Determinants of current account in Vietnam: an intertemporal approach

IN DEVELOPMENT ECONOMICS DETERMINANTS OF CURRENT ACCOUNT IN VIETNAM: AN INTERTEMPORAL APPROACH BY DO NGUYEN KHANH LINH MASTER OF ARTS IN DEVELOPMENT ECONOMICS HO CHI MINH CITY, DEC

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

DETERMINANTS OF CURRENT ACCOUNT

IN VIETNAM:

AN INTERTEMPORAL APPROACH

BY

DO NGUYEN KHANH LINH

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HO CHI MINH CITY, DECEMBER 2013

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

DETERMINANTS OF CURRENT ACCOUNT

IN VIETNAM:

AN INTERTEMPORAL APPROACH

A thesis submitted in partial fulfilment of the requirements for the degree of

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

BY

DO NGUYEN KHANH LINH

Academic Supervisor:

NGUYEN VAN NGAI

HO CHI MINH CITY, DECEMBER 2013

Table of Contents

Acknowledgement i

Table of Contents ii

List of tables iv

List of figures iv

Abbreviations v

Abstract vi

CHAPTER 1: INTRODUCTION 1

1.1 Problem Statement 1

1.2 Research objectives and questions 3

1.3 Justification of the thesis 3

1.4 Organization of the research 4

CHAPTER 2: LITERATURE REVIEW 5

2.1 Introduction 5

2.2 Definition of Current Account (CA) 5

2.3 Theoretical Literature Review 5

2.3.1 The old approaches 6

2.3.2 The intertemporal approach 8

2.3.3 Determinants of CA balance based on the intertemporal approach 23

2.4 Empirical researches 27

2.5 Conceptual framework 40

2.6 Chapter summary 41

CHAPTER 3: DATA AND RESEARCH METHODOLOGY 42

3.1 Data 42

3.2 Research Methodology 43

3.2.1 Analytical Framework 43

3.2.2 Model Specification 44

3.2.2.2 Cointegration Test 45

CHAPTER 4: FINDINGS AND DISCUSSIONS 50

4.1 Descriptive statistics 50

4.2 The dynamics of variables 51

4.3 The correlation matrix 53

4.4 The regression result 54

4.4.1 Unit Root Test 54

4.4.2 ARDL Test 55

CHAPTER 5: CONCLUSION AND POLICY IMPLICATION 59

5.1 Conclusion 59

5.2 Policy implication 60

5.3 Limitations and directions for further studies 60

5.3.1 Limitation 60

5.3.2 Further studies 61

REFERENCES 62

APPENDICES 68

List of tables

Table 2-1: Empirical studies about determinants of CA balance 33

Table 3-1: The measurements and expected signs of variables 43

Table 3-2: Critical values for Dickey Fuller test 45

Table 3-3: Critical values bounds for testing the hypothesis of no – cointegration relationship 47

Table 4-1: Descriptive statistics 50

Table 4-2: The correlation matrix of variables 53

Table 4-3: The values of stationary tests of all variables 54

Table 4-4: The summary of result 55

Table 4-5: Results of cointegration (m = 1) 56

Table 4-6: ARDL equations chosen with lowest AIC and SIC’s values 56

Table 4-7: Estimated long – run coefficients using the ARDL approach 57

List of figures Figure 1-1: Current account of Vietnam from 1999 to 2012 2

Figure 2-1: Temporary fall in income 15

Figure 2-2: Current account acts like a shock absorber 22

Figure 2-3: The conceptual framework 41

Figure 3-1: The Analytical framework 44

Figure 4-1: The dynamics of variables 52

Abbreviations

ARDL : Autoregressive Distributed Lag

CA : Current Account

CPI : Consumer Price Index

ECM : Error Correction Model

GMM : Generalized Method of Moments

GSO : General Statistics Office of Vietnam

IFS : International Financial Statistics

IMF : International Monetary Finance

LDC : Least Developed Countries

NFA : Net Foreign Assets

OECD : Organization for Economic Co-operation and Development REER : Real Effective Exchange Rate

USD : United States Dollar

VND : Vietnam Dong

Abstract

This study attempts to explain theoretically and examine empirically influences of macroeconomics factors on the current account behavior in Vietnam from 2000Q1 to 2010Q2 The study employs the intertemporal approach to build the theory of Vietnam current account behavior’s determinants and applies Autoregressive Distributed Lag method to estimate the model Theoretically, intertemporal approach was an approach broadly employed to decide determinants

of current account recently It treats current account as an absorber for the economy

in facing a shock in income The study discovers that initial net foreign assets, trade openness, real effective exchange rate and inflation volatility have significant relationships with current account in Vietnam

