The probability of a firm making a takeover bid: An empirical analysis of Australian firms

Abstract

This study develops a prediction model to investigate the probability of firms making a takeover bid. This model draws upon a number of firm characteristics, which can be categorized under three types of theory: agency costs, hubris behaviour and synergy motives. By examining a sample of 316 Australian publicly listed firms over the period 1999–2010, this study finds that bidding firms are significantly different from non-bidding firms in terms of their cash level, leverage, capital expenditure (long-term productivity growth) and management overconfidence level. Bidding firms are predisposed to enter into takeover activities because their management’s investment decisions are primarily driven by agency, hubris behaviour and synergy motives. The findings show that hubris behaviour is more dominant in influencing managers’ investment decisions in a takeover bid rather than agency or synergy motives.

Keywords

Agency bidder hubris behaviour non-bidder organic growth sample selection model synergy motives takeover announcement

1. Introduction

Takeovers constitute one of the largest investment decisions and corporate restructure changes made by corporations. While there is a continued, and even growing, interest in takeover research to understand whether the market for corporate control holds, it is necessary to explore the channels through which the market for corporate control works. This is because there is a lack of awareness of what factors (e.g. industry and firm characteristics) influence a firm’s investment decision-making process, given that takeovers constitute only one source of growth for firms and the alternative is organic growth. Hence, there is very limited awareness as to what defines bidders, given that a large amount of the literature reports that bidders on average earn significant negative return (Andrade et al., 2001; Bhagat et al., 2005; Cornett et al., 2011; Fuller et al., 2002; Masulis et al., 2007; Moeller et al., 2004; Porter and Singh, 2010; Travlos, 1987).¹ Bid announcements have the attributes of option contracts because the bidding firms are not under any obligation to continue with the deal if management choose to opt out as a bidder. The non-random nature of the firm’s investment decisions suggests management holds private information and makes it enticing to question what defines bidding firms.²

Examining a sample of 316 Australian publicly listed firms from the period 1999 to 2010, this study identifies the probability of a firm making a takeover bid vs no bid using a probit model. It is important to distinguish between bidding firms and non-bidding firms as failure to do so can lead to biased estimates in evaluating the true economic wealth effect to bidders’ shareholders around a bid announcement. The results show that bidding firms are significantly different from non-bidding firms in terms of their empire-building hypothesis, leverage, organic growth structure and hubris behaviour. Consistent with Porter and Singh (2010) and Cornett et al. (2011) their study finds that the bidding firm’s management decisions are primarily predisposed by agency cost and hubris behaviour, rather than by synergy motives. The results are robust and show that hubris behaviour is predominant compared to agency or synergy motives in influencing management bid decisions.

The current study differs in a number of ways from previous studies in takeover activities, including from the recent and sophisticated work of Cornett et al. (2011). First, the present study explores the candidacy of bidding firms using a different methodology: the probit model. Although alternative methodologies have been proposed (Eckbo et al., 1990; Espahbodi and Espahbodi, 2003), this paper uses the Heckman (1979) sample selection (probit) model in conjunction with appropriate independent variables as it is a good approach for predicting the likelihood of takeover (Cremers et al., 2009; Zhu, 2013). The Heckman model is a superior model than other models to address self- selection issues when it is possible to identify and properly incorporate both event (bidding) and non-event (non-bidding) firms of the population sample correctly (Prabhala, 1997). Further, the benefit of the probit model is that it contributes to the finance literature by acknowledging the alternative decision-making process in takeover investment; that is, organic growth (Hess, 2010).

Second, this study uses a different set of variables to define bidders as compared to those used by Cornett et al. (2011). The existing research consulted employs US data and is thus reflective of the characteristics of US firms and markets. Little is known about investment decisions outside the US market. This is despite the fact that the Australian takeover market is one of the largest in the Asia-Pacific; total takeover announcement deals alone amounted to over AUD$77.5 billion in 2010, compared to AUD$57.3 billion in 2009.³

The paper is structured into the following six sections: Section 2 provides an overview of the takeover bid motives literature and debates the relationship between organic and takeover investment. Section 3 discusses the probability of firms launching a takeover bid. Section 4 presents the data of the study. Section 5 outlines the methodology, in particular the Heckman (1979) binary model. Section 6 provides results under descriptive and multivariate analysis. Section 7 shows robustness test results and section 8 concludes.

2. Literature review

2.1. Investment decisions of firms

Research on alternative investment options to takeover strategies in corporate investment decision-making was limited until Lambrecht’s (2004) theoretical foundation gave rise to research on this topic in the finance discipline (Maksimovic and Philips, 2001; Warusawitharana, 2008).⁴ In particular, Warusawitharana (2008) suggests that productivity-driven decisions are an important component of the firms’ choice between internal (organic) investment and takeover investment growth. The distinction between the two investment types is necessary, as investment decisions have important implications for firm survival in the long-run, market participants including shareholders, managers, takeover regulators and the broader society (Delmar et al., 2003). While takeover investment is often seen as the fastest approach to alter industry structure, and internal investment is seen to be slower, more stable growth (Hess, 2010; Mitchell and Mulherin, 1996), continuous growth via takeover is impossible in the long term, even for those firms that have previously experienced successful takeover acquisition performance (Moeller et al., 2005).⁵ There is no clear consensus in the existing literature on the comparative benchmark performance between internal (organic) and takeover investment. In particular, there is no clear understanding on whether the two investment options are complementary investment choices for firms, or exclusive of one another. The next section elaborates on the dynamics of takeover and internal investment.

2.2. Relationship between organic and takeover investment

Based on compatibility of takeover and organic investment strategies, Andrade and Stafford’s (2004) investigation on the determinants of takeover and organic investment during the three decades 1970s, 1980s and 1990s suggests that, under the economic expansionary period, mergers and internal investments were positively related to capacity utilization, and were more complementary.⁶ Hence, takeover and organic growth are complementary rather than substitute. Conversely, in economic contraction periods, merger activity is negatively related to capacity utilization, making takeovers a substitute for organic growth. Taken together, it seems that organic growth and takeover growth are inter-related. From an investment policy perspective, growth maximization is partially in line with the profit-maximization objective in the sense that growth is required for developing new long-term potential for profit. While profit is the precondition to financial growth in business, it is sensible not to separate the objective of the firm from the nature of the firm since the firm exists to serve the objective of their principles (Pitelis and Teece, 2009).

Hess (2010) argues that smart growth companies are more likely to engage in takeover growth opportunities due to strategic reasoning rather than earning games.⁷ This reaffirms the conventional view of information asymmetry in agency theory between management and shareholders in the corporate governance mechanism as the primary drivers of takeover activities in corporate governance (Jensen, 1986). Considering that firms that bid are not randomly deciding to be a subgroup of the firm population (Li and Prabhala, 2007; Prabhala, 1997) suggests that such decisions are timed carefully, which leads to a sample selection problem. Despite takeover bids being carefully timed, takeover bid have been empirically shown to be performance decreasing for bidder shareholders (Bhagat et al., 2005; Cornett et al., 2011; Fuller et al., 2002; Masulis et al., 2007; Moeller et al., 2004). There is no clear consensus in the existing literature as to what drives firms to enter into a takeover bid versus organic growth strategies. Cai et al.’s (2011) study of the bidders’ anticipation hypothesis argues that it is possible for rational investors to anticipate a firm’s propensity to bid prior to the public bid announcement. Thus, to determine the probability of a firm making a takeover bid, a researcher needs to observe a firm’s industry channels; that is, investigate observable factors relevant to firm and industry characteristics.

3. Prediction of takeover bid likelihood

The conventional view of takeovers suggests they are mainly due to three motives: the synergy, behavioural and agency motives. The theory and the empirical findings underlying these three motives are discussed below.⁸

Table 1.

Variable definitions and their expected effects on bidder’s candidacy.

