Does the inclusion of an Over-Allotment Option reduce IPO
underpricing?
– An Australian case study –Amsterdam Business School
Student Willem-Pieter Berger Number 10440062
Study Economics & Business Field Finance
Supervisor Ieva Sakalauskaite Completion 15 June 2015
Abstract
I examine the role of the Over-Allotment Option in IPO underpricing. The sample used consists of 621 IPOs on the Australian Stock Exchange for the period between 2005-2010. My findings are that the OAO has no significant effect on IPO underpricing, although the OAO seems to slightly reduce IPO underpricing. This is in line with previous research, in which Hansen et al. (1987) and Franzke & Schlag (2003) find similar results. Moreover, it can be concluded that the OAO is used in relatively risky IPOs. This extends the vision of Carter & Dark (1990), who find that OAO inclusion is most effective for young and small firms.
Statement of Originality
This document is written by Willem-Pieter Berger who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is 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.
Table of contents
1. Introduction 3
2. Literature Review 5
2.1 Basics of the OAO 5
2.2 Benefits and costs of the OAO 6
2.3 Reasons why the OAO reduces underpricing 7
2.4 Legal framework OAO in Australia 8
3. Hypotheses, Methodology and Data 9
3.1 Hypotheses and Methodology 9
3.2 Data 11 3.3 Sample 12 3.4 Descriptive statistics 12 3.5 Correlation 16 3.6 Heteroskedasticity 17 4. Empirical results 17
5. Conclusion and discussion 19
5.1 Conclusion 19
5.2 Limitation 20
5.3 Further research 20
Reference list 21
1. Introduction
IPO underpricing has been an extensively researched topic in finance literature, searching for an explanation why initial return is often high. High initial return can be costly for the issuing firm and can discourage a firm from going public. Ellul & Pagano (2006) find that IPO underpricing is caused by asymmetric information and price risk. The Over-Allotment Option (OAO), or Greenshoe Option, can reduce underpricing by decreasing price risk, since the option gives the underwriter a tool to execute aftermarket price stabilization. Also, the OAO enables the underwriter to bring extra liquidity into the market.
Normally, when share demand for an IPO turns out to be high, underwriters want to borrow shares to meet this extra demand. The underwriter can borrow these shares from two different parties; primary shares can be borrowed from the issuer, whereas secondary shares can be borrowed from former shareholders. If no OAO is granted, the underwriter needs to take a ‘naked’ short position to meet this extra demand. The short position forces the underwriter to hand the shares back to the issuer or former shareholders. This can be very costly, since most IPOs are underpriced, meaning that share price most often moves upwards. As a result, a loss is frequently incurred on a ‘naked’ short position, so underwriters use it prudently. (Zhang, 2005) This is where the OAO becomes useful. The OAO is the right to buy the overallotted shares at offer price from the issuer or former shareholder within 30 days after the IPO or secondary offering, thereby diversifying all risks away from an increase of the share price. The OAO is an option that must be included in the IPO prospectus, disclosing all specific details. The OAO is in most countries limited to an extra placement of 15% of the initial amount of shares issued. If the share price increases, the underwriter will exercise the OAO and thereby buy the shares from the issuer or former shareholder at offer price, depending on which party granted the option. When share price falls below offer price, the underwriter will close his short position by buying shares in the secondary market. This leaves the underwriter with a trading profit from its short position, equal to the offer price minus the share price. Conclusively, the OAO diversifies all risk on unfavorable price movements away. Therefore, the underwriter always benefits from an OAO. (Franzke & Schlag, 2003)
The effects of the OAO on underpricing are highly versatile, whereby two arguments tend to have the largest impact. Firstly, in a firm-commitment IPO, the risk of unsold shares is shifted to the underwriter. Since IPO investors are allowed to renege on their shares, the underwriter will reduce its offer price to make sure all shares are sold. But whenever an OAO is included, the underwriter has a tool to better manage the share supply, thereby decreasing
reneging risk. Therefore, Hansen et al. (1987) find that the underwriter has less incentive to underprice the IPO if an OAO is included. Secondly, Schultz & Zaman (1994) argue that the OAO reduces IPO underpricing because of its aftermarket price support tool. When a share price falls below the offer price, the IPO is considered to be a bad IPO. This leads to
disappointed investors and underwriters suffering from reputational damage. (Welch, 1992) Therefore, to minimize the chance of a stock price falling below share price, underwriters will underprice IPOs. However, the OAO enables the underwriter to provide aftermarket price support and stabilization when closing its short position in the secondary market, thereby permanently decreasing share supply. This decrease in share supply will increase the share price. So, the OAO enables the underwriter with more tools to provide better risk
management of a share price falling below offer price. Therefore, the OAO should reduce underpricing.
In this thesis I want to examine whether the inclusion of an OAO significantly reduces IPO underpricing. Although previous studies (Hansen et al., Franzke & Schalg) have not found a significant reduction in underpricing, they show signs that the OAO reduces underpricing. Moreover, there are well-grounded reasons in favor of my hypothesis, as summarized in previous paragraph.
Even though the OAO is extensively used in most Western markets, research on the usage and effects of the OAO is limited. To examine the effect of the OAO on underpricing, samples of IPOs with and without OAO are needed. Since all Western markets, except for Australia, use the OAO in almost all IPOs, I focused my research on the Australian IPO market. Due to lack of IPO information before 2005, the tested period is from 2005 till 2010. Multiple regressions are executed where other explanatory variables are added to correct for extraneous influences. I hope to provide up-to-date insights on the effect of the OAO on underpricing, since previous studies are outdated.
