Issuing and repurchasing: The influence of mispricing, corporate life cycle and financing waves

J. of Multi. Fin. Manag. 22 (2012) 66–81

Contents lists available at SciVerse ScienceDirect

Journal of Multinational Financial Management journal homepage: www.elsevier.com/locate/econbase

Issuing and repurchasing: The influence of mispricing, corporate life cycle and financing waves Bruce Seifert a,∗, Halit Gonenc b,1 a Department of Business Administration, College of Business and Public Administration, Old Dominion University, Norfolk, VA 23529-0221, United States b Department of Economics, Econometrics and Finance, Centre for International Banking, Insurance and Finance, Faculty of Economics and Business, University of Groningen, P.O. Box 800, 9700 AV Groningen, The Netherlands

a r t i c l e

i n f o

Article history: Received 16 February 2011 Accepted 13 February 2012 Available online 21 February 2012 JEL classification: G32 G15 Keywords: Equity financing Debt financing Repurchases Mispricing Life cycle Financing waves

a b s t r a c t This paper examines the impact of mispricing, corporate life cycle, and financing waves on the debt/equity decision when firms (1) acquire funds and (2) repurchase funds by using a large international data set from 47 countries for the period 1984–2006. Our results support the mispricing hypothesis and the corporate life cycle hypothesis for both the acquisition of new funds and the repurchase of funds. However, our findings are consistent with the financing wave hypothesis only for repurchases of firms residing in common law countries as well as market-based countries. © 2012 Elsevier B.V. All rights reserved.

1. Introduction For a long time the trade-off and pecking order theories were the leading explanations for firms’ financial policies. More recently, a number of other hypotheses have been advanced. This paper examines three of these relatively new approaches—mispricing, corporate life cycle, and financial waves on the decisions (1) to issue equity or debt and (2) to repurchase equity or debt. ∗ Corresponding author. Tel.: +1 757 683 3552; fax: +1 757 683 5639. E-mail addresses: [email protected] (B. Seifert), [email protected] (H. Gonenc). 1 Tel.: +31 0 50 363 4237; fax: +31 0 50 363 7356. 1042-444X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.mulfin.2012.02.001

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

67

Of the three theories, mispricing has been studied the most, suggesting that managers take advantage of temporary deviations of a firm’s stock price from its “true” value to implement financing policies. Managers issue (repurchase) stock when the stock price is viewed as abnormally high (low). Researchers have noted high stock returns prior to stock issues and low returns after stock equity issues. Both facts are consistent with firms taking advantage of mispricing (Hovakimian et al., 2001; Baker and Wurgler, 2000, 2002; Kim and Weisbach, 2008; Loughran and Ritter, 1995).2 Kim and Weisbach (2008) also note that firms with high market to book ratios keep a higher percentage of the proceeds from a seasoned equity offering in cash than firms with low market to book ratios. This suggests that firms with high market to book ratios are taking advantage of mispricing by issuing stock and using some of the funds later. Additional support for mispricing comes from the survey done by Graham and Harvey (2001) where executives state the importance of mispricing in equity issuance. Researchers have documented low returns prior to repurchases and high returns subsequent to repurchases (Dittmar, 2000; Hovakimian et al., 2001; Skinner, 2008; Peyer and Vermaelen, 2009), results consistent with mispricing. The survey research by Brav et al. (2005) also documents the importance of mispricing to executives in the repurchase decision.3 While most of the evidence is in favor of mispricing, some studies minimize its importance. Schultz (2003) and Carlson et al. (2006) present models that explain the return behavior of IPOs that are not based on mispricing. DeAngelo et al. (2010) argue that firms issue SEOs to overcome liquidity issues and not principally to exploit mispricing. McLean (2011) posits that firms increasingly are issuing stock to save cash as their precautionary needs have increased. The financial wave approach has been articulated as another theory to explain corporate financial policies. Dittmar and Dittmar (2008) propose that stock issues and repurchases are part of the same economic expansion period and have very little to do with misvaluation. Economic growth results in an increase in demand for funds and a relative (to debt) reduction in the cost of equity. Firms respond by issuing equity which occurs early during economic growth. In contrast, repurchases happen near the end of the same cycle when cash flows are still high but the need for new funds is generally less. The corporate life cycle theory (discussed in DeAngelo et al., 2010) suggests that firms generally issue and repurchase at certain times of their life cycle. Young firms with lots of investment opportunities but little earnings issue stock. Older firms with fewer investment opportunities but greater cash flows will tend to fund internally and distribute funds as dividends and/or repurchases. DeAngelo et al. (2010) find some support for the corporate life cycle theory but they emphasize that this theory is not a stand-alone theory that can explain firms’ issuance decisions. Studies examining mispricing, financing waves, and corporate life cycle on share issuance or repurchase usually examine whether external equity increases or decreases as a result of certain external or internal events. We investigate mispricing, financing waves, and corporate life cycle hypotheses using a different approach. We ask instead whether external equity increases or decreases relative to changes in debt. Does the share of new external equity to total net new external funds change as a result of economic growth, prior or past returns, or the age of the firm? We feel that if mispricing, financing waves or the corporate life cycle can help to explain firms’ issuance and repurchase decisions then they should be able to explain not only an absolute increase in equity but more importantly a relative increase in equity. If both debt and external equity generally increase during periods of economic growth, our findings can easily differ from previous studies if debt increases more than equity. Our approach is as follows: We first divide all yearly firm observations into two separate groups—(1) firms that are net issuers of funds (the sum of net debt and net equity is greater than zero) or (2) firms that are net repurchasers of funds (the sum of net debt and net equity is less than zero). It is important to know whether a firm in a given year is a net issuer of funds or a net repurchaser because the motivations are clearly very different. In one case, the firm is acquiring additional funds and in the

2 A related line of research studies the ability of stock issues to predict subsequent returns. Pontiff and Woodgate (2008) examine this issue for US firms while McLean et al. (2009) and Wang (2011) use an international sample. Collectively the research indicates that share issuance has a negative relationship with future returns. This line of research concentrates on the issuance/subsequent return while mispricing generally looks at returns prior to as well as post issuance or repurchase. 3 For a recent study of European stock repurchases (abnormal returns and determinants of the size of the repurchases) for French, German, Italian, and British firms see Lee et al. (2010).

68

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

other case, the firm is returning funds to its suppliers. In our study, a firm for a particular year, for example, could be a net issuer of stock yet be classified as a net repurchaser of funds because the amount of debt repurchased was larger than the amount of equity issued. Our study also investigates whether these theories apply in different legal regimes (common and civil law). Legal regimes vary in their laws and the enforcement of those laws regarding the protection of minority shareholders and creditors. If minority shareholders and creditors believe that the funds they provide will not be expropriated or wasted and their rights in general (such as those when the firm has financial difficulties) are well protected, then these suppliers of funds will be more willing to provide funds and at a cheaper cost. Firms will benefit from having more funds available and also at cheaper rates. Beck and Levine (2002) document that economic growth is favorably impacted by legal regimes that protect outside investors well. Common law countries (as opposed to civil law countries) have in general more favorable laws for minority shareholders and creditors (La Porta et al., 1998).4 We also address whether these theories apply in both market-based and bank-based countries. We control for endogeneity. The amount of investment influences the amount of new funds raised and quite possibly the mix of funds raised. For empirical support see Gatchev et al. (2009, 2010), and Kim and Weisbach (2008). Likewise, repurchases often substitute for dividends (see Skinner, 2008). Our data set reveals that firms are often borrowing new long-term funds at the same time they are reducing the amount of other long-term borrowings. When firms are net issuers of funds, firms repurchase relatively little equity or preferred stock. However, when these firms are net repurchasers of funds, they frequently issue new common equity and/or preferred stock at the same time. Firms often are raising new funds via equity (debt) at the same time they are retiring/converting/repurchasing debt (equity). We find that firms in all of our samples issue relatively more stock after periods of high-adjusted returns and, with the exception of the US, issue relatively more equity prior to low-adjusted returns. Companies generally repurchase more equity after low-adjusted returns and before high-adjusted returns. These results are consistent with mispricing. We find little evidence that stock issues occur more often in periods of economic expansions. In fact, we find just the opposite. However, we do observe that stock repurchases for firms residing in common law countries and also in market-based countries occur relatively more at the end of economic expansions, a fact consistent with Dittmar and Dittmar (2008). We find support for the life cycle hypothesis in that firms generally issue equity relatively more in their early years and repurchase equity more in their mature years. Our findings suggest that increases in investment are funded relatively more with debt. We also see that firms in both their issue and repurchase decisions move toward target capital structures. In particular, we observe that firms move toward industry norms. As expected, larger firms issue relatively more debt and repurchase more equity. The rest of the paper is as follows: In Section 2, we present our hypotheses and in Section 3, we discuss data sources and methodology. In Section 4, results are given and finally, in Section 5, conclusions are offered. 2. Hypotheses This paper tests whether mispricing, corporate life cycle and financing waves can explain part of the issuing and repurchasing decisions of firms. In particular, we inquire whether these theories can account for the relative change in equity in the issuing and repurchasing decisions. If mispricing is an important influence on the decisions to issue and repurchase, then there should be a positive relationship between prior stock returns and the percent of issuing equity and a negative relationship between prior returns and the percent of equity repurchases. Post returns should be negatively associated with the percent of issuing equity and positively related to the percent of equity repurchases.

4 Differences in laws and the enforcement of those laws for minority shareholders and creditors have also been linked to valuations (La Porta et al., 2002), dividends (La Porta et al., 2000), and ownership concentrations (La Porta et al., 1999).

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

69

If the corporate life cycle theory is correct, we should see a negative (positive) relationship between the stage of the firm’s development and stock issuance (repurchase). We use the age (natural log) of the firm as our measure of the life cycle of the firm. On the other hand, if financing waves are the primary influence of equity repurchases and issues, we should find a positive relationship between GDP growth and equity issues and between GDP growth and repurchases. Dittmar and Dittmar (2008) argue that stock issues occur early in the expansionary cycle and repurchases later in the same cycle. Therefore, we should see a positive association between issues and the first year or two of an expansionary period and a positive relationship between repurchases and the latter years of an expansionary period. We believe that if these hypotheses are true, then they should apply whether the firm is located in a bank or market-based system or whether the firm is domiciled in a common or civil law country. It may be that more total equity will be issued under one economic regime or under one legal system than another or that the total amounts of funds raised will be different between economic regimes or legal systems. We do not believe, however, that if these hypotheses are correct that they should be applicable only to some economic regimes (market or bank-based) or only to some legal systems (common or civil) but not to others. 3. Data and methodology 3.1. Definition of net issues and net repurchases We use the Equity Share in Net Issues (ESNI) to represent issues as well as the Equity Share in Net Repurchases (ESNR) to represent repurchases. In both cases, this equals the ratio Et /(Et + Dt ) where Et is the net amount of equity in year t and Dt is the amount of net debt in year t. Net equity is the difference between the “Net proceeds from the Sale/Issue of Common and Preferred” and “Stocks Purchased, Retired, Converted, and Redeemed”. Net Debt is the difference between “Long-Term Borrowings” and “Reduction in Long-Term Debt” (definitions from Worldscope). We use net numbers as opposed to gross numbers because, for example, if the firm issues 10 million dollars of new stock and repurchases 2 million dollars of stock, the firm has the use of only 8 million dollars (the net number). Baker and Wurgler (2000) use a similar ratio [equity issues/(equity issues + debt issues)] but they use both gross equity and gross debt instead of net numbers. For ESNI, we require the denominator (Et + Dt ) to be greater than zero (a net inflow of external funds) and for ESNR, the denominator (Et + Dt ) must be less than zero (a net outflow of funds). It is important to note that when the denominator is positive (negative) it does not mean that both net equity and net debt are positive (negative). For the case of ESNI (ESNR) either net debt or net equity can be negative but the sum must be positive (negative). Our metric is not bounded in either direction. In the case of ESNI, when net equity is negative and net debt is positive, the ratio will be negative since the denominator is positive. In theory, the ratio could be a large negative number if the positive value of debt is just slightly bigger in absolute value than the negative value of equity. On the other hand, when net debt has a negative value and net equity has a positive value the ratio will be positive and again is not bounded. A similar situation occurs for the case of ESNR (the denominator in this case is negative). As explained below, both ESNI and ESNR are winsorized to avoid extreme outliers in our regressions. Table 1 gives the joint frequencies of net equity and net debt for all observations (Panel A), for ESNI (Panel B) and for ESNR (Panel C). The idea that firms raise or repurchase all their funds using only debt or equity in a given year is not a realistic picture. Of the 79,593 cases in Panel A where the firm issued/repurchased debt or equity or both, only 40% of the time did the firm confine itself to just equity or debt (when there was a non zero amount in one market and a zero amount in the other).5 In other words, 60% of the time firms are active in both debt and equity markets in the same year. The percentage of observations where firms are issuing (repurchasing) both equity and debt is 17%

5 When E is not zero, but D is zero, the total frequency is 9398 (753 + 2287 + 4541 + 1817). When D is not zero, but E is zero, the total frequency is 22,612 (3122 + 9927 + 5576 + 3987). The total number of these cases is 40% [(9398 + 22,612)/79,593].

70

Table 1 Joint frequencies of net equity and net debt. Net debt (D)

E ≤ −0.05

Net equity (E) N

%

N

E=0

E ≥ 0.05

0 > E < 0.05

N

%

N

%

N

%

N

%

3.7 0.3 1.2 0.9 0.6 1.4

15,121 1418 6168 2287 3171 2077

19.0 1.8 7.7 2.9 4.0 2.6

22,612 3122 9927 0 5576 3987

28.4 3.9 12.5 0.0 7.0 5.0

29,608 3130 11,137 4541 5673 5127

37.2 3.9 14.0 5.7 7.1 6.4

9308 1489 3048 1817 1265 1689

11.7 1.9 3.8 2.3 1.6 2.1

0.8

4280

10.2

9563

22.7

19,178

45.5 9.1 10.8 13.5 12.2

8820 1001 3048 1817 1265 1689

20.9 2.4 7.2 4.3 3.0 4.0

27.9 8.4 19.5

488 488

1.3 1.3

0.8 7.0 0.6 2.6 2.0 1.2 0.6

2203 2077

5.2 4.9

5576 3987

13.2 9.5

3837 4541 5673 5127

10,841 1418 6168 2287 968

29.0 3.8 16.5 6.1 2.6

13,049 3122 9927

34.9 8.3 26.5

10,430 3130 7300

This table presents the frequencies for combinations of net equity and net debt. The data is collected from Worldscope for the period 1984–2006. E (Net Equity) is the difference between ‘Net Proceeds from Sale/Issue of Common and Preferred Stocks’ and ‘Stocks Purchased, Retired, Converted, Redeemed’. D (Net Debt) is the difference between ‘Long Term Borrowings’ and ‘Reduction in Long-term Debt’. E and D are scaled by Total Assets. ESNI (Equity Share in Net Issues) is the ratio of net equity to the sum of net equity and net debt, when the total is positive. ESNR (Equity Share in Net Repurchases) is the ratio of net equity to the sum of net equity and net debt, when the total is negative.

