Catching Up or Agency Problem: Dynamics of Post-IPO Executive Compensation. James Ang Florida State University

Transcription

1 Catching Up or Agency Problem: Dynamics of Post-IPO Executive Compensation James Ang Florida State University Ansley Chua* University of Texas Pan American June 29, 2010 Abstract Utilizing the IPO event, we empirically test two issues in executive compensation. First, we investigate a novel area; the implicit promise of increased compensation based on previous pay sacrifice. Second, we estimate the amount of free cash flow paid to executives. We find one year s worth of previous pay sacrifice is made up in the first year after IPO. And, free cash flow account for $312,000 of first year s pay increase. The long term cost to shareholders is great. Each percent free cash flow distributed as executive pay is associated with 1.2% decline in market value within two years. Keywords: Initial Public Offerings, Executive Compensation, Agency, Corporate Governance JEL Codes: G32, J33, M13, M52 *Corresponding Author: chuaac@utpa.edu Phone:

2 I. Introduction In the year of IPO, we find CEO median pay increase by 59.3% and median pay for the rest of top executive also increase by 37.2%. The amount and timing of this increase in pay makes the IPO event as a natural experiment to study two issues in executive pay. The first, an investigation of implicit promise, is new. Entrepreneurs and managers hired into start up firms that are cash poor received two types of non cash pay: equity or options that could be converted to equity, and delayed pay that may be contingent on firm receiving additional cash funding such as the IPO event. The later is an implicit promise of increase in pay. This type of pay has not been analyzed before, in particular, new questions of interest includes: Does implicit promise of future pay to executives exist? How much of the promised amount is paid at IPO? How do the answers to these questions differ for CEO versus the rest of the top executives? The second issue, extracting pay from free cash flow, is known but never been satisfactorily calibrated, except for few special cases. The main issue is a measurement problem what and how to measure free cash flow from a going concern firm? IPO also provides the natural experiment as funds from IPO are so large relative to what the firm could expend, free cash flow is easy to identify and measure. Here, we make use of this empirical convenience to calculate extra pay executives received from free cash flows and the equity value dissipated by these cash wasting firms. We quantify the level of compensation increases into two categories: Promise based on the implicit promise of firms to the executives for future pay and Agency based on the executives ability to expropriate funds. We contribute to the overall understanding about compensation by documenting that both the implicit promise of

3 post-ipo pay and agency variables affect compensation levels. We find that 30% of the compensation increases is due to the principal/agent factors. We further contribute to the literature by documenting that this compensation due to agency is a significant economic cost to investors. We find that for every 1% of IPO funds misappropriated for compensation leads to 1.2% lower two year post-ipo returns. This is in contrast to pay increases based on the implicit promise. Compensation for the implicit promise of pay does not affect the two year post-ipo returns of firms. The question of executive compensation has been of interest since Jensen and Meckling (1976) formalized the relationship of the separation of ownership and control. The theoretical framework that agents work not to maximize shareholder wealth, rather to maximize their own utility is well understood. However, what is not well understood is how this separation between principal and agent affects the compensation levels of the managers. Jensen and Murphy (1990) find that though there is a direct relationship between executive compensation and firm performance, the CEO s wealth only marginally increases and a large portion of the wealth increase comes from the executive s equity ownership. They find that the majority of the compensation is independent of firm performance. The author does not, however contend that the compensation package is unjustified, but rather may be due to the labor forces of the executive managers. Hall and Liebman (1998) find that the relationship between pay and performance is higher because of the increase in use of equity options as compensation. Core, Holthausen and Larcker (1999) find that corporate governance is important in determining compensation levels. They find that CEOs receive greater compensation in

4 firms with less effective governance, implying that agency issues play a role in the determination of compensation. Blanchard, Lopez-de-Silanes and Shleifer (1994) find large free cash flows from lawsuit windfalls generally lead to greater agency issues. However, their finding is limited by their small sample set. The IPO event offers us a unique opportunity to test the implication to the principal/agent separation of large windfalls. The Blanchard, et al. study is limited to 11 firms and the median cash flow to the firm is $8.9M. The IPO event in our sample, by comparison, averages IPOs per year, or a total of 637 firms with a median free cash flow to the firm of $87.5M. This allows us a greater opportunity to examine the effects of free cash flows on the level of executive compensation. During the IPO the firm receives a large influx of capital from outside investors and these investments are earmarked for value building investments for the firm. However, we find that the median increase in compensation from pre and post IPO is $105,683 per executive. What drives this change in compensation and what the effect of these compensations to the long term profitability of the firm is the purpose of this study. We expect executives in pre IPO firms, anticipating a large payoff for taking greater risk and entrepreneurship, are willing to postpone part of their normal compensation. Pinfold (2001) finds that entrepreneurs starting their new venture, on average, expect 16% annual dividends from their investment and a six fold increase of the value of their investment in five years (47.6% annualized return). However, Hamilton (2000) finds that the actual entrepreneur s pay is lower than the pay for a matching wage earner. Additionally anecdotally, we know that entrepreneurs sacrifice for the success of their firms, such as the owner and president of Primary Freight Services, Inc., have both

