The Relative Explanatory Power of the Stock Market, Economic Development, and Interest Rates on the Bid Premium in Corporate Takeovers

The Relative Explanatory Power of the Stock Market, Economic

Development, and Interest Rates on the Bid Premium in Corporate

Takeovers

MASTER THESIS BY HANS HORTENSIUS

ABSTRACT

This paper discusses the explanatory power of three aggregate waves on the bid premium in corporate takeovers and if there are patterns over time. The three waves discussed in this paper are the overall stock market valuation wave, the macro-economic development (as measured by industrial production), and the interest development (as measured by the Euro Aggregate Corporate Bond Yield). Acquisition data used in this paper is obtained from four major European countries. The results indicate that the stock market and the interest rates are positively related to the bid premium, whereas economic development is negatively related to the bid premium. Moreover, economic development seems to be the most significant driver of the bid premium.

growth, and interest rates and how this is likely to fluctuate over uniform time intervals. Though the three mentioned waves have a certain degree of interdependency, they do not necessarily occur exactly simultaneously. Moreover, the general patterns of the waves may be more or less the same, the magnitude and the levels of the ups and downs may be different, resulting in different influences on the bid premium. Furthermore, it might be argued that the effects of the aggregate waves on takeover frequency are different than the effects on the takeover premium in particular. This paper studies whether there are significant differences in the explanatory power of these three aggregate waves on the takeover premium and in their related pattern over time. Therefore, this paper is likely to provide more thorough understanding about the aggregate causes of the bid premium.

In order to assess the influences of the three waves on the bid premium, a cross-sectional analysis is performed comprising individual bid premiums of 984 public takeovers in France, Germany, the Netherlands, and the United Kingdom in the period January 1999 till December 2008. These bid premiums are regressed against monthly levels of the stock market (i.e. the main indices of the respective countries), economic development (as measured by industrial production), and interest rates (as measured by the corporate bond yield), one month before the announcement date. In order to assess whether the results hold on a uniform time interval and to assess if there are patterns over time, a time-series analysis is performed comprising monthly average bid premiums of the same dataset. These average bid premiums are also regressed against monthly levels of the three independent variables, one month before the announcement date. The time series analysis provides insight into time patterns in the explanation of the variables.

This paper finds that the stock market is likely to have a positive impact on the bid premium, economic development is likely to be negatively related to the bid premium, significantly, and interest rates are positively related to the bid premium. The results of the latter, however, are less significant that the former two relationships. Moreover, economic development seems to be the most important driver of the bid premium. Finally, it is likely that the explanatory power of the stock market occurs in waves of approximately three years, the explanatory power of economic development in waves of 18 months, and interest rates in waves of 13 months.

I. Literature review and hypothesis development

In order to assess the explanatory power of the three waves on the takeover premium, this section starts with a discussion of the literature concerning the bid premium in general. Next, literature concerning the three aggregate waves and corporate takeovers will be discussed along with their assumed relationship with the bid premium. Further, this section presents a literature review concerning the relationships between the three waves and the consequences of these interrelationships on the bid premium. Finally, with the literature review in mind, a set of hypotheses is developed in this section that will be tested in the rest of the paper.

In the field of corporate finance, the bid premium itself is a widely discussed topic. Many studies emphasize the fact that target shareholder wealth is increased significantly during corporate takeovers (for example, see De Jong et al, 2007). In a study concerning large intra-European takeover bids, Goergen and Renneboog (2004) present several factors that affect the takeover premium significantly. They argue that, amongst others, the method of payment (stock or cash), the target market-to-book ratio, and the geographic scope (domestic or cross-border acquisition) all affect the premium paid in corporate takeovers.

(2004) posit that when the target market-to-book ratio is higher, the bid premium tends to be higher as well. They argue that a high market-to-book ratio is associated with increased growth opportunities of the target, resulting in a higher bid premium. However, in literature, there is no consensus about this relationship. Dong et al (2006), for example, find that a higher target price-to-book ratio is associated with a lower bid premium. Their argument is that there is more room for improvement within lower valued targets, resulting in a higher premium offered. Moreover, as argued by Simonyan (2005), targets know their true value and are induced to sell around the true value of their company, which has an increasing effect on the bid premium during times of low stock valuation. These arguments and findings are strong. Thirdly, Goergen and Renneboog (2004) find that the bid premium tends to be lower when the target is from a different country than the acquirer, i.e. cross-border bids are likely to generate lower takeover premiums than domestic acquisitions. They argue that this contradicts with the foreign direct investment (FDI) theory, which predicts that multinational companies do have a competitive advantage as compared to local companies. However, they present strong results that domestic acquisitions generate more target shareholder wealth as compared to cross-border acquisitions.

Beside these determinants of the bid premium, Eckbo (2009) finds that size and industry relatedness are important drivers of the bid premium. He finds that the bid premium in corporate takeovers tends to be lower the greater the market capitalization of the target, and that there is no effect of whether the takeover is horizontal or vertical (i.e. no effect of industry relatedness). In literature, more determinants of the bid premium are discussed. For example, when bidding management owns equity stakes in the company, the premium tends to be lower (Healy et al., 1992; Agrawal and Mandelker, 1987). Furthermore, Officer (2003) finds that the inclusion of a target termination fee in the merger contract is likely to give rise to the bid premium. Chatterjee et al (2008) find that premiums are larger the greater the disagreement in earning forecasts among analysts following the target. Relevant factors of influence from the discussion in this subsection will be used in this paper to assess the bid premium development more specifically.

The next subsections will relate the bid premium to the three mentioned aggregate waves in the economy, followed by a discussion of their interrelationships.

1.1 Stock Market Valuation Waves and the Bid Premium

affect takeover activity. They see investment decisions as a response to market mispricing and present explanations for the occurrence of the merger waves of the 1960s, 1980s, and 1990s. Though they dedicate much of the paper to relative firm-specific valuations, they present a study of Verter (2002) who argues that high levels of merger activity are associated with a higher dispersion in valuations, which is likely to result in a higher takeover frequency. Moreover, there seems to be consensus in literature about the positive relationship between high market-to-book ratios and merger waves (Rhodes-Kropf et al, 2005; Ang and Cheng, 2006; and Dong et al, 2006). However, recent evidence of Dittmar and Dittmar (2008) shows that it seems unlikely that waves in aggregate valuation are driving corporate events. This issue will be discussed more thoroughly in the last subsection of this literature review.

Looking more specifically at the relationship between stock valuation waves and the bid premium, it was found that the takeover premium is negatively correlated to stock valuation, as founded by Dong et al (2006). As stated in the previous section, target companies know their fundamental value and will demand a price in line with this value, which have to result in a higher bid premium in times of severe undervaluation. Conversely, when stock prices are high, the bid premium tends to be relatively low.

This line of thinking can easily be extended to the relationship between the method of payment and the bid premium. It is well known that in times of market overvaluation, there is an increasing use of stock as the method of payment. Companies tend to use stock as an additional currency. Rhodes-Kropf and Viswanathan (2004), for example, argue that during high valuation markets, targets are prone to overvalue stock offers, they tend to underestimate market-wide overvaluation when the market is overvalued, and as a result, may be more willing to accept the offer (which was in fact too low). Moreover, Shleifer and Vishny (2003) argue that only target companies that are overvalued themselves as well will accept stock bids of overvalued bidders, and Dong et al (2006) find that a higher target price-to-book ratio is associated with a lower bid premium as compared to low price-to-book ratios. Furthermore, it is argued by Eckbo (2009) that bid premiums tend to be lower when the method of payment is stock. This may limit the level of the premium during times of overvaluation, since in high valuated markets the method of payment is mainly stock. Conversely, in low valuation markets, more cash is used as the method of payment. In the same way, Shleifer and Vishny (2003) find that targets in cash acquisitions are undervalued, as compared to targets in stock acquisitions.

Eckbo, 2009). From the bidders point of view, in overvalued markets there are less financial constraints to strengthen the offer with cash (or in some situations stock). However, there seems to be consensus in literature that on average, premiums in lower valuation markets are higher, as compared to premiums in higher valuation markets, making this arguments weighting less than the former arguments.

1.2 Economic Growth and the Bid Premium

A less, though still much, emphasized topic in corporate finance is the relationship between macro-economic development and corporate takeovers. The first one who discussed the topic was Gort (1969) by arguing that economic shocks increase the dispersion of valuations and, as a result, the frequency of mergers. Recently, Dittmar and Dittmar (2008) found evidence that macro-economic growth drives stock issuance, and stock repurchases. They relate this evidence more directly to mergers by arguing that merger frequency shows the same development as equity issues and stock repurchases. They posit that merger activity is positively related to the level of excess cash and tends to increase in later periods of the business cycle. They argue that GDP growth (i.e. the development of the business cycle) is associated with improved cash flow (more surplus cash) and investment opportunities, resulting in a higher level of corporate investments.

This line of thinking can easily be extended to the bid premium. On the one hand, higher cash flow is likely to make bidders financially less constrained during corporate takeovers (for example see Harford, 2005), making them more prone to increase the bid, which is likely to result in a higher average bid premium. On the other hand, improved uncertainty may induce bidders to reserve cash for less flourishing periods, which is likely to be a negative factor to the level of the bid premium. Furthermore, Harford (2005) finds that economic, technological, and regulatory shocks drive merger waves, which is referred to as the neoclassical explanation of mergers. More specifically, he argues that merger waves are driven by the higher capital liquidity as a result of economic growth. He discusses several indicators that accompany an economic expansion (cash flow margin on sales, asset turnover, research and development, capital expenditures, employee growth, return on assets, and sales growth), which can be reliable proxies for the sensitivity of a firm towards an economic movement. Referred to the level of the bid premium, it could be argued that since economic growth accompanies higher capital liquidity, firms are less constraint in assessing their offer premium, possibly resulting in a higher bid.

imperfections. Furthermore, Maksimovic and Phillips (2001) find evidence that, amongst others, shocks in economic demand have a positive influence on the likelihood that assets will be sold. They argue that during economic upturns, inefficient firms are induced to sell and that firms are more likely to be buyers when they are efficient, which is in line with the neoclassical model of merger waves. Related to the bid premium, one can argue that when firms are induced to sell they are likely to accept a lower price than when they are not, resulting in a lower premium during economic upturns. These findings are strong and significant. As such, the number of transactions is likely to be procyclical, while the bid premium itself is not necessarily.

Summarizing, there seems to be disunion in literature about the relationship between economic growth and the bid premium, though there is no doubt that economic upturns are positively related to the frequency at which mergers occur. Several arguments exist in favor of a higher premium, including easier access to capital and enhanced competition in the takeover market. Nevertheless, evidence seems to exist that the bid premium is lower during economic upturns as a result of inefficient firms that are induced to sell, and that the bid premium is higher during economic downturns as a result of capital market imperfections that make firms more likely to be undervalued.

1.3 The Interest Wave and the Bid Premium

The third wave to be discussed in this paper is the interest wave. Economists all over the world agree that sharper credit standards will limit the degree of investment activity, as can be found in any finance textbook. More specifically, Harford (2005) recently finds evidence that the credit easiness can be directly related to the frequency of corporate takeovers. However, less and more dispersed evidence is present about the correlation of this interest wave with the level of the bid premium in particular. Officer (2007) finds that discounts on the acquisition premium tend to be greater when debt is more difficult to obtain. Moreover, Yook (2000) argues that acquisitions for cash are mainly obtained through new debt issues. So, when interest rates are lower, more cash can be generated for the acquisitions and there will me more room to increase the premium. Extending this view to the bid premium, when interest rates are low, capital is more easy to obtain and as a result, the bid premium can be established with more flexibility in financing, which is likely to have a increasing effect on the bid premium.

As such, there is no complete consensus in literature concerning this relationship. Common sense may relate debt easiness to more opportunities to increase bids, while it might well be the case that looser credit standards may favor internal investments above external investments (e.g. acquisitions), which may have a decreasing effect on the bid premium. However, there seems to be stronger evidence concerning the theory that predicts a negative relationship between interest rates and the bid premium.

