Transfer of event stigma: Reactions of Japanese regional governments to criticism on bid rigging

An organization may be stigmatized for the nature of its activities (core stigma), or for misconduct (event stigma). Stigma damages an organization's legitimacy and threatens its survival. Therefore, organizations respond by taking actions to mitigate stigma and restore legitimacy. However, there has been insufficient empirical analysis of responses to transferred event stigma resulting from other organizations' misconduct. This study quantitatively analyzes Japanese regional government entities' responses to criticism from mass media and political directives issued by the central government, after they had been stigmatized for a series of bid-rigging events. This study's major theoretical contribution is its finding that organizations are sensitive to event stigma that results from others' misconduct. Additionally, this study is among the first to examine the effects stigma has on government organizations, in contrast to prior research that has focused on private companies.


| INTRODUCTION
Society may stigmatize an organization for the nature of its activities (core stigma), or for misconduct (event stigma) (Hudson & Okhuysen, 2009). Stigma damages an organization's legitimacy and threatens its survival. Therefore, when stigmatized for its misconduct, an organization takes steps to mitigate the stigma with the objective of restoring its legitimacy.
Possible sources of event stigma are not limited to an organization's own misconduct. Misconduct by an organization may be widely dispersed (Jonsson, Greve, & Fujiwara-Greve, 2009) and damage the legitimacy of other organizations. However, prior research has mainly examined organizational responses to core stigma and event stigma that resulted from the focal organization's own misconduct. There has been insufficient empirical analysis of responses to transferred event stigma that resulted from other organizations' misconduct.
To bridge this critical research gap, this study analyzes quantitatively (1) the actions taken by Japanese regional governments to mitigate criticism from the mass media on the regional governments' involvement in a series of bid-rigging events; and (2) Japanese regional governments' responses to political directives issued by the central government to minimize further bid-rigging activities. Thus, the present study extends stigma theory since it is among the first to analyze organizational reactions to transferred stigma resulting from others' misconduct.
We begin this paper by reviewing the theoretical background of organizational stigma. Next, we explain Japan's public procurement system and describe the hypotheses and methods used to conduct this research. The discussion and conclusions are outlined in the last two sections.

| THEORETICAL BACKGROUND
This study focuses on actions organizations take to mitigate stigma and restore legitimacy (Zimmerman & Zeitz, 2002) after being damaged by social criticism triggered by misconduct. Devers, Dewett, Mishina, and Belsito (2009, p. 157) define stigma on organizations, or organizational stigma, as "a label that evokes a collective stakeholder group-specific perception that an organization possesses a fundamental, deep-seated flaw that deindividuates and discredits the organization." There has been a recent increase in stigma research (Hudson & Okhuysen, 2014).
Organizational stigma is classified into core stigma and event stigma (Hudson, 2008). Core stigma refers to "organizational-level stigma [that] occurs when some social audiences discount or discredit an organization because of core attributes, such as outputs, routines, or customers, that are in perceived violation of social norms" (Hudson & Okhuysen, 2009, p. 134). For example, the general public stigmatizes men's bathhouses-which serve as social venues for gay men-and subjects them to curious scrutiny for the nature of their services, even if they comply with laws and regulations. Meanwhile, event stigma refers to stigma caused by "the result of an unusual or anomalous event" (Hudson & Okhuysen, 2009, p. 134), and occurs when the general public or investors have lost confidence in an organization as the result of its misconduct, bankruptcy, or product deficiencies.
Stigma damages an organization's legitimacy and threatens its survival. Therefore, when stigmatized for misconduct, an organization takes various actions to mitigate the stigma and restore its previous aura of legitimacy. In this section, we review the literature on organizational actions (mitigations) taken in response to core stigma and event stigma, respectively.
2.1 | Mitigating core stigma Hudson and Okhuysen (2009) examined how organizations mitigate core stigma for their own survival and found that stigmatized organizations adopt "boundary processes." Such boundary processes include (1) "isolating" from public view; (2) "integrating" favorable organizations as insiders; (3) "dramaturgy," in which a stigmatized organization and a regulatory authority jointly play a role acceptable to outside audiences; (4) "associational" processes, which try to establish favorable relationships with stakeholders; and (5) "conventional" processes, which refer to ordinary actions that organizations generally take regarding their relationships with others.
Meanwhile, Durand and Vergne (2015) examined how firms in the arms industry responded to media attacks against core stigma.
They examined the extent to which media attacks on three types of parties-the focal firm, peers in the same industry subcategory, and peers in different subcategories-influenced firms to divest themselves of their arms industry assets in order to reduce the stigma associated with their firms. They found that attacks on focal firms are the most consequential, followed by attacks on peers in the same industry subcategory, and then by attacks on peers in different subcategories.

