FOREIGN DIRECT INVESTMENTS AND SHAREHOLDER WEALTH: THE SINGAPORE EVIDENCE. David K. Ding Qian Sun*

FOREIGN DIRECT INVESTMENTS AND SHAREHOLDER WEALTH: THE SINGAPORE EVIDENCE David K. Ding Qian Sun* Division of Banking & Finance Nanyang Business School Nanyang Technological University Singapore 639798, Singapore E-mail: AQSUN@NTUVAX.NTU.AC.SG. Centre for Research in Financial Services Working Paper No. 96-04 Abstract This study examines the stock return responses to the announcements of foreign direct investments (FDI) by Singapore companies. The standard event study methodology is used to ascertain the abnormal returns around the announcement day (day 0). The study covers the period from 1989 to 1994 with a sample size of 70 events. The announcement effect is positive and significant around the announcement day. The average abnormal return is 0.4913 percent on day 0, and the two-day (days 0 and 1) cumulative abnormal return is 0.9642 percent. However, the abnormal return is unequally distributed across the sample firms. A cross-sectional analysis reveals that the two-day cumulative abnormal return of a firm is statistically significantly related to (1) the industry the FDI is in, and (2) whether the FDI is independent in nature or is in the form of a joint venture. It is, however, found to be unrelated to the country of investment. The evidence further shows that investors who trade on the information regarding a company's impending foreign investment can earn abnormal returns, net of transaction costs, by buying the stock before the event period and selling it five days after the announcement date. * We thank Thomas McInish and C. F. Lee for helpful discussion and comments I. Introduction

This study focuses on the market evaluation of foreign direct investment (FDI) announcements by Singapore companies. Four fundamental questions are addressed: (1) Do FDI announcements contain new information? (2) Do shareholders benefit from their firms' FDI decisions? (3) Can market participants earn abnormal returns by trading on FDI announcements? (4) What affects the cross-sectional distribution of abnormal returns, if any, around the announcement day? An event study is carried out to answer the first three questions and a crosssectional analysis is conducted to answer the last one. It is hypothesized in many studies that news releases have valuation implications. Based on an examination of firm-specific news reported in the 1983 Wall Street Journal Index, Thompson, Olsen, and Dietrich (1988) show that the distribution of Wall Street Journal news-conditional residual returns is different from the distribution of returns when such news is absent. Numerous other studies have examined the various types of firm-specific news such as earnings (e.g., Zarowin (1989)), mergers/acquisitions (e.g., Bradley, Desai, and Kim (1988)), capital structure changes (e.g., Masulis (1980)), business relocation decisions (e.g., Chan, Gau, and Wang (1995)), etc. and found that there are abnormal returns associated with the news release. Using data for a sample of Singapore firms, this paper adds to the existing literature by examining an additional type of firmspecific news -- FDI announcements. Ariff and Finn (1989) extensively analyse the effects of earnings, dividends, and capital changes announcements on share prices in Singapore. They find that, while abnormal returns are associated with these announcements, investors cannot trade at the time of announcement and earn an abnormal profit after taking transaction costs into account. Thus, if the FDI announcements, like those studied by Ariff and Finn, contain new information, the market price should change upon the release. If the market views the FDI announcements favorably, the stock prices of firms associated with these FDIs should increase upon such announcements. A prudent firm will not be involved in a FDI unless the project is expected to have a positive net present value. Hence, the FDI decision should increase shareholder wealth. However, this may not hold true for all firms for several reasons. First, the firm's estimation of the expected future cash flows may differ from the market's estimation either because the firm's estimates are overly optimistic, or because the firm may have some private information. Second, as pointed out by Jensen and Meckling (1976), due to the agency problem, managers may maximize their own wealth rather than shareholders' wealth. A number of empirical studies (see Ciscell and Carroll 1

