The announcement effect of share repurchases on the stock price of publicly listed U.S. firms

Bachelor Thesis

Economics and Business

Specialisation in Finance

Share repurchases

The announcement effect of share repurchases on

the stock price of publicly listed U.S. firms

Tim de Boer

11202351

University of Amsterdam

Faculty of Economics and Business

Solomon G. Zori

30-06-20

2

Abstract

After the enactment of the Tax Cut and Job act, share repurchases increased substantially in the United States in the year onwards. As literature mostly report statistically significant abnormal returns following the announcement, in this thesis the abnormal returns are examined during the one-day announcement period. Share repurchase announcements made in the years 2017 to 2019 are used to investigate the movement in stock prices as a result from the announcement effect. I have found insignificant effects on stock prices following the announcement. This result suggests that the impact of the repurchase announcement on a firm’s share price is null. However, for future research more complex models might capture the effect more accurate and yield more precise results.

3

Table of contents

Introduction ... 4

Literature review ... 6

Hypotheses ... 10

Data description and Methodology ... 10

Results ... 15

Conclusion ... 18

Discussion ... 19

References ... 21

4

I. Introduction

On December 22, 2017 Donald Trump signed the Tax Cuts and Job act (TCJA) which is the largest gross tax cut in American history. The Tax Cut and Job act has two key provisions: (i) the reduction in corporate income tax rate from 35% to 21%, and (ii) a one-time cut in repatriation tax from foreign subsidiaries from 35% to 15.5%. Before the signing of the TCJA, companies kept their profits abroad because these profits were only taxed until repatriated. With the one-time reduction in repatriation tax, pre-2018 foreign profits were taxed at a rate of 15.5%.

Furthermore, the US corporate tax system also altered from a worldwide tax system to a territorial tax system in which foreign profits are exempted from domestic taxation. Thus, the Republican tax reduced the incentive for US corporations to hold cash overseas. The reduction in both the income- and repatriation tax rate created a cash windfall for corporates in the United States after the enactment of the TCJA in 2018.

What were the effects of these tax cuts on US corporations? The White House’s portrayal of low corporate taxes was to stimulate capital spending and invest in the long-term growth. Critics, however, have argued that the tax cuts led to an increase in repurchasing stock by US corporations as a result of the cash windfall. This was based in part on a $800 billion record in share repurchases on December 2018. In the first three quarters of 2018 U.S. stock buybacks were 52.6% higher than in 2017. The stock repurchase continued with total repurchases in 2019 being just below 2018’s record in July. A share repurchase refers to the re-acquisition by a company of its own stock from shareholders. Since the mid-1980’s share repurchase activity by U.S. firms have risen. Previous literature has investigated the short-term effect of stock prices following an open-market share repurchase announcement. Stock performance is investigated by testing the significance of the abnormal returns. Plenty studies find a significant positive abnormal return different from zero, which imply a positive stock price reaction following a repurchase announcement.

This paper will therefore focus on the potential effect of the record-breaking share repurchases by U.S. corporations on stock returns caused by the signing of the TCJA. The stock performance after the initial repurchase announcement is examined by using an event study. The abnormal returns are determined on the day the repurchase announcement is made by a firm. Announcements made in the years from 2017 to 2019 are examined to investigate the impact of the TCJA on the number of repurchase announcements and the stock price reaction. A control sample consisting of repurchasing firms is created to examine the stock price reaction of repurchasing firms relative to non-repurchasing firms to test the significance of abnormal returns and the impact of repurchase activity on returns.

Corporations have plenty motives to announce a repurchase program. The most prevalent explanation mentioned in academic literature is the undervaluation hypothesis, also called signaling. Other explanations behind share repurchases are the free cash-flow-, optimal leverage ratio,

5 management incentive- and dividend hypothesis. The signaling hypothesis is also one of the main drivers of the significant positive stock price reaction following the announcement. This hypothesis implies that a firm’s management sends a signal to the market about their future prospects which are inaccurately valued in the stock price. Misvaluation of the share price is caused by information asymmetry between a firm’s management and investors (Dittmar, 2000). Research by Ikenberry, Lakonishok and Vermaelen (1995) finds that the market reaction in the two-day announcement period is approximately 3%. They also find strong evidence for the undervaluation hypothesis. However, due to increasing transparency, the premise of information asymmetry between a firm’s managers and investors has decreased. Furthermore, the information content of the signal can be ambiguous (Vermaelen, 1981). A company can either experience a lack in growth opportunities or, as previously described, managers may perceive their stock as undervalued.

Furthermore, recent research by Bennett, Thakor and Wang (2019) finds that a cut in the repatriation tax encourages firms to repurchase its own stock. This evidence implies that repurchase activity is mainly encouraged for firms with high foreign profits. They find that the stock buybacks by an average firm almost doubles. This is in line with the earlier mentioned free cash-flow hypothesis. This hypothesis implies that firms use share repurchases to distribute excess cash. Thus, the free cash-flow is of great relevance for this research.

