Market Performance of Sponsor-Backed Initial Public Offerings: An Empirical Analysis of the German Market

This paper investigates the overall market performance of Initial Public Offerings (IPOs) in Germany, by analyzing the short and long run performance of IPOs, utilizing the data from 2000-2013. Furthermore the study aims to distinguish and compare the performance of sponsor backed IPOs to non-sponsor backed IPOs, by placing a special focus on the value creating abilities of financial sponsors. The examined data set consists of 286 IPOs out of which 46 can be considered as IPOs which were backed by financial sponsor. The study suggests that, on average, IPOs significantly underperform their benchmarks. Furthermore, the evidence implies significant differences across the IPO groups with regard to performance and operational indicators. The multivariate regression shows that in the long run, private equity firms outperform their counterparts, signified by greater buy-and hold abnormal returns respectively recorded within the three-year period after the IPO.

As soon as the fund closes, its investment phase is inducted, which lasts for about five years, considering a classical fund lifetime of about ten years. During this essential time frame, the private equity company invests the capital of the fund into various (Note 4) companies that comply with its investment strategy. The takeover (majority stake acquisition or full ownership) of the respective investment targets is typically financed by a huge proportion of debt, usually 60-90%, and is considered as a Leveraged Buyout (LBO). Whereas the smaller equity portion is derived from the fund, debt investors, such as banks, finance the debt portion.
The last phase of a private equity transaction is the exit phase, which basically determines the success of all the previous phases and the entire private equity process. Due to the fact that private equity investors require a high return on their equity (usually targeting an IRR (Note 5) between 20-30%), the exit opportunities of the potential investment are already evaluated during the investment phase by the assumption of future exit valuation. The most common three exit routes for private equity firms, which have already been mentioned briefly, are the sale to a strategic buyer (trade sale), a financial buyer (secondary buyout) or the exit via an IPO.
In a trade sale, the private equity firm sells all its complete shares to a strategic buyer, which is usually a company that operates in the same industry as the target company. As the strategic buyer is often able to realize synergies through the purchase, the exit valuation and the related sale are often priced the highest among the three exit options (HVCA, 2016).
In a secondary buyout, the held portfolio company is sold to another private equity firm. The reason for the acquisition interest in the target is that the buy-side private equity firm still sees upside potential for eventual improvements or other value enhancement measures. In both stated exit routes, the ownership of the private equity firm is typically fully liquidated.
The exit route via an Initial Public Offering (IPO), which represents the focus of the analysis and will be discussed in detail within the next section, takes the company public (Note 6). The private equity firm hereby liquidates its shares by selling them on the stock market. Lock-up periods, which prohibit the sale of the entire shares immediately after going public, together with other regulative guidelines and strategic considerations, ensure that the private equity firm still maintains a significant share of the target after the exit (Levis, 2008). This limited flexibility together with the fact that IPOs are time consuming and involve huge transaction costs are among others reasons why trade sales or secondary buyouts are typically more preferred exit routes than an IPO.

Venture Capital
Venture capital financing has rapidly grown within the last two decades due to the fact that it represents the specifically targeted financing instrument enabling to close the funding gap for young entrepreneurial firms.
Venture capital investments can be defined as "professionally managed and independent pools of capital, which focus on the dedicated provision of equity-or equity-linked investments in privately held, most cases accumulated within a VC-fund and derives from several sources such as wealthy and high net worth individuals, institutions or companies that possess a venture capital arm.
The involved risk, in any of the described venture capital investment types, is generally higher compared to private equity investments due to higher volatility of the success of young start-up firms.
Consequently, expected returns are higher than in private equity investments, approximately amounting to 20%.
After the acquisition of respective promising, high-growth entities, the venture capitalists often take a very active role in the start-up firms (managing phase), not only supporting the management in strategic decisions, but even actively deciding which future positioning and strategy the underlying portfolio company will pursue. Therefore VCs aim to add value to portfolio firms by close monitoring, managerial-and strategic advice, aiming to accelerate the growth trajectory of their selected entrepreneurial companies.
The average holding period of an investment into a start-up company by a venture capitalist ranges from about five to ten years. The returns for VC-investors are generally capitalized when the venture capital firms exits its investment entity (Spierings, 2011). The two prevalent exit routes of venture capitalists are either a trade sale or an IPO. In a trade sale, a larger company acquirers the underlying portfolio company of the VC either by cash or shares. Usually, only the most successful companies are being taken public via an IPO. Due to lock-up agreements the VCs are however unable, and often also unwilling (e.g., exit acts as a negative signal) to directly exit the company just after the IPO (Berger & Udell, 1998). The VC investments however, that pursued an IPO as its exit route, will represent the focus within the underlying study.
However, it can be generally said that the time it takes to develop a venture usually exceeds the typical investment horizon of a venture capital firm, so that venture capitalists usually specialize in certain investment types and stages, leading to high variations in the risk-return relationship amongst VCs.

IPOs-Performance Investigation
The performance analysis of initial public offerings is basis to a huge amount of research that has been conducted so far since the 1960s. Two dominant phenomena have been observed empirically that are extensively discussed in literature: short-run under-pricing and long-run underperformance. The following section will list and discuss important theories that are related to these two phenomena.

