How the ASEAN Exchanges integration affects firms’ position on the ASEAN Link and abnormal returns: evidence from

Presented by Nattawoot Koowattanatianchai

1 http://www.bus.ku.ac.th Authors

Nattawoot Koowattanatianchai* Kasetsart Business School, Kasetsart University, Bangkok, Thailand Email: [email protected] Tel: +66-87-539-3525; Fax: +66-2-579-0946

Kulkunya Prayarach Fiscal Policy Office, Ministry of Finance, Bangkok, Thailand Email: [email protected]

*Corresponding author at Kasetsart Business School, 50 Ngam Wong Wan Rd. Ladyao, Jatujak, Bangkok 10900, Thailand. Tel: +66-87-539-3525; fax: +66-2- 579-0946, email: [email protected]

2 http://www.bus.ku.ac.th Introduction

• Under the ASEAN Economic Community (AEC) Blueprint, ASEAN will move towards achieving regional economic integration by 2015.

• ASEAN countries have agreed on a development plan for financial and capital market integration, called ASEAN Exchanges, to harmonize ASEAN capital markets.

3 http://www.bus.ku.ac.th Introduction

Another key initiative is to streamline ASEAN access through security link by integrating seven exchanges of ASEAN countries, these being Bursa , Stock Exchange, Ho Chi Minh City Stock Exchange, Stock Exchange, the Philippine Stock Exchange Inc., Exchange, and the Stock Exchange of Thailand (SET).

4 http://www.bus.ku.ac.th Introduction

• ASEAN Trading Link (ATL) is an inter-connected exchange hub platform or a network gateway that connects ASEAN exchange markets. In essence, the ASEAN Trading Link makes it as simple for investors to trade in other ASEAN capital markets as it is to trade in their own domestic market.

• “ASEAN Stars” has been introduced so as to provide investors with an easily identifiable reference made up of the thirty most strongly performed stocks from each exchange, ranked in terms of market capitalization and liquidity. ASEAN Stars could be traded by local and international investors via the ATL.

5 http://www.bus.ku.ac.th Introduction

We look at the the impact of ASEAN Exchanges integration on the position of Thai firms in the ATL, which is one of economic cooperation initiatives developed by ASEAN member states in an attempt to realize the ASEAN Economic Community in 2015.

6 http://www.bus.ku.ac.th ATL VS Cross-listing

Similarity • Both aim to link securities from one country to another so as to obtain numerous benefits, such as lower cost of capital and greater liquidity, and allow investors to trade foreign securities conveniently.

Main difference • Cross-listing is governed by the host country’s regulations and other legal frictions (e.g., taxes and financial statement standards), while having a position on the ATL is not.

7 http://www.bus.ku.ac.th Hypotheses

H1: Stronger capital market integration leads to higher linked probability. • Investors are expected to adopt a regional investment policy; that is, they will take into account information from both domestic and other regional markets when making investment decisions. • Investors will have more incentive to spread their investments across a high-quality portfolio of ASEAN companies with highly different risk and reward profiles. To take advantage of these expected behavioral changes, firms will attempt to link their shares to other regional markets, in addition to their own domestic market, to increase their ability to raise funds in the region of linked stock markets.

8 http://www.bus.ku.ac.th Hypotheses

H2: Joining the ATL results in fewer abnormal returns. • Cross-listed securities are likely to be traded by price incentive based public information, leading to low or no abnormal returns (Korezak and Lasfer, 2008). • Cross-listing in overseas capital markets helps to reduce market segmentation (Foerster and Karolyi, 1999; Miller, 1999; Wang et al., 2008). Increasing market integration or reducing market segmentation (i.e., separate stock markets move together with high correlation) reduces the benefits from portfolio diversification across countries (Click and Plummer, 2005; Janor et al., 2007; Karim and Karim, 2012). • Markets with homogenous products produced and purchased by many buyers and sellers would earn a normal profit. This also pertains to equity markets, implying that a negative relation between expected equity’s returns and industry competitive position exists (Lyandres and Watanabe, 2011)

9 http://www.bus.ku.ac.th Hypotheses

H3: The linkage and announcement effects of joining the ASEAN Link are not statistically significant. • Previous empirical results pertaining to cross-listing phenomenon have been ambiguous. However, Wang et al. (2008) found no listing effects using sampled companies in Thailand that cross-list on the US stock market since these cross-listed firms concentrate on certain exchanges and the return patterns of these companies affect the returns patterns of these exchanges.

