The Underwriter Effect: The Case of Special Purpose Acquisition Companies (SPACs)
Special Purpose Acquisition Companies (SPACs) are making a comeback. SPACs are empty shells that obtain their public listing with the sole purpose of acquiring an oftentimes privately held
company. SPACs do not have great reputation, therefore, the boom in activity in the SPAC market is remarkable. This paper aims to investigate what some of the drivers behind this boom might be. The main focus of this paper is to see if an underwriter effect, evident in IPOs,
is also present in the SPAC market. I find that an underwriter effect is present in SPACs. SPACs that included a prestigious underwriter in their public offering post higher alphas than SPACs that did not include a prestigious underwriter in their public offering.
Student: B.P. Greuter Student Number: 11227753 Supervisor: A. Kaakeh
MSc Finance, Corporate Finance track
2 Statement of originality
This document is written by B.P. Greuter who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Table of contents
1 Introduction ... 4
2Literature review ... 7
2.1 What is a SPAC and how are they structured? ... 7
2.2 SPAC performance ... 9
2.3 Corporate finance and market conditions ... 10
2.4 Prestigious underwriters ... 11
3 Methodology section ... 13
3.1 Data summary ... 13
3.2 Logistic regression model to test the likelihood of merging with a SPAC ... 13
3.3 Underwriter reputation and announcement returns ... 14
4 Data……….. 16
4.1 Data and sources... 16
4.2 Summary statistics ... 17
4.2.1 Dataset 1 ... 17
4.2.2 Dataset 2 ... 20
5 Market conditions and SPAC activity ... 22
6 Prestigious underwriters and SPACs ... 25
7 Robustness checks ... 30
7.1 Robustness checks for the logistic regresion model ... 30
7.2 Robustness checks for the relation between prestigious underwriters and SPAC performance ... 32
Literature list ... 36
4 1 Introduction
Special Purpose Acquisition Companies (hereafter SPACs) have been making a comeback over the last few years. This gradual comeback started gaining momentum and eventually exploded during 2020 and the first few months of 2021. The SPAC market report by Duff & Phelps (September 2020) shows that SPACs raised $74 billion in the first nine months of 2020. The report by Duff & Phelps also shows a very clear increase in SPAC activity in terms of both initial public offerings (IPO) and merger activity. According to the New York Times, 242 SPACs raised cash on public exchanges in 2020, which is four times the 2019 number (Merced, 2020). There are currently no indications this boom is slowing down. According to another article by The New York Times, SPAC IPOs had $42.7 billion in total proceeds in the first month and a half of 2021 (Kurutz, 2021). A SPAC is a shell company that is listed on a public exchange. The sole purpose of this company is to acquire another, oftentimes privately held, company, before a specified deadline and bring that company to the public markets.
For the acquired company this process is a shortcut to a public listing. The target company saves itself the monetary cost of an IPO as well as a lot of time. Furthermore, the target company avoids the lengthy process and scrutiny by the SEC and investors. For the early SPAC investors, it is oftentimes a profitable endeavor. The management, also called sponsors, of the SPAC usually wind up with a significant stake in the post-SPAC, also called DeSPAC, firm. However, problems loom for smaller investors that are unable to access the primary market, and therefore must resort to the secondary markets. Finally, the fact that SPACs do not have a great reputation makes the boom we are currently observing remarkable.
SPACs are often associated with similar vehicles like the blank check companies involved in reverse mergers in the 1980s. These companies were involved in several pump and dump schemes in the past (Rodrigues & Stegemoller, 2013). Financial experts express their skepticism about the current SPAC boom. Since the review process of a SPAC target is less thorough than for a regular IPO firm it is easier for a company to get its acquisition approved than it would be going public the traditional way. Additionally, the fact that the threshold to go public by merging with a SPAC is lower than the threshold for an IPO, feeds the assumption that firms of lesser quality, or firms in an earlier stage of the life cycle, merge with a SPAC . This is because these firms would not be able to get their listing through a traditional IPO, so they resort to alternative ways of obtaining a listing - enter the SPAC. Empirical evidence also suggests that companies who obtain their public listing by merging with a SPAC heavily underperform the market, more so than regular IPO firms do. This aligns with the assumption that firms of lesser quality tend to merge with SPACs.
In previous empirical research, the current SPAC boom is not included. The majority of the papers on SPACs were written before 2018, and do not include the years that the current SPAC wave picked up speed. Therefore, this paper first conducts an analysis that is comparable to previous
5 research, but over a broader sample set. The goal of this analysis is to investigate if conclusions from past SPAC papers still hold true when the 2020-2021 wave is included. However, this is not the core analysis of this paper. The main focus of the paper lies on the role underwriters have in the current SPAC wave. Empirical literature writes about an ‘underwriter effect’ in IPO performance. That is, IPOs which are backed by investment banks that are perceived as prestigious, tend to perform better than their counterparts which do not work with such investment banks. This research has not been conducted for SPACs yet. Since the assumption is that investment banks care about their reputation - as it is tied to their future revenue - one would expect that investment banks do not back just any SPAC. Therefore, the main research question of this paper is:
‘Does the involvement of a prestigious underwriter in the SPAC IPO affect the performance of the SPAC, as measured by acquisition announcement returns?’
In a model inspired by Kolb & Tykvová (2016), this paper sheds light on market conditions that affect SPAC activity. The model measures the likelihood of a private firm choosing a merger with a SPAC over a traditional IPO as a manner to obtain a public listing. The variables of interest are volatility and market wide valuations. As documented in several papers (Schill (2003), Kolb &Tykvová (2016)) it is believed that volatility negatively affects IPO volume and that the opposite holds true for SPAC volume. Surprisingly, this paper finds that, when including the sample years 2016-2020, this does not seem to be the case. In a logistic regression model that is used to analyze the effect of volatility, with the VIX index as a proxy for volatility, it is concluded that volatility does not have a significant effect on the likelihood of a SPAC merger. Robustness checks show that the VIX has a significant coefficient when excluding the years 2016-2020. Therefore, one could conclude that something is different in the SPAC boom we currently observe.
One factor that does seem to be different in the current SPAC wave is the profile of the underwriters in SPAC public offerings. The reputation of an investment bank, as measured by market share (Megginson & Weis (1990), is positively related to IPO performance. Since banks care about their reputation, one would expect that they do not back just any SPAC. When analyzing the mean rankings of the banks that acted as underwriters in SPAC offerings one thing becomes clear, the underwriters active in the SPAC market seem to be bigger than they used to. The mean ranking in the period 2003 – 2015 was 87.449, in the period 2016 – 2020 the mean ranking is equal to 51.615.
Here, the best rank a bank can have is 1 and the worst rank is 101. Rankings are determined by looking at the dollar volume that a bank took on as underwriter in public offering deals. The fact that bigger banks now seem to be involved further strengthens the assumption that something is
6 fundamentally different in this SPAC wave. The involvement of bigger investment banks could be interpreted as a vote of confidence that legitimizes SPACs as an asset class for investors.
The main analysis of this paper shows that SPACs which are backed by a prestigious underwriter outperform their counterparts that are not backed by such an underwriter. A group composed of these ‘prestigious SPACs’ posts significant alphas ranging from 0.866% to 2.36%. These are all announcement returns on a short-term basis, ranging from 3 days till 10 days after the announcement date of the target. The ‘non-prestigious SPACs’ post considerably lower returns, and these returns seem to be less statistically significant than the prestigious SPACs. This all points to the presence of an ‘underwriter effect’ in SPACs.
