The Structural Change of Stock Returns and Volatility Subsequent to the Deregulation for Foreign Investment in Taiwan Markets

Chiao Da Management Review Vol. 29 No. 1, 2009

pp.175-208

外資政策對台灣股市報酬率與波動性

影響之結構性變遷分析

The Structural Change of Stock Returns and Volatility

Subsequent to the Deregulation for Foreign Investment in

Taiwan Markets

蔡璧徽1 Bi-Huei Tsai

國立交通大學 管理科學研究所

Dep訂tmentofManagement Science, National Chiao Tung University

摘要﹒過去文獻鮮少探討開放外資投資限制之後，外資的行為對股市的增額影 響。本文採用結構性變遷的 GARCH 模型捕捉民國九十二台灣全面開放外資 前後，台灣股市報酬率與波動性的轉變，本研究主題為:一、政府開放外資 後，外資賈(責)超對台灣股票報酬率的影響是否產生結構性變化。二、比 較政府開放外資前後，台灣股市波動性的差異。三、針對外國證券投資與外 國直接投資比重高的公司，探討政府全面開放外資投資後，該公司外資賈(賣) 超對其報酬率或波動性的影響，並和其他外資投資比重低的公司加以分析比 較差異之處。 本研究中模型配適度的檢定證實結構性變遷 GARCH 模型的解釋能力 較傳統模型為佳，室主結構性變遷模型實言查結呆顯示外資買賣超影響台灣股市 報酬率，而在政府全面放寬外資管制之後，外資交易行為對股市報酬率影響 程度呈現結構性下降的趨勢;研究結果隱含台灣投資人會參的外資的投資策 略作投資，而當台灣證券市場投資環境更為開放時，外資進入台灣股市交易 頻繁，外資買賣行為對投資大眾指標意義下降，故外資交易行為對台灣股市 邊際影響力因而隨之縮小。尤其，外人詮券投資與外人直接投資比重高的股

I Corresponding author: Dep缸師lentofManagement Science, National Chiao Tung Universi旬， Hsinchu Ci旬，Taiwan. E-mail: bhtsai@faculty.nctu.edu.tw

176 The Structural Change ofStock Returns and Volatility Subsequent

的 theDeregulationfor Foreign Investment in Taiwan Markets

票君手組股價波動性隨著放寬外資限制制度而顯著縮小，本研究結果隱合開放 外資來台從事證券投資可穩定台灣證券交易市場，由於新興國家(例如中國 大陸)仍管制外國人到其資本市場從事外人證券投資，該實吉查研究可作為這 些政府開放證券市場的政策參考。 關鍵詞:動能假說;資訊效采假說; GARCH 模型;結構性變遷;買賣超; 機構投資

Abstract : Since 2003, Taiwan has relaxed the restrictions for foreign investment and permitted foreign investors to freely invest in Taiwan stock markets. The purpose behind enacting this policy was the be!ief that foreign investors could promote Taiwan investors to come up with sound investment strategies through greater awareness of the fundamental information of public companies. How巴ver，

previous studies did not meticulously investigate the incremental impact of foreign behavior on market returns nor did they look into the volatility many years after the deregulation. Hence, the purpose of this paper is to compare the effect of foreign buy-sell difference on the stock returns and the volatility prior and subsequent to Taiwan's deregulation of the foreign investment. The aims of the research are to examine whether structural change is present in both stock returns and volatility in the Taiwanese markets after the restriction relaxed. First,

this paper explores whether foreign buy-sell difference is positively related to stock returns in the Taiwan Stock Exchange. Second, this article examines whether smaller correlation exists between foreign buy-sell difference and stock returns after the lifting of foreign portfolio investment. In addition, we analyze whether volatility significantly decreases after the foreign investment restriction has relaxed. The results show that foreign buy-sell difference is positively related to stock returns in the Taiwan Stock Exchange. This suggests that foreign institutions may lead Taiwanese investors to buy or sell Taiwanese stocks, which may in turn affect the demand for Taiwanese stocks. Such relations declined after the lifting of foreign investment limit, suggesting that the inf1uence of foreign

mv臼tment strategies on investment decisions of many Taiwanese investors mitigated after foreign investment restrictions were relaxed. Furthermor巴， the long-term volatility substantially decreased for firms with larger foreign

Chiao Da Man.α'gemenl Review Vol. 29 No. 1, 2009 177

ownersh哼， which points out the stabilization of stock markets during post-deregulation periods

Keywords: Buy-Sell difference; Structural change; CARCH model; Institutional investors; Momentum hypothesis; Information effect hypothesis

1.

Introduction

Since 1990, Taiwan has gradually relaxed the restrictions for foreign investment and permitted foreign investors to 企'eely invest in Taiwan stock markets. In 1990, regulations that limited Qualified F oreign Institutional Investments (QFII) were still in place. For instance, the ownership ratio for individual firms was limited to 5% while the total shares of QFII were controlled to a maximum of 10%. It was only until 2003, did the Executive Yuan pass the

“

Regulations Goveming Se氾urities Investment by Overseas Chinese and F oreign Investors" amendment and abolished the restrictions to QFII, allowing an open market for foreign investments to enter into Taiwan. From then

,

foreign institution investments do not need to attain govemment authorization to exchange in the stock market; they could finally purchase their stocks directly 企om Taiwan's stock exchange. As a result, the transfer of foreign investment into Taiwan has progressively grown over the years. In 2006, foreign investment amounted to an accumulated value of 120,121 million US dollars. In 1994, the market value of listed companies with foreign investments amounted to merely 3,774 million US dollars, but by August 2006 it had already increased to 155,224 million US dollars Furthermore, the accumulated foreign investment in Taiwan Stock exchange grew from 1.54% of the overall market value in 1994 to 33% in August 2006. These figures evidently show that there has been a change in Taiwan 's stock market investor structure after th巴 relaxation of foreign investment policies. Thus, the behavior of foreign investment and its influence on Taiwan's stock market is an important lssue

The motivation behind establishing a 企ee capital market is firstl y, for globalization, namely, for Taiwan to join the World Trade Organization (WTO); and secondly to raise the number of qualified foreign investors investing in

178 The Structural Change 01 Stock Returns and Volatility Subsequent

的 theDeregulationfor Foreign Investment 的 TaiwanMarkets

Taiwan's stock market. Individual investors stand for the majority of investments in Taiwan's stock market. This raises governmental concerns, as the government is worried 吐latthere may be a lack in qualified investors roaming the stock market. F oreign investors are mostly composed of institutional investors, who have the professional skill and time to gather and analyze a firm's basic information. Thus,

if Taiwan is able to attract foreign investor, it may as a result stimulate and encourage its investors in implementing investment strategies.

Studies have been done on the restrictions on foreign investments and their influence on the stock market (Bailey and Jagtiani, 1994; Bailey, Chung, and

Kan耳， 1999; Hargis, 2002; Tsai, and Li, 2004). According to Boner-Neal et al.

(1990), the relaxing of foreign investment regulations may cause the premium of

cIosed-end. coun甘y fundsωdecrease. Although these researches discuss about the open markets influence on the stock market, they do not probe into the investment behavior of the national incremental effort on stock returns. F or this reason, this research uses the GARCH structural model to investigate whether the open market effectively lowers the volatility of Taiwan's stock market and whether there is any influence on Taiwan's stock returns.

