Biases in Estimating Long Run Abnormal Returns and Conditional Measures of Performance: The Evidence on Takeovers Re- Examined

Using the UK as a “natural experiment” that allows examination of the effect of the form of payment separately from bidder hostility in a way not easily achievable in the US, we show that the form of payment hypothesis interacts with bid hostility, or “disciplinary bidding” in explaining acquirer wealth effects. Equity financed acquisitions and non-hostile bids generate negative returns, and their combination in a bid is particularly detrimental. Overall, neither cash financing nor bidder hostility generate positive abnormal returns. However, there is some evidence that the combination of cash financing and hostility that uniquely generates positive shareholder returns. In using this experimental arena, we note that recent UK research has produced conflicting evidence on the previously documented stock market ‘anomaly’ of acquiring firms exhibiting substantial negative abnormal returns following acquisitions. This raises the intriguing question as to whether or not the UK experience is different from that of the US, where negative abnormal returns to acquirers is now well-established (Agrawal et al [1992], Loughran and Vijh [1997]). We show that when tests are used which control for the skewness and bias in the estimation of long run abnormal returns, as prescribed in Lyon et al (1999), the negative performance of UK acquiring companies documented by Limmack (1991), Kennedy and Limmack (1996), Gregory (1997) and Conn et al (2004) is confirmed. The evidence here shows that the magnitude of these abnormal returns may have been under-estimated by previous studies. In line with previous research, we find support for the form of payment hypothesis (Aggrawal and Jaffe, 2000), but also show the combination of form of the bid and form of payment is a critical factor in determining post bid returns. We also show that these results are robust to specifications which control for performance conditional upon time-varying risk and expected return measures and for returns calculated in calendar rather than event time.