1.4 Interrelationships between the Three Waves

A couple of studies do interrelate the aggregate waves. Harford (2005), for example, examines the relationship between the neoclassical explanation of mergers waves and the behavioral explanation of merger waves. He argues that market timing (i.e. the behavioral explanation) has low explanatory power relative to a model build on economic development (i.e. the neoclassical explanation), because the relationship between stock market valuation and merger waves is driven by a higher capital liquidity that accompanies an economic expansion, rather than by market timing explanations. He argues that for the behavioral model to be true, economic shocks need to precede merger waves, which is in fact, according to his results, not true. Of course, market valuation itself is an important determinant of capital liquidity, making it an important condition for a wave to occur. However, economic growth is the major driver, not market valuation, as he argues. Extending this line of thinking to the level of the bid premium, it could be argued that in times of high economic growth, buyers are financially less constrained in the capital markets, allowing prices to be approaching fundamental values (e.g. see Shleifer and Vishny, 2003), which is likely to have a decreasing effect on the bid premium.

Another author who interrelate the different waves is Simonyan (2005), who finds that interest rates are the most significant driver of the takeover premium, when compared to the explanatory power of stock market valuation and economic growth. He assumes that economic growth and stock market valuations behave more or less the same and have a negative impact on the bid premium (as discussed earlier, in downturns, targets still know their fundamental value). However, the effect of interest rates is stronger and more significant. In times of economic growth, high interest rates increase funding costs of takeovers, but the effect of increased competition (due to unattractiveness of internal investments) dominates the effect on the bid premium, making interest rates the most accurate forecaster of the bid premium, as he argues.

To conclude, though there is enough discussion, recent literature seems to favor economic motives above market timing or stock market driven motivations. This is because it is assumed that economic development is the underlying factor of influence of the stock market, and therefore it has to be the most powerfull explanation. Furthermore, there seems to be little consensus in the relative explanatory power of the interest rate development, though Simonyan (2005) found evidence that interest rates are the most significant influencer of the bid premium. His findings of the direction of influence, however, seem to contradict the main findings in literature.

1.5 Hypothesis development

The literature review above leads to the development of a set of hypotheses. First, the directions of the three aggregate waves need to be assessed. The first hypothesis relates to overall stock market valuation. In this paper, it is assumed that, during low valuation markets, targets are low priced, but still know their fundamental value, and are likely to demand a higher takeover premium as opposed to current market prices. During high valuation markets, the reverse will happen. As such, hypothesis 1 (H1) will be:

H1: Overall stock market valuation is negatively related to the average level of the bid premium in corporate takeovers

directions to a certain extent, the latter explanation will be used as assumption for hypothesis 2 (H2):

H2: Economic development is negatively related to the average level of the bid premium in corporate takeovers

The third hypothesis relates to the direction of influence of interest rates. Though there is some doubt in literature, the most convincing evidence is present about a negative relationship between interest rates and the bid premium. As such, this paper assumes that interest rates negatively affect the level of the bid premium because of sharper credit standards and more difficulty to obtain capital to increase the bid premium. As such, hypothesis 3 (H3) will be:

H3: The development of interest rates is negatively related to the average level of the bid premium in corporate takeovers

The fourth hypothesis relates to the relative explanatory power of the three waves in terms of significance. Because of the fact that literature seems to favor economic motives above market timing motives, and interest rates are still a point of discussion, this paper assumes that economic development will be the most important driver of the bid premium in terms of significance. As such, hypothesis 4 (H4) will be:

H4: Economic development is the most significant driver of the bid premium, as compared to stock market valuation and interest rates.

II. Methodology & Dataset

2.1 Firm Level Bid Premium

First, to test the direction of influence on the takeover premium of market valuation, economic development, and interest rates, respectively, regression models 1, 2, and 3 are established. i i i

STOCK

BIDPREM

=

β

+

β

+

ε

ECO

BIDPREM

=

β

+

β

+

ε

INTEREST

BIDPREM

=

β

+

β

+

ε

where

BIDPREM

β

β

prior to the announcement of the acquisition, is the average level of industrial

production of the four countries four weeks prior to the announcement, is the level of the Barclays Euro Aggregate Corporate Bond Yield (see section 2.3) four days prior to the announcement. i

STOCK

ECO

INTEREST

A conventional approach for calculating the bid premium is performing an event study by using target cumulative abnormal stock returns. However, as argued by Eckbo (2009), this might give an incorrect picture of the true value of the offer, because these stock returns also reflect other information, e.g. the probability of competition and bid failure. He shows that the choice of the calculation method can result in premium differences of more than 20%. For this reason, this paper will use actual offer premiums as provided by the International Mergers and Acquisitions Database (IMA) of SDC Platinum of Thomson Reuters. To assess this bid premium, Eckbo (2009) recommends to measure the target share price two or three months prior to the announcement date, in order to overcome information leakage problems. However, since SDC Platinum only provides the share price four weeks before the announcement (or shorter), this period is used as a proxy for the share price prior to possible information leakage. As such, the bid premium will be measured as

1 1 − − − t t it p p y  

where is equal to the absolute bid price of the offer as provided by SDC Platinum at the

After the assessment of the directions, the most accurate forecaster of the bid premium needs to be found. In order to do this, the basic model 4 is established.

i i i

i

i

STOCK

ECO

INTEREST

BIDPREM

=

β

+

β

+

β

+

β

+

ε

Firm-level bid premiums possess the advantage that they can be controlled for firm and deal characteristics. As discussed in the literature review section, several factors are influencing the bid premium. Main factors of influence that exist regardless of the three main variables need to be controlled for in order to increase the validity of the model. As such, this paper includes the method of payment (cash versus stock), the geographic scope (domestic or cross-border acquisition), the degree of industry relatedness (whether or not the acquirer and the target have comparable SIC codes), and the target enterprise value (i.e. the size of the target) as control variables. These factors of influence were discussed before in section I. As a result, model 5 will be the final model for the firm-level bid premium regressions.