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

Panel A: frequencies for all observations Total 79,593 100.0 2944 9374 11.8 215 D ≤ −0.05 31,261 39.3 981 −0.05 > D < 0 9398 11.8 753 D=0 16,130 20.3 445 0 > D < 0.05 D ≥ 0.05 13,430 16.9 550 Panel B: frequency of ESNI {[E/(E + D)] when (E + D) > 0} 42,160 100.0 319 Total 1001 2.4 D ≤ −0.05 6885 16.3 −0.05 > D < 0 6358 15.1 D=0 14,717 34.9 0 > D < 0.05 13,199 31.3 319 D ≥ 0.05 Panel C: frequencies of ESNR {[E/(E + D)] when (E + D) < 0} 37,433 100.0 2625 Total 8373 22.4 215 D ≤ −0.05 24,376 65.1 981 −0.05 > D < 0 3040 8.1 753 D=0 1413 3.8 445 0 > D < 0.05 231 0.6 231 D ≥ 0.05

−0.05 > E < 0 %

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

71

(11%).6 An interesting observation from Panel A of Table 1 is the frequency that firms are issuing one security at the same time they are retiring another. Our data has 25,0477 instances (31%) of positive amounts of one security and negative amounts of another. This makes sense if the firm is trying to alter its capital structure. Our metrics of ESNI and ESNR allow us to study all the different positive and negative combinations of equity and debt. The only observations omitted are those where the firm did not have a net issue or a net repurchase (the denominator was zero). In contrast, some studies model the choice between debt and equity and firm observations are classified as either debt or equity, but not both. Cases where firms issued (repurchased) both debt and equity in the same year are omitted (see, for example, Hovakimian et al., 2001). 3.2. Data sources and models Financial data are gathered from Worldscope from 1984 to 2006 for 47 countries. We eliminate financial firms and utilities, and thus avoid issues about regulatory influence on these firms. To calculate a firm’s age, we first use the information from Worldscope concerning the year the company was founded. However, this information is limited for all countries, including the US and the UK. Therefore, we search for founding years from online sources, company websites, etc. for firms with missing founded years. For the final sample, we exclude those firms without a founded year. Countries are classified by legal systems and market structure based on Demirguc-Kunt and Levine (2001) and Demirguc-Kunt and Maksimovic (2002). Given that Worldscope makes occasional reporting errors and that there are some extreme values for some of our variables, we have chosen to winsorize our data. The top and bottom 1% of the values for a particular variable are set equal to the values for the 99 and 1%, respectively for that variable. To examine the impact of mispricing, life cycle, and financing waves on ESNI, we estimate the following equation: ESNIit = a + bBUSCONj,t−1 + cMISPi,t−1 + dMISPi,t+1 + f LNAGEit + gPINVit + hCONT + eit

(1)

This model examines the effects of business conditions (BUSCON), mispricing (MISP) in the previous year (t − 1) and the year after (t + 1), the logarithm of firm’s age (LNAGE), predicted investments (PINV), and several control variables (CONT) on ESNI. Table 2 lists all the variables used in this study and gives exact definitions for each of them. We do not use investments in Eq. (1) since we view the firm as making joint financing and investment decisions (for empirical support see Gatchev et al., 2009, 2010). If financing and investment decisions were jointly determined then using investments in Eq. (1) would result in the error term being correlated with investments. We also conducted endogeneity tests and they indicated that investment (as opposed to predicted investment) in Eq. (1) was an endogenous variable (see Wooldridge chapter 16, 2000). The predicted values for investment (PINV) were estimated by the following equation: INVi,t = a + bBUSCONj,t−1 + cMISPi,t−1 + dMISPi,t+1 + f LNAGEi,t + gCFLOWi,t−1 + hMtBi,t−1 + iCONT + ui,t

(2)

INV is a function of the lag of the following variables: cash flows (CFLOW), the market to book ratio (MtB), BUSCON, MISP (both for years t − 1 and t + 1), and LNAGE and control variables.

6 Both E and D are positive for 13,754 (5673 + 5127 + 1265 + 1689) observations, and both E and D are negative for 8782 (215 + 981 + 1418 + 6168) observations. The corresponding percentages are 17% (13,754/79,593) and 11% (8782/79,593), respectively. 7 When E is positive and D is negative, there are 18,804 (3130 + 11,137 + 1489 + 3048) cases, and when D is positive and E is negative, there are 6243 (445 + 550 + 3171 + 2077) cases. The total is 25,047 and that, corresponds to 31% of the total number of observations.

72

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

Table 2 Variables definitions. Definitions Variables E

D ESNI ESNR ST Debt INV PINV DIV PDIV BUSCON H1 H2 H3 MISP LNAGE DEVFT SIZE MtB CFLOW TANG CASH Country classifications Market and bank-based Common and civil

Net Equity (new equity issues − repurchases)/total assets or [(net proceeds from sale or issue of common and preferred stocks) − (stocks purchased, retired, converted, redeemed)]/total assets Net Debt [(long-term borrowings) − (reduction in long-term debt)]/total assets Equity Share in Net Issues [E/(E + D)], where (E + D) > 0 Equity Share in Net Repurchases [E/(E + D)], where (E + D) < 0 Short-term Borrowings (increase or decrease in short-term debt and current portion of long-term debt/total assets) Investments (capital expenditures/total assets) Predicted Investments [Predicted value of the ratio (capital expenditures/total assets) using Eq. (2)] Dividends (cash dividends/total assets) Predicted Dividends [Predicted value of the ratio (cash dividends/total assets) using Eq. (4)] Business Conditions (the annual growth rate of real GDP) Dummy variable for the first year of high expansionary period Dummy variable for the second year of high expansionary periods Dummy variable for the third year of high expansionary periods Mispricing (the difference between the annual return for the firm and the local market return) The natural logarithm of firm age (observation’s year − founded year of the firm) Deviation from Target Leverage (firm leverage − 4 digit industry average leverage Leverage (total shareholders’ equity/total assets)) The natural logarithm of total assets in US dollars Market to Book Ratio (market value of equity + book value of debt)/(total assets) Cash Flows (net Income + depreciation)/(total assets) Tangibility (net property, plant, and equipment/total assets) Cash Ratio (cash and short term investments/total assets) Classifications of economic systems (Demirguc-Kunt and Levine, 2001; Demirguc-Kunt and Maksimovic, 2002) Classifications of legal systems (Demirguc-Kunt and Levine, 2001; Demirguc-Kunt and Maksimovic, 2002)

In a similar fashion, we examine the impact of financing waves, corporate life cycle and mispricing on ESNR in Eq. (3). ESNR i,t = a + bBUSCONj,t−1 + cMISPi,t−1 + dMISPi,t+1 + f LNAGEi,t + gPDIVi,t + hCONT + ei,t

(3)

Since the dividend and repurchase decisions are likely jointly determined, it would be incorrect to use dividends in Eq. (3). Using dividends in Eq. (3) would have resulted in the error term being correlated with the dividend variable.8 We estimate the following equation for dividends (Eq. (4)) to get the predicted values for dividends (PDIV). DIVi,t = a + bBUSCONj,t−1 + cMISPi,t−1 + dMISPi,t+1 + f LNAGEi,t + gCFLOWi ,t−1 + hCONT + ui,t

(4)

We use a variety of controls in our models: Our first control variable is DEVFT (deviation from target), which is the industry adjusted firm’s leverage. The industry average is used as a proxy for the firm’s target equity ratio. If firms try to move toward their target equity ratio, we would expect the coefficient for this variable to be negative in Eq. (1) and positive in Eq. (3).

8

Endogeneity tests indicate that dividends (as opposed to predicted dividends) are an endogenous variable in equation 3.

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

73

SIZE (size) is our second control variable. In Eq. (1) we expect the coefficient to be negative since previous research has shown that larger firms use more debt.9 We include size in Eq. (3) because Dittmar (2000) found that large firms were more likely to repurchase than small firms.10 We also include TANG (tangibility) as a control variable. The more tangible the assets are, the more collateral the firm has and the more likely firms will issue debt. Firms with more tangible assets should also be more willing to repurchase stock. In addition to those three control variables, we include CASH (cash) in the ESNR equation only. Holding investment constant, presumably the more cash a firm has the more it could repurchase. We estimate our equations using panel-data models with the fixed-effects (within) regression estimator, which allows for endogeneity. Estimations depend on two-stage least-squares generalizations of simple panel-data estimators for exogenous variables. 4. Results 4.1. Sample statistics Table 3 provides descriptive statistics for all our independent variables. Panel A looks at the observations for ESNI while Panel B does the same for ESNR. US observations account for a little less than half of all observations for both ESNI and ESNR. Our samples reveal that our average firms are healthy and average about 45 years old. Country growth rates average a little over 3% per year. Most of our observations (both net issues in Panel A and repurchases in Panel B) occur in the later years of our study. The mean of ESNI (Panel A) is 0.522. In unreported results, we find that when firms are net issuers, net equity averages 5.6% of total assets and net long-term debt is 4.3%. Internal funds are also 5.6% of total assets. ESNR has a mean of 0.063 (Panel B). When firms are net repurchases, it is long-term debt (mean = −0.034) that is primarily repurchased and not equity (mean = −0.008) that is repurchased (results not reported). An examination of our data reveals that firms are consistently issuing new debt and reducing the amount of other debt all in the same year. This fact holds whether they are net issuers of funds or net repurchases of funds. The pattern is different for equity. When firms are net issuers of funds, they rarely repurchase equity (in a relative sense). On the other hand, when companies are net repurchasers of funds, they repurchase as well as issue equity at the same time. 4.2. Regression results Tables 4 and 5 present our main regression results. In Table 4, we give our findings for the regressions involving ESNI and for ESNR in Table 5. Panel A presents the results based on the full sample of observations and Panel B gives the findings for various samples. In model 1 of Panel A, we give the simple regression statistics using GDP growth for the proxy for the variable BUSCON as the single independent variable. In model 2, we present the regression findings where we isolate those periods that were clearly periods of high expansionary activity (in the top 25% of real GDP growth rates). We label these periods H1–H3. H1 stands for the first year of high expansionary periods and H3 for the third year of the expansionary periods. If Dittmar and Dittmar (2008) are correct, then we should see a significant positive coefficient for H1 (possibly H2) in the regression involving ESNI because these authors argue that issues occur early in the expansionary periods. Similarly in the model involving ESNR, H3 (possibly H2) should be significantly positive since repurchases should take place at the end of the expansionary periods. Models 3 and 4 show the findings for the regression that use as the sole independent variables our measures of mispricing (previous year’s stock returns and next year’s stock

9 According to the trade-off theory, larger firms are generally more diversified and should therefore have less bankruptcy risk. 10 An alternative view of size is that it is a proxy for information asymmetry. Small firms presumably have more information asymmetry and would be less likely to issue stock and more likely to repurchase stock.

74

Table 3 Descriptive statistics of variables. All samples

US

Market-based

Bank-based

Common-law

Civil-law

N = 42,160

N = 19,540

N = 33,870

N = 8290

N = 30,615

N = 11,545

0.635 3.268 0.070 0.184 0.052 36.15 0.024 5.124 2.347 0.026 0.282

0.638 3.640 0.049 0.019 −0.065 23.00 0.031 5.002 1.632 0.085 0.221

0.590 3.380 0.073 0.161 0.049 40.44 0.014 5.208 2.082 0.050 0.325

0.351 3.640 0.052 0.018 −0.051 26.00 0.017 5.104 1.478 0.093 0.279

0.246 3.219 0.069 0.102 0.043 50.93 −0.009 6.482 1.483 0.077 0.338

0.000 2.821 0.052 0.024 −0.025 44.00 −0.008 6.320 1.200 0.076 0.314

0.619 3.362 0.073 0.170 0.051 39.44 0.018 5.117 2.164 0.318 0.047

0.487 3.640 0.051 0.024 −0.050 25.00 0.022 4.993 1.536 0.264 0.094

0.266 3.313 0.071 0.095 0.041 50.65 −0.012 6.366 1.435 0.352 0.078

All sample

US

Market-based

Bank-based

Common-law

Civil-law

N = 37,433

N = 16,920

N = 28,003

N = 9430

N = 25,103

N = 12,330

Panel B: ESNR {[Et /(Et + Dt )] when (Et + Dt ) < 0} ESNRt 0.063 0.000 3.020 3.209 BUSCONt−1 (%) 0.014 0.007 DIVt MISPt−1 0.063 −0.010 0.124 0.018 MISPt+1 40.00 AGE 50.71 −0.006 −0.006 DEVFTt−1 5.740 5.649 SIZEt−1 MtBt−1 1.481 1.175 0.083 0.089 CFLOWt−1 TANGt−1 0.333 0.303 0.134 0.083 CASHt−1

0.147 3.161 0.012 0.038 0.140 47.21 −0.004 5.575 1.690 0.079 0.301 0.142

0.000 3.640 0.001 −0.038 0.019 35.00 −0.001 5.446 1.315 0.094 0.261 0.071

0.056 3.250 0.015 0.047 0.136 49.38 −0.008 5.491 1.558 0.086 0.335 0.129

0.000 3.467 0.007 −0.032 0.023 35.00 −0.006 5.340 1.227 0.096 0.301 0.069

0.081 2.339 0.011 0.109 0.088 54.68 −0.001 6.480 1.254 0.072 0.329 0.148

0.000 2.026 0.006 0.045 0.008 52.00 −0.005 6.335 1.064 0.068 0.308 0.117

0.058 3.266 0.015 0.052 0.140 48.50 −0.009 5.437 1.604 0.328 0.087 0.131

0.000 3.540 0.007 −0.028 0.026 34.00 −0.006 5.270 1.261 0.290 0.098 0.067

0.073 2.520 0.012 0.085 0.091 55.21 −0.002 6.357 1.232 0.345 0.074 0.141

0.000 2.833 0.054 0.009 −0.033 42.00 −0.009 6.201 1.150 0.334 0.077

0.000 2.268 0.007 0.021 0.006 50.00 −0.005 6.216 1.044 0.326 0.072 0.108

This table reports mean and median values of variables. The data is collected from Worldscope and Datastream for the period 1984–2006. ESNIt (Equity Share in Net Issues) is the ratio of E (Net Equity) to the sum of E and D (Net Debt), when (E + D) is positive. ESNRt (Equity Share in Net Repurchases) is the ratio of E to the (E + D), when (E + D) is negative. INVt and DIVt are the ratios of capital expenditures and cash dividends to total assets, respectively. All other variables are lagged with respect to the variables defined above. BUSCONt−1 is the growth rate of real GDP. MISPt−1 and MISPt+1 measure mispricing and are the differences between the return for the firm and the local market index return. DEVFT is the ratio of total shareholders’ equity to the sum of shareholders’ equity and total liabilities minus the 4-digit industry average. SIZEt−1 is the natural logarithm of total assets in US dollars. MtBt−1 is the ratio of the market value of equity and the book value of debt to total assets. CFLOWt−1 is calculated as Net Income plus Depreciation and is scaled by total assets. TANGt−1 is the ratio of net property, plant, and equipment to total assets. CASHt−1 is the ratio of cash and short term investments to total assets. Firm AGE is calculated by the difference between founded year and year of the first available data.

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

Panel A: ESNI {[Et /(Et + Dt )] when (Et + Dt ) > 0} 0.522 0.140 ESNIt BUSCONt−1 (%) 3.348 3.467 INVTt 0.073 0.052 MISPt−1 0.149 0.019 MISPt+1 0.048 −0.044 42.51 29.00 AGE 0.010 0.011 DEVFTt−1 SIZEt−1 5.459 5.373 MtBt−1 1.964 1.404 CFLOWt−1 0.056 0.088 TANGt−1 0.328 0.286

Table 4 Fixed effect 2SLS regressions on net issues. Panel A: total sample

PINV BUSCON

1

2

3

4

5

6

7

8

9

10

0.691a [0.044] −1.454b [0.700] −0.019a [0.004]

0.667a [0.047] −1.852a [0.653]

0.815a [0.049] −4.174a [0.674]

0.678a [0.049] −2.135a [0.665]

1.169a [0.161] −2.263a [0.696]

1.724a [0.175] −4.689a [0.842] −0.010b [0.004]

1.719a [0.175] −4.832a [0.779]

1.485a [0.187] −2.470a [0.929] −0.016a [0.004]

1.528a [0.190] −2.808a [0.953] −0.015a [0.004]

1.481a [0.188] −2.831a [0.859]

−0.019 [0.016] −0.067a [0.021] −0.121a [0.024]

H1 H2 H3

0.067a [0.007]

MISPt−1

−0.017b [0.008]

MISPt+1

−0.144a [0.036]

LNAGE

0.066a [0.007] −0.018b [0.008] −0.250a [0.039]

−0.021 [0.016] −0.041c [0.022] −0.093a [0.025] 0.065a [0.007] −0.018b [0.008] −0.253a [0.038]

DEVFT SIZE TANG R2 # of Obs.

0.015 42160

0.011 42160

0.016 42160

0.012 42160

0.056 42160

0.059 42160

0.056 42160

0.049a [0.007] −0.023a [0.008] −0.086b [0.035] −0.546a [0.054] −0.071a [0.011] −0.179a [0.060] 0.085 42160

0.057a [0.008] −0.017b [0.008] −0.211a [0.041] −0.523a [0.055]

−0.145b [0.060] 0.054 42160

−0.022 [0.016] −0.055a [0.021] −0.105a [0.024] 0.049a [0.007] −0.021a [0.008] −0.090b [0.035] −0.539a [0.052] −0.072a [0.011] −0.160a [0.060] 0.078 42160

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

Intercept

75

76

Table 4 (Continued ) Panel B: several sub-samples

Intercept PINV BUSCON

H2 H3 MISPt−1 MISPt+1 LNAGE DEVFT SIZE TANG R2 # of Obs. (E + D) ≥ 0.05 (E) ≥ 0.05

US

1.321a [0.168] −0.732 [0.832] −0.014a [0.003]

1.879a [0.276] −4.938a [1.510] −0.019a [0.007]

1.297a [0.171] −0.956 [0.782]

−0.013 [0.013] −0.041b [0.017] −0.055a [0.019] 0.052a 0.052a [0.006] [0.005] −0.035a −0.033a [0.006] [0.006] −0.090a −0.090a [0.031] [0.031] −0.649a −0.646a [0.047] [0.046] −0.094a −0.095a [0.008] [0.008] −0.103b −0.093b [0.046] [0.047] 0.213 0.207 19814 19814 19814 8171

Non-US 1.832a [0.282] −5.174a [1.413]

−0.050c [0.026] −0.035 [0.029] −0.121a [0.030] 0.062a 0.060a [0.010] [0.010] −0.014 −0.014 [0.011] [0.012] a −0.158 −0.158a [0.057] [0.057] −0.392a −0.394a [0.077] [0.075] −0.052a −0.049a [0.015] [0.015] −0.242b −0.236b [0.100] [0.100] 0.091 0.088 19540 19540 9390 4560

1.280a [0.245] −0.123 [1.054] −0.016a [0.004]

1.317a [0.239] −0.644 [0.953]

0.019 [0.020] −0.080b [0.036] −0.019 [0.052] 0.029a 0.032a [0.011] [0.011] −0.037a −0.036a [0.012] [0.012] −0.047 −0.054 [0.044] [0.043] −0.775a −0.765a [0.074] [0.072] −0.103a −0.109a [0.017] [0.017] −0.112 −0.083 [0.072] [0.072] 0.071 0.069 22620 22620 10484 4230

Market based

Bank based

1.572a [0.198] −3.053a [1.016] −0.022a [0.004]

1.137b [0.538] 3.486 [2.155] −0.006 [0.007]

1.555a [0.200] −3.540a [0.949]

−0.011 [0.018] −0.045c [0.023] −0.119a [0.027] 0.049a 0.049a [0.008] [0.008] −0.023b −0.021b [0.009] [0.009] −0.089b −0.093b [0.039] [0.039] −0.499a −0.492a [0.057] [0.056] −0.066a −0.067a [0.012] [0.012] −0.143b −0.121c [0.066] [0.066] 0.082 0.073 33870 33870 16425 7929

1.128b [0.511] 3.519c [1.833]

−0.119a [0.042] −0.150b [0.060] −0.003 [0.060] 0.042b 0.044b [0.020] [0.019] −0.037b −0.037b [0.019] [0.019] −0.105 −0.098 [0.084] [0.084] −1.226a −1.227a [0.160] [0.152] −0.115a −0.120a [0.036] [0.034] −0.013 0.004 [0.182] [0.181] 0.004 0.003 8290 8290 3389 891

Common-law

Civil-law

1.549a [0.198] −2.916a [1.062] −0.021a [0.005]

1.136b [0.525] 1.395 [1.793] −0.015a [0.006]

1.544a [0.199] −3.343a [0.994]

−0.036c [0.019] −0.060b [0.024] −0.119a [0.028] 0.050a 0.049a [0.008] [0.008] −0.022b −0.020b [0.009] [0.010] −0.071c −0.078b [0.040] [0.039] −0.505a −0.497a [0.060] [0.059] −0.070a −0.071a [0.012] [0.012] −0.177b −0.157b [0.069] [0.069] 0.078 0.072 30615 30615 14831 7465

0.039b [0.015] −0.027c [0.015] −0.138c [0.081] −0.910a [0.128] −0.078b [0.031] 0.182 [0.173] 0.011 11545 4983 1355

1.188b [0.503] 0.614 [1.555]

0.052c [0.031] −0.019 [0.048] −0.04 [0.053] 0.042a [0.015] −0.026c [0.015] −0.139c [0.080] −0.889a [0.123] −0.084a [0.029] 0.172 [0.170] 0.015 11545

Estimations depend on two-stage least-squares generalizations of simple panel-data estimators for exogenous variables. Dependent variable is ESNI (Equity Share in Net Issues) is the ratio of E (Net Equity) to the sum of E and D (Net Debt) when (E + D) is positive. INV is an endogenous variable and the ratio of capital expenditures to total assets. PINV is predicted investment. All other variables (except MISPt+1 ) are lagged with respect to the variables defined above. BUSCON is the growth rate of real GDP. H1, H2, and H3 stand for the first, second, and third year of high expansionary periods, respectively. MISPt−1 and MISPt+1 measure mispricing and are the differences between the return for the firm and the local market index return. LNAGE is the natural logarithm of firm’s age. DEVFT is the ratio of total shareholders’ equity to the sum of shareholders’ equity and total liabilities minus the 4-digit industry average. SIZE is the natural logarithm of total assets in US dollars. TANG is the ratio of net property, plant, and equipment to total assets. The robust standard errors are reported below the estimated coefficients. a, b and c denote significance levels at 1%, 5% and 10% respectively.

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

H1

(E + D) ≥ 0.05

Table 5 Fixed effect 2SLS regressions on net repurchases. Panel A: total sample

Intercept

BUSCON

2

3

4

5

6

7

8

9

10

−0.144a [0.037] 12.147a [2.767] 0.012a [0.004]

−0.130a [0.038] 12.610a [2.686]

−0.167a [0.038] 16.923a [2.714]

−0.149a [0.040] 14.743a [2.797]

−0.817a [0.124] 13.252a [2.681]

−1.018a [0.126] 17.509a [2.907] 0.010a [0.004]

−1.035a [0.126] 17.507a [2.810]

−1.149a [0.137] 6.373c [3.501] 0.013a [0.004]

−0.894a [0.134] 7.043b [3.510] 0.012a [0.004]

−1.155a [0.137] 6.532c [3.385]

0.038c [0.021] 0.120a [0.027] 0.266a [0.031]

H1 H2 H3

−0.101a [0.010]

MISPt−1

0.054a [0.011]

MISPt+1

0.193a [0.033]

LNAGE

−0.098a [0.010] 0.050a [0.011] 0.224a [0.033]

0.056a [0.021] 0.127a [0.028] 0.237a [0.031] −0.091a [0.010] 0.051a [0.011] 0.232a [0.034]

DEVFT SIZE TANG CASH R2 # of Obs.

0.004 37433

0.005 37433

0.005 37433

0.004 37433

0.007 37433

0.007 37433

0.008 37433

−0.084a [0.010] 0.057a [0.011] 0.099a [0.036] 0.921a [0.080] 0.124a [0.014] 0.005 [0.088] 0.118 [0.090] 0.023 37433

−0.095a [0.010] 0.048a [0.011] 0.236a [0.033] 0.857a [0.080]

−0.082 [0.088] 0.063 [0.091] 0.019 37433

0.068a [0.021] 0.134a [0.027] 0.234a [0.031] −0.078a [0.010] 0.058a [0.011] 0.106a [0.036] 0.927a [0.079] 0.125a [0.014] −0.008 [0.088] 0.110 [0.090] 0.024 37433

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

PDIV

1

77

78

Table 5 (Continued ) Panel B: several sub-samples (E + D) ≤ −0.05 Intercept PDIV BUSCON

H2 H3 MISPt−1 MISPt+1 LNAGE DEVFT SIZE TANG CASH R2 # of Obs. (E + D) ≤ −0.05 (E) ≤ −0.05

US

−0.293 [0.128] 24.001a [4.189]

0.013 [0.020] 0.106a [0.025] 0.133a [0.026] −0.007 −0.004 [0.009] [0.010] 0.045a 0.045a [0.011] [0.011] −0.008 −0.007 [0.035] [0.035] 0.626a 0.621a [0.088] [0.088] 0.037a 0.035a [0.013] [0.013] −0.053 −0.046 [0.078] [0.078] a 0.381a 0.385 [0.079] [0.079] 0.064 0.067 10409 10409 10409 2306

b

Non-US

−1.957 [0.265] 36.895b [14.351] 0.046a [0.007] a

−1.769 [0.268] 41.081a [14.061] a

0.107a [0.035] 0.154a [0.037] 0.265a [0.041] −0.104a −0.098a [0.018] [0.018] 0.090a 0.092a [0.019] [0.019] 0.104c 0.078 [0.057] [0.057] 0.809a 0.772a [0.187] [0.185] 0.201a 0.197a [0.023] [0.023] 0.021 0.001 [0.155] [0.156] 0.191 0.152 [0.167] [0.166] 0.034 0.034 16920 16920 5718 2271

−0.621 [0.188] −2.397 [2.298] 0.002 [0.003]

a

Market based −0.668 [0.189] −2.491 [2.250]

a

0.028 [0.025] 0.112b [0.045] 0.142b [0.070] −0.033a −0.032b [0.013] [0.013] 0.040a 0.039a [0.013] [0.013] 0.118b 0.126a [0.048] [0.048] 0.732a 0.738a [0.093] [0.093] 0.050a 0.054a [0.019] [0.019] −0.151 −0.158 [0.102] [0.102] c −0.192 −0.201c [0.116] [0.115] 0.013 0.012 20513 20513 4691 354

−1.214 [0.161] 9.911b [4.463] 0.020a [0.004]

a

Bank based

−1.190 [0.163] 10.487b [4.307]

a

0.066a [0.024] 0.142a [0.030] 0.247a [0.035] −0.097a −0.091a [0.012] [0.012] 0.068a 0.069a [0.013] [0.013] 0.080c 0.086b [0.041] [0.042] 0.870a 0.873a [0.094] [0.093] 0.139a 0.137a [0.016] [0.016] −0.032 −0.045 [0.104] [0.104] b 0.229 0.217b [0.106] [0.105] 0.021 0.023 28003 28003 8693 2607

−1.340 [0.339] −3.742 [3.954] −0.008c [0.005]

a

Common-law

−0.015 [0.017] 0.014 [0.017] 0.288a [0.085] 1.016a [0.148] 0.064b [0.028] 0.099 [0.155] −0.355b [0.165] 0.004 9430 0 0

Civil-law

−1.457 [0.338] −5.019 [3.914]

−1.208 [0.170] 8.561c [5.004] 0.023a [0.005]

0.085c [0.046] 0.03 [0.066] 0.052 [0.075] −0.015 [0.017] 0.014 [0.017] 0.306a [0.084] 1.031a [0.148] 0.070b [0.029] 0.099 [0.156] −0.351b [0.165] 0.003 9430

0.078a [0.026] 0.157a [0.032] 0.245a [0.036] −0.100a −0.094a [0.013] [0.013] 0.066a 0.068a [0.014] [0.014] 0.072 0.079c [0.044] [0.044] 0.893a 0.894a [0.099] [0.098] 0.144a 0.142a [0.017] [0.017] −0.021 −0.04 [0.111] [0.111] b 0.258 0.243b [0.114] [0.113] 0.021 0.023 25103 25103 7888 2508

a

a

−1.176 [0.173] 9.275c [4.813]

a

−1.013a [0.281] 3.675 [3.344] −0.002 [0.004]

−1.045a [0.282] 3.233 [3.306]

0.014 [0.034] −0.037 [0.053] 0.114c [0.065] −0.029c −0.028c [0.015] [0.015] 0.035b 0.035b [0.015] [0.015] 0.218a 0.223a [0.070] [0.070] 0.882a 0.892a [0.123] [0.123] 0.042c 0.044c [0.025] [0.025] 0.053 0.054 [0.134] [0.134] a −0.390 −0.392a [0.139] [0.139] 0.005 0.004 12330 12330 2521 117

Estimations depend on two-stage least-squares generalizations of simple panel-data estimators for exogenous variables. Dependent variable is ESNR (Equity Share in Net Repurchases), which is the ratio of E (Net Equity) to the sum of E and D (Net D), when (E + D) is negative. DIV is the ratio of cash dividends to total assets and is an endogenous variable. PDIV is the predicted dividend. All other variables (except MISPt+1 ) are lagged with respect to the variables defined above. BUSCON is the growth rate of real GDP. H1, H2, and H3 stand for the first, second, and third year of high expansionary periods, respectively. MISPt−1 and MISPt+1 measure mispricing and are the differences between the return for the firm and the local market index return. LNAGE is the natural logarithm of firm’s age. DEVFT is the ratio of total shareholders’ equity to the sum of shareholders’ equity and total liabilities minus the 4-digit industry average. SIZE is the natural logarithm of total assets in US dollars. TANG is the ratio of net property, plant, and equipment to total assets. MtB is the ratio of the market value of equity and the book value of debt to total assets. CASH is the ratio of cash and short-term investments to total assets. The robust standard errors are reported below the estimated coefficients. a, b and c denote significance levels at 1%, 5% and 10% respectively.

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

H1

−0.321 [0.126] 23.523a [4.294] 0.012a [0.004]

b

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

79

returns). In model 5 we test the corporate life cycle hypothesis and expect to observe a significant negative coefficient in Table 4 (issues should occur early in the life cycle of a firm) and a positive coefficient in Table 5 (repurchases should occur in the later stages of the firm’s life cycle). In models 6 and 7, all three hypotheses are combined into one regression. Models 8–10 present tests of the three hypotheses using our full set of controls. In Panel B, we present the results only for the full model for ESNI or ESNR for the following samples: The amount of net issues or repurchases was at least 5% of total assets of the firm [(E + D) ≥ 5% or (E + D) ≤ −5%], the US, non-US countries, market-based countries, bank-based countries, common law countries, and civil law countries. The first model for the results in each sample presents the results using GDP growth as the proxy for Business Conditions while the second model uses dummies for years 1–3 for business expansions. 4.2.1. Results for ESNI The results from Panels A and B of Table 4 provide little support for the notion that equity issues occur relatively more in times of high GDP growth/expansionary periods. In fact, the results suggest just the opposite. BUSCON has a negative sign (usually significant), the opposite of what would be expected if the financial wave hypothesis is correct. The coefficients associated with the different years of an expansion are almost always either significantly negative or insignificant. Panel B clearly shows that these results are the rule across the various samples. The findings for common law countries illustrate this point clearly. All the coefficients for the various years are significantly negative as well as the coefficient for BUSCON. We find strong support for the corporate life cycle hypothesis. Issues should occur early in the life cycle of the firm and thus we should see a negative coefficient on the life cycle variable (LNAGE). The coefficient is always negative and almost, always significantly negative. Our variables for mispricing (MISPt−1 and MISPt+1 ) indicate that equity issues occur more often after periods of high-adjusted stock returns and before periods of low-adjusted stock returns. The coefficient for past returns is always significantly positive. The coefficient for future returns is always significantly negative for all samples except one (US sample). Our overall results in Table 4 indicate that increases in investment (PINV) are funded relatively more with debt than equity. All other things being equal, increases in investment result in decreases in ESNI. The control variables are, in general, in line with prior expectations. The negative coefficient on DEVFT suggests that firms try to move toward their target leverage ratios. The negative coefficient on SIZE suggests that larger firms issue relatively more debt than equity, which is consistent with larger firms being more diversified and having in general less bankruptcy risk. Since SIZE and LNAGE are positively correlated (0.36) we ran our full model (model 9) without size and the results were very similar. The coefficient for TANG is usually significantly negative which indicates that greater the tangibility of the assets, the more debt is used. 4.2.2. Results for ESNR In Panels A and B of Table 5, we present the regression statistics for ESNR. One problem with tests of our hypotheses is that for civil law countries there are relatively few large equity repurchases. For bank-based countries, there are none (see the bottom of Panel B). Our results using the total sample as well the findings for common law countries and marketbased countries support the financial wave theory in that repurchases should occur more often in high economic growth periods. This finding, however, is not observed for bank-based countries. For most of our samples, years two and three have significant positive coefficients that provide further support for the idea that repurchases often occur near the end of the growth period.11

11 The evidence is mixed concerning civil law countries. The coefficient for BUSCON is insignificant but the coefficient for H3 is significantly positive.

80

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

Our findings also support the mispricing hypothesis. According to this theory, equity repurchases should occur more after low-adjusted returns and before high-adjusted returns. This pattern is observed for most samples with the bank-based countries being an exception. If the corporate life cycle hypothesis is correct for repurchases, we should observe a positive coefficient on LNAGE since repurchases should occur mostly for mature firms. All of the coefficients for LNAGE are significantly positive in Panel A and the coefficients in Panel B are generally positive but not always significant. This is the case for common law countries. The impact of dividends on ESNR is significantly positive in Panel A and mixed in Panel B. A positive coefficient for predicted dividends indicates that an increase in dividends is associated with more equity being repurchased. The control variables have in general the expected signs. DEVFT has a positive and significant coefficient across all the samples, which indicates that when firms repurchase they move toward their target debt equity ratios. The coefficient for SIZE is significantly positive in Panel A and in all samples in Panel B and this indicates that larger firms repurchase relatively more equity. The impact of TANG is generally insignificant on repurchases and the sign for CASH varies considerably. 4.3. Inclusion of changes in short-term debt We define issues (repurchases) to be the sum of net equity and net long-term debt in a given year when this sum is greater (less) than zero. We experimented with this definition by defining issues and repurchases to be equal to the sum of net equity and net long-term debt plus any net changes in short-term debt. In unreported results, we find that none of our major conclusions change because of this alternative definition. 4.4. Changing the winsorizing level One possible concern with our study is that outliers could easily affect our results. This could happen if the denominator in our ratios is close to zero (i.e., the increase (decrease) in debt almost equals the decrease (increase) in external equity). To deal with this problem, we winsorize both ESNI and ESNR at the 99 and 1% levels. We redid our analysis using 95 and 5% levels for winsorization. Our main conclusions did not change suggesting that outliers are not driving our results. 5. Conclusions It is well known that market imperfections (for example, taxes, bankruptcy costs, agency costs, and information asymmetries) make capital structure relevant and important. We seek to understand better how firms make capital structure decisions. Our paper examines three relatively new theories of capital structure (mispricing, financial waves, and corporate life cycle) to see if they can help explain firms’ issuance and repurchase decisions. Our approach for testing these three theories differs from previous research. Instead of asking whether mispricing, financial waves or the corporate life cycle are associated with changes in equity, we inquire instead whether these three theories are related to increases or decreases in equity relative to debt. Our results show that mispricing influences both the decision to issue equity and the decision to repurchase equity. Our findings suggest that some executives time their equity issues and repurchases. These managers issue after high-adjusted returns and before low-adjusted returns and repurchase after low-adjusted returns but before high-adjusted returns. We also find support for the corporate life cycle hypothesis. Younger firms issue more equity and repurchase less equity than older firms. On the other hand, our results are not consistent with the financing wave hypothesis. Firms do not issue relatively more equity during economic expansion periods. Only for common law countries and market-based countries do we find some evidence for the financial wave hypothesis and that is limited to repurchases. Overall our findings are fairly robust across different samples (all, large changes, US, non-US, common law, civil law, market-based, and bank-based). With one noticeable exception (repurchasing on the financial wave hypothesis), our results do not depend on legal regimes or financial structure. Our

B. Seifert, H. Gonenc / J. of Multi. Fin. Manag. 22 (2012) 66–81

81

conclusions are also not based on US observations alone as the findings from non-US observations generally mirror those from the US. Future research might investigate under what circumstances these hypotheses hold. For example, does the amount of new investment needed (large or small) influence whether mispricing, financing waves, or corporate life cycle hold? We could subdivide the observations into those where the firm had unusually large capital expenditures as a percent of total assets versus those where the ratio of capital expenditures to total assets was smaller and see whether the three hypotheses hold in both groups. Our empirical design would have to change to address this question. Similarly, observations could be grouped according to how large the amount of cash and short-term investments the firm holds. References Baker, M., Wurgler, J., 2000. The equity share in new issues and aggregate stock return. Journal of Finance 55, 2219–2257. Baker, M., Wurgler, J., 2002. Market timing and capital structure. Journal of Finance 57, 1–32. Beck, T., Levine, R., 2002. Industry growth and capital allocation: does having a market- or bank-based system matter? Journal of Financial Economics 64, 147–180. Brav, A., Graham, J., Harvey, C., Michaely, R., 2005. Payout policy in the 21st century. Journal of Financial Economics 77, 483–527. Carlson, M., Fisher, A., Giammarino, R., 2006. Corporate investment and asset price dynamics: implications for SEO event studies and long-run performance. Journal of Finance 61, 1009–1034. DeAngelo, H., DeAngelo, L., Stulz, R.M., 2010. Seasoned equity offerings, market timing, and corporate life cycle. Journal of Financial Economics 95, 275–295. Demirguc-Kunt, A., Levine, R., 2001. Financial structures across countries: stylized facts. In: Demirguc-Kunt, A., Levine, R. (Eds.), Financial Structure and Economic Growth. MIT Press, Cambridge, MA, pp. 81–140. Demirguc-Kunt, A., Maksimovic, V., 2002. Funding growth in bank-based and market-based financial systems: evidence from firm-level data. Journal of Financial Economics 65, 337–363. Dittmar, A., 2000. Why do firms repurchase stock? Journal of Business 73, 331–355. Dittmar, A., Dittmar, R.F., 2008. The timing of financing decisions: an examination of the correlation in financing waves. Journal of Financial Economics 90, 59–83. Graham, J., Harvey, C., 2001. The theory and practice of corporate finance: evidence from the field. Journal of Financial Economics 60, 187–243. Gatchev, V., Pulvino, T., Tarhan, V., 2010. The interdependent and intertemporal nature of financial decisions: an application to cash flow sensitivities. Journal of Finance 65, 725–763. Gatchev, V., Spindt, P., Tarhan, V., 2009. How do firms finance their investment? The relative importance of equity issuance and contracting costs. Journal of Corporate Finance 15, 179–195. Hovakimian, A., Opler, T., Titman, S., 2001. The debt–equity choice. Journal of Financial and Quantitative Analysis 36, 1–24. Kim, W., Weisbach, M.S., 2008. Motivations for public equity offers: an international perspective. Journal of Financial Economics 87, 281–307. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1998. Law and finance. The Journal of Political Economy 106, 1113–1155. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Corporate ownership around the world. Journal of Finance 54, 471–517. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 2000. Agency problems and dividend policies around the world. Journal of Finance 55, 1–33. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 2002. Investor protection and corporate valuation. Journal of Finance 57, 1147–1170. Lee, C.I., Ejarab, D.D., Gleasonc, K.C., 2010. An empirical analysis of European stock repurchases. Journal of Multinational Financial Management 20, 114–125. Loughran, T., Ritter, J.R., 1995. The new issues puzzle. Journal of Finance 50, 23–51. McLean, R., 2011. Share issuance and cash savings. Journal of Financial Economics 99, 693–715. McLean, R., Pontiff, J., Watanabe, A., 2009. Share issuance and cross-sectional returns: international evidence. Journal of Financial Economics 94, 1–17. Peyer, U., Vermaelen, T., 2009. The nature and persistence of buyback anomalies. Review of Financial Studies 22, 1693–1745. Pontiff, J., Woodgate, W., 2008. Share issuance and cross-sectional returns. Journal of Finance 63, 921–945. Schultz, P., 2003. Pseudo market timing and the long-run underperformance of IPOs. Journal of Finance 57, 483–517. Skinner, D.J., 2008. The evolving relation between earnings, dividends, and stock repurchases. Journal of Financial Economics 87, 82–609. Wang, C., 2011. Why can managers time the market in issuing new equity? The global evidence. Journal of Multinational Financial Management 21, 151–164. Wooldridge, J., 2000. Introductory Econometrics: A Modern Approach. South-Western College Publishing.