Printer Friendly

Fare thee well? An analysis of buyout funds' exit strategies.

This paper analyzes exit strategies of buyout funds in portfolio companies following initial public offerings (IPOs). We use a data set of 222 buyout-backed IPOs in the United States from 1999 to 2008, including hand-collected data about each exit process, to draw a detailed roadmap of buyout investors' divestment processes. Using this data, we document the timing and aggressiveness of the exit strategies, and analyze to which degree a multitude of possible determinants influence the choice of a given exit strategy. Our results indicate that buyout funds remain invested in their portfolio companies for a substantial period of time after the IPO, and that the choice of a given exit strategy depends not only upon the characteristics of each respective portfolio company, but also on the financial success of the deal from the perspective of the buyout investor.

**********

How do buyout investors exit their portfolio companies? Generally speaking, there is very little evidence regarding the exit behavior of private equity (PE) firms. Although there is ample evidence concerning the choice of exit channel, such as trade sales versus initial public offerings (IPOs) versus liquidations (examples include Das, Jagannathan, and Sarin, 2003; Schmidt, Steffen, and Szabo, 2010; Harford and Kolasinski, 2014), very little is known about the actual way in which PE investors dispose of their shares and board seats held in portfolio companies. Although previous research has developed a solid understanding of PE investors' share sales at portfolio companies' IPOs (Barry et al., 1990; Lerner, 1994; Lin and Smith, 1998; Gompers and Lerner, 1998; Cao, 2011; Krishnan et al., 2011), the existing papers primarily focus on venture capital (VC) firms or disregard the exit process in the time after the IPO or following the end of lockup periods. As a consequence, there exists no detailed documented knowledge about when and how PE investors, especially buyout funds, exit their portfolio companies following IPOs, both in terms of share sales and board exits. There is also little documented evidence as to what the buyout fund managers' decisions for a certain exit strategy are driven by, how capital markets react to these strategies, and what determines the capital markets' reactions.

We argue that examining buyout investors' exit behavior is important due to the likely large impact of a successful portfolio company exit on value creation by the fund managers. The business model of buyout funds consists of purchasing large equity stakes in companies, and in restructuring these companies' operations and funding structures. For this business model to create value, buyout funds must exert a tremendous amount of influence on the portfolio companies in order to perform the desired restructurings and monitor these restructurings closely. This is usually done in two ways: 1) holding the majority of the portfolio companies' voting rights and 2) placing buyout managers on the companies' boards of directors. Toward the end of the investment period, the buyout fund is presented with the difficult challenge of divesting the portfolio company successfully to generate investor value through the sale price of its equity stake. The divestment process is crucial for the buyout funds' performance. Only a profitable sale of the portfolio companies' shares will realize return value for the funds' investors. The investor thereby faces two challenges: 1) choosing the optimal timing for its divestment and 2) especially in the case of an IPO, minimizing the adverse market impact the divestment has. Thus, a successful exit is an integral part of the buyout investment process.

It is the goal of this paper to shed more light on this process. Specifically, we attempt to answer three distinct questions. First, when and how do buyout firms dispose of their portfolio firms' shareholdings and leave the companies' boards? In addition, what factors (such as timing, length, and aggressiveness) influence the exit process characteristics? And third, do capital markets react to these exits and, if so, what are these reactions driven by? To perform our analyses, we use a data set of 222 buyout-backed IPO companies in the United States from 1999 to 2008. We direct our focus on the IPO exit route of buyout investments. The reason is transparency. As we will explain in more detail below, each step of the buyout funds' exit processes in listed companies has to be reported publicly by means of Securities and Exchange Commission (SEC) filings. These filings allow us to trace every single step of the exit process, both in terms of share sale transactions at and following the IPO, as well as in terms of the board exits of the buyout fund managers in their portfolio companies. Our analysis is comprised of three consecutive steps. As a first step, we construct a detailed exit "roadmap" for each buyout investment in our sample. This roadmap contains the exact date and size of each share sale transaction through a buyout investor, and the exact date of each board exit through a buyout fund manager. Next, we analyze the influence that a wide range of factors might have on the buyout funds' exit strategies. Special focus is given to microeconomic factors, such as the characteristics of buyout funds and portfolio companies, and macroeconomic factors, such as the capital market environment. Also, we analyze to what degree the different steps of the exit process influence each other. As a third and final step, we calculate capital market reactions in the form of cumulative abnormal returns (CARs) to each single exit event to determine how investors perceive the buyout funds' divestments and the cross-sectional determinants of these perceptions.

We obtain three major results. First, buyout funds appear to dispose of their ownership stakes and board seats gradually over time. Instead of selling the ownership stakes early to exit the companies shortly after the IPO, we find that buyout funds stay invested in their portfolio companies for an average of 2.8 years after the IPO. Although the funds sell some shares at the IPO itself, the majority of the sales transactions occur in the period following the post-IPO lockup period. This contradicts the general perception that buyout funds realize returns as quickly as possible by a sudden divestment of their total shareholdings at or around the expiration of the lockup period. In addition, different factors drive different periods of the divestment process. The length of the period from the initial investment to the IPO is primarily driven by portfolio company-specific factors, such as portfolio company profitability and the intensity of the restructuring process. In contrast, the post-IPO divestment process is largely driven by buyout fund-specific variables, such as size, past profitability, and the financial success of the deal from the perspective of the buyout investor. Moreover, although capital market reactions to buyout funds' exits are negative, stock investors distinguish between different exit strategies and different stages of the divestment processes in their negative reactions. We find that stock investors' reactions to the buyout fund exits are mainly driven by the financial success of the deal from the buyout funds' perspective.

Overall, our results indicate that the financial success of a leveraged buyout (LBO) is one of the major determinants influencing the exit process itself, as well as the reactions to it. We contribute to the existing body of research in this field by offering a detailed roadmap of buyout funds' divestment processes following the IPOs of their portfolio companies. In determining each single step of the exit process, we also evaluate factors influencing buyout funds' exit strategies and the capital market reactions to these exits.

I. Background on PE Exits

The initial research regarding the exits of PE investors in their portfolio companies focused almost solely on the VC segment. In one of the first papers on the subject, Barry et al. (1990), analyze 433 IPOs by VC-backed companies from 1978 to 1987 and find that VC investments are maintained until well after the IPO, demonstrated by the relatively high level of shareholdings retained by VC investors after the IPO. Their study shed some light on the IPO exit route for VC investments and the way VC firms actually dispose of their holdings. Similar studies and results were later published by Lin and Smith (1998) and Brav and Gompers (2003). Lerner (1994) and Gompers (1996) also focus on the IPOs of VC-backed companies, but from a more strategic point of view. Lerner (1994) finds that VC firms' exit behavior is a highly strategic and meticulously planned part of the investment process. His results indicate that IPO exits only take place when stock market valuations are high. In times of stock market turmoil or low valuations, VC investors tend to choose secondary or trade sales as exit scenarios for their investments. In a direct follow-up paper, Gompers (1996) confirms that the exit is a strategically important part of the investment process for VC investors. Gompers and Lerner (1998) give detailed insight into the unwinding process of the shareholdings of VC firms. Instead of analyzing actual share sales at or around the IPO of portfolio companies, they focus on the distributions of VCs' holdings in their portfolio companies to the fund investors after the IPO. They document the timing of the distributions by demonstrating that the typical distribution takes place 20 months after the IPO, and that the median distribution takes place about a year after the IPO. In a recent paper on VC exit decisions, Ozmel, Robinson, and Stuart (2013) analyze the funding decisions of biotech startups. They also offer insight into the exit decisions of VC companies by arguing that VC firms make investments with the goal of going public or selling the startup to another company. In order to exit on attractive terms, it is the VC firm's goal to professionalize the startup during the period of their investment. The paper demonstrates that startups with VC investments, in general, and with investments from better networked VCs, in particular, are more likely to go public.

When compared to the detailed understanding of VC exit processes, the evidence on buyout exits is much more scarce and superficial. Kaplan (1991) analyzes the characteristics of 183 LBOs in the United States from 1979 to 1986. He finds that the LBO-backed companies remain private for about 6.7 years before going public. Ljungqvist and Richardson (2003) expand the scope of the exit analysis. Instead of focusing on IPOs only, they analyze all exit scenarios for a data set covering 3,800 PE portfolio companies from 1981 to 2001. As well as analyzing the actual investment period, they additionally analyze the determinants of the investment period. They determine the exit based on the capital returns to investors by looking at the time it takes the fund to return certain portions of the committed capital. The paper indicates that investment periods are shorter when there are more investment opportunities and that they are longer when competition for deal flow is tougher. Stromberg (2007) builds on this research by looking at buyout investment activity from 1970 to 2007. He finds that only 42% of the PE funds' investments are exited within five years. LBOs sponsored by more experienced funds are exited earlier. Only about 3% of deals are exited within the first 12 months. In a recent paper on reverse leveraged buyout (RLBO) exit behavior, Cao (2011) analyzes the market timing abilities of RLBO managers with special regard to the portfolio companies' operating performance and (IPO) market conditions from 1980 to 2006. He analyzes the lengths of the investments before the IPO and provides some information as to whether or not the buyout funds remain invested and on the boards in the portfolio companies following the IPOs. His main finding concerning the exit process is that fund managers drastically shorten the investment period if favorable market conditions allow for a lucrative exit via the IPO route.

II. Data Set

A. Data Collection

Our core data set is a sample of 222 US-based companies that publicly listed stock on the New York Stock Exchange (NYSE) or the NASDAQ through an IPO from 1999 to 2008 (referred to as "portfolio companies" for the purposes of this paper). All of these companies had at least one buyout fund as a shareholder at the time of going public. We focus on US buyouts for reasons of data availability. It is pivotal for the purposes of our analyses to obtain all of the information on the buyout funds' exits in their portfolio companies. The US regulatory standards for IPO companies offer high quality, holistic, and publicly available (and, therefore, verifiable) data on these exit processes, making this market most suitable for our analysis. We obtain the data for the IPOs from the Thomson Reuters Securities Data Corporation (SDC) data base of IPOs. We determine whether or not an IPO company had a buyout investor at the time of going public by using the SDC database's "buyout investor"-flag. This is cross-checked by manually analyzing the IPO companies' shareholder structures. To do so, we make use of the IPO firms' offer prospectuses (S-1 and 424-B) filed with the SEC. The shareholdings and board seats of buyout investors before and at the time of the IPO are determined through the S-1 and 424-B stock offer prospectus as well. To obtain the dates and magnitudes of all post-IPO share sales, we make use of various mandatory post-IPO filings in which public companies are required to disclose their ownership structure to the SEC. These filings are publicly available through the SEC's EDGAR website. Specifically, we use S-4 filings for mergers and acquisition transactions, 2) SC 13G/D/A filings for ownership structure updates of 'beneficial' and 5% shareholders, 3) DEF14A proxy statements for annual ownership structure statements, and 4) S-1 and S-3 statements for pre-sale share registrations. (1) All of these statements contain shareholdings either at a fixed point in time (DEF 14A statements) or changes in ownership when they occur (S-1/3/4 and SC 13G/D/A filings). For the exact dates and magnitudes of insider share sales we use SEC Forms 4. Board exits are determined through information in the proxy statements. If buyout fund managers exit the portfolio company's board, we determine the exact dates (and magnitudes, in case several directors leave the board contemporaneously) of the exits using 8-K SEC filings and proxy statements. We collect all exit event data from 1999 to 2012. (2) This means that the shortest exit period covered in the paper is four years for the IPOs in 2008. To obtain a homogeneous data set, we exclude all portfolio companies that were not fully exited at the end of our observation period. (3) Additional data, such as the lockup periods, are obtained through the S-1 and 424-B filings. The portfolio companies' balance sheet and profit & loss (P&L) data is obtained through the S-1 and 424-B filings and Compustat. For all stock data we use the Center for Research in Security Prices (CRSP). Four further control variables used in the regressions as described below (historic buyout firm profitability, fundraising volumes, buyout fund, and firm vintage years) are taken from Preqin.

It should be noted that even though we focus our analysis on US buyouts, we believe the data set to be representative for other nondomestic US buyouts as well. The discussed prior literature indicates that buyout characteristics are similar in different economic regions. Also, the US buyout market has been the largest LBO market worldwide, making our data and the results economically meaningful. (4) Additionally, even though our core data set is comprised of US domestic buyouts only, the invested funds do not stem from the United States alone, but also host a number of international investors. Therefore, the results we find in this data set should also be representative for large non-US buyouts. In addition, we focus our data set on IPO buyouts only. We acknowledge that there might be a possibility that this data set only captures the most successful buyouts as only these tend to go public (Schmidt et al., 2010). A potential consequence of this bias is that our results might only be representative for a subgroup of buyouts. However, IPO and LBO data for our observation period show that the buyouts that were exited through IPOs represent 19.7% of the total buyout market volume and 18.6% of the IPO market volume in the United States. As such, our data set captures a substantial fraction of the US buyout market. (5) We believe this to be further indication that our results are economically meaningful.

B. Summary Statistics

The summary statistics for our final data set are presented in Table I. The average in-sample deal is syndicated with 1.4 funds as investors. At the time of the portfolio firms' IPOs, the invested funds have an average age of 4.5 years (in relation to the fund vintage year), while the managing buyout firms have an average age of 16.2 years. Across all funds and since the firms' inceptions, the firms have raised an average of $6.4 billion in investor capital. The portfolio companies hold about $790 million in total assets, generating about $726 million in revenue at the time of the IPO with a net income of $11 million. The companies' leverage, calculated as debt over equity, is 1.5, on average. Management holds about 10% of the outstanding shares prior to the IPO, a stake that is reduced to 7.4% after the IPO. Finally, Panel B of Table I provides an overview of the industries of the sample portfolio companies.

III. Methodology

A. Determining the Exit Strategies

In the first step of our analysis, we begin by determining the characteristics of the full LBO investment lifecycle by measuring the overall period it takes the buyout funds to sell all shares and give up all board seats. As illustrated in Figure 1, we model the exit process on two different levels: 1) the actual period in which the fund investors lead the portfolio company to the IPO, and 2) the period following the IPO. The latter period can be further separated into the pre-and post-lockup period. For purposes of this paper, we label the pre-IPO period the "investment period" and the post-IPO period the "divestment period." The investment period is calculated as the time in years from the initial investment until the IPO date. The divestment period is calculated as the time in years from the IPO until the final recorded share sale of the buyout fund in the respective portfolio company. We calculate the third indicator, the duration, as a weighted average of the time between each share sale, much like the regular bond duration. We use this variable as a value-weighted time indicator to account for differences in buyout firms' exit behavior. (6) We obtain the start dates for the investment lifecycle from S-1 stock offering prospectuses (or, in case of RLBOs, from S-4/DEF 14A filings), which list the date on which the initial equity investment of the buyout fund is made in the respective portfolio company. The end points of the investment lifecycles are obtained through the post-IPO SEC filings, as described in Section II.A. We use the date of the last share sale transaction of each invested buyout fund in each portfolio company as the ending date of the investment lifecycle. The same procedure is used for the time spent on boards. The dates of the board tenures' beginnings are listed in the S-1 prospectus, while the final board exit date is taken from 8-K filings and/or proxy statements.

To obtain a deeper understanding of the strategies behind the exit process, we also determine the timing and aggressiveness of the post-IPO exit process. Both are measured using data on the actual dates and volumes of share sale transactions and board exits. The data collection process allows us to obtain the exact date, size, transaction value (for share sales: number of shares sold multiplied by the share price), and the relative value of the shares and board seats in relation to the shares and board seats held in the companies after the IPO takes place. This data reveals when an exit transaction takes place and by how much each exit reduces the shareholdings or number of board seats held at the time of the portfolio companies' IPOs. Using this data, we construct two additional indicators of the exit strategy that serve as proxies for the aggressiveness with which the buyout funds sell their shareholdings and exit the boards of their portfolio companies. The first variable we construct measures the average (mean) time between each single share sale transaction, starting with the first post-IPO share sale. By disregarding the overall holding period and focusing on the timing of the share sales relative to each other, this variable indicates whether buyout funds "pump" their shares quickly and in short frequencies into the market, or if they choose to sell the shares slowly, but constantly, leaving longer intervals between each sale transaction. We use this variable as an indicator for the overall aggressiveness of the exit strategy.

The second variable we construct is a dummy variable that indicates whether or not the buyout fund managers exit the board seats held in their portfolio companies in unexpected resignations. Generally, there are two ways in which directors can vacate their board seats in companies: 1) by not standing for re-election and letting their contracts expire at the end of the predetermined tenure of their board membership, and 2) through an unexpected resignation from the board during their tenure. While the re-elections are announced in the proxy statements sent out to the shareholders in advance of the annual meetings, unexpected resignations are announced in 8-K filings with the SEC. We make use of these unexpected board resignations to gain further insight into the active structuring of the exit process. Although the time component, as used to determine the aggressiveness of the share exits, is disregarded here, we believe that looking at unexpected board resignations is a useful proxy for the aggressiveness of a board exit. An unexpected board resignation during the board tenure is unusual and may be regarded as a clear sign for the exit of the buyout fund. An unexpected exit may be a sign of an aggressive exit strategy. To obtain a more precise understanding of the aggressiveness of unexpected board exits, we determine the actual reasons for each single unexpected board exit by collecting the relevant information from the 8-K filings or the corresponding press releases. This allows us to categorize the unexpected board exits into unexpected, but not aggressive exits (e.g., caused by the personal situation of a director) and unexpected and aggressive exits (e.g., directors leaving the board to turn their focus to newly made investments of their PE fund). The variable we construct takes a value of one (i.e., indicating an aggressive exit) for all unexpected board resignations that occur: 1) due to the director leaving because they must devote time to other investments in the buyout fund or to make new investments (i.e., generally due to investment pressure), 2) in connection with a share sale of the exiting buyout fund, and 3) all of those unexpected exits for which no reasons were mentioned indicating that the unexpected exit did not happen aggressively. The dummy takes a value of zero (i.e., indicating an unaggressive exit) for all expected board resignations and for all unexpected board resignations that occur due to: 1) personal reasons of the exiting director, 2) reasons of governance compliance, and 3) exits that did not occur due to serious disagreements between the exiting director and the remaining board members or managers of the company. By choosing this specification, we try and distinguish those unexpected board resignations that were clearly not caused by an aggressive exit strategy from those unexpected exit reasons that were, or might have been, caused by an aggressive exit strategy or were part thereof. (7)

In the final analysis of the exit process, we use all exit indicators as presented above to analyze changes in the buyout funds' exit behavior over time in accordance with varying market environments. It is possible that buyout funds adjust their exit behavior to those varying market environments. After all, the strength or weakness of stock markets at the time of a buyout's exit is crucial for the financial success of the sale of the buyout fund's shareholdings in the portfolio company. Lerner (1994) finds that this holds true for the timing of the IPOs of VC-backed firms, and Cao (2011) confirms this for RLBOs. Ljungqvist and Richardson (2003) and Cao (2011) also determine that PE investment periods vary with changing economic circumstances, especially when relating to fundraising opportunities. We try and extend this body of research by demonstrating that different market environments may also play a role in the exit strategies of buyouts after IPOs. We place our emphasis on different equity market periods, as opposed to economic growth or financial crisis periods. We measure bull and bear stock market periods using the common methodology in the existing literature. That is, stock market peaks and troughs in six-month windows, separating stock market upward and downward movements (Pagan and 1 Sossounov, 2003; Candelon, Piplack, and Straetmans, 2008; Chen, 2009). (8) We choose the respective peak and trough dates in line with the general consensus in the literature as to what the major turning points of the stock market were in our observation period (Siegel, 2008). This methodology yields five major market periods including two bull and three bear markets (as later demonstrated in Panel D of Table II). We use this segregation for the sample split and to construct dummy variables for the stock bull and bear markets that are used in the multivariate analysis.

B. Analyzing Determinants of Exit Strategies

In the next step of our analysis, we determine the possible factors that may influence the exit behavior of buyout funds. In a first step of this analysis, we are interested in the relationship between the different exit indicators. It might be the case that buyout funds structure their pre-and post-IPO exit periods jointly, or structure their post-IPO exit strategy in accordance with the outcome of the pre-IPO exit period. To detect possible joint relationships between the exit indicators, we perform two analyses. We run a correlation analysis in which we try and determine how strongly the different exit indicators are linked to each other without having to postulate the causalities between them. In addition, we control for the possible relationships among the indicators in our multivariate regression framework, as presented subsequently.

In the second step of the analysis, we run two sets of regressions to understand which factors influence the exit strategies of buyout funds. We employ a multivariate cross-sectional ordinary least square (OLS) regression model, as well as a multivariate probit regression, examining the influence that a broad variety of factors might have on buyout funds' exit strategies. We run four specifications of the OLS model and one specification of the probit model, each using different dependent variables. For the OLS model, we use the investment period, the divestment period, the duration, and the timing (each calculated as explained above) as dependent variables. For the probit model, we use the "unexpected board resignation" dummy variable. For each model, we use three separate groups of explanatory variables, as presented in the Appendix. The variables are a combination of buyout fund-, portfolio company-, and external market environment-specific variables. As demonstrated in the previous literature, PE investment and divestment decisions are driven by the type of buyout fund, especially its experience or reputation (Gompers, 1996; Lin and Smith, 1998; Stromberg, 2007) and the specifics of the portfolio companies, especially their (financial) success (Schmidt et al., 2010). Additionally, we control for corporate governance variables, such as the composition of the board or the ownership by the management. By including special dividend payments and fees, we are also able to control for the (cash) success of the investment; that is, the degree to which the buyout investors are able to extract cash from their portfolio companies before the IPO that can subsequently be distributed to the fund managers and their investors. The latter is strongly related to and extends the work of Cao (2011) who finds that the profitability of the portfolio company is pivotal for the exit timing of the buyout funds. However, we extend his analysis by looking at the actual success of the deal from the perspective of the buyout investor. Therefore, our results may shed some light on the profitability of the portfolio company and/or market environments or whether the financial success of the deal for the buyout fund is the biggest driver in the exit process.

C. Analyzing Capital Market Reactions

In the third major step of our analysis, we analyze the capital market reactions to buyout exits. To do so, we perform a two-step procedure. First, we calculate the CARs to the buyout exits and then run a multivariate regression model using the CARs as dependent variables to better understand what the possible determinants of the market reactions might be. In order to calculate the CARs, we use event-study methodology as suggested by MacKinlay (1997). We use an estimation window from 105 to 6 trading days prior to the event and calculate abnormal stock returns with the market model approach, following Brown and Warner (1980), in a three-day window around the event. (9) We perform separate CAR analyses for all post-IPO share sales and for all unexpected board exits (as defined in Section III.A). As events, we use the dates of the share sales and the dates on which the unexpected board resignations are announced. We deliberately disregard expected board exits as they are announced in the portfolio companies' proxy statements, which also contain information about the companies' financial situations, economic outlook, and additional personnel decisions. As such, we cannot clearly disentangle the market reactions to the different kinds of information given in the statements. As unexpected board resignations are announced in one-off 8-K filings, we obtain a clear date at which the markets should react to the board exits, unaffected by any other material information. To obtain additional insight into the drivers of stock market reactions, we re-run the regression model as presented in Section III.C. As dependent variables, we use CARs calculated in the three-day event window around the very first post-IPO share sale. We look at the very first post-IPO share sales as we believe (and as supported in Section IV.C) that the market should react strongest to the initial sign of the buyout funds' exit. Thus, the market reaction to the first exit should offer the most information for the purposes of our analysis. We run this analysis as in-depth knowledge regarding portfolio company and buyout fund-specific drivers of stock market reactions to buyout exits are still lacking in the existing body of literature. Based on our results presented in Section IV of the paper, we have reason to believe that capital markets should react differently to varying buyout fund investors, deal specifics, and portfolio company characteristics. We find that the exit process is a meticulously planned part of the strategy of the buyout deal, which is actively managed by the buyout funds. As the deals themselves, as well as the exit processes, are approached differently, we conjecture that capital markets may react in accordance with these differences.

IV. Results

A. Exit Events of Buyout Funds

Table II presents the investment characteristics and offers detailed insight into single share sales and board exit transactions. Panel A provides a general overview of the length of the investment lifecycle. We first review the time from the initial investment to the IPO (the investment period) and compare it to the time from the IPO to the final exit (the divestment period). We find that both of these periods are almost equal in length. The investment period lasts just over three years, while the divestment period is, on average, 2.8 years. This is an interesting finding, contradicting the general perception and assumptions of most of the prior literature, which generally believes that buyout funds quickly exit their portfolio firms following an IPO. Judging from our results, this does not seem to be the case. The length of the buyout managers' board tenures further supports this finding. On average, buyout firms remain active on their portfolio companies' boards for about 3.3 years after the IPO. Given the long post-IPO ownership periods, this is not surprising. The buyout firms retain some kind of insight into or control over their portfolio companies as long as they are actively invested in them. The duration of the divestment period is only marginally shorter at about 2.3 years, indicating that fund managers sell more of their shares toward the end of the divestment period. Panel A also indicates that each buyout fund in our sample holds close to 50% of the ownership rights in their portfolio companies prior to the IPO, along with approximately one-third of the board seats. While the shareholdings are reduced to about 33% at the IPO, the board seats remain roughly the same as prior to the IPO. In the three years following the IPO, both shareholdings and board seats are gradually reduced. Two things are of interest here. First, the divestment process is not performed in a few single steps, but gradually over time. In addition, the buyout investors still retain about 18% of the shares and 14% of the board seats three years after the IPO. Apparently, buyout investors are keen on remaining active investors in their portfolio companies for longer periods of time.

In Panel B, we examine the timing and magnitude of the actual share sales and board exits in relation to the IPO. The results in the table offer five main findings. First, buyout funds reduce their total shareholdings (as recorded on the IPO date) by 24.95% per sales transaction. In addition, there is about one-half year (0.45 years) between each sale. Moreover, we find that it takes buyout funds 1.75 years for the first share sales transaction to take place after the IPO. In this transaction, 46% of the buyout fund's total remaining ownership stake in the portfolio company is sold. We also find that the last share sale takes place 2.81 years after the IPO with 49.58% of the holdings sold in this transaction. Finally, the single highest share sale transaction takes place rather late in the divestment process, at an average of 2.36 years. Funds sell 57.78% of the remaining holdings in their portfolio company in this transaction. These findings provide more depth to and support for the outlined exit strategies of buyout funds presented in Panel A of Table II. Buyout funds are in no rush to dispose of their shareholdings following an IPO, and hold on to larger stakes of their ownership rights for longer periods of time. In this exit process, each sales transaction is used to gradually reduce the ownership in the portfolio companies. Again, a similar pattern emerges in regard to board seats. We find that once the buyout funds start exiting the board, they reduce their seats by half and hold on to their final board seats until 3.36 years after the IPO.

Panel B also contains information regarding the share sales transactions and board exits around the lockup dates. The lockup day sales are important as the unlock day is the first opportunity for a buyout fund investor to dispose of shares in the open market following an IPO. However, the results suggest that buyout funds do not make use of this opportunity. As the first post-IPO sale takes place 1.75 years after the IPO, and given that the usual lockup periods are 180 days, buyout funds tend to wait a substantial period of time after the end of the lockup period to initiate their first share sale. Our results indicate that buyout funds sell shares in only two portfolio companies at the end of the lockup period. Boards are never existed at this time. Our analysis reveals a second interesting finding regarding exits at or around the end of the lockup periods. Buyout funds sell shareholdings during the lockup period in 37 portfolio companies. This means that the IPO underwriters of 37 portfolio companies must have allowed the fund managers to sell shares despite preexisting lockup agreements. (10) Taking these results together, the exit process seems to be a very complex and strategic part of the investment lifecycle that does not end with the IPO or the end of the lockup period.

Panel C reports the results on unexpected director exits. We find that 125 managers of the buyout funds vacate their board seats in their portfolio companies expectedly, while 225 do so unexpectedly. In a very general sense, this indicates that boards are predominantly exited in an unexpected manner. On average, unexpected exits take place considerably closer to the IPO (1.86 years) as compared to expected exits (2.87 years). As a result of this quicker exit process, directors who leave unexpectedly also spend less time on the boards of the portfolio companies (5.38 years vs. 6.22 years). These numbers should suggest that the board exit strategy is in line with the overall divestment strategy of buyout funds. As soon as the exit process commences, the directors of the buyout fund also leave the portfolio companies' boards. If an exit takes place before the contractual tenure of the director ends, their contracts are terminated early.

However, to further understand board exit strategy, we also analyze the reasons for the unexpected board exits in the bottom portion of Panel C. These reasons should indicate to what degree unexpected exits are truly part of an active and aggressive exit strategy. Unfortunately, the 8-K filings and corresponding press releases do not mention specific reasons for 116 of the unexpected board exits. For the remaining exits, we find five different groups of reasons for leaving the boards unexpectedly: 1) "Personal Reasons," pertaining to either health issues or family commitments, 2) "Focus on Responsibilities with the PE Firm," including making new investments, raising funds, or joining the boards of other portfolio companies, 3) "Exit in Connection with PE Share Sales," implying that the board seat is vacated unexpectedly because the respective buyout fund sold a major stake of the shares in the portfolio company, 4) "Governance Compliance," to infer that the stock exchange asked the portfolio company to include more outside/independent directors on the board, and 5) "Unexpected Exits w/o Disagreements," which are cases in which no specific reason for the unexpected exit was mentioned. It was only stated that the exit did not occur because of a disagreement with other board members or the portfolio company itself. We determine that, primarily buyout fund managers leave their portfolio companies' boards unexpectedly in connection with the sale of shares and without disagreements. Personal reasons, the focus on other responsibilities within the PE firm, and compliance reasons are rather unimportant exit reasons. Thus, the numbers indicate that not every unexpected exit is part of an aggressive exit strategy. However, we do find that it can be common practice to terminate board contracts early in an exit process. In addition, buyout funds might be inclined to explicitly state that unexpected exits are not based on disagreements with the portfolio company in order to smooth their exit. Markets may react more favorably to unexpected exits if they know that an unexpected director exit is the result of the regular exit process as opposed to conflicts within the portfolio company. We will control for this in our subsequent analyses, especially with regard to stock market reactions to unexpected exits.

Finally, Panel D provides a descriptive indication that buyout funds adjust their exit behavior over time. Our results suggest that buyout funds exit their portfolio companies quicker and slightly more aggressively in bull markets than in bear markets. There are fewer single share sale transactions of lower magnitude in bear markets (2.36 with an average of $50.8 million in proceeds) as compared to bull markets (with 2.93 transactions and $95.4 million in proceeds). As a consequence, the time between each share sale is shorter in bull markets (0.41 years) than in bear markets (0.57 years). Shortening the exit process in bull market periods makes sense. High equity market valuations might allow for share sales at a premium to the original purchase price. Moreover, bull market periods are usually accompanied by phases of strong PE fundraising. (11) Therefore, fund managers must divest existing portfolio companies to invest newly raised funds. Things are different in equity bear markets, where fundraising is more difficult, valuable exit opportunities are scarce, and fund managers may be more inclined to wait for the markets to turn around before they sell their shares in the portfolio companies. We also find that more unexpected board exits occur in bull market periods when compared to bear market periods. Apparently, the stock market environment plays a pivotal role in buyout funds' exit decisions. This factor will be thoroughly controlled for in later analyses, as it is the first sign of an active exit structuring by buyout fund managers.

B. Determinants of Exit Strategies

In the first step of our analysis regarding the determinants of exit strategies, we run a correlation analysis of the different exit indicators, as presented in Table III. We find two major results. First, there is a high correlation between the divestment period and the divestment duration, and between the divestment period and the exit timing. This is perhaps unsurprising for two reasons. First, our descriptive results suggest that in different market environments, buyout funds adjust their overall post-IPO exit strategies. The joint movement of the post-IPO exit indicators we see in the correlation analysis lends support to this finding. The exit process appears to be a strategized and planned part of the overall investment. However, this result is also unsurprising for technical reasons. The divestment period and duration are calculated similarly (see Table AI for a description of the variable construction). Also, a shorter average time between share sales should also cause the overall divestment period to be shorter, hence the rather strong positive correlation. This also explains the correlation of 0.24 between the duration and the exit timing. (12) The variables are, therefore, technically correlated. Our second major result is that there are no significantly high correlations between the other exit indicators. We find low and insignificant correlations between the pre-IPO investment period and the main post-IPO exit indicators. This result is striking as it might be suspected that the length of the pre-IPO period should have some relationship with the subsequent post-IPO strategy of the exit. Especially since the remaining results point to the fact that the exit process is actively planned and executed. Longer periods prior to the IPO may call for a quick post-IPO unwinding to create distributions to the funds' investors, just as shorter pre-IPO periods may allow the fund to exit the companies in an unaggressive manner. To determine whether there is truly no influence on the pre-IPO period and the post-IPO exit strategy, we include the investment period as an explanatory variable in our regression model, as explained in greater detail below. Interestingly, we find that the results are confirmed in a multivariate setting. The length of the pre-IPO investment period seemingly has no influence on the post-IPO exit indicators. While other results suggest that the exit process is strategically planned, we cannot confirm that the pre-IPO period has any influence on the length of the post-IPO period or the aggressiveness of the exit. As we will explain in further detail below, other factors drive this part of the exit.

Table IV presents the results of our multivariate cross-sectional OLS and probit models. Looking first at the determinants of the investment and divestment periods, we find that the buyout fund-specific variables are much more important in explaining the post-IPO exit process than the pre-IPO investment period. The opposite is true for the portfolio company-specific variables. Apparently, the characteristics of the portfolio companies are more important in explaining how long it takes a buyout fund to take the respective portfolio company public, while the characteristics of the buyout fund and deal are more important when explaining how quickly the portfolio companies are exited following the IPO. Model I indicates that the investment period is primarily driven by the size and profitability of the portfolio company, as well as by the intensity of the restructuring process prior to the IPO, measured by the amount of merger and acquisition (M&A) transactions the portfolio company engages in. We find that the investment period is longer for smaller and more profitable portfolio companies with more M&A deals before the IPO. The latter result is no surprise since successful M&A transactions, including a potential integration phase of the target or the merging parties, are time consuming. The fact that portfolio companies with higher profitability at the IPO and greater profitability growth prior to the IPO also have longer investment periods supports this finding. A successful restructuring process of the portfolio companies takes more time, thus the longer investment period. Perhaps more experienced and historically more profitable buyout funds can shorten this restructuring process, as indicated by the negative and significant coefficients of both variables. Overall, these results reflect typical LBO investment process patterns and are, therefore, largely unsurprising.

In contrast, the divestment period seems to be driven by a very different set of determinants. First, we find that the strongest portfolio company-specific determinant for the post-IPO exit process is the leverage of the portfolio company at the IPO. Apparently, a larger leverage of the portfolio company at the IPO (i.e., already including the IPO-raised equity) induces a longer post-IPO divestment period of the buyout fund. Prolonging the divestment period in companies with more leverage in order to wait for a stronger debt reduction benefits the valuation of the companies' equity, thereby allowing the buyout fund a higher share price at the time of exit. Next to the restructuring process prior to the IPO, this might be another indicator that the exit process is aimed at value creation of the buyout fund managers for the fund investors. In addition, we find that plenty of buyout fund-specific features influence divestment period and duration. Historically more profitable, smaller, and younger buyout funds exit their portfolio companies more quickly after the IPO.

However, the most interesting revelation of Models II and III is demonstrated by a proxy variable for the financial success of the deal. An especially strong explanatory power has the dummy variable indicating whether or not 100% of the initial (equity) investment of the buyout fund is returned through dividends at the IPO. The direction of influence here is negative. Funds that have at least a 100% investment return at the IPO due to dividend payments prior to the IPO have shorter post-IPO divestment periods. This is a new and revealing result. Buyout funds returning their initial investment at the IPO are at a "break-even" point in their investment with a cash multiple of at least one. The consequence is a high and (desirably) early payoff to the fund investors, as well as the participation of the general partners in the return. As the buyout fund managers receive their profit share in the form of carried interest and the fund investors receive a return on their investment, the desire to stay invested for longer periods of time is muted. Since a dividend alone is responsible for this value creation, all proceeds generated through the remaining share sales will turn the investment into a success. As a consequence, the buyout fund does not necessarily have to remain invested in the portfolio company to create any economic value from which its share values could potentially benefit. Instead, the fund can use the freed up proceeds for further investments. We believe this to be further support for the premise that buyout exit processes are primarily driven by value creation.

The timing of the share sales, the first aggressiveness indicator of the exit tested in Model IV, also negatively depends upon the deals' proceeds. The coefficients indicate that more successful deals from the buyout fund's perspective have a shorter period between the single share sale transactions post-IPO. Buyout funds tend to sell their shares more quickly and in shorter intervals into the market if the investments already benefit the fund financially. We find additional support for this result with another variable, the fee investment return. This variable measures the amount of deal-level fees (transaction fees, advisory fees, and termination fees) the general partner receives from the portfolio company during the investment period (i.e., until the IPO), scaled by the total equity investment in the portfolio company. The model indicates that higher fees significantly shorten the exit timing. Since these fees are only paid until the IPO, there is no (financial) reason for the general partner to remain invested in the portfolio company after the IPO. Thus, a quick exit is the logical choice. (13) In the last specification, we focus on the decision to unexpectedly resign from the portfolio companies' boards. Interestingly, and unlike the other exit indicators, we do not find strong and persistent patterns of explanatory variables that influence the decision of a buyout fund manager to leave a portfolio company's board unexpectedly. The only exception is the dummy variable for a post-IPO takeover exit (as defined in the Appendix) that is expected to demonstrate a statistical influence. In the event of a takeover of the portfolio company, the directors are likely to resign from their director positions after completion of the deal. The only notable result is that we find the fee proxy variable, which exhibited a significant impact on the exit timing, influences the decision to resign unexpectedly and aggressively from portfolio companies' boards. Higher deal-level fees to the general partner trigger more unexpected board exits. This result is in line with the prior findings regarding the influence the financial success of a deal has on the exit strategy. If the general partner generates higher income from a deal through fees until the IPO, the exit is performed more quickly and aggressively.

C. Capital Market Reactions to Exits

The results of the capital market's reaction to the buyout divestment process are provided in Table V. The first and most obvious result is that capital markets react significantly and negatively to sales of the buyout investors' shares. The fact that capital markets react negatively to the exit of institutional investors in general is documented in the existing literature (Ofek and Richardson, 2003; Bozcuk and Lasfer, 2005; and, for VC-backed companies, see Bradley et al., 2001 and Field and Hanka, 2001). Our results confirm this finding and extend it to buyout investors in particular. The valuation of portfolio firms is lower without the (active) presence of buyout investors. This is no surprise. From principal agency theory we know that external institutional shareholders can solve agency conflicts or freeriding problems by monitoring the incumbent management (Jensen and Meckling, 1976; Shleifer and Vishny, 1986). In doing so, they increase managerial efficiency and create economic value. This is especially true for buyout investors, who closely monitor their portfolio companies' performance and remove existing inefficiencies by restructuring the operating business, governance structures, or strategic orientation (Kaplan and Stromberg, 2009; Acharya et al., 2013). However, we offer further insight by demonstrating that in spite of the general negative reaction to the exits, capital markets are able to distinguish between exits in different stages of the divestment process.

Specifically, there are five things to note. First, the strongest negative reaction can be found in the first exit. Second, the adverse reaction to the exits drops notably for the second share sale transaction. Third, the reaction to the last exit is almost as negative as to the very first exit. Fourth, the reactions to all of the sales in between are weakest. Fifth, the reactions differ strongly in different economic periods. CARs are more negative in bear markets than in bull markets. What do these numbers tell us? Stock investors are educated in the sense that they do not react in the same way to any kind of buyout investor share sale. This is interesting given that all single exits represent a divestment. It is clear that buyout funds must dispose of their shareholdings at some point in time following the IPO. However, since we do find that the divestment periods vary greatly, and that some buyout funds stay invested over a substantial period of time following the IPO, it is perhaps unsurprising to note that market participants react strongly once the exit process begins and once it reaches a conclusion. Interestingly, these results are not mirrored by the CAR reactions to unexpected board exits. We do not find the CARs to be economically and statistically significant in a persistent manner across models.

D. Determinants of Capital Market Reactions to Exits

In a last step of our analysis, we seek to better understand the factors that influence stock reactions to buyout exits. Note that we only include the CARs of the share sales in the regressions as board exit CARs are statistically and economically insignificant. All of the results are presented in Table VI. The strongest influence factor seems to be the financial success indicator already introduced and discussed in Table IV, the investment returned through a dividend dummy variable. This variable has a highly significant and positive influence on the CARs of the first buyout fund share sale post-IPO. The buyout funds' first signs of exit are met with less negative reactions if the buyout fund is able to return its full investment in the portfolio company at the time of the IPO through a dividend payment. We believe there are two possible explanations for this. First, the deal was a success and the portfolio company was financially able to pay out this dividend. Therefore, the dividend may be viewed as a positive signal to the market that the portfolio company is in financially sound condition after the LBO restructuring process. The fact that the buyout fund exits the portfolio company is met with muted skepticism, especially since the restructuring process is believed to be complete. In addition, stock investors may perceive the dividend as a negative sign of cash draining. Buyout funds are widely known to use the generated cash flows of their portfolio companies not only for debt paydown, but also as distributions to their investors in the form of dividends. Outside stock investors may view this critically as cash drained-portfolio companies might not be able to invest in positive NPV projects and/or pay out dividends to other investors in the future. If this were true, the exit of a buyout fund known for cash draining its portfolio company may be seen as beneficial to the company.

To determine which of these reasons are responsible for the positive effect on the CARs, we further analyze this issue by introducing three additional variables to the regression: 1) a dummy variable indicating whether the dividend was financed with debt in a recapitalization scheme, 2) a variable indicating the relative magnitude of the dividend in relation to the revenues of the portfolio company, and 3) an interaction term of the investment returned dummy and the return on assets of the portfolio company at the time of the IPO. (14) These variables allow us to determine whether investors trade their stocks on the positive signaling effect or on the negative cash draining effect. The profitability aspect should be supported by the interaction term, while the cash draining aspect should be indicated by the recap and size-effect variables. The intuition behind the latter is that a dividend financed with debt instead of existing cash generated through operations can be seen as cash draining thereby imposing an additional debt burden on the portfolio company and leaving it without liquid assets. The same holds true if the dividend has a greater magnitude in relation to the size of the company, also imposing a heavy financial burden on the company. In Models II and III of Table VI, we find that the dividend recap dummy is insignificant, as is the interaction term between the investment returned dummy and profitability. In contrast, the relative size effect of the dividend has a highly significant and positive impact on the CARs. In our opinion, these results indicate that stock investors neither trade on the positive signaling effect of a dividend, nor on the recap effect, but only on the actual size of the dividend. The larger the dividend in relation to the size of the company (as indicated by the revenue at the IPO), the more positive the CAR reaction to the buyout exits. Since we exclude other reasons for the CAR reaction, we believe that investors only favor the fact that buyout funds, which are known for paying large dividends to themselves and their investors, leave the portfolio companies.

Interestingly, two of the factors that we believed may play a decisive role in predicting CARs, monitoring and market environment proxy variables, have little or no influence on our results. This is surprising, especially since our descriptive results in Table V indicate that in a univariate setting, the CARs vary strongly in bull and bear markets. Our results therefore indicate that the reasons for negative CARs upon buyout exits are not predominantly rooted in the stock market surroundings if other factors are controlled for. Thus, these results extend the body of literature regarding the relationship between stock market environments and institutional shareholder exits, as in Chiyachantana et al. (2004), by demonstrating that stock market reactions to buyout exits differ to some degree from reactions to the exits of other nonbuyout institutional investors. The same is true regarding the presumed influence the agency cost theory should have on CARs upon the exit of buyout investors, as discussed in Section III.C. The different variables proxying for this assumption all indicate little or no influence on the CARs. Only the two monitoring proxy variables (chairman from the buyout fund and management fees over revenues) have a very slight and negative effect on the CARs. However, we believe the results are too weak to interpret them as major determinants of shareholder reaction to buyout exits. Our findings do not support the literature concerning institutional investors and agency costs, as previously discussed in Section IV.C, and provide no conclusive evidence that investors are aware of the monitoring through buyout funds and react in accordance with it. Regarding the relationship between the financial success of the deal and the CARs, the results in Table VI indicate that none of the portfolio company-specific variables have any influence on the CARs. Only the M&A deal indicator demonstrates minor and inconsistent significances. Thus, buyout funds' that engage in M&A deals to restructure their portfolio company receive a slightly more favorable CAR reaction to their exit. However, we believe that this result is too statistically and economically weak to reveal any striking information about market perceptions in this regard.

V. Robustness Tests

To validate our findings, we perform a number of robustness tests. In our first robustness test, we address a potential sample selection issue in our analyses. For the main analyses, we use a sample of 222 portfolio companies that went public from 1999 to 2008 that were fully exited by the invested buyout funds at the end of our observation period (year-end 2012). An additional sample of 57 LBO portfolio companies also went public from 1999 to 2008, but had not fully exited at the end of 2012. (Note that we can only analyze 47 of the original 57 nonfully exited companies due to missing or imprecise data). We omit these companies from our original sample to obtain a homogenous group of companies that all underwent full exit processes. In order to understand whether our results are supported by the group of nonfully exited companies, we rerun our main analyses including this additional sample. As demonstrated in Table All, the subsample of nonfully exited companies share very similar features with our main sample. Also, we find that the characteristics of the portfolio companies and of the invested buyout funds support the findings of our main analysis regarding the drivers behind the length of the divestment process. The nonfully exited companies have higher leverage and lower return on assets at IPO, lower management ownership, and are backed by less profitable, but larger buyout firms and funds. These same characteristics are also among the drivers behind the length of the divestment period in our main sample. In addition, the buyout funds in the nonfully exited sample hold more shares and board seats than in the exited subsample. This may be another explanation as to why these firms had not fully exited at the end of our observation period. Interestingly, 13 of the 47 subsample firms did not have any buyout fund share sales following the IPO. To check for the robustness of our main results in a statistical framework, we use this sample to run five additional regression models, all presented in Table AIII. In addition to rerunning our main regression models, as presented in Tables IV and VI including the previously omitted subsample of nonfully exited portfolio companies, we also run a Cox proportional hazard model following Cox (1972). The model is right-censored as the exit as failure event can only be recorded for 222 portfolio companies, while the 47 nonfully exited companies' exit processes end with our observation period at year-end 2012. All of the results from our main analyses are confirmed in the regression models. This allows us to rule out that the subsample of nonfully exited firms may have different exit criteria than those that were included in the main sample. The Cox (1972) model confirms that our main indicator variable, the financial success of the deal, is a relevant factor as to the length of the divestment period.

We perform a number of additional robustness tests, which for reasons of brevity are only discussed and not reported in tables. To obtain a homogeneous and unbiased data set, we specifically control for: 1) delistings and/or M&A transactions of the portfolio company following the IPO, which may represent an exit due to external causes, as opposed to deliberate exits by the funds, 2) RLBOs that may be governed by different goals or investment strategies as compared to first time IPOs, and 3) share conversion rights. We control for the first two factors by running all of our analyses with and without delisted companies and RLBOs. Although we always measure the relative magnitude of the share sales in relation to the buyout funds' shareholdings at the time of the IPO, we also account for share convertibles. We determine the actual amounts of share conversions using the aforementioned SEC reports and find that only a few buyout funds convert minor amounts of shares in the post-IPO process. Most share conversions happen at the IPO (as reported in the S-l and 424-B filings), which we control for using the shareholdings including all conversions as our main reference value for the at-IPO shareholdings of the funds. All reported values regarding the "percentage of shares sold" are the number of shares sold in each sales transaction in relation to the buyout fund's total shareholdings, including all conversions at the time of the IPO.

To validate the results we find in different market periods, we use a different classification of bull and bear market periods, as suggested by Chen (2009). We measure stock bull and bear market periods using the 250-day moving averages of the Dow Jones Industrial Index (~DJI). The 250-day moving average is calculated as the mean of the Dow Jones Industrial Average Index values of the previous 250 trading days. Bull markets are defined as periods in which daily end-of-day DJI values are above the moving average, while bear markets are defined as periods in which daily end-of-day DJI values are below the moving average. As an additional test, we use the methodology of "up" and "down" months in the stock market environment, as proposed by Fabozzi and Francis (1977). Finally, we re-run our main regression model as presented in Table IV to control for the general stock market environment by including variables proxying for the number of IPOs and the general underpricing during periods in which the buyouts were performed and taken public (in line with Loughran and Ritter, 2004). Our results do not change in any of the aforementioned settings. We also use several robustness tests to validate our event study results. First, we employ different indices to calculate the abnormal returns (value weighted and equal weighted CRSP Index). We also use different event study methodologies to calculate the abnormal returns (market, index (Lakonishok and Vermaelen, 1990), and average return (Masulis, 1980) model). Furthermore, we account for dependence in the residuals. As a test statistic, we use the standard deviation of the whole time series in the estimation window. We run all event studies with and without the dotcom stock market period and with different event periods as an additional robustness check. All of the results are robust to these changes.

Finally, to determine whether our results are also valid for non-US buyouts, we presented our results to representatives of two major European buyout funds. (15) Although this is not a formal robustness test, conducting these interviews with PE managers is not uncommon in this research area (also performed by Metrick and Yasuda, 2010), especially because it is a rather opaque industry and obtaining reliable information through data sets alone is oftentimes insufficient. The fund managers confirmed that our results are also common for European buyout deals, especially regarding the exit behavior and its determinants. Although this is not a valid statistical robustness test, we believe that it is an indication that our results do not suffer from a selection bias based on our US-only data sample.

VI. Conclusion

This paper analyzes how buyout fund investors exit their portfolio companies following IPOs. We specifically analyze when and how the buyout funds divest their shareholdings and give up board seats in their portfolio companies, as well as which factors determine these decisions. To do so, we use a data set comprised of222 buyout investments that went public in the United States from 1999 to 2008.

We find three major results. First, buyout funds dispose of their portfolio companies gradually and steadily over time, instead of selling the majority of their shareholdings at or shortly after the IPO. The fund investors deliberately choose to stay invested for an average of three years following the IPO. The same is true for board seats that are retained even longer than the share ownership. This result is especially interesting in light of buyout funds' institutional backgrounds. Generating investor returns within the shortest possible period of time is among the main goals of buyout funds. Therefore, it could be assumed that buyout funds try to dispose of their investments as quickly and swiftly as possible. However, this does not seem to be the case. In addition, the choice of an individual exit strategy in a portfolio company seems to depend upon a variety of scenarios. While the length of the period from the initial investment to the IPO is primarily driven by portfolio company-specific reasons, such as the restructuring process or profitability, the post-IPO divestment strategy largely depends upon buyout fund-specific

factors. The dominant factors include the size of the buyout firm, its overall profitability, and, most importantly, the financial success of the deal in question. Buyout investors do not appear to be keen to hold onto investments longer if they are able to return their initial investment in the deal prior to the IPO through dividend payments. Thus, the deal's financial success appears to be a pivotal factor in the choice of a divestment process. Moreover, even though capital markets react predominantly negatively to the exit of buyout funds in their portfolio companies, stock market investors distinguish between different exit strategies and exit events. To some degree, stock investors also react to the financial success of the deal from the perspective of the buyout fund. These results offer new and valuable insight into buyout funds' investment and divestment strategies.

Table AI. Variable Descriptions

The following table displays detailed explanations and measurement
units for all of the variables used in this paper. The variables are
grouped by main exit strategy indicators and further explanatory
variables used in the regression models.

Variable Name                 Unit    Description

Exit Strategy Indicators
  Investment Period           Years   Length between the date of the
                                        first (initial) investment
                                        made by a buyout fund in a
                                        portfolio company and the
                                        IPO date.
  Divestment Period           Years   Length between the IPO date and
                                        the final exit, measured as
                                        the last (recorded) share
                                        sale of the buyout fund in
                                        the portfolio company.
  Divestment Duration         Years   Value-weighted (with the
                                        relative cash flows) length
                                        of the divestment period,
                                        comparable to a bond duration.
  Exit Timing                 Years   Average (mean) length between
                                        the single share sale
                                        transactions of buyout funds
                                        in portfolio companies after
                                        the IPO. Calculations start
                                        with the first post-IPO share
                                        sale and encompass all share
                                        sales until the last recorded
                                        share sale.
  Unexpected Board            Dummy   Indicator showing if one or
  Resignation                           more board seats held by the
                                        buyout fund in a portfolio
                                        company are vacated through
                                        an unexpected resignation
                                        during the divestment period,
                                        as indicated in an 8-K filing
                                        (1 = At least one unexpected
                                        resignation in portfolio
                                        company, 0 = No unexpected
                                        resignation in portfolio
                                        company). Note that we
                                        distinguish between
                                        unexpected resignations that
                                        happened for reasons
                                        unrelated to the exit
                                        strategy (such as personal
                                        reasons or governance
                                        compliance of the portfolio
                                        company), and unexpected
                                        resignations that were part
                                        of the exit strategy (such as
                                        an unexpected exit in
                                        connection with a share
                                        sale). The former group is
                                        not classified as an
                                        unexpected exit, so that only
                                        unexpected board exits are
                                        included that were part of
                                        the exit process. A detailed
                                        explanation is given in Part
                                        III. A. of the paper.
Buyout Fund Variables
  Historic IRR                %       Net historic Internal Rate of
                                        Return (IRR), calculated as
                                        the historic realized average
                                        net IRR of the buyout firm in
                                        all its previous funds.
  Historic Capital Raising    Mn. $   Historic buyout firm capital,
                                        calculated as the total
                                        historic investor capital
                                        raised from the firms'
                                        inception to the IPO year of
                                        the respective portfolio
                                        company.
  Fund Vintage Age            Years   Buyout fund vintage age,
                                        calculated as the difference
                                        in years between the fund's
                                        vintage year and the
                                        portfolio company's IPO year.
  Firm Vintage Age            Years   Buyout firm age, calculated as
                                        the difference in years
                                        between the buyout firm's
                                        founding year and the
                                        portfolio company's IPO year.
  Deal Syndication            Dummy   Indicator for deal syndication
                                        (i.e., if more than one
                                        buyout fund invested in the
                                        deal at the time of the LBO,
                                        then 1 = Syndicated deal, 0 =
                                        No syndication).
  Buyout Ownership            %       The percentage of shares held
                                        by the buyout fund in the
                                        portfolio company before the
                                        IPO.
  Investment Returned at      Dummy   Indicator showing whether the
  IPO                                   initial investment made by
                                        the buyout fund in the
                                        portfolio company (in US
                                        Dollars) is already fully
                                        returned at the time of the
                                        IPO, either through special
                                        dividends or share
                                        redemptions (1 = Fully
                                        returned, 0 = Not fully
                                        returned).
  Dividend Recap              Dummy   Indicator showing whether the
                                        paid-out dividend from the
                                        portfolio company to the
                                        buyout fund pre-IPO is
                                        financed in a
                                        recapitalization scheme
                                        (i.e., funded through
                                        additional debt that is
                                        repaid from the IPO
                                        proceeds).
  Dividend Over               %       Total paid-out dividend from
  Revenues                              the portfolio company to the
                                        buyout fund pre-IPO in
                                        relation to the revenues of
                                        the portfolio company (as
                                        recorded on the balance
                                        sheet) at the time of the
                                        IPO.
  Chairman from Buyout        Dummy   Indicator showing whether the
  Fund                                  chairman of the board of the
                                        portfolio company is a buyout
                                        fund manager (1 = Yes, 0 =
                                        No).
  Percent Board Buyout        %       The percentage of seats on the
  Directors                             portfolio company's board of
                                        directors held by buyout fund
                                        managers.
  Fee Investment Return       %       The percentage of the initial
                                        (equity) investment of the
                                        buyout fund in the portfolio
                                        company (in US Dollars) that
                                        was returned solely through
                                        fees at the time of the IPO
                                        (fees include those fees that
                                        the portfolio company pays to
                                        the general partners for
                                        management advisory,
                                        transactions, and contract
                                        terminations).
  Management Fee Over         %       Total Management Fee paid from
  Revenues at IPO                       the portfolio company to the
                                        buyout fund during the
                                        investment period (for
                                        management, advisory, and
                                        monitoring services) in
                                        relation to the portfolio
                                        company's revenues (as
                                        recorded on the balance
                                        sheet) at the time of the
                                        IPO.
Portfolio Company Variables
  M&A Deals                   Dummy   Indicator showing whether the
                                        buyout fund managers used M&A
                                        deals to restructure the
                                        portfolio company before the
                                        IPO; that is, whether or not
                                        the portfolio company
                                        acquired other companies (or
                                        parts thereof) and/or
                                        sold/spun-off/carved-out own
                                        their company parts (1 = Yes,
                                        M&A deals, 0 = No M&A deals).

  Leverage Reduction          %       Reduction in leverage
  Pre-IPO                               (calculated as debt over
                                        market equity as recorded on
                                        the balance sheet) in the
                                        portfolio company from the
                                        LBO to the IPO.
  Leverage at IPO             %       Leverage (calculated as debt
                                        over market equity as
                                        recorded on the balance
                                        sheet) of the portfolio
                                        company at the IPO.
  Return on Assets at IPO     %       Return on Assets (calculated as
                                        net income over total assets,
                                        as recorded on the balance
                                        sheet and P&L statement) of
                                        the portfolio company at the
                                        time of the IPO.
  Increase in Return on       Dummy   Indicator showing whether the
  Assets                                portfolio company increased
                                        or decreased the Return on
                                        Assets (calculated as net
                                        income over total assets, as
                                        recorded on the balance sheet
                                        and P&L statement) from the
                                        LBO to the IPO (1 = Increase
                                        in RoA, 0 = Decrease in RoA).
  Total Assets at IPO         Mn. $   Total assets of the portfolio
                                        company (as recorded on the
                                        balance sheet) at the time of
                                        the IPO.
  Technology Firm             Dummy   Indicator showing whether the
                                        portfolio company is a
                                        technology firm (1 =
                                        Technology firm, 0 = No
                                        technology firm) based on
                                        Loughran and Ritter (2004).
  Ownership by                %       Percentage of shares of the
  Management                            portfolio company held by its
                                        top-level management (all
                                        company officers) excluding
                                        buyout officers/directors.
  Percent of Board            %       Percentage of board seats in
  Outside Directors                     the portfolio company held by
                                        outside directors (outside
                                        directors defined as
                                        directors with no relation to
                                        the portfolio company other
                                        than being directors).
  Takeover Exit               Dummy   Indicator showing whether the
                                        post-IPO remaining ownership
                                        of each buyout fund in the
                                        portfolio company is sold
                                        en bloc to a secondary
                                        purchaser in an M&A
                                        transaction (1 = M&A, 0 = No
                                        M&A).
Market Conditions
  First Day Underpricing      %       Underpricing on the first
                                        trading day, calculated as
                                        the stock closing price minus
                                        the stock offer price,
                                        divided by the stock offer
                                        price.
  Bear Market Investment      Dummy   Indicator showing whether the
                                        buyout fund investment in the
                                        portfolio company was made
                                        during the time of a bear
                                        stock market.
  Bull Market IPO             Dummy   Indicator showing whether the
                                        IPO of the portfolio company
                                        took place during a bull
                                        stock market.

Table AII. Summary Statistics on Nonexited Subsample

The table presents selected summary statistics on the subsample of 47
nonfully exited portfolio companies at the end of our observation
period (year-end 2012). Note that we initially record 57 nonfully
exited portfolio companies; however, we have to exclude 10 companies
due to missing or imprecise data. We report characteristics of the
invested buyout funds and the companies themselves, as well as on the
exit process. All exit indicators (as recordable) are in line with
the main sample. Note that 13 of the subsample portfolio companies do
not have any post-IPO share sales and are, therefore, still fully
invested at year-end 2012. For all of the other portfolio companies,
we record an average of 1.3 share sales after the IPO in which an
average of 7.4% of the holdings is given up. 29 of all of the
subsample portfolio companies had an exit of a buyout fund employed
director. Of these board exits, 14 portfolio companies had buyout
directors leave the board unexpectedly. Finally, 10 of 47 subsample
LBOs were made in stock bear market periods (following our definition
as presented in Part III A.), while 43 of 47 IPOs were made in stock
bull market periods.

Characteristics                                      Mean     Median

Buyout Fund
  Number of Invested Buyout Funds                      1.65     1.00
  Historic IRR (%)                                    16.09    16.10
  Historic Capital Raising                          8,970     3,705
  Fund Vintage Age                                     4.17     3.00
  Firm Vintage Age                                    13.05     8.00
Portfolio Company
  Ownership Management--Pre-IPO (%)                    6.43     2.91
  Ownership Management--at IPO (%)                     4.82     2.00
  Board Size                                           7.91     7.00
  Total Assets at IPO                               1258.1    522.5
  Revenues at IPO                                    993.2    487.6
  Income at IPO                                       38.24    12.10
  Leverage at IPO                                      3.58     1.08
  Return on Assets at IPO (%)                          3.24     1.60
Exit Characteristics
  Investment Period                                    3.85     2.88
  Length IPO to End-of-Period (Year-End 2012)          7.42     7.11
  Buyout Ownership--Pre-IPO (%)                       75.82    67.60
  Buyout Ownership--at IPO (%)                        54.86    58.55
  Board Seats Before IPO (%)                          33.07    33.30
  Board Seats at IPO (%)                              33.07    33.30
  Ownership at the End of IPO Year (%)                50.91    51.85
  Ownership at the End of IPO Year + 1 Year (%)       45.84    50.01
  Ownership at the End of IPO Year + 2 Years (%)      43.88    48.65
  Ownership at Year-end 2012 (%)                      38.49    38.62

Table AIII. Robustness Tests

The table presents the results of multivariate regression models used
for robustness testing of the paper's main results. All displayed
regression models use the full data set of 269 LBO-backed IPO
portfolio companies in the United States from 1999 to 2008. These 269
companies now include the 47 companies that did not have a full exit
of the invested LBO firms at the end of our observation period in
December 2012 which we exclude in previous analyses. We actually
record 57 nonfully exited companies. However, 10 companies are
excluded due to missing/imprecise data. Models I and III-V use the
same specification as the corresponding models presented in Tables IV
(for Models I, III, and IV) and VI (Model V). Models III, IV, and V
now use the specification including share sales, unexpected board
exits, and CARs of the first share sales in those 47 companies
without full exit at year-end 2012. Model II is a right-censored Cox
proportional hazard model. The failure event is the exit. Thus, we
obtain 222 failed (= exited) observations and 47 nonexited
observations, yielding an incident rate of 21%. The time-at-risk is
measured as the time from IPO to eventual exit (for exited firms) and
to year-end 2012 (for nonexited firms) in years. In Models I and
III-V, we report t-values, while Model II displays the coefficients
and corresponding hazard rates (HR). Unreported Wald Chi-2 tests
confirm the validity of the hazard model (a Chi-2 value of
150.95 ***). All robustness results confirm the main findings.

                                     (I)                 (II)
                                  Investment           Cox Hazard
                                    Period               Model

Variables                     Coef.     t-stat     Coef.       HR

Buyout Fund Variables
  Log(Historic IRR)           -1.03    -3.18 ***    0.71    2.04
  Log(Historic Capital         0.03     0.65       -0.22    0.80 ***
  Raising) Log(Fund Vintage                        -0.43    0.64 ***
    Age)
  Log(Firm Vintage            -0.18    -2.06 **     0.32    1.37 *
    Age)
  Deal Syndication            -0.95    -0.26       -0.25    0.77
  Buyout Ownership             0.46     0.49       -0.56    0.57
    at IPO
  Investment Returned          0.52     1.31        0.83    2.30 ***
    at IPO
  Chairman from               -0.19    -0.45       -0.44    0.63 *
    Buyout Fund
  Percent of Board             0.15     0.11        1.79    1.04 *
    Buyout Directors
  Fee Investment               0.16     0.84        2.06    1.88 ***
    Return

Portfolio Company
    Variables
  M&A Deals                    0.98     2.52 **     0.05    1.05
  Leverage Reduction           0.22     1.59       -0.00    0.99
    Pre-IPO
  Leverage at IPO              0.45     1.27       -0.07    0.92 ***
  Return on Assets at          0.29     1.81 *      2.44    1.58 *
    IPO
  Increase in Return           0.77     2.95 ***    0.21    1.23
    on Assets
  Log(Total Assets at         -0.43    -3.83 ***    0.01    1.00
    IPO)
  Technology-Firm             -0.70    -1.26       -0.08    0.92
  Ownership by                 0.31     0.30        1.46    2.31 ***
  Management at IPO
  Percent of Board             0.93     0.68       -0.01    0.98
    Outside Directors
  Takeover Exit                                     0.18    1.09

Market Conditions
Bear Market                    0.64     1.82 *      0.21    1.23
    Investment
  Bull Market IPO             -0.09    -0.99        0.10    1.11
  Investment Period                                -0.10    0.90
  Dividend Recap
  Dividend over
    Revenues
  Management Fee
  over Revenues at
    IPO
  First Day
    Underpricing
  Constant                     3.19     4.69 ***

  Used Sample                     All fully            All fully
                                  exited and           exited and
                                   nonfully             nonfully
                                    exited               exited

  Adj. [R.sup.2]                     0.25                 0.15

                                    (III)                 (IV)
                                    Exit            Unexpected Board
                                    Timing                Exit

Variables                      Coef     t-stat     Coef.     t-stat

Buyout Fund Variables
  Log(Historic IRR)           -0.80    -2.02 ***    -9.08   -0.17
  Log(Historic Capital         0.12     2.91 ***     3.28    1.08
  Raising) Log(Fund Vintage    0.10     0.84         1.68    0.19
    Age)
  Log(Firm Vintage            -0.06    -1.00        -6.70   -1.54
    Age)
  Deal Syndication            -0.08    -1.08         5.44    0.58
  Buyout Ownership            -0.15    -1.58         0.13    0.38
    at IPO
  Investment Returned         -0.46    -3.01 ***     0.38    0.05
    at IPO
  Chairman from                0.18     2.05 **      4.22    0.85
    Buyout Fund
  Percent of Board            -0.22    -1.36         0.19    0.90
    Buyout Directors
  Fee Investment              -0.05    -2.45 **      0.02    2.80 ***
    Return

Portfolio Company
    Variables
  M&A Deals                   -0.02    -0.40        -7.62   -0.93
  Leverage Reduction          -0.10    -0.84       -18.50   -2.49 **
    Pre-IPO
  Leverage at IPO              0.12     1.55         6.02    0.45
  Return on Assets at         -0.33    -1.05        -5.28   -0.21
    IPO
  Increase in Return          -0.08    -0.86        -7.38   -0.91
    on Assets
  Log(Total Assets at         -0.17    -2.38 **      4.01    1.85 *
    IPO)
  Technology-Firm             -0.08    -1.12        -6.33   -0.92
  Ownership by                -0.28    -1.20        -0.18   -0.65
  Management at IPO
  Percent of Board            -0.04    -0.06         0.00    0.02
    Outside Directors
  Takeover Exit                                    -24.04   -3.50 ***
Market Conditions
Bear Market                    0.16     1.54         5.08    0.50
    Investment
  Bull Market IPO              0.19     1.81 *       9.08    1.00
  Investment Period           -0.13    -1.04         1.66    0.24
  Dividend Recap
  Dividend over
    Revenues
  Management Fee
  over Revenues at
    IPO
  First Day
    Underpricing
  Constant                     2.16     2.94 ***    -1.62   -1.94 *
  Used Sample                      All fully           All fully
                                  exited and           exited and
                                   nonfully             nonfully
                                 exited with          exited with
                                 > 1 post-IPO         unexpected
                                 share sales          board exits
  Adj. [R.sup.2]                     0.22                 0.12

                                     (V)
                                 First Share
                                  Sale CARs

Variables                     Coef.     z-stat

Buyout Fund Variables
  Log(Historic IRR)           -0.18    -0.04
  Log(Historic Capital        -0.45    -1.39
  Raising) Log(Fund Vintage   -1.11    -1.01
    Age)
  Log(Firm Vintage             1.19     1.29
    Age)
  Deal Syndication            -2.00    -1.40
  Buyout Ownership             0.02     0.38
    at IPO
  Investment Returned
    at IPO
  Chairman from               -1.51    -1.35
    Buyout Fund
  Percent of Board            -0.02    -0.32
    Buyout Directors
  Fee Investment
    Return
Portfolio Company
    Variables
  M&A Deals                    2.12     1.95 *
  Leverage Reduction          -0.00    -0.31
    Pre-IPO
  Leverage at IPO             -0.00    -0.22
  Return on Assets at          0.01     0.00
    IPO
  Increase in Return           1.42     1.49
    on Assets
  Log(Total Assets at          0.74     1.38
    IPO)
  Technology-Firm              0.22     0.24
  Ownership by                -0.13    -1.84 *
  Management at IPO
  Percent of Board             0.00     0.31
    Outside Directors
  Takeover Exit
Market Conditions
Bear Market                    2.10     1.94
    Investment
  Bull Market IPO             -0.71    -0.46
  Investment Period
  Dividend Recap               0.14     0.11
  Dividend over                3.54     3.20 ***
    Revenues
  Management Fee              -0.00    -1.74 *
  over Revenues at
    IPO
  First Day                   -0.02    -0.80
    Underpricing
  Constant                      --     -1.39
  Used Sample                      All fully
                                  exited and
                                   nonfully
                                 exited with
                                 >0 post-IPO
                                 share sales
  Adj. [R.sup.2]                     0.11

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.


References

Acharya, V, O. Gottschalg, M. Hahn, and C. Kehoe, 2013, "Corporate Governance and Value Creation: Evidence from Private Equity," Review of Financial Studies 26, 368-402.

Barry, C.B., C.J. Muscarella, J.W. Peavy III, and M.R. Vetsuypens, 1990, "The Role of Venture Capital in the Creation of Public Companies: Evidence from the Going-Public Process," Journal of Financial Economics 27, 447-472.

Bozcuk, A. and A.M. Lasfer, 2005, "The Information Content of Institutional Trades on the London Stock Exchange," Journal of Financial and Quantitative Analysis 40, 621-644.

Bradley, D.J., B.D. Jordan, I.C. Roten, and H. Yi, 2001, "Venture Capital and IPO Lockup Expiration: An Empirical Analysis," Journal of

Financial Research 24, 465-493.

Brav, A. and P.A. Gompers, 2003, "The Role of Lockups in Initial Public Offerings," The Review of Financial Studies 16, 1-29.

Brown, S.J. and J.B. Warner, 1980, "Measuring Security Price Performance," Journal of Financial Economics 8, 205-258.

Candelon, B., J. Piplack, and S. Straetmans, 2008, "On Measuring Synchronization of Bulls and Bears: The Case of East Asia f Journal of Banking & Finance 32, 1022-1035.

Cao, J., 2011, "IPO Timing, Buyout Sponsors' Exit Strategies, and Firm Performance of RLBOs f Journal of Financial and Quantitative Analysis 46, 1001-1024.

Chauvet, M. and S. Potter, 2000, "Coincident and Leading Indicators of the Stock Market," Journal of Empirical Finance 7, 87-111.

Chen, S.-C., 2009, "Predicting the Bear Stock Market: Macroeconomic Variables as Leading Indicators," Journal of Banking & Finance 33, 211-223.

Chiyachantana, C.N., P.K. Jain, C. Jiang, and R.A. Wood, 2004, "International Evidence on Institutional Trading Behavior and Price Impact," Journal of Finance 59, 869-898.

Cox, D.R., 1972, "Regression Models and Life Tables," Journal of the Royal Statistical Society, Series B 34, 187-220.

Das, S.R., M. Jagannathan, and A. Sarin, 2003, "Private Equity Returns: An Empirical Examination of the Exit of Venture-Backed Companies," Journal of Investment Management 1, 1-26.

Fabozzi, F.J. and J.C. Francis, 1977, "Stability Tests for Alphas and Betas Over Bull and Bear Market Conditions," Journal of Finance 32, 1093-1099.

Field, L.C. and G. Hanka, 2001, "The Expiration of IPO Share Lockups," Journal of Finance 56, 471-500.

Gompers, P., 1996, "Grandstanding in the Venture Capital Industry, "Journal of Financial Economics 42, 133-156.

Gompers, P. and J. Lemer, 1998, "Venture Capital Distributions: Short Run and Long Run Reactions," Journal of Finance 53, 2161-2183.

Harford, J. and A. Kolasinski, 2014, "Do Private Equity Returns Result from Wealth Transfers and Short-Termism? Evidence from a Comprehensive Sample of Large Buyouts f Management Science 60, 888-902.

Jensen, M.C., and W.H. Meckling, 1976, "Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure," Journal of Financial Economics 3, 305-360.

Kaplan, S.N., 1991, "The Staying Power of Leveraged Buyouts," Journal of Financial Economics 29, 287-313.

Kaplan, S.N. and P. Stromberg, 2009, "Leveraged Buyouts and Private Equity," Journal of Economic Perspectives 23, 121-146.

Krishnan, C.N.V, VI. Ivanov, R.W. Masulis and A.K. Singh, 2011, "Venture Capital Reputation, Post-IPO Performance, and Corporate Governance," Journal of Financial and Quantitative Analysis 46, 1295-1333.

Lakonishok, J. and T. Vermaelen, 1990, "Anomalous Price Behavior Around Repurchase Tender Offers," Journal of Finance 45, 455-477.

Lemer, J., 1994, "Venture Capitalists and the Decision to Go Public," Journal of Financial Economics 35, 293-316.

Lin, T.H. and R.L. Smith, 1998, "Insider Reputation and Selling Decisions: The Unwinding of Venture Capital Investments During Equity IPOs," Journal of Corporate Finance 4, 241-263.

Ljungqvist, A. and M. Richardson, 2003, "The Investment Behavior of Private Equity Fund Managers," NYU Working paper.

Loughran, T. and J. Ritter, 2004, "Why Has IPO Underpricing Changed Over Time," Financial Management 33, 5-37.

MacKinlay, A.C., 1997, "Event Studies in Economics and Finance," Journal of Economic Literature 35, 13-39.

Masulis, R.W., 1980, "The Effects of Capital Structure Change on Security Prices: A Study of Exchange Offer," Journal of Financial Economics 8, 139-178.

Metrick, A. and A. Yasuda, 2010, "The Economics of Private Equity Funds," The Review of Financial Studies 23, 2303-2341.

Ofek, E. and M. Richardson, 2003, "DotCom Mania: The Rise and Fall of Internet Stock Prices," Journal of Finance 58, 1113-1137.

Ozmel, U., D.T. Robinson, and T.E. Stuart, 2013, "Strategic Alliances, Venture Capital, and Exit Decisions in Early-Stage High-Tech Firms," Journal of Financial Economics 107, 655-670.

Pagan, A.R. and K.A. Sossounov, 2003, "A Simple Framework for Analyzing Bull and Bear Markets," Journal of Applied Econometrics 18, 23-46.

Preqin, 2013, Preqin Private Equity Fund Terms Advisor 2012, London, UK, Preqin.

Schmidt, D., S. Steffen, and F. Szabo, 2010, "Exit Strategies of Buyout Investments: An Empirical Analysis," Journal of Alternative Investments 12, 58-84.

Shleifer, A. and R.W. Vishny, 1986, "Large Shareholders and Corporate Control," Journal of Political Economy 94, 461-488.

Siegel, J.J., 2008, Stocks for the Long Run, 4th Ed., New York, NY, McGraw Hill.

Stromberg, P., 2007, "The New Demography of Private Equity," Swedish Institute for Financial Research Working paper.

Sven Furth and Christian Rauch ([dagger])

* For valuable comments and suggestions, we would like to thank an anonymous referee and Raghavendra Rau (Editor), as well as Einar Bakke, Dirk Bursian, Walid Busaba, Stoyu I. Ivanov, Andras Marosi, Mark Wahrenburg, Uwe Walz, Andrew Winton, David Yermack, and participants in the Global Conference on Business and Finance 2012, the Brown Bag Seminar of Goethe University Frankfurt, the 2012 Southern Finance Association Annual Meeting, and the 2012 Northern Finance Association Annual Meeting. Sven Fuurth gratefully acknowledges the financial support of "Vereinigung der Freunde und Foorderer der Goethe Universitaat Frankfurt" and of "Commerzbank Stiftung Frankfurt." Christian Rauch gratefully acknowledges the financial support of the SAFE Center of Excellence at Goethe University Frankfurt. All remaining errors are our own.

([dagger]) Christian Rauch is an Assistant Professor in the Finance Department and SAFE Center of Excellence at Goethe University Frankfurt in Frankfurt, Germany. Sven Furth is an Assistant Professor in the Finance Department at Goethe University Frankfurt in Frankfurt, Germany.

(1) According to SEC rules, all trades by directors, officers and "beneficial" owners (following definitions of SEC Rule 16(a)) must be disclosed no later than the second business day following a change in ownership. This rule allows us to track the date and volume of all share sales through buyout funds above a 10% ownership threshold after the IPO. An exemption to the rule states that in case the inside shareholder (in our case the buyout fund) still has a board membership in the relevant firm, share sales must be reported regardless of the amount of shareholdings, allowing us to track share sales below the 10% threshold for those buyout investors with board seats in their companies.

(2) We deliberately restrict our sample period to 1999 as consistent digital insider trading data is primarily available from 1999 onward.

(3) Originally, we obtain 279 IPOs of which 57 still had an active buyout investor at the end of our observation period. As explained in the robustness section (Section V) of the paper, we also test our main result for this originally omitted subsample. All results are confirmed.

(4) Based on Preqin "Buyout Deals Analyst" data (www.preqin.com) from 2002 to 2013 period, the US buyout market alone accounts for 52.7% of all global buyout deals and 58.1% of the total buyout deal volume worldwide. 56.7% of all capital was committed to US-based buyout funds, and six of the ten top-performing GPs (in terms of average historic IRR) are from the United States.

(5) Data is taken from Preqin "Buyout Deals Analyst" (buyout volumes) and Thomson Reuters SDC (IPO volumes).

(6) The duration is calculated as: [ZIGMA] Time x Transaction Value/[ZIGMA] transaction Values, where the time factor is measured in years and the transaction value is measured as the amount of shares sold times the share price. Two buyout firms might have identical divestment periods; however, one firm might sell most of its shareholdings at an early point in time after the IPO, while a competing firm might hold onto the largest portion of its holdings until the end of the divestment period. Following our calculation, the latter firm would have a higher duration indicating a longer active investment in the portfolio company.

(7) It should be noted that we assume that all unexpected exits, for which no specific reason was mentioned, were part of an active exit strategy, which is why we assume them to be "aggressive."

(8) The literature actually proposes a broader view as to what bull and bear stock markets are. Chauvet and Potter (2000) talk about generally rising or declining markets, while a more refined view in the modern literature is presented by Pagan and Sossounov (2003) who address an increase or decrease of at least 20% to 25%.

(9) We keep the estimation period deliberately short so we do not bias the CAR results by including severe events, such as the unlock day or a stock run-up following the IPO.

(10) It is the underwriter who can allow single pre-IPO shareholders to sell shares in spite of the existing lockup agreement.

(11) Empirical evidence is provided by Preqin Private Equity Research in a 2011 market overview ("Preqin Special Report: US Private Equity") or by the Private Equity Growth Capital Council (PEGCC) under www.pegcc.org/research.

(12) Note that due to these high technical correlations, we cannot jointly include the variables in the regression framework.

(13) These fees are paid in addition to fund-level management fees and carried interest by the portfolio company to the general partner for certain services the general partner performs, such as restructuring, monitoring, and advising. To a large degree, these fees are not passed on to the limited partners. As such, the general partner is the sole beneficiary of these fees. These fees are only paid until the IPO in which the advisory contract between the portfolio company and the general partner (i.e., the contract that sets forth the payments of these fees) is terminated. For a detailed description of these fees, see Metrick and Yasuda (2010) or Preqin (2013).

(14) So-called "Dividend Recap"-schemes are frequently used tools of buyout investors. In these schemes, dividend payouts are funded with additional debt the portfolio companies take on. The debt-funded dividend is then paid out to the buyout fund investors as a regular fond distribution. Usually, these transactions are performed shortly before an IPO, as the equity generated through the IPO can be used to repay the debt. Prominent examples for these schemes are, e.g., Burger King, Warner Music, or Domino's Pizza, which are all included in our data set.

(15) Both funds wished to remain unnamed.

Table 1. Summary Statistics

The following table displays the summary statistics for the 222
buyout-backed IPO firms in our sample. Panel A reports the summary
statistics on general information about the buyout funds and the
portfolio companies. The upper part of the table presents several
buyout fund-specific indicators, as explained in the appendix of the
paper. The lower part of Panel A reports portfolio company-specific
information: 1) the ownership by the management of each portfolio
company, 2) the average joint ownership of all buyout investors in
each portfolio company, 3) the portfolio company board size, 4) total
assets of the portfolio companies at the IPO, 5) revenues of the
portfolio companies at the IPO, 6) net income of the portfolio
companies at the IPO, and 7) leverage (calculated as debt over
equity) of the portfolio companies at the IPO. Panel B provides the
industry distribution based on the one-digit SIC. Note that the
wording "at IPO" refers to the point in time when the primary equity
market offering has taken place, but secondary trading in the stock
has not yet commenced. "Pre IPO" refers to the point in time before
the primary market offering takes place.

Panel A

Characteristics                         Mean     Median       SD

Buyout Fund
  Number of Invested Buyout Funds        1.40      1.00        0.67
  Historic IRR (%)                      18.57     17.97       13.98
  Historic Capital Raising             6,356      2,815      8,317
  Fund Vintage Age                       4.55      4.00        2.65
  Firm Vintage Age                      16.18     14.00       10.69
Portfolio Company
  Ownership Management--pre IPO (%)      9.87      4.45       15.83
  Ownership Management--at IPO (%)       7.39      3.28       12.17
  Buyout Ownership--pre IPO (%)         67.85     74.45       25.72
  Buyout Ownership--at IPO (%)          46.13     45.90       18.13
  Board Size                             6.97      7.00        1.86
  Total Assets at IPO                  787.50    387.10     1209.02
  Revenues at IPO                      726.23    322.72     1331.69
  Income at IPO                         11.22      6.67       64.40
  Leverage at IPO                        1.48      0.98        1.82
  Return on Assets at IPO                4.22       2.1        6.70

Panel B

                                       Buyout
Portfolio Company Industry             Firms    IPO Firms

Agriculture, Forestry, and Fishing        1        0.45%
Mining                                   17        7.65%
Construction                              1        0.45%
Manufacturing                            78       35.14%
Wholesale Trade                           9        4.05%
Retail Trade                             24       10.81%
Services                                 62       27.93%
Communications                           30       13.51%

Total                                   222      100.0%

Table II. The Exit Roadmap: Share Sales and Board Seats

This table displays the timing and magnitude of exit strategies
(share sale transactions and director exits) for the 308 funds and
222 portfolio companies in our sample. Panel A presents the summary
statistics of general ownership indicators. We report the investment
period, divestment period, divestment duration (all calculated as
explained in Part III. A and the appendix), and the total time the
buyout fund managers spend on boards of their portfolio companies (in
years). All numbers are mean, median, and standard deviation (SD)
values calculated on the buyout fund level. In Panel B, we report the
actual exits in which the buyout funds sell shares and give up board
seats after the IPO. We report all exits post-IPO, the first single
exit after the IPO, as well as the last and single highest exit. For
each exit, we provide the magnitude (in form of the transaction
value), the time between sales and board exits, and the relative
reduction in holdings the exit triggers. We further report values on
the exits in relation to the end of the lockup period after the IPO.
In Panel C, we present information on expected and unexpected board
exits. We also report the exit specifics for both groups, as well as
additional information regarding the reasons for the unexpected
exits. Panel D contains information about the changes in exit
behavior in different (stock) market environments. We define bull and
bear markets using stock market peaks and troughs (in six-month
windows) represented by the index values of the Dow Jones Industrials
Average (^DJI) stock index. We also report the mean values of all of
the variables across the two aggregated bull and bear market periods.
In addition, difference-in-means values are reported, comparing the
means across the sample for the combined bull and bear markets,
respectively. The last variable, unexpected exits (%), is calculated
as the percentage of unexpected board exits in relation to the number
of all board exits. Note that the wording "at IPO" refers to the
point in time when the primary market equity offering has taken
place, but secondary trading in the stock has not yet commenced.

Panel A. Ownership Summary

                                 Mean   Median     SD

Investment Overview
  Investment Period              3.07     2.44    2.32
  Divestment Period              2.81     2.38    1.95
  Divestment Duration            2.28     1.73    1.70
  Total Time on Boards           3.35     3.00    2.12
    post IPO
Ownership (%)
  Before IPO                    48.91    43.49   29.33
  at IPO                        33.25    29.33   20.11
Ownership at the end of
  IPO year                      27.07    23.56   21.87
  IPO + 1 year                  22.16    16.00   21.41
  IPO + 2 year                  19.63    12.58   20.75
  IPO + 3 year                  18.27    11.04   20.22
Board Seats (%)
  Board Seats before IPO        32.02    28.57   15.69
  Board Seats at IPO            31.34    28.57   15.82
Board Seats at the end of:
  IPO Year                      26.43    25.00   16.27
  IPO + 1 year                  20.51    16.67   17.09
  IPO + 2 year                  16.97    14.29   16.12
  IPO + 3 year                  14.56    12.50   16.14

Panel B. Divestment Process

                               # Firms    Mean    Median     SD

All Share Sales post IPO         222
  Reduction in Holdings                  24.95%   13.76%   28.29%
  Time Between Sale                       0.45     0.18     0.85
    Transactions
  Transaction Value                       84.61   29.38    156.58
First Share Sale post IPO        222
  Years from IPO                          1.75     1.16     1.61
  Reduction in Holdings                  46.00%   35.95%   35.21%
  Transaction Value                      138.73   62.80    228.63
Last Share Sale post IPO         222
  Years from IPO                          2.81     2.38     1.95
  Reduction in Holdings                  49.58%   42.88%   35.56%
  Transaction Value                      137.10   61.03    216.19
Highest Share Sale post          222
    IPO
  Years from IPO                          2.36     1.65     1.95
  Reduction in Holdings                  57.78%   51.13%   29.95%
  Transaction Value                      168.77   82.38    238.47
Share Sale pre-Lockup             37
  Reduction in Holdings                  22.10%   11.58%   27.07%
  Transaction Value                       49.95   11.90    90.27%
Share Sale at Lockup-End          2
  Reduction in Holdings                  22.37%   5.86%    37.47%
  Transaction Value                       67.19   12.55    124.74
All Board Exits post IPO         190
  Reduction in Board Seats               58.68%   50.00%   29.26%
  Time Between Board Exits                1.59     1.19     1.44
First Board Exit post IPO        190
  Years from IPO                          2.29     1.87     1.89
  Reduction in Board Seats               68.88%   50.00%   29.97%
Last Board Exit post IPO         190
  Years from IPO                          3.36     3.00     2.12
Board Exit pre-Lockup             15
  Reduction in Board Seats               53.77%   50.00%   24.72%
  Board Exit at Lockup-End        0
Reduction in Board Seats                  0.00%   0.00%    0.00%

Panel C. Information on Unexpected Board Exits

                                Number    Mean    Median     SD

Expected Board Exits
  # Directors Leaving            125
    Boards Expectedly
  Reduction in Board Seats               66.35%   50.00%   30.22%
  Years from IPO to First                 2.87     2.42     1.89
    Expected Exit
  Time Spent on Boards                    6.22     5.85     2.72
    Expected Exits
Unexpected Board Exits
  # Directors Leaving            225
    Boards Unexpectedly
  Reduction in Board Seats               53.46%     50%    27.7%
  Years from IPO to First                 1.86     1.35     1.75
    Unexpected Exit
  Time Spent on Boards                    5.38     5.09     2.65
    Unexpected Exits

                                Number        Percent

Reasons for Unexpected Exits
  No specific Information        116          51.56%
  Personal Reasons                11           4.89%
  Focus on                        10           4.44%
    Responsibilities with
    PE Firm
  Exit in Connection with         31          13.78%
    PE Share Sales
  Governance Compliance            9           4.00%
  Unexpected Exits w/o            48          21.33%
    Disagreements

Panel D. Exits in Different Market Environments

                                 Bull          Bear           Bull
                               Market 1      Market 1       Market 2
                             01/01/1999-    01/15/2000-   10/09/2002-
                              01/14/2000    10/08/2002     10/09/2007

Share Sales
  Number of Share Sales           1.33          2.66           2.74
  Reduction in Holdings          35.68         17.69          26.58
    (%)
  Transaction Value              64.40         30.15         104.64
  Time Between Share Sales        0.13          0.55           0.32
Board Seats
  Number of Board Exits           1.50          1.26           1.83
  Reduction in Board            100.00         62.94          57.55
    Seats (%)
  Time Between Board Exits        --            2.36           1.09
  Unexpected Exits (%)            0.00         16.67          44.51

                                 Bear          Bull
                               Market 2      Market 3
                             10/10/2007-    03/10/2009-       Bull
                              03/09/2009    12/31/2012    Markets Mean

Share Sales
  Number of Share Sales          2.18           2.78          2.93
  Reduction in Holdings         22.63          24.87         26.27
    (%)
  Transaction Value             67.40          68.95         95.40
  Time Between Share Sales       0.59           0.61          0.41
Board Seats
  Number of Board Exits          1.53           1.44          1.89
  Reduction in Board            57.96          60.26         58.45
    Seats (%)
  Time Between Board Exits       1.37           2.72          1.56
  Unexpected Exits (%)          29.79          20.18         37.30

                                 Bear       Difference-
                             Markets Mean    in-Means

Share Sales
  Number of Share Sales          2.36         0.57 *
  Reduction in Holdings         20.80         5.47 ***
    (%)
  Transaction Value             50.80        44.60 ***
  Time Between Share Sales       0.57        -0.16 **
Board Seats
  Number of Board Exits          1.44         0.45 ***
  Reduction in Board            59.58        -1.13
    Seats (%)
  Time Between Board Exits       1.70        -0.14
  Unexpected Exits (%)          25.48        11.82 **

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table III. Correlations of Different Exit Indicators

This table provides the results of a correlation analysis of the
different exit indicators. Statistically significant correlation
coefficients at the 99% level are indicated with *.

                         Investment   Divestment   Divestment
                           Period       Period      Duration

Investment Period           1.00
Divestment Period          -0.09        1.00
Divestment Duration        -0.12        0.95 *       1.00
Exit Timing                -0.03        0.40 *       0.24 *
Unexpected Board Exits      0.01        0.08         0.03

                          Exit    Unexpected
                         Timing   Board Exits

Investment Period
Divestment Period
Divestment Duration
Exit Timing               1.00
Unexpected Board Exits    0.07       1.00

Table IV. Cross-Sectional OLS Regression Results--Determinants
of Exit Strategies

The table presents the results of a multivariate OLS regression model
and a probit regression model used to determine factors influencing
buyout funds' exit behavior. The dependent variables are the five
main exit indicators, calculated as presented in Part III. A and the
appendix of the paper. We run the model separately for the (log)
investment period, the (log) divestment period, the (log) divestment
duration, exit timing, and unexpected (and aggressive) board exit
variables. Note that we do not test all of the variables for each
model. Vintage age is not included for the investment period as
almost all of the funds are incepted in the same year the investment
started. For the investment period, leverage and return on assets are
measured at the LBO, while buyout and management ownership are
measured pre-IPO. In addition, we do not include the post-IPO exit
indicators as explanatory variables in the regressions due to the
high (technical) correlations of the variables among each other, as
shown in Table III. For the probit regression model, we estimate the
marginal effects using the delta method. Note that we only include
industry-adjusted balance sheet and P&L variables. We use two-digit
SIC codes to cluster the industries. Based on these industry
classifications, we use median industry numbers for each included
variable taken from the Compustat database (including all of the
companies listed in Compustat within each industry). We deduct this
data from our actual balance sheet and P&L data, such that the new
variables reflect only the deviation from the median industry values.
Standard errors were corrected for heteroscedasticity and clustered
at the buyout fund level.

                                               (I)
                                            Investment
                                            Period OLS

Variables                               Coef.     t-stat

Buyout Fund Variables
  Log(Historic IRR)                     -0.70    -2.02 **
  Log(Historic Capital Raising)          0.05     1.21
  Log(Fund Vintage Age)
  Log(Firm Vintage Age)                 -0.11    -2.05 **
  Deal Syndication                      -0.03    -0.25
  Buyout Ownership at IPO                0.16     0.72
  Investment Returned at IPO             0.00     0.00
  Chairman from Buyout Fund             -0.13    -1.08
  Percent of Board Buyout Directors      0.04     0.17
  Fee Investment Return                  0.00     1.02
Portfolio Company Variables
  M&A Deals                              0.31     3.31 ***
  Leverage Reduction pre IPO             0.35     1.55
  Leverage at IPO                        0.01     0.04
  Return on Assets at IPO                0.37     1.99 **
  Increase in Return on Assets           0.21     2.17 **
  Log(Total Assets at IPO)              -0.22    -6.09 ***
  Technology-Firm                       -0.13    -1.15
  Ownership by Management at IPO        -0.18    -0.67
  Percent of Board Outside Directors    -0.26    -1.51
  Takeover Exit
Market Conditions
  Bear Market Investment                 0.21     1.97 **
  Bull Market IPO                       -0.02    -0.21
  Investment Period
  Constant                               4.88     6.78 ***
  Observations                                 222
  Adj. [R.sup.2]                               0.22

                                               (II)
                                            Divestment
                                            Period OLS

Variables                               Coef.     t-stat

Buyout Fund Variables
  Log(Historic IRR)                     -1.35    -2.48 **
  Log(Historic Capital Raising)          0.11     2.32 **
  Log(Fund Vintage Age)                  0.21     1.81 *
  Log(Firm Vintage Age)                 -0.10    -1.33
  Deal Syndication                       0.05     0.42
  Buyout Ownership at IPO                0.14     0.72
  Investment Returned at IPO            -0.53    -4.00 ***
  Chairman from Buyout Fund              0.43     3.57 ***
  Percent of Board Buyout Directors     -0.66    -2.42 ***
  Fee Investment Return                 -0.00    -0.81
Portfolio Company Variables
  M&A Deals                             -0.09    -0.94
  Leverage Reduction pre IPO            -0.02    -0.13
  Leverage at IPO                        0.61     3.85 ***
  Return on Assets at IPO               -0.70    -1.67 *
  Increase in Return on Assets           0.10     0.95
  Log(Total Assets at IPO)               0.01     0.27
  Technology-Firm                       -0.09    -0.89
  Ownership by Management at IPO        -1.03    -3.18 ***
  Percent of Board Outside Directors     0.01     0.06
  Takeover Exit                          0.00     0.02
Market Conditions
  Bear Market Investment                 0.13     1.02
  Bull Market IPO                       -0.24    -2.35 ***
  Investment Period                     -0.15    -1.42
  Constant                               0.31     0.32
  Observations                                 222
  Adj. [R.sup.2]                               0.16

                                              (III)
                                            Divestment
                                           Duration OLS

Variables                               Coef.     t-stat

Buyout Fund Variables
  Log(Historic IRR)                     -1.08    -1.90 *
  Log(Historic Capital Raising)          0.08     2.05 **
  Log(Fund Vintage Age)                  0.20     1.64
  Log(Firm Vintage Age)                 -0.07    -1.12
  Deal Syndication                       0.05     0.40
  Buyout Ownership at IPO                0.22     1.13
  Investment Returned at IPO            -0.46    -3.58 ***
  Chairman from Buyout Fund              0.31     2.80 ***
  Percent of Board Buyout Directors     -0.51    -1.98 *
  Fee Investment Return                 -0.00    -0.14
Portfolio Company Variables
  M&A Deals                             -0.03    -0.32
  Leverage Reduction pre IPO            -0.05    -0.34
  Leverage at IPO                        0.52     3.40 ***
  Return on Assets at IPO               -1.01     2 82
  Increase in Return on Assets           0.06     0.57
  Log(Total Assets at IPO)               0.05     1.02
  Technology-Firm                       -0.06    -0.64
  Ownership by Management at IPO        -0.78    -2.64 ***
  Percent of Board Outside Directors    -0.02    -0.14
  Takeover Exit                          0.08     0.65
Market Conditions
  Bear Market Investment                 0.09     0.72
  Bull Market IPO                       -0.30    -3.00 ***
  Investment Period                     -0.17    -1.53
  Constant                              -0.51    -0.57
  Observations                                 222
  Adj. [R.sup.2]                               0.16

                                               (IV)
                                               Exit
                                            Timing OLS

Variables                               Coef.     t-stat

Buyout Fund Variables
  Log(Historic IRR)                     -0.90    -2.12 **
  Log(Historic Capital Raising)          0.07     2.51 **
  Log(Fund Vintage Age)                  0.12     0.94
  Log(Firm Vintage Age)                 -0.06    -1.11
  Deal Syndication                      -0.11    -1.32
  Buyout Ownership at IPO               -0.23    -1.38
  Investment Returned at IPO            -0.24    -2.84 ***
  Chairman from Buyout Fund              0.33     2.79 ***
  Percent of Board Buyout Directors     -0.25    -1.36
  Fee Investment Return                 -0.00    -2.05 **
Portfolio Company Variables
  M&A Deals                             -0.06    -0.92
  Leverage Reduction pre IPO            -0.26    -1.00
  Leverage at IPO                        0.21     1.83 *
  Return on Assets at IPO               -0.52    -1.45
  Increase in Return on Assets          -0.06    -0.86
  Log(Total Assets at IPO)              -0.08    -2.45 **
  Technology-Firm                       -0.12    -1.24
  Ownership by Management at IPO        -0.38    -1.50
  Percent of Board Outside Directors    -0.09    -0.80
  Takeover Exit
Market Conditions
  Bear Market Investment                 0.14     1.40
  Bull Market IPO                        0.17     1.79 *
  Investment Period                     -0.10    -1.02
  Constant                               2.04     2.68 ***
  Observations                                 157
  Adj. [R.sup.2]                               0.14

                                               (V)
                                           Unexpected
                                        Board Exit Probit

Variables                               Coef.     z-stat

Buyout Fund Variables
  Log(Historic IRR)                     -10.57   -0.25
  Log(Historic Capital Raising)           4.64    1.43
  Log(Fund Vintage Age)                   1.75    0.18
  Log(Firm Vintage Age)                 -10.21   -2.60 **
  Deal Syndication                        7.08    0.61
  Buyout Ownership at IPO                 0.08    0.44
  Investment Returned at IPO              0.40    0.04
  Chairman from Buyout Fund               5.93    0.65
  Percent of Board Buyout Directors       0.17    0.90
  Fee Investment Return                   0.00    2.01 **
Portfolio Company Variables
  M&A Deals                              -9.46   -1.14
  Leverage Reduction pre IPO            -22.72   -2.38 **
  Leverage at IPO                         6.32    0.48
  Return on Assets at IPO                -5.88   -0.20
  Increase in Return on Assets           -7.42   -0.92
  Log(Total Assets at IPO)                7.35    2.18 **
  Technology-Firm                        -6.74   -0.96
  Ownership by Management at IPO         -0.15   -0.55
  Percent of Board Outside Directors      0.00    0.03
  Takeover Exit                         -28.05   -3.48 ***
Market Conditions
  Bear Market Investment                  5.06    0.51
  Bull Market IPO                        10.14    1.14
  Investment Period                       1.85    0.22
  Constant                               -1.34   -1.93 *
  Observations                                 189
  Adj. [R.sup.2]                               0.08

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table V. Capital Market Reactions to Buyout Exits

The table displays equity capital market reactions to buyout funds'
exits in their portfolio companies. As a proxy for these reactions,
we compute cumulative abnormal stock returns (CARs), calculated using
the market model with the CRSP Value-Weighted Index as the benchmark.
Panel A reports the average (mean) CARs for different share sales
including an average (mean) across all post-IPO share sales, the
first post-IPO share sale, the second post-IPO share sale, the
average (mean) CARs for all share sales between the first and the
last share sale, the last post-IPO share sale, and the single highest
post-IPO share sale. Panel B contains the CAR reactions to the
unexpected board exits, separately reported for aggressive (i.e.,
part of an active exit strategy) and nonaggressive (due to personal
reasons, etc.) exits, as defined in the appendix and Part III.A.
t-values are in the parentheses.

                              Whole Sample   Bull Market   Bear Market
Share Sales                       All            All           All

Panel A.

All                             -1.38 ***     -1.03 ***     -2.50 ***
                               (-7.11)       (-5.05)       (-5.24)
First                           -2.14 ***     -1.35 ***     -4.84 ***
                               (-4.74)       (-3.13)       (-3.78)
Second                          -1.71 ***     -1.45 ***     -2.57 *
                               (-3.72)       (-3.25)       (-1.93)
In Between                      -0.84 ***     -0.62 **      -1.49 ***
                               (-3.40)       (-2.19)       (-3.00)
Last                            -2.07 ***     -1.69 ***     -3.66 ***
                               (-5.63)       (-4.68)       (-3.21)
Flighest                        -1.83 ***     -1.28 ***     -4.05 ***
                               (-4.31)       (-2.91)       (-3.56)

Unexp. Aggressive Board       Whole Sample   Bull Market   Bear Market
Exits                             All            All           All

Panel B.

All                               0.35           0.31          0.59
                                 (0.58)         (0.44)        (0.47)
First                             1.36           1.05          3.02 **
                                 (1.57)         (1.05)        (2.36)
Last                              0.43          -0.00          2.78
                                 (0.39)        (-0.00)        (1.43)

Unexp. Nonaggressive Board
Exits                             All            All           All

All                              -0.57          -0.88          0.55
                                (-0.89)        (-1.60)        (0.25)
First                            -0.38          -0.07         -1.30
                                (-0.48)        (-0.12)       (-0.48)
Last                             -1.46          -1.30         -2.47
                                (-1.41)        (-1.20)       (-0.64)

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VI. Determinants of Cumulative Abnormal Returns (CARs)

The table presents the results of a multivariate OLS regression model
used to determine factors influencing the capital market reactions to
buyout funds' share sales in their portfolio companies post-IPO. We
measure the capital market reactions as cumulative abnormal returns
(CARs), as reported in Table V and described in Part III.C. All CARs
used in the regressions are calculated in the three-day event window
around the exits. We use the CARs at the very first post-IPO share
sale of the buyout funds as dependent variables. The explanatory
variables are identical with the regression model as presented in
Table IV. All of the variables are explained in the appendix. We
alter the set of explanatory variables in five instances. We replace
the investment returned at IPO-variable with 1) a variable measuring
the pre-IPO paid-out dividends in relation to the portfolio company's
revenues at the IPO, 2) with an interaction term of the investment
returned dummy and the return on assets of the portfolio company at
the IPO, and 3) a dummy indicating whether the pre-IPO dividend was
financed within in a recapitalization scheme (as explained in the
appendix). We also replace the fee variable of Table IV (fees over
initial investment) with a variable indicating the ratio of fees to
revenues at the time of the IPO. Finally, we include the first day
underpricing as an explanatory variable to account for price
corrections that might affect the CARs. Standard errors were
corrected for heteroscedasticity.

                                          First Share Sale
                                                CARs

                                                (I)

Variables                                Coef.     t-stat

Buyout Fund Variables
  Log (Historic IRR)                      -0.97   -0.26
  Log(Historic Capital Raising)           -0.59   -1.12
  Log(Fund Vintage Age)                   -1.19   -1.47
  Log(Firm Vintage Age)                    1.07    1.45
  Deal Syndication                        -2.40   -1.77 *
  Buyout Ownership                         0.01    0.37
  Investment Returned at IPO               3.35    2.91 ***
  Dividend Recap                           0.13    0.10
  Dividend over Revenues
  Investment Returned at IPO x
    Return on Assets at IPO
  Chairman from Buyout Fund               -2.06   -1.75 *
  Percent Board Buyout Directors          -0.01   -0.26
  Management Fee over Revenues at IPO     -0.00   -1.85 *
Portfolio Company Variables
  M&A Deals                                2.00    1.97 *
  Leverage Reduction pre IPO              -0.00   -0.54
  Leverage at IPO                          0.00    0.18
  Return on Assets at IPO                 -0.02   -0.36
  Increase in Return on Assets             1.46    1.21
  Log(Total Assets at IPO)                 0.76    1.38
  Technology-Firm                          0.61    0.53
  Ownership by Management                 -0.14   -1.83 *
  Percent of Board Outside Directors       0.00    0.32
Market Conditions
  First Day Underpricing                  -0.02   -0.42
  Bear Market Investment                   1.94    1.73 *
  Bull Market IPO                         -0.57   -0.46
  Constant                               -13.00   -1.25
Observations                                    165
Adi. [R.sup.2]                                  0.11

                                          First Share Sale
                                                CARs

                                                (II)

Variables                                Coef.     t-stat

Buyout Fund Variables
  Log (Historic IRR)                      -0.22   -0.05
  Log(Historic Capital Raising)           -0.69   -1.34
  Log(Fund Vintage Age)                   -1.34   -1.61
  Log(Firm Vintage Age)                    1.17    1.61
  Deal Syndication                        -2.20   -1.60
  Buyout Ownership                         0.01    0.28
  Investment Returned at IPO
  Dividend Recap                           0.18    0.12
  Dividend over Revenues                   2.83    2.89 ***
  Investment Returned at IPO x
    Return on Assets at IPO
  Chairman from Buyout Fund               -1.46   -1.19
  Percent Board Buyout Directors          -0.01   -0.36
  Management Fee over Revenues at IPO     -0.00   -1.75 *
Portfolio Company Variables
  M&A Deals                                2.00    1.90 *
  Leverage Reduction pre IPO              -0.00   -0.39
  Leverage at IPO                          0.00    0.24
  Return on Assets at IPO                 -0.00   -0.00
  Increase in Return on Assets             1.72    1.39
  Log(Total Assets at IPO)                 0.79    1.45
  Technology-Firm                          0.24    0.21
  Ownership by Management                 -0.14   -1.82 *
  Percent of Board Outside Directors       0.00    0.29
Market Conditions
  First Day Underpricing                  -0.02   -0.51
  Bear Market Investment                   2.18    1.88 *
  Bull Market IPO                         -0.72   -0.56
  Constant                               -12.64   -1.22
Observations                                    165
Adi. [R.sup.2]                                  0.09

                                          First Share Sale
                                                CARs

                                               (III)

Variables                                Coef.     t-stat

Buyout Fund Variables
  Log (Historic IRR)                      -1.20   -0.32
  Log(Historic Capital Raising)           -0.61   -1.16
  Log(Fund Vintage Age)                   -1.23   -1.52
  Log(Firm Vintage Age)                    1.06    1.44
  Deal Syndication                        -2.37   -1.74 *
  Buyout Ownership                         0.01    0.41
  Investment Returned at IPO               2.96    2.42 **
  Dividend Recap                          -0.04   -0.03
  Dividend over Revenues
  Investment Returned at IPO x             0.09    0.73
    Return on Assets at IPO
  Chairman from Buyout Fund               -1.99   -1.70 *
  Percent Board Buyout Directors          -0.01   -0.22
  Management Fee over Revenues at IPO     -0.00   -1.90 *
Portfolio Company Variables
  M&A Deals                                2.00    1.98 **
  Leverage Reduction pre IPO              -0.00   -0.60
  Leverage at IPO                          0.00    0.07
  Return on Assets at IPO                 -0.04   -0.60
  Increase in Return on Assets             1.36    1.13
  Log(Total Assets at IPO)                 0.81    1.44
  Technology-Firm                          0.60    0.52
  Ownership by Management                 -0.14   -1.80 *
  Percent of Board Outside Directors       0.00    0.33
Market Conditions
  First Day Underpricing                  -0.02   -0.53
  Bear Market Investment                   2.03    1.78 *
  Bull Market IPO                         -0.67   -0.53
  Constant                               -13.54   -1.29
Observations                                    165
Adi. [R.sup.2]                                  0.10

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.
COPYRIGHT 2015 Financial Management Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Furth, Sven; Rauch, Christian
Publication:Financial Management
Article Type:Report
Geographic Code:1USA
Date:Dec 22, 2015
Words:20193
Previous Article:Liquidity benefits from IPO underpricing: ownership dispersion or information effect.
Next Article:The volatility of bid-ask spreads.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters