Exchange rate pass through to vietnams import and domestic prises

OPI: oil price index Crude Oil price index, Brent UK, NEER: nominal effective exchange rate, FPI: foreign exporters' price index, IMP: import price index, GAP: output gap, CPI: consumer

VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

EXCHANGE RATE PASS-THROUGH TO VIETNAM'S

IMPORT AND DOMESTIC PRICES

BY

CAIBAOHIEU

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HO CHI MINH CITY, DECEMBER 2012

VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

EXCHANGE RATE PASS-THROUGH TO VIETNAM'S

IMPORT AND DOMESTIC PRICES

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

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

By

CAl BAO HIEU

Academic Supervisor:

Dr DINH CONG KHAI

HO CHI MINH CITY, DECEMBER 2012

Contents

Chapter 1: Introduction 1

1.1 Problem statement 1

1.2 Research questions 2

1.3 Research objectives 3

Chapter 2: Overview of Vietnam's exchange rate, import activities and inflation 4

2.1 Exchange rate arrangement in Vietnam 4

2.2 Overview of Vietnam's import activities 7

2.3 Overview of Vietnam's inflation 8

Chapter 3: Literature review 1 0 3 1 Linear approarch 1

3.2 Vector Error Correction Model approach 11

3.3 Vector Autoregression approach 11

Chapter 4: Research methodology, empirical framework and data description 14

4.1 Research methodology 14

4.2 Empirical framework 16

4.3 Data description 17

Chapter 5: Empirical results 20

5.1 ADF tests 20

5.2 Optimal lag length 20

5.3 VAR regression 20

5.4 Impulse response function: 24

5 4.1 Response to exchange rate of foreign price index 24

5.4.2 Response to exchange rate of import price index 25

5.4.3 Response to exchange rate of consumer price index 26

• 5.4.4 Response to foreign price index of import price index 27

5.4.5 Response to foreign price index of consumer price index 28

5.4.6 Response to output gap of money supply 29

5.4.7 Response to output gap of consumer price index 30

5.4.8 Response to money supply of consumer price index 31

5.4.9 Response to import price index of money supply 32

5.4.10 Response to import price index of consumer price index 33

5.4.11 Response to consumer price index of money supply 34

5.5 Variance decomposition 35

5.6 Granger causality test 36

5.7 VARstable 37

5.8 Lagrange multiplier test 38

Chapter 6: Conclusion, Policy Recommendation and Future Work 39

6.1 Conclusion 39

6.2 Policy Recommendation 40

6.3 Future Work 41

References 42

Appendices 45

-List of figures Figure 1: REER, NEER and Exchange rate 1995-2012 6

Figure 2: Import by commodity group ofVietnam 1995-2010 7

Figure 3: Inflation ofVietnam 1996-2010 8

Figure 4: NEER, IMP and CPI of Vietnam 1999-2011 9

Figure 5: Response to exchange rate of foreign price index 25

Figure 6: Response to exchange rate of import price index 26

Figure 7: Response to exchange rate of consumer price index 27

Figure 8: Response to foreign price index of import price index 28

Figure 9: Response to foreign price index of consumer price index 29

Figure 10: Response to output gap of money supply 30

Figure 11: Response to output gap of consumer price index 31

Figure 12: Response to money supply of consumer price index 32

Figure 13: Response to import price index of money supply 33

Figure 14: Response to import price index of consumer price index 34

Figure 15: Response to consumer price index of money supply 35

Figure 16: VAR stable 37

List of tables

Table 1: VAR regression for import price index 21

Table 2: V AR regression for consumer price index 23

Table 3: Larange multiplier test 38

Table 4: ADF test of dlneerl 46

Table 5: ADF test of dlimp 46

Table 6: ADF test of dlcpi 47

Table 7: ADF test of dlm2 47

Table 8: ADF test ofdlopi 47

Table 9: ADF test of dlpi 47

Table 10: Optimal lag length 48

Table 11: Granger causality test 48

Table 12: Vector Autoregression results 50

Table 13: Impulse- Response Function 57

Table 14: Variance decomposition 69

"\

Abbreviations

Acknowledgement

I would like to give my sincerest thanks to my supervisor, Dr Dinh Cong Khai, who has provided me valuable support throughout the process of this thesis My study would not have been possible without his patience, encouragement, insightful comments and correction in all time of doing the research and writing the paper

I am also deeply grateful to Dr LeVan Chon for his dedicated, generous, enthusiasm and thoughtful helps about Stata technique and Econometrics model Thank to those important guidance and useful advice, I have overcome many obstacles to complete the research work

Last but not least, I would like to give a big gratitude to my dear family and friends who always encourage and stand beside me during my difficult moments Their love have inspired me a lot and lifted me up; hence my mere expression of thanks does not suffice

