Merger Means of Payment and Analyst Recommendation Change

Proceedings of 4th Global Business and Finance Research Conference

25 - 27 May 2015, Marriott Hotel, Melbourne, Australia

ISBN: 978-1-922069-76-4

1

Merger Means of Payment and Analyst Recommendation Change

Yiling Zhang*

We find a strong evidence that compared to pure stock deal in merger and acquisition (M&A), cash only deal is more likely to lead analysts upgrade their recommendations over the acquirer stocks within 90 days window of deal announcement date. An event study within three days window of deal announcement date in the M&A context allows us to disentangle merger means of payment impact on short term abnormal return of acquirer stocks from other impacts. We find 1.06%(-1.05%) cumulative abnormal return for the acquirer stocks of cash only (Pure stock) deals during this time frame. Our findings extend the merger means of payment hypothesis, suggesting compared to pure stock deals, cash only deal in a short term will receive more favorableness from analysts to upgrade their recommendations about the acquirer stocks. Market immediately reacts to this difference and result in a difference of cumulative abnormal return about acquirer stocks.

Track: Finance JEL Codes: G34, G14

1. Introduction Broad literature has done the research on the merger means of payment long term impact on the cross sectional stock return for the acquirer. Loughran and Vijh(1997) find that with a cash tender offer merger deal, the acquirer is able to earn a significant positive long run abnormal return post the merger transaction. But with a stock only merger, the acquirer earns a significant negative long run abnormal return post the merger transaction. Rau and Vermaelen(1998) as well as Agrawal and support Loughran and Vijh’s view by finding the same patterns of returns for acquirer stocks in cash tender offer merger deal and stock only merger deal after controlling for the size and book-to-market factors as suggested by Fama and French (1992). In addition, Shleifer and Vishny (2003) developed a simple model of acquisition to support the empirical evidence listed by previous scholars that long run returns to acquirers are negative in stock acquisitions and positive in cash acquisitions. Furthermore, they come up with the means of payment hypothesis, which states that if acquirer firms’ managers are better informed about the target’s prospects than the market, managers acquire the target with only stock when the acquirer stock price is overpriced from a long run perspective and managers will acquire the target with only cash when the acquirer stock price is underpriced from a long run perspective. This hypothesis indicates why the stock only merger earns the acquirer a negative abnormal return in the long run but the cash only merger earns the acquirer a positive abnormal return in the long run. ______________________________ *Mr. Yiling Zhang, Department of Finance and Real Estate, University of Texas at Arlington, USA. Email Address: [email protected]



ISBN: 978-1-922069-76-4

2

As we have seen that merger means of payment have significant impact on the acquirer stock abnormal return over the long run, would analyst recommendation change about acquirer stocks react differently to various merger means of payment within a short term of deal announcement and would market reacts to this short term analyst recommendation change difference, finally resulting in a difference of cumulative abnormal return for acquirer stocks for various merger means of payment? This question is a very important empirical research question. First, because most prior studies, such as Loughran and Vijh(1997), Jaffee et al. (2001) and Savor et al. (2009), focused on the merger means of payment play an important role in explaining the long run abnormal return for the acquirer stock and few literature interpret the merger means of payment impact on acquirer stocks’ short run abnormal return. This paper complements this field of research. Second, because this question helps us understand merger means of payment impact on analyst recommendation change. Our examination of the informativeness of merger means of payment contributes to extant research on the information content of analyst’s stock recommendation. The major contribution of this paper lies in our answer to the research question above, firstly our answer provides a good extension to the merger means of payment hypothesis by Shleifer and Vishny (2003). Merger means of payment hypothesis predicts that cash only deal delivers a positive long run abnormal return and negative long run abnormal return for acquirer stocks. Our extension lies in that compared to pure stock deals, cash only deal in a short term will receive more favorableness from analysts to upgrade their recommendations about the acquirer stocks. Market immediately reacts to this difference and result in a significant positive cumulative abnormal return for acquirer stocks of cash only deal and a significant negative cumulative abnormal return for acquirer stocks of pure stock deal. Second, our empirical finding complements Mitchel et al. (2004). They find that merger arbitrage short selling causes short run downward price pressure for acquirer in stock financed mergers around deal announcement period. our finding shows that cash only deal’s impact on analyst recommendation change over acquirer stock is translated into 1.06% cumulative abnormal return for the acquirer stock during the three days window of the deal announcement, suggesting analyst recommendation change because of cash only impact create an upward price pressure.