(1980) for a survey) further show a positive relation between firm size and the top managers' pay levels and power. Therefore, top managers have an incentive to invest in projects that may even generate negative net present values. Thirdly, some firms may suffer from the bandwagon effect - going abroad simply because others before them have gone successfully. This is more likely to occur in Singapore during our period of study when FDIs became popular among firms of all sizes. In addition, the Singapore government encourages domestic firms to go abroad. Therefore, the FDI announcement effect should be different across firms even if the announcements may benefit shareholders in general. This paper further examines the relationship between the distribution of abnormal returns around FDI announcements across the firms and various FDI characteristics. We do not have direct information to judge whether a firm s FDI is aimed to maximize shareholders wealth or not, but it is reasonable to argue that the agency problem and the bandwagon effect would be more serious when a firm has more free cash flow for its managers to play with. The debt-asset ratio (long-term debt over total assets) of a firm can be used as a rough proxy for its free cash flow. Usually, the higher the debt-asset ratio, the less the free cash flow. Also when the debt-asset ratio is high, the management of a firm is under high pressure to do more accurate capital budgeting before carrying out a FDI. Hence, the abnormal return associated with a FDI announcement should be positively related to the firm s debt-asset ratio. The record of a firm s return on assets (ROA) can provide information on how efficient a firm manages its assets. If market participants, lacking more specific information, use a firm s historical ROA as a benchmark to evaluate the firm s FDI, then the abnormal returns associated with the FDI announcements would be positively related to the ROA. Previous studies also show that the wealth effect of FDIs are associated with FDI characteristics. Collins (1990) finds that stockholders of US MNCs do not benefit from the firms' diversifying into developing countries. Harris and Ravencraft (1991) examine FDIs by investigating shareholder wealth gains for 1,273 U.S. firms acquired during the period 1970-1987. Their findings indicate that cross-border takeovers are more frequent in industries that are research and development intensive, and targets of foreign buyers have significantly higher wealth gains than do targets of domestic firms. Culem (1988) shows that the host market size, growth rate, and tariff barriers are significant locational determinants of FDIs among industrialized countries. An interesting revelation is that although companies prefer the lowest cost foreign location, FDIs are usually not found to be motivated by a search for lower unit labor costs. 2

We examine three FDI characteristics: the country, the industry, and the way the FDI is conducted. There are pros and cons to invest in either developing or developed countries. However, investing in developing countries may be more beneficial to Singapore firms than investing in developed countries. Most Singapore firms still do not posses the superior technology or know-how that western firms have. So it may be difficult to compete with western firms in their home countries. However, Singapore firms technology and know-how may be more adaptable in developing countries and they may be more familiar with the business environment in these countries so that they can successfully compete with other firms there. It could be interesting to see if FDIs in developing countries are associated with higher abnormal returns around the announcements. It is plausible to hypothesize that the gain from FDIs is different across industries. Singapore is a small open economy heavily dependent on international trade. As the economy matures, living standards have dramatically increased and so did production costs. This affected the competitiveness of tradable goods more than non-tradable goods produced in Singapore. Assuming FDIs can lower the production cost, a FDI involving the manufacture of tradable goods may benefit shareholders more than a FDI involving non-tradable goods. We catagorize the FDIs into two sectors: manufacturing and service. Roughly speaking, manufacturing is associated with producing tradables (almost all manufacturing goods produced in Singapore are tradables) and service is associated with producing non-tradables (and some tradables). The abnormal returns around the announcement should be higher if the FDI is in the manufacturing sector. There are several ways to conduct a FDI. A firm can conduct a FDI independently and start everything from the beginning. This may be a slow process but the firm has full control over the FDI. The firm can also conduct a FDI via a joint venture or an acquisition. The process may be much faster but, in Singapore, one often hears complaints about joint ventures associated with FDIs. Many problems arise between the partners due to different management styles, etc. For acquisitions, it has been documented in the literature that the target shareholders gain in an acquisition, while the shareholders of acquiring firms gain little, if not lose. Therefore, we hypothesize that the abnormal returns associated with independent FDIs would be higher than that with a joint venture or an acquisition. The remainder of this paper is organized as follows. Section II discusses FDIs by Singapore companies. Section III contains a description of the data set and the research methodology 3