Abstract

This paper estimates the exchange rate pass through to Vietnam's import prices and domestic inflation with monthly data from January 1999 to October 2011 By applying Vector Auto Regression methods, Impulse Response function, Variance Decomposition as well as combining with Granger causality, V AR stable, Lagrange multiplier tests, the results show that exchange rate passed through almost fully to import price index immediately and strongly at the first month, after that it fluctuates

to go up and down next months The transmission from exchange rate to consumer price index is smaller than to import price index but it penetrates with longer periods around 10 months Expanding money supply just explains for variance of inflation 2.48% But it is not the surprising result, because the State Bank of Vietnam conducts monetary policy timely and appropriately The results of variance decomposition, however, reveal that the most important factor affecting on both import and consumer prices is domestic demand pressure (output gap) with 32% and 14% respectively Therefore the policy recommendation should focus on controlling domestic demand pressure as a priority to cut down high inflation Government also can be confident with easing monetary policy due to its small effect on inflation Nonetheless, the economy still needs the stable exchange rate policy in order to keep the trust on domestic currency

Chapter 1: Introduction

1.1 Problem statement

Exchange rate is one of the most important factors in macro-economic management

It affects supply and demand of foreign currencies, exports and imports, national debts and reverses When a domestic currency appreciates; the country's exports become more expensive and its imports become cheaper in foreign markets and vice versa for the depreciation of domestic currency Appreciation will make balance of trade lower, while depreciation will increase it

Vietnam is following the export-led growth strategy; therefore export activities need

to receive a priority In recent years, import activities, however, have increased significantly; reducing net import is also a main concern now Exchange rate adjustment to depreciate domestic currency will be the useful tool to achieve those two goals

However, when the State Bank depreciates domestic currency; this will cause pressure on inflation through three channels: the import prices increase will directly affect on domestic prices, export activities are pushed up thus the supplies for domestic market decrease and export prices at domestic market increase, and aggregate demand for non-tradable goods increases also put a burden on inflation Furthermore, depreciation has a common aspect like easing monetary policy; it means the economy will encounter with inflation risk In order to balance this effect, the State Bank needs to manipulate tightening monetary policy such as increasing reserve ratio or capital adequacy ratio Definitely, there is a close relationship between money supply and exchange rate; and government should consider this connection to control inflation effectively

We have a definition of ERPT as the percentage change in local currency import prices of importing country reacting to one percent change in exchange rate between trade partner countries (Goldberg and Knetter 1996)

It is important to understand ERPT for several reasons: first, in order to assess monetary policy transmission on prices and forecast inflation, we need to know the degree and timing of pass-through Second, knowledge of the size and speed of exchange rate pass-through into inflations is useful for the inflation targeting adoption (Lian, 2006)

ERPT has been studied since 1980s by Dombursh (1987), Krugman (1986), Knetter (1989), Taylor (2000), Toh and Ho (2001), Campa and Goldberg (2002), Otani, Shiratsuka and Shirota (2005), Ghosh and Rajan (2009) In the early stage, most of researchers focused on examining the ERPT in United States, Euro Area and developed countries They also had a debate on limited pass-through of exchange rate across countries due to deviation from the law of one price Recently, there have been papers studying this subject in emerging markets, new small open economies and developing countries The transmission of exchange rate into disaggregate industries has been focused as well

1.3 Research objectives

There have been, however, a few studies regarding ERPT in Vietnam It is necessary to measure ERPT to know how it affects on the economy, how it responses to government's policies This study aims to estimate the degree and speed of ERPT to distribution chain from import prices to domestic inflation (consumer prices) in Vietnam The study also examines the relationship between inflation and monetary policy as well as other factors

This research approaches the issue under Vector Auto Regression (V AR) method combining with the analysis of Impulse Response Functions (IRF) and Cholesky Variance Decomposition After that I will check Granger causaliy, VAR stable and

do Larange multiplier test

The null hypothesis will follow the hypothesis of many previous researches that the exchange rate has no effect on import and domestic prices (in other words, the ERPT is zero)

My research also has limitations Due to data availability, I cannot obtain Producer Price Index to establish the systematic distribution chain from Import Prices Index

to Producer Prices Index and finally to Consumer Price Index I also ignore the structural change when Vietnam joined WTO in 2007 This paper does not cover ERPT in disaggregated industries, and this will leave for researches in future

This paper includes 6 parts After the introduction is part 2 with the overview of Vietnam's exchange rate, import activities and inflation Part 3 is literature review; part 4 is research methodology, empirical framework and data description The empirical results are presented in part 5 Part 6 will conclude what I found in estimation process and then will suggest policy recommendation