This paper adds to the empirical literature in two ways. First, it extends the limited research on SPACs by including years of interest beyond the years typically covered in SPAC papers up until now. With the real boom in SPAC activity occurring in 2020 and 2021, this is one of the first papers that covers the most recent wave of SPACs. Second, it adds to our understanding of the signaling tool of including prestigious underwriters in securities offerings. Underwriter prestige is associated with better quality firms in IPOs, but in this paper it is shown that underwriter prestige also matters for alternative investment vehicles like SPACs. According to my knowledge, this paper is the first paper that links underwriter reputation to SPACs.
This paper is structured as follows. Subsequent to this introduction section, the most closely related articles are summarized in the literature review. This literature review covers what a SPAC is and how their performance has been investigated over the years. Furthermore, the literature review summarizes the articles on the relation between underwriter reputation and IPO performance. The literature review also includes the two formal hypotheses of this paper. The literature review is followed by a methodology section. This section is followed by the data section which describes the data and data providers used in this study. This is then followed by a result section which reports the logistic regression including the effect of volatility and market wide valuations. Hereafter, in section 6 one can find the main analysis of this paper, linking underwriter reputation to SPAC performance.
The last two sections contain robustness checks and concluding remarks.
7 2Literature review
2.1 What is a SPAC and how are they structured?
The Securities and Exchange Commission (SEC) describes SPACs as entities without operational activities that place the IPO proceeds into a trust/escrowed account. The cash is reserved for the acquisition of at least one privately held company. The only activity of the SPAC is to look for a target firm to acquire (SEC, 2020). In other words, a SPAC is a publicly traded shell with the sole purpose of acquiring a privately held firm before a predetermined deadline. If the SPAC does not acquire a firm before this deadline, it is liquidated and the proceeds minus the costs are returned to the
SPACs are similar to the shell companies involved in reverse mergers as described by Floros
& Sapp (2011). In a reverse merger, the private firm acquirers a public listed shell firm, thereby obtaining its public listing. Not being listed before the merger means the acquiring firm must comply with less disclosure requirements than a publicly listed firm would. With a SPAC transaction, it is the other way around. Since the SPAC is the acquiring party and is publicly listed, there are more reporting obligations. Therefore, Chatterjee et al. (2016) argue that, although reverse mergers and SPACs might be comparable, SPACs need to comply with stricter regulations. Therefore, some concerns regarding reverse mergers do not apply to SPACs. Chatterjee et al. (2016) also provide a clear description of the security design of a SPAC. A SPAC public offering is typically a unit initial public offering. These units consist of one share and one, or more, warrant(s). The founders (sponsors) of the SPAC usually commit to buying a significant number of warrants or stock (Hale, 2007). The warrants cannot be exercised before the SPAC finalizes the acquisition of a company (Hale, 2007, p3.). If a target is not found and the SPAC is liquidated, the warrants expire worthless.
SPACs are occasionally described as publicly traded private equity funds. Rodrigues &
Steggemoller (2013), among others, make this comparison. Dimitrova (2017) makes a similar comparison, calling SPACs the ‘poor man’s private equity fund’. Rodrigues & Steggemoller (2013) provide an analysis of the legal framework of a SPAC. They describe the role of the sponsors (founders) of SPACs in detail. The sponsors buy a small number of shares and typically buy a large number of warrants exercisable when the SPAC successfully acquires a company. The shares of the sponsors are usually escrowed and can be released after a successful deal. This serves as an incentive for the sponsors to actively pursue a target. After the SPAC consummates a deal, the sponsors typically own a large part of the acquired company either through stock or warrants. If the SPAC fails to consummate a deal, the warrants expire worthless. The escrowed shares of the sponsor cannot participate in the liquidation either.
Raising capital on a public exchange is the first step in a SPAC life cycle. After the SPAC raises its capital, it starts looking for a suitable target. When it finds a private company it wants to acquire,
8 the shareholders get the chance to vote (Rodrigues & Steggemoller, 2013). The threshold required for an acquisition to be accepted by the shareholders is specified in the prospectus. As Rodrigues &
Steggemoller (2013) show, the mean conversion threshold in their sample set of 243 SPACs was 27.2%. In the proxy vote for approving the SPAC, the percentage of shareholders voting against the acquisition of the proposed target cannot be greater than the specified threshold. Therefore, a supermajority of the shareholders needs to be in favor of the acquisition for it to occur. Rodrigues &
Steggemoller (2013) show that this conversion threshold drastically increases in what they called
‘the third generation SPACs’, that is, SPACs that emerged after the financial crisis of 2008. A higher conversion threshold makes it easier for a SPAC to get its target approved by the shareholders.
A major concern in SPAC security design is the dilution for the shareholders due to the exercisable warrants, owner shares and redeeming shareholders. Lakicevic & Vulanovic (2013) calculate the diluting effect these three factors have. When taking the owner shares and warrants into account, they calculate the potential dilution to be 29.30%. If 20% of the shareholders redeem their shares, this potential dilution rises to 43.34%. The dilution is positively related to the
conversion threshold. If we combine that fact with what Rodrigues & Steggemoller (2013) point out about the increasing conversion thresholds, the potential dilution for current SPAC investors is higher than it ever was. This is shown by Klausner et al. (2020) that put the dilution for 46 SPACs that completed a merger between January 2019 and June 2020 at approximately 50%. As they put it in context, this means that every share worth $10 is covered by only $6.67 in cash, when taking the full potential dilution into account. This risk is oftentimes overseen by retail investors.
Lastly, an important concern with regards to the security design is that it can create the wrong kind of incentives for the sponsor. As explained before, the sponsor of the SPAC usually loses a substantial amount, if not all, of their investment in case the SPAC does not consummate the deal.
This could imply that, if the situation arises that the SPAC has no target and it approaches the deadline, the sponsors will pursue an acquisition at all costs. This is what Dimitrova (2017) points out. Dimitrova describes the same concern about the security design and the strong incentive to pursue a deal at all costs. He shows how SPACs that complete an acquisition just before their deadline perform worse than their counterparts that announce a target in an earlier stage.
9 2.2 SPAC performance
Another aspect of SPACs that seems remarkable, given the activity we are observing in the SPAC market now, is the strong underperformance in the long term. Although research specifically focused on SPAC performance is still limited, there are multiple papers that point to a significant underperformance of firms that chose to go public by merging with a SPAC. Kolb & Tykvová (2016) show that SPAC firms strongly underperform the market in a buy and hold strategy for 6, 12, 24 and 60 months after the completion date of the acquisition. This conclusion also holds in a five factor model analysis. The SPAC portfolio has a statistically significant negative alpha. As often documented (Ritter (1991), Loughran & Ritter (1995)), IPO firms underperform the broader market as well.
However, SPACs perform substantially worse after the completion of their acquisition as pointed out by Kolb & Tykvová (2016). They matched comparable SPAC firms and IPO firms and found that SPAC firms underperform comparable firms that debuted on the public markets through the traditional IPO route.
Apart from showing that SPACs that announce a target just before the deadline perform worse than other SPACs, Dimitrova (2017) also documents underperformance for the broader SPAC market on the long run. Dimitrova shows an average negative performance of more than 40% in both one year and two years after the acquisition completion date. In comparison with other benchmarks underperformance is occasionally in excess of -50%. Although Dimitrova acknowledges that, firms involved in mergers usually underperform the market, he points out that SPAC firms perform considerably worse in the long run. Moeller et al. (2003) investigate mergers and their conclusions are essential to put the SPAC underperformance in perspective. They find that firms that acquire a private firm have an abnormal buy and hold return over a three year holding period after the acquisition of -26.5%. As SPACs usually acquire privately held firms this puts the negative return in excess of -40% found by Dimitrova in perspective. However, it is clear that SPACs underperform even these firms involved in related transactions. A review by Dimitrova (2017) of the operating performance of SPAC firms after the merger further adds to the belief that firms of lesser quality tend to merge with SPACs. Klausner et al. (2020) show that the post-merger underperformance is correlated to the degree of dilution in the SPAC.
Although it is documented that SPACs underperform the market in the long run, Dimitrova (2017) shows that in the short run, the opposite seems to be the case. Dimitrova finds a statistically significant abnormal return of 1.5% in the three days surrounding the acquisition announcement.
The announcement returns are a variable of interest for this paper. As explained in further detail in a later stage, underwriter reputation matters, that is why I expect to find that the announcement returns for SPACs that have a prestigious underwriter will be higher than SPACs that do not.
10 2.3 Corporate finance and market conditions
Section 2.2 points out that SPACs underperform the market. Therefore, one could ask the question, why would a firm choose to merge with a SPAC instead of a traditional IPO? Next to their research into SPAC performance, Kolb & Tykvová (2016) use a model that investigates the variables that could contribute to the answer to this question.
The first obvious factor is that being acquired by a SPAC is cheaper than an IPO. The private firm spares itself the oftentimes steep underwriter fees. Furthermore, Rodrigues & Stegemoller (2014) document that the firms also face less indirect costs related to underpricing and the time- consuming nature of an IPO process. Ritter & Welch (2002) report an average first day return of 18.8% for IPO firms, this is what they call “money left on the table” for the company. This could be characterized as an indirect cost of IPOs. Furthermore, SPAC acquisitions have less uncertainty than IPOs. As Gleason et al. (2005) argue, IPOs may be cancelled if the market does not accept the IPO.
Furthermore, IPO proceeds can be uncertain as well. Acquisitions that are approved by shareholders of both firms are seldom aborted. Therefore, an acquisition by a SPAC provides the private firm with more certainty.
The second factor is less evident but has to do with market conditions. Schill (2004) shows that IPO volume is negatively related to market volatility. If we assume that the need for financing for private firms is unrelated to the market volatility, SPACs and reverse mergers might be a good alternative for firms that want to go public in volatile markets. This is exactly what Kolb & Tykvová (2016) show. In their logistic regression model, market volatility is positive and significantly related to the choice for firms to merge with a SPAC. This could be the first hint to the reason why we observe the current boom in the SPAC markets.
Lowry & Schwert (2002) show that high initial returns of IPOs are related to subsequent IPO waves. Klausner et al. (2020) show that the mean return for a SPAC sponsor in three months after the merger was roughly 400%. If the conclusions of Lowry & Schwert (2002) hold true for SPACs too, these exceptional returns might provide an incentive for accredited investors to start a SPAC of their own. When we tie this to the significant market volatility, we observed on the financial markets in 2020 and the implications from the conclusions by Kolb & Tykvová (2016) it seems that the market conditions in the last few years brew a perfect storm for SPACs to reemerge. This part of the literature section is used to formulate hypothesis one which is as follows.
H1: Volatility and market valuations are positively related to the probability a private firm chooses a merger with a SPAC to go public.
11 2.4 Prestigious underwriters
Empirical literature points to an underperformance of IPO firms. For example, Ritter (1991) and Loughran & Ritter (1995) investigate the long run performance of IPOs and show that these firms underperform comparable firms in both the three and five years following the IPO. In an IPO process underwriters have a crucial role. However, research on the role of the underwriters in IPOs shows that not all IPOs are equal. The job of an underwriter is to guarantee liquidity. Corwin & Schultz (2005) investigate the role of underwriter syndicates and state that almost all IPOs include more than one bank as underwriter. Reputable underwriters demand a bigger fee, so why would a firm choose to include a reputable underwriter if that comes at a steeper cost? One of the potential reasons Corwin & Schultz (2005) give is that it might certify the quality of the issuing firm. This could then act as a signaling device to investors. There has been substantial research on the relation between underwriter reputation and IPO performance. Johnson & Miller (1988) investigate the degree of underpricing in relation with the reputation of the underwriter. They show that, the level of prestige of the underwriter is negatively related to the risk of the issuing firm and to the degree of underpricing in the offering. If we assume this holds true for SPACS as well, prestigious investment banks could back SPACs of better quality that have better chances of buying a good target in the end. Carter et al. (1998) investigate the relation between underwriter reputation and the long run performance of IPO firms. While most papers which link underwriter reputation and IPOs focus on the initial announcement returns, Carter et al. wrote the first paper focused on long run
performance. Carter et al. (1998) show that firms whose IPO was underwritten by a reputable investment bank underperform the market less than their counterparts which did not have a reputable underwriter. This is further evidence that the reputation of the underwriter matters.
Chemmanur and Fulghieri (1994) show that investors judge the underwriter's reputation by looking at the returns of the issuances the company has participated in. This shows that
underwriters should care about the firms they back because when they back firms that do considerably worse than the market, this might affect their reputation and future business. This corresponds with the evidence by Johnson & Miller (1998) that prestigious investment banks are associated with less risky firms. Fang (2005) also writes about reputation as a valuable asset in the underwriting industry. She shows that prestigious investment banks are related to superior
underwriting services in the bond market. If we relate that to Chemmanur and Fulghieri (1994) this would mean the investment bank would be judged by the investors to be better. The conclusion Fang (2005) draws is that the involvement of a prestigious investment bank is a powerful signaling tool to the market.
In the financial literature the Carter-Manaster (1990) method to determine underwriter ranking is a widely accepted standard to measure reputation. The approach they use is analyzing
12 what they call the ‘tombstone announcement’. As described before, most IPOs have a syndicate of banks that act as underwriters in the issuance. The order in which these banks are listed on the announcement is what Carter & Manaster (1990) use in their methodology to determine
underwriter rankings. The problem with this process is that it takes a long time because one would need to analyze all the IPO announcements. Therefore Megginson & Weiss (1990) proposed to take the market share in the IPO market of an investment bank as proxy for their reputation. Megginson
& Weiss and several other papers (Fang (2005), Carter et al. (1998)) report a high positive correlation between the relative market share and the Carter-Manaster (1990) methodology. Given how this method is easier to use and the fact that it does not assume the underwriter rankings stay the same this paper uses market share as a proxy for underwriter rankings. The underwriter effect that is described in this section of the literature review is used to formulate the second hypothesis, which is as follows:
H2: Prestigious underwriters are positively related to SPAC acquisition announcement returns.
13 3 Methodology section
This methodology section is divided into several subsections. The first section summarizes data categories. For a detailed description of the dataset and the data providers that were used, please consult section 4. Section 3.2 describes the logistic regression that is used to analyze the likelihood of a SPAC merger. Section 3.3 describes the approach used in the main analysis of this paper.
3.1 Data summary
To test the two hypotheses of this paper, data was retrieved from several data sources. First, a list of SPACs that completed an acquisition was needed. As in previous research, the year 2003 is often used as the starting point of the SPAC as we know it today. Therefore, this is the starting point of the dataset. The dataset includes SPACs that announced and subsequently completed an acquisition between 2003 and 2020. Completion can be after December 2020, but no later than March 2021 since data for this paper was collected in April 2021. Especially these last 5 years will be of interest since these years have not been investigated in the extent that the period 2003-2015 has been investigated. To complete the dependent variable of the logistic regression model a list of regular IPOs in the United States is also needed.
