Research " DYNAMIC ANALYSIS ON ASEAN STOCK MARKETS " ppt

This dissertation investigates the underlining characteristics of stock markets in Indonesia, Malaysia, the Philippines, Singapore, and Thailand through empirical analyses.. The finding

Doctor of Philosophy Degree

® UMI

UMI Microform9982126

Copyright 2000 by Bell & Howell Information and Learning Company All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code

Beil & Howell Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, Mi 48106-1346

Souther Illinois University

April 4, 2000

I hereby recommend that the dissertation prepared under my supervision by

Praphan Wongbangpo

Entitled Dynamic Analysis on ASEAN Stock Markets

be accepted in partial fulfillment of the requirements for the

DOCTOR OF PHILOSOPHY degree

PRAPHAN WONGBANGPO, for the Doctor of Philosophy degree in Economics, presented on April 4, 2000, at Southern Illinois University at Carbondale

TITLE: Dynamic Analysis on ASEAN Stock Markets

MAJOR PROFESSOR: Subhash C Sharma

The countries of the Association of the Southeast Asian Nations (ASEAN) had

experienced the skyrocketing growth in both real and financial sectors over the past two

decades During that period, stock markets in these ASEAN countries have boomed and

become investment icons for international investors This dissertation investigates the

underlining characteristics of stock markets in Indonesia, Malaysia, the Philippines,

Singapore, and Thailand through empirical analyses

Chapter one reviews the economic cooperation and financial development in the

ASEAN countries The basic characteristics of their stock markets are also discussed

Chapter two indicates that ASEAN stock markets, except for the Philippines,

share a long-run co-movement, implying that an effective long-term diversification of an

investor’s portfolio among these stock markets cannot be achieved The finding reveals

that the stock markets of Malaysia and Singapore are under-performed relative to their

long-term trends and classified as trend-dominated markets, whereas those of Indonesia

and Thailand are cycle-dominated since the market indices outperform their trends

Chapter three verifies that the ASEAN’s stock price indices and nominal

exchange rates are bound together in a long-run relationship A positive stock

price/exchange rate relation through the inflationary effects is observed for Indonesia,

Malaysia, and the Philippines, while the asset view of the exchange rate justifies the

negative relation for Singapore and Thailand The causality from stock prices to

observed only for Malaysia, Singapore, and Thailand Thus, foreign investors are

recommended to adopt different investment strategies across ASEAN financial markets

because of the diverse relationships between stock prices and exchange rates

Chapter four reveals that the ASEAN stock markets are generally linked to their

macroeconomic variables in both long and short run Moreover, the ASEAN stock price

indices cause and are caused by their macroeconomic variables in the Granger sense

Thus, we note that ASEAN stock markets are inefficient Therefore, investors are

encouraged to thoroughly observe economic performances in the ASEAN region to

maximize their earning opportunities from these stock markets

Regarding policy implications, policy makers should be cautious in their

implementation of foreign exchange and macroeconomic policies since they have

ramification effects on respective stock markets, and vice versa Moreover, the finding of cointegrated ASEAN stock markets means that there is a need for policy coordination

among the ASEAN to mitigate the impacts of financial fluctuations Greater policy

coordination, including the reduction or removal of trade and investment barriers, will be essential if these countries are to exploit the advantages of financial interdependence

Chapter 1 ASEAN: Economies and Financial Development

Ll

1.2

1.3

1.4

L5

Introduction .-. -.QQQ HQ nen nhe sư nưet l

Economic Growth and Financial Liberalization in ASEAN 3

The ASEAN Stock Markets 4

Outline and Objectives of the Study ccằŸnằằằằằ 6 Significance of the Study .c cee ccceeeceeeee see cee cee eeecee nee eeeereeeeteene 9 Chapter 2 Long Term Trends and Cycles in ASEAN Stock Markets 2.1 Umtroduction mm h- ỐỐÔ 12 22 Methodology ¬ ress ecuesnee esse 15 2.2.1 Umit Root Tests -oQẶ cà nà nhe 15 2.2.2 Johansen Cointegration Approach 17

2.2.3 Trend and Cycle Decomposition - 18

2.3 Empirical Analysis ẶẶ Sàn he nhe nhe 20 ;” ` KA"› a a=a 20

2.3.2 Test for Stationarity ào sen 21 23.3 Long Run Relationship -.Ặ-Ặằ nà ere eeeeees 21 2.3.4 Testing Restrictions on the Cointegrating Vector 22

2.3.5 Bivariate Cointegration Analysis - 24

2.3.6 Japanese and ASEAN Stock Markets: Multivariate Framework 25

24 — Trend and Cycle Decomposition 26

2.5 Comclusion .a.- 30

Chapter 3 Stock Price and Exchange Rate Dynamic Interactions: Evidence fom ASEAN Countries 3.] — Introduction TQ HH nh vn re 33 3.2 Methodology .ẶĂẶẶĂĂ se 37 3.2.1 Unit Root Tests 37

3.2.2 Johansen Cointegration Approach 38

3.2.3 Granger Causality Test - 40

3.2.4 Variance Decomposition Analysis 41

3.3 Empirical Analysis 0.0 cce ccc eee cee cece rene cee eee eee enenneeeenees 41 ::m)', 1a 41

33.2 Test for Stationarity 42

3.3.3 Johansen Long Run Relationship 43

3.3.4 Granger Causality Tests 44

3.3.5 Variance Decomposition Analysis 47

3.4 Conclusion and Discussion - 48

¬— -.: 1 ^ệ 50

4.2 The Hypothesized Model ọoọ cece cee cec cee cee cee ceu see ceeceneees 53 4.3 Mcthodologỵ Q.QQQQQQQQQQQHQ HH ky nh sen 56 4.3.1 Unit Root Tests 2.222 0.0 ccc ccc cee cee cee cee ceeeee tec eeesunenseees 57 4.3.2 Johansen Cointegration Approach 58

4.3.3 VECM and Granger Causality Test 60

3.2.4 Innovation Accounting Analysis 61

4.4 Empirical Analysis 5 61

aadđaiiaiaiaiaaaáẢẢ 61

4.42 Test for Stationaritỵ 63

4.4.3 Long Run Relationship 63

4.4.4 Granger Causality Tests 67

4.4.5 Innovation Accounting Analysis 69

4.4 Conclusion and Policy Implications 71

Chapter 5 Conclusion and Discussion 5.1 | Summaries of the Study .0 0.0.0 coc eee eccecccncecccecececcuecscececeeees 76 5.2 Implications of the Study « 2.2 occ cc cece ccc ece cee ccceeeee cea eencnaae 77 5.3 Sustaining Economic Growth Policỵ 79

Reference «20.0.0 cece cee cee cee eee cee cee cuecesectersee ves cencescceeuctecesveetensesetersees 122

VITẠ oon ce cee eee cee eee cee cence cee tees eecessecusceceteeescrsseaerecaecerereecrstcusteness 130

ASEAN: Economies and Financial Development

1.1 Introduction

The plunge in the Thai baht during July of 1997 severely shook not only her own financial market but the world’s as well This downturn in the Thai financial sector

immediately affected her Southeast Asian neighbors creating a collective crisis The

contagious nature of this financial crisis has raised an economic concern about financial

market interdependence among the Southeast Asian region This is a concern of utmost

importance when one considers the prospects for a successful continuation of the

financial liberalization being experienced by the Association of Southeast Asian Nations

(ASEAN) Hence, for ASEAN countries the impact of financial integration on their

markets and their macroeconomic performance has become a very important issue

The Association of Southeast Asian Nations (ASEAN) was established in August

1967 with the signing by five original member countries, namely Indonesia, Malaysia, the Philippines, Singapore and Thailand’ The three main objectives of ASEAN are: to

promote the economic, social and cultural development of the region through cooperative

programs, to safeguard the political and economic stability of the region against high

powered rivalries, and to serve as a forum for the resolution of intra-regional differences

The ASEAN economies are highly diverse economically and socially in terms of

size, historical background, resource endowment, stages of economic development, and

" Brunei was later accepted into the Association in 1984, Vietnam in 1995, Laos and Myanmar in 1997.

is that they are all market-based economies with a high degree of export dependence The

economic integration of ASEAN is rationalized by that a large ASEAN market, in the

form of ASEAN Free Trade Area (AFTA), would encourage industrial development and

intra-regional trade A large market would enable firms to reap economies of scale, lower

unit costs, and encourage new investment in industrial projects designed to cater to the

entire ASEAN markets In addition, increased intra-regional trade would force firms to

become more efficient through competition from intra-regional imports

In the early years, however, ASEAN was more preoccupied with the stability and

security of the region, especially with the territorial disputes between Indonesia and

Malaysia, and the escalating war in Vietnam Even though the economic cooperation in

the past was limited, a new path has been undertaken to take advantage of resource

pooling and market sharing (Naya and Plummer, 1991) The initiation of the AFTA, established by the Fourth ASEAN Summit Meeting held in Singapore in 1992, was the

most significant step in enhancing the economic cooperation and trade in this region The

AFTA aims to make the region borderless by dismantling trade barriers among member

countries, hoping that this will lead to lower production costs through efficiency gains

Besides the economic cooperation in the AFTA, ASEAN growth triangles have also been established in the late 1980s to combine competitive strengths of its constituent areas’

Cooperation in investment, transportation, communications and services are also part of

? There are four ASEAN growth triangles, which are the Southern Indonesia-Malaysia-Singapore Growth

Triangle (IMS-GT), the Northern Indonesia-Malaysia-Thailand Growth Triangle (IMT-GT), the Eastem Brunei-Indonesia- Malaysia-Philippines East ASEAN Growth Area (BIMP-EAGA) and the Greater Mekong Subregion (GMS) (Chia and Pacini, 1997, p.63-61).

force

1.2 Economic Growth and Financial Liberalization in ASEAN

A remarkable record of consistently high economic growth for ASEAN in the last

decade deserves attention ASEAN has been one of the fastest growing regiona! groups in

the world In the period 1987-1992, the growth rate of real GDP of the ASEAN-5S

(Indonesia, Malaysia, the Philippines, Singapore, and Thailand) averaged 7.3% This

average growth rate was significantly above that experienced by developed market

economies, 2.8% as a group; exceeded that achieved by North America, 2.5%; and

surpassed that realized by the world, 2.2% (Tongzon, 1998, p 16) Individually, the ASEAN ’s average annual real GDP growth rate during the period 1987-1995 was around

9% for Malaysia, Singapore and Thailand, while Indonesia and the Philippines achieved

6.6% and 3.3% respectively (Chia and Pacini, 1997, p 9) Table 1.1 summarizes this

growth rate of real GDP for the period 1990-1995

Ariff (1996) and Chia and Pacini (1997) point out that the ASEAN economic boom could be partly attributed to the financial liberalization policies undertaken in

earnest in this region since the early 1970s These reforms attempted to expand the

domestic capital mobility, as well as, to induce inflow of international capital

Liberalization in the financial sector of Malaysia and Singapore started as early as 1973

The reform of their financial sectors was executed by opening markets to international

influences, removing restrictions on the deposit interest rate, and abolishing the current and capital account restrictions Similar reforms were undertaken a decade later in

steps toward full financial liberalization by removing the interest rate ceiling and opening

the current and capital accounts In the early 1990s, the ASEAN countries have

implemented another wave of liberalization policies Both Malaysia and Singapore

focused on the growth in capital markets; an investor protection was provided in

Malaysia, and a preferential tax treatment for investors was provided in Singapore In

1988, Indonesia adopted a policy to develop financial institutions and improve deposit

yields Thailand embraced a banking reform in 1990 by expanding the scope of banking

business and liberalizing branches Concurrently, the Philippines passed a new foreign

investment law that allowed up to 100% foreign ownership in selected companies and

eliminated all regulations on foreign exchange restrictions In all, sequenced and gradual

reforms in both real and financial sectors have improved the efficiency of the financial

systems in these ASEAN countries

1.3 The ASEAN Stock Markets

The globalization of the capital forming process introduces a unique challenge to

the ASEAN capital markets In the early 1980s, when the ASEAN stock markets took

off, funds from depressed European countries and recession-bound Japan were tempted

by the high profit margin from these markets Between 1993 and 1996, money managers

from the United States,.Europe, and Japan pumped short-run credit into the ASEAN

region and, consequently, heated the economies of Thailand, Indonesia, and Malaysia

