Acquisition Premiums, Subsequent Workforce Reductions and Post-Acquisition Performance

Authors


Hema A. Krishnan, 406 Schott Hall, Department of Management, Xavier University, Cincinnati, OH 45207, USA (Krishnan@xavier.edu).

Abstract

abstract  This study suggests that paying acquisition premiums leads to workforce reductions in the merged firm, which in turn results in poorer post-acquisition performance. This issue is important to scholars and practising managers given the pervasiveness and importance of knowledge and human capital to competitive advantage. In a sample of 174 major related acquisitions completed in the period 1992–98, results show a positive relationship between the premium paid for an acquisition and subsequent workforce reductions, controlling for a number of alternative explanations. Additionally, workforce reduction mediates the negative relationship between premiums and post-acquisition performance. The results suggest that the effects of workforce reductions following large premiums paid for the acquired firm can be detrimental to the interests of the organization.

INTRODUCTION

The mergers and acquisitions (M&A) wave of the 1990s involved a large number of firms acquiring businesses with which they often shared substantial similarities (Business Week, 1999; Carey, 2000). Additionally, the acquiring firms frequently paid large premiums (additional price paid for an acquired company over its 30-day pre-acquisition value) (Sirower, 1997), and engaged in large-scale workforce reductions (Hayward and Hambrick, 1997; O'Shaughnessy and Flanagan, 1998). The merged firm must create adequate synergy to produce returns that exceed the premium paid and workforce reductions help to gain economies and reduce costs. However, they also involve the loss of human capital. Therefore, there is need for research on the effects of workforce reductions after acquisitions to better understand their long-term implications (Cascio, 2002).

Managers often justify the payment of premiums based on the potential synergies from the integration of the two firms. However, when large premiums are paid to acquire firms, creating the needed synergy becomes more challenging because of the higher returns required to cover the costs of making the acquisition (Sirower, 1997). To maximize the value created, managers are motivated to reduce costs where possible and to attempt to increase the returns. Some expenses are difficult to control but human capital costs are often among the easiest to reduce, particularly in the short term. Yet, because of the value of human capital for competitive advantage, excessive workforce reductions may be detrimental to organizational performance (Hitt et al., 2001a, 2006; Nixon et al., 2004; Wright et al., 2001).

There are legitimate reasons for large-scale workforce reductions following acquisitions in which premiums are paid. In highly related acquisitions (e.g. horizontal acquisitions), workforce reductions are generally necessary to eliminate redundancies and thereby extract operational synergies (i.e. economies of scale) (Conyon et al., 2002; O'Shaughnessy and Flanagan, 1998). However, identifying the precise number of employees to lay off and achieve these synergies is difficult because of the information asymmetries that often exist between acquiring and acquired firms and the complexity involved in integrating the two operations. These conditions, coupled with managerial hubris, encourage managers to overestimate the workforce reduction needed to achieve synergies, especially when faced with the financial pressures imposed by large acquisition premiums (Jemison and Sitkin, 1986; Pound, 1992; Useem, 1993). While some downsizing is required to gain the economies of scale, reducing the workforce more than needed especially when no asset sale is involved can reduce the effectiveness of the integration and harm the longer-term capabilities of the merged firm (Cascio, 2002; Krishnan and Park, 2002).

For the reasons mentioned above, we suggest that in related acquisitions, workforce reduction serves as a partial mediator of the relationship between premiums paid and post-acquisition performance. Post-acquisition performance is a highly important issue given the corporate emphasis on the acquisition strategy. As such, this study fills an important gap, contributing to our knowledge of the determinants of post-acquisition performance. The remainder of this paper is organized as follows. The next section provides the conceptual framework and the hypotheses, following which we discuss the data and methodology. This is followed by the results section. In the final section, we discuss the implications of our study.

CONCEPTUAL FRAMEWORK

The underlying motives for acquisitions suggest reasons that organizations pay premiums and the actions that executives take to recoup them. Previous research coupled with the anecdotal evidence on acquisitions suggests that synergy is the primary motive for many acquisitions (Hitt et al., 2001b). Synergy can be created with acquisitions in several ways. For example, it is commonly argued that acquisitions produce efficiency; that is, economies of scale or scope are achieved via consolidation, elimination of redundancies and inefficiencies, and improvements in value chain activities. Synergy can also be achieved through integration of complementary resources from the two firms. Complementarity is based on different but mutually supportive resources that when integrated can help produce more valuable products (Chatterjee, 1992; Ravenscraft and Scherer, 1987; Hitt et al., 2001b). Additionally, synergy can be derived from market power achieved when the combined entity has greater bargaining power in the marketplace (i.e. because of larger size and economies), especially with key constituencies such as suppliers and customers.

Acquisitions can also result from agency problems based on executives' desire to enhance their personal power or to maximize their personal wealth often at the expense of the firm's shareholders (Trautwein, 1990). Acquisitions can result from CEOs' hubris i.e. overconfidence in their personal capabilities to achieve synergy. While these theoretical propositions are difficult to test, some research supports both explanations (i.e. Finkelstein, 1992; Hayward and Hambrick, 1997). An executive willing to pay a high premium because the acquisition will increase his/her personal wealth and/or power represents managerial opportunism. Alternatively, an executive willing to pay a high premium for an acquisition because he/she is convinced of the personal capabilities needed to realize synergy represents managerial hubris. In fact, Hiller and Hambrick (2005) argue that executives with a strong core evaluation (often produces hubris) are likely to make quick strategic decisions with little analyses, engage in large-scale strategic actions (e.g. large acquisitions) and have extreme confidence that they can implement the decision successfully.

Certainly, competition for acquiring the target can also contribute to premiums but in such competition, premiums are likely to be higher if the acquisition is motivated by managerial hubris or opportunism (Haunschild, 1994; Slusky and Caves, 1991; Varaiaya and Ferris, 1987). Nonetheless, potential synergies in an acquisition make it easier to justify premium payments because they can potentially enhance performance via operational efficiencies and complementarities (Slusky and Caves, 1991).