2. Data and Variable Descriptions M & A Deals We obtain U.S. domestic M&A transaction data from Securities Data Corporation (SDC) Platinum for years 1993 to 2013. Our sample is only consisted of statutory mergers and acquisitions of assets. We include cash only, stock only and cash & stock combination completed unconditional deals in our sample. We exclude from our sample buybacks, acquisitions of certain assets, acquisitions of partial interest, recapitalizations, spin-offs, split-offs, exchange offers and acquisitions of remaining interest. We require target and acquirer are both public trading companies and at least one advisor has been retained by



ISBN: 978-1-922069-76-4

3

the target or the acquirer. Finally, to make sure the merger and acquisition is a significant deal, we exclude that those deals that its target market value is less than 5% of combined acquirer and target market value. After applying all these criteria, we are left with 11,863 deals. Merging deals with CRSP To obtain the full 9 digit cusip codes for the acquirer stocks, we start to obtain 6,973 unique acquirer 6 digit Ncusip codes from those 11,863 deals in SDC platinum database. Then use those unique Ncusip codes to merge with the CRSP monthly stock entire database which dated from December 1925 to December 2013. From this step above, we can obtain 3,594 unique full 9 digit acquirer stocks cusip codes which cover the 8,071 out of original 11,863 deals. Merging deals with analyst recommendation changes We upload those 3,594 unique full cusip codes into the I/B/E/S recommendation detail database, select the database period from December 1992 to June 2014, we are left 2,751 unique full acquirer stock cusip codes which cover 5,978 out of original 11,863 deals. We are emphasizing on the analyst recommendation change instead of recommendation levels because Womack (1996) suggested the analyst recommendation change is economic meaningful. For each analyst, we obtain all available current analyst recommendation scores (ranged from 1 to 5, 1 indicates strong buy and 5 indicates strong sell) about acquirers issued within + 90 days of the M&A deal announcement date and take the average of those scores for the current analyst recommendation score. Then, we obtain last available recommendation scores issued within -90 days of the M&A deal announcement date and take the average of those scores for the last analyst recommendation score. We define a dummy variable, Upgrade, and set it equal to 1 if the current recommendation score is less than last recommendation score improved and 0 if current recommendation score is greater or equal to last recommendation score. We cannot compute Upgrade variable if the we are lack of last analyst recommendation score, then we discard this observation from the sample due to lack of last analyst recommendation score. The above criteria leaves 5,978 deals in the sample of recommendation changes. Merging deals with analyst forecasts

We upload those 2751 unique full acquirer stock cusip codes in the previous step into I/B/E/S detail history database and select the entire database period from January 1970 to June 2014, we choose the observations with long term growth forecast. This criteria leaves us only 1541 unique full acquirer stock cusip codes which covers 4,028 out of 11,863 original deals. We focus on analysts long term growth forecasts because lin and McNichols (1998) indicated analysts long term growth forecasts plays an important role in influencing on the investment recommendation.

http://www.sciencedirect.com.ezproxy.uta.edu/science/article/pii/S0165410198000160



ISBN: 978-1-922069-76-4

4

M & A descriptive statistics Table 1 presents descriptive statistics on the value and number of the M&A deals from year 1993 to year 2013 in the full sample as described in I A. We define the VALUE as the total nominal amount of consideration paid by the acquirer. Those deals are categorized into three methods of payment: Cash only, pure stock and Mixed. The mixed subset includes all the acquisitions in which payment is stock and cash combination. We tend to see that stock deals are usually larger in size than Cash only deals. The number of cash only deals is greater than that of stocks financed deals.

