HR downsizing

J. Account. Public Policy 36 (2017) 239–257

Contents lists available at ScienceDirect

J. Account. Public Policy

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j a c c p u b p o l

Full length article

Disclosure strategies and investor reactions to downsizing announcements: A legitimacy perspective

http://dx.doi.org/10.1016/j.jaccpubpol.2017.03.003 0278-4254/� 2017 Elsevier Inc. All rights reserved.

⇑ Corresponding author at: Schulich School of Business, York University, 4700 Keele Street, Canada. E-mail addresses: [email protected] (E. Nègre), [email protected] (M.-A. Verdier), [email protected] (C

[email protected] (D.M. Patten).

Emmanuelle Nègre a, Marie-Anne Verdier b, Charles H. Cho c,⇑, Dennis M. Patten d a University of Montpellier, France b University of Toulouse 3 Paul Sabatier, France c Schulich School of Business, York University, Canada d Illinois State University, United States

a r t i c l e i n f o

Article history: Available online 17 April 2017

Keywords: Disclosure strategies Downsizing operations Impression management Legitimacy theory Press releases

a b s t r a c t

In this paper, we focus on a relatively underexplored aspect of sustainability—workforce reductions. We investigate the determinants and consequences of the decisions made by French firms to use press releases in order to announce downsizing operations. We also examine whether the content of press releases has an impact on investor reactions to downsizing announcements. Particularly in the French context, downsizing operations reflect negatively on corporate social responsibility with respect to employees, and we anticipate that French managers will use disclosure strategies to counter a potential legit- imacy threat. Our sample consists of 227 downsizing operations announced between 2007 and 2012 by 119 French listed firms. We find that the disclosure of press releases is driven by both contextual and legitimacy factors. We also find that press releases are associated with more negative reactions to downsizing announcements than when there is no press release, particularly in the case of proactive operations (i.e., implemented by firms with improving performance). A content analysis of press releases indicates that firms, on aver- age, engage in a reactive impression management strategy in their disclosure that consists of attributing downsizing operations to external factors. Moreover, investors penalize the use of proactive arguments, particularly when they are used to justify proactive operations. Overall, our results show that, in the French case, disclosure strategies and their conse- quences on the financial markets relate to a legitimacy perspective.

� 2017 Elsevier Inc. All rights reserved.

1. Introduction

Organizational sustainability includes the economic-financial, environmental, and social aspects of organizations (e.g., Jabbour and Santos, 2008). However, to date, most research in the sustainability domain focuses on the environmental aspect of sustainability as opposed to its social dimension (Sharma and Ruud, 2003). And while recent literature on integrated reporting in sustainability accounting (e.g., Baboukardos and Rimmel, 2016; Melloni et al., forthcoming) does consider dif- ferent CSR dimensions, specific categories are not examined in-depth.

.H. Cho),

240 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

In contrast, in this paper we explore the social issue of workforce reductions. Consistent with Wilkinson (2005), we argue that downsizing operations1 have an adverse impact on sustainability and we note that these operations have been the subject of much controversy. More specifically, several studies note that downsizing operations are perceived by society as a breach of the social contract between organizations and society (e.g., Mäkelä and Näsi, 2010; Van Buren, 2000; Vuontisjärvi, 2013). Accordingly, the issue of workforce reductions falls within the domain of corporate social responsibility (CSR)/sustainability research since the operations have adverse societal consequences, particularly for employees. From an ethical perspective, downsizing operations can also be seen as a questionable business undertaking (Vuontisjärvi, 2013) and are generally imple- mented in order to improve firm efficiency and to maximize value for shareholders. However, for employees and the local com- munity, downsizing operations usually mean significant losses that can be difficult to support (Mäkelä and Näsi, 2010).

Such a breach of the social contract can have severe economic consequences for firms—for instance because of potential strikes or boycotts (e.g., Hunter et al., 2008), which could destroy value for shareholders through lost customers and rev- enues.2 This would be especially true for ‘‘proactive” downsizing operations as they are considered less ethically justifiable (Van Buren, 2000) because they are not justified by apparent financial needs (Love and Kraatz, 2009). Downsizing operations can thus be viewed as negative social events (e.g., Barclay et al., 2005; Flanagan and O’Shaughnessy, 2005; Leana and Feldman, 1988) that create legitimacy threats for organizations. Consequently, companies facing legitimacy threats may use dis- closure strategies to alter perceptions about the legitimacy of the organization (e.g., Beelitz and Merkl-Davies, 2012; Cho, 2009; Cho and Patten, 2007). If this is the case, we would expect to find that, when companies disclose the logic behind downsizing operations, they would be more likely to make them appear more reactive (or less proactive) than they are in reality.

In addition, prior research examines the impact of downsizing announcements on the reaction of financial markets (e.g., Chalos and Chen, 2002; Elayan et al., 1998; Hillier et al., 2007; Lee, 1997) and generally shows a negative investor reaction (e.g., Chen et al., 2001; Elayan et al., 1998; Hillier et al., 2007; Lee, 1997; Lin and Rozeff, 1993; McKnight et al., 2002; Ursel and Armstrong-Stassen, 1995; Worrell et al., 1991). Evidence also suggests that the reason why (proactive or reactive) down- sizing operations occur influences financial market reactions. However, while these prior studies primarily focus on code-law countries and adopt an economic perspective to explain results, we argue that the investigation of downsizing operations in more stakeholder-oriented countries requires consideration of other perspectives such as their potential threat to corporate legitimacy.3

In this study, therefore, we examine the determinants and consequences of the decisions made by French firms to use press releases in order to announce downsizing operations, through a national lens and by adopting a legitimacy perspective. We also examine whether the content of the press releases has an impact on investor reactions to downsizing announce- ments because the analysis of disclosure strategies aligned with such announcements highlights the potentially important role of disclosure in influencing investor response to downsizing announcements. We specifically focus on the French con- text where the need to justify and legitimate downsizing operations through disclosure strategies seems particularly strong. The French legal system protects workers and makes it difficult and costly for firms to dismiss workers (Cascio, 2005), and France is historically known for its strong conflicts between managers (or shareholders) and employees. While cooperation between both parties has been improving to some extent, the country’s protest culture – documented by a high strike rate – is still prevalent and leads to high exposure of downsizing operations in the media.4 Further, and importantly from a legit- imacy perspective, Jung et al. (2015, p. 2064) argue that ‘‘the nature of employment reduction in France is distinct from ‘offen- sive’ layoffs more common in the USA in the last three decades.” They note that in the U.S. firms ‘‘increasingly rely on employee layoffs [. . .] to improve financial performance in the context of increased pressures from shareholder value-driven institutional investors,” while in France, poor performance is the key driver to employment reduction (Jung et al., 2015, p. 2062–2063). Accordingly, proactive operations in France are often viewed as an ‘‘injustice” creating political tensions and generating a need for communication. France is thus a unique setting to examine downsizing operations as these events constitute significant threats to firms. Examining 227 downsizing operations implemented between 2007 and 2012 by 119 French listed companies, we find that the disclosure of press releases to announce downsizing operations is driven by both contextual and legitimacy factors. With respect to the former, we find that press releases are more likely for firms that implement a downsizing for the first time during the period studied, and less likely when downsizing operations are implemented through layoffs rather than voluntary measures (e.g., early retirements, voluntary redundancy plans). With respect to legitimacy-related factors, we

1 We use the term ‘‘downsizing operations” and ‘‘workforce reduction” interchangeably in the paper as we consider them synonymous. As we note in the methods section below, our sample of events includes downsizing operations related to both layoffs and voluntary reduction measures (such as voluntary retirement). We control for differences in these types of reductions in the analysis.

2 An illustration is the 2001 consumer boycott on Danone, a French multinational food and beverage firm. While it was generating high profits, Danone announced a cut of 3000 jobs in Europe, including 1700 in France. This operation was seen as a ‘‘public outrage” (Hunter et al., 2008, p. 338) even by some politicians who ordered hospitals and schools to stop buying Danone products. The Danone logo was modified to include the slogan ‘‘Human beings are not yogurts”. Despite Danone denying the impact of the boycott and strikes on firm performance, a serious decline in the company’s sales and market capitalization has been reported by financial analysts and the media.

3 In the legitimacy perspective, the target audience for disclosure is wider than in the economic perspective and includes several stakeholders (e.g., customers, employees, etc.).

4 For example, on October 5th, 2015, Air France-KLM’s angry employees interrupted a meeting in which managers and employee union representatives were discussing a new large downsizing operation. This operation caused both violent protests and actions against Air France-KLM’s CEO and Human Resources Director and this incident was subject to much international media exposure. French President François Hollande denounced this violence as ‘‘unacceptable and bad for France’s image”. He added that ‘‘social dialogue is important, and when it is interrupted by violence and disputes take on an unacceptable form, it can have consequences for the image and attractiveness [of the country]”.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 241

find that the probability of the use of press releases increases for firms that belong to socially sensitive industries, and decreases when firms implement downsizing operations during a financial crisis period. We also report a negative association between downsizing announcements via press releases and market reaction. With respect to proactive operations, we find that despite an overall positive (but not significant) market reaction to proactive operations, the reaction is, on average, negative when firms announce such operations through press releases. Finally, content analysis of the 66 press releases issued by sample firms to announce downsizing operations shows that numerous French firms appear to adopt a reactive impression management strat- egy, in that the language of the press releases make the downsizing operations appear more reactive (or less proactive) than they are in reality. This suggests that firms consider legitimacy concerns when adopting disclosure strategies. We also find that the more firms use proactive arguments to justify downsizing operations, the more investors in the French market react neg- atively to such operations. This result contrasts with the one of prior studies conducted in common-law countries that docu- ment a positive market reaction to downsizing announcements if the reasons stated within such announcements are proactive (e.g., Abraham, 2004; Elayan et al., 1998). In addition, the use of proactive arguments in press releases disclosed to announce proactive operations is negatively related to market reaction. Therefore, investors do not penalize the use of impres- sion management. On the contrary, they react negatively to disclosures that could increase a potential legitimacy threat. Overall, our results show that in the French case, disclosure strategies and their consequences on the financial markets relate more to a legitimacy perspective.

