Master Thesis
Share repurchases and corporate cash holdings: Cross-country analysis of the free cash flow hypothesis
June 8, 2018 Student name: Minqi Tu
Student number: S3262510 Study Programme: MSc IFM Supervisor: Dr. Adri de Ridder Co-assessor:Dr. W. Westerman
Abstract: This thesis examines the relationship between corporate cash holding (and free cash flows) and the likelihood of buying back shares by performing a logit regression analysis.
Using a sample of 152,655 company-year observations representing 14,938 companies from 38 countries from 2000 to 2016, supportive evidence is found, showing that firms with more cash and free cash flows are more likely to repurchase shares. Furthermore, the positive relationship between cash holding (and free cash flows) and the likelihood of share repurchase is alleviated when firms have higher growth opportunities. This positive relationship is stronger when firms are located in higher shareholder protection countries.
Firms in common law countries are more likely to repurchase shares than firms in civil law origin.
Keywords: Share repurchase, cash holdings, free cash flows, growth opportunity, shareholder
protection, law origin
1
1. Introduction
Today, a prevalent mechanism for allocating cash to shareholders is share repurchasing.
There are many potential reasons for why companies purchase their shares back, including the agency cost-based theory (Michaely & Grullon, 2002; Litzenberger & Lang, 1989; Jensen, 1986), the undervaluation (signaling) theory (Stephens & Jagannathan, 2003; Ikenberry &
Grullon, 2000; Vermaelen, 1981), the capital structure ( Michaely & Grullon, 2004; Dittmar, 2000) and catering incentives ( Kulchania, 2013; Wurgler & Baker, 2004). According to Ikenberry and Grullon (2000), different motivations can determine payout decisions. This thesis mainly focuses on the agency-based explanation of the motive for firms to repurchase shares, and especially agency theory and free cash flow hypothesis.
According to recent studies, since external financing is costly because of the asymmetry information of the market is high, an increasing number of companies pay attention to reserving significant amounts cash to make investments in the future (Opler et al., 1999). As a high degree of cash holding can offer managers the opportunity to maximize their private benefits and enhance their discretion, company managers might suffer from a serious agency problem if they have too much cash. Company payouts are regarded as an essential mechanism for governance in addressing the agency problem of free cash flow by distributing cash to external shareholders (La Porta et al., 2000; Jensen, 1986; Easterbrook, 1984).
Companies can pay a dividend and repurchase shares to distribute cash to investors. If they do not, it will flow into the pocket of the manager or become an overinvestment.
Prior research shows that share buyback has already replaced dividends as the most
essential payout approach, particularly in America. For instance, according to Michaely and
Grullon (2002), industrial companies have expended more cash on share buyback than on
dividend payment in America. Since the late 1980s, stock repurchase programs have been
prevalent in America. Since the late 1990s, more nations have repurchased stock by
legitimation, such as Asian and European nations. As a result, stock repurchase programs are
considered international activities. According to Haw et al. (2013), the percentage of
repurchasing companies has stably developed, while the fraction of companies paying
dividends has obviously decreased in Europe. A similar tendency can be observed in East
Asia.
2
There are sparse studies on global share repurchase programs in spite of the increasing trend for industrial companies to use stock repurchases as a payout approach. Most studies of share repurchase are established using an American sample (e.g., Chan et al., 2004; Michaely
& Grullon, 2002; Dittmar, 2000; Harford & Guay, 2000; Ikenberry et al., 1995). In consideration of the different institutional environments in different nations, findings from U.S. companies might not be internationally generalizable. Although studies on share repurchase from non-U.S. nations have recently increased, these studies focus mainly on one specific non-U.S. jurisdiction. For example, Young and Oswald (2008) analyze companies domiciled in Britain, showing that share buyback can replace the retention of cash that would otherwise prove to be expensive for shareholders. Bremer and Tong (2016) focus on the repurchases of stock in Japan, while De Ridder (2009, 2015) emphasizes the Swedish share repurchase market. Global studies on the repurchase of shares are fairly meager. Megginson and Von Eije (2008) state that they are not aware of any global studies that investigate share repurchases. Suh and Lee (2011) have found global evidence for share repurchase and cash holding, but they have not tested cross-country differences and only include seven developed nations in their study. Moreover, most research focuses on the signaling theory as a determinant for companies to purchase shares back (Chan et al., 2004; Stephens &
Jagannathan, 2003). Such studies look at abnormal stock returns around the date of announcement in order to test whether a company can generate value via the announcement of share repurchase. Few studies have focused on the agency cost-oriented explanation of share repurchases (Liang & Fenn, 2001).
In this paper, I explore the relationship between corporate cash holding, free cash flows and share repurchases. The overwhelming majority of the studies conducted in the existing literature are based on single-country analyses, which are mostly derived from the U.S.
market. The rules that apply to share repurchases distinguish significantly across countries,
rendering it precarious to draw the same conclusions between countries regarding share
repurchases. In order to fill the void, this study uses a multinational sample. I develop
arguments based on the theoretical analysis of motives for firms to buy back shares on agency
cost-based, especially the free cash flow hypothesis. I analyze whether firms with higher cash
holdings and free cash flows are more likely to pay out to their shareholders by using a
3
multinational sample. I also analyze whether a firm’s growth opportunities influence its payout policy. Furthermore, since different countries have different institutional environments, I also analyze the influence of country-level investor protection and legal law origin on the relationship between cash holdings and the probability of share buyback. I aim to better explore the link between corporate cash holdings and payout policies and to better understand the incidence and rationale used for managerial implications. Therefore, the main research questions in this paper are as follows: What is the influence of cash holding (and free cash flows) on the probability of a firm to buy back its shares? What is the effect of a firm’s growth opportunities on the likelihood of a firm to repurchase shares? What is the influence of shareholder protection and legal law origin on the relationship between cash holding (and free cash flows) and the likelihood of a firm to repurchase shares?
By using 152,655 firm-year observations from 2000 to 2016 across 38 countries, I find that the US, Japan and the U.K. are the top-three biggest countries where share repurchasing is prevalent. Additionally, the logit regression model shows that there is a positive relationship between cash holdings (and free cash flows) and the likelihood of firms to buy back shares, which is consistent with Dittmar (2000), Lee and Suh (2011), and Oswald and Young (2008).
This relationship is alleviated when a firm has higher growth opportunities, which is consistent with findings by Brav et al. (2005). In contrast, this positive relationship is stronger when firms are located in higher shareholder-protection countries, which is consistent with the outcome model by La Porta et al. (2000), which indicates that share repurchasing can be regarded as an outcome of legal protection of shareholders. Moreover, by splitting countries by legal region, I find that common law countries are more likely to pay out by repurchasing than civil law countries, especially when it comes to German civil origin and Scandinavian civil law.