employed. The results of the study are presented and analyzed in section IV. The last section summarizes the paper and offers some concluding remarks. II. Foreign Direct Investments By Singapore Companies Traditional multinational corporations (MNC) started the internationalization process as raw material seekers, market seekers, or cost minimizers. As raw material seekers, their goal was to capitalize on the raw materials that were found abroad. A market seeker is one that goes to a foreign land to produce and sell whereas cost minimizers seek out and invest in lower-cost production facilities overseas in order to remain cost-competitive both at home and abroad. Shapiro (1996) points out that becoming multinational is not a matter of choice but, rather, one of survival. This is especially true for many companies in Singapore, a small city state. It is well known that Singapore has few natural resources and a small domestic market. As its economy matures, the costs of production increases. In order to survive, the Singapore firm must continually reduce its costs, exploit economies of scale, engage in multiple sourcing, seek new knowledge, and retain its domestic customers. Therefore, the wealth effect associated with FDIs may be larger on shareholders in Singapore than elsewhere. The typical Singapore firm's FDI is concentrated in its neighboring countries. Recently, however, it has expanded to include many other countries. Regionalization, or FDIs in a country's geographic vicinity, is growing in importance to Singapore s economy. The Singapore government in recent years has encouraged the local companies to expand their businesses to those foreign economies which have high growth potential. Countries such as Vietnam and China welcome foreign investment as part of their efforts to open up their economies. They offer vast opportunities to Singapore firms willing to venture overseas. Other countries which are closer in proximity to Singapore, such as Indonesia and Malaysia, are also increasing their efforts in economic development. The changes in the economic patterns of the emerging economies represent both a threat and an opportunity to Singapore s development. As a threat, it may be viewed that Singapore may not be able to compete with the emerging economies because of their abundant availability of raw materials and low labour costs. Foreign manufacturing companies may shift their manufacturing facilities from Singapore to those countries. As an opportunity, however, these emerging economies provide enormous benefits to Singapore's multinationals as they give the Singapore 4

companies an expanded market. Singapore firms also possess the much-needed, technological know-how and management skills that are required by theoe countries and are essential to their continued growth. Foreign governments, especially those among the developing nations, will thus be very receptive of Singapore companies for the capital and skills that they provide. The companies involved in a FDI will thus enjoy the benefits of lower production costs and access to new markets for their products. Panel A of Table 1 shows that the FDIs of Singapore companies have increased dramatically. It has almost doubled from 1990 to 1993. By the end of 1993, Singapore companies in the private non-financial sector had invested some $50.1 billion abroad. This was $7.9 billion or 18.6% higher than the amount invested in 1992. Of this, total direct investments amounted to about $28.2 billion which represents a significant growth of 25.5% over the amount invested in 1992. The growth in direct investments was the highest compared to all other types of investments. Panel B of Table 1 shows the FDIs by Singapore firms according to the geographic region. Singapore MNCs exhibit a strong preference for investments in Asia. By the end of 1993, more than half of their investments abroad (54%) are in Asia. Although the ASEAN region remains attractive to Singapore investors (accounting for more than a quarter of Singapore s investment abroad), the trend also shows that more are venturing into other parts of Asia, especially in China and India, which are two of the most populous countries in the world. These two countries provide a large market for their products as well as cheap labour costs. With the regionalization drive in Singapore, and with the continued attraction and growth of investment opportunities in many Asian countries, the strong growth in FDI is expected to continue. III. Data and Methodology Event Study This study investigates the effects of FDI announcements on shareholders' wealth using a sample of companies listed on the Stock Exchange of Singapore (SES). An event study methodology (Brown and Warner,1985) is employed to ascertain whether there are any abnormal returns associated with the FDI announcement. In the period from 1989 to 1994, 120 FDI announcements are made by SES-listed firms to invest abroad. These announcements, or events, are identified from the SES Journal and the Business Times. Daily data on stock prices, dividends paid, and the SES all-share index are 5