CHAPTER 1: INTRODUCTION

This study attempts to explain theoretically and test empirically influences of macroeconomics factors on the current account behavior in Vietnam The study employs the intertemporal approach to build the theory of Vietnam current account behavior’s determinant and applies autoregressive distributed lag method to estimate the model Theoretically, intertemporal approach was an approach broadly employed to decide determinants of current account recently It treats current account as an absorber for the economy in facing a shock in income Determinants

of current account will be determined based on this idea

1.1 Problem Statement

The current account is a crucial indicator of a country’s economic performance (Knight and Scacciavillani 1998) Current account deficit could damage the economy in many ways First, current account deficit that accompanies

by budget deficit leads to the increasing of debts (Krugman, 1987) According to Ministry of Finance, public debt reached 52.6% of GDP, foreign debt hit 38.8% of GDP in 2010 and was equal 56.6% compared to GDP in 2011 Second, in an economy that applies fixed exchange rate regime and has weak foreign reserve, a prolonged current account deficit may raise the probability of a monetary crisis (Edwards, 2002) The Asian financial crisis in 1997 is a good example for this situation Third, a serious current account deficit raises people’s expectation of depreciation that increases the volatility of the foreign exchange market Altogether, maintaining the balance of current account at suitable level is one of the majority missions of macroeconomics policymakers Consequently, strong understanding about the determinants of current account balance becomes an important subject that requires careful researches (Yang 2011)

Like most developing countries in the world, Vietnam mostly suffers from the current account deficit over the past 12 years There are two deficit periods of

current account: the low deficit period in 2002-2006 and the high deficit period in 2006-2010 The first deficit period started in 2002 and reached the lowest value (-1.93 billion USD) in the year 2003 After 2003, it turned to improve until getting the balance in 2006 However, the situation became worse after 2006 It reached -6.953 billion USD in the year 2007 and -10.823 billion USD in the year 2008 Current account had improved and had an increasing tendency since 2008 It got the surplus position in the year 2012 (see Figure 1-1) Altogether, the situation of current account is still very unpredictable As observations mentioned in the previous paragraph, the erratic current account level could have a negative impact

on the stability of Vietnamese macroeconomic, which has become instability for long time owing to the impact of global crisis and the products of many overlap policy Therefore, it is very essential to identify determinants of Vietnam’s current account balance so that the country could restore external balance by controlling those causal factors through suitable macroeconomic policy options Hence, the aim

of this paper is to shed the light on finding the determinants of current account in intertemporal approach perspective

Figure 1-1: Current account of Vietnam from 1999 to 2012

1.2 Research objectives and questions

The study is attempted to accomplish two major objectives: (1) To find the determining factors that theoretically influence the current account based on intertemporal approach; and (2) To examine those factors statistically explaining the current account behavior in Vietnam The central question of the research is: “What are the determinants of Vietnamese current account?” To give an answer, this question is divided in to two sub- questions: (1) What is the factors theoretically influencing Vietnamese current account? (2)

Do these factors statistically explain the current account in Vietnam? By examining the theoretical literature review and empirical studies this study will get the answer the first one The second one will be answer by employing econometric method to analyze the determinants which obtained by the first sub-question

1.3 Justification of the thesis

The study gives main contributions in two aspects as follows

First, there has been a numerous quantity of empirical researches on the determinants of current account Many of them work in the case of developing countries However, there are still no published researches about this topic in the context of Vietnam Thus, this field in Vietnam still continues being an unexploited subject although the demand for strongly concentration on it is considerable Hence, this research contributes to the literature as the first comprehensive study about this subject in the Vietnamese case

Second, this research provides useful and reliable understanding of the nature

of the Vietnamese current account behavior according to intertemporal approach during the past ten years empirically for macroeconomics policymaker Consequently, they have dependable evidences to propose and implement better policies for current account

1.4 Organization of the research

The reminder of the thesis research design is organized into four sections as follows: after introduction in the first section 1, literature review is discussed in the second section, the third section is about data and research methodology, the fourth section points out the findings and discussions, and the final section contains the conclusions and policy recommendations

CHAPTER 2: LITERATURE REVIEW

2.1 Introduction

In this chapter, theories and studies of current account behavior will be reviewed This chapter will particularly concern about intertemporal approach theory due to its central role in this study The determinants of Vietnamese current account will be introduced after learning carefully intertemporal approach In addition, other empirical research will be concern to attain the general viewing Conceptual framework will be introduced in the last section of this chapter

2.2 Definition of Current Account (CA)

The CA is one of the two main elements of the balance of payments According to Madura (2007), CA is the sum of the trade balances (exports value minus imports value), factor income payments (earnings on foreign investments minus payments for foreign investors)

CA = Trade Balance + Income Payment

2.3 Theoretical Literature Review

Finding determinants of CA balance has always been the attractive topic of economist for a long time Initially, many economists such as Caprio and Howard (1984), Baxter and Crucini (1993) try to treat CA balance as a static factor This leads to the using of static approaches to determine what factors have impacts on

CA balance Although those approaches have an advantage in providing a straightforward and easy perspective, they were claimed to be not effective and not reasonable in explaining CA balance behavior (Obstfeld and Rogoff, 1995; Yang, 2011) The newer approach is dynamic approach that was named intertemporal The intertemporal approach to the CA was initially recommended by Sachs (1981) and further extended by Obstfeld and Rogoff (1995) This approach is still extended by many researches such as Debelle and Faruqee (1996), Yang (2011) and Brissimis,

