Share repurchase announcements

Share repurchase announcements The influence of firm performances on the share price impact Master Thesis Finance Student name: Administration number: Study Program: Michiel (M.M.T.) van Lent S166433 Finance Supervisor name: Second reader: Faculty: Dr. F. Castiglionesi Prof. Dr. F. de Jong School of Economics and Management Date: 28 November 2013

Abstract The objective of this study was to examine the effect of share repurchase announcements on share prices over the period 2008-2012. It was assessed whether the financial crisis, which began in 2007, influenced the level of abnormal returns after a share repurchase announcement. The sample is including firms from the NASDAQ, NYSE and LSE. The markets are tested jointly and individually, to see whether there are differences between these markets in the short term price reaction. It is found in the event study that on average, when looked at the markets together, share repurchase announcements have a statistically significant positive abnormal return of 0.61% on the announcement day. Over the window (-1, +1) the cumulative abnormal returns of 1.13% are also statistically significant. When looking at each market separately, it was found that on average the cumulative abnormal returns of the NASDAQ (2.29%) and the LSE (1.93%) were positively statistically significant over the event window (-1, +1) while this was not the case for the NYSE (0.21%). Keywords: share repurchase, repurchase methods, repurchase motivations, event study, cumulative abnormal return

Table of Contents 1. Introduction... 1 2. Theoretical Framework and Hypotheses development... 5 2.1 Share Repurchase... 5 2.1.1 Dividends versus share repurchase... 5 2.1.2 Methods of share repurchase... 6 2.1.3 Motives for share repurchase... 7 2.1.4 Effects of share repurchase announcements... 9 2.2 Literature Review... 11 2.3 Hypotheses development... 11 3. Data and Descriptive Statistics... 15 3.1 Data... 15 3.1.1 Data collection... 15 3.1.2 Data clean up and sample selection... 15 3.2 Independent variables for regression analysis... 17 3.3 Descriptive statistics... 20 4. Methodology... 21 4.1 Event Study... 21 4.1.1 Event and estimation window... 21 4.1.2 Measuring normal performance... 22 4.1.3 Determining abnormal and cumulative abnormal returns... 23 4.1.4 Testing and analyzing (cumulative) abnormal returns... 24 4.2 Regression analysis... 24 5. Empirical Results... 25 5.1 Event study results... 25 5.1.1 Abnormal returns... 25 5.1.2 Abnormal returns by market... 28

5.2 Share price performance across firm and repurchase size... 29 5.3 Regression analysis results... 30 6. Conclusion... 34 6.1 Conclusion and Discussion... 34 6.2 Limitations and future research... 36 REFERENCES... 37 APPENDIX... 40 Appendix A: Cumulative Average Abnormal Returns (CAAR)... 40 Appendix B: Graphs CAAR (-30, +30) and AAR (-30, +30) LSE, NASDAQ and NYSE... 41 Appendix C: Other OLS regression results... 42 Appendix D: OLS regression results NASDAQ and LSE... 43

1. Introduction At the end of the 20 th century, share repurchases increased enormously in popularity. In the United States, share repurchases counted for 11.82% of all payouts in 1971, while in 2000 44.42% of all payouts could be subscribed to share repurchases (Dittmar & Dittmar, 2002). In Europe however, share repurchases only became popular at the end of the 20 th century, as it was restricted or prohibited by governments in most European countries (Ginglinger & Hamon, 2007). As share repurchase, particularly the open-market share repurchase method, has become a popular tool for managers to payout excess cash to shareholders, it is interesting to look at the motives and the coherent preferred share price impact. Several studies have been performed on the share price impact of a share repurchases or share repurchase announcements. Most studies that examined the influence of share repurchase announcements have not been performed recently. For example, Vermaelen (1981) looked at the effects share repurchase announcements for the period 1970 to 1978. It is interesting to determine whether the influences of share repurchase announcements on share price for the period 2008 to 2012 are comparable to these previous empirical results. The expectation is that the abnormal returns will still be positive, but somewhat lower as a result of the loss of confidence and the more cautious behavior of investors as an aftermath of the financial crisis that started in 2007. This loss of confidence resulted from banks being rescued by national governments and large financial institutions being on the edge of bankruptcy. The purpose of this thesis is to investigate whether share repurchase announcements influence share prices, focusing on whether they are cause positive abnormal returns. Firms from three different markets, the NASDAQ, NYSE and LSE, will be selected. To examine the influence of the financial crisis on the share price reaction, share repurchase announcements from January 2008 until December 2012 are included. It is expected that the undervaluation of shares and the distribution of excess cash towards shareholders are still important motives for firms to make share repurchase announcements, as these motives may not to be directly related to the financial crisis. The influence of share repurchase announcements on the share price is examined by performing an event study. The estimation window in this study will be 130 to 30 days before the event day (-130, - 30), with a three day event window (-1, +1). In addition to this event window, other windows have been examined; (-3, +3) and (-5, +5). After the impact of share repurchase announcements on the share price has been determined, an Ordinary Least Square (OLS) multivariate regression analysis is performed to examine the factors associated with share price impact. The dependent variable is the cumulative abnormal return for the event windows (-1, +1), (-3, +3) or (-5, +5), and the independent variables include Book-to-Market ratio, Size, Return-on-Equity, Return-on-Assets, and Gross Profit Margin. The main research question of this study is therefore: How are share repurchase announcements affecting the share price and in what way do firm performances influence this share price impact? 1