The total foreign investment inflow into the “Asian—Gang of Five” (Indonesia, Malaysia,

the Philippines, Thailand and Korea) skyrocketed during the period of 1988-1991 and

of foreign investment, as well as financial liberalization and deregulation undertaken

domestically, had created a boom in ASEAN capital markets

Table 1.2 illustrates the basic characteristics of the ASEAN stock markets All of

the market trading systems involve continuous trading; all listed securities are available

for trading during the entire time the market is open for trading This allows market

participants to obtain more accurate information regarding the price quotations and

trading volume Short sales and insider trading are not permitted on any of the ASEAN

markets Traded securities do not result in immediate delivery Instead, the number of days required for settlement, delivery and payment of a market trade varies from three

business days in Thailand, to seven business days in Malaysia and Singapore Regarding

foreign investment, each market maintains separate regulations governing foreign

investment, including the repatriation of income and capital There are no significant

restrictions on foreign investors purchasing shares in Malaysia and Singapore Some

restrictions on foreign investment in Indonesia and Thailand indicate that a special

restriction regarding registration ownership applies to foreign investors In the

Philippines, foreign investors are allowed to trade exclusively in a certain class of stocks

designated for foreigners

During the last two decades, the rapid economic growth in the ASEAN countries

was accompanied by an incredible increase in the size of stock markets The ASEAN

security capitalization from 1990 to 1996 is reported in Table 1.3 Over the seven-year

period, Indonesia experienced the greatest percentage increase in market value of 816

percent Malaysia’s growth rate for the same period was 360 percent, the Philippines 638

capitalization arose from both price appreciation of listed firms and significant increases

in the number of firms listing their securities on the exchanges This immense growth in

market capitalization can be characterized as the emerging market phenomenon

However, the ASEAN stock market crash in July 1997 brought all markets in this region

into a collective financial crisis This contagious effect suggests that stock markets in the

ASEAN countries are closely linked

Table 1.4 presents the risk ratings from International Country Risk Guide (ICRG) The ranking takes into account political, financial, and economic issues At the end of

1992, the Philippine stock market was considered the least attractive with respect to the aforementioned considerations, followed by the Indonesian and Thai markets Palac-

McMiken (1997) also derives ASEAN stock market risks and returns from the investment return and market volatility As shown in Figure 1.1, he points out that the Indonesian

stock market bears the highest risk and return The returns to both the Philippine and Thai

stock markets are roughly at the same level but the Thai market carries much less risk,

making it comparatively more attractive to investors The low risk in the more advanced

markets of Malaysia and Singapore is accompanied by low return In this study, the rank

for risk and return across ASEAN stock markets calibrated by Palac-McMiken (1997)

will be referred to as the characteristics of these stock markets

1.4 Outline and Objectives of the Study

in the recent years, the performance of the ASEAN countries has been subject to

renewed debate While their impressive growth record has been sustained well, new

international economic environment, involving the global integration of goods and capital

markets The ASEAN countries have been among the first to take considerable advantage

of the new opportunity afforded by these trends

The expeditious growth in the ASEAN's economies and stock markets in the last

two decades can be considered as an indication of the successful financial reform

employed by these countries Explaining the successes of ASEAN countries, at least to

some degree, in the growth of real and financial sectors has become a challenging task

To this end, three hypothetical concerns regarding the ASEAN stock markets and their

economies are as follows

1) The recent financial crisis has caused a widespread financial disaster in stock markets

of ASEAN countries The contagious nature of this financial crisis has raised an

economic concern about stock market interdependence among countries in this

region In other words, do the ASEAN stock markets integrate? Do these markets

share similar characteristics or movements in the short and long run? How can the

market structure of each ASEAN stock market be analyzed in comparison with those

and long-run relationships between stock price indices and foreign exchange rates?

Are these relationships identical across ASEAN countries?

real and financial sectors In this circumstance, do ASEAN countries provide any evidence on the relationship between the real and financial sectors? Particularly, do

the stock market and relevant macroeconomic variables in each ASEAN country

move together in the long run? What is the short-run relationship between them? Are

the short and long-run relationships identical across ASEAN countries?

This dissertation proposes to examine the stock markets in the ASEAN region during

the period from 1985 to 1996 through empirical investigations of the interconnection

among stock markets, the interaction between stock and foreign exchange markets, and

the relationship between the stock market and macroeconomic factors It is the objective

of this dissertation to shed light on the economic issues regarding the ASEAN stock

markets Due to the unavailability of the relevant data of some countries’ economies, this

dissertation only focuses on five countries in the ASEAN: Indonesia, Malaysia, the Philippines, Singapore and Thailand The outline of the dissertation is as follows

Chapter two investigates the long-run relationship among stock markets of five

ASEAN countries In order to demonstrate a unique long-run equilibrium relationship

among these markets, the structural economic relation underlying this long-run

relationship is examined Moreover, each stock price index is decomposed into

permanent (trend) and transitory (cycle) components so that market structures of each

stock price series can be characterized individually and comparatively analyzed in

relation to others in its group

Chapter three explores the long-run equilibrium and short-run dynamic

relationships between stock and foreign exchange markets in the five ASEAN countries

stock and foreign exchange markets is investigated for each ASEAN country In addition,

the short-run dynamic relationship between the price variations of these financial markets

(stock prices and exchange rates) will be analyzed through the Granger causality tests and

variance decomposition analysis

Chapter four examines the relationships between the stock market and economic

factors in each of ASEAN countries The long-run co-movement between stock prices

and selected macroeconomic variables; the gross domestic product, the consumer price

index, the money supply, the nominal interest rate, and the exchange rate, will be

investigated Equilibrium relations experienced by these stock markets then will be

compared and contrasted to those of the more advanced and efficient stock markets in

Japan and the United States In addition, the relationship between stock markets and

selected macroeconomic variables will be examined under the short-run dynamic

framework of Granger causality, impulse response function, and variance decomposition

Chapter five draws together the results from chapters two through four Some

policy implications toward ASEAN stock markets will be discussed

1.5 Significance of the study

Even though the stock markets in these five ASEAN countries have become

investment icons for international investors as emerging financial markets, few empirical

studies have examined these ASEAN economies This dissertation attempts to examine

the underlining characteristics of ASEAN stock markets through the empirical analysis