According to Warther (1995) and Edelen and Wamer (2001), as the unexpected cash inflow increase for stock mutual funds, so cash inflow into stock market increases and inspire the stock retums. Furthermore, the liquidity effect hypothesis (e.g. Scholes, 1972), price pressure hypothesis (e.g. Shleifer, 1986) and information effect hypothesis (e.g. Close, 1975) inte吵鬧 the influence of the institutional investor's large scale, abnormal transaction volume on other investor's institution s甘ategy and consequentIy cause an substantial elevations in stock returns. Table 1 depicts the three major institutional investors (foreign institution, dealer and mutual fund) in Taiwan's stock market. From the table,

foreign institution holds for the largest amount of the total market value and has increased from a 19.52% in 2001 (pre-restriction relaxation period) to 33.40% in 2006 (post-restriction relaxation period). It can be seen that the influence of foreign transactions on Taiwan's stock returns and volatility can not be ignored.

Chiao Da Management Review Vol. 29 No. 1, 2009 179

Table 1

Market Value Proportion of Foreign Institution

,

Dealer and 1\直utualFund in Taiwan Stock Exchange

Foreign institution (%) Mutual fund (%) Dealer(%)

90 19.52 1.64 。 27 91 16.88 1.79 。 35 92 22.89 1.93 。 36 93 23.24 1.85 0.41 94 31.90 1.65 。.3 9 95 33.40 1.20 。 27

Source : Taiwan Economic Journal (TEJ) datab且e

In addition, according to the momentum hypothesis proposed by Lakonishok, Shleifer, and Vishny (1992), Cooper, Gutierrez, and Hameed (2004),

and George and Hwang (2004), foreign investors play the role of

“

news watcher" since they have the ability to gather and analyze information. On the other hand,

the general individual investors in Taiwan take the foreign investors investment strategies and use it as a source for their own investment direction. Therefore,

when foreign investor's buy-sell difference is positive, Taiwan investors will consider buying in those stocks, resulting in an increase in demand of the stock. This in turn increases the stock return. Similarly, if the buy-sell difference is negative, then it will result in a drop of stock retums. In other words, foreign investors are expected to lead individual investment strategies. Since Taiwan loosened the foreign investment restrictions, the leading influence of foreign investors on local individual investors is likely to change. It is in our interest to investigate whether foreign investment restriction relaxations will cause structural changes in foreign investors' leading positions after Taiwan relaxes foreign investment restrictions. Particularly, the estimation of the relations between stock retums and foreign investment may be biased when ignoring the foreign restriction factors. Thus, owing to 伽'0 reasons mentioned above, this study proposes to utilize the GARCH structural change model to investigate the difference b巳tween the stock returns before and after 血e deregulation of foreign investments. The first purpose of this paper is to investigate whether foreign

180 The Structural Change ofStock Returns and Volatility Subsequent to theDe間'gulationforForeign lnve叫mentin Taiwan Markets

investment restriction relaxations will cause structural changes in foreign investors' leading roles after Taiwan relaxes foreign investment restrictions.

In addition to causing a structural change to the stock re個口芯 foreign

investment restriction relaxation may also stimulate a change to the volatility of Taiwan's stock market. According ωDe Long et al. (1990) and Cutler, Poterba,

and Summers (1990), “news watchers" may prompt “momentum traders" to follow news watchers' lead, and buy in more stock thus causing the so-called

“momentums仕ategy". These

“

news watchers" are then able to sell their stock at a higher price in the following time period. Under the

“

momentum strategy" the

“

news watchers" may disregard the firm's performance and yet benefit from their investments, thus becoming

“

rational spectators". From the point of view of the momentum hypoth的時， once the amount of

“

rational spectator" increases the volatility will in tum increase.

However, from the perspective of informatìon asymmetry, a free' market may enable individual investors to gain access to research reports of foreign institutional investors and as a result construct their own investment s甘ategies.

EI-Gazzar (1998) and Tsai and Li (2004) report that intemational media focuses on firms with higher foreign investment ownership. 買lis means that the public is more easilyable to access the information conceming firms with higher foreign ownership. Foreign institutional investors are more likely to utilize professional

skills and d巴votetime into gathering and analyzing data. As such, once a firm has high foreign investment ownership the market will releas巳 moreof its information.

This will not only decrease information asymmetry between investors and the firm but also decrease stock volatili可'. From this perspective, restriction loosening has the positive effect of stabilizing the stock market. Whether market deregulations enhance or stabilize the market volatility has long been a critical issue in academic fields. Thus, our second research aim is to use a dummy variable to capture and see if there is a difference in Taiwan's stock volatility pre-and post market deregulation.

The stock exchange behavior depends on the foreign ownership. Foreign institutions investing in Taiwan may be one of two types: foreign portfolio investment (FPI) or foreign direct investment (FDI). FDI investors refer to the

Chiao Da Management Review Vol. 29 No. 1, 2009 181

investors who bring money into Taiwan by setting up factories, recruit local employees and engage in production and manufacturing activities. As such investors of this type will not exit or withdraw their capital from Taiwan easily. In most cases, FDI investors are often the major shareholders who take a keen interest in running the business. Based on Company Law2 and the Securities Exchange Ace, there are regulations that major shareholders have to abide when buying or selling their stock. These regulations prohibit FDI investors 企om

engaging in short-term stock trading on the share of their firm while promoting long-run stock investment. FDI investors prefer involving long-term investment in Taiwanese firms.

By contrast, the priorityof FPI investors is to obtain the benefits from Taiwan's stock. Thus, they are very sensitive to Taiwan's overall economic

indicato悶， name1y interest and exchange rates, so they prefer short-term security trading. FPI investors prefer short-term security trading. Local Taiwan investors can follow FPI investors' short-termtrading strategies more often than they follow FDI investors' long-term ones since FDI investors do not trade as

企equent1y as FPI investors do. Fluctuations are likely to be caused by this

short-t耳口ntrading strategy. Hence, it is worthwhile in examining the relationship between foreign investment behaviors with that of stock returns and volati1ity. The third purpose of this paper focuses our investigation on companies with high FDI and FPI to see whether their investment behavior has any different inf1uence 2 Article 197 of Company Act regulates

“

each director shall, after having been elected, declare to

也e competent authority the number and arnount of the shares of the company being held by M虹1Iher at the time when he/she is elected. In case a director of a company whose shares are issued to the public that has transfe叮'ed，during the tenn of office as a director, more than one half of the company's shares being held by himlber at the time he/she is elected, he/she shall, ipso facto, be discharged from the 0伍ceofdirector." According to article 197 ofCompany Act, each director reduces short-tenn trading on the share ofthe company in order to obey this article and stabilize their positions

3 Article 157 of Securities and Exchange Act regulates“In the event that any director, supervisor, managerial officer, or shareholder holding more than ten percent of the shares of a company sel1s the listed securities within six months after its acquisition, or repurchase the securities within six months after its sale, the company shall claim for the disgorgement of any profit realized from the sale and purchase." The regulations reduce the short-tenn trading for the director, supervisor, managerial offic吭 or shareholder holding more than ten percent of the shares of a company.

182 The Structural Change of Stock Returns and Volatility Subsequent to the Deregulationfor Foreign Inv目tmentin Taiwan Markets

on Taiwan's stock returns and volatility.