| Mitigating event stigma
Organizations also react to event stigma. Sutton and Callahan (1987) assessed the impacts of the stigma that followed the bankruptcies of US firms on the focal firms and their management, as well as the types of stigma-management strategies they used. They observed that these strategies comprised five categories: "concealing," "defining" (explaining the facts), "denying responsibility," "accepting responsibility," and "withdrawing" (refusing to participate in any mitigation activity). McDonnell and King (2013) analyzed the relationship between event stigmas caused by an organization's own misconduct and its mitigation activities.
They found that boycotted firms increased their prosocial claims. It was also found that firms whose reputations were being threatened by more media attention were more likely to increase their prosocial claims than firms faced with less media attention.
An organization is not only stigmatized for its own misconduct, but the misconduct of one organization may also damage the legitimacy of other organizations in the same category. Such stigmatization occurs in the following way. An organization's legitimacy depends on the audiences that evaluate it positively (e.g., Castelló, Etter, & Nielsen, 2016). Similarly, event stigma is premised on stakeholders' perceptions and evaluations (Devers et al., 2009). In this process, stakeholders generalize their evaluation to encompass all organizations with similar characteristics (Jonsson et al., 2009), and oversimplify the situation (Yu, Sengul, & Lester, 2008). Audiences assume that all organizations in the same category are likely to be involved in similar misconduct. As a result, stakeholders impose common sanctions on organizations not involved in the misconduct (Barnett & King, 2008).
Therefore, organizations stigmatized for other organizations' misconduct must then attempt to mitigate the transferred event stigma.
With regard to organizational responses to transferred stigma, Carberry and King (2012) examined the behavior of US Fortune 500 firms after the Enron scandal. They examined how audiences' challenges (federal investigations, shareholder proposals, and media exposure) to firms influenced the firms' reactions and found that firms targeted in federal investigations were more likely to adopt stock option expensing (SOPEX) than organizations that were not targeted.
It was also found that firms that received higher levels of media exposure were more likely to adopt SOPEX. Clearly, organizations react when stigmatized for the misconduct of other organizations.

| Problem
As discussed above, researchers have examined organizational reactions to core stigma and event stigma caused by the focal organization's own misconduct. However, as claimed by Yu et al. (2008), there is a dearth of analysis on organizational reactions to transferred event stigma resulting from the misconduct of other organizations. Carberry and King (2012) analyzed the relationships between the audiences' challenges and their likely reactions, but they did not consider the misconduct of other organizations in their analysis.
Whereas organizations' behavior is influenced by the external pressure (e.g., Ashforth & Gibbs, 1991;DiMaggio & Powell, 1983;Meyer & Rowan, 1977), evaluation by external actors, a source of stigma, is not only influenced by misconduct of a focal organization, but also by other organizations' misconduct. However, there is a paucity of research empirically analyzing this phenomenon (Yu et al., 2008).
In addition, as the demand for accountability has intensified in modern society, there is an increased practical need to examine organizational reactions to stigma resulting from others' misconduct. In particular, under the new public management (NPM) that has prevailed globally since the 1990s, public organizations are increasingly being required to be accountable (Dubnick, 2005). Private companies are also being asked by shareholders to improve corporate governance and be accountable. In these circumstances, organizations are being strictly scrutinized by external audiences and must account for their activities in more detail. The stigma that results from the misconduct of others, as well as their own misconduct, intensifies the general scrutiny they will experience. Thus, to adequately reflect modern society, researchers should also focus on the stigma caused by other organizations and organizational reactions.
To bridge this research gap, this study examines how organizations respond to stigma inflicted on them as a result of other organizations' misconduct and outsiders' evaluations in light of these events.
To answer this question, we analyzed quantitatively how Japanese regional governments reacted to stigma imposed on them as a result of a series of bid-rigging events and related political directives issued by the central government.