collected for the period starting 231 trading days prior to the announcement date (t = 0) till 10 trading days after the announcement date for a total of 242 days. The stock price and index information are obtained from the Stock Exchange of Singapore. Dividend information is extracted from the SES Journal. A 21-day period around the announcement date (from t = -10 to t = 10) is designated as the event period (or event window). Daily stock and market returns, adjusted for dividends and other capital changes, are computed starting on day -230 through day 10. The parameters of the market model are estimated during the estimation period of 220 days which run from days -230 through -11. R jt = α j + β j R mt + e jt (1) where R jt is the dividend-adjusted return of security j on day t; R mt is the return of the market on day t proxied by the return on the SES All Share Index; α j is a stable component of security returns and is constant over time; β j is the beta or a measurement of the systematic risk of security j and is assumed to be stable over time; and e jt is the random error term. Companies that do not meet the selection criteria are excluded from the study. These include those with insufficient observations and those with overlapping event periods. Multiple announcements by the same company are treated as separate observations if the subsequent announcements occur two or more weeks after the previous announcement. In all, 70 announcements by 44 companies remain after the elimination process. Table 2 shows the chronological distribution of the 70 FDI announcements. The abnormal returns (AR) of each company's stock are then determined over the event period (t = -10 to t = 10) as follows: AR jt = R jt - E( R jt ) (2) If announcements of foreign investments have no impact on stock prices, then, on average, one should expect abnormal returns to be zero. Three common methods of aggregation of abnormal returns have been employed in the finance literature. They are, namely, the cumulative average abnormal returns (CAAR), the abnormal performance index (API), and the standardised cumulative prediction error (SCPE). Both the CAAR and API start by averaging prediction errors across firms on day t. After this day, the CAAR and API are the cumulative sum and geometric product of the abnormal returns, 6

respectively. Due to its ease of testability, CAAR is more widely used. However, recent practice has been to standardize the prediction errors before aggregation. This is because the excess returns of the securities may not be identically distributed across the sample. The standardized prediction error (SPE) is calculated by dividing the abnormal returns of stock j on each day t by the standard error of the forecast (S jt ): SPE jt = AR S jt jt (3) where S jt = S j 1 + 1 k + k i=1 ( R - R ) mt m 2 ( R - R ) mi m 2 (4) and S = j k 2 (e jt - µ j ) t=1 (5) k - 1 where k is number of days in the estimation period; e jt is as in equation (1); and µ j is the average residual over the estimation period. However, the estimate of the standard deviation on a one-day return is not S j since this would assume that variation in the market during the event period is essentially the same as it was during the estimation period. Furthermore, S j does not adjust for the number of observations in the estimation period. The estimate S jt in equation (4) originally proposed by Patell (1976) to account for these considerations was popularised in the finance literature by Dodd and Warner (1983). The term 1/k in equation (4) adjusts for the length of the estimation interval where the greater k is, the less would be the error in the out-of-sample forecast (i.e., assuming there is stability in the relationship). SPE is approximately unit normal. The standardised cumulative prediction error (SCPE) for firm j over a time period of an event is determined by aggregating the SPEs. It is essentially the 7

sum of the SPE j between any 2 days, t 1 and t 2, adjusted for the number of days (m) within the window period. SCPE = j t 2 t 1 SPE m jt (6) The purpose of doing this is to see if announcements of foreign investments by SES-listed firms have a significant impact on the companies' stock prices. Thus, if the impact is significant, one should expect to see a significant rise in the abnormal return on the day of the investment announcement. To determine if the average daily abnormal returns are significant, the test statistic Z on any day t in the event window for all n stocks is constructed Z SPEt n SPE i = = 1 n i (7) The corresponding test statistic for the SCPE is Z SCPEt n SCPE = i = 1 n i (8) Both statistics in equations (7) and (8) follow a standard normal distribution. If the FDI announcements do not lead to abnormal returns, the Z-statistics should not be significantly different from zero. Therefore, we can test the null hypothesis, H 0 : FDI announcements have no statistically significant impact on stock prices, against the alternative hypothesis, H 1 : FDI announcements have a statistically significant impact on stock prices. To further check whether market participants react to the FDI announcements, we compare the daily average trading volume across all firms in the event window with the overall 8