+

+

+

+

+

i

ECO

INTEREST

PAY

GEO

TOCK

+

=

S

BIDPREM

β

β

β

β

β

β

SIZE

REL

β

ε

β

+

+

where is a dummy for the method of payment of acquisition i (it takes a value of 1 for acquisitions where more than 50% of the deal value was paid in cash, and it takes a value of 0 for acquisitions where more than 50% of the deal value was paid in stock), is a dummy for the geographic scope of acquisition i (it takes 1 for a domestic acquisition and 0 for a cross-border acquisition), is a dummy for the industry relatedness of acquisition i (it takes 1 when the first character of the SIC code is the same for acquirer and target, and it takes 0 otherwise), and is the enterprise value of the target in acquisition i (as measured by the target enterprise value).

i

PAY

GEO

REL

SIZE

Therefore, a separate regression will be performed for acquisitions where the target is situated in the UK. Finally, as a check of robustness, the bid premium will also be measured with a pre-announcement period of one week.

2.2 Time Series Analysis

To get a better understanding of the influence of the fluctuation of the three waves over uniform time intervals on the bid premium, the firm-level bid premiums are averaged on a monthly basis and regressed against the same Euro Zone wave variables (waves of four countries under review). With this transformation, a time series analysis is possible. As such, a multivariate time series regression will be performed as presented in model 6.

i t t

t

t

STOCK

ECO

INTEREST

BIDPREM

=

β

+

β

+

β

+

β

+

ε

where is the average monthly bid premium at time t, is the

one-month lagged average value of the main stock indices of the respective countries, is

the one-month lagged average industrial production, and is the one-month lagged Euro Aggregate Corporate Bond Yield. To assess certain waves in the model, an opposite regression will be performed where the three main independent variables will be measured at time t+1, as opposed to the measurement at time t-1. If the directions of the coefficients do not change, the model will be measured at t-X, where X is the number of months at the point the coefficient of the main variables changes direction (from positive to negative values or vice versa). In this way, waves in the explanation power of the independent variables can be assessed.

t

BIDPREM

STOCK

ECO

INTEREST

With this time-series analysis, controlling for firm and deal level characteristics is no longer possible. However, it is possible now to control for time relevant control variables. More specifically, Simonyan (2005) uses capacity utilization as a control variable. This is because this factor may have a great impact on the bid premium, regardless the impact of overall stock market valuation, economic growth, or corporate bond yield. For this reason, the time series model has to be controlled for this potential disturbing effect, by performing the regression as presented in model 7.

+

+

+

+

STOCK

ECO

INTEREST

where is the one-month lagged average capacity utilization rate. After the assessment of the final model (7), the same robustness checks as in the cross-sectional analysis will be performed regarding the pre-announcement period of one week and the particular case of UK acquisitions.

1 −

t

CAPUTI

2.3 Dataset

Since the tradability of a security will strengthen the reliability of the assessment of the sensitivity to the respective waves, only the most traded securities of Western European countries will be included in this study, which are represented in the main stock indices of the respective countries. Therefore, this paper will include public takeovers where the target was listed on the main indices of France, Germany, the Netherlands, and the United Kingdom. The period measured will be January 1st, 1999 till December 31st, 2008, a time span of ten subsequent years.

Transaction data is obtained from the International Mergers and Acquisitions Database (IMA) from SDC Platinum as provided by Thomson Reuters. The sample consists of public takeovers during the period January 1999 until December 2008 and meets the following criteria: 1) target companies were listed on the stock indices of France, Germany, the Netherlands, or the United Kingdom; 2), the acquirer changed ownership in the target company from less than 50% to more than 50% (in order to filter out small marginal investments); 3) the takeover was completed (i.e. not pending) in order to filter out unrealistic bid premiums; and 4) data about the premium offered and the control variables was present. In order to obtain a realistic dataset, outliers were eliminated as well as unrealistic negative bid premiums (it is assumed that targets are not likely to accept bids well below their market price).

Filtering for the above criteria and eliminating extreme and unrealistic outliers, results in a sample size of 974 public takeovers. Bid premiums were provided by SDC Platinum and were recalculated as a check by dividing the price paid in the host currency minus the target closing price in the host currency four weeks before the announcement by the target closing price in the host currency four weeks before the announcement. This yielded the same results.

are used as a proxy for economic growth, since quarterly results on a relative short time period of ten years would result in too few observations. This data on industrial production was obtained from the Organization for Economic Cooperation and Development (OECD). As a next step, aggregate waves are generated by averaging the data of the four countries for both the stock index prices as well as the industrial production. Finally, to measure the debt easiness for companies in the Euro zone, daily yield prices of the Euro-Aggregate Corporate Bond are obtained from Barclays Capital Live as provided by Barclays Bank PLC.

Concerning the control variables in the model, data on the firm and deal size characteristics (i.e. method of payment, geographic scope, industry relatedness, and size) is also obtained from the IMA from SDC Platinum. The capacity utilization rates for the Euro Zone are derived from the Datawarehouse of the European Central Bank (ECB).

Table I presents summary statistics for the relevant variables in the model. As can be concluded of this table, the mean bid premium in the sample equals 33.4 with a standard deviation of 24.9. The standard deviations of the other variables are relatively smaller. During the period under review, the bid premium went as high as 100.0 in April 2007 (UK target), and as low as -29.9 in March 2000 (French target). Average industrial production was lowest in February 1999 and reached its top in April 2008. The lowest corporate bond yield (3.32) was observed in September 2005, the highest (6.27) in September 2000. Furthermore, a correlation matrix is presented in Appendix 1. It can be seen that the highest correlation (0.51) is observed between the average stock index and the corporate bond yield. This might be explained by the fact that when obtaining debt is relatively expensive, the stock market becomes more attractive as a method of acquiring capital. The second best correlation (0.30) exists between the stock market and economic development. This correlation coefficient might be explained by the fact that stock markets react on economic development, as argued by Harford (2005).

III. Results

3.1 Cross-sectional Regression Results

Returning to the model established in Section II, at first the direction of influence on the takeover premium needs to be assessed for each respective wave using models 1, 2, and 3. The next step is to integrate these three waves (model 4) and to integrate the control variables (model 5).