Table 1 Descriptive statistics on deal value (billions of dollars)

N Mean Std Dev Min 25% Pctl Median 75%Pctl Max Cash only 6584 0.38 1.22 0.0008 0.0318 0.1000 0.2950 41.0050 Stock 3025 0.87 5.27 0.0004 0.0281 0.0785 0.2869 164.7240 Mixed 2254 1.23 4.67 0.0002 0.0383 0.1370 0.6149 72.6710

Table 2 presents the descriptive statistics on the number of acquisitions by calendar year. The sample consists of 11,863 U.S. domestic completed deals described as in the section I A. Panel A describes the number of acquisitions and aggregate value of all deals by calendar year. Panel B describes the mean deal value of Categorized payments (cash only, stocks and mixed) deals by calendar year.



ISBN: 978-1-922069-76-4

5

Table 2 Panel A: Number of Acquisitions by Calendar Year

Year Number of Number of Number of Aggregate value Total number of Cash deals Stock deals Mixed deals of all deals (billions $) all deals 1993 204 136 59 94.3 399 1994 235 250 91 137.7 576 1995 259 259 81 204.8 599 1996 293 267 116 285.4 676 1997 344 367 192 475.2 903 1998 333 382 157 854.3 872 1999 306 337 155 748.8 798 2000 293 327 141 836.7 761 2001 279 150 139 396.4 568 2002 306 63 114 188.4 483 2003 383 56 97 225.9 536 2004 355 67 102 326.7 524 2005 377 58 135 499.5 570 2006 454 49 142 465.9 645 2007 395 44 110 335.4 549 2008 271 30 73 308.6 374 2009 210 49 58 334.2 317 2010 334 34 60 275.6 428 2011 326 38 61 343.3 425 2012 333 30 87 242.8 450 2013 294 32 84 300.0 410 Total 6584 3025 2254 7879.9 11863



ISBN: 978-1-922069-76-4

6

Panel B: Mean Deal Value of Categorized Payments by Calendar Year (billions $) Year Mean value of Mean value of Mean value of Cash deals Stock deals Mixed deals 1993 0.118 0.257 0.596 1994 0.285 0.159 0.340 1995 0.220 0.423 0.473 1996 0.243 0.470 0.765 1997 0.326 0.525 0.888 1998 0.322 1.308 1.576 1999 0.384 0.968 1.968 2000 0.382 1.541 1.567 2001 0.309 0.676 1.502 2002 0.236 1.307 0.298 2003 0.241 1.320 0.617 2004 0.445 1.917 0.747 2005 0.361 1.658 1.979 2006 0.398 1.659 1.436 2007 0.542 0.508 0.899 2008 0.629 2.594 0.828 2009 0.332 0.448 4.181 2010 0.441 1.181 1.471 2011 0.537 0.884 2.205 2012 0.431 0.594 0.937 2013 0.668 0.797 0.931

Descriptive Statistics on Analyst Recommendation and Forecast Because we are more interesting to see whether cash only deal has more favorableness over pure stocks in analyst recommendation upgrade within the 90 days window of deal announcement date as described in the section I C, we exclude the mixed subsample in our final sample described in the section ID. Table 3 presents the analyst recommendation change on our final sample, we present analyst recommendation change on our final cash only subsample and pure stock subsample in this table.

