Thông tin tài liệu
AN EMPIRICAL STUDY ON MACROECONOMIC DETERMINANTS OF
EXCHANGE RATE AND STRESS TESTING APPLICATION ON VIETNAM
NON – FINANCIAL CORPORATE SECTOR
In Partial Fulfillment of the Requirements of the Degree of
MASTER OF BUSINESS ADMINISTRATION
In Finance
By
Ms: Le Tran Nguyen Nhung
ID: MBA06022
Advisor: Dr. Ho Diep
International University - Vietnam National University HCMC
August 2014
AN EMPIRICAL STUDY ON MACROECONOMIC DETERMINANTS OF
EXCHANGE RATE AND STRESS TESTING APPLICATION ON VIETNAM
NON – FINANCIAL CORPORATE SECTOR
In Partial Fulfillment of the Requirements of the Degree of
MASTER OF BUSINESS ADMINISTRATION
In Finance
by
Ms: Le Tran Nguyen Nhung
ID: MBA06022
International University - Vietnam National University HCMC
August 2014
Under the guidance and approval of the committee, and approved by all its members,
this thesis has been accepted in partial fulfillment of the requirements for the degree.
Approved:
---------------------------------------------Chairperson
--------------------------------------------Committee member
---------------------------------------------Committee member
--------------------------------------------Committee member
---------------------------------------------Committee member
--------------------------------------------Committee member
Acknowledge
To complete this thesis, I have been benefited from the following people.
Firstly, I would like to express my deepest gratefulness to my advisor, Dr. Ho
Diep, for choosing me, giving many dedicated and enthusiastic instructions and
offering me good opportunity to develop myself during the six – month period of my
research. I also send my sincere gratitude to the professors and lecturers who teach me
valuable knowledge during the MBA course.
Secondly, I would like to present my special thanks to Mr. Vinh, Mr. Linh and
Ms. Ai from Vietcombank Fund Management (VCBF) for spending their time to
participate in many discussions with Dr. Diep and me and helps me gain a great deal of
useful information and innovative ideas.
Finally, to my family and best friends, thank you for your endless love and
constant support which inspires me to overcome any difficulties and discouragement to
achieve my MBA degree.
Le Tran Nguyen Nhung
- i-
Plagiaris m State ments
I would like to declare that, apart from the acknowledged references, this thesis
either does not use language, ideas, or other original material from anyone; or has not
been previously submitted to any other educational and research programs or
institutions. I fully understand that any writings in this thesis contradicted to the above
statement will automatically lead to the rejection from the MBA program at the
International University – Vietnam National University Hochiminh City.
- ii -
Copyright Statement
This copy of the thesis has been supplied on condition that anyone who consults
it is understood to recognize that its copyright rests with its author and that no quotation
from the thesis and no information derived from it may be published without the
author’s prior consent.
© Le Tran Nguyen Nhung/ MBA06022/ 2014
- iii -
Table of Contents
1. Introduction ............................................................................................................................ 1
1.1. Background ...................................................................................................................... 1
1.2. Research Problems ........................................................................................................... 2
1.3. Research Objectives ......................................................................................................... 5
1.4. Research Questions .......................................................................................................... 6
1.5. Research Methodology .................................................................................................... 6
1.6. Research Scope and Limitation........................................................................................ 6
1.7. Implications ..................................................................................................................... 7
1.8. Research Structure ........................................................................................................... 7
2. Lite rature Review .................................................................................................................. 9
2.1. Exchange Rate .................................................................................................................. 9
2.2. Exchange Rate Regimes ................................................................................................... 10
2.3. Approaches to Exchange Rate Determination ................................................................. 12
2.4. Determinants of Exchange Rate ....................................................................................... 16
2.5. Stress Testing Theory ....................................................................................................... 21
2.6. Value at Risk Model......................................................................................................... 26
2.7. Theoretical Framework .................................................................................................... 30
3. Overvie w of Exchange Rate in Vietnam .............................................................................. 31
3.1. Vietnam Exchange Rate Regimes .................................................................................... 31
3.2. The Development of Exchange Rate................................................................................ 34
3.3. Vietnam’s Exchange Rate Policy in Comparison with Other Asian Countries ............... 45
3.4. The Relationship between Exchange Rate and Its Determinants .................................... 49
4. Data Collection and Research Methodology........................................................................ 64
4.1. Data Collection................................................................................................................. 64
- iv -
4.2. Research Methodology .................................................................................................... 64
5. Data Analysis .......................................................................................................................... 75
5.1. Descriptive Statistics ........................................................................................................ 75
5.2. Regression Model for Exchange Rate’s Macro Determinants ......................................... 76
5.3. Macro Stress Testing for Exchange Rate Exposure ......................................................... 88
6. Conclusion and Recomme ndation ........................................................................................ 98
6.1. Conclusions ...................................................................................................................... 98
6.2. Recommendations ............................................................................................................ 102
6.3. Suggestions for Future Research...................................................................................... 104
References .................................................................................................................................... 105
- v-
List of Tables
Table 2.1: The strengths and weaknesses of VaR approaches
Table 3.1: Exchange rate regimes in Vietnam from 1989 to 2013
Table 3.2: Changes in the trading band around the interbank rate
Table 3.3: The SBV’ measures for managing exchange rate in 2012
Table 3.4: The exchange rate regime and monetary policy framework
Table 4.1: Expected signs of independent variables
Table 5.1: Data description
Table 5.2: Autocorrelation test
Table 5.3: Heteroscedasticity test
Table 5.4: ARCH test
Table 5.5: Correlation matrix
Table 5.6: Omitted variables test
Table 5.7: Redundant variables test
Table 5.8: Hausman test
Table 5.9: Regression model
Table 5.10: VaR using the variance – covariance approach
Table 5.11: VaR using the historical simulation approach
Table 5.12: VaR using Monte Carlo simulation
Table 5.13: Predicted EXR under ∆ GDP shock using historical simulation
Table 5.14: VaR under ∆ GDP shock
Table 5.15: Predicted EXR under ∆ CUR shock using historical simulation
Table 5.16: VaR under ∆ CUR shock
Table 5.17: Predicted EXR using historical simulation
Table 5.18: VaR under ∆ GDP and ∆ CUR shock together
- vi -
List of Figures
Figure 2.1: Stress testing and scenario
Figure 2.2: Framework for stress testing of individual portfolios
Figure 2.3: VaR at 95% confidence level
Figure 2.4: Theoretical framework
Figure 3.1: Official VND/USD rate, 1989 – 2013
Figure 3.2: Monthly fluctuation of VND/USD, 2008 – 2013
Figure 3.3: The volatility of the exchange rate in 2012
Figure 3.4: Nominal exchange rate index of several Asian currencies against USD
Figure 3.5: The relationship between exchange rate and GDP growth rate
Figure 3.6: The relationship between exchange rate and inflation rate
Figure 3.7: The relationship between exchange rate and interest rate
Figure 3.8: The relationship between exchange rate and current account
Figure 3.9: The relationship between exchange rate and FDI
Figure 3.10: The relationship between exchange rate and foreign exchange reserves
Figure 3.11: The relationship between VND/USD and VND/EUR, VND/JPY
Figure 5.1: Normality test
Figure 5.2: Stability test
- vii -
Abstract
Due to the fact that the exchange rate fluctuation can affect business performance
but Vietnamese firms have not paid much attention to gauging or managing their foreign
exchange exposure, this research examines the relationship between the exchange rate and
its macroeconomic determinants, and then develops a framework to conduct macro stress
testing by combining modern techniques such as Monte Carlo simulation and Value-atRisk (VaR) approaches to predict the level of the exchange rate variation as well as the
expected losses that companies could suffer. This provides Vietnamese companies with a
systematic way to calculate, and to reserve risk capital for their currency exposures.
This thesis proceeded by showing that the exchange rate in Vietnam is negatively
impacted by the GDP growth rate and current account balance, and positively influenced
by the exchange rate (t-1) lagged, where t is time. Based on the resultant regression model,
VaRs of the exchange rate variation and the expected loss are computed by using three
different approaches, (i) variance – covariance matrix, (ii) historical simulation and (iii)
Monte Carlo simulation; in which the Monte Carlo gives the highest and most
conservative loss that companies should reserve for. Macro stress testing is applied to
create adverse scenarios and VaR is re-calculated in the stressed market conditions. In this
study, we used a hypothetical case of a Vietnamese company with a short position of USD
1 million (the methodology can be easily adapted to any Vietnamese enterprise), under the
more conservative Monte Carlo simulation, the minimum capital reserve for the
Vietnamese company to cover the abnormal stressed loss for the next quarter and at 99%
confidence is about VND 680 million (or approximately USD 31,000).
Keywords: Exchange rate, macroeconomic determinants, stress testing, value-at-risk,
capital adequacy.
- viii -
- ix -
CHAPTER 1
INTRODUCTION
This chapter provides an overview of the study by starting with the background and
the motivation to conduct this kind of research in Vietnam. After that, the research
objectives, questions, methodology, scope and limitation are specified together with
its significance. Finally, the general framework of the study is proposed at the end of
this chapter.
1.1. Background
Exchange rate is a key element of a country and is managed under each
country’s particular context at a specific period. Exchange rate policy is always
considered as one of the crucial monetary policies, and has unpredictable effect on
other different macroeconomic goals. For example, in an open economy, real
exchange rate affects relative prices between domestic goods and importing products.
The changes in prices will lead to the changes in product demand, and then affect the
country’s inflation. Foreign exchange is also one of the most important indicators of a
country’s state of economy along with budget deficit, trade deficit, inflation, and
interest rate, etc. (Atif et al., 2012). It affects significantly level of investment and
trade activities in the economy, which are critical factors for every country. Therefore,
exchange rate is among the most observed, analyzed, and governmentally controlled
economic variables (Van Bergen, 2010).
Since Vietnam is in the process of integrating its economy into the global
economic activities, the foreign exchange will also be suffered the impact from many
factors in both positive and negative ways. For example, increasing interest rate can
attract more capital from abroad, which in turn will appreciate the Vietnam Dong
-1-
(VND) due to the higher interests paid to VND deposits. However, because of higher
cost of investment capital and debt, the increase in interest rate quite often depresses
the stock market and the growth prospective of the economy, which makes the
currency depreciated. Therefore, it is very challenging to evaluate and analyze the
influence of macroeconomic factors on the exchange rate’s fluctuation that policy
makers really care about. Due to the complicated and diversity of implementing the
exchange rate policy, this becomes a problem which is discussed frequently in
Vietnam, especially since the uncertainty of the macro economy in 2008. Managing
the exchange rate risk needs to be done suitably in order to control the inflation,
stabilize the economy, stimulate export and improve trade balance, which creates
good business environment for domestic as well as foreign enterprises.
1.2. Research Proble ms
The exchange rate in Vietnam has seen significantly volatility in recent years,
especially since the financial crisis in 2008. According to UNDP (2013), from 2007 to
the first quarter of 2008, the supply of USD had increased considerably, which
enriched the foreign exchange reserves of Vietnam and reduced the exchange rate.
However, due to high inflation and the global economic recession, the exchange rate
went up significantly during the last two quarter of 2008. The official exchange rate
VND/USD increased 5.6% at the end of 2009 compared to the same period in 2008.
The unofficial market of VND/USD continued to increase to such a degree that the
State Bank of Vietnam (SBV) had to devaluate VND of 3.3% in 2010. Moreover,
SBV announced the highest amount of VND devaluation of 9.3% after a long time
holding its value at the beginning of 2011. At the same time, it imposed the ceiling
interest rate for foreign currency deposit, and tightened the operations on the
unofficial market and the gold market. Due to the tightening operations, from 2012 to
-2-
2013, the foreign exchange rate became more stable than previous years, with a fixed
fluctuation band of 1% as SBV’s commitment.
The exchange rate fluctuations can certainly cause a great impact on firm value,
especially firms with foreign operations, foreign revenues or foreign denominated
debt(s), because the movement in exchange rate may affect a company’s future cash
flows, revenues, risk management and investment projects. Many studies have
examined the relationship between exchange rate volatility and firm value in different
markets in the world. For example, Choi and Prasad (1995) find that the stock returns
of US firms are significantly influenced by exchange rate movements. It also exists a
relationship between exchange rate and stock prices in India, Korea, and Pakistan in
either the short-run or the long-run (Abdalla and Murinde, 1997).
In Vietnam, there is very little research, which can fully describe the foreign
exchange exposure of Vietnamese companies. For instance, Huynh and Nguyen (2013)
collected monthly data from 2007 to 2012 to study the relationship between exchange
rate, interest rate and stock prices. Although there were little statistically significant in
correlation, the result still indicates that stock prices are negatively impacted by the
exchange rate volatility. However, the influence of the exchange rate movement o n
the Vietnamese firms is not remarkable. During a long period of time, although
Vietnam had registered a regime of managed floating exchange rate to the
International Montary Fund (IMF), in fact, it pursued pegged exchange rate regime
with narrow fluctuation band (Nguyen Thi Thu Hang et al., 2010). Therefore, most of
the domestic companies do not pay much attention to the exchange rate risk as well as
its hedging methods. In case of strong volatility of the exchange rate, the firms with
borrowings in USD but their revenue in VND tend to be more dangerous than the
other ones due to their responsibility for paying debts.
-3-
Nowadays, global economic integration puts pressure on all countries to change
their exchange rate policies so that their currencies can be flexibly converted in either
trade activities or investment. Since Vietnam has already joined in the World Trade
Organisation (WTO) for several years, it has to accept a more flexible exchange rate
regime, which means more unstability. Moreover, the larger the amount of portfolio
investment capital pouring into the domestic market is, the more severe the implicit
risks will be that could cause to financial crisis to the country. One more thing that
needs to be considered, according to the Circular 03/2012/TT-NHNN and Circular
29/2013/TT-NHNN of the SBV about borrowing in foreign currency, companies that
do not have any sources of foreign currency collection can only borrow in VND; and
if they pay for their imported goods, they must buy USD from banks. In addition to
high interest rate, the biggest risk that the borrowers have to face is the fluctuation of
the exchange rate, especially when USD appreciates. To be specific, the firms are
forced to pay the principal and the interest in VND plus the difference caused by the
higher exchange rate at the payment time compared to the one at the time of signing
contract.
If the exchange rate is guaranteed to be fluctuated withtin a trading band by the
central bank as Vietnam has been implementing, the firms might become subjective
and dependent on the government, which are also exposed to devaluation risks, and
then do not use any risk hedging instruments. As a result, if there are any economic
shocks happened in the market, these companies would likely be panic and overreact
since they do not know exactly how to deal with those circumstances. In other words,
there is a chance that Vietnam would change into the more flexible exchange rate
regime, probably with wider trading band; or loosening the restrictions on borrowing
in a foreign currency. At that time, the domestic companies have to struggle with
-4-
many adverse shocks beyond their withstanding while they do not prepare for any risk
controlling strategies. Thus, they need an effective risk – estimated measures to meet
capital requirements in order to cope with the foreign exchange exposure.
Additionally, stress testing is an important risk management tool used broadly
in the worldwide financial systems. It is also required as part of banks’ internal
operating analysis model under Basel II and III requirements. However, because of its
not – too – complicated techniques, stress testing could be considered to apply to
non–financial enterprises so that they can protect themselves against severe economic
situations. Particlarly, in the age of globalisation in all aspects, although the foreign
exchange risk has either direct impact on the firm’s revenue or indirect influence
through the suppliers and customers, the awareness of market risk administration of
Vietnamese companies is still low (Vu Minh, 2013). Hence, stress testing becomes
more crucial than ever before.
In general, as high volatility and sudden changes in the exchange rate is one of
the barriers for a successful macroeconomic policy as well as the companies’ business
performance, modeling volatility is a controversial research topic that needs to be
solved. This research strives to examine the sensitivity of the exchange rate to
macroeconomic factors and develop a macro stress testing framework for the foreign
exchange exposure of Vietnam domestic enterprises.
1.3. Research Objectives
General objectives
This study is to clarify the relationship between macroeconomic variables and
the exchange rate (EXR), and then conduct stress testing using value-at-risk (VaR)
approaches for the exchange rate exposure to Vietnamese firms.
-5-
Specific objectives
To review theories and studies about EXR and describe its policy in Vietnam.
To identify the key determinants as well as analyze their significance to EXR.
To apply stress testing combined with VaR model to Vietnam corporate sector.
To make conclusions and propose recommendations for Vietnam companies
about risk management and capital adequacy.
1.4. Research Questions
Based on the above objectives, this study is designed to answer some questions:
-
What are the significant macroeconomic determinants of the exchange rate?
-
How is the exchange rate expected to fluctuate in the next quarter under VaR
approaches and macro stress testing?
-
What is the expected loss of a companny under normal conditions as well as
adverse scenarios?
1.5. Research Methodology
Both qualitative and quantitative methods are used in this research. The
qualitative measure with visual results like tables and graphs is utilized to assess
preliminarily the connection between the exchange rate and its macro indicators. Then,
the quantitative one is applied to run regression model, statistical tests and calculate
VaR in normal as well as stressed markets.
1.6. Research Scope and Limitation
This study focuses on analyzing the fluctuation and macro determinants of the
exchange rate of VND/USD from the first quarter of 2005 to the first quarter of 2014.
It assumes an foreign exchange position of a hypothetical company in order to employ
the stress testing technique and calculate VaR by using three different approaches.
-6-
In terms of the limitation, firstly, the time period of this study is rather short,
just cover 37 quarters that may not demonstrate fully the comprehensive relationship
between the exchange rate and its determinants. Secondly, the research only uses
secondary data which could contain errors when collecting and calculating. Thirdly,
since it is difficult to collect quarterly data during the study period, some variables
like government expenditure and money supply are not included in the regression
model as macroeconomic determinants of the exchange rate. Finally, for its simplicity,
this study makes a assumption about a hypothetical firm’s foreign exchange position
used to conduct stress testing. Data and results do not represent any specific company,
but the tools and methodologies can be applied to any firm for managing exchange
rate risk in Vietnam.
1.7. Implications
This research is strived to identify the key macroeconomic indicators of the
exchange rate of VND/USD. The stress testing technique is also applied to the foreign
exchange risk of the non- financial companies in Vietnam so that we can estimate the
minimum capital reserve which in turn will help them survive over the unexpected
economic shocks. A clear understanding of the capital adequacy for currency
exposure could provide a foundation for the companies to manage their reserves
efficiently. Hopefully, this study can create a suitable framework with significant
variables and assumptions, which can be made reference by future research.
1.8. Research Structure
The first chapter of Introduction has described the overview of this study about
the motivation, the main purposes, the scope and limitation, etc. The content of the
remaining chapters are presented as follows.
-7-
Chapter 2 – Literature review: This chapter discusses some relevant theories
about the exchange rate, different kinds of regimes and approaches to exchange rate
determination. Furthermore, several empirical studies of the exchange rate’s macro
determinants as well as the stress testing theories are revised.
Chapter 3 – Overview of exchange rate in Vietnam: There are four main
sections included in the chapter that are the regimes of Vietnam exchange rate over
time, the development of the exchange rate po licies and its comparison to several
Asian countries, and the qualitative measurement of the relationship between the
exchange rate and its determinants.
Chapter 4 – Data collection and research methodology: The quantitative
approaches are conducted in this chapter, including the regression model to examine
the relationship between the exchange rate and its independent variables, the steps for
VaR calculation as well as macro stress testing to estimate the exchange rate variation
and the exposure.
Chapter 5 – Data analysis: Chapter 5 illustrates the empirical results of the
regression model, the value of VaR computed by using three different approaches
including variance – covariance method, historical simulation and Monte Carlo
simulation, and the application of the macro stress testing in order to measure the
abnormal loss.
Chapter 6 – Conclusion and recommendation: This chapter summarizes the
findings of this study which meet the initial objectives and answer the three research
questions. It also makes some recommendations for Vietnam companies in building
risk management tools.
-8-
CHAPTER 2
LITERATURE REVIEW
First, in this chapter, the relevant theories on exchange rate, categories of exchange
rate regimes and approaches to exchange rate determination are discussed. Next,
several studies to detemine the macroeconomic indicators of exchange rate are
reviewed, and the stress testing theory is assessed. Finally, the theoretical framework
is built which all other parts of the study must be followed.
2.1. Exchange Rate
Exchange rate is an important macroeconomic variable, which is used as a
parameter for determining international competitiveness of any currency o f any
country (Danmola, 2013). Many scholars explain the concept of exchange rate in
different dimensions. According to Hache (1983), exchange rates are relative prices of
national currencies, and under a floating rate regime they may naturally be determined
by the interplay of supply and demand in foreign exchange markets. More simply,
Kalra (2005) considers exchange rate as a national currency’s quotatio n in respect to
foreign ones. Similarly, foreign exchange rate is described as the price of one
currency expressed in terms of another currency by Eiteman et al. (2010).
It is essential to distinguish between two kinds of exchange rate. According to
Danmola (2013), nominal exchange rate (NER) is a monetary concept which
measures the relative price of the two currencies, while real exchange rate (RER) is
regarded as real concept that measures the relative price of two tradable goods
(exports and imports) in relation to non – tradable goods (domestic trade only). In
other ways, Madura (2006) simply defines RER as the actual exchange rate adjusted
for inflation effects in the two countries concerned. More clearly, RER is clarified as
-9-
the nominal exchange rate adjusted by the ratio of the foreign price level (P*) to
domestic price level (P) in the long run as the following formula.
If changes in the exchange rate just offset the differential inflation rates, the RER
index would stay at 100. If the exchange rate strengthened more than is justified by
the differential inflation, its index would rise above 100, and then the currency is
considered “overvalued” from a competitive perspective. On the other hand, the index
value below 100 would suggest an “undervalued” currency (Eiteman et al., 2010).
Additionally, the tendency for an exchange rate to fluctuate is called foreign
exchange volatility, which implies that while a spot exchange rate is observable, the
future exchange rate for any currency will not be known ahead of time (O’Brien,
2006). Furthermore, the risk that future exchange rate uncertainty poses to companies,
which conduct international business and even those that do not, is decided by both
foreign exchange volatility and the companies’ foreign exchange exposure, which is
the sensitivity of their financial results to foreign exchange rate changes. The level
and frequency of the exchange rate volatility is influenced by a country’s exc hange
rate regime discussed in the next part.
2.2. Exchange Rate Regimes
An exchange rate regime is the way an authority manages its currency in
relation to other currencies and the foreign exchange market (Wikipedia). A country
may choose its currency regime relied on its national priorities of the economy such
as economic growth, inflation, interest rate level, trade balance, etc. Hence, the choice
between fixed and flexible exchange rates may change over time as the priorities
change. The International Monetary Fund (IMF) classifies all exchange rate regimes
- 10 -
into eight specific categories, which span the spectrum of exchange rate regimes from
rigidly fixed to independently floating (Eiteman et al., 2010, p.56-57).
1. Exchange arrangements with no separate legal tender: The currency of
another country circulates as the sole legal tender or the member belongs to a
monetary or currency union in which the same legal tender is s hared by the
members of union.
2.
Currency board arrangement: A monetary regime based on an implicit
legislative commitment to exchange domestic currency for a specified foreign
currency at a fixed exchange rate, combined with restrictions on the issuing
authority to ensure fulfillment of its legal obligation.
3. Other conventional fixed peg arrangements: The country pegs its currency
(formally or de facto) at a fixed rate to a major currency or a basket of
currencies (a composite), where the exchange rate fluctuates within a narrow
margin or at most ± 1% around a central rate.
4. Pegged exchange rates within horizontal bands: The value of the currency is
maintained within margins of fluctuation around a formal or de facto fixed peg
that are wider than most ± 1% around a central rate.
5. Crawling pegs: The currency is adjusted periodically in small amounts at a
fixed preannounced rate or in response to changes in selective quantitative
indicators.
6. Exchange rates within crawling pegs (crawling bands): The currency is
maintained within certain fluctuation margins around a central rate that is
adjusted periodically at a fixed preannounced rate or in response to changes in
selective quantitative indicators.
- 11 -
7. Managed floating with no preannounced path for the exchange rate: The
monetary authority influences the movements of the exchange rate through
active intervention in the foreign exchange market without specifying, or
precommitting to, a preannounced path for the exchange rate.
8. Independent floating: The exchange rate is market – determined, with any
foreign exchange intervention aimed at moderating the rate of change and
preventing undue fluctuations in the exchange rate, rather than establishing a
level for it.
Vietnam has officially accepted the “managed floating” regime, while it is
claimed that its exchange rate regime is actually a type of “crawling peg” (Nguyen
Tran Phuc et al., 2009; Nguyen Thi Thu Hang et al., 2010). Besides, the exchange rate
fluctuation is explained in several theoretical approaches presented in the next section.
2.3. Approaches to Exchange Rate Determination
An objective of theoretical approaches of exchange rate determination is to
provide a clear understanding of the economic mechanisms governing the actual
behavior of exchange rate in the real world and of the relationships between exchange
rate and other important economic variables (Mussa, 1984). This study just presents
three main approaches in determining the exchange rate including purchasing power
parity, balance of payment, and asset market approach.
2.3.1. Purchasing Power Parity Approach
Purchasing Power Parity (PPP) is the most widely accepted theory of exchange
rate determination. This concept has been commonly used to measure the equilibrium
values of currencies and is also a relationship which underpins other exchange rate
approaches (MacDonald, 2007). There are two kinds of PPP:
- 12 -
(1) Absolute PPP indicates that the equilibrium exchange rate is equal to the
ratio of the price levels in the two countries. The shortcomings of absolute PPP are
that not all products made by a country can be traded internationally, the existence of
transportation costs and trade barriers, and the internatio nal capital flow is also
influence on the exchange rate.
(2) Relative PPP holds that the relative change in prices between two countries
leads to the proportional change in exchange rate in a certain period of time. To be
specific, as Eiteman explains, “if the spot exchange rate between two countries starts
in equilibrium, any change in the differential rate of inflation between them tends to
be offset over the long run by an equal but opposite change in the spot exchange rate”
(Eiteman et al., 2010, p.167).
Many scholars have tested PPP and they conclude that (i) PPP performs well in
the long run but poorly in the short or medium term, and (ii) PPP holds better for
countries with high inflation rate and underdeveloped capital markets. For example, if
a country experiences inflation rate relative higher than those of its most trading
partners, its exports become less competitive with overseas products, and its imports
from abroad become more price – competitive with domestic ones. These price
changes lead to a deficit on current account that causes to the variation of exchange
rate unless it is offset by capital and financial inflows (Eiteman et al., 2010).
2.3.2. Balance of Payments Approach
The PPP approach incorporates implicitly through trade, demand and supply
factors in the determination of the exchange rate. For example, if price in a foreign
country is lower than the domestic one, then domestic demand for foreign goods will
increase, and the foreign currency will appreciate. The balance of payments (BOP)
approach can be seen as encompassing the PPP (Rauli Susmel, undated).
- 13 -
The BOP, which consists of all financial flows of a country during a given
period, is an important indicator of pressure o n a country’s foreign exchange rate.
This approach involves the supply and demand for foreign exchange as determined by
the flows of currency from international transactions. This approach affirms that the
equilibrium exchange rate is established by matching the net inflow (outflow) of
foreign exchange arising from current account activities and the net outflow (inflow)
of foreign exchange arising from financial account activities (Eiteman et al., 2010). In
other words, the flows of currency related to the BOP comprise trade in goods and
services, portfolio investment, direct investment, etc. Equilibrium exchange rate is set
when the BOP is in equilibrium.
According to this approach, a current account deficit (surplus) tends to lead to a
depreciation (appreciation) of the exchange rate. However, as foreigners might be
willing to finance the current account imbalance by lending or borrowing, which in
turn produces a capital account surplus (deficit), the exchange rate depreciation
(appreciation) might not occur in the short run (Rauli Susmel, undated). Thus, the
BOP approach only has more precise predictions in the long run because the current
account imbalance can not be financed forever. The long-run exchange rate would
move to balance the current account. Another limitation of this approach is that it just
focuses on flows of currency and capital rather than stocks of money or financial
assets (Eiteman et al., 2010). That means relative stocks of money or financial assets
has no contribution in determining exchange rate in this theory.
2.3.3. Asset Market Approach
Rather than the traditional view of the exchange rate adjusting to equilibrate
international trade in the BOP, an asset market model emphasizes on financial – asset
markets. This approach states that the exchange rate is determined by the supply and
- 14 -
demand for a wide variety of financial assets. Since prices of physical goods adjust
slowly compared to prices of financial assets which are traded more frequently every
business day, the shifts in the supply and demand for different countries’ financial
assets will be likely to change the exchange rate. Moreover, changes in monetary and
fiscal policy alter expected returns and perceived relative risks of financial assets,
which in turn alter the exchange rate (Eiteman et al., 2010).
The asset market approach also assumes that whether foreigners are willing to
hold domestic assets depends on a set of investment considerations including relative
real interest rate, prospects of economic growth and profitability, capital market
liquidity, economic and social infrastructure, political safety, corporate governance
credibility, contagion, and speculation. Thus, the approach can be used to forecast
future spot exchange rate. There are two basic groups within this model including the
monetary approach and the portfolio – balance approach.
Monetary Approach
Under this approach, the exchange rate is determined by the supply and demand
for national monetary stocks, as well as the expected future levels and growth rates of
monetary stocks (Eiteman et al., 2010). Other financial assets like bonds are not
considered relevant for exchange rate determination. According to Chinn (2013), the
monetary model provides two key implications that are (1) higher relative income
leads to a higher money demand relative to supply, and hence a stronger currency, and
(2) higher relative interest rate causes a lower money demand against money supply
which induces a weaker currency. The approach includes two versions: (i) flexible –
price monetary approach by Frenkel (1976); (ii) sticky – price monetary approach by
Dornbrusch (1976) and Frankel (1979). PPP holds in both the long run and short run
in the former approach, while it only holds in the long run in the latter one.
- 15 -
Portfolio Approach
This approach claims exchange rate affected by different portfolios of financial
assets. It assumes that assets denominated in different currencies do not substitute
perfectly, which is the main difference from the monetary approach. The imperfect
substitution of international assets is due to foreign exchange risk perceived by
investors. According to Crowder (undated), the portfolio approach has risk premiums
in the forward exchange rate that are a function of relative assets supplies. As the
supply of domestic bonds rises relative to foreign bonds, there will be an increased
risk premium on the domestic bonds that will cause the domestic currency to
depreciate in the spot market. If the spot exchange rate depreciates today, and if the
expected future spot rate is unchanged, the expected rate of appreciation (depreciation)
over the future increases (decreases).
These above approaches of exchange rate determination has been tested in many
studies to find out what are the most important factors affecting a country’s exchange
rate. The next section discusses the exchange rate’s determinants in previous research.
2.4. Determinants of Exchange Rate
According to Kanamori (2006), an exchange rate of a country is determined by
macroeconomic factors, speculative factors, and economic expectations. In other
words, factors affecting exchange rate can be economic, political, and psychological
as well as in the short run or long run (Saeed et al., 2012). A study of a cross – section
of 81 countries carried out by Canales-Kriljenko & Habermeier (2004) observes that
high inflation and fiscal deficit have a significant correlation with higher volatility,
while foreign exchange reserves of a country and current account deficit appeared to
be insignificant. This study follows the categorisation by countries studied in order to
- 16 -
evaluate the sensitivity of structural differences among the country groupings. The
categories include developed and developing countries.
2.4.1. Empirical Literature from Developed Countries
McMillin and Koray (1990) examine the effects of the market value of both US
and Canadian government debt on the real CAD/USD exchange rate. The model
indicates significant effects of debt on the exchange rate. To be specific, debt shocks
lead to a short – lived depreciation of the US Dollar rather than to an appreciation.
MacDonald (1998) builds a model that features productivity differentials, terms
of trade, fiscal balances, net foreign assets and real interest rate differentials as key
determinants of the real exchange rate of the US Dollar, Yen and the Deutschmark.
He finds that all the variables have a positive relationship with the real exchange rate
in both the short run and long run.
Wilson (2009) examines the effective exchange rate of US Dollar based on the
weighted average trading partner of USA. Money supply is positively related to the
effective exchange rate, which means increase in money causes decline in the value of
currency. In contrast, interest rate, government expenditure and defic it to GDP are
negatively related with the effective exchange rate.
Atif et al. (2012) demonstrates the relationships between Australian exchange
rate and its economic and non-economic determinants. It is suggested that Australia’s
trade components and macroeconomic indicators like net exports, GDP and money
supply play a significant role in determining its exchange rates. However, interest rate
and inflation appear insignificant in this relationship.
2.4.2. Empirical Literature from Developing Countries
The exchange rate’s determinants are also focused on in developing nations.
- 17 -
Siddiqui et al. (1996) estimates the determinants of real exchange rate for
Pakistan. The result is that increase in government expenditures leads to depreciation
in real exchange rate, as well as excess domestic credit creation and openness
significantly contributes to real exchange rate appreciation. Similar to the previous
research, Hyder and Mehbood (2005) find that increase of governmental expenditure
brings depreciation in the real exchange rate. However, they identify that degree of
openness and capital account balance also causes the real exchange rate’s depreciation.
Moreover, Saeed et al. (2012) undertakes an analysis of determinants of exchange rate
between US Dollar and Pakistani Rupee under the monetary approach from 1982 to
2010. The results confirm that the relative terms of stock of money, total debt, and
foreign exchange reserves are significant factors of PKR/USD exchange rate.
Aron et al. (2000) investigate the short-run and long-run determinants of the real
effective exchange rate for South Africa. They find that an appreciation of the real
exchange rate is led by an increase in terms of trade, price of gold, tariffs, capital
inflows, official reserves, government share in GDP, technological progress and
domestic credit, while a depreciation of the real exchange rate is produced by an
increase in openness and nominal depreciation. Furthermore, MacDonald and Ricci
(2003) estimate the equilibrium real exchange rate for South Africa using the cointegration estimation procedure from 1970 to 2002. The result shows that the longrun real exchange rate is explained by real interest rate differentials, relative GDP per
capita (productivity), real commodity prices (terms of trade), trade openness, fiscal
balance and net foreign assets. The behavior of South Africa Rand against US Dollar
and Euro is also analyzed by Egert (2010) from 2001 to 2007. The empirical results
show that changes in gold prices and innovations in exchange ra te of Dollar and Euro
are two of four factors affecting South Africa’s exchange rate.
- 18 -
Takaendesa (2006) also analyzes the determinants of real exchange rate of
South Africa in the period of 1975 to 2005, and finds that the terms of trade, real
interest rate differential, domestic credit, openness and technological progress have a
long-run relationship with the real exchange rate. Among other determinants, the
terms of trade explain the largest proportion of the exchange rate’s variation. The real
exchange rate fluctuations are primarily equilibrium responses to monetary shocks
rather than fiscal policy shocks.
Bahmani and Kara (2000) examine the exchange rate of Turkish Lira against
US Dollar using monthly data from 1987 to 1998. They find that the negative sign of
change in real income indicates the relative growth in the real income in Turkey
relative to USA appreciates Lira. Interest rate differential and inflation differential
have positive signs and are statistically significant.
Zhang (2001) adopts equilibrium real exchange rate approach and co-integration
techniques to identify that investment, government consumptions, growth rate of
exports, and degree of openness of the economy to trade are main explanatory factors
for the China’s RMB’s long-term equilibrium path from the mid-1950s to the mid1990s.
Karfakis (2003) tests the monetary model for Romanian Lei and US Dollar
exchange rate and concludes that increase in money is the source of depreciation in
the domestic currency. Real income is negatively related with the value of currency,
while inflation has positive connection with the value of Lei against Dollar. Moreover,
Nucu (2011) explores the relationship between exchange rate of EUR/RON and key
macroeconomic indicators including GDP, inflation rate, money supply, interest rate,
and balance of payments for Romania. The outcome is that GDP has a negative
relationship with EUR/RON, whereas inflation and interest rate relate positively with
- 19 -
the exchange rate. Both money supply and balance of payments are statistically
insignificant.
Joyce and Kamas (2003) investigate the factors determining the real exchange
rate in Argentina, Colombia and Mexico. Its co- integration analysis demonstrates that
the exchange rate has an equilibrium relationship with terms o f trade, capital flows,
productivity and government share of GDP. An increase in all these variables
appreciates the exchange rate.
Karim et al. (2007) uses quarterly data of macroeconomic variables for New
Zealand and its trading partners including Australia, Japan, and USA to demonstrate
the result that implementation of tight monetary policy causes both nominal and
effective exchange rate to appreciate.
Hsieh (2009) studies the behavior of Indonesian Rupiah per unit of US Dollar.
The result shows that a relatively higher domestic interest rate, or a relatively more
expected inflation rate causes real depreciation for Indonesia Rupiah. Conversely,
higher ratio of governmental spending to GDP, or higher stock prices leads to real
appreciation in IDR/USD exchange rate.
Sinha and Kohli (2013) study the influence of some economic factors on the
exchange rate of Indian Rupee against US Dollar over the period of 1990 to 2011.
They find that economic variables like inflation differential, lending interest rates, and
current account deficit (as a percentage of GDP) have significant effect on the
USD/INR exchange rate.
2.4.3. Conclusion
Overall, many studies about the relationship between the exchange rate and its
macro determinants in both developed and developing countries have shown many
- 20 -
different empirical results regarding to the way of influence of each factors to the
exchange rate. Nevertheless, it can be seen that there are several main indicators that
usually have impact on exchange rate including productivity, interest rate differential,
inflation differential, terms of trade, capital inflows, government expenditure, current
account, technological progress, official foreign reserves, money supply and degree of
openness. Thus, GDP, interest rate differential, inflation differential, current account,
foreign direct investment (FDI) and foreign exchange reserves are chosen as the key
determinants of VND/USD.
However, this study adds two additional indicators that are the exchange rate of
VND/EUR and VND/JPY to the list to test their correlation with VND/USD. If the
number of transactions using EUR and JPY increases, Vietnam would have a surplus
supply of EUR and JPY that causes to the reduction in the price of VND/EUR and
VND/JPY. But since VND is a controlled and non- freely tradeable currency, other
currencies such as EUR and JPY are converted to VND via USD peg. Thus, these two
exchange rates are considered to be highly correlated with VND/USD. Due to this
high correlation, later we will see in the multicollinearity test (section 5.2.4), EUR
and JPY will be dropped out of our regresstion model. Note that the regression on the
above macroeconomic factors is an intermediate step of our analysis, and we will be
using the resultant regression model in our Value-at-Risk (VaR) model and Stress
Testing analysis.
2.5. Stress Testing Theory
The second stage of this study is to conduct stress testing for Vietnamese firms.
Thus, the subsequent sections will introduce about the concept of stress testing
including its definition, conducting framework as well as two main approaches.
- 21 -
2.5.1. Definition of Stress Testing
Stress testing is one of the important tools for effective risk management and
macro prudential oversight. IAA (2013) clarifies a stress test as a projection of the
financial condition of a firm or economy under a specific set of severely adverse
conditions that may be caused by several risk factors over several time periods with
severe consequences that can extend over either months or years as presented in
Figure 2.1. On the other hand, a scenario is a possible future environment which
involves changes and interactions among many risk factors over time. A scenario with
significant or unexpected adverse consequences is considered as a stress scenario
(IAA, 2013). That means the probability of the scenario underlying a stress test has
been referred to as extreme but plausible..
Figure 2.1: Stress testing and scenario
Source: Stress Testing and Scenario Analysis, IAA (2013)
According to Blaschke et al. (2001), the objective of a stress test is to make
risks more transparent by estimating the potential loss on a portfolio in abnormal
markets as well as to evaluate the strength or stability of institutions. Moreover, stress
- 22 -
tests can be applied for either an individual institution or aggregate portfolios, and
most often used to measure market risk – as in our case.
There are two kinds of stress testing including micro versus macro stress test.
Micro stress testing is conducted by individual institutions as part of their risk
management and often ignores behavior of competitors, whereas macro stress testing
refers to a range of techniques used to assess the vulnerability of a system to
exceptional but plausible macroeconomic shocks (Sorge, 2004).
2.5.2. Framework for Stress Tests
Despite what kind of stress test, a sequence of the different decision elements of
a stress test shown in Figue 2.2 needs to be followed when conducting this test.
The type of risks is specified first, and then the appropriate models are
considered to use. Stress tests can focus on individual risks or encompass multiple
risks. The most widely used ones are: (1) market risk, which consists of four standard
risk factors are interest rate risk, exchange rate risk, equity risk and commodity risk, is
defined as the risk of losses on a portfolio arising from movements in market prices;
(2) credit risk is the risk of loss associated with debtor’s default of a loan or any other
lines of credit principal or interest or both (Chopra, 2009); (3) other forms of risk
includes liquidity risk and operational risk. We will be focusing solely on market risk
(i.e. exchange rate risk).
The next element is the range of factors to include, followed by the specification
of scenarios. Stress tests can involve estimating the impact of a change in a single risk
factor called a sensitivity test, or the effect of a simultaneous move in a group of risk
factors named a scenario analysis. According to Kalfaoglou (2007), the scenario
analysis technique is more demanding in terms of application and requires the use of
- 23 -
sophisticated econometric models. In contrast, the sensitivity analysis is not as
realistic as the former one because there is definitely more than one risk factor that is
affected in times of shock. Thus, it is mostly useful for analysis in the short run. We
will be conducting both sensitivity test and scenario analysis for our study.
Figure 2.2: Frame work for stress testing of individual portfolios
Source: Blaschke et al., IMF Working Paper (2001)
- 24 -
Thirdly, the parameters to be shocked are decided. The type of shocks are
designed to integrate both movements in individual market variables such as prices,
interest rates, etc. and changes in the underlying relationship between different asset
markets represented by their respective volatilities and correlations. We will be
shocking the changes in the statistically significant macro deteminants.
Fourthly, the type of scenario used to conduct the stress test is critical to the
analysis. Stress testing can also be based on historical scenarios, which employ shocks
that occurred in the past as a benchmark for future analysis, or based on hypothetical
scenarios that are considered as plausible changes in circumstances that have no
historical precedent. Moreover, Monte Carlo simulations use techniques to look
jointly at the sensitivities and probability distributions of various input variables
(Chopra, 2009). We will be setting some hypothetical shocks for market parameters
and running Monte Carlo simulation as well.
Finally, the core assets are to be shocked with what the relevant risks are, how
much to stress them and over what time period. Once specified, the scenarios are
applied to the portfolio to determine the potential change in the present value
(Blaschke et al., 2001).
2.5.3. Main Approaches to Macro Stress Testing
Macro stress testing is used to quantify the link between macroeconomic
variables and the health of either a single institution or the business secto r as a whole.
According to Sorge (2004), there are two main methodological approaches to macro
stress testing including the piecewise approach and the integrated approach.
A piecewise approach that evaluates the vulnerability of the corporate sector
to single risk factors by forecasting several financial soundness indicators, such as
- 25 -
non-performing loans, capital ratios and exposure to exchange rate or interest rate
risks, under various macroeconomic stress scenarios. A direct economic relationship
is estimated using historical data between the macroeconomic variables and the
various risk measures. This approach is commonly used because of its ease in
implementation. However, it just focuses on the linear relationships between
corporate risk and macro fundamentals as well as lacks an ability to characterize the
entire loss distribution (Chopra, 2009).
An integrated approach combines the analysis of the sensitivity of a system
to multiple risk factors into a single estimate of the probability distribution of
aggregate losses that could happen under any given stress scenarios. Value at risk
(VaR), which is a method based on the probability of deviation from anticipated profit
(Yildirim, 2012), is the most popularly used summary statistic of this distribution. The
VaR approach allows the risk parameters to be state or time dependent which helps
address concerns of parameter instability. It also permits non-linear relationships
between macroeconomic shocks and risk measures. Nevertheless, the non-additive
across portfolios is a problem of the VaR approach. As a result, such studies focus on
an aggregate portfolio cannot be likely to taken into account due to the domino effects
among individual institutions (Sorge, 2004).