Hondroyiannis, Papazoglou, Tsaveas and Vasardani (2012) This approach is based

on the consideration that CA balance is a dynamic result that must be determined not only by the recent events but also by the past and the future ones Therefore, solely concentrating on the recent events could fail to find the determinants of CA balance Instead, intertemporal approach try to decide the CA balance’s determinants based on investigating effect factors in multiple periods Hence, Obstfeld and Rogoff (1995), Yang (2011) claim that the intertemporal approach is suitable for deciding determinants of CA balance than the old others Consequently, this thesis tries this approach to find out the determinant of CA balance

2.3.1 The old approaches

In the beginning of open economy macroeconomics, CA balance’s behavior was recognized as the consequence of impacts of static determinants Indeed, many theories were developed to explain CA balance’s determinants in this viewpoint The most prominent of those are elasticity approach and absorption approach

A concentrating on the CA balance as the net export balance guided many economists to think that vital determinant of CA could be its relative international prices The "elasticity approach" to the current account was born based on this idea According to elasticity approach, static price elasticities of supply and demand decide the net international flow of capital while the determinants of international expenditure and incomes levels are held fixed in the background (Obstfeld and Rogoff, 1995) This approach was chiefly built on the study of price elasticities of demand for exports and import, with respect to changes in RER (Yang, 2011) This approach typically states that the CA balance is mostly decided by RER, domestic income and foreign income It has the advantage in straightforward empirical calculations, which are frequently found to be useful in investigating the short-run implications of exchange rate changes on the CA balance Nonetheless, the most important weakness of elasticity approach is that it merely concentrates on the role

of RER and the trade flow (Goldstein and Khan, 1985) Moreover, this approach

only looks at the traded goods market and ignores the interaction of other various markets in an economy (Yang, 2011)

The absorption approach treats the CA balance as the gap between savings and investment Hence, this approach is similarly recognized as the saving-investment balance approach It claims that if a country produces less than it spends (i.e investment exceeds saving), it must import from other economies for its redundant consumption and spending This country thus runs a CA deficit Because the sum of CA and capital account must equal zero in a flexible exchange rate regime, shocks that happen first in capital account will evidently affect CA, and vice versa Hence, the absorption approach maintains that it is necessary to involve determinants of capital account balance when constructing the CA model This approach also claims that the RER is unimportant in CA adjustments (Krugman, 1987) It has the advantage when concerning two side, saving and investment, of the current account problem However, the critical weakness of this theory was that its analyze CA position based on the level of determinants in a given point of time Therefore, Sachs (1981) calls it one period theory It means that the determinant’s value before and after the time t has no role in adjustment CA This obviously makes no sense As Sachs (1981, p.212) appropriately put it, “[a] one period theory

of the CA that describes a static balance of imports and exports makes as much sense as a one-period theory of saving and investment” As he mentioned in his paper in 1981, saving is, by definition, an intertemporal choice whereby an agent is willing to offer consumption today for larger future consumption He conclude that this one-period approach is not effective enough to adjust saving Hence, a static approach is thus primarily inaccurate in analyzing CA balance

To conclude, the old static approaches’ techniques should bring the inappropriate results in examining the CA behavior

2.3.2 The intertemporal approach

In contrast with the standard static approach, the modern intertemporal optimizing approach presents a framework appropriate for a positive and normative examination of CA dynamics (Razin, 1995) According to Sachs (1982), a dynamic approach seems to be the best theory to explain the movement of CA balance Intertemporal approach was born owing to this consideration

2.3.2.1 Overview

This model shares the saving – investment theory with absorption approach However, intertemporal approach is a microeconomics-oriented approach Different from other macroeconomics-oriented approaches, it analyzes CA balance of an economy by treating that economy as an agent and using mathematical modeling with the utility function to analyze its behavior

Consumption smoothing was the core theory of intertemporal approach According to Williamson (2005), consumption smoothing is a principle in the response of consumption to changes in income That is, there are natural forces that induce consumers to maintain a stable consumption path over time in spite of the path of income (Vegh, 2013) This principle comes to substitute the opinion that consumer had a marginal propensity to consume and hence present consumption was tied to present income For instance, James, who is currently wins $1 million in

a lottery, could be an unambiguous example for consumption smoothing principle James could consume all of his lottery prizes on consumption within the present period and save nothing However, it would appear more sensible if he spent a small part of his winning in the present period and saved a considerable fraction in order

to consume more in the future Therefore, his consumption just change a small value compare with the increase in his income That behavior of agents was named consumption smoothing At all, consumption smoothing is the tendency of consumers to maintain a consumption path over time that is smoother than income (Friedman, 1957) This theory involves an economic trade-off decision between

current and future consumption and saving, which will be called intertemporal decision (Williamson, 2005)