Our research extends the empirical literature on both downsizing and sustainability in four different ways. First, most prior research defines proactive and reactive operations according to the reasons stated within downsizing announcements (e.g., Abraham, 2004; Elayan et al., 1998; Lee, 1997; Worrell et al., 1991). However, we provide evidence of impression man- agement strategies in such announcements; that is, firms appear to justify downsizing operations by using reactive argu- ments—they establish a link between the decision to downsize and external events (e.g., bad market or sector conditions) and internal financial difficulties potentially in order to offset any legitimacy threats. Second, we take into account how downsizing operations are brought to the public’s attention (i.e., by the media or directly by the firms through press releases). By issuing press releases, firms keep control of their communication and thus could manipulate a particular mes- sage and make it easier to accept. However, our results suggest that disclosures made via these releases are more likely to damage firm organizational legitimacy. This stands in contrast to Griffin and Sun’s (2013) findings of positive market reac- tions to U.S. company press releases related to carbon emissions. Third, we examine the market reaction to downsizing announcements in a code-law country such as France in which legitimacy considerations are strong and very relevant regarding labor issues (Harris et al., 1994; Mora and Sabater, 2008). In contrast, most prior research has been conducted in common-law countries and finds that when managers mention proactive arguments to justify downsizing operations, investors react positively or less negatively than when reactive arguments are given (e.g., Abraham, 2004; Elayan et al., 1998; Gunderson et al., 1997; Hahn and Reyes, 2004; McKnight et al., 2002). Our results show that, in France the market perceives badly the use of proactive arguments presumably due to concerns with potentially damaged legitimacy. Finally, despite the large number of studies on sustainability issues to date, little attention has been paid to the social dimension of CSR activities and their relation to the concept of sustainability (Kent and Zunker, 2013; Mäkelä and Näsi, 2010). By inves- tigating the disclosure strategies and market reactions to downsizing announcements from a legitimacy perspective, this study helps to address this gap in sustainability-related research.

The remainder of this paper is organized as follows. Section 2 presents the theoretical framework of the study and devel- ops the hypotheses. Section 3 describes the variables and the sample. Sections 4 and 5 present the empirical findings. Sec- tion 6 discusses the main results and concludes.

2. Theoretical framework, literature review and hypotheses development

2.1. Legitimacy and the social contract

As noted by Gray et al. (1995), a large body of social and environmental accounting research is grounded in legitimacy theory. Dowling and Pfeffer (1975, p. 122) define organizational legitimacy as the ‘‘congruence between the social values asso- ciated with or implied by their activities and the norms of acceptable behavior in the larger social system of which they are a part”. The concept of a social contract established between organizations and society is central to legitimacy theory (e.g., Cho, 2009; Deegan, 2002; Deegan and Blomquist, 2006; Hooghiemstra, 2000; Patten, 1992). Shocker and Sethi (1973) argue that all organizations are linked to society by a social contract, and it can be viewed as an ‘‘ethical floor below which firms cannot fall and still be considered socially legitimate” (Van Buren, 2000, p. 210). Therefore, organizational legitimacy and social con- tract compliance go hand in hand (Deegan et al., 2000), and a breach of the contract may lead to a perception by society that the organization is not legitimate.

According to Suchman (1995), prior literature defines organizational legitimacy from two different perspectives. First, from a strategic perspective, legitimacy is considered to be a resource on which an organization is dependent for survival (Dowling and Pfeffer, 1975; Deegan, 2002), and legitimacy can also be influenced or manipulated by organizations to gain societal support. Second, from an institutional perspective, society generates ‘‘cultural pressures that transcend any single organization’s purposive control” (Suchman, 1995, p. 572) and thus managers’ practices are constructed by external institu- tions. Suchman (1995) concludes that both strategic and institutional perspectives need to be considered in order to

242 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

understand the complexity of organizations. Organizational legitimacy is thus the result of both managers’ active strategies and managers’ passive responses to external pressures.

Prior research suggests that firms facing legitimacy threats engage in specific disclosure strategies to alter perceptions about the legitimacy of the organization (e.g., Beelitz and Merkl-Davies, 2012; Cho, 2009; Cho and Patten, 2007). Legitimacy theory has been extensively used in the previous literature to explain firms’ disclosure strategies in non-routine situations (Merkl-Davies and Brennan, 2007) such as (1) corporate scandals (e.g., Breton and Côté, 2006; Elsbach, 1994), (2) environ- mental disasters (e.g., Beelitz and Merkl-Davies, 2012; Cho, 2009; Deegan et al., 2000; Hooghiemstra, 2000; Patten, 1992), and (3) restructuring firms (e.g., Arndt and Bigelow, 2000; Mäkelä and Näsi, 2010; Ogden and Clarke, 2005). Disclosures are regarded as responses to both public pressure and increased media attention (Hooghiemstra, 2000), which are particu- larly strong in the context of downsizing operations (Henderson et al., 2010).

2.2. Downsizing operations

Proactive downsizing operations aim to improve the efficiency and profitability of firms (Freeman and Cameron, 1993; Pouder et al., 2004; Sheaffer et al., 2009) and/or to maintain competitiveness (Lee, 1997), and they are generally imple- mented without apparent financial needs (Love and Kraatz, 2009). In contrast, reactive operations are carried out because of poor or declining firm performance (Lee, 1997; Worrell et al., 1991). Several studies (e.g., Abraham, 2004; Chen et al., 2001; Elayan et al., 1998; Hillier et al., 2007; Lee, 1997; McKnight et al., 2002; Worrell et al., 1991) document that the market reaction is more negative when downsizing operations are reactive compared to when they are proactive, while others report a positive reaction to proactive operations (Abraham, 2004; Elayan et al., 1998; Gunderson et al., 1997; Hahn and Reyes, 2004; McKnight et al., 2002). From an economic perspective, it is argued that proactive operations lead to a positive market reaction because they convey good news to investors about future performance, whereas reactive operations high- light the severity of the financial distress, and thus convey bad news to investors. In order to lessen the impact of bad news, managers can use disclosure strategies when they provide the logic behind reactive operations. According to Lee (1997), management could make downsizing operations appear proactive rather than reactive in their disclosures (i.e., proactive impression management strategy), leading to a more favorable market reaction to downsizing announcements. Lee’s implicit assumption is that investors are the main target audience for disclosure.

We argue that legitimacy considerations should be more prevalent in code-law countries compared to common-law countries due to the strong social differences between these two systems of governance. For example, both Ball et al. (2000) and Simnett et al. (2009) note that code-law countries such as France, Italy, and Spain, are characterized by a stake- holder governance model in which accounting earnings are presumed to be divided among shareholders, governments, man- agers, employees, and other stakeholders. Therefore, when firms announce downsizing operations, employees can perceive the operations as a way for managers to transfer wealth from employees to firms, leading to a breach of the social contract between firms and society. In contrast, labor relations in common-law countries are characterized by ‘‘more flexible hire and fire arrangement” (Munoz-Bullon and Sanchez-Bueno, 2011, p. 2925). Further, France, in particular, relies on a system of code law where the power of unions is stronger than in common law countries (Ball et al., 2000; Garcia Lara et al., 2005; Harris et al., 1994). When the power of unions is high, the issue of the distribution of wealth is more likely to lead to social protests in the form of strikes or boycotts, potentially increasing the costs associated with such operations.

2.2.1. The decision to use press releases: determinants and stock market reaction Beelitz and Merkl-Davies (2012) argue that depending on whether there is congruence or incongruence between manage-

ment’s and organizational audiences’ interpretations of the downsizing, the managerial decision to reduce the workforce is either accepted or rejected. Rejection of the managerial decision could create legitimacy threats for organizations. Therefore, managers have incentives to resolve the conflicting interpretations of the event by using disclosure. In the context of down- sizing operations, one way to preserve organizational legitimacy is to use press releases to announce and justify these oper- ations. However, we assume that the probability of conflicting interpretations of the event is different according to the reason (proactive or reactive) for the operation. The firm’s decision to downsize is more likely to lead to conflicting interpre- tations in the case of proactive operations. Indeed, managers and shareholders may perceive proactive operations positively as they convey a positive signal on future performance. In contrast, particularly in the French context, other stakeholders including employees and the public may perceive these operations negatively as they appear unnecessary regarding the firm’s current performance. Therefore, we assume that for French companies, the probability of disclosing a press release is higher for proactive rather than reactive operations. Further, because the use of press releases could enable firms to resolve the potential conflicting interpretations about the downsizing operations, we expect market reactions to downsizing oper- ations to be less negative when companies proactively announce the workforce reductions via press releases.

We formally state our first two hypotheses as:

H1: The use of press releases by French firms to announce downsizing operations is more likely for proactive than reactive operations.

H2: The market reaction to downsizing operations by French firms will be less negative when press releases are used to announce workforce reductions.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 243

2.2.2. Content of the press releases and stock market reaction As noted above, prior studies of investor reactions to downsizing operations assume, from an economic perspec-

tive, that proactive operations will be perceived more positively. Based on this, Lee (1997), for example, argues that when managers provide the logic behind the downsizing operations, they have an incentive to use a proactive impression management strategy to make downsizing operations appear more proactive (or less reactive) than they are in reality.

However, in the French context, such an argument may not be valid. In France, downsizing operations are far more likely to be seen as threatening events because they potentially involve a violation of values (Ashforth and Gibbs, 1990). Further, the risk of legitimacy loss is higher in the case of proactive operations because these operations are not justified by apparent financial need (Love and Kraatz, 2009), and thus, they are less ethically justifiable (Van Buren, 2000). Therefore, from a legitimacy perspective, French firms have an incentive to highlight the role of external factors such as demand slump, economic crisis or bad market conditions that are beyond the control of managers (e.g., Brockner, 1990; Stevens and Kristof, 1995) when attempting to justify the downsizing operation. In other words, we expect French firm managers to use a reactive impression management strategy to make downsizing operations appear to be more reactive (or less proactive) in their press releases than they are in reality. Finally, we expect the type of impres- sion management used in the press releases to influence market reactions to the announcements. In contrast to the tra- ditional belief that investors will react more positively to the use of proactive arguments to justify workforce reductions, in the French context, we expect the opposite due to the increased potential for such claims to lead to legitimacy concerns. If investors believe the cost of lost legitimacy outweighs the economic benefit of the downsizing operations, the announcements will lead to reduced market value (Groening and Kanuri, 2013), and we anticipate reactions to be even more negative where proactive arguments are used to justify proactive operations as this reinforces the potential legiti- macy threat.

We formally state these hypotheses as follows:

H3: French firms are more likely to use reactive rather than proactive arguments in press releases announcing downsizing operations.

H4: The market will react negatively to French firms’ use of proactive arguments to justify downsizing operations.

H5: The market reaction will be more negative when proactive arguments are used by French firms to justify proactive oper- ations compared to reactive operations.

3. Research method

3.1. Dependent variables

In this study, we first investigate the determinants of firms’ decisions to issue press releases to announce downsizing operations. Second, we explore the relation between press release disclosure and the stock market reaction to downsizing announcements. We obtained press releases from the Factiva database or from companies’ websites, and in the first equation (Eq. (1)), the dependent variable (DISCPR) equals one when the firm discloses a press release to announce the downsizing and zero otherwise. In the second equation (Eq. (2)), the dependent variable (CARi) is the cumulative abnormal return for firm i over the event period from day 0 until day 2. The event date (t = 0) is the date of the downsizing announcement (by firms themselves through press releases or by the media). We estimate market model parameters using ordinary least squares regression with an estimation period ranging from days �120 to �20 prior to the event date. We consider the CAC All-Tradable index as our proxy for the market portfolio, and we obtain all financial data (stock prices and market index data) from the Datastream database. Appendix A provides additional details on the event study methodology used in this study.