Research suggests that over the last 20+ years, acquiring firms have paid, on average, a 40–50 per cent premium for target firms (Haunschild, 1994; Hayward and Hambrick, 1997). Unfortunately, research also suggests a negative relationship between premiums paid and the returns to the acquiring firm's shareholders (Datta et al., 1992; Haunschild, 1994; Hayward and Hambrick, 1997; Varaiaya and Ferris, 1987). One of the primary reasons for this negative association is that firms are unable to realize the potential synergies and their operating performance often deteriorates after the acquisition is completed. Payment of a premium to acquire a firm increases the pressure on the acquiring firm's top executives to create returns for the shareholders beyond the cost of the acquisition. If potential synergies cannot be achieved, even a small premium could be considered excessive (Hitt and Pisano, 2003). Thus, executives who pay a premium may feel pressure to reduce costs with the hope of achieving greater returns. Sometimes they sell off assets of the acquired firm but selling assets separately is unlikely to achieve returns equal to the value of those assets in the acquired firm's asset portfolio (in other words, they may have to sell them at lower than true market value). Another strategy to reduce costs and maintain shareholder value is workforce reduction in the merged firm (Jensen, 1991; Johnson et al., 1993; Nixon et al., 2004; Useem, 1993; Zuckerman, 2000; Vaara et al., 2005).

Premiums and Subsequent Workforce Reduction

While firms attribute synergy as a predominant motive for premium payment, workforce reductions following the payment of large premiums may be motivated by synergy and other factors (Bethel and Liebeskind, 1993; McKinley et al., 2000). Workforce reduction, undertaken during the acquisition integration process, streamlines operations and reduces redundancy, thereby contributing to synergy creation (Bethel and Liebeskind, 1993; Bowman and Singh, 1993; Conyon et al., 2002; Gutknecht and Keys, 1993; Harrison, 1999; Ravenscraft and Scherer, 1987). Streamlining involves more efficient use of resources, whereas reductions in redundancy limit extra costs due to unnecessary duplicate resources. By eliminating redundant activities, firms should gain operational synergies arising out of economies of scale and thereby improve profitability (Capron, 1999; O'Shaughnessy and Flanagan, 1998; Porter, 1987). Announced workforce reductions provide a signal to the market of the potential synergies the acquiring firm should achieve (Bowman and Singh, 1993; Cascio et al., 1997; Demuse et al., 1994).

Yet, capturing the potential synergies can be a challenging task. Information asymmetries often exist between the acquiring and acquired firms. The asymmetries may be especially acute if hubris is a reason for the acquisition because it reduces the likelihood of effective due diligence (Hitt et al., 2001b). Information asymmetries make it difficult to identify the specific actions needed to achieve the synergies and to effectively integrate the two firm's operations. Furthermore, integration is a complex process (Haspeslagh and Jemison, 1991) and identifying precisely the number of employees to lay off in order to achieve the economies is difficult.[1]

Theory suggests that the framing of the acquisition implementation decisions is important in the actions chosen by the acquiring firm managers. This work is based on the commonly accepted notion of bounded rationality. Because managers are unable to process complete information for any one decision, biases may affect it (Wiseman and Gomez-Mejia, 1998). When managers frame a decision in potential gains, they take less risky decisions designed to avoid losses (March and Shapira, 1987). If managers pay a large premium to acquire a firm, they clearly expect important performance enhancing synergies. A major workforce reduction is likely perceived as a relatively low-risk alternative (low uncertainty because the cost savings can be easily calculated) to ensure a positive outcome (Morrow et al., 2005). Loss aversion may only be greater when the pressure by shareholders on managers to achieve synergies after acquisitions is high. Such loss aversion is likely to lead them to overestimate rather than underestimate the workforce reduction needed. The higher the premium paid, the greater the pressures to achieve synergy (Sirower, 1997) (or take actions to avoid losses). Executives managing the acquisition, search for ways to economize and to reduce costs (in order to avoid losses and hopefully enhance profits). Sirower (1997) argues that large acquisition premiums often translate into specific required performance targets that are difficult to realize. Large premiums impose greater pressures to generate free cash flows (Zuckerman, 2000). Hitt et al. (1994) interviewed managers about decisions to make large layoffs. They found that managers frequently overestimated the size of the layoff needed with the intent of signalling to the shareholders and investors of the efficiency of their operations and the likelihood of earning higher returns.

Reducing costs contributes to the acquiring firm's ability to avoid losses due to the large premium payment and achieve positive returns from the acquisition in a shorter period of time (Abowd et al., 1990; Gibbs, 1993; Harrison, 1999; Johnson, 1996; Useem, 1993). Further, it is generally easier to implement cost reductions (costs are more controllable) than to implement actions that enhance revenues (Carey, 2000). Finally, workforce reductions are often targeted by managers because they represent one of the easier costs to control and the reductions can be made more quickly than many other means of decreasing costs.

The arguments presented above suggest that there is a positive relationship between the size of acquisition premiums paid and the amount of workforce reductions in the merged firm, after controlling for the reductions necessary to eliminate redundancies, create efficiencies and achieve economies of scale. Thus, the following hypothesis is proposed.

Hypothesis 1: There is a positive relationship between the premium paid for an acquisition and subsequent workforce reductions in the merged firm.

Workforce Reductions and Post-Acquisition Performance

Since the 1990s there has been considerable interest among researchers centred on the linkage between workforce reductions and firm performance. However, much of this research focuses on restructuring the firm's assets where workforce reduction is a secondary outcome (i.e. Johnson et al., 1993). In support, Love and Nohria (2005) found downsizing to have a positive effect on performance, but only when the firms have high slack, when the planned downsizing is broad (such as in restructuring of assets) and when it is proactive (rather than in response to other strategic actions such as acquisitions). Most of the recent research on downsizing has concluded that workforce reductions often lead to lower performance unless they are planned and implemented carefully, with special sensitivity to human resources and the manner in which these resources are managed (Chadwick et al., 2004; Nixon et al., 2004). In fact, reactions of the stock market to workforce reduction announcements are likely to be negative unless the reductions are well planned and the negative effects on employees are limited (Chadwick et al., 2004; Nixon et al., 2004). Scholars argue that workforce reduction may be detrimental to organizational performance because it erodes the resource base of the combined firm (Cascio et al., 1997; Collis and Montgomery, 1995). This leads to the conclusion that workforce reductions are unlikely to have positive effects on firm performance unless they represent well planned proactive and independent strategic initiatives (with particular sensitivity to the protection of valuable human capital).