-Chapter 2: Overview of Vietnam's exchange rate, import activities and inflation

2.1 Exchange rate arrangement in Vietnam

All we know that Vietnamese economy achieved many successes thanks to the Renovation or "Doimoi" program in 1986 The economy step by step changed from the centrally planned to the market-oriented system With this comprehensive reform package, the economy became more stable, open and free During the procedure, the financial system also was renovated towards market-oriented; therefore the role of exchange rate was very important According to Vo et al (2000) Vietnamese authorities has considered exchange rate control as a central macroeconomic instrument for many targets: to ensure low inflation rate, to keep financial system stable, to support exports, to control imports, and to push economic growth

Before 1989 there were three-tier exchange rate systems: trading, non-trading and internal exchange rate The trading exchange rate was used to do payments Non-trading exchange rate was applied for inwards remittance and intangible goods transactions such as tourism, education, health among socialist countries The rest one, internal exchange rate, served for business transactions between domestic banks and domestic companies using foreign exchange; internal exchange rate also was used when there had foreign aids coming from former Soviet Union In 1989 exchange rate was unified by a sharp devaluation of the official rate from VND225/USD to VND4500/USD and pegged to the United State Dollar during 1990s Commercial banks can set their own exchange rate transaction within +/-5% around the official rate This depreciation brought a very positive impact on exports and economic activities during the period of 1990-1991 However according to Vo

et al (2000) this instrument has been used with reserve and passively During the period of 1993-1996, the nominal exchange rate stayed relatively stable; and as inflation cut down, there appeared concerns about an overvalued Dong It is

estimated that over this period the real exchange rate and the real effective exchange rate appreciated about 20% and 15%, respectively This appreciation was the reason for current account deficits at that time Although exports grew at a high rate they still remained lower than imports, following very high current account deficits (peaking in 1996 at 12% ofGDP)

With the pressure of current account deficits due to the reduction of foreign direct investment as well as the East Asian financial crisis, export goods of Vietnam were less competitive with other countries, therefore the State Bank of Vietnam allowed the VND to devalue modestly and gradually similar to an adjustable fixed rate regime Vietnam also imposed stronger controls over imports and current account transactions As a result, trade and current account deficits decreased to about 4-5% ofGDP in 1997-1998 (Vo et al., page xi)

From 2000-2003, the trend of exchange rate continued depreciating to support for export-led growth strategy of government However, in 2004-2010 the inflation rate started to increase at faster rate than the rate of exchange rate depreciation, thus the nominal effective exchange rate and real effective exchange rate started to separate Vietnam Dong, actually, turned to appreciate and the comparativeness of Vietnamese goods was weaker than that of other foreign exporters The State Bank effort to devalue Dong in 2009 and 2010 to improve situation but the Vietnam Dong still relatively appreciated against the currencies of Vietnam's major trading partners

Figure 1: REER, NEER and Exchange rate 1995-2012

Source: Le Hong Giang (2012)

During 2000-2011 when Vietnamese economy grew quickly and more integrated in the world market The economy of Vietnam will not only cope with unstable domestic macro-economy but also suffer the global financial crisis Therefore exchange has continued the trend of depreciation through years

According to Nguyen (2011) in the sub period 2008-2011, we have ever witnessed the extreme fluctuation of Vietnam Dong vis-a-vis US dollar with many depreciation adjustments together with the wide band During this time the economy of Vietnam was hit by the global financial crisis, the unpredictable increase of gold prices as well as the big gap of official exchange rate and speculation "black market" rate By the end of 2009, VNDIUSD rate had increased 5.6% compared to that at the end of 2008 and in Feburary 2010, the State Bank of Vietnam increased the official rate from 17,941 VND/USD to 18544 VNDIUSD, 3.3% devaluation with the band +/-3% Next, Sate Bank declared an

•

unprecedentedly high rate of devaluation of VND 9.3% in early February 2011, increasing the official rate to 20,693 VNDIUSD and narrowed the band to +/-1 %

2.2 Overview of Vietnam's import activities

Vietnam has been suffering trade deficit for a long time and it still has had upward trend The proportion of import to GDP is increasing continuously from 50% in

2000, 84% in 2007, 93% in 2008, 79% in 2009, 88% in 2010 and 87% in 2011 (World Bank, 2012) We can see more about the proportion of disaggregated import goods The mean of production goods always takes the biggest proportion, hence it will be a difficult mission of the State Bank to use exchange rate as a tool to push export up and pull import down It may have reverse effect when depreciation currency, the import volume may not much reduce but the import value may strongly increase and cause burden for inflation