The purposes of this research can be stated in twofold. First，也lS paper investigates whether foreign investment restriction relaxations will cause structural changes in foreign investors' leading roles after Taiwan relaxes foreign investment restrictions. We employ the structural change GARCH model to detect the s甘uctural declining relations between Taiwan's stock return and foreign trading, after the deregulation of foreign investment in Taiwan. Second, this article uses a dummy variable to capture and see if there is a difference in Taiwan's stock volatili句 pre- and post market deregulation. From our research,

the results of the goodness of fit demonstrated that the structural change GARCH model exhibited superior analysis compared to that of the conventional GARCH model. We also find that a market deregulation will reduce the significance of foreign investors as trading indicators for local investors，的 local investors 訂E

less inclined to stake their investment according to foreign trading practices. In addition, companies with high foreign shareholding ratios show a decline in stock volatility. This suggests that companies with high proportions of FPI and FDI have reduced the information asymmetry between outsider investors and firms particularly after foreign investment res仕ictions relaxations. Henc巴， the stock

volatility 吐rops substantially during the post-deregulation periods. Market deregulations are proven to be favorable to stabilize the firm's stock volatility.

2.

Research Hypothesis

2.1 Relationship Between Foreign Investor Trading Behavior and Stock Returns

According to Scholes (1972) liquidity effect hypothesis, as opposed to the general individual investors' huge dealings, institutional investors deal in massive amounts that will inf1uence the markets stock return value. In addition, according to Shleifer (1986) price pressure hypothesis, the law ofsupply and demand st泌的

that large- volume trading of institutional investors will result in price change; in other words institutional investments has a fair impact on the mark仗's stock

ChiaoDaM扭扭'gementReview Vol. 29 No. 1, 2009 183

reωm. If the buy-sell difference of institutional investors is positive

,

individual investors will t耳keinto account its investment s仕的egyand buy stock; the demand in stock will increase and so will the retum on equity. On the oth巳r， if institutional investors' buy-sell difference is negative, the demand in stock will decrease, and so will the retum on equity.

Furthermor巴， Arbel, Carvell, and Strebel (1983) and Brous and Kini (1994) also pointed out that institutional investors have a rich source of information; they have professionals that collect and analyze data to prepare investment strategies. Hence, according to scholars in support of information 巴ffect hypothesis,

institutional investors are more objective and rational than average investors.

Therefore, when market investors notice abnormal or unusually large 仕ansactions

of institutional investors, this usually implies important investments and causes market investors to reevaluate their investment strategy; stock prices (stock retums) will thus be affected.

Scholars of momentum hypothesis (such as Lakonishok, Shleifer, and Vishny, 1992; Chan and Lakonishok, 1993; Daniel, Hirshleifer, and Subrahmanyam, 1998; Hong and Stein, 1999) pointed out that there are two types of investors in the market:

“

news watchers" and “momentum 仕aders". “News

watchers" act and trade in accordance to the firm's fundamental information,

while “momen仙m 訂aders"， who neither gather nor analyze data conceming the firm's operati凹， may only follow investment strategies set by “n巳ws watchers",

i.e. institutional investors. '‘Momentum traders" are said to gain profit in their

invcstments through the momentum strategy (Jegadeesh and Titman

,

1993

Rouwenhor鈍， 1998 ; Jegadeesh and Titman, 2001 ; Lui, Strong, and Xu, 1999) Following from the momentum hypothes詣， institutional investors possess the ability to collect and analyze data, playing the role of news watchers; individual investors then observe their investment strategies and use their (news watchers') investment policies to determine the dir巴ction of investment. Hence

,

the buy-sell difference of foreign institutional investors is likely to positively related to the market stock retums. Based on the above theoretical basis, we put forth our first hypothesis:

184 TheSl伽cturalChange ofStock Returns and Vòlatility Subsequent

的 theDeregulation for Foreign Investment 的 TaiwanMarkets

Hl: There is a positive relationship be，削'een the buy-sell dijJerence of foreign institutional investors and the market告 stockreturns.

2.2 The Effect of Foreign Trading Behaviors on Stock Returns After Market Deregulation

As of October 2 in 2003, the Executive Yuan passed the amendments of

“

Regulations Governing Securities Investment by Overseas Chinese and F oreign Investors". This enabled foreign institutional investors to trade in Taiwanese stock market without the government's approval. In 2006, the number of registered foreign investment in Taiwan's stock market raised up to 2,294 cases. By the end

ofM訂ch in 2007, the total value of foreign investment in Taiw徊's stock market accounted for 31 % of the total market value. It is clear that the proportion of individual investors, who mostly operate short-term security trading, has reduced; while the proportion of institutional investors, who mostly operate long-term security trading, has increased. There is no doubt that the investor's structure has changed in Taiw妞 's stock market as result of the governmental deregulation policy. In the past, scholars (such as, Bailey and Jagtiani, 1994; Bailey, Chung,

and Kang, 1999; Hargis, 2002; Ts缸， and Li, 2004) have studied the impact of deregulation on foreign investment in the stock market. Boner-Neal et al. (1990) found that during the initiating period when foreign trade controls were said to be relaxed, c1osed-country fund premiums shrank.

During the implement耳tion period of qualified foreign institutional investment before October 2, 2003, Taiwan only approved quali日ed forei伊

institutional investors (QFIIs)ωengage in the investment in Taiwan stock markets under speci日c regulations. There are s仕ict controls that monitor foreign investment accounts and capital remittance. F or instance, all capital gains and dividends of foreign investors had to be audited before they could be .converted

企om new Taiwan dollars to foreign cu叮ency. As a result, foreign investment institutes investing in Taiwan will make the best use of their professional investment analysis. Thus, the foreign investors' trading behaviors will provide as a trading strategy reference point for other investors. The impact of foreign institutional trading on local individual investor (“momentum traders") strategy is

Chiao Da Management Review Vol. 29 No. 1, 2009 185

obvious. However, after the government abolished restrictions on foreign investment, large volumes of foreign capital f10wed into Taiwan. The possible outcome is that the marginal inf1uence of the increas巴吐 foreign institutional trading activities on local individual investors is likely to decrease. In addition,

Taiwan 's listed companies may set up subsidiary companies overseas and remit their funds back to the local parent companies, thus causing

“

fake foreign investment". This generates a false impression of foreign investment. As such,

general individual investors no longer take foreign investments to act as significant indicators

,

as they did prior to market deregulation. The impact of foreign institutional 訂ading on Taiwan stock returns is reduced; the relations between the foreign institutional investor behaviors and the market's stock returns decline. From the above, the following hypothesis is established

H2: After the deregulation of foreign investment, the relationship between the

buy-sell difference of foreign institutional investors and the market甘 stock returns decline.

Foreign investment in Taiwan enterprises may be categorized into two categories: FPI and FDI. The main aim of FPI is to profit from the investment in Taiwan's capital market. Once investors (usually institutional investors) locate a potential investment opportunity, there will be an inf1ux of funds. However, once their expected profits are realized, these foreign investors will instantly withdraw their capital out of the Taiwan stock exchange. Such FPIs are sai位 to have rapid

f10w and depend on the current local economic status. F or instance, interest rates,

investment environment, policies and regional stability are factors that may cause FPI to decide to retain or take away their capital in Taiwan stock markets. Thus,

FPI capital f10ws are often referred to as

‘

Hot Money'\Enterprises with a high proportion of FPI are profit-oriented, their stock prices are very sensitive to macroeconomic factors and their investment strategy is a good investment indicator for Taiwan 's individual investors. In other wor侮， the trading behaviors of FPIs (“news watchers") can be regaτded as significant trading indicators to local Taiwan individual investors (“momentum traders"). However, after Taiwan opened its stock markets, an increasing number of FPIs have f10wed into Taiwan

186 TheSl的cturalChange ofStock Returns and Volatility Subseq間1t

to the Deregulationfor Foreign Investment in Taiwan Markets

and trades are now done much more frequently. This decreased the information

asymme甘y be制reen firms and outside investors, especially for firms with high FPI holdings. The FPI's leading position is no longer as app缸ent in post-deregulation period as it is in pre-der巴gulation period. The marginal impact of FPI trading on Taiwan stock retums has substantially been reduced; the relationship between FPI behaviors and the market's stock retums has dramatically de心lined. F ocusing on firms with high FPI, our research devises the third hypothesis:

H3: After the deregulation of fo阿拉n investment

,

the impact of foreign

institutional buy-sell difference on stock returns is reduced for firms with

high FPI holdings.