Investigating ASEAN countries, which are small developing and export-oriented

economies, on the concerning issues differentiates this study from others Moreover, the

resolutions of this study are likely to have policy implications for other small emerging

economies

Some important issues can be addressed given the results of this study The

cointegration among ASEAN stock markets implies that these equity markets share a

long-run co-movement, and thus an effective long-term diversification of an investor’s

portfolio among ASEAN stock markets cannot be achieved Moreover, cointegrated

ASEAN stock markets have implications for regional financial stability A given

country’s economy cannot be effectively, in the long run, insulated from foreign

influences if they are cointegrated Destabilizing effects transferred from others through

the environment of an integrated capital market will make pursuing an independent

monetary policy difficult Consequently, integration of ASEAN stock markets facilitates

harmonized monetary policies within the region

As the choice of currency domination adds an important concern to investors’

overall portfolio decisions The study of dynamic relationships between stock and foreign

exchange markets provides information regarding the relationship between two financial

markets, which then could be used to enhance the ability of a multinational portfolio

management regarding rates of return and profit opportunities Given observed relations,

the government will also be able to choose appropriate measures to boost or remedy bearish stock markets

However, in order to avoid an unpleasant outcome on the real sector, government

should be concerned with the nature of the relationship, whether it is a short-run or long- run phenomenon, between the stock market and key macroeconomic variables For

example, stock market intervention in the money market may have little or no effect in

the short run but will eventually have an effect in the long run if the relationship between

stock prices and money supply is in fact a long-term relationship The results from this

study will address this issue

Moreover, the observed short and long-run relationships between the stock market

and selected macroeconomic variables yield empirical evidence for economists A

causality that goes from macroeconomic variables to the stock prices confirms the ability

of a stock price to perform its fundamental role in recognizing changes in economic

conditions A reverse causality (i.e., from the stock price to selected macroeconomic

variables) indicates that the stock market’s performance can be considered as signaling

the future performance of the macro-economy

Long Term Trends and Cycles in ASEAN Stock Markets

2.1 Introduction

The growing globalization of security markets has brought increased attention

from investors and academic scholars to stock markets throughout the world An

extensive body of empirical research has been produced attempting to justify and to

describe the ways in which equity markets integrate The stock market crash of October

1987 provides one example in which the global nature of the event reveals evidence of

global stock markets reacting one to another A second example is the 1997 stock market

crash in ASEAN countries Recent literature has utilized various econometric techniques

to investigate whether or not stock markets across countries systematically converge in

either the short or the long run

The stock market co-movement among G7 countries and other industrialized

countries has been the primary focus of past research Lai, Lai and Fang (1993) observe both the short-run and the long-run feedback relationships between the New York and

Japanese stock markets, suggesting that the US stock market does not necessarily

dominate the other stock markets Kasa (1992), by using price and dividend indices, finds

a single common trend driving the stock markets in the G7 countries Recently, Serletis and King (1997), following the methodology used by Kasa (1992), conclude that the ten

European Union stock markets tend to move toward a long-run relationship Moreover,

Koutmos (1996) investigates the dynamic first and second moment interactions of major

European stock markets, and concludes that these markets are cointegrated in the sense

that they react not only to loca! news but also to news generated from other countries A

long-run relationship among European stock markets has been observed for the period of

the 1920s and 1930s by Choudhry (1996) On the other hand, Choudhry (1994) does not

support the existence of common stochastic trends among G7 countries Gallagher (1995)

supports Choudhry’s (1994) claim and notes that the long-run relationship among Irish,

British and German stock markets do not exist in either the stock price level or stock

return For the Scandinavian stock markets, Booth, Martikainen and Tse (1997), by using

an Exponential Generalized Autoregressive Conditional Heteroscedasticity (EGARCH)

model, observe that these markets are weakly related to each other

Some researchers have investigated the long-run relationship of stock markets in

the Asian and Pacific countries Engle and Susme! (1993), by using an Autorgressive

Conditional Heteroscedasticity (ARCH) model, test for common volatility driving

eighteen nations’ stock markets and find two closed-boundary groups that share similar

volatility characteristics Masih and Masih (1997) study four Asian Newly Industrializing

Countries (NICs); Taiwan, South Korea, Singapore and Hong Kong, and investigate how the well-established markets propagate these markets; Japan, USA, UK and Germany The result indicates that the well-established markets, regardless of which established market is being considered, drive the fluctuations in NICs’ stock markets Other

researchers discuss different Asian stock market phenomenon; for example, Hung and

Cheung (1995) find a long-run relationship among five Asian Emerging Stock Markets

(ESMs); Hong Kong, Korea, Malaysia, Singapore and Thailand However, Kwan, Sim

and Cotsomitis (1995) observe that four Asian ESMs (Hong Kong, Singapore, Korea and

Taiwan) are not cointegrated among themselves but they are cointegrated with G7

countries On the other hand, Chan, Gup and Pan (1992 and 1997), in a bivariate

cointegration analysis, find support for the international diversification in five major

Asian stock markets and in European Union (EU) stock markets Furthermore, Corhay,

Rad and Urbain (1995) address the significance of the regional aspects of the common

stochastic trend in the stock markets among Pacific-Basin countries They find that in the

long run there exists a geographical separation between the Asian and the Pacific

markets In the context of Southeast Asian or ASEAN countries, Palac-McMiken (1997) applies the bivariate Engle and Granger cointegration analysis into the ASEAN stock

markets, concluding that all the markets are linked together with the exception of

Indonesia However, based on the weekly data of the period 1988-1995, Roca,

Selvanathan, and Shepherd (1998) examine the linkage in stock markets of these five

countries and conclude that these markets are related in the short term, but not

significantly linked in the long run

Compared to studies of developed countries’ stock markets, there are a limited

number of studies for the ASEAN region A detailed investigation of the ASEAN stock

markets is of interest because of the increased economic cooperation in accordance with

the ASEAN agreement, the successful financial reform, and the distinguished structure of

emerging stock markets Thus, the main purpose of this study is to investigate the

characteristics of the ASEAN stock markets and provide explanations for the market

movements There are two objectives of this study First, we investigate whether there is

a common long-term relationship among the stock price indices between Indonesia, Malaysia, the Philippines, Singapore and Thailand This will shed some light on the

understanding of ASEAN countries in terms of financial liberalization and international

diversification Second, given that these stock price indices are cointegrated, we analyze

the degree of co-movement of ASEAN stock indices by decomposing each series into

trend and cycle The trend will be analyzed as the permanent component, while the cycle

can be considered as transitory component that causes the fluctuation from trend By

decomposing these five ASEAN stock price indices into trends and cycles, we shall

examine the characteristics of these markets across the region

The study is structured as follows Section 2.2 discusses the methodologies used

in the study Data and empirical findings on the study of cointegration are elaborated in

section 2.3 The study of trend and cycle decomposition is presented in section 2.4