5 stopped receiving pay during the current economic downturn (Zwahlen, 2009). Therefore if the entrepreneurs are rational, we would expect that if firms become successful and undergo an IPO their ex-ante expectations and previous sacrifices must be compensated. We test the promise and agency components by separating variables previously shown to affect compensation. For the promise variables, we include variables that proxy for the implicit promise of pay for executives: Founder, as Kihlstrom and Laffont (1979) shows that the firm founders generally undertake significant risk and sacrifice in order to undertake a new venture; Tenure, because the length of time the executive has been with the firm measures the duration of their sacrifice prior to IPO; The amount of underpayment as compared to peer firms, based on industry and market size, because this approximates how much more the executive should be paid; And underpayment interacted with tenure, because this estimates the total amount of underpayment for each executive s pre-ipo tenure. Additionally, we included the log of market value one year after IPO as Smith and Watts (1992) documents that the level of executive compensation increases with the size of the firm. Because Jensen and Murphy (1990) showed that the change in firm value related to the executives compensation, we also included the change in firm value from IPO issue date to one year after. For the agency variables we include the variables that have been shown to affect compensation due to the separation of principal and agent as documented by Beatty and Zajac (1994). We included executives characteristics such as: Board membership status, chairmanship status and CEO status. We also included firm characteristics such as: Number of board members, venture capital backing, percentage of executive board member and the free cash flow. Free cash flow is defined as; funds not used in the first

6 year to increase investments in either: capital expenditures, research and development and advertising. We utilize this measure because it accurately measures how much of the IPO fund is not used and therefore can easily be appropriated to compensation. We postulate that post IPO compensation is affected by both previous sacrifice and agency issues, specifically, we hypothesize that: 1) Catching up hypothesis is supported. That is, variables that support the sacrifice will be positively correlated to the increase in compensation. 2) Agency motivated hypothesis is supported. The variables that support the agency theory will also correlate to the increase in compensation. i. Chairmanship, CEO, Board status, percentage of executive board members and free cash flows, factors which increases the agency issue, will be positively correlated to the increase in compensation ii. Venture capital backing and total number of board members, factors that mitigate the agency issue, will be negatively correlated to the increase in compensation 3) Misappropriated funds will be funds that cannot be used for value building investments. Therefore the IPO firm will realize lower long term growth. This study finds that the amount of underpayment significantly drives the underpaid managers increase in compensation. This finding shows that a significant portion of the increase in executive pay was due to the firms implicit promise pre-ipo to the executives and the increase in difficulty of running a public firm. We also consistently find that free cash flow is strongly correlated to the increase of executive compensation. We find for every one million dollars of free cash flow, corresponds to an increase of $2,100 for each underpaid executive. Though this sounds trivial, the mean amount of free cash flow from IPO funds in the first year is $155M, which translates to an economically significant increase of $325,500 per executive. For those executives whose pay is already above the matched peer pay, the increase in pay is less pronounce, but still receive an economically significant $1,600 per $1M of free cash flow, translating to an increase in compensation of $248,000 per executive of which are already

7 overcompensated. CEOs, for their roles, receive an estimated $614,000 increase in compensation. Additionally, we find that total ownership strongly and positively correlates with the level of compensation increase. As expected, the increased ownership stake creates a larger agency issue because of the ability of the manager to expropriate funds. We find for every $1M of pre-ipo ownership, the executive s post-ipo compensation increases by $3,800. With an average ownership of $18M, this translates to an increase of $68,000 post-ipo. These findings support that the changes in compensation are also attributable to the separation of principal and agent. These finding may not be relevant to investors if there were no economic effects to their investments. However, we document that the percentage of IPO funds that firms used to increase the unjustified component of the compensation strongly correlates to the returns to the outside investor. We find that for every 1% of IPO funds used for the agency factors, leads to 1.2% decrease in the two year post-ipo returns. However, we find that pay increases based on the firms implicit promise does not affect the firms two year post-ipo returns. II. Data The data used for this study comes from firms undergoing the Initial Public Offerings from 2002 to 2006 as reported by SDC. Because of the Execucomp only tracks the top 1500 largest firms and most of the IPO firms are relatively small, the pre-ipo executive compensation data was hand collected. Firms are required to disclose in the prospectuses the five highest compensated executives of firm earning more than $100,000. Additionally, we collected executives characteristics: Tenure with the firm, role as CEO, role as Chairman, role as a founder. We also collected firm characteristics,

8 such as: Number of board members, venture capital backing and number of executive board members. REITs, closed-end funds and foreign firms were excluded from the sample. Based on these requirements we collected 3011 executives and their compensations from 637 firms. Post-IPO executive compensation was also hand collected from the firms proxy statement the next fiscal year after IPO. We collected 3034 executives and their compensations from 622 firms. Because there is a significant amount of executives who did not transition from pre-ipo to post-ipo and a significant of new executives on the sample, based on matching by executive name, we were able to obtain 2467 executives with compensation pre and post IPO representing 615 firms. In order to obtain balance sheet information, the sample was merged with Compustat. Because quarterly data is unavailable for the pre-ipo period, annual Compustat was used to obtain the pre-ipo levels of investments in capital expenditures, research and development and advertising. Quarterly Compustat was utilized to obtain the 4 quarter post IPO levels of expenditures in the three respective investments. In order to obtain peer matched compensation levels, Execucomp was utilized. Table I shows the summary statistics of the sample of executives and Table II shows the correlations between the variables. The correlations among the promise variables are weak. The correlations for the agency variables are likewise weak, except for the relationships between chairmanship, board members and CEO, which is expected. III. Methodology Because the aim of the study is to decompose the increase in compensation to the promise and agency component, we utilized firm and executive characteristics that affect