Foreign direct investors (FDI) refer to foreign firms that invest in Taiwan's fixed assets (e.g. factories, land，仕ansportation devises) and set up factories in

Taiwan. Because most FDI investors 訂e large sharehol卸的 or hold manger positions, they are able to obtain private insider information easily. Therefor巴， the trading behaviors of FDIs (“news watchers") can be regarded as significant trading indicators to local Taiwan individual investors (“momentum traders"). Deregulation of foreign investment reduced information asymmetry between the firm and its investors be氾ause more FDI investors provided information. The marginal inf1uence of FDI trading activities on the firm's retum rate is thus reduced. Hence

,

we formulated our fourth hypothesis:

H4: After the deregulation of foreign investment, the impact of foreign

institutional buy-sell difference on stock returns is reduced for firms with

high FDI holdings.

2.3 The Effect of Foreign

Tr

ading Behaviors on Stock Volatility After Market Deregulation

According to Froot et al. (2001) research findings, foreign investors have the tendency to implement th巴 momentum strategy. When there is a large amount of foreign funds f10wing 尬， it is predictable that, in the near future, the local stock

Chiao Da Man，α'gement Reνiew Vol. 29 No. 1, 2009 187

market will yield positive returns. This type of scenario may be seen p訂ticularly

m emergl時 markets. Research findings by Harris and Pisedtasalasai (2006) reveal that foreign investors in emerging Asian markets prefer to implement the momentum strategy, simply because foreign investors have obvious short-terrn impacts on local stock markets. On the other hand, according to EI-Gazzar (1998)

的 well as Tsai and Li (2004), intemational media is more likely to report messages conceming companies with a high proportion of foreign institutional investors. The inforrnation release decreases inforrnation asymmetry between firrns and outside investors. After the govemment abolished restrictions on

foreign 仕ade， foreign investors allowed foreign institutions to more smoothly engage in FDIs and FPIs in Taiwan. The increasing number of foreign institutional trading in Taiwan enables general individual investors to gain access to firrn's inforrnation more readily because institutional investor will seek professional skills to gather or analyze data related to the firrn and release those messages. This reduces inforrnation asymmetry between the enterprise an社 its investo悶， thus creating a more stable market with low stock volatility. This leads to our fifth hypothesis:

H5: After the deregulation of foreign investment, the stock volatility of Taiwan stock exchange will decrease.

Once Taiwan relaxes foreign investment restrictions, the market will release fair amounts of inforrnation regarding enterprises with high proportions of FPI.

This will not only reduce inforrnation asymmetry between the firrn and its investors, but also prevent trading behaviors such as

“

herding" in the market and thus decrease stock volatility. Focusing on firrns with high proportions ofF肘， we forrnulated the sixth hypothesis:

H6: After the deregulation of foreign investment

,

the stock volatility of firms

with high proportions of FPI will decrease.

F oreign direct investors (FDI) referωforeign firrns that invest in Taiwan 's fixed assets (e.g. factories, land, transportation devises) and set up factories in Taiwan. FDI investors are less sensitiv巴 to interest rates and exchange rates. As

188 The Si師lcturalChange ofStock Returns and Volatility Subsequent to the Deregulationfor Fo叫fgnInvestment in Taiwan Markets such, they will not bring in capitals to Taiwan or take away their funds out of Taiwan merely because of short-term changes in interest rates or temporary environmental changes. In addition, most FDI investors are often the major shareholders who take a keen interest in running the business. Under Company

L在w and the Securities Exchange Act, the regulations prohibit FDI investors from engaging in short司term stock trading on the shares of their firm. FDI investors prefer long-term investment in Taiwanese firms, as a resuIt the increasing FDI proportion in Taiwan markets decrease stock volatility. Additionally, the deregulation of foreign investment increases the FDI proportion in Taiwan, so the stock volatility is further reduced. This forms our next hypothesis:

H7: After the deregulation of foreign investment

,

the stock volatility of firms with highproportions of FDI wil/ decre的ι

3. Data and Sample

We focus on listed firms in Taiwanese Stock Exchange (TSE) to explore th巴

impact of foreign investor buy-sell difference on the stock retums and volatility. Since the daily institutional investor buy-sell difference is composed of the buy-sell diffi巴rence of foreign institutional investors, dealers and mutual funds,

this study simuItaneously takes into account the impact of the three types of institutional investors' behaviors. This paper selects the rate of retums and the variations of stock retums from Taiwan Stock Exchange (TSE) weighted indices. Because nearIy 70% of foreign investment capitals invest in Taiwanese electronics firms, the Taiwanese Electronics Industry weighted indices is also investigated in this study. In this articIe, Taiwan Stock Exchange (TSE) weighted indices and Taiwanese Electronics Industry weighted indices 訂e collected from the Taiwan Economic Journal (TEJ) database. In addition

,

the TEJ database was also the source for the institutional investor data. Such institutional investor data incIudes buy-sell difference 企om foreign institutions, mutual fund and dealers Taiwan Stock Exchange (TSE) weighted indices range from August 1弋 1995 to March 30'\ 2007, while Taiwanese Electronics Industry weighted indices range

Chiao Da Management Review Vol. 29 No. 1. 2009 189

企om December 12曲， 2000 to March 30th, 2007 since the Taiwanese Electronics Industry weighted indices are only available from the beginning ofDecember 1

2'

h in 2000.

In addition, this study further divides foreign investment into FPI and FDI to explore the impact of their behaviors on stock markets. Three groups of stocks are chosen according to 也e foreign investment amount in this research: FPI, FDI,

and control sample group. The three groups are chosen only 企om Taiwanese electronics firms because the majority of foreign investors invest in ele咄咄cs firms. Besides, the electronics firms with capital smaller than 300 mil1ion are deleted because the stock prices of these small firms are easily manipulated by investors. Next, the firms are classified into the three groups in keeping with the following criteria: The firms whose foreign market value ratio is over 1 % are selected. Foreign market value ratio refers to the foreign ownership market value ofunderlying securities for the individual firm deflated by total foreign ownership market values across all firms in the Taiwan Stock Exchange. Among these firms whose foreign ownership market value ratio is over 1 %, firms are classified as the

“

FPI group" only if its average QFII ownership 企om 1995 to 2001 accounts for 50% of the foreign ownership. On the other hand, firms are only classified as the

“

FDI group" 4 if its average QFII ownership during the period does not account

for 50% of the foreign ownership. Finally

,

the firms whose foreign market value ratio is under 1 % are classified as

“

control sample group".