The independent variables Return-on-Equity, Return-on-Assets, and Gross Profit Margin are examined to determine the influence of firm performances on share price impact. The results indicate how the profitability of a firm, the firm performances, influences the share price impact after a share repurchase announcement. Share repurchase announcements (event days) are collected from the ZEPHYR database. The daily share price returns and other information regarding firm performance are retrieved from the DataStream and Worldscope Fundamentals database. As a benchmark for the market returns, the S&P500 Composite for the NASDAQ and NYSE, and the FTSE All-share Index for the LSE are used, to determine the abnormal return. Initially, 743 share repurchase announcements for 484 firms were collected. It is possible that there are firms with multiple announcements for a share repurchase. When two announcements were followed each other within one month, only the first one was taken into account. Share repurchase announcements in which less than 1% was announced to be repurchased are excluded from the sample, as it was expected that the effect on the share price would be negligible for very small repurchases. After excluding these firms and firms with insufficient data or missing values, the total number of share repurchase announcements over the time period 2008-2012 was 683 for 449 firms. From the event study it turns out that on average, share repurchase announcements had a statistically significant positive abnormal return of 0.61% on the announcement day a statistically significant positive cumulative average abnormal return and over the event window (-1;+1) of 1.13%. This is in line with expected positive reaction of share prices after a share repurchase announcement. Furthermore, the results are also coherent with empirical results from other studies, such as Ikenberry, Lakonishok, and Vermaelen (1995). When looking at firm size and number of shares announced to repurchase, it has been found that on average the share price reaction is more positively for smaller firms and larger share repurchases. When looking at the markets separately, the results showed that the abnormal returns around the event date (-1, +1) for the American markets, NASDAQ and NYSE, and for the British market, the LSE, are positive. The abnormal returns for the NYSE were not statistically significant. However, there has not been found that the cumulative abnormal returns for the LSE are on average negative in the period prior to the announcement. Although the abnormal returns are not negative for the LSE, there is some reaction due to the share repurchase announcement as the abnormal returns were more positive around the announcement date. Therefore, it can be assumed that a share repurchase announcement has a positive effect on the share price. Furthermore, the results are showing that the cumulative abnormal returns for the NASDAQ are on average higher than for the NYSE and the LSE in every event window. Furthermore, a regression analysis was performed to assess whether and how the independent variables mentioned have influenced the share price reaction. The undervaluation hypothesis was tested by looking at the book-to-market ratio and firm size, the log of total assets. Firms with higher book-to-market ratios, so called value firms, are expected to use share repurchases sooner to signal the market that their shares are undervalued. The variable Size was included for testing the 2

information asymmetry between insiders and outsiders. It is expected that bigger firms have more publicly available information, meaning less information asymmetry between insiders and outsiders. Therefore, bigger firms should be better valued than smaller firms. The results show that the coefficients of the variables, that tested the undervaluation hypothesis, were as expected. To test whether the profitability of a firm was related to the abnormal returns the return-on-equity (ROE), the return-on-assets (ROA) and the gross profit margin (GPM) were tested. It is expected that less profitable firms, with lower ROE and ROA, are more likely to use the excess cash to repurchase shares. Therefore, a negative relation between abnormal return and the ROE and ROA was expected. However, when all independent variables were included in the regression analysis with the CAR (-1, +1), the coefficients for ROE and ROA were positive. In the regression analysis with the CAR for the event windows (-3, +3) and (-5, +5), the ROA was negative, while the ROE remained positive. A positive relation between the GPM and abnormal returns was expected. As high GPM is assessed as positive by the market, abnormal returns would be higher if firms with a high GPM announce a share repurchase. As expected, the results show a positive coefficient for the GPM. With exception of the coefficients for ROE, the results for the variables related to firm performance suggest that for less profitable firms the abnormal return is higher, as the market responds more positively towards announcements made by less profitable firms. The most frequently mentioned reason for making a share repurchase is to send a signal towards the market that the shares are not correctly valued or undervalued (Dittmar, 2000; Grullon & Ikenberry, 2000). By providing the market more and potentially new information, the information asymmetry between insiders and outsiders is reduced. When firms give positive information regarding future performances with a share repurchase announcement, this may result in a correction of the share price and firm value according to Vermaelen (1981). The purpose of this study was to examine if share repurchase announcements cause positive share price reactions, and to assess whether and how firm performances are influencing the share price reaction. This study confirms that on average, share prices increased after share repurchase announcements. As the abnormal returns were statistically significant positive. This indicates that share repurchase announcements send a signal towards the market, which is consistent with the results of Vermaelen (1981). The results are somewhat low when compared to other literature where the cumulative abnormal returns for U.S. firms varied from 1.5% till 3.5% for the event window (-1, +1). But this was expected due to the financial crisis. The main conclusion regarding the factors that are associated with the share price impact is that not all results were in line with the expectations. However, it has been found in the regression analysis that undervaluation of shares, tested by the book-to-market ratio and firm size, may still be an important motive for firms to perform a share repurchase announcement. Furthermore, there are a few limitations regarding this study. This due to the fact that there were firms with multiple share repurchase announcements. Therefore, only the first share repurchase announcement was taken when two followed each other within one month. Otherwise both share repurchase announcement were taken. Since most firm data was quarterly based, there is a risk that the data related to these firms is biased. Besides that, there were some firms with not enough data 3