3.2. Independent and control variables

3.2.1. Determinants of the use of press releases To the best of our knowledge, no studies to date focus on the determinants of voluntary disclosure in the social context of

downsizing operations. Therefore, we rely on both CSR disclosure and downsizing literature to identify factors potentially explaining the disclosure decision in this context. We introduce two sets of determinants—contextual and legitimacy factors.

The first contextual factor is the reason for the operation (H1). The empirical literature often defines proactive and reac- tive operations according to the reasons stated within the downsizing announcements (e.g., Capelle-Blancard and Tatu, 2012; McKnight et al., 2002). However, we argue that if managers are using impression management in these announce- ments, the underlying cause of the layoffs may be obscured. Accordingly, we follow Cascio et al. (1997), Love and Kraatz (2009), and Sheaffer et al. (2009) and instead use performance indicators to classify proactive and reactive operations. Sheaffer et al. (2009, p. 260) argue return on assets (ROA) is a ‘‘key financial indicator indicative of firms’ pre-downsizing financial situation” (p. 960). Consistent with Cascio et al. (1997), we compute ROA as earnings before interest, taxes, depre- ciation, and amortization (EBITDA) divided by total assets, and rely on this measure to identify the underlying reason for the

244 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

downsizing operations we explore.5 More precisely, we define operations implemented by firms as proactive (reactive) follow- ing an increase (a decrease) in ROA for the previous year. We use a dummy variable (REASON) coded one for proactive oper- ations and zero otherwise to measure the impacts of the reason for the downsizing on disclosure.

Also related to contextual factors, we introduce a dummy variable (FIRST) taking the value of one for firms implement- ing a downsizing operation for the first time over the period studied and zero otherwise. Lee (1997) argues a first-time downsizing operation usually conveys more new information to stakeholders than subsequent actions, potentially increas- ing the probability of the use of a press release to explain the operation to the public. Further, we include the size of the operation (DOWNSIZE) measured as the percentage of employees laid off as a contextual factor (see, e.g., Nixon et al., 2004; Ursel and Armstrong-Stassen, 1995). Large operations, presumably due to greater visibility, should exhibit a higher probability of disclosure. Finally, we introduce a dummy variable (LAYOFF) that equals one for downsizing operations car- ried out through layoffs and zero for operations implemented through voluntary measures (e.g., early retirements, volun- tary redundancy plans). The unions’ demands and specific actions are likely to be stronger in layoffs than in other types of downsizing operations as layoffs may be perceived as less fair for employees (Greenhalgh et al., 1988), increasing the need for disclosure.

The use of press releases to announce downsizing operations can also result from legitimacy-related factors. Related to these, we first include a dummy variable (SENSIND) that serves as a proxy for socially sensitive industries such as oil and gas, basic materials, aerospace and defense, utilities, and health. Sensitive industries are industries associated with social externalities such as alcohol and tobacco, guns and defense, natural resources, and health (Brammer and Millington, 2005; Hong and Kostovetsky, 2012) or industries that caused numerous social issues in the past. Legitimacy theory suggests sensitive industries are more likely to face legitimacy issues as they are closely monitored by non-governmental organiza- tions, environmental lobby groups (Deegan and Gordon, 1996) and politicians (e.g., Cowen et al., 1987), and accordingly we expect companies within these industries to use disclosure to address these exposures. We next introduce a mimicry factor (MIMETPR) that identifies the number of press releases announcing downsizing operations issued by firms within the same industry for the previous year. Lee (1997) notes that firms within a given industry may feel pressure to mimetically follow the trend of layoffs implemented by competing firms. Further, we include a variable related to labor pressures (LABOR) given that they can increase the need for legitimation and thus the probability of disclosure. Aligned with Depoers (2000) and Missonier-Piera (2004), we use the ratio of labor charges on net sales.6 Finally, we expect the need for legitimation of the downsizing is reduced in times of financial crisis, which could decrease the probability of disclosure. We study the downsizing announcements over the period 2007–2012 and this timeframe includes the financial crisis of 2008–2009, which was charac- terized by negative Gross Domestic Product (GDP) growth rates. In order to take into account the differing economic context, we use a dummy variable (CRISIS) that equals one for downsizing announcements in 2008 and 2009 and zero otherwise.

We control for ownership structure by including a variable related to the percentage of shares held by foreign investors (FORINV). Several studies (e.g., Cormier and Magnan, 2003; Hannifa and Cooke, 2005) show that social and environmental disclosures are positively related to foreign investors. Next, we include a variable (FRSALES) that measures the percentage of net sales realized in France as an increase in this percentage should reinforce negative consequences for firms of potential boycotts. Finally, we control for the firm’s size (FIRMSIZE) as measured by the logarithm of the total assets, and the book-to-market ratio (BM).

3.2.2. Stock market reaction to the use of press releases The test variable related to H2, DISCPR, equals one when the firm issues a press release to announce the downsizing and

zero otherwise (Eq. (2)). We also take into account other factors identified in the prior literature as determinants of market reaction to downsizing announcements (REASON, FIRST, DOWNSIZE, LAYOFF, CRISIS, MIMET). First, prior literature shows that the market reacts more negatively to reactive than proactive operations (e.g., Abraham, 2004; Chen et al., 2001; Elayan et al., 1998; Hillier et al., 2007; Lee, 1997; McKnight et al., 2002; Worrell et al., 1991) because such operations convey a negative signal on performance. Second, previous studies document that the market reacts more negatively to the first downsizing than to subsequent downsizing operations (e.g., Capelle-Blancard and Tatu, 2012; Ursel and Armstrong-Stassen, 1995). Third, the market reaction is more negative in large operations (i.e., when a large percentage of total employees is affected by the downsizing operation) than in small operations (Elayan et al., 1998; Hillier et al., 2007; Lee, 1997; Nixon et al., 2004; Pouder et al., 2004; Ursel and Armstrong-Stassen, 1995; Worrell et al., 1991). Next, McKnight et al. (2002) find that the market reacts more negatively to layoffs than to downsizing operations carried out through voluntary measures. Indeed, layoffs could reflect a strong need to cut costs and be perceived as ‘‘emergency mea- sures” (McKnight et al., 2002, p. 89). Further, Capelle-Blancard and Tatu (2012) find that the market reaction to downsizing announcements is less negative in times of financial crisis than in non-crisis periods, potentially because these operations are less surprising in crisis periods. We add a variable (MIMET) that represents the number of downsizing operations

5 Love and Kraatz (2009) and Sheaffer et al. (2009) also use ROA measures to capture changes in firm profitability. In non-tabulated sensitivity tests, we alternatively identified the underlying reason for the downsizing operations using what Love and Nohria (2005) refer to as the return on market-valued assets. Calculated as EBIT divided by the market value of equity, Love and Nohria argue this metric captures performance relative to current and future performance expectations. Results using this alternative measure for classification of operations are qualitatively unchanged from those we report in the paper.

6 We recognize that the level of unionization can be a good indicator of the labor force power. However, this data is not available for French firms.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 245

announced by firms within the same industry for the previous year. We expect the market reaction to be less negative when downsizing operation announcements are common within the given industry (Lee, 1997).

We control for sensitive industries as investors should react more negatively to downsizing operations announced by firms in such industries due to the strong risk of legitimacy damage. Next, the percentage of shares held by foreign investors can be considered as a proxy for the market discipline imposed by active investors who constantly demand that firms imple- ment efficient operations to improve their performance (Nishitani and Kokubu, 2012). These investors should thus perceive downsizing operations more positively than local investors. We also take into account firm size because it affects company visibility and public scrutiny (Aerts and Cormier, 2009) which could in turn impact the market reaction. Finally, we include the book-to-market ratio to control for the impact of the firm’s growth prospects on the market reaction.

We analyze newspaper articles and press releases available on the Factiva database to identify the characteristics of downsizing operations (e.g., LAYOFF, DOWNSIZE). We use the Infinancials7 database to obtain data for ROA, total assets, indus- try, and book-to-market ratio. The other variables are collected manually from annual reports for the year before the announcement.

3.3. Models

The explanatory model used to identify the determinants of the disclosure of a press release to announce a downsizing is stated as:

7 See 8 We

Log½PrðDISCPR ¼ 1Þ=ð1 � PrðDISCPR ¼ 1Þ� ¼ b0 þ b1 REASON þ b2 FIRST þ b3 DOWNSIZE þ b4 LAYOFF þ b5 SENSIND þ b6 MIMETPR þ b7 LABOR þ b8 CRISIS þ b9 FORINV þ b10 FRSALES þ b11 FIRMSIZE þ b12 BM þ ei ð1Þ

With DISCPR = disclosure of a press release; REASON = proactive/reactive operations; FIRST = first downsizing imple- mented by the firm over the period studied; DOWNSIZE = size of the downsizing; LAYOFF = downsizing operations imple- mented through layoffs; SENSIND = sensitive industries; MIMETPR = number of press releases disclosed by firms of the same industry to announce downsizing operations; LABOR = labor pressures; CRISIS = economic crisis/expansion periods; FORINV = foreign investors; FRSALES = net sales realized in France; FIRMSIZE = firm size; BM = book-to-market ratio; b0 = intercept; b1. . .12 = regression coefficients; ei = residual term.

In addition, we estimate the following model to analyze the relation between press release disclosure and the stock mar- ket reaction to downsizing announcements:8

CARi ¼ b0 þ b1 DISCPR þ b2 REASON þ b3 FIRST þ b4 DOWNSIZE þ b5 LAYOFF þ b6 SENSIND þ b7 MIMET þ b8 CRISIS þ b9 FORINV þ b9 FIRMSIZE þ b10 BM þ ei ð2Þ

With CARi = cumulative abnormal return for firm i over the event period from day 0 until day 2; DISCPR = disclosure of a press release; REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; LAYOFF = downsizing operations implemented through layoffs; SENSIND = sensitive industries; MIMET = number of downsizing operations announced by firms within the same industry; CRISIS = economic crisis/expansion periods; FORINV = foreign investors; FIRMSIZE = firm size; BM = book-to-market ratio; b0 = intercept; b1. . .10 = regression coefficients; ei = residual term.

Table 1 summarizes the independent and control variables.

