Market returns to acquirers of substantial assets.

Abstract:

Does poor post-acquisition performance characterise firms that make non-M&A acquisitions? We investigate the wealth effects of substantial asset acquisitions (i.e. acquisitions that cost over $10 million) on acquiring firms' shareholders. We find significant abnormal positive market reaction to asset acquisition announcements and, contrary to findings for firms undertaking M&As, the acquiring firms perform exceptionally well post-acquisition. Our findings are robust to the research method weaknesses common to many studies of long-term performance and we control for free-cash-flow as well. Our results contradict the hubris hypothesis of acquisitions and lend weight to the argument that the auction-style process that characterizes corporate takeover bids contributes to overpayment.

Keywords:

ASSET ACQUISITIONS; MARKET EFFICIENCY; LONG-TERM PERFORMANCE.

1. Introduction

Asset acquisitions are frequent and economically significant events. In Australia alone, more than ten-thousand asset acquisitions took place during the 1990s. The number of asset acquisitions with a deal value of greater than A$10 million, which took place over the period extending from 1995 to 1999, substantially exceeds the number of takeovers. (1) However, despite their evident economic importance, substantial asset acquisitions have received little academic attention and the wealth effect of asset acquisitions on shareholders of acquiring firms is not wholly understood.

We address two questions: (a) Is there a significant market reaction to announcements of substantial asset acquisitions? And, (b) do the acquirers experience predictable abnormal performance post-acquisition? Answers to these questions are interesting in light of widespread evidence that announcement period returns to acquirers in takeovers are generally insignificant and that their post-acquisition performance is abnormally poor, on average. Evidence of similar returns to acquirers of substantial assets would indicate that the anomalous drift in abnormal returns is a robust phenomenon beyond the takeover market and add to the impressive body of research that suggests the market may take a considerable period of time to absorb value relevant information. (2)

Not everyone is convinced that markets are inefficient to the extent implied by some of the research to date (eg, Fama 1998), and indeed our results show that asset acquisitions are typically value-adding investments. Importantly, our research method incorporates controls for the now well-known biases and problems in evaluating long-run performance.

The rest of the paper proceeds as follows. Section 2 shows how our research fits the literature. Given the extensive body of work on M&A activity, we focus on the salient findings on sharemarket returns to acquirers in takeovers to show how they condition our expectations about the returns to acquirers of substantial assets. Section 3 describes the data and research method. Section 4 details and discusses the results. A summary and conclusion comprise section 5.

2. Sharemarket Returns to Acquirers: Salient Findings

2.1 Pre-Bid Performance

Empirical evidence consistently shows that most firms that make a takeover bids have a recent history of exceptional performance (Asquith 1983; Walter 1984; Bishop, Dodd & Office 1986; Schwert 1996). Roll (1986) observes good performance in the pre-bid period is consistent with the hubris hypothesis, that acquirers with a history of good performance suffer from excessive confidence and therefore overpay for their target firms. However, positive pre-bid abnormal returns are also consistent with the argument that managers who have performed well in the past do so because of superior managerial skills (Morck, Shleifer & Vishny 1990) and look to takeovers as a profitable avenue of investing their surplus funds (Dodd 1980). There is no reason why this should not also apply to the acquisition of substantial assets. Firms that are performing well are more likely to seek out new investments irrespective of the investment being in the form of a listed company or other assets. In short, the evidence to date gives rise to a strong expectation that firms acquiring substantial assets are likely to be firms that have performed exceptionally well.

2.2 Announcement Period Returns

Although target shareholders gain from takeover announcements (Jensen & Ruback 1983; Jarrell, Brickley & Netter 1988; da Silva Rosa 1994), this is less clear for shareholders of the acquiring firms. Average returns to bidding shareholders are at best slightly positive in some studies, and significantly negative in other studies (Morck, Shleifer & Vishny 1990).

Jarrell, Brickley and Netter (1988) observe that there has been a secular decline in the returns to bidding firms over the 1960s, 1970s and 1980s and posit that part of the reason is the passage of legislation which has strengthened the bargaining position of targets and increased competition among bidders. Bradley, Desai and Kim (1988) report evidence consistent with this hypothesis. They document a significantly positive average abnormal return to 236 acquirers. However, they find the positive returns only accrued to shareholders of acquiring firms in single-bidder tender offers.

The largely unexceptional returns to bidders on announcement of a takeover is not confined to US bidders. Australian studies by Walter (1984) and Bugeja and Walter (1995) also report similar results. It is significant then that Australian law on takeovers for listed companies is similar to the US regulations that promote auction-style contests.

Given that takeover regulations make it difficult for acquirers to capture any economic rent, it may be that returns on announcement of takeovers of private companies and of new capital expenditures provide a more accurate pointer to the returns we can expect when firms announce substantial asset acquisitions.

da Silva Rosa, Limmack, Supriadi and Woodliff (2004) document that successful bids for private targets in Australia are associated with significantly positive abnormal returns to bidders over the announcement period. They conclude that lower competition for private targets allows acquirers to capture more of the economic rent from takeovers. It is reasonable to presume that the market for substantial asset acquisitions is characterised by about the same level of competition as the market for unlisted private target companies.

Several studies examine the market impact of announcements of new capital expenditures and they all report a significantly positive market reaction, on average (McConnell & Muscarella 1985; Woolridge 1988; Szewczyk, Testsekos & Zantout 1996; Chen & Ho 1997; Chung, Wright & Charoenwong 1998). However, there is significant variation in market reaction and the two main explanations that have been advanced are the growth opportunities hypothesis and the free cash flow hypothesis. These are discussed next.

The growth opportunities hypothesis is that firms with higher growth opportunities are more likely to have positive NPV projects. Support for the hypothesis comes from Chung, Wright and Charoenwong (1998), who review the association between Tobin's q (3) and market reaction to capital expenditure decisions. They find that for firms with high Tobin's q, the decision to increase capital expenditure results in a significantly positive abnormal return but for firms with low Tobin's q, increased capital expenditure results in negative abnormal returns. Brailsford and Yeoh (2004) also examine the valuation effect of capital expenditure announcements by Australian listed firms. Using Signal-G data they identify 170 announcements of expenditure on physical assets made between July 1995 and December 1997. Brailsford and Yeoh find that acquiring firms earn insignificant announcement period returns. However, when they partition their sample on the basis of growth opportunities (proxied by the market-to-book ratio), they find support for the investment opportunities hypothesis. Firms in the highest market-to-book quartile experience, on average, significantly positive returns. On the other hand, firms in the lowest market-to-book quartile experience significantly negative abnormal returns. Brailsford and Yeoh conclude that the market reacts positively to the capital expenditure announcements made by firms with high growth opportunities.

Another variable that conditions the market reaction to acquisitions is 'free cash flow'. Jensen (1986) defines free cash flow as any 'cash flow in excess of what is required to fund positive NPV projects'. Thus, any acquisition by firms with free cash flow is, by definition, a negative NPV project. The presence of free cash flow is a manifestation of an agency problem because the excess cash is not returned to the shareholders. Brailsford and Yeoh (2004) report that firms with high level of cash flow and low growth opportunities experience a significant negative abnormal return of 2.002% when announcing capital expenditure.

2.3 Post-Acquisition Returns

Unlike the announcement period returns, the evidence on post-takeover performance is more consistent, with a clear majority of studies indicating a downwards drift in bidder performance (Jensen & Ruback 1983; Bradley, Desai & Kim 1983; Agrawal, Jaffe & Mandelker 1992; Loughran & Vijh 1997; da Silva Rosa, Izan, Steinbeck & Walter 2000).

Controlling for the now widely acknowledged biases in assessments of long-term sharemarket performance significantly ameliorates the secular decline in post-acquisition performance (Brown & da Silva Rosa 1998) and it is reasonable to conclude that the under-performance documented in earlier studies has been overstated. However, post-bid under-performance in the long-term remains predictable for some acquirers such those who offer shares rather than cash (da Silva Rosa, Izan, Steinbeck & Walter 2000) and 'glamour' acquirers (i.e. acquirers with a high market-to-book ratio) (Rau & Vermaelen 1998).