4 With regard to FPl (foreign portfolio investment), QFIIs can trade through the Taiwan Stock

Exchange (TSE) 組d thus acquire direct ownership of Taiwanese finns. With regard to FDI (foreign direct investrnent), forei伊 institutionsgain approval from the Taiwan government to bring in capital and the technological know-how to construct factories, employ sta缸;and operate production in Taiwan. The FDI ownership is not the QFII ownership through Taiwan stock exchange (TSE). For example, Philip corp. is the main FDI investors and its FDI ownership accounts for 25.49% of the total ownership for Taiwan Semiconductor Manufacturing C。中­ (TSMC) before 1995. After TSMC is listed in TSE in 1995, the average QFII ownership only accounts for 9.82%, which is smaller than the foreign ownership through FDIs. Therefo間， TSMC belongs to the “FDI group" in this research.

190 The Structural Change of Stock Re的rnsand Volatility Subsequent

的 theDeregulationfor Foreign Investment 的 TaiwanMarkets

In this paper, the data conceming stock retums and buy-sell difference from foreign investors, mutual fund and dealers of these three groups are applied since all firms in these three groups are listed in the sample. Thus, the three groups range 企omJuly23曲， 2001 to March 30'\ 2007.

4.

Methodology

4.1

GARCH Model

Initially, we used the Froot and Rogoff (1995) Augmented Dickey-Fuller (ADF) tests 5 to check whether the two series of the rate of retums from

Taiwanese stock exchange indices and Taiwanese Electronics Indus甘y weighted indices, were unit roots, 1(1). Next, the GARCH model was constructed. The conventional time-series model assumed the constant conditional variance, which implied that risk did not vary with time. However, Engle (1982) proposed the Autoregressive Conditional Heteroscedasticity (ARCH) model to quantify the risk and uncertainty. Bollerselev, Chou, and Kroner (1 992) 如此her illustrated that the GARCH (Generalized Autoregressive Conditional Heteroscedasticity) model could reliably explain the financial time-series and upgrade the model's

叫lanatorypower. For this reason, this 的自 chooses GARCH (1,1) model to examine the relations between foreign institutional buy-sell difference and stock retums as equation (1) :

RtJtα。 +αIFi，H + α 2B_{i，H + α3IZJ I+EZJ'}

吭，1 ~ N(O, hi,t),

hυ=ω。 +ωIEf卜1+ω九一1

,

i=TSE

,

TSEE

,

、'，

J

1

/t

、、

5 Augrnented Dickey-Fuller unit root tests:

One unit root--ð. X, = ζ+ζIX，-l+ 2:;=1;1 月 X，-i+&!.l

Twounitro叫s--ð,

2x

, =;2 +ζ2axz l+EiJJua2XJGJ

xd臼enot岫e自s t伽he祖era叫te自s ofs恥t岫oc仗k叮r間叫e前t仇阻閒u1汀盯1m口ms祖s.w，恥ee間xamine閑et伽he祖叫e

Chiao Da Management Review Vol. 29 No. J, 2009 191

where RrsE,t and RrsEE,1 are Taiwan Stock Exchange index (i=TSE) stock

returns an吐 Taiwan Stock Exchange electronics industry index (i=TSEE) stock returns, respectively, on the tth day. F7呵，1-1 and FTSEE.I_I are foreign institutional buy-sell difference detlated by market values for Taiwan Stock Exchange listed firms and electronics firms listed in Taiwan Stock Exchange, respectively, on the

t-lth day. BTSE,I_I and B扭扭，1-1 are dealer buy-sell difference detlated by market

values for Taiwan Stock Exchange listed firms and electronics firms listed in Taiwan Stock Exchange, respectively, on the t-lth day. l_{TSE •,_,} and IrsEE.I-' are mutual fund buy-sell difference detlated by market values for Taiwan Stock Exchange listed firms and electronics firms listed in Taiwan Stock Exchange,

respectively, on the t-lth day

Institutional investors are composed of foreign institutional investors,

mutual fund and dealers, so we take account of the trading behaviors 企om

institutional investors, mutual fund and dealers in equation (1). If the buying orders are more (less) than the selling orders, the phenomenon is defined as possessing a positive (negative) buy-sell difference, namely, a buy-sell surplus (deficit). Because the buy-sell difference of foreign institutions is 吐isclosed

everyday, individual investors can follow their trading strategies on a daily bases. Hence, equation (1) regresses the stock return rates on these three types of institutional trading behaviors at the previous day to investigate the impact of institutional trading on returns and volatility in Taiwan Stock Exchange. In equations (1), coefficients 吭， α2 ， α'3 denote the previous day's marginal

effi巳ct of the buy-sell difference on stock returns for the foreign institution, the dealer, and the mutual fund, respectively. If the coe位icientsα" 凹， α3 缸G

positive, it implies that the foreign institution芯，dealer's and mutual fund's trading behavior for the day before enhanced stock returns. This paper employs the Taiwan stock exchange listed firms and electronics firms listed in Taiwan Stock Exchange to run regression (1).τ~statistiωis employed to examine whether the

coefficients 吭，門， α'3 are significantly positive, so we can compare the impact of these thr的 kinds of institutional investors on stock returns. Particular勻，

this artic\e can confirm the hypothesis H_IO: 科=0 to see whether foreign institutional investor's strategy is indeed critical to Tai

192 TheSJ的cturalChange 01 Stock Returns and Volatil砂 Subsequent

10 the Deregulationfor Foreign Investment in Taiwan Markets

4.2 Structural Change GARCH Model 4.2.1 Structural Change Analysis of Stock Returns

Studies have been done on the influence of institutional investors' capital flows into stock market on the market returns (Warther, 1995; Froot, O'Connell,

and Seasholes, 2001; Edlelen and Warner, 2001). However, they do not probe into the market deregulation, which leads to the relation changes between market retums and foreign institutional investors' capital flows. If the market deregulation is ignored in the model, the estimation of coefficients will be biased. For this reason, this research uses the structural change GARCH model to investigate whether the open market eff巴ctively lowers stock volatility of the Taiwan's stock market and whether there is any change to Taiwan's stock retums.

To incorporate the structural change caused by the market deregulation, we use dummy variables to divide our time line into two periods, namely, pre-deregulation and post-deregulation periods. The dummy can capture the difference in stock retums and volatility between pre-deregulation and post-deregulation periods. We modified equation (1) as equation (2)

1\,t = ßo + β武t-1

+

ß2(F;,t-1 xDi

/)+

ß3

B

i,t-1

+

ß4

I

i,t-1 + 丸 (2)

令，t ~ N(O, _{hu )},

hi,t = YO +

Yléi~t-l

+ Y2hi,t-1 +

Y3(é~t-l

xDi,t)

+ 凡的，t~l

X Di.t ) ,

i=TSE. TSEE

where D;,t = 1 if the data is during post-deregulation period (period before Oct. 1,

2003); Di,t= 0 ifthe data is during pre-deregulation period (period a缸erOct. 2, 2003).