This thesis makes several contributions to the literature. First, this research applies agency theory to give an explanation of why firms repurchase shares. Second, I extend previous studies to an international level with different institutional settings, which analyze the moderating effect of both firm-level factors such as corporate growth opportunity and cross-country differences such as shareholder protection and legal regions on share buybacks.
The remainder of this thesis is organized as follows. Section 2 provides a review of the
4
literature concerned with share buybacks. Section 3 subsequently develops the hypotheses.
Section 4 describes the process of data collection and the research sample. Section 5 details the definition of the variables and the regression model used in this thesis. Section 6 discusses the empirical findings and relates the findings to previous studies. Section 7 concludes the main findings.
2. Literature review
2.1. The basic of share repurchases
Share repurchases, or share buybacks, refers to a firm’s reacquisition of its own stock, which is an alternative way to allocate cash to shareholders. Share repurchases is a flexible method of cash distribution, as they are not corporate commitments. This is the major advantage of share repurchases. In contrast, dividends are performed as corporate obligations as the market penalizes companies that omit or decrease dividend payments (Michaely &
Grullon, 2002). The substitution hypothesis has been proposed by Michaely and Grullon (2002), who argue that actual corporations are substituting share repurchases for dividends.
Nonetheless, according to some empirical studies, repurchasing firms do not reduce dividend payments (e.g., Suh & Lee., 2011; Harford & Guay, 2000; Dittmar, 2000). There are four major approaches for a company to buy back shares: (1) fixed-price tender offers, (2) Dutch auctions’ tender offers, (3) privately negotiated repurchases and (4) open-market repurchases program (Vermaelen, 2005; Ikenberry & Grullon, 2000).
Utilizing the approach of fixed-price tender offer, a firm can make a tender offer to
purchase back a particular number of shares (the target number of shares) at a fixed price
from shareholders at a premium of the stock price of the present market. Shareholders can
choose to submit the number of shares they have the willingness to sell back to the company
or hold them. The tender offer is valid for a limited time period. A company can purchase
back any number of shares between the number of tendered shares and the shares’ target
number, when all shareholders are treated equally if the offer is excessively subscribed
(Vermaelen, 2005). The company might commit itself to purchasing all the shares back or
extend the offering period if the amount of tendered stocks is smaller than the targeted
number of stocks.
5
The method of Dutch auctions’ tender offers is similar with fixed-price tender offers.
Rather than providing a single fixed-offer price, Dutch auctions set up a series of prices for shareholders, and the minimal price is higher than the market price. Shareholders offer their minimal acceptable selling price as well as the amount of shares they want to sell to the repurchasing firm. Subsequently, the offering company identifies a minimum price to buy back the desired number of shares from shareholders and pays the price to all qualified bidders (Vermaelen, 2005). The transaction could be completed in a short time period.
Repurchases that are privately negotiated are also referred to as targeted repurchases. A firm gains access to negotiations on a share price with bigger shareholders (enterprise or individual investors) and repurchases the shares from them. The price is often at a premium in comparison with the market price. According to Vermaelen and Peyer (2005), repurchases negotiated privately are unique, as the initiative to buy back shares is not taken by the company. Instead, the seller can be a proactive agent during the repurchasing process.
Different from other repurchase forms, targeted repurchases’ premiums are greatly determined by the seller and by the company’s bargaining strength, which shows that private repurchases rely on different price behaviors and motivations than repurchases of other types, such as fixed-price tender offers, Dutch auctions’ tender offers and open-market repurchases.
Open-market repurchases program is the most universal approach to purchasing stock from the market. Brokers are instructed by a company to conduct the transaction. As opposed to tender offers, the open market repurchases are an option instead of an obligation. Therefore, open-market repurchases are not corporate commitments. After an announcement, the firm has the option to terminate the program or make changes based on the company’s situation;
the firm does not have any legal obligation to finish the repurchasing program after it is announced. Open-market repurchases program is considered to be the most prevalent approach among those four methods, due to its flexibility and cost effectiveness. According to Ikenberry and Grullon (2000), around 91% of all repurchase announcements are made on the U.S. open market.
Because a majority of buybacks are carried out in the open market, this paper focuses on
only the open-market repurchase programs.
6
2.2. Why do corporate allocate cash via share repurchase?
According to Modigliani and Miller (1961), the payout policy of a company is irrelevant when the company’s business policy is held constant in a frictionless market. The payout policy does not have any influence on the wealth of total shareholders since higher payout results in lower capital gain and retained earnings, and vice versa. Ever since their dividend irrelevance theory, variety of studies has struggled to explain why companies distribute cash in general and why they choose between share repurchase and dividends as a particular form of payout. According to these studies, explanations can be divided into three types: the undervaluation (signaling) theory (Stephens & Jagannathan, 2003; Ikenberry & Grullon, 2000;
Vermaelen, 1981), agency problems ( Michaely & Grullon, 2002; Litzenberger & Lang, 1989;
Jensen, 1986) and catering incentives (Kulchania, 2013; Wurgler & Baker, 2004). This thesis uses the agency-cost-oriented explanation. Thus, this is the discussion’s focus.
2.2.1. Agency theory
The agency theory describes the relationship between principals (shareholders) and agents (managers in a company). A principal delegates an agent to operate business in the relationship. In fact, how to set objectives for agents and principals to reduce conflicts has been a long-standing topic in the literature of corporate finance. The major conflict centers on how to allocate the internal funds of a company, meaning what proportion of the corporate profit should be paid in the forms of share repurchases and what proportion of profits should be retained for development in the future.
According to the agency theory, managers try to maximize their own utility at the expense of corporate shareholders in the imperfect markets of labor and capital. As they have a better understanding than shareholders about whether they can meet the objectives of a shareholder (Easterbrook, 1984), managers might tend to operate in their own self-interest instead of considering the company’s best interest because of uncertainty and asymmetric information. For instance, myriad factors contribute to the final outcome, and it is not obvious whether an agent can directly lead to a certain result, no matter whether it is negative or positive.
Under the agency theory, the assumptions derived from Miller and Modigliani’s (1961)
7
dividend irrelevance theory can be moved aside for two reasons (La Porta et al., 2000). First, the business policy cannot be regarded as independent of its payout policy, as the cash payout might reduce the corporate marginal investment’s effectiveness. Second, we cannot ensure that company profits can be allocated to investors on a pro rate basis, as managers might have preferential treatment via theft, diversion of corporate property and transfer prices even keeping business policy unchanged. Managers’ power can be reduced, as a payout to shareholders reduces the company resources under their control. The payout approach of share buyback can be considered a tool to reduce the conflicts between principals and agents (La Porta et al., 2000; Jensen, 1986; Easterbrook, 1984).
2.2.2. Free cash flow hypothesis
The free cash flow hypothesis is first described by Jensen (1986). According to Jensen, free cash flow is described as cash flow excessive of that needed to finance all projects with a positive net present value (NPV) while being discounted at the related capital cost. Jensen considers that companies with cash in excess of that needed to finance projects of positive NPV face severe issues of agency, as free cash flow exacerbates the conflict of benefit between principals and agents. An essential implication of Jensen’s free cash flow theory is that companies with higher free cash flow have high probability to initiate investment and acquisition in value-decreasing projects. If the company retains the free cash flow, the decreasing marginal utility of the available investments will result in the deterioration of the stock price and returns. Therefore, corporate managers with free cash flow are pressured to pay it to investors instead of reinvesting cash in negative NPV projects. As payouts to shareholders increase, the stock price will increase correspondingly.
In sum, Jensen’s (1986) free cash flow hypothesis expresses the idea that limiting a
manager’s freedom of action plays an essential function in decreasing the equity’s agency
costs, and it is recommended that the decrease of free cash flow can help address the agency
issue. According to this, share repurchase can be considered an effective mechanism of
control to reduce the expropriation of minority shareholders since they draw down the
reserves of corporate cash that are available to shareholders in control.
8
2.2.3. Other motives for share repurchase
According to Ikenberry and Grullon (2000), there is no single major incentive for companies to buy back share, and managers might be encouraged by many factors to repurchase simultaneously. Besides the motive by the free cash flow, the signaling theory believes to be the determinant of the share repurchase studied by many studies. According to the signaling hypothesis, a payout policy might offer signals about a company’s present performance and future prospects under asymmetric information between investors and managers. The asymmetry of information means that insiders have more private knowledge regarding present mispricing and the future expectation of the firm’s shares than both potential investors and external shareholders do. According to Dittmar (2000), when companies consider their stock prices are undervalued in the market, companies are companies are more likely to buyback. Thus, the action of repurchasing is a signal of misevaluation in the market. Furthermore, according to Wurgler and Baker (2004), companies cater to the demands of investors, indicating that the propensity to pay out is reduced when the proxy for the dividend premium of the stock market is negative and vice versa. According to Kulchania (2013), companies repurchase shares when an investor places a stock price premium of repurchasing firms, and a company would pay a dividend when an investor places a higher value on the dividend-paying companies.
There are a number of other motivations that drive firms to buy back their shares.
Evidence is provided by Liang and Fenn (2001) that stock option activities can affect decisions regarding share repurchases. Billett and Xue (2007) consider that companies utilize share repurchases to protect themselves from takeovers, confirming the hypothesis of takeover deterrence. In accordance with the rivalry-based theory of imitation, Agrawal and Adhikari (2018) have found that peer influence on repurchases can posit a company’s peers’
actions in order to maintain competitive parity. In general, a variety of motivations can determine payout decisions at the same time.
2.4. Cash holding and share repurchase
Based on agency theory and free cash flow hypothesis, cash holding is an important
determinant when firms consider share repurchase. Harford (1999) shows that companies
9
having high level of cash are more likely to take acquisitions even the acquisition cannot create value when compared to acquisitions undertaken by companies with normal cash levels.
This result implies that firms with more free cash flow are more likely to waste resources, which leads to higher agency costs. As suggested by the free cash flow hypothesis, firms with large cash balances and more free cash flows tend to repurchase shares to reduce this potential agency conflict by reducing the amount of cash available.
Many empirical studies find evidence to support the free cash flow hypothesis. Stephen and Weisbach (1998) show there is a positive relationship between the levels of free cash flow and share repurchases, which supports this hypothesis. Dittmar (2000) uses cash and cash flow as proxies for testing the free cash flow hypothesis and finds that the median cash flows and cash of repurchasing firms are significantly greater than those of non-purchasing firms.
Additionally, his results of the Tobit model also show positive and significant coefficients on either cash or cash flow, which indicates that firms repurchase to distribute cash or cash flow to shareholders. Evidence has been found by Young and Oswald (2008) that the reacquisition of shares replaces the retention of cash, which would prove to be expensive for shareholders, according to a research sample from Britain. Lee and Suh (2011) use cash and short-term investment to represent corporate cash holdings. They find that large cash holdings at the beginning of the year are significantly positively related to the amount of share repurchase by analyzing seven developed countries: the U.S, Australia, Canada, Germany, France, Japan and the U.K. They further show that large cash holdings represent free cash flows in repurchasing firms. Megginson and Von Eije (2008) have examined 15 EU countries, finding that larger cash holdings decrease the probability of paying cash dividends yet enhance repurchasing probability. Nonetheless, Opler et al. (1999) have found contradicting evidence about free cash flow hypothesis, namely that there is little evidence that excessive cash has influence on payouts to shareholders by trading American companies publicly. To be more specific, they show a rise in the free cash flow result in a small increase of payouts to shareholders and that typically companies decrease free cash flows by covering losses instead of making acquisitions or making investments on new projects. Thus, there is barely any evidence to show that free cash flow is transferred to the pockets of managers.
Previous studies focus mainly on single jurisdictions and especially U.S. samples to
10
analyze the free cash flow hypothesis, and international studies are scarce. Therefore, one needs to be cautious generalizing the results globally.
2.5. Growth opportunity, free cash flows and share repurchase
Myers (1977) considers that opportunities for growth are one component of corporate market value (the other is the value of assets in place). The lower fraction of corporate value represented by assets in place tends to result in greater corporate investment opportunities or growth opportunities. According to Merton and Mason (1985), companies with growth opportunity have projects with more expansion capacity, new equipment, maintenance and replacement of current property, and acquisition of other firms.
Previous studies argue that there is a negative relationship between growth opportunity and payout. According to French and Fama (2001), dividend payers tend to be more profitable, have larger companies and have less valuable opportunities for growth compared to non-payers, and Ferris et al. (2009) find similar results. Fatemi and Bildik (2012) examine international payout policies and show that dividend payers are associated with low investment opportunities. Gugler (2003) finds that Austrian firms which have higher total payout ratios are equipped with low growth opportunities. Thus, company with low (high) growth opportunities should have relatively high (low) corporate payout.
Based on the agency theory, firms that hold free cash flow have a propensity to make an investment even if the project cannot create a positive NPV, since the manager’s interest is to maximize their personal value at the cost of shareholders (Jensen, 1986). Cash distributions through dividend and share repurchases remove corporate resources and therefore help to eliminate agency costs of free cash flows. This assumes that companies without profitable projects to invest are more likely to pay out to shareholders rather than undertake negative NPV projects. Boudry et al. (2013) indicate that the amount of free cash flow available has a positive and significant influence on repurchase only for firms with poor growth opportunities.