When an open economy is considered as an agent, consumption smoothing principle is still appropriate for explaining its behavior In fact, this principle insists that because agents in an economy wish to smooth their consumption, this economy has the same desire (Yang, 2011) Different from a closed economy, an open economy can borrow from abroad during bad times and pay back in good times By lending and borrowing, that open economy can weaken the connection between present consumption and present production Hence, it has an ability to smooth entirely its consumption despite a vacillating endowment path (Vegh, 2013) To smooth its consumption, when shocks happen, agents in the economy have a tendency to adjust their saving (Blanchard and Fisher, 1989) Hence, shocks will affect CA owing to the fact that CA equals saving minus investment This ability, thus, leads the CA acts as a shock absorber, improving in good times and worsening

in bad times, to avoid consumption of that economy being affected by shocks (Vegh, 2013)

Lastly, this approach is suitable for the developing economies that are usually experiencing high level of investment and running the CA deficit (Dumitru and Dumitru, 2009) It is due to the fact that developing economies characteristically face larger shocks, which suggests that they should rely more on

CA imbalances to smooth consumption over time This offers a theoretical rationale for efforts intended at ensuring that developing economies have admission to external finance during bad times (Vegh, 2013)

2.3.2.2 Mathematical model

As mentioned above, the intertemporal approach for examining CA’s determinants is based on microeconomics principles Therefore, to build the model, the research will starts from the basic microeconomics concepts

a) The basic model

This research considers Vietnam as a small open economy inhabited by a large number of identical infinitely lived households We assume that there is no uncertainty (households have perfect foresights), no investment and no government

in the economy We also assume that there is only one good, tradable and storable one Because the economy is small, it takes the tradable good’s price that is set by the rest of the world Households assume to borrow/lend in international capital markets as much as they want at an exogenously given real interest rate, rt rt

non-is assume to be constant over time at the value r

Utility

In basic microeconomic theory, the household utility function is In

which, each consumption level at time t ( ) results in the corresponding level of utility This instantaneous utility function is assumed continuously differentiable, strictly increasing, and strictly concave We have the discounted factor of the continuous utility:

(1) where > 0 is the subjective discount rate

Therefore, the household’s lifetime utility is represented by a utility

functional, U, which assigns utility U(c t ) to each consumption path c t according to:

Households are assumed to maximize their lifetime utility subject to

their budget constraint

Household budget constraint

The budget constraint in a period ∆t is the net foreign assets that household accumulates in that period Let represent net foreign assets (NFA) denominated

in terms of the tradable good held by the household at time t, represent the

endowment flow received by the household at time t, and denotes taxes According to (Vegh, 2013), the household budget constraint is given by:

Equation (3) implies that the change in NFA1 equals his/her total income (interest receipts and endowment) minus his/her consumption and taxes Notice that the

stock of NFA, b, is a predetermined variable in the sense that, barring some

exogenous change, it is a continuous function of time The household was start with

some NFAs (i.e., b 0)

Household maximizing problem

As usual, households want to maximize their utility under their budget constraint

The condition lim → ( ≥ 0 typically mentions a “no-Ponzi games

condition” and implies that the household not “end” with debt Because the instantaneous utility function is strictly increasing in consumption (i.e., there is no satiation point), it will never be optimal for the agent to “die” with assets because increasing consumption at some point in time would always produce higher lifetime utility It means that, at an optimum, lim*→ ( ≤ 0 Linking this condition with

the constraint, given by (4), that the household will not “end” with assets yields the condition

- –-.0 ∆

∆ = lim∆ → ∆-∆.=

lim → ( = 0 (5) The process is proceeded here is to convert the issue into one that can be solved with standard Lagrange multiplier methods To this end, we will reformulate this maximization problem subject to an uncountable number of constraints (i.e., budget constraints in each point of time) into a maximization issue subject to one constraint (i.e., an intertemporal budget constraint)

Intertemporal budget constraint

To derive the household’s intertemporal budget constraint, rewrite equation (3) as

Solution for the maximization problem

The household’s issue can then be restated as taking ∈ [0, ∞]to maximize

(1) subject to (10) We can examine this problem by standard Lagrange-multiplier method The Lagrangean is given by:

where ; is the Lagrange multiplier It can be represented that the first order

condition with respect to is given by2,3

“pointwise optimization” The “pointwise optimization” captures the concept that, instinctually, optimal consumption was being chosen by households at each point of time Practically, this concept allows us to

“ignore” the integral when differentiating Because consumption was being chosen at each point of time, this

Interpretation of the Lagrange multiplier

To have insightful interpretation of Lagrange multiplier, we multiply both sides of equation (12) by ( and integrate forward to obtain4:

@

The term λ/r can then be considered as the (relative) price of an asset that

would give the household a “dividend” of ’ at every point in time λ would imply the annuity value of the present discounted value of dividends If were constant overtime ( = ̅ , then λ would imply that it equals the dividend ’ ̅

According to basic finance principle, a stock price λ/r will not be affected by

anticipated changes in dividend, but by “news”, that is unanticipated events Similarly, the multiplier λ will not be affected by anticipated fluctuations in income,

but by unanticipated changes in income

Friedman (1957), he argued that current consumption should not be the outcome of current income as Keynes had claimed He asserts that current consumption will depend on long – term expected income that was called permanent income