related to the regression analysis. For that reason the regression analysis was performed with fewer firms than the event study. For future research, there are also some suggestions. It might be interesting to look at the share price performance after actual share repurchases, looking at the long-term, and compare these to the returns after share repurchase announcements. Subsequently, the abnormal returns of firms that actually repurchased shares can be compared with the abnormal returns of firms that only made announced a share repurchase, to assess the differences in share price performance on the long run. Furthermore, there can be looked at the differences in motives of these firms. As it is expected that a firm that only announces a share repurchase, has other motives than the firms that actually repurchases its shares. The rest of this study is organized as follows: Chapter 2 will provide relevant background information regarding share repurchases, especially what the different methods, motives and effects are of share repurchases. Furthermore, chapter 2 will convey the literature review, in which empirical results of other papers are summarized, and the hypotheses to be tested with the event study. Chapter 3 and 4 will deal with the data collection, independent variables and their descriptive statistics, and the methodology used in research. In chapter 5 the empirical results from the event study and the regression analysis are presented and finally in chapter 6 the main conclusions, limitations and future research are drawn. 4

2. Theoretical Framework and Hypotheses development Before considering any analytical tests, it is useful to have a better understanding regarding share repurchase announcements and the corresponding abnormal share price performance. Therefore, some theoretical background is given in the next sections regarding share repurchases. Furthermore, this chapter provides an overview of empirical results of other studies concerning share repurchase announcements. In the last section the hypotheses to be tested by an event study, based on previous literature, will be described. 2.1 Share Repurchase In this section the theory regarding share repurchases is described. A look will be given at the increasing popularity of share repurchases as method for the distribution of excess cash or profit towards shareholders, and why it has become a more frequently used method than the payment of dividends. In addition the different methods and motives of share repurchases are described, together with the accompanying effects on share price performance. Furthermore, it is often the case that an announced share repurchase program is not exercised, or not completed within a reasonable timeframe. Since there is no obligation for firms to actually repurchase the shares in the most frequently used method, the open-market share repurchase method. Therefore, it is interesting to look at the effects of share repurchase announcements on the share price. But first, what is a share repurchase exactly? According to De Cesari, Espenlaub, Khurshed, and Simkovic (2012) a share repurchase is the reacquisition of own shares by firms, the distribution of corporate cash towards shareholders in return for a fraction of the outstanding equity. A share repurchase distributes capital back to the shareholders if they are prepared to sell their shares (De Cesari et al., 2012). It reduces the number of shares outstanding on the market, with the consequence that the price of the remaining outstanding shares increases. 2.1.1 Dividends versus share repurchase There are basically two ways for firms to distribute excess capital towards shareholders, the repurchase of shares and the payment of dividends (Dittmar, 2000). Until share repurchases got popular in the 80s, particularly open-market share repurchases, cash dividends were the most frequently used method by firms for distributing excess capital to the shareholders (Grullon & Ikenberry, 2000). Therefore, the differences between the repurchase of shares and dividends will be discussed, and why share repurchases are more frequently used for distributing excess capital to shareholders. One of the reasons why share repurchases are more popular than dividends, is that they are considered as more flexible for firms. This because in an open-market share repurchase, the most frequently used method, firms have no obligation to actually repurchase the shares (Dittmar, 2000). 5

Whereas in a dividend program firms are required to payout a certain amount of dividends to all shareholders, once announced by firms (Thein, 2013). Moreover, there is no expectation that firms repurchase shares on a regular basis, unlike the payment of dividends (Dittmar, 2000). A share repurchase allows firms to be more flexible, while shareholders are more flexible when dividends are paid out 1. According to Grullon and Michaely (2002) share repurchases and dividends are practically economic equivalents, when firms are fairly valued. However, there are two important differences: their tax treatment and the investor s income. The tax provisions for share repurchases appear to be more advantageous than for payment of dividends (Barclay & Smith Jr, 1988), because capital gains are generally lower taxed than dividends. Furthermore, the investors that choose to sell their shares incur only a capital gain tax in case of a share repurchase, while the investors that are holding their shares receive a pro-rata increase in their ownership. Since the number of shares outstanding is reduced after the repurchase of shares. However, the investors have not the obligation to pay the taxes immediately (Grullon & Michaely, 2002). The big difference between dividends and share repurchases is that the shareholders retain their right as the owners of the entity after the payment of dividends, and they can expect the firm to pay dividends for the coming years till liquidation or until they sell their shares (Thein, 2013). Lastly, share repurchases are better than dividends when the shares are undervalued and worse when the shares are overvalued. 2.1.2 Methods of share repurchase As it was explained in the previous section what a share repurchase program is, there will now be taken into account the different methods for performing a share repurchase. It is important to look at these methods, because it makes the rest of the study more understandable. Especially when looking at the effects and consequences of share repurchase on firm value and share prices. For the repurchase of shares there are according to Stephens and Weisbach (1998) three methods that are most frequently used: Fixed-price tender offers, Dutch Auctions and the open-market stock repurchase program (or just repurchase program). The first two methods are most frequently used when the intention is to retire a large amount of shares in a relative short time period. Fixed-price tender offer In a tender-offer firms typically announce the number of shares they wish to repurchase and the price they are prepared to pay for the shares. Furthermore, the time period in which the shareholders can offer and sell their shares is announced by the firms. The duration of the period is typically limited for a tender offer (Grullon & Ikenberry, 2000). When shareholders are not offering the desired amount of shares the firm wants to repurchase, the firm is free to extent the period to 1 Retrieved from the website of Dividend Monk: Disciplined Dividend Stock Research http://dividendmonk.com/ 6

