Based on the signaling theory (Spence, 2002), we view the NPS’s “vote no” pre-announcement as a signal sent to the stock market. When the stock market receives a signal, it interprets the signal, and its interpretation is reflected in stock price changes. If the stock market views “vote no” announcements positively, then the stock price will rise, increasing shareholder wealth. On the contrary, if the stock market views “vote no” announcements negatively, then the stock price will decline, decreasing shareholder wealth.

The market interprets the NPS’s “vote no” pre-announcements positively if it expects that target firms will improve their governance structure or performance. The stock market can expect future improvements for two reasons. First, management is sensitive to reputational damage caused by the NPS’s targeting. The NPS’s veto disclosure implies that the NPS is dissatisfied with the current management’s decisions (Del Guercio et al., 2008; Grundfest, 1992). Hence, “vote no” announcements impair management’s reputation and positive image associated with having membership on the target firms’ board (Grundfest, 1992). Management has elite status owning to their titles (D’Aveni and Kesner, 1993; Mills, 2000). Managers are sensitive to reputational damage and feel shame on the “vote no” announcements because, to them, non-pecuniary rewards such as pride, reputation, and honor are as valuable as pecuniary rewards (Grundfest, 1992). Hence, the target firm’s managers will make an effort to rebuild their reputation and avoid being targeted in the future.

Second, management is likely to improve governance structures and performance due to their concern for career paths. Management is less likely to be invited to other boards or other prestigious positions when targeted by activists (Grundfest, 1992). The literature on outside directors shows that vigilant directors are assessed as good monitors, which contributes to a positive reputation. This type of reputation confers them additional board seats and prestigious positions. In contrast, negligent directors lose reputation and are unlikely to be appointed to other boards (Fama and Jensen, 1983; Fich and Shivdasani, 2007). Because NPS targeting implies that management made shareholder value-destroying decisions, the target firm’s management would experience a decline in its reputation. Therefore, it is unlikely to be appointed to other boards or other prestigious positions. To restore its reputation and prevent jeopardizing future career opportunities, the target firm’s management would improve corporate governance structure. All things considered, we can reason that the stock market would react positively to the NPS’s veto pre-announcements.

A competing argument posits that shareholders do not expect future improvements and are concerned about the negative signal of the NPS’s veto pre-announcements. The management ignores “vote no” announcements and not initiate any reforms (Grundfest, 1992) if they think the impact of a veto announcement is negligible. In many Korean firms, controlling shareholders’ opinion is more influential than institutional investors’ opinion. Even if controlling shareholders and their families own small amounts of shares, they maintain control using pyramidal, or circular patterns of ownership (Joh, 2003). Hence, investors might assume that the NPS does not have enough power to bring about changes in target firms. Investors might emphasize the fact that the firm is engaging in shareholder value-destroying behavior. From this point of view, the stock market would interpret the NPS’s veto pre-announcements as a negative signal that indicates the NPS’s dissatisfaction with management. Thus, the stock market would react negatively to the NPS’s “vote no” pre-announcements (Kim and Yon, 2014; Sohn and Bae, 2018). Since both positive and negative responses are expected, the paper empirically tests whether the stock market responds positively or negatively to the NPS’s “vote no” pre-announcements.

The media is an external constituent that shapes public opinion. Hence, the influence of media needs to be examined when studying stock market responses. Publicity over pension fund activism spreads information about the NPS’s pre-announcements, but there are conflicting views about the effects of publicity. From one perspective, extensive media coverage about the NPS’s veto disclosure could increase shareholder wealth by promoting target firms to improve governance structure and firm performance. If shareholders assume that media coverage initiates such improvements, the stock market will show a positive reaction to news about the NPS’s veto pre- announcements. The “vote no” announcement implies that the NPS is dissatisfied with the current management’s decisions and signals that the firm is engaging in shareholder value-destroying behavior (Del Guercio et al., 2008; Ertimur et al., 2010; Kim and Yon, 2014; Sohn and Bae, 2018). If the NPS’s “vote no” pre-announcements receive considerable media attention, the management feels a sense of shame because their prestige is damaged. Since directors and managers are sensitive to reputational damage (Grundfest, 1992), they might voluntarily reform corporate governance structure. The degree of shame would heighten as publicity increases. Therefore, when publicity is high, shareholders expect improvements in the governance structure, and such expectations lead to a positive stock market response. An interview with Richard Koppes, a former Chief Counsel of CalPERS, supports this view. Koppes stated that publicity is a powerful tool to gain leverage with management and that the media’s attention to governance promoted market-wide changes (Del Guercio and Hawkins, 1999).

Not only a shame but also pressure from the public might cause corporate governance reform. The media renders public opinion by drawing people’s attention to certain issues that would otherwise go unnoticed (McCombs and Shaw, 1972). Through public opinion, the media shapes the public image of corporations. Firms conform to social norms and behave in a way that is considered desirable to avoid being negatively reported in the media and having a negative public image (Dyck and Zingales, 2002). News about the NPS’s veto pre-announcements denotes the NPS’s negative evaluation of target firms and can harm the target firms’ public image. If the NPS’s “vote no” pre-disclosures are extensively covered in the media, the management feels pressure to reform governance structure in a legally and socially appropriate way. Previous studies found that media reports about a firm substantially influence its managers’ decisions. For instance, media coverage about a firm can change its strategy (Bednar et al., 2013) and make the firm more responsive to shareholders’ demand (Dyck and Zingales, 2002). To sum up, when the NPS’s veto pre-announcements are covered in the media, the management will feel a sense of shame and feel pressure to reform governance structure in a socially and legally appropriate way. Hence, shareholders might expect that the management will improve corporate governance, and such expectation would cause a positive stock market reaction.

Another view states that publicity decreases shareholder wealth because news about the NPS’s “vote no” pre-announcements spreads negative information about target firms and because shareholders do not expect that the target firms’ governance structure will be improved. A signal spreads further in the public when it is covered extensively in the media (Andrews and Caren, 2010). If the media works as a mechanism to diffuse negative information about target firms (i.e. the NPS’s dissatisfaction with the management), then extensive media coverage regarding the NPS’s “vote no” pre-disclosure will result in negative stock market reactions. A past study has found that negative news about a firm decreases its market value. DeAngelo et al. (1994) showed that negative media coverage of First Executive Corporation’s investment in junk bonds caused a sharp decrease in its market value, even though the market value of similar firms that were also invested in junk bonds did not sharply decline.

Moreover, shareholders might not expect future improvements despite the NPS’s veto pre-announcements being covered in the media. Media attention lasts for a short period of time and hence might not exert enough pressure on the target firm’s management to change governance structures or increase performance. Time and effort are required to improve corporate governance, and the duration of media attention would not long enough to enhance corporate governance (Kim et al., 2014). Therefore, it’s possible that publicity is negatively associated with stock market reactions. Because the stock market response is predicted in both positive and negative directions, the paper examines which direction is correct.

We use an event study methodology to measure the stock market reaction to the NPS’s “vote no” pre-announcement. The event day is defined as the day after a “vote no” pre-announcement. The NPS announced its voting decision in the late afternoon or at night, after the stock market had closed. Hence, the opening price on the day after the announcement contains the effect of the NPS’s proxy voting announcement without any contaminating factors. We first collected each firm’s daily adjusted opening price and used the market adjusted model to estimate abnormal returns. Because most firms hold annual shareholder meetings in late March, events are clustered in our analysis. When events are clustered, the market adjusted model gives less biased results than the market model (Kim and Yon, 2014).

Market adjusted abnormal returns (AR) are calculated by subtracting daily market returns from daily stock returns of each company. Cumulative abnormal returns (CAR) are the sum of the ARs during the event window. The event window in our study is four days, including the event day itself plus the three preceding days. Many firms in our sample held annual shareholders’ meetings soon after the NPS announcements. If the event window included the days after the shareholders’ meetings, decisions made at the meetings would affect the event study result. To single out confounding factors, the authors do not include the days after the announcement in the event window. The specification to test the hypotheses is as follows:

Our main independent variable is publicity, which measures the degree of media coverage at the firm level. We collected newspaper articles about the NPS’s vote no pre-announcements and counted the number of articles published between the “vote no” announcement time and the subsequent day’s stock market opening time. Only news articles published during this time affect the subsequent day’s opening price. Publicity is a dummy variable that takes the value of one if the number of news articles is higher than the median and zero otherwise.

Control variables include firm size, the NPS ownership ratio, past performance, dividend ratio, and debt ratio. Firm size is the log of total assets, and NPS ownership is a dummy variable that takes the value of one if the NPS ownership is higher than the median and zero otherwise. Past performance is the past five-year average ROE, and dividend ratio is a dummy variable that takes the value of one if the cash dividend divided by asset is higher than the median value and zero otherwise. The debt ratio is the past five-year average of debt divided by asset. All explanatory variables are lagged by one year.

In addition to the major analysis, the authors conducted the event study using the market model. We estimate the market model parameters over the 200-day period, ending 20 days prior to the NPS’s announcements. The results are in <Table 7> and are similar to the market adjusted model results. As for the stock market response, the event study using the market model also indicates that the stock market reacts negatively to the NPS’s “vote no” announcements. The first row of Panel A in <Table 7> shows that the CAR of the average four-day event window (-3, 0) is -1 percent, which is significantly different from zero with p values below five percent.

Regarding the effect of publicity on the stock market, the market model results also indicate that media coverage alleviates the negative stock market response and increases shareholder return. The second row of Panel A in <Table 7> reports that the average CAR (-3, 0) of target firms that received little media attention is -2 percent, which is significantly different from zero at the one percent level. The third row shows that the average CAR (-3, 0) of target firms that received extensive media attention is 0.3 percent, which is statistically insignificant. The results presented in Panel B of <Table 7> illustrate that the average CAR (-3, 0) of low publicity group firms and high publicity group firms are significantly different at the five percent level.

The paper further examines whether the negative stock market response is driven not by the NPS’s pre-disclosure to “vote no,” but just by the pre-disclosure. An event study is conducted using a sample of firms for which the NPS disclosed proxy voting decisions ahead of 2019 annual meetings but did not announce to veto. The 48 firms are in the sample, and two firms are excluded because they were not listed during the estimation period. The final sample includes 46 firms. The market model parameters are estimated over the 200-day period, ending 20 days prior to the NPS’s announcements. In the third column of <Table 8>, the average CAR (-3, 0) of the market adjusted model is 0.4 percent, and the average CAR (-3, 0) of the market model is 0.2 percent, both of which are statistically insignificant. The results suggest that the pre-announcement itself does not affect the stock market reaction. The first column of <Table 8> shows the event study result of the total firms that are the subjects of pre-announcements. The average CAR (-3, 0) is not significant, which also indicates that the results of our main analysis is not caused by the pre-disclosure but caused by the NPS’s pre-disclosure to “vote no.”

The fourth column of <Table 8> presents the difference between the average CAR (-3, 0) of the firms the NPS vetoed agendas and that of the firms the NPS did not veto. The difference is statistically significant at the ten percent level when we used the market adjusted model methodology. However, when we used market model methodology, the difference between the two groups is statistically insignificant. The result that the two groups have different average CAR (-3, 0) confirms the key assumption of this paper - people regard voting yes as an empty ritual (Grundfest, 1992) and are indifferent about voting yes.