Temporary fall in income

In this part, we will examine the effect of a temporary fall in income on consumption We assume that income is flat, is at high level ( F) when t < 0;

falls in period [0, T) at lower level ( G), and goes back to high level ( F) after that

(see Figure 2-1) In brief, we have:

Figure 2-1: Temporary fall in income

With unanticipated shock, the income equation is:

+ K G ( + F (

The household thus maximizes its lifetime utility subject to the intertemporal constraint (15) As before, the corresponding first-order condition implies that consumption will be flat along the new perfect foresight equilibrium path Hence, from (15), it follows that5:

Because O = − ̅ , when facing a

temporary fall in income, household needs to reduce savings to maintain constant consumption in a lower level

The bigger is J, the larger permanent income will fall and, therefore, the

larger consumption will fall For a small value of J, the fall in consumption will be

very little The impacts of the fall in income on consumption and saving depend on its duration The shorter is the duration of the disturbance, the larger is the decrease

in saving and the smaller is the decrease in consumption carried out by households

b) The Aggregate Model

Because of the no investment assumption, the CA was equal to savings In that circumstance, we examined how a transitory fall in the endowment cause lower saving and therefore to a CA deficit In fact, a temporary fall in income could cause both a decrease saving (as the consumption smoothing principle above) and a decrease in investment This section will introduce investment into the basic model and show how the relative strength of both influences depends on the duration of the disturbance

While the basic model is formulated in continuous time, switching it to discrete time was more convenient It is due to the fact that variable was performed

in discrete time (i.e quarterly or annually) In theory, it has to employ many mathematics techniques such as derivative or integral, which need to use in continuous time In fact, there are many empirical research build the discrete time

model

Household budget constraint

We can rewrite the household budget constraint equation in discrete type as follows:

Household’s intertemporal budget constraint

Iterating forward for the equation above and performing the same process as described in basic model:

The process of finding the solution for maximization problem under this discrete intertemporal budget constraint was proceeded as same as the previous one (Vegh, 2013) Specifically, household chooses to maximize its lifetime utility

∑ ZV subject to the intertemporal budget constraint The Lagrangean is applied again to solve this maximization problem:

ℒ = S Z

V

+ ; [ 1 + + S T1 + U1 − S T1 + U1 [ + + R ]

V V

We can assume that Z 1 + = 1, with the same reason which is already

discussed in the continuous case Therefore, the condition above can be presented as:

As the outcome in our continuous model without investment (equation (12)), equation (20) indicates that, along a perfect foresight equilibrium path, consumption will be flat9 in spite of the path of income

The government budget constraint

According to Batdelger and Kandil (2012), the government budget constraint

at time is:

+L

where c is the government bonds issued at time − 1 and maturing at time t; c

is the interest payment for the bond; is the government taxes; and d is the

government consumption

If the right-hand side of the equation (19) is negative, it implies the budget deficit And the opposite is true if it is positive The left-hand side term describe that the budget deficit was financed by issuing more bonds

The government’s intertemporal budget constraint

Performing the same process applied in household’s intertemporal budget constraint by iterating forward for the equation (19), we have the intertemporal budget constraint of the government:

⇒ − 1 + c = S T1 + U1 [ − d − Rc]

V

⇒ 1 + c + ∑ XV L+(L Y = ∑ XV L+(L Y [d + Rc] (22) This government intertemporal budget constraint maintains that the government’s expenditure and investment (right-hand side) have to be equal to the government’s wealth (left-hand side), presented by government bond plus the present discounted value of their taxes

The process of finding the solution for maximization problem under this government intertemporal budget constraint was proceeded as same as the previous one By using the Lagrangean, the maximization problem is solved:

The condition (21) suggests in the short-run, similar to households, government also optimally borrows, lends and invests in order to smooth their consumption Under consumption smoothing, the equation (19) suggests that government can finance their current deficit either by issuing more government bonds or by increasing taxes but they have to reverse it after that

The aggregate budget constraint

According to Batdelger and Kandil (2012), to obtain an aggregate model for the whole economy, we combine the household and government model As discuss above, the household and government budget constraints can be presented respectively as below:

+L

c − c = c+ − d − Rc

Let e and f denote NFA and investment of the whole economy

respectively Both of they are defined as the sum of household and government part Then, we have:

households and government smooth their consumption ( ̅ , d̅ ) to maximize their

lifetime utilities Because consumption and government spending should be smooth overtime and e is a predetermined factor, saving will suffer the fluctuation of

income (Batdelger and Kandil, 2012) It means that, in a shock of income, saving will fluctuate to maintain the aggregate consumption unchanged This implies that,

in good times, the economy increases its lending from the rest of the world The opposite is true in bad times Saving, hence, act as a shock absorber

According to equation (27), CA has two components: saving and investment

As discussed above, equation (26) shows that the saving component actually play as the shock buffer Since CA equal saving minus investment, it acquires this characteristic This means that CA will act as a shock absorber in the face of income

shock to allow the economy to smooth consumption and maximize welfare (Obstfeld and Rogoff, 1995) By this, when income changes from low to high, the