Variable name	Variable description	Expected sign
SIZE	Size of bidding firm relative to the target firm	Positive
CASH	Cash Reserve available in firm prior to the corporate announcement date	Positive
LEV	Total Debt firm holds prior to announcement	Positive
TQ	The profitability level of firm prior to its takeover announcement	Positive
ROE	Return on Equity is the profit earned in relation to equity resources invested from equity holder’s perspective	Positive
S_MISV	The pre-bid measure of bidder firm to identify whether management believes bidder is overvalued or undervalued	Positive/Negative
ORG	Firm long-term productivity growth, capturing intangible assets and capital expenditure, which is a proxy for measuring organic growth	Positive/Negative
AGROWTH5	Annual Sales growth level prior to bid announcement	Positive
LCYC	Life stage of firm in terms of mature, young and old	Positive/Negative
TOE	Percentage of shares that bidders hold of target prior to bid announcement	Positive
D_Size	Market capitalization of bidding firm	Positive
D_EQ	Bidder directors’ ownership of bidding firm prior to takeover announcement	Negative
D_NEXE	Number of non-executive directors serving on bidder’s board	Positive/Negative
D_DUAL	Number of directors holding directorship position in other firms	Positive/Negative
CEO_TURNVR	1 if CEO gets replaced by a new CEO prior to the announcement	Positive
CEO_CHAIR	1 if CEO is also acting as Chair prior to the announcement	Positive
I_CON	The level of competitiveness of the firm within the industry	Positive
I_GROWTH	The past five years’ sales growth of the industry prior to the bid announcement	Positive
I_SHK	Shock to the industry that measures absolute difference between an industry’s five years’ sales growth rate across all industries	Positive

3.1. Agency characteristics

3.1.1. Firm free cash flow, size and leverage

The most widely cited theory in takeover bid decisions is the agency motive. The agency motive argues that, due to the existence of excess cash flow, empire-building hypothesis (Jensen, 1986), size (Jensen, 1986), and optimal leverage (Stulz, 1990), firms engage in takeover-related activities. However, there is no consensus view on these theories. For example, according to the free cash flow hypothesis, large firms’ management have a tendency to overinvest in wealth-destroying projects owing to those firms’ internal cash reserve capacity (Harford, 1999; Harford et al., 2008). However, Huyghebaert and Luypaert (2010) argue that a firm with fewer available cash reserves, debt capacity and internal capabilities cannot easily achieve growth via organic investment and is more likely to enter into takeover bid.

According to optimal corporate financial structure, managers will not over-expand their empire by reinvesting excess cash unwisely if the firm disciplines managers using optimal capital structure (Hart, 1995). Similarly, Stulz (1990) proposes an optimal leverage framework that trades off the leverage in discouraging managers from empire building when free cash flows is high. This indicates that the level of leverage can be treated as a mechanism of anti-takeover protection, making high leveraged firms more likely to launch a takeover bid (Moeller et al., 2004). However, a number of empirical studies show that bidders are significantly less leveraged before mergers than a control sample of firms (Morellec and Zhdanov, 2008).

The literature on size effect of firms suggests that large firm size increases the likelihood of a firm to engage in takeover bid activity (Andrade et al., 2001; Moeller et al., 2004, 2005). In contrast, Offenberg (2009) argues that larger firms are more likely to be the target of a disciplinary takeover than are smaller firms because large firms have high CEO turnover. This increases management inefficiency, increasing the firms’ chance of being disciplined by the market for corporate control.⁹

3.2. Firm behavioural characteristics

3.2.1. Managerial myopic, over confidence and stock misevaluation

According to ‘managerial myopic’ behaviour theory, growing institutional holdings and takeover threats cause managers to sacrifice long-term growth by investing in short-term projects (Stein, 1988). This behaviour is more common when managers are under extreme pressure from the capital market to demonstrate good financial performance. Further, the myopia hypothesis suggests that investment in fixed capital and R&D would result in long-term productivity growth, which would in turn operate as a takeover defence in the takeover market, increasing the probability of a firm becoming a bidder (Harford et al., 2008; Madden, 2007).

Roll’s (1986) theoretical work on the ‘hubris hypothesis’ postulates that positive profitability in acquiring firms builds overconfidence in bidding management’s decision-making, leading to an increasing probability of such firms making a takeover bid. A number of researchers on behavioural finance suggest that there is a strong presence of behavioural bias during the decision-making process of management in takeover-related activities (Malmendier and Tate, 2008; Moeller et al., 2005). Thus takeover announcements follow periods of exceptionally good performance experienced by firms, and that engaging in multiple takeover acquisitions within a short period tends to overestimate management ability to select profitable investments and synergy gains associated with a takeover.

Based on the overconfidence hypothesis, the misvaluation hypothesis theory of takeovers states that market inefficiency, such as irrational shifts in investor sentiment, affects the takeover decisions of firms and has important effects on takeover activity (Dong et al., 2006; Shleifer and Vishny, 2003; Rhodes-Kropf et al., 2005; Rhodes-Kropf and Viswanathan, 2004). In particular, Rhodes-Kropf et al. (2005) and Dong et al. (2006) postulate that, in a period of high takeover activity (such as during a takeover wave), firms are more likely to be misvalued because valuations are mostly driven by market participant beliefs. The market being optimistic during the presence of good news, in combination with positive recent performance, boosts management’s overconfidence, increasing the volume of stock issued and giving rise to opportunistic stock bids by overvalued potential bidders because overpriced bidders can afford to pay high premiums when the market is more optimistic about potential synergies.

3.3. Firm synergy characteristics

3.3.1. Industry shock

Theory on industry shock includes both external shock (economy, regulation or technology: Faria, 2008) and internal shocks (growth and industry concentration: Jensen, 1986). Shock is an important determinant of the level of corporate restructuring. A number of empirical studies support the external shock argument. For example, Harford’s (2005) findings suggest that it is not the market timing (misvaluation effect) but rather the clustering of shock-driven industry merger waves that causes aggregate takeover waves. In a recent study, Madura and Ngo (2008) argue that the expected synergy of a takeover reflects the takeover premium paid. Further, Harford (2005) argues economic and/or industry shocks allow more potential for synergies in a particular period and industry is more relevant here. Industry shock is a perfect proxy to capture expected synergy of a takeover bid.

Andrade and Stafford’s (2004) study of firms during the 1970–1994 sample period shows that a firm’s internal investment stays fairly stable over time, while merger activity intensifies in response to changing industry conditions, in turn prompting broad restructuring (Andrade and Stafford, 2004; Mitchell and Mulherin, 1996). Consistent with the industry shock argument, a recent empirical study in Australia has found that synergy is a dominant motive behind takeover, as compared to hubris and agency motives (Porter and Singh, 2010).

4. Data

4.1. Sample period

This study examines takeover bid announcements made by Australian publicly listed companies during the period 1 January 1999 to 31 December 2010. The selection of takeover bid announcement dates is between three databases: Connect4, Securities Data Corporation (SDC) and Aspect Huntley DatAnalysis. The sample only comprises 10 years for a number of reasons. First, announcements data prior to 1999 are unavailable. Second, for comparative purposes, this study needed to overlap at least in part with US studies. The US-based studies suggest that both bidder size and takeover transaction values increased significantly during 1999 and 2000, a period generally noted as the ‘bubble’ period (Masulis et al., 2007; Moeller et al., 2004).¹⁰ In addition, this study chose to capture incremental growth in takeover activity during the pre-GFC and post-GFC years. The sample period mainly takes place after the change in takeover regulation in Australia, as given in Chapter 6 of the Corporations Act (2001): Section 606(1) and (2). In 1999, part of the Corporate Law Economic Reform Programme’s reform of the law and policy governing takeovers in Australia scrapped the mandatory bid rule. Under the current regulations, the inability of bidders to acquire greater than 20% of a firm in advance of a public takeover announcement bid means that takeover bids have to be launched before the bidder holds a controlling stake in the target. From the bidding firm’s perspective, this makes takeover bids for Australian companies considerably riskier and more costly due to uncertainty.

4.2. Sample selection and sample characteristics process

According to McWilliams and McWilliams (2000), correct announcement dates are essential to an effective event study method. The sample firms selected had to meet the following criteria:

Bid announcement comprises successful, unsuccessful, withdrawn or pending takeover

Both bidder and target firms are Australian firms

The deal value disclosed on the Connect4 database is more than AUS$1 million

The bidder owns less than 50% of the target’s shares prior to the bid

Annual financial data is available from Connect4 and Aspect Huntley Databases.

The primary sources of data were the databases Connect4, SDC and Aspect Huntley (DatAnalysis and FinAnalysis). The takeover announcement data was extracted from the Connect4 database, which provides an in-depth summary of every takeover of all ASX-listed companies in Australia from 1999 onwards.¹¹ Connect4 also provides comprehensive bid information, including takeover bid value, revised bid details, detailed offer descriptions, director’s recommendations report, company sector information and final status of takeover transaction. To ensure the validity and consistency of the takeover announcement dates, the dates in Thomson Reuters SDC database are randomly checked and matched against the data in Aspect Huntley’s DatAnalysis database. The accounting data of each firm is extracted from financial annual reports from Connect4 and the Aspect Huntley databases.

The initial sample comprised 384 takeover bid announcements made by 338 listed companies over the sample period. Following the filtering procedures, the sample consisted of the takeover bid announcements of 162 companies. The current study excludes firms within the financial and utility industries. Further, the study excludes companies with unavailable financial annual reports. The final sample includes 316 observations of companies, 162 of which are sample firms, while the remaining 154 are control firms. The control firms are lessened by 10 firms because of data unavailability. Table 2 depicts the distribution of the annual sample of bidding firms. The final sample incorporates bid announcements in the form of either a scheme of arrangement or a takeover. In Australia, these forms are essentially the same. Schemes of arrangement are friendly, with the acquirer firm usually acquiring 100% of the target firm with the approval of the target firm’s board. Takeovers can either be hostile (unsolicited) or friendly (non-contested), depending on the board of the target firm.

Table 2.

Distribution of annual sample of bidding firms.

Year	Number of bidders	Bid status (W)	Bid status (U)	Bid status (S)	Bid status (P)	Bid type (BS)	Bid type (SA)
1999	3	1	1	1	0	3	0
2000	2	1	0	1	0	2	0
2001	7	2	0	5	0	5	2
2002	8	0	1	7	0	5	3
2003	12	2	0	10	0	7	5
2004	8	0	0	8	0	6	2
2005	13	1	0	12	0	11	2
2006	20	3	1	16	0	9	11
2007	28	7	3	18	0	18	10
2008	19	6	2	11	0	16	3
2009	24	9	2	12	1	9	15
2010	28	5	3	12	8	19	8
Total	172	37	13	113	9	110	61

Notes: Table 2 shows the number of bid announcements made between 1999 and 2010. Table 2 also reports the type of bid launched by each bidder at the initial bid announcement date and the bid status. Bid type (BS) represents bidder’s statement from the Connect4 database; Bid type (SA) represents scheme of agreement between bidders and target management. In the bid status columns: (W) denotes ‘withdrawn’, (U) denotes ‘unsuccessful’, (S) denotes ‘successful’ and (P) denotes ‘pending’ bids.

4.3. Methods for computation of control firm

Empirical literature suggests that, in general, target firms are relatively smaller than the bidding firms (Moeller et al., 2004, 2005). This emphasizes the need for an appropriate benchmark (Barber and Lyon, 1997; Fama and French, 1992). To control for any potential bias in the selection of control firms, the present study needs to examine non-bidding firms (control firms). The control firms represent the matched control firm approach following Barber and Lyon (1997).¹² The rationale behind the matched control firm method is that the sample firms are matched to control firms such that they establish an appropriate benchmark against which to compare and distinguish any difference between the characteristics of the two sets of firms (control vs sample). The benefit of the matched sample approach is that it can control for latent factors, although it also imposes a cost of reduced sample size.

To generate the industry-size matched control group, the following steps are involved. First, there was an attempt to pair each sample firm with a control firm with the same industry type. This required an industry name filtering option. Second, each sample firm was paired with a control by ranking firms by their market capitalization. The present study introduces the term ‘one clone firm’ to represent one sample firm. The rationale for using a clone firm is that it is often impossible to identify control firms that are perfect matches for all sample firms matching on market capitalization. Therefore, each clone firm denotes an average of three control firms, where the average closely matches each sample firm within the same industry. Following this filtering procedure, the pool of potential matches had a mean difference in size between the sample firms and control (clone) firms of less than 25% three years preceding the takeover announcement date. Table 3 summarizes the difference between the industry-size characteristics of the sample and clone groups for the time points: the announcement year, two years previous and three years previous to the announcement year. Table 3 presents these three models, showing that the market capitalization of the sample firms exceeds the respective means of the control (clone) groups at the 0.05 level (two-tail t-test). This reveals that the sample firms are biased towards larger firms at the 0.05 level, but not at the conventional 1% level for a two-tailed test. The control firms selected had to meet the following criteria:

It is an Australian firm

Annual data for at least three years prior to the announcement date is available

Market capitalization data available (Connect4 and Aspect Huntley databases)

A control firm is used only once in a control clone group

Control firms matched on size (market cap) and industry type.

Table 3.

Comparison of the sample firms and the industry-size matched control group.

Items	Industry – size matched
	Announcement year		2 years prior to announcement		3 years prior to announcement
	Sample	Control	Sample	Control	Sample	Control
Number of sample firms to be matched	340		340		340
Total number of observation	324		314		308
Firms matched	162	162	157	157	1554	154
Average market capitalization ($billions)	$3.56b	$1.08b	$2.57b	$0.65b	$2.58b	$0.81b
Mean difference between control and sample firms ($billions)	$2.49b		$1.91b		$1.77b
t-stat	2.2159**		2.3323**		2.3066**

Denotes significantly different from zero at the level of 5% for two-tail tests.

4.4. Measurement of control variables

The measurement of each of the explanatory variables is based on the last annual financial balance report of the firms prior to the announcement of the takeover bids. In cases in which the takeover announcement bid came before the release of the company’s annual report, but after or before the financial reporting date, the previous available annual report is used, based on the principle that these data represent the best available market information at the time of the announcement. Therefore, each of the measurement variables represents figures from company’s annual reports, rather than interim reports. The present study uses a number of accounting proxies for the control variables to assist in identifying the determinants of firms becoming potential bidders. Table 3 gives all variable names, proxy measures and sources of measurement.

5. Methodology

The firm’s probability of making a takeover bid is calculated using a two-stage method analogous to the popular Heckman (1979) binary model estimator. The binary model is ideal to distinguish event related firms (bidding firms) from non event related firms (non bidding firms) in estimating announcement returns in corporate finance. Research has long been aware of the fact that corporate finance decisions are not random and that they typically reflect the deliberate decisions of firm managers and owners to self-select into their preferred choices (Li and Prabhala, 2007; Prabhala, 1997). The notion that firms’ investment decisions are driven purely by asymmetrical information problems between managers and shareholders suggests that firm management self-selects to be among those firms that bid. Firms’ corporate financial decisions are based on their attributes and characteristics, and are made after rigorous consideration of the options and their suitability to the firm. Thus, given that firms that launch a bid do so consciously and well-timed, failing to account for self-selection bias in an event study could lead to potentially bias estimates and the conclusions rendered inaccurate.

It is probable that takeover establishment is a function of a firm’s strategic decision-making process, which is driven by a wide range of characteristics within which the firm operates. The multinomial probit model specification is an appropriate framework for addressing the self-selection problem, as the model requires the computation of choice probabilities of each firm within the population of firms (Prabhala, 1997). Following Li and Prabhala (2007), equations (1)–(3) are developed. Equation (1) represents the population regression in which Yi is a function of some X variables. Note that equation (1) must be established such that it incorporates a subsample of firms that were subject to being self-selected into choice B (Bid). Thus, selection is specified using a probit model in which firm i chooses B if the net benefit from doing so is positive (Wi), this is shown in equation (2), and equation (3) represents non-bidding firms.

Y_{i} = X_{i} β + ε_{i}

(1)

P = B \equiv W_{i} = Z_{i} λ + μ_{i} > 0

(2)

P = N B \equiv W_{i} = Z_{i} λ + μ_{i} \leq 0

(3)

Where, Y_i is the dependent variable representing the outcome; X_i is the variable outlining the outcome; ε_i is the error term; W_i is a scalar; Z_i denotes publicly available information; λ is the vector of the probit coefficient and µ_i is orthogonal to Z_i; P=B is the probability that the firm will bid; P=NB is the probability that the firm will not bid. ε_i and µ_i are bivariate normal, it is possible to derive the likelihood function and the maximum likelihood estimator for the equations (1)–(3).

Heckman models estimate two equations. It is essential to separate firm behavioural parameters (selection model) from those parameters determining outcome function (regression model). Due to the current study’s objectives, this paper only concentrates on Heckman’s first model, also known as the probit model. The probit model predicts whether the strategic choice variable (firm characteristics, industry characteristics, corporate governance characteristics and organizational characteristics) influences in any way the probability of a firm engaging in a takeover bid.¹³ The first stage specification is presented below:

\Pr (Y_{i} = 1 | b * > 0) = Φ ({X^{'}}_{i} β) + ε_{i}

(4)

b^{*} = z_{i} δ + μ_{i}

(5)

Where, in equation (4),Y_i is the dichotomous dependent variable taking the value 1 to represent propensity to be included in the sample (Bid), and 0 represents propensity to not be included in the sample (not bid). Φ is the cumulative normal distribution function, X_i is the vector of parametric estimates or covariates, β is the vector of coefficients, and ε_i is the error term (heterogeneity term). In equation (5), z_i is the information set (the instrument that influences the firm’s bid decision), δ is the vector of unobservable determinants, µ_i is the error term and b* (endogenous variable) is the unobservable latent variable. Thus, the observed bid is modelled such that:

\begin{array}{l} B = 1, i f Z_{i} λ + μ_{i} > 0 \\ B = 0, o t h e r w i s e \end{array}

Where B= 1 is a bidder and 0 is a non-bidding firm.

