BLANK CHECK ACQUISITIONS

Transcription

1 BLANK CHECK ACQUISITIONS Anh L. Tran * November 16, 2009 Abstract Special Purpose Acquisition Companies (SPACs), a particular type of blank check firms, have risen dramatically since 2003 and account for one third of the U.S. IPO market in Compared to other public acquirers, SPACs tend to make focused acquisitions, target mostly private companies, are less likely to do cash only deals or tender offers, and less likely to complete acquisitions. SPACs exhibit a mean cumulative abnormal return of 1.7% upon deal announcement and a monthly excess return of 1.5% from deal announcement until closing. I show that SPACs make better acquisitions: they negotiate an additional 7.6% discount compared to other public acquirers that bid for private targets. The results highlight the role of specialization, ownership structure, and independent long-term institutional blockholders monitoring as important corporate governance mechanisms in SPACs. JEL Classification: G34 Keywords: Blank check; Acquisitions; Bidder returns * Ph.D. candidate (expected graduation June 2010). Contact information: LeBow College of Business, Drexel University, 3141 Chestnut St, Philadelphia, PA Phone: (215) anh@drexel.edu. I am deeply indebted to Eliezer Fich, my advisor, for providing invaluable feedback and support. I also thank Jie Cai, Naveen Daniel, Daniel Dorn, Jacqueline Garner, Jarrad Harford, Lalitha Naveen, Ralph Walkling and seminar participants at Drexel University and the 2009 FMA Doctoral Consortium in Reno for their helpful comments. I am grateful to Brian Bushee for sharing his institutional investor classification data. All errors are my responsibility.

2 BLANK CHECK ACQUISITIONS Abstract Special Purpose Acquisition Companies (SPACs), a particular type of blank check firms, have risen dramatically since 2003 and account for one third of the U.S. IPO market in Compared to other public acquirers, SPACs tend to make focused acquisitions, target mostly private companies, are less likely to do cash only deals or tender offers, and less likely to complete acquisitions. SPACs exhibit a mean cumulative abnormal return of 1.7% upon deal announcement and a monthly excess return of 1.5% from deal announcement until closing. I show that SPACs make better acquisitions: they negotiate an additional 7.6% discount compared to other public acquirers that bid for private targets. The results highlight the role of specialization, ownership structure, and independent long-term institutional blockholders monitoring as important corporate governance mechanisms in SPACs. JEL Classification: G34 Keywords: Blank check; Acquisitions; Bidder returns

3 1. Introduction Special purpose acquisition companies or SPACs for short are public shell companies that raise capital from the general investing public for the purposes of implementing a merger, asset acquisition or similar business combination of an operating company. SPACs are regarded as a "blind pool" of capital, as investors do not know the operating company that the SPAC managers will ultimately invest their money in. SPACs are also referred to as blank-check companies since investors essentially give the SPAC managers a blank check with which to make an acquisition within a limited time frame. In many respects, SPACs are similar to reverse mergers in that a blank shell company looks to acquire a standing operating company. Like other recent financial innovations, SPACs have grown more popular since 2003 and developed into one of the most important sectors of the U.S. IPO market. They account for one third of the IPO market in 2008 in terms of both the number of offerings and the total money raised. Such market share is a significant increase from 2007, a year in which SPACs account for 20% of the U.S. IPO market. In 2007, the total proceeds from SPAC IPOs exceed 11 billion dollars. Giving this rapid surge in SPAC importance, banking giants such as Deutsche Bank, Citigroup, and Merrill Lynch are now active participants in underwriting and advising SPACs. Moreover, in recent years, investment companies, advisors, and funds such as Stanford Investment Group, Millennium Management, Baupost Group, Pequot Capital Management, and Amaranth Advisors have raised their ownership stakes in SPACs. 1 Given the large amount of public capital invested in SPACs in recent years and because the sole purpose of SPACs is to acquire a standalone firm, in this paper, I investigate whether SPACs buy targets at better terms. Put differently, my goal is to test whether the bidder gains from acquisition differ if the bidding firm is a SPAC. 1 Stanford Investment Group is not affiliated with Allen Stanford or his companies. Mr. Stanford was charged by the Securities and Exchange Commission (SEC) for orchestrating a fraudulent multibillion dollar investment scheme related to an 8 billion dollar CD program (Washington D.C., Feb 17, 2009). In addition, Amaranth was liquidated in October