CA balance improves and the opposite is true in bad times (see Figure 2-2) This is the basic idea of intertemporal approach Based on this idea, maintaining current account balance is even more serious issue for developing economies since they usually experience larger shocks, which infers that they will depend more on current account to smooth consumption overtime (Vegh, 2013)

Figure 2-2: Current account acts like a shock absorber

2.3.2.3 Conclusion

Intertemporal approach contributes two implications for examining current account First, it provides two channels that factors could affect current account Those are saving and investment It maintains that current account equals saving minus investment Therefore, if it improves saving, it will improve current account

In contrast, it will worsen current account when it improves investment

Second, it suggests how factors affect saving through income It maintains that because economy desires to smooth its consumption, saving will act as an absorber When faced to a shock in income, saving will suffer and therefore, current account will suffer This is also the main principle of intertemporal approach

, , d

̅ , d̅

CA surplus

0

2.3.3 Determinants of CA balance based on the intertemporal approach

Before explaining the estimation approach, our research will first identify the main determinants of CA deficit Since the literature on CA modeling is vast and numerous specifications are available, we proceeded by selecting standard variables that are typically affect CA according to intertemporal approach Other existence determinants are not considered because they do not affect CA in intertemporal approach’s viewpoint

Because temporary shock and permanent one has different impact on CA, to analyze the disturbance of factor’s effect on CA, we must ask whether that this disturbance’s effect on income is temporary or permanent If this effect on income

is a temporary one, it influence saving and CA without changing consumption On the contrary, a permanent one influence saving ratio and CA by changing consumption

2.3.3.1 Initial Stock of Net Foreign Assets (INFA)

This variable was used in Debelle and Faruqee (1996), Chinn and Prasad (2003), Gruber and Kamin (2007), Chinn and Ito (2008), Clower and Ito (2011), Yang (2011) and Morsy (2012)

According to Morsy (2012), the level of the initial NFA affects the current account in two perspectives First, countries with high initial NFA usually attract higher foreign income flows, leading to a positive relationship between initial NFA and the current account The first effect often is expected to dominate for most of empirical studies Second, in terms of an intertemporal perspective, the stock of initial net foreign assets (INFA) serves as an important initial condition Specifically, a country remains solvent if it has high levels of initial net foreign assets Indeed, from a buffer stock savings perspective, countries with a relatively high level of NFA can have the funds to run current account deficit for an extended

period and still remain solvent Therefore, this leads to a negative association between NFA and current account (Yang, 2011)

2.3.3.2 Trade Openness

This variable was used in Chinn and Prasad (2003), Chinn and Ito (2008), Kwalingana and Nkuna (2009), Ang and Sek (2011),Clower and Ito (2011), and Yang(2011)

This variable represents the economy’s openness degree to international trade The openness variable could specify the capability to attract foreign investment of a country (Kwalingana and Nkuna, 2009) International trade provides an important channel for transferring technology to developing economies (Yang, 2011) Therefore, economies with higher openness degree tend to be comparatively more attractive to foreign capital, letting them to acquire more investment and to finance the resulting current account (Chinn and Prasad, 2003)

In general, a high trade openness degree could lead to an improving investment through this, a worsening current account balance Therefore, expected sign of the relationship between this factor and CA balance is negative

2.3.3.3 Real Effective Exchange Rate (REER)

This variable was used in Bergin and Sheffrin (2000), Herrmann and Jochem (2005), Moccero, (2008), Kwalingana and Nkuna (2009), Ang and Sek (2011), Yang (2011), Brissimis et al (2012)

The REER can influence the CAB in two ways On one hand, decreases in REER (appreciations) influence the CAB in the positive direction The Mundell-Flemming model predicts that an appreciation in the REER can adversely affect a country’s competitiveness position, leading to a worsening trade balance and, through this, a worsening current account balance As a result, a decrease in REER

(appreciations) could have a positive impact on current account balance (Herrmann and Jochem, 2005)

On the other hand, in the intertemporal perspective, a decrease (appreciation)

in the REER could improve country’s CAB The consumption smoothing hypothesis maintains that the CA performs as a buffer to smooth consumption in the face of temporary shocks to national cash flow Therefore, in response to a temporary decrease (appreciation) in the REER which leads to a temporary increase

in income, an open economy would desire to rise saving and run a CAS rather than permit consumption to rise (Yang, 2011) As a result, a domestic currency appreciation could cause an improvement of the CA, which means that relationship

is negative (Herrmann and Jochem, 2005)

Additionally, Bayoumi (1993) maintains that the process of financial development will relax the borrowing constraint experienced by households

Therefore, this process is generally associated with lower levels of saving ratio (Brissimis et al., 2012) Meanwhile, financial development allow banks to lend at a lower cost and offer more financial services to individuals, for instance for house purchases or consumption This cause a significantly increase in consumption and, hence, a decrease in saving (Lehmussaari, 1990; Bayoumi, 1993; Ostry and Levy,