Following the above, a probit model of a firm’s decision to bid takes the following specification:

P (B_{i} = 1 | Z_{i t}) = Φ (\begin{array}{l} α_{0} + β_{1} s i z e_{i t} + β_{2} c a s h_{i t} + β_{3} l e v_{i t} + β_{4} l i q_{i t} + \\ β_{5} r o e_{i t} + β_{6} s_m i s v_{i t} + β_{7} t q_{i t} + β_{8} o r g_{i t} + \\ β_{9} a g r o w t h 5_{i t} + β_{10} l c y c l_{i t} + β_{11} t o e_{i t} + β_{12} d s i z e_{i t} + \\ β_{13} d n e x e_{i t} + β_{14} d d u a l_{i t} + β_{15} d e q_{i t} + β_{16} c e o_c h a i r_{i t} + \\ β_{17} c e o_t u r n v r_{i t} + β_{18} I . s h o c k_{i t} + β_{19} I . g r w t h_{i t} + β_{20} I . c o n_{i t} + ε_{i t} \end{array})

(6)

Where, the dependent variable, B_i, is a binary model – an indicator variable that takes the value 1 for firms that make a bid and zero for control firms that decide not to bid. Φ is the cumulative normal distribution function. The variables in parentheses are instruments that explain the firm’s decision to bid. The quadratic hill-climbing approach provides the parameter estimates, while to correct for possible misspecification of the distribution of B, the Huber White robust standard errors and covariance method serves. The results are discussed in detail in the next section.¹⁴

6. Empirical analysis

6.1. Descriptive statistics on sample and control firms

This study investigates the motive behind a firm’s decision to announce a takeover bid by examining a range of variables to determine those that are the most influential. Table 5 presents the mean and median values for each variable of the univariate analysis for bid firms and control firms. The statistics show that, on average, bidding firms, as compared to non-bidding firms, significantly outperform the control in terms of cash holding, leverage, firm lifecycle, non-executive directors serving on the board, director duality, CEO as Chair and toehold level. However, bidding firms fall short of control firms in organic growth. Thus, with reference to the literature review, the findings are consistent with agency and organization theory.

Table 4.

The following table presents all the variable descriptions, proxy measures and references.

Variable name	Agency element definitions and proxy measurements	Theory and empirical support
ROE	Evaluates profit earned in relation to equity resources invested from equity holder’s perspective. This provides a more immediate measure of corporate performance in comparison to firm’s price to book (or market to book) ratio $R O E = \frac{net profit before abnormals}{shareholders equity}$	Owen and Yawson, 2010
SIZE	Measures the relative size of the bidding firm as compared to the target firm $S i z e = L n [\frac{M k t C a p_{B i d d e r}}{M k t C p_{T \arg e t}}]$	Jensen, 1986; Bhagat et al., 2005
TQ	Measures the relative value (profitability) of a company compared to its book value $T Q = \frac{Enterprice value of firm}{BV of Assets}$	Moeller et al., 2004
CASH	Cash reserve available at the firm prior to the corporate announcement event $C a s h = \frac{S a l e s}{T A}$	Harford et al., 2008
LEV	The amount of debt in the firm’s capital structure $D E B T = \frac{T o t a l D e b t}{T A}$	Stulz, 1990; Dong et al., 2006; Huyghebaert and Luypaert, 2010
S_MISV	The PE ratio differential is a common measure to assess pre-bid value of a bidding firm. It reveals whether management perceives their firm to be overvalued or undervalued $S_MISV = ({PE ratio}_{firm} - {PE ratio}_{industry})$	Rhodes-Kropf et al., 2005
LIQ	The propensity to convert assets into cash $L I Q = \frac{C a s h + M k t . S e c u r i t y}{T A}$	Owen and Yawson, 2010
ORG Variable name	A firm’s long-term productivity growth can capture the firm’s internal reinvestment rate and thus its organic growth rate. Capital expenditure is a proxy for organic growth because it represents the funds used for additions to property, plants and equipment, excluding amounts arising from acquisitions (e.g. fixed assets of purchased companies). This item includes the property and equipment expenditures of the company Agency element definitions and proxy measurements $O R G = \frac{Capital Expenditure}{T A}$	Higgins and Rodriguez, 2006; Harford, 2008 Theory and empirical support
LCYC	Measures at which stage (mature, old or young) firms are more likely to engage in takeover activity. Two proxies are used to capture firm lifecycle $\frac{Retained Profit}{TA} & \frac{Retained Profit}{TE}$	Owen and Yawson, 2010
AGROWTH5	A firm’s annual growth in terms of sales over the five-year period prior to bid announcement	Owen and Yawson, 2010
I_SHK	Absolute difference between an industry’s last five years’ sales growth and the average last five years’ sales growth rate across all industries	Mitchell and Mulherin, 1996; Andrade and Stafford, 2004;
I_GRW	Measures the past five years’ sales growth of the industry prior to the bid announcement	Owen and Yawson, 2010
I_CON	Measures how competitive a firm is within an industry. Using a HHI index, the following formula is conducted: $H H I = {\sum_{i = 1}^{n} [\frac{S a l e s_{i, t}}{\sum_{i = 1}^{n} S a l e s_{i, t}}]}^{2}$	Gorton et al., 2009.
D_SIZE	Total number of executive and non-executive directors serving on the company board	Owen and Yawson, 2010.
D_DUAL	Number of directors holding a directorship position in other firms
D_NEXE	Total number of non-executive directors serving on the board
D_EQ	Directors’ ownership in the firm. This is used instead of management ownership because the former is the primary party involved in the major investment decision-making processes regarding a firm’s business operation activities	Owen and Yawson, 2010.
CEO_CHAIR	CEO also acting as Chair on the board of a firm. This is a dummy variable. 1 if the CEO is also the Chair; otherwise, 0	Masulis et al., 2007
CEO_TURNVR	CEO gets replaced by a new CEO on the board. This is a dummy variable. 1 if the CEO of the previous year is the same one at the time of the announcement; otherwise, 0	Harford and Li, 2007
TOE	Percentage of shares that bidders hold of the target firm prior to the announcement of a takeover bid. The set of sample firms that do not have any toehold in the target firm are assigned a ‘zero value’, similar to the set of control firms	Baixauli and Fernandez, 2009

Table 5.

Mean and median values and t-stats (mean values) of control and sample firms.

Financial items	Mean		Median		Std. Dev.		Skewness		t-stat
	Sample	Control	Sample	Control	Sample	Control	Sample	Control
SIZE	2.18	1.78	1.82	1.34	2.17	1.86	0.29	0.89	1.58
CASH	37.83	7.61	0.72	0.17	86.04	40.52	3.22	9.07	4.07
LEV	0.52	0.29	0.17	0.21	1.07	0.37	2.74	2.88	2.29
LIQ	0.23	0.20	0.18	0.10	0.17	0.23	0.74	1.35	1.11
ROE	0.26	0.21	0.17	0.13	0.35	0.25	5.18	3.43	1.57
S_MISV	15.60	15.99	12.86	10.72	163.49	92.16	−2.41	4.09	0.08
TQ	2.79	1.92	1.73	1.45	3.60	2.00	4.12	3.49	1.32
ORG	0.11	0.13	0.09	0.08	0.07	0.16	0.92	2.64	−2.68
AGROWTH5	0.25	0.21	0.17	0.15	0.82	0.62	2.52	0.96	0.48
LCYC	1.03	0.10	0.33	0.10	4.94	1.57	11.85	−0.59	2.30
TOE	9.87	0.00	0.00	0.00	20.36	0.00	5.24	11.63	6.24
D_SIZE	6.02	6.19	6.00	6.00	1.46	2.03	0.61	1.09	−0.69
D_EQ	0.19	0.17	0.12	0.05	0.26	0.28	4.18	3.05	−0.18
D_NEXE	3.99	4.74	4.00	5.00	1.62	2.09	0.37	0.66	−3.54
D_DUAL	3.60	4.49	3.00	5.00	1.40	2.27	0.87	0.41	−4.61
CEO_TURNVR	0.94	0.85	1.00	1.00	0.23	0.36	−3.88	−1.97	1.07
CEO_CHAIR	0.09	0.02	0.00	0.00	0.29	0.14	2.81	6.95	2.62
I_CON	0.16	0.16	0.15	0.15	0.09	0.09	1.08	1.10	0.01
I_GRWTH	0.12	0.12	0.10	0.13	0.13	0.12	0.29	0.31	0.02
I_SHK	0.10	0.10	0.10	0.08	0.07	0.07	0.59	0.69	0.20

Note: See Table 4 for the definitions of each of the variables.

6.2. Multivariate analysis

As mentioned in the previous section, the probit model regression provides evidence on the predictors of firms making takeover bid announcements. This section conducts a thorough analysis of the variables mentioned in the binomial probit model of the form specified in equation (6). The result of the multivariate regression framework is presented in Table 6. Model 1 of Table 6 shows the independent variables’ explanatory power on the probability of firms making a bid. As expected, the coefficient on size is insignificant; however, the size coefficient becomes significant when a different measure for size is used (see Tables 7 and 8). The coefficient of CASH or FCF is positive (0.0042) and statistically significant at the 5% level. Consistent with empirical studies, firms with more cash reserves are more likely to launch a takeover bid (Malmendier and Tate, 2008; Moeller et al., 2004). The coefficient of leverage (LEV) is also positive (0.2869) and statistically significant at the 5% level. This suggests that firms that are large in size have more debt, which in turn serves as an anti-takeover defence mechanism. Such firms are more likely to finance large investments (i.e. takeovers) at a high premium due to entrenched management optimism (Moeller et al., 2004).