10 http://www.bus.ku.ac.th Methodology: Two-stage regression system

Stage 1: cross-sectional probit regression model

11 http://www.bus.ku.ac.th Methodology: Two-stage regression system

Stage 2: VAR model

12 http://www.bus.ku.ac.th Variables

Pr(TL)/PROBTL The probability that company has a position on the ASEAN Link CTSAR The cumulative total standardized abnormal return TL A binary response variable that equals to one for the linked companies, and zero otherwise. PE price-earnings ratio ROE Return on equity OC ownership concentration AEC ASEAN Economic Cooperation OIL Oil price INT Interest rate FX Foreign exchange rate

13 http://www.bus.ku.ac.th Data

Sampled firms • The first group consists of 30 Thai companies that have linked their shares on the ASEAN Link gateway. The second group contains 30 Thai companies that have not linked their shares on the ASEAN Link gateway.

Variables • Daily return data on each sampled stock was collected during the event study period. • Daily data on financial and economic variables that constitute independent variables in our two-stage regression system were retrieved from several sources; for instance, the Computer Enterprise and Investigations Conference (CEIC), Reuters, DATASTREAM, and other official websites. Daily data on CTSAR and AEC must be computed using a particular methodology

14 http://www.bus.ku.ac.th Computation of CTSAR

Seiler’s (2004) methodology • The event period is 15 trading days before the first trading day through 15 trading days after the first trading day (-15, 15) in order to eliminate the impact that is not caused by factors influencing the overall market, and to cover unintended effects owing to leakage of news and investors’ trading delays. The first public announcement of the first trading day is also included in this event window in order to investigate the announcement effect.

15 http://www.bus.ku.ac.th Computation of CTSAR

Seiler’s (2004) methodology

16 http://www.bus.ku.ac.th Computation of AEC

Cross-sectional standard deviation of country index returns around a benchmark index (Solnik and Roulet, 2000; Adjaoute and Danthine, 2004)

17 http://www.bus.ku.ac.th Descriptive statistics of the variables in stage 1

18 http://www.bus.ku.ac.th Probit model marginal effects

19 http://www.bus.ku.ac.th Probit model marginal effects

H1 is supported by a significant negative marginal effect of AEC throughout the event period. • Thai securities that are highly integrated with foreign securities competing in the same industrial sector in the ASEAN region (or low standard deviation) are about 0.7 to 4.4 percent more likely to have a position on the ATL.

PE is not always positive and not significant. • Securities are placed on the ATL even when their expected future earnings are not high. This implies that increasing the ability to raise more capital to cope with expected growth in future earnings is one of many reasons that firms earn from linking.

20 http://www.bus.ku.ac.th Probit model marginal effects

ROE has a statistically significant effect on the linked probability for 8 days during the event period. • Thai securities with high ROE are about 0.4 to 0.6 percent more likely to be placed on the ATL. Our findings mean that higher- quality firms are more likely to link.

OC is not always positive and not significant. • The theory espoused by Abdallah and Goergen (2008) and Ayyagari and Doidge (2010); that is, firms with a high level of controlling shareholders can enjoy many benefits from cross-listing. A different case is found in the context of ATL.

21 http://www.bus.ku.ac.th Probit model marginal effects

The mean predicted probability of linking lies between 49.3 to 51.4 percent during the event window. • This implies that our selected independent variables influence the likelihood that Thai firms will have the position on the ASEAN Link by approximately 49.3 to 51.4 percent. The correctively predicted test demonstrates that our probit model is well justified in predicting the value of a binary model outcome, with the percentage of correctly predicted values being higher than 68 percent for every regression during the event window.