In equation (2), the slope of the foreign institutional buy-sell difference is

ßl

in pre-deregulation periods, but becomes (βt+β'，) in post-deregulation periods. This paper uses the t-statistic to examine the hypothesis H₂₀ ß2 = 0 to investigate the incremental effect of the deregulation on the relations between buy-sell difference and market stock returns. If the coefficient ß2 is negative, it suggests that the foreign institutional influence on stock returns dec1ines as the market opens for

Chiao Da Management Review Vol. 29 No. 1, 2009 193

foreigners. In addition, to investigate the foreign institutional effect on stock returns,

this paper uses the t-statistic to examine whether foreigninstitutional investor affect Taiwan stock retums before market deregulation, namely, Ho βI = 0 . Also, Wald statistics is used to examine whether foreign institutional investor affect Taiwan stock returns after ma如tderegulatio

n,

namely, Ho 為 +β2=0

4.2.2 Structural Change Analysis of Volatility

To incorporate the volatility structural change caused by the market deregulation, we use dummy variables to divide the timeline into two periods,

pre-deregulation and post-deregulation periods in the variance equation of structural change GARCH model. Hence, the dummy variable can c通pture the difference in volatility between pre-deregulation and post-deregulation periods. In the variance equation, if the coefficients y

,

and 乃 are negative, it suggests that the short-term and long-term volatility declined, respectively. This paper use the t-statistic to examine the hypothesis H₅₀ y

,

= 0 and H₅₀

_几

=0 to investigate the incremental effect of market deregulations on the volatility.

4.2.3 Goodness of Fit Test

This paper employs a model's likelihood ratios. It compares the goodness-of-fit of the structural change GARCH model (model 2) 叫出自己

conventional GARCH model (model 1). The likelihood ratio is stated as equation

(3):

D=-2log(?)=-2(logA-log 丸)， (3)

where Lo is the likelihood function values of GARCH model (model 1); Lm is the likelihood function values of structural change GARCH model (model 2). The likelihood ratio D can be converted into %2 dis仕ibution， so this article can employ %2 statistics in which the degree of 命eedomis the parameter number. If the %2 statistics test results significantly reject the null hypothesis, it suggests that the structural change GARCH model (model 2) contain incremental explanatory power and that it performs better than the conventional GARCH model (model 1).

194 The Structural Change of Stock Returns and Volatility Subsequent to the Deregulationfor Foreign Inv，目的冊11in Taiwan Markets

4.3 Analyzing the difference in stock returns and volatility of firm with differential foreign investment characteristics

After the deregulation of foreign trade, the proportions of foreign institutional investors have increased in the Taiwan stock market. The market will also release more information regarding companies with high proportions of foreign investment. However, it is found that although more information has been released, inforτnation is also merely targeted to firms with higher FD1 and FP1 holding. 1nformation 企om mutual funds and dealers remain scarce. 1t seems that Taiwan's open market has singularly benefited the firms with high proportions of foreign investment.

For the aforementioned reasons, we extend our investigation by dividing the sample companies into three categories: firms with high FP1 proportions (FP1 group)

,

firms with high FDI proportions (FDI group) and firms with little foreign investment (control sample group). 百le weighted average rate of stock retum is calculated for each group and this article runs the equation (2) with the three groups (i= FP1 group, FDI group, and control sample group) to compare the stock retum changes between pre-deregulation and post-deregulation periods to examine the third and fourth hypothesis. 1n addition, we further test whether the coefficients Y3 and Y4 in the variance equation are negative for the three groups

firms to examine the sixth and seventh hypothesis

5.

Empirical Results

5.1

Results of GARCH Regression Model

From our sample, the average buy-sell difference for foreign investment is positive. This means that within the time of our investigation, there was an average surplus in foreign investment in Taiwan stock market. Particularly, after the deregulation policy, foreign investment has continued to supply funds to Taiwan markets. However

,

it was found that dealers and mutual funds fell short of the supply funds to Taiwan markets. Daily average buy-sell difference is negative for dealers and mutual funds.

Chiao Da Management Review Vol. 29 No. 1, 2009 195

In addition, before the deregulation poIi c其 the average daily buy-seII difference of foreign investment was a surplus of NT 602 miIIion doIIars Meanwhile, once the govemment opened up the marl主仗， the average daiIy

buy-seII difference of foreign institutions increased to NT 1,893 miIIion dolIars

These figures indicate that the deregulation policy had a positive effect of encouraging foreign investment. Similar勻" the average daily buy-selI difference of dealers went from a negative value of NT 14 miIIion doIIars before market deregulation to a positive buy-selI difference ofNT19 miIIion doIIars after market deregulation. This shows the positive capital inf10w e宜ect of the open market However

,

the average daily buy-seII differences of mutual funds show a different story. It ranges 丘。m a negative buy-seII difference of NT 75 miIlion doIIars to a negative buy-seII difference ofNT 180 miIlion doIIars.

Observing the stock market from the time-series point of view, it was noticed that no matter in which point in time, after the deregulation policy, the standard deviation of foreign institutions were significantIy greater than those of dealers and mutual funds. This is possibly caused by the “hot money" characteristic of foreign investments, whereby intemational funds wilI flow

rapidly 丘。mone market to the next. Thus, the opening ofthe market only sped up the cuπency f10w and increased the standard deviation of foreign investments.

The Augmented Dickey FuIIer (ADF) Test results showed that the variables were stable; we then ran GARCH regression analysis. The regression results of the Taiwan Stock Exchange index retums and Taiwan Stock Exchange electronic

indus仕y index returns on the three institutional investors' trading behavior are listed in Table 2. From Table 2, it is seen that the coe宜icients of foreign institutional buy-seII difference and those of mutual funds 位巴 2.7983 (α1) and 4.7209 (α3 ) ,respectively. The T-statistic shows that the coefficient of buy-selI

difference (α1 ) is significantIy positive at 1% level for foreign institutional investors and 趴ecoefficient of buy-selI difference (α3 ) is significantly positive at 5% level for mutual funds. This means that the buy-selI difference of foreign institutions and mutual funds have a positive effect on stock retums. 甘le findings suggest that if there is a surplus in the buy-selI difference of foreign institutions and mutual fun心， general individual inves

196 The Struclural Change ofSlock Returns and Volatility Subsequenl 10 Ihe Deregulation for Fo月ignInvestment in Taiwan Markets

S訂-ategies to buy more stocks, thus increasing the demand and raising stock retums. Inversely, if foreign institutional buy-sell di在erence was to have a negative value, general individual inv官stors are inclined to seIl their stocks,

resulting in an inc凹的巴 in supply and a dωrease in stock returns. 甘le above analysis results support hypothesis 1 (Hl).

Table 2

Results of GARCH Regression

R， .= αA+α， F. ， + α吋 B" ， + α勻!" ， +ε i,t ~U' ~~l~ i,l-l ' ...2~i ，t-l . ...3~ i,(-1 ' ...i,l'

εZJ~N(0， hiJ) ，

hir= 的+叫Efr l+ 叫hU_I ﹒

、'，

J

l

f1

、、

Taiwan Stock Exchange index Taiwan Stock Exchange

Dependant variable retums electronics indust立 indexreturns

Coefficient T-statistics Coefficient τ:'statistics

F

i.t_l 2.7983 5.3720'" 1.7838 3.709

7'''

B

i.t-l 2.9727 0.9330 0.2178 。 0547 IU_I 4.7209 2.2163" 7.5149 2.7997'" Interpret (αo ) 0.0384 1.6470 。.0368 0.9526 Interp目t( lÐ o) 0.0253 4.2447'" 0.0206 2.5466"

片{2

,t-l 0.0746 11.2233'" 0.0531 6.8103'" hU_l 0.9157 126.7220'" 0.9401 111.1559'"

N。但s:*Significant at the 10 percent level.料 Significant at 也e5 percent level. ***Significant at 血e1 perc叩tlevel.