4 To test this research question, I analyze a sample of 3130 acquisitions from 2004 to This sample constitutes the entire universe of acquisition offers in the Securities Data Company s merger database that meet the criteria outlined in Moeller, Schlingemann, and Stulz (2005) and Masulis, Wang, and Xie (2007). This sample includes 108 offers (about 3.5%) by SPACs and 3022 offers by different publicly traded bidders. 2 My analyses show that SPACs acquisitions have characteristics different from those by other public acquirers. In particular, SPACs make focused acquisitions, target private companies over 90% of the time, are 22% less likely to do cash only deals, and 33% less likely to complete a merger. In addition, SPACs exhibit a mean three-day cumulative abnormal return (CAR) of 1.7% upon the acquisition announcement, which is significantly greater than the market adjusted return of 0.33% by other public acquirers. I estimate the mean value appreciation over this three-day announcement window to be about 5 million dollars for SPACs shareholders. 3 Given the short-term nature of the event study results and to further examine the difference in performance between SPACs and other public acquirers, I form portfolios according to the type of bidder as follows. A bidder enters the portfolio two days after the acquisition announcement and exits the portfolio once the acquisition is completed or withdrawn. Daily returns are calculated for each of the two portfolios and benchmarked against the four-factor model. 4 Based on this analysis, I find that SPACs exhibit an average four-factor excess return of 1.5% per month, which is significantly larger than that by other bidders (0.08% per month). Moreover, when compared to other anomalies in the literature such as the 0.8% per month return from merger arbitrage in Baker and Savasoglu (2002) or the -0.5% per month return from IPO underperformance in Brav and 2 The 108 offers by SPACs that I analyze come from the entire set of 161 publicly SPACs operating as of June 19, SPACs are dropped from the sample because these companies have not made an acquisition offer as of the cutoff date. 3 I estimate this value following Moeller, Schlingemann, and Stulz (2005). To obtain this value, I multiply the raw return over the three-day announcement window with the market value of equity for each SPAC and then average this product across all SPACs. 4 See Fama and French (1996) and Carhart (1997). 2

5 Gompers (1997), the magnitude of the abnormal returns accruing to the sample SPACs in this study is meaningful. The acquisition announcement returns as well as the portfolio analysis suggest that when compared with other public bidders SPACs make better acquisitions. Nonetheless, I investigate this issue further. I notice that 90% of the recipients of the SPAC related acquisition bids are private targets. Officer (2007) finds that acquirers buying private targets pay lower premia than those buying public targets in similar transactions. Using the comparable industry transaction technique in Officer (2007), I find that SPACs obtain an acquisition discount that is 7.6% higher than that of other public acquirers that also bid for private targets. The average economic effect associated with this additional discount obtained by SPACs is 23 million dollars which is captured by SPAC shareholders. This discount is non-trivial given that the average size of acquisitions by SPACs is 230 million dollars. This result provides direct evidence that SPACs make better acquisitions. Next, I study how and why SPACs obtain this additional discount. I find that the unique feature of SPACs in making focused acquisitions and their ownership structure, with higher managerial and independent long-term institutional blockholder stakes, partly explain why SPACs make better acquisitions. In addition, I find that the time constraints and the employment compensation-ownership arrangement unique to SPAC managers might prompt the SPAC management into making unsuitable acquisitions. However, such perverse incentives are partially mitigated under the monitoring by independent long-term institutional blockholders. Overall, my results contribute to the literatures related to the recent financial innovations, to the conditions in which mergers are profitable, and to the importance of the ownership structure in acquiring firms. The paper also highlights the role of independent long-term institutions as a crucial component of corporate governance mechanisms in SPACs. The paper proceeds as follows. Section 2 provides background on SPACs. Section 3 develops the hypotheses. Section 4 describes the data. Section 5 contrasts the deals structured by 3

6 SPACs with those by other public bidders. Section 6 tests whether SPAC bidders make better acquisitions. Section 7 examines the sources of efficiency gains by SPACs. Section 8 studies the SPAC deal pressure. Section 9 addresses some robustness issues. Section 10 concludes. 2. Background on SPACs 2.1. SPACs as publicly traded firms SPACs are typically sold by way of an IPO, structured as a sale of units consisting of both common stock and "in the money" warrants that can be converted within a previously specified time frame. 5 As with the common stock, the warrants can be traded after the initial offering period. A minimum of 85% of the cash generated in the IPO is placed in a trust that cannot be touched by the SPAC s management until an acquisition is made. Until then, this money is typically invested in U.S. short-term government securities. These funds are released upon the earlier of the completion of a business combination or the liquidation of the SPAC. The remaining proceeds are used for fees, expenses and working capital. SPAC managers do not receive cash compensation. However, they are awarded a 20% ownership stake of the acquisition deal if such transaction is completed. 6 Once the IPO is completed, (1) SPACs have to file a Current Report on Form 8-K with the Securities and Exchange Commission (SEC) to disclose the IPO proceeds on an audited balance sheet, and (2) SPAC s securities may begin to trade in public stock exchanges. While most SPACs have been listed on the OTC Bulletin Board (OTCBB) and the American Stock Exchange 5 For instance, a common share may be initially offered for $6 or $8 per unit, and for every common share offered there will be either one or two warrants issued, with each reflecting the right to purchase a common share in four years. Each warrant is then exercisable beginning on the later of the acquisition completion and one year after the IPO, and ending four years after the IPO. If the common stock is trading at a substantial premium (approximately 140%) for 20 trading days within a 30 day trading period, the warrants are redeemable for one cent each when first exercisable. 6 To be precise, the founding stockholders purchase the SPAC s common stock for nominal consideration and generally retain 20% of the SPAC s common equity after the completing the IPO. Due to potential conflicts of interest, industry norms have formed. Consequently, a SPAC cannot form a business combination with a target where an insider has a financial stake, in order to protect the interest of all shareholders (an exception is if a fairness opinion is issued by a third party to qualify the deal). 4