Finally, summaries and concluding remarks are provided in section 2.5

recovered using the methodology of Vahid and Engle (1992)

2.2.1 Unit Root Tests

In this study, both the Augmented Dickey Fuller test (ADF, Said and Dickey, 1984) and Phillips and Perron test (PP, Phillips and Perron, 1988, and Perron, 1988) are employed to test for the unit roots in the underlying stock price series The ADF tests are

based on the following equations, which account for the presence of a non-zero mean in

equation (2.1) and a non-zero mean with linear trend in equation (2.2)

where Xj represents the ASEAN stock price indices (Xi = Indonesia, Xx, = Malaysia, Xx

= the Philippines, X= Singapore, and Xs, = Thailand), p and B are non-zero mean and linear trend terms, respectively The unit root process is tested under the null! hypothesis

ofH.: œ =0 by using tas test statistics in (2.1) and H,: a = 0 by ta- test statistics in (2.2) The critical values are given in Fuller (1976, pp 371 and 373)

The PP tests are based on the following equations

X = pt aXe + &, j=1,2, 5 — (1A)

Xp

where Xj represents the time series as indicated above, T is the number of observations,

w+ Bt-T/2) + aXe + &, j=1,2, 5 (2.2A)

ụ and B are non-zero mean and linear trend terms, respectively In (2.1A) the null

hypothesis that H,: œ = | is tested by using the Z(œ`) and Z(tas) test statistics and Ho: u=

0 and a= 1 is tested using Z(®,) test statistic In (2.2A) the null hypothesis that H,: a” =

1 is tested by the test statistics Z(a) and Z(ta-) and that H.: B = 0 and a = | by using test statistics Z(®) and that H,: B= 0 and ði=O and a = | by using test statistic Z(®2) The

adjusted Z test statistics are given in detail in Perron (1988, pp 308-309) The critical

values of these adjusted Z test statistics are given in Fuller (1976, pp 371 and 373) and

Dickey and Fuller (1981, p 1063).

2.2.2 Johansen Cointegration Approach

The maximum likelihood cointegration approach introduced by Johansen (1988,

1991) is used to determine for the number of cointegrating vectors The procedure begins

by expressing the stochastic variables in a (nx!) vector, X;, as the unrestricted vector

autoregression (VAR) The VAR model used in this study is

where X, = [Xn, Xx, Xx, Xa, XsJ’ is a (5x1) vector of ASEAN stock price indices, A; is

a (5x5) parameter matrix, c is a (5x1) constant vector, is the (5x1) vector of random error terms with zero mean and constant variance, and p is the lag length Following

Johansen (1988) and Johansen and Juselius (1990), the system of equations (2.3) can be

written in its first difference form:

AX, = PAX, + PAX + + [prAXtpe + MWXip + &

The existence of the long-run relationship among the ASEAN stock price indices

is suggested by the rank of IT matrix, r, where r is 0 <r <n The two matrices @ and B

with dimension (nxr) are such that af’ = IT The matrix B contains the r cointegrating vectors and has the property that B’X, is stationary a is the matrix of the error correction

presentation that measures the speed of adjustment in AX,

The likelihood ratio (LR) test statistics for the hypothesis that are the following:

A

A

where A,;’s are the n —r smallest squared canonical correlations between the residuals of

Xpand AX, series, corrected for the effect of the lagged differences of the X process, and

T is the number of observations The A-trace test statistic tests the existence of at least r

cointegrating vectors against a general alternative, while the null hypothesis of r against

r+1 cointegrating vectors is tested by À-max

A test for the linear restriction on B reveals information regarding the structural

economic relationship underlying the long-run model We use the LR test developed by

Johansen (1991) to test this economic restriction The hypothesis for a linear restriction in the matrix of cointegrating vector can be set up as:

where B is a (nxr) cointegrating matrix, H is a (nxs) matrix with n-s restrictions, and ọ is

a (sxr) matrix The LR test statistics is as the following:

LR: TL [In(i-4) - In(-ÀA)] ~ xÃdÐ (2.8)

il

where df = r(n-s) is the number of degree of freedom, A; and A; are the eigenvalues based

on restricted and unrestricted eigenvectors, respectively

2.2.3 Trend and Cycle Decomposition

The short-run co-movement in a time series is called a common cycle, while common trend characterizes the co-movement in the long run (Engle and Issler, 1992) A common cycle exists in the I(1) variables when the linear combination of their first

differences conditional on all observed information prior to time t is unpredictable and so

called ‘cycle free’' Therefore, searching for a common cycle is equivalent to searching

for the zero (statistically insignificant) correlation of the linear combination between the

first difference of the variables and their past information set

Given that the series are cointegrated, Vahid and Engle (1992) propose a test statistic to test for a common cycle using the canonical correlation analysis Vahid and

Engle (1992), Engle and Issler (1992), and Issler and Vahid (1992) show that the test for common cycle is in fact the test for zero statistically insignificant canonical correlation

between AX, and (B'X:.1, AX:-1, AX:2, ., AXtp+1) The notion that there exist at least s zero canonical correlations, i.e., the s dimension of the cofeature space, is tested under

the null hypothesis The test statistic proposed by Vahid and Engle (1992) is

C(ps) = -(T- pz In (1-A;7) ~ x? with (np‘s+rs-ns+s*) degree of freedom (2.9) where 1,’ are the s smallest squared canonical correlations between AX, and (B’X:.1, AX.i, AXv2, , AXtp+1), T is the number of observations, P is the lag length of VAR system in difference, i.e., p-l, and r represents the number of cointegrating vector Vahid

and Engle (1992) condition the existence of common cycle upon the existence of linearly independent linear combination:s such that the system shares n-s common cycle

In the special case of co-movement when n = r+s, the permanent and transitory components of each series can be recovered through the trend and cycle decomposition

by jointly exploiting short and long run co-movement restrictions (Vahid and Engle,

1992) In our trend and cycle decomposition analysis, X;, is a (nx1) vector of non-

stationary stock price indices from ASEAN countries Let a (nxn) matrix A = [a”'a’]’