| PUBLIC PROCUREMENT IN JAPAN
This study examines a series of bid-rigging events that occurred in Japan's public procurement processes and the countermeasures implemented by regional governments. The general public scrutinizes Japan's public procurement processes strictly and evaluates whether these processes are being conducted appropriately. Public criticism became especially fierce when bid-rigging or bribery scandals were revealed. Therefore, the data that describe these circumstances will be useful in analyzing the relationship between stigma and organizational responses.
In the history of public procurement in Japan, discrepancies have surfaced since the 1990s. The series of bribery scandals involving the construction industry and politicians in the 1990s was one such event.
One response to this scandal argued that selective bidding, which had previously been the primary bidding method, facilitated bid rigging. industry and, subsequently, to a drop in their stock prices (Barnett & King, 2008). As more examples of misconduct in organizations become apparent, audiences apply more rigorous scrutiny and stigmatize them more severely, based on their evaluation (Devers et al., 2009;Jonsson et al., 2009). Therefore, organizations are forced to respond not only to their own misconduct but also to that of others (Carberry & King, 2012;Desai, 2011). Therefore, our first hypothesis, based on this notion, is as follows: Hypothesis 1 The more often the misconduct of other organizations in the same category is revealed, the more likely an organization in the same category is to take action to mitigate its stigmatization due to others' misconduct.

| Pressure from organizations at higher levels in the hierarchy
Organizations respond to questions about their legitimacy from external audiences (Oliver, 1991). Typically, organizations may automatically accept the regulatory authority's order as a rationalized myth (Meyer & Rowan, 1977). For organizations, such acquiescence (Oliver, 1991) is an easy reaction-by adopting this strategy, they are exempted from demonstrating their accountability. As a result, organizations accept superior organizations' orders as ceremonial conformity (Ashforth & Gibbs, 1991). Therefore, our second hypothesis is as follows: Hypothesis 2 The more an organization is exposed to pressure to accommodate its regulatory authority's requirements, the more likely the organization is to take action to mitigate the stigma it has incurred as a result of the misconduct of other organizations in the same category.

| An organization's own misconduct
An organization's perceived legitimacy is damaged when it is involved in misconduct. To mitigate the resulting stigma, organizations attempt to restore their lost legitimacy (e.g., Ashforth & Gibbs, 1991). For example, companies boycotted for their activities increase prosocial claims by sending messages to the general public (McDonnell & King, 2013). Thus, once an organization has been involved in misconduct, it takes remedial actions to mitigate the resulting stigma to recover its former legitimacy, as hypothesized below: Hypothesis 3 An organization that is guilty of misconduct is more likely to take action to mitigate the resulting stigma than an organization that is not involved in misconduct.

| Ranking
How organizations are ranked influences their behavior. Chatterji and Toffel (2010) found that, with regard to ratings on firms' environmental performance, firms with a low environmental rating were more likely to improve their environmental performance than other firms.
Thus, low-ranking organizations aim to portray images to their audiences by adopting measures favorable to their audiences.
Our study examines stigma-mitigating behaviors that send positive messages to the general public. The general public will consider lowering the OBT as a favorable measure, as it will extend the application of open biddings, which is considered favorable for fairness. Thus, low-ranking organizations (procurement agencies) will be more likely to take remedial action (lowering the OBT) to create a positive image.
Hypothesis 4 Organizations with low rankings are more likely to be proactive in mitigating stigma than organizations with high rankings.

