and monetary policy', XII Symposium of Moneda y Credito, Madrid.
Central Bureau of Statistics. (2013). Seasonal adjustment, Central Bureau of Statistics,
viewed 19/5/2013, www.census.gov.
Charoenseang, J. and Manakit, P. (2007). 'Thai monetary policy transmission in an
inflation targeting era', Journal of Asian Economics, 18(1): 144-157.
Chowdhury, K. (2011). 'Modelling the Balassa-Samuelson effect in Australia',
Australasian Accounting Business and Finance Journal, 5(1): 77-91.
Chowdhury, K. (2012a). 'Modelling the dynamics, structural breaks and the
determinants of the real exchange rate of Australia', Journal of International
Financial Markets, Institutions and Money, 22(2): 343-358.
Chowdhury, K. (2012b). 'The real exchange rate and the Balassa-Samuelson
hypothesis in SAARC countries: an appraisal', Journal of the Asia Pacific
Economy, 17(1): 52.
Clarida, R. H., Gali, J. and Gertler, M. (2000). 'Monetary policy rules and
macroeconomic stability: evidence and some theory', Quarterly Journal of
Economics 115(1): 147-180.
225

Claus, I. (2011). 'Inside the black box: how important is the credit channel relative to
the interest and exchange rate channels?', Economic Modelling, 28(1): 1-12.
Cleveland, W. S. and Devlin, S. J. (1980). 'Calendar effects in monthly time series:
detection by spectrum analysis and graphical methods', Journal of the American
Statistical Association, 75(371): 487-496.
Cushman, D. O. and Zha, T. (1997). 'Identifying monetary policy in a small open
economy under flexible exchange rates', Journal of Monetary Economics, 39(3):
433-448.
Deardorff, A. (2010). Glossary of international economics, viewed 11/7/2011,
http://www-personal.umich.edu/~alandear/glossary/.
Del Negro, M. and Schorfheide, F. (2004). 'Priors from general equilibrium models for
VARs', International Economic Review, 45(2): 643-673.
Dennis, R. (2003). 'Exploring the role of the real exchange rate in Australian monetary
policy', Economic Record, 79(244): 20-38.
Diebold, F. X. (2007). Elements of forecasting, Thomson South-Western, Mason, OH.
Du, L., He, Y. and Wei, C. (2010). 'The relationship between oil price shocks and
Chinas macro-economy: an empirical analysis', Energy Policy, 38(8): 41424151.
Dungey, M. and Fry, R. (2009). 'The identification of fiscal and monetary policy in a
structural VAR', Economic Modelling, 26(6): 1147-1160.
Dungey, M. and Pagan, A. (2000). 'A structural VAR model of the Australian
economy', Economic Record, 76(235): 321-342.
Dungey, M. and Pagan, A. R. (2007). Extending a structural VAR model of the
Australian economy, NCER Working Paper Series, No. 21, National Centre of
Econometric Research, Australia.
Elbourne, A. and de Haan, J. (2006). 'Financial structure and monetary policy
transmission in transition countries', Journal of Comparative Economics, 34(1):
1-23.
Elbourne, A. and de Haan, J. (2009). 'Modeling monetary policy transmission in
acceding countries: vector autoregression versus structural vector
autoregression', Emerging Markets Finance and Trade, 45(2): 4-20.
Enders, W. (2004). Applied econometric time series, J. Wiley, Hoboken, N.J.
Eo, W. S. (2010). Seasonal adjustment of monthly economic time series in Korea: a
comparison of X-12-ARIMA and TRAMO-SEATSYonsei Economic Research
Institute.
Fleming, J. M. (1962). 'Domestic financial policies under fixed and under floating
exchange rates', International Monetary Fund Staff Papers, 9(3): 369.
226