3.4. Sample selection and descriptive statistics

We initially considered all the downsizing operations announced by French listed firms during the period from 2007 to 2012. Using the Factiva database, we identified all newspaper articles related to French listed firms containing the words layoff, job cut, downsizing, or redundancy plan. We used both singular and plural forms of the words. Our search revealed a total of 318 downsizing operations. We then excluded from our sample (1) financial and insurance firms (8 firms), (2) all observations with missing values for any of the variables used in the empirical analysis (5 firms) and (3) outlier values (2 firms). Finally, we required, for each firm, that the first downsizing in the period studied be preceded by 120 days in which no downsizing operations were announced. This criterion resulted in the exclusion of an additional 76 downsizing opera- tions. Our final sample thus consists of 227 downsizing operations implemented by 119 French listed firms during the period from 2007 to 2012. Table 2 presents the sample selection procedure.

Table 3 presents descriptive statistics. As identified in the table, 30% of the downsizing operations are announced by firms through press releases, while 70% are announced by the media. Moreover, in our sample, the majority of downsizing oper- ations was reactive (60%), implemented through layoffs (55.1%), and was not the first such event implemented by firms over the period studied (55%). Further, 25% of the downsizing operations were announced by firms that belong to a sensitive

http://www.infinancials.com/ for addition information. conduct Hausman tests to check for any endogeneity issues and they indicate none. As such, we conduct OLS regressions.

Table 1 Definition of the independent and control variables.

Variables Definitions Eq. (1)

Eq. (2)

Disclosure of a press release (DISCPR)a

1 if the firm discloses a press release to announce the downsizing; 0 otherwise X

Reason for the downsizing (REASON)

1 if the downsizing is proactive (i.e., implemented by a firm with an increase in the ROA for the previous year); 0 otherwise

X X

First announcement (FIRST) 1 if the firm implements a downsizing for the first time during the period studied; 0 otherwise X X Size of the downsizing

(DOWNSIZE) % of employees laid off X X

Type of downsizing (LAYOFF) 1 if the downsizing is implemented through layoffs; 0 otherwise X X Sensitive industries (SENSIND) 1 if the firm belongs to a sensitive industry (oil and gas, basic materials, health care, utilities); 0

otherwise X X

Mimicry (MIMETPR) Number of press releases announcing downsizing operations issued by firms within the same industry for the previous year

X

Mimicry (MIMET) Number of downsizing operations announced by firms within the same industry for the previous year

X

Labor pressures (LABOR) Labor charges / net sales X Crisis period (CRISIS) 1 for downsizing announcements in 2008 and 2009 (i.e., years with negative GDP growth rates); 0

otherwise X X

Foreign investors (FORINV) % of shares held by foreign investors X X French net sales (FRSALES) % of net sales realized in France X Firm size (FIRMSIZE) Logarithm of the total assets X X Book-to-Market (BM) Book value of firms / market value of firms X X

a The variable DISCPR is an independent variable in Eq. (2) but a dependent variable in Eq. (1).

Table 2 Sample selection procedure.

Number of downsizing operations

All the downsizing operations implemented by French listed firms during the period from 2007 to 2012 318

- financial and insurance firms - missing values - outlier values - operations in the estimation period

- 8 - 5 - 2 - 76

Final sample 227

246 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

industry. The percentage of employees laid off announced is, on average, 3.39%, and 50% of the downsizing operations occurred during times of financial crisis. Labor charges represent more than one third of the total net sales (35%). Foreign investors held on average 13% of the firm’s shares and the average percentage of firms’ net sales realized in France is 48%. Finally, the average book-to-market is 68.54%.

4. Results

4.1. Determinants of the use of press releases to announce downsizing operations

We first compare the reason for the operation between disclosing and non-disclosing companies and Table 4 identifies the frequency distributions and the result of the Pearson’s Chi-Squared test (with asymptotic significance between brackets). The results reveal no significant difference in the use of press releases across proactive and reactive downsizing operations. This is not in line with our first hypothesis (H1).

We next investigate the determinants of firms’ decisions to use press releases to announce downsizing operations using binary logistic regressions. Because two of our intended explanatory variables are highly correlated (see Appendix B),9 we estimate models separately including FIRST and FIRMSIZE. The results of our estimations are reported in Table 5. As identified in the table, the models, based on Chi-Squared test statistics, are statistically valid, and the explanatory power (indicated by

9 The correlation matrix reveals that there is one correlation coefficient higher than 0.4. We run two different regression models to ensure that correlated independent variables (FIRST and FIRMSIZE) are regressed in separate models. We define (i) two specifications of Eq. (1), named models 1 and 2; (ii) two specifications of Eq. (2), named models 3 and 4.

Table 3 Descriptive statistics.

Variables Frequency Mean (Standard deviation)

DISCPR Press release 30.0% No press release 70.0%

REASON Proactive operations 40.0% Reactive operations 60.0%

FIRST First downsizing 45.0% Multiple downsizing operations 55.0%

DOWNSIZE % of employees laid off announced 0.0339 (0.071)

LAYOFF Layoffs 55.1% Voluntary measures 44.9%

SENSIND Sensitive industries 25.0% Other industries 75.0%

MIMETPR Number of press releases announcing downsizing operations issued by firms within the same industry for the previous year

2.2100 (2.406)

MIMET Number of downsizing operations announced by firms of the same industry in the previous year 10.2400 (8.743)

LABOR Labor charges / net sales 0.3546 (1.215)

CRISIS Downsizing operations in crisis periods 50.0% Downsizing operations in expansion periods 50.0%

FORINV % of shares held by foreign investors 0.1320 (0.185)

FIRMSIZE Logarithm of the total assets 15.1400 (2.415)

FRSALES % of net sales realized in France 0.4818 (0.298)

BM Book-to-market ratio 0.6854 (0.611)

Table 4 Difference in reason for the operation between disclosing and non-disclosing firms.

DISCPR = 1 DISCPR = 0

REASON Proactive 29.67% 70.33% Reactive 29.41% 70.59%

Pearson Chi-Squared: 0.002 (0.541)

DISCPR = disclosure of a press release; REASON = proactive/reactive operations.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 247

Nagelkerke-R2) for the two models is 21.50% and 31.40%, respectively. Model 1 correctly classifies 75.4% of the observations while Model 2 correctly classifies 77.7% of the cases. Each of these is higher than the 70.2% classification rate for a naïve model (i.e., all firms decide not to disclose a press release).

Contrary to H1, the variable REASON is not statistically significant. One potential explanation is that disclosure related to reactive operations can address potential concerns from investors about the firms’ financial situation. Results also reveal that the disclosure of a press release is more likely for firms that implement a downsizing for the first time during the period studied (model 1: p < 0.01). The probability of the disclosure of press releases decreases when downsizing operations are implemented through layoffs rather than voluntary measures (model 1: p < 0.05; model 2: p < 0.01). Disclosure is a means for firms to protect their legitimacy by showing that they do their best to prevent negative consequences of downsizing oper- ations for employees. Conversely, the probability of disclosing a press release is higher for firms that belong to socially sen- sitive industries than for firms in other industries (model 1: p < 0.1; model 2: p < 0.05). This result is in accordance with the previous literature suggesting that firms in sensitive industries are more likely to face legitimacy issues than firms in other industries because their activities are likely to have impact upon society, which result in more voluntary disclosures (e.g., Patten, 1992; Adams et al., 1998). In contrast, the probability of disclosure decreases when the operation is announced in a crisis compared to an expansion period (models 1 and 2: p < 0.01). One potential reason for this latter finding is that firms’ financial difficulties increase in crisis periods and thus the need for legitimation is lower than in economic expansion periods.

Table 5 Determinants of the disclosure of press releases to announce downsizing operations.

Variables Model 1 Model 2

Coefficients Wald (p-value) Coefficients Wald (p-value)

REASON 0.123 0.128 (0.720) �0.007 0.000 (0.984) FIRST 1.132 9.429 (0.002)***

DOWNSIZE 2.161 0.733 (0.392) �1.217 0.200 (0.655) LAYOFF �0.682 3.840 (0.050)** �1.249 9.686 (0.002)***

SENSIND 0.634 2.996 (0.083)* 0.814 4.442 (0.035)**

MIMETPR 0.024 0.115 (0.734) �0.055 0.489 (0.484) LABOR 0.540 0.654 (0.419) 0.183 0.340 (0.560) CRISIS �0.921 6.723 (0.010)*** �1.092 8.460 (0.004)*** FORINV 2.273 6.886 (0.009)*** 2.983 9.595 (0.002)***

FRSALES 0.334 0.302 (0.583) �0.455 0.456 (0.499) FIRMSIZE �0.517 23.831(0.000)*** BM �0.325 1.139 (0.286) �0.224 0.515 (0.473) Intercept �1.438 8.103 (0.004)*** 7.815 17.176 (0.000)*** Nagelkerke R2 0.215 0.314 Chi-Squared (p-value) 36.509 55.535

(0.000) (0.000) Classification rate (overall%) 75.4% 77.7%

REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; LAYOFF = downsizing operations made through layoffs; SENSIND = sensitive industries; MIMETPR = number of press releases disclosed by firms of the same industry to announce downsizing operations; LABOR = labor charges/turnover; CRISIS = economic crisis/expansion periods; FORINV = foreign investors; FRSALES = net sales realized in France; FIRMSIZE = firm size; BM = book-to-market ratio. *** Indicate significance at the 1% levels, respectively. ** Indicate significance at the 5% levels, respectively. * Indicate significance at the 10% levels, respectively.

248 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

Results also indicate that the variables DOWNSIZE, MIMETPR, LABOR are not statistically significant. These findings sug- gest that the probability of disclosing a press release to announce a downsizing is not explained by the will of the firm to mimic the disclosure practices of other firms. Similarly, the visibility of the operation proxied by its size does not influence the firm’s decision to disclose a press release to announce a downsizing, as well as labor pressures.

With regards to control variables, we find a positive relation between foreign investors and the probability of issuing a press release (models 1 and 2: p < 0.01). We also find that the disclosure of a press release is negatively related to firm size (model 2: p < 0.01). One possible explanation for this result is the high correlation between the variables FIRST and FIRMSIZE, which means that the probability of implementing a downsizing for the first time during the period studied is higher for small firms than for large firms (and the use of press releases is due to the first-time nature of the event). Alternatively, because information asymmetry is generally greater for small firms than for large firms, small firms receive less media atten- tion and analyst coverage than large firms (Lang and Lundholm, 2000). Accordingly, they may be more likely to use discre- tionary disclosure to influence stakeholder perceptions.

In order to provide an economic interpretation of the results, we compute marginal effects (untabulated results) deter- mined as the means of the independent variables. We find that all the statistically significant variables are also economically significant. A first downsizing implementation increases the probability of disclosure by 22.65% (p < 0.01), and membership in a socially sensitive industry increases the probability of disclosure by 13.28% (p < 0.1). In contrast, downsizing operations implemented through layoffs decrease the probability of disclosure of 13.56% (p < 0.1) and events occurring during a crisis period decrease the probability of disclosure by 18.02% (p < 0.01). Finally, with respect to foreign investors and firm size, at the mean level, an increase of one standard deviation respectively increases the probability of disclosure by 44.79% (p < 0.01) and decreases the probability of disclosure by 6.9% (p < 0.01).

4.2. Effect of the use of press releases on the stock market reaction

In the second stage of our analysis we focus on the market reaction to the downsizing announcements, and overall, we find a negative, but not significant, average market reaction of �0.15% over the three-day event window. In contrast to our expectations, univariate analysis, presented in Panel A of Table 6, indicates that the average market reaction is lower when firms issue press releases to announce downsizing operations than when there is no press release. Indeed, for firms issuing a press release, the market reaction is negative (�0.89%) and statistically significant (p < 0.01), whereas the market reaction is positive (0.16%), but not significant, when there is no press release. One possible explanation for this result is that the dis- closure of press releases to announce downsizing operations increases the visibility of such operations. Given that downsiz- ing operations are generally perceived negatively by the market (e.g., Worrell et al., 1991; Lin and Rozeff, 1993; Ursel and

Table 6 Factors influencing the market reaction to downsizing announcements.

N Mean t-student Median Z Wilcoxon

Standard deviation t-statistic (p- value)

Wilcoxon statistic (p-value)

Panel A: Univariate comparisons of event-period returns DISCPR Press release 67 �0.0089 �0.0031 0.0253 �2.970 �3.270

�2.890*** �2.930*** No press release 160 0.0016 0.0013 0.0239 (0.003)*** (0.001)***

0.834 1.378

REASON Proactive operations 91 0.0020 0.0014 0.02437 1.779

(0.077)* 1.833 (0.067)*

0.792 1.124 Reactive operations 136 �0.0039 �0.0011 0.02475

�1.838* �1.555

Panel B: Multivariate analyses of event-period returns Variables Model 3 Model 4 Model 5

Coef. t (p-value) Coef. t (p-value) Coef. t (p-value)

DISCPR �0.144 �2.082 (0.039)** �0.199 �2.674 (0.008)***

REASON 0.113 1.712 (0.088)* 0.115 1.722 (0.086)* 0.131 1.644 (0.102) FIRST �0.190 �2.657 (0.008)*** �0.191 �2.660 (0.008)*** DOWNSIZE �0.023 �0.340 (0.734) �0.076 �1.033 (0.303) �0.021 �0.313 (0.755) LAYOFF �0.022 �0.323 (0.747) �0.084 �1.134 (0.258) �0.022 �0.325 (0.745) SENSIND 0.050 0.755 (0.451) 0.059 0.873 (0.383) 0.052 0.789 (0.431) MIMET 0.123 1.809 (0.072)* 0.130 1.864 (0.064)* 0.128 1.848 (0.066)*

CRISIS 0.036 0.523 (0.602) �0.025 �0.357 (0.721) 0.034 0.496 (0.621) FORINV 0.092 1.379 (0.169) 0.120 1.748 (0.082)* 0.091 1.364 (0.174) FIRMSIZE �0.053 �0.600 (0.549) BM 0.070 1.022 (0.308) 0.077 1.110 (0.268) 0.066 0.960 (0.338) Proactive*DISCPR �0.126 �1.660 (0.098)* Reactive*DISCPR �0.102 �1.405 (0.162) Intercept �0.005 �1.109 (0.269) 0.003 0.165 (0.869) �0.005 �1.166 (0.245)

N 225 225 225 Adjusted R2 (Fisher) 0.082 (2.996)*** 0.053 (2.257)** 0.049 (2.056)**

Durbin-Watson 1.854 1.851 1.857

DISCPR = disclosure of a press release; REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; LAYOFF = downsizing operations made through layoffs; SENSIND = sensitive industries; MIMET = number of down- sizing operations announced by firms within the same industry; CRISIS = economic crisis/expansion periods; FORINV = foreign investors; FIRMSIZE = firm size; BM = book-to-market ratio; Proactive*DISCPR = proactive operations announced by press releases; Reactive*DISCPR reactive operations announced by press releases. *** Indicate significance at the 1% levels, respectively. ** Indicate significance at the 5% levels, respectively. * Indicate significance at the 10% levels, respectively.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 249

Armstrong-Stassen, 1995; Lee, 1997; Elayan et al., 1998), the negative market reaction to these operations should be higher in the case of high visibility. The univariate analyses also reveal that the market reaction to proactive operations is positive (0.20%) but not significant, whereas the market reaction to reactive operations is negative (�0.39%) and statistically signif- icant (p < 0.1).

Results of the regression analysis focusing on explaining differences in the market reactions, presented in Panel B of Table 6, confirm the univariate analyses in that the press release variable (DISCPR) is negative and statistically significant (model 3: p < 0.05; model 4: p < 0.01) when controlling for other potential determinants of the market reaction to such oper- ations. This is contrary to H2. Results also indicate that the market reaction to downsizing announcements is positively asso- ciated with proactive operations (models 3 and 4), although the REASON variable is significant at only p < 0.1. Further, when cases where the downsizing related to merger and acquisition (six observations) are removed, this relation is no longer sig- nificant at conventional levels.10 Thus, in the French context it is not clear whether, as in prior investigations in other settings (e.g., Abraham, 2004) investors positively value proactive downsizing operations. Second, our results confirm that investors have a more negative reaction to a firm’s first downsizing than to subsequent downsizing operations (model 3: p < 0.01). This is in line with Lee (1997) and Capelle-Blancard and Tatu (2012) who posit that in the case of multiple downsizing operations,

10 Other results reported in the paper do not change with the removal of the merger and acquisition-related downsizing observations.

250 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

the first downsizing contains more information, and thus, should have a greater impact on the market reaction. Consistent with Lee (1997), we obtain a positive relation between the market reaction and the number of downsizing operations announced by firms within the same industry (models 3 and 4: p < 0.1). Finally, our results show that the market reaction is positively correlated to the percentage of shares held by foreign investors (model 4: p < 0.1). All the remaining factors are not significant (DOWNSIZE, LAYOFF, SENSIND, CRISIS, FIRMSIZE and BM).

To better understand why investors react negatively to downsizing operations announced by press releases, we allow for impacts to differ across press releases disclosed to announce proactive vs. reactive operations (Proactive⁄DISCPR; Reac- tive⁄DISCPR). The results (model 5) reveal that the market reaction is more negative in the case of press releases disclosed to announce proactive operations (p < 0.1). More precisely, the disclosure of a press release leads to a reversal in the market reaction related to proactive operations. When firms disclose press releases to announce proactive operations, investors per- ceive these operations negatively, whereas they perceive proactive operations positively when there is no press release. This result suggests that investors perceive positively proactive operations but not the disclosure made by firms to announce such operations.

5. Content of press releases and consequences on the market reaction

In the final stage of our investigation, we conduct additional analyses examining the content of the press releases announcing downsizing operations and how the content relates to market reactions. The purpose of the content analysis is threefold. First, we provide evidence on the content of the press releases issued to announce downsizing operations, and then, we investigate whether the content contributes to the more negative market reaction to downsizing announce- ments disclosed through press releases. Finally, we examine whether the content of the press releases leads to a reversal in the market reaction related to proactive operations. The subsample consists of the 66 downsizing operations announced by press releases by 54 French listed firms.

5.1. Methodology

Although manual content analysis of the press releases is labor-intensive (Beattie et al., 2004), the relatively small size of the subsample allowed us to conduct an in-depth analysis of disclosures. Length of press releases varies from a minimum of 89 to a maximum of 2986 words, with an average of 679 words per press release.

Similar to Files et al. (2009), we observe varying degrees of prominence of the disclosure related to downsizing operations within press releases. Among the 66 press releases, 38 (58%) are entirely dedicated to downsizing operations, whereas 28 (42%) include additional subjects. In this latter case, the company describes the downsizing operations in the body of the press release, but the headline refers to other subjects. We take into account the prominence of the disclosure related to downsizing operations by using a variable (PROMWORD) that measures the total number of words related to the downsizing operation divided by the total number of words included in the press release. On average, 63.11% of the total number of words included in the press release is dedicated to the downsizing operation.

We next identify the reasons stated for downsizing operations. We rely on schemes used in prior studies (e.g., Capelle-Blancard and Tatu, 2012; Elayan et al., 1998; Lee, 1997) to distinguish proactive and reactive arguments and we pro- vide examples of these in Table 7. We use statements as the unit of analysis because they are more reliable than other units of analysis such as words or sentences. According to Milne and Adler (1999), individual words have no meaning to provide a sound basis for coding disclosures without a sentence or sentences for context. Given the possibility of multiple themes within the same sentence, we do not use sentences as the unit of analysis. Press releases were separately assessed by two independent coders.11 The Kappa index is 0.8 and indicates sufficient reliability (e.g., Fogarty and Rogers, 2005), and all coding differences between the two coders were resolved through discussions.

The content of press releases (PRCONT) is defined in terms of proactive and reactive arguments. Following Henry (2008), we calculate this variable as the count of proactive statements minus the count of reactive statements, divided by the sum of proactive and reactive statement counts. Therefore, a positive (negative) score indicates a higher number of proactive (reac- tive) arguments than reactive (proactive) arguments. The first specification of Eq. (2) (model 3) becomes Eq. (3):12

11 A c 12 Bec

(LAYOF

CARi ¼ b0 þ b1 PROMWORD þ b2 PRCONT þ b3 REASON þ b4 FIRST þ b5 DOWNSIZE þ b6 SENSIND þ b7 MIMET þ b8 FORINV þ b9 BM þ ei ð3Þ

With CARi = cumulative abnormal return for firm i over the event period from day 0 until day 2; PROMWORD = number of words related to the downsizing operation, divided by the total number of words in the press release; PRCONT = count of proactive statements minus count of reactive statements, divided by the sum of proactive and reactive statement counts;

opy of the coding instructions is available from the authors on request. ause of the small sample size for this analysis, we reduce the number of independent variables. Compared to model 3, we exclude one contextual factor F) and one legitimacy-related factor (CRISIS) because of the lack of significance of these variables in the previous analyses.

Table 7 Examples of proactive and reactive arguments.

Proactive arguments Reactive arguments

� Performance increase � Cost reduction � Improvement of customer satisfaction � Development of new products � Implementation of new technologies

� Financial distress � Bad market conditions � Bad sector conditions � Changes in regulatory requirements

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 251

REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWN- SIZE = size of the downsizing; SENSIND = sensitive industries; MIMET = number of downsizing operations announced by firms within the same industry; FORINV = foreign investors; BM = book-to-market ratio; b0 = intercept; b1. . .9 = regression coefficients; ei = residual term.

5.2. Results

As highlighted in Table 8 and in support of H3, we find no significant difference between proactive and reactive opera- tions in terms of arguments used to justify downsizing operations. Both proactive and reactive operations are mainly justi- fied using reactive arguments (see negative scores). The finding that proactive operations are being justified using reactive arguments is consistent with legitimacy theory. French firms appear to engage in impression management, presumably in an effort to make proactive operations appear more acceptable for employees, unions and the public.

Next, we examine whether the content of press releases issued to announce downsizing operations influences market reactions. Table 9 reports results for two specifications of Eq. (3): models 6 and 7. Results show that the content of press releases (PRCONT) is significantly related to differences in market reactions (model 6: p < 0.05). More precisely, we find a negative relation between the market reaction and the content of press releases. Supporting H4, we find the use of more proactive arguments to justify downsizing operations is significantly associated with more negative investor reactions. In contrast, the prominence of the downsizing operation in the press release is not significantly associated with the market reaction.

Finally, to test H5, we distinguish in model 7 the content of press releases disclosed to announce proactive vs. reactive operations (Proactive⁄PRCONT; Reactive⁄PRCONT). In support of H5, we find that the more firms use proactive arguments to justify proactive operations in the press release, the more the market reacts negatively (p < 0.05). This result is also in line with the legitimacy perspective, in that investors appear to perceive negatively the use of proactive arguments to justify proactive operations, potentially because it increases the risk of additional legitimacy costs being imposed on the firms.

5.3. Sensitivity tests

5.3.1. Endogeneity concerns We first check that results are not driven by selection bias introduced when we focus on disclosing firms. To address this

endogeneity concern, we employ the Heckman (1979) selection model. We use a two-stage regression with the two speci- fications of Eq. (1) (models 1 and 2) to estimate the likelihood of disclosing a press release that announces the downsizing, and then compute an inverse Mills ratio using the parameters of this model. The inverse Mills ratio, when included in the two specifications of Eq. (3) (models 6 and 7), is not statistically significant at conventional levels, suggesting our results are not driven by endogeneity bias.

5.3.2. Controlling for the overall tone of the press releases Next, we examine whether content results are influenced by the overall tone (TONE) of press releases issued by firms to

announce downsizing operations. Consistent with Henry (2008), we measure the variable TONE by the number of positive statements minus the number of negative statements, divided by the total number of statements. Non-tabulated univariate analyses show that the market reaction is not associated with the tone of the press release (p > 0.941). The results of mul- tivariate analyses, reported in Table 10, confirm that the variable TONE does not influence the market reaction, suggesting that our primary results are not driven by confounding information disclosed in press releases.

5.3.3. Further analysis of relations during crisis period As noted above, our test period overlaps with the financial crisis of 2008 and 2009, and our results show that while sam-

ple companies were less likely to use a press release to announce their downsizing operations during those years, market reactions did not differ significantly across the crisis period. In our last set of sensitivity tests, we explore whether the impact

Table 8 Univariate comparisons of the content of press releases between proactive and reactive operations.

N Mean t-student Median Z Wilcoxon Standard deviation t-statistic (p-value) Wilcoxon statistic (p-value)

Proactive operations 27 �0.2947 �0.3333 0.6151 �0.590 �0.560 �2.490** �2.406**

Reactive operations 40 �0.2045 �0.3333� 0.6125 0.557) (0.575) �2.112** 1.831*

Proactive (reactive) operations are implemented by firms with an increase (a decrease) in the ROA for the previous year. PRCONT = count of proactive statements minus count of reactive statements, divided by the sum of proactive and reactive statement counts.

** Indicate significance at the 5% levels, respectively. * Indicate significance at the 10% levels, respectively.

Table 9 Content analysis and multivariate analysis of event-period returns [�120; �20].

Variables Model 6 Model 7

Coef. t (p-value) Coef. t (p-value)

PROMWORD 0.082 0.654 (0.516) 0.035 0.279 (0.781) PRCONT �0.271 �2.059 (0.044)** REASON 0.057 0.472 (0.638) �0.028 �0.215 (0.831) FIRST �0.203 �1.450 (0.153) �0.177 �1.275 (0.208) DOWNSIZE �0.140 �1.120 (0.267) �0.115 �0.922 (0.360) SENSIND 0.163 1.277 (0.207) 0.199 1.557 (0.125) MIMET 0.135 1.085 (0.282) 0.151 1.230 (0.224) FORINV 0.168 1.364 (0.178) 0.145 1.188 (0.240) BM 0.199 1.626 (0.110) 0.200 1.658 (0.103) Proactive*PRCONT �0.336 �2.562 (0.013)** Reactive*PRCONT �0.080 �0.606 (0.547) Intercept �0.021 �2.375 (0.021)** �0.020 �2.251 (0.028)**

N 66 66 Adjusted R2 (Fisher) 0.117 (1.958)* 0.140 (2.055)**

Durbin-Watson 2.391 2.311

PROMWORD = number of words related to the downsizing operation, divided by the total number of words in the press release; PRCONT = count of proactive statements minus count of reactive statements, divided by the sum of proactive and reactive statement counts; REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; SENSIND = sensitive industries; MIMET = number of downsizing operations announced by firms within the same industry; FORINV = foreign investors; BM = book-to-market ratio; Proactive*PRCONT = content of press releases disclosed to announce proactive operations; Reactive*PRCONT = content of press releases disclosed to announce reactive operations. ***Indicate significance at the 1% levels respectively.

** Indicate significance at the 5% levels respectively. * Indicate significance at the 10% levels respectively.

252 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

of our independent variables differed during the financial crisis relative to periods of more normal economic operations. To examine this issue, we construct interaction terms using CRISIS and all combinations of the other independent variables used in the prior analyses. Our results (non-tabulated) indicated only one of the crisis period interaction terms was statistically significant at conventional levels. The CRISIS⁄PRCONT variable was negatively signed statistically significant (p < 0.05) in the analysis of differences in market reaction across press release content (original model 6). This finding indicates that investors reacted more negatively to the use of proactive arguments in the press releases during the financial crisis than during other sample years, supporting the position that investors see the use of that tactic as increasing legitimacy threats even more dur- ing periods of financial downturn.

6. Discussion and conclusions

In this study we investigate the determinants and consequences of French firms’ decisions to use press releases to announce downsizing operations. We also examine the content of press releases and its influence on investor reactions to downsizing announcements. The French context is a unique setting to study downsizing operations as it is characterized by extensive legal protections for workers, a strong protest culture, and a high strike rate. From a legitimacy point of view,

Table 10 Content analysis and market reaction after controlling for the tone of the press release.

Variables Model 6 bis Model 7 bis

Coef. t (p-value) Coef. t (p-value)

TONE 0.168 1.208 (0.232) 0.145 1.040 (0.303) PROMWORD 0.093 0.742 (0.461) 0.048 0.380 (0.705) PRCONT �0.362 �2.395 (0.020)** REASON 0.057 0.474 (0.637) �0.021 �0.163 (0.871) FIRST �0.221 �1.579 (0.120) �0.195 �1.394 (0.169) DOWNSIZE �0.143 �1.151 (0.255) �0.119 �0.961 (0.341) SENSIND 0.170 1.336 (0.187) 0.202 1.582 (0.119) MIMET 0.126 1.016 (0.314) 0.143 1.155 (0.253) FORINV 0.166 1.356 (0.181) 0.146 1.192 (0.238) BM 0.179 1.453 (0.152) 0.183 1.500 (0.139) Proactive*PRCONT �0.377 �2.755 (0.008)*** Reactive*PRCONT �0.153 �1.021 (0.312) Intercept �0.023 �2.587 (0.012)** �0.022 �2.427 (0.019)**

N 66 66 Adjusted R2 (Fisher) 0.124 (1.923)* 0.141 (1.969)**

Durbin-Watson 2.396 2.322

TONE = number of positive statements minus the number of negative statements, divided by the total number of statements; PROMWORD = number of words related to the downsizing operation, divided by the total number of words in the press release; PRCONT = count of proactive statements minus count of reactive statements, divided by the sum of proactive and reactive statement counts; REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; SENSIND = sensitive industries; MIMET = number of downsizing operations announced by firms within the same industry; FORINV = foreign investors; BM = book-to-market ratio; Proactive*PRCONT = content of press releases disclosed to announce proactive operations; Reactive*PRCONT = content of press releases disclosed to announce reactive operations. *** Indicate significance at the 1% levels respectively. ** Indicate significance at the 5% levels respectively. * Indicate significance at the 10% levels respectively.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 253

these operations thus constitute highly threatening events for firms, particularly when they are proactive. We find that French firms are more likely to issue press releases when they implement a downsizing for the first time during the period studied and when they belong to sensitive industries. In contrast, the probability of disclosure decreases for downsizing operations implemented through layoffs and during crisis periods. Overall, these results show that the French company use of press releases to announce workforce reductions is driven by both contextual and legitimacy factors. Next, we find that the average market reaction to downsizing operations announced by press releases is significantly more negative than the average market reaction to announcements where there is no press release (i.e., downsizing operations are announced by the media). We also provide evidence that the market reaction is more negative where the press releases announce proactive as opposed to reactive operations.

In order to better understand why the market reacts negatively when firms issue press releases, we examine the content of the press releases in terms of proactive and reactive arguments. First, we find that French firms, on average, adopt a reac- tive impression management strategy in their press releases as they mainly use reactive arguments to justify these opera- tions, even when the downsizing operations are proactive. This evidence suggests that firms may be using impression management to maintain organizational legitimacy. As noted by Love and Kraatz (2009), proactive operations are less well received by employees, unions and the public because they are implemented without apparent financial need. The need for legitimation is thus stronger in proactive than in reactive operations, and our results suggest firms use reactive arguments in these cases to potentially reduce their legitimacy threats. We further show that investors react more negatively to the use of proactive arguments for proactive operations indicating that French investors interpret the use of reactive arguments in such cases as reducing the exposure to the potential legitimacy costs of proactive downsizing operations. Overall, we show that in the French case, the disclosure strategies and their consequences on the financial markets relate to a legitimacy perspective. Both proactive and reactive operations are mainly justified by reactive arguments and the market reacts negatively to the use of proactive arguments presumably because they fail to reduce the risk of legitimacy damage.

Like all studies, this examination is subject to certain limitations. First, we focus on the study of disclosure at the time of the announcement of downsizing operations. However, it could also be interesting to examine the subsequent disclosures made by firms, for instance during the downsizing process. This should lead us to consider disclosure as a dynamic and mutual influence process. Indeed, it is likely that a firm’s disclosure interacts with both the unions’ disclosure and the media. Second, our investigation is limited to an analysis of archival data. Interviews with managers, heads of investor relations, or heads of financial communication could potentially reveal additional insight into companies’ use of press release disclosure at the time of downsizing operations. Of course, too, we examine only the French context in this study. Examining the extent

254 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

to which companies in other countries sensitive to employee issues use disclosure related to workforce reductions would prove valuable. Overall, by providing evidence of disclosure strategies and their impact on market reactions, this study con- tributes to the literature on CSR/sustainability and encourages future research on accounting choices in social contexts.

Acknowledgments

The authors wish to thank Editors Lawrence Gordon and Martin Loeb, Salvador Carmona, one anonymous reviewer, two anonymous participants (A and B) of the JAPP Conference, Isabelle Martinez, Niamh Brennan, Amy Hageman (discussant at the 2016 Conference of the American Accounting Association’s Public Interest Section Mid-Year Meeting) and the partici- pants of the 36th French Accounting Association Annual Meeting, the 3rd French Conference on Social and Environmental Accounting Research, the 38th European Accounting Association Annual Congress, the 2016 Conference of the American Accounting Association’s Public Interest Section Mid-Year Meeting for their valuable comments and suggestions provided on earlier versions of this paper. We also thank the ‘‘Labex Entreprendre” (Responsible Management and Entrepreneurship Chair), Montpellier Research in Management and the ESSEC Research Centre (CERESSEC) for financial support.

Appendix A. Event study methodology

The abnormal return for firm i on day t is equal to the difference between the observed return and the expected return for firm i on day t. Expected return is defined as that expected if the event did not take place.

ARit ¼ Rit � EðRitÞ

With ARit = abnormal return for firm i on day t; Rit = observed return for firm i on day t; E(Rit) = expected return for firm i on day t.

The observed return is:

Rit ¼ log½ðPit þ DitÞ=Pit � 1�

With Rit = return on the share price of firm i on day t; Pit = share price of firm i on day t; Dit = dividend paid on firm i’s shares on day t; Pit � 1 = share price of firm i on day t � 1.

For the estimation of the expected return, two benchmark models are used: the index model (1) and the market model (2). The expected return for firm i on day t is:

EðRitÞ¼ Rmt ðA:1Þ

or

EðRitÞ¼ ai þ biRmt ðA:2Þ

With E(Rit) = expected return for firm i on day t; Rmt = return on the CAC All-Tradable index on day t. In Eq. (A.2), the parameters ai and bi are OLS values that are calculated using an estimation period running from trading

days �180 to �20 prior to the event period (t = 0 is the event date). Average abnormal returns for each relative day are calculated by:

AARt ¼ð1=NÞ XN

i¼1 ARit

With ARit = abnormal return for firm i on day t. The cumulative average abnormal return over the event period from k days until l days is given by the equation below:

CAARk;l ¼ Xl

t¼k AARt

With AARt = Average Abnormal Return on day t.

Appendix B. Correlation matrix

DISCPR REASON FIRST DOWNSIZE LAYOFF SENSIND MIMETPR MIMET LABOR CRISIS FORINV FRSALES FIRMSIZE BM

DISCPR 0.003 0.173⁄⁄⁄ 0.147⁄⁄ �0.017 0.138⁄⁄ �0.033 �0.091 0.135⁄⁄ �0.148⁄⁄ 0.172⁄⁄⁄ 0.082 �0.292⁄⁄⁄ �0.136⁄⁄ REASON 0.003 0.02 0 �0.02 0.003 �0.110⁄ �0.092 0.081 0.185⁄⁄⁄ 0.097 �0.09 �0.018 0 FIRST 0.173⁄⁄⁄ 0.02 0.249⁄⁄⁄ 0.228⁄⁄⁄ �0.033 �0.098 �0.084 0.106 0.226⁄⁄⁄ �0.073 0.247⁄⁄⁄ �0.478 �0.028 DOWNSIZE 0.165⁄⁄ �0.043 0.231⁄⁄⁄ 0.130⁄ �0.045 0.011 �0.008 0.296⁄⁄⁄ 0.033 0.006 0.344⁄⁄⁄ �0.449⁄⁄⁄ 0.092 LAYOFF �0.017 �0.02 0.228⁄⁄⁄ 0.109 0.013 �0.046 �0.037 0.103 �0.067 0.038 0.214⁄⁄⁄ �0.389⁄⁄⁄ �0.012 SENSIND 0.138⁄⁄ 0.003 �0.033 �0.160⁄⁄ 0.013 �0.043 �0.128⁄ 0.001 �0.094 0.104 �0.104 0.091 �0.052 MIMETPR 0.001 �0.184⁄⁄⁄ �0.153⁄⁄ �0.003 �0.049 0.037 0.863⁄⁄⁄ �0.075 0.088 0.03 �0.092 �0.1 0.236⁄⁄⁄ MIMET �0.121⁄ �0.112⁄ �0.111⁄ �0.025 �0.06 �0.139⁄⁄ 0.742⁄⁄⁄ �0.08 0.027 �0.018 0.005 �0.05 0.293⁄⁄⁄ LABOR 0.111⁄ �0.109 0.164⁄⁄ 0.260⁄⁄⁄ 0.216⁄⁄⁄ 0.017 0.041 �0.037 �0.082 0.031 0.186⁄⁄⁄ �0.176⁄⁄⁄ �0.053 CRISIS �0.148⁄⁄ 0.185⁄⁄⁄ 0.226⁄⁄⁄ 0.094 �0.067 �0.094 �0.02 0.124⁄ �0.041 �0.042 �0.044 �0.164⁄⁄ 0.044 FORINV 0.123⁄ 0.026 �0.116⁄ �0.036 0.01 0.114⁄ 0.014 �0.082 �0.127⁄ �0.077 �0.147⁄⁄ 0.101 �0.133⁄⁄ FRSALES 0.046 �0.083 0.233⁄⁄⁄ 0.253⁄⁄⁄ 0.203⁄⁄⁄ �0.125⁄ �0.107 0.068 0.205⁄⁄⁄ �0.02 �0.214⁄⁄⁄ �0.435⁄⁄⁄ 0.104 FIRMSIZE �0.290⁄⁄⁄ �0.028 �0.501⁄⁄⁄ �0.425⁄⁄⁄ �0.383⁄⁄⁄ 0.093 �0.108 �0.083 �0.434⁄⁄⁄ �0.188⁄⁄⁄ 0.175⁄⁄⁄ �0.335⁄⁄⁄ 0.005 BM �0.137⁄⁄ 0.035 �0.049 0.025 �0.008 �0.024 0.181⁄⁄⁄ 0.246⁄⁄⁄ �0.008 0.079 �0.152⁄⁄ 0.052 0.047

Pearson (Spearman) correlation coefficients are presented below (above) the diagonal. The coefficients in bold are all higher than 0.5. DISCPR = disclosure of a press release; REASON = proactive/reactive operations; FIRST = first downsizing implemented by the firm over the period studied; DOWNSIZE = size of the downsizing; LAYOFF = downsizing operations implemented through layoffs; SENSIND = sensitive industries; MIMETPR = number of press releases disclosed by firms of the same industry to announce downsizing operations; MIMET = number of downsizing operations announced by firms of the same industry; LABOR = labor pressures; CRISIS = economic crisis/expansion periods; FORINV = foreign investors; FRSALES = net sales realized in France; FIRMSIZE = firm size; BM = book-to-market ratio.

E .

N ègre

et a

l./J. A

cco u

n t.

P u

b lic

P o

licy 3

6 (2

0 1

7 )

2 3

9 –

2 5

7 2

5 5

256 E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257

References

Abraham, S.E., 2004. Layoff announcements and employee guarantee announcements: how do shareholders respond? Int. J. Manpow. 25 (8), 729–740. Adams, C.A., Hill, W.-Y., Roberts, C.B., 1998. Corporate social reporting practices in Western Europe: legitimating corporate behavior. Br. Account. Rev. 30 (1),

1–21. Aerts, W., Cormier, D., 2009. Media legitimacy and corporate environmental communication. Acc. Organ. Soc. 34 (1), 1–27. Arndt, M., Bigelow, B., 2000. Presenting structural innovation in an institutional environment: hospitals’ use of impression management. Adm. Sci. Q. 45 (3),

494–522. Ashforth, B.E., Gibbs, B.W., 1990. The double-edge of organizational legitimation. Organ. Sci. 1 (2), 177–194. Baboukardos, D., Rimmel, G., 2016. Value relevance of accounting information under an integrated reporting approach: a research note. J. Account. Public

Policy 35 (4), 437–452. Ball, R., Kothari, S.P., Robin, A., 2000. The effect of international institutional factors on properties of accounting earnings. J. Account. Econ. 29 (1), 1–51. Barclay, L.J., Skarlicki, D.P., Pugh, S.D., 2005. Exploring the role of emotions in justice perceptions and retaliation. J. Appl. Psychol. 90 (4), 629–643. Beattie, V., McInnes, B., Fearnley, S., 2004. A methodology for analyzing and evaluating narratives in annual reports: a comprehensive descriptive profile and

metrics for disclosure quality attributes. Account. Forum 28 (3), 205–236. Beelitz, A., Merkl-Davies, D.M., 2012. Using discourse to restore organizational legitimacy: CEO-speak after an incident in a German nuclear power plant. J.

Bus. Ethics 108 (1), 101–120. Brammer, S., Millington, A., 2005. Corporate reputation and philanthropy: an empirical analysis. J. Bus. Ethics 61 (1), 29–44. Breton, G., Côté, L., 2006. Profit and the legitimacy of the Canadian banking industry. Account. Auditing Account. J. 19 (4), 512–539. Brockner, J., 1990. Scope of justice in the workplace: how survivors react to co-worker layoffs. J. Soc. Issues 46 (1), 95–106. Capelle-Blancard, G., Tatu, D., 2012. Stock market reaction to layoff announcements: European evidence (2002–2010). Working paper, University of Orléans. Cascio, W.F., 2005. Strategies for responsible restructuring. Acad. Manage. Exec. 19 (4), 39–50. Cascio, W.F., Young, C.E., Morris, J.R., 1997. Financial consequences of employment-change decisions in major U.S. Corporations. Acad. Manage. J. 40 (5),

1175–1189. Chalos, P., Chen, C.J.P., 2002. Employee downsizing strategies: market reaction and post announcement financial performance. J. Bus. Fin. Account. 29 (5),

847–870. Chen, P., Mehrotra, V., Sivakumar, R., Yu, W.W., 2001. Layoffs, shareholders’ wealth, and corporate performance. J. Empir. Fin. 8 (2), 171–199. Cho, C.H., 2009. Legitimation strategies used in response to environmental disaster: a French case study of Total’s Erika and AZF incidents. Eur. Account. Rev.

18 (1), 33–62. Cho, C.H., Patten, D., 2007. The role of environmental disclosures as tools of legitimacy: a research note. Acc. Organ. Soc. 32 (7/8), 639–647. Cormier, D., Magnan, M., 2003. Environmental reporting management: a continental European perspective. J. Account. Public Policy 22 (1), 43–62. Cowen, S.S., Ferreri, L.B., Parker, L.D., 1987. The impact of corporate characteristics on social responsibility disclosure: a typology and frequency-based

analysis. Acc. Organ. Soc. 12 (2), 111–122. Deegan, C., 2002. The legitimising effect of social and environmental disclosures – a theoretical foundation. Account. Auditing Account. J. 15 (3), 282–311. Deegan, C., Blomquist, C., 2006. Stakeholder influence on corporate reporting: an exploration of the interaction between WWF-Australia and the Australian

minerals industry. Acc. Organ. Soc. 31 (4/5), 343–372. Deegan, C., Gordon, B., 1996. A study of the environmental disclosure practices of Australian corporations. Account. Bus. Res. 26 (3), 187–199. Deegan, C., Rankin, M., Voght, P., 2000. Firms’ disclosure reactions to major social incidents: Australian evidence. Account. Forum 24 (1), 101–130. Depoers, F., 2000. A cost-benefit study of voluntary disclosure: some empirical evidence from French listed companies. Eur. Account. Rev. 9 (2), 245–263. Dowling, J., Pfeffer, J., 1975. Organizational legitimacy: social values and organizational behavior. Pac. Sociol. Rev. 18 (1), 122–136. Elayan, F.L., Swales, G.S., Maris, B.A., Scott, J.R., 1998. Market reactions, characteristics and the effectiveness of corporate layoffs. J. Bus. Fin. Account. 25 (3/4),

329–351. Elsbach, K.D., 1994. Managing organizational legitimacy in the California cattle industry: the construction of effectiveness of verbal accounts. Adm. Sci. Q. 39

(1), 57–88. Files, R., Swanson, E.P., Tse, S., 2009. Stealth disclosure of accounting restatements. Account. Rev. 84 (5), 1495–1520. Flanagan, D.J., O’Shaughnessy, K.C., 2005. The effect of layoffs on firm reputation. J. Manage. 31 (3), 445–463. Fogarty, T.J., Rogers, R.K., 2005. Financial analysts’ reports: an extended institutional theory evaluation. Acc. Organ. Soc. 30 (4), 331–356. Freeman, S.J., Cameron, K.S., 1993. Organizational downsizing: a convergence and reorientation framework. Organ. Sci. 4 (1), 10–29. Garcia Lara, J.M., Garcia-Osma, B., Mora, A., 2005. The effect of earnings management on asymmetric timeliness of earnings. J. Bus. Fin. Account. 32 (3/4),

691–726. Gray, R., Kouhy, R., Lavers, S., 1995. Corporate social and environmental reporting. Account. Auditing Account. J. 8 (2), 47–77. Greenhalgh, L., Lawrence, A., Sutton, R., 1988. Determinants of workforce reduction strategies in declining organizations. Acad. Manage. Rev. 13 (2), 241–

254. Griffin, P.A., Sun, Y., 2013. Going green: market reaction to CSRwire news releases. J. Account. Public Policy 32 (2), 93–113. Groening, C., Kanuri, V.K., 2013. Investor reaction to positive and negative corporate social events. J. Bus. Res. 66 (10), 1852–1860. Gunderson, M., Verma, A., Verma, S., 1997. Impact of layoff announcements on the market value of the firm. Ind. Relat. 52 (2), 364–381. Hahn, T.W., Reyes, M.G., 2004. On the estimation of stock-market reaction to corporate layoff announcements. Rev. Fin. Econ. 13 (4), 357–370. Hannifa, R.M., Cooke, T.E., 2005. The impact of culture and governance on corporate social reporting. J. Account. Public Policy 24 (5), 391–430. Harris, T.S., Lang, M., Moller, H.P., 1994. The value relevance of German accounting measures: an empirical analysis. J. Account. Res. 32 (2), 187–209. Heckman, J.J., 1979. Sample selection bias as a specification error. Econometrica 47 (1), 153–161. Henderson, B.H., Masli, A., Richardson, V.J., Sanchez, J.M., 2010. Layoffs and chief executive officer (CEO) compensation: does CEO power influence the

relationship? J. Account. Auditing Fin. 25 (4), 709–748. Henry, E., 2008. Are investors influenced by how earnings press releases are written? J. Bus. Commun. 45 (4), 363–407. Hillier, D., Marshall, A., McColgan, P., Werema, S., 2007. Employee layoffs, shareholder wealth and firm performance: evidence from the UK. J. Bus. Fin.

Account. 34 (3/4), 467–494. Hong, H., Kostovetsky, L., 2012. Red and blue investing: values and finance. J. Financ. Econ. 103 (1), 1–19. Hooghiemstra, H., 2000. Corporate communication and impression management: new perspectives why companies engage in corporate social reporting. J.

Bus. Ethics 27 (1/2), 55–68. Hunter, M.L., Le Menestrel, M., De Bettignies, H.-C., 2008. Beyond control: crisis strategies and stakeholder media in the Danone boycott of 2001. Corporate

Reput. Rev. 11 (4), 335–350. Jabbour, C.J.C., Santos, F.C.A., 2008. The central role of human resource management in the search for sustainable organizations. Int. J. Hum. Resour. Manage.

19 (12), 2133–2154. Jung, D.K., Aguilera, R., Goyer, M., 2015. Institutions and preferences in settings of causal complexity: foreign institutional investors and corporate

restructuring practices in France. Int. J. Hum. Resour. Manage. 26 (16), 2062–2086. Kent, P., Zunker, T., 2013. Attaining legitimacy by employee information in annual reports. Accounting, Auditing and Accountability Journal 26 (7), 1072–

1106. Lang, M.H., Lundholm, R.J., 2000. Voluntary disclosure and equity offerings: reducing information asymmetry or hyping the stock? Contemporary

Accounting Research 17 (4), 623–662. Leana, C.R., Feldman, D.C., 1988. Individual responses to job loss: perceptions, reactions and coping behaviors. J. Manage. 14 (3), 376–389.

E. Nègre et al. / J. Account. Public Policy 36 (2017) 239–257 257

Lee, P.M., 1997. A comparative analysis of layoff announcements and stock price reactions in the United States and Japan. Strateg. Manage. J. 18 (11), 879– 894.

Lin, J.-C., Rozeff, M.S., 1993. Capital market behavior and operational announcements of layoffs, operation closings, and pay cuts. Rev. Quant. Financ. Acc. 3 (1), 29–45.

Love, E.G., Kraatz, M., 2009. Character, conformity, or the bottom line? How and why downsizing affected corporate reputation. Acad. Manage. J. 52 (2), 314–335.

Love, E.G., Nohria, N., 2005. Reducing slack: the performance consequences of downsizing by large industrial firms, 1977–93. Strateg. Manage. J. 26, 1087– 1108.

Mäkelä, H., Näsi, S., 2010. Social responsibilities of MNCs in downsizing operations. Account. Auditing Account. J. 23 (2), 149–174. McKnight, P.J., Lowrie, A., Coles, C., 2002. Investor reactions, social implications and layoff announcements in UK: a comparison between periods. J. Manage.

Governance 6 (1), 83–100. Melloni, G., Caglio, A., Perego, P. (forthcoming). Saying more with less? Disclosure conciseness, completeness and in integrated Reports. J. Account. Pub.

Policy. Merkl-Davies, D.M., Brennan, N.M., 2007. Discretionary disclosure strategies in corporate narratives: incremental information or impression management?

J. Account. Literature 26, 116–196. Milne, M.J., Adler, R.W., 1999. Exploring the reliability of social and environmental disclosures content analysis. Account. Auditing Account. J. 12 (2), 237–

256. Missonier-Piera, F., 2004. Economic determinants of multiple accounting method choices in a Swiss context. J. Int. Fin. Manage. Account. 15 (2), 118–144. Mora, A., Sabater, A., 2008. Evidence of income-decreasing earnings management before labour negotiations within firms. Investigaciones Economicas 32

(2), 201–230. Munoz-Bullon, F., Sanchez-Bueno, M.J., 2011. Does downsizing improve organisational performance? An analysis of Spanish manufacturing firms. Int. J.

Hum. Resour. Manage. 22 (14), 2924–2945. Nishitani, K., Kokubu, K., 2012. Why does the reduction of greenhouse gas emissions enhance firm value? The case of Japanese Manufacturing Firms. Bus.

Strategy Environ. 21 (8), 517–529. Nixon, R.D., Hitt, M.A., Lee, H.U., Jeong, E., 2004. Market reactions to announcements of corporate downsizing actions and implementation strategies.

Strateg. Manage. J. 25 (11), 1121–1129. Ogden, S., Clarke, J., 2005. Customer disclosures, impression management and the construction of legitimacy. Account. Auditing Account. J. 18 (3), 313–345. Patten, D.M., 1992. Intra-industry environmental disclosures in response to the Alaskan oil spill: a note on legitimacy theory. Acc. Organ. Soc. 17 (5), 471–

475. Pouder, R.W., Hindman, H.D., Cantrell, R.S., 2004. How unions affect shareholder wealth in firms announcing layoffs. J. Labor Res. 25 (3), 495–502. Sharma, S., Ruud, A., 2003. On the path to sustainability: integrating social dimensions into the research and practice of environmental management. Bus.

Strategy Environ. 12, 205–214. Sheaffer, Z., Carmeli, A., Steiner-Revivo, M., Zionit, S., 2009. Downsizing strategies and organizational performance: a longitudinal study. Manage. Decis. 47

(6), 950–974. Shocker, A.D., Sethi, S.P., 1973. An approach to incorporating societal preferences in developing corporate action strategies. California Manage. Rev. 15 (4),

97–105. Simnett, R., Vanstraelen, A., Chua, W.F., 2009. Assurance on sustainability reports: an international comparison. Account. Rev. 84 (3), 937–967. Stevens, C.K., Kristof, A.L., 1995. Making the right impression: a field study of applicant impression management during job interviews. J. Appl. Psychol. 80

(5), 587–606. Suchman, M.C., 1995. Managing legitimacy: strategic and institutional approaches. Acad. Manage. Rev. 20 (3), 571–610. Ursel, N., Armstrong-Stassen, M., 1995. The impact of layoff announcements on shareholders. Ind. Relat. 50 (3), 636–649. Van Buren, H.J., 2000. The bindingness of social and psychological contracts: toward a theory of social responsibility in downsizing. J. Bus. Ethics 25 (3),

205–219. Vuontisjärvi, T., 2013. Argumentation and socially questionable business practices: the case of employee downsizing in corporate annual reports. Scand. J.

Manage. 29 (3), 292–313. Wilkinson, A., 2005. Downsizing, rightsizing or dumbsizing? Quality, human resources and the management of sustainability. Total Qual. Manage. 16 (8/9),

1079–1088. Worrell, D.L., Davidson III, W.N., Sharma, V.M., 1991. Layoff announcements and stockholder wealth. Acad. Manage. J. 34 (3), 662–678.