average trading volume (average across all firms and days) in the estimation period. The overall average trading volume in the estimation period is k n V = ( V ) kn t = 1 j = 1 jt (9) where V jt is the trading volume for firm j on day t in the estimation period. The average trading volume for a particular day in the event window is V t n V j = = 1 n jt (t = -10, -9,... 9, 10) (10) The relative trading volume on a particular day in the event window is simply V Vt. If market participants do react to FDI announcements, we should observe that the relative trading volume is greater than one, especially on the days around the announcement. Cross-Sectional Analysis From the previous discussion, we identify five variables that may affect the abnormal returns across firms upon FDI announcements. They are ROA (return on assets), debt-asset ratio (D/A), and three dummy variables reflecting the FDI characteristics (country, type, and industry). The ROA and debt-asset ratio for each firm are retrieved from the financial database at Nanyang Business School. The ROA used in the analysis is the average of three years before the announcement year. It indicates how efficient the firm managed its asset in the recent past. The debt-asset ratio is calculated as long-term debt over total assets using the data from the year before the announcement year. It roughly proxies the free cash flow a firm has. Table 3 dichotomizes the 70 announcements according to each of the three FDI characteristics. The industry dummy is assigned a value of one if the FDI is primarily in the manufacturing business; otherwise, a zero value is assigned. If the FDI is a joint venture with, or an acquisition of, a company in the foreign country, the type dummy is assigned a value of one; if it is an independent investment, a value of 9

zero is assigned. The country dummy receives a value of one if the MNC invests in a country other than an OECD country or a in a newly industrialized economy (NIE), and a value of zero if it does. The cross-sectional analysis is conducted via the following regression D CAAR [0, 1] = β + β Industry + β Type + β Country + β 4ROA + β5( ) + ε A i 0 1 i 2 i 3 i j j j (11) where the CAAR j [0,1] is the cumulative abnormal return over day 0 and day 1 for announcement j, β i 's are the regression coefficients for the five variables mentioned above, and ε is a random Gaussian error term. IV. Results and Analysis Table 4 shows the results of the event study. It reports the average abnormal returns (AAR) across all events (announcements), the cumulative average abnormal returns (CAAR) and the relative volume during the event period [-10, 10]. The Z-statistics for the standardized prediction error (SPE) and the standardized cumulative prediction error (SCPE) are also reported. In addition, it shows the number of positive versus negative AARs in each day during the event window. On the announcement day, shareholders of companies that make a foreign investment announcement earn, on average, a statistically significant one-day return of 0.4913%. The Z SPE on this day is significant at the 0.05 level. On day 1, the abnormal return stands at 0.4729% with a Z SPE that is significant at the 0.10 level. The trading volume is also 38 to 68 percent higher during the days immediately around announcements than during the estimation period. On day zero the relative volume reaches the highest level at 1.6823. These show that FDI announcements do contain new information and thus, move the market. Also, investors generally react favorably to the internationalization efforts of SES-listed firms as evidenced by the significant positive reaction of the market to announcements of investments in foreign countries. On the whole the current shareholders of these firms can benefit from their firms FDI announcements. Over the 21-day event period, an investor earns an average return of 2.7284%. The Z SCPE over this period is significant at the 0.01 level. This means that an investor who buys the stock of 10

an announcing firm on day -10 will earn a return of 2.7284%, on average, if he holds on to the stock till day +10. The two-day announcement period CAAR, which includes days 0 and +1, is 0.9642%. In Singapore, a round-trip buy and sell transaction involves commissions and taxes of about 2.26%. All AARs in the study report values that are well below one percent. Thus, on average, no profit opportunity is possible from trading shares using a buy today and sell tomorrow strategy. These results are, therefore, very much consistent with market efficiency. However, as the CAARs reveal, it is possible for an astute investor, either through sound analysis of a firm's fundamentals or through the receipt of privileged information, to earn, on average, a modest profit of 0.47% that is net of commissions and taxes over the 21-day event period. In fact, after taking into consideration the transaction costs, a profit opportunity is possible only if investors bought the stock before day -10 and liquidate their positions after day 5. Figure 1 displays a plot of the cumulative average abnormal returns over the 21-day event period of the foreign investment announcement. It is interesting to note that the CAAR starts to increase at about five days prior to the announcement date. This can be attributed to the leakage of information to the market. The CAAR follows a general uptrend even after the announcement date. The uptrend continues till about eight days after the announcement. Such a phenomenon shows that either (1) the market did not fully reflect the value of the foreign investment announcement or that (2) speculators begin to chase the stock after the announcement is made. The slight dip in CAAR on days +2 and +4 could be due to investors taking profit on their investments that were made based on the information leakage prior to the announcement date. As expected, not all announcements generate positive abnormal returns during the event period. The column of positive versus negative abnormal return in Table 4 indicates that there are quite a number announcements with negative AARs on each day during the event window. The number of positive versus negative signs is 37:33 on day 0 and 41:29 on day 1. The shareholders of those firms with negative abnormal returns view the FDI announcement by their companies as one that would decrease their wealth. Such firms may have managers that have the incentive of simply increasing the value of assets under their control or they may be suffering from the bandwagon effect. In addition, the AARs may be influenced by the FDI characteristics. As such, we further examine the possible factors that may affect the abnormal returns across the FDI announcements. Cross-sectional Regression Results 11