Agrawal, Jaffe and Mandelker (1992) observe that one possible explanation for the predictability of returns is the inability of the market to react immediately to merger events. Thus, abnormal long run post merger performance could 'reflect that part of NPV of merger to acquirer which is not captured by announcement period returns' (p. 1620). However, Agrawal, Jaffe and Mandelker find a negative relation between the market reaction to an announcement and subsequent performance. They conclude that market inefficiency is not a plausible explanation for the anomalous drift but it is not obvious why inefficiency is ruled out, particularly in light of the other evidence showing that some subsets of acquirers exhibit predictable drift in their abnormal returns. Notwithstanding this point, Agrawal et al.'s results indicate, if more evidence is needed, that it is not sufficient to review announcement period returns to estimate the total sharemarket consequences of acquisitive activity. A complete study must take into account long-term returns as well.

3. Data and Research Method

3.1 Sources

To identify our sample of ASX firms that made substantial asset acquisitions over the period 1 January 1990 to 31 December 1999, we rely primarily on the Signal G (4) database made available by the Securities Industry Research Centre of Asia-Pacific (SIRCA). Signal G is the most comprehensive source of company announcements in Australia. The operation of the Continuous Disclosure Regime in Australia means that the database ostensibly captures all material asset acquisitions and disposals by listed companies since it began collecting these in late 1991. The SDC Platinum database is used identify asset acquisitions announcements made prior to Signal G's operation. SDC Platinum database is a popular source of takeover data in Australia, and has been used by previous Australian studies e.g. da Silva Rosa, Limmack, Supriadi and Woodliff (2004).

The day end share prices for each of the listed firms in the sample are obtained from the Core Research Database (CRD) maintained by SIRCA. The daily share prices data are used to calculate the short run returns associated with market reaction to the announcement of asset acquisitions. The All Ordinaries Accumulated Index data are also obtained from CRD.

All other data are sourced from Aspect Financial. These include data on financial and operating firm characteristics such as total assets, market-to-book, and cash flow.

3.2 Sample Selection

To identify firms announcing substantial asset acquisitions, we search across the entire Signal G database using keywords (e.g. 'acquisitions', 'assets'). This keyword search yields over ten thousands announcements. Determining which announcements are relevant entails reading every announcement and we adopt the following procedure.

Several filters are applied. First, the transaction size must be greater or equal to A$10m. While the minimum deal size of A$10m is an arbitrary figure, it is imposed to ensure that only asset acquisitions of economic significance are included in the sample. (5) Our deal size is obtained from the Signal G acquisition announcements. Where the Signal G announcement does not contain a deal value, we use the deal value as reported in SDC.

Second, related announcements are identified and the date of the earliest related announcement is taken as the announcement date. Where both SDC and Signal G overlap, the earlier of the two dates (where they differ) is deemed the announcement date

Announcements which make passing reference to the possibility of implementing an acquisition program at some (unspecified) time in the future are excluded. We also exclude announcements relating to the option of acquiring an asset. However, announcements relating to the exercise of an option to acquire an asset are included in the sample at the date that the option is announced to be exercised. Announcements of contracts rewarded are also excluded.

Also excluded are announcements of the acquisition of an asset for the purpose of sale-lease back arrangement, as they are more in the nature of financing for the leasing firm rather than for the growth of the acquiring firm. Asset acquisitions by a subsidiary from the parent entity or vice versa are also excluded. The final sample ('full sample') contains 624 announcements of substantial asset acquisitions.

A potential problem with the full sample is that it incorporates a very broad definition of asset acquisition. A number of acquisitions are purchases of complete lines of business and include whole companies. Other announcements relate to the acquisition of shares. Given that the market may be differentially efficient in valuing these kinds of assets we apply two more filters to the full sample to derive a subset comprising pure asset acquisitions and a subset of pure physical asset acquisitions.

3.2.1 Pure Assets Sub-Sample This sub-sample excludes all acquisitions of financial assets, companies, businesses or divisions. Also excluded are assets acquired through mergers and takeovers. However, assets need not be tangible to be included in this sub-sample. Acquisitions of intangible assets such as brand names are also included. This results in a sub-sample of 333 announcements of substantial asset acquisitions by 184 firms.

3.2.2 Physical Assets Sub-Sample The physical assets sub-sample is a subset of the pure assets sub-sample. To be included in the physical assets sub-sample the announcement must relate to the purchase of tangible assets. Such acquisitions include purchases of plant, machinery, property, equipment, and shopping centre. Capital expenditures associated with the construction of a physical asset such as a building or new plant are excluded. Also excluded are acquisitions of assets via purchase of an interest in a joint venture. Announcements of permit, rights, distribution and acquisitions of interest in projects by mining companies are also excluded from the sample of physical asset acquisitions. However, an acquisition of a mine is regarded as an acquisition of a physical asset. The physical assets subsample comprises 225 announcements by 116 firms.

3.3 Short Run Abnormal Returns

The short run abnormal performance of our sample firms is estimated over several periods of days centred on the acquisition announcement date. We report average abnormal returns. The abnormal return to each sample firm for a given period is calculated by subtracting the return to the All Ordinaries Accumulation Index (AOAI) from the sample firm's return expressed as a price relative. The use of the AOAI as the proxy for expected returns is consistent with earlier Australian takeover studies (Bellamy & Lewin 1992; Bugeja & Walter 1995; da Silva Rosa, Izan, Steinbeck & Walter 2000).

The calculation of abnormal return (AR) is as follows (illustrated with reference to the period [-5,+5] days relative to the acquisition announcement date):

A[R.sub.i[-5,+5] = ([P.sub.5]/[P.sub.-6]) - ([I.sub.5]/[I.sub.-6])

where: [P.sub.5] = the share price of firm i at end of day 5;

[P.sub.-6] = the share price of firm i at end of day -6;

[I.sub.5] = is the All Ordinaries Cumulative Index at end of day 5; and

[I.sub.--6] = is the All Ordinaries Cumulative Index at end of day -6.

Given the evidence that market reaction to announcements of asset acquisitions are conditioned by firms' growth opportunities and free cash flow we sort our samples into quartiles on the basis of growth opportunities and free cash flow. Acquiring firms are also sorted into quartiles according to relative size of the transaction to facilitate the assessment of the economic significance of the results.

3.4 Long Run Abnormal Returns

Our estimates of abnormal returns over longer periods are calculated on the same basis as our short-run returns. (6) Note that this requires our sample firms to survive over a given period since we effectively estimate the average abnormal return to a buy-and-hold strategy. We favour the use of buy-and-hold returns because they more closely approximate the actual return to investors. To minimise the impact of survival bias, we ensure that the firms included in our market portfolio proxy also survive over the given period. For instance, Pacific Dunlop's abnormal return over the period spanning, say, June 1993 to June 1995 would be its buy-and-hold return over this period less the simple average buy-and-hold return to all firms with price data available over the full period. In short, we define the market portfolio as comprising all firms that with sufficient share price data available to calculate a buy-and-hold return over the period June 1993 to June 1995. (7)

The simple average buy-and-hold return to the market portfolio yields a systematically misleading estimate of the return to the market portfolio since small stocks are over-weighted. We therefore also calculate a value-weighted return to the market portfolio, using each firm's market capitalisation at the beginning of the period to weight its return. Finally, in order to control for the well-known systematic association between firm size and share returns, we calculate a third market return proxy: a size-decile return. The size-decile return for a given experimental sample firm is obtained by identifying all firms in the market portfolio with buy-and-hold returns over the given event-window, ranking them on their market capitalisation at the beginning of the event-window, sorting them into size-deciles (with decile one comprising the 10% smallest firms), and calculating the equally-weighted buy-and-hold return to the decile to which the sample firm belongs.