7 (AMEX), some recent SPACs have been approved for listing on the New York Stock Exchange (NYSE). Both AMEX and NYSE require SPACs to satisfy their listing standards SPACs vs. blank check companies Blank check companies have been around for several years, mostly operating under Rule 419 of the Securities Act of 1933 and the Penny Stock Reform Act of Recent SPAC deals are exempt from Rule 419 because they are not penny stock IPOs. Moreover, unlike firms subject to Rule 419, SPAC units (a combination of shares and warrants) sell at more than 5 dollars each and after their IPOs SPACs have more than 5 million dollars in net tangible assets. Units under a SPAC offering may begin trading after the prospectus date while those under Rule 419 cannot legally trade until after the completion of a business combination. Moreover, SPAC IPO proceeds may only be invested in U.S. government securities, while those under Rule 419 can be invested in several other securities. 8 Some similarities between SPACs and companies operating under Rule 419 are as follows. Both are subject to provisions that help protect investors by keeping the IPO proceeds in an escrow account until after the completion of a business combination. If a business combination goal is not met within a specified amount of time, the money deposited in the escrow account is returned to the investors. In addition, the initial target company considered for a potential business combination must have a fair market value of at least 80% of the SPAC s net assets at the time of the acquisition (or at least 80% of the maximum offering proceeds under Rule 419) Acquisition negotiation process SPACs are created with a sole purpose to make business combinations. Given that the time for SPACs to complete an acquisition is within 18 to 24 months from their own IPO, the search 7 Given the similarity between SPACs and the blank check companies of the 1990s that were the subject of pump-and-dump scams, suspicion is understandable. In 1997, securities regulator the NASD fined underwriter GKN Securities (predecessor to EarlyBird Capital) more than 2 million dollars for controlling after-market trading in the blank check warrants and artificially inflating prices. Since that time, regulators have been nervous about similar offerings. 8 For example, money market funds meeting conditions of the Investment Company Act of 1940 or securities that are direct obligations of or are guaranteed by the United States 5

8 process for a suitable target is very intense. 9 As part of the efforts to identify potential targets, the SPAC s management typically seeks the advice from investment bankers, private equity professionals, business brokers, business owners, and lawyers. During this process, SPACs generally consider a large number of potential target candidates. To narrow the field, SPACs solicit confidential information from the most promising target firms. In my sample of 108 SPACs that go public during and that formalize a bid for a target as of June 19, 2009, I find that the average SPAC signs confidentiality agreements and receives confidential information from 30 different potential targets. After reviewing this confidential information, SPACs typically submit preliminary and non-binding acquisition proposals to about 5 potential targets. These firms are those that the SPAC s management believes to be sufficiently attractive to warrant further consideration. In the end, SPACs identify the most suitable target firm to pursue. Once the merger is executed, the deal has to be approved by the SPAC s shareholders. The terms of a typical SPAC offering allow stockholders to vote on a proposed business combination, even though such approval may not be required under state law. Depending on the SPAC, a majority of votes (60%-80% of its shareholders) is needed in order for the merger to be approved. This high threshold creates a higher than usual rate of deals that are not completed Academic studies on SPACs To date, the academic literature on SPACs is sparse. Most papers in the subject provide an overview of SPAC s features. For instance, Hale (2007) describes the SPAC as a financing tool that has something for management, investors, and target companies. Hale argues that this vehicle provides the SPAC s management with a way to capitalize on their expertise, target companies with access to potential public investors, and investors with a new investment choice. Davidoff (2008) also describes the characteristics of SPACs on the corporate side while Sjostrom (2008) shows how SPACs are different from other blank check companies on the legal side. Finally, 9 An extension is possible in order to complete a deal if a SPAC already enters into a definitive agreement with a certain target company. 6

9 Berger (2008) provides a clinical study of SPACs with three case studies. The goal of these cases is to illustrate how SPACs can offer an alternative way for a company to access the public markets. In terms of empirical financial research on SPACs, to my knowledge, only three papers examine some financial characteristics of SPACs. Jog and Sun (2007) study 62 SPAC IPOs and find a relatively small IPO underpricing. They also document negative post-ipo returns for SPAC investors and positive returns for SPAC management. Lewellen (2008) finds that SPAC s return patterns are highly predictable and highly unusual, and concludes that SPACs are structured and behave unlike any other asset class in public equity markets. Jenkinson and Sousa (2009) propose a trading rule using SPACs that is associated with low risk profits. Unlike any of these papers, my study examines whether SPACs make better acquisitions than other public acquirers do, and if so why. 3. Hypotheses 3.1. Specialization One characteristic that distinguishes the acquisitions executed by SPACs from those by different bidders is the fact that the combined firm resulting from a SPAC acquisition always operates in a single industry. As soon as the acquisition consummates, the SPAC s SIC code 6770 disappears and the combined firm retains the target firm s SIC code. Consequently, unlike acquisitions by different bidders, SPAC deals are always structured as focused transactions. Jensen and Ruback (1983) suggest that focused transactions can create value from the replacement of less efficient with more effective managers. Through focusing, firms can also enhance value by (1) reducing cross-subsidization among business segments (Meyer, Milgrom, and Roberts, 1992), (2) decreasing overinvestment in lines of business with poor investment opportunities (Stulz, 1990), and (3) curtailing value-decreasing investments by rent-seeking managers (Jensen, 1986, and Scharfstein and Stein, 2000). In the context of mergers, Mørck, Schleifer, and Vishny (1990) find 7

10 that focused mergers create positive abnormal returns for bidders while diversified deals destroy value. 10 The value SPACs bring to their shareholders might depend on the expertise of the SPAC management team. SPAC founders often have an established record of career success. They generally are financial sponsors, asset managers, or corporate executives with extensive networks of contacts and prominent industry experience that help them find and organize specialized transactions that are difficult to replicate. 11 Previous literature documents that individuals with superior abilities and track records add value. For instance, Laderman (1994) reports that mutual fund managers graduated from Ivy League schools outperform their non-ivy counterparts. Chevalier and Ellison (1999) find that mutual fund managers attending higher-sat undergraduate universities generate superior portfolio returns. The related industry experience and managerial expertise of SPAC managers might be associated with the search for a suitable target and the execution of a value-enhancing acquisition. Since SPACs are formed with the sole purpose of making business combinations, the specialization hypothesis predicts that the sources of SPAC s specialization in executing a focus acquisition and in managerial expertise enhance shareholder value. 10 However, firms created through diversifying mergers can also enhance value by lowering the cost of capital with an effective internal capital market (Weston, 1970). Diversification can also lead to a lower cost of capital for the firm due to the availability of more uncorrelated investment projects (Stein, 1997), economies of scale (Chandler, 1977), or coinsurance through greater debt capacity (Lewellen, 1971). 11 An example of such executives is Richard J. Heckmann of Heckmann Corp. In 1971, Mr. Heckmann founded and became chairman of the board of Tower Scientific Corp. From 1978 to 1979, he was the associate administrator of the Small Business Administration in Washington, D.C. In 1990, Mr. Heckmann founded United States Filter Corp and was its CEO. In 2002, he became CEO and chairman of K2 Inc. He also served as chairman of the Listed Company Advisory Committee of the NYSE from 2001 to 2003 and was appointed to its Corporate Accountability and Listing Standards Committee in In addition, Mr. Heckmann served on the board of directors of many companies including MPS Group Inc, Philadelphia Suburban Corp, Station Casinos Inc, United Rentals Inc, and Waste Management Inc. Mr. Heckmann has extensive experience with business acquisitions. During his time with US Filter and K2, US Filter made over 150 acquisitions with a maximum deal value of $1.7 billion and K2 consummated over 20 acquisitions with the largest deal valued at $150 million. 8

11 3.2. Deal pressure The SEC imposes a time limit for SPACs to complete an acquisition. Specifically, SPACs have a maximum of 24 months from their own IPO to acquire another company. Extensions are possible if a SPAC can demonstrate that it has already entered into a definitive merger agreement with a target company. Otherwise, once the 24 month time limit elapses, SPACs have to liquidate. 12 Given the time constraints, there are obvious concerns that the pressure to complete an acquisition might prompt the SPAC management into making unsuitable acquisitions. Moreover, during the 24 months, SPAC managers are not officially compensated and they are to receive an ownership stake in the combined firm (typically 20%) only if such combination comes to fruition. This employment compensation-ownership arrangement unique to SPAC managers might also pressure them to execute an unsuitable acquisition. Based on the above discussion, I also test the deal pressure hypothesis. This hypothesis predicts that the distinct features of SPACs of having to complete deals in 24 months as well as the unique employment arrangements of its managers might prompt them to execute unsuitable acquisitions Hypotheses testing One possible implication of the time pressure hypothesis is that, as the 24 month deadline to complete an acquisition approaches, SPAC managers might be more likely to bid for an unsuitable target. However, under the specialization hypothesis one would expect managers to do a thorough (and perhaps time consuming) due diligence on a prospective target which might explain a late bid. Moreover, even if SPAC bids are more likely to be bad or closer to the 24 month deadline, it is possible that such potentially bad bids might not materialize if SPAC managers are appropriately monitored. In this vein, Amihud and Lev (1981) find that, unless closely monitored by large block shareholders, managers of bidding firms will attempt to reduce their employment 12 In addition, SPACs cannot form a business combination with a target where an insider has a financial stake in order to protect shareholder interests. 9

12 risk through unsuitable acquisitions. Moreover, Chen, Harford, and Li (2007) argue that independent institutions with a long-term investment goal will closely monitor and influence efforts on the management of bidding firms. The above discussion yields several implications for my hypotheses testing. First, I will examine whether bidder specialization in deal-making and bidder ownership structure are related to the wealth effects to bidding shareholders. Second, I will evaluate whether timing of a SPAC acquisition bid is related to the bid s market announcement returns and the bid s ultimate outcome Econometric issues As I will explain in detail later in the data section of this paper, I analyze a sample of 108 SPAC bids together with a sample of 3,022 bids by different publicly traded acquirers in the SDC database from January 2004 to May Moreover, I follow the existing literature and match my SPAC bidders with other acquirers and use the matching sample for robustness tests. However, because the 24 month deadline to complete an acquisition and the employment arrangements to SPAC managers are unique to SPACs, in some tests I have to analyze the SPAC firms in isolation. This issue creates several empirical challenges related to the statistical power of my tests. However, several authors provide helpful guidance on how to address similar situations. Wilcoxon (1945) proposes a signed-rank nonparametric test that can be used when the population cannot be assumed to be normally distributed. These nonparametric tests are particularly useful when the sample size is small and hence the sample distribution cannot be inferred (Kruskal, 1952). Moreover, Efron (1979) develops nonparametric bootstrapping methods to estimate the sampling distribution of a statistic empirically without making assumptions about the form of the population and without deriving the sampling distribution explicitly. Freedman (1981) validates the bootstrapping regression models by showing that the coefficients distribution converges to the normal under the assumptions of predetermined variables and i.i.d. errors Previous literature also relies on nonparametric tests to address this sample size issue. For example, Chance, Kumar, and Todd (2000) study 53 instances of executive stock option repricing during

13 4. Data I analyze a sample of 3130 acquisitions from January 2004 to May This sample constitutes the entire universe of acquisition offers in the Securities Data Company (SDC) merger database that meet the criteria outlined in Moeller, Schlingemann, and Stulz (2005) and Masulis, Wang, and Xie (2007). Like those authors, I require (1) that the bidder controls less than 50% of the target s shares prior to the announcement and owns 100% of the target s shares after the transaction; (2) that the bidder has stock return data during 20 trading days before the acquisition announcement date; and (3) that deal value is at least 1 million dollars. As of June 19, 2009, there are 114 SPACs that, after bidding for a target, either complete their acquisitions or liquidate. From this group, I lose 5 observations in which SPACs could not find a target and 1 observation in which information on the target is not available. Consequently, in my sample of 3130 acquisition offers, 108 bids (or 3.5%) are by SPACs. 14 I collect data on SPACs from the EDGAR database for public companies listed under SIC code 6770 and from the SPAC Market Update Reports by Morgan Joseph & Co. Inc. In addition, I track all SPACs in my sample from the time they become public companies with information provided in the IPO section of SDC. I supplement my SPAC data with information in the S-1 (Prospectus), S-4 (Registration/Proxy), 8-K (Current Report), 10-K (Annual Report) and other proxies from the SEC. I obtain accounting data from Compustat and stock return data from the Center for Research in Security Prices (CRSP) and FinancialContent Services Inc. In addition, I collect managerial ownership data from proxies filed by the acquirers with the SEC once the offer They use nonparametric signed rank tests to show that their results are robust to the small sample size. Chatterjee, Dhillon, and Ramirez (1995) examine a sample of 46 public workouts for distressed high-yield debt during They employ the Mann-Whitney-Wilcoxon nonparametric tests to investigate the differences between 16 firms that use tender offers and 30 firms that use exchange offers. 14 The first acquisition by a SPAC in the sample happens in March 2004 for the SPAC that goes IPO in The last acquisition by a SPAC in the sample happens in May 2009 for the SPAC that goes IPO in Only two SPACs that go IPO in 2008 complete their acquisitions as of June 19, I exclude 12 SPAC IPOs that trade on international exchanges (Euronext and Alternative Investment Market, a submarket of the London Stock Exchange). 11

14 is made. I use the Thomson Reuters Institutional Holding database (formerly CDA/Spectrum) to obtain ownership by institutions at the quarter before merger announcement. Table 1, Panel A provides information on firms that go IPO in the U.S. over the period. I observe that 66 SPACs go public in 2007, accounting for about 20% of the total number of IPOs in the U.S. during that year. Moreover, as a fraction of total IPOs in the U.S. market SPACs exhibit an annual increase over time. The total capital raised by SPACs in 2007 is almost 12 billion dollars, which accounts for 14% of the total proceeds of all IPOs in the market during that year. In fact, the average proceeds raised by a SPAC at IPO increase every year during the sample period. Panel B of Table 1 shows the acquisition status of SPACs as of June 19, For instance, among 66 SPACs that go IPO in 2007, 14 complete their acquisitions, 22 liquidate, 12 announce their potential deals, and 18 are still looking for a target. Among 161 SPACs that go IPO from 2003 to 2008, 108 either complete their acquisitions or liquidate, 6 are lost as described earlier, and the fate of 47 others is unknown as of this cutoff date. 5. Deal structure 5.1. Financial characteristics Table 2 Panel A reports the financial characteristics of parties in acquisition deals involving SPACs vs. those involving other public acquirers at the end of the fiscal year before the merger announcement date. The mean (median) value of deals involving SPACs as bidders is 230 (107) million dollars. This mean value is significantly smaller than that of deals by other public acquirers. However, the median deal value is not statistically different between the two groups. The relative size between targets and acquirers is significantly greater in SPAC deals. 15 The average 15 As in Moeller, Schlingemann, and Stulz (2005) and Masulis, Wang, and Xie (2007), when deals involve nonpublic targets, relative size is defined as deal value divided by the acquirer s market value of equity. In the robustness section, I show that the relative size difference between deals by SPACs and those by other public acquirers does not qualitatively drive the results. 12

15 (median) SPAC has a market capitalization of 134 (79) million dollars, which is significantly smaller than that of other public acquirers. SPACs Tobin s Q and return on assets (ROA) are significantly smaller than those of other public acquirers. The difference is significant at the 1% level. The latter result is not surprising because SPACs only earn the interest on the cash deposited in their escrow accounts Deal characteristics In this section, I examine the deal characteristics of my sample firms sorted by bidder type. These characteristics include: the type of target firm, the form of the deal, deal attitude, the method of payment, and the deal outcome. Panel B of Table 2 reports the univariate results related to these characteristics Univariate tests From the information reported in Panel B of Table 2, I observe that among 108 targets approached by SPACs, only 9 (or 8%) are publicly traded. This incidence is statistically lower than the 42% of other public acquirers seeking private targets at the 1% level using a nonparametric test for the difference in proportions. 16 Given that over 90% of SPAC deals involve non-publicly traded targets, the methodology I later use to estimate the acquisition premia and the wealth effects to bidder shareholders in these deals addresses this issue. Tender offers account for 8% of deals by other public acquirers. In contrast, a majority of SPAC deals (99%) are structured as merger and acquisition (M&A) transactions. It is possible that the goal of organizing the transaction as an M&A is that the target receives the SPAC s stock as the form of payment. In this vein, Panel B of Table 2 shows that 77% of SPAC deals involve stock while that proportion is 34% for other public acquirers. Conversely, 42% of other public acquirers pay for the transaction entirely in cash. Only 19% of SPAC bidders pay entirely in cash. Such difference is significant at the 1% level. 16 The z statistic is estimated as follows: z P 1 P 2 P(1 P) P(1 P) N 1 N 2 with P 1 N 1 P 2 N P 2 N 1 N 2 13

16 From Panel B of Table 2, I note that all offers by SPACs are focused acquisitions while 64% of those by other public acquirers fall into this category. Before the acquisition, all SPACs operate under SIC code As soon as the acquisition consummates, this SIC code disappears and the combined firm retains the target s SIC code. In addition, the proportion of friendly deals made by SPACs is 96% while that by other public acquirers is 95%. The z statistic shows an insignificant difference between the two groups with respect to deal attitude. Moreover, SPAC deals are completed only 62% of the time. In contrast, the completion rate for deals by other public acquirers is 93%. The difference is statistically significant at the 1% level Multivariate tests In the multivariate analysis reported in Table 3, I test whether SPACs are less likely to use only cash as the medium of exchange. I perform year- and industry-fixed effect logistic regressions in which the dependent variable is one if the deal is paid entirely in cash and zero otherwise. The independent variables are similar to those in Baker, Coval, and Stein (2007) and Officer, Poulsen, and Stegemoller (2008) and defined in the Appendix. These authors also examine the method of payment in acquisitions. The independent variable of interest is a SPAC (0,1) indicator that takes the value of one if the bidder is a SPAC and zero if it is not. The results show that the SPAC (0,1) indicator exhibits a negative coefficient of that is significant at the 1% level. The marginal effect associated with this variable indicates that SPACs are 22 percentage points less likely to engage in all-cash deals. This result may reflect the risk sharing argument in Hansen (1987). He contends that stock is more useful than cash in resolving information asymmetry when deal related uncertainty is due to the target firm. Consequently, Hansen states that the use of stock as a form of payment forces target shareholders to share in the risk of the acquisition Other estimates in the regression are similar to those in the existing literature. For example, the relative size variable, which is defined as the deal value divided by the bidder s market value before the announcement date, exhibits a significant negative coefficient. This finding is consistent with Baker, Coval, and Stein (2007) that cash is preferred for smaller targets. 14

17 The previous univariate test shows that SPACs are less likely to complete a merger. I test if this result holds in a multivariate context. In Table 4, I report the estimates of four fixed effect logistic regressions in which the dependent variable is one if the deal is completed and zero otherwise. Walkling (1985), Bates and Lemmon (2003), and Officer (2003) estimate similar models. Therefore, the control variables (defined in the Appendix) in these regressions are similar to theirs. As with earlier tests, the key independent variable is the SPAC (0,1) indicator. Consistent with the existing literature, in model (2), I find that deals are about 6 percentage points more likely to be completed if there is a target termination fee (Officer, 2003). In addition, tender offers are 2 percentage points more likely to go through. A deal is also more likely to be completed if it is paid for in cash or if the deal attitude is friendly (Walkling, 1985). The SPAC (0,1) variable in model (3) exhibits a negative coefficient equal to that is significant at the 1% level. The marginal effect associated with this variable indicates that SPACs are 33 percentage points less likely to complete the deal. Walkling (1985) and Schwert (2000) show that deal attitude is the best single predictor of merger success. Given that the proportion of friendly deals by SPACs is about 96%, the incidence of SPAC acquisitions that fail to complete is not only puzzling, but also noteworthy. The sample statistics presented earlier document that only 14 SPACs that go IPO in 2007 complete their acquisitions while 22 liquidate. Many of these withdrawals happen during the economic downturn in It is possible that during this period a large set of institutional investors bail from the market after getting hit with waves of redemptions. These conjectures suggest that SPACs might be more likely to complete a merger under favorable stock market conditions. To test whether it is indeed easier for pending SPAC acquisitions to complete in up markets, in model (4) of Table 4, I include an interaction variable between an up market (0,1) indicator and the SPAC (0,1) variable. Following Cooper, Gutierrez, and Hameed (2004), I define the up market (0,1) variable to be one if the return on the S&P500 during 36 months prior to the acquisition announcement is non-negative and zero otherwise. In model (4) of Table 4, the joint effect of the up market (0,1) variable and the interaction term shows that under up market 15

18 conditions, the rate of withdrawn deals involving SPACs is cut by 42 percentage points. This result indicates that SPAC deals are more likely to be completed in up markets. This finding supports the existing view that stock market health affects mergers (see, for example, Shleifer and Vishny, 2003, and Rhodes-Kropf and Viswanathan, 2004). 6. Do SPAC bidders make better acquisitions? 6.1. Acquirer s announcement returns I now test whether the performance of SPAC-related acquisition is better than the performance of deals by other public bidders. In this section, I examine bidder shareholder returns upon the acquisition announcement. I calculate the cumulative abnormal return (CAR) over a three-day window surrounding the acquisition announcement date using the value-weighted CRSP return as the benchmark. Following Moeller, Schlingemann, and Stulz (2005), I also compute the raw return over this three-day window, defined as the percent difference in stock prices between day -2 and day +1. Panel A of Table 5 reports the univariate results. Upon the acquisition announcement, SPACs exhibit mean and median CARs of 1.70% and 0.48%, respectively. Both are statistically significantly different from zero. In contrast, the mean and median CARs accruing to other public acquirers are 0.33% and 0.05%, respectively. Using a two-sided t test for the mean and a Wilcoxon test for the median, I find that the announcement return to SPACs is significantly higher than that to other public acquirers at the 1% level for the mean and at the 5% level for the median. The results also hold when I use raw returns instead of CARs. Following the approach in Moeller, Schlingemann, and Stulz (2005), I estimate that the value appreciation over this three-day announcement window is about 5 million dollars for SPAC shareholders. In constrast, shareholders of other public acquirers lose about 56 million dollars over the same period As explained earlier, the method in Moeller, Schlingemann, and Stulz (2005) involves multiplying the raw return over the three-day announcement window with the market value of equity for each SPAC and then averaging this product across all SPACs. 16

19 Since the literature identifies several factors that potentially affect the acquisition announcement returns to the bidders, I examine whether the results from my univariate tests also hold in a multivariate setting. Following Moeller, Schlingemann, and Stulz (2005) and Masulis, Wang, and Xie (2007), I control for bidder, deal, and market characteristics in two OLS regressions (reported in Panel B of Table 5) in which the dependent variable is the CAR(-1,+1) in model (1) and the raw return(-1,+1) in model (2). In both models, the key independent variable is the SPAC (0,1) indicator. The estimates for this variable are positive and significant in both models. Model (1) shows that SPACs earn a cumulative abnormal return that is 1.5% higher than other public acquirers over three days around announcement. Model (2) indicates that the raw return over the same period is 1.3% greater for SPACs. Other control variables are consistent with the previous literature. For example, the positive relative size is in line with Moeller, Schlingemann, and Stulz (2004) who find that small bidders gain more upon announcement Four-factor portfolio excess returns Given the short-term nature of the announcement returns, in this section I examine longer term returns by bidder type. To do so, I form a portfolio of SPACs and a portfolio of other public acquirers to estimate the daily excess portfolio returns accruing to each group. I use a procedure similar to that in Baker and Savasoglu (2002) to form portfolios and estimate portfolio excess returns. I include a merger or acquisition in the portfolio beginning two days after announcement and remove it when the deal is either completed or withdrawn. These criteria are aimed at explicitly excluding the higher announcement return associated with SPAC bidders. 19 Using the four-factor model (Fama and French, 1996, and Carhart, 1997), I estimate four OLS regressions of daily excess portfolio returns on the market risk premium, size, book-to-market, and momentum factors in calendar time from January 2004 to May I obtain all the factors from Kenneth French s website. During each day, I calculate the return for all active deals within each portfolio and then 19 In the case of Baker and Savasoglu (2002), they exclude the announcement period to exclude the takeover premium accruing to the targets. 17

20 average the returns across all active deals to get a daily portfolio return within each portfolio. For the purpose of completeness, I evaluate both value-weighted and equal-weighted combinations of portfolios related to deals by SPACs and deals by other public acquirers. In each case, I check whether the intercept is positive and statistically significant. Table 6 reports the results of the four-factor portfolio regressions. SPACs earn a valueweighted (equal-weighted) excess return of 0.077% per day (0.072% per day). The value-weighted portfolio excess return is equivalent to 1.55% per month, or 20.29% per year, and significantly different from zero at the 5% level. The equal-weighted portfolio excess return is equivalent to 1.45% per month, or 18.86% per year, and significant at the 10% level for the equal-weighted estimates. 20 In contrast, for the portfolio of other public acquirers, the four-factor related intercepts are statistically indistinguishable from zero at conventional levels. A test for the difference between the intercepts shows that SPACs earn an equal-weighted portfolio excess return that is 0.068% per day (or 1.37% per month) statistically higher than that by other public acquirers. The alphas of value-weighted portfolios yield similar inferences. To put these results in perspective, I compare my findings to other return anomalies documented in the literature. Baker and Savasoglu (2002) find abnormal returns from a merger arbitrage strategy of holding target firms stock two days after announcement until the completion or withdrawn date to be 0.8% per month during Brav and Gompers (1997) employ a similar portfolio approach to IPO firms over the period and find that the IPO portfolio underperforms by 0.5% per month. The abnormal return accruing to my sample of SPACs is 1.37% per month, which suggests that the performance of SPACs during the period I examine is associated with a non-trivial return anomaly. 20 I use daily calendar time portfolio return regressions instead of monthly ones because the sample time period is just over 5 years. Monthly regressions result in a portfolio excess return of 1.72% per month (valueweighted, p-value=0.041) or 1.50% per month (equal-weighted, p-value=0.074) for SPACs. Although the results are qualitatively similar (slightly higher), caution must be exercised in these cases since the number of observations is just over

21 6.3. Acquisition discount Despite the superior stock market performance of SPAC bidders, it is not ex-ante evident that when compared to other public bidders SPACs buy targets at better terms. Therefore, in this section, I directly examine whether SPACs negotiate better acquisition deals. I adapt the comparable industry transaction technique in Kaplan and Ruback (1995) and Officer (2007) that helps compare the premia paid to private targets with those paid to public targets in comparable transactions. Specifically, following Officer (2007), for each SPAC s target, I form a portfolio of comparable acquisitions of public targets from SDC, where comparable acquisitions are those in which (1) the matching target is in the same two-digit SIC code as the SPAC s target, (2) the matching deal is valued within 20% of the deals structured by the SPAC, and (3) the announcement date is within the three calendar-year window centered on the announcement of the SPAC s acquisition. 21 The goal of forming the portfolio of public targets in comparable transactions is to predict the acquisition multiple each private target in the sample should earn. As in Officer (2007), I also use four acquisition multiples: (1) price to book value of equity, (2) price to earnings per share, (3) deal value to EBITDA, and (4) deal value to sales. For each multiple, the acquisition discount is the percent difference between the multiple for the SPAC s target and the average corresponding multiple for the portfolio of comparable targets. The acquisition discount is a negative number if the acquisition multiple for the SPAC s target is less than the average multiple in acquisitions of comparable targets. It is a positive number if the acquisition multiple for the SPAC s target is greater than the average multiple in acquisitions of comparable targets. For each deal, I estimate the average acquisition discount as the equal-weighted average across all 21 Since all acquisitions by SPACs and other public acquirers that bid for private targets happen from 2004 to 2009, the matching pool consists of deals by public acquirers bidding for public targets that are announced from 2001 to One concern is that deals announced towards the end of the sample period (from 2007 to 2009) might have smaller pools of comparable transactions (due to shorter time period used to identify matching firms). The result from a Chow test shows that there is no significant structural break of acquisition discounts at the year 2007 (F-value = 1.16). As documented in Officer (2007), most results are qualitatively similar when the window for finding comparable transactions of publicly traded targets is restricted to three years preceding the sample acquisition. 19

22 acquisition discounts computed using each multiple. As in Officer (2007), all acquisition discounts used later in multivariate tests refer to this equal-weighted average acquisition discount measure. I use the same method to estimate the acquisition discount for private targets of non-spac public acquirers. Officer (2007) finds that acquirers buying private targets pay lower premia compared to deals that involve industry-and-size comparable public targets. Consistent with his finding, in the univariate tests presented in Table 7, I also find that compared with public acquirers that buy public targets, non-spac public acquirers get a discount of -19% when targeting private companies. In Table 7, I also document that SPACs get a discount of -26% compared with public acquirers that buy public targets in similar transactions. As shown in the table, the 19% and 26% discounts are both statistically different from zero at the 1% level. Since SPACs target mostly private firms, a more appropriate comparison is between SPACs and other non-spac public acquirers that also bid for private targets. In this context, I find that the discount SPACs are able to obtain exceeds that obtained by other public acquirers bidding for private targets. The excess discount is 7.6% and statistically different from zero. The economic effect associated with this additional discount is 23 million dollars. 22 This discount appears substantial given that the average size of acquisitions by SPACs is 230 million dollars. I view this result as direct evidence that SPACs pay lower prices for their targets. 7. Sources of efficiency gains by SPACs Given the results in Table 7, in this section, I investigate the sources of the additional takeover discount negotiated by SPACs. The natural place to begin is to exploit the differences in the types of deals SPACs execute when compared to those executed by other public bidders. I will also focus on the differences in the quality of monitoring and ownership of SPACs and non-spac bidders. 22 These estimates are obtained at the mean transaction value million dollars. 20