Figure 2: Import by commodity group ofVietnam 1995-2010

Imports by commodity group 1-995-2010

2.3 Overview of Vietnam's inflation

In the period 1985-1989, the average CPI of Vietnam was 263%, the failure of the

so call "General Adjustment of Price, Wage and Money" resulting the persistent hyper-inflation through years In order to solve this problem, the State Bank of Vietnam had to aggressively tighten monetary policy, raised monthly interest rate and pegged VND strictly to USD These policies included in the "Doi moi" program and brought the big successes to Vietnam's economy Inflation started to fell significantly to below 20% in 1992 and close to 10% in 1995 (Nguyen and Nguyen, 2010)

During the Asian financial crisis 1996-1998, the inflation rate was low and even negative in 2000-2001 as reported -0.5% The deflation in this period can be explained by the remarkably decline of world prices as well as aggregate demand After that the recovery period came in 2002, the inflation was around 5% during period 2002-2003 It started accelerating in 2004 with 9.5%, jumping to 12.7% in

2007 and hit the peak of23.1% in 2008 According to Nguyen and Nguyen (2010) there have been many reasons explaining for this strong comeback of inflation such as: the large increasing minimum wage, the rising of international commodity prices, the inflexible monetary policy, the slow response of exchange rate management or the joining of Vietnam to WTO at the of 2006

In 2009, tightened monetary policy combined with the world economic depression, inflation slowed down again It came back to 11.8% in 2010 and 18.7% in 2011 Figure 3: Inflation ofVietnam 1996-2010

CPI of Vietnam 1996-2011

Source: Author's calculation from World Bank (2012)

Now I combine data of exchange rate, here I use nominal effective exchange rate (NEER), import price and consumer price in one graph so we can have a general overview of them Nominal effective exchange rate and inflation are rather consistent; they have the same movement through time Import price index, however, likes an outlier in the graph with its own remarkably and continuously increase year by year

Figure 4: NEER, IMP and CPI ofVietnam 1999-2011

NEER, IMP and CPI of Vietnam

1999-2011

1999200020012002200320042005200620072008200920102011

Source: Author's calculation from Datastream

Chapter 3: Literature review

Generally, there are two approaches on ERPT effect Some economists such as Dombursh (1987), Feinberg (1986) and Krugman (1986) analyzed ERPT basing on microeconomics behavior of firms Another approach is based on macro variables

to find the degree of ERPT effect on economy or group of economies, we can name here Taylor (2000), McCarthy (2002), Campa and Goldberg (2002), Marazzi et al (2005)

There have been 3 methods to measure ERPT: linear, Vector Error Correction Model (VECM) and V AR

3.1 Linear approarch

First, using the linear econometrics model approach, Ihrig et al (2006) conducted the research for ERPT in G7 countries (the United States, the United Kingdom, Japan, Italy, Germany, France, and Canada) for both import prices and consumer prices with quarterly data from 1975Ql to 2004Q4 For each country authors applied two versions of linear equation, one for estimating import prices and one for consumer prices pass-though The results showed that in the last 15 years, by 10 percent depreciation in the local currency, the import prices increased by 4 percent and consumer prices increased by 2 percent on average The research also stated that ERPT declined for almost G-7 countries

Another research of Campa and Goldberg (2005) analyzed ERPT to import prices across countries and product categories in the Euro Area The results showed that ERPT to import prices was high, but not complete, and different across industries and countries in the short run The transmission in the long run was higher and closed to one The research tested for structural changes in pass-through rate when the Euro currency was introduced; there was a decline trend for two-thirds of the industries However, the evidence was not statistically significant except manufacturing industries The authors predicted a wider decline in pass-through may be happening but it was too early to conclude the structural change explaining

- - -

for such declines Because the research conducted in 2005, at that time, the history

of Euro currency was still very short; the small sample is one of disadvantage for testing structural

3.2 Vector Error Correction Model approach

Second, the VECM approach, there are many researchers choosing this method to measure ERPT in case of cointegration in model VECM has an advantage to measure long-term ERPT based on stable cointegration among variables; conditions, however, to apply this method is stronger than VAR approach We can name here many researchers applied VECM to measure EPRT such as Kim (1998), Beiner & Bijsterbosch (2009), Dahl & Lo (2005), etc

Beiner and Bijesterbosch (2009) applied the cointegrated V AR and impulse response from the VECM for a research about ERPT in Central and Eastern European member states Their results showed that ERPTs for domestic consumer prices were 0.6 and 0.5 for the cointegrated VAR and impulse response, respectively The paper found that ERPT in countries adopting fixed exchange rate regime is higher than in countries adopting floated exchange rate regime

There also has a combination between linear and VECM methods, Hanshilin (2006) conducted his research for New Zealand, he recognized that pass-through elasticity was incomplete in the short run and it varied across 8 disaggregated industries; in the long run, the pass-through elasticity was higher than short run and half of the sample of industries experienced complete exchange rate pass-through Finally, he concluded that the exchange rate pass-through rate was endogenous to inflation performance at disaggregate level under New Zealand's economy