As the enactment of TCJA resulted in reduction of tax rates, U.S. firms’ cash flows increased substantially after signing. The portrayal in which companies used the cash windfall to invest in long-term growth, employment and R&D resulted, however, in stock buybacks by publicly listed U.S. firms. Previous research by Blouin and Krull (2009), Dharmapala, Foley and Forbes (2011) and Bennett et al. (2019) find that a cut in the repatriation tax encourages firms to repurchase its own stock. Thus, in line with findings of previous literature, this paper also investigates firm-specific financial data to examine the main drivers of repurchasing firms and the relationship with the stock price reaction after an announcement. The research question of this paper is therefore: What is the short-term effect of

open-market share repurchase announcements on the stock price of publicly listed U.S. firms between 2017 and 2019, and can the cumulative abnormal returns be explained by the announcement effect of a share repurchase? I find that the cumulative abnormal returns are only significant in the event window [0, 0]

with a mean return equal to 0.4419%. Furthermore, within the OLS framework the estimate on the announcement effect is not statistically significant.

Firstly, the literature review section will discuss the different methods a firm can use to repurchase its own shares followed by previous evidence in academic literature on the stock price reaction after the initial announcement and the list of motives behind stock buybacks. Subsequently, the methodology and data section deals with data collection and the empirical framework that is used in this paper. It continues by addressing the hypotheses based on previous studies. The results section will describe the main findings of this research and review the stated hypotheses. Finally, the conclusion is formulated and suggestions for future research will be discussed.

6

II. Literature review

This section will provide literal background for understanding the analysis on share repurchases. It will focus on the history of open-market share repurchases, the motives behind share repurchases and stock price performance following an announcement by using academic literature. There is plenty academic literature about the stock price performance after a share repurchase announcement. Most studies report significant abnormal returns, in both the short- as well as the long-term event windows.

A share repurchase is a decision by a firm’s management to acquire its own stock on the marketplace. The most common methods to repurchase shares are in the open-market, through a negotiated transaction, also called a dutch-auction, or using a fixed-price tender-offer. With a traditional tender-offer a single purchase price, a number of shares sought and an expiration date is specified in advance by the offering firm (Stonham, 2002). Thereafter, company shareholders can decide the number of shares they are willing to tender. The price offered by the repurchasing firm is usually above the market price so shareholders have a stronger incentive to sell their shares (Stonham, 2002). Furthermore, the maximum number of shares firms promise to repurchase is equal to the amount initially specified in the offer. If the number of shares investors are willing to sell exceeds the shares initially sought by the firm, the stock tendered is repurchased proportionally from each shareholder (Dann, 1981). The offering firm has the right to extend the repurchase beyond the announced expiration date, mostly common if the number of shares tendered by investors is less than sought by the firm.

Compared to a fixed-price offer, in a Dutch-auction the number of shares sought is also specified in advance, however, the repurchasing firm sets a range of prices within each shareholder can choose at which minimum price they are willing to sell their own shares. Subsequently, the offering firms takes in all bids to determine the highest price at which all shares tendered can be bought. This price is then paid to all investors who tendered their stock at a price equal to, or lower than, the determined price (Comment and Jarrell, 1991). If the amount of shares tendered by company shareholders is more than the number of shares initially sought, the Dutch-auction is similar to an over-subscribed fixed-price offer.

In an open-market share repurchase, firms buyback their own shares in the market. Unlike the other two forms of share repurchases, open-market repurchases do no obligate firms to actually buyback their own shares. In fact, Stephens and Weisbach (1998) find that the number of shares actually acquired are only a fraction of the shares initially sought or firms do not even complete the repurchase at all. Besides the number of shares, the actual dollar value of the program is also uncertain since there is no obligation to repurchase. Furthermore, Cook, Krigman and Leach (2004) find that U.S. firms acquire their shares at a discount in the market using open-market repurchases, while a higher price relative to the market price is offered in a fixed-price tender-offer.

Over the last decades, open-market share repurchase activity have become common for companies in the United States. Stock repurchases by U.S. firms have grown in popularity as

7 expenditures on share repurchase programs (relative to total earnings) rose from 4.8% in 1980 to 41.8% in 2000 (Grullon and Michaely, 2002). In the following years open-market share repurchases of U.S. public firms have grown further. Resultingly, as stated by Farre-Mensa, Michaely and Schmalz (2014), the fraction of publicly listed firms repurchasing their own stock peaked in 2012, with more than 45% of U.S. firms engaging in a stock buyback. Figure 1 shows the relation between the number of firms contributing in pay-out policies and the different pay-out methods used by publicly listed firms. It demonstrates the increase in share repurchases over the last decades, which resulted in share repurchase activity surpassing dividend payments in 1997.

After the rise in stock buyback activity, academic literature studied the price behaviour of firms’ shares after a repurchase announcement. Most studies find strong evidence on significant abnormal returns following the announcement. Research by Ikenberry et al. (1995) report an average market reaction of 3.5% over a two-day announcement period. They find that firms primarily used share repurchases for information signalling. However, as stated by Ikenberry et al. (1995), new information provided to the market is not fully incorporated. Resultingly, they find an average abnormal buy-and-hold return of 12.1% in the four year following the initial announcement. The long-term stock performance, however, depends on market level and is highest for firms with high book-to-market ratio.

A firm’s decision to repurchase their own shares is affected by many factors. The most common motives to buy back stock include the undervaluation-, free cash-flow, optimal leverage ratio-, management incentive- and takeover deterrence hypothesis (Dittmar, 2000). Although all these motives are plausible, any literature concerning share repurchase report that the undervaluation hypothesis is one of the most prevalent explanations behind share repurchases. Previous research by Dann (1981), Ikenberry et al. (1995), Stephens and Weisbach (1998) and Chan, Ikenberry and Lee (2004) find strong evidence for this hypothesis.