The results of the cross-sectional regression across all events are reported in Table 5. The dependent variable is the individual firm s two-day (day 0 and day 1) cumulative abnormal return (CAAR) for each announcement. The independent variables are ROA, D/A, and dummy variables that account for the FDIs industry (manufacturing or others), the type of investment (joint venture and acquisition or independent), and country of investment (non-oecd/nie or others). Columns 1 through 5 show the results of regressing CAAR against each of the five independent variables, one at a time. Column 6 shows the results for the full model (equation 11) regression. The results reveal that only the industry and the type of investment are important explanatory variables of the two-day announcement CAAR. The industry dummy is positive and significant at the 0.01 level in all the relevant regressions. The market tends to react favorably if the announced FDI is in the manufacturing business. This is consistent with rising manufacturing costs in Singapore and it may also imply that Singapore does not have many comparative advantages in manufacturing at home. The type dummy is negative and significant at nearly the 0.01 level in all the relevant regressions. This negative relationship suggests that investors prefer firms that make independent FDIs to those that are involved in joint ventures or acquisitions. This is consistent with many unpleasant experiences associated with joint ventures abroad. It also echoes the previous findings that crossborder acquisitions do not bring wealth gains to the shareholders of acquiring firms. On the other hand, the market seems to be indifferent to the type of countries that MNCs invest in. This is evidenced by the statistically insignificant t-value of the country dummy variable in all the relevant regressions. It seems contrary to the common belief that Singapore firms may gain more when they invest in developing countries. However, it is similar to the finding by Collins (1990). The coefficients of ROA and D/A are never significant in all relevant regressions. The easiest explanation is that agency problems and bandwagon effects are not serious among Singapore firms such that they would not affect the distribution of abnormal returns among firms upon FDI announcements. However, it is possible that the average three-year ROA and the debtasset ratio may not be good proxies for capturing the agency problem and bandwagon effects. For most firms in Singapore, the debt-asset ratio is quite low so that it may not be a good proxy for the free cash flow. The historical ROA may also not be a good proxy for whether the current FDI is a good one or not. V. Summary and Concluding Remarks 12

This study has examined the effects of announcements of foreign investments by Singapore companies listed on the Stock Exchange of Singapore on shareholder returns over a period of six years, from 1989 through 1994. The analysis of abnormal returns indicates that the announcements have an impact on shareholder wealth. On the announcement day, shareholders earn, on average, a one-day abnormal return of 0.4913 percent. The two-day announcement cumulative abnormal return is 0.9642 percent. The evidence shows that investors who trade on the information regarding a company's impending foreign investment can earn abnormal returns, net of transaction costs, by buying the stock at least 10 days before and selling it five days after the announcement date. The study also finds that shareholders of MNCs in the manufacturing business that make independent foreign direct investments in any country tend to earn higher abnormal returns as a result of the investment announcement. In conclusion, this study has found that the internationalization efforts of SES-listed companies are viewed favorably by the market. However, the actual amount that investors may earn as a result of the announcement has to depend invariably on the timing of the purchase and sale of the stock. 13