3.3 Vector Autoregression approach

Final, the V AR approach is the most popular method which has been used by many researchers all over the world There are a wide range of papers concerning about ERPT in Euro Area, emerging countries, Latin countries, comparing those together

or just ERPT of a specific country such as United States, United Kingdom, Switzerland, China, etc

Faruquee (2006) estimated ERPT in Euro Area with monthly data from 1990 to

2002 The proportion of exchange rate shock in his research included: import prices, producer prices, consumer prices, wages and term of trade Based on the estimation

of V AR, impulse response functions and the Cholesky decomposition he concluded that the effect of ERPT to prices was small Prices were very sticky in local currency in response to depreciation in the euro exchange rate Over time, the degree of ERPT generally raise with the greatest response to import prices The wholesale producer prices had tendency to rise more than retail consumer prices Twelve to eighteen months after the shock, ERPT to export and import prices are about one-half and one, respectively Finally, the terms of trade for the euro area declined in response to exchange rate depreciation

Zori et al (2007) measured ERPT in emerging markets based on 12 countries in Latin America, Asia and Central and Eastern Europe Their results confirmed that ERPT declined across the pricing chain The analysis also demonstrated that ERPT was higher in emerging than in developed countries For emerging economies with one-digit level of inflation, ERPT was low and similar levels in the developed economies Doing more research, the paper found that there was a positive correlation between the degree of the ERPT and inflation Finally, the research pointed that although level of inflation was controlled, the relationship between import openness and ERPT is weaker than that between inflation and ERPT

One study of Marazzi et al (2005) focused on ERPT in US; the study pointed out that the ERPT to import prices in US declined from 0.5 in 1980s to somewhere neighborhood of 0.2 and it explained for this decline by "the composition of imports has shifted toward goods whose prices are less sensitive to exchange rate movements" (p 3) together with "China's increasing presence in the U.S market"

(p 4)

Jin (20 1 0), with ERPT in China study, explored that nominal effective exchange rate appreciated 1% will decline 0.132% of domestic consumer prices and 0.495%

of producer prices, respectively Jin also illustrated that exchange rate regime affected on inflation, particularly; CPI pass through in fix exchange rate regime was higher than in flexible exchange rate regime Combining above results, he suggested Chinese government should pursue more flexible exchange rate policy Furthermore, the research also examined the aspect of price control, basket and weight of Chinese price indices, distribution cost, non-tradable share and imported input in order to find the reason for low EPRT to consumer price inflation

Vo (2009) in his study about ERPT to inflation in Vietnam figured out the degree of ERPT to inflation in Vietnam is 0.61 He also compared with ERPT level of other countries such as Indonesia (0.53), Korea (1.59), Thailand (1.27), Singapore (0.59),

23 OECD countries (0.46) and 10 euro area economies (0.47) and concluded that ERPT to inflation in Vietnam is moderate In his study, however, he did not count exporters' cost (foreign price index) in the model My study will compute this variable and expand the sample size from M1:1999 to M10:2011 comparing to M1:2001 to M2:2007 The nominal exchange rate and import prices in his study are weighted of 9 main trading countries, but in my paper I generate these variables from weighted trade of 20 countries I also apply more V AR stable and Larange multiplier tests for regression model Therefore the result is rather different with Vo' s result

Under perfect competition market with no trade barriers, the PPP states that the home price of goods in home currency should be equal to the foreign price of goods

in foreign currency multiplied by the exchange rate between those two currencies

We have an equation as follow:

Where h denotes the home country, f is foreign country Pf:l is the import price measured in home country currency Pf is the export price measured in foreign currency, E is the nominal exchange rate between home currency vis-a-vis foreign currency (home currency/foreign currency)

According to Hooper and Man ( 1989), Goldberg and Koetter ( 1997) and Campa and Goldberg (2002), the exporting firms set the prices depended on the markup

(Markup x) on marginal cost of production of foreign exporting firms (CJ)

Hooper and Man ( 1989) argue that an exporting firms set up their markup basing on the market demand pressure in both foreign and home markets ( Y ), and competitive pressure in the home country market On the other hand, the competitive pressure in the importing market is measured by the profit margin, i.e price over production costs Therefore, the markup is presented as follows (Vo, 2009):

Markupx = [:;;]" yP (3)

Where [ Phx]a represents competitive pressure in the home market and yP represents

ECf

demand pressure in both home and foreign market and 0 <a <1, ~ > 0

From (1 ), (2) and (3) we have the import price as follows:

P,;" = E [:;;]" yPCf (4)

Simplify equation (4) we get:

(5)

We take logarithm equation (5):

We can rewrite equation (6) as simple linear equation by generating variables as follows:

Equation (7) shows that import price is affected by the exchange rate e, marginal

cost of production of foreign firms c[ , the home country price level Ph and market

demand for both home and foreign country y

From equation (7) we can know the elasticity of import price with respect to the change of exchange rate (ERPT) of home country is (1- a) If a = 0, we have complete pass-through effects If a = 1, we have zero pass-through and 0 < a < 1,

we have limited or incomplete pass-through

;

We will arrange equation (7) to obtain a, and then we can analyze the movement of ERPT Because ERPT is measured by coefficient (1 - a), ERPT has an inverse relationship with a

a= -~ ~

From equation (8) we have the movement of ERPT as follows:

If demand in both home and foreign markets (y) increase, a decreases and hence ERPT increases and vice versa for case of a decreases

If Ph increases, raising inflation in the home market, a decreases and ERPT increases and vice versa;

If p~ increases, a increases and ERPT decreases and vice versa;

If home currency depreciates (ore increases), a will decrease and ERPT increases

and vice versa

If c[ increases, a will increase leading to the decrease of ERPT and vice versa

4.2 Empirical framework

This study will apply the Vector Auto Regression (V AR) approach to measure the level of ERPT According to Faruquee (2006), using the V AR approach to examine exchange rate pass-through has several advantages compared with single-equation methods First, the single equation method estimates pass-through into a single price import or consumer prices separately It does not further distinguish between the types of underlying exchange rate shocks that may be arriving By investigating exchange rate pass-through into a set of prices along the pricing chain, the V AR analysis characterizes not only absolute but relative pass-through in upstream and downstream prices (p.64, 65) Second, after running V AR model, we can apply Cholesky variance decomposition to identify specific structural shocks affecting the

system Using this identification scheme, one can map the empirical results into a well-defined shock in an economic model of incomplete pass-through

One study of McCarthy (2002) used VAR method to examine the transmission of exchange rate to import price, producer price and to domestic price However data

of producer price of Vietnam is not available; this paper can only apply V AR method for import price (IMP) and then domestic price (CPI)

We have matrix of V AR approach as follows:

n

Yt = Wo + L /3i Yt-i + Ut

i=l

Where Yt is the 7 vectors of variables [OPI, NEER, FPI, IMP, GAP, CPI, M2], w 0

is intercept, f3 i is coefficient of matrices 7x7 and ut is error term

OPI: oil price index (Crude Oil price index, Brent UK), NEER: nominal effective exchange rate, FPI: foreign exporters' price index, IMP: import price index, GAP: output gap, CPI: consumer price index, M2: broad money

4.3 Data description

Data to estimate ERPT to import price index and domestic price index is monthly time series from January 1999 to October 2011 with 154 observations; January

1999 is the base All data are extracted from International Financial Statistics (IFS)

of IMF organization, Asia Regional Integration Center (ARIC) of Asia Development Bank (ADB), Datastream and General Statistics Office (GSO) of Vietnam Specifically, oil price index, broad money (M2) are extracted from IFS; import price index and NEER, REER, foreign price index, CPI are extracted and calculated from Datastream; IPI is extracted from ARIC combining with GSO

All variables will be generated under natural logarithm form Since they have characteristic of time series, we need to do Augmented Dickey-Fuller test (ADF) to find the existence of unit root except output gap variable Then taking first

differences and ADF test one more time will be applied respectively The results show that all variables are integrated at I( 1)

Oil price index (OPI) is used to represent for foreign demand pressure OPI will be exogenous variable because it affects all the rest of variables and it is not affected

by any variable in the econometric model Therefore, we need to figure out the lag length of var regression model and then applying for oil price index

Nominal effective exchange rate (NEER) stands for exchange rate e; NEER is used because NEER exhibits the exchange rate of Vietnam with trading partners correspondingly NEER is calculated basing on trade weight of 20 mainly countries trade partners with Vietnam However basing on the formula NEER is quoted as foreing currency/domestic currency, to take it easy I will generate a new variable NEERl = 1/NEER to quote to VND/foreign currency With this change we can see that when VND depreciates, nominal exchange rate increases, and this change also

is consistent with the definition of the exchange rate used in the pass through literature

Domestic demand pressure is presented by GAP (Output Gap = Real GDP Potential GDP) Due to data availability, I cannot obtain monthly GDP Therefore, like many previous empirical studies, I will use Industrial Production Index (IPI) to proxy for real GDP Then I apply Hodrick-Prescott filter to compute output gap; it

-is the residuals from a regression of the log of industrial production index on a constant plus linear and quadratic trend With this method the variable GAP does not need to be seasonally adjusted because Hodrick-Prescott filter extracted data into two part smoothing and error term This error term is output gap and it is stationary as well

In this research I calculate foreign exporters' price index (FPI) to get one more control variable for model and FPI also represents for the adjustment mark-ups of exporting firms to the change in the exchange rate