See the detailed discussion of the common cycle in Vahid and Engle (1993).

where a is a (rxn) matrix of cointegrating vector, œ is a (sxn) matrix of a co-feature vector Then A is full rank and will have an inverse Partition the columns of the inverse accordingly as A’ = [a |a’] and get the trend and cycle decomposition as

X = (TAX = a (aX) + a(a’'X%) (2.10)

T + G

| Trend + Cycle where T, contains only trend because œ ”X: is a random walk and hence cycle free, and C, characterizes cycle since a’X, is [(0) and serially correlated

2.3 Empirical Analysis

2.3.1 Data

The end-of-month closing share price indices of the Jakarta composite stock price index (JCSPI) for Indonesia, the Kuala Lumpur stock exchange composite index (KLSE) for Malaysia, the Philippine stock exchange composite index (PSE) for the Philippines, the Stock Exchange of Singapore index (SES) for Singapore, and the stock exchange of Thailand index (SET) for Thailand from January 1986 to December 1996? are used in this study These stock price indices obtained from DataStream (Thailand) The starting date

is appropriately chosen to correspond to market transactions when the ASEAN stock markets emerged from the era of financial regulation Extending the study to a longer period poses the risk of altering the financial scenario Prior to the empirical analysis, the series are transformed into natural logs

2.3.2 Test for Stationarity

The results of unit root tests are given in Table 2.1 Both the ADF and PP tests

fail to reject the null hypothesis of the existence of a unit root when the log level series

are used, but reject the same null hypothesis for the log first difference of the series

Thus, each stock price index series is integrated of order 1, or I(1)

2.3.3 Long Run Relationship

To proceed with the long-run cointegration analysis, the order of the VAR system

needs to be determined The various likelihood ratio (LR) tests are performed and

examined for the exclusion of the (p-1)" lag We begin with twenty-four lags The

general-to-specific procedure yields a VAR model of nine lags As reported in Table 2.2,

no autocorrelation in the residuals is found for the system at nine lags

Johansen (1992) suggests a systematic test procedure for the cointegration model

specification, by jointly testing both the rank order and the deterministic component’ As

indicated in Table 2.3, the A-trace statistics at the 5 percent significant level yields the

simultaneous finding of the model with an intercept term and one cointegrating vector

Tests for the properties of individual series‘ suggest that the Philippines can be omitted from the cointegrating relation; the Philippine stock market does not share a

long-run equilibrium relationship with the rest of ASEAN markets Economically, the

Philippines possesses the smallest degree of trading relations with other ASEAN-5

3 See Harris (1995, p 97) for the test procedure

“ This procedure provides useful information, particularty the basic features of the data, in the exploratory phrase of the empirical analysis (Hansen and Juselius 1995, p 64).

countries’ The characteristics of the Philippine stock market also raise the question of its

insignificance relative to others The Philippine stock market has experienced formidable

obstacles Over the years, political turmoil, scandals and economic hardship have stunted

the development of the Philippine stock market Regarding the market capitalization, the

Philippines is considered the smallest relative to other ASEAN stock markets Moreover,

a small number of big Philippine stocks comprising roughly up to 87 percent of the

market capitalization dominates and tends to dictate the movement of market indices As

of the end of 1992, only 10 out of 170 listed stocks were considered attractive to foreign

investors (Price, 1994, p 293) Hence, the argument of the long-run exclusion in the case

of the Philippine stock market is credible The LR test, R, in Table 2.4, fails to reject the

null hypothesis that the Philippines can be excluded from the cointegration space

Therefore, the restricted model of ASEAN-4, i.e., X, = [Indonesia, Malaysia, Singapore, Thailand]’, is estimated As indicated in Table 2.5, the ^-max and À-trace statistics at the

5 percent significant level recommend one cointegrating vector (r=1) for this restricted

system

2.3.4 Testing Restrictions on the Cointegrating Vector

To obtain a unique and economically meaningful cointegrating relation, we

proceed by imposing and testing restrictions on the cointegrating relationship A robust

relationship between the stock markets of Malaysia and Singapore is hypothesized responding to the fact that their economies are the most closely linked The bold

distribution of the inward Foreign Direct Investment (FDI) flow and the strength in trade

5 From the 1993 to 1996, the Philippines has the smallest share, 5-6 percent in import and 2-4 percent in

export, of the total trade in the intra-ASEAN parameters (hup://www.asean.or.id, as of June 4, 1999).

intensity indexes between these two economies (Tongzon, 1998, pp 28 and 49), in

association with their geographical proximity and cultural factors elucidate the strong

relationship between these two markets In fact, they had developed from the very same

root as The Stock Exchange of Malaysia and Singapore, and became independent entities

in 1973 Despite the market separation, at the end of 1994, more than 110 Malaysian

stocks were actively traded as well as a few other foreign issues, accounting for more

than 50 percent of the total market capitalization in Singapore (World Stock Exchange

Fact Book, 1996) The long-run interdependence between the Malaysian and Singaporean

stock markets is also documented by Palac-McMiken (1997), and supported by the

analysis in section 2.3.5 of this study Accordingly, the hypothesis of proportionality

between the stock markets of Malaysia and Singapore is imposed on the cointegrating

vector As indicated by R2 in Table 2.4, the null hypothesis of the homogeneous relation

with opposite sign between Malaysia and Singapore clearly can not be rejected,

indicating that they systematically enter the cointegrating relation together, along with

Indonesia and Thailand

Besides their strong relationship, the stock markets of Singapore and Malaysia are

by far the best equipped in Asia, outside of Tokyo and Hong Kong, to absorb foreign

investment Malaysia and Singapore have received approximately two thirds of the FDI

flows in the ASEAN region (Tongzon 1998, p 146) Moreover, the capitalization in both

markets accounts for roughly 75 percent of the total stock market capitalization in this

region The fact that stock markets of Malaysia and Singapore are highly developed

relative to other ASEAN markets justifies the notion that these variables are weakly

exogenous to the system and can enter the right hand side of the vector error correction

model (VECM) The LR tests of joint hypotheses, R3, R4 and Rs in Table 2.4, strongly

indicate that Malaysia and Singapore should be considered exogenous to the system