Fountas, S. and Segredakis, K. N. (2002). 'Emerging stock markets return


seasonalities: the January effect and the tax-loss selling hypothesis', Applied
Financial Economics, 12(4): 291-299.
Friedman, M. (1957). A theory of the consumption function, Princeton: Princeton
University Press.
Gambacorta, L. (2011). 'Do bank capital and liquidity affect real economic activity in
the long run?: a VECM analysis for the US', Economic notes, 40(3): 75-91.
General Statistics Office of Vietnam. (2010). Trade and prices, viewed 4/9/2012,
http://www.gso.gov.vn/default.aspx?tabid=433&idmid=3.
Glynn, J., Perera, N. and Verma, R. (2007). 'Unit root tests and structural breaks: a
survey with applications', Revista de mtodos cuantitativos para la economa y
la empresa, 3: 63-79.
Gottschalk, J. (2001). An introduction into the SVAR methodology: identification,
interpretation and limitations of SVAR models. Kiel Working Paper, Kiel
Institute of World Economics. 1072: 42.
Goujon, M. (2006). 'Fighting inflation in a dollarized economy: the case of Vietnam',
Journal of Comparative Economics, 34(3): 564-581.
Gujarati, D. N. and Porter, D. C. (2009). Basic econometrics, McGraw-Hill Irwin,
Boston.
Hespeler, F. (2013). 'A VECM evaluation of monetary transmission in Uzbekistan',
Economic change and restructuring, 46(2): 219-253.
Hoa Binh Securities. (2011). Report on Vietnam national petroleum group (in
Vietnamese), viewed 19/5/2012, www.stoxplus.com/download.asp id=2710.
Hoang, V. Q. (2009). The Vietnam's transition economy and its fledgling financial
markets: 1986-2003, SSRN Working Paper Series, No. 04/032.
Hoang, V. Q. and Dung, T. T. (2009). 'A note on studies of monetary policy and
implementation in Vietnam', SSRN Working Paper Series.
Ireland, P. N. (2006). The monetary transmission mechanism, FRB Working Paper.
Ishii, S. (2007, 30/8/2011). IMF statements at donor meetings, Consultative group
meeting for Vietnam, viewed 20/12/2011,
http://www.imf.org/external/np/dm/2007/121007.htm.
Johansen, S. (1988). 'Statistical analysis of cointegration vectors', Journal of Economic
Dynamics and Control, 12(2-3): 231-254.
Johansen, S. and Juselius, K. (1990). 'Maximum-likelihood-estimation and inference
on cointegration - with applications to the demand for money.', Oxford Bulletin
of Economics and Statistics, 52(2): 169-210.
227

Jorgenson, D. W. (1963). 'Capital Theory and Investment Behavior', The American


Economic Review, 53(2): 247-259.
Kamin, S., Turner, P. and Dack, J. V. (1998). The transmission mechanism of
monetary policy in emerging market economies: an overview, Policy papers,
No. 3, Bank for International Settlements, Switzerland.
Kilian, L. (2011). Structural vector autoregressions, CEPR Discussion Papers, No.
DP8515, Centre for Economic Policy Research.
Kim, S. and Roubini, N. (2000). 'Exchange rate anomalies in the industrial countries: a
solution with a structural VAR approach', Journal of Monetary Economics,
45(3): 561-586.
Kubo, A. (2008). 'Macroeconomic impact of monetary policy shocks: evidence from
recent experience in Thailand', Journal of Asian Economics, 19(1): 83-91.
Le, H. V. and Pfau, W. D. (2009). 'VAR analysis of the monetary transmission
mechanism in Vietnam', Applied Econometrics and International Development,
9(1): 165-176.
Lee, J. and Strazicich, M. C. (2003). 'Minimum lagrange multiplier unit root test with
two structural breaks', The Review of Economics and Statistics, 85(4): 10821089.
Lee, J. and Strazicich, M. C. (2004). Minimum LM unit root test with one structural
break, Working Paper, No. 04-17, Department of Economics, Appalachian State
University.
Li, Y. D., can, T. B. and Xu, K. (2010). 'The impact of monetary policy shocks on
stock prices: evidence from Canada and the United States', Journal of
International Money and Finance, 29(5): 876-896.
Loayza, N. and Schmidt-Hebbel, K. (2002). 'Monetary policy functions and
transmission mechanisms: an overview',in Monetary Policy: Rules and
Transmission Mechanisms, edited, Central Bank of Chile, Santiago, Chile.
Lucas, R. E. (1976). 'Econometric policy evaluation: A critique', Carnegie-Rochester
Conference Series on Public Policy, 1(1): 19-46.
Lumsdaine, R. L. and Papell, D. H. (1997). 'Multiple trend breaks and the unit-root
hypothesis', The Review of Economics and Statistics, 79(2): 212-218.
Mankiw, N. G. (2010). Macroeconomics, Worth Publishers, New York.
Masina, P. P. (2006). Vietnam's development strategies, Routledge, London.
Mello, L. and Pisu, M. (2010). 'The bank lending channel of monetary transmission in
Brazil: a VECM approach', The quarterly review of economics and finance,
50(1): 50-60.