9 either component. From the prospectus and proxy, the pre-ipo level of compensation is computed as: (1) Stock value is either the company assigned stock value as given in the prospectus or proxy, or as the number of shares granted * IPO offer price. Because the values of the options are not given until 2006 we must estimate the pre-ipo value of the option grants. In order to estimate the value, we assumed that the strike price for the pre-ipo options is 20% of the offer price. We assumed that the executives knew of the impending public offering and expected offer price, therefore we used the offer price as the underlying asset price. In order to estimate the volatility of the underlying asset, we determined the one year prior s monthly returns of all listed stocks in the firms industry based on Fama and French s 49 industry breakdown. We assumed the standard 10 year employee stock option to expiration. Because of the uncertainty of the firm actually going public, we discounted the value of the pre-ipo option by 25% based on the probability a firm will go public in this late stage consistent with Reid, O Connor and Shapiro (2006). Post-IPO option values were assumed to be at the money, with 10 years left to expiration. Volatility is calculated the similarly to the pre-ipo options. Change in compensation is calculated as the difference between the pre-ipo compensation and the post-ipo compensation. Pre-IPO executive characteristics were also obtained from the prospectus. A dummy variable was assigned to the executive to

10 individually denote whether the executive is: Chairman of the Board, CEO, founder, board member and the backed by venture capitalists. The tenure of the executive was calculated based on the year of the issue minus the year the executive joined the firm. Total number of board members during IPO was also obtained. Insider Board is calculated as a percentage of executives who are board members / total board members. Log of Size is the log of the market value one year after IPO. Change in Firm Value is the total market value change from IPO issuance to the end of the first calendar year. Because IPO investors will generally only finance companies with growth potential and not fund distressed companies trying to stay afloat, we can presume that funds generated from IPO are for the purposes of increasing the investment levels in wealth generating investments in the three investment categories. We assume that the pre-ipo levels of investments are the levels that the company has been investing with their own funds. Thus the post-ipo increase in levels of investments is what the IPO funds are utilized for and this is the point of interest of this study. The increase in R&D, capital expenditure and advertising level is calculated as the difference between the current quarter s expenditure and the pre-ipo level. This estimates the amount of IPO funds that is utilized for each of these investment types per quarter. (2) (3) (4) We then summed the incremental increases in each of the investment categories to determine the amount of incremental expenditures for each quarter.

11 (5) In order to determine the amount of IPO funds not spent in the first year, we summed the incremental increases in the four quarters after IPO and subtracted the sum from the funds generated in the primary market. We define this underutilization as the free cash flow from IPO funds. (6) In order to estimate the level of underpayment for each executive, we utilized compensation levels from Execucomp. We matched the firms with the Execucomp based on industry and size. We grouped Execucomp firms and our sample firms into industries based on 49 industry breakdowns according to Fama and French. Following Loughran and Ritter (1995) and Ritter (1991), we utilized the market value as the basis for size matching. We matched firms based on the market value at the end of the IPO fiscal year to the fiscal year end of other same industry firms. However, unlike the prior two papers that match firms one to one, we matched the IPO firms with firms where the end of fiscal year market value is +/- 30% of the end of fiscal year market value of the IPO firm. We ranked each executive based on their compensation levels and matched based on the compensation ranking. The highest compensated executive was matched with the highest compensated executive in the matched firms. The median level of compensation for the matched firm was determined and the level of underpayment was computed as the difference between the IPO executive and the median compensation levels.

12 (7) Because we do not have the yearly compensation levels for entire tenure of an executive, we estimate the overall sacrifice of an executive as the level of underpayment at the time of IPO multiplied by the tenure of the executive. This gives us an estimate of their entire sacrifice when working with the company. In order to control for firm size, we found the natural logarithm of the firms market value one year from IPO issuance. The market value is calculated as the number of shares outstanding * equity price the last trading day on or prior to the anniversary of IPO issuance. The change in firm value is calculated as the difference between the market value one year from IPO issuance and market value at the day of IPO. Market value of firm during IPO is calculated as the number of shares outstanding * offer price of the IPO. In order to determine the overall effects of each component on the change in compensation we regress: (8) Because one of our measures of sacrifice is the level of payment relative to a median industry pay, there are some pre-ipo executives where their level of

13 compensation is higher than the industry median. These executives are, by our definition did not undertake sacrifice on their pay. In order to test the differential effects of the variables we included interaction terms. We regress the following relationship: (9) IV. Results Table III documents the relationship of the change in compensation with the promise and agency components of the increase in compensation. In the first model we find that the level of underpayment relative to the median to be a significant driver of the change in compensation. This supports our hypothesis that previous executive sacrifices are compensated post-ipo. The coefficient is close to one for underpayment and implies that there is a strong adjustment to the medians for the post-ipo compensation levels. We also find, consistent with Smith and Watts (1992), that log of firm size to be a significant driver of compensation. This also supports our hypothesis that the largeness of the firm and managing a larger firm will be compensated during post-ipo. Contrary to the Jensen and Murphy (1990) we find a negative relationship between the change in shareholder wealth and change in compensation through IPO. This relationship may need further examination but, this difference may be due to: 1) IPO firms are well understood to