Figure 1 Pay out policies by firms in last decades

8 The undervaluation hypothesis is the most commonly used motive why managers decide to repurchase its own stock. The hypothesis states that information asymmetry arises between a firm and the market resulting in company’s stock price being misvalued. When a firm announce an open-market share repurchase, this implies that the value of their stock is trading below its value. If perceived undervaluation is indeed a motivation for firms to repurchase, the perception might thus be based on management’s privately held information. However, as stated by Vermaelen (1981), the information signal provided by a firm might be ambiguous because a company might also experience a lack in growth opportunities. Previous research by Leng and Noronha (2013), however, find that firms indeed provide private information to the market by using share repurchase announcements. Furthermore, repurchase activities and prior stock return are negatively correlated, suggesting that firms announce buybacks if perceived undervaluation arises (Stephens and Weisbach, 1998).

As earlier described, stock price reaction is differently impacted based on a firm’s book-to-market level. Firm-specific characteristics impact the returns following an announcement and therefore indicating whether a firm is more misvalued relative to other firms. Dittmar (2000) finds that the median book-to-market ratio is significantly higher for repurchasing firms relative to a non-repurchasing sample, indicating that repurchasing firms are more likely to be misvalued. Furthermore, smaller sized firms are more likely to be misvalued because they are less covered by analytics compared to larger companies (Zhang, 2002).

The free cash-flow hypothesis is another reason for firms to announce a share repurchase. When a company’s free cash flow exceeds the funds needed to finance positive NPV projects, managers might have the incentive to waste the excess cash in unprofitable investments (Jensen, 1986). Stephens and Weisbach (1998) find a positive relation between the level of cash flows and repurchases. Thus, to reduce the agency costs of free cash-flows, firms decide to distribute the excess cash to company shareholders. Firms therefore mitigate potential over-investment by management. The market reaction to share repurchase announcement is more positive among firms that are more likely to overinvest (Grullon and Michaely, 2004).

The free cash-flow hypothesis is of interest in this study as the enactment of the TCJA resulted in substantially higher cash flows for U.S. firms. The rise in share repurchase activity in the U.S. in 2018 might be an indication of the free cash-flow hypothesis. Bennett, Thakor and Wang (2017) find that the main driver of surge in repurchases was the repatriation tax cut. Firms with high foreign profits repurchase relatively more shares than low foreign profit firms. Thus, the free cash-flow hypothesis might be of great relevance in this study.

Another plausible motive for firms to announce an open-market share repurchase is the optimal leverage ratio hypothesis. When firms distribute excess cash to shareholders by using repurchases, a firm’s equity will be reduced resulting in a higher leverage ratio. Companies may use repurchases to alter the capital structure to achieve an optimal leverage ratio (Dittmar, 2000). Thus, when a firm’s leverage ratio is below their optimal level, it is more likely that the firm will repurchase its own stock.

9 Firms can also use debt to fund the repurchase, resulting in an increase in debt and a decrease in equity. However, as stated by Bennett et al. (2017), the reduction in income-tax rates decreases the benefit of borrowing which makes it less attractive for companies to fund repurchases with debt.

Previous research by Dittmar (2000) finds that leverage ratio relative to a target has significant, but, minor effects on a firm’s decision to announce an open-market share repurchase. The leverage ratio of repurchasing firms is mostly less than those of non-repurchasing firms. This hypothesis, however, might be less relevant in this study as Bennett et al. (2017) states that low-leverage firms increase investments after the income-tax cut resulting from the TCJA. However, a firm without or with less investment opportunities may use repurchases either as a tool to de-leverage to achieve an optimal leverage ratio or to distribute excess cash to shareholders.

Another motivation on share repurchases with less relevance for this study is the management incentive hypothesis. When a firm compensates the executives with a large number of stock options, managers may find it more attractive to use share repurchases because this does not result in dilution of the per-share value of stock. Therefore, stock options encourage managers to use shares repurchases instead of dividends. Dittmar (2000) finds that the use of stock options by firms significantly increases the probability of a firm to repurchase its own stock.

Finally, the takeover deterrence hypothesis states that a firm may use a share repurchase as a takeover defence. Stock buyback increases the debt-to-equity ratio and decreases the number of shares outstanding, resulting in a takeover being less valuable for other firms. Bagwell (1991) shows shareholders’ valuation can differ and that the supply curve of corporate equity is upward-sloping. Thus, a company that is a potential target can increase the cost of the takeover by repurchasing stock. Previous research by Billet and Xue (2007) finds a positive relation between takeover probability of a firm and its repurchase activity, which implies that firms may increase repurchases when a firm is a potential target.

Most academic literature focus on stock price behaviour following a share repurchase announcement on either a short- or long-term period. However, as this paper also focus on stock price performance, it differs from the existing literature in that it compares the stock performance of repurchasing firms relative to a non-repurchasing sample. The research sample of previous studies mainly consists of repurchasing firms only to examine the price impact on their stock. In this research, the announcement effect of a share repurchase on the stock price is captured. Second, as Bennett et al. (2017) find a significant increase in share repurchases after the enactment of the TCJA, this paper contributes to existing literature because share repurchases are examined prior to and following a period with substantial tax cuts resulting in greater buyback activity. Therefore, this research aims to contribute to existing academic literature on the stock price performance of repurchasing firms in order to get an clear picture of whether the abnormal returns are of significance.