From the GARCH regression results, the coe缸icient of mutual funds' buy-sell difference (α3 = 4.7209) is greater than those offoreign institutions' (a

,

=

2.7983). These findings imply that mutual funds have more a significant influence on Taiwan's Stock Exchange Index stock retums th組 foreign investors. Furthermore, dealers are perceived to be insignificantly positive with Taiwan Stock Exchange Index stock retums.6 The reason for this may be owing to the

Chiao Da Management Revi，叫，Vol. 29 No. 1, 2009 197

fact that from 2001 to 2006 the proportion of foreign institutional investors were between 16.88% and 33.40%, but no more than 0.5% for dealers. As such, the trading behaviors of dealers would merely hold limited influence on stock retums,

so 也ecoefficient of buy-sell difference for dealers is insigni自cant.

From the regression results of Taiwan Stock Exchange elec仕onic industry index stock re個rns on the three institutional investors' buy-sell difference as depicted in Table 2, the coefficient of foreign investors, dealers and mutual funds are 1.7838, 0.2178 and 7.5149 respectively.τbeτ~statistic shows that the coefficient of foreign investments buy-sell difference (α1 ) and that of mutual funds (α3 ) are significantly positive at 1 % level. This means that the buy-sell difference of foreign investments and mutual funds have a positive relationship with stock returns of the elec甘onicsindustry. This in turn implies that if there is a buy-sell surplus of foreign investments and mutual funds in Taiwan Stock Exchange elec仕onics industry stocks, then general individual investors will follow their investment strategy and buy electronics stocks

,

thus increasing the demand of stock in the electronics industry and raising its stock returns. Conversely, if foreign inves個lents' buy-sell difference was to be negative in the electronics indus仕y， general individual investors would sell their electronics stocks, resulting in an increase in electronics industry stock supply and a decrease in stock returns of electronics stocks. From the above analysis, hypothesis 1 (Hl) is supported.

In addition, for both Taiwan Stock Exchange Index and Elec仕onic Indus仕y

Index，也ecoefficient of mutual fund instiωtional investors' buy-sell difference is always greater (α3 ) than those of foreign investors (α1 )的 well. The reason behind this could b巳 because mutual fund investors and their related-party members periodically release their analysis of Taiwan stock trends to the media,

allowing their investment strategies to become one of th巳 leading indicators to Taiwan's general investors. The trading behavior of mutual fund investors is hence directly related to stock trends. Furthermore

,

dealers are observed to have a

the standard deviation of the coefficientα2 is also much greater 由an that of the coefficient

198 The Structural Change ofStock Returns and Volatility Subsequent

的 theDe阿~gulationforForeign Investment in Taiwan MG1如ts

much lower average buy-sell difference in our sample. The extent of dealers' influence on Taiwan Stock Exchange Index and Taiwan Stock Exchange Electronics Industry Index is substantially minor compared to the other two major institutional investors. Thus, we concluded the unobvious relations between

dealer法甘adingbehavior and stock returns.

5.2

Results of the Structural Change GARCH Model 5.2.1 Structural Change Analysis of Stock Returns

This research investigated the structural change in stock returns and volatility for five major groups when the government enacted the deregulation policy in 2003. The five major groups include firms in Taiwan Stock Exchange,

electronics firms in Taiwan Stock Exchange, firms with high FPI proportions,

firms with high FDI proportions, and firms with little foreign investment (control sample firms). The structural change GARCH regression results are displayed in Table 3. The results ofthe Taiwan Stock Exchange index stock returns and Taiwan

Elec仕onicIndustry index stock returns, from Table 3, show the coefficients ofthe interaction terms (F,.t-1 x Dt) between foreign investments' buy-sell difference and the dummy variable are negatively related to stock returns. After the markets opened, the influence of foreign investment buy-sell difference on stock returns diminished. This supports our fourth hypothesis (H4): after the deregulation of foreign investment, the impact of foreign trading activities on stock returns is reduced in firms high in FD I.

Table 3 depicts the results of the Wald statistics. After the policy deregulation, the coe在icient of foreign investors' buy-sell di臨時nce on the Taiwan Stock Exchange Index and Electronic Industry were 1.9561 7 and 1.51258

respectivel弘 Wald Statistic exhibits the two coefficients significantly positive at 5% level. From our research findings, we observed that despite the fact that market deregulation may reduce the impact of foreign investments on stock returns; however, there remains significant impact of foreign institutional buy-sell

7β\+β'2 =6.5406-4.5845=1.9561

ChiaoDaManα'gement Review Vol. 29 No. 1, 2009 199

difference on the Taiwan's stock returns. Ifthere is a buy-sell surplus, local

Table 3

Results of Structural Change GARCH Regression

l\,

t =ßo+ βlffr1+ 屁(F;，1-1 X Dj

,

t) + ß3B，小1 + ß4I j

,

t-l + 鈣，μ

&i

,

t ~ N(O

,

hj

,,),

(2)

hi

,

t

- 几十九&j:1-1

+Y2 hj

,

t-1

+Y3(ε立1 × DLr)+Y4(hzt1 × DEJ).

Dependant variable

Taiwan Stock Exchange index retums

Taiwan Stock Exchange electronics industry index Coefficient T-statistics

Fj

,t-l (FU _ 1 X D,)

B

i.(-l

Iht一 l

Int呻阻，t (β。)

Interpret (

r

0 ) 2 &j~，-l

hi

,t-l

(

〈品←J

呵l戶xD

叭丸

j

，t)

)

(彷h小1 xDυ Goodnes臼s-ofι~f缸itresults Likelihood ratio Walds泣ta剖訓t封18泣組t缸lC臼s Ho: β1+β2

=

0 2 Chi-square (正) 6.5406 -4.5845 2.6983 3.6006 0.0341 0.0382 0.0802 0.9077 -0.0172 -0.0032 5.1481**' -3.2102'" 0.8576 1.6233 1.4519 4.6672'" 8.6327'" 95.5978'** -1.3251 -0.2923 23.54'" 1.9561 7.4357'" retums Coefficient T-statistics 3.4166 -1.9041 0.6512 6.4790 0.0310 0.0393 0.0479 0.9431 0.0038 -0.0161 2.3642** -1.2336 0.1633 2.3441** 0.7905 3.1719'" 3.3718'"

64717。"*

0.2413 -1.2925 12.44'" 1.5125 6.4737"

Notes: *Significant at the 10 pe阻entlevel 紳副研 ificant 前也e5 pe四 ent level 料 *Significa刮目

the 1 percent le、 el.

200 TheS，的'cturalChange of Stock Returns and Volatility Subsequent

10 the Deregulationfor Foreign Investment in Taiwan Markets

continue to soar. In short, foreign investment strategies remain a significant indicator in Taiwan markets even after market deregulations.

The research results 企om Taiwan Stock Exchange Index and Electronic

Indus叮 Indeximply that before the markets opened, during the regulation period,

the amount of foreign investments were scarce. As a result, foreign institutional (news watchers') trading behaviors b的ame significantly meaningful to Taiwan's local investors (momenωm traders). However, after the deregulation went through,

not only did the amount of foreign investments dramatically increased

,

but momentum traders were able to access information through other channels. In addition, in order to make a company seem successful, some firms may 甘Y to create a false image of high foreign investment by manipulating buying their stocks from their own branch offices abroad. As such, trading behaviors of foreign institutions (news watchers) are no longer critical indicator in the eyes of Taiwan's local investors (momentum traders). 明白， inadvertently, decreases the influence of foreign investors' buy-sell difference on stock returns.