| Sample and data source
This study examines the relationships between stigma and organizational reactions using data on Japan's public procurement processes between the fiscal year (FY) 2006and 2015(from April 1, 2006to March 31, 2016. Following a series of bribery scandals involving the construction industry and politicians in the 1990s and early 2000s, the general public applied much stricter scrutiny to bidding processes and to public procurements in general, requiring the procurement organizations and contractors involved to take steps to mitigate stigma. The GOJ took action in response to such scrutiny in the early 2000s, as described previously. Therefore, the period after FY 2006 was determined to be appropriate for our analysis. FY 2015 is the most recent year for which data were available at the time of our analysis. The procurement agencies-that is, the actors to be analyzedinclude Japan's 47 prefectural governments. We examine the relationship between the stigma that resulted from a series of bid-rigging events and the mitigation actions taken by the prefectural governments. The mitigation action we considered was "lowering the OBT" We acquired OBT data from a "A report on the study on successful bid rate (SBR) and suspicions about bid rigging," compiled by the Japan Ombudsman Liaison Office (JOLO) based on its survey of regional governments (JOLO, n.d.). Bid-rigging event data were acquired from a database compiled from articles in Asahi Shimbunone of Japan's major newspapers-following the method adopted by previous studies (e.g., Durand & Vergne, 2015).
We extracted all articles in which the titles or texts included the words "bid rigging" (dango in Japanese), and selected events we regarded as bid rigging (including those initiated by procurement agencies). The final number of events analyzed was 263. The types of procurement agencies that were involved in the events were prefectures (56); GOJ (28); regional governments other than prefectural governments, such as governments in cities, towns, and villages (159); unspecified agencies (4); and public corporations, such as the electric power company (16).

| Variables
Dependent variables. We observed the actions prefectural governments took to mitigate stigma that resulted from misconduct. Our dependent variable was therefore the mitigation reactions by prefectural governments, and we quantified it as "OBT lowering." OBT lowering (dummy variable) in FY n takes the value 1, if OBT at the end of FY n is less than that at the beginning of the FY and 0 otherwise. Independent variables. As independent variables, we applied "own stigma," "transferred stigma," "pressure from organizations at higher levels in the hierarchy," and "ranking of the focal organization." For own stigma, we used "misconduct of the focal organization" as a dummy variable, taking 1 if any of the focal organization's bid-rigging events were reported in a newspaper in the FY and 0 otherwise.
For transferred stigma, we measured the misconduct of other organizations in the same category by counting the total number of bid-rigging events the newspapers reported in the FY. It is assumed that prefectural governments are not only stigmatized for the bidrigging events of other prefectural governments, but also by those of the GOJ, cities, towns, villages, unspecified agencies, and public corporations. Therefore, we included all these public procurement agencies' events, except for those of the focal government.
As pressure from an organization at higher levels in the hierarchy, we used political directives issued by MIC to prefectures' governors.
During the period of analysis, the MIC, which is the regional governments' regulating authority, issued political directives to lower the OBT in the FYs 2006FYs , 2007FYs , 2011FYs , and 2014 under the name of its Minister (MIC, n.d.). We used ministerial directives as a dummy variable, which took the value 1 if the MIC issued a directive during the FY and 0 otherwise. There was no FY in which multiple directives were issued.
For ranking of the focal organization, we applied the weighted mean of the SBR of a government for a FY. The JOLO regards a high SBR as a symptom of bid rigging (JOLO, n.d.). Therefore, our study applied the weighted means of the SBR of a government for a FY as our independent (reverse) variable indicating a government's ranking; a high SBR indicates a low ranking. Hence, we calculated "a government's SBR for a FY" as the ratio of the sum of the successful bid prices of all procurement events to the sum of the estimated prices set by procurement agencies before all the events in a FY. Thus, Control variables. We controlled for "organization size," "earthquake disaster," "governor change," and "governor election." As organization size might influence organizations' behavior (e.g., Sherer & Lee, 2002), we applied a "standard financial scale" for each prefecture for FY 2014 to control the influence of organizational size. A standard financial scale is an indicator of a regional government's current general revenue in a normal situation, which is calculated as being the sum of standard tax revenues and ordinary tax grants from the central government. This number indicates the size of each prefectural government's operation and is equivalent to private companies' annual sales. The numbers were logged.