Table I

Summary Statistics Main Variables

This table shows summary statistics of the main variables of the model, i.e. the variables that are subject of review in this paper. Moreover, it includes the results of normality tests on the residuals. BIDPREM is the bid premium as measured with a pre-announcement date of four weeks, STOCK is the one-month lagged average value of the

stock indices, ECO is the one-month lagged average industrial production, and INTEREST is the one-month lagged corporate bond yield.

Number of observations 974

BIDPREM STOCK ECO INTEREST

Mean 33.41 4242.55 98.97 4.83 Median 31.01 4251.44 98.93 4.97 Maximum 100.00 5466.89 107.93 6.27 Minimum -29.93 2119.77 91.96 3.32 Std. Dev. 24.87 734.39 3.58 0.89 Skewness 0.31 -0.57 25.31 -0.05 Kurtosis 2.81 2.81 2.63 1.63 Jarque-Bera 16.91 55.21 25.31 77.78 Probability 0.00 0.00 0.00 0.00

the three regressions do not reveal violations on the normality assumption. However, all three regression results show relative low adjusted R-squared values, violating the predictive power of the models. Nevertheless, the results give an indication of the direction to which the waves are exposed to the bid premium.

It can be seen that it is likely that there is a positive relationship between the overall value of the stock market and the bid premium when measured on a standalone basis. In terms of significance, it can be seen that this effect seems to be the strongest of all univariate regressions. Hence, it might be argued that a change of 1% in the aggregate value of the stock markets, leads to a premium increase of 0.002%. These results are significant at the 5% level. This result is surprising, since a negative coefficient was expected in hypothesis H1, arguing that target companies know their fundamental value, and will demand a price in line with this value, which have to result in higher premiums during low valuation markets. However, the founded positive coefficient might be explained by the fact that during periods of high stock valuation, takeover frequency is usually high, increasing competition in the market for corporate takeovers (Eckbo, 2009). Moreover, the founded positive relationship is in line with the findings of Shleifer and Vishny (2003), who finds that investment decisions seem to be a response to market mispricing.

Table II

Cross-sectional Regression Results on Bid Premium

This table shows the regression results with BIDPREM as dependent variable, where BIDPREM is the bid premium as measured with a pre-announcement date of four weeks, STOCK is the one-month lagged average value of the stock indices, ECO is the one-month lagged average industrial production, and INTEREST is the one-month lagged corporate bond yield, dPAY is a

dummy for the method of payment (it takes 1 for acquisitions where more than 50% of the deal value was paid in cash, and it takes 0 if more than 50% of the deal value was paid in stock), dGEO is a dummy for the geographic scope of the acquisition (it takes 1 for a domestic acquisition and it takes 0 for a cross-border acquisition), dREL is a dummy for the industry relatedness of

the acquisition (it takes 1 when the SIC codes of the acquirer and the target begin with the same number, and it takes 0 otherwise), and SIZE is equal to the enterprise value of the target.

Number of observations 974

Model 1 Model 2 Model 3 Model 4 Model 5 Weekprem UK-targets Intercept 24.12 71.01 27.25 75.90 90.76 73.97 141.84 (5.20)*** (3.24)*** (6.20)*** (3.45)*** (4.14)*** (3.60)*** (5.88)*** STOCK (t-1) 0.00 0.00 0.00 0.00 0.00 (2.03)** (2.07)** (2.15)** (2.05)** (2.60)*** ECO (t-1) -0.38 -0.57 -0.77 -0.66 -1.32 (-1.71)* (-2.47)** (-3.32)*** (-3.04)*** (-5.15)*** INTEREST (t-1) 1.27 0.56 0.96 1.71 1.78 (1.42) (0.54) (0.93) (1.78)* (1.52) dPAY 9.17 8.27 9.81 (4.68)*** (4.51)*** (4.33)*** dGEO -4.58 -2.63 -6.81 (-2.56)** (-1.57) (-3.17)*** dREL -3.15 -2.25 -4.28 (-1.94)* (1.48) (-2.35)** SIZE -0.00 -0.00 -0.00 (-0.69) (-1.03) (-1.85)* Adjusted R-squared 0.00 0.00 0.00 0.01 0.04 0.04 0.08

*** significant at the 1% level

** significant at the 5% level

* significant at the 10% level

and is an important variable in ex

results are in line with the expectations of hypothesis 2 (H2). As such, it might be argued that the fact that companies are induced to sell during economic expansions seems stronger than the effect that economic development has on capital easiness. Moreover, it might be explained by the finding of Simonyan (2005) that during periods of low economic growth, capital market imperfections have a decreasing effect on the bid premium. Moreover, it might be argued that strong economic conditions enhance uncertainty and may induce firms to reserve cash for less flourishing periods, which might have a decreasing effect on the bid premium.

Since the results of the univariate regressions of the interest rates on the bid premium are not significant at any of the three levels, inferences need to be taken with caution. However, it might be argued that there seems to be a positive relationship between interest rates and the bid premium, thereby violating hypothesis H3. This result might be explained by the findings of Simonyan (2005), arguing that a higher interest rate favors external investments above internal investments and thus increases competition in the takeover market. As a consequence, these results contradict with the findings of Officer (2007), Yook (2000), and Harford (2005), all arguing that low interest rates are stimulating the level of the bid premium in a positive direction, because of a greater accessibility to capital.

As a next step, the multivariate model of equation 4 (see section II) is created by executing a multivariate regression with the three independent variables integrated. This leads to the estimation results of Model 4 in Table II1. As can be seen from the table, the value of the adjusted R-squared increased as compared to the univariate regression results. This would imply that the three waves together improve the explanation power of the model as compared to the waves on a standalone basis.

The most significant result in the renewed model seems to be economic development. An increase of industrial production of 1% now causes a change of -0.57% of the bid premium and is significant at the 5% level. The stock market coefficient now reveals a lower t-statistic than the economic development coefficient. Thus, where the stock market is the most important driver of the bid premium on a standalone basis, economic development seems to be more important in an integrated model. This is in line with the arguments of Dittmar and Dittmar (2008) and Harford (2005) that economic motives are stronger than stock market motives in explaining corporate investment activity and is in favor of hypothesis H4 in this paper. Nevertheless, the coefficient of the stock market is still significant at the 5% level plaining the bid premium. In this way, the results of Shleifer

1

In this paper, aggregate values are used for the separate waves. When regressing bid premiums against stock market and industrial production on a per country basis, the small sample sizes lead to

nreliable outcomes. u

and Vishny (2003) and Rhodes-Kropf and Viswanathan (2004) are not violated and the stock market remains a factor of influence. A striking result of this multivariate regression model, however, is the insignificance of the corporate bond yield coefficient. This is remarkable, since Simonyan (2005) found that interest rate development is the most significant factor in explaining the bid premium. As such, inferences are dangerous to make since the low t-statistic.

The next step is to include the control variables as discussed in section II: the method of payment (>50% cash acquisitions vs >50% stock acquisitions), the geographic scope (domestic or cross-border acquisitions), industry relatedness (whether the acquirer and the target have similar SIC codes or not), and size (as measured by the enterprise value). As mentioned earlier, the first three control variables will be entered as dummies in the model. For the estimation results of this final firm-level cross-sectional model, see Model 5 in Table II. As can be seen, when controlling for the method of payment, geographic scope, industry relatedness, and size, the three main independent variables all become more significant. Moreover, the adjusted R-squared value increased significantly to 0.04, indicating a higher explanation power of the model when controlled for the respective variables.

The effect of industrial production on the bid premium in particular seems to be strengthened by including the control variables (strongest increase of the t-statistic compared to Model 4). In fact, the economic development coefficient is now significant at the 1% level and its influence increased as opposed to model 4, now causing a change of -0.77% of the bid premium for each 1% increase in industrial production. The t-statistic of the stock market coefficient increased slightly, just as the t-statistic of the interest rate coefficient. The latter, however, is still not significant at any of the three levels and seems to be of almost no importance in forecasting the bid premium. Another improvement of the model is that heteroscedastic residuals are less likely. In short, the model itself and two of the three main variables became remarkably stronger by including the four control variables in the model.

acquisitions comprise a hidden, non-transparent currency (stock), and are based on sentiment, and thus difficult to measure.

As discussed earlier, the geographic scope of an acquisition is likely to be of influence when assessing the bid premium, in that a lower bid premium is expected when the acquisition is a cross-border transaction (see literature section). The regression results of Model 5 contradict with these earlier findings of literature. It can be seen that the GEO dummy takes a coefficient value of -4.57 and is significant at the 5% level. In this way, it is likely that 1) geographic scope is an important control variable referring to the high t-statistic, and 2) domestic acquisitions generate lower bid premiums than cross-border acquisitions. In this way, the results contradict with the findings of Goergen and Renneboog (2004), who argue that domestic acquisitions do generate more target shareholder wealth than cross-border transactions. This can be explained by the earlier discussed foreign direct investment theory (FDI) as violated by the findings of Goergen and Renneboog (2004). Clearly, this theory cannot be violated by the results of this paper.

The third control variable to be discussed is the industry relatedness of the acquirer and the target (whether or not their SIC codes begin with an identical number). Though not significant at the 5% level, but on the 10% level, the results indicate that the fact that the acquirer and the target are in the same industry (i.e. begin with the same first number in their SIC codes) does have a decreasing effect on the bid premium. While Eckbo (2009) did not find any relationship between the industry relatedness of the acquisition and the level of the bid premium, this paper does clearly. On the other hand, though his findings are not strong, Simonyan (2005) finds that unrelated acquisitions do result in higher bid premiums than related acquisitions. The results in this paper seem to confirm his findings. It might be argued that target shareholders of non-related acquisitions have less confidence in the success of the takeover (possibly because of unfamiliarity of the acquirer with the industry) and, as a result, are demanding a higher takeover premium.

Finally, previous research (e.g. Eckbo, 2009) showed that the greater the market capitalization of the target, the lower the bid premium. Table II also presents the regression results of enterprise value against the bid premium. As can be seen from the results, there seems to be, as was expected, a negative relationship between the bid premium and the enterprise value. However, the coefficient is far from significant, so it is not valid to make inferences out of these results. In short, it is likely that there is no substantial size effect in this study that may damage the validity of the results.

coefficient and the economic development coefficient became less significant. The most striking result, however, is that the level of the corporate bond yield is significant for the first time as compared to all former regressions. It might be argued that the level of interest rates (i.e. the corporate bond yield) is an important factor in short-term investment decisions, while stock market levels and economic development are more important in assessing the bid premium on a longer term.

The last step in the cross-sectional analysis is assessing whether there is a UK effect in the results or not. As discussed before, Goergen and Renneboog (2004) argue that a higher bid premium is expected when a UK target is involved, as compared to acquisitions where no UK target is involved, building on the idea of higher shareholder protection and an advanced market for corporate control. Column 7 of Table II shows the result of the final cross-sectional model on only UK targets. It can be seen from the results that all coefficients (including control variables) became more significant. More specifically, all factors but size and interest rate development are now significant at least at the 5% level. Moreover, all coefficients became stronger (in value), as well as the adjusted R-squared value (it almost doubled), indicating stronger influences from macro economic developments for acquisitions where UK targets are involved. These results indicate that shareholders of UK companies might be more sensitive to movements in market forces than shareholders of non-UK companies, possibly caused by more developed high-liquid equity markets (Goergen and Renneboog, 2004), as discussed earlier. These results are significant.

an acquisition, and the industry relatedness of an acquisition. Furthermore, the explanation power of interest rates seems to increase when decreasing the pre-announcement period to one week. Another remarkable finding is the significant improvement of the model when only looking at acquisitions where a UK target is involved.

3.2 Time Series Regression Results

After assessing the importance of the aggregate waves on the firm level bid premium and controlling for several firm and deal characteristics, time series analysis is performed to assess whether the effects hold over uniform time intervals and whether there are certain patterns in the influence of these waves on the aggregated premium. For this reason, bid premiums and independent variables are averaged into monthly data. To identify waves in these patterns, the bid premium is regressed by measuring the main independent variables at t-1 as well as t+t-1.

The results of the basic time series model in this paper (see equation 6) are presented in Table III. It can be seen from the table that the directions of the waves are identical to the directions founded with the cross-sectional analysis. This is not surprising, since the same acquisition data was used to establish the average monthly bid premia. Also the coefficients of the stock market and the industrial production show similar values as the former analysis. The influence of interest rates, however, seems stronger when looking at uniform time intervals. It is now likely that a 1% increase in interest rates causes a rise of 2.11% in the monthly average bid premium, as compared to an insignificant 0.96% in the cross-sectional results. This might be explained by the fact that interest rates have a more pattern-like development over time, as opposed to the other waves under review. Moreover, these time-series regression results are more in line with the findings of Simonyan (2005), who finds that interest rates are the most significant driver of the bid premium. Another important finding of the model is that the model at t+1 seems to have more explanatory power than the model at t-1, as can be seen in column 3 of Table III. More specifically, the coefficient of the interest rate development is now significant at the 5% level (this was not achieved at the cross-sectional regressions). Moreover, the adjusted R-squared of the model is significantly higher, now approaching a value of 0.11. These results may indicate that predictions about future corporate bond yields are more important in assessing the bid premium as opposed to historical figures of corporate bond yields, compared to the differences between t-1 and t+1 concerning the other two independent variables.

Table III

Time-Series Regression Results on Bid Premium

This table shows the regression results with BIDPREM as dependent variable, where BIDPREM is the average monthly bid premium as measured with a pre-announcement date of four weeks, STOCK is the one-month lagged average value of the stock indices, ECO is the one-month lagged average industrial production, and INTEREST is the one-month lagged corporate bond yield,

and CAPUTI is the one month lagged capacity utilization rate.

Number of observations 120

Model 6 Model 7 t+1 Weekprem UK-targets

Intercept 75.91 -39.59 74.41 33.58 23.83 (2.76)*** (-0.42) (2.59)** (1.14) (2.46)** STOCK (t-1) 0.00 0.00 0.00 0.00 0.00 (1.87)*** (0.23) (1.98)*** (1.06) (2.05)** ECO (t-1) -0.65 -0.79 -0.66 -0.31 -0.13 (-2.26)** (-2.56)** (-2.21)** (-1.00) (2.23)** INTEREST (t-1) 2.11 1.88 2.50 4.40 1.91 (1.84)*** (1.63) (2.22)** (3.59)* (1.38) CAPUTI (t-1) 1.67 (1.29) Adjusted R-squared 0.08 0.09 0.11 0.14 0.07

*** significant at the 1% level

** significant at the 5% level

* significant at the 10% level

t+1, where X is the number of months at the point the coefficient of the main variables turns around (from a positive value to a negative value or vice versa). It turns out that the coefficient of the stock market stays more or less the same over a longer period, and turns around for the first time at t-39, indicating that the stock market wave on the bid premium last for more than three years. The value of the economic development however, becomes positive for the first time at t-18 and the value of the development of the corporate bond yield becomes negative for the first time at t-13. This may indicate that economic development occurs in waves of one and a half year, whereas interest rates occur in waves of approximately one year.

The next step is to include the time series control variable of the capacity utilization rate (as included in the framework of Simonyan, 2005). This leads to the estimation output of column 2 at Table III. As can be seen, capacity utilization does not add predictive power to the model and does not seem to be a good forecaster of the bid premium. As such, it is no valid control variable and should be left out of the model. It should be said, however, that capacity utilization figures were only available on a quarterly basis and may not provide a good indication and a proper estimate. As such, results should be interpreted with care.

So far, the model seems valid. However, as in the cross-sectional analysis, a robustness check needs to be performed on the pre-announcement period. As such, bid premiums as calculated with a one week pre-announcement period are averaged into monthly data and the same regression is performed as the first time series regression in this paper. The results are presented in column 4 of Table III. It can be seen that when a one week pre-announcement period is used, the coefficient of the interest rates becomes significant at the 1% level for the first time and causes a rise in the average monthly bid premium of 4.40% for every 1% increase in interest rates. This is almost similar to what happened in the cross-sectional analysis when using a one week premium and might be explained by the fact that interest rates are most important at assessing the bid premium in the short term. However, the coefficients of the stock market and economic development become insignificant at all levels, thereby violating the model severely.

differences in the influence of economic development. The interest rate coefficient becomes insignificant again. Moreover, the adjusted R-squared decreases. These results may lead to the conclusion that forecasting the bid premium for UK firms is not remarkably different than forecasting the bid premium for the whole sample when looking at uniform time intervals. This is in slight contrast with the cross-sectional analysis and might be explained by the fact that averages are used instead of individual acquisition data.

To conclude, the results of the time series analysis in this paper are not completely different from the results of the cross-sectional analysis. However, the model seems to be more valid according to the relatively high value of the adjusted R-squared (0.08 compared to 0.04 for the cross-sectional results). The largest difference seems to be that the coefficient of the interest rate factor is more significant now, indicating that bid premiums are better explained by interest rates at uniform time intervals, than they are explained at the cross-sectional dataset. This may be because interest rates behave more time consequent. Nevertheless, the directions of the waves remain the same within the times series analysis and economic development remains the most significant driver of the bid premium. As such, the same conclusions concerning the hypotheses can be drawn as in the cross-sectional analysis. Another interesting result is that the model at t+1 seems to have more explanatory power than the model at t-1, indicating that economic projections are more important in assessing the bid premium rather than historical figures. Looking at fluctuations over time, it seems likely that the influence of the stock market does change over time every three years. However, it is likely that the explanation of economic development occurs in waves of approximately 18 months, whereas the explanation of the corporate bond yield occurs in waves of approximately 13 months.

Conclusion and discussion

The aim of this paper is to find the directions of influence and the relative explanatory power of the stock market, economic development, and interest rates on the bid premium in corporate takeover. As such, the direction of influence of the three separate waves, the relative predictive accuracy in terms of significance, and the fluctuation over time, are assessed successively. In order to perform this study, a cross-sectional analysis and a time series analysis are established.

effect on the bid premium. This was expected after the findings of Maksimovic and Phillips (2001) that economic flourishing times induces inefficient firms to sell and after the findings of Simonyan (2005) that economic less flourishing times make firms too much undervalued caused by capital market imperfections. Thirdly, the outcomes show that the development of interest rates (as measured by the corporate bond yield) has a positive effect on the bid premium. However, these results are only significant at the time series model and when measured with a one week pre-announcement period, and need to be interpreted with caution. This likely positive influence of interest rates is rather surprising since strong findings in literature that lower interest rates provide easier excess to capital and are driving up bid premiums (Officer, 2007; Yook, 2000). However, it might be explained by the fact that low interest rates also make internal projects of the firm more attractive, thereby decreasing competition in the external takeover market (Simonyan, 2005). As such, hypothesis H1 that the stock market is negatively related to the bid premium, needs to be rejected. Hypothesis H2 that economic development is negatively related to the bid premium can be accepted, and hypothesis H3 that interest rates are negatively related to the bid premium needs to be rejected.

Looking at the relative explanatory power, the main multivariate models in this paper find that economic development is the most significant driver of the bid premium (this does not hold when measuring the independent variables at t+1 and when using a one week pre-announcement period). This result is in line with Dittmar and Dittmar (2008) and Harford (2005) who find that corporate investments are driven by economic motives, rather than stock market motives. As such, hypothesis H4, stating that economic development is the most important driver in assessing the bid premium, can be accepted.

The time series model seems to have a higher explanatory power than the cross-sectional model in assessing the bid premium. Another notable result of this paper is the fact that the time series model for t+1 seems to have more explanatory power than the time series model for t-1, suggesting that projections are more important than historical observations in assessing the bid premium. Looking at possible fluctuations over time, this paper finds that the explanatory power of the stock market coefficients seem to fluctuate over time intervals of approximately three years, the economic development force on the bid premium seems to occur in waves of 18 months, where interest rates fluctuate over a period of 13 months.

remarkable result of this paper is that the cross-sectional model seems to increase in explanatory power when only regressing UK targets, in that coefficients become more significant.

REFERENCES

Agrawal, A, and Mandelker, G, 1987, Managerial Incentives and Corporate Investment and Financing Decisions, Journal of Finance, 42, 823–37

Ang, J.S, and Cheng, Y, 2006, Direct Evidence on the Market-driven Acquisition Theory, The

Journal of Financial Research, 29, 199-216

Chatterjee, S; John, K; Yan, A, 2008, Takeover Premium and Divergence of Opinions, Working Paper, School of Business, Fordham University

Dittmar, A.K, and Dittmar, R.F, 2008, The Timing of Financing Decisions: An Examination of the Correlation in Financing Waves, The Journal of Financial Economics, 90, 59-83

Dong, M, Hirshleifer, D, Richardson, S, and Teoh, S.H, 2006, Does Investor Misvaluation Drive the Takeover Market?, The Journal of Finance, 61, 725-762

Eckbo, B.E, 2009, Bidding Strategies and Takeover Premiums: a Review, Journal of

Corporate Finance, 15, 149-178

Goergen, M, and Renneboog, L, 2004, Shareholder Wealth Effects of European Domestic and Cross-border Takeover Bids, European Financial Management, 10, 1: 9-45

Gort, M, 1969, An economic Disturbance Theory of Mergers, Quarterly Journal of

Economics, 83, 624–642

Harford, J, 1999, Corporate Cash Reserves and Acquisitions, The Journal of Finance, 54, 6: 1969 - 1997

Harford, J, 2005, What Drives Merger Waves?, Journal of Financial Economics, 77, 529-560

Healy, P; Palepu, K; and Ruback, R., 1992, Does Corporate Performance Improve after Mergers?, Journal of Financial Economics, 31, 135–75

Jong de, A., Poel van der, M., Wolfswinkel, M., 2007, Corporate Governance and Acquisitions: Acquirer Wealth Effects in the Netherlands. Erasmus University Rotterdam

Jovanovic, B, and Rousseau, P, 2001, Mergers and Technological Change: 1885-2001, unpublished working paper, Vanderbilt University

Maksimovic, V, and Phillips, G, 2001, The Market for Corporate Assets: Who Engages in Mergers and Asset Sales and Are There Efficiency Gains?, 56, 6: 2019 - 2065

Officer, M.S, 2003, Termination Fees in Mergers and Acquisitions, Journal of Financial

Economics, 69, 431-467

Targets, Journal of Financial Economics, 83, 571 - 598

Rhodes-Kropf, M, and Viswanathan, S, 2004, Market Valuation and Merger Waves, The

Journal of Finance, 59, 2685-2718

Rhodes-Kropf, M, Robinson, D.T., and Viswanathan, S, 2005, Valuation Waves and Merger Activity: The Empircal evidence, Journal of Financial Economics, 77, 561-603

Rossi, S, and Volpin, P.F, 2004, Cross-Country Determinants of Mergers and Acquisitions,

Journal of Financial Economics, 74, 277-304

Shleifer, A, and Vishny, R.W, 2003, Stock Market Driven Acquisitions, The Journal of

Financial Economics, 70, 295-311

Simonyan, K, 2005, The Dynamics of Takeover Premia: an Empircal Analysis, Sawyer School of Management, Suffolk University Boston

Verter, G, 2002, Timing Merger Waves, Harvard University Mimeo

Appendix 1

Correlation Matrix Main Variables

Overview of correlation coefficients between variables used in this paper, where BIDPREM is the bid premium as measured with a pre-announcement date of four weeks, STOCK is the one-month lagged average value of the

stock indices, ECO is the one-month lagged average industrial production, and INTEREST is the one-month lagged corporate bond yield.

Number of observations 974

BIDPREM STOCK ECO INTEREST

BIDPREM - 0.06 -0.05 0.05

STOCK 0.06 - 0.30 0.51

ECO -0.05 0.30 - 0.19