The reading level for this article is Expert
We study the long-term performance of 267 Canadian mergers and acquisitions that take place between 1980 and 2000, using different calendar-time approaches with and without overlapping cases. Our results suggest that Canadian acquirers significantly underperform over the three-year post-event period. Further analysis shows that our results are consistent with the extrapolation and the method-of-payment hypotheses, that is, glamour acquirers and equity financed deals underperform. We also find that cross-border deals perform poorly in the long run.
The international wave of mergers and acquisitions (M&As) of the last decade has been exceptional, despite the recent slowdown. In the US alone, the years 1999 and 2000 each saw more than 9,000 transactions worth more than US$1 trillion (Mergerstat Review).
Canada has been no stranger to this trend. In 2000, a record-breaking year, Crosbie & Co. reported close to 1,300 transactions for a total value of almost 235 billion dollars. Nevertheless, there is a growing concern among financial economists over the prices that are being paid for M&As and how these major transactions may impact future corporate performance.
Significant accounting and finance research resources have focused on understanding the value creation process related to these transactions, as well as to the factors that drive it. Recently, numerous long-term event studies question the ability of stock markets to fully and rapidly interpret the consequences of major transactions such as mergers and acquisitions when they are announced. These studies document negative abnormal returns over the three to five years following the transactions that overwhelm the positive abnormal returns documented over short-term windows, making the net wealth effect negative (Andrade, Mitchell, and Stafford, 2001).
In this article, we undertake an out-of-sample investigation of the post-M&A performance using Canadian data. We include in our sample the M&A wave of the late 1990s, the most significant ever, both in terms of numbers and value. We also examine potential determinants of post-acquisition abnormal performance to understand better the sources of value creation or destruction arising from Canadian M&As. This article is one of the few studies available that examine Canadian acquisitions.
Further, we make a number of methodological choices. Time frameworks (event-time or calendar-time), abnormal return metrics, benchmarks, and weighting procedures vary across long-term post-acquisition studies, making comparisons difficult (Bruner, 2002; Agrawal and Jaffe, 2000) and the measurement of long-term abnormal performance complex (Mitchell and Stafford, 2000).
Several recent studies address methodological concerns with tests of long-term abnormal returns, which are inevitably joint tests of stock market efficiency and a model of market equilibrium. Although there is no consensus over the method that dominates, we prefer a calendar-time portfolio approach to deal with the cross-sectional dependence problem inherent in M&A studies since M&As occur in waves and within a wave, cluster by industry (Andrade et al., 2001). Further, to mitigate the low power of the calendar-time portfolio approach to detect abnormal performance because it averages over months of high and low M&A activity, we use a WLS procedure rather than an OLS procedure.
Our main findings are as follows. First, we find that in most cases, Canadian acquirers significantly underperform over the post-event period. Second, our results support the extrapolation and the method-of-payment hypotheses. We also find that cross-border deals perform poorly in the long run.
The paper proceeds as follows. Section I discusses previous literature. Section II discusses data and method used to estimate the average long-run abnormal performance of Canadian acquirers. Section III presents results. Section IV presents potential determinants of long-run performance, and Section V presents our conclusions.
One of the basic issues that remains unclear in finance is the poor long-term performance of acquiring firms. Franks, Harris, and Titman (1991) study 399 acquisitions during the 1975-1984 period. After adjusting for systematic risk and size, but not for the book-to-market ratio, they find positive and significant long-term abnormal returns only for small transactions.
Loderer and Martin (1992) study 304 mergers and 155 acquisitions that take place between 1965 and 1986. They observe a negative but insignificant abnormal return over the five subsequent years (significant when measured over three years) for the mergers and positive but insignificant abnormal return for the acquisitions. However, they note that the long-term abnormal returns over three years are statistically significant only in the 1960s.
In contrast, Agrawal, Jaffe, and Mandelker (1992) find negative and significant abnormal returns for 937 mergers over the five subsequent years, and positive but insignificant abnormal returns for 227 tender offers that occurred between 1955 and 1987.
Recently, Loughran and Vijh (1997) and Rau and Vermaelen (1998) improve the measure of long-term performance by adjusting for systematic risk, size, and book-to-market. Loughran and Vijh find that five-year buy-and-hold abnormal returns are -15.9% (t = -2.36) for mergers, but 43% in the case of tender offers (t = 1.67). They also perform the same tests on a sample in which there is no overlapping of events by the same firm and find long-term abnormal returns of-14.2% (t = -1.69) for mergers and 61.3% (t = 1.86) for tender offers. Rau and Vermaelen (1998) use a three- (rather than a five-) year window and find that long-term abnormal returns are respectively negative and significant for mergers (-4.04%), but positive and significant for acquisitions (8.85%).
Mitchell and Stafford (2000) and Ikenberry, Lakonishok, and Vermaelen (2000) perform a more general examination of the long-term financial performance of three types of events: share repurchases, equity offerings, and M&As. Mitchell and Stafford (2000) analyze 2,068 transactions announced between 1961 and 1993 and report negative mean abnormal monthly returns over three years of -0.04% and -0.03% for equal-weighted and value-weighted M&A portfolios respectively, using calendar-time abnormal returns based on the Fama-French three-factor model. Ikenberry et al. (2000) examine a sample of 27 acquisitions in Canada between 1989 and 1995, a third of which are financed by shares. They find that the three-year abnormal returns are negative, but not significantly different from zero.
There are few studies on the long-run performance of acquiring firms in Canada. Eckbo and Thorburn (2000) examine a sample of 1,800 domestic and foreign successful acquisitions in Canada during the 1964-1983 period, before the introduction of substantive Canadian antitrust laws governing acquisitions. They find that Canadian bidders earn significant and positive average announcement-period returns, but US bidder returns are not significantly different from zero.
Eckbo, Giammarino, and Heinkel (1990) analyze the abnormal returns to the shareholders of 182 acquiring companies between 1964-1982. They report abnormal returns of 5.7% when acquisitions are financed with cash and stocks, and 2.7% when acquisitions are financed by stock alone. Abnormal returns are not significant when they are paid in cash. Eckbo (1986) finds insignificant results for successful Canadian bidders, as do Betton and Williams (2001) for a more recent sample.
II. Data and Method
We first describe our data and then the method.
We obtain our data sets of Canadian M&As from the Thomson Financial Securities Data Corporation (SDC) Worldwide Mergers and Acquisitions database. Our data must meet the following criteria: 1) Observations are for 1980-2000; 2) Deals are completed; 3) Deals are mergers, exchange offers, or acquisitions of majority interest; 4) Companies with several M&As during the period are included; 5) Only transactions greater than US$10 million are included; 6) Companies have market data and financial statements available in the Research Insight Compustat database over the April 1980-April 2001 period. Our final sample comprises 267 events (176 companies).
Figure 1 and Table I present descriptive statistics. Figure 1 plots the number of acquisitions and the total dollar value of the transactions by year.
[FIGURE 1 OMITTED]
Panel A of Table I reports the annual numbers, aggregate values, and mean values of acquisitions completed during the 1980-2000 period. Our sample comprises 267 acquisitions, with a market value of over US$100 billion. We note a trend in the data: few transactions take place before 1994, and the number of acquisitions continuously increases during the late 1990s.