228

Microsoft (2009a). 'Chapter 11. Series',in EViews 7 Users Guide II, edited,
Quantitative Micro Software, LLC, Irvine, CA.
Microsoft (2009b). 'Chapter 15. Vector autoregression and error correction models',in
EViews 7 Users Guide II, edited, Quantitative Micro Software, LLC, the United
States.
Mishkin, F. S. (1995). 'Symposium on the monetary transmission mechanism', The
Journal of Economic Perspectives, 9(4): 3-10.
Mishkin, F. S. (2010). The Economics of Money, Banking, and Financial Markets,
Addison Wesley Longman Ltd, Boston.
Miyazaki, M. (2010). IMF statements at donor meetings Consultative group meeting
for Vietnam, viewed 30/8/2011,
http://www.imf.org/external/np/dm/2010/120610.htm.
Modiglina, F. and Brumberg, R. (1954). 'Utility Analysis and the Consumption
Fucntion',in Post-Keynesian Economics, edited by K. Kurihana, Rutger
University Press: 338-436.
Mundell, R. A. (1963). 'Capital Mobility and Stabilization Policy under Fixed and
Flexible Exchange Rates', The Canadian Journal of Economics and Political
Science / Revue canadienne d'Economique et de Science politique, 29(4): 475485.
Nakahira, K. (2009). 'A structural VAR analysis of the monetary policy stance in
Japan', The International Journal of Economic Policy Studies, 4(5): 78-103.
Narayan, P. K. and Narayan, S. (2010). 'Modelling the impact of oil prices on
Vietnams stock prices', Applied Energy, 87(1): 356-361.
Narayan, P. K., Narayan, S., Mishra, S. and Smyth, R. (2012). 'An analysis of Fiji's
monetary policy transmission', Studies in Economics and Finance, 29(1): 52-70.
Narayan, P. K. and Popp, S. (2010). 'A new unit root test with two structural breaks in
level and slope at unknown time', Journal of Applied Statistics, 37(9): 14251438.
Narayan, S. (2013). 'A structural VAR model of the Fiji islands', Economic Modelling,
31: 238-244.
Nelson, C. R. and Plosser, C. I. (1982). 'Trends and random walks in macroeconomic
time series: some evidence and implications', Journal of Monetary Economics,
10(2): 139.
Ngoc, P. M. (2008). 'Sources of Vietnam's economic growth', Progress in
Development Studies, 8(3): 209-229.
Nguyen, H. M., Cavoli, T. and Wilson, J. K. (2012). 'The determinants of inflation in
Vietnam, 2001-09', ASEAN Economic Bulletin, 29(1): 1.
229