10

III. Hypotheses

As described in the introduction, this research aims to measure the stock price reaction on securities after an initial share repurchase announcement. As the evidence on the price impact in academic literature is rich, implying positive abnormal returns, my hypothesis will hence be:

𝐻0: 𝐶𝐴𝑅 = 0

In case of the alternative hypothesis, the 𝐶𝐴𝑅 > 0 indicates a positive relationship between the abnormal returns and the announcement of a repurchase program. Thereby, the null-hypothesis is statistically insignificant.

Furthermore, the announcement effect of repurchases on stock price performance is investigated. As such, based on academic literature regarding stock price reaction after the initial announcement, my hypothesis will be:

𝐻0: 𝑅𝑒𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒 = 0

IV. Data description and methodology

In this research the stock price reaction after an open-market share repurchase announcement in the United States is examined. Besides the share price reaction, it also investigates the differences in firm characteristics between repurchasing and non-repurchasing firms. The research sample used consists of share repurchase announcements made by firms in the years January 1, 2017 to December 31, 2019. This period is taken with the intention to capture the impact of the TCJA on share repurchases and the following stock price reaction. The data for all the U.S. repurchasing firms is drawn from Zephyr. The Zephyr database contains data for all U.S. share repurchases, but also has information on Merger & Acquisitions, IPO’s, etc..

To investigate the movement in stock price following a repurchase announcement, a sample is created consisting of both repurchasing- and non-repurchasing firms. Following Leng and Noronha (2013), non-repurchasing firms are matched to a repurchasing firms based on the total asset size in the fiscal year-end prior to the announcement. First, a 1% range of a repurchasing firm’s total assets is used to determine matching non-repurchasing firms. A non-repurchasing firm’s total asset size that match a repurchasing firm’s asset size most closely is then added to the sample. If there are no matching firms within the 1% range, it is expanded to 2% and so forth. The only requirement for non-repurchasing firms is that the firm did not announce any share repurchase programs in the same year as and three years prior to the repurchase announcement made by a sample firm.

11 The initial repurchasing sample consists of 411 open-market share repurchase announcement made by U.S. publicly listed firms in the research period. First, only the repurchasing firms listed on either the NASDAQ or New York Stock Exchange (NYSE) are included in the sample. Firms that are listed on the OTC Bulletin Board are excluded because the share prices of those firms are usually below a threshold of $3 per share or a firm’s financial information is missing. Furthermore, announcements that were made by firms that were delisted or unlisted during the research are excluded from the sample. The control sample is drawn from the Orbis database between the period January 1, 2014 to December 31, 2019. The initial control sample consists of approximately 15,000 firms, however, after selecting only the non-repurchasing firms for each three years based on the conditions mentioned above, about 2,500 firms are left. Subsequently, all control firms are matched to a repurchasing firms which result in the same number of non-repurchasing firms relative to repurchasing firms.

Besides stock market requirements, all repurchasing and non-repurchasing firms without CRSP or Compustat data are eliminated from either sample. The CRSP database mainly contains information on stock prices. The Compustat database contains data on the financial, statistical and market information of global firms. After elimination, both samples contain 278 firms, either repurchase- or non-repurchase firms.

Both data sources, CRSP and Compustat, are used to gather all the stock and financial data of both the event- and control firms. The stock data, containing the abnormal and cumulative abnormal returns, are calculated by the WRDS Event Study itself. The following firm-specific data used in the regression is collected by using Compustat: Market-To-Book ratio (MTB), Operating Net Cash Flow, Leverage, Size, Return on Assets (ROA), Return on Equity (ROE) and Dividend. The logarithm of the total assets is used as a proxy for firm size. The descriptive statistics of both the repurchasing- and non-repurchasing sample can found in Table 1.

Table 1 Sample Distribution

This table reports the distribution by year for the sample firms that announced open-market share repurchases.

Year No. of obs. Fraction of sample (%) No. of completed repurchases Value of repurchases (in $ millions) Mean value of repurchases (in $ million) Shares sought (%) Total assets (in $ thousand)

Mean Median Mean Median

2017 25 17.99 20 71.297 1.517 6.62 7.03 9,641.62 2,373 2018 83 59.71 27 397.788 3.825 8.44 6.80 32,592.32 8,149 2019 31 22.30 8 77.940 1.694 8.41 6.74 20,404.69 5,823 All years 139 100.00 231 547.025 7.036 7.82 6.86 20,879.54 5,448.33

12

Methodology Event studies

An event study can be used to measure the impact of a specific event on the value of a firm using firm financial data. To examine the stock price reaction after an open-market share repurchase, the Event Study in WRDS is used to calculate the cumulative abnormal returns. Subsequently, the cumulative abnormal returns are computed over the following event windows: [0, 0], [-1, +1], [-5, +5], [-10, +10] and [+2, +10].

At first, the normal return needs to be calculated to determine the abnormal returns. The normal returns are calculated by using the Market Model method. According to this model, a linear relationship exists between a firm’s share price and the market return. The model uses the following formula:

𝑅𝑖,𝑡 = 𝛼𝑖+ 𝛽𝑖 𝑅𝑚𝑡+ 𝜀𝑖

Where 𝑅𝑖,𝑡 represents the return on stock of firm i in a specific time period t. The 𝑅𝑚𝑡 is the return on

the market portfolio in time period t. 𝛽𝑖 refers to firm-specific risk, which is also captured by the model.

Finally, 𝜀𝑖 is the error of stock i.

Thereafter, the WRDS Event Study calculates the abnormal returns for each event window used in this research. The abnormal returns are computed on a daily basis where the length of days depends on the event window. The following formula is used to determine the abnormal returns:

𝐴𝑅𝑖,𝑡 = 𝑅𝑖,𝑡− 𝐸(𝑅𝑖,𝑡)

In the formula 𝐴𝑅𝑖,𝑡 represents the abnormal returns on stock i in a pre-defined event window. The 𝑅𝑖,𝑡

is the actual return whereas 𝐸(𝑅𝑖,𝑡) is the expected return on stock i in time period t. Followed by the

abnormal returns, the CARs are calculated by adding up the daily abnormal returns. MacKinlay (1997) uses the following formula:

𝐶𝐴𝑅𝑖(𝜏1, 𝜏2) = ∑ 𝐴𝑅𝑖 𝜏₂ 𝜏= 𝜏₁

Where 𝜏1 and 𝜏2 are the first- and last day of the event window respectively. Finally, to test whether

the cumulative abnormal returns are significantly different from zero following a share repurchase announcement, the t-statistics are determined to test the hypotheses.

13

Regression analysis

In this research an OLS framework is used to examine the cumulative abnormal returns (CARs) and the different firm characteristics affecting the CARs. As I am interested in the announcement effect of an open-market share repurchase announcement, the variable of interest will be the Repurchase coefficient. This will be formalised by means of a dummy variable named Repurchase which is the announcement of a share repurchase program.

After determining the CARs, the regression analysis is used to research the relationship between the cumulative abnormal returns and the announcement effect of repurchases. The regression model is in the following form:

𝐶𝐴𝑅

= 𝛼 + 𝛽

𝐵𝑜𝑜𝑘𝑀𝑎𝑟𝑘𝑒𝑡 + 𝛽

𝑂𝑝𝑒𝑟𝐶𝐹/𝐴𝑠𝑠𝑒𝑡𝑠 + 𝛽

𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 + 𝛽

𝐹𝑖𝑟𝑚𝑠𝑖𝑧𝑒

+ 𝛽

𝑅𝑂𝐸 + 𝛽

𝑅𝑂𝐴 + 𝛽

𝑆ℎ𝑟𝑆𝑜𝑢𝑔ℎ𝑡 + 𝛽

𝑃𝑟𝑒𝑅𝑒𝑡𝑢𝑟𝑛 + 𝛽

𝐷𝑖𝑣𝑖𝑑𝑒𝑛𝑑

+ 𝛽

𝑅𝑒𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑒 + 𝜀

The following Table displays the variables used in the regression with corresponding meaning, measurement and data source:

Table 2 Variables explanation

Variable Meaning Measurement Data source CAR A firm’s i cumulative

abnormal return.

The event window used is [0, 0] as this window only shows significant CARs. Four other event windows with corresponding CARs are shown in the result section.

The Event study on WRDS calculates the CARs in pre-determined event window.

BTM A sample firm’s

book-to-market ratio (BTM).

This ratio is calculated by dividing a firm’s book value at the previous fiscal year-end prior to the announcement by the closing price of a firm’s stock.

Compustat

OperCF/Assets The net cash flow of operating activies scaled by a firm’s total assets.

Net cash flows of operating activities divided by total assets.

Compustat

Leverage A ratio that measures how much capital comes in the form of debt.

A firm’s total liabilities is divided by total assets.

Compustat

Firmsize Variable to control for the difference in size between sample firms.

A firm’s size is determined by taking

14 the natural logarithm

of the total assets.

ROE Formula to calculate the return on equity.

The return on equity is calculated by dividing the net income by the market value of stock.

Compustat

ROA Formula to calculate the return on assets.

The return on assets is calculated by dividing the net income by total assets.

Compustat

ShrSought The fraction of shares sought relative to total share outstanding in share repurchase program. The percentage of shares sought is determined by dividing the number of shares sought by total shares outstanding.

Zephyr

PreReturn A firm’s return prior to the announcement in the event window [-20, -2].

The pre-return is calculated by using the Event Study on WRDS.

CRSP

Dividend Dividend is a dummy variable that takes value one if firm conducts a dividend payment and zero otherwise.

Dummy variable takes value 1 if firm’s dividend on common stock is greater than 0 in the year of the announcement.

Compustat

Repurchase Repurchase is a dummy variable that indicates either value 0 or 1, 0 indicating a firm not announce a share repurchase program and 1 otherwise.

For firms in sample that announced share repurchase program, the repurchase variables takes value 1.

Compustat

Before running the analysis, the OLS assumptions are tested as potential complications may arise. First, the normality of the residuals is tested. The unstandardized residuals are used to test for normality. All tests show evidence for normality in the residuals. Besides, the histogram displays the normal distribution of the residuals. Furthermore, the linearity in parameters is tested by using scatterplots. Resultingly, all parameters in the model show linearity.

Multicollinearity can arise when an independent variable can be linearly predicted by one or more other variables in the regression model. In the Appendix section, the correlation matrix displays some variables that are significantly correlated. This might result in insignificant coefficients after running the regression analysis. Another measure to test for multicollinearity in the model is called VIF. When the value of the VIF is higher than 10 it indicates for collinearity. However, as the VIF table displays, all values for the variables are below 2, which indicates no multicollinearity in the model.