Short Run Performance of IPOs
The most commonly observed performance phenomena, considering the time frame immediately after an IPO, is called under-pricing, in which an IPO is priced below its market value. This is the fact, when the offer price of the stock is below its closing price on the first trading date. This under-pricing shows high variations in distinctive markets as well as in the magnitude, but can be broadly estimated to be in the range of 10-20% on average. This under-pricing implies that the potential maximum amount of equity that could have been raised by firms in an IPO is, on average, significantly subdued (Ibbotson, The key parties involved in an IPO transaction are the issuing firm (Note 7), the underwriter, which is usually an investment bank that is marketing and underwriting the deal and investors (Ljungqvist, 2004). Most of the well established theories and models that explain under-pricing are based on asymmetric information, assuming that one of these parties has superior information over the others. a) Under-Pricing-Based on Asymmetric Information The "winners curse" theory relies on the adverse selection problem, which assumes that there are two types of investors with different information about the fair value of the shares: uninformed and well-informed investors. Whereas the uninformed investors also bid for overpriced shares on the capital market, well-informed investors merely bid for under-priced IPOs. Therefore under-pricing is not only a tool to keep uniformed investors in the market by compensating for their losses, but also increases the attractiveness of the respective shares for informed investors (Rock, 1986).
The theory established by Baron (1982) is based on a principal agent problem, which assumes that the underwriter has a better informational access to demand conditions on the market than the issuing firm.
Therefore the underwriter uses under-pricing in order to encourage the maximum selling efforts, which obviously creates agency costs (Bessler & Thies, 2008).
Another theory however, which is based on the signalling effect, reveals that the issuing firms have superior information over the firms intrinsic value so that they are able to use the under-pricing mechanism to signal their worthiness. Therefore valuable and well-performing firms can reduce the information asymmetry for investors by signalling their true value via under-pricing (Welch, 1989). Benveniste and Spindt (1989) established a further theory that argues that well-informed investors, which truthfully disclose their estimation of the intrinsic stock value, when the IPO is marketed, are compensated by under-pricing as they help underwriters to reduce the potential pricing inefficiencies (Ljungqvist, 2004).

b) Under-Pricing-Based on Firm-Specific Characteristics
With regard to firm-specific characteristics, research assumes that company size and under-pricing are inversely related. This originates from the fact that larger companies have a higher reputation on capital markets that demand a lower under-pricing magnitude compared to smaller companies (Ritter, 1984).
Furthermore Clarkson and Merkley (1994) identified that the offering size of an IPO is also inversely related to the magnitude of under-pricing. Investors apply higher speculative efforts among smaller offer sizes, implying a greater risk for them, in turn, leading to a greater compensation need by stock under-pricing (Lammi, 2016).

c) Under-Pricing-Based on Market-Specific Characteristics
Considering the empirical evidence that the markets provide us with, it can be said that the magnitude of under-pricing is related to variations in the business cycle (Loughran & Ritter, 2004). While average under-pricing reached its peak during the internet bubble in 1997-2000, with under-pricing levels of approximately 65%, it has declined to moderate levels quite after the bubble period (12% on average). underwriting companies with a significant underwriting reputation. High-level management of the issuing companies consequently traded a part of the under-priced shares internally in personalized accounts (Lammi, 2016).

Long Run Performance of IPOs
In addition to the studied short-run performance based on the phenomenon of under-pricing, extensive research has also been conducted with regard to the long-run performance of IPOs.
According to Ritter (1991), who has studied the long-run performance of IPOs in the US, IPOs significantly underperform (Note 8) an appropriate benchmark (e.g., S&P 500) over a studied three-year period. The IPO performance was exceptionally poor during bull market conditions, in which IPO volumes were significantly high and equity valuations were at its peak. This hypothesis is in close alignment with the "windows of opportunity" within the market-timing theories, in which companies seem to exploit the high equity valuation levels of investors on the capital market.
Other studies by Schultz (2003) or Brav and Gompers (1997), however reveal that the underperformance is actually caused by calculations without value-weighted matching. Therefore they identified, when matching respective benchmarks with underlying book-to-market ratios and size of the observed companies, by value-weighting the IPO returns, underperformance is significantly diminished (Lammi, 2016).
Despite the challenging research findings and hypotheses of Brav andGompers (1997) andLerner (2003), the predominant view through the identified pattern is, that in the long run, IPOs underperform.
One reason might be that investors are too optimistic about the long-run performance of firms and adjust their actions over time by being more realistic (Bessler & Thies, 2008). Another possibility is that the pricing of IPOs is fair in the primary market but overpriced instead from the first trading day on the secondary market. According to this hypothesis, by using the offer price as the starting point, IPOs are correctly valued in the long run but overpriced instead in the secondary market. Therefore the whole theory of short-run under-pricing would be twisted towards a short-term overpricing phenomenon (Bessler & Thies, 2008).
Conclusively it can be suggested that information asymmetries as well as agency problems, which have been examined in the short run, also exist in the long run and thus embody a prevailing factor to explain the underperformance of IPOs in the long run.

Short-Run Performance of Sponsor-Backed IPOs
When sponsor-backed (Note 9) companies decide to go public, the involvement of a sponsor within the issuing firm should be seen as a positive sign that provides the firm with a reliable certification of quality when going public (Bessler & Seim, 2012). Consequently, this may results in a decreased magnitude of under-pricing for sponsor-backed IPOs compared to their non-backed counterparts (Megginson & Weiss, 1991). any under-pricing differences between sponsor-backed and non-backed IPOs. The researcher Wahal and Lee (2004) even argue that although VCs exit from their portfolio firms in rather favourable market conditions, they do not time their exit optimally, as their willingness of liquidating ownership outweighs maximization of their equity stake.
Apart from the fact that the majority of literature ascertains that sponsor-backed IPOs show a lower under-pricing magnitude, it should be noted that most of the research that has been conducted is based on the US market, justifying the underlying investigation of the German IPO market, as a logical next step in order to enrich the research field within the European market.

Long Run Performance of Sponsor-Backed IPOs
One of the most famous studies that compare the long run performances of sponsor-backed and non-sponsor backed IPOs were established by Brav and Gompers (1997). The results of the research of the US market imply that sponsor backed IPOs perform superior to their non-sponsored counterparts, considering equal weighted returns (Note 10). They argue that the reason for the outperformance originates from the managerial competencies of private equity firms such as a close monitoring, an optimized capital structure and efficient operational performance. Furthermore they identified that as soon as the private equity or venture capital firm exits the company, the magnitude of underperformance directly rises. This again illustrates that the sponsors possess extensive insights into the company and its respective valuation, exiting the company when stock prices are deemed to be overvalued (Bessler & Kurth, 2003).
Another theory suggests that renowned private equity firms diligently select and filter their portfolio companies according to strict quality criteria in the first place (Krishnan et al., 2009). Consequently private equity backed IPOs in the US show higher performance levels than their non-backed counterparts.
Considering the European research landscape, Frederikslust and van der Geest (2001) have identified that on the Amsterdam Stock Exchange, private equity backed IPOs did not necessarily underperform the set benchmarks in the long-run, whereas non-backed IPOs showed significant levels of underperformance. They trace this performance difference back to the fact that private equity firms acquire their portfolio companies only if their target returns are reached. As these returns are based on superior company performance during the holding period, investors judge it as a quality signal that leads to a high demand of the stock on the capital markets (Lammi, 2016).
Another comprehensive and recent study over the European market performance of private equity backed IPOs has been established by Levis (2009). It identified that on the London Stock Exchange, IPOs with private equity background show superior long run aftermarket-and operating performances.
He signifies that the degrees of ownership by financial sponsors as well as the leverage level are two levers that are closely and positively correlated to the performance of the stock (Levis, 2011). degree of controversy about the performance levels of the respective IPO-forms and high variations among national equity market conditions

The German Equity Market-Setting for IPOs and Private Equity
The characteristics and regulation of the German Stock Exchange set an important framework for firms going public. The legal-, banking-and corporate governance system and structure are quite distinctive to that of the United States, assuming also differences in the performance results of IPOs (Bessler & Kurth, 2003). Whereas in the US, IPOs of growth-and high-tech firms were a regular phenomenon, companies that went public in Germany, used to be rather mature and well-established companies, in the past (Bessler & Thies, 2008).
A further important characteristic to consider is the existence of a lock-up period of at least six months after going public, in which investors or shareholders are not allowed to sell shares. This period is not only limiting the flexibility of financial sponsors, but also makes the time window around the expiration date of the lock-up period of specific interest, especially with regard to the actions by the private equity firms. A possible exit or an extension of the holding period may indicate a signal for outside investors about the attractiveness of the stock.
Not only by considering the diverging regulative frameworks for IPOs, but also by looking at the significant volume differences of buyout-backed exits of private equity investors (US: $182bn; Europe: $107bn), it can be said that results may be deviating from research based on the IPO market of the United States.

Hypotheses
The previously discussed literature regarding the performance of IPOs suggests that especially in Europe, research results show huge variations in IPO performances. As a clear performance pattern can consequently not be identified, the aim of the paper is to close this research gap to a certain degree. As Germany can be identified as one of the most important and biggest IPO markets in the European research sphere, the performance of IPOs on the German market will be investigated in detail for the time span of the year 2000-2013.

Hypothesis 1: German IPOs Underperform Their Benchmarks (Note 11) in the Long Run
The extensively discussed literature advocates that IPOs, on average, used to underperform the equity markets and their non-issuing counterparts. It is argued that the favourable conditions for companies to go public (windows of opportunity) imply overvaluations on the equity markets, leading to a significant under-pricing for the first-trading day. It is suggested however, that equity markets absorb this under-pricing in the long run by mean-reverting performances, which hypothesizes that, in the long run, IPOs underperform the equity markets in Germany. How the expected negative abnormal returns are calculated, will be explained in the next section.

Run
The arguments for an over-performance of sponsor-backed IPOs compared to their non-backed counterparts have been raised by various researches, which have especially been conducted in the US market. It is argued that the strong managerial expertise paired with financial power and strategic excellence of private equity firms, enables superior performance, indicated by higher returns, compared to companies that go public without the backing of a financial sponsor.
The financial sponsors will be divided into subsamples consisting of private equity and venture capital in order to establish a higher degree of detail in the investigation. Due to the fact that venture capitalists invest into companies with a higher risk profile and a lower maturity, it is expected that they underperform the IPOs backed by private equity investors, ceteris paribus.

Data-Country Selection & Time Frame
The overall focus of the market performance analysis is the measurement of the long-run performance of sponsor-backed IPOs in comparison to their selected benchmark as well as to their non-backed counterparts. As already stated, this analysis will be conducted for the German IPO market, which has been selected as the focus country for the forthcoming performance analysis. It is assumed that Germany, despite the fragmentation of the European market, illustrates a representative sample for the European market, which is sufficiently extensive to draw reliable conclusions.
As long-run abnormal return analyses are conducted by looking at the performances including 36-months after the initial public offering, the last available data that can be used in order to properly analyse the long-run returns are derived from the year 2013. Therefore the time frame, from which IPO data has been selected for the analysis ranges from the year 2000-2013, thereby representing the latest available data set for IPOs in Germany.
Consequently, the data set used for the analysis, consists of 324 IPOs on the German market between the year 2000-2013, after the exclusion (Note 12) of re-listings, market transfers, property funds and investment trusts.

Data-Sources
The major sources for the IPO data (e.g., stock-prices), data for the index benchmarks as well as data that enable the classification of the type of sponsor an IPO is backed with (none, private equity or Zephyr, which is a platform that offers comprehensive M&A data with integrated and detailed company information, has been used to define ownership of the IPOs. Accordingly, From the 286 identified IPOs, 46 IPOs have been identified as sponsor backed (PE = 21 companies; VC = 25 companies).
Nevertheless it has to be mentioned that the establishment of a strict separation between venture capital-and private equity investments is often not clear because the respective firms partly operate in both investment markets.

Methodology-Calculation of Returns
The two most important aspects in identifying whether stocks outperform or underperform is on one hand to identify the appropriate method to calculate the abnormal returns, and on the other hand to compare the respective results to a suitable benchmark for the distinctive companies that went public (Bessler & Thies, 2008). Accordingly, empirical findings can frequently be explained by either of the two levers.

Calculation of Long Run Abnormal Returns
Whereas in early research papers, the Cumulative Abnormal Return method (CAR) used to be the prevailing technique to calculate abnormal returns, more recent papers increasingly make use of the Buy-and-Hold Abnormal Return method (BHAR). It is argued that the BHAR method is the most appropriate and accurate method to measure long-run performances as it depicts the experience of investors of holding stocks for an extended period (Conrad & Gautam, 1993).
BHAR are measured by compounding capital gains of a specific investment vehicle for a given period of time (Johansson, 2011). The way buy-and-hold abnormal returns are computed is demonstrated in the following formula: In the BHAR equation, r it is the raw return of an IPO-company (i) and r bt represents the raw return of the benchmark (b) at month (t). N denotes the number of companies.
These BHARs will be investigated for the entire sample in order to identify the IPO performance against the benchmark as well as the performance differences between sponsor-backed IPOs and their non-sponsor backed counterparts.

Calculation of the Skewness-Adjusted T-Test
As the BHAR method is subject to a compounding effect, which may lead to positively skewed returns in the long-run, Johnson (1978) and Lyon (1999), adjusted the t-test towards a skewness-adjusted t-test.
This adaptive model eradicates the bias, when using the BHAR-method and is calculated in the following way: AR t and σ(AR t ) represent the sample mean and the samples standard deviation of abnormal returns.
Whereas N denotes the number of companies within the sample, y signifies an estimate of the skewness coefficient (Lyon et al, 1999).

Selection of Suitable Benchmarks
Apart from the selection of the appropriate method to measure abnormal returns (BHAR-method), the selection of suitable benchmarks is highly important as results change quite sensitively to the choice of respective benchmarks (Loughran & Ritter, 1995). In order to provide a very detailed analysis, exonerated from biases, various benchmarks are applied, with the aim of giving a comprehensive foundation of the right argumentation of the results.
The first benchmark that will be applied is the MSCI Germany index (Note 13), which is designed to measure the large and mid cap segments of the German IPO market, covering about 85% of the equity universe in Germany. This country-specific index aims to give an overall insight into the sample of identified IPOs.
The second benchmark that will be utilized adjusts the IPOs according to their size, with regard to market value on their offering date. In Germany the respective size adjusted benchmarks are the SDAX (small-cap), MDAX (mid-cap) and the DAX30 (large-cap). Any IPO of the sample will be categorized into one of the three indexes, according to their market value.
The third benchmark that will be applied classifies the IPOs according to their industry (FTSE industry classification (Note 14)), which in turn will be matched to their respective MSCI industry sector indexes.

Variables
The variables used in this section not only aim to define important characteristics of respective IPOs but also intend to analyse the performance differences across the sample. Apart form measuring abnormal returns via the BHAR method, a multiple regression will be run that uses dummy variables in order to compare the long-run IPO performances within the sample. In order to be able to analyse the data via a regression in section 5 (5.3), the initial dependent-, control-and dummy variables have to be defined. The selection is in close alignment of the quite recent studies of Levis (2009)  In order to interpret the results of the stock returns, a Wealth Relative (WR) is used to evaluate the performance of respective stock against its benchmark. This WR is equally weighted over the 36-month period and represents the dependent variable of the regression analysis (Levis, 2009). As a benchmark the MSCI Germany index has been selected.
Again, the natural logarithm of the WR was selected in order to decrease the potential bias by outlying performances.

Total Assets (Firm Size)
In corporate finance, total assets measure total resources of a company and are often used as a proxy for the size of a company. The total assets of companies are measured at the moment they go public. As already discussed, previous literature suggests that the size of a firm is positively correlated to superior long-run performances. It is argued that the greater the firm, the more accurately the initial valuations (offer price) is considered as fair, due to lower informational asymmetries. In order to mitigate the effect of outliers, again, the natural logarithm of total assets is used.

Boom-Bust Period (Business Cycle)
Due to the fact that the sample extends over a period in which different phases of the business cycles were respectively present, the investigation has to take into account the historically prevailing conditions of the market. Hence, he boom-bust period represents a dummy variable that is (=1), if the firm went public during a boom period, and will equal (=0) if the IPO took place during moderate or "bust" periods. By analogy of the mentioned renowned study of Levis (2009)

Price-to-Book Ratio (Firm Type)
The price-to-book ratio is a variable that compares the stock's market value of a company to its book value when the firm goes public. The ratio therefore expresses whether a companies stock is over-or undervalued. Whereas high price-to book ratios can be categorized as growth stocks, low price to book values indicate a value stock. Literature suggests that value stocks outperform growth stocks so that it 4.4.6 Asset Turnover (Operational Efficiency) Asset turnover is a widely used ratio that represents the value of a firms' sales generated in relation to its asset value. Therefore it is often used as an indicator of operational efficiency with that a firm is deploying its assets in generating revenue: Asset Turnover Ratio= Net Sales Total Assets It is predicted that the asset turnover ratio is positively correlated to the long-run performance of IPOs as firms that superiorly deploy their assets are expected to perform better in the overall market.

Margin (EBITDA)
The most widely used indicator of the profitability of a company is the EBITDA (Note 15)-margin as it very suitable for performance comparisons due to the fact that it is independent of capital structure.
Margins are always calculated by dividing them by net sales: Similarly to the asset turnover ratio, the EBITDA-margin variable is expected to have a positive correlation towards long-run IPO performance, as companies with a higher profitability are expected to perform simultaneously better.

Debt-to-Assets Ratio (Leverage Ratio)
The leverage ratio is a variable that represents with how much debt (in %) a company is financed. As a high leverage is a key aspect within an LBO valuation, the variable should be definitely considered in the forthcoming valuation. According to literature however, views about the effect of leverage on long-run IPO performance are obscure. Whereas some researchers (e.g., Levis) argue that a higher leverage as a rule increases the stocks earnings per share and is therefore positively correlated to IPO performance, other researchers investigated opposite findings.

Type of Ownership
The overall sample and the investigation of data are conducted by the separation of IPOs into three

Methodology-Multivariate Regression
Apart from the investigation of IPO performance against a certain benchmark by applying the BHAR method, the investigation targets the evaluation and analysis of performance difference amongst IPO groups and the reasoning behind that. Therefore the previously stated variables have been identified which may all have a certain influence on the initial valuation on the companies that went public on the stock market which in turn may influence the IPO performances in the long run.
Therefore multivariate regressions for the entire sample will be deployed that target to identify the relationship between distinctive variables and according long-run performance. In total two regressions will be conducted, whereas the respective regressions are split into two slightly different equations, thereby applying the same logic and methodology as in the study of Levis (2009 Whereas regression 1 SB analyses the aftermarket performance of sponsor-backed IPOs, in general, regression 1 PE&VC analyses the performances even more filigreed as it distinguishes among the type of financial sponsoring, thereby subdividing the performances into private equity and venture capital sponsoring.
The second regression aims to integrate operational performance metrics into the performance analysis.
Therefore the three identified variables with operational characteristics (asset turnover, EBITDA-margin, leverage ratio) will be added to the previous equations. Hence, the investigation target is to identify whether the operational efficiency, capital structure and economic performance has an influence on the long-run performance of IPOs within the German market: Once again, the second regression consists of two equations. Regression 2 PE&VC subdivides the general term of sponsor-backing into private equity and venture capital sponsoring in order to investigate performance patterns across different IPO groups.   Overall it can be said that this depiction is already in alignment with the theory of "Market-Timing Theories".

Industry Classification of IPOs in Germany
Apart from a division and classification of the sample in total numbers over the investigated period (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011), the sample can be classified into the industries defined by the FTSE industry classes in Table 2: By analysing the industry distributions within the distinctive financial sponsoring alternatives, it can be ascertained that IPOs with private equity backing especially prevail in the consumer goods-and industrial sector (28.6% respectively). This pattern is a sign for firms, which are usually considered as rather mature, and foreshadows a potential strategy of a rather short investment horizon, typical for private equity investors. IPOs with the financing background of venture capital firms on the other hand, tended to invest especially in the technology-(32%), industrial-(24%) and health care (20%) sector. Especially the health care and technology sector are typically industries that demand strong investments into research and development, which are filled by a strong venture capitalist that aims to finance these firms already in their early phases.

Firm & Operational Characteristics of IPOs in Germany
Apart from classifications according to industry and general annual IPO distributions, the data set has been examined according to the operational and firm characteristics of the respective financial IPO groups. This distinction is considered as important with regard to a detailed analysis especially targeting to distinguish among the existing IPO types. Therefore, apart from the operational variables identified in section 4.4 (total assets, asset turnover, EBITDA-margin, leverage-ratio), firm data that is capable of giving an estimate about firm size (number of employees, net sales, market cap) has been added to the analysis (Table 3).  Considering the comparison of private equity backed IPO firms and their non-backed counterparts, a similar pattern can be observed regarding net sales and total assets in which PE-backed IPOs are 22x and 10x larger than their non-backed counterparts. When comparing net sales and total assets of the venture capital-backed IPOs to their non-sponsored counterparts, the differences are significantly lower (1.6x and 1.4x correspondingly), implying again similar company sizes of the two IPO groups.
When analysing operational characteristics, it can be seen that the EBITDA-margin is highest (12.8%) for IPOs that have been backed by a private equity firm although VC-backed (9.6%) and non-backed IPOs (11.1%) just slightly underperform in this category. Furthermore it can be seen that companies that went public, backed by a private equity firm show a higher operational efficiency in terms of asset turnover (1.2) compared to VC-backed (0.6) and non-backed (0.5) IPOs.
The leverage ratio on the other hand shows significant differences amongst the respective IPO groups.
Whereas companies that went public, financed by private equity firms showed a median leverage ratio of 25.3%, non-backed IPOs and VC-backed IPOs showed drastically lower levels of about 5% and 1% respectively. This unexpected result implies that companies that went public, financed by venture capital firms, have nearly no debt. The reason behind that might be that the investments of venture capitalists that generally have a high-risk profile are too risky for the debt capital markets, so that banks are rather reluctant to finance these IPOs in the past, so that they solely rely on equity.
By looking at the price-to-book ratio, it can be identified that PE-backed IPOs show the lowest ratio (2.8) compared to the VC-backed (3.6) and non-backed (3.1) counterparts. This rather unexpected result might be explained by the fact that all the IPOs that were conducted during the technology bubble in 2000 have not been backed by a financial sponsor. Therefore the overvaluation during that time leads to a positively skewed data set of non-backed IPOs.
Conclusively it can be said that PE-backed IPOs are distinctively larger in terms of size compared to their VC-backed and non-backed counterparts. Furthermore, IPOs that went public with the sponsoring of a private equity firm tend to perform, on average, operationally more efficient than the two other IPO groups and have a higher leverage ratio, whereas VC-backed and non-backed IPOs show rather comparable results with respect to size and operational characteristics.

Sample Analysis-Short Run Performance of IPOs
As discussed in the literature review, the magnitude of IPO under-pricing, a phenomenon that was approached by various researchers in the past, depends on numerous factors such as asymmetric information as well as on firm-and market-specific characteristics (see section 2.4.1(a)-2.4.1(c)).
Similar to the positive correlation between the states (Note 18) of the economy and the number of companies that go public, the magnitude of IPO under-pricing in Germany also shows a high fluctuation over time (see Figure 4). under-pricing levels are quite correlated to the economic states, as apart from the technology bubble, high under-pricing levels can also be discovered just before the financial crisis (2004)(2005)(2006)(2007). Therefore it can be said that the identified under-pricing data is confirming the findings of the previously discussed literature, stating that under-pricing magnitude follows conditions of the economy, expecting it to increase steadily since 2011.

Figure 4. Average Annual Under-pricing of German IPOs (2000-2013)
In order to provide a more detailed view on the under-pricing levels, especially in the form of a summarized comparison between the respective IPO groups, Table 4 provides data that represents the averages and medians of under-pricing on the German market for the distinctive IPO groups. Conclusively it can be said that, ceteris paribus, in Germany the under-pricing phenomena of IPOs is existent and with respect to the market-timing theories in alignment with movements of the German economy following the business cycle, leading to higher under-pricing levels during boom and bubble periods.

Sample Analysis-Long Run Performance of IPOs
In order to enable the confirmation or rejection of Hypothesis 1-2, the long-run performance of IPOs within the German market was investigated. The time frame under investigation, aiming to identify the equally weighted as well as the value weighted buy-and-hold abnormal returns of IPOs, starts after the first trading date, and persists until 36 months month after the firm went public. The equally-and value weighted BHAR were calculated against various benchmarks. These applied benchmarks are country-(MSCI Germany), size-(DAX30, MDAX, SDAX) and industry-specific (FTSE) and set the performance of the IPOs under investigation into a context towards the respective indices.
The result is a post-launch performance analysis that aims to depict the various performance patterns of the distinctive IPO groups over the stated three-year horizon and is demonstrated by Table 5.
Considering the entire IPO sample (Sample 1), it can be seen that German IPOs, on average, heavily underperform the selected benchmarks. By equally weighting the BHARs within the three-year period, the negative returns range from -29% to -34% (36 months after the IPO). Additionally, the results from Sample 1 show a significance level (Note 19) of 1%, indicating that the results are highly significant even after the skewness adjusted t-statistics (Johnson, 1978).
Due to the fact that the majority of the data stems from non-sponsored IPOs (84%), the underperformance of returns for the non-backed sample (Sample 2) is expectedly very similar to www.scholink.org/ojs/index.php/jepf Journal of Economics and Public Finance Vol. 4, No. 1, 2018 Published by SCHOLINK INC.
sample 1, ranging from -29% to -35%. Despite the circumstance that Sample 3, which summarizes the performances of sponsor-backed IPOs, also shows negative returns ranging from -10% to -25%, the performance is considerably better compared to the non-sponsored IPOs.
Hence, comparing the equally weighted performance of sponsor-backed and non-sponsor backed IPOs, it can be said that although both samples show a negative performance, sponsor-backed IPOs perform about 4% to 25% better than their non sponsor-backed counterparts.
An interesting difference between private equity backed (Sample 4) and non sponsor-backed IPOs (Sample 2) that can be observed is the performance over the entire event window. Whereas Sample 2 shows a quite steady underperformance level over the three-year period (12 months: -24% to -27%; 24 months: -32% to -33%; 36 months: -29% to -35%), the performance of Sample 4 improves over the investigated time horizon (12 months: -12% to -17%; 24 months: -15% to -25%; 36 months: -10% to 6%). This performance improvement may indicate that the private equity firms are able to reduce informational asymmetries for investors over time that increase the returns on the stock exchanges.
Sample 4 and Sample 5 investigate and separate the performances of financial sponsoring of IPOs. In fact the performances between private equity backed and venture capital backed IPOs highly varies.
While PE-backed IPOs show return figures around zero, after 36 months, which amount to -10% to 6%, venture capital backed IPOs significantly underperform with negative return figures ranging from -38% to -22%. Therefore it can be said that the negative performance of sponsor-backed IPOs is primarily driven by venture capital backed IPOs, despite the fact that, on average, they still show lower negative returns than non-sponsor backed IPOs.
When considering value-weighted returns instead of equally weighted BHARs, the entire IPO sample (Sample 1) shows significantly better results, with a reduced magnitude of negative returns, ranging from -5% to -15%. A similar pattern holds true when looking at the value weighting of private equity-backed IPOs, which show a performance range between -3% to 10%. By looking at the venture capital sample (Sample 5) however, the value-weighted results still imply highly negative BHARs, which range from -21% to -33%. Hence it can be said that the majority of negative IPO performances originates from smaller companies, which are primarily backed by venture capital firms or no sponsor.
Therefore the theory prompted by Brav and Gompers (1997), stating that value-weighting leads to more consistent data, in which large companies dominate, seem to be visible in the underlying data set.
Finally it can be said, that the samples mirror the tenor of previous literature, an underperformance of IPOs compared to the respective benchmark indices. Therefore Hypothesis 1 can be confirmed by the underlying data set. Moreover, the sample indicates that especially the private equity backed IPOs seem to outperform the non sponsor-backed sample although this may not hold true for the venture capital backed IPOs. However as sponsored IPOs overall slightly outperform non-sponsor backed IPOs, it can be ascertained that Hypothesis 2, exactly assuming this pattern, can be confirmed as well.
Published by SCHOLINK INC.
generally tend to cluster around bull market conditions. Furthermore it is assumed that companies even postpone IPOs in order to issue their stocks on the equity market during favourable market conditions during the favourable market conditions (Lucas & McDonald, 1990).  This stated phenomenon might be explained by the study of Weild and Kim (2010), which suggests that financial sponsors increasingly decide to exit their most promising portfolio enterprises via a trade sale instead of an IPO because of changes within the regulatory environment, leading to a so-called "IPO crisis". The financial sponsors see a higher chance of a successful exit through a trade sale, so that the general behaviour of private equity and venture capital firms have shifted away from choosing IPOs as a exit strategy to sell their stake of equity. A further reason for the diminishment of IPO activity, especially backed by a financial sponsor, might be the fact that young and dynamic growth firms have already been acquired by mature and established firms which consequently leads to a lower IPO activity (Bessler, 2012). In order to extend the observed relationship between the issuing frequency and the index value towards a performance based IPO analysis, the long-run performance of IPOs was investigated on an annual basis. Table 5 illustrates the BHARs for the 36 months period of IPOs in Germany on an annual basis, ranging from the year 2000 to 2013. The distinctive returns are calculated by the monthly compounding of returns for three years, starting for the individual IPOs after they went public. As a performance benchmark, the MSCI Germany was selected. and 2013. Therefore it can be said that on the first view, sponsor-backed and non-sponsor backed performance may seen to be equalized. When however considering the IPO groups in more detail it can be seen that, on average, PE-backed IPOs outperform their counterparts, whereas VC-backed IPOs perform similarly to non-sponsor backed IPOs or even worse.
One obvious limitation of the table is the unequal distribution of IPO numbers and IPO group inequalities, which skew the results in a first sight. Therefore it has to be ascertained that most of the IPOs were conducted during the technology bubble and most of the IPOs have a non-sponsored background, distorting the findings especially when looking at total numbers.

Sample Analysis-Multivariate Regression Analysis for IPOs with Financial Sponsoring
Background So far the underlying data analysis has identified that sponsor-backed IPOs have outperformed non-sponsor backed in the analysed three-year period after going public. This superior performance primarily originates from private equity backed IPOs, which, on average, consist of rather large companies with comparatively great performances compared to the IPOs, backed by venture capitalists.
To enable the identification of factors that cause the respective differences in performances, one methodology, discussed in section (4.5), is the establishment of a multivariate regression that targets to identify the relationships between the IPO performance and selected company characteristics, depicted by Table 6.
Whereas Regression 1 captures general IPO characteristics that were prevalent at the time of offer (initial return, logarithm of total assets, price-to-book ration, boom-bust period), Regression 2 extends the stated characteristics by operating levers (asset turnover, EBITDA-margin, leverage ratio). The two regressions are both separated into Sponsor-Backed (SB) performances and a more detailed view by separating the sponsor backing into private equity and venture capital (PE&VC).

Regression Results-Financial Sponsoring
By looking at the regression results it can be seen that Table 6 confirms the previous findings, stating that sponsor-backed IPOs perform superiorly to their non-backed counterparts. Both regressions (1SB high aftermarket performance (36 months).
When going into the more detailed regressions (1PE&VC & 2PE&VC), it is visible that especially the sponsoring by private equity firms has a very positive and highly significant effect on the aftermarket performance of IPOs. Therefore it can be said that the analysis is in close alignment to the previously identified tendencies, confirming a positive correlation between financial sponsoring and IPO long-run performance.

Regression Results-General Characteristics
Considering the coefficient of the initial average IPO return on the first trading date, it can be seen that the figure is negative within each equation of the respective regressions. This implies a negative correlation between the initial return and the aftermarket performance. This would consequently strengthen the research findings by Ritter (1991) and other researchers, depicting the mean reverting characteristic of IPOs: the higher the first-day under-pricing, the greater the underperformance in the long run, reverting towards the mean. The logarithm of total assets, a characteristic that aims to represent the size of an IPO, shows a positive and significant relationship towards the long run performance of IPOs. That would imply that the larger a company, the better the performance in the aftermarket. Exactly this phenomenon has been discussed in the study of Brav and Gompers (1997), which states that larger companies are less volatile than smaller companies and therefore perform better. In the underlying sample this is precisely the case with private equity backed IPOs that are on average larger firms that perform comparatively better than the rest of the overall sample.
A further coefficient under investigation in the regression is the price-to-book ratio, which is negatively correlated to long-run IPO performance, according to the sample (Table 6). Whereas high price-to-book ratios imply a growth stock, low price-to-book ratios indicate a value stock. Due to the fact that high price-to-book values already entail a high valuation based on rather risky determinants, the performance is often lower due to a higher stock volatility, leading to the discussed negative relationship.
The boom-bust period variable has also a negative coefficient with regard to the IPO long-run performance, demonstrating that within boom periods, such as the technology bubble in the year 2000 or the years 2006-2007 just before the financial crisis, were times in which stock valuations, on average, were heavily overpriced. Therefore in these "windows of opportunity", high under-pricing and the following gradual mean reverting characteristic of IPOs lead to poor levels of IPO performance, justifying the negative relationship.

Regression Results-Operational Characteristics
By looking at the operational characteristics of companies that went public, the following characteristics can be ascertained: The asset turnover ratio, which represents an approximation for the operational efficiency of firms, is positively related to the aftermarket performance of IPOs. Due to the fact that higher operational efficiency should normally lead to a superior performance, the regression result sounds reasonable and is in close alignment to economic theory.
A further operational variable, the EBITDA-margin, also has a positive effect on the performance of IPOs with a significance level at 1%, implying high reliability. As the EBITDA-margin is an indicator the data is meaningful and highly relevant.
The leverage ratio, a key feature that is especially considered by private equity firms, when evaluating the attractiveness of an investment of a potential target, pronounces the positive relationship between leverage and long-run performance of IPOs. Levis (2011) already discussed the role and positive influence on leverage within his study of European IPO performance.
In summary, it can be said that the underlying results of the multivariate regression show a very close alignment with findings of previous literature by various researcher. The data simply suggests that a financial sponsor, with regard to the German market, especially in form of a private equity investor, has a positive influence on the investigated long-run IPO performance of three years, following the data of going public. This positive influence is backed by a high statistical significance, underlining the validity of the analysed data on the IPO equity capital market in Germany.

Result Discussion
The result discussion section aims to create a linkage between the discussed findings by various  With regard to the selected country-, size-, and industry benchmarks, on average, all investigated IPOs on the German market, underperform their respective market indices in the long run, confirming Hypothesis 1 of this study. Overall, long-run underperformance of IPOs, measured by BHARs, was highest during the technology bubble, followed by the time span just before the financial crisis.


On average, IPOs that are backed by a financial sponsor perform superior to firms that go public without financial sponsoring, thus supporting and confirming Hypothesis 2. The greater performance is primarily driven by IPOs that are backed by a private equity firm, which represent in terms of size and operational performance the most distinctive and highest results. Financial sponsoring by a venture capitalist however, shows clearly lower performances, comparable to companies that went public without being backed by a financial sponsor.
 By looking at the overall developments of the past years, it may be assumed that financial sponsors, and firms in general, tend to dissociate from choosing IPOs as their exit route, inter alia, due to very high involved transaction costs and the inflexibility of financial sponsors to exit the company compared to other exit channels.

Conclusion
The underlying paper investigated the overall market performance of initial public offerings in Germany by analysing the long run performance of IPOs by utilizing the latest (Note 21) available data, ranging from 2000-2013. The analysed data set overall consists of 286 IPOs, out of which 46 can be regarded as IPOs that were backed by a financial sponsor (PE: 21; VC: 25). The aim was not only to identify how IPOs perform against their benchmark categories (country, industry, and size), but also to depict and distinguish between performance patterns for any of the examined IPO groups, with a special focus on the IPOs backed by financial sponsoring. Apart from the measuring of initial first-day returns, in order to identify whether IPOs are under-priced, and the compounding of buy-and-hold abnormal returns to analyse the 36 months aftermarket performance of IPOs, a multivariate regression has been run attempting to understand the influence factors (e.g., operational and size characteristics) that led to the respective results of the study.
Furthermore the empirical findings advocate that, in essence, German non-sponsor backed IPOs underperform their financially sponsored counterparts, measured by BHARs within the investigated three year period after the issuance of the particular stocks. Dependent on the benchmark applied, the superior performance of sponsor-backed IPOs ranges between 4% and 25%. Hence, it can also be said, that these findings also enable the confirmation of Hypothesis 2.
Another major finding of the study is the fact that companies deliberately postpone and target to issue their stocks to the public during bull market conditions, visible by soaring IPO numbers during the technology bubble (2000) and just before the financial crisis (2006)(2007). Surprisingly, it can be ascertained that under-pricing levels are among the highest in the stated "hot market" periods. Due to the mean reverting characteristic of IPO returns, the long run performances during bull market conditions are among the worst within the investigated data sample, confirming the findings in the study of Ritter (1991). Of certain interest is the fact that the under-pricing levels and related negative performances flattened in the second boom period (2006)(2007) compared to the technology bubble.
This may imply that investors and financial sponsors have learned from previous market phenomena and related failures and include this knowledge when assessing new investment opportunities.
Conclusively it can be said that in the recent years, choosing IPOs as an exit channel has lost certain attractiveness, visible by stagnating numbers despite recovered economic situation on a global scale.
With regard to financial sponsors, Weild and Kim (2010) even suggest an "IPO crisis" due to the fact that especially financial sponsors assess the attractiveness of a trade sale or secondary buyout nowadays as distinctively higher, than the exit via an initial public offering. Main reasons for that tendency are the highly involved transaction cost, a time consuming and strict regulatory requirement as well as multiple IPO performances that fell short their expectations. Although Snapchat has in fact not made any profits so far, its recent successful IPO, which was the biggest US technology IPO since the Alibaba Group Holding Inc in 2014 (Note 22), signifies that IPOs will always be a common and logical next step of a company in order to acquire "fresh" equity.
It remains to be seen how IPO numbers will develop in general, what role financial investors will play in the future with regard to IPOs, and how the under-pricing magnitudes as well as the long run performance of IPOs will evolve on a global-, European-and German level.

Limitation and Suggestions of Further Research
Despite the fact that the study definitely contributes to the vague and divergent research sphere of the European IPO market, it should be noted that this study solely analyses the performances within Germany. Although Germany is unquestionably a driving financial force in Europe, this study is unable to provide a ré sumé for the IPO performances of the entire European market. This determination can be made due to the fact that Europe is highly fragmented in terms of regulatory requirements, economic strength, size and significance of the national financial markets so that IPO performances cannot be just generalized or reconciled.
The fact that a great part of IPOs (39%) derives from the year 2000, in which the technology bubble was prevalent, might skew the results towards higher overall under-pricing and lower long run performances. Therefore future research might isolate and compare the effects of the technology bubble.
Furthermore, in order to depict potential differences and fragmentations of the European IPO market, future studies could compare and highlight the differences between two or more European countries (e.g., Germany vs. France).
Although this study depicts financial sponsoring in detail, by separating among private equity and venture capital sponsorships, a clear division of the two sponsorship types is often very difficult as the financing types are closely related and often financial sponsors operate within both types of funding.
Finally, it can be said that although this study not only confirms previous findings by various academics but also contributes to the German and European research landscape of IPOs, several global and European markets have not yet been sufficiently analysed. Especially, the role and performance determinants of IPOs backed by financial sponsors, as well as the general development of the attractiveness of IPOs as a future exit channel, could be interesting topics for further research.