buy additional number of shares. Firms also have the right to change to amount of shares it intends to repurchase, but in general the announced amount of shares to be repurchased is the limit. Moreover, in most tender are managers and directors of a firm prohibited from tendering their own shares. Dutch auction share repurchase An alternative for the fixed-price tender offer is the Dutch auction share repurchase. The big difference according to Bagwell (1992) is that in a fixed-price tender offer firms only use one price, while in a Dutch auction share repurchase firms make use of a range of prices. Firms also have to mention the number of shares they intend to repurchase and the period in which the repurchase will take place. Furthermore, the firm is specifying a price range in which he can repurchase the desired amount of shares from the shareholders and, unlike most other options, the price for which the shares are repurchased is the lowest price at which the shareholders are willing to sell their shares (Bagwell, 1992). The shareholders that have tendered at or below the final price will become part of the repurchase program and receive all the same price (Grullon & Ikenberry, 2000). Open-market share repurchase Lastly there is the open-market share repurchase program, which has become one of the most frequently used methods for repurchasing shares in many countries over the last decade (Ginglinger & Hamon, 2007). Since it is mentioned in the paper of Grullon and Ikenberry (2000) that 91% of the total value of share repurchase announcements in the United States over the period 1980-1999 can be attributed to open-market share repurchases. In an open market share repurchase a firm buys back its own shares in the market over a period that can vary from months to years (Oded, 2005). Firms are, unlike the other commonly used methods, not required to actually repurchase the shares. In general, the repurchase of shares via this method are not fully executed, as an open-market share repurchase costs more time to complete than the other two methods (Michel et al, 2010). Despite the fact that in most instances only a small proportion is actually repurchased by firms, there is according to Oded (2005) empirical evidence that the market is seeing the announcement of these programs as good news. Furthermore, there are according to Barclay and Smith Jr (1988) two fundamental and related issues regarding open-market share repurchases, the fact whether managers use private information to time the share repurchase and whether they influence the firms liquidity. 2.1.3 Motives for share repurchase It is important to look at the motive of a firm to perform a share repurchase, as the effects of share repurchase announcements on the share price are examined in this study. There is however not a single motive for firms to perform a share repurchase, therefore a look will be given to the most important motives according to the existing literature. The majority of these different motives can be classified in three groups according to Oded (2005): taxation, agency problems, and asymmetric information. 7

Dittmar (2000) and Grullon and Ikenberry (2000) described that the decision to perform a share repurchase is influenced by a firm's distribution, investment, capital structure, corporate control and compensation policies. Basically there are five different motives why firms decide to repurchase shares. The most important motives for performing a share repurchase are nevertheless to send signal to the market, with respect to optimistic future firm performances, and the distribution of excess capital to shareholders. These motives will be explained in more detail, as well the other frequently mentioned motives: the prevention of a hostile takeover which has been mentioned to be important by Thein (2013), the optimal leverage hypothesis and the management incentive hypothesis. Excess Capital Hypothesis The first motive to be explained for performing a share repurchase is the availability of excess capital in a firm. When firms have excess capital they can make the decision to either hold the cash within the company or to distribute it to the shareholders (Jensen, 1986). There are two ways for firms to distribute excess capital towards shareholders, repurchasing shares and the payment of dividends (Dittmar, 2000). Share repurchases are nevertheless more preferred by firms, due to the fact that under the open-market share repurchase method firms are not committed to actually repurchase shares. Furthermore, in contrary to paying dividends, it is not expected that firms repurchase shares on a regular basis (Dittmar, 2000). Since firms have the option, and not the obligation, to repurchase shares they have more flexibility. The repurchase of shares is also more preferred due to the personal-tax-rate advantage of capital gains (Dittmar, 2000), as capital gains are generally lower taxed than dividend income. The use of excess capital has also a positive influence on the earnings per share, as long as the number of shares outstanding decreases more than the decrease in earnings (Grullon & Ikenberry, 2000). Undervaluation Hypothesis The second important motive for performing a share repurchase is the undervaluation hypothesis. Firms send signals to investors and the market at the moment it is thought that their shares are undervalued. This hypothesis, examined among others by Dittmar (2000), is based on the fact that there exists an information asymmetry between investors (outsiders) and the management (insiders), which causes that the firm and the shares are not correctly valued by the market. By announcing a share repurchase, the management of a firm is showing their disagreement with the valuation of the shares by the market. According to Vermaelen (1981) and Comment and Jarrell (1991), the positive share price reaction after the share repurchase announcement should bring the share price back at the right level. As a result that the signals, given by a share repurchase announcement, can convey information regarding the firm s expectations on future earnings and cash flows, by giving good news the pricing discrepancy would be eliminated (Grullon & Ikenberry, 2000). However, according to Ikenberry et al. (1995) undervaluation is a more important motive for firms with out-of-favor shares, for which it is likely that the book-to-market ratio is high. Furthermore, Ikenberry and Vermaelen (1996) argue that for small firms, with low market capitalization, the possibility of being undervalued is bigger. This is also concluded by Brockman and Chung (2001), as they mention that repurchasing firms are more likely to have lower share prices, lower market valuation (capitalization) and trading volumes than non-repurchasing firms. 8

