گستاخی در میان پیشنهاد دهندگان ژاپنی

We consider the issue of hubris in the Japanese mergers and acquisitions (M&A) market. Although prior research suggests that hubris should be and is less severe in the Japanese market, our findings suggest that it still has a significant presence. Using past (excess) market return as a proxy for the likelihood of hubris, we find that high hubris bidders frequently have negative event period abnormal returns, while low hubris bidders have positive event period abnormal returns. Given the importance of keiretsu in Japan and their similarity to Korean chaebols we consider these results in light of the alternative hypothesis of tunneling, which explains similar results in Korea. Finally we consider whether the results are driven by better performing bidders using acquisitions to pay in stock while poorer performing bidders choose to pay in cash. Our results are largely consistent with the hubris hypothesis where over-confident managers may engage in value-destroying M&A.

Behavioral corporate finance may be approached from two different perspectives. One is the “irrational investors” perspective and the other is the “irrational managers” perspective (Baker et al., 2004). Roll (1986) develops the idea of “irrational managers” and proposes a “hubris theory” to explain some well-documented findings in the merger and acquisition (M&A) literature. To date, hubris is one of the most cited reasons for M&A. Roll (1986) points out that managers may be optimistic and overconfident in their valuation of target firms. Consequently they tend to engage in frequent transactions and they are inclined to overpay when acquiring targets. Both prior research (Moeller et al., 2004 and Jensen and Ruback, 1983), which has shown little or no value creation after mergers, and observed acquirer announcement effects are consistent with the hubris hypothesis. It is important to identify the presence of hubris in the M&A market because it helps to predict the efficiency effects of mergers (Hartman, 1996) and to motivate CEOs to progressively correct their overconfidence (Aktas et al., 2005). Many studies have examined the U.S. market (see, for example, Malmendier and Tate, 2003, Roll, 1986 and Hayward and Hambrick, 1997) and concluded the presence of hubris amongst U.S. bidders. Some have also looked at international markets and examined the impact of hubris on firm value. For example, Raj and Forsyth (2003) study United Kingdom (U. K.) bidders and find that hubris bidders significantly lose value on the announcement of a bid. This paper focuses on the Japanese market and examines the M&A announcement effects for hubris bidders. Kang et al. (2000) study M&A activities between 1977 and 1993 and show a two-day abnormal return of 1.2% for acquiring firms. Subsequently, Van Schaik and Steenbeek (2004) update the sample and find the abnormal return to be around 1.5% from 1993 to 2003. Kester (1991) suggests that when large-scale mergers take place in Japan, they are invariably friendly deals among companies with established relationships. Consistent with this, Kruse et al. (2004) argue that hubris is less likely to occur in the Japanese market because most of the mergers are friendly. Although, the magnitude of hubris may be smaller for friendly mergers compared to that in hostile takeovers, there still exists a greater likelihood for a “hubris manager” to engage in more merger transactions. To date, there has been no study that empirically examines the extent of hubris in the Japanese M&A market. Using a sample between 1989 and 2003, we use various proxies for bidder hubris. We find that various measures of past bidder profitability have little relationship to bidder cumulative abnormal returns (CARs) around their bids. When we use past bidder stock market performance as a proxy for hubris, we find that the lowest past-performing quintiles (representing firms whose managers have the lowest hubris) have consistently positive and significant event period CARs, while the highest past-performing quintiles (representing firms whose managers have the highest hubris) have consistently negative and significant CARs for longer event periods and insignificant CARs for shorter periods. Invariably the difference in CARs for low versus high hubris bidders is positive and significant. We also explore the possibilities that these results can be explained by either tunneling or some Japanese bidders choosing to exchange their (overvalued) stock for that of targets; neither hypothesis explains our results. Our study extends the theory of behavioral corporate finance to the Japanese market. It quantifies the impact of hubris in the Japanese M&A market and contributes to the understanding of hubris in international markets. The remainder of the paper proceeds as follows. In Section 1, we provide some background. Section 2 discusses the data and methodology. Section 3 presents hypothesis and empirical results. Finally, Section 4 offers concluding remarks.

We study the effect of hubris in the Japanese M&A markets. Although prior research suggests that hubris is less severe in the Japanese market, our findings suggest that it still has a significant presence. Using past excess market returns to proxy the likelihood of hubris, we show that high hubris bidders have negative abnormal returns over longer event windows, while low hubris bidders have positive abnormal returns. We find that the CAR differences between low and high hubris bidders, especially for longer CAR measurement periods, are quite large economically, with the 12-month excess returns providing particularly powerful results. Alternative explanations are explored. We examine acquisitions where both participants are in a keiretsu to consider the hypothesis of tunneling. In contrast to Bae et al. (2002), we find that this subset of firms has quite positive announcement window CARs, a result inconsistent with tunneling. We also compare stock versus non-stock offers and we do not find that acquirers utilizing their (overvalued) stock drive our results. Explanatory regressions of CARs are consistent with the earlier analyses and indicate that excess returns are important during subperiods of our study. While our results are consistent with Kang et al. (2000) in most aspects, a notable exception is that we find prior excess returns (significantly) negatively impact on CARs and they find that prior excess return (significantly) positively impact on CARs. A full explanation of this important difference is not provided. However, we note that their analysis is of mergers from 1979 through 1993; ours is of a rather different period of 1989 through 2003; the intersection is only about 4 years. When we look at our pre-bubble period we do find that excess return is far less important than in the later periods. One might speculate that over time, as the Japanese economy and market changed, the relevance of this variable might also have changed. We conclude that our results are consistent with the hubris hypothesis where over-confident managers may engage in value-destroying M&A.