14 decline in absolute value, and 2) the post-ipo pay of executives is usually determined pre-ipo with the help of the venture capitalists and investment bankers. The coefficient for ownership is positive and highly significant, implying that the increase in ownership leads to higher increases in pay. Based on the compensation for sacrifice theory, a priori, we expect a negative coefficient because the previous sacrifices would be compensated in the ownership of equity. But this relationship seems to support the conflict between principal and agent in regard to the effects of ownership on the executives compensation. We also find that CEO dummy and free cash flows are both strongly significant in driving the increase in compensation. We find support that the separation of principal and agent allow the executives to expropriate funds from the firm. IV. A. Under and Over Compensated Contrary to expectations that all pre IPO executives take sacrifices in pay, we find a significant portion of these executives have compensation levels greater than the matched pay of median firms. Of the 2,475 executives with match pay information 902 had pay higher than the matched pay. Under the sacrifice hypothesis, these executives did not sacrifice prior to the IPO. Therefore the drivers of the changes in compensation should be different for the two groups of executives, those that sacrificed and those that did not. In the second model of Table III, we included the interaction term for the FCF and Underpayment. We find that there is a significant difference in the way overpaid and underpaid executives are compensated. We find that the underpaid executives expropriate a larger amount of funds from the free cash flows. While the overpaid executives receive $1,594 for every $1M of free cash flow, we find that the underpaid executives receive $2,584. We also find that the adjustment to the median pay is lower for the underpaid

15 executives. We find that the post-ipo pay only partially adjusts by 73.4%, but the overpaid executives have a tendency to fully adjust to the median. Utilizing the coefficients as point estimates the average contribution of the justifiable variables is calculated as: And the contribution of the unjustified variables is calculated as: (10) (11) Figure I illustrates the contribution to the change in compensation by the justified and unjustified. We find that an average of $1.6M is attributable to the sacrifice and managing a larger firm and $709K or approximately 30% is attributable to the separation of principal and agent. We further investigate if the pre-ipo rank of the executives affects the change in compensation of the executives. In Table IV we separate the sample by the ranking of the executives. We classified the top two pre-ipo compensated executives per firm, usually the CEO and CFO, as the Top Rank executives. We find similar pattern for drivers of their compensation with the full sample findings. We find that level of underpayment is still consistently compensated post-ipo. We find that the free cash flow is still an important factor in the change in compensation. However comparing the two subsamples,

16 we see that the increase in pay derived from the free cash flow is more than three times larger for the higher ranked executives. This highlights the agency issues because higher ranking executives are more able to capture the unutilized funds that should be earmarked for value building investments but otherwise appropriate it for their pay. We also find that lower ranking executives derive more pay from ownership. This further supports the agency finding because for the lower ranking executives, their source of control of their pay must derive from their ownership stake. Therefore having a larger ownership stake allows them to increase their post-ipo pay. These findings further support that the principal / agent separation is an integral driver in the compensation of executives. Because we hypothesize that the greater control the executives have over their firm the easier it would be for them to expropriate funds, we tested the differential effects of being a large or small owner. We sorted the full sample of executives by the value of stocks where they were assigned as the beneficial owners. We grouped the executives above the median as the high ownership and the executives below the median as low ownership. As expected, we find that executives with higher ownership see a greater increase in their pay due to the free cash flows. We find this relationship especially true for those executives with high ownership and were previously underpaid. We find for the underpaid executives with high ownership were able to receive $3,500 per $1M of free cash flow. This translates to an increase in compensation of $542,500 for each executive. This further supports that agency issues are significant determinants on the increase in executive compensations. VI. B. Post-IPO Returns

17 We document that previous sacrifices and managing a larger firm was compensated post-ipo. We also document that executives also expropriate funds because of the separation of principal and agent. However, for this to be relevant to the IPO investors, we need to understand the implication of the expropriation of funds to the investment returns. To test the effects of the compensations attributable to principal/agent, we test the effects on the two year post-ipo returns of these expenditures. We utilize the predicted compensation for each executive from the sacrifice variables as the amount of compensation for the executives sacrifice and greater responsibility. The predicted compensation for each executive is estimated as: (12) We estimated the value of each executive s compensation associated with the firms principal/agent as the difference between the actual change in compensation and the predicted amount based on the justified variables: (13) We estimated the per firm total amount of funds expropriated because of the separation of principal and agent by summing each firm s executives estimated Incentive pay. (14) The per firm percentage of IPO funds is computed as: (15)

18 We utilized four measures of the two year post-ipo buy and hold returns. We used the raw, CRSP value weighted adjusted abnormal returns, CRSP equally weighted adjusted abnormal returns and size/industry adjusted abnormal returns. As standard in most finance literature, we exclude the first day underpricing in computing the buy and hold abnormal returns. In order to calculate the abnormal returns we computed: (16) In order to calculate the size/industry return, we sorted all stocks based on the Fama and French 49 industries. We then computed the market value of each stock at each year end. We matched each IPO firms to those firms in the same industry and whose market capitalization at year end prior to the IPO is +/- 30% from the first day closing of the IPO firm. We utilize the mean returns of the matched firms as the expected returns. Panel A of Table VI documents our finding. For all four of the measures, we find there is a very statistically and economically significant relationship between the percentage of funds used and the two year post-ipo returns of the firms. We find that for every 1% of IPO funds used for the unjustified compensation leads to a 1.2% decrease in the two year returns. We tested to relationship to see if the percentage of funds used for justified compensation also leads to a decrease in two year post-ipo returns. We estimated the value of each executive s compensation associated with the firms sacrifice as the difference between the actual change in compensation and the predicted amount based on the unjustified variables: (17) Where the predicted compensation based on the principal/agent conflict is:

19 AgencyComp XIV * i X BoardDummy VII FreeCashFlow i TotalOwnership f XI TotalBoard XVI i f VIII XIII FreeCashFlow ChairDummy f PercentInsider i f IX * UnderpaidDummy CEODummy i i (18) We estimated the total amount of funds expropriated because of the separation of principal and agent by summing each firm s executives estimated Incentive pay. I i * FirmTotal _ Sacrifice f Sacrifice i, f (19) 1 The per firm percentage of IPO funds is computed as: FirmTotal _ Sacrifice f Agency _ Per _ Sacrifice f (20) IPOproceeds f Panel B of Table VI documents our findings. We find that the percentage of IPO funds used to compensate executives does not have an impact of the two year post-ipo returns in the four returns measures. Therefore it seems that the compensation for past sacrifices does not affect the future returns of the firm. These findings suggest that the sacrifice compensation and compensation for managing a larger firm is compensation is justified and doesn t decrease the firm value and that pre-ipo sacrifices of executives are compensated but not at the expense of the new shareholders. However, the amount of post-ipo compensation that is derived because of the separation of principal and agent is unjustified and is value destroying for the equity holders. V. Robustness Table VII documents some robustness checks. Instead of examining the free cash flows, we instead used the total amount funds raised during IPO. We find the same relationship. We find, consistent with the agency hypothesis, that for every $1M of IPO fund generated, each executive receives on average $980. With an IPO proceeds sample

20 mean of $187M, each executive receives an average $183,000 increase in compensation post-ipo. Examining Table VIII we find the same relationship with the two year post- IPO returns. We find that every 1% of IPO funds used for the unjustified variables correlates with a decrease of 1.36% in the matched firm abnormal returns. VI. Conclusion In this paper we study a previously unexamined area: Whether firms implicitly promise to repay the pre-ipo sacrifices of their executives. We find strong and consistent empirical support that post-ipo, pre-ipo sacrifices of executives are repaid. In the first year post IPO one year s worth of sacrifice is repaid. Additionally, we support previous studies that agency issues affect compensations. Free cash flow is consistently significant in contributing to the increase of executive pay. Beneficial ownership of equity also allows the executives to increase their compensation. We show that repayment for pre- IPO sacrifices is not value destroying for the firm but compensation increases because of the principal/agent separation is. For every 1% of free cash flows that is expropriated because of the separation of principal and agent lead to a 1.2% decrease in the two-year post-ipo returns.

21 Table I: Sample Descriptive Statistics Panel A shows the descriptive statistics of the executives. Pre-IPO compensation is the total compensation of the executives. Post-IPO compensation is the total compensation of the executives. Change in pay is the difference in compensation for the executives where compensation levels for both pre and post IPO were available. CEO is the ratio of the sample designated as CEO. Chairman is the ratio of the sample designated as Chairman. Founder is the ratio of the sample designated as founders. Board is the ratio of sample designated as board members. Tenure is the number of years an executive has been with the firm, calculated as the year of the IPO minus the year the executive joined the firm. Panel B shows the descriptive statistics of the firm. Underpayment is the difference between the executives' pre-ipo pay and median pay of similar ranking executive matched by industry and size. Total Board is the number of board members of the firm during the IPO. Free Cash Flow is the amount of funds, in millions, not utilized in the first year after IPO. Log of Size is the natural log of the market value of the firm one year after IPO. VC Backed is the ratio of firms backed by venture capitalists. Panel C shows the breakdown of IPO based on year. Panel A: Descriptive Statistics - Executives N Mean Median Std Dev Pre-IPO Compensation 2959 $ 1,388,408 $ 482,326 $ 4,156,774 Post-IPO Compensation 2458 $ 1,556,496 $ 715,215 $ 2,785,470 Change in Pay 2458 $ 65,304 $ 105,683 $ 4,487,902 Founder Tenure Underpayment $428,891 $148,439 $ 4,436,403 Ownership Value ($M) 2309 $18.05 $2.36 $ Chairman CEO Board Panel B: Descriptive Statistics - Firms N Mean Median Std Dev Log of Size Total Board VC Backed Percent Insider % 22.2% 22.4% Free Cash Flows ($M) 559 $ $84.69 $ Value Change 607 $312,292 $62,939 $1,738,182 Panel C: Number of IPOs per Year Year N Percentage % % % % % Total %

22 Table II: Correlations Panel A shows the correlations between the factors which may justifiably increase executive compensation, including ownership. Panel B shows the correlatins between factos which unjustifiablity may increase executive compesation, including ownership. Founder Dummy Panel A: Correlations - Justifiable Variables Tenure Underpayment Ownership Ln Size (0.000) (0.412) (0.000) (0.000) Tenure (0.462) (0.002) (0.000) Underpayment (0.000) (0.000) Ownership Panel B: Correlations - Unjustifiable Variables CEO Board Total Board VC % Insider Underutilization Ownership Chair Dummy (0.000) (0.000) (0.343) (0.078) (0.215) (0.241) (0.000) CEO Dummy (0.000) (0.000) (0.271) (0.331) (0.709) (0.360) (0.000) Board Dummy (0.022) (0.110) (0.000) (0.026) (0.000) Total Board (0.403) (0.000) (0.000) (0.040) VC Backed (0.000) (0.000) (0.219) Percent Insider (0.001) (0.182) Free Cash Flow (0.027)

23 Table III Relationship of Change in Executive Compensation and Independent Variables Table III shows the relationship between the increase in the executive compensation and the dependent variables. Chair Dummy is 1 if the executive is the Chairman of the Board. CEO Dummy is 1 if the executive is the CEO. Board Dummy is 1 if the executive is a member of the board. Total Board is the number of board members. Percent insider is the percentage of board members that are executives of the firm. Free Cash Flow is the amount in Millions of IPO not used in the first year after IPO. Founder Dummy is 1 if the executive is a founder. Tenure is the number of years the executive has been with the firm, in years. Underpayment is the difference between the industry, size and rank matched median pay and the executive pay. The dependent variable is the change in compensation pre and post IPO. The number in parenthesis is the significance level. * 5%, ** 1% and *** <1% Model 1 Model 2 Founder Dummy * * (0.029) (0.035) Tenure (0.842) (0.810) Underpayment *** *** (0.000) (0.000) Underpay*Tenure *** *** (0.000) (0.000) Log of Size ** ** (0.002) (0.003) Change in Market Value ** * (0.002) (0.013) Ownership *** *** (0.000) (0.000) Chair Dummy (0.859) (0.755) CEO Dummy * ** (0.012) (0.002) Board Dummy * (0.055) (0.042) Total Board (0.646) (0.597)

24 Table III - Continued VC Backed (0.356) (0.198) Percent Insider (0.709) (0.746) Free Cash Flow *** *** (0.000) (0.000) Underpaid Dummy * (0.014) FCF*UnderpaidDummy * (0.014) Underpay*UnderpayDummy *** (0.000) Constant ** ** (0.002) (0.008) Observations R

25 Table IV Relationship of Change in Executive Compensation and Independent Variables - by Rank Table III shows the relationship between the increase in the executive compensation and the dependent variables. Chair Dummy is 1 if the executive is the Chairman of the Board. CEO Dummy is 1 if the executive is the CEO. Board Dummy is 1 if the executive is a member of the board. Total Board is the number of board members. Percent insider is the percentage of board members that are executives of the firm. Free Cash Flow is the amount in Millions of IPO not used in the first year after IPO. Founder Dummy is 1 if the executive is a founder. Tenure is the number of years the executive has been with the firm, in years. Underpayment is the difference between the industry, size and rank matched median pay and the executive pay. The dependent variable is the change in compensation pre and post IPO. The number in parenthesis is the significance level. * 5%, ** 1% and *** <1% Top Rank Bottom Rank Underpaid Overpaid Founder Dummy ** (0.317) (0.118) (0.001) (0.843) Tenure (0.926) (0.277) (0.929) (0.447) Underpayment *** *** *** *** (0.000) (0.000) (0.000) (0.000) Underpay*Tenure *** ** *** (0.000) (0.004) (0.065) (0.000) Log of Size * *** (0.014) (0.104) (0.000) (0.789) Change in Market Value *** (0.052) (0.129) (0.001) (0.814) Ownership *** *** (0.533) (0.000) (0.000) (0.778) Chair Dummy * (0.457) (0.037) (0.765) (0.780) CEO Dummy * ** (0.242) (0.022) (0.002) (0.066) Board Dummy * * (0.014) (0.931) (0.021) (0.673)

26 Table IV - Continued Total Board * (0.509) (0.848) (0.672) (0.026) VC Backed * (0.231) (0.145) (0.781) (0.028) Percent Insider *** (0.833) (0.242) (0.179) (0.000) Free Cash Flow *** ** *** (0.000) (0.002) (0.000) (0.320) Underpaid Dummy ** (0.261) (0.007) FCF*UnderpaidDummy (0.203) (0.067) Underpay*UnderpaidDummy ** (0.007) (0.240) Top Rank (0.055) (0.151) Underpay*Top Rank (0.063) (0.982) FCF*Top Rank *** ** (0.000) (0.003) Constant ** *** (0.010) (0.392) (0.000) (0.409) Observations R