Takeover Deterrence Hypothesis The next motive that is declared to be important for performing a share repurchase, is the prevention of a hostile takeover (Dittmar (2000); Thein (2013)). According to Thein (2013), a hostile take-over may be only a cautious anxiety, as it is not necessarily the case that the threat of being targeted is real. D. T. Brown and Ryngaert (1992) and Bagwell (1992) show that shareholder inequalities exists and that there is an upward slope of the supply curve. If the supply curve is upward sloping, the costs for the buying party will increase, when shares are repurchased first. The repurchase is ensuring the buying party a higher price, as the lowest price for which the shares could be repurchased becomes higher (Dittmar, 2000). However, the share price will only be higher for a certain period when the timing is correctly according to Dittmar (2000) and Thein (2013). For that reason a share repurchase can be used by firms as a takeover defense strategy (Bagwell, 1992). Consequently, firms with a higher potential of becoming a takeover target will have more tendency to repurchase shares (Dittmar, 2000). Optimal Leverage Ratio Hypothesis Since firms use share repurchases to distribute excess cash towards shareholders, the amount of equity will decrease and the leverage ratio will consequently be higher (Dittmar, 2000). If it is assumed that an optimal leverage ratio exists, firms may use share repurchases to establish this optimal leverage ratio according to Bagwell and Shoven (1989) and Opler and Titman (1998). Firms are therefore more likely to repurchase shares when the leverage ratio is lower than their optimal leverage ratio. For that reason it can be assumed that the capital structure is partly determining the firm s decision to repurchase shares (Dittmar, 2000). Management Incentive Hypothesis The last important motive is the management incentive hypothesis. The repurchase of shares can be a good alternative for managers to distribute excess cash, as the par value of shares is not influenced negatively. For managers that own stock options it is in particular interesting to maintain the share price, as stock options are not giving managers the right of earning dividends (Dittmar, 2000). This will cause that managers, who own a lot of stock options, are more inclined to repurchase shares than paying dividends according to Dunsby (1995), Jolls (1996) and Fenn and Liang (1998). 2.1.4 Effects of share repurchase announcements The last section will discuss the effects of share repurchase announcements on share price, firm performance and the liquidity of a firm. Where there will be looked at the short and long term effects. In most studies it is concluded that in the short term share repurchase announcements are influencing the share price and firm value. Regarding the long term effect there is more discussion. Ikenberry et al. (1995) conclude that, from the motives to perform a share repurchase, signaling is the most frequently mentioned motive, and that in an efficient market shares are no longer undervalued after the announcement. The results of De Cesari et al. (2012) also suggest that firms announce a share repurchase when shares are undervalued by the market. As investors assess a share repurchase announcement as signal of undervaluation, share prices should increase in the 9

period after the share repurchase announcement. A negative drift was found by Ikenberry et al. (1995) in the period prior to the announcement (-5.52%) and positive drifts after the announcement (5.31%). This is consistent with the signaling hypothesis, and therefore they conclude that it is not always necessary for firms to actually perform a share repurchase. Furthermore, Lie (2005) also found evidence that share prices are influenced by the announcement of a share repurchase program. It is therefore likely that share repurchase announcements convey favorable information regarding the firm performances towards the market. Massa, Rehman, and Vermaelen (2007) suggest that share repurchase announcements, in particular open-market share repurchases, not only contain information about the firms themselves, but that they also contain information about the competitors. This means that when firms send a positive signal to the market regarding their own performances or their future prospects, by announcing a share repurchase, they also send negative signals to the market about their competitors (Massa et al., 2007). According to Lie (2005), the positive effect on share prices and firm performances lasts only for the long term when the shares are actually repurchased, otherwise the effect on performance and share price will only be temporary. Which is in contrast to what was concluded by Ikenberry et al. (1995), that an actual share repurchase is not always necessary. Similarly was concluded by Zhang (2005), around actual share repurchases, as a result that the CAAR (-20, -1) was -1.84% and the CAAR (0, +2) was 0.429%. This suggests that firms repurchase shares after price drops and that the market is responding positively in the short term to actual share repurchases. Firms are also more likely to repurchase shares when the share prices are relatively underperforming the market according to the results of Zhang (2005). However, in the paper of Grullon and Michaely (2002), where was looked at all share repurchases between 1980 and 1997 in the United States, it was found that analysts forecasts of future earnings tend to go down after share repurchase announcements. This is in contradiction to the hypothesis that managers are signaling good news about future earnings and cash flows, via share repurchase announcements. Just as Ikenberry et al. (1995) argued, Zhang (2005) concluded that the market reacts more positively to a share repurchase announcement when they are made by small firms and firms with high bookto-market values ( value firms ). Therefore, the undervaluation motive would be more important for smaller and value firms. This suggests further that positive information regarding the future is more likely to be important for small firms for taking advantage, while it is more likely that large firms make a share repurchase when the share prices have been declined (Zhang, 2005). Furthermore, Ginglinger and Hamon (2007) investigated share repurchases of French firms on their timing and influence on share price and the firm s liquidity. In contrast to Zhang (2005), no significant rises in share prices were found after actual share repurchases. Also the effects of share repurchases on firm liquidity were examined. The conclusion was that when firms make a share repurchase the market's liquidity for the relevant shares is considerably reduced, what is suggesting that the benefits of share repurchases are limited (Ginglinger & Hamon, 2007). Cook, Krigman, and Leach (2004) have mentioned that it can be suggested that open-market share repurchases provide firms an important liquidity advantage. 10