Based on Campa and Goldberg (2002), the exporters' price index is calculated as follows:

Import price index is calculated by weight of 20 mainly countries being trade partners with Vietnam The formula is below:

Consumer price index (CPI) is used to represent for domestic inflation ADF test showed that CPI has unit root, therefore it needs to take first difference to get stationary

Details regarding variables NEER and REER computation will be explained in Appendices

Chapter 5: Empirical results

The results of ADF test show that all variables (except GAP) have unit root Therefore we need to take first difference of them and then apply ADF test again for those new variables The results tell us that we will reject Ho (variable has a unit root) at I(l)

I do not apply ADF test for GAP; because HP filter removed cyclical component during its processing Hence, all variables are now stationary The results of these tests will be available in Appendix

5.2 Optimal lag length

There are many lag length selection criteria such as: Aikaike's information criterion (AIC), Schwarz information criterion (HQIC), Hannan-Quinn criterion (HQC), etc Stata will support us to apply all these criteria to find the optimal lag length Moreover, theoretically, with relative large sample 120 or more observations, HQIC

is a good choice Combining both empirical test and theory we have the optimal lag length is 2

5.3 V AR regression

The oil price index is the exogenous variable; it affects all variables and is not affected by other variables The variables will follow the order: [ dlneer 1, dlfpi, gap, dlm2, dlimp, dlcpi] with exogenous dlopi

The results of V AR regression for import price index are in table 1:

•

Table 1: V AR regression for import price index

*, ** and *** denote statistical stgmficance at 1%, 5% and 10% level, respectively

There only, however, foreign price index of lag order 1, import price index of lag orders 1 and 2 and oil price index of lag order 2 have statistical significance at 1% level Output gap of lag order 1 has statistical meaning at 5% level Exchange rate

of lag order 1 is statistical significant at 10% level but exchange rate of lag order 2 does not have statistical meaning

The results of regression show that when exchange rate at the first lag increases 1 %, import prices will over react to 1 011% This means ERPT to import prices is bigger

than 1 and we have over complete pass-through effect When exchange rate

changes, it will transfer to import prices immediately and completely at lag order 1 Hence the result of exchange rate at time order 2 is insignificant

Return to Figure 4, the graph about nominal effective exchange rate, import price and consumer price also partly tells us that import price index is increasing out of control Moreover, the structure of Vietnam's import is mostly for intermediate goods, necessary goods for production When VND is depreciated to push export; firms still need to import more to produce finished products Thus the import will continue increasing both volume and value It happens like a spiral circle till

Vietnamese companies step by step establish the industrial production for input

The response of import prices is now easy to understand with foreign price index

We have the same situation of above exchange rate The import prices response extremely to the change of foreign price index The result of foreign price index will tell us the stubborn resistance of import prices When the change of foreign price index at time lag 1 increases 1 %; the change of import prices at present period will increase 3.122% This procedure, however, happens very quickly just within one month The foreign prices change at period lag order 2 does not have statistical meanmg

The world oil price also impacts on import prices with 1% change of oil price

leading to 0.41% change of import prices This number reflects the dependence of Vietnam' economy on the world price deeply

The results ofVAR regression for consumer price index are in table 2:

Table 2: VAR regression for consumer price index

*,**and*** denote statistical significance at 1%,5% and 10% level, respectively

All independent variables are significant 5% level except money supply of lag order

2 is significant at 1 0% level

The effect of exchange rate to inflation is bigger at first and is smaller later, specifically; with 1% change of exchange rate will cause 0.104% and 0.068% change of consumer price index at the first and second lag orders respectively We can see that the transfer of exchange rate to import prices is quick and strong, while that of to consumer prices is rather moderate and gradual

The result of foreign price index also affect to consumer price index with average 0.11 % In this V AR regression result the effects of output gap, money supply and import prices to inflation are rather small However I will analyze them intensely in next impulse-response and variance decomposition processes so we can know the extent as well as the importance of each variable through time

The oil price does not have statistical meaning, this can be understood that oil prices transferred to import prices but not to consumer prices

All results of V AR regression will be presented in Appendices; I just extract two results of import prices and consumer prices in here

5.4 Impulse response function:

With impulse response outcome we can trace out how typical shocks will affect a variable through time In this paper, I apply impulse response function for 24 months

5.4.1 Response to exchange rate of foreign price index

As mentioned in the previous sections, foreign price index variable presents for the response of foreign exporters to the change of exchange rate The result of impulse response function shows that one percent increase in exchange rate will increase 0.0198 percent in foreign price index at the first month This can be understood that when currency is depreciated, the price of foreign currency becomes more expensive and thus raising positive response as well Going through to second month, however, foreign exporters adjust markup to keep their market shares; hence the response of foreign price index become negative (-0.394%) But foreign exporters only keep this decrease in very short time After occupying market share, they start to increase foreign price index (0.394%) and then slightly decrease prices

in next months The response ends within sixth months However the decreases are almost near zero hence, in generally, the shock of exchange rate does not affect on the prices of foreign exporters The results are reasonable because in the world market, the power of foreign exporters is stronger than that of Vietnamese