Nguyen, P. L. (2010). Monetary transmission mechanism in Vietnam under the


quatitative approach (in Vietnamese), The State Bank of Vietnam, viewed
13/7/2011, http://www.sbv.gov.vn.
Nguyen, Q. T. and Bui, T. (2011). Changes in economic structure in the past 10 years
(in Vietnamese), viewed 4/9/2011,
http://www.csdp.vn/Download.aspx/7E52053FB6DF4446930155D21F429DF6/
1/No5-2011-Thay_doi_cau_truc_KT_10_nam_gan_day.doc.
North, D. C. (1994). 'Economic performance through time', The American Economic
Review, 84(3): 359-368.
Norway, S. (2008). Seasonal adjustment: general information, Statistics Norway,
viewed 20/3/2013,
www.ssb.no/english/metadata/methods/seasonal_adjustment.pdf.
Oliver, H., Peter, W. and Andreas, W. (2004). 'Bank Lending and Monetary Policy
Transmission: A VECM Analysis for Germany: Bankkredite und geldpolitische
Transmission: Eine VECM Analyse fr Deutschland', Jahrbcher fr
Nationalkonomie und Statistik, 224(5): 511.
Pagan, A. (1995). Three econometric methodologies: an update, Oxford: Basil
Blackwell.
Pahlavani, M., Wilson, E. and Valadkhani, A. (2006). 'Structural changes in the
Iranian economy: an empirical analysis with endogenously determined breaks',
International Journal of Applied Business and Economic Research, 4(1): 1-8.
Perron, P. (1989). 'The Great crash, the oil price shock, and the unit root hypothesis',
Econometrica, 57(6): 1361-1401.
Pham, M. C. and Vuong, Q. H. (2009). Vietnam economy: the ups and downs, and
breakthroughs, Knowledge Publishing, Hanoi.
Phan, T. N., Harvie, C. and Tran, V. H. (2006). Vietnam's economic transition:
policies, issues and prospects, ICFAI University Press, Hyderabad, India.
Phuc, N. T. and Duc-Tho, N. (2009). 'Exchange rate policy in Vietnam, 19852008',
ASEAN Economic Bulletin, 26(2): 137-163.
Poon, C.-c. and Wong, F.-k. (2011). 'China's monetary policy and its transmission
mechanisms before and after the financial tsunami', Chinese Economy, 44(3): 84108.
Popp, S. (2008). 'New innovational outlier unit root test with a break at an unknown
time', Journal of Statistical Computation and Simulation, 78(12): 1143-1159.
Raghavan, M. and Silvapulle, P. (2008). 'Structural VAR approach to Malaysian
monetary policy framework: evidence from the pre- and post-Asian crisis
periods', The 49th Annual Conference of the New Zealand Association of
Economists, Wellington.
230

Romer, D. (2012). Advanced macroeconomics, McGraw-Hill/Irwin, New York.


Safaei, J. and Cameron, N. E. (2003). 'Credit channel and credit shocks in Canadian
macrodynamics - a structural VAR approach', Applied Financial Economics,
13(4): 267-277.
Sbordone, A. M., Tambalotti, A., Rao, K. and Walsh, K. (2010). 'Policy analysis using
DSGE models: an introduction', Economic policy review, 16(2): 23-43.
Sellon, G. H. J. (2002). 'The changing U.S. financial system: Some implications for the
monetary transmission mechanism', Economic Review - Federal Reserve Bank of
Kansas City, 87(1): 5-35.
Sen, A. (2003). 'On unit-root tests when the alternative is a trend-break stationary
process', Journal of Business and Economic Statistics, 21(1): 174-184.
Sharifi-Renani, H. (2010). 'A structural VAR approach of monetary policy in Iran',
International Conference on Applied Economics, Athens, Greece.
Shrestha, M. B. and Chowdhury, K. (2005). A sequential procedure for testing unit
roots in the presence of structural break in time series data, Economics Working
Paper Series, No. University of Wollongong.
Shuja, N., Lazim, M. A. and Wah, Y. B. (2007). 'Moving holiday effects adjustment
for Malaysian economic time series', Journal of the Department of Statistics, 1:
35-50.
Sims, C. (1986). 'Are forecasting models usable for policy analysis?', Federal Reserve
Bank of Minneapolis. Quarterly Review - Federal Reserve Bank of Minneapolis,
10(1): 2.
Sims, C. A. (1992). 'Interpreting the macroeconomic time-series facts - the effects of
monetary policy', European Economic Review, 36(5): 975-1000.
Sims, C. A. and Zha, T. A. (1998). Does monetary policy generate recessions?,
Working Paper, No. 98-12, Federal Reserve Bank of Atlanta.
Sultan, J. (2012). 'Options on federal funds futures and interest rate volatility', Journal
of Futures Markets, 32(4): 330-359.
Suzuki, T. (2004). 'Is the lending channel of monetary policy dominant in Australia?',
Economic Record, 80(249): 145-156.
Tang, S. (2011). A general theory of institutional change, Routledge, New York.
Taylor, J. B. (1995). 'The monetary transmission mechanism: an empirical framework',
The Journal of Economic Perspectives, 9(4): 11-26.
Taylor, J. B. (2002). 'The monetary transmission mechanism and the evaluation of
monetary policy rules',in Monetary Policy: Rules and Transmission Mechanisms,
edited, Central Bank of Chile, Santiago, Chile.
231