Additionally, firms must consider their need to fund growth when setting a payout policy.
Black (1976) indicates that non-paying firms can demonstrate that they have attractive
investment projects that might be missed if the firm chooses to pay dividends. If a firm
11
undertakes these attractive projects, the share prices might rise up. This increased value of shares may more than the values result from the lost dividends. If that happens, its shareholders can gain benefits. Similarly, Brav et al. (2005) survey 348 financial executives about their perspectives on corporate payout policy. They find that firms opt to repurchase when they hold free cash flows or have a lack of growth opportunities, and firms usually choose to satisfy liquidity and growth needs first and then think about whether or not to repurchase.
In summary of the previous studies, there is a closed relationship between growth opportunity, free cash flow and firms’ payout decisions. Growth opportunity can influence a firm’s payout decision since managers tend to trade off the deployment of internal funds.
Corporate managers need to think whether to disgorge cash to profitable projects or distribute funds to the shareholders.
2.6. Shareholder protection and share repurchases
The discussion above presents the application of agency theory in firm-level analysis. As there are cross-country differences, I also consider the agency theory on a country-level basis.
The effectiveness of stock repurchase in a decreasing agency conflict greatly relies on the environment of country-level investor protection. The problem of free cash flow might be more serious in nations with institutions that have poor investor protection since managers can transfer cash reserves into their pockets at a low expense (Haw et al., 2011). La Porta et al.
(2000) have conducted a study of 33 nations by using the index of shareholder protection to identify the minority shareholder rights’ level. It was found that higher dividend payouts and stronger minority shareholder rights are related. Their finding supports their outcome agency model of dividends, which indicates that a dividend is an outcome of legal protection of shareholders. This result implies that greater rights of minority shareholders (such as preventing the expropriation of wealth and voting for directors) tend to lead to more cash extracted from company insiders.
Similarly, Jiraporn (2006) reveals that the likelihood of a repurchase program is
positively significantly related to the extent of shareholder rights: the better protected the
shareholder rights, the more payout there is by repurchase shares. Jiraporn’s result implies
12
that the managers of firms in strong shareholder protection countries have a lower capability of retaining cash, which could be a waste for their own private interest and are thus forced to disgorge cash in the form of share repurchases. Dittmar and Mahrt-Smith (2007) indicate that weak corporate governance leads to higher cash holdings with less disbursement of cash to outside shareholders. According to Oswald and Young (2008), the degree of an external minority shareholder’s rights can affect the corporate payout decision to distribute free cash flows to external shareholders. Harford et al. (2008) demonstrate that firms with weaker shareholder protections tend to choose to repurchase instead of increasing dividends when distributing cash to shareholders because these firms can avoid future payout commitments.
In other words, managers are likely to arbitrarily expropriate the benefits of minority shareholders if the interest of external shareholders cannot be protected well. This paper considers that companies in nations with weaker shareholder protections are less likely to allocate the cash via share repurchase.
2.7. Cross-country legal regimes and the agency problem
As is widely known, the conflict of interest between external investors, such as minority shareholders, and company insiders, such as managers, should be taken into consideration when payout decisions are analyzed. The insiders who manage the company assets can use the resources for many different purposes harmful to external investors. For instance, a company insider can divert company assets, such as cash, to his or her own pocket. According to La Porta et al. (2000), the law is the principal remedy for agency issues. A nation’s law provides external investors, including shareholders, certain power in order to protect their investments from being controlled by company insiders.
According to La Porta et al. (1997, 1998), outside investors’ legal protection is different across nations. The authors divide nations into two classes: common law nations and civil law nations. The civil legal tradition stems from Roman law, with constantly updated, comprehensive legal codes and other legal materials. Legal scholars subdivide civil law into three categories: Scandinavian, French and German. The opposite of civil law, common law is not established on legal statutes and the comprehensive compilation of rules.
La Porta et al. (1998) examine whether law can govern the protection of investors in 49
13
nations, finding that companies with legal rules in the common law tradition tend to have better investor protections than those in nations with legal rules in the civil law tradition, particularly French civil law. Scandinavian civil law and German civil law play an intermediate role in protecting investors. In consideration of legal law differences, Megginson and Von Eije (2008) study share repurchases in 15 EU nations, finding that companies with headquarters in a civil law nation tend to repurchase shares, as opposed to companies with headquarters in a common law nation.
2.8. Summaries of previous studies
There are three pervasive theories that explain the motivation of share buybacks:
undervaluation, or signaling theory; agency theory; and catering theory. The majority of studies focus on the undervaluation theory, and scholars explain that firms are more likely to buy their shares back when they perceive their shares are undervalued by the market (Vermaelen, 1981; Grullon and Ikenberry, 2000; Dittmar, 2000; Jagannathan and Stephens, 2003). The catering theory emphasizes that managers care about demands of investors, and therefore firms tend to repurchase their shares when the repurchase premium is high (Baker and Wurgler, 2004; Jiang et al., 2013; Kulchania, 2013). However, this thesis concentrates on agency theory.
Grounded in agency theory, Jensen (1986) raises the free cash flow hypothesis. Later studies find empirical evidence that the level of cash holdings (and free cash flows) is highly related to the probability of firms buying back their shares (Lee & Suh, 2011; Oswald & Yong, 2008; Grullon & Michaely, 2002; Stephens & Weisbach, 1998; Lang & Litzenberger, 1989).
Notably, a firm’s growth opportunity also plays an important role in corporate payout policy.
Previous studies indicate that firms tend to pay out more when they face less growth opportunity (Ferris et al. 2009; Brav et al., 2005; Fama and French, 2001; Black, 1976).
Additionally, La Porta et al. (1997, 1998) point out that monitory shareholder protection and legal region are two important factors that influence corporate payout decisions.
3. Hypotheses development
The first hypothesis of this study is about the relationship between the level of free cash
14
flow and the likelihood of firms to conduct open market share repurchase programs.
According to the agency theory and the free cash flow hypothesis, discussed above, it is known that managers tend to waste corporate resources on unprofitable projects rather than returning the capital to shareholders. Managers with sufficient funds can increase payouts, dividends or repurchased stock to shareholders that would otherwise be wasted on low-return investments. Free cash flow is reduced through buyback shares, and therefore the likelihood of unprofitable and expansion-oriented investment being undertaken decreases. In order to fill the research gap, this study examines the free cash flow hypothesis on an international basis and we predict the following:
Hypothesis 1: Firms with a high level of cash holding (and free cash flow) are more likely to buy back shares.