Table 3 Descriptive statistics on analyst recommendation change N % Upgrade %Flat %Downgrade Final sample 25,601 35.03% 32.07% 32.90% Cash Only 16,133 35.97% 30.38% 33.65% Stocks 9,441 33.39% 35.01% 31.63%



ISBN: 978-1-922069-76-4

7

3. Probit Analysis of Analyst Recommendation Change In this section, we use probit model to analyze short term dynamic link between merger means of payment and analyst recommendation changes over the acquirer stocks around the M&A deal announcement date. We use the probit analysis instead of ordinary least squares regression because recommendation changes are discrete and ordinal. In all cases, we compute standard errors by clustering on calendar month to ensure the robustness to heteroscedasticity and make sure there exists arbitrary cross sectional and intra month serial correlation error. our first probit model tests whether cash only deal are more likely than pure stock deals to lead analysts to upgrade their recommendations about the acquirer stocks, as opposed to downgrade their recommendations about them or leaving their recommendations unchanged. Our first null hypothesis is that compared to pure stock deal, cash only deal will more likely lead the analysts to immediately leave their recommendations about the acquirer stocks unchanged or downgrade their recommendation about the acquirer stocks within 90 days window of the deal announcement date. According to Kolasinski and Kothari (2008), variables deal value, days between analyst recommendation change date and deal announcement date, analysts experience covering stocks, and market size for the acquirer stocks are key controlled variables for the analyst recommendation change. According to Luo et al. (2010), analyst’s previous recommendations dispersion has played an important role in influencing on the analyst recommendation changes. Combining the key controlled variables from the analyst recommendation change literatures, we have come up with the probit model as the following model M (1)

𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖 = 𝛽0 + 𝛽1 × 𝐷𝑖 + 𝛽2 × ln(𝑉𝑎𝑙𝑢𝑒𝑖) + 𝛽3 × ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗) + 𝛽4 ×

ln(𝑆𝑖𝑧𝑒𝑖) + 𝛽5 × ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) + 𝛽6 × ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) + 𝛽7 × ln(𝐷𝑎𝑦𝑠𝑗) + 𝜀𝑖…M (1)

Variable Upgrade is a variable that takes on one of two values: 1 if the analyst upgrade the acquirer stock 0 if the analyst downgrade or leave the recommendation unchanged.

The process of determining value of Upgrade is described in the section I C. 𝐷𝑖 is a dummy variable for the acquirer stock i on deal announcement date if the merger deal is

paid by cash only, then 𝐷𝑖 = 1, otherwise𝐷𝑖 =0. 𝑉𝑎𝑙𝑢𝑒𝑖 is the total nominal consideration the acquirer i paid for the transaction, including all cash, securities, and assumed debt on deal announcement date. 𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗 measures the analyst j ‘s recommendation

dispersion about the acquirer stock within -90 days of deal announcement date. 𝑆𝑖𝑧𝑒𝑖 values the nominal market capitalization of the acquirer i on deal announcement date. 𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗 is the number of years between the deal announcement date and the analyst

j’s first recommendation date in I/B/E/S;𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗 is the number of analysts j either

recommending or issuing forecast about the acquirer in the same calendar month as deal announcement month. 𝐷𝑎𝑦𝑠𝑗 is number of days between the date of the first



ISBN: 978-1-922069-76-4

8

recommendation date after deal announcement and an deal announcement date and 𝜀𝑖is the cross sectional error term. Table 4 provides the descriptive statistics as well as the on the variable definitions.

Table 4: Variable Definitions and Descriptive statistics

Upgrade 1 if the analyst upgrade the acquirer stock 0 if the analyst downgrade or leave the recommendation unchanged. Used as an

dependent variable in model M(1) and table 5.

𝐷𝑖 Indicates whether deal is cash only deal or pure stock deal. 𝑉𝑎𝑙𝑢𝑒𝑖 Total nominal consideration the acquirer i paid for the transaction. 𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗 Analyst j ‘s recommendation dispersion about the acquirer stock

within -90 days of deal announcement date. 𝑆𝑖𝑧𝑒𝑖 Nominal market capitalization of the acquirer i on deal announcement date. 𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗 Number of years between the deal announcement date and the

analyst j’s first recommendation date in I/B/E/S. 𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗 Number of analysts j either recommending or issuing forecast about

the acquirer in the same deal announcement calendar month. 𝐷𝑎𝑦𝑠𝑗 Number of days between the first recommendation date

after deal announcement and an deal announcement date Mean Std Dev Min 25% Pctl Median 75%Pctl Max

Upgrade 0.350 0.477 0.000 0.000 0.000 1.000 1.000 𝐷𝑖 0.631 0.483 0.000 0.000 1.000 1.000 1.000 ln(𝑉𝑎𝑙𝑢𝑒𝑖) 5.648 1.661 -2.590 4.544 5.561 6.798 11.398 ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗) -0.233 0.181 -1.903 -0.306 -0.235 -0.132 1.039

ln(𝑆𝑖𝑧𝑒𝑖) 16.543 1.825 10.731 15.204 16.634 18.110 20.025 ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) 7.125 1.061 1.099 6.494 7.286 7.964 8.907

ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) 3.634 1.061 0.000 2.944 3.829 4.554 4.984

ln(𝐷𝑎𝑦𝑠𝑗) 3.229 1.031 0.000 2.639 3.526 4.060 4.499

Table 5 presents the results for M(1) model. We model the probability that an analyst will upgrade the acquirer stock recommendation within 90 days window of the deal announcement date as a function of the merger means of payment and other controlled variables exhibited in table 4.

Table 5: Merger means of payment on Analyst Recommendation Change



ISBN: 978-1-922069-76-4

9

(Dependent Variable: 𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖)

Constant 0.2433 (6.21)***

𝐷𝑖 0.0199 (2.85)*** ln(𝑉𝑎𝑙𝑢𝑒𝑖) 0.0049 (2.22)**

ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗) -0.0997

(-5.60)*** ln(𝑆𝑖𝑧𝑒𝑖) -0.0048 (-1.87)* ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) 0.0227

(7.43)*** ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) 0.0142

(3.43)*** ln(𝐷𝑎𝑦𝑠𝑗) -0.0221

(-6.96)*** Notes: *** indicates standard error is significant at 1% significance level ** indicates standard error is significant at 5% significance level *indicates standard error is significant at 10% significance level Table 5 shows that compared to pure stock deals, the cash only deals will have 2% more chance to lead analysts to upgrade their recommendations about acquirer stocks within 90 days of deal announcement date. This strongly indicates that merger means of payment has a significant impact on the analyst recommendation change within 90 days of deal announcement date. We see cash only deal gains more favorableness than pure stocks deal over the analyst recommendation upgrade about the acquirer stocks. According to Kolasinski and Kothari (2008), analyst recommendation change is sensitive to the length of the time between deal announcement date and the analyst’s first recommendation after the deal announcement date. It is very interesting to examine day length impact, cash dummy impact and their interaction impact on the analyst recommendation change. Therefore, we carry out the M(2) model to examine those impacts. M(2) model is identical to M(1) model except M(2) model includes the day dummy and interaction between daydummy and cash dummy.

𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖 = 𝛽0 + 𝛽1 × 𝐷𝑎𝑦𝑑𝑢𝑚𝑚𝑦𝑖 + 𝛽2 × 𝐷𝑖 + 𝛽3 × (𝐷𝑖 × 𝐷𝑎𝑦𝑑𝑢𝑚𝑚𝑦𝑖) + 𝛽4 ×ln(𝑉𝑎𝑙𝑢𝑒𝑖) + 𝛽5 × ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑖) + 𝛽6 × ln(𝑆𝑖𝑧𝑒𝑖) + 𝛽7 × ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) + 𝛽8 ×

ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) + 𝜀𝑖…M (2)

We define daydummy in two cases. In a 7 day case, we require daydummy=1 if 𝐷𝑎𝑦𝑠𝑗 ≤

7, otherwise daydummy=0. In a 30 day case, we require daydummy=1 if 𝐷𝑎𝑦𝑠𝑗 ≤ 30,

otherwise daydummy=0. All other variables are defined in the table 5.



ISBN: 978-1-922069-76-4

10

As table 6 shows the impact of cash only deal on analyst recommendation change is sensitive to the length of time between the first recommendation date after deal announcement and its deal announcement date.

Table 6 days length impact on Analyst Recommendation Change

(Dependent Variable: 𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖) 7days 30days

Constant 0.1255 0.121 (3.46)*** (3.32)***

𝐷𝑖 0.0157 0.0083 (2.09)** (0.89)

𝐷𝑎𝑦𝑑𝑢𝑚𝑚𝑦𝑖 0.0152 0.0346 (1.16) (3.32)***

𝐷𝑎𝑦𝑑𝑢𝑚𝑚𝑦𝑖 × 𝐷𝑖 0.0313 0.0274 (1.87)** (2.10)** ln(𝑉𝑎𝑙𝑢𝑒𝑖) 0.0041 0.0041 (1.91)* (1.90)* ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗) -0.1114 -0.1112

(-6.33)*** (-6.30)*** ln(𝑆𝑖𝑧𝑒𝑖) -0.0031 -0.0019 (-1.22) (-0.77) ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) 0.0250 0.0243

(8.36)*** (8.15)*** ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) 0.0145 0.0114

(3.61)*** (2.81)*** According to Loh and Stulz(2011), recommendation changes are more likely to be influential when experiencing a high forecast dispersion firm. Therefore, it is interesting to examine the previous dispersion impact, cash dummy impact and their interaction impact on the analyst recommendation change. Therefore, we carry out the M(3) model to examine those impacts. M(3) model is identical to M(1) model except M(3) model includes the predispersiondummy and interaction between predispersion dummy and cash dummy.

𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖 = 𝛽0 + 𝛽1 × 𝑝𝑟𝑒𝑣𝑖𝑜𝑢𝑠𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑑𝑢𝑚𝑚𝑦𝑖 + 𝛽2 × 𝐷𝑖 + 𝛽3 × (𝐷𝑖 ×𝐷𝑎𝑦𝑑𝑢𝑚𝑚𝑦𝑖) + 𝛽4 × ln(𝑉𝑎𝑙𝑢𝑒𝑖) + 𝛽5 × ln(𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑖) + 𝛽6 × ln(𝑆𝑖𝑧𝑒𝑖) + 𝛽7 ×ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) + 𝛽8 × ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) + 𝜀𝑖…M (3)



ISBN: 978-1-922069-76-4

11

According to Jegadeesh and Kim (2009), 0.75 is a roughly the average dispersion in a general sample. Therefore, we define predispersiondummy=1 if 𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑗 > 0.75,

otherwise predispersiondummy =0.

Table 7 presents analyst previous dispersion impact on analyst recommendation change. As table 7 shows that the pure stock deal group with their analyst previous dispersion within -90 days of deal announcement date greater than 0.75 will have 3.16% negative impact on analyst recommendation change. This indicates that for pure stock deals, analysts are less likely to upgrade the recommendation about the acquirer stocks when there was above average dispersion across the analyst opinions. This is consistent with Jegadeesh and Kim (2009)’s finding that analysts are less likely to herd when there is a large dispersion across analysts’ opinion. In addition, the impact of cash dummy variable

on analyst recommendation change is 5.08%-4.69%×predispersiondummy. In this case, we can see as the predispersiondummy varies from 0 to 1, impact of cash dummy on analyst recommendation change varies from 5.08% to 0.39%. Therefore, the impact of cash only deal on analyst recommendation change is sensitive to the previous analyst recommendation dispersion about the acquirer stocks within -90 days of deal announcement date.



ISBN: 978-1-922069-76-4

12

Table 7 days Analysts Previous Dispersion Impact on Analyst Recommendation Change

(Dependent Variable: 𝑝(𝑢𝑝𝑔𝑟𝑎𝑑𝑒)𝑖) Constant 0.3088 (7.79)***

𝐷𝑖 0.0508 (4.17)***

𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑑𝑢𝑚𝑚𝑦𝑖 -0.0316 (-2.73)***

𝑝𝑟𝑒𝑑𝑖𝑠𝑝𝑒𝑟𝑠𝑖𝑜𝑛𝑑𝑢𝑚𝑚𝑦𝑖 × 𝐷𝑖 -0.0469 (-3.23)*** ln(𝑉𝑎𝑙𝑢𝑒𝑖) 0.0048 (2.19)** ln(𝑆𝑖𝑧𝑒𝑖) -0.0059 (-2.29)** ln(𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒𝑗) 0.0222

(7.28)*** ln(𝐹𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔𝑗) 0.0166

(4.00)*** ln(𝑑𝑎𝑦𝑠𝑗) -0.0234

(-7.40)***

4. Event Study and Instrumental Analysis of Cumulative Abnormal Return As table 5 show us that within 90 days of deal announcement date, cash only deal will have a higher chance than pure stocks to lead analysts to upgrade their recommendations about acquirer stocks We are interesting to examine whether this recommendation change difference is caused by different merger means of payment will be reflected in a short term market reaction. Therefore, we conduct an event study to examine the short term cumulative abnormal return of cash only deal and that of pure stock deal separately within three days window of deal announcement. To better measure the short term abnormal return, we employ the common technique of calculating cumulative abnormal return relative to a beta benchmark. To calculate abnormal returns based on market beta, we use the procedures documented in Boehmer et al (2002, P44). According to Mitchel et al. (2004), to disentangle price pressure and information effects, we will use 3 day cumulative abnormal return around deal announcement date t. To be consistent with Mitchel et al. (2004), market model parameters are estimated over a 150 day window beginning 21 days after deal announcement date, where value weighted CRSP included dividends index proxies for



ISBN: 978-1-922069-76-4

13

market portfolio. After a deal announcement date for acquirer stock I, we compute 3 day buy-and-hold abnormal returns ABRi (t-1, t +1) as model M(4)

𝐴𝐵𝑅𝑖(𝑡 − 1, 𝑡 + 1) = ∏ (1 + 𝑅𝑖,𝜏) − ∏ (1 + 𝑅𝑚,𝜏)

𝑡+171

𝜏=𝑡+21

𝑡+1

𝜏=𝑡−1

𝐶𝐴𝑅𝑖(𝑡 − 1, 𝑡 + 1) = ∑ 𝑎𝑣𝑒𝑟𝑎𝑔𝑒(𝐴𝐵𝑅𝑖)𝑡+1𝑡−1 ....M (4)

Where 𝑅𝑖,𝜏 and 𝑅𝑚,𝜏 are the return on acquirer stock i and the value-weighted index return,

respectively. We choose the [-1 1] or from one day prior to one day after deal announcement date for the event window. We choose the [21 171] or from 21 days to 171 days after deal announcement date as our market estimation window. Market portfolio returns are collected from value weighted CRSP included dividends index proxies returns from CRSP. These returns within [21 171] are then used as benchmarks to calculate the abnormal performance. Abnormal returns are calculated for each firm relative to its beta benchmark in [-1 1] time frame. Cumulative abnormal returns are calculated by averaging across acquirer firms every day and then summing those averages over time. Table 8 reports cumulative abnormal returns for cash only deal and pure stock deal within 3 days window of deal announcement date [-1 1]. 3 days window can disentangle merger means of payment impact on short term abnormal return of acquirer stocks from other impacts. As presented in table 8, we can clearly see cash only deal delivers a significant positive cumulative abnormal return or 1.06% for acquirer stocks and pure stock deal delivers a significant negative cumulative abnormal return or -1.05% for acquirer stocks. This cumulative abnormal return difference reflect market prefers a cash only deal to a pure stock deal during short time after the deal announcement. But would their respective abnormal return explained by the analyst recommendation change? If find the answer is yes, then we can see merger means of payment impact on analyst recommendation change will be translated into a short term abnormal return for acquirer stocks.

Table 8 Cumulative Abnormal Returns for Cash Only and Stocks Deals Period Cash only Stocks (N=2,700) (N=1,213) [−11] 1.06% -1.05% (8.97)*** (-3.84)***



ISBN: 978-1-922069-76-4

14

5. Summary and Conclusions We find that compared to pure stock deals, cash only deal will receive more favorableness from analysts to upgrade their recommendations about the acquirer stocks within 90 days window of deal announcement date. In particular, compared to pure stock deals, the cash only deals will have 2% more chance to lead analysts to upgrade their recommendations about acquirer stocks within 90 days of deal announcement date. In addition, we find that the impact of cash only deal on analyst recommendation change is sensitive to the length of time between the first recommendation date after deal announcement and its deal announcement date. The impact of cash only deal on analyst recommendation change is sensitive to the previous analyst recommendation dispersion about the acquirer stocks within -90 days of deal announcement date. We also find that analyst recommendation change will have a significant positive impact on the short term cumulative abnormal return for the acquirer stocks within 3 days of deal announcement date. As event study enables us to show that cash only deal delivers a significant positive cumulative abnormal return or 1.06% for acquirer stocks and pure stock deal delivers a significant negative cumulative abnormal return or -1.05% for acquirer stocks. Our key results above indicate that merger means of payment has a direct impact on analyst recommendation change and market immediately reacts to this impact, resulting different cumulative abnormal returns for acquirer stocks by various merger means of payment.

Reference Agrawal, A., Jaffe, J.,2000. The post-merger performance puzzle. In: Cooper, C., Gregory,

A., (Eds.), Advances in Mergers and Acquisitions, Vol.41 no.1, pp. 7–41. Boehmer, E., Broussard, J. P., & Kallunki, J. 2002. Using SAS in financial research. Fama, E.F. 1992, "The cross-section of expected stock returns", The Journal of Finance,

vol. 47, no. 2, pp. 427-465. Jegadeesh, N. & Kim, W. 2010, "Do Analysts Herd? An Analysis of Recommendations

and Market Reactions", The Review of Financial Studies, vol. 23, no. 2, pp. 901-937.

Kolasinski, A.C. & Kothari, S.P. 2008, "Investment Banking and Analyst Objectivity: Evidence from Analysts Affiliated with Mergers and Acquisitions Advisors", Journal of Financial and Quantitative Analysis, vol. 43, no. 4, pp. 817-842.

Loh, R.K. & Stulz, R.M. 2011, "When are analyst recommendation changes influential?", The review of financial studies, vol. 24, no. 2, pp. 593-627.

Loughran, T.J. 1997, "Do long-term shareholders benefit from corporate acquisitions?", The Journal of Finance, vol. 52, no. 5, pp. 1765-1790.



ISBN: 978-1-922069-76-4

15

Lin, H. & McNichols, M.F. 1998, "Underwriting relationships, analysts' earnings forecasts and investment recommendations", Journal of accounting & economics, vol. 25, no. 1, pp. 101-127.

Rau, P.R. 1998, "Glamour, value and the post-acquisition performance of acquiring firms", Journal of Financial Economics, vol. 49, no. 2, pp. 223-253.

Savor, P.G. & Lu, Q. 2009, "Do stock mergers create value for acquirers?", The journal of finance, vol. 64, no. 3, pp. 1061-1097.

Shleifer, A. & Vishny, R.W. 2003, "Stock market driven acquisitions", Journal of Financial Economics, vol. 70, no. 3, pp. 295-311.

Womack, K.L. 1996, "Do brokerage analysts' recommendations have investment value?", The Journal of Finance, vol. 51, no. 1, pp. 137-167.

Merger Means of Payment and Analyst Recommendation Change

Documents