We test for significance using a re-sampling technique that relies on the construction of an empirical distribution of returns from 1002 control portfolios matched with the experimental portfolio on one or more attributes. To explain the construction of the 1002 control portfolios, and demonstrate the matching process, we describe the procedure assuming the event window of interest is the period spanning [-1,+1] months relative to the asset acquisition announcement month. The four steps below are repeated for each experimental sample bidder that has price data over the period [-1,+1] months relative to the announcement month.

Identify the set {LS} of all listed firms that survived over the period [-1,+1] months defined relative to the sample firm's announcement month.

Calculate the size (i.e. price per ordinary shares at the beginning of month [-1] multiplied by number of issued shares) of each firm in set {LS} and identify the size decile of the experimental firm.

Select a firm from the same size decile as the experimental firm, drawing randomly and sampling with replacement, and allocate it to the first 1001 control portfolios.

Repeat step 3 one thousand times, each time moving on to the next control portfolio in the sequence of 1001 control portfolios.

The performance of the experimental sample firms are then compared against the distribution of the performance measures of the 1002 control portfolios. An important advantage of this method is that significance tests are free of parametric assumptions since there is no reliance on formal test statistics.

4. Findings and Discussion

Section 4 reports the sharemarket performance of firms that engaged in substantial asset acquisitions over the period 1990-1999.

4.1 Preliminary Observations

Asset acquisitions are clearly frequent and economically significant events. Over the sample period, the aggregate dollar value invested in asset acquisitions by the sample firms exceeds A$69.6bn (see fig. 1). Table 1 reveals that of the 624 asset acquisitions with a deal value of equal or greater than A$10m, 511 occurred after 1995. This highlights the increasing popularity of asset acquisition as a form of external investment (although the figures are not deflated). This observation is consistent across the pure assets and physical asset sub-samples.

Not only do asset acquisitions occur frequently, they are also large in size. Table 2 reports an average dollar value of asset acquisition, for the full sample, of $112.654m. The large deal size is also present in the physical asset sub-sample, which recorded an average deal size over $88m and a median close to $40m. Table 2 also reports the relative deal value of the asset acquisition transactions. While the figures indicate that the average size of an acquisition is approximately half the size of the acquiring firm (the relative deal size for the full sample, pure assets sub-sample, and the physical asset sub-sample are 45.55%, 51.73% and 57.09% respectively), this result is probably driven by outliers (maximum relative deal size is over 44 times the size of the acquiring firm). The typical size of asset acquisition would probably be closer to 10% of the acquiring firm's size, as suggested by the median.

4.2 Announcement Period Returns

Table 3 reports abnormal announcement returns. Consistent with our prediction that the lower level of competition in the asset market allows the acquirer to capture more of the economic gains associated with an asset acquisition, the full sample of acquirers performed significantly better than the market on announcement of asset acquisitions across all event windows. This finding is consistent with those reported in the capital expenditure literature (Woolridge 1988; Chen & Ho 1997; Chung, Wright & Charoenwong 1998).

However, acquirers in the physical asset sub-sample experienced no exceptional performance over the announcement period. The significantly positive abnormal returns decreases from 1.56% (column 2), for the [-5,+5] day event window, to 0.74% (column 3) as you move from the full sample to the pure assets sub-sample, and disappears for the physical asset sub-sample (column 4). The unexceptional performance by the physical sub-sample may be related to the fact that this sub-sample consists largely of listed property trusts. These trusts frequently buy and sell assets as part of their normal business operation. Thus, the market may anticipate the asset acquisitions by these acquirers and this may explain the lack of significant market reaction on the announcement of acquisitions.

The higher positive abnormal performance for the full sample, compared to the pure assets sub-sample, suggests that some of the exceptional performance is driven by assets acquired indirectly through the acquisition of shares. This finding is consistent with the evidence in the takeover literature that cash offers for equity are view positively by the market (Jensen & Ruback 1983).

4.2.1 Differences in Returns to Firms in the Top and Bottom Quartiles Tables 4 and 5 report the mean abnormal returns to firms ranked in the top and bottom quartiles on relative size of acquisition (table 4) and on returns along with the difference in mean abnormal performance for firms acquiring substantial assets over the announcement period, comparing firms in the top and bottom quartiles for a particular characteristic. Abnormal performance is estimated over [-5,+5] days.

4.2.1.1 Relative Size of Acquisition The results in table 4 indicate that the market only reacts to announcements that have a potentially significant impact on acquirer's value (Brailsford & Yeoh 2004). Firms in the highest relative deal value quartile earn, on average, a significant positive return of 4.07% over the [-5,+5] day window, while firms in the other quartiles exhibit unexceptional performance. The difference in mean abnormal returns between acquirers in the top and bottom quartiles is significant at the 0.01 level (panel A of table 4). This result is robust across the different return windows.

However, this result does not hold for the physical asset sub-sample. Unlike the full sample, the only event window for which the highest relative deal value quartile display exceptional performance is the [-5,+5] day window. Further, panel B of table 4 shows that the difference in performance between the top and bottom quartiles is not significant. The absence of a positive relationship between abnormal returns and the relative size of asset acquisition for the physical asset sub-sample is probably due the fact that, on average, this sub-sample does not experience abnormal performance over the announcement period.

4.2.1.2 Growth Opportunities and Free Cash Flow As noted earlier, past research generates the expectation that firms with high growth opportunities will experience a significantly higher abnormal return on the announcement of substantial asset acquisition. We use the market-to-book value of total equity to proxy for growth opportunities.

Table 5 shows that acquirers in the top market-to-book quartile do not exhibit abnormal performance However, firms in the lowest market-to-book quartile have a significant positive abnormal return of 3.34% over the [-5, +5] day return window.

We investigate whether agency considerations condition the market's response to asset acquisitions by grouping firms on the basis of free cash flow. We use, as our proxy for free cash flow, undistributed cash flow defined as operating income before depreciation minus interest expense, taxes and dividends, scaled by total assets. (8)

Table 6 reveals that firms with highest level of undistributed cash flow significantly outperform the market over the various return windows. Over the [5,+5] day relative to announcement day, the full asset acquisition sample earn an average abnormal announcement return of 2.08%. However, a similar significant positive abnormal performance (2.01%) is also reported for acquirers with the lowest level of undistributed cash flow.

These results do not generate evidence to support or contradict the free cash flow hypothesis. It may be that undistributed cash flow is not a valid proxy for free-cash-flow. However, our results do indicate that a high level of cash flow is not necessarily penalised by the market. The positive abnormal returns of 2.08% over the [-5,+5] day window for the top cash flow quartile indicates that market does not treat asset acquisition by firms with a high cash flow level as non-value maximising behaviour.

Notwithstanding the above results, the agency problems of free cash flow problem are most likely to manifest in firms with a high level of undistributed cash flow and limited growth opportunities. Thus, testing for free cash flow effect requires the investigation of the interaction of undistributed cash level and growth opportunities.

Table 7 provides the mean BHARs for each of the four cells (each cell presenting a combination of cash flow and market-to-book level) over the [-5,+5] day window. Cell A of table 7 (i.e. the interaction of high cash flow level and low market-to-book ratio), is the most likely combination that would result in a free cash flow environment. High (low) cash flow firms are firms with a cash flow ratio equal or greater (less) than the sample mean. Whereas, high (low) growth firms is defined as firms with a market-to-book value of total assets of equal or greater (less) than 1.

Contrary to the prediction of the free cash flow hypothesis, the market does not punish acquisitions by firms with free cash flow. Firms with free cash flow (i.e. observations in cell A) outperform the market, with an average abnormal positive return of 4.99% over the [-5, +5] day window. Firms in cells C and D also experience exceptional performance. Perhaps the most surprising finding is the insignificant return in cell B (i.e. the combination of high growth prospect and high cash flow, which should in theory send a positive signal to the market as growth options are being exercised). (9)

The significantly positive abnormal returns for observations in cell A is consistent with the notion that investment by firms with free cash flow may minimise the loss that investors otherwise expect as a result of the high level of free cash flow (Lang, Stulz & Walking 1991).

4.3 Long-Run Returns." Pre-, Post- and Around Acquisition Announcement

The analysis below starts with a description of the principal features of table 8, which summarizes our findings. A commentary on the findings and discussion of their implications follows.

4.3.1 Principal Features of Table 8 Table 8 includes a summary of the average performance of the acquiring firms over various event-windows expressed in months relative to the month in which each acquisition was announced. For instance, the period [-1,+1] refers to the three month period centered on the announcement month. The cumulative average return to the sample firms is reported in column two, column three reports the average market (equal weighted) adjusted return to the sample firms (i.e. the average of each firm's return less the market portfolio return), column four reports the average market (value-weighted) adjusted return, and column five includes the average decile-adjusted return (i.e. the average of each firm's return less the return to the firms in its decile). The number in bold immediately below each return is the number of control portfolios, out of 1002, that had a higher return than the experimental sample portfolio. Given the arguments advanced in Brown and da Silva Rosa (1998) and elsewhere, we place most weight on the decile-adjusted results in our analysis, particularly over the wider event-windows where the biases associated with market-adjusted results become more severe.

The 10 event-windows listed in the first column of table 8 review performance over three distinct periods: the pre-announcement period, the announcement period and the post-announcement period. This categorization of periods is standard in event-studies of takeovers.

The last column in table 8 includes the number of sample firms with available share return data over each event-window.

Table 8 is divided into three panels. Panel A includes summary results for the whole sample of acquiring firms. Panel B reports summary results for firms that made acquisitions only of physical assets. Panel C contains results for firms that made acquisitions of all assets excluding physical assets. One reason for reviewing acquisitions of physical assets separately is that they are arguably easier to value and integrate into operations than non-physical assets. Given this assumption, it may be that the sharemarket is differentially efficient with respect to assessing the impact of physical and non-physical assets. As seen below, there is significant support for this proposition.

One important point needs to be borne in mind when reviewing the summary results. Brown and da Silva Rosa (1998) note because the market and decile portfolios consist of firms that have survived over the event-window, it is not possible to interpret the market and decile adjusted returns as an estimate of abnormal performance attributable to a feasible investment strategy. The 'look-ahead' bias imposed by conditioning on survival means that our measures of abnormal performance are not attainable in practice. Conditioning on survival facilitates tests relative performance because survival related biases are controlled but the abnormal returns should not be interpreted as an estimate of economic growth (or loss) attributable to the acquisition of substantial assets. For this reason, point estimates of abnormal performance over various event-windows are not much discussed below. Only relative performance is assessed. Where point estimates are discussed, the focus is relative difference rather than absolute value.

4.3.2 Pre-Announcement Period Performance-All Acquiring Firms A striking finding from panel A of table 8 is that the abnormal performance of acquirers of substantial assets in the pre-announcement period is similar to that documented in most takeover event-studies. That is, the acquirers outperform substantially over the months leading up to announcement of an acquisition.

A possible explanation is that investors anticipate the benefits of the expansionary program that the acquirers undertake and incorporate this into share prices. However, our (other, unreported) results indicate the acquirers' exceptional performance begins up to two years prior to the acquisition and so we conclude a more likely explanation is that acquirers of substantial assets are typically firms that are doing well and seek to build on that performance by expanding via acquisition of assets.

An interesting finding from panel A of table 8 is that the market (value-weighted) adjusted returns are consistently much higher than the market (equal-weighted) adjusted returns across all event-windows. This result is not surprising given that over two-thirds of our sample firms are drawn from size deciles nine and ten (i.e., the two deciles comprising the largest 20% of firms). The equally-weighted market portfolio returns tend to be driven by the extreme positive returns recorded by small firms (i.e. the well attested 'size-effect') and so market (equally weighted) adjusted returns to our sample of predominantly large firms are typically much lower than the corresponding market (value-weighted) adjusted returns. Panel A also shows that the decile-adjusted returns are in line with the market (value-weighted) adjusted returns. Again, this result is not unexpected given that the value-weighted return to the market portfolio will be overwhelmingly influenced by the returns to the firms in the top two deciles, which is where most of our sample firms reside.

A pertinent question that has been raised in the context of takeovers is whether managers of firms that have performed well are justified in seeking to continue to outperform by engaging in takeovers. As noted earlier, Roll (1986) has argued that managers of firms doing well may well be afflicted by hubris, that is, they may over-estimate their own abilities and over-reach by expanding too rapidly. The abnormal decline in merged firms' performance over the post-bid period that has been documented in many studies is consistent with Roll's hypothesis. However, takeover event-studies have been unable to control for the role played by takeover regulations in encouraging over-payment by acquiring firms. That is, given that takeover bids for listed companies have to be open bids, the regulations encourage a public auction-like process that increases the probability that the winning firm will have paid too much. Acquisitions of substantial assets are not subject to the stringent public auction-inducing regulations which are enforced for takeover bids and so the hubris hypothesis can be tested more rigorously. (10)

4.3.3 Announcement Period Performance All Firms Judging by the abnormal performance of the acquiring firms over the announcement period [-1,1] months (see panel A), the sharemarket does not seem to share the belief that the managers of these firms suffer from hubris. The 630 firms with available data over this period exhibit an average decile-adjusted return of around 2.40%, a performance better than the average decile-adjusted return to any of the 1001 control portfolios. In short, the sharemarket typically strongly endorses the acquisitive behaviour of the firms in our sample.

4.3.4 Post-Announcement Period Performance-All Firms An implication of the efficient market hypothesis is that it is not possible to earn persistent, predictable abnormal returns by selecting firms using a selection rule based on easily observable public information. Firms that have announced acquisitions of substantial assets arguably constitute an easily observable set. In this context, it is pertinent to observe that the sample firms' post-acquisition performance is only somewhat consistent with the efficient markets hypothesis. Over the three year period following asset acquisition, the sample firms earn an average of 16.56%, which is about 5% less than the 21.02% average return they posted in the 12-months leading to acquisition The extent of the reversal of fortune can be seen by comparing the market (equal weighted) adjusted performance of the sample firms in the pre- and post-announcement periods. The market (equal weighted) adjusted returns are heavily biased against revealing positive abnormal performance from our sample of predominantly large firms but even so the pre-announcement period average adjusted return is positive. However, it turns sharply over the post-announcement, with well over 90% of all control portfolios outperforming the experimental sample.

The poor relative performance of the experimental firms using market (equally weighted) adjusted returns contrasts dramatically with their performance when decile adjusted returns are assessed. Over all three years post-acquisition, the experimental sample firms earn an average decile-adjusted return higher than over 1000 of the 1002 control portfolios. This result, using our preferred measure of performance, directly contradicts the hubris hypothesis but poses a similar challenge to market efficiency as would evidence of underperformance on the same scale. If asset acquisitions are value increasing, on average, it is puzzling that the market should take such a long period of time to appreciate the magnitude of the gains. Of course, it is possible that our results are sample specific, however, we explore the possibility that the market finds it difficult to assess the value of asset acquisitions by reviewing separately the returns to acquisitions of purely physical assets and other assets.

4.3.5 Abnormal Performance-Acquisitions of Physical & Non-Physical Assets Panels B and C of table 8 show average sharemarket returns after dividing our sample of firms into those that made acquisitions comprising solely physical assets and those that made acquisitions that included intangible assets. Panel B reports the results for the acquirers of physical assets while panel C includes results for all other acquirers. Our categorization is based on the assumption that physical assets are easier to value than other assets and also easier to integrate into operations. If true, our premise implies that the sample of firms comprising acquirers of physical assets will display less evidence of market inefficiency in pricing, for instance, the post-announcement period performance of these firms will be unexceptional, indicating that the effects of the acquisitions were fairly and accurately incorporated into share prices around the announcement period.

The results in panels B and C of table 8 do not support for the above hypothesis. The relative performance of acquirers of both purely physical and other categories of asset acquisitions is exceptionally good over all three years.

The sharemarket does not appear differentially efficient in valuing tangible and intangible assets.

5. Summary and Conclusions

Asset acquisitions are frequent and economically significant occurrences that, on average, significantly enrich the shareholders of the acquiring firms, at least in the short term. Acquiring firms are typically firms that have performed exceptionally well in the months leading to announcement of acquisition of a substantial asset. The significantly positive market reaction to announcement of news of an asset acquisition indicates that investors generally agree acquisitions are positive-NPV investments. Contrary to what has been widely documented in the existing literature, firms' level of free cash flow is not a relevant variable in explaining the market reaction to the announcement of asset acquisitions. That is, asset acquisitions are evidently not seen by investors as being symptomatic of agency problems.

Over the long-term, in the post-announcement period, acquiring firms earn lower returns relative to those earned in the pre-acquisition performance but their relative performance remains exceptionally good, on average. If future studies analyzing different samples and time periods confirm the finding of continued exceptional performance in the post-acquisition period it will pose a challenge to the efficient market hypothesis. We do not find support for the proposition that sharemarket performance over the long-term is systematically related to the degree of difficulty that investors may experience in unbiasedly assessing the value of assets. Overall, we conclude that while acquisitions of substantial assets rival the takeover market in level of activity and economic substance, the outcomes from the market do not support the hypothesis that acquisitions are motivated by managerial hubris. The sharemarket evidence indicates acquisitions of substantial assets are value-increasing for equity investors, on average. We leave for future research investigation of the extent the auction-style bidding process inherent in the

Corporations Law's takeover regime contributes to over-payment in corporate takeovers and thus account for the anomalously post-acquisition returns to aquirers in takeover transactions.

(Date of receipt of final transcript: April, 2003. Accepted by Garry Twite, Special Issue Editor.)

Table 1
Sample Distribution by Year

Time series distribution of substantial asset acquisitions announced
each year from January 1990 to December 1999. The full sample consists
of all assets acquisitions by an Australian listed firm with a deal
value over $10 million. Acquisitions of financial assets, companies,
businesses or divisions were excluded to derive the pure assets
sub-sample. Assets acquired through merger and takeovers are also
excluded from this sub-sample. The physical assets sub-sample is a
subset of the pure assets sub sample and consists of acquisitions of
physical assets, including acquisition of plant, machinery, property,
equipment, mine, and shopping centre. Capital expenditures associated
with the construction of a physical asset such as a building or new
plant are excluded.

Year          All Acquisitions   Pure assets    Physical Assets
               (Full Sample)     (Sub-Sample)    (Sub-Sample)

1990                  7                0                0
1991                  8                0                0
1992                  9                3                3
1993                 17                9                6
1994                 30               13                9
1995                 42               20               11
1996                106               60               37
1997                130               79               46
1998                121               76               66
1999                154               73               47
Total               624              333              225
No of Firms         325              184              116

Table 2
Descriptive Statistics

Deal value and relative deal size of asset acquisitions. Deal Value is
the total value offered for the assets in millions. It includes cost
incidental to the acquisition such as contract costs. Relative Deal
Size is calculated by dividing the deal value by the acquiring firm's
book value of total assets as obtained from the financial statements
immediately prior to announcement date.

                      All Acquisitions   Pure Assets   Physical Assets

Deal Value ($'000s)        N = 618          N = 329         N = 224
  Mean                     112,654           84,691          88,258
  StdDev                   245,259          169,469         190,234
  Min                       10,000           10,000          10,000
  Median                    40,000           36,650          40,000
  Max                    2,800,000        2,200,000       2,200,000

Relative Deal Size         N = 520          N = 278         N = 189
Mean                        0.4555           0.5173          0.5709
StdDev                      2.2006           2.8671          3.4331
Min                         0.0001           0.0001          0.0001
Median                      0.0872           0.0974          0.0955
Max                        44.7958          44.7958         44.7958

Table 3
Announcement Period Returns

Abnormal (market-adjusted) buy-hold announcement returns across
various trading day windows expressed relative to the earliest
announcement date for substantial asset acquisitions. Student
t -statistics, indicating whether the mean return is significantly
different from zero, and 'n', the total number of observations in
each sample, are reported in the parentheses respectively.

Return      All Acquisitions     Pure assets     Physical Assets
Window       (Full Sample)      (Sub-Sample)       (Sub-Sample)

(-1, 0)        0.6150 ***         0.4488 **           0.1847
            (3.51, n = 622)    (1.98, n = 332)   (0.94, n = 224)
(-1, 1)        1.0845 ***        0.8513 ***           0.3326
            (4.57, n = 616)    (2.63, n = 330)   (1.05, n = 222)
(-2, 2)        1.1989 ***         0.5697 *            0.2218
            (4.56, n = 615)    (1.74, n = 330)   (0.66, n = 222)
(-5, 5)        1.5248 ***         0.7137 *            0.4709
            (4.23, n = 610)    (1.65, n = 325)   (1.03, n = 217)
(-5, 1)        1.5931 ***        1.0234 ***         0.8226 **
            (5.35, n = 613)    (2.71, n = 328)   (2.16, n = 220)
(-10, 1)       1.8710 ***         0.8524 *            0.4972
            (5.17, n = 611)    (1.95, n = 327)   (1. 10, n = 219)
(-10, 10)      1.7934 ***          0.5301             0.2039
             (4.27, n -608)    (1.00, n = 324)   (0.34, n = 216)

Note: *** Significant at the 0.01 level;

** Significant at the 0.05 level; and

* Significant at the 0.10 level.

Table 4
Announcement Returns by Relative Deal Value

Mean abnormal (market-adjusted) buy-hold announcement returns
calculated over various trading day windows relative to the earliest
announcement date for the full sample (panel A) and physical asset
sub-sample (panel B), sorted by relative deal value quartiles.

Return Windows                               (-1,0)        (-1,1)

Panel A: All Acquisitions (Full Sample)

Highest Relative deal Value Quartile (N)   129           126
  Mean BHAR (%)                              1.596 ***     2.777 ***
  StdDev                                     6.469         9.256
  t-statistic                                2.8016        3.3638
Second Quartile (N)                        130           130
  Mean BHAR (%)                              0.393         0.854 **
  StdDev                                     3.388         4.950
  t-statistic                                1.323         1.969
Third Quartile (N)                         129           129
  Mean BHAR (%)                             -0.004         0.433
  StdDev                                     2.794         3.620
Lowest Relative Deal Value Quartile (N)    130           129
  Mean BHAR (%)                              0.237         0.329
  StdDev                                     2.223         2.778
  t-statistic                                1.217         1.347
Difference in Mean BHARs Between the Highest and Lowest Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]      1.358 **      2.444 ***
  t-statistic                               -2.256        -2.842

Panel B: Physical Asset Acquisitions (Sub-Sample)

Highest Relative Deal Value Quartile (N)    47            45
  Mean BHAR (%)                              1.028 *       1.523
  StdDev                                     4.215         8.636
  t-statistic                                1.672         1.183
Second Quartile (N)                         48            48
  Mean BHAR (%)                             -0.372        -0.389
  StdDev                                     2.088         2.484
  t-statistic                               -1.233        -1.084
Third Quartiles (N)                         46            46
  Mean BHAR (%)                             -0.484 *      -0.228
  StdDev                                     1.749         2.618
  t-statistic                               -1.876        -0.591
Lowest Relative Deal Value Quartile (N)     47            47
  Mean BHAR (%)                              0.469         0.590
  StdDev                                     2.446         3.238
  t-statistic                                1.316         1.249
Difference in Mean BHARs Between the Highest and Lowest Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]      0.559         0.933
  t-statistic                               -0.786        -0.680

Return Windows                             (-2,2)        (-5,5)

Panel A: All Acquisitions (Full Sample)

Highest Relative deal Value Quartile (N)   126           124
  Mean BHAR (%)                              3.410 ***     4.074 ***
  StdDev                                    10.292        12.370
  t-statistic                                3.7188        3.6675
Second Quartile (N)                        130           129
  Mean BHAR (%)                              0.611         1.220
  StdDev                                     6.105        10.678
  t-statistic                                1.141         1.297
Third Quartile (N)                         129           129
  Mean BHAR (%)                              0.682 *       0.874 *
  StdDev                                     4.512         5.942
Lowest Relative Deal Value Quartile (N)    129           128
  Mean BHAR (%)                              0.208        -0.139
  StdDev                                     3.336         4.653
  t-statistic                                0.707        -0.338
Difference in Mean BHARs Between the Highest and Lowest Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]      3.202 ***     4.213 ***
  t-statistic                               -3.326        -3.557

Panel B: Physical Asset Acquisitions (Sub-Sample)

Highest Relative Deal Value Quartile (N)    45            43
  Mean BHAR (%)                              1.704         2.442 *
  StdDev                                     8.214         9.363
  t-statistic                                1.392         1.710
Second Quartile (N)                         48            47
  Mean BHAR (%)                             -1.259 **     -1.622
  StdDev                                     3.699         7.105
  t-statistic                               -2.358        -1.565
Third Quartiles (N)                         46            46
  Mean BHAR (%)                             -0.137         0.478
  StdDev                                     3.518         4.444
  t-statistic                               -0.265         0.730
Lowest Relative Deal Value Quartile (N)     47            46
  Mean BHAR (%)                              0.655         0.677
  StdDev                                     4.140         4.898
  t-statistic                                1.084         0.938
Difference in Mean BHARs Between the Highest and Lowest Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]      1.049         1.765
  t-statistic                               -0.768        -1.103

Note: *** Significant at the 0.01 level;

** Significant at the 0.05 level; and

* Significant at the 0.10 level.

Table 5
Announcement Returns by Growth Opportunities

Mean abnormal (market-adjusted) buy-hold announcement returns
calculated over various trading day windows relative to the
earliest announcement date for the full sample (panel A)
and physical asset sub-sample (panel B). sorted by
market-to-book quartiles.

            Return Windows                  (-1,0)       (-1,1)

Panel A. All Acquisitions (Full Sample)

Highest Market-to-Book Quartile (N)       106          106
  Mean BHAR (%)                             0.322        0.326
  StdDev                                    3.083        4.475
  t-statistic                               1.077        0.749
Second Quartile (N)                       106          104
  Mean BHAR (%)                             0.473        0.917  **
  StdDev                                    3.265        4.047
  t-statistic                               1.490        2.310
Third Quartile (N)                        106          105
  Mean BHAR (%)                             0.164        1.023  **
  StdDev                                    3.083        5.270
  t-statistic                               0.548        1.990
Lowest Market-to-Book Quartiles (N)       105          104
  Mean BHAR (%)                             1.170  *     2.472  ***
  StdDev                                    6.127        8.541
  t-statistic                               1.958        2.951
Difference in Mean BHARs Between the Highest and Lowest
    Market-to-Book Ratio Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]    -0.848       -2.146  **
  t-statistic                               1.268        2.275

Panel B. Physical Asset Acquisitions (Sub-Sample)

Highest Market-to-Book Quartiles (N)       36           36
  Mean BHAR (%)                            -0.330       -0.945
  StdDev                                    2.722        3.918
  t-statistic                              -0.728       -1.447
Second Quartile (N)                        36           34
  Mean BHAR (%)                            -0.006       -0.140
  StdDev                                    1.713        1.890
  t-statistic                              -0.023       -0.433
Third Quartile (N)                         36           36
  Mean BHAR (%)                            -0.248       -0.119
  StdDev                                    1.999        2.674
  t-statistic                              -0.743       -0.268
Lowest Market-to-Book Quartiles (N)        34           34
  Mean BHAR (%)                             0.686        2.102
  StdDev                                    4.500        9.231
  t-statistic                               0.8884       1.3276
Difference in Mean BHARs Between the Highest
    and Lowest Market-to-Book Ratio Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]    -1.0152      -3.0465 *
  t-statistic                               1.1345       1.7790

            Return Windows                  (-2,2)        (-5,5)

Panel A. All Acquisitions (Full Sample)

Highest Market-to-Book Quartile (N)       106           105
  Mean BHAR (%)                             1.323  **     1.345
  StdDev                                    5.785         8.561
  t-statistic                               2.354         1.610
Second Quartile (N)                       104           104
  Mean BHAR (%)                             1.380  **     1.184
  StdDev                                    6.941         9.361
  t-statistic                               2.027         1.290
Third Quartile (N)                        105           105
  Mean BHAR (%)                             0.503         0.691
  StdDev                                    5.865         6.442
  t-statistic                               0.879         1.099
Lowest Market-to-Book Quartiles (N)       104           101
  Mean BHAR (%)                             2.090  **     3.340  ***
  StdDev                                    8.652        11.956
  t-statistic                               2.463         2.808
Difference in Mean BHARs Between the Highest and Lowest
    Market-to-Book Ratio Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]    -0.767        -1.995
  t-statistic                               0.753         1.372

Panel B. Physical Asset Acquisitions (Sub-Sample)

Highest Market-to-Book Quartiles (N)       36            35
  Mean BHAR (%)                            -0.459        -1.463
  StdDev                                    5.679         7.993
  t-statistic                              -0.485        -1.083
Second Quartile (N)                        34            34
  Mean BHAR (%)                            -0.619        -0.061
  StdDev                                    3.094         3.807
  t-statistic                              -1.167        -0.094
Third Quartile (N)                         36            36
  Mean BHAR (%)                            -0.454        -0.568
  StdDev                                    3.304         4.328
  t-statistic                              -0.824        -0.788
Lowest Market-to-Book Quartiles (N)        34            31
  Mean BHAR (%)                             1.387         2.450
  StdDev                                    8.993         9.503
  t-statistic                               0.8991        1.4354
Difference in Mean BHARs Between the Highest
    and Lowest Market-to-Book Ratio Quartiles
  BHA[R.sub.highest]--BHA[R.sub.lowest]    -1.8454       -3.9124 *
  t-statistic                               1.0198        1.7973

Note: Market-to-Book is the firms' market capitalisation three
months prior to announcement divided by the book value of equity
for the last financial period prior to announcement.

*** Significant at the 0.01 level;

** Significant at the 0.05 level; and

* Significant at the 0.10 level.

Table 6
Announcement Returns by Undistributed Cash Flows

Mean abnormal (market-adjusted) buy-hold announcement returns and the
differences between mean BHARs calculated over various trading day
windows relative to the earliest announcement date for the full sample
(panel A) and physical asset sub-sample (panel B), sorted by
undistributed cash flow quartiles. Undistributed cash flow is defined
as operating income before depreciation minus interest expense, taxes,
and dividends, scaled by total assets.

            Return Windows                  (-1,0)        (-1,1)

                                         Panel A. All Acquisitions
                                               (Full Sample)

Highest Undistributed CF Quartiles (N)   130           128
  Mean BHAR (%)                            0.6694 **     1.1904 ***
  StdDev                                   3.9263        4.9692
  t-statistic                              1.9439        2.7103
Second Quartile (N)                      129           127
  Mean BHAR (%)                            0.4157        1.3888 ***
  StdDev                                   3.0492        4.8159
  t-statistic                              1.5484        3.2499
Third Quartile (N)                       130           130
  Mean BHAR (%)                            0.4810        1.1788 ***
  StdDev                                   3.8091        5.3327
  t-statistic                              1.4399        2.5204
Lowest Undistributed CF Quartiles        128           128
  Mean BHAR (%)                            0.6893        0.6760
  StdDev                                   5.3257        7.5594
  t-statistic                              1.4643        1.0117
Difference in Mean BHARs Between the
    Highest and Lowest Undistributed
    CF Ratio Quartiles
  BHA[R.sub.highest]-BHA[R.sub.lowest]    -0.0199        0.5144
  t-statistic                              0.0340       -0.6433

                                          Panel B. Physical Asset
                                         Acquisitions (Sub-Sample)

Highest Undistributed CF Quartiles (N)    45            43
  Mean BHAR (%)                           -0.3399       -0.0547
  StdDev                                   2.4946        2.7801
  t-statistic                             -0.9141       -0.1290
Second Quartile (N)                       47            47
  Mean BHAR (%)                            0.2244        0.6696
  StdDev                                   2.8131        3.7174
  t-statistic                              0.5468        1.2348
Third Quartile (N)                        47            47
  Mean BHAR (%)                            0.2386        0.0251
  StdDev                                   2.0066        2.6221
  t-statistic                              0.8153        0.0657
Lowest Undistributed CF Quartiles (N)     46            46
  Mean BHAR (%)                            0.4143        0.7035
  StdDev                                   3.7223        8.2792
  t-statistic                              0.7549        0.5763
Difference in Mean BHARs Between the
    Highest and Lowest Undistributed
    CF Ratio Quartiles
  BHA[R.sub.highest]-BHA[R.sub.lowest]    -0.7542       -0.7582
  t-statistic                              1.1377        0.5867

            Return Windows                  (-2,2)         (-5,5)

                                         Panel A. All Acquisitions
                                               (Full Sample)

Highest Undistributed CF Quartiles (N)   128            128
  Mean BHAR (%)                            1.1051 *       2.0755 **
  StdDev                                   6.9539         9.8996
  t-statistic                              1.7979         2.3720
Second Quartile (N)                      127            127
  Mean BHAR (%)                            1.5678 ***     0.7560
  StdDev                                   5.4206         7.4029
  t-statistic                              3.2595         1.1509
Third Quartile (N)                       130            130
  Mean BHAR (%)                            1.3555 **      1.3498 **
  StdDev                                   6.7254         7.0834
  t-statistic                              2.2980         2.1727
Lowest Undistributed CF Quartiles        128            124
  Mean BHAR (%)                            0.8714         2.0135 **
  StdDev                                   7.5440        11.4640
  t-statistic                              1.3068         1.9559
Difference in Mean BHARs Between the
    Highest and Lowest Undistributed
    CF Ratio Quartiles
  BHA[R.sub.highest]-BHA[R.sub.lowest]     0.2336         0.0620
  t-statistic                             -0.2576        -0.0459

                                          Panel B. Physical Asset
                                         Acquisitions (Sub-Sample)

Highest Undistributed CF Quartiles (N)    43             43
  Mean BHAR (%)                           -0.8095        -1.1417
  StdDev                                   5.0407         6.5504
  t-statistic                             -1.0530        -1.1429
Second Quartile (N)                       47             47
  Mean BHAR (%)                            0.7341         1.4366
  StdDev                                   4.8870         6.3207
  t-statistic                              1.0298         1.5582
Third Quartile (N)                        47             46
  Mean BHAR (%)                           -0.3431         0.4870
  StdDev                                   3.9498         5.8759
  t-statistic                             -0.5955         0.5622
Lowest Undistributed CF Quartiles (N)     46             43
  Mean BHAR (%)                            1.0914         0.8888
  StdDev                                   6.8950         8.3578
  t-statistic                              1.0736         0.6973
Difference in Mean BHARs Between the
    Highest and Lowest Undistributed
    CF Ratio Quartiles
  BHA[R.sub.highest]-BHA[R.sub.lowest]    -1.9009        -2.0305
  t-statistic                              1.4915         1.2539

Note: *** Significant at the 0.01 level;

** Significant at the 0.05 level; and

* Significant at the 0.10 level.

Table 7
Announcement Returns by Market-to-Book and Undistributed Cash Flow

Mean abnormal (market-adjusted) buy-hold announcement returns for the
full sample (panel A) and the physical asset sub-sample (panel B). The
value in each cell is mean BHAR over the (-5,+5) day window relative
to the announcement date. Observations are grouped according to the
market-to-book ratio and the scaled cash flow ratio (cash flow divided
by total assets). Low cash flow firms are firms with cash flow ratio
of less than the sample median. Firms with a market-to-book ratio of
less than 1 is classified as low market-to-book firms. Student
t-statistics, indicating whether the mean return is significantly
different from zero, are reported in the parentheses along with the
number of observations.

                    Market-to-Book

Cash Flow         Low               High

Panel A. All Acquisitions (Full Sample)

High            4.988 *            0.975
             (1.861, n=25)     (1.545, n=180)
Low            3.406 **          1.176701 *
            (2.312, n = 65)   (1.746, n = 139)

Panel C. Physical Asset Acquisitions (Sub-Sample)

High             1.934             -1.092
            (1.129, n = 19)   (-1.171, n = 49)
Low              0.835             -0.187
            (0.466, n = 26)   (-0.239, n = 40)

Note: Market-to-Book ratio is defined in this table as MB3 (calculated
as firms' market capitalisation three months prior to announcement
divided by the book value of equity for the last financial period
prior to announcement).

*** Significant at the 0.01 level;

** Significant at the 0.05 level; and

* Significant at the 0.10 level.

Table 8
Long-Run Abnormal Returns

Abnormal (market-adjusted) buy-hold announcement returns for various
trading windows expressed in months relative to the announcement
month. Cumret is the average buy-and-hold return to the sample firms.
Emktadj is the average market adjusted return to sample of firms,
where the market portfolio is equal-weighted. Vmktadj is the average
market adjusted return to sample of firms, where the market portfolio
is value-weighted. Decadj is the average decile adjusted return to
sample firms, where the return to the decile to which the sample firm
belongs is calculated on an equal-weighted basis. Sample No is the
number of sample firms with available data over the event-window. The
number in bold below each return is the number of control portfolios,
out of 1002, that have a mean return higher than the experimental
sample portfolio.

Return Window   Cumret    Emktadj    Vmktadj   Decadj   Sample No

                    Panel A: All Acquisitions (Full Sample)

[-12,-1]         21.02%      9.19%    16.99%   14.74%      587
                  0          1         0        0
[-1,+1]           2.86%      0.53%     2.56%    2.40%      630
                  0        175         0        0
[0,+12]           5.25%    -10.06%     1.76%    1.49%      615
                242       1002       198        0
[0,+24]           9.27%     -8.06%     2.97%    2.17%      577
                227       1000       151        1
[0,+36]          16.56%     -5.95%     6.76%    5.07%      519
                122        939        71        0

                   Panel B: Physical Acquisitions (Sub-Sample)

[-12,-1]         13.70%      6.45%    10.00%    9.67%      204
                  3         22         3        1
[-1,+1]           1.87%     -0.23%     1.47%    1.55%      221
                 66        580        92        2
[0,+12]           8.65%     -7.59%     5.38%    4.25%      217
                142        991        99        1
[0,+24]          12.50%     -8.52%     6.05%    3.31%      204
                265        942       157       17
[0,+36]          16.36%     -4.17%     6.54%    3.94%      188
                269        703       180       52

                       Panel C: Other (Non-Physical)
                          Acquisitions (Sub-Sample)

(-12,-1)         25.10%     10.78%    20.90%   17.64%      382
                  0          1         0        0
(-1,+1)           3.43%      0.98%     3.18%    2.90%      408
                  0        136         0        0
(0,+12)           3.38%    -11.41%    -0.22%    0.04%      397
                473       1002       493        0
(0,+24)           7.64%     -7.74%     1.45%    1.71%      371
                303        986       308        1
(0,+36)          16.82%     -6.88%     7.04%    5.86%      330
                125        910       103        3

(1.) Over the period 1995-1999, there were a total of 572 asset acquisitions made by ASX-listed companies where the transaction value is equal or greater than A$10m, compared to 193 takeovers of publicly listed targets. While the total value of asset acquisitions did not necessarily exceed the value of assets acquired in takeovers, the comparison shows that asset acquisitions are frequent and economically significant events.

(2.) For instance, Rau and Vermaelen (1998) show that glamour bidders under-perform value bidders, after controlling for the effects of method of payment and form of acquisition. In Australia, Da Silva Rosa, Izan, Steinbeck and Walter (2000) show that acquirers who offer equity exhibit negative post-acquisition returns in the long-term while acquirers who offer cash experience unexceptional returns. This is after controlling for the biases known to affect estimates of long-run performance.

(3.) Tobin's q is generally defined as the ratio of market value of firm's assets to their replacement costs. It is a proxy for growth opportunities. Firms with high Tobin's q are presumed to have high growth opportunities.

(4.) Signal G is the data feed provided by the Australian Stock Exchange (ASX) to facilitate the communication of company announcements and disclosure to brokers and investors. It is an instrument to keep the market informed of company specific developments that are value-relevant. Company announcements fed through Signal G include takeover announcements, announcements of assets acquisitions and disposals, dividends, and periodic reports.

(5.) A minimum deal value of $10m is consistent with that employed in numerous studies.

(6.) Our approach is modelled on Brown and da Silva Rosa (1998)

(7.) Our method does not yield an estimate of the abnormal return to a feasible investment strategy since our experimental samples and market portfolios incorporate a deliberate 'look-ahead' bias. However, this does not matter since our principal interest is not in estimating precise point estimates of abnormal return but in detecting if our experimental sample firms exhibit significant abnormal performance. Brown and da Silva Rosa (1998) discuss this issue in more depth.

(8.) Lang, Stulz and Walkling (1991) use this measure, among others, as a proxy for free cash flow. They report their results do not alter substantially when different cash flow measures are used.

(9.) We also performed another sensitivity analysis, which categorised the observations based on sample median for both cash flow and market-to-book ratio. The results are essentially the same.

(10.) Notwithstanding this point, it should be borne in mind that there is no compulsion on firms to make takeover bids. Further, it is not obvious why the market should take so long to penalise the hubristic managers.

References

Agrawal, A., Jaffe, J.F. & Mandelker, G.N. 1992, 'The post-merger performance of acquiring firms: A re-examination of an anomaly', Journal of Finance, vol. 47, pp. 1605-21.

Asquith, P. 1983, 'Merger bids, uncertainty, and stockholder returns', Journal of Financial Economics, vol. 11, pp. 51 83.

Bellamy, D. & Lewin, W. 1992, 'Corporate takeovers, method of payment, and bidding firms; shareholders returns: Australian evidence', Asia Pacific Journal of Management, vol. 9, pp. 137-49.

Bishop, S., Dodd, P. & Officer, R.R. 1986, Australian Takeovers'." The Evidence, The Centre for Independent Studies, St. Leonards, NSW.

Bradley, M., Desai, A. & Kim, E.H. 1988, 'Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms', Journal of Financial Economics, vol. 21, pp. 3-40.

Brailstford, T. and & Yeoh, D. 2004, 'Agency costs and capital expenditure', Journal of Business, vol. 77, no. 2, pp. 223-57.

Brown, P. & da Silva Rosa, R. 1998, 'Research method and the long-run performance of acquiring firms', Australian Journal of Management, vol. 23, pp. 23-38.

Bugeja, M. & Walter, T. 1995, 'An empirical analysis of some determinants of the target shareholder premium in takeovers', Accounting and Finance, vol. 35, pp. 33-60.

Chen, S.S. & Ho, K.W. 1997, 'Market response to product-strategy and capital expenditure announcements in Singapore: Investment opportunities and free cash flow', Financial Management, vol. 26, pp. 82-8.

Chung, K., Wright, P. & Charoenwong, C. 1998, 'Investment opportunities and market reaction to capital expenditure decisions', Journal of Banking and Finance, vol. 22, pp. 41-60.

da Silva Rosa, R. 1994, 'The economic consequences of Australian corporate takeovers', PhD thesis, Department of Accounting and Finance, University of Western Australia.

da Silva Rosa, R., Izan, H.Y., Steinbeck, A. & Walter, T., 2000, 'The method of payment decision in Australian takeovers: An investigation of causes and effects', Australian Journal of Management, vol. 25, pp. 67-94.

da Silva Rosa, R., Limmack, R., Supriadi, S. & Woodliff. D. 2004, 'The equity wealth effects of method of payment in takeover bids for privately held firms', Australian Journal of Management, vol. 29, pp.93-110.

Dodd, P. 1980, 'Merger proposals, management discretion, and stockholder wealth', Journal of Financial Economics, vol. 8, pp. 105-37.

Fama, E. 1998, 'Market efficiency, long-term returns, and behavioural finance', Journal of Financial Economics, vol. 49, pp. 283-306.

Jarrell, G.A., Brickley, J.A. & Netter, J.M. 1988, 'The market for corporate control: The empirical evidence since 1980', Journal of Economic Perspective, vol. 23, pp. 49-67.

Jensen, M.C. 1986, 'Agency costs of free cash flow, corporate finance and takeovers', American Economic Review, vol. 76, pp. 323-9.

Jensen, M.C. & Ruback, R.S. 1983, 'The market for corporate control: The scientific evidence', Journal of Financial Economics, vol. 11, pp. 5-50.

Lang, L., Stulz, R. & Walkling, R. 1991, 'A test of free cash flow hypothesis: the case of bidder returns', Journal of Financial Economics, vol. 29, pp. 315-36.

Loughran, T. & Vijh, A.M. 1997, 'Do long-term shareholders benefit from corporate acquisitions?' Journal of Finance, vol. 5 l, pp. 1765-90.

McConnell, J. & Muscarella, C. 1985, 'Corporate capital expenditure decisions and the market value of the firm', Journal of Financial Economics, vol. 14, pp. 339-422.

Morck, R., Shleifer, A. & Vishny, R. 1990, 'Do managerial objectives drive bad acquisitions?' Journal of Finance, vol. 45, pp. 31-48.

Rau, P.R. & Vermaelen, T, 1998, 'Glamour, value and the post-acquisition performance of acquiring firms', Journal of Financial Economics, vol. 49, pp. 223-53.

Roll, R. 1986, 'The hubris hypothesis of corporate takeovers', Journal of Business, vol. 59, pp. 197-216.

Schwert, G.W. 1996, 'Markup pricing in mergers and acquisitions', Journal of Financial Economics, vol. 41, pp. 153-92.

Szewczyk, S., Testsekos, G. & Zantout, Z. 1996, 'The valuation of corporate R&D expenditures: Evidence from investment opportunities and free cash flow', Financial Management, vol. 25, pp. 105-10.

Walter, T.S. 1984, 'Australian Takeovers: Capital market efficiency and shareholder risk and returns', Australian Journal of Management, vol. 9, pp. 63-118.

Woolridge, R. 1988, 'Competitive decline and corporate restructuring: Is a myopic stock market to blame?' Journal of Applied Corporate Finance, vol. 1, pp. 26-36.

Raymond da Silva Rosa ([dagger]) Thuy Nguyen ([section]) Terry Walter ([double dagger])

([dagger]) UWA Business School, University of Western Australia, 35 Stirling Highway, Nedlands, WA 6009. Email: [e-mail].

([section]) Macquarie Bank, 20 Bond St, Sydney, NSW, 2000. Email: [e-mail]

([double dagger]) School of Banking and Finance, The University of New South Wales, Sydney, 2052 and Capital Markets Cooperative Research Centre Limited, Level 2, 9 Castlereagh Street, Sydney NSW 2000 Email: [e-mail]