The Effect of Investment Type on Industrial Information Transfer of Foreign Direct Investment Announcement
4. Research design
4.1 Research period and sample selection criteria
This research covered the period from 2001 to 2007 and examined the electronics companies in Taiwan (listed on the TAIEX or OTC board) that disclosed their investment plans in China in the said period. We collected our sample and searched for related announcements on Market Observation Post System (MOPS), using key-words and combinations such as“investmentin theMainland”,“investment in China”,“capitalinvestmentin China”and so on in the“announcement”page of the system. Information of thesamplefirms’accounting, financial, and stock prices was gathered from Taiwan Economic Journal (TEJ). The data period of our variables (such as company size) extends from 2000 to 2006 since we have to use the beginning accounting and financial information at the year of announcement to analyze the information transfer effect. In addition, we eliminate companies that announced other major events three days prior to or after the announcement of the investment in China, so as to prevent potential effect from these events. Firms with incomplete data were also excluded from our empirical analysis.
Based on the criteria, we collected 98 companies and 208 announcements of investment plans in China, with 11 announcements in 2001, and 70, 36, 31, 23, 22,
The Securities and Exchange Law article 36 stipulates that within two days from thedateofoccurrenceofmajoreventsthathave amaterialimpacton shareholders’ equity or securities prices (ex. investments in China), companies shall publicly announce and register with the Competent Authority. In general, acompany’sdecision and announcement pertaining to investments in China follow the following procedure.
First, the board of directors decides on the date of investment plan (not yet approved by the Investment Commission at this point). Then the company files an application and obtains approval from the Investment Committee. After the approval, the company can use the un-finished quota within the investment capital to invest in China. Based on the efficient market hypothesis, we assume that after a firm announces the approval of its investment plan in China on MOPS, investors can immediately obtain the information and affect the share price reactions. As such, we define the event date as the date when companies announce the approval of their investment plans by the Investment Committee. However, after examining the announcing dates of the sample firms, we found that over 50% of the companies disclosed their investment plans after the stock market closed. For these companies, the“eventdate”was the next trading date of their announcements.
4.3 Categorization of the sample firm industries
When analyzing theeffectofacompany’sFDI announcementon rivals’stock prices, a proper categorization of the sample firms will significantly enhance the accuracy of the research. Therefore, wecategorizethesamplefirmsbased on TEJ’s second industry categorization criteria (3-digit Standard Industrial Classification; SIC) and divide the companies into eleven subsectors; those sub-industries are PC system, motherboards, optronics, electronic components, wireless local area network (WLAN) equipment, semiconductor, electronic equipment, communication equipment,
computer channel, consumer electronic, and software service provider (e.g., application service providers; ASP).3
4.4 Investment types of announcing firms
In this study, we define investment types based on a company’s investing products (core-business/non-core business) as well as the investing industry (whether it’satthesamevalue chain position as the mother company) in China. Based on the criteria, we found that 121 firms in our sample belong to the core-business investment type while the rest 84 firms invest in non-core businesses. With respect to the investing industry, afirm’sinvestmentisregarded ashorizontalFDI if it belongs to the same value chain as the mother company; otherwise, it is considered to be a vertical FDI. In our sample, 162 companies have horizontal type of investments while the rest 43 belong to the vertical FDI type.
4.5 Methodology and empirical models
4.5.1 Information content effect of China investment announcements
We briefly describe the key features of the event study methodology and how to estimate the abnormal returns (see Brown and Warner, 1985, for details). Various models have been proposed to estimate abnormal returns. We adopt the market model to estimate the daily stock returns of FDI announcements because the model is the best specified model and controls for the systematic risk of stock.
it mt i i
it
R
R
(1)is event period,
t 0
is the announcement day of China investment, andt
is estimation period.4R
itis the rate of return on the stock price of firm i on day t.R
mtis the rate of return on a market portfolio of stocks on dayt. is the systematic risk of stock i.
iis error term.
itFrom estimation of equation (1), we derive estimates of daily abnormal returns (AR) for the i firm using equation (2).
mt
where
ˆ
i and
ˆ are the ordinary least squares (OLS) parameter estimates obtainedi from the regression ofR
itonR
mtover an estimation period preceding the China investment announcement (t { 180 , 179 ,..., l 1 }
). The abnormal returns (AR )
it represent returns earned by the firm after the investors have adjusted for the “normal”return process. That is,
AR
it is the difference between the actual return of firm i on day t (R
it) and the expected return (Rˆ). Averaging the abnormal returns across the
it sample firms on any day t, the daily mean abnormal return (AR ) can be expressed as
t equation (3).where n is the number of sample observations on day t. The mean cumulative abnormal return over a given event period (
t { l ,..., 0 ,..., k }
), is the sum of the daily mean abnormal returns and is expressed as equation (4).
We use t statistic obtained from standardized residual method to make statistical inference. The significance of the mean (cumulative) abnormal return allows us to infer that the event (China investment announcement) has an essential impact on the value of the firms.
4 We also employ different estimation periods, t{210,...,l1} , t{180,...,20} and
t , to rerun empirical analysis. We obtain the similar results.
4.5.2 Information transfer effect of China investment announcements
We refer some related researches, for example Lang and Stulz (1992), Erwin and Miller (1998), Akhigbe and Martin (2000), Chen et al. (2005), and Chen et al. (2007), to build the empirical model:
t
where
Ccar
i,t( l , k )
represent the cumulative abnormal returns of home-market rival firm i in the event (China investment announcement) window ( l
,k
).) ,
,
( l k
Acar
jt
is the cumulative abnormal returns of China investment announcing firmj in the event window ( l
,k
).Ams
j,t1is the market share of announcing firm j at the fiscal year prior to the announcement.Apb
j,t1is growth opportunity of announcing firm j at the fiscal year prior to the announcement.R
ij,t1represents announcing firm j degree of relatedness to the industry at the fiscal year prior to the announcement.1 ,t
HI
ij is the degree of industry competition.Cs
i,t1 is the relative size of home-market rivals at the fiscal year prior to the announcement.Cpb
i,t1is the growth opportunity of rivals at the fiscal year prior to the announcement.Clev
i,t1 is the financial leverage of rivals at the fiscal year prior to the announcement.Crdi
i,t1 is the R&D intensity of rivals at the fiscal year prior to the announcement.Cfc
i,t1 is the free cash flow of rivals at the fiscal year prior to the announcement.time
j,t represents the year that firm j announce to invest in China. is a disturbance term.
itThe extent of the announcement effects of FDIs on home-market rivals is affected by the extent to which the announcement is expected (Chaney et al., 1991). If an announcement of event is “big news,”there should be a larger effect on the announcer and a larger intra-industry effect on the rivals (e.g., Hertzel, 1991).
the information transfer effect of China investment announcement to home-market rivals. A positive relationship between the abnormal returns of the announcing firms and those of the home-market rivals (
is significantly positive) suggests that, on
1 average, the contagion intra-industry effect of FDI announcement dominates the competitive effect. A negative relationship suggests that, on average, the competitive intra-industry effect of FDI announcement dominates the contagion effect.In the following section, we confer on the potentially relevant variables that could explain the cross-sectional variation of the wealth impact of FDI on home-market industry rivals.
（1）Dependent variable
Cumulative abnormal returns of home-market rival firms (
Ccar
i,t( l , k )
)This paper uses the cumulative abnormal return rate of event window
( l , k )
(l
days before andk
days after the date of announcement) to proxy stock price reactions for the home-market rival firms to the announcement of investment in China.We use different event windows, for example (-5, 5), to implement the empirical analysis to avoid the confounding effects of windows.
（2）Independent variable