2.2 Literature Review In table 1 the empirical results of other papers have been summarized. Share repurchase have become a popular management device in Europe the last decade, where it was already very popular in the United States since the 1980 s. For that reason it makes sense that most literature existing is mainly focusing on share repurchase announcements made in the United States. Furthermore, most studies focused on the short term share price effect. In this part the empirical results from other papers on the topic of share repurchase announcements will be compared to each other. Table 1 gives an overview of some empirical results, it can be seen that all studies found a positive cumulative abnormal return after share repurchase announcements. For the U.S. the table shows that the cumulative abnormal return (CAR) is varying between 1.5% and 3.5%, for the studies performed on non-u.s. countries the CAR is between 0.3% and 6.0%. Furthermore, the results of the long-term show a much larger CAR than for the short term, with 8.7% and 14.9%. Short-term Table 1: Empirical results of abnormal returns from share repurchase announcements Country of Author(s) of study Sample period Number of Event Window CAR research observations (Days) U.S. Vermealen (1981) 1970-1978 243 (-1, +1) 3.0% Comment and Jarrel (1991) 1985-1988 1197 (-1, +1) 2.3% Ikenberry et al. (1995) 1980-1990 1239 (-1, +1) 3.5% Grullon and Michealy (2002) 1980-1997 4443 (-1, +1) 2.7% Kahle (2002) 1993-1996 712 (-1, +1) 1.5% Lie (2005) 1981-2000 4729 (-1, +1) 3.0% U.K. Rees (1996) 1981-1990 882 (-2, +2) 0.3% Rau and Vermaelen (2002) 1985-1998 126 (-2, +2) 1.1% Canada Ikenberry et al. (2000) 1989-1997 1060 (-15, +15) 0.9% Japan Zhang (2002) 1995-1999 126 (-1, +2) 6.0% Long-term Country of Author(s) of study Sample period Number of Event Window CAR research observations (Years) U.S. Ikenberry et al. (1995) 1980-1990 1239 3 8.7% Canada Ikenberry et al. (2000) 1989-1997 1060 2 14.9% 11

2.3 Hypotheses development In this section the hypotheses, based on previous literature, are described. But first it is important to look again at the main research question: How are share repurchase announcements affecting the share price and in what way do firm performances influence this share price impact? One of the reasons for the management of a firm to perform a share repurchase announcement is that it believes that their shares are undervalued by the market, which results from the fact that not all information regarding the current and future firm performances are available for the market (Grullon & Ikenberry, 2000). With a share repurchase announcement, the firm can give a signal to investors and the market that find their shares are believed to be undervalued. This signal can also provide information about the optimism of the firm regarding the future performances. The signals will therefore have a positive influence on the share price and imply positive returns after the announcement. To assess the influences of a share repurchase announcements on the share price, there will be looked at the abnormal returns of shares prior, during and after the share repurchase announcement. As it is thought that undervaluation of shares is one of the most important reasons for firms to consider a share repurchase program (Ikenberry et al., 1995), this should mean that share prices have been declined in the period prior to the announcement. The financial crisis, which began in 2007 2, may also have caused or at least influenced the undervaluation of shares. Because many firms came into financial trouble, investors lost confidence and as a result became more cautious. It is likely that the American markets are affected more by the financial crisis than the British market, due to the fact that the burst of the housing bubble in the United States in 2006 basically started the financial crisis 3. Therefore it is expected that share prices in the period prior to the announcement have declined more for firms listed on the NASDAQ and the NYSE, and that thus the preannouncement returns would be more negative. The first hypothesis is therefore: H1. The decline of share prices in the period prior to a share repurchase announcement is bigger for firms listed on the NASDAQ and NYSE than for firms listed on the LSE As it is believed that share repurchase announcements convey information regarding the firm s optimism on their future performances, the undervalued shares would be affected positively when firms make share repurchase announcements. The abnormal returns should be somewhat lower than what has been found in previous studies, as a result of the financial crisis. The NASDAQ and NYSE are expected to have higher announcement returns than the LSE, because it is expected that the share prices have been dropped more on these markets. This will lead to the second hypothesis: 2 Retrieved from http://www.nber.org/cycles.html 3 Holt, J. (2009) A Summary of the Primary Causes of the Housing Bubble and the Resulting Credit Crisis: A Non-Technical Paper 12