15 over an amount of time. In this case, the different values the Repurchase variable might indicate, 0 or 1, show different variances. Therefore, in this research a robust regression is used allowing for heteroscedasticity. Another measure to test for heteroscedasticity is the Breusch-Pagan / Cook-Weisberg test, shown in the Appendix. The null hypothesis of this test, indicating constant variances, is not rejected within the model. Therefore, the problem of heteroscedasticity does not arise in this research.

V. Results

This section will analyse and discuss the results of both the event studies and OLS estimates. First, the CARs in five different event windows are exhibited. Second, the regression analysis section will briefly discuss the results and examine a potential relationship between the cumulative abnormal returns and firm-financial data.

Event studies

The event windows [0, 0], [-1,+1], [-5+5], [-10, +10] and [+2, +10] are used to calculate the cumulative abnormal returns prior to and following an open-market repurchase announcement. The total sample consisting of both repurchasing- and non-repurchasing firms contains 278 observations and is used to determine the stock price reaction. The market model that is based on CAPM is used.

Based on academic literature, we expect the CARs to show a positive reaction after a repurchase announcement is made. Keeping all variables constant for non-repurchasing firms, the CARs should exhibit a flat curve. Table 3 presents the CARs for the previously described event windows. As Table 3 shows, all CARs calculated in the last four event windows are not statistically significant at a 10% level. Only the event window [0, 0] shows significantly positive abnormal returns at 10% level with a mean cumulative abnormal return of 0.442%. The p-value of 0.05109 rejects the null hypothesis that CARs are not different from zero. Although the event window [-5, +5] has the highest calculated cumulative abnormal return, it still indicates no statistical significance with a reported p-value of 0.1968. The results indicate that repurchasing firms only experience a positive stock price reaction in the one-day event window of the announcement date.

Table 3 CARs calculated using five different event windows.

Event window CARs P-value t statistic

[0, 0] 0.4419% 0.05109* 1.63981*

[-1, +1] 0.2085% 0.39023 0.27899

[-5, +5] 0.5105% 0.19684 0.85431

16

[+2, +10] -0.1830% 0.35559 -0.37066

* Significant at 0.1 level ** Significant at 0.05 level *** Significant at 0.01 level

Figure 2 shows the calculated CARs in the event windows [0, 0], [-1, +1], [-5, +5], [-10, +10], [+2, +10]. As presented in Figure 2, the CARs are negative and slightly decreasing prior to the announcement date, however, increase onwards. The evidence of decreasing stock prices prior to the announcement is in line with the undervaluation hypothesis because firms perceive their shares as undervalued and decide to announce a repurchase program. However, the stock price reaction is largest during the announcement date and one day afterwards. Thereafter the CARs stagnate and slightly decrease. Although the stock prices increase after the announcement date the positive short-term effect is only significant for the event window [0, 0]. This result might imply that the information asymmetry between managers and investors relatively decreased compared to previous studies. A firm’s motive to use share repurchases as a signal to the stock market might be less prominent in this sample.

Regression analysis

The results of the regression analysis on the CARs following the repurchase announcement are presented in Table 4. The independent variable is placed above the findings in the second and third column. The regression result with standard errors are shown below (1) and the robust regression results beneath (2). The coefficients of the explanatory variables are displayed as well as the corresponding t-statistics indicating whether values are significant. Furthermore, both the R-squared and the number of observations are displayed for both regression results.

17

Table 4 Regression on CARs

Variables Coefficients Coefficients (1) CAR [0, 0] (2) CAR[0, 0] BookMarket -.0023576 (-0.54) -.0023576 (-0.49) Leverage .0001107 (1.62) .0001107*** (3.12) Size -.0023132 (-1.01) -.0023132 (-1.03) OperCF/Assets .0689281*** (2.88) .0689281*** (2.73) ROE -.0089993 (-0.79) -.0089993 (-0.85) Prereturn -.019275 (-1.38) -.019275 (-1.16) Shrsought .0642212** (2.03) .0642212* (1.80) ROA -.0549796*** (-3.78) -.0549796*** (-8.83) Dividend -.0018558 (-0.49) -.0018558 (-0.43) Repurchase -.0041791 (-0.95) -.0041791 (-0.90) _cons .0107313 (1.19) .0107313 (1.22) R-squared 0.1460 0.1460 N 258 258 t statistics in parentheses * Signaficance at 0.1 level ** Significance at 0.05 level *** Significance at 0.01 level

As presented in Table 4, the variable of interest in this study, the coefficient of the Repurchase dummy, shows no significant evidence in both standard and robust regression results at the 10% level. This result suggests no or a small impact on a firm’s stock price due to a share repurchase announcement. The coefficient, however, is similar in both regression results which confirms the absence of heteroscedasticity. The value of the Repurchase coefficient is equal to -.00418 indicating that the cumulative abnormal returns are decreasing with -.00418% after a firm announced a share repurchase program. However, as the reported p-value is equal to 0.369 the null-hypothesis is not rejected and, resultingly, findings are not concluding. Finally, the R-squared, the proportion of the variance that is explained by the explanatory variables in the model, is 0.1460 in both cases.

Regarding the other explanatory variables, some coefficients show a significant relationship with the cumulative abnormal returns. Firstly, the OperCF/Assets control variable shows significant

18 evidence in both the standard and robust regression analysis at the 1% level. Therefore, Table 4 shows a positive relation between a firm’s cash-flow divided by assets and share price. This implies that the share price reaction to a repurchase announcement increases with increasing generated cash-flows. The result provides evidence for the free cash-flow hypothesis.

The positive coefficient of the Shrsought variables shows significance at the 5% level in the standard regression. The similar coefficient in the robust regression, however, is only significant at the 10% level. This result implies a positive relationship between the number of shares sought and the cumulative abnormal returns. Thereby, the market reaction is more positive for firms with larger repurchase programs.

For the ROA control variable, both regression analyses report a negative relationship between the return on assets and stock price reaction. The coefficient shows negative significance at the 1% level implying that market reaction after a repurchase announcement decreases with increasing returns on assets. Finally, the Leverage variable is only significant in the robust regression analysis at a 1% level. Table 4 reports a positive corresponding coefficients indicating that the market reacts more preferable to firms with higher leverage in the one-day announcement period. This finding is not in line with the leverage hypothesis.

VI. Conclusion

In this research the share repurchase announcement of U.S. firms listed on either the NASDAQ or the New York Stock Exchange have been investigated in the years from 2017 to 2019. Due to the enactment of the Tax Cut and Job Act in 2018, repurchase activity increased substantially afterwards. This paper aimed to examine the announcement effect on the cumulative abnormal returns of repurchasing firms. In order to investigate the potential effects of repurchase announcements, the announcements made by firms are analysed in the described time set using an OLS regression analysis. I found a statistically insignificant result on the Repurchase effects suggesting limited reaction in the share price resulting from the announcement made by a firm.

First, share repurchase announcements are not followed by significant abnormal returns, except from the event window [0, 0]. Analysing the evolution of the CARs in the longest announcement period ([-10, +10]), the returns presented a significant upward slope in the first day after firms announced a repurchase program. Thereafter the abnormal returns stagnated and became slightly negative from the second day onwards. In the days prior to the announcement firms’ stock prices showed a negative abnormal return which provides evidence for the undervaluation hypothesis. The hypothesis implies that repurchase announcements are followed by positive CARs as the announcement is used as signalling device. The statistical significance of the stock price reaction, however, is only exhibited in the one-day period.

19 repurchases showed no significant result. This finding is not in line with evidence from previous literature, as research by Vermaelen (1981) and Ikenberry et al. (1995) find on average a positive market reaction in a two-day announcement period. The evidence of this study on the announcement effect could mean that investors perceive a repurchase announcement as an ambiguous signal. As stated by Vermaelen (1981) the signal might contain information about a lack of growth opportunities or NPV projects.

In future research, a more complex model is likely to yield more accurate results. As the OLS analysis only incorporates variables in line with the hypotheses described in academic literature, the regression does not yield precise results on the effect of announcement in a period with tax cuts. By using a more extensive model, it would be possible to capture both historical- and future firm financial data. Therefore, the results generated are more likely to generalized.

VII. Discussion

According to theory mentioned in the literature review, the initial market reaction during the announcement period shows positive abnormal returns. Strong evidence on this is find by previous literature, however, research also states that the market does not fully incorporate the information signalled through a repurchase announcement. Thereby, this might result in positive abnormal returns in the following years after the announcement. However, as this research did not find significance on the repurchase announcement effect, it is not conform to the results in previous studies.

Firstly, insignificance of the variable might be a result of the usage of the cumulative abnormal returns of non-repurchasing firms. In this research, I suppose the CARs of these firms exhibit a flat curve in the event windows. However, even if these firms do not announce a repurchase program, they might experience other activities that impact the share price. As, in this research, the CARs of non-repurchasing firms are not filtered on activities like M&A’s, and IPO’s, the estimate of the Repurchase coefficient might be different. Further research can filter on those activities to determine the coefficient more accurately. However, as literature states that the information signalled to the market is not fully conveyed, the market reaction to the announcement might also be incomplete.

Furthermore, the OLS framework used in this research might be incomplete and not fully explain the values of the dependent variable. As the goodness of fit of my model is equal to 0.1460 only 14.60% of the variance in CARs in the event window [0, 0] is explained by the explanatory variables. Omitted variable bias can occur which might result in the insignificance of the Repurchase variable. Previous research on tax cuts in relation to share repurchases find that reduction of the repatriation tax leads to an increase in share repurchase. Therefore, a control variable such, for example, foreign profit can capture this evidence in the regression analysis in future research.

Contrary to prior research, this study indicates no upward movement in stock prices as a result from a share repurchase announcement. Whereas other studies find positive evidence on the abnormal

20 returns following an announcement, this paper shows that the impact on stock prices as a result form the announcement is small and only observable in the first day. Therefore, this might implicate that repurchase announcement are a less useful tool to increase share price when a firm perceive their stock as undervalued.

However, this research shows some limitations regarding the regression analysis. As described, a more complex model might be useful for further research to examine the impact of the announcement effect on share prices. Furthermore, the condition used for matching non-repurchasing- to repurchasing firms was limited to only one in this paper. In future research, the number of conditions can be expanded to match firms more precisely. Resultingly, the result of the announcement effect within the regression analysis is more accurate.