1995 and Brissimis et al., 2012) Additionally, Brissimis et al (2012) also maintains that the credit conditions and private agents’ borrowing behavior associated with the sharp increase in asset prices, mostly house prices This will lead to an increase in consumption, therefore, lower saving ratio Therefore, the process of financial development is expected to have negative relationship with saving ratio (Brissimis

et al., 2012)

At all, a large body of literature forecast that domestic credit has a positive relationship with investment and negative relationship with saving Hence, this variable is expected to have a negative impact on current account

of volatile future income flows Nevertheless, there is many empirical evidences that high inflation volatility could cause less saving, because it enhances expenditure in time by building a climate of insecurity that performs in the opposite direction, that is favoring present (relative to future) consumption As a result, the relation between CA balance and volatility of inflation variable is indecisive and can only be decided empirically (Brissimis et al., 2012)

2.4 Empirical researches

Debelle and Faruqee (1996) investigate both short-run and long-run effects

of net foreign assets, relative income, government debt, demographic factors, fiscal surplus, terms of trade, inflation, real interest rate, and capital controls on current account behavior for 21 industrial economies from 1971 to 1993 Their research use panel technique and apply intertemporal approach’s perspective The results present that demographic factors, government debt and relative income have statistically significant long-run impacts on the CA In contrast, the coefficients of capital controls, terms of trade and fiscal surplus are statistically insignificant

Bergin and Sheffrin (2000) employing a panel data of 3 countries (Canada, Australia and United Kingdom) and GMM estimation of equation to support intertemporal model of the CA The research followed structural orientation and investigated CA behavior by observing net output, interest rate and exchange rate of the economies The paper discovers that including the exchange rate and interest rate increases the fit of the intertemporal model over what was done in previous studies

Chinn and Prasad (2003) examine the medium-term determinants of current accounts by emphasizing the roles of saving and investment Their work was built upon the work of the work of Debelle and Faruqee (1996) and used annual data for

71 developing and 18 industrial economies over the period 1971-1995 They study the current account‘s properties by both cross-section and panel regression techniques (OLS estimator) They discover that, in developing economies, indicators of openness to international trade have negative relationship while measures of financial deepening have positive relationship with CA balance Additionally, they also discover that government budget balances and initial stocks

of NFA have positive impacts on CA balance Other theoretically important variables seem to have weak relationship with CA balance

Bussière and Müller (2004), applying GMM, Least Squares Dummy Variable and Two Stage Least Squares estimator, estimate the intertemporal model for annual data for 33 economies, including Bulgaria, Romania, 10 acceding economies and 21 OECD economies over the period from 1980 to 2002 The macroeconomic variables chosen in this paper include lagged CA, fiscal surplus, change in net output, public spending ratio, relative income and investment ratio All coefficients of these variables, except that of the public spending, are significant and have the expected sign They concluded that, the intertemporal approach could certainly explain the considerable current account deficits in acceding economies They also prospect the high potential of these economies to come up to the OECD economies in the next decades

Herrmann and Jochem (2005), using Feasible Generalized Least Squares and intertemporal approach, examines the determinants of the CA deficits in eight central and east European countries, joining the European Union in May 2004, from

1994 Q1 to 2004 Q4 This paper investigates the impact of relative per capita income, exchange rates, investment demand, fiscal deficits and the state of development of the financial systems on CA balance of these countries Its model shows that the relative per capita income has a significant positive impact on driving the deficit of CA It also indicates that the high valuation of their currencies, underdeveloped financial system, less development finance, low interest rate and high fiscal deficit has a negative impact on CA balance However, the impact of the fiscal deficits was proved slight

Gruber and Kamin (2007) examine CA imbalances, which has emerged in recent years, in United State and developing Asian countries They followed the work of Chinn and Prasad (2003) to observe a panel data of 61 countries from 1982

to 2003 They present five standard CA determinants, including relative growth rates, international openness, per capita income, demographic factors, financial crises and fiscal balance They discovered that the model successes in describing the

Asian current account surpluses, but fails in describing the American current account deficit

Chin and Ito (2008), try to describe the determinants of CA by following the study of Chinn and Prasad (2003) They examine the medium-term determinants

of current account in 69 developing and 19 industrial economies in the period 1971–

2004 They discover that the standard determinants, like income and demographics variables, employed in Chinn and Prasad (2003) cannot describe the upward trend

of Asian current account alone Hence, they expand Chinn and Prasad (2003) specification by factors of legal environment and financial development that influence economic growth and saving - investment behavior They discover that the government budget balance is a significant determinant of the industrial countries current account balance Their outcome also state that improvements of Asian current account were not explained by excess saving but the lack of investment over the last decade

Moccero (2008) tries to decide determinants of Argentina CA from 1985 to

2002 by applying VAR model and intertemporal approach to the current account

He introduces five influenced factors of CA balance including consumption and net output, nominal and real interest rates, wholesale price index, real exchange rate, population and working-age population The intertemporal model does not describe precisely the actual CA from 1885 to 2002 in Argentina, but some evidence in favor

of the intertemporal approach for the period 1885-1930

Dumitru and Dumitru (2009) assess the sustainability of the CA deficit in Romania by estimating its structural components, based on LSDV, IV, GMM estimation and intertemporal perspective They declared that previous value of the