Table 6.

Probit model estimation.

This table reports the agency elements’ (including firm, industry and corporate governance characteristics) influence on the likelihood of firms making a takeover bid. All variables in both models 1 and 2 are identical except for the lifecycle variable, as explained below. SIZE is the log ratio of the bidding firm’s market cap to the target firm’s market cap. LEV is the ratio of total debt to total assets. ROE is the ratio of net profit before extraordinary items to the sum of common and preferred equity. CASH is net operating cash flow from operations minus capital expenditure scaled by total assets. LIQ is the ratio of cash and marketable security to total assets. S_MISV represents the investor misvaluation hypothesis of a firm’s stock measured as the difference between firm PE ratio and industry average PE ratio. TQ denotes management overconfidence level measured as the total enterprise value of the firm to the book value of its assets. ORG denotes capital expenditure to total assets. Among the control variables in Model 1, LCYC is defined as the ratio of retained earnings to total assets, while in Model 2 lifecycle is defined as the ratio of retained earnings to total equity. AGROWTH5 is the firm’s sales growth over the five years prior to the takeover announcement. TOE is the percentage of target’s shares bidders hold prior to announcement. D_SIZE symbolizes the total number of directors. D_NEXE denotes total number of non-executive directors (independent directors). D_EQ indicates total shares held by directors. D_DUAL specifies the total number of directors holding multiple directorial positions in other companies. CEO_TURNVR is a dummy variable, taking the value 1 if the CEO prior to the takeover announcement is the same as in the previous year (0 is otherwise assigned). CEO_CHAIR denotes a dummy variable, taking the value of 1 if the CEO also serves as the Chair of the company (or 0 otherwise). I_CON is the industry concentration level measured using the HHI index. I_GRWTH is industry sales growth, calculated using the previous five years of sales. I_SHK is industry sales shock, calculated as the absolute difference between an industry’s five-year growth rate in sales and the average five-year growth rate in sales across all industries. The standard errors are in brackets. Total observation = 316 (Observation with Dep=1, 154 and observation with Dep = 0, 162).

	Model 1		Model 2
	Coefficient	z-Statistic	Coefficient	z-Statistic
Agency Elements
SIZE	0.0295	0.6098	0.0272	0.5633
	[0.0483]		[0.0483]
CASH	0.0042	1.9154	0.0041	1.8475
	[0.0022]**		[0.0022]*
LEV	0.2869	2.4171	0.2944	2.5691
	[0.1188]**		[0.1146]***
LIQ	−0.6681	−1.4659	−0.5523	−1.1916
	[0.4557]		[0.4635]
ROE	0.2783	0.8005	−0.0049	−0.0123
	[0.3476]		[0.3967]
S_MISV	−0.0005	−0.5924	−0.0005	−0.5970
	[0.0008]		[0.0008]
TQ	0.1047	3.2182	0.1079	3.2668
	[0.0325]***		[0.0330]***
ORG	−2.1488	−2.6792	−2.0224	−2.5194
	[0.8021]***		0.8027]***
I_SHK	1.3745	0.9617	1.2487	0.8688
	[1.4292]		[1.4373]
Control Variables
Firm characteristics
AGROWTH5	−0.0638	−0.4876	−0.0835	−0.6353
	[0.1310]		[0.1314]
LCYC	0.1978	1.7033	0.1579	2.0863
	[0.1161]*		[0.0757]**
TOE	44.9790	2.6067	45.9043	2.6811
	[17.2551]***		[17.1215]
Corporate Governance characteristics
D_SIZE	0.2408	2.9212	0.2459	2.9527
	[0.0824]***		[0.0833]***
D_EQ	−0.2781	−0.7995	−0.3101	−0.8791
	[0.3478]		[0.3528]
D_NEXE	−0.1752	−2.5668	−0.1739	−2.5323
	[0.0683]		[0.0687]***
D_DUAL	−0.2075	−3.5282	−0.2201	−3.6612
	[0.0588]***		[0.0601]***
CEO_TURNVR	0.4652	1.5399	0.4741	1.5296
	[0.3021]		[0.30999]
CEO_CHAIR	0.9556	2.4147	0.9552	2.4047
	[0.3957]***		[0.3972]**
Industry characteristics
I_CON	−0.4076	−0.3809	−0.3702	−0.3442
	[1.0700]		[1.0754]
I_GRWTH	0.1483	0.1852	0.2036	0.2525
	[0.8011]		[0.8064]
C	−0.8991	−1.6567	−0.8842	−1.6220
	[0.5427]		[0.5452]
McFadden R-squared	0.40		0.41

***

Denotes statistical significance at the 1% level; ** Denotes statistical significance at the 5% level; * Denotes statistical significance at the 10% level.

Table 7.

Alternative measures of firm characteristics.

Variable name	Definition and proxy measurement	Theory and empirical support
CASH	Cash reserve available at the firm prior to the corporate announcement event $C A S H * = \frac{E B I T D A - (\begin{array}{l} I n t . \exp . d e b t + \\ i n c o m e . t a x + \\ o r d . d i v + p r e . d i v \end{array})}{T A}$	Moeller and Schlingemann, 2005
SIZE	The log of firm’s total assets $S i z e * = L n (Total Assets)$	Owen and Yawson, 2010
TQ	Measures the relative value (profitability) of a company compared to its book value $T Q * = \frac{(\begin{array}{l} Book value of Assets - \\ Book Value of Equity + \\ Market value of Equity \end{array})}{Book value of Assets}$	Moeller et al., 2005; Masulis et al., 2007
S_MISV	Market to book value of firms is the ratio of the current stock price to the book value per share. It measures how much a company is worth now in comparison to the amount of capital invested by the shareholders. The purpose of this ratio is to show the value investors place on the company. Thus, it summarizes an investor’s view on the company’s performance and its future prospects $S_MISV * = (\frac{Market value of firm equity}{Book value of firm equity})$	Harford, 2005

Table 8.

Probit estimation using different variable measures.

This table reports agency element characteristics’ (firm, industry and corporate governance) influence on the likelihood of firms making a takeover bid. All variables in both models 1 and 2 are identical except for the lifecycle variable, as explained in Table 4. All the definitions of the variables are similar to those in Table 4 except for size, Tobin-Q, cash and stock misvaluation. The definitions of each of these variables are outlined in Table 7. The standard errors are in brackets. Total observation = 315 (Observation with Dep=1, 153 and observation with Dep = 0, 162).

	Model 1		Model 2
	Coefficient	z-Statistic	Coefficient	z-Statistic
Agency Elements
SIZE	−0.1607	−2.6704	−0.1632	−2.7252
	[0.0602]***		[0.0599]***
CASH	−1.6367	−1.9172	−1.5793	−1.8789
	[0.8537]*		[0.8405]*
LEV	0.2616	2.3107	0.2727	2.4385
	[0.1132]**		[0.1118]**
LIQ	−0.6321	−1.1918	−0.5634	−1.0490
	[0.5304]		[0.5371]
ROE	0.7347	1.3111	0.4949	0.8596
	[0.5603]		[0.5758]
S_MISV	−0.0433	−1.2919	−0.0588	−1.7792
	[0.0335]		[0.0331]*
TQ	0.1314	2.3921	0.1601	2.6188
	[0.0549]**		[0.0611]***
ORG	−2.0698	−2.4552	−1.9936	−2.3272
	[0.8430]**		[0.8567]**
I_SHK	1.5677	1.1054	1.4602	1.0227
	[1.4182]		[1.4278]
Control Variables
Firm characteristics
AGROWTH5	−0.0175	−0.1263	−0.0247	−0.1797
	[0.1384]		[0.1373]
LCYC	0.2326	1.7664	0.1902	2.2547
	[0.1317]*		[0.0844]**
TOE	56.4157	3.0701	55.3583	3.0349
	[18.3756]***		[18.2405]***
Corporate Governance characteristics
D_SIZE	0.3011	3.6009	0.3161	3.6799
	[0.0836]***		[0.0859]***
D_EQ	−0.3027	−0.8859	−0.3527	−1.0231
	[0.3417]		[0.3448]
D_NEXE	−0.1504	−2.1173	−0.1554	−2.1530
	[0.0711]**		[0.0722]**
D_DUAL	−0.2000	−3.3004	−0.2148	−3.4782
	[0.0606]***		[0.0618]***
CEO_TURNVR	0.5907	1.9014	0.6250	1.9227
	[0.3107]*		[0.3251]*
CEO_CHAIR	0.7825	1.8050	0.7820	1.7599
	[0.4335]*		[0.4443]*
Industry characteristics
I_CON	−0.2710	−0.2513	−0.2499	−0.2319
	[1.0786]		[1.0774]
I_GRWTH	−0.0624	−0.0801	−0.0172	−0.0218
	[0.7794]		[0.7883]
C	1.7534	1.5239	1.7489	1.5247
	[1.1506]		[1.1471]
McFadden R-squared	0.41		0.41

Note: * denotes significance at 1%; ** denotes significance at 5%; * denotes significance at 10%.

Similarly, the coefficient TQ (indicating Tobin-Q), which reflects management overconfidence based on a firm’s prior performance, is positive (0.1047) and statistically significant at the 1% level. This indicates that cash availability encourages hubris behaviour, which cultivates overconfidence among management, which in turn influences management investment decision-making approaches (Malmendier and Tate, 2008; Stein, 1988). The idea of overconfidence in this context is that management makes mistakes in evaluating potential target firms by either overvaluing or undervaluing them. This mistake is derived from management’s belief that they have made superior decisions in terms of the past investments of the firm and, as a result, they rely more on their own analytical skills and instincts rather than on the intrinsic and long-term value of the investment. Consistent with Rolls’ (1986) view, the extreme version of hubris is evident in bidder firms’ management behaviour, due to the superior record of those firms’ prior performance. Accounting for a number of coefficients including liquidity (LIQ), return on equity (ROE), stock misvaluation (S_MISV) and industry shock (I_SHK), all are insignificant; this affirms that there is no fundamental difference between control and sample firms when considering such characteristics.

The organic growth coefficient (ORG) was negative and significant at the 5% level. This means firms with a high level of organic growth strategies are less likely to make a bid in the capital market and more likely to grow via internal investment strategies. This supports Stein’s (1988) myopic hypothesis, which views organic investment acts as a takeover defence and a turn to takeover only as a means of investment growth strategies and acts as a substitute (Andrade and Stafford, 2004; Harford et al., 2008). The overall findings identify agency and hubris behaviour in bidding firms as the dominant factors driving the probability of firms engaging in takeover-related activities.

6.3. Control variables

In Table 6, among the set of control variables, as expected none of the industry characteristics variables are significant. From the corporate governance variables perspective, all the variables are significant except for director shareholding of the company and CEO turnover. The positive coefficient of both board size and CEO also acting as Chair (CEO_CHAIR) shows that there is a positive and highly significant association between these factors and the likelihood of a firm making a takeover bid. This is consistent with the prior empirical study of Masulis et al. (2007). Regarding the CEO and Chair role combined into one position, Jensen and Murphy (2010) claim such roles solidify CEOs as powerful individuals on the board, making other board members reluctant to question CEO investment motives in order to protect their personal reputations and positions as senior executives on the board. This implies that CEOs that serve as a Chair display overconfidence behaviour in their investment decision-making processes due to their optimism concerning their prior knowledge, skills and capacity.

The coefficient of non-executive directors on the board and director duality (D_DUAL) is significantly negatively related to likelihood of a firm engaging in a takeover bid. This suggests that outside board members are more experienced in judging potential investment opportunities, and that their presence reduces the propensity towards overinvestment that typically coincides with excess cash availability. Similarly, based on the reputational hypothesis, directors holding positions on multiple boards are less likely to make value-reducing investment decisions that could damage their reputation. Firms with such directors are more cautious about investment decisions, and they use a wide range of expertise in assessing potential investment strategies. The coefficient of director shareholdings, non-executive directors and CEO turnover is not significant. Taken together, these results suggest that the corporate governance characteristics variables (as compared to industry, firm or organizational characteristics) have a paramount affect in increasing the likelihood of a firm becoming a takeover bidder.

7. Robustness tests

This section reports the results of the robustness checks of the findings reported in the previous section. A number of studies have identified and applied different measures to proxy for firm characteristics variables in the specification of equation (6). Table 7 presents alternative measures of a firm’s FCF, performance level, stock misvaluation and size variables to determine whether the results exhibited in the previous section hold. Using a series of alternative variables in Table 8, the results of equation (6) become slightly different. For example, the size variable (SIZE) is negative and significant. The new measure of size does not support empirical studies that larger firms with high leverage are more likely to make takeover bids (Moeller et al., 2004). The results of the coefficients on Cash and Size, are significant, but with the sign of the coefficients inversely related to the probability of a firm making a bid. The cash coefficient does not support the empire-building hypothesis of cash-rich firms (Harford, 1999). Thus, in contrast to agency theory, a fall in cash could induce firms to make bids in the expectation of fast firm growth when management is aware that its innovative ideas are lacking. The differing result could be due to valuation proxies, or difference in time could reflect an improved/unimproved corporate governance mechanism (Dong et al., 2006).

The economic interpretation of the cash coefficient is that a 1% increase in FCF induces a fall in the probability of firms launching a takeover bid of 1.64%. This result is consistent with the work of Huyghebaert and Luypaert (2010), who argue that, when firms’ ideas for innovation are not fruitful, and when they have access to less cash, firms will feel forced to grow via takeover investment opportunities – the fastest option for growth. In addition, such firm’s management could become overconfident with their decision-making capacity. This is supported by the Tobin-Q coefficient being positively significant. This shows that the probability of firms making a takeover bid increases as firm management succumbs to increasing hubris (Malmendier and Tate, 2008; Moeller et al., 2005).

A particularly interesting finding is the coefficient on size, which indicates that as firms increase in size they become less likely to make a takeover bid. According to Offenberg (2009), this result can be interpreted as larger firms being more likely to be the target of a disciplinary takeover than are smaller firms. This could be due to high CEO turnover in large firms increasing management inefficiency. Large firms would thus be more likely to be disciplined by the market for corporate control than small firms would be. This suggests that the large target firms were once bidding firms. During their bidding stage, these firms made some bad acquisitions and, as a result, they became a target to potential bidders in the market for corporate control. Thus, smaller firms are more likely to be potential bidders. This is inconsistent with both the agency theory paradigm and organization theory. Overall, the results are robust. However, one must be cautious in interpreting the effect of cash and size on bid probability, since different variable measurements can yield negative, positive or even insignificant effects.

Despite the above findings, two possible problems could influence the results reported in Table 6. First, it is possible that multicollinearity could exist due to the large number of regressors used to represent the characteristics variables. Multicollinearity between the regressors can dampen the explanatory power of the independent variables on the probability of firms making a bid. To explore whether multicollinearity is influencing the results, an examination of the correlations between the independent variables is given in Table 9. Some correlation is apparent between the variables. This study concentrates on those elements under agency costs, with the notable exception of the correlation between cash and LIQ and ROE and LIQ. To tackle the multicollinearity problem, a robustness test is conducted. Although the results are not reported, they do not differ from those results reported in Table 6.

Table 9.

Pearson product moment correlation coefficients.

This table reports the correlations matrix between the independent variables used in the sample selection model in Table 5. The sample used in this model consists of 162 announcements made by 155 firms during the period from 1 January 1999 to 31 December 2010. All the announcement data were obtained from the Connect4 database and were randomly checked with two other databases (the SDC database and the Data Analysis database) to ensure the validity of data. The correlation coefficients for each of the variables were obtained from E-views group statistics correlations. Please refer to Table 2 for definitions of each of the variables. The t-statistic for each of the correlation coefficients was calculated using the following formula to test the hypothesis about the true correlation coefficient. In the following formula, r is the correlation coefficient and n is the total number of observations. The t-statistic values are reported in brackets in the table below. Please note the correlation table is divided into three parts as the whole table does not fit within a page.

$t = \frac{r}{\sqrt{\frac{(1 - r^{2})}{(n - 2)}}}$
	SIZE	CASH	LEV	LIQ	ROE	S_MISV	TQ	ORG	AGROWTH5
SIZE	1
CASH_1	−0.0198	1
	[−0.3516]
LEV	−0.1024	0.0105	1
	[−1.8238]	[0.18656]
LIQ	−0.1141	0.2466	0.0912	1
	[−2.0343]**	[4.5084]***	[1.6229]
ROE	−0.0936	0.0265	0.0129	0.1890	1
	[−1.6663]	[0.47019]	[0.2285]	[3.4105]***
S_MISV	0.0607	−0.0908	0.0425	−0.1395	0.0368	1
	[1.0776]	[−1.6152]	[0.7534]	[−2.4966]**	[0.6527]
TQ_1	0.0957	0.1172	−0.0732	0.1161	0.0717	−0.0408	1
	[1.7031]	[2.0910]**	[−1.3014]	[2.0711]	[1.2742]	[−0.7229]
ORG	0.0367	0.0148	0.0275	0.1231	0.0301	−0.0581	0.0947	1
	[0.6504]	[0.2614]	[0.4877]	[2.1975]	[0.5341]	[−1.0310]	[1.6862]*
AGROWTH5	−0.0230	−0.0598	0.0462	−0.0611	0.1119	0.0929	−0.0084	−0.0466	1
	[−0.4083]	[−1.0617]	[0.8196]	[−1.0842]	[1.9950]**	[1.6540]*	[−0.1494]	[−0.8261]

***

Denotes statistical significance at the 1% level; ** Denotes statistical significance at the 5% level; * Denotes statistical significance at the 10% level.

	SIZE	CASH	LEV	LIQ	ROE	S_MISV	TQ	ORG	AGROWTH5
LCYC	0.0189	0.0011	0.0694	0.0906	0.4676	−0.0158	0.0100	−0.0105	−0.0103
	[0.3358]	[0.0187]	[1.2332]	[1.6125]	[9.3746]***	[−0.2800]	[0.1781]	[−0.1868]	[−0.1834]
TOE	0.0404	0.0151	0.0483	−0.0112	0.0822	0.0520	−0.0561	−0.0247	0.0104
	[0.7156]	[0.2678]	[0.8571]	[−0.1992]	[1.4608]	[0.9230]	[−0.9956]	[−0.4381]	[0.1842]
D_SIZE	0.1142	−0.1045	−0.1434	−0.3500	−0.1077	0.0330	−0.0167	−0.2042	−0.0218
	[2.0362]**	[−1.8628]**	[−2.5670]**	[−6.6208]***	[−1.9202]	[0.5849]	[−0.29613]	[−3.6970]***	[−0.3870]
D_EQ	−0.0413	−0.0740	0.0430	−0.0317	−0.0653	−0.0160	0.0787	0.0185	−0.0041
	[−0.7324]	[−1.3151]	[0.7623]	[−0.5622]	[−1.1593]	[−0.2834]	[1.3986]	[0.3287]	[−0.0733]
D_NEXE	0.1222	−0.1232	−0.0613	−0.3569	−0.1360	0.0269	−0.0527	−0.1294	−0.0281
	[2.1812]**	[−2.1995]**	[−1.0882]	[−6.7700]***	[−2.4325]**	[0.4763]	[−0.9346]	[−2.3130]	[−0.4989]
D_DUAL	0.0279	−0.1344	−0.1193	−0.3029	−0.1653	0.0575	−0.0547	−0.1997	−0.0665
	[0.4950]	[−2.4033]**	[−2.1294]**	[−5.6319]***	[−2.9705]***	[1.0209]	[−0.9712]	[−3.6106]	[−1.1809]
CEO_TURNVR	0.0460	0.0465	−0.0795	−0.1128	−0.1004	−0.0252	0.0132	−0.0804	−0.0926
	[0.8152]	[0.8251]	[−1.4129]	[−2.0119]**	[−1.7888]*	[−0.4465]	[0.2345]	[−1.4297]	[−1.6479]
CEO_CHAIR	0.0633	0.0810	−0.0057	0.0009	−0.0282	0.0764	−0.0630	−0.0220	0.0339
	[1.1231]	[1.4406]	[−0.1009]	[0.0157]	[−0.4995]	[1.3581]	[−1.1181]	[−0.3895]	[0.6008]
I_CON	0.1400	−0.0212	0.0440	−0.0860	−0.0096	0.0172	−0.0493	−0.0064	0.0448
	[2.5063]***	[−0.3758]	[0.7809]	[−1.5290]	[−0.1706]	[0.3047]	[−0.8745]	[−0.1139]	[0.7945]
I_GRWTH	0.0669	−0.0566	−0.0887	−0.0573	−0.0611	0.0753	−0.0066	−0.0856	0.0864
	[1.1880]	[−1.0047]	−1.5782	[−1.0163]	[−1.0841]	[1.3386]	[−0.1177]	[−1.5229]	[1.5365]
I_SHK	0.0119	0.0229	−0.0902	0.1074	−0.0111	−0.0097	−0.0297	0.1351	0.1232
	[0.2114]	[0.4061]	[−1.6046]	[1.9143]**	[−0.1963]	[−0.1711]	[−0.5261]	[2.4166]**	[2.2003]**

***

Denotes statistical significance at the 1% level; ** Denotes statistical significance at the 5% level; * Denotes statistical significance at the 10% level.

	LCYC	TOE	D_SIZE	D_EQ	D_NEXE	D_DUAL	CEO_TURNVR	CEO_CHAIR	I_CON	I_GRWTH	I_SHK
LCYC	1
TOE	0.0242	1
	[0.4282]
D_SIZE	−0.0998	−0.0264	1
	[−1.7775]*	[−0.4681]
D_EQ	0.0005	0.0296	−0.0649	1
	[0.0095]	[0.5256]	[−1.1527]
D_NEXE	−0.0708	−0.0645	0.6470	−0.0951	1
	[−1.2580]	[−1.1454]	[15.0362]***	[−1.6937]*
D_DUAL	−0.0822	−0.0896	0.6340	−0.1351	0.4485	1
	[−1.4614]	[−1.5933]	[14.5275]***	[−2.4155]**	[8.8931]***
CEO_TURNVR	0.0396	0.0547	−0.0278	0.0390	−0.0258	−0.0589	1
	[0.7026]	[0.9714]	[−0.4925]	[0.6913]	[−0.4572]	[−1.0456]
CEO_CHAIR	−0.0050	0.0233	0.0009	0.0080	−0.0748	−0.0760	0.0372	1
	[−0.0882]	[0.4138]	[0.0166]	[0.1423]	[−1.3292]	[−1.3498]	[0.6603]
I_CON	−0.0252	0.1708	0.0901	−0.0507	0.1967	0.0317	0.0146	−0.0944	1
	[−0.4463]	[3.0721]***	[1.6037]	[−0.8996]	[3.5548]***	[0.5626]	[0.2591]	[−1.6810]*
I_GRWTH	−0.0430	0.0945	0.0426	−0.0291	0.1315	0.0377	0.0404	−0.1119	0.1735	1
	[−0.7628]	[1.6822]*	[0.7549]	[−0.5158]	[2.3502]**	[0.6694]	[0.7165]	[−1.9945]**	[3.1222]***
I_SHK	−0.0157	0.0607	0.0078	−0.1213	−0.0516	0.0251	0.0055	0.0450	−0.0027	0.2058	1
	[−0.2786]	[1.0781]	[0.1383]	[−2.1663]**	[−0.9155]	[0.4445]	[0.0983]	[0.7990]	[−0.0478]	[3.7270]***

***

Denotes statistical significance at the 1% level; ** Denotes statistical significance at the 5% level; * Denotes statistical significance at the 10% level.

The second possible problem that could influence the results reported in Table 6 is that the presence of firm-specific instruments excludes the potential problem of endogeneity presence in the specification of equation (6). This problem is universal because analysis of the efficacy of corporate control instruments on firms’ decision-making requires that these instruments are strictly exogenous (Demsetz and Lehn, 1985). An important assumption of the ordinary least square method is that the error term in equation (6) is uncorrelated with each of the regressors. This implies that the explanatory variables are all pre-determined or are determined outside the system. Hence, for the explanatory instruments to be satisfied, they must be uncorrelated with the errors $[h e n c e : E (X_{t,} ε_{t}) = 0] .$

However, in practice, it is possible that a change in a firm’s decision-making strategy causes changes in that firm’s agency characteristics. A causal relation may exist between firm behaviour (characteristics or instruments) and action (the binary model) such that firms with unique behaviour may be more likely to pursue certain actions. If firms are self-selecting themselves with respect to creating a takeover bid based on their characteristics and on some other unobservable attributes, then self-selection will likely bias the estimates of the strategy choice variable due to the presence of endogeneity. Lag of explanatory variables could resolve the endogeneity problem, such that they correspond to annual reports prior to the annual report that precedes the announcement event. Therefore, the specification of equation (6) becomes:

P (B_{i} = 1 | Z_{i t}) = Φ (\begin{array}{l} α_{0} + β_{1} s i z e_{i t, t - 1} + β_{2} c a s h_{i t, t - 1} + β_{3} l e v_{i t, t - 1} + \\ β_{4} l i q_{i t, t - 1} + β_{5} r o e_{i t, t - 1} + β_{6} s_m i s v_{i t, t - 1} + \\ β_{7} t q_{i t, t - 1} + β_{8} o r g_{i t, t - 1} + β_{9} a g r o w t h 5_{i t, t - 1} + \\ β_{10} l c y c l_{i t, t - 1} + β_{11} t o e_{i t, t - 1} + β_{12} d_s i z e_{i t, t - 1} + \\ β_{13} d_n e x e_{i t, t - 1} + β_{14} d_d u a l_{i t, t - 1} + β_{15} d_e q_{i t, t - 1} + \\ β_{16} c e o_c h a i r_{i t, t - 1} + β_{17} c e o_t u r n v r_{i t, t - 1} + β_{18} I . s h o c k_{i t, t - 1} + \\ β_{19} I . g r w t h_{i t, t - 1} + β_{20} I . c o n_{i t, t - 1} + ε_{i t, t - 1} \end{array})

Although the result of equation (7) is not reported, the findings did not alter the result reported in Table 6. Thus, equation (7) reports the same result as equation (6).

8. Conclusion

This paper establishes that, by using the sample selection model, there are significant differences between bidder and non-bidding firms. This study employs a sample of 316 Australian publicly listed firms from the period 1999–2010. The results suggest that the bidding firms are significantly different from non-bidding firms in terms of their cash, leverage, organic growth structure and hubris behaviour. These findings are consistent with prior takeover studies and suggest that bidding firms’ management decisions are predisposed to enter into takeover activities due to being driven primarily by agency and hubris behaviour, rather than synergy motives. In the agency context, although the results support the empire-building hypothesis and the agency argument of large firm size, caution is required in interpretation as the relation can be positive, negative or even insignificant depending on type of variable measurement. The findings indicate that hubris behaviour is predominant in effecting management bid decisions compared to agency motives. The results also show that long-term productivity growth of a firm has a negative effect on management’s decision to engage in a takeover bid. This indicates that takeover and organic investments are mutually exclusive.

The paper defines bidders from non-bidding firms and the implication of this study is that management investment strategies have significant consequences across the economic, managerial and investment outcome levels. This study invites future research to extend its findings to incorporate both private firms and offshore takeover bid announcements to determine whether significant differences exist between private and public firms in the market for corporate control.

Footnotes

Final transcript accepted 8 April 2014 by Kathleen Walsh (AE Finance).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

References

Akhtar

(2013) Conditional returns to shareholders of bidding firms: An Australian study. PhD Dissertation, UQ Business School, The University of Queensland.

Andrade

Stafford

(2004) Investigating the economic role of mergers. Journal of Corporate Finance 10: 1–36.

Andrade

Mitchell

Stafford

(2001) New evidence and perspectives on mergers. Journal of Economic Perspectives 15: 103–120.

Baixauli

Fernandez

(2009) Toeholds and takeover probability: Implications for investment strategies. Studies in Economics and Finance 29: 69–86.

Barber

Lyon

(1997) Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of Financial Economics 43: 341–372.

Bhagat

Dong

Hirshleifer

. (2005) Do tender offers create value? New methods and evidence. Journal of Financial Economics 76: 3–60.

Cai

Song

Walking

(2011) Anticipation, acquisition and bidders returns: Industry shocks and the transfer of information across rivals. Review of Financial Studies 24: 2242–2285.

Cornett

Tanyeri

Tehranian

(2011) The effect of merger anticipation on bidder and target firm announcement period returns. Journal of Corporate Finance 17: 595–611.

Cremers

KJM

Nair

John

(2009) Takeovers and the cross-section of returns. Review of Financial Studies 22: 1409–1445.

10.

Delmar

Davidsson

Gartner

(2003) Arriving at the high-growth firm. Journal of Business Venturing 18: 189–216.

11.

Demsetz

Lehn

(1985) The structure of corporate ownership: Causes and consequences. Journal of Political Economy 93: 1155–1177.

12.

Dong

Hirshleifer

Richardson

. (2006) Does investor misevaluation drive the takeover market? Journal of Finance 3: 725–762.

13.

Eckbo

Maksimovic

Williams

(1990) Consistent estimation of cross sectional models in event studies. Review of Financial Studies 3: 343–365.

14.

Espahbodi

(2003) Binary choice models and corporate takeover. Journal of Banking & Finance 27: 549–574.

15.

Fama

French

(1992) The cross-section of expected stock returns. Journal of Finance 47: 427–465.

16.

Faria

(2008) Mergers and the market for organisation capital. Journal of Economic Theory 138: 71–100.

17.

Fuller

Netter

Stegemoller

(2002) What do returns to acquiring firms tell us? Evidence from firms that make many acquisitions. Journal of Finance 57: 1763–1793.

18.

Gorton

Kahl

Rosen

(2009) Eat or be eaten: A theory of mergers and firm size. The Journal of Finance 64: 1291–1343.

19.

Harford

(1999) Corporate cash reserves and acquisitions. Journal of Finance 54: 1969–1997.

20.

Harford

(2005) What drives merger waves? Journal of Financial Economics 77: 529–560.

21.

Harford

(2007) Decoupling CEO wealth and firm performance: The case of acquiring CEOs. Journal of Finance 62: 917–949.

22.

Harford

Mansi

Maxwell

(2008) Corporate governance and firm cash holdings in the US. Journal of Financial Economics 87: 535–555.

23.

Hart

(1995) Corporate governance: Some theory and implications. Economic Journal 105: 678–689.

24.

Heckman

(1979) Sample selection bias as a specification error. Econometrica 47: 153–161.

25.

Hess

(2010) Smart growth: Creating real long term value. Applied Corporate Finance 22: 74–82.

26.

Higgins

Rodriguez

(2006) The outsourcing of R&D through acquisitions in the pharmaceutical industry. Journal of Financial Economics 80: 351–383.

27.

Huyghebaert

Luypaert

(2010) Antecedents of growth through mergers and acquisitions: Empirical results from Belgium. Journal of Business Research 63: 392–403.

28.

Jensen

(1986) Agency costs of free cash flow, corporate finance and takeovers. American Economic Review 76: 323–329.

29.

Jensen

Murphy

(2010) CEO incentives: It’s not how much you pay, but how. The Journal of Applied Corporate Finance 22: 64–76.

30.

Lambrecht

(2004) The timing and terms of mergers motivated by economies of scale. Journal of Financial Economics 72: 41–62.

31.

Prabhala

(2007) Self-selection models in corporate finance. In: Eckbo

(ed.) The Handbook of Corporate Finance: Empirical Corporate Finance. New York: Elsevier, North-Holland, pp. 37–86.

32.

Loughran

Ritter

(2004) Why has IPO underpricing changed over time? Financial Management 33: 5–37.

33.

Madden

(2007) For better corporate governance, the shareholder value review. Journal of Applied Corporate Finance 19: 102–114.

34.

Madura

Ngo

(2008) Clustered synergies in the takeover market. Journal of Financial Research 31: 333–356.

35.

Maksimovic

Philips

(2001) The market for corporate assets: Who engages in mergers and asset sales and are there efficiency gains?. Journal of Finance 56: 2019–2065.

36.

Malmendier

Tate

(2008) Who makes acquisitions? CEO overconfidence and the market’s reaction. Journal of Financial Economics 89: 20–43.

37.

Masulis

Wang

Xie

(2007) Corporate governance and acquirer returns. Journal of Finance 62: 1851–1889.

38.

McWillliams

McWilliams

(2000) Another look at the theoretical issues in event study methodology. Journal of Applied Business Research 16: 1–11.

39.

Mitchell

Mulherin

(1996) The impact of industry shocks on takeover and restructuring activity. Journal of Financial Economics 41: 193–229.

40.

Moeller

Schlingemann

Stulz

(2004) Firm size and gains from acquisition. Journal of Financial Economics 73: 201–228.

41.

Moeller

Schlingemann

(2005) Global diversification and bidder gains: A comparison between cross-border and domestic acquisitions. Journal of Banking and Finance 29: 533–564.

42.

Moeller

Schlingemann

Stulz

(2005) Wealth destruction on a massive scale? A study of acquiring-firm returns in the recent merger wave. Journal of Finance 60: 757–782.

43.

Morellec

Zhdanov

(2008) Financing and takeovers. Journal of Financial Economics 87: 556–581.

44.

Offenberg

(2009) Firm size and the effectiveness of the market for corporate control. Journal of Corporate Finance 15: 66–79.

45.

Owen

Yawson

(2010) Corporate life cycle and M&A activity. Journal of Banking & Finance 34: 427–440.

46.

Pitelis

Teece

(2009) The new nature and essence of the firm. European Management Review 6: 5–15.

47.

Porter

Singh

(2010) What factor drive takeover in Australia? International Journal of Business and Economics 9: 87–103.

48.

Prabhala

(1997) Conditional methods in event studies and an equilibrium justification for standard event-study procedures. Review of Financial Studies 10: 1–38.

49.

Rhodes-Kropf

Viswanathan

(2004) Market valuation and merger waves. Journal of Finance 59: 2685–2718.

50.

Rhodes-Kropf

Robinson

Viswanathan

(2005) Valuation waves and merger activity: The empirical evidence. Journal of Financial Economics 77: 561–603.

51.

Roll

(1986) The hubris hypothesis of corporate takeovers. Journal of Business 59: 197–216.

52.

Shleifer

Vishny

(2003) Stock-market driven acquisitions. Journal of Financial Economics 70: 295–312.

53.

Stein

(1988) Takeover threats and managerial myopia. Journal of Political Economy 96: 61–80.

54.

Stulz

(1990) Managerial discretion and optimal financing policies. Journal of Financial Economics 26: 3–27.

55.

Travlos

(1987) Corporate takeover bids, methods of payment and acquiring firms’ stock returns. Journal of Finance 42: 943–963.

56.

Warusawitharana

(2008) Corporate asset purchases and sales: Theory and evidence. Journal of Financial Economics 87: 471–497.

57.

Zhu

(2013) Determinants of long-term debt issuing decisions: An alternative approach. Australian Journal of Management 38: 429–439.