Workforce reductions that go beyond those needed to achieve economies can be harmful to the merged firm because of the loss of knowledge and skills held by the terminated employees (Walsh, 1988). Tacit knowledge held by human capital helps firms implement valuable strategic actions, thereby allowing them to earn returns that recoup the premiums paid for the acquired companies. Often the potential synergies from mergers are based on integrating the intangible assets of the two firms. Integration and use of the knowledge embedded in the two firms' human capital is necessary to create the synergies (Hitt et al., 2001a). Losses of such knowledge reduce the potential to create the synergies. Therefore, the loss of valuable human capital is likely to have a negative effect on organizational performance.

In addition, large-scale workforce reductions are likely to create disgruntled employees who lack the motivation to facilitate resource sharing and transfer. After large workforce reductions, problems in the relationships between managers and employees are magnified (Leana and Feldman, 1992; Lee, 1997). Additionally, research has shown that survivors of major layoffs experience guilt that in turn negatively affects their productivity (Brockner et al., 1992). Customer service may also be adversely affected in the long term with the loss of employee skills and the survivors may be less willing to invest effort in actions to satisfy the customers. Additionally, managers under pressure due to large premiums may also attempt to speed the integration process to consolidate the operations of the two firms, producing unsatisfactory results. These arguments suggest that large premiums lead to workforce reductions, which in turn lead to lower firm performance. Premiums may also affect other operations in the firm (e.g. expense reductions – postponing required maintenance of assets) and therefore we suggest that workforce reductions act as a partial mediator of the premium–performance relationship. That is, the premiums paid partly affect performance through the proximate variable, workforce reduction. Thus, the following hypotheses are proposed.

Hypothesis 2: There is a negative relationship between workforce reductions and post-acquisition performance.

Hypothesis 3: The relationship between premium and post-acquisition performance is partially mediated by workforce reduction. Specifically, as a higher premium is paid, more workforce reduction occurs, thereby decreasing the post-acquisition performance of the organization.

METHODOLOGY

Data and Sample

The initial sample was drawn from the Securities Data Corporation (SDC) Database on mergers and acquisitions and comprised firms that satisfied several criteria. The acquisitions were completed during the period 1992–98, and both the acquiring and acquired firms were US based, publicly held, and broadly related at the two-digit SIC level. An overwhelming majority of the acquisitions that occurred in the 1990s were related at the two-digit SIC level. This marked the continuation of a trend witnessed in the previous decade (1980s) with firms increasingly focused on strengthening their core businesses via divestments and selected acquisitions (Haynes et al., 2002). Financial institutions were excluded from our sample because they have different asset characteristics and objectives concerning financial risk. Additionally, they are generally subject to more regulatory oversight and restrictions (e.g. their financing decisions and capital structure). Only acquiring and acquired firms that had revenues of at least $200 million at the time of the acquisition were included in the sample because acquisition premiums often are greater and integration concerns are more profound in larger acquisitions (Sirower, 1997). Also, large acquisitions have a stronger potential impact on market valuation. Third, it is easier to obtain secondary data on larger firms, which partly explains their popularity in empirical research on acquisitive activity. Next, transactions were identified in which the acquiring firm had not acquired additional targets with revenues of $200 million or more for at least two years following the year studied. This criterion was imposed to ensure that workforce reductions were not a result of multiple acquisitions and therefore related only to the acquisition (and premium) under study. The final criterion stipulated that the acquiring firm should not have divested the acquired firm for at least two years after the acquisition. A total of 353 transactions met these criteria. Of these acquisitions, complete data were unavailable on 175 transactions (e.g. lack of workforce reduction announcement information in the Wall Street Journal and other publications; lack of performance data on the acquiring firm in Compustat). Complete data were available for 178 of these transactions for inclusion in the study. The mean revenues between firms excluded from the final sample and firms included were compared. t-Tests revealed no statistically significant differences in the two groups, suggesting no exclusion bias.

The period 1992–98 was chosen for two reasons. First, following the recession of the late 1980s and early 1990s, this period represented a time of strong growth in the US economy, making workforce reductions less likely because of recessionary pressures (Merger Stat Review; Workforce, 1996). Second, Merger Stat Review reports that following the recession of the late 1980s, the period 1992–98 was particularly active in M&As in the USA. Hence, 1992–98 was a particularly robust period in the US economy. In our preliminary analysis, we used dummy variables to control for the year-by-year effects of the period studied, 1992–98. However, none of these variables was statistically significant. Because of the lack of effects and to avoid reducing the degrees of freedom in the statistical models, we eliminated them from subsequent analysis.

Preliminary analysis employing Cook's test revealed four firms to be outliers. As a result, these four firms were eliminated from the final sample and data from the remaining 174 firms were used in the analyses. These outliers are explained in greater detail in the discussion section.

Primary Variables

Independent variable (time t).  Consistent with previous research in both strategy and finance (Haunschild, 1994; Hayward and Hambrick, 1997; Sirower, 1997; Slusky and Caves, 1991), an acquisition premium was defined as the percentage difference between the final purchase price for the target (acquired firm) and the trading price of the target's stock 30 days before the first announcement of the acquisition made in the Wall Street Journal. The 30-day period is used to control for information leakage. That is,

Premium = (Purchase Price − 30-day Pre-takeover Price) / 30-day Pre-takeover Price

Additionally, the premium was adjusted for the market movement in the Standard & Poor's Index. A 30-day period is typically used in M&A research because it reflects the stock price of the target before leakage of information about the acquisition to the market. These data were obtained from CRSP tapes (Center for Research in Security Prices). The average premium in this data set was 0.39 (i.e. 39 per cent) with a standard deviation of 0.33, consistent with other studies on premiums (Haunschild, 1994; Hayward and Hambrick, 1997).