CA, fiscal balance as % of GDP , relative income and net output as % of GDP have positive impacts on CA balance of Romania On the contrary, investment rate, public spending rate, real effective exchange rate have negative relations with CA balance of Romania This paper emphasized an excessive CA deficit of Romania in

the years from 1980 to 2007 It also found that the main drivers of this structural CA deficit are the relatively low GDP and the high rate of investment

Kwalingana and Nkuna (2009), applying Malawian annual data from 1980 to

2006, intertemporal approach and OLS estimator, empirically identify the influences that have impact on the long run and short run behavior of Malawi’s CA Its empirical model for estimating was built on the work of Debelle and Faruqee (1996), and Chinn and Prasad (2003) Its findings present that trade openness, terms

of trade, accumulation of external debt and the real exchange rate basically determine the CA balance in Malawi The relation between CA balance and trade openness is negative while the rest factors affect CA balance positively Furthermore, this paper finding shows that these deficits of CA have been, largely, persistent

Ang and Sek (2011), employing a panel data of 10 countries, intertemporal approach and GMM method to compare if the determinants on the CA balance in these two groups of economies the economies with CA deficit and the economies with CA surplus The economies with CA surplus consist of Germany, Japan, Singapore, Norway and Switzerland The economies with CA deficit include United States, Italy, Australia, Cyprus and Portugal It test 7 candidates of determinants on

CA balance which consist of exchange rate, consumer price index (CPI), the oil price of the world, interest rate, terms of trade, trade openness and reserve accumulation Its results present that CA balance determinants are different between two groups of economies The main determinants to Ca in economies with CA deficit are CPI, world oil price, interest rate and exchange rate Oppositely, the macroeconomics factors that have central impact on CA in economies with CA surplus consist of CPI, trade openness and terms of trade Other factors only display significant influence on CA balance in certain economies but not the others

Clower and Ito (2011), employing a panel data of 70 countries, Probit analysis, OLS analysis, and intertemporal approach hypothesis, provide evidence to

investigate the medium-term determinants of the CA They recommended 9 candidate of CA persistency Those are exchange rate regime, trade openness, open financial markets, the size of CA imbalances, government debt, financial development, stage of development, net foreign assets and currency crisis They discover that trade openness, government debt, financial development, stage of development, and net foreign assets contribute to the current account persistence’s degree

Yang (2011) investigates the current account behavior of 8 emerging Asian countries by following the theory of intertemporal approach and applying the cointegrated Vector Autoregression methodology The eight sample economies include Hong Kong, Singapore, India, Thailand, Malaysia, Korea, Philippines and China The candidates of current account determinants are relative income, initial stock of NFA, real exchange rate and degree of openness to trade on CA balance This paper specifies that current account balances in sample economies are not homogeneous This paper also indicates that domestic relative income, initial stock

of NFA, REER and trade openness strongly have significant impacts on current account in long run

Batdelger and Kandil (2012) investigate intertemporal approach and Vector Error Correction Models (VECM) to study the behavior of the CA balance in the United States from 1960 to 2004 This paper introduced six factors that could determine CA balance of the US They included US real GDP, GDP deflator in the

US, energy price, nominal effective exchange rate, domestic interest rate, and weighted average of real GDP in major trading partners of the US They found that higher exchange rate (appreciation) worsens the current account balance in the long run Moreover, an increase in the interest rate improves the current account balance significantly in the long run

Brissimis, Hondroyiannis, Papazoglou, Tsaveas and Vasardani (2012), using intertemporal model and fully modified OLS, built an empirical model for

explaining Greek current account behavior over the period from1960 to 2007 Determinants of Greek current account are specified as inflation volatility, domestic real GDP per capita, the dependency ratio, the real effective exchange rate, the real interest rate, government fiscal balance, credit to the private sector and private investment Its results suggest that all the variables except REER have statistically insignificant coefficient in the period before 1991 It means that, in this period, the current account was mainly affected by the exchange rate policy In contrast, after

1991, all the variables except fiscal balance and REER have statistically significant coefficient in long run

Morsy (2012) use dynamic panel estimation techniques to examine the determinants of the medium-term current account balance for 74 countries including

28 oil-exporting countries for the period of 1970-2009 He introduces candidates of current account balance based on intertemporal approach They are fiscal balance, demographic factors, net foreign assets, oil balance, economic growth, oil wealth and level of oil production–related imports The results state that factors that matter

in determining the equilibrium current account balance of oil-exporting counties are fiscal balance, oil wealth, oil balance, economic growth, age dependency, and degree of oil production–related imports

Generally, there has been a widespread quantity of study work on the determinants of CA balance in many contexts, but there are not any analyses that do the same job in the case of Vietnam

The table below summarizes all empirical studies, which are reviewed above: