Quality vs. Greed:
Drivers of Swedish IPO Pricing
A Comprehensive Analysis of IPO Valuations and the Impact of
Institutional Ownership
Edvin Andersson & Thomas Pietsch
Master’s Thesis in Finance
Graduate School
Spring, 2024
Supervisor: Christer Ljungwall
Abstract
This study investigates the effect of different pre-IPO institutional investors on the pricing
of Swedish initial public offerings(IPOs). Specifically, an event study and Ordinary Least
Squares regressions are used to compare initial short-term returns of Private Equity and
corporate-backed IPOs to non-backed listings. The primary findings show that Swedish
IPO issues are underpriced, with private equity backed listings priced higher than non-
backed IPOs, while Corporate-backed listings appear to be priced lower. The results
contest the use of signaling theory, indicating that alternative theoretical frameworks,
such as agency-related explanations, should be used to describe the relationship between
IPO pricing and pre-IPO investors.
JEL Classifications: G12, G14, and G32
Keywords: Initial Public Offering, IPO Valuation, Institutional Ownership, Ownership Structure, Private
Equity, Venture Capital, Underpricing, Event Study, OLS Regression, Signaling Theory, Principal-Agent
Theory
Acknowledgements
We would like to thank our supervisor, Christer Ljungwall, for his continuous
encouragement and support throughout the writing process. Additionally, we would
like to express our gratitude to Per Hulthén at Chalmers University of Technology and
Science. His previous work within the field has served as an inspirational source for the
development of this study. Lastly, we would like to thank Christopher Fägerskiöld, Håkan
Juserius, and Per Wassén for sharing their valuable real-world insights gained from their
professional experiences with IPOs.
Contents
1 Introduction 1
2 Background and Theoretical Framework 3
2.1 IPO Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 IPO Underpricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3 Money Left on the Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Information Asymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 Literature Review 6
3.1 Private Equity and Corporate Involvement . . . . . . . . . . . . . . . . . . . . . . 6
4 Research Question 9
4.1 Application of Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.2 Research Questions and Null Hypothesizes . . . . . . . . . . . . . . . . . . . . . . 9
4.3 Expected Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5 Methodology 11
5.1 Event Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5.2 OLS Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5.3 OLS Regression Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6 Data 22
6.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6.2 Historical Prices and Listing Dates . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6.3 Company Specific Control Variables . . . . . . . . . . . . . . . . . . . . . . . . . 24
6.4 Investor Type Categorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6.5 General Adjustments to Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7 Results and Analysis 27
7.1 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.2 Event Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7.3 OLS Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8 Discussion 36
8.1 Main Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
8.2 Additional Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
9 Conclusions 39
References 41
Appendix 45
A Winsorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
B Statistical Tests for OLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
C Robustness Check Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
List of Abbreviations and Synonyms
Abbreviation Definition
CAAR Cumulative Aggregated Abnormal Return, Underpricing
IPO Initial Public Offering
IPO price Listing price, issue price, offer price
LLC Limited Liability Company
OLS Ordinary Least Squares
OMXSPI OMX Stockholm Price Index
PE Private Equity
R&D Research and Development
SME Small and Medium Enterprises
SPAC Special Purpose Acquisition Company
VC Venture Capital
1 Introduction 1
1 Introduction
The pricing of Initial Public Offerings (IPOs) is a well-researched topic within finance, with
a consensus suggesting that firms generally underprice their stocks during IPOs relative to
aftermarket prices (Engelen et al., 2020). From the perspective of an investor, it is of vital
interest to identify factors that influence the degree of undervaluation. Understanding
these drivers enables more informed investment decisions and enhances the potential
for maximizing returns from IPO investments. Consequently, this study explores the
underpricing phenomenon and its determinants by comparing empirical data with previous
research findings.
Several theories attempt to explain why companies consistently price IPOs below
aftermarket levels. One prominent theory revolves around the "fear of lemons" as
described by Akerlof (1970) and Ritter and Welch (2002). This theory suggests that
aftermarket investors are aware of pre-IPO investors’ superior knowledge about a firm’s
true value but cannot easily differentiate between high-quality ("peach") and low-quality
("lemon") firms. Because market participants cannot easily identify whether an IPO is a
peach or a lemon, issuers must demonstrate that they are of high quality. High-quality
firms often signal their value by pricing their shares below market value, essentially, as
Ritter explains, "demonstrate that they are high quality by throwing money away". On
the other hand, poorer quality firms are believed to be financially unable to afford such
discounts and, therefore, keep their issue prices closer to their expected market values
(Ritter and Welch, 2002).
This common practice of pricing IPOs below potential market value imposes chronic
pressure on all new listings, driven by aftermarket discount expectations. This is not
beneficial from a fundraising perspective, as firms must balance the need to attract
investors with the goal of maximizing their capital raised (Darrough and Rangan, 2005).
Firms are thus motivated to minimize the discount to a level that still effectively signals
quality. They can do this by actively working with transparency to reduce the information
asymmetry gap or by involving external actors that serve a monitoring role, such as
analysts, banks, or external investors. A commonly studied factor that seems to be able
to decrease the size of the discount is the presence of Private Equity (PE) and Venture
1 Introduction 2
Capital (VC) firms (Barry et al., 1990; Folus and Boutron, 2015; Michala, 2019).
However, there is no consensus within the literature on why PE actors have the
abovementioned effect on IPO pricing. The two main explanations stem from arguments
based on either signaling theory (Engelen et al., 2020) or principal-agent conflicts (Arthurs
et al., 2008; Ritter and Welch, 2002). Additionally, PE involvement in the Swedish
IPO landscape is relatively unexplored. Since the effect is expected to differ between
geographical regions due to differences in legislative frameworks (Bruton et al., 2010),
the presence of an underpricing discount in the Swedish market due to PE involvement
remains uncertain.
This study aims to investigate the effect of pre-IPO investor type involvement on Swedish
IPO pricing between 2004-01-01 and 2023-12-31. This is done by testing for a general
presence of underpricing and the effect of PE involvement. Subsequently, the study
investigates if the quality-signaling theory holds for different investor types (PE, VC,
and Corporate actors) that can be assumed to be associated with sending quality signals.
Lastly, the principal-agent theory is applied to the setting to potentially explain why IPOs
are priced differently depending on the incentives of different pre-IPO investors.1
1The research questions are presented in Section 4.2
2 Background and Theoretical Framework 3
2 Background and Theoretical Framework
2.1 IPO Process
From a firm’s perspective, the primary motivation for pursuing an IPO is to raise additional
capital (Ritter and Welch, 2002).2 This can be achieved through issuing new equity or
selling existing shares to public market actors. The amount of new capital raised, commonly
referred to as proceeds, depends on two main factors: The initial listing price and the
number of shares issued. Both are determined by the company during the pre-money
valuation phase, also known as the price discovery phase. This process involves three main
actors:
(1) The pre-existing pool of equity owners: This group may include founders,
early external investors such as angel investors, and institutional actors like PE and
VC firms or corporate entities.
(2) Underwriter: Typically an investment bank hired by the company to assist
throughout the whole IPO process. The underwriter provides market insights,
valuation assistance, and crucial access to a network of new issue investors.
(3) New issue investors: These are investors who participate in and purchase shares
(subscribe) before the official listing date, during the IPO new issuance phase. Their
involvement is crucial for estimating demand and setting the price for the company’s
stock.
During the price discovery phase, the company aims to establish a fair market value
that sufficiently satisfies all stakeholders. The process involves addressing information
asymmetry and integrating factors such as market environment, risk sentiment, and
valuations based on peer multiples or discounted cash flows. The most common method
used for this is book building, where the underwriter explores market demand and
appropriate pricing through interactions with potential new issue investors, a process
sometimes referred to as pilot fishing (Ellis et al., 1999).
2For readers of this study, a detailed understanding of every aspect of an IPO process is not necessary.
However, knowledge of the most central actors and steps in the process will enhance the clarity and
understanding of the main terms and concepts used in the study.
2 Background and Theoretical Framework 4
Once the issue price is set, the underwriter helps sell the issued shares to new issue
investors. These shares are sold either via subscription rights or through unit rights. A
unit typically consists of a fixed number of shares and potentially call options bundled
together at a fixed price primarily determined by the number of shares in each unit.
2.2 IPO Underpricing
In previous research within IPO valuation, underpricing is the most commonly used
measure for short-term IPO pricing. Despite slight variations in its definition across
different studies, this concept generally refers to the first trading day return of a stock
(Ritter, 1991; Ritter and Welch, 2002; Lowry and Schwert, 2002; Vismara et al., 2012).
While these returns can be left unadjusted, they are often market-adjusted using a
benchmark index (MacKinlay, 1997).
Historical studies over the past half-century have consistently found that IPOs tend to
close their first trading day at higher valuations than their initial listing prices, suggesting
a systematic undervaluation by issuers (Carter et al., 1998; Miller and Reilly, 1987; Ritter,
1984; Tiniç, 1988). Underpricing is therefore just a term used to describe the relationship
between the listing and close price of the first trading day (Reilly and Hatfield, 1969;
Logue, 1973; Ibbotson and Jaffe, 1975).
A stock is considered underpriced if it shows a positive adjusted return from its listing
price on the first day, indicating that it was initially valued lower than what the market
was willing to pay. Conversely, a negative adjusted return suggests an overpricing scenario.
Underpricing serves as an indicator to evaluate whether a stock was over or undervalued
during its pre-money valuation compared to its post-money market valuation. Positive
initial returns suggest low listing valuations at IPO, while negative returns indicate higher
valuations. This framework helps measure different investor types’ relative effects on IPO
valuations. A clear understanding of underpricing is crucial for interpreting the findings
of this study accurately and meaningfully.
2 Background and Theoretical Framework 5
2.3 Money Left on the Table
The concept of Money left on the table refers to the additional capital an IPO issue could
have raised if the issue price was set at a higher level, closer to its first-day closing price
(Loughran and Ritter, 2000). For instance, consider an IPO issued at $10 per share that
closes at $15 on its first trading day. This indicates that the market valued the firm higher
than the issuers anticipated, suggesting they theoretically could have priced the shares at
$15. The issuer fails to capitalize on an additional 50% in potential proceeds by setting a
low initial price. This unclaimed capital is referred to as money left on the table.
This concept is essential for understanding the financial implications of underpricing an
IPO and serves as a critical measure of how effectively a firm capitalizes on its market
entry. It reflects the opportunity cost of conservative pricing strategies and underscores
the financial stakes in IPO pricing decisions.
2.4 Information Asymmetry
In the paper “Theory of The Firm”, Jensen and Meckling (1976) developed a managerial
problem setting that has since been conventionally applied by researchers and industry
actors. The problem revolves around two parties: the principal, who is the initial owner
and controller of an asset. The principal then hires an agent to manage and partially
control parts of the asset while still maintaining total ownership and liability obligations
related to the asset. A problem can occur when the incentives and goals of the agent
are not in line with the incentives of the principal. For the abovementioned situation to
actually turn into a problem setting, two additional theoretical concepts must also be
present. The first is the concept of asymmetric information coined by Akerlof (1970) in
his paper “The Market for Lemons”. Secondly, the agent also needs to act on his own
interests, which are prioritized above the incentives of the principal. This concept of
agent behavior is called moral hazard and is an extensively developed theory in economics
(Arrow, 1963).
3 Literature Review 6
3 Literature Review
3.1 Private Equity and Corporate Involvement
An IPO can be seen as a way for an existing investor to exit their position and realize
capital gains. Since this is a natural part of the business cycle for PE and VC firms, it is
reasonable to believe that these types of actors would want to maximize their investment
returns (Michala, 2019; Arthurs et al., 2008). PE actors can realize gains via different exit
strategies; among the most common are strategic sales, secondary buyouts, and partial or
full exits via IPO (Folus and Boutron, 2015). Given this, PE’s relation to IPOs has grown
into a well-studied field. The findings differ between observed markets and time periods.
Jain and Kini (2005) and Lowry et al. (2010) find that PE-backed IPOs tend to be more
underpriced than non-backed IPOs, while Michala (2019) finds no effect. However, several
authors find that PE and VC actors seem to be able to reduce the underpricing typically
seen in IPOs (Barry et al., 1990; Folus and Boutron, 2015; Bruton et al., 2010; Megginson
and Weiss, 1991). Additionally, the reason these actors successfully issue their IPOs at
higher valuations still needs to be determined.
3.1.1 Quality Signaling Theory
The quality signaling theory states that PE involvement provides a monitoring role, which
the aftermarket perceives as a signal of high quality (Barry et al., 1990). This perception
leads to the acceptance of lower issuance discounts for PE-backed IPOs, as the presence
of PE is seen as indicative of better oversight and value. Consequently, this contributes
to why PE firms often manage to issue their IPOs at higher valuations (Engelen et al.,
2020; Megginson and Weiss, 1991; Barry et al., 1990). Additionally, Ginsberg et al. (2011)
and Chemmanur and Loutskina (2009) find that Corporate VC involvement shows similar
signaling traits. While these insights partially explain the reason to how PE firms can
manage to minimize discounts, they do not fully explain why these actors frequently issue
their IPOs at higher valuations compared to their peers. Due to this, there are arguments
to combine theoretical frameworks when studying drivers of IPO pricing. As is the case
for Ritter and Welch (2002) who states:
3 Literature Review 7
"We argue that asymmetric information is not the primary driver of many IPO
phenomena. Instead, we believe future progress in the literature will come
from non-rational and agency conflict explanations."
3.1.2 Principal-Agent Theory
The principal-agent theory includes several levels of perspectives and uses incentive-based
arguments. A frequently discussed aspect is the short-term investment horizon of PE
actors, who inevitably plan for a future exit, contrasted with the long-term aspirations
of firms aiming for perpetual existence. These different time horizons lead to differing
incentives (Barry et al., 1990; Ritter and Welch, 2002; Pratt and Foreman, 2000). Fischer
and Pollock (2004) finds that PE-backed firms are more likely to fail compared to their
non-backed counterparts, suggesting that PE actors may exhibit short-term opportunistic
behavior at the expense of the firm’s long-term performance.
Industry standards such as investor lockup periods have been implemented to bridge
incentive differences. These lockups prevent PE owners from fully exiting immediately
through IPOs, compelling them to retain a portion of their original stake in the firm.
Depending on the duration of the lockup period, this situation could incentivize PE actors
to initially accept higher degrees of underpricing. However, Engelen et al. (2020) finds that
VCs often sell their shares faster and more assertively than other shareholders after the
IPO lock-up period ends. Additionally, with lockup periods usually ending within a year,
the timeframe is still relatively short and unlikely to fully resolve the divergence in time
horizons and incentives between PE actors and firms. According to Folus and Boutron
(2015), PE actors generally prefer exits through direct sales, which are less costly and
administratively burdensome than IPOs and allow for immediate liquidity. Yet, Folus and
Boutron (2015) notes that PE actors tend to choose IPO exists during periods of positive
market sentiment, or "Hot markets", as defined by Ritter (1984). Despite not allowing for
immediate full exits, these periods make IPOs significantly more profitable and attractive
as an exit route. Thus, PE actors can still be viewed as agents with short-term incentives
that deviate from the principal’s.
Another commonly mentioned phenomenon is window dressing, or as Ross and Hopkins
(2011) describes it, putting lipstick on a pig. This notion refers to the practice by which
3 Literature Review 8
PE firms enhance their portfolio companies’ short-term financial outlook. While these
actions might not align with the firm’s long-term interests, they aim to make the firm
appear more attractive to less informed external investors. (Michala, 2019; Chou et al.,
2006; Fischer and Pollock, 2004). Strategies include reporting accrual cash inflows, selling
highly depreciated assets, and temporarily adjusting capital expenditures to improve cash
flows instead of recognizing costs. Window dressing is unethical and illegal; however
Ross and Hopkins (2011) observes that PE actors often engage in such practices within
legislative boundaries before public issues. This behavior indicates that PE actors have
incentives different from those of the firm itself in the IPO setting.
According to Cho and Lee (2013), PE firms are aware of their superior information
and understanding of firm-specific projects, particularly in R&D. This knowledge allows
them to exploit the information asymmetry between themselves and aftermarket investors,
granting them the opportunity to construct compelling narratives about the potential of the
firm’s R&D projects and consequently increase short-term IPO valuations. Furthermore,
Hull et al. (2013) finds evidence that that PE-backed firms typically have higher R&D
expenditures just before their IPOs, indicating that PE actors behave hazardously as
agents during the IPO process.
4 Research Question 9
4 Research Question
4.1 Application of Theory
In this study, the firm itself is considered the principal, driven by the primary incentive to
grow sustainably over time. To assure aftermarket investors of its viability and quality,
the firm chooses to underprice its IPO. This strategy helps demonstrate that the firm is
not a lemon, intending to establish trust and attract investment from the market. Acting
as agents are various external investor types: PE, VC, and Corporate actors. PE and
VC firms aim to maximize capital gains from an IPO exit (Michala, 2019; Arthurs et al.,
2008). As a result, these investors tend to prefer lower degrees of underpricing, seeking to
optimize their return on investment through the IPO exit. Hence, this study classifies
both PE and VC investors as PE investors. Conversely, corporate actors typically do not
have a primary incentive to maximize profits through high IPO valuations. According to
Masulis and Nahata (2009) and Hellmann (2002), their goals may be more aligned with
those of the principal, focusing on the long-term stability and growth of the firm. They
might even favor a higher degree of underpricing to mitigate any potential negative market
reactions that could arise from an overvalued IPO. The diverse incentives concerning
IPO valuations and investment horizons between the firm (principal) and these external
investors (agents) provide a foundation for the principal-agent problem explored in this
study.
4.2 Research Questions and Null Hypothesizes
• Is IPO underpricing present in the Swedish IPO market?
• Is underpricing affected by private equity and corporate ownership involvement?
• Can the quality signaling theory explain underpricing for different investor types?
This is tested through contradiction with the following null hypothesizes:
(1) H0 : Underpricing is not present in the Swedish IPO landscape
(2) H0 : Private Equity backing does not affect the level of underpricing.
4 Research Question 10
(3) H0 : Corporate backing does not affect the level of underpricing.
By comparing the results of hypotheses (2) and (3), this study aims to evaluate the
quality-signaling theory and explain pricing differences via a principal-agent perspective
instead.
4.3 Expected Results
Relative to Organic IPOs, this study expects to find that Private Equity backed IPOs
exhibit higher pre-money issuance valuations. Given that PE firms view the IPO as an exit
route from a portfolio company, they are expected to advocate for higher valuations during
price discovery discussions. Consequently, IPOs backed by PE actors are anticipated to
be valued higher than Organic IPOs, where no such actor pushes for increased valuations.
Conversely, corporate actors do not invest in portfolio companies solely to achieve a
capital return from an IPO exit (Hellmann, 2002). Instead, they seek to invest in potential
acquisition targets, focusing on operational, strategic, and synergistic value. A corporate
actor is presumed to pursue an IPO for their portfolio company under two circumstances:
(i) the investor sees future potential in the target firm but is unwilling to invest more capital,
and (ii) the corporate actor wishes to divest its stake. In both scenarios, maximizing
short-term proceeds from the venture investment is not the primary concern for the larger
corporate entity (Masulis and Nahata, 2009; Ivanov and Xie, 2010; Chemmanur and
Loutskina, 2009). Instead, they focus on ensuring they do not participate in an overvalued
IPO. That can result in negative brand attention from having a history of overvalued
IPOs. Therefore, the corporate actor is expected to actively advocate for modest issuance
valuations and should have lower or similar issuance valuations to Organic IPOs.
Due to the expected results of hypotheses (2) and (3), this study expects to find that a
principal-agent perspective must complement quality-signaling theory to fully describe
the relationship between IPO pricing and pre-IPO investors.
5 Methodology 11
5 Methodology
This chapter outlines the method and models used to investigate the research question and
hypotheses. An event study is performed to assess the general presence of underpricing in
the Swedish IPO landscape and for each investor type individually. The output from the
event study is further analyzed with OLS regressions to determine if there are significant
relationships between the type of pre-IPO investors and the issue price.
5.1 Event Study
Event studies are a standard methodology for assessing the impact of specific economic
events on the value of a firm. The time frames used for this analysis are referred to as
the estimation window, event window, and post-event window, respectively. Each window
serves a distinct purpose: the estimation window helps establish a baseline for the firm’s
performance, the event window captures the immediate effects of the event, and the
post-event window examines the subsequent impact on firm value. When applying this
methodology to study IPO underpricing, a few common adjustments are typically made
to adapt to the unique characteristics of IPOs.
• Removing the estimation window: Typically, the estimation window establishes
baseline parameters for the subsequent analysis. However, in the context of IPOs,
where pre-market price data is scarce or nonexistent, it is hard to establish such
parameters. Therefore, eliminating the estimation window is a logical adjustment
for IPO underpricing studies, as there is insufficient data to calibrate any model
parameters before the IPO (MacKinlay, 1997).
• Adjusting the start of the event window: In traditional event studies, the event
window might start before the actual event date to capture early effects. However,
for IPOs, the lack of prior market prices necessitates a different approach. The event
window should start on the day the issue price is set, which is the final day investors
can buy shares at the initial price. Starting the event window on this day directly
focuses on the initial reactions to the IPO pricing before the shares are traded on
the open market. This method captures the immediate investor response to the
5 Methodology 12
issue price, providing a clear picture of the initial valuation dynamics of the IPO
(MacKinlay, 1997; Ritter, 1991).
A multi-period event window is used to analyze each firm’s daily stock price changes
during its first month of trading. This thesis defines a trading month as the first 21 days
of trading.
5.1.1 Unadjusted Return
The unadjusted return for each firm i on event day τ involves calculating the simple return
ri,τ and the cumulative holding period return Ri,τ . The simple return is defined as:
ClosePricei,t
ri,τ = − 1, for t = [0, 1, . . . , 21] (Eq. 1)
ClosePricei,t−1
The cumulative return is then calculated by:
∏τ
Ri,τ = (1 + ri,t), for τ = [1, 2, . . . , 21] (Eq. 2)
t=1
Here, t = 0 represents the issue date, hence ClosePricei,0 is the issue price. This
methodology assumes a buy-and-ho∑ld strategy, slightly deviating from Ritter (1991),
who aggregate simple returns using ri,τ . Employing cumulative returns, while more
computationally intensive, ensures a more precise measurement of stock performance by
accounting for the compounding effect of daily returns, compared to summing the returns,
which implicitly assumes active portfolio rebalancing.
5.1.2 Abnormal Return and CAAR
To accurately measure the impact of an IPO event, it is crucial to calculate the abnormal
returns, which account for deviations from expected or "normal" returns. Normal return,
denoted as E[Ri,τ |Xτ ], represents the return a firm would generate if it was not its first
5 Methodology 13
day of trading. The abnormal return for firm i on event day τ is defined as:
ARi,τ = Ri,τ − E[Ri,τ |Xτ ]
= Ri,τ − (α + βRm + ε) (Eq. 3)
= Ri,τ −Rm − ε
Here, E[Ri,τ |Xτ ] represents the normal return, which is the expected return absent the
event. The normal return is estimated using a market model, typically assuming a linear
relationship stemming from the Capital Asset Pricing Model. Given the lack of pre-IPO
pricing data, historical beta calculations are challenging; thus, like Ritter (1991), beta is
constrained to one and alpha to zero, using the market return Rm as a proxy for normal
return. Alternative approaches include setting a constant mean return, but MacKinlay
(1997) argues that using a market model may reduce the variance of abnormal returns by
accounting for market-related variations.
Abnormal returns ARiτ are aggregated cross-sectionally to draw inferences about
underpricing across multiple firms.
N
1 ∑
CAARτ = ARi,τ (Eq. 4)
N
i=1
This produces the Cumulative Aggregated Abnormal Return CAARτ , which provides a
consolidated measure of abnormal returns across the sampled firms for each event day τ .
5.1.3 Underpricing Measurement
The focus of this study is to investigate whether certain stakeholder groups tend to
overvalue or undervalue their companies during the IPO issuance process. To assess this,
underpricing will be measured using Cumulative Aggregated Abnormal Returns (CAAR),
which indicates relative evaluation differences between pre- and post-money valuations.
Throughout this study, the terms CAAR and underpricing are used interchangeably.
T-tests with robust standard errors are employed to statistically test the significance
of underpricing. These tests compare the mean CAAR of the sample to a theoretical
mean (zero in the case of no underpricing) to ascertain whether the observed average
5 Methodology 14
underpricing is statistically different from zero. This parametric test assumes that the
underlying distribution of returns is normal.
5.2 OLS Regressions
After examining the presence of underpricing in the Swedish IPO market, this study
explores the relationship between the degree of underpricing and different investor types
using cross sectional OLS regressions. The CAAR from the event study is employed as
the dependent variable, and each investor type is assigned a dummy variable, excluding
the non-backed (Organic) companies to avoid the dummy variable trap. Consequently, all
variable coefficients are interpreted relative to the underpricing experienced by non-backed
companies. Notably, the Other investor type can be viewed as a control variable, acting
as a catch-all for investors with diverse incentives and no straightforward interpretation.
Its primary function is to ensure that other investor type categories are more distinctly
defined and pure in their classification.
To ensure the regression results are robust and not affected by factors such as the absence of
first-day trading volume or other market-specific circumstances, three types of regressions
are conducted:
(1) One-Day Cumulative Regression: This regression uses the IPO date return
as the dependent variable to assess the immediate difference between pre- and
post-money valuations.
(2) One-Week Cumulative Regression: This follows up with an evaluation of
cumulative returns one week (five trading days) after each IPO, including the
previous day’s trading volume as an independent variable.
(3) One-Month Cumulative Regression: Extending the analysis further, this
regression measures cumulative returns one-month (21 trading days) post-IPO, also
considering the previous days’ trading volume.
These regressions aim to determine whether the degree of underpricing observed at the
IPO persists in the short term and how it relates to different types of investors. They
incorporate several historical control variables commonly used in IPO underpricing studies
5 Methodology 15
to reduce endogeneity stemming from omitted variable bias. This is sensible because the
extent of a company’s underpricing might not be fully explained by its pre-issue investor
type alone. Therefore, additional variables are included to offer a broader perspective on
the various factors influencing underpricing. Further details on these variables and their
potential impacts will be explained in the subsequent sections.
CAARi,t = β0 + β1PE/V C + β2Corporate+ β3Other + β4OMXS Lagi
+ β5Solvencyi + β6R&Di + β7Sizei + β8Exchange Type (Eq. 5)
+ β9V olumei,t−1
Table 1: Summary of Variables
Variable Name Description
Dependent Variables
CAARt Cumulative Aggregated Abnormal Return, after t
day(s)
Variables of Interest
PE/V C Dummy for Private Equity or Venture Capital backed
companies
Corporate Dummy for Corporate or Corporate-VC backed
companies
Control Variables
Other Dummy for companies with alternate types of
external equity financing
OMXS Lag (%) Cumulative OMXSPI Index return 3 months before
IPO
Solvency (%) Total Equity over Total Assets
Size Natural logarithm of Total Assets
R&D Capitalized R&D over Total Assets
Exchange Type Dummy for First Tier Exchange listings
V olumet−1 (M) Previous Days reported total trading volume.
Omitted in first-day regression (Model 1) due to
availability.
5 Methodology 16
5.2.1 Investor Types
Following a categorization methodology similar to Vismara et al. (2012), this study uses
independent dummy variables to classify each company into mutually exclusive investor
type groups. Due to the similarities in incentives, firms backed by Private Equity and
Venture Capital are grouped under PE/VC. IPOs supported by Corporate or Corporate
Venture Capital are classified as Corporate. Companies financed through other means,
such as Angel Investors or Grants, are categorized as Other. Companies without significant
external equity investors are labeled as Organic and serve as the benchmark for all other
groups in the regression analysis.
5.2.2 OMXS Lag
To capture the market sentiment around the time of each IPO, as Ritter (1991) and Jain
and Kini (2005) captured with their "Hot Market" variable, each firm i is assigned a
unique value representing the three-month cumulative return of the OMXSPI before its
IPO:
OMXSPIi,t
OMXS Lagi = − 1 (Eq. 6)
OMXSPIi,t−3m
Where t is the unique IPO date for each firm i and 3m is three calendar months. The lag
interval of 3 months is chosen to capture the short-term market sentiment and reflect the
current IPO climate, providing a snapshot of the market conditions leading up to each
offering (Folus and Boutron, 2015). The use of this market return lag also incorporates a
forward-looking aspect of market sentiment (Suresha et al., 2023). The decision to use a
continuous measure rather than a "Hot/Cold Market Dummy" is based on its availability
before each IPO. Additionally, the continuous variable approach enables a more dynamic
and nuanced view of market conditions compared to the binary classification used in prior
studies.
5 Methodology 17
Figure 1: IPO Frequency and OMXSPI Index Level
In Figure 1, the relationship between the number of IPOs and various lagged versions of
OMXSPI is displayed. This shows how the IPO market co-moves with market conditions
and highlights the importance of employing a detailed measure of market sentiment in
IPO studies, capturing the clear connection between market trends and IPO activities.
5.2.3 Solvency
As Jain and Kini (2005) notes, a leverage ratio serves the purpose of capturing a relative
risk level associated with each company. In this study, leverage is measured using the
Solvency Ratio (Equity to Asset Ratio), which Swedish companies commonly report:
Total Equityi
Solvencyi = (Eq. 7)
Total Assetsi
Even though an interest-bearing debt ratio might be preferable to avoid biases due to
differences in capital structures, the Solvency Ratio should still capture each company’s
relative risk level. The reasoning behind this decision is mainly based on data accessibility,
further discussed in Section 6.3.
5 Methodology 18
5.2.4 Size
Among the most frequently used control variables in research on IPO underpricing is
a proxy for firm maturity. Studies often employ either the age or size of the firm, or a
combination of both (Carter et al., 1998). However, Ritter (1991) argues to only use one
of these measures to avoid multicollinearity.
The decision to choose size over age stems from the complications in accurately measuring
a firm’s age. Specifically, the maturity proxy can be biased for companies that remain
passive for an extended period after their formal registration. Thus, the natural logarithm
of Total Assets, as reported in the last fiscal annual report (Suresha et al., 2023) is selected
as a proxy for maturity:
Sizei = ln(TotalAssetsi) (Eq. 8)
This approach is selected to provide a consistent measure of firm maturity across various
companies, avoiding the potential biases arising from inactiveness following a company’s
formal registration and other complexities associated with determining a precise founding
date.
5.2.5 R&D
Another variable commonly found to impact underpricing is a measure related to R&D
expenditures. Cho and Lee (2013) argue that pre-money insider investors, having access to
superior information, can more accurately appreciate the future potential of R&D projects
compared to outside market investors. Consequently, startups and companies that operate
within R&D-intensive industries, such as medicine and some technology sectors, often
face higher degrees of information asymmetry and underpricing (Darrough and Rangan,
2005; Jain and Kini, 2005). To account for these variations, this study includes each
firm’s capitalized R&D over total assets as a control variable. This variable is intended to
capture the uncertainty differences for companies where capitalized R&D constitutes a
significant part of the firm’s potential enterprise value (Hull et al., 2013).
5 Methodology 19
5.2.6 Exchange Type
Differences in listing requirements between first- and second-tier exchanges tend to attract
different types of IPOs (Vismara et al., 2012; Lowry and Schwert, 2002; Buckle et al.,
2018). According to the quality signaling theory, listing on a regulated exchange serves
as a quality benchmark, where higher transparency requirements reduce the information
asymmetry between insiders and aftermarket actors. This enables post-money investors
to more effectively screen for "lemons" compared to the less detailed prospectuses typical
of IPOs on second-tier exchanges (Reese, 1998). Thus, a dummy variable capturing this
quality distinction between exchange types is included in the study.
5.2.7 Volume
Similar to the market sentiment proxy, including a measurement for firm-specific market
demand is common. For example, Abrahamson and De Ridder (2015) and Vismara et al.
(2012) use the first trading day turnover to assess aftermarket demand for each IPO.
Suresha et al. (2023) incorporates different IPO issue subscription rates as a measure.
Following the approach of Reese (1998), this study employs trading volume as a proxy for
IPO-specific market demand. However, to maintain the historical character of the input
data, the variable is lagged by one trading day.
5.3 OLS Regression Accuracy
Verifying that certain assumptions are met is crucial to ensuring the reliability of
Ordinary Least Squares regressions. This section evaluates critical assumptions such as
multicollinearity, homoscedasticity, and the normality of error terms. These tests help
confirm the accuracy of the regression estimates and the validity of the inferential statistics
derived from the model.
5.3.1 Multicollinearity
Multicollinearity occurs when independent variables in a regression are highly correlated,
making it difficult to isolate the effect of each variable on the dependent variable. To
5 Methodology 20
ensure that multicollinearity does not bias the results, it is tested at two levels:
(i) Pairwise correlations are computed in a correlation table in Appendix B. Since the
highest absolute correlation is 0.2879, it indicates that multicollinearity is not a
concern at this level.
(ii) Linear dependence between three or more variables is assessed by computing the
Variance Inflation Factor (VIF). A VIF value of 1 indicates no multicollinearity,
with a commonly accepted threshold of 10 (Kutner, 2004). The data shows no VIF
higher than 1,3505, as seen in Appendix B.
Based on these tests, multicollinearity is not a problem when interpreting each predictors
effect as isolated and unique.
5.3.2 Homoscedastic Error Terms
Homoscedasticity, or constant variance of error terms, is essential for the efficiency of
estimators and the reliability of standard errors. However, the OLS estimator remains
unbiased and asymptotically normal even if this assumption is violated. Homoscedasticity
is tested through:
(i) Visual inspection by plotting the residuals against fitted values.
(ii) The Breusch-Pagan test for heterogeneity is preferred over the White test due to
its robustness in non-normal distributions. The results of the Breusch-Pagan test
indicate a rejection of homoscedasticity as documented in Appendix B.
To address the observed heteroscedasticity, the model employs robust standard errors,
thereby discarding the assumption of homoscedasticity to ensure more reliable inference.
5.3.3 Normality of Error Terms
Ensuring that residuals are normally distributed is desirable for making valid statistical
inferences and conducting accurate hypothesis testing. The normality of residuals is
assessed through the following methods:
5 Methodology 21
(i) Visual Inspection with a QQ-plot and histogram to check for normality. Ideally, the
scatter in the QQ plot should align closely with the positive diagonal, representing a
normal distribution, and the histogram should resemble a standard normal density
curve. The observations indicate deviations from these patterns in the tails, as
detailed in the plot in Appendix B.
(ii) Jarque-Bera Test: This test compares the skewness and kurtosis of the residuals
with those expected under a normal distribution. The rejection of the Jarque-Bera
test in Appendix B confirms that the residuals are not normally distributed.
To address these deviations from normality, the analysis incorporates bootstrap regressions
alongside traditional OLS regressions, similar to the method used by Ritter (1991).
Bootstrap regressions involve repeatedly running regressions on multiple random samples
(with replacement) from the original dataset. This method allows for the drawing of the
same observation more than once per sample, helping to ensure robustness in the face of
non-normality. The bootstrap regression, consisting of 1000 draws, aggregates the results
to provide more reliable standard errors and beta coefficients without making any strict
form assumptions on the population (Fox and Weisberg, 2011).
The bootstrapped regression outputs, available in the appendix, serve as a robustness
check against the standard OLS results. If the bootstrap analysis does not show significant
deviations in estimators, significance levels, or standard deviations, the original OLS
regression result will be considered unbiased and valid.
6 Data 22
6 Data
In this section, the data collection process is fully presented. This study uses a unique
dataset of 739 observations compiled 3 from several databases and online resources. This
section also discusses the adjustments and limitations impacting the dataset.
6.1 Data Collection
The foundational data for this study is sourced from the Refinitiv Eikon database4. The
company screening is set to include both delisted and currently listed companies on
Swedish first and second tier stock exchanges5, with IPO dates ranging from January
1:st 20046 to December 31:st 2023. Key data points such as listing exchange, IPO date,
trading volume, and OMXSPI index returns are collected simultaneously. To ensure that
the data accurately reflects regular firm IPOs, several types of listings are excluded:
• SPAC-company listings: Excluded due to their valuations primarily being derived
from cash holdings rather than operational business performance.
• Preference share listings: Excluded as they are often considered hybrid securities
with characteristics similar to bonds rather than being pure equity instruments
(Liberadzki and Liberadzki, 2019).
• Previously publicly traded listings: Excluded because their listing price is
determined from existing market values, either from foreign exchanges or from
situations where a holding company goes public and simultaneously delists a
subsidiary.
After applying these exclusion criteria, the dataset is refined to 793 IPOs (636 active and
93 delisted). These companies form the reference set of companies for the subsequent data
collection procedures. Including delisted firms ensures that even the worst-performing
3To combine data from different sources, ISIN and Swedish Organization numbers are used as unique
company identifiers
4www.eikon.refinitiv.com
5Primary exchange: NASDAQ Stockholm, Second tier exchanges: Spotlight, Nordic Growth Market,
NASDAQ First North
6In 2004, the first standardized regulatory frameworks for second tier exchanges (MTFs) was introduced
(European Parliament and Council of the European Union, 2004).
6 Data 23
IPOs from the observation period are included and mitigates the effect of survivorship
bias.
6.2 Historical Prices and Listing Dates
The collection of data necessary for calculating underpricing is conducted differently for
the two key variables: historical (unadjusted) closing prices and issue prices.
Unadjusted closing prices are sourced from the Refinitiv Eikon database for all companies
in the reference set. Some anomalies were identified, such as companies without a closing
price registered on their IPO date. For these cases, IPO dates and matching close prices
are manually assigned based on information from the prospectus and Skatteverket7.
Refinitiv Eikon only provides issue price data for 309 of the 793 companies in the
reference set. To increase the number of observations, additional data is sourced from
Capital IQ8 and Affärsvärldens IPO Guide9. This inclusion raises the total number of
companies with available issue price data to 722. The remaining companies’ issue prices
are determined by combining data from Skatteverket, nyemissioner10, and prospectuses
from finansinspektionen11. Skatteverkets records include fixed price issues and details on
IPO-related unit offerings. The issue price in this study is defined as either the fixed share
issue price or the unit price divided by the number of shares in each unit, excluding the
value of any options included in the units.
Cross-referencing manually gathered issue prices with those from databases revealed
inconsistencies, notably where Refinitiv Eikon occasionally reports the unit price as the
per share issue price. Since the underpricing calculations use the per-share closing price,
such discrepancies can significantly skew results. To address this, all issue price data is
cross-referenced among databases and verified through Skatteverket.se to ensure accuracy.
7www.skatteverket.se/
8www.capitaliq.com
9www.affarsvarlden.se/ipo-guiden
10www.nyemissioner.se
11www.fi.se/sv//sv/vara-register/prospektregistret/
6 Data 24
6.3 Company Specific Control Variables
Company specific data from financial statements are obtained from Business Retriever12.
The data used is from each company’s last fiscal report before their IPO.
Pre-IPO data ensures that the inflow of new equity capital does not influence the reported
values. This is of interest as post-IPO financials could be significantly affected by the
proceeds raised, which would not accurately reflect the company’s operational maturity at
the time of the listing. For instance, while two companies might have comparable Total
Assets pre-IPO, their financials could diverge substantially post-IPO based on the amount
of proceeds raised. Since the purpose of this variable is to capture the maturity of the
firm pre-money, focusing on pre-issue data helps maintain consistency.
Data from the Business Retriever Database covers 659 of the 793 companies in the
reference set. However, data quality and availability vary. For example, the Solvency
is available for 631 companies, while the interest-bearing debt ratio is only available for
425. This variation influences the selection of variables, prioritizing those with more
comprehensive data coverage. This has led to the complete exclusion of some variables
initially intended to be a part of the model. Among the variables that had to be dropped
are IPO subscription rates, exercised over-allotment options, and percentage of investor
type ownership.
6.4 Investor Type Categorization
Investor backing data for each company is primarily sourced from Pitchbook13, which
details the "First Financing Deal Type". These deal types are then manually categorized
into one of the four investor types used in this study, as shown below:
12www.retrievergroup.com
13www.pitchbook.com; Pitchbook definitions (a); Pitchbook definitions (b)
6 Data 25
Table 2: Investor Type Assignment
Organic PE/VC Corporate VC Other Removed
IPO Early Stage VC Corporate Asset Purchase Angel (individual) Other Transaction Types
PIPE Later Stage VC Corporate Grant Bankruptcy: Admin/Reorg
Seed Round Equity Crowdfunding Reverse Merger
Accelerator/Incubator Debt - General Merger/Acquisition
Convertible Debt Corporate Asset Purchase
Out of the 793 observations in the reference sample set, 739 companies are assigned
an investor type. Due to difficulties accessing data related to pre-IPO investors the
uncategorized firms are dropped from the data set.
6.5 General Adjustments to Data
6.5.1 Winsorization
Extreme outliers can cause a non-representative sample, which skews the results and
potentially distorts the regression analysis (Daszykowski et al., 2007). Because the return
outliers are cross-checked to be legitimate extreme values and not data errors, winsorization
is employed rather than truncation or trimming. This technique replaces extreme values
with those at the 1st and 99th percentiles. Specifically, in a sample size of 739, the
eight lowest observed CAARs are replaced by the 1st percentile value and the eight
highest CAARs by the 99th percentile. Winsorization helps ensure that the analysis
is not unduly influenced by rare extreme outliers, leading to more robust and reliable
statistical inferences and enhancing the generalizability and interpretability of the findings
(Reifman and Keyton, 2010). Appendix A displays summary statistics before and after
winsorization.
6.5.2 Mean Imputation
There are various strategies for managing missing data. A simple approach is to drop
all observations with missing data points. Alternatively, mean imputation can be used,
where missing values are replaced with the mean of the available data. Although this
6 Data 26
method has been criticized for potentially creating biased estimates (Baraldi and Enders,
2010), Donders et al. (2006) argues that mean imputation, while possibly affecting the
distribution’s standard deviation, still yields unbiased estimators.
This study utilizes mean imputation on the control variables to maximize the use of
available observations. To validate the robustness of the model outputs, results are cross-
verified with regression results performed with a dataset trimmed of any companies with
missing observations. If no major differences in estimators, significance levels, or standard
deviations are observed, the results treated with mean imputation will be considered
unbiased and valid.
7 Results and Analysis 27
7 Results and Analysis
This section presents and analyses descriptive statistics, the results from the Event study,
and the three regression outputs.
7.1 Descriptive Statistics
Table 3: Summary Statistics
obs mean std min 25% 50% 75% max
1d CAAR(%) 739 8.83 34.27 -57.85 -10.88 2.79 21.52 152.04
1w CAAR(%) 738 12.14 46.31 -62.35 -13.65 1.66 24.98 232.81
1m CAAR(%) 738 13.64 56.41 -66.95 -19.45 0.88 30.02 289.06
OMXS Lag (%) 739 2.52 7.33 -25.96 -1.20 3.32 6.73 33.04
Solvency 631 0.52 0.39 -5.82 0.32 0.54 0.77 1.00
Size1 643 1183.48 7204.08 0.05 11.00 34.23 236.82 128349.08
R&D1 653 0.15 0.25 0.00 0.00 0.00 0.22 0.98
Volume 42 739 468.80 1811.18 0.00 20.65 83.11 318.63 25566.67
Volume 202 739 122.85 563.29 0.00 4.94 24.58 83.18 10396.68
Exchange Type 739 0.23
1 in million SEK, 2 in thousand shares
The summary statistics provided in Table 3 describes the shape of the data from which
the results are derived. There are a few notable observations to be made from this table.
Most of the underpricing comes from a minority of the firms since the 1d CAAR shows a
mean of 9.58% with a median of 2.79%. The mean Exchange Type of 0.23 is interpreted
as the proportion of firms listed on NASDAQ Stockholm. This shows that a majority of
the firms in the data set are SME firms listed on the secondary exchanges. This fact is
also apparent by observing the difference between the mean (1.2 billion) and median (34
million) sizes. 14
14 Additional observations: The negative minimum observation for solvency is correct and stems from
the company Jumpgate AB, which had a majorly negative result in 2015. This value is deemed to still
capture the risk aspect related to the firm. The minimum size of 50 000 SEK is correct and reflects the
previous minimum amount of equity capital needed to start a LLC company in Sweden. The highest
three-month return of the OMXS index(+33%) is observed for the company Biofrigas Sweden AB, listed
on the 18th of June 2020. The lagged period represents the rapid rebound of the stock market after the
initial index crash of Covid-19. The worst three-month return of the OMXS index(-25,96%) is observed
for the company Cloetta AB, which was listed on the 8th of December 2008, and the lagged period
happened to include the initial crash of the financial crisis of 2008.
7 Results and Analysis 28
Figure 2: Investor Type Distribution
The frequency distribution of the sample’s different investor types is visualized in Figure 2.
Over 50% of all IPOs from the past 20 years are backed by either PE or VC firms. This
high representation is advantageous from a research perspective since it enables more
observations to be used, providing a more robust dataset for analysis. This is particularly
significant for this study given PE and VCs role as one of the primary variables of
interest. Conversely, Corporate actors only represent 6.9% of the IPOs in the sample. This
renders the corporate group more sensitive to extreme outliers and statistical assumptions
regarding normality. Additionally, 30.4% of the IPOs are conducted by companies without
significant external equity financing, categorized under the Organic group. With 225
new issues, the Organic group is sufficiently large to mitigate concerns related to small
sample sizes. This robust sample size is crucial, as all other coefficient interpretations are
compared against the estimates of this group, ensuring that the baseline for comparison is
stable and representative.
7 Results and Analysis 29
Figure 3: Underpricing and IPO Frequency Over Time
Figure 3 illustrates the degree of underpricing over time. The bars in the chart represent
the number of listings every year, and the line charts represent the annual average
underpricing. The black line chart is adjusted for outliers.15 The "Cold" IPO markets,
post the financial crisis of 2008 and 2009, show among the lowest degree of underpricing.
There is a pattern where the degree of underpricing foreshadows the number of IPOs per
year by approximately two years. This is most apparent when connecting the underpricing
spikes in 2015 and 2019 with the frequency spikes in 2017 and 2021, respectively. This
indicates that there are periods when IPO is desirable and less desirable. This is consistent
with prior studies that refer to years with high IPO frequency, such as 2017 and 2021, as
Hot markets or Cluster periods (Ritter, 1991; Jain and Kini, 2005). These periods are
typically characterized by positive market sentiment, naturally making them attractive
for IPO issuance.
15Small sample sizes are susceptible to outliers. Therefore, 2004 and 2005 are dropped, 2007 is adjusted
for Diadrom Holding AB (+120%) and Keynote Media Group AB (+92%), and 2011 is adjusted for Ecomb
AB (+133%) and Enzymatica AB (+118%). These adjustments enhance the clarity of the relationship
between underpricing and IPO frequency.
7 Results and Analysis 30
7.2 Event Study
Table 4: Event Study with CAAR
Event
window Whole PE&VC Corporate VC Organic
[0:1] 0.1066 ∗ ∗∗ 0.0750 ∗ ∗∗ 0.1845 ∗ ∗ 0.1592 ∗ ∗∗
(0.0192) (0.0203) (0.0808) (0.0470)
[0:2] 0.1229 ∗ ∗∗ 0.0911 ∗ ∗∗ 0.1839 ∗ ∗ 0.1781 ∗ ∗∗
(0.020) (0.0219) (0.0763) (0.0488)
[0:3] 0.1327 ∗ ∗∗ 0.0984 ∗ ∗∗ 0.2071 ∗ ∗∗ 0.1857 ∗ ∗∗
(0.0214) (0.0239) (0.0747) (0.0500)
[0:4] 0.1376 ∗ ∗∗ 0.0991 ∗ ∗∗ 0.2335 ∗ ∗∗ 0.1816 ∗ ∗∗
(0.0229) (0.0243) (0.0760) (0.0499)
[0:5] 0.1413 ∗ ∗∗ 0.0994 ∗ ∗∗ 0.2807 ∗ ∗∗ 0.1819 ∗ ∗∗
(0.0232) (0.0255) (0.0823) (0.0509)
[0:6] 0.1460 ∗ ∗∗ 0.1065 ∗ ∗∗ 0.2820 ∗ ∗∗ 0.1895 ∗ ∗∗
(0.0231) (0.0266) (0.0865) (0.0523)
[0:7] 0.1507 ∗ ∗∗ 0.1116 ∗ ∗∗ 0.2955 ∗ ∗∗ 0.2040 ∗ ∗∗
(0.0243) (0.0279) (0.0972) (0.0569)
[0:8] 0.1403 ∗ ∗∗ 0.1038 ∗ ∗∗ 0.2818 ∗ ∗∗ 0.1888 ∗ ∗∗
(0.0243) (0.0296) (0.1009) (0.0543)
[0:9] 0.1541 ∗ ∗∗ 0.1226 ∗ ∗∗ 0.3310 ∗ ∗ 0.1914 ∗ ∗∗
(0.0270) (0.0355) (0.1238) (0.0560)
[0:10] 0.1495 ∗ ∗∗ 0.1142 ∗ ∗∗ 0.3642 ∗ ∗∗ 0.1847 ∗ ∗∗
(0.0253) (0.0305) (0.1337) (0.0537)
[0:11] 0.1480 ∗ ∗∗ 0.1138 ∗ ∗∗ 0.3450 ∗ ∗∗ 0.1901 ∗ ∗∗
(0.0247) (0.0297) (0.1192) (0.0542)
[0:12] 0.1409 ∗ ∗∗ 0.1097 ∗ ∗∗ 0.3162 ∗ ∗∗ 0.1818 ∗ ∗∗
(0.0244) (0.0289) (0.1175) (0.0543)
[0:13] 0.1432 ∗ ∗∗ 0.1116 ∗ ∗∗ 0.3249 ∗ ∗∗ 0.1877 ∗ ∗∗
(0.0243) (0.0282) (0.1170) (0.0546)
[0:14] 0.1435 ∗ ∗∗ 0.1141 ∗ ∗∗ 0.3228 ∗ ∗∗ 0.1857 ∗ ∗∗
(0.0243) (0.0282) (0.1114) (0.0552)
[0:15] 0.1463 ∗ ∗∗ 0.1154 ∗ ∗∗ 0.3244 ∗ ∗∗ 0.1914 ∗ ∗∗
(0.0248) (0.0289) (0.1170) (0.0558)
[0:16] 0.1558 ∗ ∗∗ 0.1206 ∗ ∗∗ 0.3391 ∗ ∗∗ 0.2060 ∗ ∗∗
(0.0253) (0.0292) (0.1207) (0.0566)
[0:17] 0.1600 ∗ ∗∗ 0.1276 ∗ ∗∗ 0.3678 ∗ ∗ 0.2010 ∗ ∗∗
(0.0258) (0.0293) (0.1380) (0.0573)
[0:18] 0.1591 ∗ ∗∗ 0.1230 ∗ ∗∗ 0.3726 ∗ ∗ 0.2078 ∗ ∗∗
(0.0261) (0.0297) (0.1412) (0.0580)
[0:19] 0.1549 ∗ ∗∗ 0.1176 ∗ ∗∗ 0.3546 ∗ ∗ 0.2072 ∗ ∗∗
(0.0260) (0.0296) (0.1405) (0.0577)
[0:20] 0.1512 ∗ ∗∗ 0.1126 ∗ ∗∗ 0.3559 ∗ ∗ 0.2005 ∗ ∗∗
(0.0255) (0.0290) (0.1340) (0.0570)
[0:21] 0.1524 ∗ ∗∗ 0.1103 ∗ ∗∗ 0.3589 ∗ ∗∗ 0.2058 ∗ ∗∗
(0.0255) (0.0290) (0.1327) (0.0567)
* p < 0.1, ** p < 0.05, *** p < 0.01
Table 4 presents the event study results, with each row detailing the Cumulative Aggregated
Abnormal Returns (CAAR) for different event days. The first row depicts the underpricing
7 Results and Analysis 31
on the first day; the second row shows cumulative underpricing across the first two days,
and so forth. The columns display these results across the entire sample and various
subsets categorized by investor types.
The results show that the degree of underpricing significantly differs from zero for all
days of the event window and for all investor subgroups. This means that, on average,
pre-money valuations are lower than the post-money valuations set by the market. In
other words, the price of a stock tends to be lower at issuance and increase when it starts
trading on public markets; thus, underpricing is present in the Swedish IPO landscape.
Any slight deviations from normality are overshadowed by the sheer magnitude of the
level of underpricing.
Figure 4: Underpricing (CAAR) by Investor Types
Figure 4 illustrates the differences in underpricing among various investor types. It shows
that Corporate VC firms exhibit the highest degree of underpricing, which tends to
increase over time. In contrast, Organic firms maintain a consistently lower degree of
underpricing throughout the event horizon. PE firms demonstrate the lowest degree of
underpricing, indicating that these investors are more aggressive in their IPO valuations.
These findings provide a basis for deeper analysis in Section 7.3.
7 Results and Analysis 32
7.3 OLS Regressions
Table 5: OLS Regression
(1a) (2a) (3a) (1b) (2b) (3b)
1d CAAR 1w CAAR 1m CAAR 1d CAAR 1w CAAR 1m CAAR
PE/VC -0.0677** -0.0780** -0.0830*
(-2.35) (-2.07) (-1.69)
Corporate VC 0.0217 0.137 0.141
(0.37) (1.57) (1.28)
Other -0.0613 -0.0692 -0.107
(-1.31) (-1.18) (-1.57)
OMXS Lag 0.491*** 0.477** 0.590** 0.478*** 0.473** 0.577**
(3.02) (2.23) (2.29) (2.98) (2.27) (2.29)
Solvency 0.0615** 0.0572* 0.125*** 0.0701** 0.0664** 0.137***
(2.12) (1.73) (3.09) (2.34) (2.00) (3.26)
(LN) Size -0.00321 -0.00945 -0.0168* -0.00286 -0.0104 -0.0179*
(-0.59) (-1.26) (-1.75) (-0.52) (-1.35) (-1.79)
R&D -0.0118 -0.0158 -0.0411 0.00434 0.00145 -0.0191
(-0.18) (-0.18) (-0.38) (0.07) (0.02) (-0.17)
Exchange Type 0.0700** 0.0299 0.0813* 0.0816*** 0.0500 0.101**
(2.38) (0.85) (1.72) (2.74) (1.39) (2.15)
Volume 0.579*** 0.977** 0.595*** 0.975**
(2.77) (2.23) (2.86) (2.35)
Constant 0.0644 0.151* 0.214** 0.0918 0.187** 0.258**
(1.02) (1.79) (1.98) (1.43) (2.24) (2.39)
Observations 739 739 738 739 739 738
R-square (adjusted)1 0.0205 0.0604 0.0259 0.0303 0.0763 0.0389
F-statistic 4.07 3.11 3.98 3.38 3.39 3.49
* p < 0.1, ** p < 0.05, *** p < 0.01
1OLS regressions when studying underpricing tend to have relatively low adjusted R-squared (Ritter,
1991; Teoh et al., 2009)
In table 5, the OLS regression outputs with underpricing as a dependent variable are
displayed. Models labeled "1,2,3" correspond to different dependent variables: first-day
return (1), first-week CAAR (2), and first-month CAAR (3). The regressions labeled "a"
explores the relationship between underpricing and control variables alone, while those
labeled "b" incorporate the variables of interest, specifically the investor types.
7 Results and Analysis 33
7.3.1 Variables of Interest
The results show that PE backing has a significant negative impact on the degree of
underpricing, consistent with the implications of the event study and with previous
research findings by Bruton et al. (2010). The effects estimated remain negative across
all "b" models, ranging approximately between -6.8% and -8.3%. This demonstrates a
significant reduction in the first day and first week underpricing in models 1b and 2b,
with a weakly significant effect observed in the first-month model (3b). These results
suggest that IPOs backed by PE firms exhibit first-day returns that are 6.8 percentage
points lower compared to their Organic counterparts, indicating that PE investors have a
positive effect on pre-money valuations, which leads to a relatively lower degree of initial
returns.
The observed decrease in significance in the first-month model (3b) might be attributed
to the control variables not fully accounting for external market events post-IPO. For
instance, ZAZZ Energy AB went public on February 7, 2022, and reported a positive first-
day return of 6.32%. However, its first-month return plummeted to -46.89%, indicating
a substantially overpriced IPO. This severe fluctuation could be explained by external
factors, such as the market impacts from the Russian offensive in Ukraine, which occurred
shortly after the IPO and significantly influenced the stock’s performance.
The regression results indicate no significant relationship between the level of underpricing
and investments from Corporate and Other investor types. Although the sign of the
Corporate variable aligns with expectations, suggesting that Corporate VC backing might
lead to increased underpricing, the lack of statistical significance implies that there is no
substantial main effect influencing underpricing in this category. This result could be
explained by the discussion in Section 7.1 and merits further research.
The Other category aggregates various types of investors not classified under PE, VC, or
Corporate and serves as a "catch-all" to preserve the integrity of the specified investor
categories. Due to its broad and diverse nature, the significance and interpretability of
results related to this group are limited and not particularly interesting. This category’s
heterogeneous composition complicates any straightforward analysis or interpretation,
diminishing the potential for meaningful insights.
7 Results and Analysis 34
7.3.2 Control Variables
Similar to the findings of Cho and Lee (2013) and Suresha et al. (2023), the market
sentiment variable OMXS Lag consistently shows a significant positive relation to
underpricing with coefficients around 0,5. This means that IPOs conducted during
periods of positive market sentiment (reflected by a positive three-month OMXS lag)
tend to exhibit higher levels of underpricing. Conversely, when the market sentiment is
less favorable, the degree of underpricing decreases. To illustrate the magnitude of this
effect: if the OMXS index returned 10% in the three months preceding a firm’s IPO,
the underpricing would be five percentage points higher compared to a ceteris paribus
scenario where the index return is 0%.
The Exchange dummy variable has a significant positive effect on underpricing in both
the one-day and one-month models. This means that listings on NASDAQ Stockholm
seem to be priced at 8 to 10 percent lower levels than listings on second-tier exchanges.
This result does not align with the results of Lowry and Schwert (2002) and Reese (1998),
who find negative relationships between first-tier exchange listings and underpricing on
the US markets. The US findings align with the reduced information asymmetry theory
associated with the more strict listing requirements for the larger exchanges.
Solvency has a positive significant coefficient of about 0.6 in the short term and 0.13 in
the one-month model. The positive solvency relationship implies that lower degrees of
leverage lead to more underpricing and translate to firms with less leverage being priced
lower than highly leveraged firms. These results directly contradict the findings of Vismara
et al. (2012) and Bruton et al. (2010), who suggest that higher leverage leads to lower
issue prices. Additionally, this contradicts the intuition that higher leveraged firms are
associated with higher risk and, hence, higher risk premiums. This deviation could stem
from specific dynamics in the Swedish IPO landscape, further discussed in Section 8.2.
Volume shows a positive relationship with underpricing, consistent with Reese (1998)
findings regarding volume. This is also in line with the findings for other measurements of
firm-specific market demand used by Abrahamson and De Ridder (2015), Vismara et al.
(2012), and Suresha et al. (2023).
The LN Size variable exhibits weakly significant effects on underpricing in the one-month
7 Results and Analysis 35
model. The negative coefficient indicates a negative relationship between company size
and the degree of underpricing. This suggests that larger companies, possibly perceived
as less risky and more stable, are less underpriced at IPO. The varying significance of this
variable is particularly surprising given its frequent use and established importance in
IPO literature. Previous research consistently highlights the negative impact of size on
underpricing (Carter et al., 1998; Cho and Lee, 2013; Suresha et al., 2023) The discrepancy
to the consistently significant findings in prior research may suggest nuances in the Swedish
market or specific industry characteristics that affect how size influences IPO underpricing,
further discussed in Section 8.2.
The R&D yields non-significant coefficients across all models, indicating that its effect
on underpricing is unclear. It could potentially be positive or negative. This aligns
with the mixed results in the existing literature: while Jain and Kini (2005) found R&D
investment to have non significant effects on underpricing, Cho and Lee (2013) reported
a significant positive impact. This inconsistency suggests that the relationship between
R&D spending and IPO performance may vary depending on other contextual factors or
market conditions.
7.3.3 Robustness
The additional robustness regressions16 used to enhance the validity of the OLS model
show no significant deviations in estimators, significance levels, and standard deviations
for the main variables. Hence, the results of the OLS model will be considered unbiased
and valid.
16See Appendix C
8 Discussion 36
8 Discussion
8.1 Main Findings
The findings of this study confirm the presence of IPO underpricing in Sweden, consistent
with consensus from other markets (Lowry and Schwert, 2002; Reese, 1998). This means
that firms tend to systematically underprice their IPO listings and consequently leave
money on the table (Loughran and Ritter, 2000). This pattern among issuers aligns with
the "fear of lemon" hypothesis, as formulated by Ritter and Welch (2002), suggesting that
firms offer discounts to signal quality to the aftermarket.
Regarding the mitigating effect of PE actors on the size of the quality discounts, this
study finds significant differences between PE-backed and non-backed IPOs, contrary to
Michala (2019). PE involvement in Swedish IPOs decreases the degree of underpricing,
consistent with findings from other markets (Bruton et al., 2010; Ritter and Welch, 2002;
Engelen et al., 2020). However, the impact of PE on underpricing is less pronounced than
theorized by Barry et al. (1990), which proposes that PE-backed IPOs would diminish
the underpricing tendency. The results still demonstrate that PE involvement benefits
firms from a fundraising perspective by reducing the amount of money left on the table.
According to the quality signaling theory, this stems from the aftermarket perception of
PE backing as an indicator of quality. Meanwhile, the principal-agent theory explains
this effect through the firm’s and PE’s differing incentives, driven by varying investment
horizons.
To further assess whether the quality signaling theory can fully explain the effect of PE
involvement, it is compared to the corporate actor’s effect. According to Chemmanur
and Loutskina (2009) and Ginsberg et al. (2011), corporate backing should show similar
quality signals as PE involvement. If the reduction in underpricing stemmed solely from a
quality signal sent to aftermarkets, the effect on underpricing should be similar for both
the PE/VC and the Corporate investor groups. In other words, corporate-backed IPOs
should also be priced higher relative to Organic IPOs. Interestingly, the results from
the event study in Section 7.2 suggest that corporate-backed IPOs are priced at even
lower levels than non-backed listings. This discrepancy indicates that the quality-signaling
8 Discussion 37
theory does not fully explain why different investors have different effects on the pricing
of new issues. However, from the principal agent perspective, this difference is explained
by the deviating incentives between PE and corporate actors.
Due to the lack of significance for the corporate actor, the results from the regression
analysis in Section 7.3 cannot fully support the idea that corporate and PE actors have
different effects on underpricing. For such argumentation to be valid, the results must
show that corporate involvement impacts underpricing significantly differently than the
PE actor. Such results would greatly support the need to use alternative explanatory
theories instead of the signaling theory. This could possibly be achieved by using a larger
sample size for the corporate-backed investor group.
However, within the data set are some statistical traits that support the argument of
using the principal-agent theory to explain the PE effect on underpricing. For instance,
PE-backed firms make out 67% of the delisted companies. Based on the findings of Fischer
and Pollock (2004), this is interpreted as an indication of PEs not enhancing the long-term
performance of the firm. This implies the presence of some incentive difference between
the firm and the PE actor. Furthermore, in accordance with the findings of Hull et al.
(2013) and Cho and Lee (2013), the PE-backed firms in the data set report the highest
average ratio between capitalized R&D expenditures and total assets the year before each
IPO. The positive correlation between PE involvement and high R&D expenditure ratios
enables the discussion regarding PE actors’ eventual tendency to window dress portfolio
companies before IPOs.
8.2 Additional Findings
The result that listings on Nasdaq Stockholm have lower issue valuations than firms listed
on the second-tier exchanges contradicts the arguments of Lowry and Schwert (2002) and
Reese (1998), stating that first-tier exchanges are associated with higher transparency, thus
lower information asymmetry and lower degrees of underpricing. This could be explained
by Ritter and Welch (2002) theory about high-quality firms demonstrating their quality
by throwing money away. If lower-quality firms cannot afford to leave money on the table
to signal quality, the outcome would result in relatively higher degrees of underpricing in
first-tier exchanges, given that second-tier exchanges has a higher proportion of low-quality
8 Discussion 38
firms.
The high proportion of second-tier listings included in this study could help explain
deviations in empirical findings for certain control variables if there are differences in
dynamics between first- and second-tier listings. Given the drastic difference in maturity,
size, and risk between exchange listings, it is reasonable to believe that factors that affect
issue prices differ heavily depending on the type of IPO. For instance, Suresha et al.
(2023) and Cho and Lee (2013) find that more mature firms on first-tier exchanges set
higher issue prices. However, considering that startup firms with barely any history can
manage to list on second-tier exchanges, it is fair to deduce that maturity plays a smaller
role when pricing second-tier listings. This reasoning could explain the non-significant
and deviating result of the size variable in this study, considering the high proportion of
secondary listings.
Another deviating finding that can be explained by differences in dynamics between
first- and second-tier listings is the positive relationship between leverage and pre-IPO
valuations. At first sight, this might appear non-sensical and wrong, given that previous
findings by Bruton et al. (2010) and Vismara et al. (2012) show the opposite relationship.
However, due to the high degree of information asymmetry associated with second-tier
listings, the presence of a loan implies that a bank has monitored the firm. From the
aftermarket perspective, this can be seen as an indication of quality that justifies higher
initial issue prices. Consequently, the effect of higher leverage could be reversed in the
case of second-tier listings and be seen as something reassuring, contrary to the common
approach to associate leverage with higher risk17
17An example of a company for which a bank loan could send quality signals is the SME Edyoutec
AB. They report 93% of their total assets consisting of capitalized R&D expenditures and 24 million in
accumulated losses. A bank loan would signal that the bank has trust in their future potential.
9 Conclusions 39
9 Conclusions
This study uses a unique dataset of 739 Swedish IPOs issued between 2004 and 2024 to
analyze how different pre-IPO investor types influence IPO issue pricing. The relative
level of IPO valuations and the impact of each investor type are analyzed using an event
study combined with three OLS regression models. The results are then compared to
findings from previous literature and current theoretical frameworks.
The concept of underpricing is used to assess relative valuation pre- and post-IPO issuance.
Positive initial returns are interpreted as indicating "low-valued IPOs", while negative
returns suggest "high-valued IPOs". Underpricing, measured by Cumulative Aggregated
Abnormal Returns (CAAR), serves as the dependent variable in the OLS regressions,
with the involvement of Private Equity and Corporate investors measured using dummy
variables. Conclusions regarding the effect of each investor type are drawn based on the
sign of the coefficients.
The study finds that companies in the Swedish markets restrictively price their listings and
systematically underprice IPOs. This aligns with tendencies observed by several previous
studies covering IPO pricing in different geographical regions (Ritter, 1991; Ritter and
Welch, 2002; Lowry and Schwert, 2002; Vismara et al., 2012; Carter et al., 1998; Miller
and Reilly, 1987; Tiniç, 1988). Additionally, there is evidence that PE-backed firms price
their initial listings higher than non-backed firms, proving that ownership involvement
from private equity affects the level of underpricing. Conversely, corporate-backed IPOs
appear to be valued lower, suggesting that different types of pre-IPO investors affect firms’
IPO valuations differently. The observed differences underline the difficulties of explaining
different investor types’ effect on IPO pricing purely through the quality signaling theory,
as suggested by Ritter and Welch (2002). Furthermore, there appear to be anomalies in
the effect of the variables commonly used when studying IPO pricing. Leverage appears
to have a positive relation to listing prices for less mature IPOs on second-tier exchanges.
The findings of this study both enhance investor insight and contribute to the current
discussion regarding drivers of IPO pricing. Awareness of the tendency for PE-backed
IPOs to be priced higher at IPO compared to non-backed and corporate-backed listings
allows investment professionals to make more informed decisions. The difficulties in
9 Conclusions 40
explaining the perceived differences purely through the quality signaling theory suggest
that future research should explore principal-agent driven explanations instead.
Future research should develop the methodology by incorporating more IPO-specific
variables. For instance, exercised overallotment options manipulate the aftermarket
valuation, and accounting for them could mitigate bias in the underpricing measurement.
Additionally, when studying the first-day returns, subscription rates could be incorporated
to measure the firm-specific market demand. Finally, the ownership percentage of pre-IPO
investor types would better reflect differences in the magnitude of ownership and could
provide more nuanced interpretations. These factors are expected to affect IPO pricing
but could not be included in this study due to time restraints and lack of data access.
Moreover, future studies on underpricing are encouraged to further investigate the reversed
effect of leverage on second-tier exchanges and explore additional differences between
second-tier and primary exchange listings.
References 41
References
Abrahamson, M. and De Ridder, A. (2015). Allocation of shares to foreign and domestic
investors: Firm and ownership characteristics in swedish ipos. Research in International
Business and Finance, 34:52–65.
Akerlof, G. A. (1970). The market for "lemons": Quality uncertainty and the market
mechanism. The Quarterly Journal of Economics, 84(3):488–500.
Arrow, K. J. (1963). Uncertainty and the Welfare Economics of Medical Care Notes On
Contributors, pages 81–136. Columbia University Press, New York Chichester, West
Sussex.
Arthurs, J., Hoskisson, R., Busenitz, L., and Johnson, R. (2008). Managerial agents
watching other agents: Multiple agency conflicts regarding underpricing in ipo firms.
Academy of Management Journal, 51.
Baraldi, A. N. and Enders, C. K. (2010). An introduction to modern missing data analyses.
Journal of School Psychology, 48(1):5–37.
Barry, C. B., Muscarella, C. J., Peavy, J. W., and Vetsuypens, M. R. (1990). The role of
venture capital in the creation of public companies: Evidence from the going-public
process. Journal of Financial Economics, 27(2):447–471.
Bruton, G. D., Filatotchev, I., Chahine, S., and Wright, M. (2010). Governance, ownership
structure, and performance of ipo firms: The impact of different types of private equity
investors and institutional environments. Strategic Management Journal, 31(5):491–509.
Buckle, M., Chen, J., Guo, Q., and Li, X. (2018). The impact of multilateral trading
facilities on price discovery. Financial markets, institutions & instruments, 27(4):145–
165.
Carter, R. B., Dark, F. H., and Singh, A. K. (1998). Underwriter reputation, initial
returns, and the long-run performance of ipo stocks. The Journal of finance (New York),
53(1):285–311.
Chemmanur, T. and Loutskina, E. (2009). How do corporate venture capitalists create
value for entrepreneurial firms? SSRN Electronic Journal.
Cho, J. and Lee, J. (2013). The venture capital certification role in r&d: Evidence from
ipo underpricing in korea. Pacific-Basin finance journal, 23:83–108.
Chou, D.-W., Gombola, M., and Liu, F.-Y. (2006). Earnings management and stock
performance of reverse leveraged buyouts. The Journal of Financial and Quantitative
Analysis, 41(2):407–438.
Darrough, M. and Rangan, S. (2005). Do insiders manipulate earnings when they sell
their shares in an initial public offering? Journal of accounting research, 43(1):1–33.
Daszykowski, M., Kaczmarek, K., Vander Heyden, Y., and Walczak, B. (2007). Robust
statistics in data analysis — a review: Basic concepts. Chemometrics and Intelligent
Laboratory Systems, 85(2):203–219.
References 42
Donders, A. R. T., van der Heijden, G. J., Stijnen, T., and Moons, K. G. (2006). Review:
A gentle introduction to imputation of missing values. Journal of Clinical Epidemiology,
59(10):1087–1091.
Ellis, K., Michaely, R., and O’Hara, M. (1999). A guide to the initial public offering
process. Corporate Finance Review, 3:14–18.
Engelen, P.-J., Heugens, P., van Essen, M., Turturea, R., and Bailey, N. (2020). The
impact of stakeholders’ temporal orientaton on short- and long-term ipo outcomes: A
meta-analysis. Long Range Planning, 53(2):101853.
European Parliament and Council of the European Union (2004). Directive 2004/39/ec
of the european parliament and of the council of 21 april 2004 on markets in financial
instruments and amending council directives 85/611/eec and 93/6/eec and european
parliament and council directive 2000/12/ec and repealing council directive 93/22/eec.
Official Journal of the European Union. L 145/1-44.
Fischer, H. and Pollock, T. (2004). Effects of social capital and power on surviving
transformational change: The case of initial public offerings. The Academy of
Management Journal, 47:463.
Folus, D. and Boutron, E. (2015). 215 Exit Strategies in Private Equity. In Private Equity:
Opportunities and Risks. Oxford University Press.
Fox, J. and Weisberg, S. (2011). Bootstrapping regression models in R. In An R Companion
to Applied Regression, chapter Appendix. SAGE, Los Angeles, 2 edition. Last revision:
10 October 2017.
Ginsberg, A., Hasan, I., and Tucci, C. L. (2011). Unpacking the effects of corporate
venture capital investor ties on the reduction of price discounting among ipo firms.
Entrepreneurship Research Journal, 1(2):1–29.
Hellmann, T. (2002). A theory of strategic venture investing. Journal of Financial
Economics, 64(2):285–314.
Hull, R., Kwak, S., and Walker, R. (2013). IPO Valuation and Insider Manipulation of
R&D. Managerial Finance, Volume 39:888–914.
Ibbotson, R. G. and Jaffe, J. F. (1975). “hot issue” markets. The Journal of Finance,
30(4):1027–1042.
Ivanov, V. I. and Xie, F. (2010). Do corporate venture capitalists add value to start-
up firms? evidence from ipos and acquisitions of vc-backed companies. Financial
Management, 39(1):129–152.
Jain, B. A. and Kini, O. (2005). Industry clustering of initial public offerings. Managerial
and Decision Economics, 27(1):1–20.
Jensen, M. C. and Meckling, W. H. (1976). Theory of the firm: Managerial behavior,
agency costs and ownership structure. Journal of Financial Economics, 3(4):305–360.
Kutner (2004). Applied linear statistical models. McGraw-Hill, New York, NY, 5 edition.
References 43
Liberadzki, M. and Liberadzki, K. (2019). Corporate Hybrid Securities and Preferred
Shares, pages 167–210. Springer International Publishing, Cham.
Ljungqvist, A., Nanda, V., and Singh, R. (2006). Hot markets, investor sentiment, and
ipo pricing. The Journal of Business, 79(4):1667–1702.
Logue, D. E. (1973). On the pricing of unseasoned equity issues: 1965-1969. The Journal
of Financial and Quantitative Analysis, 8(1):91–103.
Loughran, T. and Ritter, J. (2004). Why has ipo underpricing changed over time?
Financial Management, 33(3):5–37.
Loughran, T. and Ritter, J. R. (2000). Why don’t issuers get upset about leaving money
on the table in ipos? The Review of Financial Studies, 15(2):413–443.
Lowry, M., Officer, M. S., and Schwert, G. W. (2010). The variability of ipo initial returns.
The Journal of finance (New York), 65(2):425–465.
Lowry, M. and Schwert, G. W. (2002). Ipo market cycles: Bubbles or sequential learning?
The Journal of Finance, 57(3):1171–1200.
MacKinlay, A. C. (1997). Event studies in economics and finance. Journal of economic
literature, 35(1):13–39.
Masulis, R. W. and Nahata, R. (2009). Financial contracting with strategic
investors: Evidence from corporate venture capital backed ipos. Journal of Financial
Intermediation, 18(4):599–631. Financial Contracting: Theory and Evidence.
Megginson, W. L. and Weiss, K. A. (1991). Venture capitalist certification in initial public
offerings. The Journal of finance (New York), 46(3):879–903.
Michala, D. (2019). Are private equity backed initial public offerings any different? timing,
information asymmetry and post-ipo survival. Journal of Corporate Finance, 59:31–47.
Institutions, Governance and Finance in a Globally Connected Environment.
Miller, R. E. and Reilly, F. K. (1987). An examination of mispricing, returns, and
uncertainty for initial public offerings. Financial Management, 16(2):33–38.
Pratt, M. G. and Foreman, P. O. (2000). Classifying managerial responses to multiple
organizational identities. The Academy of Management Review, 25(1):18–42.
Reese, W. (1998). Ipo underpricing, trading volume, and investor interest. Available at
SSRN 115348.
Reifman, A. and Keyton, K. (2010). Winsorize. In Salkind, N. J., editor, Encyclopedia of
Research Design, pages 1636–1638. SAGE Publications, Inc., Thousand Oaks, CA.
Reilly, F. K. and Hatfield, K. (1969). Investor experience with new stock issues. Financial
Analysts Journal, 25(5):73–80.
Ritter, J. R. (1984). The "hot issue" market of 1980. The Journal of Business, 57(2):215–
240.
References 44
Ritter, J. R. (1991). The long-run performance of initial public offerings. The Journal of
finance (New York), 46(1):3–27.
Ritter, J. R. and Welch, I. (2002). A review of ipo activity, pricing, and allocations. The
Journal of Finance, 57(4):1795–1828.
Ross, D. and Hopkins, M. (2011). What influences investors’ perceptions of private equity
firm window dressing at ipo? Int. J. of Behavioural Accounting and Finance, 2:259 –
272.
Sahlman, W. A. (1990). The structure and governance of venture-capital organizations.
Journal of Financial Economics, 27(2):473–521.
Suresha, B., Nayak, S., Krishna T A, T. A., and Thomas, R. (2023). Determinants of
book built ipo underpricing - differential issue size and market momentum approach
revisited. Cogent economics & finance, 11(2).
Teoh, S., Yang, G., and Zhang, Y. (2009). R-square and market efficiency. SSRN Electronic
Journal, pages 14–18.
Tiniç, S. M. (1988). Anatomy of initial public offerings of common stock. The Journal of
Finance, 43(4):789–822.
Vismara, S., Paleari, S., and Ritter, J. R. (2012). Europe’s second markets for small
companies. European Financial Management, 18(3):352–388.
References 45
Appendix
A Winsorization
Table 6: Data Before and After Winsorization
obs mean std min 25% 50% 75% max
1d CAAR 739 9.58 39.34 -70.51 -10.88 2.79 21.52 408.33
1d CAAR(w)1 739 8.83 34.27 -57.85 -10.88 2.79 21.52 152.04
1w CAAR 738 13.16 53.29 -76.18 -13.65 1.66 24.98 571.19
1w CAAR(w)1 738 12.14 46.31 -62.35 -13.65 1.66 24.98 232.81
1m CAAR 738 14.33 60.61 -74.11 -19.45 0.88 30.02 484.31
1m CAAR(w)1 738 13.64 56.41 -66.95 -19.45 0.88 30.02 289.06
1Winsorized
B Statistical Tests for OLS
Table 7: Correlation Matrix
PEVC Corporate Organic Other OMXSLag Solvency (LN)Size R&D Volume4 Volume20
PEVC 1.0000
Corporate -0.2981 1.0000
Organic -0.6414 -0.1812 1.0000
Other -0.3858 -0.1090 -0.2345 1.0000
OMXSLag 0.0178 -0.0250 0.0512 -0.0702 1.0000
Solvency 0.0331 0.0027 -0.1021 0.0875 -0.0008 1.0000
(LN)Size 0.1263 0.0563 -0.0879 -0.1233 0.0691 -0.1342 1.0000
R&D 0.0198 -0.0179 -0.0915 0.1007 -0.0380 0.0396 -0.2954 1.0000
Volume4 0.0209 -0.0103 -0.0516 0.0433 0.0235 0.0354 0.0533 0.0584 1.0000
Volume20 0.0483 0.0288 -0.0505 -0.0312 0.0149 0.0158 0.0681 0.0005 0.5215 1.0000
References 46
Since Volume on trading day 4 and volume on trading day 20 are never used in the same
regression, their correlation does not matter.
Table 8: Statistical tests for OLS regression
1d CAAR 1w CAAR 1m CAAR
Breusch-Pagan Test Statistic 21.7905 54.1731 24.8700
Breusch-Pagan Test p-value 0.0053 0.0000 0.0031
Jarque-Bera Test Statistic 635.1311 1288.9683 1892.2118
Jarque-Bera Test p-value 0.0000 0.0000 0.0000
VIF (const) 33.6682 33.6773 33.6851
VIF (PE&VC) 1.3655 1.3705 1.3687
VIF (Corporate VC) 1.1566 1.1566 1.1586
VIF (Other) 1.2609 1.2628 1.2609
VIF (OMXS lag) 1.0121 1.0123 1.0123
VIF (Solidity) 1.0340 1.0345 1.0344
VIF (Size) 1.3939 1.3965 1.3962
VIF (R&D) 1.0373 1.0376 1.0388
VIF (OMXS) 1.3528 1.3538 1.3529
VIF (VOLUME 4d) NaN 1.0080 NaN
VIF (VOLUME 20d) NaN NaN 1.0109
Figure 5: 1d CAAR residuals
References 47
Figure 6: 1w CAAR residuals
Figure 7: 1m CAAR residuals
References 48
C Robustness Check Regressions
Table 9: Bootstrapped Regression
(1) (2) (3)
1d CAAR 5d_CAAR 21d_CAAR
PE/VC -0.0677** -0.0780** -0.0830*
(-2.23) (-2.03) (-1.72)
Corporate VC 0.0217 0.137 0.141
(0.37) (1.57) (1.25)
Other -0.0613 -0.0692 -0.107
(-1.33) (-1.18) (-1.52)
OMXS Lag 0.478*** 0.473** 0.577**
(2.94) (2.26) (2.18)
Solvency 0.0701* 0.0664 0.137***
(1.88) (1.53) (2.58)
(LN) Size -0.00286 -0.0104 -0.0179*
(-0.51) (-1.32) (-1.78)
R&D/Size 0.00434 0.00145 -0.0191
(0.07) (0.02) (-0.17)
Exchange Type 0.0816*** 0.0500 0.101**
(2.81) (1.39) (2.15)
Volume 4 0.595**
(2.34)
Volume 20 0.975
(0.79)
Constant 0.0918 0.187** 0.258**
(1.39) (2.14) (2.36)
Observations 739 739 738
* p < 0.1, ** p < 0.05, *** p < 0.01
Table 10: Truncated Regression
(1) (2) (3)
1d CAAR 5d_CAAR 21d_CAAR
PE/VC -0.0663** -0.0908** -0.121**
(-2.02) (-2.15) (-2.17)
Corporate VC 0.0267 0.130 0.108
(0.43) (1.43) (0.93)
Other -0.0370 -0.0648 -0.103
(-0.70) (-0.99) (-1.32)
OMXS Lag 0.499*** 0.466** 0.612**
(2.97) (2.15) (2.33)
Solvency 0.0685** 0.0653* 0.146***
(2.27) (1.91) (3.19)
(LN) Size -1.23e-09 -9.07e-09 -1.29e-08
(-0.19) (-0.81) (-1.18)
R&D -0.00212 -0.0160 0.000410
(-0.03) (-0.18) (0.00)
Exchange Type 0.0623** 0.000654 0.0412
(2.23) (0.02) (0.93)
Volume 4 0.964***
(3.76)
Volume 20 0.946**
(2.26)
Constant 0.0643** 0.0856** 0.0933*
(2.14) (2.20) (1.84)
Observations 631 631 630
* p < 0.1, ** p < 0.05, *** p < 0.01