H2. The effect of a share repurchase announcement is more positively for firms listed on the NASDAQ and NYSE than firms listed on the LSE Subsequently there will also be looked at whether share prices have been increased in the period after the repurchase announcement. As it is assumed by Vermaelen (1981) that share repurchases should correct undervalued shares and firms, it is expected that a share repurchase announcement already may cause an increase and therefore positive correction of the share price. For that reason I assume that the share prices are increasing in the period after the announcement. But due to the financial crisis, the correction in share prices may take longer, as investors became more cautious. As the decline in share prices for the NASDAQ and NYSE is expected to be bigger, the share price reaction should be larger than for the firms on the LSE. The third hypothesis is therefore: H3. The increase of share prices in the period after a share repurchase announcement is bigger for firms listed on the NASDAQ and NYSE than for firms listed on the LSE Furthermore, there could be looked towards the size of firms. As Ikenberry and Vermaelen (1996) concluded that smaller firms, with lower market capitalizations, are more likely to be mispriced by the market. When a small firm makes a share repurchase, the market should react therefore more positive. Zhang (2005) on the other hand, concluded that small firms are more likely to perform a share repurchase for taking advantage of positive information regarding the future, while declining share prices are more likely to be the reason for big firms to repurchase shares. The expectation is however that share price of a smaller firm is affected more by share repurchase announcements, than the share price of a big firm. The fourth hypothesis then formulated as: H4. Share repurchase announcements have a bigger effect on the share prices of small firms For this hypothesis there will be looked at the total assets of a firm at the date of the announcement (t=0). As a reason that there are firms with multiple share repurchase announcement during the period 2008-2012. Besides looking at the firm size, it is also interesting to examine whether the amount of shares to be repurchased is affecting the manner in which share prices are influenced. Zhang (2005) also looked to repurchase size to assess whether there are differences between small and large amounts of shares to be repurchased. From these results it was however not concluded that share prices react differently to large or small amounts of shares to be repurchased. I expect however that when the amount of shares to be repurchased is smaller, in percentage of total share outstanding, the abnormal returns will also be lower. The fifth hypothesis will therefore be noted as: H5. Share repurchase announcements have a bigger effect on share prices when the repurchase size is bigger 13

These hypotheses will be tested by performing an event study. The event study is examining whether and in how share prices are influenced by the announcement of a share repurchase, by looking at the abnormal returns. Furthermore, the influence of firm performances (profitability ratios) on the price impact will be examined, by looking at the Return-on-Equity, Return-on-Assets and the Gross Profit Margin. These and the other independent variables included in the regression analysis are described in section 3.2, where also the expected relation is formulated between the independent variables and the abnormal returns around the event windows (-1, +1), (-3, +3) and (-5, +5). For testing the variables a multivariate regression analysis is performed. 14

3. Data and Descriptive Statistics This study focuses on the share price reaction after a share repurchase announcement and how firm performances influence the share price reaction. This section will describe how the data was collected and how the sample has been cleaned to get the final sample. Furthermore, there will be mentioned which variables are used in the regression analysis, how they have been established and what their expected relation is with the abnormal return. At last the descriptive statistics are shown in section 3.3. 3.1 Data 3.1.1 Data collection The data used for the effect of a share repurchase announcements is obtained from several databases, like Zephyr, DataStream and Worldscope Fundamentals 4. The list extracted from Zephyr comprises share repurchase announcements, dates of completion, industry of the firm and the amount of shares to be repurchased as percentage of the total shares outstanding for firms from the NASDAQ, New York Stock Exchange (NYSE) and the London Stock Exchange (LSE). This for the period April 2008 to December 2012, due to the fact for the calculation of the normal performance needed for the event study. From DataStream data was collected regarding the share prices, market values and total assets for the firms that are listed on one of the three markets. For calculating the market returns for the LSE and the NASDAQ/NYSE data has been obtained from DataStream. For the calculation of the market returns, the price-index of the FTSE All-share Index 5 and the S&P500 Composite 6 has been used as benchmark. In order to measure firm performances, the return on equity (ROE), return on assets (ROA), gross profit margin, book-to-market values and total assets (used as measure for size) have been obtained from Worldscope Fundamentals available in DataStream. The data should be available for the period from January 2008 to December 2012. The firms listed on the LSE are recorded in British pound sterling, and the firms listed on the NASDAQ and NYSE recorded in US dollar. 3.1.2 Data clean up and sample selection At the start of the data collection process it was the intention to use daily data ranging from the 2000 to. As this resulted in an enormous bulk of daily returns, it has been decided to only analyze abnormal returns and firms performances over the period of 2008 to 2012. 4 Available through DataStream 5 The FTSE All-share Index is a capitalization-weighted index, which is consisting of more 2000 companies listed on the London Stock Exchange (LSE). 6 The S&P 500 Composite index is a market capitalization-weighted index, based on the average performance of common stock of the 500 largest firms in the US listed on the NYSE and NASDAQ. 15