Share repurchase, on one hand, can remove resources from the company and thus help reduce the agency cost of free cash flow. It is assumed that companies without opportunities for profitable investment pay higher dividends rather than undertake projects with negative NPV (Litzenberger & Lang, 1989; Jensen, 1986). On the other hand, managers would stipulate that dividend policies conform to the opportunities of corporate investment. As companies with opportunities for high growth tend to be less flexible in payout policy and have lower free cash flow, they tend to repurchase fewer or even no shares. Companies might repurchase fewer shares in order to decrease their dependence on expensive outside financing.
This reasoning has been supported (Brav et al., 2005; Black, 1976). Correspondingly, low-growth firms are likely to have relatively higher payout because they do not need to retain cash to fund growth.
The literature suggests that firms’ growth opportunities are closely related to share repurchase decisions since growth opportunities are associated with the deployment of internal funds. Managers need to decide what proportion of profits to reinvest in new projects and what proportion of profits to return to shareholders. Therefore, the second hypothesis is as follows:
Hypothesis 2: The positive relationship between cash holding (and free cash flow) and
the probability of buying back shares is alleviated when firms have high growth opportunities.
15
In addition to a firm-level analysis, I also consider cross-country differences in the development of a hypothesis. Legal regimes and shareholder protection are two important remedies for agency problems (La Porta et al., 2000). Without an effective legal control, it is hard to persuade self-interested managers to distribute cash reserves to external shareholders.
Previous studies mostly focus on the effect of shareholder protect on dividends. However, the impact of shareholder protection on the decision to pay out through an irregular payout method such as share buyback is underexplored. Based on the argument discussed above, I raise the following hypotheses:
Hypothesis 3: The positive relationship between the level of free cash flow and the
likelihood of a firm to buy back shares is stronger in countries with better shareholder protection.
Hypothesis 4: The positive relationship between the level of free cash flow and the
likelihood of a firm to buy back shares is stronger in countries with a common law legal origin than in those with a civil law legal origin.
4. Data
4.1. Sample collection
The sample for this study was collected from Compustat annual files and Worldscope.
These two databases record financial information on worldwide publicly held companies. The data in this paper was collected from 2000 to 2016. The period starts in 2000 because restrictions on open-market share repurchases in non-U.S. countries were abolished starting in the late 1990s. For instance, Japan abolished the restriction in 1995, Finland and Poland in 1997, Germany and France in 1998, Norway in 1999 and Sweden in 2000, as indicated by De Ridder (2009).
I focused on the repurchases’ actual amount rather than the announcement of share
buyback, since an announcement cannot show a company’s commitment to implement the
repurchase of a share. Thus, this paper estimates an actual stock repurchase in a certain year
as Compustat data item #115 (total expenditures on the purchase of preferred and common
stocks) minus any decrease in the preferred stock’s value (the preferred stock’s redemption,
Compustat item #56). According to Banyi et al. (2008), even though there are some errors in
16
the Compustat-based measure, it is the most precise estimate of actual repurchase. Most prior studies use Compustat to estimate share repurchases’ actual numbers for American companies (Grullon & Michaely, 2002, 2004; Dittmar, 2000; Weisbach & Stephens, 1998).
However, the Compustat-based measurement of a share repurchase amount is only feasible for Canadian and American companies. For the international sample, the Compustat Global files do not offer an item corresponding to the redemption value, so we could not identify the cash amount used to decrease the preferred share. Besides, few companies have information available about the number of preferred and common stocks they have bought in the Compustat Global files. Thus, I chose to use Worldscope to gather the financial information of non-American enterprises, following Haw et al. (2011, 2013) and Suh and Lee (2011), who use Worldscope to estimate the number of share buybacks for an international sample. Worldscope data item #04751 (Common/Preferred Redeemed, Retired, Converted, etc.) corresponds to Compustat #115. However, there is no appropriate item corresponds to Compustat #56 in the Worldscope database. Lee and Suh (2011) and Haw et al. (2011, 2013) provide an alternative way to deal with the lacking redemption value; they drop firms whose preferred shares declined during the fiscal year, and Lee and Suh (2011) show that they lose only a small number of the firms in the sample by excluding those firms.
I excluded all financial services firms whose Standard Industrial Classification (SIC) codes are between 6,000 and 6,999. I excluded firms that had no share repurchase programs between 2000 and 2016. Following Lee and Suh (2011), I winsorized extreme values at the 1st and 99th percentiles to reduce the effect of outliers. I excluded countries that do not have a shareholder protection (anti-self-dealing) index. The final sample included 152,655 firm-year observations representing 14,938 firms from 38 countries.
4.2. Sample distribution
Table 1 reports the sample distribution by year. It shows the number of firms with share
repurchases and the number of firm without share repurchases in each year. The number of
repurchasing firms shows an increasing tendency from 2000 to 2016 which indicates that
share repurchases have become a popular distribution method as time goes by. There are
152,655 of firm-year observations in the total.
17
Table 1. Sample distribution by year.
Year Number of firm without share repurchases
Number of firm with share repurchase
Total number of firms % of total observations
2000 2001 2002
3,997 4,760 5,111
2,643 2,651 2,801
6,640 7,411 7,912
4.350 4.855 5.183
2003 5,323 2,947 8,270( 5.471
2004 5,623 2,852 8,475 5.552
2005 5,433 3,157 8,590 5.627
2006 5,424 3,454 8,878 5.816
2007 5,471 3,823 9,294 6.088
2008 2009
5,110 6,083
4,430 3,669
9,540 9,752
6.249 6.388 2010
2011 2012 2013 2014 2015 2016 Total
6,263 5,883 5,934 5,564 5,255 5,005 5,000 91,239
3,472 3,749 3,740 4,294 4,530 4,646 4,558 61,416
9,735 9,632 9,674 9,858 9,785 9,651 9,558 152,655
6.377 6.310 6.337 6.458 6.410 6.322 6.261 100
Table 2 reports the sample distribution by legal region and it includes English, French, German and Scandinavian origins. The final sample was made up of 38 countries. The US had the majority of the total observations with 30.076% representing 4,947 firms. The second most came from Japan, with 22.172% of the total observations representing 2,660 firms, and the third was the U.K., with 5.467% of the total observations representing 822 firms.
Table 2. Sample distribution by legal region
Country No. of
firms
Firm-year obs. with no repurchases
Firm-year obs.
with repurchase
Total obs. % of total obs.
English-origin Australia Canada Hong Kong India Ireland Israel
506 573 61 607
33 117
3969 3155 661 4461
263 747
1097 1839 151 785 114 304
5066 4994 812 5246
377 1051
3.319 3.271 0.532 3.437 0.247 0.688
18
Table 2 (continued)
Country No. of
firms
Firm-year obs. with no repurchases
Firm-year obs.
with repurchase
Total obs. % of total obs.
Malaysia New Zealand Singapore Thailand U.K.
U.S.
371 53 259
73 822 4947
2894 430 2146
839 5485 22,950
1529 141 653 95 2860 22962
4423 571 2799
934 8345 45,912
2.897 0.374 1.834 0.612 5.467 30.076 French-origin
Belgium Brazil Chile France Greece Indonesia Italy Mexico Netherlands Philippines Portugal Russian Turkey Spain
71 146
36 477
90 90 167
81 108
87 37 113
48 124
543 949 453 3526
801 750 1302
543 789 831 254 504 300 677
290 460 63 1956
227 146 479 393 468 215 170 244 50 610
833 1409
516 5482 1028 896 1781
936 1257 1046 424 748 350 1287
0.546 0.923 0.338 3.591 0.673 0.587 1.167 0.613 0.823 0.685 0.278 0.490 0.229 0.843 German-origin
Austria China Germany Japan Korea Poland
49 18 295 2660 1027 154
465 148 2671 17,008
5,069 1009
146 23 852 16,838
2318 295
611 171 3523 33,846
7387 1304
0.400 0.112 2.308 22.172
4.839 0.854
Switzerland 192 1179 1118 2297 1.505
Scandinavian-origin Denmark
Finland Iceland Norway Sweden
94 90 9 121 132
727 843 31 773 1094
445 279 22 380 399
1172 1122 53 1153 1493
0.768 0.735 0.035 0.755 0.978
Total 14938 91239 61416 152655 100
19
5. Methodology
5.1. Dependent variable
Based on the actual number of share buybacks each year, I used a binary variable to indicate whether a firm buys back shares or not, which is equal to 1 if the firm buys back shares and 0 if the firm does not buy back shares. This binary variable is the dependent variable in this paper.
5.2. Major independent variables
5.2.1. Measuring cash holdings and free cash flows
Previous studies such as Dittmar (2000), Oswald and Young (2008) and Lee and Suh (2011) use more than one indicator to test the free cash flow hypothesis. Based on their research, I include two indicators to represent the level of cash holdings, namely cash holding and free cash flows. Based on Dittmar (2000) and Lee and Suh (2011), I measure cash holding as the ratio of cash and short-term investments to total assets at the end of the year prior to share buybacks. I also follow previous work by Lehn and Poulsen (1989) and define free cash flow as follows:
FCF = INC − 𝑇𝐴𝑋 − 𝐼𝑁𝑇𝐸𝑋𝑃 − 𝑃𝐸𝐷𝐷𝐼𝑉 – 𝐶𝑂𝑀𝐷𝐼𝑉, (1)
where INC is operating income before depreciation; TAX is total income taxes minus change in deferred taxes from the previous year to the current year; INTEXP is gross interest expense on short- and long-term debt; PFDDIV is the total amount of preferred dividend requirement on cumulative preferred stock and dividends paid on noncumulative preferred stock; and COMDIV is the total dollar amount of dividends declared on common stock.
This free cash flow definition is also used in studies on open-market repurchases by Kahle (2002) and Chan et al. (2004). After calculating the amount of free cash flow, I use the ratio of free cash flow to total assets at the beginning of the year (Dittmar, 2000; Kahle, 2002;
Lee and Suh, 2011). As indicated by Jensen (1986), free cash flow is the undistributed cash
flow in excess of that needed for a positive NPV project.
20
5.2.2. Growth opportunity
According to Kahle (2002), the market-to-book ratio is a proxy for a firm’s growth opportunities. A negative relationship is expected between share repurchase and growth opportunities based on agency costs of corporate payouts, since companies with good growth opportunities tend to maximize shareholders’ value by investing cash flow in projects rather than disgorging cash flow to shareholders. The market-to-book ratio is measured by market value of equity plus book value of debt, scaled by book value of assets at the end of the year prior to share repurchase (Dittmar, 2000; Kahle, 2002; Denis and Osobov , 2008).
5.2.3. Investor protection
Based on Haw et al. (2011), I used the anti-self-dealing (ASD) index to proxy for country-level investor protection. The ASD index provides the level of legal protection of minority shareholders against expropriation by corporate insiders and was developed by Djankov et al. (2008). The higher this index, the stronger the investor protection is. As indicated by Haw et al. (2011) the ASD index works better than the commonly used revised anti-director rights index (ADR)
1to explain a variety of stock market outcomes. I used the ADR as an alternative measurement of shareholder protection in order to provide a robust check. ADR shows the degree of a legal protection of minority shareholders in the corporate decision-making process such as voting process.
5.2.4. Legal regimes
Based on La Porta et al. (1998), I divided the sample according to law origin. I used dummy variables to indicate the legal origin (OGN) of the countries. An OGN equals 1 if the origin of the company law is common law and is equal to 0 if the origin of the company law is civil law (including French, German or Scandinavian civil law). La Porta et al. (1998) find that common law countries generally have the strongest legal protections for investors relative
1 The anti-director rights index is extensively used to proxy for country-level shareholder protection, and it was firstly developed by La Porta et al. (1998). Spamann (2010) has recently published a paper raising concerns over the anti-director rights index’s reliability and argues that this index should be improved.
Djankov et al. (2008) further improve the anti-director rights index with the revised anti-director rights index, and I use the revised anti-director rights index for a robustness check in this paper.
21
to civil law countries; therefore, firms in common law countries are more likely to pay out.
5.3. Control variables
Based on previous studies, we included a number of firm-specific variables that may influence a firm’s open-market share repurchase decision.
Firm size (SIZE) has been used in previous research as a proxy for information asymmetry and financing costs (Dittmar, 2000; Chay and Suh, 2009; Haw et al., 2013). It is argued that larger firms have lower information asymmetry and financing expenditures, thus more stable cash flows. The lower financing costs enable a firm to pay out more cash to shareholders since they can raise funds relatively inexpensively in the future. SIZE can be measured as the logarithm of total assets at the end of the year prior to share buyback. Dittmar (2000) finds that large firms are more likely to buy back shares.
Leverage ratio (LEV) is another factor that also influences a company’s payout decision.
As indicated by Kahle (2002), higher debt ratio can serve as a proxy for financial distress, and thus firms with higher LEV are unwilling to distribute cash through repurchase shares.
Dittmar (2000) finds that firms may repurchase to adjust their LEV. Lee and Suh (2011) indicate that firms with low LEV may use share repurchase for recapitalization purposes.
Third, the annual stock return from the prior year (RET) is also incorporated. Stephens and Weisbach (1998) indicate that stock repurchases are negatively related to prior stock price performance. Firms that choose to repurchase shares have lower stock returns prior to the repurchases if managers believe that the stock is undervalued (Jagannathan et al., 2000). RET is also a proxy used to test the undervaluation hypothesis.
According to Kahle (2002), firms with high capital expenditure (CAPEX) should have both better investment opportunities and less free cash flow, and thus should pay out less.
Therefore, CAPEX is an important factor influencing the payout policy. I also include market
capitalization (MC), which is measured by the stock price at the end of the year multiplied by
the number of shares outstanding. The MC value can be a proxy for financing costs, based on
Kahle. Firms with high MC are more likely to disgorge cash since they can easily raise money
with fewer financing costs. Additionally, non-operating income (NOPER) is included.
22
Jagannathan et al. (2000) indicate that repurchases are more likely to be used by firms with higher non-operating cash flows, which also refers to temporary cash flow.
Considering that stock repurchase is a corporate payout method, the cash dividend (DIV), operating profitability (ROA) and operating profit volatility (ROAVOL) need to be taken into consideration since they are closely related to corporate payout strategy. Grullon and Michaely (2002) argue there is a substitutive relationship between share repurchases and dividends, and thus a DIV may influence the share repurchasing decision. DIV is measured as cash dividends paid to shareholders scaled by total assets at the beginning of the year (Lee and Suh, 2011). ROA is calculated as earnings before interest and taxes (EBIT) divided by total assets at the beginning of the year (Lee and Suh, 2011). Chay and Suh (2009) indicate that ROAVOL, a proxy of cash flow uncertainty, is an important factor for firms to decide whether to pay out and how much to pay. They find that firms with lower ROAVOL are more likely to initiate dividends in the following year, and therefore managers incorporate cash flow uncertainty into their dividend decisions. Following Chay and Suh (2009) and Lee and Suh (2011), ROAVOL is measured as the standard deviation of return on assets over the most recent four years, including the current fiscal year.
In addition to the firm-level characteristics, cross-country differences also need to be controlled. Following Haw et al. (2013), the logarithm of Gross Domestic Product (GDP) per capita (LnGDP) is incorporated to control for cross-country differences in economic development. This variable can be accessed on the World Bank website.
Table A.1 provides detailed definitions, measurements and sources for all the variables included in this paper.
5.4. Logistic regression analysis
The main purpose of this paper is to explore the relationship between the level of cash
holdings (and free cash flow) and the probability of a firm buying back stocks. Moderators are
included to test whether a firm’s growth opportunity and investor protection can moderate this
relationship. Since the dependent variable in this research is binary, which means it can only
take the value of 0 or 1, it is not appropriate to use a standard panel data regression, so the
appropriate model for this research is logistic regression. The logit regression model of this
23
study takes the following form:
𝑃
𝑖𝑡= 𝑃𝑟𝑜𝑏(𝑌
𝑖𝑡= 1|𝑋
𝑖𝑡−1) =
𝑒𝑙𝑜𝑔𝑖𝑡(𝑝𝑖𝑡)1+ 𝑒𝑙𝑜𝑔𝑖𝑡(𝑝𝑖𝑡)
=
11+𝑒−𝑙𝑜𝑔𝑖𝑡(𝑝𝑖𝑡)
(1)
𝑙𝑜𝑔𝑖𝑡(𝑝
𝑖𝑡) = 𝑙𝑛 (
𝑝𝑖𝑡1−𝑝𝑖𝑡
) = 𝛽
0+ 𝛽
1𝑋
𝑖𝑡−1+ 𝜀
𝑖𝑡(2) Where 𝑝
𝑖𝑡the probability that firm 𝑖 repurchases shares in year 𝑡 , 𝑌
𝑖𝑡is the binary
dependent variable which equal to 1 if firm 𝑖 repurchases shares in year 𝑡 , 𝑋
𝑖𝑡−1is explanatory variables in this study, 𝛽 represents the estimated coefficients, 𝜀
𝑖𝑡𝑖𝑠 the error term.
In order to test the hypothesis, the logistic regression model is composed of main independent variables, cash and free cash flows, and other firm characteristics as well as country-level control variables. Therefore, the logistic regression model takes the following forms. Equation (3) is used for testing the hypothesis 1, which predicts 𝛽
1and 𝛽
2to be positive.
𝑙𝑜𝑔𝑖𝑡(𝑝
𝑖𝑡) = 𝛽
0+ 𝛽
1𝐶𝑎𝑠ℎ
𝑖𝑡−1+ 𝛽
2𝐹𝐶𝐹
𝑖𝑡−1+𝛽
3𝑆𝑖𝑧𝑒
𝑖𝑡−1+ 𝛽
4𝑅𝑒𝑡
𝑖𝑡−1+ 𝛽
5𝐿𝑒𝑣
𝑖𝑡−1+ 𝛽
6𝑀𝐶
𝑖𝑡−1+ 𝛽
7𝐶𝐴𝑃𝐸𝑋
𝑖𝑡−1+ 𝛽
8NOPER
𝑖𝑡−1+ 𝛽
9𝐷𝐼𝑉
𝑖𝑡−1+ 𝛽
10𝑅𝑂𝐴
𝑖𝑡−1+ 𝛽
11ROAVOL
𝑖𝑡−1+ 𝛽
12LnGDP
𝑖𝑡−1+ ∑ 𝛽
𝐾𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + ∑ 𝛽
𝑡𝑌𝑒𝑎𝑟 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + 𝜀
𝑖𝑡, (3) In equation (4), the firm-level moderator, growth opportunity, is added. Equation (4) is
used for testing the hypothesis 2, which predicts 𝛽
3, 𝛽
4and 𝛽
5to be negative. Besides, I will further partition the MTB into low, moderate and high groups based on the 1/3 annual MTB value.
𝑙𝑜𝑔𝑖𝑡(𝑝
𝑖𝑡) = 𝛽
0+ 𝛽
1𝐶𝑎𝑠ℎ
𝑖𝑡−1+ 𝛽
2𝐹𝐶𝐹
𝑖𝑡−1+ 𝛽
3𝑀𝑇𝐵
𝑖𝑡−1+ 𝛽
4𝐶𝑎𝑠ℎ
𝑖𝑡−1∗ 𝑀𝑇𝐵
𝑖𝑡−1+ 𝛽
5𝐹𝐶𝐹
𝑖𝑡−1∗ 𝑀𝑇𝐵
𝑖𝑡−1+ 𝛽
6𝑆𝑖𝑧𝑒
𝑖𝑡−1+ 𝛽
7𝑅𝑒𝑡
𝑖𝑡−1+ 𝛽
8𝐿𝑒𝑣
𝑖𝑡−1+ 𝛽
9𝑀𝐶
𝑖𝑡−1+ 𝛽
10𝐶𝐴𝑃𝐸𝑋
𝑖𝑡−1+ 𝛽
11NOPER
𝑖𝑡−1+ 𝛽
12𝐷𝐼𝑉
𝑖𝑡−1+ 𝛽
13𝑅𝑂𝐴
𝑖𝑡−1+ 𝛽
14ROAVOL
𝑖𝑡−1+ 𝛽
15LnGDP
𝑖𝑡−1+ ∑ 𝛽
𝐾𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + ∑ 𝛽
𝑡𝑌𝑒𝑎𝑟 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + 𝜀
𝑖𝑡, (4)
24
In equation (5), two country-level moderators, shareholder protection and law origin, are added. Equation (5) is used for testing the hypothesis 3 and 4, which predicts 𝛽
3, 𝛽
4, 𝛽
5, 𝛽
6, 𝛽
7and 𝛽
8to be positive.
𝑙𝑜𝑔𝑖𝑡(𝑝
𝑖𝑡) =𝛽
0+ 𝛽
1𝐶𝑎𝑠ℎ
𝑖𝑡−1+ 𝛽
2𝐹𝐶𝐹
𝑖𝑡−1+ 𝛽
3𝐴𝑆𝐷
𝑖𝑡−1+ 𝛽
4𝐶𝑎𝑠ℎ
𝑖𝑡−1∗ 𝐴𝑆𝐷
𝑖𝑡−1+ 𝛽
5𝐹𝐶𝐹
𝑖𝑡−1∗ 𝐴𝑆𝐷
𝑖𝑡−1+ 𝛽
6𝑂𝐺𝑁
𝑖𝑡−1+ 𝛽
7𝐶𝑎𝑠ℎ
𝑖𝑡−1∗ 𝑂𝐺𝑁
𝑖𝑡−1+ 𝛽
8𝐹𝐶𝐹
𝑖𝑡−1∗ 𝑂𝐺𝑁
𝑖𝑡−1+ 𝛽
9𝑆𝑖𝑧𝑒
𝑖𝑡−1+ 𝛽
10𝑅𝑒𝑡
𝑖𝑡−1+ 𝛽
11𝐿𝑒𝑣
𝑖𝑡−1+ 𝛽
12𝑀𝐶
𝑖𝑡−1+ 𝛽
13𝐶𝐴𝑃𝐸𝑋
𝑖𝑡−1+ 𝛽
14NOPER
𝑖𝑡−1+ 𝛽
15𝐷𝐼𝑉
𝑖𝑡−1+ 𝛽
16𝑅𝑂𝐴
𝑖𝑡−1+ 𝛽
17ROAVOL
𝑖𝑡−1+ 𝛽
18LnGDP
𝑖𝑡−1+ ∑ 𝛽
𝐾𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + ∑ 𝛽
𝑡𝑌𝑒𝑎𝑟 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 + 𝜀
𝑖𝑡, (5) The variables mentioned in the equations above are defined as before and the detailed information presents in Table A1.
6. Empirical results and analysis 6.1. Descriptive statistics
Table 3 reports the summary statistics of the variables used in this thesis. Panel A shows
the mean, median, maximum, minimum and standard deviation of all non-dummy variables
based on total observations. In panel B, the sample is bifurcated into repurchasing firms and
non-repurchasing firms, and the t-test and z-test for the difference of the mean and median
between repurchasing and non-repurchasing firms is shown, respectively. The average cash
holdings and free cash flow for repurchasing firms are 0.173 and 0.063, which is significantly
higher than for non-repurchasing firms, with 0.166 and 0.030 at a 1% level. The
non-repurchasing firms have a significantly higher market-to-book ratio (MTB) than
repurchasing firms both in the average and the median. In addition, repurchasing firms have
significantly larger SIZE, ROA, NOPER and MC than non-purchasing firms both in both the
average and the median. In contrast, repurchasing firms have significantly lower RET, LEV,
and ROAVOL in both the mean and the median than non-purchasing firms. Panel C shows the
correlation matrix of variables used on the logit regression model to check whether there is
problem of multicollinearity. The coefficients in panel C are all smaller than 0.5, which
indicates that the variables are not highly correlated, suggesting no evidence of
multicollinearity.
25
Table 3. Summary statistics
Panel A: Descriptive statistics of variables
Variable Mean Median Max. Min. Std. Dev.
𝐶𝑎𝑠ℎ
𝑡−10.168 0.112 0.999 0.000 0.173
𝐹𝐶𝐹
𝑡−10.043 0.057 0.846 -2.341 0.135
𝑀𝑇𝐵
𝑡−12.069 1.374 19.568 0.000 2.221
𝑅𝑒𝑡
𝑡−10.168 0.029 15.982 -1.000 0.822
𝐷𝐼𝑉
𝑡−10.015 0.005 0.746 0.000 0.033
𝐶𝐴𝑃𝐸𝑋
𝑡−10.049 0.031 0.798 0.000 0.059
𝑆𝑖𝑧𝑒
𝑡−15.931 5.817 13.680 0.035 1.990
𝐿𝐸𝑉
𝑡−10.203 0.179 0.882 0.000 0.176
𝑀𝐶
𝑡−15.502 5.338 13.525 0.030 2.119
𝑅𝑂𝐴
𝑡−10.046 0.057 1.503 -2.635 0.146
𝑅𝑂𝐴𝑉𝑂𝐿
𝑡−10.045 0.023 2.206 0.000 0.094
𝑁𝑂𝑃𝐸𝑅
𝑡−10.002 0.002 0.783 -0.827 0.048
𝑙𝑛𝐺𝐷𝑃
𝑡−110.300 10.549 11.543 6.081 0.867
ASD 0.586 0.650 1.000 0.170 0.186
Note:
This table presents descriptive statistics for the full sample of 152,655 firm-year observations. Cash is cash and short-term investment to total assets in the year prior to share buybacks, FCF is calculated as Lehn and Poulsen (1989), MTB is market value of equity plus book value of debt divided by book value of assets. Size is the logarithm of total assets in U.S. dollar. Ret is annual stock return. Lev is total debt divided by total assets. MC is market price at the year-end times common shares outstanding in U.S dollars.CAPEX is capital expenditure divided by total assets. NOPER is non-operating income divided by total assets. DIV is cash dividend divided by total assets. ROA is earning before interest and taxes divided by total assets. ROAVOL is the standard deviation of ROA over the most recent four years including the current year. LnGDP is the logarithm of the per capita Gross Domestic Product in U.S dollars. ASD is an index defined by Djankov et al. (2008).All independent variables are calculated at the end of the prior fiscal year and displayed with a subscript of t-1. The detailed definitions of the variables are provided in Table A.1