•

importers Foreign exporters are price leaders and Vietnamese importers are price takers

Figure 5: Response to exchange rate of foreign price index

result1: dlneer1 -> dlfpi

5.4.2 Response to exchange rate of import price index The shock of one percent of exchange rate will lead to positive responses of 0.970 percent of import price index at the first month and negative response to - 0.008 percent at the second month and then turning to -0.395% at the third month However, in the fourth and the fifth months the change of import price index becomes positive again; after that it reduces gradually to zero beginning from the tenth month The response of import prices can explain as when there a shock of change in exchange rate happens, almost the shock will complete pass through to the change of import price immediately after one month The shock continues penetrating in second month and reduces in third month But when economy starts

being familiar with new status of exchange rate, the import starts going back to its increase trending in fifth and sixth months and ending from tenth month

Figure 6: Response to exchange rate of import price index

result1: dlneer1 -> dlimp

5.4.3 Response to exchange rate of consumer price index

After going through transmission chanel from exchange rate to import prices, the shock of change in exchange rate transfers to the change of consumer prices is rather small With 1% increase of exchange rate, there will always have positive change in consumer price index 0.109% in the first month, 0.118% in the second month and 0.061% in the third month and so on The transmission process begins moving to zero from the twelfth month Although the transmission from exchange rate to inflation is smaller than to import, the resistance of exchange rate to inflation

is longer than that of to import

Figure 7: Response to exchange rate of consumer price index

result1: dlneer1 -> dlcpi 2

5.4.4 Response to foreign price index of import price index

The 1% positive change of foreign prices causes significant change of import prices instantly 3.06% at the first month This response tells us that import reacts to the change of foreign market very quickly and fully It is one again confirms that import activities of Vietnam depending on the fluctuation of foreign market deeply The second and third months, import price index reduces remarkably with -1.95% and-1.02% However, import price index, in fourth and fifth months, reverses to increasing trending The procedure commences to die out at the eleventh month

Figure 8: Response to foreign price index of import price index

result1: dlfpi -> dlimp

5.4.5 Response to foreign price index of consumer price index

The shock of foreign price index leads to positive response of consumer price index during the whole 24 periods At the first month inflation increases 0.121%, the second month inflation increases 0.185%, the third and fourth months the responses are 0.098% and 0.081% respectively The procedure is going on and it becomes extinct from the eleventh month The shock of foreign price index transferring to inflation is more stable and gradual than to import It is logical, as we are examining the pass-through by distribution chanel, import price index is a buffer to receive the shock before foreign price index affects on consumer price index

Figure 9: Response to foreign price index of consumer price index

.3

.2

step 95%CI impulse response function (irf)

5.4.6 Response to output gap of money supply

In Figure 10, we can see that when output gap is up, money supply is down The negative results of money supply responding to positive shock of output gap reveal that the change or the difference of money supply later period is smaller than previous period This reflects the tightening monetary policy of the State Bank The State Bank of Vietnam perceives that inflation increases due to output gap increases; therefore it shrinks money supply timely When real GDP is bigger than potential GDP; output gap increases It means economy will suffer inflation; we need to contract money supply to reduce inflation accordingly

Figure 10: Response to output gap of money supply

result1: gap -> dlm2 1

5.4 7 Response to output gap of consumer price index

The results show that increasing in output gap will follow the increasing in consumer price index When output gap increases 1%, consumer price index increases 0.039% at the first month and 0.013 at the second month The response turns to negative in the third month but it returns to positive in the fourth (-0.0046%) and all months after Although the degree of output gap affecting to inflation is small, combining with the results of variance decomposition in next parts we can confirm that output gap is the second factor most impacted on the movement of inflation

Figure 11: Response to output gap of consumer price index

result1: gap -> dlcpi

5.4.8 Response to money supply of consumer price index

When the State Bank increases 1% of money supply, consumer price index will increase 0.039% at the first month, 0.067 at the second month Consumer price index is negative in next of the third and the fourth months and positive in the rest

of months The result of consumer price index responses to the shock of money supply is small It means expanding money supply will not affect so much on inflation Therefore the State Bank can be confident to control monetary policies flexibly This can be explained that the State Bank of Vietnam conducted monetary policy following the change of gap to avoid inflation as we discuss in the part 5.4.6 This helped minimizing the effect of money supply to inflation and the State Bank

is doing well with monetary policy