2.6. Value At Risk (VaR) Model
Stress testing is designed to explore the tails of the distribution of losses beyond
the threshold used in the VaR analysis. According to Blaschke et al. (2001), VaR of a
portfolio is a statistical measure that summarizes the maximum expected loss that the
portfolio most probable to suffer over a specified period of time for a given level of
confidence. VaR can be used by any entity to measure its potential loss in value of its
traded portfolios from adverse market movements, which can then be compared to its
- 26 -
available capital and cash reserves to ensure that the losses can be absorbed without
putting the firm at risk (Damodaran, undated). The following figure shows the curve
of a hypothetical profit and loss probability density function with 5% VaR.
Figure 2.3: VaR at 95% confidence level
Source: Wikipedia
There are three key elements of VaR including a specified level of loss in value,
a fixed time period over which risk is assessed and a confidence interval. In VaR
estimation, a confidence level and size of shocks need to be determined in advance.
The level of confidence is decided by the risk appetite of the firms and the amount of
its economic capital, which is the amount of risk capital that a company estimates in
order to remain solvent at a given confidence level and time horizon (Yildirim, 2012).
If the objective of estimating VaR is to forecast capital adequacy, the confidence level
should be high enough so that there is very little probabiltiy of failure (Gupta and
Liang, 2004). Bank for International Settlement (BIS), for example, suggests that
market risk computations under VaR models should use 10-day holding period and 99%
confidence level in their Basel II and Basel III regulations.
- 27 -
Three basic approaches commonly used to compute VaR comprises variance –
covariance method, historical simulation and Monte Carlo simulation.
Variance – covariance approach: This approach is based on the assumption
that the underlying market factors have multivariate normal distribution, and hence,
the distribution of mark – to – market portfolio profits and losses could be determined
to be normal. The potential loss or VaR is calculated from the volatility of the market
factors, which can be estimated from the historical market data for the respective ones
as well as their correlations. Consequently, VaR can be expressed as a function of the
standard deviations of market returns and the covariance between them (Cassidy and
Gizycki, 1997).
Historical simulation: The aim of this simulation is to determine what
profits or losses would be incurred if a market price development from the past were
to occur today (SAP Website). It is a simple approach since it requires no assumption
about the statistical distributions of the underlying indicators. Instead, this method
involves using historical changes in market rates and prices to construct a distribution
of potential portfolio profits and losses, and then come up with VaR as the loss that is
exceeded a particular significance level (Linsmeier and Pearson, 1999).
Monte Carlo simulation: It is the most widely used method among these
three approaches. The simulation of returns scenarios is based on the generation of
independent and identically random variables (Mina and Xiao, 2001). In other words,
a statistical distribution that effectively captures the possible changes in the market
factors could be chosen, and a random number of hypothetical changes in these
factors are generated in order to set up the distribution of possible portfolio profits and
losses (Linsmeier and Pearson, 1999).
- 28 -
We will cover all three methods above in this thesis. In general, although these
approaches for measuring VaR have the same limitation due to the fundamental
assumption that is the future risk could be estimated from the historical distribution of
returns, they also possess their own strengths and weaknesses which are summarised
in Table 2.1 as below.
Table 2.1: The strengths and weaknesses of VaR approaches
Approach
Variance – covariance
Advantages
Disadvantages
• No need for extensive
historical data, only volatility
and correlation matrix are
required
• Less accurate for nonlinear
portfolios or for skewed
distributions
• Be used only when the
• Fast and simple calculation portfolio is less diversified
Historical simulation
• No
need
to
make • Requires a significant
distributional assumptions
amount of historical data
• Provides a full distribution • Difficult to scale far in the
of potential portfolio values
long horizons
• Faster calculation
Monte Carlo since
scenarios are used
than • Coarse at high confidence
less level (e.g. 99% and beyond)
• Incorporates tail risk only
if historical data set includes
tail events
Monte Carlo simulation • No need for extensive • Computionally intensive
historical data
and time – consuming for
revalue the portfolio under
• Permits use of various
each scenarios
distributional assumptions
• Provides a full distribution • Quantifies fat – tailed risk
of potential portfolio values only if market scenarios are
from
the
(not just a specific percentile) generated
appropriate distributions
Source: RMG, Risk Management
- 29 -
2.7. Theoretical Frame work
This study includes two main stages of analysis described in the following
Figure 2.4. The first one examines the relationship between exchange rate and its
macro determinants in order to forecast the future volatility of the exchange rate. The
second one is to conduct macro stress testing associated with the VaR model to
estimate the expected loss and determine capital adequacy.
Figure 2.4: Theoretical frame work
- 30 -
CHAPTER 3
OVERVIEW OF EXCHANGE RATE IN VIETNAM
This chapter consists of an introduction to Vietnam exchange rate system, the
development of exchange rate policies over some time periods, the comparison of
these policies to other Asian countries, and the qualitative analysis about the
relationship between the exchange rate and its factors in recent years.
3.1. Vietnam Exchange Rate Regimes
In an open economy, the exchange rate policy typically has substantial impact
on the financial stability, trade performance, international competitiveness and the
functioning of the foreign exchange market (Frieden, Ghezzi and Stein, 2001; Ho and
McCauley, 2003). In Vietnam, the purpose of the exchange rate management is to
control the inflation rate, promote the macro economic growth, stabilize the financial
system and improve the trade balance are always an important policy, especially when
Vietnam is still trying to escape from the effect of the global crisis in 2008.
Since putting an end to the system based on administrative subsidies in 1989,
Vietnam has had a lot of adjustment in the exchange rate regime. It has gradually
evolved from a system of multiple exchange rates to a single announced fixed rate,
then to the current system incorporating a narrow adjustable band around the official
rate, which is itself set on a daily basis reflecting the interaction of market forces
(Nguyen Tran Phuc and Nguyen Duc Tho, 2009). However, in its nature, all these
changes turn around to the conventional fixed peg exchange rate in which USD is
considered as a key nominal anchor. The State Bank of Vietnam (SBV) announces the
exchange rate of VND/USD, and then commercia l banks set up the exchange rate
between VND and other foreign currencies relied on the international exchange rate
- 31 -
between USD and those currencies. International as well as domestic observers have
encouraged the Vietnamese authorities to allow greater flexibility in the exchange rate,
as a means to manage external shocks more effectively and preserve competitiveness
on global markets (IMF, 2006).
The exchange rate regimes applied in Vietnam in the period of 1989 to 2013 are
classified based on the IMF criteria in the Table 3.1. During the time with strong
fluctuation of the economic activities, such as the removal of the administrative
subsidies in 1989, the Asian financial crisis in 1997, or the global crisis in 2008, the
SBV had adjusted the fluctuation band as well as the official rate, but only in a certain
limit. After these impacts, the exchange rate came back to the regime of conventional
fixed peg arrangements or crawling peg (Nguyen Thi Thu Hang et al., 2010).
Table 3.1: Exchange rate regimes in Vietnam from 1989 to 2013
Time period
1989 – 1990
Exchange rate regime
Crawling bands
Characteristics
+ Official exchange rate (OER) was
established and adjusted by SBV due to the
specific situation of inflation, interest rate,
BOP, etc.
+ Commercial banks were allowed to quote
within a trading band of ±5% around OER
+ Close control of using foreign currency
1991 – 1993
Pegged exchange rate
with horizontal bands
+ Two foreign exchange trading floors were
opened in Ho Chi Minh and Hanoi
+ Official foreign exchange reserves was set
up to stabilize the exchange rate
+ The trading band of ±0.5% around OER
1994 – 1996
Conventional fixed peg
arrangement
+ The trading floors we replaced by an
interbank foreign exchange market whose
transactions were interfered strongly by
SBV, which plays an important participant
in the market.
- 32 -
+ OER was announced based on the
interbank exchange rate
+ The trading band of ±0.5%, then increased
to ±1% in November, 1996
+ OER was kept stable at 11,100 VND/USD
1997 – 1998
Crawling bands
+ The trading band was expanded from ±1%
to ±10% in 1997, and then reduced to ±7%
in 1998
+ OER was adjusted to 12,998 VND/USD in
August, 1998
+ OER was replaced by the interbank rate
without adjustment of the previous working
day announced by the SBV in 1999
1999 – 2000
Conventional fixed peg
arrangement
+ Commercial banks were again allowed to
decide their own rates within a stipulated
band, which was fixed at +0.1%
+ OER was kept stable at 14,000 VND/USD
+ Commercial banks freely decided interest
rate for USD based on transactions on the
Singapore interbank market (Sibor) in 2001
2001 – 2007
Crawling peg
+ OER increased from 14,000 VND/USD in
2001 to 16,100 VND/USD in 2007
+ The trading band was adjusted three times,
to ±0.25%, ±0.5%, and ±0.75% in 2002,
2006, and 2007 respectively
2008 – 2011
Crawling bands
+ OER increased from 16,100 VND/USD at
the beginning of 2008 to 18,544 VND/USD
in the same period of 2010, and then
devaluated 9.3% in 2011
+ The trading band was adjusted six times
from 2008 to 2011 and remained at ±1%
since February, 2011
+ OER was managed relatively stable
2012 – 2013
Crawling peg
+ The trading band stayed at ±1% during the
years
Source: Updated from Nguyen Thi Thu Hang et al. (2010),
and the exchange rate policies of SBV
- 33 -
Nevertheless, the current exchange rate has been described by the authorities as
a managed floating system. According to Nguyen Tran Phuc et al. (2009), in the case
of Vietnam, the term “managed floating” means that the SBV just simply informs the
average interbank rate based on the market supply and demand of the preceding
business day, and does not set the OER any more, as well as the exchange rate under
the administrative controls can only move within a stipulated band.
Another difference in the Vietnam exchange rate regime is that Vietnam has had
a duo – track system, which has coexisted, the official and unofficial rate. In other
words, although in reality the SBV just applies the OER for all the transactions in the
country, the exchange rate of the parallel black market (free exchange rate – FER)
still exists along with the former one. There is also likely a case of discrimination
against private enterprises in approaching for foreign capital from banks, who do not
have any foreign business activities or are not encouraged to use foreign currency to
import luxury goods as well as those kinds of goods could be produced domestically.
Due to this reason, the black market continues d eveloping in a relative large-scale in
Vietnam.
3.2. The Development of Exchange Rate
The nominal bilateral exchange rate of VND/USD is typically the one that
displays prominently the policy discussion in Vietnam. The official VND/USD tends
to follow a clear cycle including two stages: (1) VND has devaluated strongly during
the crisis or economic regression; (2) VND is pegged rigidly to USD when the
economy is relatively stable (UNDP, 2013). Figure 3.1 shows the annual average
official exchange rate of VND/USD which increased dramatically from 1989 to the
beginning of the 1990s. It remained stable until 1997 when it began to climb again to
around 15,350 VND/USD in 2002. This exchange rate again stayed firmly in the
- 34 -
period of 2003 to 2007, and then went up substantially from 16,977 VND/USD in
2008 to 20,828 in 2012. During this five years, the average devaluating rate of the
VND value is rather high, 5.29% per year, especially the decline of 9.3% in VND
value in 2011. The exchange rate rose steadily with the rate of 1% as the commitment
of the SBV in 2013.
Figure 3.1: Official VND/USD rate, 1989 – 2013
25,000
20,000
15,000
10,000
5,000
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
-
Source: SBV and World Bank (WB)
In general, stage 1 of the cycle included (i) 1989 – 1992 when Vietnam’s
economy was in its comprehensive economic renovation started with the removal of
the administrative subsidies; (ii) 1997 – 2000 with the strong impact of the Asian
financial crisis; and (iii) 2008 – 2011 with the influence of the global economic crisis.
In this stage, the gap between the OER and the FER are at their highest. The pressure
from the market forced the SBV to expand the trading band and raised the OER,
which made VND depreciated considerably. Moreover, stage 2 of the cycle included
(i) 1993 – 1996; (ii) 2001 – 2007; (iii) 2012 – 2013 when the economy remained
relatively stable. The FER stayed calm and followed close behind the OER. This is
- 35 -
because the OER had increased significantly in the preceding periods and was equal
to the FER at the end of these periods. To clearly understand, the movement of the
exchange rate in Vietnam from 2001 to 2013 will be analyzed in the following
sections.
3.2.1. The period of 2001 – 2007
During these years, the exchange rate policy mainly focused on increasing
steadily the interbank rate announced by the SBV on a daily basis. The interest rate
was set on the agreement between the banks and their customers at 8.5% - 9.5% per
year in 2002. The FER increased much higher than the nominal rate that caused to the
strain on the foreign exchange market, which in turn forced many companies to
officially buy USD on the free market. Therefore, the SBV had to expand the trading
band to ±0.25% in July, and give the commercial banks more freedom to action in
order to stabilize the market. The banks were also allowed to carry out currency
options, and the remittance rate applied for the companies was reduced to zero in
2003. Since then, the SBV has gradually changed the peg of VND with USD to a peg
with a currency basket which includes the currencies of its major trade partners with
the weights reflecting the geographical distribution of trade, services and capital flows
(Anh Ngoc Nguyen and Nicholas Sarantis, 2008).
In addition, the SBV had moved the upper limits of forward rate in all terms
before permitting the commercial banks to deal with their customers relied on the
difference between the basic rate of the SBV and the FED’s interest rate of USD on
May, 2004. This was an crucial development of the foreign exchange derivatives
market because it would probably strengthen the investors’ expectations about future
exchange rate fluctuations. On 8th December, 2004, the SBV promulgated the
Decision No.1452/2004/QĐ-NHNN about the adjustment in the foreign exchange
- 36 -
transactions of the financial institutions. According to this Decision, forward, swap
and option transactions could be carried out between financial institutions, economic
entities and individual investors. Transactions without documents to improve the
purpose of foreign currency were allowed, except for those in which foreign currency
is bought from commercial banks through spot and forward contracts. Moreover,
more instruments for preventing risks were utilized on the options market. As a result,
the foreign exchange market was developed more flexibly, and the exchange rate
reflected more effectively the relationship between the supply and demand.
From 2005 to 2006, since the United States declared war against any
organisations or countries related to terrorism and nuclear weapons, USD was
depreciated against major currencies like EUR, JPY, etc. but appreciated against VND
so that the import activities of Vietnam were at a disadvantage and the gold price
increased unusually high. However, the SBV kept the trading band at ±1% that
affected badly to the flexibility in the managment of the exchange rate. In the middle
of the year 2006, the government promulged the Ordinance on Foreign Exchange
which liberalized the interest rate in the options market and repealed the upper limit of
the forward rate. The decisions were in conformity with international practices and
made it more easily for Vietnam in the progress of joining WTO. Furthermore, the
foreign portfolio investment in the stock market rised considerably as well as the
individual investors exchanged USD into VND for their investment brought to the
balance between the bid rate and ask rate of the commercial banks in the last two
quarters of 2006. The SBV decided to widen the trading band from ±0.25% to ±0.5%
followed the Decision No. 2554/QĐ-NHNN in 2006.
From the beginning of 2007, the foreign capital and remittance from overseas
went up significantly along with the difference in interest rate between VND and
- 37 -
USD got bigger. That made the exchange rate of VND/USD stable, the bid and ask
rate of the banks were often at the lowest band. The instruments of the monetary
policy were utilized effectively to control the money supply which limited the
pressure on appreciating the value of VND, such as buying foreign currency on the
open market, selling foreign currency for the commercial banks with negative short
position of 5% instead of 10%, abolishing the ceiling interest rate of USD, and
restricting the credit growth rate, etc. The exchange rate of VND against USD
increased in August, 2007 due to FED and the state banks of other countries enforced
several financial recovery measures as well as foreign investors withdrew their capital
to balance their home financial positions. However, if compared to 2006, the
VND/USD just increased a very small percentage, 0.14%. It can be seen that the
exchange rate in this period has not changed as much as expected, and thus, not
demonstrate correctly the market signal. The next period had far stronger fluctuation
than this one.
3.2.2. The period of 2008 – 2011
Since the end of the year 2007, the exchange rate had strong and fast variations.
The phenomenon of the surplus of USD has happened again. The short-term capital
came into Vietnam surged in the first three months of 2008 which led to the banks’
exchange rate (BER) reduced to the lower limit of the trading band. However, the
financial crisis had made the world economy as well as Vietnam economy more
complicated. The inflation and interest rate went up extraordinarily, some times over
21% per year, as well as the reverse direction of the indirect investment. Therefore,
the interbank rate increased 5.36% in 2008 compared to 2007, and the trading band
was adjusted more frequently than ever before, including ±1%, ±2%, and ±3% on
March, June, and November 2008 respectively as shown in Table 3.2. In addition, the
- 38 -
SBV applied the tightened monetary policy, issued compulsory SBV bill of VND
20,300 billion, increased the administrative interest rates, and controlled the credit of
foreign currency closely. Nevertheless, the SBV began carefully loosing the policy,
decreasing the interest rate and required reserves rate, selling a large amount of USD
from the foreign exchange reserves to the market in order to support the liquidity for
the commercial banks in the last quarter of 2008.
Table 3.2: Changes in the trading band around the interbank rate
Time
Trading band
01/07/2002
±0.25%
31/12/2006
±0.5%
24/12/2007
±0.75%
10/03/2008
±1%
27/06/2008
±2%
06/11/2008
±3%
24/03/2009
±5%
01/12/2009
±3%
11/02/2011
±1%
Source: Truong Dinh Tuyen et al. (2011)
Although the exchange rate of 2008 fluctuated constantly, even lower than the
OER during the year, the BER in 2009 usually reached the upper limit of the trading
band. Because of the market pressure which was caused by the maturity of enterprises’
foreign currency loans and import payments, the big gap between the domestic gold
price and international price, the individuals’ psychology of holding foreign currency
as well as speculative behaviours, the SBV had to devaluate strongly VND, which
means increasing the exchange rate considerably from 16,977 in 2008 to 17,941 in
2009, together with adjusting the trading band without giving notice, so that the
- 39 -
market could become more liquid. Furthermore, the SBV raised the basic interest rate
from 7% to 8% per year. The Decision No. 131/QĐ-TTg in 2009 allowed an interest
rate assistance policy of 4% per year which enabled companies to reduce their lending
expenses, and then their products’ prices. These policies were considered to be
reasonable but tardy. The VND continued depreciating on the black market with the
FER of about 19,400 VND/USD at the end of 2009.
Similar to 2009, the BER was set at the upper limit of the trading band, and the
gap between the OER and the FER was increasing incredibly high, especially in the
last few months of 2010. To deal with the intensive market pressure, the SBV
adjusted the exchange rate from 17,941 VND/USD to 18,544 VND/USD, which was
equivalent to devaluation of 3.3% on February, 2010. It also performed series of
administrative measures such as reducing required reserves rate for foreign currency
deposits, extending the objects of foreign currency lending, terminating gold trading
on overseas accounts, etc. As a result, the credit of foreign currency increased 27%,
while the credit of VND just grew 4.6% in the first half of 2010. At the same time, the
overseas remittance as well as the amount of FDI and ODA also rose sharply and that
helped to increase the supply and decrease the demand of foreign currency, which in
turn shortened the gap between OER and FER in the second and third quarter of the
year as in Figure 3.2. Then, the market continued expecting the depreciation of VND
due to several reasons: (i) the increase in the foreign currency supply was mainly
because of the difference between the interest rate of VND and USD deposits; (ii) the
SBV’s policies were only the temporary ones; (iii) the investors’ speculative
behaviours were caused by the decline in confidence of VND value. Thus, the SBV
raised the exchange rate to 18,932 VND/USD (devaluation of 2.1%) in August.
- 40 -
Additionally, the surplus of foreign currency resulted to the strong increase of
the FER from 20,500 VND/USD in October to a recording level of 21,500 on
November, 2010. The SBV immediately sold USD to the market for importing
necessary goods, controlled tightly the lending of gold and raised the import quota for
gold as well in order to reduce the liquidity stress on the gold market. Simultaneously,
the basic interest rate was also increased to 9% in November. However, the FER just
decreased slightly in the last month of 2010, and the n rose again in the beginning of
the year 2011. The first reason was that the foreign exchange reserves had declined
continuously since 2009, thus the SBV’s announcement about supplying USD for the
market was not completely guaranteed. Second, the domestic inflation climbed up
significantly. Finally, the trade deficit was really high, about USD 12.4 billion. These
unfavourable factors made the SBV’s depreciating action less effective and the
exchange rate kept being unstable.
Figure 3.2: Monthly fluctuation of VND/USD, 2008 – 2013
Source: Citibank, Securities and Fund Services (2013)
- 41 -
The volatility of the exchange rate continued happening in 2011, even more
severely than in 2010. On 11th February, 2011, the SBV decided to raise the OER to
20,693 VND/USD and narrow the trading band from ±3% to ±1%. This was the
highest amount of devaluation of VND by increasing the interbank rate of 9.3%. The
FER shot up to 22,100 VND/USD just a few days after the decision. Since the SBV
tightened the operations on the black market and imposed a low ceiling interest rate of
foreign currency deposits in March, the supply of USD went up when both enterprises
and individuals exchanged USD into VND for the interest rate difference. Therefore,
either the FER or BER dropped in the following months. It can be seen that during the
period of 2010 and 2011, the SBV had been more flexible and reacted more
effectively based on the signal of the market. Nevertheless, the SBV should have
created an environment where the market could operate by itself instead of using
many administrative methods. Since the fourth quarter of 2011, the foreign exchange
market gradually took back its stability, the decreasing pressure on the exchange rate
helped to strengthen the investors’ confidence in the VND value. That had created a
firm base of the exchange rate management for the next two years.
3.2.3. The period of 2012 – 2013
The tendency of stability of the exchange rate in 2012 and 2013 completely
contrasted with the exchange rate’s fluctuation in the preceding years. The OER was
kept changing with a small percentage of 1% in the fo urth quarter of 2011, and then
maintained at 20,828 VND/USD throughout the year 2012. In terms of BER, it
increased slightly in the first half of 2012 but changed the direction to decrease in the
second half of the year. As a result, the BER declined nearly 1% for the whole year of
2012. To be specific, from January to May, the commercial banks reduced the
exchange rate lower than the upper limit of the trading band after holding for a long
- 42 -
time. However, they raised the exchange rate to the highest value again in June. The
fluctuation of FER sticked close on the BER so that the gap between two markets
became much smaller than the one in 2011, which was about 20 – 70 VND/USD (as
in Figure 3.3). One difference in 2012 compared to other years was that in February,
the bid rate of the SBV’s Foreign Exchange was adjusted to higher than the one of the
commercial banks, which encouraged these banks to sell foreign currency to the SBV
so that the country’s foreign exchange reserves could be increased.
Figure 3.3: The volatility of the exchange rate in 2012
21,050
21,000
20,950
20,900
OER
20,850
BER
20,800
FER
20,750
20,700
Source: SBV and Tran Thi Luong Binh (2013)
In the last six months of 2012, the BER decreased gradually, except for the one
in August. Although the exchange rate had a tendency of fluctuating strongly along
with the big gap between the OER and FER in the last few months of the previous
years, this phenomenon disappeared in 2012. It can be explained by three reasons.
First, the SBV committed to adjust the trading band not over 2 – 3% for 2012 at the
beginning of the year. Second, the trade balance and overall balance were surplus
which supported efficiently for the commitment of the SBV. Finally, the gold market
- 43 -
was controlled closely so that it could not have much negative effect on the black
exchange market as before. There are several administrative measures for managing
the exchange rate in 2012 summarized in the Table 3.3.
Table 3.3: The SBV’s measures for managing exchange rate in 2012
Number
Measure
Content
Adjust directly the
interbank rate
The interbank rate was fixed at 20,828
VND/USD
Sell and buy foreign
currency on the market
Long position of about USD 10 billion
3
Control the credit of
foreign currency
The customers were restricted as the
inhabitant borrowing foreign currency for
payments in the short, medium, and long
term must have the source of income in
foreign currency from their businesses
4
Adjust the discount
interest rate
The discount interest rate was reduced from
13% to 7% through six times of adjustment
5
Change the financial
institutions’ foreign
exchange position
The foreign exchange position was
narrowed from 30% to 20% of the financial
institutions’ charter capital (Circular No.
07/2012/TT-NHNN)
1
2
6
Tighten the operations
on the gold market
+ The SBV is the representative of the
government in managing the gold market.
The sale and purchase of gold bars are
carried out only at financial institutions and
businesses that are permitted by the SBV
+ The deposit and credit of gold were
prohibited since 25th November, 2012
7
Others
+ A committment of adjusting the exchange
rate in the range of 2 – 3% was exercised to
put the investors’mind at ease
+ Strict control on the operations of the
parallel market
Source: Tran Thi Luong Binh (2013)
- 44 -
Similar to 2012, the SBV also engaged in keeping the exchange rate fluctuation
of about 2–3% in 2013 and the trading band of ±1%, combining with the flexibly sale
and purchase activities on the foreign exchange market in accordance with the interest
rate situation. The exchange rate continued to be remained stable until the mid- year
when it was officially increased 1%, which was lower than the target set by the SBV.
The OER was modified marginally from 20,828 VND/USD to 21,036 VND/USD on
June, 2013. The SBV’s ask rate was changed three times, from 20,950 VND/USD to
21,246 VND/USD in July, to reflect the supply and demand of the market during the
year. On the other hand, the bid rate was managed to be higher than the OER in order
to exhort the financial institutions to sell USD. It was kept continuo usly at 20,850
VND/USD and increased to 21,100 VND/USD in August. In addition, a series of
supportive methods were performed, such as the ceiling interest rate of deposits with
no term or term under 6 months was set at 7%, the interest rate of USD deposit was
assigned 0,25% per year to companies and 1.25% per year to individuals. These
policies helped to maintain the difference of returns in holding VND or USD so that
the adjustment of the official rate did not cause to the strong variation in the market
(as shown in Figure 3.2). Since then, the foreign exchange market was stable, and the
SBV continued buying foreign currency to enrich the foreign reserves.
3.3. Vietnam’s Exchange Rate Policy in Comparison with Other Asian Countries
The exchange rate regime and monetary policy in some Asian countries in 2008
and 2012 are summarized in the Table 3.4. According to IMF, in 2008, most of the
East Asian countries had adopted an exchange rate policy commonly referred to as
“managed floating” such as Cambodia, Indonesia, Laos, Malaysia, Singapore,
Thailand, etc. They allowed their currencies to fluctuate in foreign exchange markets
as long as the volatility in value did not violate the country’s other economic goals
- 45 -
like inflation limits or money supply constraints. The Phillipines followed the policy
of “free floating” as well as some other Asian countries that are Japan, South Korea,
etc. However, the exchange rate policy of several countries changed completely in
2012. For example, Indonesia and Thailand had applied the free floating policy, while
Malaysia, Singapore and Myanmar had exercised the other managed arrangement.
Table 3.4: The exchange rate regime and monetary policy frame work
2008
2012
Country
EXR regime
Monetary policy
EXR regime
Monetary policy
Cambodia
Managed floating
USD anchor
Managed floating
USD anchor
Indonesia
Managed floating
Inflation target
Free floating
Inflation target
Laos
Managed floating
USD anchor
Stabilized
arrangement
Other
Malaysia
Managed floating
Other
Other
managed Other
arrangement
Philippines
Free floating
Inflation target
Free floating
Singapore
Managed floating
Composite
Other
managed Composite
arrangement
Myanmar
Managed floating
USD anchor
Other
managed Other
arrangement
Thailand
Managed floating
Inflation target
Free floating
Inflation target
Vietnam
Conventional
fixed peg
USD anchor
Stabilized
arrangement
Composite
China
Crawling peg
USD anchor
Crawl- like
arrangement
Monetary
aggregate target
South Korea
Free floating
Inflation target
Free floating
Inflation target
Source: IMF Annual Reports on Exchange Arrangements and
Exchange Restrictions (2008, 2012)
Inflation target
Although IMF classified Vietnam’s exchange rate regime as a conventional
fixed peg arrangement in 2008 and stabilized arrangement in 2012, Vietnam has
officially accepted the “managed floating” policy, while some experts claimed that it
- 46 -
is actually a type of “crawling peg” such as Nguyen Tran Phuc et al. (2009) and
Nguyen Thi Thu Hang et al. (2010). Furthermore, China’s exchange rate policy had
been categorized as “crawl- like arrangement” because its interbank exchange rate was
widened to 1% since April, 2012, as well as the spread between China Yuan (CNY)
and USD selling and buying prices may not exceed 2% of the central parity.
Additionally, as shown in Figure 3.4, Vietnam’s NER 1 increased year over year,
especially increased strongly since 2008.
Figure 3.4: Nominal exchange rate index of several Asian currencies against USD
(2004 is the base year - Unit: %)
140
120
100
VND/USD
MAR/USD
80
SGD/USD
60
THB/USD
CNY/USD
40
KRW/USD
20
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Source: Calculation based on the data of WB (2013) and
Federal Reserve System (2014)
According to UNDP (2013), the level of devaluation was about 0.4% per quarter
during the period of 2000-2007, and then it rised to 1.8% per quarter in the following
periods starting from the beginning of 2008. From 2005 to 2013, Vietnam was the
only country whose currency uninterruptly depreciated of 32.94%, while other
1
No minal exchange rate
- 47 -
currencies like MAR 2 , SGD 3 , THB4 , CNY5 , and KRW 6 appreciated against USD. Due
to the effect of the global financial crisis, the exchange rate of these four countries,
except for China, climbed considerably in 2009, especially the KRW and the MAR
were devalued 15.87% and 5.66% compared to 2008 respectively. CNY was the only
currency appreciated continuously over the period.
It is clearly that Vietnam’s exchange rate policy has been more rigid than the
other Asian countries. The fluctuation of the exchange rate depends on the subjective
viewpoint of the SBV rather than what actually happens in the market. However, this
trend has only been applied for the NER of VND because the RER7 has gone to the
opposite way in recent years (UNDP, 2013). It grew slightly in the first four years of
the 21st century, and then decreased with the amount of 1.5% per quarter from 2004
to 2010. The gap between NER and RER became more and more extremely large,
particularly in 2008 and 2009. Although VND was also appreciated following to the
general tendency in Asia, the pace of the VND’s real appreciation was higher than
four neighbouring countries’ currencies including SGD, MAR, CNY and KRW,
which has deteriorated Vietnam’s international competitiveness and trade balance
(UNDP, 2013). That could be partially explained by the higher and more volatile
inflation of Vietnam compared to these Asian economies. For example, if compared
to 2000, the CPI of Vietnam increased 120% in 2010, whereas the CPI of the United
States just went up 26.7% in the same period. According to WEO (2014), the average
inflation of Vietnam in 2008 was about 23.12% which was larger fo ur times than the
ones of Malaysia, Thailand and China, and nearly five times than South Korea’s. In
2
Malaysia Ringgit
Singapore Do llar
4
Thailand Baht
5
Ch ina Yuan
6
South Korea Won
7
Real exchange rate
3
- 48 -
2010, in spite of the decelaration, Vietnam’s inflation was still higher than the other
countries’ three times. This reflects the weaknesses in the macroeconomic policy
framework of our country.
On the other hand, from 2011 to 2013, the RER went up again to shorten the
gap with the NER. With the base year of 2004, the RER index of VND/USD was
89.34 and 86.90 in 2012 and 2013 respectively. This index was higher than the ones
of the four comparative countries, except for South Korea, which means that VND
had appreciated against USD much less than the appreciation of other currencies. That
could enhance the competitiveness of Vietnam in exporting goods, and the n improve
the trade balance. For instance, Vietnam stood in the third place among the South East
Asian countries in the value of exports to the US, which was just lower than that of
Thailand and Indonesia, and held the highest positive amount of trade balance to the
US only in the area in 2012 (Tran Luong Thanh, 2013). Apart from the inflation, the
exchange rate fluctuation can be clarified by other factors such as GDP, interest rate,
current account, etc. which are discussed in the next section.
3.4. The Relations hip between Exchange Rate and Its Determinants
3.4.1. Exchange Rate and GDP Growth Rate
Vietnam has often been one of the countries with the highest economic growth
rate in the world. The following figure shows Vietnam’s GDP growth rate in relation
with the exchange rate. Note that GDP in 2012 and 2013 was calculated based on the
comparative price of 2010, while the others was relied on the price of 1994.
The growth rate of Vietnam’s GDP is ranging from 5% to 8.5% during the
period of 2005 to 2013. Before the global crisis, the average GDP growth rate was
about 8.37% between 2005 and 2007. However, it reduced significantly from 8.48%
- 49 -
in 2007 to 6.78% in 2010 and 5.42% in 2013 because of the impact of the economic
downturn. In the first quarter of 2014, GDP was estimated to grow 4.96% compared
to the same period in 2013, which was larger than the increase in the first quarter of
2013 and 2012.
Figure 3.5: The relationship between exchange rate and GDP growth rate
25,000
9%
8%
20,000
7%
6%
15,000
5%
4%
10,000
3%
2%
5,000
1%
0
0%
2005 2006 2007 2008 2009 2010 2011 2012 2013
VND/USD
GDP Growth Rate
Source: SBV and GSO
As mentioned in the literature review, many studies found that GDP growth has
a negative relationship with exchange rate. In case of Vietnam, Figure 3.5 shows that
the exchange rate of VND/USD kept going up, no matter what the GDP growth rate
increased or decreased. Moreover, the exchange rate in the first four months of 2014
was maintained at the same rate as 2013 and was guaranteed to fluctuate around 2% 3% by the SBV, while the GDP growth rate had had the highest level in the last three
years. Nevertheless, there are two periods that the GDP growth and the exchange rate
follow the pattern in which GDP growth goes down and the exchange rate goes up
that are 2007 – 2009 and 2010 – 2012.
- 50 -
3.4.2. Exchange Rate and Inflation Rate
Inflation is always a stinging problem that Vietnam has to confront year over
year, especially since the year 2007. The inflation rate rose sharply between 2007 and
mid – 2008 reflecting the increase in international commodity prices, the heavy
investment by state – owned enterprises and the outpouring of foreign investments
into Vietnam on the approach to the World Trade Organisation (WTO). From the
peak of 22.97% in 2008, the inflation declined dramatically to 6.88% in 2009 by the
weakening domestic demand and decreasing food and energy prices. Since the SBV
just exercised the tightened monetary policy to control inflation until the first half of
2010, and then loosened its policy by increasing money supply to support for the fall
of interest rate, the inflation rate began accelerating again when it reached another
peak of 18.58% in 2011. Due to the impact of a package of strict monetary and fiscal
measures announced by the government known as “Revolution 11”, the inflation
considerably went down to 9.21% in 2012 and 6.60% in 2013. Besides, the inflation
rate of the quarter one of 2014 increased 4.83% in comparison with the same period
last year. This is the lowest level of increase in the inflation in the last ten years.
As in previous studies, inflation rate positively relates with exchange rate. The
link between these two factors of Vietnam illustrated by Figure 3.6 is that despite the
ups and downs of the inflation rate, the exchange rate still kept its rising trend. In
2008, while the inflation attained the highest rate since 1993, the VND/USD just grew
5.36% compared to 2007. The reason for this was that the SBV had increased the
money supply of VND in order to buy foreign currency from the foreign investments
and oversea remittance. However, it had not carried out any effective measures to
neutralize the amount of VND supplied to the market. From 2010 to 2011, the annual
average exchange rate grew about 10.19% including the devaluation of VND of 9.3%
- 51 -
in February. The SBV had to raise the official rate as well as narrow the trading band
due to the escalation of the inflation rate at the end of 2010. This period reflects the
most precisely the expected relationship between the inflation rate and the exchange
rate among other nine years. Because of the government’ good control to the inflation,
the exchange rate became relatively stable in 2012 and 2013.
Figure 3.6: The relationship between exchange ra te and inflation rate
25,000
25%
20,000
20%
15,000
15%
10,000
10%
5,000
5%
0
0%
2005
2006
2007
2008
VND/USD
2009
2010
2011
2012
2013
Inflation Rate
Source: SBV and GSO
According to Nguyen Duc Thanh et al. (2009) and Luong Thi Nga et al. (2013),
the main causes for the soaring inflation rate in some periods in Vietnam are the
excess of money supply and credit growth. The exchange rate policy did not affect
strongly and directly to the inflation, but it only had a resonant impact to make the
inflation higher. This is the reason why in 2007 – 2008, the exchange rate was settled
while the inflation rate still increased steeply. However, in a period when the inflation
was controlled properly like from the end of 2011 to 2013, it would contribute to
stabilize the exchange rate effectively.
- 52 -
3.4.3. Exchange Rate and Interest Rate
Among many kinds of interest rates, lending interest rate has direct influence on
the enterprises’ manufacturing activities as well as their competitiveness in the market.
According to World Bank (WB), lending rate is the bank rate that often meets the
private sector’s short- and medium-term needs and is differentiated based on
creditworthiness of borrowers and objectives of financing. In Vietnam, the lending
interest rate is usually led by the market forces, but not exceed 150% of the base rate
announced by the SBV in each period. The relationship between exchange rate and
interest rate examined in previous research can be either positive or negative. This
connection which is displayed in Figure 3.7 had followed the two patterns including (i)
positive relationship in 2005 – 2008 and 2010 – 2011, (ii) negative link in 2008 –
2009 and 2012 – 2013.
Figure 3.7: The relationship between exchange rate and inte rest rate
25,000
20%
18%
20,000
16%
14%
15,000
12%
10%
10,000
8%
6%
5,000
4%
2%
0
0%
2005 2006 2007 2008 2009 2010 2011 2012 2013
VND/USD
Lending Interest Rate
Source: SBV, Stockplus,OECD
From 2005 to 2008, money supply and credit growth increased substantially
adding to the high returns of the stock market and real estate that made the SBV
- 53 -
exercise several tightened policies such as increasing required rate, limiting total
credit, issuing compulsory bills, etc. The result was that the market faced the risk of
illiquidity which caused to lending rate surged in the middle of 2008. Especially, in
the third quarter of 2008, the lending rate climed up to over 21% that brought to t he
high average rate of 17.08% for this year. Thus, the exchange rate increased in order
to reduce the pressure on VND.
In 2009, when the growth of the economy had a sign to decline due to the
lagged effect of the crisis, the SBV loosened its administrative interest rate to reduce
the lending rate and extended the credit to provide more capital to the market. Since
the high inflation arised in the last six months of 2010, the strict monetary policy was
carried out, the prime rate was raised to 9% instead o f 8% unchanged in 10 months.
The suddenly changing direction of the monetary policy led to many difficulty in
liquidity for the enterprises at the moment of year – end, which in turn put more stress
on raising the official exchange rate. In 2011, the SBV was consistent with controlling
inflation and credit growth rate by setting a ceiling interest rate for deposit and
intensifying managment on non – manufacturing objectives borrowing from banks.
With the interest rate reaching a peak of 18.14% in 2011, the government came
under heavy pressure to cut the lending rates to help the real estate sector. Since the
beginning of 2012, the SBV managed the interest rate positively and sticked close on
the sequence of economic events to help businesses extricate from t he difficulty in
operation and production. In 2013, the SBV reduced 2% per year of the administrative
interest rates, 3% per year of short-term lending interest rate, and 1% per year of
deposit interest rate of VND. The lending rate was fairly stable, around 7% - 10.5%
for the short term and 11% - 13% for the long term. The SBV kept reducing the
administrative interest rate and the overnight lending rate in the interbank market on
- 54 -
March, 2014. That created a good condition for commercial banks to decrease their
lending interest rate. The interest rates have steadily declined in order to stabilize the
monetary market and reduce the price of inputs for businesses. The exchange rate has
not also had any strong fluctuation from the third quarter of 2013 to the first four
months of 2014.
3.4.4. Exchange Rate and Current Account
Current account balance is defined as the sum of net exports of goods and
services, net primary income and net secondary income (WB). The current account is
expected to have negative relationship with the exchange rate which means that the
current account deficit will lead to the increase in the exchange rate due to the rising
demand of foreign currency for importing and vice versa. In Figure 3.8, this
relationship is totally agreed with the assumption in the period of 2005 to 2010, while
it is not true for the most recent three years. Although the NER rose annually, the real
appreciation of VND had caused to the decrease in Vietnam goods’ competitive
competency in the international markets. This was expressed in the state of the current
account’ large deficit for a long time. Particularly, after Vietnam had joined in WTO
in 2006, the excess of imports over exports became much bigger than ever before
which was nearly USD 18 billion and USD 12.3 billion in 2008 and 2009 respectively.
As curbing inflation became the country’s priority rather than maintaining growth, the
government tried to reduce the amount of imports to cut the current account deficit,
which in 2009 was USD 6,608 million.
However, the acceleration of the demand of foreign currency was not
counterbalanced timely and sufficiently by the depreciation of domestic currency that
put the foreign exchange market into a strained situation and made the black market
as well as currency speculation grow strongly. An illustration for this phenomenon
- 55 -
was the period from the end of 2006 to the beginning of 2008 when the exchange rate
was hold firmly even if the large deficit lasted constantly for months. In contrast, from
the middle of 2008 to 2009, although the deficit reduced significantly, the VND was
still devalued considerably.
Figure 3.8: The relationship between exchange rate and current account
25,000
15,000
10,000
20,000
5,000
15,000
0
10,000
-5,000
5,000
-10,000
-
-15,000
2005 2006 2007 2008 2009 2010 2011 2012 2013
Current Account (mil USD)
VND/USD
Source: SBV, WB, WEO
Additionally, despite the surplus current account since 2011, the exchange rate
continued rising rather than falling as expectation. While the current account was
surplus of USD 236 million after many years, the exchange rate had had the biggest
devaluation of 9.3% in 2011. That represented the lagged effect of the government’s
policy when it tried to control the inflation. Despite the high excess of imports over
exports of USD 9.8 billion in 2011, 2012 was the first year that had trade surplus of
approximately USD 780 million. The exports of garment and electronics increased
while the imports of machines and raw materials decreased in this year. In spite of
trade surplus, the net income balance was in the large – scale deficit which was about
USD 3.8 billion in 2012. The main deficit financing was the oversea remittance,
- 56 -
which was USD 10 billion, added to the foreign currency collected from labour export.
Therefore, the surplus current account of 2012 reached USD 9.06 billion.
Moreover, the export turnover of 2013 set a new record of USD 132.2 billion
which increased 15.4% compared to 2012, while the import turnover achieved USD
131.3 billion. Thus, the trade surplus was about USD 863 million in 2013, which
helped the current account balance grow bigger than the one of 2012 1.25 times. The
balance of trade continued to be surplus in the first quarter of 2014, and the overseas
remittance was estimated about USD 2.3 billion. Meanwhile, the exchange rate still
remained stable during these two years. It could likely have no clear link between the
exchange rate and the current account balance in this period.
3.4.5. Exchange Rate and Foreign Direct Investment
According to WB, foreign direct investment (FDI) is the net inflows of
investment to acquire a lasting management interest in an enterprise that includes
equity capital, reinvestment of earnings, other short- and long-term capital. In general,
FDI can either increase the supply of capital or facilitate technology transfer. Thus, it
contributes to impressive economic growth of developing countries. Vietnam began
attracting significant FDI from 1988 due to the promulgation of a liberal foreign
investment law in 1987 (Sajid Anwar and Lan Phi Nguyen, 2010).
In terms of the relationship between exchange rate and FDI, many studies have
stated that FDI has negative impact on exchange rate because of the increase in the
supply of foreign currency. However, as shown in Figure 3.9, the exchange rate and
FDI of Vietnam did not have this kind of relationship but the reverse one. That means
both indicators increased together despite the small decline of FDI in some years.
There had been a rapid rise in FDI in Vietnam starting from 2005 because of some
- 57 -
policies of the government such as the enforcement of a new enterprise law, the
opening of a stock market and the resumption of IMF lending, etc. During Vietnam’s
accession to WTO, the FDI accelerated in 2007 with the doubled amount against 2006
and reached a peak of USD 11.5 billion in 2008. It fell slightly in 2009 due to the
global crisis, and then went up again in 2010.
Figure 3.9: The relationship between exchange rate and FDI
25,000
14,000
12,000
20,000
10,000
15,000
8,000
6,000
10,000
4,000
5,000
2,000
0
0
2005 2006 2007 2008 2009 2010 2011 2012 2013
FDI (mil USD)
VND/USD
Source: SBV, GSO, MPI
As the foreign capital was restrained in the domestic market by the context of
comsuming inconsiderably and manufacturing at a indifferent level in 2012, the
registered FDI attained USD 13 billion while the implemented FDI was about USD
10.46 billion, which was less than the amount of capital in 2011 over USD 500
million. Newly registered FDI capital in 2013 increased to USD 21.6 billion, an
increase of 54.5% over the same period of 2012. The FDI disbursement in 2013 was
estimated at USD 11.5 billion, up 9.9% from 2012. The capital flows focused mostly
on the processing and manufacturing industry that occupied about 76.9% of the total
registered FDI (VFS, 2014) instead of the real estate in previous years. For the first
- 58 -
four months of 2014, the disbursed FDI approximates USD 4 billion increased 6.7%
compared to the same period in 2013. Besides, the countries with the highest amount
of FDI for Vietnam are usually Japan, Singapore, South Korea, Canada, etc.
3.4.6. Exchange Rate and Foreign Exchange Reserves
A country’s foreign exchange reserves can enable its central bank to intervene
against volatile fluctuations in currency by affecting the exchange rate and increasing
the demand for and value of the country’s currency (Sargeant, 2009). In other words,
the central bank usually uses the reserves to maintain a steady exchange rate against
factors that can negatively affect the exchange rate. Thus, the foreign reserves and
exchange rate relate positively to each other that means when domestic currency faces
a pressure to depreciate, the central bank will buy foreign currency to keep the
exchange rate stable, and then the foreign reserves is increased.
Figure 3.10: The relationship between exchange rate and FX reserves
25,000
35
30
20,000
25
15,000
20
10,000
15
10
5,000
5
0
0
2005
2006
2007
2008
2009
FX Reserves (bil USD)
2010
2011
2012
2013
VND/USD
Source: SBV, ADB, CIA
In case of Vietnam, the positive relationship between exchange rate and foreign
reserves was only specified in the period of 2005 – 2008 and 2011 – 2013 as shown in
- 59 -
Figure 3.10. The reserves surged to USD 23.74 billion in 2007, which was more than
doubled from 2005, and continued rising to over USD 24 billion in 2008. This
acceleration was caused by the large amount of FDI capital overflowed into Vietnam
since its accession to WTO in 2006. Nevertheless, the national foreign reserves
reduced sharply about USD 11.25 billion just in two years 2009 – 2010 as the
government defended VND. According to WB, foreign currency was hoarded as
expectations of a devaluation in VND caused a shift into assets such as USD and gold,
and then led to a shortage of USD under control of the SBV.
Along with the surplus balance of payments, the official foreign exchange
reserves increased substantially from USD 17 billion in 2011 to USD 23.5 billion in
2012, which met a demand of 12 weeks of imports. The amount of foreign currency
bought for the reserves in 2012 was about USD 12 billion that made up the one sold
in 2009 – 2010 when the economy had suffered the large BOP deficit (Nguyen Duc
Thanh et al., 2013). In 2013, the low inflation, stable macro economy, surplus trade
balance, and well – controlled foreign exchange market had bolstered the investors’
confidence in the sustainable development of the country. With the stabilized
exchange rate, the SBV reported to buy many billions worth of foreign currency to
feed the national reserves which was expected to surge to nearly USD 34 billion, as
twice as the amount in 2011.
In the first four months of 2014, the exchange rate remained stable, the SBV
bought over USD 10 billion so that the foreign reserves was gained more than USD
35 billion. According to the Governor, Nguyen Van Binh, this is the net reserves that
can be used any time, while the potential foreign exc hange reserves of Vietnam is
around USD 45 billion.
- 60 -
3.4.7. Exchange Rate of VND against USD, EUR and JPY
In addition to USD, Euro (EUR) and Japanese Yen (JPY) are the two main
currencies used in the international transactions. More and more Vietnam companies
choose EUR and JPY as their payment currencies in foreign trade because they want
to diversify the exchange rate risks, rather than relying on only one currency like USD.
Different from the VND/USD, there have no official rate for the VND/EUR and
VND/JPY announced daily by the SBV. These exchange rate s are established by the
market supply and demand. In other words, they depend on the agreement between
banks and their customers as well as banks’ foreign exchange positions.
Therefore, it can be stated that if the number of transactions of EUR and JPY
increases in the domestic market, it could probably affect the exchange rate of
VND/USD since the demand for USD might be reduced. However, the direction of
change of VND/USD is not clearly defined if just relying on the violation of
VND/EUR and VND/JPY because it is also influenced by the relationship between
USD and EUR, JPY in the international markets. As a result, the exchange rate of
VND/EUR and VND/JPY can relate either positively or negatively to VND/USD.
The Figure 3.11 shows the relationship of the three pairs of main exchange rates
in Vietnam. Overall, all of them followed a rising trend from 2005 to 2013. However,
the exchange rate of VND/EUR and VND/JPY fluctuated more frequently than
VND/USD. That means while VND/USD grew constantly for years, VND/EUR and
VND/JPY had the ups and downs during the last nine years.
- 61 -
Figure 3.11: The relationship between VND/USD and VND/EUR, VND/JPY
35,000
300.00
30,000
250.00
25,000
200.00
20,000
150.00
15,000
100.00
10,000
50.00
5,000
0
0.00
2005 2006 2007 2008 2009 2010 2011 2012 2013
VND/USD
VND/EUR
VND/JPY
Source: http://www.oanda.com/currency/average
In terms of VND/EUR, it kept going up with the average rate of 6.54% per year
since 2005, especially this exchange rate increased 12.81% in 2011 and 9.12% in
2008 compared to the preceding years (2010 and 2007). 2008 and 2011 were also the
two years that witnessed the strongest depreciation of VND/USD in many years. It
declined considerably in 2012, and then climbed up again in 2013.
Additionally, after a slight fall in 2006 and 2007, VND/JPY began increasing
steeply with the rate of 16.97% per year from 2008 to 2011 and reached a peak of
261.14 VND/JPY in 2012. This exchange rate also rose in the period when VND was
robustly devalued against USD. Nevertheless, it just grew modestly in 2012 and fell
drastically of 17.70% in 2013, whereas VND/USD was rather settled.
3.4.8. Conclusion
The relationships between exchange rate and several macro determinants chosen
from the literature review are examined by looking at their development in the past.
Since the exchange rate is a sensitive issue that the SBV attempts to control rather
- 62 -
strictly, it sometimes has had movements that do not reflect the signs of the market.
However, these above variables have influenced the exchange rate as expected in some
periods, and the management of the SBV for the exchange rate is presented through
the policies about them. Obviously, from 2005 to 2013, the exchange rate relates
positively to the foreign exchange reserves and negatively to the current account that
are specified most of the time. The possible positive and negative links between
VND/USD and the interest rate as well as VND/EUR and VND/JPY are easily seen
during the years. However, the exchange rate is likely to have a positive connection
with the FDI against what is expected in the theory. The GDP growth rate and
inflation rate show not – clearly – defined correlation with the exchange rate because
despite their ups and downs, the exchange rate still keeps rising over time.
- 63 -
CHAPTER 4
DATA COLLECTION AND RESEARCH METHODOLOGY
The main research strategy for this chapter is quantitative approach. After choosing
variables and collecting data, the regression model for analyzing the relationship
between those variables are conducted. Some tests will be done to find the optimal
regression model for forecasting the exchange rate in the next period. Macro stress
testing and VaR approach for the foreign exchange exposure are also introduced in
this section.
4.1. Data Collection
This study focuses on a time-series data spanning from 2005 to 2014. The
quarterly data for the exchange rate VND/USD and its the deter minants including
GDP growth rate, inflation rate, interest rate, current account, etc. are collected from
the General Statistics Offices, State Bank of Vietnam, World Bank, International
Financial Statistics (IFS), Asian Development Bank (ADB) and so on. The data for
these indicators related to the US are drawn from the Federal Reserve System and the
US Bureau of Economic Analysis, etc. The computer software used for analysis are
Eviews 8.0, Excel 2007 and Fronline Solvers 2014.
4.2. Research Methodology
4.2.1. Regression Model
Multiple regression analysis is used when there are more than one independent
variables to explain variance in dependent variable. It provides a means of objectively
assessing the degree and the character of the relationship be tween the independent
variables and the dependent variable through regression coefficients (Sekaran, 2010).
In this research, the nine indicators chosen from the literature review that can affect
- 64 -
the exchange rate of Vietnam Dong against the US Dollar are presented in the
following equation:
Where
EXRt : The exchange rate of VND/USD at time t;
GDPt : The GDP growth rate of Vietnam at time t (in percentage);
INFt : The difference between Vietnam’s inflation rate and the US inflation rate
at time t (in percentage);
INTt : The difference between Vietnam’s interest rate and the US interest rate
at time t (in percentage);
CURt : The current account balance at time t (in million dollars);
FDIt : The foreign direct investment at time t (in billion dollars);
FXRt : The foreign exchange reserves at time t (in million dollars);
EURt : The exchange rate of VND/EUR at time t;
JPYt : The exchange rate of VND/JPY at time t;
EXRt-1 : The exchange rate of VND/USD at time t-1.
4.2.1.1. Dependent Variable
EXRt is the exchange rate at time t that is measured in terms of Vietnam Dong
(VND) per US Dollar (USD). That means VND appreciates relative to USD if the
ratio of VND/USD falls, and it depreciates with an increase in the ratio.
4.2.1.2. Independent Variables
GDPt is the growth rate of Vietnam’s gross domestic product at time t. A
country with high GDP tends to have a strong currency compared to others with low
GDP. It can be seen that the higher the country’s GDP is, the stronger its currency
- 65 -
will be. GDP shows a significantly negative relationship with the exchange rate
(Nuku, 2011; Atif et al., 2012). Thus, GDP is expected to negatively relate to EXR.
INFt is the difference between Vietnam’s and the US inflation at time t. A high
inflation rate should be accompanied by the depreciation of the exchange rate. A
country with consistently lower inflation rate exhibits a rising currency value as its
purchasing power increases relative to other currencies (Sinha and Kohli, 2013). The
country with higher inflation rate may have its currency depreciated more deeply in
relation to its trading partners. So INF probably has a positive relationship with EXR.
INTt is the difference between Vietnam’s and the US interest rate at time t.
The currency with the higher interest rate is predicted to appreciate because the higher
interest rate attracts capital from abroad that results to increasing demand for that
currency, and causes the exchange rate of the currency to increase (O’Brien, 2006).
However, according to Kalra (2005), an increase of interest rate quite often depresses
the stock market, leading to the sales of shares by foreigners. It also discourages the
economy’s growth perspective, and then stops the foreign direct investment inflows.
That causes to the depreciation of the currency with higher interest rate. Therefore,
the relationship between INT and EXR might be positive or negative.
CURt is Vietnam’s current account at time t. The current account includes all
international economic transactions with income or payment flows occurring within
the year. Current account deficit occurs when a country’s total imports of goods,
services and transfers are greater than its total exports of goods, services and transfers.
This means the country is spending more foreign trade than it is earning, and that it is
borrowing capital from foreign sources to make up the deficit (Van Bergen, 2010).
The excess demand for foreign currency makes the co untry’s currency depreciated,
which means the exchange rate will increase. In other words, the surplus of the
- 66 -
current account results to the excess supply for foreign currency, which causes the
decrease in the exchange rate. Thus, CUR is likely to be negatively related with EXR.
FDIt is Vietnam’s foreign direct investment at time t. FDI is a direct
investment into production or business in a country by foreign investors, either by
buying a company in the target country or expanding operations of an existing
business in that country (Wikipedia). The more capital of FDI come into Vietnam, the
more amount of foreign currency the country will have, which helps appreciate VND
against USD. As a result, FDI might have negative influence on EXR.
FXRt is Vietnam’s foreign exchange reserves (FX reserves) at time t. FX
reserves are supposed to be the amount of foreign currency deposits that a country’s
central bank holds, mostly in US Dollars. If the supply of foreign currency increases
which makes the exchange rate declined, the central bank tends to purchase foreign
currency to keep its currency value unchanged. That action leads to the increase in the
FX reserves. In other words, the exchange rate is likely to go up due to the high
amount of the FX reserves. Therefore, FXR is predicted to relate positively with EXR.
EURt is the exchange rate of VND/EUR at time t. If the VND/EUR declines,
VND appreciates against EUR. However, the relationship between EUR and USD
could not certainly be known. If USD depreciates or appreciates against EUR less
than VND, VND/USD will decrease. Otherwise, VND/USD will increase when USD
appreciates against EUR more than VND. Thus, EUR is expected to have either
positive or negative relationship with EXR.
JPYt is the exchange rate of VND/JPY at time t. If the VND/JPY declines,
VND appreciates against JPY. However, the relationship between JPY and USD
could be certainly unknown. If USD depreciates or appreciates against JPY less than
VND, VND/USD will decrease. Otherwise, VND/USD will increase when USD
- 67 -
appreciates against JPY much more than VND. Thus, JPY is expected to have either
positive or negative relationship with EXR.
EXRt-1 is the exchange rate of VND/USD at time t-1. This research assumes
that there is an autoregressive relationship with historical EXR. The current value of
EXR is likely relied on the immediate preceding value of EXR. Therefore, the
coefficient of this variable should have a positive sign.
Table 4.1: Expected signs of independent variables
Independent Variables
Expected Signs
GDPt
(-)
INFt
(+)
INTt
(+) or (-)
CURt
(-)
FDIt
(-)
FXRt
(+)
EURt
(+) or (-)
JPYt
(+) or (-)
EXRt-1
(+)
4.2.1.3. Statistical Techniques
There are several statistical techniques used in this research to provide the most
accurate results that can be applied for forecasting in the next stage.
-
Test for heteroscedasticity: Heteroscedasticity arises when the standard
deviations of variables monitored over a specific amount of time are non-constant
(Investopedia). The presence of this problem can invalidate statistical tests of
significance that assume the variances do not vary with the effects being modelled.
Therefore, the Breusch – Pagan – Godfrey test and White test is applied to examine
whether the residual variances in the regression model is constant.
- 68 -
-
Test for autocorrelation: Autocorrelation is a mathematical representation of
the degree of similarity between a given time series and a lagged version of itself over
successive time intervals (Investopedia). The Breusch – Godfrey test is a widely used
method of testing for autocorrelation. This test examines the null hypothesis that the
residuals from the regression are not autocorrelated up to any order. It can be worked
in cases where lagged values of the dependent variables are used as independent
variables in the model.
-
Test for multicollinearity: Multicollinearity is a statistical phenomenon in
which two or more independent variables in a multiple regression model are highly
correlated (Wikipedia). The issue of multicollinearity arises when there is an
approximate linear relationship among two or more independent variables. This study
will base on the rule of thumb that if the pair wise correlation between two regressors
is in excess of 0.8, the regression is said to have significant multicollinearity problem.
-
Test for redundant variables: Redundant variables are the independent ones
that have no relationship with the dependent variable. Thus, a redundant test is used to
examine any suspected variables in the model. Wald test is employed to test the true
value of the parameter based on the sample estimate whenever a relationship within or
between data items can be expressed as a statistical model with parameters to be
estimated from a sample (Wikipedia).
-
Test for endogeneity: In ordinary least square (OLS), errors are homoscedastic,
which means that the error term, conditional on the regressor, is constant. The
problem of endogeneity arises when a regressor is correlated with the error term,
which causes to the OLS estimates’ bias (Bascle, 2008). There are three instances
where endogeneity is present including errors in variables, omitted variables and
simultaneous causality. The errors – in – variables problem comes up when the true
- 69 -
value of a regressor is unobserved. There is an omitted variable bias when a variable,
which affects the dependent variable and is correlated with one or more independent
variables, is omitted from the regression (Wooldridge, 2009). Simultaneous causality
occurs when the causality runs in both directions: from the regressors to the
dependent variable, and from the dependent variable to the regressors (Bascle, 2008),
which is the main concern of this research. Hausman test is utilized to detect this
problem. Put in another way, this test can be used to check for the endogeneity of a
variable by comparing instrumental variable estimates to OLS estimates (Wikipedia).
4.2.2. Macro Stress Testing for Exchange Rate Exposure
The second stage of this study starts with employing the VaR models which
measure normal potential losses of a company. VaR figure is estimated based on the
historical behavior of the market determinants that influence the value of the
exchange rate. This create a basic limitation of VaR, to the extent that the past may
not be a good indicator of the future, and thus, the VaR measure may under- or overestimate market risk (Illova, 2005). Therefore, the application of stress tests, which
are designed to estimate potential economic losses in abnormal markets, are supported
as a complement.
Generally, stress tests would enhance transparency by exploring a range of
potential low – probability events when VaR bands are dramatically exceeded (RMG,
1999). The combination of them gives a more comprehensive picture of risks. Hence,
the stress testing is conducted after calculating the VaR by three approaches: variance
– covariance matrix, historical simulation and Monte Carlo simulation. The steps
included in these approaches are followed the framework of Linsmeier and Pearso n
(1999) and Berry (2012).
- 70 -
4.2.2.1. Variance – Covariance Approach
In terms of the computation required, the variance – covariance method is the
simplest one among the other two VaR approaches. It assumes that the changes in the
economic indicators follow a normal distribution that makes the steps to conduct this
approach become easy to implement. However, the assumption of normality might be
considered to be unrealistic for most financial markets.
Step 1: Calculate the historical parameters of the macro determinants ’ changes
including mean, standard deviation and correlation. The variance – covariance
procedure captures the variability through the standard deviations of the normal
distribution and the comovement of the factors through the correlation coefficients.
Step 2: Compute the mean and standard deviation of the exchange rate based on
these parameters of the changes in the values of the indicators.
Step 3: Establish the VaR for a given confidence level by multiplying the
standard deviation of the dependent variable and the relevant scaling factor, which is
derived from the standard normal distribution. For example, if a 99% level of
confidence is desired, the appropriate scaling factor is 2.33 since the probability of
occurrence of a number less than -2.33 is 1%. Scaling the standard deviation by this
amount yields a VaR which should only be exceeded 1% of the time (Cassidy and
Gizycki, 1997).
4.2.2.2. Historical Simualtion
The fundamental assumption of this methodology is that the past is the good
indicator of the near future. Thus, it uses historical changes in market factors to
compute the hypothetical profits and losses. It can be described in terms of four steps.
Step 1: Calculate the changes of all the macro determinants for the number of
last periods. Generally, it is required a long history of changes in order to get a
- 71 -
meaningful VaR. We uses a quarterly horizon from 2005 to 2014, covering 37
quarters in total.
Step 2: Apply the changes calculated to the current mark-to- market value of the
macro determinants and re-value the exchange rate and foreign exchange (FX)
position. We consider that these changes from the latest quarter back to the first
quarter of the period of time may occur in the next quarter with the same likelihood.
Then, we use them to construct hypothetical values of the determinants used in the
calculation of hypothetical exchange rate and FX exposure.
Step 3: Order the mark-to- market FX exposure from the lowest to the highest
value. This sort is implemented as VaR focuses on the tail of the distribution.
Step 4: Select the potential loss that corresponds to the desired confidence level.
This study is interested in two levels of confidence including 95% and 99%. The loss
that equals or exceeds 5% and 1% of the time is the VaR.
4.2.2.3. Monte Carlo Simulation
Unlike the historical simulation, Monte Carlo simulation does not rely on the
determinants’ past movements, but base on the generation of a large number of
possible changes that could affect the value of the FX position. Therefore, its process
allows evaluation of the effect of events that are just likely to occur as events
observed in the historical period. This method requires the following steps.
Step 1: Determine probability distributions for the changes in the independent
variables and to estimate the parameters of those distributions. This is the feature that
distinguishes Monte Carlo simulation from the other two approaches. The assumed
distribution for the variables is freely chosen, usually normal distribution, and
subjective judgments can be brought in to modify these distributions.
- 72 -
Step 2: Use series of pseudo – random numbers to generate a large amount of
hypothetical value of changes in the indicators, and then the corresponding simulated
mark-to-market values of exchange rate and FX exposure are obtained.
Step 3: Order the mark-to- market FX exposure from the lowest to the highest
value and select the potential loss that corresponds to the desired confidence level,
which represents VaR. This step is the same as in the historical simulation.
4.2.2.4. Stress testing
Macro stress testing is used to gauge the impact of extreme but plausible
predefined movements in the key market risk factors on the profit and loss of
individual firms. The applied scenarios generally represent upward and downward
movements in the risk factors to examine the sensitivity of the company’ exposure
under these circumstances. In other words, the stress scenarios need to cover a range
of factors that can create extraordinary losses, or make the firms’ control of risk very
difficult (Vohra, 2013). According to Badik (2005), there are three types of stress
scenarios including:
Relatively probable scenarios represent a change in risk factors by three to
five standard deviations.
Tests based on predictions of future development based on qualified
estimates, such as interest rate changes by monetary authorities, etc.
Catastrophic and worst – case scenarios ensuring from dramatic events like
terrorist attacks, wars, etc.
From a different aspect, a variety of scenarios could be involved in stress tests that are
(i) generating historical scenarios based on days when markets moved violently, (ii)
introducing market shocks and moving risk factors in isolation by large amounts to
- 73 -
gauge sensitivity to each risk factor, and (iii) creating anticipatory scenarios in which
many market factors are moved together to assess real volalibity of all relevant world
markets (RMG, 1999).
In this research, several hypothetical scenarios are constructed by stressing of
the group of macro determinants. These scenarios could be relative probable with the
change of three standard deviations, and the macro variables are shocked successively
as well as simultaneously. To be specific, the reversed movements of the market
factors will be applied to predict the exchange rate exposure for the next few quarters,
and then we can calculate the hypothetical expected losses that the company has to
suffer in case of unusual market conditions. To create comprehensive scenarios, we
also use historical events like the global financial crisis in 2008 as a benchmark
because they can cover various popular adverse market situations.
Additionally, according to RMG (1999), market returns usually have “fat tails”,
where extreme market moves (beyond 99% confidence) occur far more frequently
than a normal distribution would suggest. Since we would like to capture the fat tails,
we have to abandon the variance – covariance approach in this phase due to its
assumption of normality. Therefore, we only apply stress testing for the historical
simulation and Monte Carlo simulation. Especially, in Monte Carlo simulation,
although the simulated mixes of macroeconomic variab les are supported by the
estimated relationships based on historical data, this study does not assume that future
changes of these indicators will follow the same distributions as past changes, but
determine again the fitted distributions of the future volatility.
- 74 -
CHAPTER 5
DATA ANALYSIS
This chapter presents the descriptive data and analyzes the empirical results of the
regression model as well as the macro stress testing. This study applies quantitative
methods supported by some statistics softwares like Eviews 8.0 and Frontline Solvers.
5.1. Descriptive Statistics
Descriptive statistics presents simple summaries about the total observations of
each variable used in this study. Table 5.1 illustrates several commonly used measures
to describe the data sets including measures of central tendency (mean, median) and
measures of variability (maximum, minimum, standard deviation).
Table 5.1: Data description
Mean
Median
Maximum
Minimum
Std. Dev.
Obs
EXR
GDP
18,505
6.45
18,003
6.04
21,246
8.69
15,932
3.14
2,116
1.36
37
37
INF
8.36
6.03
22.39
3.16
5.64
37
INR
CUR
FDI
FXR
EUR
8.92
-629.23
2.21 21,542.43
24,842
9.25
-629.60
2.54 23,008.00
25,579
16.25 3,857.00
3.97 35,000.00
30,033
2.93 -6,139.50
0.32 7,865.00
18,954
4.52 2,506.50
0.99 7,156.42
3,212
37
37
37
37
37
Source: Eviews 8.0
This research examines the time series data from 2005Q1 to 2014Q1, and thus,
it includes totally 37 observations for each indicators. The independent variable, EXR,
has average value of 18,505 VND/USD and standard deviation of around 2,116 which
is rather low meaning that the data points tend to be close to the mean. The highest
value of EXR is 21,246 VND/USD, while the lowest one is only 15,932 VND/USD.
With respect to the macroeconomic variables, GDP has the smallest standard
deviation of 1.35% compared to the ones of INF and INR, constituting 5.64% and
4.52% respectively. The biggest difference between the inflation rate of Vietnam and
- 75 -
JPY
194.14
192.33
273.56
136.03
47.53
37
the US is 22.39%, seven times larger than the smallest one. The highest premium of
Vietnam’s interest rate is 16.25% with comparison to the US’s. Moreover, the range
of FXR is from USD 7,865 million to USD 35,000 million, which is much wider than
CUR’s from USD -6,139.50 million to USD 3,857 million or FDI’s from USD 0.32
billion to USD 3.97 billion. The average value of EUR is 24,842 VND/EUR, larger
than EXR 6,337 VND for one foreign currency. JPY’s mean value is about 194.14
VND/JPY with the highest value ever during the period is 273.56 VND/JPY.
The statistical description just helps to demonstrate the basic features of the data
in this study. The next two sections will provide as deeper understanding of the results
of the two main methodologies.
5.2. Regression Model for Exchange Rate’s Macro Determinants
This part aims to introduce several techniques to conduct the regression model
in order to not only attain the Ordinary Least Square (OLS) estimators but also verify
if the coefficients are Best Linear Unbiased Estimators (BLUE). There are six basic
tests handled in the study including serial correlation (autocorrelation), normality,
heteroscedasticity, multicollinearity, redundant variables and endogeneity.
5.2.1. Testing for Serial Correlation
One of the linear regression model’s assumptions is that the covariance between
the error terms over time is zero, i.e. cov(ui, uj) = 0 for i ≠ j. In other words, the
residuals are likely to be correlated with their own lagged values, especially in time
series regressions. If the errors are not uncorrelated with one another, they would be
autocorrelated or serially correlated that a test need to be required. The consequences
of this problem are that the coefficient estimates from OLS are still unbiased but
inefficient (i.e. not BLUE), so that the standard error estimates could be wrong.
- 76 -
There are some statistical tests including Durbin – Watson (DW), Breusch –
Godfrey (BG) and Lagrange Multiplier (LM). While DW only tests for first order
autocorrelation, which is a link between an error and its immediately previous value,
BG is a more general test that examines the connection between the residual and
several of its lagged values at the same time (Brooks, 2008). Morever, one of the
conditions for DW to be a valid test is that there must be no lags of dependent
variable in the regression, whereas LM test can be applicable whether there are lagged
dependent variables or not. As a result, this study conducts the BG serial correlation
LM test with the null hypothesis of no serial correlation up to lag order p. The number
of lags of the residual is determined based on the frequency of the data used in the
model (Brooks, 2008). Thus, p is set equal to 4 due to the quarterly data.
Table 5.2: Autocorrelation test
Breusch-Godfrey Serial Correlation LM Test:
F-statistic
Obs*R-squared
1.829700
8.986618
Prob. F(4,22)
Prob. Chi-Square(4)
0.1591
0.0614
Source: Eviews 8.0
According to Table 5.2, F-statistic is an omitted variable test for joint
significance of all lagged variables, and it is presented just for comparison purpose
since the exact finite sample distribution of the F-statistic under H0 is not known.
Obs*R-squared is the BG LM test statistic that is calculated by taking number of
observations (n) minus number of lags (p) and then time R2 from the regression. This
statistic is asymptotically distributed as a 2 (p). Therefore, if Obs*R-squared 2 (p),
we can not reject H0 . With the significance level () of 5%, 2 (4) = 9.488 > 8.986 or
p-value (0.0614) > 0.05, so the null hypothesis of no autocorrelation should not be
rejected. The F-statistic also comes with the same result.
- 77 -
5.2.2. Testing for Heteroscedasticity
The OLS assumes that the variance of the errors is constant, i.e. var(ui) = 2 <
(i = 1,2,...,n). This homoscedasticity assumption means that the variability around of
the regression line does not fluctuate when the values of the dependent variables
change. In other words, the residuals are considered to be heteroscedastic if the
variances are not constant. Similar to the consequences of serial correlation, when
there is heteroscedasticity, the OLS estimators are no longer BLUE. In addition, the
covariance matrix of the estimators is no valid that leads to erroneous inferences from
the t and F statistic based on this estimated matrix.
This research conducts two types of heteroscedasticity tests that are Breusch –
Pagan – Godfrey (BPG) and White. Both of them are based on the residuals of the
fitted model and have the form of a LM test. The BPG test only checks for the linear
form of heteroscedasticity, which means it tests whether the estimated variance of the
residuals is dependent on the independent variables. On the other hand, the White test
is more general since it tests the null hypothesis of no heteroscedasticity against
heteroscedasticity of unknown form. It regresses the squared residuals from the
original regression on all possible cross products of the regressors. However, in a
large data set with many explanatory variables, it might be difficult to include all of
these cross product terms. Hence, when the model has a lot of regressors, a simplified
version of the White test is applied in which the cross products are omitted from the
auxiliary regression, i.e. only squares of the regressors are used.
Table 5.3 illustrates the BPG test and the White test with no cross terms. In both
test, the F-statistic is a redundant variable test for the joint significance of regressors
in the auxiliary regressions. The test statistic (Obs*R-squared) is asymptotically
distributed as a 2 with degrees of freedom equal to the number of coefficients
- 78 -
(excluding the constant) in the regression. The third statistic is the explained sum of
squares from the auxiliary regression with the 2 distribution.
Table 5.3: Heteroscedasticity test
Heteroskedasticity Test: Breusch-Pagan-Godfrey
F-statistic
Obs*R-squared
Scaled explained SS
1.786648
13.75657
4.668882
Prob. F(9,26)
Prob. Chi-Square(9)
Prob. Chi-Square(9)
0.1194
0.1313
0.8622
Prob. F(9,26)
Prob. Chi-Square(9)
Prob. Chi-Square(9)
0.1297
0.1398
0.8681
Heteroskedasticity Test: White
F-statistic
Obs*R-squared
Scaled explained SS
1.741053
13.53751
4.594537
Source: Eviews 8.0
It can be clearly seen that the results from the two tests are similar to each other,
with very small differences in the value of test statistic as well as p-value. Since the
values of Obs*R-squared and Scaled explained SS are smaller than 2 (0.05,9)
(16.919), both of the tests give the same conclusion that the null hypothesis can not be
rejected, i.e. the regression has no heteroscedasticity problem.
One more thing that needs to be considered is that it is possible for the errors’
variance to change over time rather than systematically with one of the explanatory
indicators (Brooks, 2008). This phenomenon is known as autoregressive conditional
heteroscedasticity (ARCH), and the test for ARCH in the residuals is also a LM test.
This specification is motivated by the magnitude of residuals appeared to be related to
the magnitude of recent residuals in the context of financial time series. The ARCH is
conducted with the null hypothesis of no ARCH up to order q in the residuals. Like
the way to choose pth order in the autocorrelation test, q is set equal to 4 because of
the quarterly data.
- 79 -
Table 5.4: ARCH test
Heteroskedasticity Test: ARCH
F-statistic
Obs*R-squared
1.962788
7.208853
Prob. F(4,27)
Prob. Chi-Square(4)
0.1288
0.1253
Source: Eviews 8.0
Unlike the former two test, the F-statistic of the ARCH test is an omitted
variable test for the joint significance of all lagged squared variables and also is
shown for comparison purpose. The LM test statistic, Obs*R-squared, has 2 (q)
distribution under general conditions. The result is that there is no evidence for the
presence of ARCH up to order 4 since the p-value is considerably in excess of 0.05.
5.2.3. Testing for Normality
The F and t significance tests in the regression model are based on the normality
assumption of the disturbances. So it is necessary to conduct a test for the normal
distribution that is symmetric about its mean. One of the most common applied test is
Jarque – Bera (JB), which is relied on the statistic of skewness and kurtosis of the
residuals. While skewness is a measure of asymmetry of the distribution of the series
around its mean, kurtosis measures the peakedness or flatness of the distribution.
Figure 5.1: Normality test
7
Series: Residuals
Sample 2005Q2 2014Q1
Observations 36
6
5
4
3
2
1
Mean
Median
Maximum
Minimum
Std. Dev.
Skewness
Kurtosis
-2.18e-12
24.40043
311.9080
-276.2721
156.5719
0.194310
2.301341
Jarque-Bera
Probability
0.958724
0.619178
0
-300
-200
-100
0
100
200
300
Source: Eviews 8.0
- 80 -
In terms of the normality, the skewness and kurtosis are expected to be 0 and 3
respectively. Under the null hypothesis of a normal distribution, the JB statistic is
distributed as 2 (2) and would not be significant. This means that the probability
should be bigger than 0.05 to not reject the null of normality at the 5% level. As
shown in Figure 5.1, p-value (0.619) is much larger than 0.05, thus the residuals is
likely to be normally distributed. However, they are not perfectly followed normal
distribution. Because of the positive skewness and the less – than – three kurtosis, the
residuals’ distribution has a long right tail and rather flat relative to the normal.
Additionally, the linear regression embodies an implicit assumption that the
estimated coefficients are constant for the entire sample. This assumption can be
checked by parameter stability tests. Two important stability tests that are CUSUM
and CUSUMSQ are derived from the residuals of the recursive estimation. According
to Brooks (2008), the CUSUM statistic is based on a normalised version of the
cumulative sums of the residuals, whereas the CUSUMSQ test is relied on a
normalised version of the cumulative sums of the squared residuals. Both tests have
the null hypothesis of parameter stability and have a set of ±2 standard errors bands
plotted around zero.
Figure 5.2: Stability test
20
15
10
5
0
-5
-10
-15
-20
2006
2007
2008
2009
CUSUM
2010
2011
2012
2013
2014
5% Significance
- 81 -
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
2006
2007
2008
2009
2010
CUSUM of Squares
2011
2012
2013
2014
5% Significance
Source: Eviews 8.0
Generally, since the line represented the residuals and the squared ones lies well
within the confidence bands, the null hypothesis of stability is not rejected in these
tests. It is suggested that the residual variance is relative stable for the data period.
5.2.4. Testing for Multicollinearity
The independent variables using in the OLS estimation must not be correlated to
each other. In other words, if any variable is added or removed from the regression,
the values of the coefficients on the other variables would not be changed. The
problem of multicollinearity arises when there are highly correlated among two or
more explanatory variables. The multicollinearity will causes several serious concerns.
According to Brooks (2008), first, the regression has high coefficient of determination
(R2 ), which indicates how well the dependent variable explained by its independent
variables, but insignificant individual coefficients. Second, the regression becomes
very sensitive to small changes in the specification, so that adding or removing an
explanatory variable leads to large changes in the coefficient values or significances
of the other variables. Finally, confidence intervals for the parameters tend to be very
wide which in turn might make significance tests give inappropriate inferences.
- 82 -
To detect the multicollinearity problem, this study examines the correlation
matrix of the individual variables in order to find out whether there is any pair – wise
correlation which is in excess of 0.8 or not. As illustrated in Table 5.5, there are three
pairs of independent variables that have the correlation over 0.8, including JPY and
INR (0.855), JPY and EUR (0.838), EUR and FXR (0.833). That means the
regression may have significant multicollinearity. Therefore, to solve this problem,
the regressors most affected by multicollinearity are removed.
Table 5.5: Correlation matrix
GDP
INF
INR
CUR
FDI
FXR
EUR
JPY
GDP
1.000000
-0.323235
-0.639390
-0.362531
-0.516720
-0.621855
-0.643775
-0.640986
INF
1.000000
0.569377
-0.075924
0.429346
0.217290
0.288625
0.219870
INR
CUR
FDI
1.000000
0.180509 1.000000
0.705941 -0.021791 1.000000
0.581617 0.189768 0.743119
0.781315 0.013359 0.768094
0.855860 0.172029 0.596723
FXR
1.000000
0.833320
0.598173
EUR
JPY
1.000000
0.838950 1.000000
Source: Eviews 8.0
This research conducts six auxiliary regressions in pairs without one of the four
above explanatory variables so that only one regression with smallest R2 could be
chosen. Put in another way, if one particular variable is eliminated from the original
regression that causes to the largest reduction of R2 compared to the others, this
variable will be remained in the regression. The result is that R2 from the auxiliary
regressions without INR, JPY, EUR and FXR are 99.44%, 99.42%, 99.37% and 99.30%
respectively. Thus, this study keeps FXR as the main regressors as well as removes
the other three variables.
Nevertheless, the problem with this solution is that the estimators of the new
model would be biased if the original model was correct. This is likely due to the
phenomenon of omitted variables, which can make significant contribution to explain
- 83 -
the variation of the dependent variable but are not included in the regression. The test
for omitted variables is managed with the null hypothesis is that the additional set of
regressors are not jointly significant. In this case, we run a regression model without
INR, JPY and EUR, and then examine to see whether these three variables are omitted.
Table 5.6: Omitted variables test
Omitted Variables Test
Specification: EXR C EXR(-1) GDP CUR INF FDI FXR
Omitted Variables: INR EUR JPY
F-statistic
Likelihood ratio
Value
1.244479
4.830325
df
(3, 26)
3
Probability
0.3138
0.1847
Source: Eviews 8.0
In Table 5.6, the output from the test is an F-statistic and a Likelihood ratio (LR)
with their associated p-value. The F-statistic is based on the difference between the
residual sums of squares of the restricted and unrestricted regressions while the LR
statistic is the difference between the maximized values of the log likelihood function
of the unrestricted and restricted regressions. Under H0 , the LR statistic has an
asymptotic 2 distribution with degrees of freedom equal to the number of added
variables. As the p-values of these two statistic (0.313 and 0.184) are both bigger than
0.05, neither test can reject the null hypothesis, which means that INR, EUR and JPY
do not play any vital roles in explaining EXR. Hence, they could be eliminated from
the original regression.
5.2.5. Testing for Redundant Variables
In the new regression model (without INR, EUR and JPY), the problem of
multicollinearity is disappeared, but there are only three out of six indepe ndent
variables that have statistically significant coefficients with EXR, including lagged
- 84 -
EXR (EXR(-1)), GDP and CUR. Because of the insignificance of the other three
regressors (INF, FDI, FXR), this raises a concern of the redundant variable problem in
the regression. Therefore, it is necessary for a test to decide whether these regressors
all have zero coefficients and might be deleted from the equation.
Table 5.7: Redundant variables test
Redundant Variables Test
Specification: EXR C EXR(-1) GDP CUR INF FDI FXR
Redundant Variables: INF FDI FXR
F-statistic
Likelihood ratio
Value
1.520396
5.258680
df
(3, 29)
3
Probability
0.2301
0.1538
Source: Eviews 8.0
The test is conducted with the null hypothesis of this subset of variables are
redundant. The reported test statistics are the F-statistic and the LR distributed as a 2
with degrees of freedom equal to the number of excluded variables under H 0 . Both of
them has concluded that INF, FDI and FXR are redundant variables as the p-values
(0.23 and 0.15) are in excess of the critical level (5%). Thus, these variables should be
removed from the regression.
5.2.6. Testing for Endogeneity
As mentioned earlier, the endogeneity problem can be caused by three main
instances in which the simultaneous causality is mainly focused on. The significant
macro determinants of the regression model are GDP and CUR. So we test these
variables to see whether they are endogeneous ones b y using Hausman test. Firstly,
the instrumental variables are determined that are foreign direct investment (FDI) for
GDP, and exports (EXPO) and imports (IMPO) for CUR. Due to the time limitation
and the difficulty of searching for the data, these instrumental variables are the most
- 85 -
appropriate ones. Secondly, we run the two instrumental regressions with the estimate
equations presented below.
GDP = 8.020475 – 0.710831*FDI + V1
(p-value = 0.0011)
CUR = 1195.322 + 1.043768*EXPO – 1.027827*IMPO + V2
(p-value = 0.00)
(p-value = 0.00)
Thirdly, we get the residuals (V1 and V2) as extra regressions in our original model
and use their significance to test the null hypothesis that the macro determinants are
not endogeneous variables.
Table 5.8: Hausman test
Dependent Variable: EXR
Method: Least Squares
Sample (adjusted): 2005Q2 2014Q1
Included observations: 36 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
EXR(-1)
GDP
CUR
V1
C
0.962099
-130.3361
-0.026444
23.67398
1666.589
0.022419
60.73226
0.014834
62.43997
731.4171
42.91407
-2.146078
-1.782669
0.379148
2.278575
0.0000
0.0398
0.0844
0.7072
0.0297
Variable
Coefficient
Std. Error
t-Statistic
Prob.
EXR(-1)
GDP
CUR
V2
C
0.953598
-117.5570
-0.015559
-0.069250
1751.320
0.020925
29.88689
0.015059
0.038208
520.1206
45.57190
-3.933395
-1.033194
-1.812445
3.367141
0.0000
0.0005
0.3098
0.0799
0.0021
Source: Eviews 8.0
We notice that the p-values for both V1 and V2 are larger than the significance
level of 5%. As a result, we can not reject the null hypothesis that means GDP and
CUR are exogenous and the OLS remains consistent.
- 86 -
5.2.7. Final Regression Model
After taking six independent variables out of the regression due to the problems
of multicollinearity and redundant variables, there are only two macroeconomic
indicators like GDP and CUR as well as EXR(-1) that are needed to analyze the
influences on EXR. Table 5.9 demonstrates the final result for the regression model
with no serial correlation, no heteroscedasticity, no endogeneity, normality and
stability with the data set at 5% level of significance.
Table 5.9: Regression model
Dependent Variable: EXR
Method: Least Squares
Sample (adjusted): 2005Q2 2014Q1
Included observations: 36 after adjustments
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
EXR(-1)
GDP
CUR
1467.871
0.965895
-110.4883
-0.028418
503.2969
0.019789
30.37544
0.013703
2.916511
48.80895
-3.637422
-2.073872
0.0064
0.0000
0.0010
0.0462
R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
F-statistic
Prob(F-statistic)
0.992652
0.991963
188.3772
1135551.
-237.5458
1440.885
0.000000
Mean dependent var
S.D. dependent var
Akaike info criterion
Schwarz criterion
Hannan-Quinn criter.
Durbin-Watson stat
18576.11
2101.218
13.41921
13.59516
13.48062
1.258436
Source: Eviews 8.0
As we can see, the R2 value is 0.9927 meaning that nearly 99.27% of the
variation of EXR can be explained by EXR(-1), GDP and CUR. This is a very high
value since the maximum value of R2 is 1, which makes the model significantly
meaningful. Moreover, the estimated coefficients of GDP and CUR have negative
signs that are the same as this study’s expectation. Noted that the increase in Fstatistic from 517.401 to 1440.885 indicates that the new model is more preferred.
- 87 -
5.3. Macro Stress Testing for Exchange Rate Exposure
In the second stage of this research, the stress testing is conducted into the VaR
model with the use of the three particular approaches, such as historical simulation,
variance – covariance method and Monte Carlo simulation. We predict the change of
EXR in the next quarter, and then apply it to calculate the FX exposure of a
hypothetical company. Using the results from the regression analysis, the EXR is
forecasted by the following equation:
EXRt = 1467.871 + 0.965895*EXRt-1 - 110.4883*GDPt - 0.028418*CURt
We also assume that the hypothetical company currently has a shor t FX position of
USD 1 million and the exchange rate is USD/VND 21,246. Therefore, the capital
charged for this position is VND 21,246 million.
5.3.1. Variance – Covariance Approach
This method assumes that the changes in the exchange rate as well as the macro
determinants are normally distributed. Hence, we only need to estimate two factors
including mean and standard deviation to calculate VaR at the tail of the distribution.
The volatility of the exchange rate determined by ∆ GDP and ∆ CUR is given by:
where
w1 is the weighting of ∆ GDP
w2 is the weighting of ∆ CUR
1 is the standard deviation or volatility of ∆ GDP
2 is the standard deviation or volatility of ∆ CUR
1,2 is the correlation coefficient between ∆ GDP and ∆ CUR
- 88 -
VaR can be expressed as:
where
p is the weighted mean of the exchange rate
p is the standard deviation of the exchange rate
z is the left – tail percentile of a standard normal distribution
P is the FX position
In the regression model, GDP is statistically significant at both 1% and 5%
while CUR is only meaningful at 5% significance level. Probably, the changes of
GDP has more impact on the volatility of EXR than CUR’s. Thus, this study supposes
that the weightings for GDP and CUR are 70% and 30% respectively.
Table 5.10: VaR using the variance – covariance approach
Confidence level
VaR (% change)
VaR (VND million)
95%
-1.64%
-348
99%
-2.41%
-513
Source: Excel 2007
With 95% confidence, we expect that the increase of the exchange rate in the
next quarter will not exceed 1.64%, which also means that VND is predicted not to
depreciate over 1.64% compared to USD. From a different aspect, our worst quarterly
loss will not exceed VND 348 million. Put another way, the company needs no more
than VND 348 million to cover the short position of USD 1 million at 95%
confidence. Similarly, the exchange rate does not go up in excess of 2.41%, and thus,
the expected loss suffered by the company does not go beyond VND 513 million with
99% confidence.
- 89 -
5.3.2. Historical Simulation
To run historical simulation, we calculate the historical percentage changes of
GDP and CUR during the period of 2005Q1 to 2014Q1. Next, we apply these changes
to compute the current mark-to- market value of the two determinants, and then revalue the EXR. After that, the changes in hypothetical value of EXR in comparison
with the current exchange rate are sorted from the smallest to the largest. Finally, the
expected loss, which is an additional amount of money the company has to pay in
case the exchange rate increases, is computed corresponding to each value of EXR so
that we can determine VaR with the confidence level of 95% and 99%.
Table 5.11: VaR using the historical simulation approach
Confidence level
VaR (% change)
VaR (VND million)
95%
-1.49%
-321
99%
-1.80%
-390
Source: Excel 2007
As shown in the above table, in the 2014Q2, there is a 5% probability that the
exchange rate will increase more than 1.49%, or the expected loss on the short
position will be larger than VND 321 million. Likewise, there is only a 1% probability
that the exchange rate will exceed 1.80%, or the loss will go over VND 390 million.
5.3.3. Monte Carlo Simulation
In comparison with the first two approaches, Monte Carlo simulation is the most
sophisticated method to calculate VaR due to its being computationally intensive. It
involves conducting thousands (or millions) of repeated trials of the values of the
uncertain inputs based on some known probability distributions in order to produce
the distribution for the output (Cheung and Powell, 2012). In this study, the uncertain
- 90 -
variables are the percentage changes of GDP and CUR each quarter, and the outputs
are the volatitily of the exchange rate and the FX expo sure of the hypothetical firm.
With the technical support of the program Frontline Solvers 2014, we explore
that the most fitted distributions of GDP’s and CUR’s changes are Beta and Min
Extreme distribution as followed:
% ∆ GDP ~ Beta (4.464, 4.864)
% ∆ CUR ~ MinExtreme (5.242,3.013)
In our simulation, 100,000 trials of quarterly changes are conducted to generate
hypothetical values of GDP and CUR that will be used to compute the same number
of mark-to-market values of the exchange rate for the next quarter, which create the
distributions for the related outputs. Based on these distributions, we can calculate
VaR for any preferred confidence level.
Table 5.12: VaR using Monte Carlo simulation
Confidence level
VaR (% change)
VaR (VND million)
95%
-2.03%
-431
99%
-2.92%
-621
Source: Frontline Solvers Add-in Spreadsheet
The result from Table 5.12 indicates that EXR is predicted to rise less than
2.03% and 2.92% with the level of 95% and 99% confidence, respectively. Moreover,
the company incurs to loss less than VND 431 million and VND 621 million at 95%
and 99% confidence, respectively.
Comparison of the three VaR approaches
Looking back Table 5.10, 5.11 and 5.12, it is clearly that at any confidence level,
the smallest VaR is computed by the historical simulation, and the largest VaR
- 91 -
belongs to the Monte Carlo approach. Especially, with 99% confidence level, VaRs of
Monte Carlo simulation and variance – covariance matrix are much bigger than the
one of the historical simulation. Furthermore, the higher the confidence level is, the
larger the VaR number will be. That means the larger expected loss the company has
to face, the smaller percentage of the time it can happen. To be conservative in our
risk management process, we would choose the highest loss to reserve for, i.e. the
Monte Carlo approach.
5.3.4. Stress Testing
From the three previous sections, it is easy to infer that stress testing of the
foreign exchange risk is likely to be based on quantifying the size of this loss
calculated as the product of the expected change in the exchange rate and the open
position value. Consequently, the change in the capital adequacy should have been
reserved from the company’s own funds in order to cover the predefined percentage
level of its losses over the next quarter.
There is a number of stress testing techniques that measure the magnitude of
market factor changes. Sensitivity tests isolate the porfolio impact of movement in a
single factor, and thus, the stressed loss is the marginal contribution of that factor
(Basu, 2006). However, they do not take into account the correlations of the market
factors that usually move together. Scenario analysis can fix this problem since it
captures the effect of a relevant event on a group of factors. Hence, the stressed loss is
a combined impact of simultaneous movement of these factors. As a result, this study
conducts either sensitivity tests by involving ∆ GDP shock and ∆ CUR shock or
scenario analyis by shocking these two determinants’ changes at the same time.
Moreover, the time horizon of stress testing is the next four quarter s from 2014Q2 to
2015Q1, and the macro conditions in 2014Q1 are taken as the current environment.
- 92 -
5.3.4.1. ∆ GDP Shock
As GDP has negatively related with EXR, we put stress on the downward
movements of GDP for four consecutive quarters. In each scenario, when the changes
of GDP is shocked, a proxy of CUR’s changes will be calculated as the average of the
percentage change of the four corresponding previous quarters. For instance, the
change of CUR in 2014Q2 is equivalent to the average of the ones in 2010Q2,
2011Q2, 2012Q2 and 2013Q2. Moreover, the largest shock for ∆ GDP is about three
standard deviation (-3).
Table 5.13: Predicted EXR under ∆ GDP shock using historical simulation
Quarters
% ∆ GDP
% ∆ CUR
GDP (%)
CUR (mil)
EXR
2014Q2
-20%
-19.80%
3.97
1,142.92
21,518
2014Q3
35%
29.61%
5.70
1,846.95
21,197
2014Q4
-17%
17.54%
4.12
1,675.00
21,487
2015Q1
-51%
-71.83%
2.43
401.49
21,709
Source: Excel 2007
However, this calculation is just applied for the historical simulation. The unshocked CUR’s changes will be obtained randomly based on its distribution in Monte
Carlo simulation, which is more precise than the former one. Besides, we determine
again the distribution of GDP’s changes when employing these shocks that turns into
Beta General, while the distribution of CUR’s changes remains unchanged.
% ∆ GDP ~ BetaGen (5.023, 4.594, -0.553)
% ∆ CUR ~ MinExtreme (5.242,3.013)
The simulated changes of EXR and expected loss at 99% confidence level are
examined by using the historical simulation (HS) and Monte Carlo (MC) simulation
presented in the following table.
- 93 -
Table 5.14: VaR unde r ∆ GDP shock
VaR (% change)
VaR (VND million)
Confidence level
HS
99%
MC
-2.00%
HS
-2.98%
MC
-434
-634
Source: Excel and Frontline Solvers
With some stress scenarios on the percentage changes of GDP, VaRs become
bigger than the values in the normal market. Under HS, there is 1% of time that the
loss due to exchange rate is 2% or the maximum loss is about VND 434 million,
whereas the two figures under MC are 2.98% or VND 634 million. As before, the
VaRs of MC are higher than the HS’s. To be conservative in our risk management, we
would choose the higher number to reserve for.
5.3.4.2. ∆ CUR Shock
Likewise the case of the ∆ GDP shock, we also stress on the adverse fluctuation
of CUR due to its negative relationship with EXR. The biggest shock is approximate
to three standard deviation in the last quarter of the forecasted year. The un-shocked
GDP’s changes is computed as the average of the ones of the same period in the four
previous years in the historical simulation, which is similar to the estimation of
CUR’s changes in the previous section, as well as is run randomly followed its
distribution in Monte Carlo simulation.
Table 5.15: Predicted EXR under ∆ CUR shock using historical simulation
Quarters
% ∆ GDP
% ∆ CUR
GDP (%)
CUR (mil)
EXR
2014Q2
6.38%
-218.75%
5.28
1,692.00
21,454
2014Q3
8.46%
198.90%
5.38
4,259.00
21,274
2014Q4
4.19%
-116.50%
5.17
-235.01
21,425
2015Q1
-19.96%
-1,009.00%
3.97
-12,953.15
21,919
Source: Excel 2007
- 94 -
Furthermore, the new distribution of ∆ CUR under stress scenarios in Monte
Carlo simulation that is tested for a second time and the unchanged distribution of ∆
GDP are shown below.
% ∆ GDP ~ Beta (4.464, 4.864)
% ∆ CUR ~ MinExtreme (13.262, 3.178)
The final step is that VaRs for the EXR’s changes and the loss from the short
position are calculated at the level of 99% confidence.
Table 5.16: VaR unde r ∆ CUR shock
VaR (% change)
VaR (VND million)
Confidence level
HS
99%
-2.56%
MC
HS
-3.14%
MC
-560
-666
Source: Excel and Frontline Solvers
As we can see, VaRs under ∆ CUR shock are larger that VaRs under ∆ GDP
shock, especially when using the historical simulation. The reason might be that under
HS case, the assumed stress scenarios for CUR’s changes could be relatively more
severe than the ones for GDP’s changes. For example, during the study period, the
lowest GDP is 3.14% with the largest decrease of 46.69%, but the shock reduction for
GDP is only 51%. On the other hand, the worst scenarios for CUR’s deficit in this
study is about USD 12,950 million, which is nearly double the biggest deficit in the
past. In addition, we do not have to make an assumption about the proxy value for the
un-shocked determinants, but base on their fitted distributions for estimation.
Therefore, VaRs under MC case are just higher than the ones in the previous section
with an acceptable amount. Once again, Monte Carlo simulation is a more reliable
and conservative approach than the historical simulation.
- 95 -
5.3.4.3.∆ GDP and ∆ CUR Shock Together
After conducting the sensitivity tests for GDP’s and CUR’s changes, we try to
consider these macro determinants’ shocks at the same time. We simply put it all
together and the shocks would be applied simultaneously. In other way, we assume
that the stress scenarios for ∆ GDP and ∆ CUR could happen in the same period of
time. The purpose is to examine the volatility of the exchange rate and the fluctuation
of the VaRs.
Table 5.17: Predicted EXR using historical simulation
Quarters
% ∆ GDP
% ∆ CUR
GDP (%)
CUR (mil)
EXR
2014Q2
-20%
-218.75%
3.97
1,692.00
21,599
2014Q3
35%
198.90%
5.70
4,259.00
21,128
2014Q4
-17%
-116.50%
4.12
-235.01
21,541
2015Q1
-51%
-1,009.00%
2.43
-12,953.15
22,089
Source: Excel 2007
As in Table 5.17, the forecasted EXR is not changed much in the last three
quarters of 2014 since GDP decreases but not CUR, except for the third quarter when
both GDP and CUR increase in value that makes the exchange rate become smaller
than that in the two sensitivity tests. However, in the first quarter of 2015, as either
GDP or CUR reduce considerably, the EXR rises significantly that means VND
depreciates compared to USD. In addition, the distributions of ∆ GDP and ∆ CUR in
Monte Carlo simulation are presented again.
% ∆ GDP ~ BetaGen (5.023, 4.594, -0.553)
% ∆ CUR ~ MinExtreme (13.262, 3.178)
Finally, as usual, we calculate the value of VaRs for the exchange rate’s change
and the expected loss.
- 96 -
Table 5.18: VaR unde r ∆GDP and ∆ CUR shock together
VaR (% change)
VaR (VND million)
Confidence level
HS
99%
MC
-3.01%
HS
-3.20%
MC
-662
-679
Source: Excel and Frontline Solvers
It is clear that with both stress scenarios for the changes of GDP and CUR
applied at the same time, VaRs become much bigger for the case of the historical
simulation, but not for Monte Carlo simulation where these values just fairly higher
than the two previous tests. Nevertheless, the VaR in this situation is less than the sum
of the individual VaRs computing separately under each kind of shock. This is
probably because the correlation of ∆ GDP and ∆ CUR is not perfect, even though it
increase with the shocks, from 0.1036 in normal condition to 0.1461 in abnormal
circumstance. Thus, one advantage of the historical simulation and Monte Carlo
simulation against the variance – covariance approach is that there is no need to
compute any measure of dependence like correlation because the association is
implicit in the market data (Basu, 2006). The losses of the same degree due to t he
movements of each determinant are not clustered anyway. It expresses the non –
additivity characteristic of the VaR measure.
In conclusion, macro stress testing can be performed to assess the vulnerability
and risk exposure of the company’s foreign exchange open position. On the one hand,
VaR is used to measure unexpected loss in normal markets. On the other hand, stress
scenarios are examined for abnormal losses. With all the assumptions given, it is
suggested that the hypothetical company should set aside a minimum capital amount
of about VND 680 million on the short position of USD 1 million so that it can cover
99% of its losses over the next quarter.
- 97 -
CHAPTER 6
CONCLUSION AND RECOMMENDATION
This ending chapter provides conclusion about the the research’s results that meet the
overall objectives and answer the research questions at the beginning. We then makes
some suggestions to companies which desire to gauge and set aside capital reserve
for their foreign exchange exposure.
6.1. Conclusions
This study investigates the relationship between the exchange rate and its macro
determinants as well as applies macro stress testing combined with VaR calculation
for a hypothetical Vietnamese enterprise but it can also be easily applied to any
Vietnamese firms to estimate their foreign exchange exposure. These are the two
main objectives of the research.
To achieve the first aim and answer the first research question, based on the
results of several international studies, we reasonably choose nine market factors
including GDP, INF, INR, CUR, FDI, FXR, EUR, JPY, and historical EXR as the
independent variables for our model. We run the linear regression model with the
default method of Ordinary Least Squares (OLS) to examine the impact of these
macro indicators on the exchange rate. The time-series data used for analysis stretches
for 37 quarters (from 2005Q1 to 2014Q1) that can be considered as a large sample.
Before coming up with the final result, we conduct a number of various tests for
serial correlation, heteroscedascity, normality, stability, endogeneity, multicollinearity
as well as appropriate variables for our study. As a result, there are only three
variables, GDP, CUR and one – lagged EXR should be included in the model, which
provides a very good R2 value of 0.992652. That means the model is incredibly
- 98 -
significant with nearly 99.27% of the exchange rate variation explained by GDP,
CUR and EXR(-1). The resultant regression equation can be presented as:
EXRt = 1467.871 + 0.965895*EXRt-1 - 110.4883*GDPt - 0.028418*CURt
We found that all the signs of the three variables’ coefficients are the same as
our initial expectation. To be specific, the current exchange rate is positively related
to its value in the previous period. This is due to the fact that the SBV tends to adjust
the official exchange rate within the trading band based on the lagged one. Moreover,
GDP and CUR have negative signs, which mean that with all things hold constant,
any improvement either in the GDP growth rate or in the surplus current account will
lead to the decrease in the exchange rate. In other words, VND would be appreciated
in the good market environment (high GDP and large CUR). This result is consistent
with several previous research such as Wilson (2009), Nucu (2011), and Sinha and
Kohli (2013).
In addition, while EXR(-1) and GDP are statistically signficant at both 1% and
5% level of significance, CUR is only meaningful at significance level of 5%. So it
could be said that the influence of CUR on the exchange rate is rather weak compared
to the other variables. However, there is very little or no relationship between EXR
and the other indicators in the original model, which points out that the differences of
Vietnam’s interest rate and inflation rate from the ones of the US, foreign direct
investment, foreign exchange reserves, the exchange rate of VND/EUR and
VND/JPY have no impact on the VND/USD in Vietnam during the period from
2005Q1 to 2014Q1.
With respect to the second objectives and the other two research questions, we
combine macro stress testing with Monte Carlo simulation and VaR estimation to
assess the foreign exchange exposure of a hypothetical company which has a short
- 99 -
position of USD 1 million. Note that the methodology can be easily adapted to any
company. We assume that the current exchange rate is VND/USD 21,246 as well as
the value of GDP and CUR are 4.96% and VND 1,425 million respectively.
Therefore, the capital required for this positio n is VND 21,246 million.
Before stress test is performed, VaR is calculated in the normal market by using
the three different approaches including variance – covariance matrix, historical
simulation and Monte Carlo simulation. Among them, the variance – covariance
method is the simplest one, whose fundamental assumption is the normal distribution
of the market factors’ fluctuation. The historical simulation assumes that the historical
events are likely to be repeated in the future, and make no hypothesis about the
distribution. The Monte Carlo simulation differs from the historical simulation is that
it applies stochastic distribution of the variables in order to run thousands or millions
of trials to estimate the distribution of the exchange rate volatility as well as the
expected loss due to the depreciation of VND. For instance, in this study, GDP’s
changes have the Beta distribution while the one of CUR’s changes is Min Extreme.
The results of these three VaR approaches are quite similar regarding the higher
confidence level will lead to the larger VaR of the exchange rate’s depreciation and
the predicted loss. This indicates that there is small probability for the company to
deal with the substantial loss in the normal market. However, there is different in the
value of VaR calculated by these methods. To be specific, the smallest VaR is
computed by the historical simulation while the biggest VaR belongs to Monte Carlo
simulation. For example, at 95% confidence level, the exchange rate can not increase
over 1.49% or the loss does not exceed VND 321 million under the historical
approach, whereas those figures under Monte Carlo simulation are 2.03% or VND
431 million respectively.
- 100 -
The next step of the research’s second stage is to apply macro stress testing to
create adverse scenarios so that we can compute VaR in the abnormal markets. Since
we want to capture the fat tails of the market changes, we eliminate the variance –
covariance approach from this step. Hence, we just use the other two VaR measures to
evaluate the company’s risk exposure. In addition, both sensitivity test and scenario
analysis are carried out in this stage, which means we apply ∆ GDP shock and ∆ CUR
shock successtively, and then shock them simultaneously. The stress time horizon is
one year that extends from 2014Q2 to 2015Q1. Under the hypothetical scenarios, the
distributions of ∆ GDP and ∆ CUR have changed into the Beta General and Min
Extreme with different parameters respectively.
The marco stress testing provides four important findings. Firstly, as in the
normal condition, in our hypothetical case, the Monte Carlo simulation always gives
larger VaR for the exchange rate variation and the expected loss than the historical
simulation. Secondly, when we shock both ∆ GDP and ∆ CUR at the same time, the
VaRs under the HS case are much bigger than the ones in each individual shock,
while they are just relatively higher under the MC case. This is because we attempt to
compute the proxy of the un-shocked determinant’s changes as the average value of
four corresponding previous intervals in the historical simulation. In contrast, we
simply use random value of this variable based on its distribution in Monte Carlo
simulation that makes the VaR estimation more accurate.
Thirdly, the sum of VaRs in each sensitivity test is not equal to the one in the
scenario analysis even though we keep the same level of the shock. One of the reasons
for this non – additivity characteristic of VaR might be the unperfected correlation
within ∆ GDP and ∆ CUR. Thus, the losses caused by the movement of each factors
can not be clustered. Finally, under the shock of the two macro determinants’ changes
- 101 -
together at 99% confidence level, the hypothetical company’s capital reserve to cover
the abnormal loss is about VND 680 million so that it can enhance its business
stability and efficiency in the adverse scenarios.
6.2. Recommendations
The empirical results of this study indicates that Vietnamese firms should pay
attention to several macroeconomic variables like GDP growth rate and current
account balance which mainly affect the volatility of the exchange rate. They can also
use VaR approaches to calculate the percentage change of the exchange rate to see
what the maximum loss they are likely to suffer in the normal market with a certain
confidence level. Then, they need to set forth some stress scenarios to forecast the risk
exposure that they have to face in order to reserve enough capital from their own fund
to cover the abnormal loss during stressed market condition.
According to Vu Minh (2013), most of the large corporations in Vietnam have
been used the risk management services of financial consulting companies like PWC
or McKenzie which employ very complex quantitative risk models. Nevertheless,
small and medium enterprises (SME) have not still taken shape of risk controlling
thoughts in their minds since they do not understand fully the benefits of building a
mathematical approach for measuring risk exposure. This problem could be explained
by several reasons. Firstly, these firms’ subjectiveness is led by the guarantee of the
SBV with regard to the exchange rate fluctuation during a year. Moreover, the SBV
ban companies from borrowing in foreign currency if they have no foreign
denominated revenues. Secondly, the high fees of external consulting services makes
them ignore the risk managing process. Finally, the SMEs usually focus more on
achievable short-term targets than long-term strategies, and thus they tend to expose
to less foreign exchange risk in the short run than in the long run.
- 102 -
By applying macro stress testing associated with the VaR calculation, the
enterprises would be able to handle the above problems in a systematical way. By the
time the SBV accepts the more flexible exchange rate regime, the firms could be more
confident in dealing with the exposure because they already have the trustworthy
approaches to estimate the loss beforehand. For example, if the company knows its
expected loss in the future, it can choose the most appropriate derivaritves instruments
such as forward contracts and options, etc. for hedging purpose, or it just simply set
aside an adequate capital to cover the loss.
Additionally, in order to avoid high consulting fees frequently paid for other
financial organisations, the enterprises can establish a small deparment or group
which specializes in controlling risk by conducting stress testing. The employees or
group members could be sent for taking a crash course of risk managing techniques so
that they would gain professional macroeconomic knowledge and econometric skills
in performing some analytical supporting softwares like Eviews and Frontline Solvers
Spreadsheet. When the companies have a strong base of risk administration, they
should look further into the future to set a clear long-term goal for their business
development.
One benefit for the imports companies to apply stress test is that they would not
worry about the supplemental capital due to the exchange difference if the y have to
borrow VND from the banks to buy USD for trade payment. In other words, they
should prepare funds for the predicted loss calculated by the stress model. The banks
might apply a lower interest rate than usual for those enterprises which have
implemented market risk managing methods, or they might be allowed to borrow in a
foreign currency even when they have no foreign denominated revenues.
Furthermore, the exports and imports firms that effectively utilizes foreign exchange
- 103 -
risk estimating techniques are likely not only to reduce economic losses but also to
generate competitive advantages in either domestic or foreign markets, particularly in
the context of the wide and deep international economic integration nowadays.
Another challenge for stress testing application in the non – financial firms is
the quality and availability of the related data. For instance, to create an extreme but
plausible stress scenario, the company needs at least the data for one or two business
cycles, which is about 10 – 12 years (Pham Do Nhat Vinh, 2012). Therefore, the
government and the relevant ministries need to enhance its ability in the transparency
of information and provide sufficient economic data as well as statistics for a long
period of time, especially about macro variables that play an important role in setting
up the risk model and analyzing the scenarios.
6.3. Suggestions for Future Research
Firstly, this study emphasizes the significance of the macro determinants to the
exchange rate through the linear regression model from 2005Q1 to 2014Q1. Future
studies could employ nonlinear model to test these variables’ significance in a longer
period and probably with different intervals such as annually or monthly.
Secondly, as one of this paper’s limitation, a hypothetical company’s foreign
exchange position is assumed to illustrate how to conduct stress testing and calculate
VaR at any confidence level. The latter studies can apply this framework for a
particular firm with distinguished characteristics in order to gain a comprehensive
understanding of assessing the vulnerability of a real circumstance.
Finally, the future studies could focus on micro stress testing which applies
stress scenarios to a company’s internal parameters and is conducted as part of its risk
management, rather than carrying out macro stress testing as in this research.
- 104 -
REFERENCES
Abdalla and Murinde. (1997). Exchange rate and stock price interactions in emerging
financial markets: Evidence on India, Korea, Pakistan and the Philippines.
Applied Financial Economics, 7(1): 25-35.
Anh Ngoc Nguyen and Nicholas Sarantis. (2008). The equilibrium real exchange rate
of Vietnam: Determinants and Misalignments. Discussion Papers, Centre for
International Capital Markets, London Metropolitan University. Retrieved from
http://www.londonmet.ac.uk/media/london- metropolitan- university/london- metdocuments/faculties/law-governance-and- international-relations/cicm/200801.pdf
Aron, J., Elbadawi, I.A., and Kahn, B. (2000). Real and monetary determinants of the
real exchange rate in South Africa. In: Elbadawi, I. And Hartzenberg, T. (ed.).
Development Issues in South Africa. London: Macmillan.
Asian Development Bank Institute. (2006). The reminbi exchange rate revaluation:
Theory, practice, and lessons from Japan. Tokyo, Japan: Kanamori and Zhao.
Atif, Sauytbekova and Macdonald. (2012). The determinants of Australian exchange
rate: A time series analysis. Munich Personal RePEc Archive, 42309 (31).
Retrieved from http://mpra.ub.uni- muenchen.de/42309/
Badik, P. (2005). Use of the VAR method for measuring market risks and calculating
capital adequacy. Commercial Banking, 13, 17-21.
Bahmani-Oskooee, M. And Kara, O. (2000). Exchange rate overshooting in Turkey.
Economic Letters, 68, 89-93.
Bank of England. (1999). Chapter 5: Vector autoregression models. Retrieved from
http://www.bankofengland.co.uk/publications/Documents/other/beqm/1999/five
.pdf
Bascle, G. (2008). Controlling for endogeneity with instrumental variables in strategic
management research. Strategic Organization, 6(3): 285-327.
Basu, S. (2006). The impact of stress scenarios on VaR and expected shortfall. NIBM.
Retrieve from http://ssrn.com/abstract=1091462
- 105 -
Berry, R. (2012). Value at risk: An overview of analytical VaR. J.P. Morgan
Investment
Analytics
and
Consulting.
Retrieve
from
http://intinno.iitkgp.ernet.in/courses/1239/wfiles/249706
Blaschke et al. (2001). Stress testing of financial systems: An overview of issues,
methodologies, and FSAP experiences. IMF Working Paper, WP/01/88.
Brooks, C. (2008). Introductory economics for finance (2nd edition). New York, NY:
Cambridge University Press.
Canales-Kriljenko, J. & Habermeier, K. (2004). Structural factors affecting exchange
rate volatility: A cross-section study. IMF Working Paper, WP/04/147.
Cassidy, C. and Gizycki, M. (1997). Measuring traded market risk: Value at Risk and
backtesting techniques. Research Discussion Paper 9708, Bank Supervision
Department, Reserve Bank of Australia.
Cheung, Y. H. and Powell, R. J. (2012). Anybody can do Value at Risk: A teaching
study using parametric computation and Monte Carlo simulation. Australasian
Accounting Business and Finance Journal, 6 (5): 101-118.
Chinn, M.D. (2013). Macro approaches to foreign exchange determination. La
Follette School Working Paper, University of Winconsin-Madison, 2013-005.
Retrieved
from
https://www.lafollette.wisc.edu/publications/workingpapers/chinn2013-005.pdf
Choi, J.J and A.M. Prasad. (1995). Exchange rate sensitivity and its determinants: A
firm and industry analysis of US multinationals. Financial Management, 24(3):
77-88.
Chopra, G. (2009). Stress testing financial systems: A macro perspective. Retrieved
from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1529434
Citibank. (2013). Vietnam market update. Online Academy, Securities and Fund
Services.
Retrieved
from
http://www.citibank.com/transactionservices/home/about_us/online_academy/d
ocs/vietnam_market.pdf
Crowder, W.J. (n.d.). Exchange rate theories. University of Texas. Retrieved from
http://www.uta.edu/faculty/crowder/papers/Chapter18.pdf
- 106 -
Damodaran, A. (undated). Value at risk (VAR). The Stern School of Business at NYU.
Retrieved from http://people.stern.nyu.edu/adamodar/pdfiles/papers/VAR.pdf
Danmola, R.A. (2013). The impact of exchange rate volativity on the macro economic
variables in Nigeria. European Scientific Journal, 9(7): 1-14.
Dash, P. (undated). The relationship between interest rate and exchange rate in India.
Retrieved from www.igidr.ac.in/money/mfc-13/mfc_6/mfc.../dash.doc
Dornbrush, Rudiger. (1976). Expectations and exchange rate dynamics. Journal of
Political Economy, 84, 1161-76.
Economic Committee of Congress and UNDP in Vietnam. (2013).
a
-
c
sa l ch
t c
h
u t h u (Research
Report RS – 01). Tri Th c Publisher.
Egert, B. (2010). The impact of monetary and commodity fundamentals, macro news,
and central bank communication on the exchange rate: Evidence from South
Africa. Open Econ Rev, 21, 655-671.
Eiteman, D.K., Stonehill, A.I. and Moffett, M.H. (2010). Multinational Business
Finance (12th ed.). Boston, MA: Prentice Hall.
Frenkel, Jacob A. (1976). A monetary approach to the exchange rate: Doctrinal
aspects and empirical evidence. Scandinavian Journal of Economics, 78, 200224.
Frankel, Jeffrey. (1979). On the mark: A theory of floating exchange rates based on
real interest differentials. American Economic Review, 69, 610-622.
Freiden, J., Ghezzi, P. and Stein, E. (2001). Politics and exchange rates: A cross –
country approach. The Currency Game Exchange Rate Politics in Latin America,
Washington, DC: Inter-American Development Bank, 21-63.
Fuss, R. (2007). Chapter 4: Vector autoregressive models. Financial Analysis.
Retrieved from https://www.empiwifo.uni- freiburg.de/lehre-teaching-1/winterterm/dateien- financial-data-analysis/chapter4.pdf
Gupta, A. and Liang, B. (2004). Do hedge funds have enough capital? A value-at-risk
approach. EFA 2003 Annual Conference Paper No. 376. Retrieved from
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=424370
- 107 -
Hacche, G. (1983). The deteminants of exchange rate movements. OECD Economics
and Statistics Department. Working Papers, No. 7, June 1983. Retrieved from
http://www.oecd.org/finance/monetary/1915611.pdf
Ho, C. and McCauley, R.N. (2003). Living with flexible exchange rate: Issues and
recent experience in inflation targeting emerging market economies. BIS
Working Paper, Bank for International Settlement, No. 130.
Hsieh, Wen-jen. (2009). Study of the behavior of the Indonesian Rupiah/US Dollar
exchange rate and policy implications. International Journal of Applied
Economics, 6(2): 41-50.
Huynh The Nguyen
Nguyen Quyet. (2013). M i quan h gi a t gi h i o i, l i
su t v gi c phi u t i TP.HCM. Journal of Development and Integration,
11(21): 1-5.
Hyder, Z. and Mehbood, A. (2005). Equilibrium real effective exchange rate and
exchange rate misalignment in Pakistan. SBP Conference 2005. Retrieved from
http://www.sbp.org.pk/research/conf/Session_IV_Zulfiqar_Adil.pdf
Illova, L. (2005). Stress testing of bank risks. Thesis Project, Winter Semester
2004/2005. Retrieve from http:// ies.fsv.cuni.cz/default/file/download/id/1591
International Actuarial Association. (2013). Stress testing and scenario analysis.
Ontario, Canada: Insurance Regulation Committee.
International Monetary Fund. (2006). Vietnam: 2006 Article IV Consultation. Public
Information Notice on the Excutive Board Discussion and Statement by the
Excutive Director for Vietnam, IMF Country Report No. 06/22.
International Monetary Fund. (2008, 2012). Annual report on exchange arrangements
and exchange restrictions. Retrieved from http://www.imf.org/
Joyce, J.P. and Kamas, L. (2003). Real and nominal determinants of real exchange
rates in Latin America: Short-run and long-run equilibrium. The Journal of
Development Studies, 39(6): 155-182.
Kalra, S. (2005). Note on determinants of foreign exchange rates. Retrieved from
http://people.hbs.edu/mdesai/IFM05/Kalra.pdf
- 108 -
Kalfaoglou, F. (2007). Stress testing best practices and risk management implications
for Egyptian banks. Summary of speeches at the seminar.
Karfakis, C. (2003). Exchange rate determination during hyperinflation: The case of
the Romanian Lei. Applied Financial Economics, 13, 473-476.
Karim, M.S., Lee, M. And Gan, C. (2007). Exchange rate dynamics of New Zealand.
Journal of Economic Policy Reforms, 10(3): 241-260.
Linsmeier, T. J. and Pearson, N. D. (1999). Risk measurement: An introduction to
Value at Risk. Urbana-Champaign, University of Illinois. Retrieve from
http://www.exinfm.com/training/pdfiles/valueatrisk.pdf
c
Luong Thi Nga et al. (2013).
giai
2000
–
c a ch h s ch t
2012.
Banking
t
l m h t
Academy.
Retrieved
t am
from
http://nckh.hvnh.edu.vn/upload/5830/20140106/Tacdong_tygia_3.pdf
Madura, J. (2006). International financial management (8th ed.). Mason, OH:
Thomson/South-Western.
MacDonald, R. (1998). What determines real exchange rates? The long and short for
it. Journal of International Financial Markets, Institutions and Money, 8, 117153.
MacDonald, R. (2007). Exchange rate economics – Theories and evidence. Oxon, OX:
Taylor
&
Francis
e-Library.
Retrieved
from
http://www.armeconomist.com/lecture/macro/Exchange%20Rate%20Economic
s%20McDonald.pdf
MacDonald, R. and Ricci, L. (2003). Estimation of the equilibrium real exchange rate
for South Africa. IMF Working Paper, WP/03/44. Washington: International
Monetary Fund.
McMillan, W.D. & Koray, F. (1990). Does government debt affect the exchange rate?
An empirical analysis of the US – Canadian exchange rate. Journal of
Economics and Busniess, 42, 279-288.
Mina, J. and Xiao, J. Y. (2001). Return to RiskMetrics: The evolution of a standard.
RiskMetrics Group Paper, www.riskmetrics.com
- 109 -
Mussa, M. (1984). The theory of exchange rate determination. In Bilson J. F. Editor &
Marston R. C. Editor (Eds.), Exchange rate theory and practice (13-78).
Chicago: University of Chicago Press.
c
Nguyen Duc Thanh et al. (2009).
h t th
2008. VEPR.
t am 2012. VEPR, NC-30.
Nguyen Duc Thanh et al. (2013).
ua
ht
Nguyen Thi Thu Hang et al. (2010).
a ch
ch h s ch t
h
c h h c
h t . Vietnam Centre for Economic and Policy Research (VEPR), NC-21.
c
Nguyen Thi Tuyet Nga. (2012).
tr
tr
u tr h h
h
h
ca
ht t
a tr c a t
t
h
am. Economic Doctoral
Dissertation, The Banking University of Ho Chi Minh City.
Nguyen Tran Phuc and Nguyen Duc Tho. (2009). Exchange rate policy in Vietnam,
1985 – 2008. ASEAN Economic Bulletin, 26 (2):137-163.
Nuku, Anca, E. (2011). The relationship between exchange rate and key
macroeconomic indicators. Case study: Romania. Romanian Economic Journal,
41, 127-145.
O’Brien, T. (2006). International Financial Economics – Corporate Decisions in
Global Markets (2nd ed.). London: Oxford University Press.
t
Pham Do Nhat Vinh. (2012).
h
m ts
m tra s c ch u
c a h th
t am. Retrieved from www.sbv.gov.vn/portal/
RMG. (1999). Risk management – A pratical guide. RiskMetrics Group Paper,
www.riskmetrics.com
Sajid Anwar and Lan Phi Nguyen. (2010). Foreign direct investment and economic
growth in Vietnam. Asia Pacific Business Review, 16, 183-202.
Sargeant, Nicola. (2009). How do central banks acquire currency reserves and how
much
are
they
required
to
hold?
Retrieved
from
http://www.investopedia.com/ask/answers/06/reserverequirements.asp
Seed, A., Awan, R.U., Sial, M.H, and Sher, F. (2012). An econometric analysis of
determinants of exchange rate in Pakistan. International Journal of Business and
Social Science, 3(6):184-194.
- 110 -
Sinha, P. and Kohli, D. (2013). Modeling exchange rate dynamics in India using stock
market indices and macroeconomic variables. Munich Personal RePEc Archive,
45816 (4). Retrieved from http://mpra.ub.uni- muenchen.de/45816/
Siddiqui, R., Afridi, U. and Mehmood, Z. (1996). Exchange rate determination in
Pakistan: A simultenous equation model. The Pakistan Development Review,
35(4): 683-692.
Sorge, M. (2004). Stress – testing financial systems: An overview of current
methodologies. BIS Working Papers, 165.
Su, C.W. (2012). The relationship between exchange rate and macroeconomic
variables in China. Zb. rad. Ekon. fak. Rij., 30(1): 33-56.
Susmel, R. (n.d). Structural models of exchange rate determination. Bauer College of
Business. Retrieved from http://www.bauer.uh.edu/rsusmel/7386/ln4.pdf
Takaendesa, P. (2006). The behaviour and fundamental determinants of the real
exchange rate in South Africa. (Masters in Commerce dissertation). Retrieved
from http://eprints.ru.ac.za/516/1/Takaendesa_MCom.pdf
h h s ch t
Tran Luong Thanh. (2013).
h t tr
ch u
h
ch h s ch t
h D
t c ac c
c a
. Friedrich Schiller University. Retrieved
from http://vepr.org.vn/upload/533/20131218/Seminar%20VEPR%2016.pdf
Tran Thi Luong Binh. (2012).
from
h h s ch t
h
t ra. Retrieved
http://www.tapchitaichinh.vn/Trao-doi-Binh- luan/Chinh-sach-ty- gia-va-
nhung-van-de-dat-ra/22350.tctc
Truong Dinh Tuyen et al. (2011).
from the report of “T c
u
t
ch h s ch t
ng c a cam k t m c a th tr
ng trong
. Extracted
TO v FTAs
n Vi t Nam,” MUTRAP III Project.
Van Bergen, J. (2010). 6 factors that influence exchange rates. Retrieved from
http://www.investopedia.com/articles/basics/04/050704.asp#axzz24zwzWwSw
VFS. (2013). Investment Research – Macro Q4 2013. Vietfirst Jointstock Securities.
Vohra. (2013). Capital adequacy standards and risk management guidelines for
standalone primary dealers. Reserve Bank of India.
- 111 -
Vu Minh. (2013). Qu n tr r i ro t i ch nh trong c c doanh nghi p Vi t Nam hi n nay.
Science Magazine, National University of Hanoi, Economy and Business, 29 (3):
53-60.
Wilson, I. (2009). The monetary approach to exchange rates: A brief review and
empirical investigation of debt management: Evidence from United States. The
Journal of Business Inquiry, 8(1): 83-99.
Wooldridge, J. M. (2009). Introductory economics: A modern approach (4th edition).
Michigan State University, Cengage Learning.
Yildirim, I. (2012). Stress testing in risk management: An application in the Turkish
banking sector. International Journal of Trade, Economics and Finance, 3(6):
441-444.
Zhang, Z. (2001). RER misalignment in China: An empirical investigation. Journal of
Comparative Economics, 29, 80-94.
- 112 -
[...]... effective exchange rate, which means increase in money causes decline in the value of currency In contrast, interest rate, government expenditure and defic it to GDP are negatively related with the effective exchange rate Atif et al (2012) demonstrates the relationships between Australian exchange rate and its economic and non- economic determinants It is suggested that Australia’s trade components and macroeconomic. .. approaches of exchange rate determination is to provide a clear understanding of the economic mechanisms governing the actual behavior of exchange rate in the real world and of the relationships between exchange rate and other important economic variables (Mussa, 1984) This study just presents three main approaches in determining the exchange rate including purchasing power parity, balance of payment, and. .. foreign exchange volatility and the companies’ foreign exchange exposure, which is the sensitivity of their financial results to foreign exchange rate changes The level and frequency of the exchange rate volatility is influenced by a country’s exc hange rate regime discussed in the next part 2.2 Exchange Rate Regimes An exchange rate regime is the way an authority manages its currency in relation to... degree of openness and capital account balance also causes the real exchange rate s depreciation Moreover, Saeed et al (2012) undertakes an analysis of determinants of exchange rate between US Dollar and Pakistani Rupee under the monetary approach from 1982 to 2010 The results confirm that the relative terms of stock of money, total debt, and foreign exchange reserves are significant factors of PKR/USD exchange. .. the development of the exchange rate po licies and its comparison to several Asian countries, and the qualitative measurement of the relationship between the exchange rate and its determinants Chapter 4 – Data collection and research methodology: The quantitative approaches are conducted in this chapter, including the regression model to examine the relationship between the exchange rate and its independent... current account imbalance can not be financed forever The long-run exchange rate would move to balance the current account Another limitation of this approach is that it just focuses on flows of currency and capital rather than stocks of money or financial assets (Eiteman et al., 2010) That means relative stocks of money or financial assets has no contribution in determining exchange rate in this theory... 2.4 Determinants of Exchange Rate According to Kanamori (2006), an exchange rate of a country is determined by macroeconomic factors, speculative factors, and economic expectations In other words, factors affecting exchange rate can be economic, political, and psychological as well as in the short run or long run (Saeed et al., 2012) A study of a cross – section of 81 countries carried out by Canales-Kriljenko... of any currency o f any country (Danmola, 2013) Many scholars explain the concept of exchange rate in different dimensions According to Hache (1983), exchange rates are relative prices of national currencies, and under a floating rate regime they may naturally be determined by the interplay of supply and demand in foreign exchange markets More simply, Kalra (2005) considers exchange rate as a national... expenditure and money supply are not included in the regression model as macroeconomic determinants of the exchange rate Finally, for its simplicity, this study makes a assumption about a hypothetical firm’s foreign exchange position used to conduct stress testing Data and results do not represent any specific company, but the tools and methodologies can be applied to any firm for managing exchange rate risk... determinants of real exchange rate of South Africa in the period of 1975 to 2005, and finds that the terms of trade, real interest rate differential, domestic credit, openness and technological progress have a long-run relationship with the real exchange rate Among other determinants, the terms of trade explain the largest proportion of the exchange rate s variation The real exchange rate fluctuations are .. .AN EMPIRICAL STUDY ON MACROECONOMIC DETERMINANTS OF EXCHANGE RATE AND STRESS TESTING APPLICATION ON VIETNAM NON – FINANCIAL CORPORATE SECTOR In Partial Fulfillment of the Requirements of the... 3.6: The relationship between exchange rate and inflation rate Figure 3.7: The relationship between exchange rate and interest rate Figure 3.8: The relationship between exchange rate and current... effective exchange rate Atif et al (2012) demonstrates the relationships between Australian exchange rate and its economic and non- economic determinants It is suggested that Australia’s trade components
Ngày đăng: 22/10/2015, 13:17
Xem thêm: An empirical study on macroeconomic determinants of exchange rate and stree testing application on vietnam non financial corporate sector, An empirical study on macroeconomic determinants of exchange rate and stree testing application on vietnam non financial corporate sector