The Process of Going Public in Sweden

Get Complete Project Material File(s) Now! »

Method and Data

This study focuses on numerical data to examine underpricing and differences in under-pricing differences in the Swedish IPO market, thus it uses a descriptive quantitative ap-proach. Ritter and Welch (2002) claim that the typical way academics measure IPO under-pricing, is by taking the percentage difference between the issue/offer price and the first day closing price. Underpricing of IPOs is therefore often referred to first-day return or initial return. Because this definition is common practice within the field, it is also used when calculating underpricing in this study. The mathematical formula of underpricing is as follows:
The data set is composed of all firms listing on the major Swedish Stock Exchange, Nasdaq OMX Nordic Stockholm (Stockholm Stock Exchange), between 1994 and 2011 though is-suing IPOs. Firms listing on the minor official stock exchange, Nordic Growth Market Equity, and unofficial exchanges are not included in this study due to insufficient infor-mation. The IPO data consists of secondary data supplied by Nasdaq OMX Nordic, which have been verified in several instances. The issue prices have been confirmed using the Swedish Tax Agency’s stock history, prospectuses supplied by the Swedish Financial Su-pervisory Authority and issuing firms’ webpages. First day closing prices has been collected from Nasdaq OMX Nordic’s webpage. Information on dead stocks, e.g. firms that have been acquired by other firms or gone bankrupt, has been collected from microfilms of the financial newspaper Dagens Industri; the stock price information from Dagens Industri has also been used to verify the closing prices supplied by Nasdaq. Information on the acting Lead Manager for the issues has been gathered from the issuing firms prospectuses. 185 firms issued IPOs and listed on the Stockholm Stock Exchange between 1994-2011. This study examines all those 185 firms and should therefore provide a good indication of underpricing in the Swedish IPO market as a whole. Table 2 present descriptive statistics for the data. The IPOs were on average underpriced by 11.49% but the initial returns var-ied greatly and ranged from -22.35% to 241.04%. The high return variation generates a high standard deviation of 25.75% in the Swedish IPO market, which essentially means that IPO returns are highly uncertain. The results are presented further in the result and analysis chapter.
Figure 2 further illustrates the IPO data set and ranks the IPO observations from the high-est to lowest initial return. As can be seen in the figure the highest initial return is more than twice as high as the second highest initial return and is therefore an extreme value. However, it is not excluded from the data set since this thesis aims to investigate all IPOs issued on Nasdaq OMX Stockholm between 1994-2011.

Statistical significance

A null hypothesis is set up to confirm the significance of the results. The hypothesis is test-ed using a linear regression between initial IPO returns (dependent variable) and the corre-sponding OMXS30 returns (independent variable).
The null hypothesis is rejected at α = 10%, verifying that positive initial IPO returns are statistically significant. Thus, there is a 90% probability of earning abnormal returns when investing in IPOs. A 90% confidence interval is chosen since the 95% and 99% confidence intervals are unable to prove statistical significance. The regression output is found in ap-pendix 1.

Cross-sectional study

A cross-sectional study is used to examine underpricing differences between different years, market segments, sectors and lead managers. In the first analysis, IPOs are divided in to 18 subsets depending on which year between 1994-2011 the IPO was issued 1994-2011. In the second analysis, IPOs are divided into 3 subsets depending on which market seg-ment the IPOs listed in. Before Nasdaq acquired the Stockholm Stock Exchange, shares were divided between the three different lists A, O and OTC depending on the trade vol-ume of the shares. Essentially, the largest companies were listed on the A-list, the smallest on the OTC-list and the midsized companies on the O-list. After the acquisition Nasdaq made some changes to the subdivisions of shares creating the three segments Large Cap, Mid Cap and Small Cap. Thus, shares are still divided into groups based on the size of the company. Therefore, three market segments are presented, one large segment including both the A-list and the Large-Cap issues, one medium segment including the O-list and the Mid-Cap issues and finally one small segment including the OTC-list and the Small-Cap is-sues. This is to create less confusion and to make it easier to follow the reasoning. Keep in mind though that when referring to IPOs issued in the large segment it can be IPOs listed on either the A-list or Large-Cap and the same goes for the medium segment and small segment. The third analysis divides IPOs into 10 subsets depending on the issuing firms sector classifications according to the Global Industry Classification Standard (GICS), also used by Nasdaq. The different classifications and thus subsets are Energy, Materials, Indus-trials, Consumer Discretionary, Consumer Staples, Health Care, Financials, Information Technology (IT), Telecommunication Services (Telecom) and Utilities (MSCI, 2010). In the fourth analysis IPOs are divided into 29 subsets, one for each investment bank who acted as lead manager for at least one IPO transaction. In some of the IPOs two or more banks acted as joint lead managers. These IPOs will therefore appear in each lead manager’s sub-set. Some of the banks have acquired other banks that were active in the IPO market. In most cases, the acquired banks will be gathered under the acquiring bank’s name. When re-ferring to IPOs lead by Kaupthing it includes IPOs led by the acquired banks Matteus, Nordiska and Aragon. Nordea include Aros, Nordbanken and MNB Maizels. Swedbank include Föreningsbanken and Sparbanken. In 2010 Carnegie acquired HQ but since they were such a big player in the IPO market they are presented separately. Cerdit Lyonnais, Fisher Partners, Kaupthing, Lehman Brothers and S.G. Warburg were also acquired by other banks, however, none of the acquirers have been present in the Swedish market since which is why their original names are used. Finally, the results from the different analyses are compiled to present a general picture of the underpricing in the Swedish IPO market.
In the cross sectional study initial returns are adjusted for movements in the stock market index, OMXS30. If initial returns are positive after the market movement adjustment it can be concluded that IPOs outperform the market. Since closing prices are used when calcu-lating initial returns for IPOs closing prices will also be used when calculating the index re-turn. The stock market movements are calculated as follows.

Granger causality tests

The Granger causality test determines a relationship between two time series. In contrast to correlations, which simply states how strong the relationship between two variables is, the Granger causality test is a statistical hypothesis test that concludes which variable is caused by the other, or if both are caused by each other. Hence, with a significant Granger causali-ty test one variable can be used to forecast the other variable. Granger causality tests have been used to strengthen the results and analysis by confirming or rejecting relationships be-tween different variables.
The null hypothesis cannot be rejected for any of the significance levels 90%, 95% or 99%, thus OMXS30 returns does not affect initial returns.
The null hypothesis cannot be rejected for any of the significance levels 90%, 95% or 99%, thus average annual initial return in one year does not affect the number of IPO issues the following year.

Definitions 
1 Introduction 
2 Background 
2.1 The Process of Going Public in Sweden
2.2 Underpricing of IPOs
2.2.1 Information asymmetries
2.2.2 Competition in the IPO market
2.2.3 IPO risk
2.3 Implications of the Efficient Market Hypothesis
3 Method and Data 
3.1 Statistical significance
3.2 Cross-sectional study
3.3 Granger causality tests
3.4 Criticism of sources
4 Results and Analysis 
4.1 Time trends
4.2 Initial return by segment
4.3 Initial IPO return by sector
4.4 Initial return by lead manager
4.4.1 Domestic versus U.S. banks
4.5 A hypothetical example of IPO investing
4.6 A reasonable return for the risk
5 Conclusion 
5.1 Suggestions for further research
References
GET THE COMPLETE PROJECT