Mediating variable (up to time t + 1).  Workforce reduction in the merged firm, the mediating variable was measured as the percentage reduction in the number of employees in the combined organization announced after the acquisition similar to Bethel and Liebeskind (1993). Workforce reduction announcements were obtained for a period of one year following the acquisition. Consistent with previous research (e.g. Nixon et al., 2004) the data on workforce reductions were collected from the Wall Street Journal Index and Lexis-Nexis Academic Universe using the following approach. First, all major national newspapers including Wall Street Journal, New York Times, Washington Post, Los Angeles Times, and Boston Globe that reported layoff announcements in either the acquiring firm, acquired firm or combined firm were screened. Second, newspapers with a large regional presence were also screened for layoff announcements. Finally, the ratio of the announced number of employee layoffs in the merged firm to the total number of employees in its workforce was calculated. Several researchers have argued that the announced workforce reduction is a highly valid measure (Worrell et al., 1991) especially because of its immediate effect on the stock market. Announced workforce reductions, particularly those related to mergers/acquisitions, signal firms that intentionally downsize in order to capture synergies. Also, O'Shaughnessy and Flanagan (1998) argue that there are substantial organizational costs associated with a workforce reduction announcement that make it a credible commitment on the part of the organization. Large scale merger-related workforce reductions are likely to have a significant impact on the total employee base of the combined organization. Hence, as a means of validation, we compared announced workforce reductions with actual reductions using the following procedure. We obtained data on workforce reduction announcements and also data on the total employment of the two firms in the year prior to the acquisition and of the merged firm in the year after the acquisition. The data on total employment, required to be disclosed by publicly-held companies, were obtained from company annual reports. The changes in employment were positively correlated with the percentage of workforce reductions announced (r = 0.33; significant at p < 0.01). The correlation between the two variables is only modest as expected because besides merger related employee reductions, firms often undertake workforce reductions as part of a broader restructuring strategy and in other businesses in their portfolio (unrelated to the merger). Such actions are more likely in large companies.

Dependent variable (time t + 2, t + 3).  Post-acquisition performance, the dependent variable, was measured for the merged firm using two-year average industry-adjusted return on sales (ROS) consistent with previous studies on acquisitions (Hitt et al., 2001b). ROS is a more effective performance measure in acquisitions because of the often large changes in assets and equity and potentially odd valuations observed in M&As (Hitt et al., 2001b). Also, it is directly related to our central argument on synergy generation in the context of acquisition-related workforce reduction and captures actual performance conditions instead of investor expectations. Third, it is a more robust variable than market measures because it is less sensitive to unexpected economic changes and information asymmetries in the market and reflects the internal effectiveness of managing resources in the merged firm. Using Compustat business segment data at the four-digit SIC level, consistent with previous research (Barker and Mueller, 2000; Baysinger et al., 1991), we computed the weighted average industry performance based on a firm's level of sales in the industries in which it operated and subtracted it from the firm's ROS to calculate the dependent variable. This approach was used to control for potential industry effects on returns.

Control Variables

We used several control variables in the study; they are described below.

Prior performance of acquiring firm.  Firms performing poorly relative to their industry are more likely to implement workforce reductions to reduce costs (John et al., 1992; O'Shaughnessy and Flanagan, 1998). The data on prior performance of the acquiring firm were collected from Compustat and measured as a two-year average industry-adjusted ROS prior to the acquisition.

Relative organizational size.  When the acquirer and the acquired firms are of comparable sizes, perceived redundancies and duplication of activities are likely to be higher. Hence, the smaller the size differential, the greater are the challenges experienced in integrating the two firms (Haspeslagh and Jemison, 1991; Hitt et al., 2001b). The data on relative organizational size, measured as the ratio of the revenues of the acquiring firm to the revenues of the acquired firm at the time of the acquisition, were gathered from company annual reports.

Type of payment.  Cash payments are more likely to result in higher levels of workforce reduction compared to stock payments. Cash acquisitions are often partly or largely financed with debt (Hitt et al., 2001b). When the acquirer's growth opportunities are higher, the company is more likely to use stock to finance the acquisition (Martin, 1996). Consistent with previous research (Brown and Ryngaert, 1991; Kesner et al., 1994), this variable, collected from Merger Stat Review, was coded as 1 for pure cash transactions and 0 for stock and cash/stock combination transactions.

Relatedness.  Higher relatedness between the two firms suggests greater redundancies in their activities. Hence, we would expect larger workforce reductions to capture the economies of scale in the merged firm (Capron, 1999). Consistent with previous research, relatedness was measured as 1 if the two partners shared one or more four-digit SIC level commonalties in their businesses and 0 otherwise.

Prior performance of acquired (target) firm.  Acquired firms that performed poorly prior to the acquisition are more likely to experience greater workforce reductions in their employee ranks after acquisition (O'Shaughnessy and Flanagan, 1998). The data on prior performance were collected from Compustat and measured as a two-year average industry-adjusted ROS prior to the acquisition.

Multiple bidders.  The presence of multiple bidders for a target firm may have implications for workforce reduction and performance (Beckman and Haunschild, 2002). Although some researchers argue that premiums tend to be larger when there are multiple bidders for the acquired firm, several studies reveal that the effect of an acquisition premium is sufficiently independent of multiple bidders to display separate effects (Sirower, 1997). This variable was coded as 1 for multiple bidders and 0 otherwise. The data on multiple bids were obtained from the SDC Database.

Leverage.  Debt may play an important role in the outcomes of acquisitions because debt costs affect performance (Hitt et al., 1998). The need to service high interest costs, following huge debt, places pressure on management to restructure (Haynes et al., 2003). To generate cash in order to pay down debt, firms may engage in workforce reduction (Jensen, 1991). Following the approach advocated by Hitt et al. (2001b), we controlled for leverage (measured as debt/equity) of the merged firm in the year of the acquisition.

Acquisition motives.  The motive for the acquisition, in other words, the type of synergy pursued, efficiency, complementarity, or market power may have an effect on workforce reduction (Haspeslagh and Jemison, 1991). Hence, these three broad categories of motives were identified for the firms in the sample. The data were collected from Lexis Nexis academic universe using the following approach. First, press releases around the time of the acquisition announcement were collected. Next, two graduate students not connected with the study analysed the documents and coded the acquisitions using the following criteria. Acquisitions were coded as having a ‘market power’ motive if they were subjected to scrutiny by antitrust regulators. If the major motive cited by the firm was to reduce cost via elimination of duplication and consolidation, it was coded as an ‘efficiency’ motive. If the major motive cited was to extend the product line or expand geographically, the acquisition was coded as a ‘complementarity’ motive. The inter-rater reliability was 95 per cent. We used three dummy variables, corresponding to each of the three motives. These dummy variables include ‘market power’, which takes the value of 1 if the primary motive for the merger was to increase the market power of the merged entity and 0 otherwise (to disallow a unitary matrix); ‘efficiency’, which takes a value of 1 if the primary motive for the merger was to increase efficiency and 0 otherwise; and ‘complementarity’, which takes a value of 1 if the primary motive for the merger was complementarity and 0 otherwise. Market power was the omitted variable in the regression analyses (about 7 per cent of the firms had this motive).

Board composition.  The board of directors is the primary governance mechanism for firms to control agency costs (Sundaramurthy et al., 1997). Previous research (Gibbs, 1993; Jensen, 1991) suggests that when top managers engage in acquisitions involving huge premiums to increase the size and scope of their operations, the board may force the firm to restructure. Often, reductions in the workforce accompany such restructuring activity. To control for this variable, we included the ratio of inside directors plus the number of directors selected by the CEO divided by the total number of directors in the acquiring firm at the time of the acquisition. The number of insiders plus directors appointed by the CEO averaged 58 per cent of the total number of board members. The data were obtained from company proxy statements. We chose a broad measure such as board composition rather than a more narrow measure such as CEO compensation to control for agency effects because in large firms, restructuring is often motivated by the board (Jensen, 1991) and less likely by one individual such as the CEO.

RESULTS

The descriptive statistics and intercorrelations for the variables in the study are presented in Tables I and II. Results of tests revealed no significant multicollinearity problems among the major variables in the sample (Belsey et al., 1980) and suggested that the data are normally distributed. The correlation table revealed post-acquisition performance to have statistically significant relationships, in the direction expected, to workforce reduction, premium, prior performance of the acquiring firm, prior performance of the acquired firm, and leverage.

Table I.  Means for variables used in the analyses (N = 174)
VariableMean (standard deviation)
Post-acquisition performance of merged firm−1.49 (5.12)
Workforce reduction in the merged firm0.035 (0.04)
Prior performance of acquiring firm−0.14 (4.50)
Relative organizational size7.02 (11.73)
Type of payment (1 = all equity or combination, 0 = cash)0.70 (0.46)
Relatedness0.60 (0.49)
Prior performance of acquired firm−2.31 (7.56)
Multiple bidders0.24 (0.43)
Leverage (after acquisition) (debt to equity ratio)1.26 (1.95)
Acquisition motive – efficiency0.53 (0.50)
Acquisition motive – complementarity0.40 (0.49)
Acquisition motive – monopoly0.07 (0.24)
Agency0.58 (0.26)
Acquisition premium0.39 (0.33)
Table II.  Pearson correlation coefficients (N = 174)
 Variables1234567891011121314
  1. Notes: * p < 0.05,** p < 0.01,*** p < 0.001.

  2. Intercorrelation between efficiency and complementarity motives is negative and high because dummy variables were used to code them. Also, 93% of the firms in the sample espoused one of these two motives.

 1Post-acquisition performance merged firm1             
 2Workforce reduction−0.41***1            
 3Prior performance of acquiring firm0.24***0.011           
 4Relative organizational size0.12−0.24***−0.14*1          
 5Type of payment0.050.080.14*−0.21**1         
 6Relatedness−0.140.27***0.07−0.28***0.121        
 7Prior performance of acquired firm0.33***−0.27***0.23**−0.010.08−0.081       
 8Multiple bidders0.08−0.020.01−0.04−0.16*−0.060.001      
 9Leverage−0.26***0.19**−0.14*−0.08−0.15*0.13−0.08−0.031     
10Acquisition motive – efficiency−0.10−0.02−0.06−0.22**0.030.17*0.070.01−0.021    
11Acquisition motive – complementarity0.09−0.120.040.27***−0.02−0.24***0.01−0.080.03−0.8***1   
12Acquisition motive – monopoly0.0080.28***0.03−0.09−0.030.11−0.16*0.13−0.02−0.28***−0.21**1  
13Agency−0.10−0.04−0.10−0.08−0.001−0.020.050.040.16*0.08−0.03−0.101 
14Acquisition premium−0.16*0.20**−0.040.10−0.11−0.010.000.000.13−0.04−0.030.15*0.021

Mediated regression analysis was used to test the hypotheses in the study and the results are graphically depicted in Figure 1 (Baron and Kenny, 1986; MacKinnon and Dwyer, 1993). We used this technique based on the theoretical arguments presented earlier. However, caution should be exercised because workforce reduction is not necessarily the only path between premium and performance; there could be other paths that are not estimated in our study.

Figure 1.

Mediating influence of workforce reduction on the premium and post-acquisition performance relationship
Notes: * p < 0.05, ** p < 0.01, *** p < 0.001.
Numbers denote unstandardized estimates from the regression results in Tables III and IV.
Unstandardized path estimates are used for Sobel test.
Standardized estimates are in parentheses.

Mediator Analysis and Results

Workforce reduction plays a mediating role if the following conditions are met:

  • • First step: Variations in the premium paid have a significant effect on the outcome variable, post-acquisition performance. This establishes the presence of a direct relationship between the two variables.
image(1)

where Y is the post-acquisition performance, and X is the premium paid. This relationship is represented as path ‘c’ in Panel A of Figure 1.

  • • Second step: Variations in the premium paid significantly account for variations in the presumed mediator (workforce reduction). In this step, the mediator is treated as the outcome variable and premium is the independent variable and the following regression equation is estimated.
image(2)

where M is the workforce reduction, and X is the premium paid. This relationship is represented as path ‘a’ in Panel B of Figure 1.

  • • Third step: Variations in the mediator (workforce reduction) have significant effects on the dependent variable (post-acquisition performance). In this step, post-acquisition performance is the dependent variable in a regression equation and premium and workforce reduction are predictors (independent variables).
  • • Fourth step: To establish mediation, the effect of premium on post-acquisition performance controlling for workforce reduction is tested. The effects in Steps 3 and 4 are estimated in the same regression equation, which is estimated as follows.
image(3)

where Y is the post-acquisition performance, X is the premium paid, and M is the workforce reduction. This relationship is represented in Panel B of Figure 1, in the paths ‘b’ and ‘k’.

  • • For mediation to exist, once the influence of regression coefficient b is accounted for, path ‘c’ between premium paid and post-acquisition performance in panel A should reduce in significance or no longer be statistically significant. This is represented as path ‘k’ in Panel B.

The results of the mediated regression analysis are presented in Tables III and IV. Model 1 in Table III presents the results of regressing workforce reduction on the control variables. Our results reveal that the control variables explained 19 per cent of the variance (adjusted for the number of variables in the model) and the model was statistically significant. Relatedness has a positive relationship with workforce reductions as expected (O'Shaughnessy and Flanagan, 1998). Prior performance of the acquired firm, as expected, has a negative relationship with workforce reductions suggesting that firms may seek to turn around the organization in an effort to improve efficiencies. Leverage is positively related to workforce reductions consistent with arguments that firms using debt to finance their acquisitions are more likely to reduce overall costs to pay the debt expenses without reducing profits. The acquisition motive complementarity, as expected, is negatively related to workforce reduction which shows that such acquisitions are less likely subject to duplication in value chain functions and hence, these firms are less likely to reduce their workforce. The efficiency motive was statistically significant, but in the direction opposite to our arguments. In Model 2, workforce reduction and the control variables are regressed on acquisition premiums. The results show a statistically significant positive relationship between premium and workforce reduction, thus providing support for Hypothesis 1.

Table III.  Results of regression analysis; dependent variable: workforce reduction in the merged firm
Predictor(1)
Control variables
(2)
Full model
  1. Notes: * p < 0.05,** p < 0.01,*** p < 0.001.

  2. Numbers denote unstandardized estimates (standard errors in parentheses).

Intercept0.05 (0.016)0.034 (0.02)
t = 3.42***t = 2.40**
Prior performance of acquiring firm0.0004 (0.0007)0.0005 (0.0007)
t = 0.62t = 0.67
Relative organizational size−0.0005 (0.0003)−0.0006 (0.0002)
t = −1.99*t = −2.32*
Type of payment0.007 (0.007)0.008 (0.006)
t = 1.03t = 1.22
Relatedness0.013 (0.006)0.014 (0.006)
t = 2.06*t = 2.13*
Prior performance of acquired firm−0.0013 (0.0004)−0.001 (0.0004)
t = −3.21***t = −3.39***
Multiple bidders0.0014 (0.007)0.0019 (0.007)
t = −0.20t = 0.28
Leverage0.0037 (0.0016)0.0032 (0.002)
t = 2.3*t = 2.02*
Acquisition motive – efficiency−0.037 (0.013)−0.032 (0.013)
t = −2.9**t = −2.5**
Acquisition motive – complementarity−0.037 (0.013)−0.031 (0.013)
t = −2.81**t = −2.35*
Agency−0.0025 (0.012)−0.003 (0.012)
t = −0.21t = −0.29
Acquisition premium 0.025 (0.01)
 t = 2.77**
Intercept0.00.0
R20.240.27
Adjusted R20.190.22
ΔR2 0.03*
F value5.15***5.6***
N174174
Table IV.  Results of mediated regression analysis; dependent variable: post-acquisition performance of merged firm
Predictor(1)
Control variables
(2)
Full model
  1. Notes: * p < 0.05,** p < 0.01,*** p < 0.001.

  2. Numbers denote unstandardized estimates (standard errors in parentheses).

Intercept2.65 (1.96)4.22 (1.89)
t = 1.35t = 2.23*
Prior performance of acquiring firm0.19 (0.08)0.21 (0.08)
t = 2.29*t = 2.65**
Relative organizational size0.05 (0.03)0.024 (0.03)
t = 1.50t = 0.76
Type of payment0.09 (0.81)0.42 (0.77)
t = 0.11t = 0.55
Relatedness−0.71 (0.78)−0.17 (0.74)
t = −0.93t = −0.23
Prior performance of acquired firm0.19 (0.05)0.14 (0.05)
t = 3.99***t = 2.93**
Multiple bidders0.87 (0.84)0.95 (0.80)
t = 1.04t = 1.20
Leverage−0.42 (0.19)−0.29 (0.18)
t = −2.18*t = −1.57
Acquisition motive − efficiency−1.81 (1.53)−3.07 (1.48)
t = −1.18t = −2.08*
Acquisition motive − complementarity−1.16 (1.60)−2.41 (1.53)
t = −0.73t = −1.57
Agency−1.69 (1.39)−1.83 (1.32)
t = −1.22t = −1.39
Acquisition premium−2.35 (1.10)−1.34 (1.06)
t = −2.14*t = −1.26
Workforce reduction −40.02 (8.95)
 t = −4.4***
Intercept0.00.0
R20.240.32
Adjusted R20.190.27
ΔR2 0.08**
F value4.75***6.5***
N174174

Model 1 in Table IV shows the effects of regressing post-acquisition performance on the control variables and premium. Model 2 in Table IV provides a test for Hypotheses 2 and 3. Premium and workforce reduction are entered into a model as predictors of post-acquisition performance, the dependent variable. The results show a statistically significant negative main effect of workforce reductions on post-acquisition performance. Note also from our variable means depicted in Table I and our unstandardized parameters depicted in Table IV, that for every 1 per cent increase in workforce reduction, the post-acquisition performance decreases by 0.40 units. In other words, if there is aworkforce reduction of 3.5 per cent (the sample average) in the firm, the performance drops by 1.4 units. This is a significant decrease especially because industry effects on performance are controlled. These results provide support for Hypothesis 2.

To establish mediation, we test for the effects of premium on post-acquisition performance after controlling for workforce reduction. The results are presented in Figure 1. Notice that the direct path ‘c’ (Panel A) shows that the relation between premium and post-acquisition performance without including workforce reduction is statistically significant, but after the addition of workforce reduction to the regression model, the relationship between premium and post-acquisition performance, direct path ‘k’, is no longer statistically significant (Panel B). Also notice that path ‘a’, which examines the relation between premium and workforce reduction and path ‘b’, which examines the relation between workforce reduction and post-acquisition performance are statistically significant at conventional levels (Panel B). These results are consistent with Hypothesis 3 and suggest that workforce reduction mediates the relationship between premium and post-acquisition performance. Interestingly, while we hypothesized the mediating effect to be partial, the results suggest a full mediating effect. The direct relationship between premium and post-acquisition performance is statistically significant at the 0.05 level (p < 0.03). However, with the addition of the mediating variable, the relationship between premium and performance is not statistically different from zero at conventional levels (p < 0.20). Also, the 95% confidence interval for the value of ‘k’ includes zero (range is from −3.44 to 0.75). Furthermore, the actions taken to reduce the potential for specification error provide more confidence in the accuracy of the conclusion of complete mediation. To test for the statistical significance of the mediated effect, we used the approach advocated by MacKinnon and Dwyer (1993) and Sobel (1982). We calculated a z-score using the following formula:

zab = a * b/seab where a and b are the unstandardized regression coefficients.

inline image where sea and seb are the standard errors for a and b. Our calculation yielded a z score of −2.3, implying that the mediation is statistically significant at the p < 0.05 level (p < 0.02). Finally, using Cohen's (1968) method, we tested for the change in R2, gains from adding the new variables in Model 2, and found it to be statistically significant (ΔR2 = 0.08; p < 0.01).

We also conducted a post-hoc analysis to examine if workforce reduction has a curvilinear (inverted U) relationship with performance. After a related acquisition, firms may find it necessary to reduce some workforce to improve efficiency, thereby suggesting moderate levels of workforce reduction could have a positive effect on performance. In other words, at very low levels of workforce reduction, there could be an adverse impact on performance because of the failure of the firm to generate efficiencies by continuing to retain redundant and duplicate functions. Similarly, very high levels of workforce reduction are likely to result in a loss of valuable assets, thereby dissipating firm performance. Our results reveal no such relationship, thus supporting our argument that the relationship between workforce reduction and performance is linear.

DISCUSSION

Many acquisitions perform below expectations partly because their costs exceed the benefits they provide. When an acquiring firm pays a premium for the target firm, it must achieve synergies that exceed the costs of the acquisition including the premium paid. The larger the premium, the greater the challenges managers experience in producing positive returns. Integration of the assets and operation of two formerly independent firms are highly complex processes. Thus, managers search for means to create synergies that can be implemented relatively easily and are visible to investors. Workforce reductions satisfy these criteria.

Some workforce reductions could be necessary to obtain economies of scale. However, because of the information asymmetries and complex interdependencies among jobs in operations, identifying the correct amount of staff reduction is difficult. Furthermore, causal ambiguity and social complexity of firms' capabilities make it exceedingly difficult to choose whom to lay off, much less how many. Therefore, workforce reductions are challenging tasks for managers. When premiums have been paid for the acquisition, managers are likely to overestimate the workforce reductions needed. The results support these arguments as higher premiums lead to larger workforce reductions.

Our results also reveal that the effects of workforce reduction are largely negative instead of the positive effects desired by managers. To achieve the synergies usually requires the integration of intangible resources (Seth et al., 2002). One important intangible resource is human capital and the tacit knowledge embedded in it. While integration of the two firms requires integration of information systems, product and service capabilities and distribution systems, among others, such integration necessitates cooperation and the internalization of the knowledge, often tacit, held by the human capital. A potential benefit of an acquisition is the knowledge that can be learned from the acquired firm and internalized by the acquiring firm (Vermeulen and Barkema, 2001). However, some of that valuable knowledge can be lost in workforce reductions. The results of this study support these arguments as most of the workforce reductions lead to lower firm performance. Our arguments also have the potential to generalize to unrelated acquisitions, especially if the acquisition is motivated for disciplinary reasons.

Although our study reveals that workforce reduction served as a full mediator of the effect of premium paid for an acquisition on firm performance, caution needs to be exercised in interpreting these results. Our results reveal that excessive workforce reduction can have a significant effect on an organization's performance. However, premium is likely to affect other variables that in turn affect performance. While we reduced the potential for the omitted variable problem by introducing a number of important controls in the analyses, further research is needed to fully understand the potential mediators of the premium–performance relationship because there could be other paths between premium and post-acquisition performance not estimated in our study.

We examined the four outliers eliminated from our earlier analyses to understand their behaviour. There were some major differences between these firms and the remaining 174 firms in the sample. First, the workforce reductions in these four firms were 21, 22, 28 and 30 per cent, respectively, compared to the sample average of 3.5 per cent. Second, the mean revenue of the parent firm at these four firms, at $650 million, was far below the sample mean of $6.35 billion in 174 firms. Hence, even a small reduction in absolute value in the workforce results in an unusually large reduction in percentage terms. Third, these firms implemented large workforce reductions (20 per cent or greater), along with divestitures of other assets, and did not appear to experience any deterioration in their performance, thereby supporting the work of Capron et al. (2001). In these cases, the employees laid off were no longer needed because of the assets divested. Usually, firms cannot implement exceptionally large workforce reductions without closing facilities or divesting other assets. This is so because the operations require adequate staff. Additionally, some firms make acquisitions with the intent of selling off less productive assets. When they do so, the number of employees is automatically reduced as well. Because of the reductions in assets, the firm retains an adequate amount of human capital to use the remaining assets effectively.

A post-hoc multiple regression analysis included the four outliers in the sample. We found a curvilinear (U-shaped) relationship between workforce reduction (WFR) and performance. However, the plot between WFR and performance revealed some interesting insights on these four firms. The relationship between WFR and performance was linear as we had predicted until WFR reached 15 per cent, which encompassed 174 firms in the sample; thereafter, the slope began to increase. Thus, only four firms were driving the U-shaped relationship between WFR and performance. We conclude that the majority of the sample (174 firms) shows a linear relationship between these two variables. The analyses with the four outliers, we believe, do not provide an accurate test of our hypotheses. Hence, we eliminated the four outliers from the final sample.

In this study, we focused on the potential linkages between acquisition premiums, subsequent workforce reductions and post-acquisition performance. Importantly, our study reveals that the reasons for poor organizational performance following the payment of premiums may be partly attributable to excess workforce reductions. We controlled for several factors that may influence workforce reduction such as the debt of the acquiring firm, performance of the acquiring and acquired firm, relative difference in size of the acquiring and acquired firms, the presence of multiple bidders and the relatedness of the acquired firm as explained earlier. As we argued, redundancies arising out of the relatedness between the two firms, the poor performance of the target firm, and high debt prompt firms to engage in workforce reduction to reduce costs. Further, one of the acquisition motives, complementarity, has a negative effect on workforce reduction suggesting a lower likelihood of duplication in the value chain activities of the two firms. Interestingly, the efficiency motive was negatively related to workforce reduction, contrary to what we had predicted. We believe the measure of efficiency to be valid because it was positively correlated with the relatedness between the acquired firm and acquiring firm and because it was negatively related to firm performance. Hitt et al. (2001b) suggested that efficiency gains may be useful but are often insufficient to produce positive returns from acquisitions. Additionally, the negative relationship between an efficiency motive and workforce reductions strongly suggests that the workforce reductions are driven by other motives such as managerial hubris or agency issues. In fact, this finding likely suggests that the managers are trying to avoid losses (behavioural decision theory) due to the high premiums paid. Even after controlling for all these factors, acquisition premiums are associated with larger workforce reductions.

While prior research has shown that acquisition premiums are negatively related to firm performance, it did not show why the effects are negative. The contribution of this study, then, is to suggest one of the reasons for the negative effects of premiums; they often induce larger workforce reductions than are needed to achieve economies of scale. Furthermore, our research suggests the importance of human capital for achieving integration and realizing synergies from acquisitions. Unless other assets are also divested along with the large workforce reductions, the firms are unable to manage continuing operations of the merged firm with the effectiveness needed to achieve heightened performance. Thus, our research helps us to better understand the potential effects of acquisition premiums on subsequent managerial actions and the importance of human capital to achieve integration and synergies in the merged firm.

Our research shares some similarities with Conyon et al.'s (2002) recent work on the employment effects of acquisitions. Both studies suggest that increases in productivity and efficiency following a merger allow for a significant reduction in the number of employees. Both studies examined large companies, used a similar measure of relatedness, and examined the lagged effects of employee reduction in the companies following a merger. Taken together, the results of these two studies provide strong evidence for the motivation behind merger-related employee reduction in both US (our study) and UK based (Conyon et al.) firms. However, there are notable differences between the two studies, as well. While we view acquisition premium as the catalyst for excessive workforce reduction, and limit our study to related acquisitions given its heavy dominance in the USA, Conyon et al. (2002) study employment effects in both related and unrelated acquisitions and show that there is a significant decrease in employment for both types, although as expected in related acquisitions it is more pronounced. Therefore, while both studies are complementary, they also provide independent contributions to our knowledge of workforce reductions following acquisitions.

As with much organizational research, there are some potential limitations in our study. First, we must be concerned with the potential for Type II errors. To control for this concern, we examined the statistical power of our sample and found it to be greater than 0.90. Thus, the probability of a Type II error is quite low (Baron and Kenny, 1986). A second concern is the potential for reverse causality, particularly with firm performance. We controlled for this potential problem by using prior performance as a control variable and by lagging the measures of the independent and dependent variables. Endogeneity bias is also a potential problem with workforce reduction and payment of premiums for acquisitions. To examine the potential for this problem, we used the omitted variable version of the Hausman test (Greene, 2000) with a two-stage least squares regression analysis. The results showed no endogeneity bias present in the results (the complete results of these analyses are available from the first author). Third, there are potential alternative explanations for the results, especially with regard to the effects of different activities involved in the integration of the two firms on the performance of the merged firm. While the effectiveness of many of the integration activities is likely affected by the amount and quality of the remaining human capital, differences in information technology, human resource management policies etc, could also affect the ability to integrate operations and thus firm performance. Fourth, involuntary turnover was of prime importance in this study. Although our data did not allow us to distinguish between involuntary dismissals and voluntary turnover, workforce reduction announcements in an acquisition context largely signal involuntary layoffs. Likewise, our data do not allow us to separate the divestment of assets from employee reduction announcements. As noted by Haynes et al. (2003), these data often are not available from secondary sources as firms rarely disclose the sources of employee reduction when announcing a merger.

Fifth, there are other variables that could potentially impact workforce reduction and post-acquisition performance such as operational relatedness between the two firms, factors in the environment or in the organization not connected to the acquisition, the CEO compensation structure, hostile versus friendly acquisitions, and differences in organizational culture, which were not examined in this study. While potential effects of the industry environment on firm performance were controlled, the relationship between workforce reduction and post-acquisition performance may depend on differences in organizational culture (affecting the ability of the firm to achieve integration) (Haspeslagh and Jemison, 1991). However, the sample was constructed of publicly-held firms, and workforce reduction announcements likely operate as a signalling mechanism to the market. The climate of the acquisition is likely to affect workforce force reduction. While we acknowledge that hostile acquisitions are likely to generate expectations for greater workforce reductions (Conyon et al., 2002), we did not control for this variable because a majority of the acquisitions (90 per cent) undertaken in the USA in the 1990s were friendly in nature. This trend was also evident in our sample which comprised largely of friendly acquisitions (only 14 were classified as hostile). Nevertheless, we conducted a post-hoc analysis incorporating this variable and found its effects to be insignificant.

The issues addressed in this study have important implications for managers. Large workforce reductions are likely to adversely affect the long-term performance of the firm, not only because they harm its resource base; they also negatively affect the morale of surviving employees (Cameron, 1994; Cascio, 1993; Wall Street Journal, 2001). Following a high-premium acquisition, firms need to exercise caution when they undertake workforce reductions because redeployment of resources may create greater benefits than resource divestment, even allowing for the reductions in cost (Capron, 1999; Cascio, 2002; Collis and Montgomery, 1995, 1998). As Jemison and Sitkin (1986) note, premature solutions rarely result in success because the capabilities that were expected in the merged entity may be lost with the laid off employees. Additionally, there are significant limitations to attaining synergies in competitive markets relative to the requirements for recapturing acquisition premiums (Sirower, 1997). Hence, although large workforce reductions may generate short-term benefits, in the long term, they may adversely affect performance, when they go beyond the amount necessary to gain economies of scale (Krishnan and Park, 2002; Lee, 1997). Finally, the reasons that managers pay premiums may also lead them to overestimate the need for layoffs. Hence, unless there are credible growth opportunities, managers should carefully and thoroughly evaluate acquisitions that involve huge premiums.

This research adds to our knowledge of how mergers and acquisitions affect subsequent firm performance. Much of the literature on M&As assumes that they fail often because of poor integration. Indeed, the effectiveness of integration activities likely affects firm performance following acquisitions. Our results suggest that other important factors for post-acquisition performance include the payment of a premium, the size of that premium and the subsequent actions taken by managers during the integration process to achieve returns in excess of the premium paid. As a result, this study provides important implications for scholars and practitioners interested in understanding the determinants of post-acquisition performance.

NOTE

  • [1] 

    Industry growth may also affect the relationship between premium and workforce reduction. Significant premiums are more likely in high growth markets because of the potential positive returns available. Such premiums are unlikely to produce workforce reductions because of the expected positive returns and more personnel needed to service the growth. As a result, the potential effects of industry characteristics were controlled in this study.

Ancillary