References Ariff, M. and F. Finn (1989). "Announcement Effects and Market Efficiency in a Thin Market: An Empirical Application to the Singapore Equity Market," Asia Pacific Journal of Management, 6:243-265. Bradley, M., A. Desai, and E. H. Kim. (1988). "Synergistic Gains From Corporate Acquisitions and Their Division Between the Stockholders of Target and Acquiring Firms," Journal of Financial Economics, 21:3-40. Brown, S. J. and J. B. Warner (1985). "Using Daily Stock Returns," Journal of Financial Economics, 14:3-32. Business Times, Singapore. Various issues. Chan, S., G. Gau, and K. Wang (1995). "Stock Market Reaction to Capital Investment Decisions," Journal of Financial and Quantitative Analysis, 30:81-100. Ciscell, D. and T. Carroll (1980). "The Determinants of Executive Salaries: An Economic Survey,"Review of Economics and Statistics, 62:7-13. Collins, J. M. (1990). "A Market Performance Comparison of U.S. Firms Active in Domestic, Developed and Developing Countries," Journal of International Business Studies, 21:271-287. Culem, C. G. (1988). "The Locational Determinants of Direct Investments Among Industrialized Countries," European Economic Review, 32:885-904. Dodd, P. and J. B. Warner (1983). "On Corporate Governance: A Study of Proxy Contests," Journal of Financial Economics, 11:401-438. Harris, R. S and D. Ravenscraft (1991). "The Role of Acquisitions in Foreign Direct Investment: Evidence From the U.S. Stock Market," Journal of Finance, 46:825-844. Jensen, M. and W. Meckling (1976). "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure," Journal of Financial Economics, 3:305-360. Masulis, R. (1980). "The Effects of Capital Structure Change on Security Prices," Journal of Financial Economics, 8:139-178. Patell, J. M. (1976). "Corporate Forecasts of Earnings Per Share and Stock Price Behavior: Empirical Tests," Journal of Accounting Research, 14:246-276. Shapiro, A. C. (1996). Multinational Financial Management, Prentice-Hall, New Jersey, U.S.A. SES Journal. A publication of the Stock Exchange of Singapore. Various issues. Thompson, R., C. Olsen, and R. Dietrich (1988). "The Influence of Estimation Period News Events on Standardized Market Model Prediction Errors," Accounting Review, 63:448-471. 14

Zarowin, P. (1989). "Does the Stock Market Overreact to Corporate Earnings Information?" Journal of Finance, 44:1385-1399. 15

Table 1 Singapore s Investment Abroad, 1990-1993 Panel A: Type of Investment Abroad S$ in Million Year 1990 1991 1992 1993 Direct Investments 16877.9 18607.5 22442.5 28159.7 Portfolio Investments 8126.7 9583.4 7661.4 8137.8 Other Foreign Assets 10815.8 11252.1 12097.3 13758.5 Total 35820.5 39443.0 42201.1 50056.0 Percent Distribution Direct Investments 47.1 47.2 53.2 56.3 Portfolio Investments 22.7 24.3 18.2 16.3 Other Foreign Assets 30.2 28.5 28.7 27.5 Total 100.0 100.0 100.0 100.0 Panel B: Direct Investment Abroad By Major Regions S$ in Million Year 1990 1991 1992 1993 Asia 7,013.3 7,401.4 9,209.3 11,480.0 ASEAN 3,567.1 3,995.6 4,896.7 5,933.8 Europe 1,095.4 1,397.6 1,480.2 1,549.7 North America 716.4 1,322.7 1,621.3 1,866.6 Oceania 1,994.3 2,068.2 2,102.3 2,035.7 Others 2,802.3 2,993.9 3,328.2 4,308.2 Total 13,621.7 15,183.8 17,741.3 21,240.2 Source: Yearbook of Statistics, Singapore 16

Table 2 Breakdown of Sample Panel A: Classification of Sample According to Year of Announcement Year Number 1989 9 1990 12 1991 11 1992 14 1993 11 1994 13 Total 70 Table 3 Characteristics of the FDIs This table presents the distribution of the FDI s characteristics. The industry variable is classified according to the business the investments in, i.e., manufacturing or others. The type of investment is classified according to how the FDI is conducted, i.e., a joint venture (which includes mergers and acquisitions) with a company in the foreign country, or an independent direct investment. The country variable classifies an investment according to the country of the FDI is in, i.e., whether the Singapore firm invests in an OECD or newly industrialized economy (NIE), or others. Manufacturing Others Industry Type Country 23 47 Joint Venture Independent 42 28 OECD + NIE Others Total 70 Total 70 Total 70 11 59 17