Tobin, J. (1969). 'A General Equilibrium Approach To Monetary Theory', Journal of


Money, Credit and Banking, 1(1): 15-29.
Tovar, C. E. (2009). 'DSGE Models and Central Banks', Economics: The Open-Access,
Open-Assessment E-Journal, 3(2009-16): 1.
Tran, C. T., Do, L. H. and Le, N. M. (2013). Who has benefited from high rice prices
in Vietnam?, Oxfam and IPSARD, viewed 31/10/2013,
http://www.oxfamamerica.org/files/who-benefits-rice-prices-vietnam-full20131017.pdf.
Tran, K. V. (2009). Monetary policy in Vietnam: evidence from a structural VAR,
viewed 31/10/2011, http://ssrn.com/abstract=1372294.
Vietnam National Assembly (2011a). The Law on the State Bank of Vietnam. V. N.
Assembly. Hanoi, Vietnam National Assembly.
Vietnam National Assembly (2011b). Report on the socio-economic in six first months
of 2011Vietnam National Assembly, Hanoi.
Vinayagathasan, T. (2013). Monetary policy and the real economy: a structural VAR
approach for Sri Lanka, GRIPS Discussion Papers, No. National Graduate
Institute for Policy Studies.
Vogelvang, B. (2005). Econometrics: theory and applications with EViews, Financial
Times Prentice Hall, Harlow.
Vuong, Q. H. and Tran, T. D. (2009). A note on studies of monetary policy and
implementation in Vietnam. SSRN Working Paper Series.
Waal, A. and Eyden, R. (2014). 'Monetary policy and inflation in South Africa: A
VECM augmented with foreign variables', South African Journal of Economics,
82(1): 117-140.
Wilson, E. (2011). Monetary policy, University of Wollongong, viewed 20/5/2011.
World Bank (2010). Vietnam development report (VDR) 2010: modern instittutions,
viewed 15/8/2012, http://www.worldbank.org.
World Bank (2011). Vietnam development report (VDR) 2012: market economy for a
middle-income country, viewed 15/8/2012,
http://www.worldbank.org/vi/news/2012/01/12/vietnam-development-report2012-an-overview.
Zaidi, M. A. S. and Fisher, L. A. (2010). 'Monetary policy and foreign shocks: a
SVAR analysis for Malaysia', Korea and the World Economy, 11(3): 527-550.
Zaidi, M. A. S. B. (2011). 'Structural vector autoregressive analysis of monetary policy
in Malaysia', PhD thesis, The University of New South Wales.

232

Zivot, E. and Andrews, D. W. K. (1992). 'Further evidence on the great crash, the oilprice shock, and the unit-root hypothesis', Journal of Business and Economic
Statistics, 10(3): 251-270.
Zaidi, M. A. S. and Fisher, L. A. (2010). 'Monetary policy and foreign shocks: a
SVAR analysis for Malaysia', Korea and the World Economy, 11(3): 527-550.
Zaidi, M. A. S. B. (2011). 'Structural vector autoregressive analysis of monetary policy
in Malaysia', PhD thesis, The University of New South Wales.
Zivot, E. and Andrews, D. W. K. (1992). 'Further evidence on the great crash, the oilprice shock, and the unit-root hypothesis', Journal of Business and Economic
Statistics, 10(3): 251-270.

233

Das könnte Ihnen auch gefallen