27 Table V Relationship of Change in Executive Compensation and Independent Variables- by Ownership Table III shows the relationship between the increase in the executive compensation and the dependent variables. Chair Dummy is 1 if the executive is the Chairman of the Board. CEO Dummy is 1 if the executive is the CEO. Board Dummy is 1 if the executive is a member of the board. Total Board is the number of board members. Percent insider is the percentage of board members that are executives of the firm. Free Cash Flow is the amount in Millions of IPO not used in the first year after IPO. Founder Dummy is 1 if the executive is a founder. Tenure is the number of years the executive has been with the firm, in years. Underpayment is the difference between the industry, size and rank matched median pay and the executive pay. The dependent variable is the change in compensation pre and post IPO. The number in parenthesis is the significance level. * 5%, ** 1% and *** <1% High Owndership Low Ownership Underpaid Overpaid Founder Dummy * ** (0.412) (0.024) (0.004) (0.777) Tenure (0.278) (0.278) (0.992) (0.463) Underpayment *** *** *** *** (0.000) (0.000) (0.000) (0.000) Underpay*Tenure *** *** *** (0.000) (0.000) (0.063) (0.000) Log of Size *** *** (0.194) (0.001) (0.000) (0.760) Change in Market Value *** (0.157) (0.079) (0.001) (0.942) Ownership *** *** (0.000) (0.711) (0.000) (0.610) Chair Dummy *** (0.263) (0.000) (0.891) (0.756) CEO Dummy ** ** (0.061) (0.004) (0.003) (0.051) Board Dummy * (0.092) (0.610) (0.011) (0.582)

28 Table V - Continued Total Board * (0.131) (0.263) (0.854) (0.022) VC Backed * (0.580) (0.526) (0.543) (0.036) Percent Insider *** (0.895) (0.899) (0.207) (0.000) Free Cash Flow *** *** *** (0.000) (0.000) (0.000) (0.110) Underpaid Dummy (0.131) (0.073) FCF*UnderpaidDummy * (0.011) (0.691) Underpay*UnderpayDummy ** (0.002) (0.219) High Ownership Dummy * (0.012) (0.258) Underpay*High Ownership (0.414) (0.521) FCF*High Ownership *** (0.000) (0.207) Constant ** *** (0.140) (0.006) (0.000) (0.351) Observations R

29 Table VI Two Year post-ipo Returns Table VI documents the two year post-ipo buy and hold returns. Raw is the raw returns two years post- IPO. Matched Firm is the abnormal returns based on the industry and size matched firms. Value Weighted is the CRSP value weighted adjusted returns and Equally Weighted is the CRSP equally weighted adjusted returns. In Panel A, Unjustified percent is the percentage of IPO funds used for unjustified compensation. In Panel B, Justified percent is the percentage of IPO funds used for justified compensation. Retention percentage is the percentage of pre-ipo executives retained post-ipo. Log of size is the log of the firms' market value. VC backed is a dummy variable to denote if the IPO firm is backed by venture capitalists. Percent Insider is the percentage of executives who are/were executives. Total board is the total size of the board. Panel A: Unjustified Compensation Raw Matched Firm Value Weighted Equally Weighted Unjustified Percent * ** ** ** (0.018) (0.009) (0.006) (0.007) Retention Percent (0.061) (0.276) (0.224) (0.266) Log of Size *** *** *** *** (0.000) (0.000) (0.000) (0.000) VC Backed (0.318) (0.305) (0.298) (0.257) Percent Insider * (0.123) (0.067) (0.112) (0.021) Total Board ** ** ** ** (0.003) (0.005) (0.005) (0.007) Constant *** *** *** *** (0.000) (0.000) (0.000) (0.000) Observations R

30 Panel B: Justified Compensation Raw Matched Firm Value Weighted Equally Weighted Justified Percent (0.738) (0.994) (0.983) (0.990) Retention Percent * (0.024) (0.122) (0.093) (0.116) Log of Size *** *** *** *** (0.000) (0.000) (0.000) (0.000) VC Backed (0.527) (0.503) (0.501) (0.446) Percent Insider * (0.134) (0.076) (0.125) (0.024) Total Board ** ** ** * (0.005) (0.007) (0.008) (0.011) Constant *** *** *** *** (0.000) (0.000) (0.000) (0.000) Observations R

31 Table VII Relationship of Change in Executive Compensation and IPO Proceeds Table III shows the relationship between the increase in the executive compensation and the dependent variables. Chair Dummy is 1 if the executive is the Chairman of the Board. CEO Dummy is 1 if the executive is the CEO. Board Dummy is 1 if the executive is a member of the board. Total Board is the number of board members. Percent insider is the percentage of board members that are executives of the firm. IPO Proceeds is the amount in Millions of funds raised during IPO. Founder Dummy is 1 if the executive is a founder. Tenure is the number of years the executive has been with the firm, in years. Underpayment is the difference between the industry, size and rank matched median pay and the executive pay. The dependent variable is the change in compensation pre and post IPO. The number in parenthesis is the significance level. * 5%, ** 1% and *** <1% Model 1 Model 2 Founder Dummy * (0.032) (0.067) Tenure (0.765) (0.754) Underpayment *** *** (0.000) (0.000) Underpay*Tenure *** *** (0.000) (0.000) Log of Size *** ** (0.001) (0.004) Change in Market Value * (0.023) (0.098) Ownership *** *** (0.000) (0.000) Chair Dummy (0.827) (0.750) CEO Dummy * ** (0.026) (0.003) Board Dummy * (0.059) (0.030)

32 Table VII - Continued Total Board (0.339) (0.382) VC Backed (0.235) (0.137) Percent Insider (0.808) (0.827) IPO Proceeds *** * (0.000) (0.018) Underpaid Dummy *** (0.000) FCF*UnderpaidDummy *** (0.000) Underpay*UnderpayDummy *** (0.000) Constant *** * (0.001) (0.018) Observations R

33 Table VIII Two Year post-ipo Returns Table VI documents the two year post-ipo buy and hold returns. Raw is the raw returns two years post- IPO. Matched Firm is the abnormal returns based on the industry and size matched firms. Value Weighted is the CRSP value weighted adjusted returns and Equally Weighted is the CRSP equally weighted adjusted returns. In Panel A, Unjustified percent is the percentage of IPO funds used for unjustified compensation. In Panel B, Justified percent is the percentage of IPO funds used for justified compensation. Retention percentage is the percentage of pre-ipo executives retained post-ipo. Log of size is the log of the firms' market value. VC backed is a dummy variable to denote if the IPO firm is backed by venture capitalists. Percent Insider is the percentage of executives who are/were executives. Total board is the total size of the board. Panel A: Unjustified Compensation Raw Matched Firm Value Weighted Equally Weighted Unjustified Percent * ** ** ** (0.014) (0.007) (0.004) (0.005) Retention Percent (0.066) (0.286) (0.238) (0.283) Log of Size *** *** *** *** (0.000) (0.000) (0.000) (0.000) VC Backed (0.283) (0.269) (0.260) (0.223) Percent Insider * (0.122) (0.067) (0.111) (0.021) Total Board ** ** ** ** (0.003) (0.005) (0.005) (0.007) Constant *** *** *** *** (0.000) (0.000) (0.000) (0.000) Observations R

34 Panel B: Justified Compensation Raw Matched Firm Value Weighted Equally Weighted Justified Percent (0.704) (0.975) (0.986) (0.954) Retention Percent * (0.024) (0.122) (0.093) (0.117) Log of Size *** *** *** *** (0.000) (0.000) (0.000) (0.000) VC Backed (0.531) (0.506) (0.505) (0.449) Percent Insider * (0.134) (0.076) (0.125) (0.024) Total Board ** ** ** * (0.005) (0.007) (0.008) (0.011) Constant *** *** *** *** (0.000) (0.000) (0.000) (0.000) Observations R

35 Figure I Compensation Increase Breakdown Figure I shows the breakdown of increases in compensation for all executives based on the justifiable and unjustifiable variables. The contributions are based on the variable mean multiplied by the variable coefficient, if the coefficient is significantly different from zero. The justifiable variables includes contributions from Founder, Tenure, Underpayment, Underpayment interacted with Tenure and Ownership if the coefficient is negative. The unjustifiable variables includes contributions from Chairmanship, CEO, Board membership, Total number of board members, Venture capital backed, Percent Insider, Free Cash Flows and Ownership if the ownership coefficient it positive. $1,800, $1,640, $1,600, $1,400, $1,200, $1,000, $800, $709, $600, $400, $200, $- Justified Unjustified

36 References Baker, Malcom and Paul A. Gompers, 2003, The Determinants of Board Structure at the Initial Public Offering, Journal of Law and Economics Bebchuk, Lucian Arye, 2003, Executive Compensation as an Agency Problem, NBER Working Paper, Bertrand, Marianne and Sendhil Mullainathan, 2000, Agents with and without Principals, The American Economic Review Blanchard, Olivier Jean, Florencio Lopez-de-SiIanes, and Andrei Shleifer, What do Firms do with Cash Windfalls?, Journal of Financial Economics Boot, Arnoud W.A., Radhakrishnan Gopalan and Anjan V. Thakor, 2008, Market Liquidity, Investor Participation, and Managerial Autonomy: Why Do Firms Go Private? The Journal of Finance Core John E., Robert W. Holthausen, and David F. Larcker, 1999, Corporate Governance, Chief Executive Officer Compensation, and Firm Performance, Journal of Financial Economics Fama, Eugene F., 1980, Agency Problems and Theory of the Firm, The Journal of Political Economy Fama, Eugene F. and Michael C. Jensen, 1983, Separation of Ownership and Control, Journal of Law and Economics Garen, John E., 1994, Executive Compensation and Principal-Agent Theory, The Journal of Political Economy He, Lerong and Conyon, Martin J., 2004, CEO Compensation, Incentives, and Governance in New Enterprise Firms, Journal of Derivatives Accounting Jensen, Michael C., 1986, Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers, The American Economic Review Jensen, Michael C. and Kevin J. Murphy, 1990, Performance Pay and Top-Management Incentives, The Journal of Political Economy Lang, Larry H.P., Rene M. Stulz and Ralph A. Walkling, 1991, A Test of the Free Cash Flow Hypothesis, Journal of Financial Economics Murphy, Kevin, 1998, Executive Compensation, Working Paper,