22 http://www.bus.ku.ac.th Descriptive statistics of the variables in stage 2

23 http://www.bus.ku.ac.th Transformation of the original VAR model

The original VAR model is simplified to the following form:

General econometrics problems • Unit root, serial correlation and heteroscedasticity may result in biased parameter estimates or biased inferences, and therefore need to be addressed. E.g., Variable OIL appears in logarithmic form to accommodate easier economic interpretation and better estimation. Table 4 shows the descriptive statistics of these variables.

24 http://www.bus.ku.ac.th Augmented Dicker Fuller unit root test

• CTSAR, PROBTL, and LNOIL are stationary at level. • INT and FX are stationary at first difference.

25 http://www.bus.ku.ac.th Pairwise Granger-causality test results

26 http://www.bus.ku.ac.th Pairwise Granger-causality test results

PROBTL Granger causes CTSAR. • PROBTL contains useful information that helps predict the future value of CTSAR.

A statistically significant feedback effect exists between CTSAR and DFX. • This implies that independent variables and dependent variable may be both endogenous and that a VAR type of the model is therefore appropriate for representing this dynamic relationship.

27 http://www.bus.ku.ac.th VAR Lag Order Selection Criteria

Akaike Information Criterian (AIC), Schwartz Information Criterion (SC), and Hanna-Quinn Information Criterion (HQ) indicate that a lag length of 1 is most suitable.

28 http://www.bus.ku.ac.th VAR coefficients

29 http://www.bus.ku.ac.th VAR coefficients

Coefficients for the lagged value of PROBTL and CTSAR are statistically significant (at the 1% and 10% levels respectively). • H2 is supported since the coefficient of PROBTL is -1.290 and significant. This means that a 1 percent increase in the linked probability decreases firms’ abnormal returns by about 1.290 percent, although it must be noted that the impact is one-period lagging.

Interest rate is a statistically important determinant of the abnormal return. • Rising interest rates represent good news for stocks linked to the ATL.

30 http://www.bus.ku.ac.th Linking and announcement effects

31 http://www.bus.ku.ac.th Linking and announcement effects TSARs and CTSARs are not statistically significant. • H3 is supported: the official participation of the ASEAN Link by Thai firms did not convey new and relevant information to public markets. • SET is an emerging market that is not highly efficient. Premature information leakage usually occurs in this type of markets. • Linked companies subjected to higher visibility, lower information asymmetries, lower degree of market segmentation and higher liquidity should earn only market expected returns, just like their cross- listed counterparts. The pattern of TSARs over the event window is somewhat consistent with market segmentation hypothesis. • Most of TSARs in the first half of the event are positive, while TSARs in the second half of the event are negative for ten days. • This implies that more open capital markets result in smaller abnormal returns.

32 http://www.bus.ku.ac.th Conclusions and implications

Probit model • As ASEAN become a more integrated capital markets in the region, the probability that Thai firms will have a position on the ASEAN Link increases. In a financially integrated environment, firms are more likely to have a position on the ASEAN Link.

VAR model and event study • Abnormal returns of securities tend to reduce after the linkage. • ASEAN capital markets have become less speculative in nature after the launch of the ATL. ASEAN Stars are major stocks from each ASEAN exchange. The reduction in ASEAN Stars’ abnormal returns means that index movements are more likely to be reflected by stocks’ fundamentals. Market speculation is now harder. This less speculative capital market environment is particularly encouraging for value investors who tend to hold stocks of companies whose value will eventually correspond to their long-term fundamentals.

33 http://www.bus.ku.ac.th Questions

34 http://www.bus.ku.ac.th Contact me

Email: [email protected] Homepage: http://fin.bus.ku.ac.th/nattawoot.htm Phone: 02-9428777 Ext. 1218 Mobile: 087- 5393525 Office: 9th floor, KBS Building 4

35 http://www.bus.ku.ac.th