Conditioning on the weakly exogenous variables of Malaysia and Singapore, the

partial cointegration model is estimated The residuals of the above system are tested for

serial correlation and reported in Table 2.6 The Liung Box test and Lagrange Multiplier

test indicate that the residuals are not serially autocorrelated Under the interpretation for

cointegration, we cannot reject the hypothesis of a single cointegrating vector among

stock price indices of ASEAN-4 The coefficients of linear combination are arbitrarily

normalized by the first B from the cointegrating vector, or Indonesia in this case Hence

the stationary linear combination of ASEAN-4 stock price indices can be uniquely

defined as the following:

Indonesia = - 21.693 + 9.556 Malaysia - 9.556 Singapore + 3.100 Thailand (2.11)

2.3.5 Bivariate Cointegration Analysis

To reexamine the previous finding by Palac-McMiken (1997), the Johansen cointegration technique is applied to ten pairwise possibilities As reported in Table 2.7,

only four long-run relationships are found to be significant One interesting result is the

presence of cointegration between the Singaporean stock market and all the other

ASEAN markets, with the exception of the Thai market Therefore, the stock market of

Singapore can be taken as representative of the ASEAN stock markets, as it is the market

found to be the most interactive with all the others Rationalizing that Singapore is the

financial center in the ASEAN region justifies the significance of Singapore in binding

the long-run relation among ASEAN markets The result also detects the strong long-run

relationship between Malaysia and Singapore, which was also well observed in the

multivariable cointegration test

2.3.6 Japanese and ASEAN Stock Markets: Multivariate Framework

Japan has always shared a close economic link with ASEAN countries As one of

the ASEAN major trading partners, in 1995, Japan shared 14.3 percent of ASEAN's total

exports, and was the largest source of ASEAN imports, accounting for 23.1 percent of the total (Tongzon, 1998, p 118) Furthermore, Japan has been the important source of FDI

for countries in this region The ASEAN countries experienced a surge in FDI from Japan

in the 1980s, particularly in the post 1986 period when Japan's Outward investment were

pushed by the sharp yen appreciation ASEAN’s share of Japanese outward FDI rose

from 3.8 percent in 1986 to 12.4 percent by 1994 While Japan’s outward FDI world wide

fell sharply after peaking in 1989, its FDI in ASEAN region showed a less steep decline

in 1990-91 and has grown absolutely since then (Chia and Pacini, 1997)

However, the relationship between the Japanese and ASEAN stock markets is

rather inconclusive Based on the causal relationship, the Japanese stock market seems to play a moderate leading role as a regional factor for Asian emerging stock markets,

including those of ASEAN (see Cheung and Mak, 1992) Ghosh, Saidi and Johnson

(1999) investigate the relationship between the Japanese and ASEAN stock markets by applying the bivariate Engle and Granger cointegration approach into daily closing value

stock price indices for the year 1997 They find that Japanese stock market shares long- run relationships with that of Indonesia, the Philippines and Singapore, but not with that

of Malaysia and Thailand

To investigate the long-run co-movement between the Japanese and ASEAN ©

stock markets in the multivariate framework, the Tokyo Stock Price Index (TOPIX) is included in the study From Table 2.1, the ADF and PP tests indicate that the Japanese

stock price indices are [(1) series The Johansen cointegration analysis for the Japanese

and ASEAN stock markets, summarized in Table 2.8, suggests that the Philippine stock

market is still not a part of the long-run relationship The LR test also fails to reject the

null hypothesis of weak exogeneity for Japan, implying that the Japanese stock market

can be considered weakly exogenous for the model Moreover, the A-trace and A-max

Statistics indicate, at the 95 percent significant level, that the long-run structure of the

ASEAN-4 and the Japanese stock markets are bound together with two cointegrating

vectors These cointegrating vectors are found to be the long-run relationship among

ASEAN-4 themselves as previously discussed, and the relationship between Japan and

ASEAN-4 This finding is in line with those of Cheung and Mak (1992), and Ghosh, Saidi and Johnson (1999), verifying the significance of the Japanese stock market as the

regional leader, at least, in the Southeast Asia

2.4 Trend and Cycle Decomposition

The procedure for the trend and cycle decomposition begins with the test for the

common cycle As proposed by Vahid and Engle (1992), the squared canonical

correlation is obtained from the analysis of the relationship between AX, and (B'X,.:,

AXi.1, AXc2, AXcg) The null hypothesis that the s smallest squared canonical

correlation equals zero, or there exists at least s cofeature space, is tested against the

existence of (s+1) cofeature space In Table 2.9, the common cycle test rejects the null

hypothesis that the cofeature space has a dimension of four, suggesting the system

possesses three cofeature vectors and one common cycle As tabulated in Table 2.10, the number of cofeature (3) and cointegrating vectors (1) add up to the number of variables

in the system (4) Thus the trend and cycle decomposition is preformed to recover the

permanent and transitory components of ASEAN-4 stock price indices

Figure 2.1 reveals that the cyclical components of Indonesia and Thailand are

positive and move together during the entire sample period As a cyclical component is

derived from the actual series deviating from the trend, the stock market performances in

Indonesia and Thailand achieve higher market levels than do their long-term trends On

the other hand, the cyclical components of Malaysia and Singapore are below zero, suggesting that their stock market performances stay below their trends Moreover,

Indonesia, Thailand and Singapore follow the same cyclical pattern, even though they are

inversely related, while Malaysia has its own rather stable path over the entire sample

period The inverse relationship between the cyclical components of Singapore and

Indonesia/Thailand is quite interesting as it emphasizes the significance of Singapore as

the financial and investment center for the ASEAN region The significantly higher risk

and return in Indonesian and Thai stock markets relative to those of Singapore and

Malaysia allow the higher-return opportunity to occur in the short term Investors in the

stock market of Singapore seize the opportunity to gain higher return by taking advantage

of the bullish markets in Indonesia and Thailand At the same time, investors in

Indonesian and Thai stock markets could reduce their investment risk by including

Singaporean securities in their portfolio Due to the higher risk and return in Indonesia,

the cyclical components in Indonesia market exhibit more volatility than in Thailand

This arbitrage opportunity has less influence in the Malaysian stock market

Notice in Figure 2.2 that the trends from the Malaysian and Singaporean stock

markets are quite comparable They project identical patterns and directions toward the

future regardless of market bullishness or bearishness Meanwhile, the stock market of

Thailand shared the same trend movement of those in Malaysia and Singapore until early

1991, when foreign investors lost confidence in Thai market due to the military coup in

February 1991 This resulted in a huge loss in Thai stock market performance Since

foreign investors observed no fundamental change in political direction after the military

disturbance, the Thai stock market has become relatively stagnant in the international

portfolio since 1992 (Agtmael, 1993, p 125), reflecting the steady trend in Thai stock

market For the Indonesian market, the trend moved close to its peak in late 1988, and fluctuated there until the mid 1990 Following Price (1994, pp 265-266), this incident

could be explained by a series of deregulation packages, reforms, and policy adjustments

issued during the period 1987-1988 The intention of the Indonesia’s government was to

spur development of its capital market It succeeded for only a few years In mid 1990,

Indonesia experienced a two and a half-year period of difficult times Bearishness

reemerged in the market due to the fear of a tightened money supply, disappointing

cooperate performances and other concerns of market regulations, i.e., inadequacy of

supervision and a lack of reliab!e information about listed companies (Price, 1994, p

266) To resolve the problems, new attempts at upgrading the market’s professionalism had been implemented In 1991 the Capital Market Supervisory Agency was created as a

supervisory body for the Indonesian stock market (Price, 1994, p 267) Clearly, the

market policy adjustment, as well as the influx of FDI in 1993, pushed Indonesian market

indices up drastically One striking point here is that the permanent components of the

Indonesian and Thai stock markets did not register the similar leap, as did the markets of

Malaysia and Singapore One possible explanation is that the permanent component of

Indonesia and Thailand are less sensitive to the aggressiveness of foreign investors The

stock markets of Indonesia and Thailand still impose some restrictions, despite the

financial liberalization policy, toward foreign investment, i.e., foreign investors in

Indonesia and Thailand are required to ensure the repatriation rights before involving in

the markets (Park, 1993, p 11 and Price, 1994, pp 18-19) Due to rather strict market structure, the permanent components in Indonesia and Thailand are quite stable even

though the market indices are pushed upward by the inflow of FDI With this regard,

local policy makers could help sustain the stock market performances by deregulating restrictions regarding foreign investment

In Figure 2.3, we plot actual series relative to its own trend and cyclical

components for ASEAN-4 countries Their characteristics are analyzed in order to shed

some light on the nature of each stock market Generated by the profit opportunity in

ASEAN-4 scenario, the nature of these stock markets can be classified as either trend-

dominated or cycle-dominated markets In cycle-dominated markets, such as in Indonesia and Thailand, the transitory component plays a major role in clarifying the deviation of actual series from their trends The permanent components from both Indonesia and

Thailand are fairly stable and experience very little growth due to the unstable political

issues A higher return in the Indonesian and Thai stock markets attract foreign

investment, including the portfolio investment from Malaysia and Singapore, and

consequently causes the market indices to outperform market trends The increase in gap

between the actual series and their trends started to take off when Indonesia and Thailand

seriously implemented financial liberalization policies in mid-1980s The actual-trend

gap is even broader with the influx of FDI in 1993 On the other hand, Malaysia and

Singapore are categorized as trend-dominated markets Indicating by their below zero

cyclical components, the actual series in these markets are under-performing relative to

their long-term trends because portfolio investment in these markets is drawn to higher

returns in Indonesia and Thailand markets This profit opportunity creates a stronger

impact with respect to market under-performance in Singapore than in Malaysia

2.5 Conclusion

With respect to the fundamentals, the Philippine stock market is relatively small

in terms of capitalization and trading volume, and rather restricted Moreover, the

Philippines experiences very few trading transaction and economic linkage with other

ASEAN countries The likelihood ratio test confirms that the Philippines does not share a

long-run equilibrium relationship with the rest of ASEAN markets This finding implies

that the Philippine stock market complies with the international diversification, i.e.,

foreign investors could diversify by including the Philippine stock market in their

portfolio

There exists strong long-run interdependence between the Malaysian and

Singaporean stock markets The strong economic relationship between these two

economies could be verified by their geographical proximity, economic and cultural

factors, their strong trade intensity, and historical factors The concept is overwhelmingly

supported by the bivariate Engle and Granger cointegration test (Palac-McMiken, 1997), and by both the bivariate and multivariate Johansen cointegration tests in this study

Moreover, the stock markets of Malaysia and: Singapore are highly developed relative to other ASEAN markets, and comprise roughly 75 percent of the total capitalization in the

ASEAN region Hence, it can be concluded that these two markets are weakly exogenous

in the cointegrating model

This study reveals the existence of a long-run equilibrium relationship between

the ASEAN-4 markets Singapore and Malaysia are strongly bound with the economic

interdependence The higher-return opportunity ties stock markets of Indonesia and

Thailand to that of Singapore, but not to that of Malaysia The integrated ASEAN stock

markets imply that an effective long-term diversification of an investor’s portfolio among

ASEAN stock markets cannot be achieved Moreover, pursuing an independent monetary

policy is sub-optimal as the financial linkage among countries increases Therefore, the

integration of the ASEAN stock markets promotes harmonized monetary policy within

the region

The economic and financial linkage between Japan and ASEAN is conceivable, as

Japan is the major trading partner and the most significant source of FDI for ASEAN

countries Based on the cointegration analysis, the long-run relationships between the

Japanese and ASEAN stock markets are observed, while Japan is considered weakly

exogenous to the system The finding signifies Japan as the regional factor leader in the

Southeast Asian region

The stock markets of Malaysia and Singapore share similar permanent and

transitory components Both markets under-performed, relative to their long-term trends,

and are classified as trend-dominated markets On the other hand, the cycle-dominated

markets are represented by the stock markets of Indonesia and Thailand since their

market indices outperform their market trends The classification of trend- and cycle-

dominated markets in the ASEAN properly coincides with the level of stock market

development The stock markets of Malaysia and Singapore are highly developed and

mature relative to those of Indonesia and Thailand Removing the foreign investment

restrictions, as well as continuing to implement financial liberalization, would raise the

market trends in Indonesia and Thailand, upgrade their market development, and even

push them toward the trend-dominated market category Projecting the analysis to the rest

of ASEAN, the stock markets of the Philippines and new comers (i.e., Brunei and

Vietnam) are less developed and most likely fall into the cycle-dominated class,

assuming that they become emerging markets with high return Gradual implementation

of the financial liberalization and deregulation of existing foreign investment restrictions would prepare these new comers for the era of international portfolio investment and globalization of equity markets