The remainder of this thesis proceeds as follows. The next chapter describes the literature on cost and benefits of the OAO and reasons why the OAO should reduce
underpricing. The third chapter discusses the hypothesis, methodology and data used in the regression, after which the descriptive statistics have been summarized. A test for
multicollinearity and heteroskedasticity is executed to scan for errors in the model. In the fourth chapter, the empirical results have been described and discussed. The last section presents the main findings and reasons on possibilities for further research.
2. Literature review 2.1 Basics of the OAO
The OAO is an agreement between the underwriter and the issuer or the former shareholders. The OAO enables the underwriter to sell more shares to the public than originally planned, without additional risk. Therefore, the OAO gives the underwriter share supply flexibility. For example, instead of being committed to sell 1.15 million shares, the underwriter and the issuer could agree to sell 1 million shares with an additional option to sell another 150,000 shares if share demand turns out to be larger than anticipated. Whenever the underwriter overallots, or oversells the IPO, the underwriter must borrow shares from the issuer or the former
shareholders to be able to sell those ‘extra’ shares. Borrowing those extra shares creates a short forward position for the underwriter and a long forward position for the lender. These extra shares must be sold to the public at offer price, otherwise a trading profit could be made from exercising the OAO, which is forbidden by law. When the underwriter takes a short position, the underwriter has an obligation to the lenders to close the position within 30 days following the IPO. The underwriter can do this in two different ways, or in a combination of the two.
Firstly, the underwriter can pay the issuer or former shareholder back by handing back the borrowed shares. Since the underwriter sold more shares than they had in position, the underwriter first needs to acquire the shares in the secondary market. This acquisition of shares boosts the share price, since it increases the demand for the stock. This is called aftermarket price support. Whenever the underwriter closes his short position, and thereby hands back the borrowed shares, share supply is permanently decreased. This results in an increase of the share price. This alternative will be chosen when markets are bearish, since the payoff from the short position is the offer price minus the share price. If the share price is below offer price, a trading profit is made.
Secondly, the other alternative is to pay the lender back by exercising the OAO. When the OAO is exercised, the ‘extra’ shares are bought from the issuer or former shareholder, dependent on who granted the OAO. The costs for the underwriter equal the offer price minus the underwriting fee. In this case, the short position is closed by a repayment of cash instead of shares to the lender. This repayment does not affect the share supply and it does not give any aftermarket price support. This alternative will only be chosen in times of a bullish market, since underwriters want to avoid the share price falling below offer price. The underlying reason it that this causes reputational damage and disappointed investors.
As mentioned earlier, an OAO consist of a call option and a short position in a forward contract. The call option is the right to purchase the share at offer price from the issuer or previous shareholders, whereas the short forward contract is executed when the underwriter buys shares in the secondary market. The pay-off structure can be seen as:
!! = max ! − !!, !"
Where !! is the final pay-off for the underwriter, ! is the offer price, !! is the share price and ! is the gross spread. (Franzke, 2003)
Conclusively, OAO inclusion leads to a larger offer size, a higher market price for the securities, due to stabilization bids, or a combination of both.
2.2 Benefits and costs of OAO
OAOs are heavily used in most Western markets. Franzke & Schlag (2003) are of the option that the former shareholders or the issuer are the ones paying the costs of the OAO. In times of a bullish market, the OAO will be exercised and the former shareholders will receive the offer price, which is lower than the proceeds that could have been made if sold at market price. In a bearish market whereby the share price is below the offer price, the underwriter will return the overallotted shares to the former shareholders or issuer. In both alternatives, the former shareholders or issuers are the ones incurring a loss resulting from the OAO. Ellul & Pagano (2006) argue that investment banks are profit maximization firms with large market power. So it is logical to think that issuers, former shareholders and/or IPO investors may face wealth losses due to granting the OAO. Also, when the OAO is exercised, the issuer could raise more capital than initially planned. This can be costly for the issuer if the proceeds cannot be used efficiently, potentially leading to a drop in share value. (Corwin, 2003)
On the other hand, the inclusion of the OAO is often argued to be in the best interest of the IPO underwriter, issuer and its investors. (Prabhala & Puri, 1998) Aggarwal (2003) finds that aftermarket price support is beneficial for all parties,because the OAO mitigates the risk of flippers, i.e. investors who sell their shares shortly after the IPO. Benviste et al. (1998) argue that institutional investors are the ones who profit most from the price stabilization efforts of the underwriter. The aftermarket price support decreases the risk of a poorly perceived IPO, since prices are expected to be higher due to the underwriter’s short position. This is because the underwriter’s short position creates upward share price pressure. Carter and Dark (1990) find that OAOs are valuable for issuers and former shareholders because it gives the underwriter greater incentive to market the IPO aggressively. Furthermore, they find that IPO inclusion is more beneficial for riskier firms, since appropriate sales and share
allotment levels are harder to establish upfront. Besides, Carter and Dark (1990) find that the OAO can help building and maintaining a good relationship between underwriter and
investor, since the OAO enables the underwriter to satisfy more clients due to its flexibility to increase share supply in case of high demand.
All in all, since the OAO is almost always included in IPOs in Western countries and is increasingly used in developing markets, the perceived benefits of the OAO tend to outweigh the costs.
2.3 Reasons why the OAO reduces IPO underpricing
Firstly, the OAO reduces IPO underpricing due to the decreasing incentive to underprice the IPO, because of the mitigation of reneging costs. Reneging costs arise when the IPO is executed on a firm-commitment basis. In a firm-commitment IPO, the underwriter bares the risk of all unsold shares. The underwriter allots the shares before the offering date, but
investors can withdraw their buy order. IPO investors are allowed to renege on their buy order without legal penalty within five days after the subscription date. (Hansen et al., 1987) So when the effective date has arrived, share demand may be smaller. The option to renege can be seen as a long put option, because the value of reneging increases when the share price falls. (Schulz & Zaman, 1994) This leaves the underwriter with unsold shares, which leads to increased costs due to carrying costs and the risk of unfavorable price fluctuations. Therefore, the underwriter will reduce the offer price to ensure all shares are sold. This problem can be mitigated when an OAO is granted, since the option gives the underwriter share supply flexibility. The OAO enables the underwriter to reduce its initial offer size and overallot the issue, without additional risk, if share demand turns out to be large. (Schulz & Zaman, 1994) The share supply flexibility provided by the OAO enables the underwriter to price the IPO more accurately. Therefore, the incentive to underprice the IPO is reduced when an OAO is included. (Hansen et al., 1987)
Secondly, the OAO enables the underwriter to exercise aftermarket support without potential losses on his short position. Aftermarket price support can also be executed when taking a ‘naked’ short position, but in this case the underwriter will incur a large loss when the share price rises. Aftermarket price support of an IPO stabilizes and increases the share price when share price would otherwise have fallen below offer price. The support is executed when the underwriter closes his short position by buying shares in the secondary market, thereby decreasing share supply permanently. This decrease in supply will cause an increase in share price. Schultz & Zaman (1994) argue that the reason for maintaining the share price
at or above offer price is the same as the reason for initially establishing the offer price below the real share price: A negative initial return is perceived as an unsuccessful IPO. Since market reputation and its network of institutional investors are valuable assets for the underwriter, they want to ensure that initial return does not become negative. Hansen, et al. (1987) explain that the aftermarket price support enables the underwriter to provide better risk management, since is diminishes the risk of a lower share price than the offer price.
Therefore, due to the mitigation of the risk of negative initial return, the underwriter has less incentive to underprice the IPO. Furthermore, Aggarwal (2000) finds that IPO investors receive IPOs better when the underwriter is able to provide aftermarket support due to the OAO. This means that the chance on negative initial return is even smaller.
Lastly, as explained before, the exercise of the OAO can be seen as a call option, whereas the share repurchase in the secondary market is regarded as a short forward contract. Following the put-call parity, a short forward contract plus a long call option is equal to a put option and some constant payment. (Franzke & Schlag, 2003) The put increases in value when the IPO is less underpriced, since the share is then more likely to fall in-the-money. This means that the underwriter has an incentive to reduce the level of underpricing. However, underwriters are reluctant when it comes to overpricing. Main reason is that the underwriter faces potential loss of reputation when overpricing, due to unhappy IPO
investors. Furthermore, Rock (1986) finds that share demand will be low when overpricing, since the informed investor will choose not to subscribe for the IPO. Also, the OAO only has value in case the underwriter overallots the IPO. Therefore, underwriters are reluctant to overprice IPOs.
2.4 Legal framework OAO in Australia
Telstra 1 was in 1997 the first company that used an OAO in its IPO on the Australian Stock Exchange. Unlike most Western countries, there is no legislative safe harbor for market stabilization activities in Australia. Market stabilization occurs when a person or corporation takes part in a transaction that creates an artificial trading price, such as with aftermarket price support. However, the Australian Securities and Investments Commission (ASIC) may issue letters that state that no action will be conducted if parties get involved in market
stabilization. In practice, underwriters are legally safe carrying out market stabilization activities after they have fully disclosed this in the prospectus and ensured that the
stabilization has been done following the standard terms and conditions of the ASIC. The standard application for a ‘no action’ letter requires the offer to be an IPO of which the gross
proceeds must be at least $50 million and the OAO must be 15% at most. However, this is solely true for a ‘no action’ letter via the standard application. Since ‘no action’ letters are granted case-by-case, exceptions are common.
Besides, the ASIC has established strict rules regarding market manipulation. The underwriter must follow disclosure conditions, where stabilization bids must be disclosed on the ASX announcement platform. The market stabilization bid can only be made if share price is below offer price. Lastly, like most markets, the stabilization period is limited to 30 days.
3. Hypotheses, Methodology and Data 3.1 Hypotheses and Methodology
Based on the literature, there are plausible reasons to accept that OAO inclusion reduces underpricing. Although Hansen et al. (1987) and Franzke & Schlag (2003) do not find that the OAO significantly reduce IPO underpricing, they show reasons to believe that the OAO has some effect on underpricing. For example, Hansen et al. (1987) show that OAO inclusion reduces IPO underpricing with 2,5%, whereas Franzke & Schlag (2003) observe that the mean and median of the first day return are indeed higher for issues with an OAO included. However, the difference is not significant at conventional levels. Yet, my hypothesis is that OAO inclusion reduces IPO underpricing.
This paper investigates the influence of the inclusion of the OAO on market returns. This is done using OLS. The first regression is the basic regression, where the OAO is regressed on underpricing. Afterwards, more variables are added to control for market status, firm age and size, high technology firms, relative size of primary shares, underwriter- and auditor reputation on underpricing. The variables are further explained below.
!" !" = ! + !!∗ !"#$%&!"#"$"%& + !!∗ !"#$%&'$() + !!∗ !"#$%&'() + !!∗ !"! + !!∗ !"#$%"&'ℎ!"#$
+!!∗ !"#ℎ!"#ℎ + !!∗ !"#$%&%'($%)$* + !!∗ !"#$%&'()* + !
The model shown above contains all control variables. However, multiple regressions are executed be it that not all variables are always excluded.
The dependent variable is the log of underpricing. The value of underpricing is calculated as follows:
!"!=(!!− !!) !! ×100
Where !! is the closing price of the share after the first trading day, !! is the offer price and !"! is the arithmetic return. This calculation is in line with Ritter (2008) and is commonly
accepted. The underpricing is measured in log, since it is conventional to have the returns logged in a model for securities.
The OAO is a dummy variable which tells whether an Over-Allotment option is included. If an OAO is included, the value is 1 and otherwise 0. The findings for this value are most important, since my hypothesis is that the OAO reduces underpricing. So OAO is expected to have a negative correlation with underpricing.
The state of the market is measured by lnMarketStatus, because a ‘hot’ IPO market has a positive influence on IPO underpricing. Derrien (2005) shows that IPOs can be
overpriced, but still exhibit positive initial return thanks to a hot market. The more favorable the noise trader sentiment, the higher the IPO price. The number of IPOs done in the month before each IPO show how ‘hot’ the market is.
The size of the firm at the moment of IPO is measured as lnFirmSize. Corwin (2003) finds that underpricing is in general higher for small firms with higher price uncertainty. This is in line with Carter & Dark (1990), who find that riskier firms are more underpriced. So firm size is expected to be negatively correlated with underpricing, since bigger firms are less risky. The size of the firm is measured by gross proceeds of the IPO, so by market
capitalization at the time of IPO. This is calculated by multiplying shares issued with offer price. The assets at the moment of IPO or the number of employees are found to be a better measure of firm size, but unfortunately these specific data was not available.
The age of the firm is calculated by lnFirmAge. Loughan & Ritter (2004) argue that younger firms are riskier. Hence, firm age is expected to have a negative correlation with underpricing. This is also confirmed by Carter & Dark (1990). Firm age is calculated by the period between the foundation date and the date of the IPO. Firm age is measured in days.
The HighTech dummy is included for IPOs in which the firm focuses on high technology. Darrien & Womack (2003) find that high-tech firms are generally more
overpriced than non-tech firms. To control for this effect, a value of 1 is given for high-tech firms. High-high-tech enhances overpricing, so a positive correlation is expected.
The PrimaryShares is a dummy variable which shows 1 if solely primary shares are issued, 0 if solely secondary shares are issued and 0.5 if a mixture of both is used. Campbell et al. (2007) state that a relative high portion of secondary shares in an IPO sends bad signals to IPO investors, because if it is a ‘good’ IPO, the previous shareholders won’t sell their shares. Therefore, the expected correlation is positive.
The UnderwriterRep variable is to control for difference in the reputation of the underwriter. Carter & Dark (1998) find that IPOs managed by more reputable underwriters
are associated with less short-run underpricing. They introduced a tombstone ranking of reputation per underwriter. Unfortunately, this list only contains American underwriters. Although mostly the same American underwriters are active in the Australian market, their reputation in Australia might be different. Therefore, an ordinal ranking is made which
accounts for a 1 is the market share is top 10 in Australia, 2 if the market share is top 20 and 3 otherwise. A positive relation with underpricing is expected, because the better the reputation, the lower the ordinal ranking.
Lastly, the AuditorRep variable stands for auditor reputation. Beatty (1996) shows in his research that firms which have a more reputable auditor have a significantly smaller IPO return, so less underpricing. This means that auditor reputation and underpricing are inversely correlated. A dummy is created for Big 4 accounting firms. These firms are Deloitte,
PricewaterhouseCoopers, Ernst & Young and KPMG.
The explanatory variables !"#$%&'()($(*+, !"#$%&'$()!!"#!!"#$%&'() are all logged. This is done because the difference in value is more important for smaller number than for higher number. For example, the movements of firm age on underpricing has more impact if a company is 4 years old than if a company is 100 years old. The same conclusions are drawn for the other variables.
To summarize the variables and expected correlation with the dependent variable underpricing, table I is drawn.
Table I: Explanation of relationship between variable and underpricing Variable Expected sign Description
lnMarketStatus + The log of the number of IPOs done in last month lnFirmSize - The log of the gross proceeds of the IPO
lnFirmAge - The log of the number of days of foundation till IPO OAO - A dummy variable which controls for Overallotment option HighTech + A dummy variable which controls for high-tech firms Primary Shares + A dummy variable which controls for primary shares issued UnderwriterRep + An ordinal ranking which is listed following market share AuditorRep - A dummy variable which controls for BIG4-firms
3.2 Data
The data used in this thesis was not readily available. Most of the required data has been found on Thomson One Reuters, especially the IPO details whether an OAO was included, IPO and OAO size, firm industry and characteristics of the bookrunner and auditor used. Due
to incomplete information, often IPO prospectuses were needed to get additional information. Moreover, the foundation date of the firm included is mostly derived from the companies website, since Thomson One Reuters and Datastream do not have complete information. The share price movements and most other aftermarket data are derived from Datastream and connected with the data from Thomson One Reuters. The variables collected are specified in appendix I.
3.3 Sample
This thesis focuses on the Australian market because of the distribution of OAO inclusion. In other Western stock exchanges, OAOs are extensively used. To measure the effect of the OAO on underpricing, several observations are needed for IPOs with and without an OAO. Since almost all observations have an OAO included in other Western countries, measuring the OAO effect on underpricing in these markets is not possible. Alternatively, it would be interesting to focus on emerging markets, like India and China, where the OAO usage is upcoming. Unfortunately, data availability in these countries is incomplete and not possible to retrieve. The choice for the period of 2005-2010 is because of the availability of data. Before 2005, many observations had limited availability of data, especially OAO information was unavailable. The decision to include only IPOs on the Australian Stock Exchange (ASX) is made to ensure an equal comparison between IPOs in Australia, since the characteristics of companies listed on other Australian exchanges are considerably different. Although the OAO is used in secondary offerings as well, this thesis solely focuses on IPOs.
According to Thomson One Reuters, there were 830 IPOs on the ASX between 2005-2010. This information has been compared with the data from the ASX website to ensure no significant information was omitted. Of this number, 794 IPO’s were stocks of companies that were solely issued at the ASX. In addition, 59 IPO’s were excluded due to the cancelation of the IPO process. Furthermore, another 100 IPOs were erased from the sample due to the lack of prospectuses or incomplete data. Finally, the dataset included 621 IPOs, all with sufficient data.
3.4 Descriptive statistics
This section explains several descriptive statistics founded from the data sample. Table II provides an outlook on the inclusion of the OAO in all IPOs per year. The size of the gross proceeds per year is shown, so that the proceeds can be seen per year.
Table II: Details of the sample size
Year Number of IPOs IPOs with OAA Percentage with OAO Gross proceeds (million $) 2005 146 38 26.0% 68.5 2006 171 50 29.2% 43.6 2007 228 60 26.3% 44.8 2008 53 15 28.3% 18.1 2009 23 1 4.3% 112.8 Total 621 164 26.4% 287.8
Year 2007 is by far the most active IPO year, which is as expected since the credit crises started at the end of that year in Australia. The crisis started after the crisis on the American market, due to different market characteristics. After the start of the crisis, the number of IPOs significantly reduced. The inclusion of the OAO tends to be constant, around a quarter of all IPOs contain an OAO, except for the last year. In 2009 a relative drop in the inclusion of OAOs can be found. A reason for the drop could be that the sample size in 2009 is small (23), which could explain the deviation results.Remarkably, the gross proceeds of all IPOs per year are largest in 2009, whereas just 23 IPOs were issued that year. The large size is explainable by the large IPO of Myer Holdings Ltd. The gross proceeds tend to be slightly inversely correlated with the number of IPOs. Figure I shows the distribution of gross
proceeds. The gross proceeds of IPOs with OAO per year appear to be more stable than the IPOs without OAO.
Figure I: Gross proceeds per year divided by OAO inclusion
Source: Thomson One Reuters
Table III shows a detailed overview of most descriptive statistics. Gross proceeds, firm age and underpricing are divided into three groups. One group containing all IPOs, one group containing solely of IPOs with OAO and the last group only containing data from IPOs without OAO. 0! 20! 40! 60! 80! 100! 120! Without!OAO! With!OAO!
Table III Summary statistics
The sample contains all IPOs on the Australian Stock Exchange between 2005-2010 which had a complete prospectus. Panel A: Firm Data Summary Statistics
Full sample Sample With OAO Sample Without OAO
Year IPOs Total gross proceeds (million $)
Age (days)
Underpricing (%)
IPOs Total gross proceeds (million $)
Age (days)
Underpricing (%)
IPOs Total gross proceeds (million $) Age (days) Underpricing (%) 2005 146 68.5 3059 1.128 38 16.6 2220 1.123 108 51.9 3354 1.129 2006 171 43.6 1626 1.327 50 7.8 1221 1.317 121 35.8 1793 1.331 2007 228 44.8 1938 1.272 60 15.7 1589 1.217 168 29.1 2063 1.292 2008 53 18.1 1588 1.085 15 7.2 2427 1.118 38 10.9 1257 1.072 2009 23 112.8 2677 1.172 1 13.0 1296 0.762 22 99.8 2739 1.191 Total 621 287.8 164 60.2 457 227.6 Average 170 32.7 2113 1.233 32.8 14.2 1698 1.214 169 38.7 2262 1.241 Median 171 47.4 682 1.091 34.0 3.8 552 1.051 155 5.5 732 1.101 SD 57.8 130.0 3724 0.644 45.3 53.5 3266 0.727 59.5 150.1 3867 0.612 Min. 23 1570 8 0.016 1 0.2 8 0.110 22 1570 8 0.016 Max. 228 81.1 28654 8.147 60 5.8 2402 8.147 168 80.1 28654 6.198
Distribution across Fama-French 12 industry groups (621 observations) Consumer Products 30 (4.8%) Industrials 28 (4.5%) Consumer Products 8 (4.9%) Industrials 3 (1.8%) Consumer Products 22 (4.8%) Industrials 25 (5.5%) Consumer Staples 16 (2.6%) Materials 321
(51.7%)
Consumer Staples 7 (4.3%) Materials 96 (58.5%)
Consumer Staples 9 (2.0%) Materials 225 (49.2%) Energy and Power 63
(10.1%)
Media and Entertain 12 (1.9%) Energy and Power 22 (13.4%)
Media and Entertain 1 (0.6%) Energy and Power 41 (9.0%) Media and Entertain 11 (2.4%) Financials 58 (9.3%) Real Estate 17 (2.7%) Financials 12 (7.3%) Real Estate 2 (1.2%) Financials 46
(10.0%)
Real Estate 15 (3.3%) Healthcare 33 (5.3%) Retail 11 (1.8%) Healthcare 8 (4.9%) Retail 2 (1.2%) Healthcare 25 (5.5%) Retail 9 (2.0%) High Technology 25 (4.0%) Telecommunications 7 (1.1%) High Technology 2 (1.2%) Telecommunications 1 (0.6%) High Technology 23 (5.0%) Telecommunications 6 (1.3%)
Gross proceeds and firm age are expected to be smaller for firms with an OAO, since Carter & Dark (1990) find that an OAO is included in IPOs which are relatively risky. Gross proceeds and firm age are on average $14.2 million and 1698 days, respectively, for IPOs with an OAO. For firms without an OAO, gross proceeds and firm age are $38.7 million and 2262 days. These findings are exactly in line with previous studies.
My hypothesis is that OAO inclusion reduces IPO underpricing. The average
underpricing for IPOs with OAO is 21.4%, whereas the underpricing for IPOs without OAO is 24.1%. Although the difference is small, there are indications that OAO inclusion lowers underpricing. Moreover, in all years, except for 2008, the level of underpricing is lower for IPOs with an OAO included. The indications are confirmed when looking at figure II, where the line with OAO clearly is below the line without OAO. Also, the median shows a 5% decrease in underpricing if an OAO is included, which is even more than the mean.
In 2009, the average underpricing is found to be 0.763, equal to an overpricing of 23.7%. This seems to be an extreme case, since IPOs are seldom overpriced that much. However, there was just one IPO with an OAO in 2009. Therefore, due to the small sample size, the number may not be reliable.
The average returns are compared to the initial offer price in figure III. The time interval includes 30 days, the period in which the OAO can be exercised or aftermarket price support can be executed. The line of returns when an OAO is included is increasing towards the end of the period of 30 days that the OAO can be executed. This can be explained by the underwriter exercising aftermarket price support. The increase in the line for IPOs without OAO is increasing in the beginning. This can be explained by a higher level of underpricing, which is correlated with higher first week return.
Figure II: Average First Day Return Figure III: Average return compared to offer price
Source: Datastream 0,6! 0,8! 1! 1,2! 1,4! 2005! 2006! 2007! 2008! 2009! With! OAO! Without! OAO! 1,1! 1,15! 1,2! 1,25! 1,3! With! OAO! Without! OAO!
In the second part of table III, the inclusion of OAO is compared between the 12 industry groups formed by Fama and French. The size of IPOs in the materials industry is large (51.7%), but this is because Australia has a lot of rapidly growing mining companies. The OAO inclusion is expected to be larger for risky industries. However, the findings do not support this. For example, the high technology is a typical industry that is risky. Risky
companies and industries are found to be relatively more underpriced. (Kirkulak & Davis, 2005) In my dataset, the OAO is solely used in 2 (1.2% of all OAO-IPOs) high technology IPOs, whereas the high technology IPOs without OAO are 23 (5.0% of all non-OAO-IPOs). Moreover, the materials industry, which is characterized as a relative stable industry, has a larger percentage of OAO included IPOs (58.2% against 49.2%).
3.5 Correlation
Variables that are significantly correlated will cause multicollinearity. Therefore, if a variable is significant at 1%, or has a correlation above 0.30 with another variable, it will be tested in different regressions. Table IV shows the correlation between the dependent and all
independent variables. Overall, only the variable lnSize shows a high correlation with Primary Shares, Auditor- and Underwriter reputation. Therefore, a variance inflation factor (VIF) test is executed.
Table IV: Correlation matrix all variables
lnUnderpricing lnMarketSt lnSize lnAge High- Tech
OAO Auditor Under- writer Primary Shares lnUnderpricing 1.00 lnMarketstatus 0.09 1.00 lnSize -0.07 0.08 1.00 lnAge -0.08 -0.02 0.27** 1.00 HighTech -0.12 -0.05 0.08 0.08 1.00 OAO -0.04 0.05 -0.17 -0.06 -0.08 1.00 Auditor -0.01 0.13 0.42*** 0.25* -0.02 -0.08 1.00 Underwriter 0.05 -0.08 -0.62*** -0.25* -0.11 0.15 -0.30** 1.00 PrimaryShares -0.05 -0.08 -0.42*** -0.27** -0.08 0.12 -0.18 0.30** 1.00 *, ** and *** show the level of significance at a 10%, 5% and 1% level, respectively
Source: Thomson One Reuters, IPO prospectuses and Datastream
Table V shows the results of the VIF test. Whenever the VIF values are above 10, further investigation for multicollinearity needs to be conducted. As observed, the VIF value is maximum 1.99 and the mean VIF is 1.31. Therefore, it can be concluded that none of the variables suffer from multicollinearity.
Table V: VIF test for all variables
Variable VIF 1/VIF
lnMarketStatus 1.03 0.97 lnSize 1.99 0.50 lnAge 1.16 0.86 HighTech 1.03 0.97 OAO 1.05 0.96 Auditor 1.28 0.78 Underwriter 1.70 0.59 PrimaryShares 1.27 0.79 Mean VIF 1.31
Source: Thomson One Reuters, IPO prospectuses and Datastream 3.6 Heteroskedasticity
An assumption of the OLS states that variances must be equal. This means that the model may not suffer from heteroskedasticity. Hence, it is important to test for heteroskedasticity. The Breusch-Pagan test compares all variances with a chi2 test. The null hypothesis states that there is constant variance, also homoscedasticity. Whenever the p-value is smaller than 0.05, the null hypothesis is rejected and there is significant evidence that there is
heteroskecasticity. The results are shown in table VI. Since the p-value is 0.00, the null hypothesis is rejected. The model suffers from heteroskedasticity, so it is important to use a regression model that controls for robustness.
Table VI: Breusch-Pagan test
Chi2 296,21
Prob > Chi2 0,00
Source: Thomson One Reuters, IPO prospectuses and Datastream
4. Empirical results
Five different regression results are provided in table VII. The outcome of the OAO variable is most important in the findings. The first regression contains solely the dummy variable OAO. In the second regression control variables are added, namely the variables of market size, firm size, firm age and OAO. In the third regression, the variable that controls for
primary shares is added. Thereafter, a high technology dummy is included. Lastly, auditor and underwriting reputation are also added. All regressions are checked on robustness, since het Breusch-Pagan test has shown that heteroskedasticity is present.
In all different regressions, the correlation of all variables are as expected. The inclusion of OAO is found to reduce underpricing, although the value is not significant.
Moreover, the value of OAO is negative in all regressions, but it is never significant. Noteworthy, the value is most significant in the third regression where it almost equals the 10% significance. In this model, the inclusion of an OAO appears to reduce the underpricing by 5.4%, which is more than expected when compared with the descriptive statistics. The lack of significance of the OAO variable can be explained by the versatile effects of the OAO, which causes different effects per firm.
Table VII: Regressions on underpricing
(1) (2) (3) (4) (5)
Market Status (log) 0.060 (2.58)*** 0.057 (2.45)** 0.054 (2.19)** -0.053 (2.17)**
Firm Size (log) -0.021
(1.93)* -0.031 (2.63)*** -0.030 (2.46)** -0.031 (1.99)**
Firm Age (log) -0.015
(1.52) -0.019 (1.93)* -0.018 (1.77)* -0.018 (1.82)* OAO -0.039 (1.22) -0.052 (1.37) -0.047 (1.26) -0.054 (1.47) -0.054 (1.48) High Technology Firm -0.236 (1.20) -0.234 (1.21) Primary Shares -0.207 (3.05)*** -0.217 (3.05)*** -0.218 (3.03)*** Auditor Reputation 0.012 (0.40) Underwriter Reputation 0.0236 (0.04) Firm Size (log)*OAO Observations 621 621 621 621 !! 0.02 0.03 0.04 0.04
Note: *, ** and *** shows the level of significance at a 10%, 5% and 1% level Source: Thomson One Reuters and IPO prospectuses
The third regression shows that four independent variables are significant and have the highest R-squared value. The log of firm age is significant at 1%, log market status and log firm age are significant at 5% and primary shares is significant at 10%. Only the OAO- and high tech dummy are found not to be significant. All values are in line with existing literature, wherein market status enhances IPO underpricing as opposed to firm size, firm age, OAO inclusion, high tech and primary shares, which reduce underpricing. This is because these values decrease the risk of an IPO.
In the fourth regression, auditor- and underwriter reputation are included, but these values are found to be highly insignificant. This is in contrast to the findings of Beatty (1996)
and Carter & Dark (1990), who have found a significant impact of auditor- and underwriter reputation on underpricing. Also, auditor- and underwriter reputation were expected to reduce firm risk. However, both values are positive, so that a better reputation leads to more
underpricing. Although, due to the high insignificance there is no reason to assume that there is any relation between these values.
The R-squared is low in all the regressions, with the highest level in the third
regression of 0.04. This means that the relative percentage of the response variable variation that is explained by the linear model is low. This does not mean that the regressions have no meaning, since most values are found to be significant. Therefore, important conclusions can still be drawn from the results.
5. Conclusion and discussion 5.1 Conclusion
The purpose of this thesis has been to examine whether the OAO reduces IPO underpricing. The existing literature provided two arguments in favor of the OAO reducing underpricing. Firstly, Hansen et al. (1987) argue that the OAO reduces underpricing, since the potential reneging costs are mitigated due to share supply flexibility. Therefore, there is less need for the underwriter to underprice the IPO. Secondly, Schulz & Zaman (1994) find that the risk on an unsuccessful IPO (where the share price falls below offer price) is better managed, due to the aftermarket price support tool provided by the OAO. Thus, the underwriter has less risk in fairly pricing the IPO. Furthermore, the long put position from the underwriter provides incentive to price the IPO closer to be in-the-money, equal to reduce underpricing. (Franzke & Schlag, 2003)
The average underpricing is found to be 23,3% for all IPOs. When separated by OAO inclusion, the level of underpricing is 21,4% for IPOs without OAO whereas for IPOs with OAO the underpricing is 24,1%. This is line with previous studies (Hansen et al, Franzke & Schlag) who also find a smaller mean for IPO underpricing if an OAO is included. For example, Hansen et al. (1987) find that OAO inclusion reduces underpricing with 2.5%, where my findings for the Australian market are that the OAO reduces underpricing with 2,7%, so the results are similar. The level of underpricing is found to be smaller for older and bigger firms. This is in line with existing literature since those IPOs are less risky. Moreover, an OAO is included when firm size is smaller and the age of the company is smaller. This is supported by Carter & Dark (1990) who find that an OAO should be included for risky IPOs.
For IPOs with OAO, gross proceeds and firm age are $14.2 million and 1698 days, respectively. For IPOs without OAO these values are $38.7 million and 2262 days. Both values are found to be significant in the regression.
5.2. Limitations
The effects of the OAO are highly versatile, which makes it an interesting subject. Unfortunately, data availability is limited. Research on the OAO in emerging markets is practically not feasible, whereas in developed markets the OAO is used too frequently to research the OAO effects. Hence, as far as I know, Australia and Singapore are the only countries were an up-to-date research is feasible. Furthermore, the details of the price stabilization, like dates, prices and volumes, are not publicly available. This makes it
impossible to test the exact effect of the price support activities. Conclusively, since contracts between underwriter and issuer are not made public, the possibilities for OAO research are limited.
5.3 Further research
Firstly, it would be interesting if a qualitative research would be conducted focusing on the incentive for issuers to include an OAO. Since almost all Western IPOs have an OAO included, the benefits of the OAO appear to outweigh the costs obviously. However, in previous research, such benefits are not found to be that significant. Furthermore, an investigation of the effects and efficiency of price stabilization activities in the aftermarket may lead to interesting conclusions. Because these activities enable the underwriter to support the share price, this has a significant impact on the success of an IPO. For now, Schultz & Zaman (1994) are the only ones who carried out such research.
Reference list
Aggarwal, R. (2000). Stabilization activities by underwriters after initial public offerings, Journal of Finance 55(3), 1075-1103.
Benveniste, L., Erdal, S.M., & Wilhelm, W.J. (1998). Who benefits from secondary market price stabilization of IPOs?. Journal of Banking & Finance 22, 741-767.
Chowdhry, B., & Nanda, V. (1996). Stabilization, Syndication, and Pricing of IPOs. Journal of Financial and Quantitative Analysis, 25-42.
Corwin, S. A. (2003). The determinants of underpricing for seasoned equity offers. The Journal of Finance 58(5), 2249-2279.
Derrien, O. D. (2005). IPO Pricing in ‘Hot’ Market Conditions: Who Leaves Money on the Table?. The Journal of Finance 150(1), 487-512.
Ellul, A., & Pagano, M. (2006). IPO underpricing and After-Market Liquidity. Review of Financial Study, 381-421.
Franzke, S. A., Schlag, C. (2003). Over-Allotment Options in IPOs on Germany’s Neuer Markt – An Empirical Investigation -. EFMA 2004 Basel Meetings Paper.
Hansen, R. S., Fuller, B. R., & Janjigian, V. (1987). The Over-Allotment Option and Equity Financing Flotation Costs: An Empirical Investigation. Financial Management, 24-32.
Loughran, T., Ritter, J. (2004). Why has IPO underpricing changed over time?. Financial Management 33(3), 5-37.
Prabhala, N. R., Puri, M. (1998). How does underwriter support affect IPOs? Empirical evidence. Yale School of Management Working paper.
Muscarella, C. J., Peavy III, J. W., & Vetsuypens, M. R. (1992) Optimal Exercise of the Over-Allotment Option in IPOs. Financial Analyst Journal, 76-82.
Schultz, P. H., Zaman, M. A. (1994). Aftermarket support and underpricing of initial public offerings. Journal of Financial Economics, 199-219.
Sherman, A.G. (2012). The pricing of Best Efforts New Issues. The Journal of Finance 47(2), 781-790.
Rock, K. F. (1986). Why new issues are underpriced. Journal of Financial Economics 15, 187-212.
Welch, I. (1992). Sequential Sales, Learning, and Cascades. The Journal of Finance 47(2), 695-732.
Zhang, D. (2004). Why do IPO underwriters allocate extra shares when they expect to buy them back?. Journal of Financial and Quantitative Analysis 39(3), 571-595
Appendix
Appendix I: Specification of needed data
Variables Description
Foundation date The date that the company was founded Issue date The date that the company went IPO Macro description Market of the company
OAO IPO OAO inclusion in prospectus OAO used Size of the exercise of the OAO OAO option Size of the option of the OAO Primary shares Total primary shares offered in IPO Secondary shares Total secondary shares offered in IPO Bookrunner Bookrunner of the specific IPO Auditor Auditor in use in times of IPO Offer price Price of the IPO offer per share
Share price All share prices of the IPO for the year after the IPO Gross Proceeds Total proceeds of the IPO
Market status The number of IPOs done on the ASX the previous month Offering technique The offering technique used in the IPO
Source: Datastream, Thomson One Reuters and IPO prospectuses
Appendix II: Summary statistics all variables
Variable Obs Mean Std. Dev. Min Max
lnUnderpricing 621 0.241 0.395 -2.611 2.089 lnMarketStatus 621 2.265 0.654 0 5.000 lnSize 621 15.792 1.396 7.358 21.416 lnAge 621 6.593 1.522 1.327 10.263 HighTech 621 0.041 0.196 0 1 OAO 621 0.264 0.441 0 1 AuditorReputation 621 0.677 0.619 0 1 UnderwriterReputation 621 2.545 0.736 1 3 PrimaryShares 621 0.944 0.187 0 1