Table 4 Event Study Analysis This table shows the average abnormal returns (AAR), the cumulative average abnormal returns (CAAR), and the Z statistics for the SPE and SCPE during the event period from day - 10 to day +10. The number of firms with positive and negative abnormal returns in each day is summarized under the column +:-. The relative volume for each day in the event window is the ratio of the average trading volume on day t to the average trading volume in estimation period. Day AAR Z SPE CAAR Z SCPE +:- Relative Vol -10 0.4588% 1.906 0.4588% 1.906 43:27 0.9988-9 0.3919% 2.706 ** 0.8507% 3.261 ** 35:35 1.1233-8 -0.4590% -2.411 * 0.3917% 1.271 27:43 1.2365-7 -0.2084% -0.439 0.1833% 0.881 23:47 1.5832-6 -0.0527% -0.036 0.1306% 0.772 28:42 1.3116-5 0.0985% 1.159 0.2292% 1.177 32:38 1.4149-4 0.1749% 0.913 0.4040% 1.435 34:36 1.0526-3 0.4950% 1.785 0.8991% 1.974 * 39:31 1.2155-2 0.2858% 1.395 1.1849% 2.326 * 42:28 1.2639-1 -0.2150% -1.257 0.9699% 1.809 32:38 1.3884 0 0.4913% 2.286 * 1.4612% 2.414 * 37:33 1.6823 1 0.4729% 1.900 1.9341% 2.860 ** 41:29 1.4684 2-0.1422% -0.467 1.7918% 2.618 ** 35:35 1.4186 3 0.0276% 1.433 1.8194% 2.905 ** 35:35 1.0253 4-0.0537% 1.000 1.7657% 3.065 ** 30:40 0.9912 5 0.3913% 1.171 2.1569% 3.261 ** 38:32 1.0396 6 0.2921% 1.102 2.4491% 3.430 ** 39:31 1.1316 7-0.0929% -1.129 2.3561% 3.068 ** 34:36 0.9318 8 0.4997% 1.932 2.8559% 3.429 ** 42:28 0.9433 9-0.0376% -0.428 2.8182% 3.247 ** 32:38 0.8420 10-0.0898% -0.076 2.7284% 3.152 ** 32:38 0.8690 * (**) Significant at the 0.05 (0.01) level. 18

Table 5 Cross-sectional Regression Analysis This table provides the results of the two-day Cumulative Average Abnormal Return (CAAR 0,1 ) return on assets (ROA), debt-asset ratio (D/A), and three dummy variables -- industry, type, country. The Industry Dummy is assigned a value of one if the FDI is in the manufacturing sector and zero otherwise. The Type Dummy is assigned a value of one if the FDI is conducted through a joint venture, a merger, or an acquisition, and zero otherwise. The Country Dummy has a value of one if the FDI is invested in a developing country and zero if it is invested in an OECD country or a NIC (newly industrialized country). The t statistics in parentheses are heteroskedasticity-consistent. Variable (1) (2) (3) (4) (5) (6) (7) Intercept -0.0031 (-0.083) 0.0240 (3.424)** 0.0053 ( 1.365) 0.0074 (1.537) 0.0019 ( 2.041)** 0.0003 ( 0.054) 0.0120 ( 2.264)** Industry Dummy 0.0303 ( 3.394)** 0.0276 ( 3.616)** 0.0258 ( 3.290)** Type Dummy -0.0239 (-2.929)** Country Dummy 0.0051 ( 0.831) ROA 0.0469 (0.689) D/A -0.0077 (-0.419) -0.0197 (-2.978)** 0.0092 ( 1.352) 0.0373 ( 0.777) 0.0152 ( 0.922) -0.0181 (-2.663)* Adj. R 2 0.1665 0.1085-0.0116-0.0077-0.0128 0.2130 0.2216 DW 1.8559 1.5573 * (**) Significant at the 0.05 (0.01) level. 1.8957 1.8581 1.8944 1.5420 1.5790

Figure 1 Cumulative Average Abnormal Returns 3.5 3 2.5 CAAR (%) 2 1.5 1 0.5 0-10 -9-8 -7-6 -5-4 -3-2 -1 0 1 2 3 4 5 6 7 8 9 10 Day