To test the relation between volatility and the likelihood of a SPAC merger, market specific data was needed. This includes the Volatility Index (VIX), the Price Earnings ratio of the S&P 500 index and the yield on the 10-year treasury note. To complement the model, firm specific control variables were needed. To investigate if underwriter prestige matters in the SPAC market, underwriter data for SPACs that went public and subsequently completed an acquisition was needed. To determine underwriter rankings, a list of all IPOs and the underwriters in those IPOs in the US and in Europe was needed. The last data category that was needed was composed of returns for SPAC firms, the returns of the S&P 500 and the risk factors developed by Fama & French.
3.2 Logistic regression model to test the likelihood of merging with a SPAC
The first hypothesis of this paper is that volatility and market valuations are positively related to the likelihood of a firm choosing a merger with a SPAC as its way into the public markets. A logistic regression model is used to formulate an answer to this question. The dependent variable in this logistic regression model is a binary variable that takes 1 for a firm that goes public by merging with a SPAC and takes 0 for a regular IPO. Please find the exact specification of the model below.
14 In this model the VIX index and the price earnings ratio (PE) are the variables of interest. Following the hypothesis of this paper, one would expect a positive sign for both variables. All other variables are control variables and are based on the paper by Kolb & Tykvová (2016). The model adapts the proxy for volatility as compared to the model Kolb & Tykvová used. In their paper, the proxy for volatility was the average historical volatility over the 6 months preceding the announcement date.
In this model, the VIX index is used, which measures the expected volatility over a period of 30 days after the reference date. I assume this is a better proxy than historical volatility because in an IPO, the issuing firm will be more concerned with the volatility directly following the IPO as opposed to volatility in the preceding months. As the decision to go public is made way before the
announcement date, I too, take the average value of the VIX index over the 6 months preceding the announcement date.
Because of the wide array of control variables, the model should help determining a causal relation instead of only pointing to a statistical correlation. In terms of signs for these control variables, one would expect a negative sign for the yield, since acquisitions by SPACs are usually also financed by debt and a lower yield makes taking on debt easier. The expected sign for time to resolution, which is defined as the acquisition date/issuance date minus the announcement, is positive. This follows what Kolb & Tykvová (2016) found with regards to this variable. In terms of the firm specific variables, following the conclusions made by Kolb & Tykvová, one would expect the SPAC firms to be of lesser quality. This implies a negative sign for Return on Assets and a positive sign for Debt Ratio. Furthermore, with the high valuations in the market one would expect the SPAC firms to have a higher Market to Book ratio.
3.3 Underwriter reputation and announcement returns
To test the second hypothesis, that the underwriter reputation is positively related to
announcement returns, the SPACs are separated into two groups. After determining the underwriter rankings by looking at the total dollar volume of those underwriters, every SPAC is given a ranking which is dependent on its biggest underwriter. These rankings are then used to separate the SPACs into two different groups. The first contains ‘prestigious SPACs’, whereas the other group contains
‘non-prestigious SPACs’. The SPACs are divided using a dummy variable that takes value 1 when the SPAC has an underwriter from the top 10. The analysis will then focus on the alphas of the groups by using several asset pricing models. The model focuses on six different timeframes, namely 3, 5, 10, 30, 60 and 180 days after the announcement of an acquisition. The expectation is that the
prestigious SPACs will be associated with higher alphas.
15 This paper aims to answer the question if SPACs with a prestigious underwriter are a better
investment than SPACs with a non-prestigious underwriter. By looking at the alphas, one can conclude whether or not the securities had a return that was to be expected given their risk profile and relation with other factors in the model. Positive alphas mean an outperformance as compared to the expected returns for that security. The expectation is that the asset pricing models should report positive alphas in a short time frame after the announcement of an acquisition. Dimitrova (2017) shows that the market generally reacts to SPAC acquisition announcements in a positive way.
Dimitrova measures cumulative abnormal returns (CARs) to arrive to his conclusion. By choosing a different methodology this paper adds to the empirical literature instead of replicating and slightly altering the methodology used by Dimitrova. Furthermore, there is a lot of variables affecting abnormal returns. Running a regression of abnormal returns on just underwriter prestige will likely result in results harmed by a major omitted variable bias. Because determining what control variables to use in such an AR regression, for an asset class that has not been researched to a wide extent is a big challenge, this paper chose an approach using asset pricing models. By grouping the SPACs into ‘prestigious’ and ‘non-prestigious’, one can determine whether or not SPACs with a prestigious underwriter are a better investment than SPACs without by comparing the reported alphas. A negative alpha in one or both groups indicates that investors should avoid these investments altogether.
16 4 Data
This data section is structured as follows. Section 4.1 points out which data providers have been used for what specific data. Subsequently, in section 4.2, the descriptive statistics for the two main datasets can be found. This is then followed by a description of sub-sample datasets and the first things that one can already observed by looking at the descriptive statistics.
4.1 Data and sources
Because this paper conducts two separate analyses, two different datasets were needed. The first dataset is composed of SPAC acquisitions and IPOs in the United States between 2003 and 2020. For all of these firms, firm fundamental variables were collected. Furthermore market conditions were also collected for the same period. For the second analysis of this paper underwriter data for SPAC IPOs was needed. This was then matched to the list of completed SPAC acquisitions between 2003 and 2020. Underwriter rankings were determined by looking at dollar volume and subsequently merged into the underwriter dataset. To complete the analysis, returns for the SPAC firms & the S&P 500 were downloaded from CRSP along with Fama & French risk factors.
Table 1 - List of datasets and source
Data Definition Source
SPAC acquisitions All announced and
subsequently completed SPAC acquisitions of US listed SPACs between 2003-2020.
(Completion can be after 2020, but no later than March 2021)
Regular IPOs All completed IPOs between 2003-2020 on US stock exchanges, excluding SPACs
Firm fundamentals Total assets, total debt, EBIT at the time of announcement for all SPAC targets and IPO firms
Transaction values of SPAC
acquisitions Transaction value of the
10-year treasury rate 10-year T-bill rate at the
announcement date Federal Reserve Bank of Saint Louis (FRED)
VIX index Average values of the VIX
index in the 6 months preceding the announcement date
17 Table 1 - continued
Data Definition Source
S&P 500 Price/Earnings ratio PE ratio of the S&P500 index in the quarter of and the 2 quarters preceding the announcement date
Underwriters in SPAC IPOs Underwriters in the IPO of SPACs that completed an acquisition
Underwriter rankings Top 100 of underwriters in
terms of dollar volume Zephyr Returns for SPAC firms Returns for SPAC firms that
completed an acquisition, both of the SPAC and DeSPAC firm
Fama & French risk factors SMB, HML, RMW & CMA CRSP
4.2 Summary statistics 4.2.1 Dataset 1
Table 1 lists all the data and providers used to obtain this data. Table 2 below lists all the specific variables used in the logistic regression of section 5. The paper by Kolb & Tykvová (2016) is the main inspiration for the model and the specific variables but there are some modifications due to data availability issues and adaptations to the model.
Table 2 contains three types of variables. Market specific variables provide an insight in the market conditions typically observed for both SPAC firms as well as IPO firms. Following Schill (2004), and Kolb &Tykvová (2016) it is expected that volatility is an important variable in the choice for a firm to go public by means of a traditional IPO or by merging with a SPAC. However, instead of using the historical volatility, this paper uses the VIX index, which is the index that measures the expected volatility over a period of 30 days after the reference date. As historical volatility should be of lesser importance than the actual volatility at and directly following the issuance date of an IPO, I expect that the VIX index is a better proxy for volatility than the approach used by Kolb & Tykvová (2016).
The second and third sections within the table focus on the deal and firm variables. Time to resolution is the time between the announcement date of the IPO/Acquisition and the issuance date/date the acquisition is completed. Kolb & Tykvová (2016) show that the time to resolution is generally lower for IPO firms than for SPAC firms. Firm specific variables can be found in the third section. These are control variables and are fundamental variables for the IPO firm/SPAC target.
18 Table 2 is divided into three panels. Panel A lists the descriptive statistics for the full sample set, panel B lists the statistics for the SPAC firms and panel C lists the statistics for IPO firms. We observe that the market situations seem to be more volatile and that market wide valuations tend to be higher for SPAC firms as opposed to IPO firms. SPAC firms report a mean value of the VIX of 22.059 and a mean value of the price earnings ratio of 21.904, where IPOs report values of 17.85 and 18.711 respectively. Note that both differences are economically large. Furthermore, the conclusion Kolb &
Tykvová (2016) arrived at regarding time to resolution seems to hold true in this dataset too. SPAC firms report a mean time to resolution of 138 days while IPO firms only report 98 days. When it comes to firm specific variables, SPAC firms seem to be less profitable. SPAC targets report a ROA of -0.156 compared to IPO firms reporting an ROA of -0.136. Furthermore, debt ratio and market to book ratio seem higher for SPAC firms as opposed to IPO firms. Especially the difference in market to book ratios is striking at 3.041 versus 1.253. This is in line with what previous papers found
suggesting that SPAC firms usually have higher valuations. With regards to size, it seems that bigger firms tend to choose for an IPO over a merger with a SPAC as IPO firms report a mean size of 5.24 where SPAC firms report a size of 4.747. As shown in table 3, multicollinearity should not hurt the results of this paper with the most extreme correlation in either direction being –0.581.
Table 2 - Descriptive Statistics
This table reports the descriptive statistics for respectively the full dataset, the SPAC firms and the regular IPO firms. VIX is the average value of the VIX index in the 6 months preceding the announcement date. Price Earnings is the Price Earnings ratio of the S&P 500 index. Treasury yield is the 10-year US treasury yield on the date of announcement. Time to Resolution is acquisition/issuance date – announcement date. ROA is EBIT/Assets. Market to Book is deal value/assets. Debt ratio is debt/assets. Size is log(assets). All deal and firm specific variables are winsorized at 2%.
Panel A – Full Sample Obs Mean Std. Dev. Min Max
Market specific variables
VIX 3403 18.068 6.165 10.593 46.748
Price Earnings 3403 18.877 3.792 11.561 31.44
Treasury Yield 3403 2.977 1.204 .52 5.23
Deal specific variables
Time to Resolution 3188 100.518 91.373 17 579
Firm specific variables
Debt Ratio 2846 .315 .347 0 1.837
Market to Book 2788 1.333 1.698 .03 12.548
ROA 2551 -.137 .411 -2.065 .47
Size 2788 5.212 1.713 1.138 9.182
19 Table 3 - Matrix of correlations
Variables VIX Yield Time to
MtoB ROA Size
(1) VIX 1.000
(2) Treasury Yield -0.350 1.000
(3) Time to Resolution 0.022 0.250 1.000
(4) PE 0.429 -0.420 -0.263 1.000
(5) Debt Ratio -0.012 0.025 0.102 -0.006 1.000
(6) MtoB 0.014 -0.084 -0.082 0.073 -0.034 1.000
(7) ROA 0.036 0.187 0.137 -0.114 0.037 -0.484 1.000
(8) Size 0.058 -0.025 0.046 0.010 0.224 -0.581 0.474 1.000
Table 2 - Continued
Panel B – SPAC Firms Obs Mean Std. Dev. Min Max Market specific
VIX 176 22.059 8.034 10.908 46.223
Price Earnings 176 21.904 5.362 11.561 31.44
Treasury Yield 176 2.029 1.21 .55 4.94
Deal specific variables
Time to Resolution 169 138.112 75.884 24 570
Firm specific variables
Debt Ratio 157 .424 .403 0 1.804
Market to Book 125 3.041 2.948 .082 12.498
ROA 141 -.156 .405 -1.512 .414
Size 156 4.747 1.573 1.138 9.028
Panel C – IPO firms Obs Mean Std. Dev. Min Max Market specific
VIX 3227 17.85 5.971 10.593 46.748
Price Earnings 3227 18.711 3.616 11.561 31.44
Treasury Yield 3227 3.029 1.182 .52 5.23
Deal specific variables
Deal Multiple 2417 .786 14.097 -97.989 72.87
Time to Resolution 3019 98.414 91.719 17 579
Firm specific variables
Debt Ratio 2689 .309 .343 0 1.837
Market to Book 2663 1.253 1.571 .03 12.548
ROA 2410 -.136 .411 -2.065 .47
Size 2632 5.24 1.718 1.157 9.182
20 4.2.2 Dataset 2
The main analysis of this paper focuses on the effect the level of prestige of the underwriter has on the announcement returns of the SPAC. Underwriter rankings are computed using the methodology of Megginson & Weiss (1991) which looks at the dollar volume the investment bank takes on in its IPO underwriting deals. By following this methodology, the build in analysis tool in Zephyr is used to compute the top 100 investment banks for IPOs in Europe & North America. The top 10 of this list can be consulted in table 4.
The underwriter rankings are used to assign a rank to every SPAC that announced and subsequently completed an acquisition between January 2003 and December 2020. The completion can be after December 2020, but only firms that completed an acquisition at the time of the data collection (April 2021) are included. The rank is determined by looking at the bank that took the biggest share in the initial public offering of the SPAC. For example, when a SPAC has both Citigroup and Goldman Sachs as underwriter, but Citigroup was responsible for 60% of the offering and Goldman was responsible for 40% of the offering, this SPAC is given a rank of 2. When there is a tie, the highest-ranking underwriter determines the rank the SPAC gets. SPACs with an underwriter that is not in the top 100 are given a ranking of 101. Summary statistics for underwriter rankings are listed in table 5.
Table 4: Underwriter top 10
IPOs in Europe & North America for all years available in Zephyr. Deal values in Millions of USD
Underwriter Number of
Deals Total deal
value Average deal value
Morgan Stanley 601 355 680 592
Citigroup 572 349 481 611
Goldman Sachs 611 348 464 570
JP Morgan 630 328 048 521
Bank of America 640 325 395 508
Deutsche Bank 453 255 057 563
RBC Capital Markets 416 228 911 550
Wells Fargo 346 200 391 579
UBS 339 187 621 553
Credit Suisse 333 182 164 547
21 Cumming & Schweizer (2014) conclude that in their sample set, underwriters for SPAC offerings predominantly are niche investment banks. One would expect these banks to have a lower ranking, since their niche focus would give them less dollar volume. This is evident when looking at the mean ranking of 89.514 in the sample period between 2003 – 2016. However, when looking at the mean in the 2016 – 2020 sample period, it becomes clear that bigger banks seem to be involved with SPAC offerings in this period now too. Note that the difference between the two means is economically large. This could be seen as a vote of confidence for the SPAC market and could be one of the factors that explain the SPAC boom we observe in the last years.
Table5 – Mean underwriter rankings
Panel A – Full sample Obs Mean Std. Dev. Min Max
Underwriter Ranking 201 71.413 41.254 1 101
Panel B – 2003 – 2015 Obs Mean Std. Dev. Min Max
Underwriter Ranking 89 87.449 32.034 2 101
Panel C – 2016 – 2020 Obs Mean Std. Dev. Min Max
Underwriter Ranking 96 51.615 42.988 1 101
22 5 Market conditions and SPAC activity
Table 6 reports the results of the logistic model, as presented in section 3. These results shed some light on possible factors that could contribute to the boom we have observed on the SPAC market in the last few years. The model that was used is, to a certain extent, similar to the model as used by Kolb & Tykvová (2016). Therefore, the expectations with regards to the volatility, which is the main variable of interest, are the same. The extension that was added to the model is the Price Earnings ratio of the S&P 500 as it is believed that high market valuations could affect interest for alternative investments, making SPACs more attractive.
In column one, the results with only market control variables are listed. We observe signs and correlations that were expected. All three variables are significant at the 1% level. As stated by the hypothesis, a positive relation between the dependent variable and the VIX and Price earnings ratio was expected. Therefore, column one shows the first hints of evidence in favor of that hypothesis. Column two controls for deal specific variables. Time to Resolution has a significant effect and adding this variable seems to have an economically large effect on the coefficient of the Price Earnings ratio with this going up by more than 60%. All variables of interest remain statistically significant at the 1% level and are of reasonable economic magnitude.
Column three controls for firm specific control variables. Some interesting things can be observed here. First, the coefficient for the VIX index drops to about half the size it was in column one and two and it completely loses its significance. This does not provide evidence that market volatility is tied to the likelihood of a firm choosing for a merger with a SPAC over a regular IPO.
Contrary to what Kolb & Tykvová concluded, this non-significant coefficient does not support the expectations with regards to this variable. With regards to the other variable of interest, namely Price Earnings ratio as a proxy for market valuations, evidence in favor for the alternative hypothesis is found. In all three specifications the Price Earnings ratio is significant at the 1% level. The Price Earnings ratio that is used in the model is the PE ratio of the S&P 500 in two quarters prior to the announcement. That is, if the announcement for the IPO or the SPAC merger happens in Q3, the Price Earnings ratio that is used is the ratio as reported in Q1 of that year. With regards to the signs of the control variables, some things are worth pointing out. As one can see in the summary
statistics table 2, SPAC firms are associated with lower return on assets. Therefore, a positive sign for ROA is a surprising finding. This would imply that higher return on assets would have a positive effect for the likelihood a firm merges with a SPAC while we would expect the exact opposite. With regards to other control variables, the signs are less surprising. Following Dimitrova (2017) and Kolb
& Tykvová (2016) one would expect firms that choose for a merger with a SPAC to be of lesser quality. According to Kolb & Tykvová (2016), Dimitrova (2017) and the summary statistics of table 2,
23 SPAC firms usually have higher valuation multiples and typically are more levered than IPO firms.
Therefore positive signs are not surprising for Market to Book ratio and Debt ratio. With regards to Size, we observe one last notable thing. When looking at the summary statistics in section 4 and the paper by Kolb & Tykvová (2016), one would expect a negative sign for size. This is not the case as can be seen in table 6. This could be due to the fact that SPACs have become a lot larger in the last few years. If one would compose the mean size of SPACs issued in the last 4 years and compare that to SPACs issued before, it becomes evident that the SPACs of the last few years are larger than they used to be before.
Table 6 – Logistic Regression
The dependent variable is binary and takes 1 for a SPAC acquisition and 0 for a regular IPO. The sample period covers SPACs/IPOs that were announced and subsequently completed between January 2003 and December 2020. Completion can be after December 2020 but no later than March 2021. VIX is the average value of the VIX index in the 6 months preceding an acquisition/IPO announcement. PE2 is the S&P 500 price earnings ratio two quarters prior to the announcement date. Time to Resolution is the difference between the issuance date and announcement date. ROA is return on assets. Market to book is deal value/assets. Size is log assets. Debt ratio is debt/assets. Standard errors are robust and clustered at the industry level. P-values in parentheses
(1) (2) (3)
Market controls Market &Deal
controls Full model
VIX 0.0538*** 0.0537*** 0.0212
(0.000) (0.000) (0.389)
Treasury Yield -0.577*** -0.731*** -0.724***
(0.000) (0.000) (0.000)
PE2 0.0694*** 0.114*** 0.0900**
(0.004) (0.000) (0.047)
Time to Resolution 0.00535*** 0.00613***
Market to Book 0.877***
Debt Ratio 0.793*
Constant -3.780*** -4.857*** -6.992***
(0.000) (0.000) (0.000)
Observations 3,376 3,179 2,268
*** p<0.01, ** p<0.05, * p<0.1
As discussed above and further shown in the robustness checks, the volatility proxy seems to lose its significance when including the years not covered in the sample period of Kolb & Tykvová (2016).
This is an interesting conclusion since their paper pointed to a significant effect of volatility on the
24 likelihood of a SPAC merger. Future research could focus on the reason for this change. Taking the full model into account, the null hypothesis with regards to the volatility is not rejected in favor of the alternative hypothesis that volatility affects the likelihood of a SPAC merger. It does not seem that the VIX as a proxy for volatility significantly affects the probability a firm chooses a merger with a SPAC over a traditional IPO. The null hypothesis with regards to the market valuations is rejected in favor of the alternative hypothesis. The model seems to point to a positive and statistically
significant effect of the market wide valuations, as measured by the S&P 500 price earnings ratio, on the probability a firm chooses a merger with a SPAC over a traditional IPO.
The results of this logistic regression are not aligned with the results of previous research.
Previous research found a positive and significant relation between market wide volatility and both SPAC IPO and acquisition volume. One must be aware of the fact that none of the SPAC papers in the literature section include the SPAC wave of 2020-2021. Some variables point to a fundamental difference of SPACs in this time period as opposed to SPACs in earlier years. For example, as shown in table 5, the profile of the underwriters involved in the IPOs of SPACs has changed. Furthermore, it seems that the typical SPAC nowadays is larger than it used to be before. These two variables show that the current generation of SPACs might fundamentally differ from earlier waves. This might be a reason why including these years in the sample set makes the conclusions differ from other papers on the same topic.
25 6 Prestigious underwriters and SPACs
As shown in table 5, the profile of the typical underwriter seems to have changed in the years 2016- 2020 as compared to 2003-2015. Following the empirical evidence presented in section 2.4, one would expect SPACs backed by a prestigious underwriter to outperform those that did not include a reputable underwriter in their IPO. The SPACs are split into two groups using a dummy variable that takes value 1 for SPACs with a prestigious underwriter and 0 for SPACs without such an underwriter.
Prestigious underwriters are defined as underwriters from the top 10. The rankings are determined by looking at the dollar volume the bank has taken on in underwriting deals. The stock performance of these SPACs is evaluated using three asset pricing models on 6 different time frames.
Table 7, panel A reports the results of a market regression using the capital asset pricing model (CAPM). Columns one through three test the returns on three different short-term
timeframes after the acquisition announcement date T for SPACs with a prestigious underwriter.
Just as in Dimitrova (2017), we observe positive outperformance in the days following an acquisition announcement. Alphas in all timeframes for both groups have a positive sign. However, the alphas for the prestigious SPACs (columns 1 through 3) are consistently higher and the difference is of considerable economic magnitude. We observe an alpha of 1.85% in the 5 days after the announcement for the prestigious SPACs while the alpha is only 0.536% for the non-prestigious SPACs. Bear in mind that an alpha of 1.85% in a time window of 5 days is of considerable economic magnitude. Furthermore, while both groups of SPACs report positive alphas, these alphas are only significant for the prestigious SPACs. This provides empirical evidence in favor of the expectation that SPACs that are backed by a prestigious underwriter outperform SPACs that are not backed by such an underwriter.
To add to the robustness of the results, panel B and C report an analysis using a 3-factor and 5-factor model, both developed by Fama & French. In both specifications, the alphas in the first two time windows are significant and of considerable economic magnitude. In all three specifications of the model, the prestigious SPACs report a positive alpha that is statistically significant at the 5% level in the T+5 window. In the T+3 window, the prestigious SPACs report positive alphas significant at the 10% level. Especially the alphas in the 3 day window are of considerable economic magnitude , all coming in at more than 2%. This adds to the evidence that SPACs that include a prestigious underwriter in their public offering seem to outperform SPACs that do not include such an underwriter in their public offering.
26 Table 7 – Short term regressions underwriter portfolios
This table presents the results of an analysis of the excess returns. SPACs are divided into two groups using a dummy variable for a prestigious underwriter. A prestigious underwriter is defined as an underwriter in the top 10. Columns one through three report the results for the group that consists of SPACs with a prestigious underwriter. Columns 4 through 6 list the results for SPACs without such an underwriter. T is the announcement date of the acquisition. Panel A lists results for a CAPM model. Panel B lists the results for a Fama & French 3 factor model. Panel C lists the results for a Fama & French 5 factor model. Robust standard errors, p-values in parentheses.
Prestigious Underwriters Non-prestigious Underwriters
(1) (2) (3) (4) (5) (6)
Panel A -
CAPM T+3 T+5 T+10 T+3 T+5 T+10
Alpha 0.0219* 0.0185** 0.00874 0.00924 0.00536 0.00499
(0.0568) (0.0453) (0.109) (0.120) (0.198) (0.181) Market return -0.0236 -0.0153 -0.00900 0.000749 0.00145 0.000214
(0.468) (0.411) (0.424) (0.885) (0.703) (0.941)
Observations 90 131 247 128 187 348
R-squared 0.050 0.034 0.019 0.000 0.001 0.000
(1) (2) (3) (4) (5) (6)
Panel B – 3
factor T+3 T+5 T+10 T+3 T+5 T+10
Alpha 0.0201* 0.0177** 0.00879 0.00989* 0.00590 0.00535
(0.0590) (0.0357) (0.109) (0.0935) (0.158) (0.159) Market return -0.0243 -0.0162 -0.00979 0.000824 0.000593 -0.000981
(0.434) (0.436) (0.431) (0.863) (0.864) (0.726) SMB 0.00840 -0.00281 -0.00289 0.0199** 0.0126** 0.0115***
(0.562) (0.808) (0.694) (0.0126) (0.0206) (0.00946)
HML 0.0191 0.00582 0.00480 -0.00321 1.74e-05 -0.000496
(0.367) (0.716) (0.613) (0.369) (0.995) (0.783)
Observations 90 131 247 128 187 348
R-squared 0.085 0.039 0.026 0.029 0.018 0.010
(1) (2) (3) (4) (5) (6)
Panel C – 5
factor T+3 T+5 T+10 T+3 T+5 T+10
Alpha 0.0236* 0.0193** 0.00866 0.00947 0.00559 0.00501
(0.057) (0.031) (0.102) (0.123) (0.192) (0.175) Market return -0.0248 -0.0197 -0.0128 0.000762 0.000911 -0.000310
(0.415) (0.399) (0.388) (0.875) (0.800) (0.902) SMB 0.00418 -0.000323 -0.000899 0.0198** 0.0131** 0.0122**
(0.773) (0.977) (0.894) (0.012) (0.016) (0.010)
HML 0.0316 0.0128 0.00930 -0.00451 -0.00161 -0.00304
(0.260) (0.480) (0.454) (0.322) (0.660) (0.352)
RMW 0.0102 0.0153 0.00489 0.00834 0.00471 0.00418
(0.635) (0.271) (0.511) (0.638) (0.725) (0.635)
CMA -0.0861 -0.0683* -0.0315 -0.00147 0.00313 0.00821
(0.136) (0.082) (0.164) (0.897) (0.733) (0.383)
Observations 90 131 247 128 187 348
R-squared 0.137 0.083 0.047 0.032 0.020 0.012
*** p<0.01, ** p<0.05, * p<0.1
27 The only notable difference between the 3-factor regression and the CAPM is that the non-
prestigious portfolio now posts a significance outperformance in the three-day window as well.
However, the results for the prestigious SPACs remain of considerably higher economic magnitude with an alpha of 2.01% versus an alpha of 0.989%. Following Dimitrova (2017), the fact that the non- prestigious SPACs also post a significant alpha is not surprising. Dimitrova found considerable abnormal returns for all SPACS around merger announcement dates. The findings rhyme with previous literature that found a significant positive reaction in the short term for SPACs that
announce an acquisition target. Therefore, positive alphas were expected for all SPACs. However, as these market regressions point out, SPACs that included a prestigious underwriter in their public offering post higher alphas than SPACs that did not include a big bank as an underwriter.
Furthermore, where results for the prestigious SPACs are significant in all three specifications, the non-prestigious SPACs only report one significant alpha. Therefore, the null hypothesis that SPACs that include a prestigious underwriter do not outperform SPACs that include a prestigious
underwriter is rejected in favor of the hypothesis that states that SPACs that are backed by a prestigious underwriter outperform their counterparts that are not backed by such an underwriter.
Because financial literature suggests that SPACs are a bad investment in the long run, the returns of SPACs are analyzed on a longer timeframe next to the short term regressions of table 7.
Table 8 presents the result of this analysis. We observe that the prestigious SPACs lose their significant outperformance in a longer holding period. Consistent with the literature on SPACs, one would expect SPACs to underperform in the long run. This is what we observe in table 8. The prestigious and non-prestigious SPACs seem to converge to a statistically significant
underperformance of about 0.31% on the 180-day time frame. However, the non-prestigious SPACs seem to commence this underperformance in an earlier stage. The non-prestigious SPACs report a negative sign in all three timeframes whereas the prestigious SPACs still report a positive sign in the 30-day window in all three specifications. Furthermore, the underperformance of the non-
prestigious SPACs is statistically significant already in the 60-day window, whereas the prestigious SPACs only report statistically significant alphas in the 180-day window. This is consistent with the hypothesis that SPACs that are backed by a prestigious underwriter outperform their counterparts that are not backed by such an underwriter. Or in this case, it takes them longer to significantly underperform the market.
28 Table 8 – Longer term regressions underwriter portfolios
This table presents the results of an analysis of the excess returns. SPACs are divided into two groups using a dummy variable for a prestigious underwriter. A prestigious underwriter is defined as an underwriter in the top 10. Columns one through three list the results for the group that consists of SPACs with a prestigious underwriter. Columns 4 through 6 list the results for the SPACs without such an underwriter. T is the announcement date of the acquisition. Panel A lists results for a CAPM model. Panel B lists the results for a Fama& French 3 factor model. Panel C lists the results for a Fama& French 5 factor model. Robust standard errors, p-values in parentheses.
Prestigious Underwriters Non-prestigious Underwriters
(1) (2) (3) (4) (5) (6)
Panel A -
CAPM T+30 T+60 T+180 T+30 T+60 T+180
Alpha 0.00127 -0.000942 -0.00122 -0.00203** -0.00311*** -0.00313***
(0.551) (0.520) (0.300) (0.0108) (0.0000) (0.0007) Market return -0.00353 0.000295 -0.00115 0.000383 0.000448 0.000632***
(0.483) (0.770) (0.655) (0.454) (0.439) (0.00511)
Observations 688 989 1,333 1,927 3,681 5,541
R-squared 0.007 0.000 0.001 0.000 0.000 0.000
(1) (2) (3) (4) (5) (6)
Panel B – 3
factor T+30 T+60 T+180 T+30 T+60 T+180
Alpha 0.00135 -0.00120 -0.00310*** -0.000938 -0.00204** -0.00308***
(0.541) (0.318) (0.000) (0.523) (0.0102) (0.000985) Market
Return -0.00366 -0.00124 0.000461 2.00e-05 0.000266 0.000504***
(0.481) (0.667) (0.545) (0.984) (0.614) (0.000460) SMB -0.000290 -0.00135 -0.000468 0.00289* 0.00146 0.00125**
(0.903) (0.344) (0.507) (0.0902) (0.111) (0.0104)
HML 0.00224 0.00133 9.25e-05 -0.000492 -0.000384 0.000355
(0.567) (0.522) (0.918) (0.490) (0.319) (0.563)
Observations 688 1,333 3,681 989 1,927 5,541
R-squared 0.009 0.003 0.001 0.001 0.001 0.001
(1) (2) (3) (4) (5) (6)
Panel C – 5
factor T+30 T+60 T+180 T+30 T+60 T+180
Alpha 0.00138 -0.00108 -0.00311*** -0.00107 -0.00220*** -0.00311***
(0.542) (0.404) (0.000) (0.458) (0.005) (0.001)
Return -0.00490 -0.00209 0.000177 0.000418 0.00114* 0.000654***
(0.432) (0.546) (0.844) (0.657) (0.063) (0.000)
SMB 0.000332 -0.00122 -0.000718 0.00333* 0.00268** 0.00177***
(0.882) (0.392) (0.355) (0.072) (0.038) (0.004)
HML 0.00463 0.00292 0.000767 -0.00184 -2.05e-05 -0.000285
(0.441) (0.385) (0.544) (0.158) (0.983) (0.614)
RMW 0.00193 0.00110 -0.000335 0.00248 0.00275 0.00158
(0.446) (0.616) (0.753) (0.458) (0.385) (0.285)
CMA -0.0124 -0.00851 -0.00364 0.00419 0.00239 0.00241
(0.248) (0.186) (0.114) (0.232) (0.359) (0.159)
Observations 688 1,333 3,681 989 2,215 5,541
R-squared 0.018 0.008 0.002 0.003 0.006 0.001
*** p<0.01, ** p<0.05, * p<0.1
29 To conclude this results section, the empirical evidence supports the hypothesis that SPACs that are backed by a prestigious underwriter seem to outperform SPACs that are not backed by such an underwriter. The null hypothesis is rejected in favor of the alternative hypothesis that underwriter prestige matters. The results are aligned with results found in financial literature . An underwriter effect shows that prestigious underwriters are associated with higher quality IPO firms. This
underwriter effect seems to be present in the SPAC market as well. When looking at announcement returns, we can conclude that it seems that prestigious underwriters are associated with SPACs that engage in deals that are valued more by investors than SPACs with less prestigious underwriters. The implications of this, from an investor’s standpoint, is that if one would want to invest in SPACs, the underwriter of the SPAC public offering should be one of the selection criteria. A prestigious underwriter might certify the quality of the SPAC or the management of the SPAC and this might result in a higher quality target. However, underperformance on a longer time frame still seems to be evident, therefore SPACs should be treated as investments with a high risk profile. Future research could focus on the relation between prestigious underwriters in SPAC deals and the typical profile of a sponsor. For example, it might be the case that SPAC sponsors already have contacts at the investment banks and that the underwriter job is just a classic case of cronyism. SPAC sponsors usually are accredited investors, therefore it could very well be that these sponsors are alumni of the big banks on Wall Street, or that they were in business with these banks in the past. A pattern in the type of SPACs that are backed by more prestigious underwriters might be present and this could proof to be an interesting topic for future research.
30 7 Robustness checks
In this section the results of checks for robustness can be found. This section consists of two subsections. Subsection 7.1 reports the results of robustness checks for the results found in the logistic regression of section 5. Subsection 7.2 reports robustness checks for the results of section 6.
7.1 Robustness checks for the logistic regression model
As described in section 5, the conclusions with regards to the VIX index are not similar to the
conclusions of Kolb & Tykvová (2016). Although it is true that the proxy for volatility is different from the one used by Kolb & Tykvová and the model has some other adaptations, given the empirical literature, one would at least expect a statistically significant coefficient for the VIX index.
Table 9 – Logistic Regression, sample set 2003 – 2015
This table presents the results of a logistic regression. The dependent variable is binary and takes 1 for a SPAC acquisition and 0 for a regular IPO. The sample period covers SPACs/IPOs that were announced and subsequently completed between January 2003 and December 2015. VIX is the average value of the VIX index in the 6 months preceding an acquisition/IPO announcement. PE2 is the S&P 500 price earnings ratio two quarters prior to the announcement date. Time to Resolution is the difference between the issuance date and announcement date. ROA is return on assets. Market to Book is deal value/assets.
Size is log assets. Debt ratio is debt/assets. Standard errors are robust and clustered at the industry level.
P-values in parentheses.
(1) (2) (3)
Market controls Market & Deal
controls Full model
VIX 0.0656*** 0.0700*** 0.0881***
(0.000) (0.000) (0.001)
Treasury Yield -0.00713 0.0474 -0.114
(0.963) (0.790) (0.654)
PE2 -0.121** -0.114* -0.197*
(0.041) (0.080) (0.062)
Time to Resolution 0.00126 0.00393***
Market to Book 0.525***
Debt Ratio 1.279***
Constant -3.016*** -3.617*** -1.841
(0.004) (0.003) (0.360)
Observations 2,331 2,177 1,497
*** p<0.01, ** p<0.05, * p<0.1
Table 9 replicates the results presented in table 6, but with a smaller sample set. The sample set is more in line with the sample sets used in previous research on SPACs and omits the years beyond
31 2015. The results presented in table 9 provide an interesting insight and could provide interesting areas for future research. As we can observe, the VIX index has a considerably higher coefficient in this sub sample set, as compared to the full sample specification in table 6. Furthermore, where the coefficient loses its significance in the main sample set, it retains its significance at the 1% level in all three specifications in the sub sample set. This was to be expected given the empirical evidence presented in section 2.3, but nonetheless it provides an interesting question as to why the volatility index seems to lose its explanatory power when including recent years. Another interesting
observation to point out is that the sign for the PE ratio, as proxy for the overall market valuations, has flipped. Where in the main sample set, the market valuation had a positive and significant effect on the likelihood a firm chooses a SPAC over a regular IPO, in this sub-sample set, the exact opposite is true. Although the significance of the coefficients for the Price Earnings ratios is lower than in the main sample set, the fact that the sign has flipped could provide topic for future research.
Figure 1 – S&P 500 price earnings ratios
One factor that might be a starting point for such research is to investigate the behavior of the market valuations in recent years, as opposed to previous decades. Some big investors and market experts suggest that the market valuations are higher than they have ever been before. Figure 1 plots the S&P 500 price earnings ratios. We can observe that the PE ratios were in a downtrend after the bursting of the internet bubble. This downtrend continued until the financial crisis in 2008. After a major spike in market valuations, the valuations came down again around 2011-2012. It seems