Production Suspension, Corporate Governance, and Firm Value

ABSTRACT

This study investigates the effects of planned and unplanned production suspension on firm value. We find that the increased business risk caused by the production suspension negatively affects firm value. Additionally, good corporate governance alleviates the negative effect of planned suspensions, while only firm-specific factors affect unplanned suspensions. Moreover, among firms facing unplanned suspensions, those with strong governance recover faster than firms with weak governance because the former have stable managerial structures. Finally, business group-affiliated firms cope with planned production suspensions well, whereas they are more vulnerable to unplanned suspensions. The result suggests that the group-level supplier-demander relationships have potential business risk owing to low flexibility in unexpected business situations.

KEYWORDS:

• Corporate governance
• production suspension
• business risk

JEL:

• D22
• G32
• G34

Notes

1. Japan’s Nikkei 225 dropped 6.2% on the first trading day after the Tohoku earthquake and dropped 10.6% on the second trading day. Notably, the stock price of Toyota Motors dropped 14.74%, Nissan Motors dropped 12.53%, and Honda Motors dropped 10.15% in two days after the Tohoku earthquake.

2. Korean firms cannot provide information about critical accidents to any entity before a public disclosure using DART by law (Han et al. 2014; Kwon and Han 2020). This law minimizes the potential for information leaks that may affect stock returns before the public announcement.

3. Unplanned production suspensions rapidly increase business uncertainty. In highly uncertain conditions, Waldman et al. (2001) and Lam and Yeung (2010) suggest that strong CEO leadership contributes to the business’s success.

4. Although many previous empirical studies find a positive relationship between corporate governance and firm performance, some previous studies report different results. Particularly, different corporate governance measures may cause inconsistent empirical results. For example, Bhagat and Bolton (2008) find that there is no statistically significant relationship between GIM G-Index (generated by Gompers, Ishii, and Metrick (2003)) and next year’s Tobin’s Q after controlling for the endogeneity issue. Schultz, Tan, and Walsh (2010) also show no causal relationship between corporate governance and firm performance using Australian Securities Exchange data. However, many empirical studies still suggest a significant relationship between corporate governance and firm performance (Ammann, Oesch, and Schmid 2011; Bae et al. 2012; Leung and Horwitz 2010). Moreover, many previous studies using Korean samples show the significant relationship between corporate governance and Korean firm’s performance (Baek, Kang, and Park 2004; Black et al. 2015; Joh 2003; Pae and Choi 2011). We also follow these studies.

5. Many chaebol-affiliated firms in Korea construct vertically integrated networks in Korea. For example, Lee and He (2009) show the vertical integration network in Samsung group, which has three tiers – Samsung Electronics as a final assembler, Samsung Electro-Mechanics and Samsung SDI in the middle tier, and Samsung Corning as a parts supplier.

6. The just-in-time (JIT) manufacturing strategy represents modern supply chain management. For example, Sakakibara et al. (1997) show that over 700 related articles were published from 1985 to 1990.

7. Many empirical studies demonstrate negative abnormal returns. For example, McConnell and Muscarella (1985), Blackwell, Marr, and Spivey (1990), Brickley and Van Drunen (1990), and Gombola and Tsetsekos (1992) also report negative abnormal returns around the announcement of a production suspension.

8. Although internal capital markets in a business group may increase capital allocation efficiency, some previous studies suggest the cost of group-level internal capital markets (Almeida and Wolfenzon 2006). Particularly, Gopalan, Nanda, and Seru (2007)show negative spillovers in the Indian business group using firm bankruptcy samples. Therefore, if the business restructuring process is unsuccessful, the event will negatively impact the entire group.

9. Binder (1998) reported that the beginning of the event study methodology is the paper by Fama, Fisher, Jensen and Roll (FFJR) (1969), and discusses the various applications of the original FFJR technique since 1969. Particularly, both event period and benchmark model of normal returns are important to measure abnormal returns. Our study follows the methodology of Mikkelson and Partch (1986) that is cited many studies that followed it, such as Fields and Mais (1991) and Szewczyk (1992). We calculate normal returns using two overall country indices (KOSPI composite index and KOSDAQ composite index) because our samples are separated into two different stock markets. KOSPI is similar to NYSE, and KOSDAQ is similar to NASDAQ, respectively.

10. The event date 0 is the first announcement date or the first trading date after the announcement of production suspension by DART. Market investors may recognize the issue by the first announcement, and the stock market reflects this new information about the firm’s operation. This is common for previous empirical studies using event study methods, such as those of Chan, Ge, and Lin (2015).

11. Previous studies use these two databases to collect financial and ownership data in Korea. Korea Listed Companies Association provides financial and ownership data through TS2000. Black, Jang, and Kim (2006), Kang and Kim (2006), and Joh and Jung (2018) use the TS2000. The FN Dataguide is another useful database that is similar to Thomson Reuters DataStream. Recent finance studies also use this database to collect financial data, stock prices, and public event data in Korea. Lee, Park, and Shin (2009), Choe and Yang (2010), Joh and Jung (2012), and Kang and Kim (2013) use this database for their research.

12. In unreported results, we test the relationship between the board independence index, board structure index, and CARs. Although both coefficients are positive, they are insignificant.

13. Additionally, we measure BHARs of one year (−256 day, 0) before the announcement. It represents the pre-performance of production suspension.