Via Zephyr is thus data collected regarding share repurchase announcements. The data has been filtered based on whether the announcement of a share repurchase program is in the period ranging from January 2008 to December 2012. Announcements prior or after this period have been dropped. Furthermore, the dataset has been checked for the size of the share repurchase, size is measured as a percentage of the total shares outstanding. Firms and event dates with an unknown amount or where the amount of shares to be repurchased is less than 1% of the total shares outstanding will be dropped from the sample. There were also some repurchase announcements in which rumors have been expired, thus never an actual announcement has been made by the firm and these are also removed. The amount of shares (%) to be repurchased is used to test whether large repurchases have more effect on share price, or more likely to have influence, then small percentages. Thereafter data from DataStream and WorldScope Fundamentals have been obtained for the firms that had announcements with fully available data and the amount of share to be repurchased is greater than 1%. Daily data regarding the share price, market value and total assets, book-to-market ratio, return on equity, return on assets, gross profit margin, have been obtained for the period January 2008 to December 2012. The data was checked for missing values and whether firms have been merged or acquired, if so the firms have been dropped from the sample and thus also the announcement date. There have been chosen to drop also trust and fund companies from the sample, as these firms do not have the intention in general to repurchase shares with the purpose of improving share prices and firm performances. This resulted in a sample of 484 firms; 49 from the LSE, 193 from the NASDAQ and 242 from the NYSE. For these firms all data is fully available regarding the share prices and the other data for testing the firm performances for the selected period (2008-2012). The total number of share repurchase announcements for the 484 firms is 743, for the LSE 63, the NASDAQ 273 and for the NYSE 407. It is however not always the case that, when an announcement of repurchasing shares has been made, the share repurchase is completed before the end of 2012. As the dataset is now fully controlled for missing values the log-returns of the share prices and the price index of the FTSE All-share Index and the S&P500 Composite can be calculated, which are needed to perform an event study. The log-returns are calculated as follows: ri = log (P t P t 1 ). Where P t and P t-1 are closing values of the firms listed on the NASDAQ, NYSE or LSE, and for the market indexes the closing values of the price index, on days t and t-1. After this the data was ready to be inserted into STATA. After the implementation of the data into STATA the event study can be performed for determining the abnormal returns. Firms and events with less than 100 days in the estimation window will be left out of the sample. The final sample is then at last consisting, after dropping firms with too less data in the estimation window, of 449 firms with 683 events. The final sample to be used in the event study is given in table 1, which is showing besides the number of firms and share repurchase announcements per market for the period 2008-2012, also the size of the firms is presented in the table. 16

Table 2: Summary statistics of firms listed on the NASDAQ, NYSE and LSE Market Number Repurchase Size Tertile of Firms announcements 1 (small) 2 3 (large) NASDAQ 179 252 209 34 9 NYSE 225 372 331 31 10 LSE 45 59 54 3 2 Total 449 683 594 68 21 The size (total assets) textile rankings are determined relative to all firms listed on the NASDAQ, NYSE or LSE that are included in the final sample, where for firm total assets the average has been taken over the years 2008 to 2012. It can be seen in the table that the NASDAQ and NYSE, have far more repurchase announcements than the LSE. Furthermore, the table shows that the majority of firms that repurchase shares are falling into the first tertile. In table 2 the share repurchase announcements are summarized based on their size, as a percentage of total shares outstanding. In the table it can be seen that the most common amount of shares announced to be repurchased is more than 5%, but less than 10%. It can also be seen that in 75% of the share repurchase programs announced, the amount is less than 10% of total shares outstanding. Year N Table 3: Share Repurchase announcements from 2008 to 2012 Percentage of shares announced to be repurchased 1-5% 5-10% 10-15% 15-20% above 20% NASDAQ NYSE LSE NASDAQ NYSE LSE NASDAQ NYSE LSE NASDAQ NYSE LSE NASDAQ NYSE LSE 2008 91 5 19 1 7 17 2 12 9 1 7 3 0 6 1 1 2009 81 9 18 1 7 10 0 12 7 10 1 2 1 2 1 0 2010 150 15 35 1 14 29 2 10 13 12 4 3 0 7 4 1 2011 220 25 42 3 25 47 4 17 23 8 4 8 3 6 5 0 2012 141 14 20 2 23 34 1 12 13 4 5 4 0 3 5 1 68 134 8 76 137 9 63 65 35 21 20 4 24 16 3 Total 683 210 222 163 45 43 This table provides an overview of the sample, consisting of 683 share repurchase announcements from the Zephyr database. In the table there can be seen how share repurchases are divided over the years 2008-2012 and the amount of shares to be repurchased in percentage of total shares outstanding. 3.2 Independent variables for regression analysis In this section the independent variables that will be used in the second part of the study will be described. These variables are used for testing whether and how the cumulative abnormal returns are influenced by firm performances. The description of these variables is based on definitions from DataStream and Investopedia 7. Data necessary for determining the different independent variables have been collected from DataStream and WorldScope Fundamentals. The most important motives and the corresponding variables, from those described in section 2.1.3, will be examined for 7 Investopedia.com is a premiere resource for investing education, personal finance, market analysis and free trading simulators. 17