21

VIII. References

Albaity, M., & Said, D.S. (2016). Impact of Open-Market Share Repurchases on Long-Term Stock Returns: Evidence From the Malaysian Market. Sage, pp 1-12.

Bagwell, L.S. (1991). Share repurchase and Takeover Deterrence. Journal of Economics, Vol. 22. pp. 72-88.

Bennett, B., Thakor, A., & Wang, Z. (2019, October). Stock Repurchases and the 2017 Tax Cuts and Jobs Act.

Billet, M.T. & Xue, H. (2007, August). The Takeover Deterrent Effect of Open Market Share Repurchases. Journal of Finance, Vol. 62. pp. 1827-1850.

Blouin, J. & Krull, L. (2009). Bringing it home: A study of the incentives surrounding the repatriation of foreign earnings under the American Jobs Creation Act of 2004. Journal of Accounting Research

47. pp. 1027-1059.

Chan, K., Ikenberry, D.L., & Lee, I. (2004, September). Economic Sources of Gain in Stock Repurchases. Journal of Finance and Quantitative Analysis, Vol. 39, pp. 461-479.

Chan, K., Ikenberry, D.L., & Lee, I. (2007). Do managers time the market? Evidence from open-market share repurchases. Journal of Banking & Finance 31, pp. 2673-2694.

Chan, K., Ikenberry, D.L., Lee, I., & Wang, Y. (2009). Share repurchases as a potential tool to mislead investors. Journal of Corporate Finance 16, pp. 137-158.

Comment, R. & Jarrell, G.A. (1991). The Relative Signalling Power of Dutch-Auction and Fixed-Price Self-Tender Offers and Open-Market Share Repurchases. The Journal of Finance, Vol. 46, No.

4, pp. 1243-1271.

Cook, D.O., Krigman, L. & Leach, C.J. (2004, April). On the Timing and Execution of Open Market Repurchases. The Review of Financial Studies, Vol. 17. pp. 463-498.

Dann, L.Y. (1981). Common Stock Repurchases: An Analysis of Returns to Bondholders and Stockholders. Journal of Financial Economics 9, pp. 113-138.

Dharmapala, D., Foley, C.F. & Forbes, K.J. (2011). Watch what I do, not what I say: The unintended consequences of the Homeland Investment Act. Journal of Finance 66. pp. 753-787.

Dittmar, A.K. (2000). Why Do Firms Repurchase Stock? Journal of Business, Vol.73, No. 3, pp. 331-355.

Fama, E.F. (1997). Market efficiency, long-term returns and behavioral finance. Journal of Financial

Economics 49, pp. 283-306.

Farre-Mensa, J., Michaely, R. & Schmalz, M. (2014). Payout Policy. Annual review of Financial

Economics, Vol. 61(1), pp. 75-134.

Grullon, G., & Michaely, R. (2004). The Information Content of Share Repurchase Programs. The

Journal of Finance, pp. 651-680.

Grullon G. & Michaely, R. (2002). Dividends, Share Repurchases, and the Substitution Hypothesis.

The Journal of Finance, Vol.57, No. 4, pp. 1649-1684.

Ikenberry, D., Lakonishok, J., & Vermaelen, T. (1995). Market underreaction to open market share repurchases. Journal of Financial Economics 39, pp. 181-208.

22 Jagannathan, M., & Stephens, C. (2003). Motives for Multiple Open-Market Repurchase Programs.

Financial Management, Vol. 32, No. 2, pp. 71-91.

Jensen, M.C. (1986, May). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. The

American Economic Review, Vol. 72. pp. 323-329.

Kothari, S.P., & Warner, J.B. (1996). Measuring long-horizon security price performance. Journal of

Financial Economics 43, pp. 301-339.

Leng, F., & Noronha, G. (2013). Information and Long-Term Stock Performance Following Open-Market Share Repurchases. The Financial Review, Vol. 48, pp. 461–487.

MacKinlay, A.C. (1997). Event Studies in Economics and Finance. Journal of Economic Literature,

Vol. 35, No. 1, pp 13-39.

Mitchell, M.L., & Stafford, E. (2000, July). Managerial Decisions and Long‐Term Stock Price Performance. The Journal of Business, Vol. 73, No. 3, pp. 287-329.

Nai-Hui, S., & Chan-Jane, L. (2012).The Impact of Open-Market Share Repurchases on Long-Term Stock Returns: Evidence from the Taiwanese Market. Emerging Markets Finance and Trade, Vol. 48, pp. 200-229.

Skinner, D.J. (2008). The evolving relation between earnings, dividends, and stock repurchases.

Journal of Financial Economics 87, pp. 582-609.

Stonham, P. (2002). A Game Plan for Share Repurchases. European Management Journal, Vol. 20(1). pp. 37-44.

Stephens, C.P., & Weisbach, M.S. (1998, February). Actual Share Reacquisitions in Open-Market Repurchase Programs. The Journal of Finance, Vol. 53, No. 1, pp. 313-333.

Vermaelen, T. (1981). Common stock repurchases and Market signalling. Journal of Financial

Economic 9. pp. 139-183.

Von Eije, H., & Megginson, W.L. (2008, December). Dividends and Share Repurchases in the European Union. Journal of Financial Economics 89, pp. 347-374.

Werner, S.P. (2010). Short Selling Activities and Convertible Bond Arbitrage. Reichartshausen:

23

IX. Appendix

24

VIF Result

25