5.2.2 Structural Change Analysis of Stock Volatility

From Table 3, after the deregulation, the Taiwan Stock Exchange volatility regression results show that variables Y3 and Y. are no longer dominant. This

indicates that

,

in co甜甜 to our hypothesis 5 但5)， after the government allowed open markets there were no drastic changes to the stock return volatility. The results are different from Ou (1994) findings 出atrestriction relaxations of foreign investment decrease the stock volatility. The possible reason behind this could be because the market releases relatively more information merely concerning firms with high ratios of institutional investors; as such information asymmetηr IS

deceased. However, although the deregulation policy enabled more foreign investors to invest in Taiwan stock

,

increase information flow and decrease information asymmetry, these information are generaliy confined to firms with higher foreign investment. However, since a minority of Taiwanese firms is involved in foreign investment, the government's effort in opening up the market brings little effect to companies. We are then unable to solely judge the deregulation effect on market volatility using the above information alone on

ChiaoDa M.αnagement Review Vol. 29 No. 1, 2009 201

whether or not the deregulation has indeed decreased stock return volatility. In the following sωtion of our research, we propose to extend our investigation by dividing the sample companies into three categories: firms with high FPI proportions, firms with high FDI proportions and firms with little foreign investment (co帥。1sample firms) in the section 5.3.

5.2.3 Goodness團Of-FitResults

Our inv巳stigation compares the goodness-of-fit between th巳 structural

change GARCH model (modeI2) and the conventional GARCH model (modell). Table 3 depicts the results of the likelihood ratios. From Table 3, the chi-square results reject the null hypothesis. The results su臨的t that structural change GARCH model reveals better prediction outcomes; it has the superior ability of explaining the stock returns and volatility.

5.3 Return and VolatiIi

ty

Analysis of Firms With

Differential

Foreign Investment Characteristics

Table 4 depicts the structural change GARCH regression results ofthe three types of institutional investors' trading behavior, according to the three firm categories mentioned above i.e. firms with high FPI proportions (FPI group),

firms with high FpI proportionsσDI group) and firms with little foreign investment (control sample group). From table 4, the coe缸icient

ß

,

is

slgm日cantly greater than zero for FPI sub-sample firms. This means that before the government allowed open markets, firms with high proportions of FPI ha吐 a

positive relationship between buy-sell differences and stock return rates. However,

after the deregulation policy, the Wald statistic test showed that the (β， +β'2 )

results were not promir閥混， pointing to the fact that the buy-sell difference's influence on stock return rates are diminishing. These results suppo前 hypothesis3 (H3) which st刮目 that: after the deregulation of foreign investment, the impact of foreign trade on stock returns is reduced for firms high in FPI.

Also, 企om 伽 d糊 on t帥 4, ←抖efi臼0削u叮叫II

high propor此tions of FDI were significantly grea仗t巴叮r than 泊er昀z oa晶sw巴11. Before the government allowed open markets, firms with high proportions of FDI had a

202 The Structural Change of Stock Returns and Volatility Subsequent

的 theDeregulationfor Foreigη Investment in Taiwan Markets

positive relationship between foreign institutional buy-sell difference and stock returns. However, after the deregulation policy, the coefficientβ2 was significantly negative. In addition, the Wald statistics test showed that the

(ß1 + β~2) results were not as promine肘， indicating that the buy-sell

\

\

Table 4

Re2ression Results of Firms With Differentia Investment Characteristics

RiJ=β。 +β'1F:，，-1 + ß2(尺，'-1 X Ði

,,)

+ β舟，1-1 + ß41

小

1 + &i

,

t'

Ci

,

t

~N(O， hit)，

(2)

九 =yo+Y1〈叫+九九-1

+

Y3(&，ι1 xÐ

i

，，)+ 仇的，←l

xÐ

i.，)'

Sample FPI FDI Con甘01

Coefficientτ'statistics Coefficientτ:'statistics Coefficientτ'statistics

F

_{i .}I _ 1 15.4721 2.0383" 3.8546 2.0718" 14.2502 。 5920 (F

,,

'_l X D

,)

-9.5109 -1.1400 -3.9415 -1.9046" -7.4448 -0.3023

B

i .1-1 11.4209 。 3967 32.9939 2.3000" 7.0132 0.9553

Iz

t l 21.4229 1.1611 8.6048 0.8712 10.4222 2.6351'" Int呻自t(β。) 0.0757 2.0375" 。.1ll 3 2.4851" 0.1306 2.7847抖， Interpret (

r

0 ) 0.0278 4.1546'" 0.0464 2.9951" 0.0873 3.4475'" Et2 -1 0.0194 2.2116" 0.0264 2.0992" 0.0551 3.0670'"

h

i.t-1 0.9704 102.0日32 U ﹒ 0.9633 69.5641'" 0.9312 49.7734"

.

Ct~l

X

D

i.t 0.0300 3.0448'" 0.0127 0.9048 -0.0017 -0.0958 (鳥，1-1 X Ði

,

l) -0.0352 -4.0021'" -0.0225 -1.6711' -0.0176 -1.2663 Wald statistics Ho: β1+β2

=

0 5.9612 -0.0869 6.8054 2 Chi-square ( X - ) 2.3144 0.0086 1.6689

Notes: 嗯嗯nificantat the 10 P耐centlevel. **Significant at the 5 percent level 抖 *Significantat the 1 percent level

Chiao Da Management Review Vol. 29 No. 1, 2009 203

deregulations. These results support hypothesis 4 (H4) which states that: after the deregulation of foreign investment

,

the impact of foreign trade on stock retums is reduced in firms with higher FDI holdings

The con甘01 group in table 4 depicts an insignificant 疚， which implies the negligible relations between foreign institutional behaviors and Taiwan stock retums. Furthermore, the Wald statistics test also showed (ß

,

+ β~2) as not prominent. After the deregulation of foreign investment policies

,

the buy-sell difference of foreign investment did not contain incremental effect on Taiwan stock retums. Neither before nor after the market opened, had foreign institutional buy-sell difference contained effect on stock retum rates.τbis finding is unique to previous descriptions of firms with high FDI and FPI, which implies that regulation policies on foreign investment yield great impact on firms high in FDI and FPI. On the con仕a旬; companies with relatively low foreign investment seem not to have perceptible influence. Consistent with the resu1ts proposed by Chang

(1999) and Wu (2005), foreign buy-sell difference is more influential on stock retums for firms with greater foreign holdings than those with lower ones.

On the whole, table 4 supports hypothesis 3 and 4. However, we must be careful of over simpli1'ying; the resu1ts acquired 企om the structural change GARCH regression analysis do not, directly and solely, point out the fact that deregulation of trade policies lesson the role of firms high in FDI and FPI as market indicators. In fact

,

the receding role of FDI and FPI firms in the market could be as a resu1t of the free market flow. Under the 企ee market, once foreign investors have 企'ee access to Taiwan marke俗， foreign investment funds are more likely to trade according to business cycle. For instance, foreign investors sell their shares at high values to gain profit, and buy-in shares when the stock sales fall to take advantage of low stock prices. Therefore

,

overall speaking

,

changes in foreign investments decreased their relationship with Taiwan stock retums.

In regards to stock vo!atili旬， coefficient

r

₃ and

r

₄ in Tab!e 4 captures the short-term and !ong-run volatili紗， respectively. Firms with high FPI have coefficient Y4 significantIy negative. This implies that after markets were opened, long-term stock retum volatility dec1ines dramatically. Similarly, firms high in FDI had a negative coefficient Y4 value. After the govemme