SBR of a government for a FY =
We used earthquake disaster (dummy) as a control variable that takes 0 for FY 2010 or earlier and 1 for FY 2011 or later. Since the Great East Japan earthquake on March 11, 2011, more reconstruction projects have been initiated, and the costs of construction materials and labor are higher than before. As a result, the nationwide weighed mean value of the SBR of procurement events ordered by prefectural governments changed significantly; while it had monotonically decreased from 95.3% (FY 2002) to 82.9% (FY 2010), except for FY 2008, after the earthquake, it increased monotonically from 86.5% (FY 2011) to 92.7% (FY 2015. This implies that bidders' behaviors might have changed, potentially influencing the behavior of procurement agencies. We therefore incorporated this dummy variable into our model. Changes in top management frequently lead to changes in organizational strategies. Durand and Vergne (2015), for example, controlled for chief executive officer changes in analyzing arms industry firms' responses to media attacks. Hence, we also controlled for governor change as a dummy variable, taking the value 1 when a governor was changed during the FY and 0 otherwise.
Meanwhile, when a governor election is anticipated in the near future, a current governor may initiate measures to reform public procurement as a symbolic action (Ashforth & Gibbs, 1991) to appeal to voters. Therefore, we controlled for such situations by entering a governor election dummy, which takes the value 1 when an election is planned for the next FY and 0 otherwise.

| Model
We applied the panel logistic regression model. Among the independent variables, the misconduct of other organizations in the same category and the SBR were lagged by 1 year to enhance causal inference. "Elections" was lagged by 1 year in the opposite direction, based on our assumption that governors might take actions in anticipation of future elections. Elections are planned and announced beforehand, except in the event of the death of the present governor or an unplanned resignation. Therefore, we can enter "elections" in FY n + 1 to explain the reactions of governments in FY n. The misconduct of focal organizations and pressure from organizations at higher levels in the hierarchy were not lagged, as we assumed that governments would react to these factors immediately. With regard to the standard financial scale, we applied the values for FY 2014 consistently, independent of FY, as the annual changes involved are trivial. Table 1 reports descriptive statistics. Table 2 summarizes the results of our regression analyses. Model 1, as a baseline model, includes only controls. An earthquake disaster (March 11, 2011) is significantly related to organizational reactions (p < .001); during and after FY 2011, regional governments lowered their OBT much less frequently than before. This can be explained by the fact that procurement agencies might have observed that lowering SBR by lowering OBT was no longer viable. In other words, procurement agencies might have recognized that open bidding would not attract many bidders because of the tight balance between demand and supply in public works following the earthquake.

| RESULTS
Model 2 incorporates variables related to organizational category levels (transferred stigma and pressure from organizations at higher levels in the hierarchy). Prefectural governments react significantly to the misconduct of other organizations (p < .001); the greater the number of bid-rigging events revealed nationwide, the more likely the governments are to lower their OBT to mitigate the stigma. This result strongly supports Hypothesis 1. Similarly, the governments react significantly to political directives from the MIC-that is, pressure from organizations at higher levels in the hierarchy (p < .001); organizations respond to these directives by lowering their OBTs in the FYs in which the orders are issued. This result strongly supports Hypothesis 2.
Model 3 further incorporates variables related to individual organizations [own stigma and SBR as a (reverse) substitute for ranking)].
However, the influence of these variables on organizational reactions is not significant. Thus, neither Hypothesis 3 nor Hypothesis 4 is supported. Of note is the fact that Hypothesis 1 is still strongly supported (p < .001), even after incorporating these two independent variables in model 3. This implies that organizations take steps to mitigate the stigma caused by the misconduct of other organizations as well as, or even rather than, the stigma caused by the misconduct of focal organizations. Model 3 still supports Hypothesis 2 regarding how the pressure imposed by organizations at higher levels in the hierarchy affects their reactions (p < .05).

| DISCUSSION
The results indicate that organizations, Japanese regional governments in our case, are sensitive to stigma caused by the misconduct of other organizations in the same category. While prior research focused on organizational reactions to counter core stigma (e.g., Vergne, 2012) or event stigma resulting from an organization's own misconduct (e.g., McDonnell & King, 2013), there have been few analyses of organizational responses to transferred event stigma (Yu et al., 2008). Our study is among the first to address this gap.
While Durand and Vergne (2015) state that organizations react more strongly to self-induced stigma than stigma resulting from other organizations' misconduct, our study found one case that differed somewhat from their findings. It is, of course, inferred that compared to a T A B L E 1 Descriptive statistics   (Sutton & Callahan, 1987), they may hold a press conference to provide an explanation for an appearance of misconduct. They may punish an officer involved in a case of bid rigging at the agency's initiative. They may even sue the bidder, deny their responsibility (Sutton & Callahan, 1987), and present themselves as victims. If so, another suitable dependent variable could be used instead of OBT lowering. We should examine organizations' reactions to their own misconduct using another dependent variable.
We found that the governments reacted significantly to pressure-that is, the directives issued by the MIC. This fact is consistent with the basic claim of new institutionalism; organizations tend to adapt to the prevailing institutional environment (Meyer & Rowan, 1977). In our case, OBT lowering did not always improve the effectiveness of the procurement itself or even create a new problem (Nakanishi, 2017). However, regional governments rushed to implement OBT lowering as a rationalized myth (Meyer & Rowan, 1977) or as ceremonial conformity (Ashforth & Gibbs, 1991).

| CONCLUSION
Using data from Japan's public procurement processes, this study examined how stigma that resulted from other organizations' misconduct and pressure from organizations at higher levels in the hierarchy influenced the mitigation actions taken by the stigmatized organizations.
This study's primary contribution to organizational theory is the finding that organizations respond sensitively to event stigma caused by the misconduct of other organizations in the same category, whereas prior research focuses solely on a focal organization's own misconduct (Yu et al., 2008). This result is applicable to governments in other countries, as they are similarly subject to pressure from outside audiences with stricter requirements for accountability under the rise of the NPM (Dubnick, 2005). Furthermore, this finding is also applicable to private firms to some extent, as they, like public organizations, need to respond to the external pressure to acquire legitimacy (e.g., Ashforth & Gibbs, 1991;DiMaggio & Powell, 1983;Meyer & Rowan, 1977).
Our research also discusses the practical implications for public administrators. Government entities are exposed to various pressures and are urged to respond appropriately. Our study found that regional governments in Japan are sensitive to the evaluations of their audiences. This result does not suggest that their actions represent symbolic management (Ashforth & Gibbs, 1991) in the pursuit of ceremonial conformity. However, problems that have resulted from the excessive application of open bidding have already been reported (Nakanishi, 2017). We emphasize that governments should not implement countermeasures solely in response to stigma or pressure from external audiences, but should also emphasize the measures' fundamental effectiveness.
Despite its contributions, this study was also subject to the following limitations. First, organizations' reactions to their own misconduct should be analyzed more closely. In our study, the governments' responses to the bid riggings in which they were involved were not significant. However, this may be explained by the fact that our dependent variable, OBT lowering, might not be a suitable measure of the responses. We should analyze the responses of procurement agencies to the stigma they incurred as a result of their own misconduct using other types of organizational reactions as variables, such as holding press conferences. By examining the differences in organizations' reactions to different sources of stigma (their own misconduct and those of other organizations), we would reveal the detailed mechanism of reactions to stigmas. Second, our data were based on government entities. We believe that our finding can be applicable to the private sector, considering that both public and private organizations must respond to external pressure in order to acquire legitimacy.
However, we must be cautious about generalizing our findings, given that these organizations depend on different resources with different goals and motivation. Comparable research on private companies is needed. Third, this quantitative research study did not analyze the mechanisms through which stigmatized organizations take reactions.
Further research is needed on the internal mechanism through which organizations respond to stigma.

ENDNOTE
1 In Japan, procurement agencies identify contractors through a "competitive bidding" process, in principle. Competitive bidding can follow an "open bidding" or "selective bidding" process. In open bidding, any registered contractor can submit a bid, as long as it satisfies the screening criteria. In selective bidding, only invited contractors can submit bids. "Bid rigging" refers to a situation in which bidders conspire to assign the successful bidder. Procurement agencies may also initiate bid rigging, in which officer(s) in charge unfairly intervene in deciding on the successful bidder and/or bidding prices.