Insurer Information, Insiders and Initial Public Offering* M. Martin Boyer† First draft: September 6th, 2011 This draft: May 2012 Abstract Investors accept that investing in public firms is risky. Risk – and uncertainty – is even more pronounce when investing in new public firms since relatively little is known at the time of the initial public offering (IPO). One way to reduce the cost of such information asymmetry cost for new investors is through the underpricing of the security at the time of the IPO. The goal of this paper is to examine whether one source of unused and often unavailable information could be valuable to investors at the time of a firm’s IPO. In particular, I study the predictive power of an audit that insurance companies conduct when a firm seeks to subscribe to a Directors’ and officers’ liability insurance contract (D&O insurance) to protect its managers’ wealth in the event of a lawsuit against the corporation and its representatives. More precisely, I examine whether D&O insurance contracts offer a glimpse into a firm’s stock market return right after its IPO. I find that the more a corporation pays per unit of D&O insurance coverage at the time of the IPO – a signal of its risk of lawsuit – the lower and more volatile its first year return is. This effect disappears in the second year. One implication is that, because the insurer has a direct financial incentive to correctly assess the firm’s liability and governance risk, the information embedded in a D&O insurance contract is valuable to investors since it is unbiased. This information appears to be as valuable, if not more, than any debt rating. JEL Classification: G34, G22, J44, G32. Keywords: Corporate governance, Directors’ and officers’ insurance, Rate‐on‐line. *
I thank Léa Stern for very helpful discussions on a related paper. I also thank Mark Huson and seminar participants at the University of Alberta, at Copenhagen Business School and at the University of Wisconsin for very useful comments. The paper also benefited from the able research assistance of Wan Wang and Mathieu Leclerc in the collection of data. The financial support of the Social Science and Humanities Research Council is gratefully acknowledged. †
CEFA Professor of Finance and Insurance, Département de la finance, HEC Montréal (Université de Montréal), 3000, chemin de la Côte‐Ste‐Catherine, Montréal, QC H3T2A7; martin.boyer@hec.ca Insurer Information, Insiders and Initial Public Offering Abstract: Investors accept that investing in public firms is risky. Risk – and uncertainty – is even more pronounce when investing in new public firms since relatively little is known at the time of the initial public offering (IPO). One way to reduce the cost of such information asymmetry cost for new investors is through the underpricing of the security at the time of the IPO. The goal of this paper is to examine whether one source of unused and often unavailable information could be valuable to investors at the time of a firm’s IPO. In particular, I study the predictive power of an audit that insurance companies conduct when a firm seeks to subscribe to a Directors’ and officers’ liability insurance contract (D&O insurance) to protect its managers’ wealth in the event of a lawsuit against the corporation and its representatives. More precisely, I examine whether D&O insurance contracts offer a glimpse into a firm’s stock market return right after its IPO. I find that the more a corporation pays per unit of D&O insurance coverage at the time of the IPO – a signal of its risk of lawsuit – the lower and more volatile its first year return is. This effect disappears in the second year. One implication is that, because the insurer has a direct financial incentive to correctly assess the firm’s liability and governance risk, the information embedded in a D&O insurance contract is valuable to investors since it is unbiased. This information appears to be as valuable, if not more, than any debt rating. JEL Classification: G34, G22, G24. Keywords: Corporate governance, D&O insurance, Rate‐on‐line. ii 1. Introduction Litigation against a firm’s directors and officers is one of the unfortunate consequences of governance risk. As representatives of the corporation, corporate directors and officers are personally liable1 for damages caused by their corporation’s actions or absence thereof. Being exposed to such an important liability risk induces directors and officers to request insurance coverage to protect their personal wealth in the event a lawsuit is brought against them as representatives of the corporation. This insurance, known as directors' and officers' liability insurance (D&O insurance hereinafter), is extremely common in public corporations.2 Since insurers have the appropriate incentives to correctly measure the potential cost of litigation against the insured firm’s directors and officers, the structure of a D&O insurance contract should then be an unbiased measure of a firm’s governance risk. Baker and Griffith (2008) mention that the main risk directors and officers face is associated with shareholder litigation, and the major liability exposure is securities litigation on the basis of misrepresentation (see also the different Towers‐
Watson surveys3). McTier and Wald (2011) even argue that the threat of a class action lawsuit prompts firms to invest significantly in reducing the potential cost of agency problems. Provided insurers use the correct technology to transform a firm’s observable characteristics into a D&O liability insurance premium, the insurance contract should therefore provide information on the firm’s financial perspectives, the quality of its management team and its “deep governance” features (i.e. its culture and character as highlighted in Baker and Griffith, 2007a). Unfortunately the basic characteristics of D&O insurance contracts are not publicly available for most U.S. firms (see Griffith, 2006). That is why researchers such as Chalmers at al. (2002) and Kalchev (2004) rely on a one broker’s private book of business. A second strategy used by Linck et al. (2009) is to examine the 27 firms incorporated in the state of New York and the 12 S&P firms that voluntarily disclosed the information to examine how much D&O premiums each company pays; but without coverage, the value of this information is very limited. An alternative is to use excerpts from the 1
A corporate director’s duty goes beyond a simple firm value maximizing paradigm to include a fiduciary duty, a duty of loyalty and a duty of care. 2
According to different Towers‐Watson surveys (that were in the past published by Tillinghast Towers‐Perrin, and before that by Watson‐Wyatt), approximately 95% of public corporations in the United States and 75% of public corporations in Canada provide such insurance to their managers. 3
According to these surveys, 20% of U.S. firms had at least one lawsuit brought against their directors in the previous ten years, of which half, and the most costly, arise from shareholders. 1 Towers‐Watson surveys as in Cao and Narayanamoothy (2011) – only two survey years 2001 and 2002 – and Fier et al. (2010). A third approach is to use Canadian data as in Core (1997, 2000), Boyer (2003), Park Wynn (2008), Gillan and Panasian (2009), Li et al. (2011), Rees et al. (2011) and Boyer and Stern (2012) since Canadian firms typically disclose the D&O insurance coverage they purchased along with the premium that was paid. This is the route I will take. But unlike most of these studies that examine large well established firms (the notable exceptions being Chalmers et al, 2002, and Boyer and Stern, 2012), I will focus on firms that just went public though an initial public offering (IPO). Coming back to Chalmers et al. (2002), whose analysis is closest to the current one since they sudy the relationship between IPO underpricing and D&O insurance coverage for 72 American firms that purchased D&O insurance through the same broker, they find that firms with more coverage at the time of the IPO were more likely to be sued for mispricing. Consequently, insurers are able to price this opportunism by charging higher premiums prior to the IPO to companies whose post‐IPO performance will be poor. In line with Chalmers et al. (2002), I limit my study to companies that went public through an initial public offering (IPO) to see whether D&O insurance have any power in predicting market risk and returns. I also limit the study to the years between 1996 and 2006 because information in not easily available prior to 1996, whereas securities class action lawsuits were introduced in Canada only after 2005.4 In contrast to Chalmers et al. (2002), my sample of firms includes some that opted for no insurance, thus reducing the potential bias involved in using only firms that purchased D&O insurance from one particular insurance broker. Empirical results will show that the higher is a firm’s D&O insurance premium per dollar of coverage (what is known as the rate‐on‐line in the insurance industry), the higher is the firm’s stock market volatility and the lower is its stock market return in the first year after the IPO. In other words, firms that insurance companies deem riskier at the time of the IPO have a lower (higher) stock market return (volatility) in the first year post‐IPO. Consequently D&O insurers appear to possess information that could be valuable to stock market participants. My results are robust to many econometric specifications (simultaneous equations, treatment effects) and robustness checks. The predictive power of D&O insurance disappears after the first 4
Prior to 2005, securities class action suits were virtually unheard of. The first “large” settlement occurred in September 2008. Since 2006, approximately 20 class action suits have been filed. Securities Docket webpage http://www.securitiesdocket.com/2008/09/14/dimitri‐lascaris‐and‐siskinds‐leading‐the‐way‐in‐canada/ (last visited on 31 October 2011). 2 year, suggesting that market participants eventually learn information that was only available to insurers at the time of the IPO. As this information gradually permeates the market, D&O insurance does not seem to hold any predictive long‐term power, at least with respect to the stock market return. The use of D&O insurance information as a tool to predict stock return and volatility falls within the recent push to find a way to account properly for a firm’s governance structure (see Rose, 2007, Bebchuk and Hamdani, 2009, Bebchuk and Weisbach, 2010, and Adams et al., 2010, inter alia). As many governance factors are not specified properly or are completely unavailable to the general investor (see Baker and Griffith, 2007a), one has to wonder how much of the internal structure of the firm remains unknown to market participants. That is why Holderness (1990), O’Sullivan (1997), Core (2000) and Boyer and Stern (2012) argue that D&O insurance providers are more likely to be good firm monitors since they have a monetary incentive to price the contract properly. A counter point to this argument is provided by Baker and Griffith (2007b) who argue that insurers do not seem to act as corporate governance monitors. The informational problems between managers and shareholders must be even greater in new firms and in firms that are becoming public. Not only is the information incomplete, but it is not even obvious what the directional impact between governance and performance is. Bhagat and Bolton (2008) suggest that governance indices generally suffer from important measurement errors, irrespective of the necessity to account for the endogeneous relationship between governance and performance (see also Wintoki et al., 2012). A case in point is the relationship between board independence and market performance. As Bebchuk and Weisbach (2010) argue, the relationship between board independence and performance is hard to find, perhaps because firms respond to a plethora of governance problems with an optimal albeit constrained governance structure (see Adams et al., 2010). Measurement problems become arguably more important when looking at IPOs since relatively little is publicly known about the firm’s past financial stability and health. The remainder of the paper is organized as follows. Section 2 presents a short primer on directors and officers insurance contracts. I develop my hypotheses and describe the data in Section 3, and present the main results of the paper in Section 4. Section 5 is devoted to robustness checks. Finally, Section 6 concludes with a discussion. 3 2. Directors’ and officers’ insurance: A primer 2.1 D&O insurance contracts Directors’ and officers’ liability insurance contracts cover corporate directors and officers against lawsuits brought against them as representatives of the corporation. Divergence in interests mixed with asymmetric information between managers (including both directors and officers) and shareholders is the main source of conflicts, and potentially the most costly. The insurance company will reimburse the corporation and/or its managers for the costs of settling and defending the lawsuit up to the policy limit, provided the firm’s directors and officers have acted honestly and in good faith. In theory, should managers and their company have acted in a fraudulent manner, the insurance company could decide not to honor the policy. A D&O insurance policy is usually comprised of three distinct types of coverage. Side A coverage covers directors and officers whenever their corporation is unable or unwilling to indemnify them. Side A coverage is usually purchased by a corporation to attract competent outside directors and has no deductible. Side B coverage reimburses the corporation whenever it indemnifies its managers following a lawsuit brought upon them as a representative of the corporation. Side C coverage reimburses claims made against the company itself, typically concerning securities claims. Hence, Side B and Side C cover the corporation rather than its directors and officers thereby epitomizing agency costs, as mentioned by Griffith (2006). The 2006 Towers‐Watson survey mentions that 38% of public companies purchased Side A only coverage, and that this figure is on the rise. Directors and officers are increasingly aware of the risks they face when they take on the position. 2.2 The pricing of D&O coverage Insurance companies must accurately assess the potential cost of each policyholder since it ultimately bears the full cost of any mistake. As a result, D&O insurance underwriters have developed specific risk assessment tools that allow them to out‐select their clients and assess litigation risk properly. Underwriters use three sources of information: The written application that contains a full array of documentation, the analysis of public financial and accounting data, and interviews with the prospective insured’s senior management team.5 This underwriting procedure 5
Baker and Griffith (2007a) provide a valuable in‐depth account of the pricing approach used by D&O underwriters using detailed interviews of 41 D&O insurance professionals (underwriters, actuaries, brokers, risk managers, lawyers and claim process specialists). They also report that, in the United States, the average 4 means that very little asymmetric information remains between the underwriter and the insured (see Knepper and Bailey, 1998). The information gathered by the insurer about a potentially insured firm’s internal processes and structure is not divulged to other market participants. Baker and Griffith (2007a) present one approach used to measure a firm’s potential cost of litigation that is based on the firm’s financial information, the industry’s perspective and the firm’s governance risk factors. Of particular interest is the fact that insurers appear to divide the analysis of governance characteristics into two categories: Culture, which refers to the stringency of the firm’s formal and informal internal controls (how the information is disseminated in the firm), and Character, which refers to the directors’ and officers’ attitude toward risk. In essence, Culture seeks to identify the source of potential D&O litigation whereas Character aims to uncovering the managers’ sense of ethics. Consequently, as in the pricing of any insurance contract, insurers must assess properly the probability that a claim will be paid as well as the severity of such a claim. In the case of D&O insurance, the probability and severity of a claim is likely linked to the firm’s stock market performance since the most costly lawsuit are brought against the firms’ managers by shareholders. Disclosure of basic D&O insurance information has become the norm in Canada. The earliest work on Canadian public data was done by Core (1997) who shows that the main determinants of the demand for D&O insurance are the firm’s litigation risk and its cost of financial distress. In a follow‐
up paper, Core (2000) notes that the premium paid is a function of the firm’s governance structure quality, which he measures as the firm’s ownership structure, board independence, management entrenchment and business risk. Ten years earlier, Bhagat, Brickley and Coles (1987) found that the decision to purchase D&O insurance does not reduce shareholder wealth (which would have been an indication of managerial moral hazard). They conclude that D&O insurance may have a positive impact because D&O insurers exert a monitoring role that shareholders should value (see also Holderness, 1990, and O’Sullivan, 1997). Liability insurance then favors internal monitoring since one director’s negligence could lead to coverage being denied for all. Consequently, purchasing D&O insurance comes with the service of an independent investigation of a firm’s internal practices. This monitoring hypothesis has been challenged by Baker and Griffith (2007b) who conclude that insurers do not act as corporate governance monitors. settlement was $13.3 million for the period 1996‐2001, $22.3 million for the period 2002‐2005 and $33 million in 2006‐2007. 5 3 Hypothesis development, data and variables description 3.1. Hypothesis and data The main hypothesis I develop in this paper is that firms that are riskier in the eyes of D&O insurance providers have a post‐IPO stock return that is lower and more volatile. Underpinning this question is my assertion that sellers of D&O insurance use a plethora of publicly unavailable information from which a premium emerges, just like a credit score. As a result, D&O insurance information should be valuable to investors. The risk assessment conducted by insurers (see Blades, 2006), with the possible denial of coverage should information have been omitted or distorted, means that there is very little residual information asymmetry. This means the D&O insurance providers have some type of private information regarding the inner working and the governance quality of the firms that seek to protect their managers against the cost of lawsuits. If such information was to reach the stock market, we should expect that it would be embedded directly and immediately in prices. Figure 1 provides a timeline of when information is available to the public. Figure 1. Timeline of the D&O insurance information acquisition and release Firm operates as a private enterprise D&O insurance purchased or not (no public information) as a private enterprise (the insurer conducts an audit and negotiates with the firm for the coverage and the premium) Firm decides to go public Firm contacts investment banker AND D&O insurance purchased or not (no public information) as a “new” public firm. The Offering prospectus that includes a pro‐forma annual statement, the date of the IPO and the offering price, is made public. IPO occurs End of fiscal year (not necessarily 12 months after the IPO) First annual report is released First management proxy is released (first public information about D&O insurance, if any) End of second year of operations The D&O insurers’ information technology is not publicly known. And because they are usually absent from the offering prospectus, D&O insurance coverage and premium are not known at the 6 time of the IPO, even though they are revealed later in the life of the firm. My null hypothesis will therefore be that D&O insurance contract parameters have no power in explaining future risk and returns. The alternate hypothesis will be that D&O insurance contract parameters have some ability to forecast a firm’s stock market risk and returns in the first year of public life post‐IPO. The main variable I shall use to measure an insurer’s assessment of a firm’s D&O liability risk will be the “rate‐
on‐line”, which is essentially the ratio of the total premium paid to the maximum possible coverage (or the policy limit). I present the main testable hypothesis below. H10: A firm’s D&O insurance rate‐on‐line at the time of the IPO has no power in explaining the firm’s first year’s stock market risk and return. H1A: A firm’s D&O insurance rate‐on‐line at the time of the IPO is linked to higher stock market volatility and lower return in the first year of trading. Because firms that pay a high premium per dollar of coverage are more susceptible to claims, their stock market return in the first year post‐IPO should be lower and more volatile.6 If a D&O insurance contract’s rate‐on‐line conveys information, then a corollary to hypothesis H1A should be that the firm’s stock market performance much after the information is revealed should be uncorrelated with such information. The information gathered by insurers to form an opinion regarding the liability risk of an IPO firm before the IPO date becomes known to market participants some time after the first year post‐IPO. This information should then be incorporated in the stock prices so that D&O insurance information should not have any long‐term predictive power. H20: A firm’s D&O insurance rate‐on‐line at the time of the IPO has no power in explaining returns after the first year. H2A: A firm’s D&O insurance rate‐on‐line at the time of the IPO is linked to lower return in the second year of trading. 6
Boyer and Tennyson (2008) argue that similar to any insurance contract, D&O insurance premiums depend on the frequency as well as on the severity of claims. Assuming that premiums are the product of frequency and severity, and that severity can be measured by the policy limit, we have that the rate‐on‐line (the ratio of premium to policy limit) is a good proxy for the frequency of lawsuits. And given that the different Towers‐
Watson surveys report that lawsuits are more likely to occur (and more severe) following a decrease in the stock price, we have that the D&O insurance market is efficient if the premium‐over‐coverage ratio is a function of the likelihood that the stock price will decline. 7 My results will show strong support for rejecting hypothesis H10 in favor of hypothesis H1A, but no support for rejecting hypothesis H20. Consequently, I will conclude that basic D&O insurance contract parameters have some power to predict stock market returns in the first year of life post‐IPO, but not in the second year. To test my hypotheses, I used the same approach as in Huson and Pazzaglia (2007) in collecting the initial dataset. Starting with the 1400‐odd new securities issued in Canada over the period 1995‐
2005, we are left with only 272 Canadian firms that correspond to the classic definition of an IPO for the period 1995 through 2005 (see Huson and Pazzaglia, 2007, for more details). Financial data was collected from Compustat (when available) and from the firms’ annual reports that were available on SEDAR, the Canadian equivalent of EDGAR. Due to missing data, my usable sample consists of 180 observations for which a decent number of financial and governance information is available. The main reasons why the number of firms drops from 272 to 180 is that many firms’ first proxy circular or annual reports are not available on SEDAR. It is also often the case that information about the governance structure of the entity is not obvious, especially in terms of the board composition (who the chairman is, who is an independent director, etc). Moreover the intersection of missing information from the annual report and from the management proxy is such that if we want to use a firm’s assets, market value of equity, total liability, board composition and independence, the number of usable observations drops to 150. The financial variables collected are as of the end of the first complete fiscal year post‐IPO. The governance and insurance information is collected in the first available management proxy and information circular. All monetary figures are in Canadian dollars and a conversion to Canadian dollars as of the end of the firm’s fiscal year was applied when needed. 3.2 Description of variables I describe the variables used in this paper below. I start with the dependent variables, the main independent variables and finish with all the control variables that I will use. 3.2.1
Dependent variables The dependent variables I shall use reflect the firm’s observed stock market risk and return following the IPO date. Starting with my main risk measure, the Volatility variable is calculated as the standard deviation of annualized daily returns. I expect firms that are deemed riskier by insurers, and that thus pay a higher rate‐on‐line, to have higher stock market volatility. 8 In terms of returns, I will use four different measures: The First year return, the First year excess return, the Two‐year return and the Second year return. In all cases, if D&O insurance providers have private information regarding the operations and governance of insured firms, firms that have a higher risk assessment measure should have a lower return. First year return is calculated as the total return in the first year after the IPO’s completion date. First year excess return is the one year total return the firm post‐IPO minus the return on the S&P/TSX, Canada’s main stock market index, over the same period. If insurers are able to charge a higher premium to firms that will perform poorly because of their revealed poor governance and opaque operations, then the relationship between the first year’s return, calculated either as First year return or First year excess return, and my measure of D&O insurance risk will be negative. The Two‐year return and the Second year return measures will also be used to see if D&O insurers are able to predict for a period longer than a year. The first is calculated as the total compounded stock return of the firms over a two‐year horizon after the IPO completion date, whereas the second is calculated as the two‐year compounded return minus the first year return. I will also combine the risk and return variables by calculating two quasi‐Sharpe ratio variables, called Sharpe ratio and Excess Sharpe ratio. In the first case, I divide First year return by Volatility, whereas I divide First year excess return by Volatility in the second measure. If the rate‐on‐line leads to higher volatility and lower returns, it should also lead to a lower quasi‐Sharpe ratio value. IPO issue price were collected using the firms’ prospectus available on SEDAR. Values were verified using the FPInfomart database. Subsequent price information comes from Bloomberg. 3.2.2
Main Independent variables The main variable of interest in this paper is the Rate‐on‐line variable, which is calculated as the ratio of the total premium paid to the maximum possible coverage (or the policy limit). This information is usually released in the first management proxy after the firm becomes public. As D&O insurance is considered part of managerial compensation, the information that is released relates to the previous year’s protection (just as salary information relates to the previous year’s compensation). I shall measure Rate‐on‐line in two ways to make the measure more tractable. First, I will use the log of ratio of the D&O premium to one thousand dollars of coverage (ln_prem_cov) to reduce the impact of very large rate‐on‐line on the results. Second, I will use the D&O premium divided by 1000 dollars of coverage (Premcov1000). In both cases a firm that pays a higher rate‐on‐line is 9 hypothesized to be perceived by the insurer as riskier since it is paying more per unit of coverage. I will therefore test whether the rate‐on‐line has an impact on stock market volatility and returns (as well as the quasi‐Sharpe ratios) in the first year of operations post IPO. The decision to purchase insurance or not will be used in the first step regression equation of the Heckman two‐step procedure to account for the possible selection bias that the sample of firms for which we observe the rate‐on‐line is not random. In other words, we need to account for the fact that we do not observe the rate‐on‐line for firms that choose to remain uninsured. Purchase is thus an indicator variable equal to one if the firm reported that it carried D&O insurance in its first management proxy following its first annual report post‐IPO, and zero otherwise.7 3.2.3
Control variables8 I divide the control variables in three categories: Financial, Governance and Other. Each variable is presented in turn. 3.2.3.1 Financial variables Mkt1year is the one‐year post‐IPO return of the stock market as calculated by the total return of Canada’s main stock index. I use FirstDayReturn as a control variable to account for the possible underpricing of IPO shares on the first day of trading to attract risk‐averse investors (for a more thorough discussion, see amongst many others Ritter, 1987). This variable is computed as the price at the end of the first day divided by the offer price. Another motivation for using FirstDayReturn as a control variable is that Chalmers et al. (2002) find a significant negative relationship between D&O insurance coverage and IPO underpricing, suggesting that IPO underpricing acts as a substitute for shareholder lawsuits. If true, this means that there exists a negative relationship between D&O insurance coverage (the denominator of the rate‐on‐line) and the one‐day return. It is therefore important to control for the first‐day return in the regression. I will measure firm size by the log of the firm’s market value of equity at the time of the IPO (lnMVE_IPO). Large firms should be less volatile and, possibly, more successful on average because more investors will scrutinize the activities of larger firms. I use a firm’s market value of equity at the 7
As in Core (1997), Park Wynn (2008), Lin et al. (2011) and Boyer and Stern (2012), I shall assume that the lack D&O insurance information disclosure in a firm’s proxy circular means that the firm is uninsured. 8
Unless notes otherwise, all independent variables are measured using accounting information available in the firm’s first annual report post IPO. 10 time of the IPO to guarantee that this measure of size is not too much confounded with stock market returns. The lnMVE_IPO variable is calculated as the log of the product of the offer price by the number of shares outstanding on the day of the IPO. Growth measures a firm’s growth opportunities. Following Core (1997), I compute Growth as market value of equity at time of IPO  book value of liabilitie s
. A firm with a high growth book value of assets
ratio should be more profitable if the growth options turn out to be in the money. Consequently, we should expect high growth firms to have higher first year returns on average. The last financial variable I will use is the firm’s book leverage (or Debt_Ratio), which I calculate as Total Liabilitie s
. Stock market volatility should be lower if debt holders are exercising a greater Total Assets
level of monitoring. On the other hand, the more levered is the firm the higher should be its stock return volatility since it is more at risk of going bankrupt. The overall effect is confounded. 3.2.3.2 Governance variables I collected several variables related to governance: CEO and chairman of the board duality, board composition and independence, the presence of a blockholder. I will also control for the corporate structure. Duality is an indicator variable taking on the value one if the chairman of the board is also the company’s chief executive officer and zero otherwise. This particular feature of a board is usually viewed as an entrenchment red flag. If entrenchment is an issue then Duality could be associated with low volatility (because the CEO/COB does not want to risk bankruptcy) and low return (because firm resources are spent on projects that are not necessarily value creating). Because Duality affects both risk and return in the same direction, I do not expect to see much impact on the Sharpe ratio variables. Blockholder is an indicator variable equal to one if a shareholder owns 10% or more of the firm’s voting shares according to its first proxy statement. Similar to debt holders, block holders should have more at skate in monitoring the firm. Consequently, stock return volatility should be lower when a block holder is present. Independence is the percentage of unrelated directors on the board of directors as mentioned in the firms’ proxy statements. The presence of a more independent board could increase returns if it 11 prevents the entrenchment of management and if it reduces the likelihood of cash flow misappropriation. The ITCE variable is an indicator variable equal to one if the company is an income trust and zero otherwise. I include this variable since Halpern (2004), Gillen (2005), Zetzsche (2005) and Huson and Pazzaglia (2007) argue that is that income trusts are riskier than stock companies from a governance standpoint (see also Boyer and Stern, 2012). The greater governance risk of income trusts is not only due to their higher risk of cash flow misappropriation, but also due to a lack of jurisprudence regarding their directors’ and officers’ duties. On the other hand, income trusts are required to pass along more of their operating cash flows to their investors, which should reduce volatility and returns since very little earnings are reinvested, which makes capital gains infrequent. Volatility should be lower since income trusts distribute more dividends, and are older and more mature firms than stock companies. The impact of ITCE on Sharpe is therefore confounding since the numerator and the denominator are expected to decrease when the firm is an income trust. 3.2.3.3 Other variables Risky_Industry is an indicator variable equal to one if the firm belongs to one of the ten two‐digit SIC codes risky industries as identified in Bajaj et al. (2000). Firms that belong to one of these ten risky industries, that were deemed riskier based on the number of cases settled as well as the average settlement amount, should have more volatile stock returns and, possibly, higher returns on average. Firms operating in risky industries should also be more likely to purchase D&O insurance if only because of the higher frequency of lawsuits. Age measures the number of years since the start of the company’s operations at the time of the IPO announcement. I expect this variable to have a negative coefficient impact on stock market volatility since the firms becoming public should be more mature and have more stable cash flows, everything else equal. The service offered by the investment banker at the time of the IPO is represented by the variable IPOfeerat, which is calculated as the total fees paid at the time of the IPO divided by the market value of the IPO (offer price multiplied by number of shares outstanding). I expect firms that purchase a higher level of service to have higher returns in the first year and lower volatility. Float is the ratio of the number of shares issued at the IPO on the total number of shares outstanding after the IPO. I expect that firms that have a higher float should have less volatile market returns 12 because more investors are likely to follow the firm, therefore disseminating the appropriate information to the markets. With respect to the decision to purchase D&O insurance or not, Float should be positively correlated with the decision to purchase. The reason is that the more shares are issued, the greater the probability of litigation and the greater should be the expected loss should a claim arise (see Guttiérrez, 2003, and Boyer, 2003) since minority shareholders are the most likely originator of lawsuits against managers. At the same time, a greater float means that the “firm’s entrepreneur” has gotten rid of a larger portion of the firm, which should be a bad signal to markets. If such adverse selection is present, stock return should be negatively related to the float. And because a higher float means that more different market participants are interested in the firm’s activities, information should be more reliable all the time so that stock market volatility should be smaller. US_Presence is a dummy variable equal to one if the firm reports activities in the United States. The variable was collected by looking at annual reports for the year following the IPO. The United States being a more fertile environment for potential litigation, I expect this variable to be positively correlated with the decision to purchase D&O insurance. A related measure to US_Presence is US_Sales that is measured as the ratio of sales that a firm reports doing in the United States to total sales. Sales in the United States should increase stock returns since it is a proxy for the potential growth in sales and profitability of the company. Appendix A presents the detailed construction of the variables and their anticipated impact on the different dependent variables. 3.3 Sample Statistics Table I presents the number of IPOs by year over the sample period. The first line presents the number of IPOs that were initiated in that year (1995‐2005), whereas the second line presents the number of IPOs that were completed in that year (1995‐2006). The third line gives us the year when the first management proxy circular is available (1996‐2006). [INSERT TABLE I ABOUT HERE] Table II presents the main sample statistics of the variables that I will use in this paper, starting with the dependent variables, the variables related to the D&O insurance contract and finishing with the different control variables and the variables used in the treatment equation related to the decision to 13 purchase D&O insurance. Dichotomous variables in Table II are those for which the entire sample statistic is not provided. [INSERT TABLE II ABOUT HERE] It is interesting to note that of the average return in the first year is 11.5%, giving an average excess return in the first year of 3.4%, whereas the median return is only 7.5% with a median excess return of ‐3.0%. Eight companies did not reach the end of the first year of operations, and 2 companies did not trade on their first day post‐IPO (for the other 15 companies for which no initial day return is available, the data is not available for the 3 and 12 companies that completed their IPO in 1995 and 1996 respectively). The two‐year return is more than twice the return in the first year; only one company ceased its operations in the second year. In terms of the main variables of interest, we see that information about the Rate‐on‐line is available for only 103 firms out of the original 180. Of the 77 firms for which no rate‐on‐line is available, 54 are deemed not to have purchased D&O insurance whereas the other 23 purchased D&O insurance but did not give enough details to calculate the rate‐on‐line (typically, the premium information is not revealed). These 23 firms must be kept in the dataset regressions because the mere fact that they purchased insurance tells us something about their behavior toward insurance. In the end, I therefore have 126 out of the 180 companies (70%) in my usable sample that purchased D&O insurance to protect their directors and officers. The proportion of D&O insurance purchasers in my usable sample is similar to that published in the different Towers‐Watson surveys in relation to the entire Canadian market. 4
Analysis of results The theoretical econometric model has the following structure. 1‐ Firms decide to go public through an initial public offering; 2‐ Firms decide to purchase D&O insurance or not 3‐ Insurers give firms that purchase D&O insurance a price per unit of coverage (rate‐on‐line); 4‐ The stock market reacts to the flow of information during the year. The variables that determine whether a firm goes public (step 1) are, of course, not measurable since we do not have access to information for non public firms. After deciding to go public, a firm purchases a D&O insurance contract (or not) in step 2. Regarding step 3, we observe a firm’s rate‐on‐
line only if it purchased insurance. To control for the potential selection bias in steps 2 and 3, I shall 14 use a classic Heckman two‐step procedure as we will see in later sections of the paper (but there does not seem to be any selection bias in the data). Step 4 is potentially given by a system of equations that gives us risk and return simultaneously. 4.1 Preliminary results Panel A of Table III separates the observations presented in Table II between firms that have D&O insurance and firms that do not descriptive statistics for the different independent and dependent variables for the usual sample of IPO firms. Of the 70% of IPO firms revealing having purchased D&O insurance around their IPO date, the average premiums was $148,554 for an average coverage of $24 million. We also see a very wide distribution for the rate‐on‐line variable, which is reported as the premium paid per $1000 of coverage. We see that the maximum rate‐on‐line is almost 40 times larger than the minimum rate‐on‐line. The average rate‐on‐line on line is $7,227 per $1,000 of coverage, whereas the median rate‐on‐line is $6,370 per $1,000 of coverage. This provides some justification for using a log‐transformation of the rate‐on‐line. [INSERT PANEL A OF TABLE III ABOUT HERE] Although income trusts represent a larger fraction of IPO companies, there is no statistical significant difference in the underpricing of income trusts and common equity companies. The one‐year stock performance statistics show that common equity firms perform better in the year following the IPO, although the difference in not statistically significant. The annualized daily stock price volatility over the same period is much higher for common equity firms, which corroborates the fact that income trusts generate more stable cash flows. When examining the decision to purchase insurance (the Purchase variable is equal to one if the firm revealed carrying D&O insurance around its IPO and zero otherwise), the hypothesis is that this decision should be influenced by financial, governance measure and other control variables. Panel B of Table III shows the results of a simple probit model on the decision or not to purchase D&O insurance for different model specifications. [INSERT PANEL B OF TABLE III ABOUT HERE] These results are quite consistent across all model specifications. Of all the financial and governance measures, only Debt_Ratio seems to have a significant impact on the purchase of D&O insurance across all model specifications. Specification 1 includes all the financial and governance variables we have in the regression, whereas Specification 2 includes only those that were significant in the 15 univariate analysis of Panel A (except US_Sales). Interestingly, only in Specification 2 is US_Presence significant in explaining the purchase of D&O insurance, in contrast to what was found for larger firms in Rees et al. (2011). Firm size, as measured by the log of the firm’s market value of equity (lnMVE_IPO), seems to have no impact on the decision to purchase D&O insurance. This is contrary to most studies that examine the demand for D&O insurance (see Core, 1997, and Boyer, 2003). Perhaps the demand for D&O insurance by new firms that are going public through an IPO is driven by factors that are slightly different that the demand for D&O insurance by more mature and established firms. Specification 3 uses only the governance measures to see if D&O insurance is a substitute or a complement to other governance measures, albeit crude. In particular, board Independence does not seem to increase the likelihood of having D&O insurance, contrary to what has been the main argument provided to support the existence of this type of insurance contract (see Bhagat et al., 1987). Concentrating now on Specification 4, that adds whether the firm announced in its IPO prospectus that it had D&O insurance (DO_Prospectus) to the variables used in Specification 2 (i.e., only the variables that were significant in explaining the purchase or not of D&O insurance in an univariate regression), we observe again that only the firm’s debt ratio has a significant impact on the likelihood of having D&O insurance in the first year after the IPO. Surprisingly, whether the firm operates has activities in the United States (US_Presence), which is significant in Specification 2, no longer seems significant after controlling for other factors. The positive and significant impact of the DO_Prospectus variable suggests that there is some sort of inertia in the purchase of D&O insurance so that the most likely reason for a firm to purchase D&O insurance in the current year that it purchased D&O insurance in the previous year (see Kalchev, 2004, and Boyer and Tennyson, 2008, for discussions of this phenomenon). We will control later (in Table IX) for the inertia associated with the D&O insurance information included in the prospectus. We shall see that the paper’s main results are sensibly the same. Specification 4 will be the model used later as the Heckman’s selection regression (first stage) model. 4.2 The Short‐term Predictive Power of D&O Insurance We report in the five panels of Table IV the results from OLS regressions where we examine the impact of the main independent variables (the log of the rate‐on‐line or the premium per $1,000 of coverage) on five different measures of risk and/or return. All regressions also control for a set of 16 measures that have been hypothesized to have an impact on the risk and return of a firm in the first year after the IPO. Panel A presents the results using the first year total return as the dependent variable, Panel B uses the market‐adjusted first year return (i.e., the difference between the firm’s stock return and the market return), Panel C uses the volatility of the firm’s stock return in the first year, Panel D with respect to the quasi‐Sharpe ratio (i.e., the ratio of the first year total return to volatility) and Panel E with respect to the market‐adjusted quasi‐Sharpe ratio (i.e., the ratio of the first year market‐adjusted return to volatility). [INSERT PANEL A OF TABLE IV ABOUT HERE] [INSERT PANEL B OF TABLE IV ABOUT HERE] [INSERT PANEL C OF TABLE IV ABOUT HERE] [INSERT PANEL D OF TABLE IV ABOUT HERE] [INSERT PANEL E OF TABLE IV ABOUT HERE] The story in all five panels is generally consistent with the hypothesis that D&O liability insurers have information at the time of the IPO that should be valuable to stock market participants. We see that the price per unit of D&O insurance coverage (whether it is the log of the rate‐on‐line or simply the premium‐to‐coverage ratio) is significantly related to the firm’s stock market return (in Panels A and B), risk (in Panel C) and return per unit of risk (Panels D and E) in the first year after the IPO.9 These results support the idea that the technology used by D&O insurers to transform into premiums the firms’ liability risk characteristics has some power in explaining the firms’ first year basic stock market return characteristics. Interestingly, very few control variables have any statistical power in explaining the return, the risk and the return per unit of risk of newly public firms. For instance in Panel A and Panel B, where we try to explain the first‐year return (Panel A) as well as the market‐adjusted first year return (Panel B), that no firm specific characteristic has any significant impact that is robust to all econometric specification. Results with respect to the return per unit of risk (the quasi‐Sharpe ratio in Panel D and 9
The number of observations used in Specification 1 all panels is higher than in the other regressions because it includes all firms for which we have accounting and governance information whereas the four other regressions in each panel include only those firms that purchased D&O insurance and for which we have the premium as well as the coverage. 17 the excess quasi‐Sharpe ratio in Panel E) have similar features. The only explanatory power that firm‐
specific variables have is in explaining the first year’s stock market volatility (Panel C). The results presented in the first five panels of Table IV use the information that is available at the time of the IPO. One could argue that all that information is already incorporated in the stock price at the time of the IPO by the investment banker, or that it is quickly reflected in the stock price at the end of the first day. As a consequence, it is normal to observe little impact of known firm specific characteristics on the stock return in the first year. Panel F and Panel G of Table IV present regression results using only the information that is unknown to investors at the time of the IPO: The rate‐on‐
line, the market portfolio return in the first year and the first day’s return. Panel F includes year fixed effects in the regression whereas Panel G does not. [INSERT PANEL F OF TABLE IV ABOUT HERE] [INSERT PANEL G OF TABLE IV ABOUT HERE] We see in these two panels that the paper’s main hypothesis, that the rate‐on‐line conveys valuable information, is supported since returns (both total and in excess of the market) and returns per unit of risk (both total and in excess of the market) are negatively affected by either measure of the rate‐
on‐line. Although the results for the stock return volatility are not as strong, the rate‐on‐line still has some predictive power when year fixed effects are included. 4.3 Long‐term impact I present in Table V regression results for the two‐year compounded stock market return and the second year return.10 The number of observations is, of course, smaller since some firms have disappeared in the second year because of merger activities or bankruptcy. [INSERT PANEL A OF TABLE V ABOUT HERE] In every specification the long‐run return is not linked to either measures of D&O insurance rate on line. Only the total stock return over the first year appears to have a significant and persistent impact on the two‐year compounded return, which should happen by construction as the return over the first year is included in the two‐year compounded return. Focusing on the second year return only (i.e., the two‐year compounded return minus the first year’s return), the regression result of 10
The analysis using the three‐year compounded and the third year return yielded similar results. 18 Specifications 5 and 6 show that the rate‐on‐line measures remain insignificant in explaining the long‐run return in all model specifications. That being said, it seems that Growth has some significant negative impact on long‐run returns, although it had no explanatory power in the first year. The results in Table V are consistent with our hypothesis that when the D&O insurance information becomes public sometime after the first year of trading, that information is somehow incorporated in the prices so that returns in the second and third year are no longer dependent on the information related to the first year of operations. We can therefore conclude that H20 cannot be rejected so that there is no evidence that D&O insurance information has any long‐term impact on stock prices. 5
Robustness checks This section reports the results of a series of robustness tests that examine the sensitivity of the empirical results to various econometric specifications. As we will see, the results presented in this section are generally the same as those I presented in the different panels of Table IV. The first robustness check I conduct is to see whether the choice of being insured or not creates a significant selection bias in the sense that firms that choose to purchase insurance are fundamentally different from those that do not. I will control for the potential selection bias using a classic Heckman two‐step approach in all the panels of Table VI. The results of Table VII take into account the potential simultaneous determination of risk and return in the model since feedback effects could exist between volatility and return. Using a three‐stage least square approach to account for such simultaneity does not alter the results much; if anything, the results appear stronger. Table VIII combines the selection bias and the feedback effect by including the inverse‐Mills ratio in the three‐stage least square regression. Again, the results remain unchanged. The last set of results I present in Table IX includes only the firms that did not provide enough information in their IPO prospectus to calculate a rate‐on‐line. Despite a further reduction in the number of observations, the main message of the paper holds: D&O insurance appears to have some predictive power regarding the return in the first year of operations post‐IPO. We can therefore be quite confident that the paper’s main results presented in the different panels of Table IV are robust. Consequently, there is ample evidence to suggest that the firms’s first year returns are significantly predicted by the price per unit of coverage of D&O insurance, suggesting that insurers that offer directors and officers protection against liability lawsuits brought against 19 them as firm representatives know something at the time of the IPO that investors should find valuable: An audit of the firm’s governance risk. 5.1 Likelihood of carrying D&O insurance and two‐step regression It is possible that the decision to purchase D&O insurance conveys information to the market that is not captured when we are only using an OLS regression of the rate‐on‐line to explain the first years’ risk and return. Indeed, given that we do not observe the rate‐on‐line of firms that do not purchase insurance the distribution of insured and non‐insured firms is not random and could therefore bias the results. I shall use a Heckman (1979) two step approach to reduce the potential bias. The first step consists of a probit regression that assesses the determinants that pushes a firm to purchase D&O insurance around its IPO date. Similar to the regression results we presented in Panel B of Table III, the dependent variable in this regression is Purchase. We will model the firms’ decision to purchase insurance using Specification 4 in Panel B of Table III. The second step in the Heckman two‐step procedure involves an OLS regression to determine what impact the two rate‐on‐line measures have on the different post‐IPO dependent variables. Table VI presents the regressions results for different model specifications. Panel A presents the two‐step analysis results for the first year total and excess stock returns whereas Panel B presents the results for the volatility and the quasi‐Sharpe ratio. [INSERT PANEL A OF TABLE VI ABOUT HERE] [INSERT PANEL B OF TABLE VI ABOUT HERE] Starting with the first stage results (under the different “Purchase” columns) we see that only whether the firm had announced in its prospectus that it carried D&O insurance have any impact on the whether the company said it had D&O insurance in its first proxy statement. No other firm specific variable11 seems to have any explanatory power so that the most likely reason for a firm to purchase D&O insurance in the current year that it purchased D&O insurance in the previous year. With respect to the second stage regressions, we see in Panel A (first year total and excess stock returns) that the two main variables of interest, the log of the rate‐on‐line and the premium per $1,000 of coverage, have a negative significant impact on the total return of firms that went public 11
The results vary a little from specification to specification because of the preferred use of a Maximum Likelihood Estimator econometric approach instead of a two‐stage least square regression. 20 through an initial public offering. Not only do the selection bias regression results confirm the qualitative results presented in Table IV, even the value of the coefficients are extremely similar. Indeed, comparing the results of Specifications 1 and 2 of Panel A of Table VI to the results of Specifications 5 and 7 of Panel A of table IV, we see that the economic size of the coefficients remains almost the same after controlling for the potential selection bias. This should tell us that the impact of the selection bias is quite small, at least with respect to the first year’s return. With respect to Volatility in Panel B (Specifications 1 and 2), we see again that the two main variables of interest have an impact that is in line with the paper’s main hypothesis: The insurance premium per unit of coverage is positively linked to the volatility of the stock price in the first year post‐IPO. This means that insurance companies that provide firm managers with protection in the event of a lawsuit brought against them as representatives of the corporation seem to have a technology that allows them to predict the volatility of the stock price in the year following the IPO. If one believes that stock price volatility is, inter alia, the result of governance opaqueness, the results presented in Panel B of Table VI suggest that insurers are able to assess such opaqueness when they insure the firms’ managers. In terms of the other control variables, we observe that firm size at the time of the IPO (lnMVE_IPO) is negatively related to stock price volatility, and so is the income trust structure (ITCE) compared to the stock company. The first result is in line with the perception that larger firms should have a larger number of analysts following them so that stock price volatility should be smaller. With respect to income trusts, the Volatility results suggest that because income trust are mandated to distribute more to unit holders, they require stable cash flows to exist. In both cases the results are in line with our hypotheses. Interestingly, firms that have a larger proportion of their stock issued at the time of the IPO (Float) do not have a lower stock volatility, at least not significantly so. The percentage of total sales coming from the United States seems to increase stock return volatility in the first year, whereas the DebtRatio reduces volatility. This latter result is possibly due to some monitoring done by debt holders who want to make sure that there is enough money to pay the debt, thus resulting in projects that have more stable cash flows. Another possibility is that creditors accept to lend money only to those firms that have stable cash flows in the first place. Finally, when the dependent variable is the market‐adjusted quasi‐Sharpe ratios (Specifications 3, 4 and 5), the results are still sensibly similar to the results presented in Panel E of Table IV, again suggesting that the selection bias, if it exists, is relatively small and economically inconsequential. Interestingly, the results are still strong when the log of the rate‐on‐line is used (Specification 3), but 21 the not when the rate‐on‐line per se is used (Specification 4). A possible explanation is that there is some non‐linearity in the impact of the rate‐on‐line that is not picked up by the variable in itself. Specification 5 tests this non‐linear hypothesis of the impact of Rate‐on‐line on the market‐adjusted quasi‐Sharpe ratio. Results are again in line with the paper’s main hypothesis that insurance companies have information that should be valuable to market participants since the rate‐on‐line of the firms that purchase D&O liability insurance protection to protect their managers against costly lawsuits appear to have a significant predictive power on the risk‐return profiles of firms in the first year post‐IPO. The message that we can take home when we control for the possible bias associated with the possible difference between firms that purchase D&O insurance and firms that do not is consistent with the main message of the paper: D&O insurers appear able to anticipate the risk‐return profile of public firms in their first year of trading. Consequently, consistent with the paper’s main hypothesis, the higher is the price of insurance per unit of coverage, the lower will be the total stock return in the first year post‐IPO, the higher will be the stock return volatility and the lower will be the quasi‐Sharpe ratio of the firm, even after controlling for the potential bias associated with the decision to purchase insurance or not 5.2 Simultaneous (feedback) effects of risk and return A second robustness check that is important to conduct is to see what happens if the stock market risk and return are jointly determined and that each feedbacks into the other. For instance, a security that has zero return on every day will have zero volatility. I will therefore modify the econometric models to estimate simultaneously the total stock market returns of the securities and their volatility in the first year post‐IPO. The simultaneous regression results using ln_ROL as the main independent variable appear in Panel A of Table VII, whereas the results using Rate‐on‐line appears in Panel B. The econometric technique used in Table VII is either a two‐stage or three‐stage least square regression to make sure that the results presented thus far are not due to a misspecification of the econometric model. The only difference between Specifications 2 and 3 in both panels is that the year fixed effect is not included in the stock return regression of Specification 3. [INSERT PANEL A OF TABLE VII ABOUT HERE] [INSERT PANEL B OF TABLE VII ABOUT HERE] 22 As it is clearly apparent in both panels, the simultaneous equation approach (whether through a two‐
stage least‐square as in Specifications 1 of each panel or a three‐stage least‐square as in Specifications 2 and 3 of each panel) does not alter the main message of the paper: The premium per unit of coverage is a good predictor of the first years’ total stock return and volatility. Indeed, results in both panels show that the price‐to‐coverage ratio is positively associated with volatility and negatively associated with returns. With respect to the control variables, firm size at the time of the IPO (LnMVE_IPO) still has a negative impact on stock market volatility in all model specifications in both panels, and so does the incorporation as an income trust (ITCE) and the debt ratio (DebtRatio). The percentage of sales in the United States is also still significantly positive in explaining Volatility. In terms of the significant impact of independent variables on returns, only the ratio of IPO fees to the market value of equity (IPOfeerat) and Float are significant at the usual levels, but only if we include a time fixed effect in the regression. Interestingly, when the time fixed effect is not included in the regression, it seems that only the premium‐to‐coverage ratio (however it is measured) has any power in explaining the stock return in the first year. The results in Table VIII combine both the correction for the selection bias (is a firm insured or not) and the simultaneity of the risk‐return relationship in the firms’ first year of public life post‐IPO. To do so, the regressions include as an explanatory variable the inverse‐Mills ratio (MillsRatio) obtained from a first stage regression12 on the probability of being insured. [INSERT TABLE VIII ABOUT HERE] As we see, the main message of the paper remains valid. As before, controlling for the selection bias does not seem to alter the results much so that the economic impact of the independent variables in Table VIII is sensibly the same as in Table VII (as was the case when we compared the results of Table VI with those of Table IV). 5.3 Information availability In Table IX, I present a summary of the previous results, but with a dataset limited to those firms that did not reveal any information related to the D&O insurance rate‐on‐line in the IPO prospectus. This further restriction on the data results in the deletion of approximately 15 firms compared to the 12
The regression used to calculate MillsRatio is the same regression as the one used the first stage of the Heckman two‐step analysis in Table VI. 23 results presented in Table IV. The reason behind this analysis is to see if my results are due to a subset of firms for which all information related to the D&O insurance policy was incorporated by the market. Panel A of Table IX presents regression results using the full model to explain first year returns, first year excess return and stock market volatility in the first year. As we see, the results are very close those I presented before so that the negative and significant relationship between the rate‐on‐line (and the log of the rate‐on‐line) and the return in the first year post‐IPO remains even by removing the 15‐odd firms that revealed their rate‐on‐line in their IPO prospectus. The same can be said about excess returns and stock volatility. [INSERT PANEL A OF TABLE IX ABOUT HERE] Panel B of Table IX presents the regression results when we suppose that the investment bankers incorporated all the public information in the offer price. Again the results tell the same type of story so that the ratio of D&O liability insurance premium to the D&O liability coverage appears to have predictive power for the return in the first year of operations post‐IPO, but a bit less so when we examine market volatility. [INSERT PANEL B OF TABLE IX ABOUT HERE] In Panel C of Table IX, the regression tries to explain the return past the first day. In other words, if we suppose that the market incorporates all the information available in the first day of trading, no information that was available prior to the first day should have any power in predicting returns in the year after the first day. It seems, however, that the return in the first year after the first day is again negatively related to the rate‐on‐line, whichever way it is measured. The predictive power of the rate‐on‐line remains quite strong whether we measure the net of first day return as (first year – first day), in the top half of Panel C, or as (1 + first year) / (1 + first day) in the bottom half. Results hold whether we use all available data points (Specifications 1 though 4 in both halves of Panel C) or only restrict the analysis to those firms that did not report the rate‐on‐line in the IPO prospectus (Specifications 5 though 8). [INSERT PANEL C OF TABLE IX ABOUT HERE] 5.4 Summary of the robustness checks The general conclusion that follows the robustness section of the paper is that the main message of the paper is supported using many different econometric model. It is therefore reasonable to conclude that there is ample evidence to suggest that the price of D&O insurance per unit of 24 coverage is a good predictor of future stock returns and volatility in the first year of life post‐IPO of new public firms. Indeed, it appears that in the first year post‐IPO, the greater is the D&O liability insurance premium per dollar of coverage, the lower will be the total stock return, the lower will be the market‐adjusted stock return, the higher will be the stock return volatility and the lower will be the return‐to‐volatility ratio. We feel these results are quite robust to conclude that hypothesis H10 can safely be rejected. These results lend support to the underlying story that D&O liability insurers have information about the internal structure of soon‐to‐be public firms that other investors do not have. Lawsuits are more likely to occur when returns are low and the volatility is high (the downside risk is greater). As a result, when insurers decide to insure a firm’s managers against lawsuits brought against them as representatives of the corporation, they assess the potential probability and severity of such lawsuits, which should be related to the future return and volatility of the stock price. Consequently, D&O insurers have had to develop an audit technology (underwriting skill) such that firms that face a high probability and severity of a costly lawsuit are required to pay a higher price of insurance per dollar of coverage. 6 Discussion and conclusion The primary objective of this paper was to examine whether insurers, that provide firms and their managers with protection against the risk of litigation, have access to information that could be valuable to investors at the time of the firms’ initial public offering. More precisely, firms that seek to protect their managers against costly liability lawsuits may purchase protection from insurers that examine each firm’s governance structure, organizational processes and its “character and culture” to arrive at a premium that reflects the firm’s risk of lawsuits. Even though this audit is performed prior to the IPO date, investors learn this information much after, typically right after the first year of operations. Whenever insurers decide to provide coverage to any policyholder (and this is true whatever the type of insurance that is sold), they assess the potential probability and severity of the possible claim or claims they could potentially be asked to pay. D&O insurance claims are generally related to lawsuits brought against the firm’s managers by shareholders or other stakeholders. Intuitively it is easy to imagine that lawsuits are more frequent when the stock had performed poorly, and that a lawsuit’s severity is greater when volatility is higher. Consequently, insurers that sell D&O insurance must use an audit technology (known as risk underwriting in the insurance industry) that yields a higher price of insurance when the expected loss is greater, whether this is due to a higher frequency 25 of losses (i.e., lower stock returns) or a higher severity of lawsuits (higher stock volatility). The main results of this paper confirm this view of the D&O insurance world. More to the point, I find that firms that pay a high price for their directors’ and officers’ liability insurance coverage tend to underperform in their first year since they are more likely to have a lower stock return and a higher volatility. The results I present support the assertion of Baker and Griffith (2007a) that premiums charged by D&O insurers are based on business and governance risk factors that are unobservable for normal investors. It therefore appears that D&O insurance can act as an alternative to governance indices. Since lawsuits against corporate directors are potentially very expensive insurers have a monetary incentive to assess properly the amount of risk they are assuming. As a result insurers have a direct incentive to measure this liability risk as it impacts the potential cost of litigation (see Core, 2000, and Baker and Griffith, 2008). Consequently, the ratio of D&O insurance premium to D&O insurance coverage offers a measure of the governance and litigation risk that a firm faces; a measure which, if underestimated by the insurance company, will lead to unexpected losses by the entity that issues the measure. Unfortunately for American investors, the information related to the purchase of D&O insurance is usually not available in the United States (see Griffith, 2006). And even in cases where some information is available, as in the state of New York (see Linck et al., 2009), it is usually the case that only the premium information is given. If one does not know how much insurance was purchased, the premium paid does not tell investors much. The results in this paper suggest that mandating the revelation of basic D&O insurance information (premium and coverage) could be valuable to investors since it would provide them with an unbiased signal about a firm’s risk of litigation against its directors and officers, perhaps because of some managerial mishaps. As the vast majority of Canadian and American corporations purchase liability insurance on behalf of their directors and officers (see the different Towers‐Watson surveys), wouldn’t be cost efficient to have access to this measure of risk? One must be wary to make such a leap to large U.S. listed firms, however, since the results in the current paper relate only to the case of small Canadian firms that became public through an initial public offering. Nevertheless, the question is relevant and timely given the time and energy that investors and regulators devote to governance issues. 26 7. References Adams R., B.E. Hermalin and M.S. Weisbach (2010). The Role of Board of Directors in Corporate Governance: A Conceptual Framework and Survey. Journal of Economic Literature, 48: 58‐107. Bajaj, M., S. Mazumdar and A. Sarin (2000). Securities class action settlements: an empirical analysis. Working Paper, University of California, Berkeley. Available at SSRN: http://ssrn.com/abstract=258027. Baker, T. and S.J. Griffith (2007a). Predicting Corporate Governance Risk: Evidence from the Directors' and Officers' Liability Insurance Market. University of Chicago Law Review, 74: 1‐58. Baker, T. and S.J. Griffith (2007b). The Missing Monitor in Corporate Governance: The Directors’ and Officers’ Liability Insurer. Georgetown Law Journal, 95: . Baker, T. and S.J. Griffith (2008). How the Merits Matter: Directors’ and Officers’ Insurance and Securities Settlements. University of Pennsylvania Law Review, 157: 755‐832. Bebchuk, L.A., A. Cohen and A. Ferrell (2009). What Matters in Corporate Governance? Review of Financial Studies, 22: 783‐827. Bebchuk, L. A. and A. Hamdani (2009). The Elusive Quest for Global Governance Standards. University of Pennsylvania Law Review, 157 (4): 1263‐1317. Bebchuk L.A. and M.S. Weisbach (2010). The State of Corporate Governance Research. Review of Financial Studies, 23: 939‐961. Bhagat S. and B. Bolton (2008). Corporate Governance and Firm Performance. Journal of Corporate Finance, 14: 257‐273. Bhagat, S., J.A. Brickley and J.L. Coles (1987). Managerial Indemnification and Liability Insurance: The Effect on Shareholder Wealth. Journal of Risk and Insurance, 54: 721‐736. Beatty, R.P and J.R. Ritter (1986). Investment bank reputation and the underpricing of initial public offerings. Journal of Financial Economics 15: 213‐232. Blades, J. (2006). A different outlook: D&O underwriters are putting companies under more scrutiny than in the past. Best's Review, September 2006, p.107‐108. Boyer, M.M. (2003). Directors’ and Officers’ Insurance and Shareholders’ Protection. CIRANO Scientific Series, No.2003s‐64 (http://www.cirano.qc.ca/pdf/publication/2003s‐64.pdf). Boyer, M.M. and L.H. Stern (2012). Is Corporate Governance Risk Valued? Evidence from Directors’ and Officers’ Insurance. Journal of Corporate Finance 18: 349‐372. Boyer, M.M., and S. Tennyson (2008). Directors’ and Officers’ Liability Insurance, Corporate Risk and Risk Taking: New Panel Data Evidence on the Role of Directors’ and Officers’ Liability Insurance. ARIA 2008 meeting, Portland (OR). Cao, Z. and G.S. Narayanamoorthy (2011). The Effect of Litigation Risk on Management Earnings Forecasts. Contemporary Accounting Research, 28 :125‐173. 27 Chalmers, J.M.R., L.Y. Dann and J. Harford (2002). Managerial Opportunism? Evidence from Directors’ and Officers’ Insurance Purchases. Journal of Finance, 57: 609‐639. Clarkson, P. and D. Simunic (1994). The Association between Audit Quality, Retained Ownership and Firm‐Specific Risk in United States vs. Canadian IPO markets. Journal of Accounting and Economics, 17: 207‐228. Core, J. E. (1997). On the Corporate Demand for Directors’ and Officers’ Insurance. Journal of Risk and Insurance, 64: 63‐87. Core, J.E. (2000). The Directors’ and Officers’ Insurance Premium: An Outside Assessment of the Quality of Corporate Governance. Journal of Law, Economics and Organization, 16: 449‐477. Fier, S., J. Gable, K.A. McCullough and N. Mansfield (2010). The Directors and Officers Insurance Marketplace: An Empirical Examination of Supply and Demand in Uncertain Times. WRIEC 2010 meeting, Singapore. Gillan, S. and C. Panasian (2009). Reassessing what matters in Corporate Governance: Evidence from the Market for Directors’ and Officers’ Liability Insurance. FMA 2009 meeting Reno (NV). Gillen, M., 2005. Income Trust Unitholder Liability: Risks and Legislative Response. Canadian Business Law Journal 42: 325‐345. Gillen, M., 2006. A Comparison of Business Income Trust Governance and Corporate Governance: Is There a Need for Legislation or Further Regulation? McGill Law Journal 51: 327‐383. Griffith, S.J. (2006). Unleashing a Gatekeeper: Why the SEC Should Mandate Disclosure of Details Concerning Directors' and Officers' Liability Insurance Policy. University of Pennsylvania Law Review, 154: 1147‐1208. Gutiérrez, M. (2003). An Economic Analysis of Corporate Directors’ Fiduciary Duties. Rand Journal of Economics, 34: 516‐535. Heckman, J. (1979). Sample Selection Bias as a Specification Error. Econometrica, 47: 153‐161. Holderness, C.G. (1990). Liability Insurers as Corporate Monitors. International Review of Law and Economics, 10: 115 ‐129. Huson, M.R. and F. Pazzaglia (2007). Choice of Organizational Form as a Trade‐Off between Fit and Market Timing. Available at SSRN: http://ssrn.com/abstract=970268. Kalchev, G. (2004). The Demand for Directors' and Officers' Liability Insurance by US Public Companies. Available at SSRN: http://ssrn.com/abstract=565183. Knepper, W. and D.A. Bailey (1998). Liability of Corporate Officers & Directors, Seventh Edition, Matthew Bender edition. Lin, C., M.S. Officer and H. Zou (2011). Directors’ and officers’ liability insurance and acquisition outcomes. Journal of Financial Economics, 102: 507‐525. 28 Linck, J.S, J.M. Netter and T. Yang (2009). The Effects and Unintended Consequences of the Sarbanes‐
Oxley Act on the Supply and Demand for Directors. The Review of Financial Studies, 22(8): 3287‐3328. Maddala, G.S. (1983). Limited Dependent and Qualitative Variables in Econometrics, Cambridge. McTier, B.C. and J.K. Wald (2011). The Causes and Consequences of Securities Class Action Litigation. Journal of Corporate Finance, 17: 649‐665. O’Sullivan, N. (1997). Insuring the Agents: The Role of Directors’ and Officers’ Insurance in Corporate Governance. Journal of Risk and Insurance, 64: 545‐556. Park Wynn, J. (2008). Legal Liability Coverage and Voluntary Disclosure. The Accounting Review, 83: 1639‐1669. Rees,R., D. Radulescu and P. Egger (2011). Corporate Governance and Managerial Incentives: Evidence from the Market for D&O Insurance. Working paper, ETH Zurich Ritter, J. (1984). The ‘hot issue’ market of 1980. Journal of Business 57: 215–240. Ritter, J. (1987). The costs of going public. Journal of Financial Economics 19: 269‐282. Ritter, J. (1991). The long‐run performance of initial public offerings. Journal of Finance 46: 3‐27. Rose, P. (2007). The Corporate Governance Industry. Journal of Corporation Law, 32(4) Towers‐Watson (formerly Tillinghast Towers Perrin) (1999, 2000, 2002, 2003, 2004, 2005, 2006). Directors and Officers Liability Survey, US and Canadian Results. Wintoki M.B., J.S. Linck and J.M. Netter (2012). Endogeneity and the Dynamics of Corporate Governance. Journal of Financial Economics (In Press). 29 Appendix B. List of all companies in the final sample with the date of the IPO initiation and IPO completion.
For all the 180 firms in our sample, we present, in alphabetical order of their name at the time of the IPO, the date of the IPO initiation as well as the date of the IPO completion.
# Company name at the time of the IPO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1-8-0 Connect Inc.
A & W Revenue Royalties Income Fund
Absolute Software Corporation
ACS Media Income Fund
Adaltis Inc.
Addax Petroleum Corporation
Addenda Capital Inc.
ADF Group Inc.
Adherex Technologies Inc.
Aeroplan Income Fund
AEterna Zentaris Inc.
Alexis Nihon Real Estate Investment Trust
Algonquin Power Income Fund
AltaGas Utility Group Inc.
Amtelecom Income Fund
Angiotech Pharmaceuticals Inc.
Armtec Infrastructure Income Fund
Arriscraft International Income Fund
ART Advanced Research Technologies Inc.
Art In Motion Income Fund
Ascalade Communications Inc.
Associated Brands Income Fund
Atrium Biotechnologies Inc.
ATS Andlauer Income Fund
Axcan Pharma Inc.
AXXENT Inc.
Bell Nordiq Income Fund
BFI Canada Income Fund
Calpine Natural Gas Trust
Canaccord Capital Inc.
Canadian Helicopters Income Fund
Canadian Hotel Income Properties Real Estate Investment Trust
Canadian Satellite Radio Holdings Inc.
Canexus Income Fund
Cavell Energy Corporation
Celestica Inc.
Centerra Gold Inc.
Chromos Molecular Systems Inc.
Clarke Inc.
Clearwater Seafoods Income Fund
COM DEV International Ltd.
ConjuChem Inc.
Connors Bros. Income Fund
Coretec Inc.
Cossette Communication Group Inc.
Countryside Power Income Fund
Custom Direct Income Fund
DALSA Corporation
Danier Leather Inc.
DataMirror Corporation
Davis + Henderson Income Fund
DirectCash Income Fund
Duke Energy Income Fund
Duvernay Oil Corp.
Dynetek Industries Ltd.
E.D. Smith Income Fund
Electrovaya Inc.
Enbridge Income Fund
Energy Savings Income Fund
eNGENUITY Technologies Inc.
Date of IPO Date of IPO
completion # Company name at the time of the IPO
initiation
(yyyymmdd) (yyyymmdd)
20040227
20040423
61 Equitable Group Inc.
20011221
20020215
62 ExAlta Energy Inc.
20000210
20000328
63 EXFO Electro-Optical Engineering Inc.
20030306
20030508
64 FP Newspapers Income Fund
20041013
20041201
65 Freehold Royalty Trust
20051206
20060216
66 Futuremed Healthcare Income Fund
20041105
20041214
67 Gateway Casinos Income Fund
19990406
19990721
68 General Donlee Income Fund
20010301
20010605
69 Gienow Windows & Doors Income Fund
20050520
20050629
70 Gildan Activewear Inc.
19951020
19951219
71 GMP Capital Corp.
20021025
20021220
72 Golf Town Income Fund
19971016
19971223
73 Granby Industries Income Fund
20050829
20051117
74 Great Lakes Hydro Income Fund
20030116
20030306
75 Hardwoods Distribution Income Fund
19971017
19971218
76 Harvest Energy Trust
20040617
20040727
77 Heating Oil Partners Income Fund
20041104
20041214
78 Highpine Oil & Gas Limited
20000428
20000627
79 Hub International Limited
20040621
20040803
80 Husky Injection Molding Systems Ltd.
20050408
20050627
81 Hydrogenics Corporation
20020930
20021115
82 IAMGOLD Corporation
20050217
20050406
83 IAT Air Cargo Facilities Income Fund
20050826
20050930
84 IDS Intelligent Detection Systems Inc.
19951030
19951228
85 Indigo Books & Music Inc.
19990917
19991028
86 Industrial Alliance Insurance and Financial Services
20020306
20020423
87 ING Canada Inc.
20020306
20020425
88 Innergex Power Income Fund
20030825
20031015
89 InnVest Real Estate Investment Trust
20040527
20040630
90 INSCAPE Corporation
20050729
20050909
91 IPC US Real Estate Investment Trust
19970429
19970625
92 Ivanhoe Energy Inc.
20051104
20051212
93 KCP Income Fund
20050630
20050818
94 Keyera Facilities Income Fund
19960830
19961031
95 Kingsway Financial Services Inc.
19980429
19980707
96 KMS Power Income Fund
20040513
20040630
97 Legacy Hotels Real Estate Investment Trust
20000529
20000717
98 Liquor Stores Income Fund
19980122
19980317
99 Livingston International Income Fund
20020611
20020731 100 MacDonald Dettwiler and Associates Ltd.
19961021
19961210 101 Madacy Entertainment Income Fund
20000929
20001130 102 Mainframe Entertainment Inc.
20010926
20011108 103 Manitoba Telecom Services Inc.
20000804
20000926 104 March Networks Corporation
19990430
19990618 105 Maxxcom Inc.
20040220
20040408 106 Mediagrif Interactive Technologies Inc.
20030328
20030529 107 Mega Bloks Inc.
19960404
19960521 108 Menu Foods Income Fund
19971113
19980520 109 MethylGene Inc.
19961017
19961216 110 Miranda Technologies Inc.
20011108
20011220 111 Morneau Sobeco Income Fund
20041103
20041214 112 Movie Distribution Income Fund
20051103
20051220 113 Neurochem Inc.
20031212
20040203 114 Noranda Income Fund
20000417
20000921 115 Northbridge Financial Corporation
20050502
20050603 116 Nurun inc.
20001103
20001110 117 O&Y Real Estate Investment Trust
20030526
20030630 118 Oceanex Income Fund
20010309
20010430 119 OFI Income Fund
19990422
19990722 120 OnX Enterprise Solutions Inc.
Date of IPO Date of IPO
completion # Company name at the time of the IPO
initiation
(yyyymmdd) (yyyymmdd)
20040205
20040318 121 OPTI Canada Inc.
20050323
20050510 122 Osprey Media Income Fund
20000609
20000630 123 Pantera Drilling Income Trust
20020326
20020528 124 PBB Global Logistics Income Fund
19961008
19961125 125 PDM Royalties Income Fund
20051128
20060106 126 Philex Gold Inc.
20021011
20021128 127 Pizza Pizza Royalty Income Fund
20020319
20020503 128 Polaris Minerals Corporation
20040910
20041019 129 Polyair Inter Pack Inc.
19980514
19980624 130 Prime Restaurants Royalty Income Fund
20031022
20031209 131 ProMetic Life Sciences Inc.
20041014
20041112 132 Pure Energy Services Ltd.
20041108
20041216 133 Q9 Networks Inc.
19990914
19991118 134 Quadra Mining Ltd.
20040216
20040323 135 Residential Equities Real Estate Investment Trust
20020927
20021205 136 Revett Minerals Inc.
20020320
20020522 137 Richards Packaging Income Fund
20050218
20050405 138 Ridley Inc.
19981222
19990210 139 RONA Inc.
19980921
19981109 140 ROW Entertainment Income Fund
20000731
20001101 141 Royal Host Real Estate Investment Trust
19951227
19960308 142 Saputo Inc.
19970404
19970610 143 Saxon Financial Inc.
19971024
19971231 144 SCI Income Trust
19961009
19961204 145 Second Cup Royalty Income Fund
19990915
20000210 146 SFK Pulp Fund
20041028
20041210 147 Shoppers Drug Mart Corporation
20030528
20030707 148 Sierra Systems Group Inc.
20020606
20020726 149 Solium Capital Inc.
19971127
19971209 150 Spectra Premium Industries Inc.
20011113
20011220 151 Sun Gro Horticulture Income Fund
19970506
19970611 152 Sun Life Financial Inc.
20020710
20020823 153 Sunrise Senior Living Real Estate Investment Trust
20030411
20030530 154 Sun-Rype Products Ltd.
19951031
19951218 155 Superior Plus Income Fund
19970320
19970610 156 Synenco Energy Inc.
19970912
19971110 157 TECSYS Inc.
20040817
20040928 158 Teknion Corporation
20011221
20020211 159 Telesystem International Wireless Inc.
20000518
20000713 160 TGS North American Real Estate Investment Trust
20050311
20050420 161 The Brick Group Income Fund
19970508
19970617 162 The Consumers' Waterheater Income Fund
19960502
19970107 163 The Data Group Income Fund
20050309
20050427 164 The Descartes Systems Group Inc.
19991229
20000323 165 The Keg Royalties Income Fund
20000724
20001003 166 Tree Island Wire Income Fund
20020322
20020503 167 Trimac Income Fund
20020326
20020522 168 TSX Group Inc.
20040426
20040629 169 Tundra Semiconductor Corporation
20051027
20051208 170 UE Waterheater Income Fund
20050824
20050930 171 VCom Inc.
20030829
20031015 172 VFC Inc.
20000510
20000622 173 Vincor International Inc.
20020314
20020503 174 Western Oil Sands Inc.
20030411
20030528 175 WestJet Airlines Ltd.
19980227
19980422 176 Workbrain Corporation
20010424
20010627 177 Xantrex Technology Inc.
19971031
19980106 178 Xceed Mortgage Corporation
20050715
20050901 179 Xenos Group Inc.
20000229
20000417 180 Yellow Pages Income Fund
Date of IPO Date of IPO
completion
initiation
(yyyymmdd) (yyyymmdd)
20040223
20040415
20040305
20040415
20051219
20060222
20020325
20020515
20040416
20040608
19960912
19961028
20050527
20050706
20051024
20060110
19951220
19960220
20020606
20020722
19980525
19980728
20051216
20060206
20040309
20040429
20040220
20040408
19971107
19980216
20041221
20050216
20040227
20040407
19970730
19970808
20020909
20021105
20030915
20031104
19970903
19971031
19970916
19971009
20050526
20050707
19970822
19971016
20041025
20041202
20020614
20020801
20011025
20011121
19980227
19980421
20010305
20010511
19971017
19971209
20020212
20020327
20000223
20000327
20041115
20041223
19960829
19961112
19960807
19961008
20050926
20051112
19980529
19980727
19980520
19980714
19970320
19970509
20021003
20021206
20040528
20040720
20021030
20021217
20041115
20041221
19971114
19980121
20020412
20020531
20020930
20021112
20050107
20050225
20020907
20021112
19981210
19990208
20031110
20031219
20050929
20051114
20030829
20031006
19960329
19960606
20001023
20001221
19990525
19990713
20031104
20031211
20040206
20040319
20040430
20040608
19990628
19990811
20030626
20030801
29
Table I. Number of Canadian IPOs per year in sample: Year of IPO initiation, year of IPO completion and year of first management proxy
For the 180 firms in our sample (see Appendix B for the complete list), we present the number of IPOs intitiated between 1995 and 2005, and completed between 1995 and
2006. We also report the year in which the first management proxy is available, the total number of new securities issued in Canada over the sample years, the total potentially
usable proxies (excluding financial structured products, capital pools, utilities) and the total market value (in million of Canadian dollars) of new securities issued.
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Total
IPO initiated
5
11
20
11
9
15
10
25
16
31
27
-
180
IPO completed
3
12
17
13
9
17
8
27
15
31
23
5
180
Proxy
-
4
12
19
12
9
15
10
27
14
32
26
180
Total new issues
177
257
380
247
198
237
182
157
146
257
294
-
2532
Total "usable"
86
144
191
97
64
81
32
59
50
75
110
-
989
$ 5 464
$ 14 682
$ 4 896
$ 6 443
$ 7 953
$ 6 982
$ 9 434
$ 10 559
$ 14 344
$ 16 050
-
$ 100 335
$ 4 783
$ 11 549
$ 3 357
$ 1 028
$ 3 767
$ 1 540
$ 5 294
$
$
$
Total new issue value
$ 3 529
(in million of CAD)
Total "usable" value (in
$ 3 481
million of CAD)
4 161
4 715
6 642
Usable IPO is given by the total new issues minus IPO issues classified as "Capital pools", "Structured & Financial Products", and "Utilities" by FP-Infomart
$
50 318
29
Table II. Basic statistics of the sample data set
For the 180 firms in our sample (see Appendix B for the complete list), we present for each variable mentioned in the paper the number of
available data points, the mean, standard deviation and the value of the non-dummy variables for different distribution points. The table
seperates the different variables with respect to the categories in which they were presented in the paper.
Variable
Obs
Dependent variables
First year return
172
First year excess return
172
First day return
163
Two-year return
171
Three-year return
163
Volatility
174
sharpe
169
Mean
Std. Dev.
Min
5%
25%
Median
75%
95%
Max
0,115
0,034
0,044
0,276
0,218
0,424
0,572
0,534
0,519
0,100
1,077
1,151
0,320
1,400
-0,923
-0,926
-0,152
-0,970
-0,987
0,128
-1,479
-0,654
-0,762
-0,050
-0,861
-0,925
0,144
-1,074
-0,180
-0,260
-0,009
-0,364
-0,521
0,214
-0,453
0,075
-0,030
0,018
0,075
-0,034
0,314
0,220
0,314
0,234
0,083
0,544
0,533
0,495
1,241
1,238
1,058
0,503
2,414
2,355
1,063
3,133
2,592
2,111
0,563
7,000
5,800
2,600
7,307
Main independent variables
Rate-on-line (per $1000)
96
6,829
ln_ROL (per $1000)
96
1,700
Purchase
180
0,689
4,581
0,687
0,464
1,000
0,000
1,750
0,560
3,326
1,202
5,272
1,662
9,870
2,289
15,129
2,717
22,336
3,106
Financial variables
Market_return
180
lnMVE_IPO
180
IPOfeerat
163
Growth
162
Debt_Ratio
162
0,088
19,096
0,031
1,698
0,381
0,181
0,939
0,019
2,383
0,242
-0,372
16,369
0,001
0,057
0,005
-0,238
17,842
0,007
0,365
0,011
-0,053
18,445
0,014
0,802
0,192
0,129
18,958
0,026
1,037
0,354
0,216
19,575
0,050
1,522
0,532
0,318
20,741
0,060
4,567
0,868
0,563
22,384
0,062
16,908
0,944
Governance variables
Duality
180
Blockholder
180
Independence
170
ITCE
180
0,306
0,767
0,694
0,456
0,462
0,424
0,160
0,499
0,182
0,429
0,600
0,667
0,800
1,000
1,000
Other variables
Board_size
176
Risky_Industry
180
Age
176
Float
175
US_Presence
180
US_Sales
162
7,222
0,356
28,211
0,538
0,572
0,228
2,278
0,480
31,217
0,332
0,496
0,310
3,000
4,000
6,000
7,000
8,000
11,000
18,000
0,003
0,048
1,155
0,137
6,022
0,231
16,099
0,450
41,205
0,898
102,686
1,000
135,140
1,023
0,000
0,000
0,000
0,028
0,448
0,839
1,000
29
Table III Panel A. Separation and test between firms that have D&O insurance or not
We test for differences in means and median between the sample of firms that purchased D&O insurance (124 firms) and firms that did not purchase
D&O insurance (56 firms). A selection of independent control variables are presented
Return Variables
First year return
First year excess return
Two-year return
First day return
Volatility
Sharpe (return/volatility)
Sharpexs (Excess return/volatility)
Obs
51
51
51
47
53
50
50
No D&O insurance (56)
Mean
Std. Dev. Median
0,055
0,590
-0,002
-0,016
0,557
-0,078
0,157
1,108
-0,045
0,029
0,088
0,013
0,439
0,421
0,274
0,406
1,372
0,035
0,138
1,391
-0,280
Obs
121
121
120
118
121
119
119
D&O insurance (124)
Mean
Std. Dev. Median
0,141
0,509
0,090
0,055
0,503
0,006
0,327
1,064
0,111
0,043
0,124
0,021
0,417
0,265
0,327
0,641
1,412
0,294
0,343
1,438
0,026
Tests* of differences in
Mean
Median
ns
ns
ns
ns
ns
5,87%
ns
ns
ns
ns
ns
ns
ns
ns
Other independent variables
Ln (MVE at IPO)
Debt_Ratio (%)
Ipo fees
IPO fees / MVE at IPO
Risky industry
US Presence
US sales
56
55
46
46
56
56
56
18,894
0,327
6,919
0,036
0,321
0,446
0,136
0,996
0,234
6,264
0,020
0,471
0,502
0,263
18,885
0,316
4,587
0,037
0
0
0
124
124
117
117
124
124
124
19,188
0,406
7,646
0,030
0,371
0,629
0,270
0,902
0,242
9,206
0,018
0,485
0,485
0,321
18,964
0,384
5,159
0,025
0,000
1,000
0,082
3,12%
2,07%
ns
3,89%
ns
1,22%
0,20%
9,58%
4,61%
ns
9,23%
ns
2,23%
0,71%
Age
Float
Income trust
Independence
Duality
Blockholder
Growth
53
55
56
50
56
56
42
27,800
0,604
0,554
0,692
0,357
0,750
1,478
31,787
0,352
0,502
0,184
0,483
0,437
1,590
13,484
0,636
1
0,714
0
1,000
0,985
123
120
124
120
124
124
120
28,388
0,508
0,411
0,695
0,282
0,774
1,775
31,097
0,319
0,494
0,149
0,452
0,420
2,606
16,482
0,400
0
0,667
0
1
1,049
ns
4,34%
3,98%
ns
ns
ns
ns
ns
ns
7,68%
ns
ns
ns
ns
* A t‐stat was used to test the equality of the two sample means (with equal variance), whereas we used a Wilcoxon rank‐sum test for the test of equality of medians. Only differences significant at the 10% level or better are highlighted.
Table III Panel B. Marginal impact on the decision to purchase D&O insurance.
We evaluate the marginal impact of control variables on a firm's likelihood to purchase D&O insurance. The dependent variable is
Purchase, an indicator variable equal to one if the firm purchased D&O insurance in the first year post IPO. Onle the marginal
effects are reported for ease of interpretation of the estimated coefficients.
lnMVE_IPO is the log of the market value of equity at the time of the IPO. Debt_ratio is the ratio of total debt to market value of
equity at the time of the IPO. Growth is the market value of equity plus the book value of liability, divided by the book value of
assets. Duality takes on the value one if the CEO of the company is also the chairman of the board. Blockholder is an indicator
variable equal to one if a shareholder owns 10% or more of the firm’s voting shares. Independance is the proportion of directors
deamed independent in Canada. Risky_Industry is an indicator variable equal to one if the company operates in one of the
industries classified as risky in Bajaj et al. (2000). US_Presence is an indicator variable equal to 1 if the firm has activities in the
United States and zero otherwise. IPOfeerat is equal to the fees paid to the investment banker divided by the firm's market value
of equity at the time of the IPO. Float is the ratio of the number of shares available at the IPO to the total number of shares
outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust and 0 otherwise. Age is the age of the operating
firm at the time of the IPO. DO_Prospectus is equal to one if the firm revealed in its IPO prospectus that it had D&O insurance of
some sort.
VARIABLES
lnMVE_IPO
Debt_Ratio
Growth
Duality
Blockholder
Independence
Risky industry
US Presence
IPOfeerat
Float
ITCE
Age
(1)
Purchase
(2)
Purchase
-0.0147
(0.0289)
0.103***
(0.0215)
0.00417
(0.0120)
-0.123
(0.0884)
-0.0279
(0.0498)
-0.109
(0.164)
0.0343
(0.0437)
0.0752
(0.0630)
4.780
(5.665)
-0.317
(0.341)
-0.0458
(0.0754)
0.00122
(0.00103)
0.0385
(0.0385)
0.0981*
(0.0547)
(3)
Purchase
(4)
Purchase
(5)
Purchase
0.0368
(0.0328)
0.124***
(0.0348)
0.195**
(0.0772)
-0.00742
(0.0217)
0.0996***
(0.0228)
0.000320
(0.00904)
-0.0980
(0.0826)
-0.0217
(0.0388)
-0.133
(0.147)
0.0180
(0.0341)
0.0535
(0.0546)
2.428
(4.188)
-0.153
(0.250)
-0.0730
(0.0768)
0.000724
(0.000797)
0.0953
(0.0683)
0.1579
-76.727
151
0.1491
-65.312
137
-0.164*
(0.0869)
-0.00742
(0.0858)
-0.0464
(0.234)
0.0690
(0.0747)
0.142*
(0.0731)
7.694
(7.025)
-0.525
(0.416)
-0.0456
(0.0986)
-0.135*
(0.0793)
0.00102
(0.00128)
DO_Prospectus
0.111
(0.0690)
2.902
(5.547)
-0.216
(0.329)
-0.0939
(0.0882)
Constant
PseudoR2
LL
Observations
0.1164
-67.827
137
0.0955
-82.413
151
0.0341
-96.756
167
In all cases, the standard deviation is in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
Model 6 will be used later as our first stage regression in our Heckman two-stage regression
Table IV Panel A. OLS regression that measures the firms' first year total return
We evaluate the impact of price per unit of coverage on the firms' stock market return in the first year following the
IPO. The dependent variable is the first year total return (FirstYearReturn). The two main variables of interest are
Rate-on-line, calculated as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We
hypothesize that these two measures should be negatively related to the first year return.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the
first trading day. Mkt1Year is the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's
market value of equity at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is
the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed as the ratio of the
number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the
firm is an income trust. US_Sales is the percentage of sales carried out in the US. Independence is the proportion
of board members that are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of the
board and zero otherwise. Growth is the market value of equity plus the book value of liability, divided by the book
value of assets.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
First year
First year
First year
First year
First year
First year
First year
VARIABLES
total return total return total return total return total return total return total return
ln_ROL
-0.180**
(0.0792)
Rate-on-line
Mkt1Year
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.489*
(0.261)
0.906
(0.556)
0.132**
(0.0511)
9.266
(12.01)
-0.351
(0.694)
-0.155
(0.121)
-0.165
(0.127)
-0.233
(0.302)
-0.0934
(0.116)
-0.0553*
(0.0324)
-0.177**
(0.0807)
0.684**
(0.276)
-25.29**
(11.90)
0.671**
(0.283)
0.119*
(0.0693)
21.84
(16.11)
-0.960
(0.897)
0.000349
(0.180)
0.0533
(0.348)
-0.0131
(0.150)
-0.0192
(0.0247)
-0.158**
(0.0762)
0.114
(0.0696)
21.84
(15.81)
-0.969
(0.878)
-0.00481
(0.175)
0.587**
(0.261)
1.233**
(0.491)
0.111*
(0.0655)
20.44
(16.51)
-0.921
(0.908)
0.0294
(0.175)
-26.10**
(11.42)
0.573**
(0.267)
1.287**
(0.497)
0.106
(0.0653)
20.46
(16.21)
-0.933
(0.888)
0.0265
(0.166)
0.00505
(0.345)
-0.0413
(0.149)
-0.0209
(0.0251)
0.0505
(0.326)
-0.0155
(0.145)
-0.0184
(0.0201)
0.00518
(0.322)
-0.0449
(0.144)
-0.0202
(0.0204)
Year Dummy
Constant
-2.181*
(1.135)
-2.059
(1.391)
-2.045
(1.397)
-1.945
(1.314)
-1.914
(1.307)
0.819**
(0.376)
1.032*
(0.539)
0.0827
(0.0750)
24.91
(19.52)
-1.415
(1.113)
-0.0774
(0.196)
-0.199
(0.182)
0.300
(0.370)
-0.0316
(0.130)
-0.00690
(0.0175)
-23.59**
(11.23)
0.812**
(0.393)
1.066*
(0.552)
0.0794
(0.0742)
25.21
(19.24)
-1.436
(1.093)
-0.0776
(0.193)
-0.195
(0.184)
0.264
(0.364)
-0.0594
(0.133)
-0.00821
(0.0179)
YES
YES
-1.273
(1.493)
-1.273
(1.473)
Observations
143
83
83
83
83
83
83
R-squared
0.204
0.171
0.160
0.224
0.218
0.373
0.369
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at
the 1%, 5%, and 10% levels, respectively.
Table IV Panel B. OLS regression that measures the firms' first year excess return
We evaluate the impact of price per unit of coverage on the firms' market-adjusted stock return in the first year following the IPO. The dependent
variable is the first year total return minus the return of the market over the same year (FirstYearReturn - Mkt1Year). The two main variables of
interest are Rate-on-line, calculated as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two
measures should be negatively related to the firm's first year excess return.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. LnMVE_IPO is
the log of the firm's market value of equity at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is the ratio of
the IPO fees paid per million dollar of market value of equity. Float is computed as the ratio of the number of shares issued over the total number
of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust. US_Sales is the percentage of sales carried out in the
US. Independence is the proportion of board members that are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of
the board and zero otherwise. Growth is the market value of equity plus the book value of liability, divided by the book value of assets.
VARIABLES
(1)
(2)
(3)
(4)
(5)
(6)
(7)
First year excess First year excess First year excess First year excess First year excess First year excess First year excess
return
return
return
return
return
return
return
ln_ROL
-0.191**
(0.0777)
Rate-on-line
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.764
(0.539)
0.147***
(0.0536)
9.544
(12.76)
-0.332
(0.728)
-0.169
(0.117)
-0.128
(0.133)
-0.252
(0.293)
-0.0690
(0.118)
-0.0442
(0.0290)
1.010**
(0.446)
0.103
(0.0669)
17.98
(20.50)
-0.720
(1.138)
-0.0112
(0.172)
-0.210
(0.201)
0.0313
(0.330)
0.0341
(0.151)
-0.0169
(0.0180)
-0.192**
(0.0761)
-27.72**
(11.41)
1.067**
(0.445)
0.0983
(0.0669)
18.16
(20.21)
-0.741
(1.118)
-0.0135
(0.160)
-0.200
(0.205)
-0.0231
(0.325)
0.00298
(0.152)
-0.0183
(0.0185)
-0.160**
(0.0733)
1.101**
(0.451)
0.123*
(0.0678)
21.33
(17.53)
-0.932
(0.956)
0.0219
(0.168)
-28.36**
(11.01)
1.154**
(0.450)
0.117*
(0.0681)
21.39
(17.25)
-0.945
(0.937)
0.0184
(0.158)
-0.0114
(0.333)
0.0120
(0.151)
-0.0121
(0.0180)
-0.0635
(0.329)
-0.0189
(0.150)
-0.0138
(0.0184)
Year Dummy
Constant
-2.554**
(1.177)
-3.066**
(1.438)
-1.748
(1.342)
-3.495**
(1.446)
-2.141
(1.365)
1.007*
(0.519)
0.0866
(0.0744)
26.19
(19.24)
-1.472
(1.095)
-0.0862
(0.190)
-0.187
(0.178)
0.266
(0.375)
-0.0319
(0.130)
-0.00744
(0.0176)
-23.83**
(10.87)
1.040*
(0.530)
0.0835
(0.0739)
26.54
(18.98)
-1.496
(1.077)
-0.0868
(0.186)
-0.183
(0.180)
0.227
(0.367)
-0.0600
(0.132)
-0.00878
(0.0179)
YES
YES
-2.443
(1.545)
-1.341
(1.476)
Observations
143
83
83
83
83
83
83
R-squared
0.140
0.182
0.172
0.166
0.158
0.344
0.340
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table IV Panel C. OLS regression that measures the firms' first year volatility (standard deviation of stock market
return)
We evaluate the impact of price per unit of coverage on the firms' stock market volatility in the first year following the IPO.
The dependent variable is the standard deviation of the daily return during the first year. The two main variables of interest
are Rate-on-line, calculated as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that
these two measures should be positively related to volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading
day. Mkt1Year is the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity
at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid
per million dollar of market value of equity. Float is computed as the ratio of the number of shares issued over the total
number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust. US_Sales is the
percentage of sales carried out in the US. Risky_Industry is an indicator variable equal to one if the company operated in
one of the industries classified as risky in Bajaj et al. (2000). Age is the number of years separating inception from the
announcement of the IPO. DebtRatio is the ratio of total liabilities to the market value of equity at the time of the IPO.
Blockholder is equal to 1 if there is at least one independent blockholder and zero otherwise.
(1)
(2)
(3)
(4)
(5)
VARIABLES
ln_ROL
First year volatility
First year volatility
0.0261
(0.0272)
Rate-on-line
IPOfeerat
Float
ITCE
US_sales
RiskyIndustry
Age
DebtRatio
Blockholder
First year volatility
0.0357
(0.0270)
0.0637**
(0.0277)
6.025
(4.418)
Mkt1Year
lnMVE_IPO
First year volatility
-0.0690*
(0.0400)
-3.042
(3.085)
-0.0713
(0.167)
-0.196***
(0.0540)
0.177**
(0.0679)
0.0673
(0.0457)
-0.000721
(0.000528)
-0.292***
(0.0867)
0.0164
(0.0360)
-0.0689*
(0.0400)
-3.027
(3.119)
-0.0779
(0.168)
-0.198***
(0.0525)
0.176**
(0.0668)
0.0702
(0.0441)
-0.000758
(0.000505)
-0.289***
(0.0873)
0.00913
(0.0349)
12.23***
(3.782)
-0.187*
(0.104)
-0.0752**
(0.0369)
-3.131
(2.950)
-0.0767
(0.159)
-0.193***
(0.0538)
0.168**
(0.0665)
0.0593
(0.0434)
-0.000609
(0.000554)
-0.278***
(0.0814)
0.0269
(0.0346)
Year Dummy
Constant
2.127***
(0.778)
1.955**
(0.764)
First year volatility
2.305***
(0.706)
-0.0596*
(0.0315)
0.811
(3.066)
-0.210
(0.154)
-0.129**
(0.0599)
0.166***
(0.0571)
-0.00995
(0.0355)
-0.00104**
(0.000470)
-0.210***
(0.0680)
-0.0322
(0.0320)
-0.0578*
(0.0308)
1.117
(2.989)
-0.234
(0.150)
-0.129**
(0.0533)
0.164***
(0.0531)
-0.00597
(0.0331)
-0.00112**
(0.000468)
-0.204***
(0.0682)
-0.0441
(0.0302)
YES
YES
2.168***
(0.597)
1.731***
(0.598)
Observations
87
87
87
87
87
R-squared
0.647
0.652
0.665
0.770
0.782
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%,
5%, and 10% levels, respectively.
Table IV Panel D. OLS regression that measures the firms' Sharpe ratio in the first year
We evaluate the impact of price per unit of coverage on the firms' stock market volatility in the first year following the IPO.
The dependent variable is the quasi Sharpe ratio calculated as the first year total return divided by the standard deviation of
the daily return during the first year (FirstYearReturn / Volatility). The two main variables of interest are Rate-on-line,
calculated as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two
measures should be negatively related to the quasi-Sharpe ratio.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first
trading day. Mkt1Year is the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value
of equity at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is the ratio of the IPO
fees paid per million dollar of market value of equity. Float is computed as the ratio of the number of shares issued over the
total number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust. US_Sales is the
percentage of sales carried out in the US. Independence is the proportion of board members that are classified as
independent. Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market
value of equity plus the book value of liability, divided by the book value of assets. Risky_Industry is an indicator variable
equal to one if the company operated in one of the industries classified as risky in Bajaj et al. (2000). Age is the number of
years separating inception from the announcement of the IPO. DebtRatio is the ratio of total liabilities to the market value of
equity at the time of the IPO. Blockholder is equal to 1 if there is at least one independent blockholder and zero otherwise.
VARIABLES
(1)
Sharpe ratio
ln_ROL
(2)
Sharpe ratio
-0.430**
(0.214)
Rate-on-line
FirstDayReturn
Mkt1year
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
RiskyIndustry
Age
DebtRatio
Blockholder
2.360**
(1.179)
0.980*
(0.562)
0.384***
(0.139)
46.02
(28.64)
-2.787*
(1.586)
0.109
(0.397)
-0.717**
(0.327)
-0.826
(0.757)
-0.145
(0.315)
-0.0593
(0.0669)
0.214
(0.283)
0.00454
(0.00488)
0.816*
(0.461)
-0.969***
(0.357)
Year Dummy
Constant
(3)
Sharpe ratio
-5.779**
(2.874)
(4)
Sharpe ratio
(5)
Sharpe ratio
-0.315
(0.240)
3.618**
(1.541)
2.022**
(0.996)
-62.12*
(32.89)
3.743**
(1.585)
2.001*
(1.033)
YES
YES
-2.237*
(1.178)
0.425
(0.376)
2.437*
(1.420)
1.239*
(0.631)
0.315
(0.204)
67.60
(49.15)
-3.768
(2.661)
0.374
(0.574)
-0.975*
(0.509)
-0.269
(0.930)
0.126
(0.481)
-0.00583
(0.0513)
0.139
(0.426)
0.00168
(0.00902)
0.996
(0.673)
-0.788
(0.534)
(6)
Sharpe ratio
-0.279
(0.223)
-41.93
(34.44)
2.544*
(1.414)
1.195*
(0.631)
0.303
(0.202)
67.96
(48.67)
-3.824
(2.613)
0.364
(0.551)
-0.963*
(0.515)
-0.357
(0.916)
0.0755
(0.478)
-0.00693
(0.0510)
0.114
(0.426)
0.00178
(0.00908)
1.023
(0.676)
-0.789
(0.541)
2.267
(1.522)
2.377**
(1.064)
0.322
(0.230)
91.58*
(53.52)
-5.543*
(2.934)
-0.109
(0.678)
-0.840*
(0.461)
-0.0688
(1.157)
0.0443
(0.409)
0.00755
(0.0451)
0.260
(0.426)
0.00266
(0.00789)
0.842
(0.694)
-0.577
(0.527)
YES
-6.947
(4.263)
-4.701
(3.709)
-6.960
(4.534)
Observations
142
90
90
83
83
83
R-squared
0.262
0.208
0.203
0.256
0.250
0.365
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%,
5%, and 10% levels, respectively.
Table IV Panel E. OLS regression that measures the firms' Sharpe ratio (in excess return) in the first year
We evaluate the impact of price per unit of coverage on the firms' stock market volatility in the first year following the IPO. The
dependent variable is the market-adjusted quasi Sharpe ratio calculated as the market-adjuste first year return divided by the
standard deviation of the daily return during the first year (FirstYearExcessReturn / Volatility).The two main variables of interest
are Rate-on-line, calculated as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that
these two measures should be negatively related to the market-adjusted quasi-Sharpe ratio.
LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price multiplied by number of
outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed
as the ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if
the firm is an income trust. US_Sales is the percentage of sales carried out in the US. Independence is the proportion of
board members that are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of the board and zero
otherwise. Growth is the market value of equity plus the book value of liability, divided by the book value of assets.
Risky_Industry is an indicator variable equal to one if the company operated in one of the industries classified as risky in Bajaj
et al. (2000). Age is the number of years separating inception from the announcement of the IPO. DebtRatio is the ratio of total
liabilities to the market value of equity at the time of the IPO. Blockholder is equal to 1 if there is at least one independent
blockholder and zero otherwise.
(1)
(2)
(3)
(4)
(5)
VARIABLES
Excess Sharpe ratio Excess Sharpe ratio Excess Sharpe ratio Excess Sharpe ratio Excess Sharpe ratio
ln_ROL
-0.383*
(0.213)
-0.536***
(0.198)
-0.455**
(0.218)
Rate-on-line
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
RiskyIndustry
Age
DebtRatio
Blockholder
0.355**
(0.141)
43.17
(32.17)
-2.658
(1.776)
-0.165
(0.396)
-0.447
(0.373)
-0.930
(0.783)
-0.120
(0.340)
-0.0577
(0.0636)
0.187
(0.301)
0.00519
(0.00521)
0.610
(0.482)
-1.070***
(0.391)
Year Dummy
Constant
-5.040*
(2.953)
0.266
(0.218)
88.14*
(46.22)
-5.299*
(2.692)
-0.379
(0.692)
-0.846*
(0.494)
0.178
(1.137)
0.0734
(0.416)
-0.00613
(0.0536)
0.414
(0.403)
0.00139
(0.00766)
0.705
(0.676)
-0.512
(0.477)
YES
YES
-6.469
(4.353)
-2.465**
(1.062)
0.279
(0.200)
67.83
(52.11)
-3.446
(2.860)
0.0589
(0.602)
-0.926*
(0.545)
-0.469
(1.002)
0.231
(0.505)
-0.0257
(0.0555)
0.291
(0.427)
-0.00125
(0.00897)
1.007
(0.692)
-0.807
(0.527)
-59.25*
(34.23)
0.265
(0.198)
68.50
(51.52)
-3.522
(2.801)
0.0383
(0.581)
-0.913
(0.554)
-0.612
(0.992)
0.164
(0.505)
-0.0264
(0.0552)
0.262
(0.431)
-0.00109
(0.00911)
1.052
(0.695)
-0.811
(0.538)
-6.982*
(4.102)
-3.814
(3.676)
Observations
142
83
92
83
83
R-squared
0.196
0.365
0.168
0.193
0.181
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%,
and 10% levels, respectively.
Table IV Panel F. OLS regression that measures the firms' first year return statistics as a function of the information unknown at the
time of the IPO
We evaluate the impact of price per unit of coverage on the firms' stock market volatility in the first year following the IPO. The dependent
variable is one of the five return distribution statistics used in the paper (total first year return, first year excess return, first year volatility, first
year quasi-Sharpe ratio and first year excess quasi-Sharpe ratio). The two main variables of interest are Rate-on-line, calculated as the
premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively related to
the return and Sharpe ratio measures, and positively related to volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO. The Year_Dummy variables are there to take into account everything else that occured in the
year of the IPO.
(1)
VARIABLES
ln_ROL
(3)
First year
First year First year
excess
total return total return
return
-0.221***
(0.0746)
Rate-on-line
Mkt1Year
FirstDayReturn
Year Dummy
Constant
(2)
(4)
First year
excess
return
-0.221***
(0.0732)
(5)
(6)
(7)
(8)
First year
volatility
First year
volatility
Sharpe
ratio
Sharpe
ratio
0.0689*
(0.0349)
0.882**
(0.386)
1.071**
(0.470)
-30.17***
(10.23)
0.894**
(0.413)
1.016*
(0.514)
-30.24***
(10.02)
1.073**
(0.457)
1.018**
(0.501)
YES
YES
YES
YES
YES
-1.097***
(0.411)
0.275*
(0.161)
-1.101***
(0.404)
0.272*
(0.162)
0.909***
(0.199)
-0.430**
(0.214)
11.48*
(6.485)
(9)
Excess
Sharpe
ratio
(10)
Excess
Sharpe
ratio
-0.509**
(0.210)
2.022**
(0.996)
3.618**
(1.541)
-62.12*
(32.89)
2.001*
(1.033)
3.743**
(1.585)
-73.00**
(31.58)
3.817**
(1.577)
3.960**
(1.620)
YES
YES
YES
YES
YES
0.474***
(0.0450)
-2.237*
(1.178)
0.425
(0.376)
-2.646**
(1.144)
0.503
(0.375)
Observations
91
91
91
91
94
94
90
90
90
90
R-squared
0.226
0.208
0.206
0.188
0.477
0.483
0.208
0.203
0.225
0.218
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table IV Panel G. OLS regression that measures the firms' first year return statistics as a function of the information unknown at the
time of the IPO: No year dummy
We evaluate the impact of price per unit of coverage on the firms' stock market volatility in the first year following the IPO. The dependent
variable is one of the five return distribution statistics used in the paper (total first year return, first year excess return, first year volatility, first
year quasi-Sharpe ratio and first year excess quasi-Sharpe ratio). The two main variables of interest are Rate-on-line, calculated as the
premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively related to
the return and Sharpe ratio measures, and positively related to volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO.
(1)
VARIABLES
ln_ROL
(3)
First year
First year First year
excess
total return total return
return
-0.189**
(0.0734)
Rate-on-line
Mkt1Year
FirstDayReturn
Year Dummy
Constant
(2)
(4)
First year
excess
return
-0.206***
(0.0701)
(5)
(6)
(7)
(8)
First year
volatility
First year
volatility
Sharpe
ratio
Sharpe
ratio
-0.0147
(0.0405)
0.632**
(0.288)
0.994**
(0.455)
-25.17**
(10.72)
0.635**
(0.298)
0.947*
(0.503)
-28.48***
(10.28)
0.962**
(0.407)
0.915**
(0.450)
No
No
No
No
No
-0.961**
(0.376)
0.195*
(0.102)
-1.079***
(0.354)
0.189*
(0.101)
0.363*
(0.218)
-0.370*
(0.216)
-2.551
(7.829)
(9)
Excess
Sharpe
ratio
(10)
Excess
Sharpe
ratio
-0.547**
(0.210)
1.080
(0.655)
3.179**
(1.497)
-53.36
(33.69)
1.078
(0.669)
3.304**
(1.539)
-81.18**
(31.67)
3.024**
(1.353)
3.214**
(1.382)
No
No
No
No
No
0.457***
(0.0561)
-1.593
(1.100)
0.683**
(0.294)
-2.715**
(1.049)
0.670**
(0.300)
Observations
91
91
91
91
94
94
90
90
90
90
R-squared
0.149
0.131
0.107
0.090
0.001
0.002
0.096
0.092
0.099
0.093
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table V Panel A. OLS regression that measures the firms' second year return
We evaluate the impact of price per unit of coverage on the firms' stock market return in the second year
following the IPO. The dependent variable is the two-year total return or the second year return (given by the twoyear total return minus the first year total return). The two main variables of interest are Rate-on-line, calculated
as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two
measures should be negatively related to either measure of two-year return.
FirstYearReturn is our control variable the first year return. LnMVE_IPO is the log of the firm's market value of
equity at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is the ratio of
the IPO fees paid per million dollar of market value of equity. Float is computed as the ratio of the number of
shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an
income trust. Independence is the proportion of board members that are classified as independent. Duality is
equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market value of equity
plus the book value of liability, divided by the book value of assets.
(1)
(2)
(3)
(4)
(5)
(6)
Two-year total Two-year total Two-year total Two-year total Second year Second year
VARIABLES
return
return
return
return
return
return
ln_ROL
-0.00940
(0.187)
0.221
(0.162)
Rate-on-line
0.751
(26.42)
0.102
(0.140)
34.08
(35.65)
-1.714
(1.874)
-0.245
(0.305)
-0.429
(1.418)
0.130
(0.526)
-0.0937**
(0.0362)
1.609***
(0.260)
-0.0541
(0.0879)
2.106
(20.51)
-0.215
(0.898)
-0.244
(0.197)
-0.746
(1.305)
0.236
(0.437)
-0.0440
(0.0294)
0.102
(0.140)
33.94
(35.74)
-1.707
(1.880)
-0.247
(0.306)
-0.436
(1.410)
0.131
(0.528)
-0.0934**
(0.0363)
32.47
(25.04)
1.599***
(0.252)
-0.0468
(0.0857)
2.196
(20.46)
-0.208
(0.905)
-0.239
(0.196)
-0.687
(1.300)
0.271
(0.439)
-0.0421
(0.0288)
-1.213
(3.181)
3.030
(2.591)
-1.169
(2.927)
1.458
(1.790)
FirstYearReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
Independence
Duality
Growth
Constant
0.134
(0.147)
20.59
(22.48)
0.00491
(0.0895)
14.21
(23.48)
-0.783
(1.138)
-0.244
(0.190)
-0.626
(1.325)
0.196
(0.449)
-0.0628***
(0.0230)
0.00900
(0.0887)
14.09
(23.62)
-0.770
(1.151)
-0.242
(0.193)
-0.593
(1.318)
0.218
(0.448)
-0.0614***
(0.0225)
1.423
(2.426)
0.474
(1.941)
Observations
82
82
82
82
82
82
R-squared
0.054
0.481
0.054
0.479
0.068
0.068
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance
at the 1%, 5%, and 10% levels, respectively.
Table VI Panel A. Two-step regression that measures the first year's total stock return by controling for the purchase of insurance
We evaluate the impact of price per unit of coverage on the firms' total stock market return in the first year following the IPO controlling for the
information imbeded in the purchase of insurance or not (model 6 in Table 3B). The dependent variable is the total stock market return (in
models 1 and 2) and the return in excess of the market (in models 3 and 4). The two main variables of interest are Rate-on-line, calculated as
the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively
related to the first year total return.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price
multiplied by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is
computed as the ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the
firm is an income trust. US_Sales is the percentage of sales carried out in the US. Independence is the proportion of board members that
are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market
value of equity plus the book value of liability, divided by the book value of assets. Debt_ratio is the ratio of total debt to market value of equity
at the time of the IPO. US_Presence is an indicator variable equal to 1 if the firm has activities in the United States and zero otherwise.
IPO_fee_ratio is equal to the fees paid to the investment banker divided by the firm's market value of equity at the time of the IPO.
DO_Prospectus is equal to one if the firm revealed in its proxy that it had D&O insurance of some sort.
VARIABLES
ln_ROL
(1A)
First year total
return
(1B)
Purchase
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.843**
(0.354)
1.030**
(0.472)
0.0798
(0.0680)
24.35
(17.21)
-1.363
(0.996)
-0.0605
(0.168)
-0.215
(0.169)
0.262
(0.357)
-0.0333
(0.117)
-0.00687
(0.0157)
Debt Ratio
DO_Prospectus
Constant
Purchase
0.0274
(0.171)
8.245
(22.46)
-0.716
(1.298)
-0.580
(0.424)
YES
-2.251
(1.451)
(3A)
First year
excess return
(3B)
Purchase
(4A)
First year
excess return
(4B)
Purchase
-0.162**
(0.0672)
-23.83**
(10.01)
0.838**
(0.366)
1.065**
(0.483)
0.0762
(0.0673)
24.61
(16.89)
-1.381
(0.969)
-0.0588
(0.165)
-0.213
(0.167)
0.221
(0.346)
-0.0614
(0.120)
-0.00820
(0.0160)
0.783
(0.645)
0.380
(0.264)
0.634**
(0.267)
US Presence
Year Dummy
(2B)
-0.159**
(0.0675)
Rate-on-line
Mkt1Year
(2A)
First year total
return
-23.41**
(10.71)
0.0277
(0.171)
8.349
(22.33)
-0.720
(1.292)
-0.582
(0.420)
0.792
(0.636)
0.378
(0.261)
0.634**
(0.266)
YES
-0.505
(3.370)
0.0867
(0.0697)
25.37
(16.84)
-1.357
(0.991)
-0.0721
(0.181)
-0.261
(0.173)
0.258
(0.365)
-0.0300
(0.124)
-0.00775
(0.0190)
-1.140
(1.383)
0.0271
(0.171)
8.243
(22.12)
-0.718
(1.289)
-0.579
(0.416)
0.786
(0.628)
0.383
(0.258)
0.630**
(0.268)
YES
-0.511
(3.375)
0.0838
(0.0695)
25.69
(16.53)
-1.376
(0.966)
-0.0724
(0.181)
-0.260
(0.172)
0.219
(0.356)
-0.0579
(0.125)
-0.00897
(0.0193)
-2.326
(1.500)
0.0270
(0.171)
8.236
(22.01)
-0.718
(1.284)
-0.580
(0.412)
0.792
(0.624)
0.381
(0.257)
0.631**
(0.266)
YES
-0.499
(3.373)
-1.217
(1.446)
-0.497
(3.376)
Observations
127
127
127
127
127
127
127
127
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10%
levels, respectively.
Table VI Panel B. Two-step regression that measures the first year's volatility and the first year's excess quasi-Sharpe ratio by controling for
the purchase of insurance
We evaluate the impact of price per unit of coverage on the firms' stock market volatility and quasi-Sharpe ratio in the first year following the IPO
controlling for the information imbeded in the purchase of insurance or not (Model 6 in Table 3B). The dependent variable is the volatility of the first
year's return and the excess quasi-Sharpe ratio. The two main variables of interest are Rate-on-line, calculated as the premium-to-coverage ratio, and
ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be positively related to the volatility and negatively to the Sharpe
ratio.
LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price multiplied by number of outstanding shares). IPOfeerat is
the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed as the ratio of the number of shares issued over the total
number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust. US_Sales is the percentage of sales carried out in
the US. Risky_Industry is an indicator variable equal to one if the company operated in one of the industries classified as risky in Bajaj et al. (2000).
Age is the number of years separating inception from the announcement of the IPO. DebtRatio is the ratio of total liabilities to market value of equity at
the time of the IPO. Blockholder is equal to 1 if there is at least one independent blockholder and zero otherwise. US_Presence is an indicator variable
equal to 1 if the firm has activities in the United States and zero otherwise. DO_Prospectus is equal to one if the firm revealed in its proxy that it had
D&O insurance of some sort.
(1A)
(1B)
(2A)
(2B)
(3A)
(3B)
(4A)
(4B)
(5A)
(5B)
VARIABLES
Volatility
Purchase
Volatility
Purchase
Excess
Sharpe ratio
Purchase
Excess
Sharpe ratio
Purchase
Excess
Sharpe ratio
Purchase
ln_ROL
0.0638***
(0.0233)
-0.384**
(0.185)
Rate-on-line
12.24***
(3.414)
-47.72
(32.08)
(Rate-on-line)^2
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
RiskyIndustry
Age
DebtRatio
Blockholder
-0.0587***
(0.0215)
0.974
(3.636)
-0.221
(0.208)
-0.133**
(0.0586)
0.170***
(0.0515)
-0.00827
(0.0357)
-0.00104*
(0.000594)
-0.205***
(0.0740)
-0.0338
(0.0427)
0.0677
(0.145)
0.0104
(24.91)
-0.165
(1.454)
-0.543
(0.405)
0.716
(0.533)
-0.0568***
(0.0209)
1.278
(3.540)
-0.245
(0.203)
-0.133**
(0.0570)
0.167***
(0.0501)
-0.00431
(0.0347)
-0.00112*
(0.000579)
-0.199***
(0.0720)
-0.0456
(0.0419)
0.0677
(0.145)
0.0104
(24.91)
-0.165
(1.454)
-0.543
(0.405)
0.716
(0.533)
Independence
Duality
Growth
US Presence
0.363
(0.250)
0.594**
(0.254)
DO_Prospectus
Year Dummy
Constant
YES
2.125***
(0.444)
0.363
(0.250)
0.594**
(0.254)
YES
1.531
(2.965)
0.265
(0.195)
87.72**
(41.85)
-5.270**
(2.424)
-0.372
(0.601)
-0.853**
(0.430)
0.411
(0.353)
0.00136
(0.00654)
0.693
(0.624)
-0.509
(0.421)
0.172
(0.982)
0.0733
(0.361)
-0.00630
(0.0469)
1.695***
(0.437)
0.0247
(0.167)
7.574
(22.67)
-0.706
(1.337)
-0.563
(0.407)
0.717
(0.588)
0.395
(0.257)
0.646**
(0.265)
YES
1.531
(2.965)
0.258
(0.196)
88.93**
(41.35)
-5.337**
(2.374)
-0.380
(0.603)
-0.861**
(0.436)
0.378
(0.353)
0.00137
(0.00665)
0.733
(0.620)
-0.505
(0.428)
0.0778
(0.978)
0.00775
(0.363)
-0.00728
(0.0472)
-6.416
(3.956)
0.0246
(0.167)
7.565
(22.67)
-0.705
(1.337)
-0.564
(0.407)
0.717
(0.589)
0.394
(0.259)
0.646**
(0.266)
YES
-0.437
(3.301)
-25.51***
(9.367)
1.163**
(0.473)
0.299
(0.200)
83.88*
(42.84)
-5.026**
(2.490)
-0.260
(0.589)
-0.944**
(0.423)
0.495
(0.370)
0.00168
(0.00641)
0.527
(0.625)
-0.506
(0.416)
0.357
(1.034)
0.164
(0.375)
-0.00135
(0.0477)
-3.834
(3.906)
0.0241
(0.167)
7.609
(22.62)
-0.707
(1.334)
-0.567
(0.407)
0.717
(0.590)
0.391
(0.258)
0.650**
(0.264)
YES
-0.435
(3.302)
-4.221
(3.957)
-0.423
(3.302)
Observations
134
134
134
134
127
127
127
127
127
127
Coefficients are reported with their standard deviation in parentheses. Heckman two-stage regression reported in models 1 and 2 since the MLE
regressions are not converging. ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.
Table VI Panel C. Two-step regression that measures the first year's excess return and excess quasi-Sharpe ratio by controling for
the purchase of insurance
We evaluate the impact of price per unit of coverage on the firms' market-adjusted stock return (in models 1, 2 and 3) and on the marketadjusted quasi-Sharpe ratio (in models 3 and 4) in the first year following the IPO controlling for the information imbeded in the purchase of
insurance or not (Model 6 in Table 3B). The two main variables of interest are Rate-on-line, calculated as the premium-to-coverage ratio, and
ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively related firms' market-adjusted stock return
and their market-adjusted quasi-Sharpe ratio.
LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price multiplied by number of outstanding shares).
IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed as the ratio of the number of shares
issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust. US_Sales is the
percentage of sales carried out in the US. Independence is the proportion of board members that are classified as independent. Duality is
equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market value of equity plus the book value of liability,
divided by the book value of assets. DebtRatio is the ratio of total liabilities to the market value of equity at the time of the IPO. Risky_Industry
is an indicator variable equal to one if the company operated in one of the industries classified as risky in Bajaj et al. (2000). Age is the number
of years separating inception from the announcement of the IPO. Blockholder is equal to 1 if there is at least one independent blockholder and
zero otherwise. US_Presence is an indicator variable equal to 1 if the firm has activities in the United States and zero otherwise.
DO_Prospectus is equal to one if the firm revealed in its proxy that it had D&O insurance of some sort.
VARIABLES
ln_ROL
(1A)
First year
excess
return
(1B)
Purchase
-0.162**
(0.0672)
(2A)
First year
excess
return
(2B)
Purchase
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.0867
(0.0697)
25.37
(16.84)
-1.357
(0.991)
-0.0721
(0.181)
-0.261
(0.173)
0.258
(0.365)
-0.0300
(0.124)
-0.00775
(0.0190)
DebtRatio
(3B)
Purchase
-0.191***
(0.0721)
0.0271
(0.171)
8.243
(22.12)
-0.718
(1.289)
-0.579
(0.416)
0.0984
(0.0648)
17.30
(18.70)
-0.627
(1.055)
-0.0150
(0.166)
-0.279
(0.193)
0.00617
(0.327)
0.0360
(0.145)
-0.0197
(0.0200)
0.0229
(0.171)
7.889
(21.81)
-0.686
(1.286)
-0.595
(0.415)
-23.41**
(10.71)
0.0838
(0.0695)
25.69
(16.53)
-1.376
(0.966)
-0.0724
(0.181)
-0.260
(0.172)
0.219
(0.356)
-0.0579
(0.125)
-0.00897
(0.0193)
0.0270
(0.171)
8.236
(22.01)
-0.718
(1.284)
-0.580
(0.412)
0.786
(0.628)
0.814
(0.606)
0.792
(0.624)
0.383
(0.258)
0.630**
(0.268)
0.359
(0.261)
0.650**
(0.257)
0.381
(0.257)
0.631**
(0.266)
RiskyIndustry
Age
Blockholder
US Presence
DO_Prospectus
Dummy Year
Constant
YES
-2.326
(1.500)
YES
-0.499
(3.373)
-2.848**
(1.394)
(4A)
Excess
Sharpe
ratio
(4B)
Purchase
-0.384**
(0.185)
Rate-on-line
lnMVE_IPO
(3A)
First year
excess
return
-0.427
(3.372)
-1.217
(1.446)
0.265
(0.195)
87.72**
(41.85)
-5.270**
(2.424)
-0.372
(0.601)
-0.853**
(0.430)
0.172
(0.982)
0.0733
(0.361)
-0.00630
(0.0469)
0.693
(0.624)
0.411
(0.353)
0.00136
(0.00654)
-0.509
(0.421)
(5A)
Excess
Sharpe
ratio
(5B)
Purchase
-0.456**
(0.199)
0.0247
(0.167)
7.574
(22.67)
-0.706
(1.337)
-0.563
(0.407)
0.717
(0.588)
0.275
(0.187)
66.88
(48.98)
-3.379
(2.694)
0.0766
(0.554)
-0.941*
(0.499)
-0.482
(0.908)
0.231
(0.462)
-0.0262
(0.0513)
0.980
(0.669)
0.282
(0.395)
-0.00131
(0.00812)
-0.800
(0.491)
0.395
(0.257)
0.646**
(0.265)
0.0238
(0.167)
7.447
(22.64)
-0.695
(1.338)
-0.568
(0.411)
0.722
(0.592)
0.387
(0.272)
0.651**
(0.269)
YES
-0.497
(3.376)
-6.416
(3.956)
-0.437
(3.301)
-6.861*
(3.894)
-0.419
(3.304)
Observations
127
127
127
127
127
127
127
127
127
127
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table VII Panel A. Simultaneous regressions for the first year total return and volatility, and the first year excess return and volatility
We evaluate simultaneously the impact of price per unit of coverage on the firms' stock market return in the first year following the IPO. The
dependent variables are the first year total return or the first year excess return, and the first year volatility. The main variables of interest is ln_ROL,
calculcated as the log of the premium to coverage ratio. We hypothesize that this variable should be negatively related to the first year return, the
first year excess return and positively to the first year volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is the
market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price multiplied
by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed as the
ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust.
US_Sales is the percentage of sales carried out in the US. Independence is the proportion of board members that are classified as independent.
Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market value of equity plus the book value of
liability, divided by the book value of assets. DebtRatio is the ratio of total liabilities to the market value of equity at the time of the IPO.
Risky_Industry is an indicator variable equal to one if the company operated in one of the industries classified as risky in Bajaj et al. (2000). Age is
the number of years separating inception from the announcement of the IPO. Blockholder is equal to 1 if there is at least one independent
blockholder and zero otherwise.
VARIABLES
ln_ROL
(1a)
(1b)
Two-stage least square
First year total
Volatility
return
-0.160**
(0.0744)
(2a)
(2b)
Three-stage least square
First year excess
Volatility
return
(3a)
(3b)
Three-stage least square
First year excess
Volatility
return
0.0614**
(0.0243)
-0.162**
(0.0763)
0.0612**
(0.0243)
-0.192**
(0.0772)
0.0632***
(0.0242)
-0.0578***
(0.0212)
2.085
(3.712)
-0.319
(0.216)
-0.122**
(0.0564)
0.179***
(0.0498)
0.0899
(0.0652)
25.44**
(11.62)
-1.396**
(0.658)
-0.0867
(0.176)
-0.245
(0.158)
0.383
(0.333)
-0.0104
(0.114)
-0.00736
(0.0220)
-0.0577***
(0.0212)
1.996
(3.709)
-0.312
(0.216)
-0.123**
(0.0564)
0.179***
(0.0498)
0.102
(0.0696)
17.40
(12.54)
-0.671
(0.698)
-0.0357
(0.171)
-0.257
(0.167)
0.147
(0.358)
0.0499
(0.123)
-0.0191
(0.0227)
-0.0591***
(0.0211)
2.612
(3.705)
-0.363*
(0.215)
-0.129**
(0.0559)
0.179***
(0.0496)
Rate-on-line
Mkt1year
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.939***
(0.315)
0.909*
(0.469)
0.0850
(0.0639)
25.04**
(11.53)
-1.420**
(0.648)
-0.0779
(0.173)
-0.201
(0.157)
0.358
(0.332)
-0.0155
(0.112)
-0.00678
(0.0216)
DebtRatio
-0.208***
(0.0663)
-0.00665
(0.0335)
-0.00105*
(0.000588)
-0.0489
(0.0417)
RiskyIndustry
Age
Blockholder
Year Dummy
Constant
-0.206***
(0.0659)
-0.00356
(0.0332)
-0.00103*
(0.000583)
-0.0459
(0.0414)
-0.209***
(0.0666)
-0.00398
(0.0336)
-0.00105*
(0.000590)
-0.0432
(0.0419)
YES
YES
YES
YES
NO
YES
-2.459*
(1.319)
2.132***
(0.435)
-2.513*
(1.347)
2.125***
(0.434)
-3.077**
(1.461)
2.181***
(0.434)
Observations
83
83
83
83
83
83
R-squared
0.371
0.764
0.310
0.764
0.144
0.761
Coefficients are reported with their standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table VII Panel B. Simultaneous regressions for the first year total return and volatility, and the first year excess return and volatility
We evaluate simultaneously the impact of price per unit of coverage on the firms' stock market return in the first year following the IPO. The
dependent variables are the first year total return or the first year excess return, and the first year volatility. The main variables of interest is Rate-online, calculated as the insurance premium per $1000 of coverage. We hypothesize that this variable should be negatively related to the first year
return, the first year excess return and positively to the first year volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is the
market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price multiplied
by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is computed as the
ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if the firm is an income trust.
US_Sales is the percentage of sales carried out in the US. Independence is the proportion of board members that are classified as independent.
Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the market value of equity plus the book value of
liability, divided by the book value of assets. DebtRatio is the ratio of total liabilities to the market value of equity at the time of the IPO.
Risky_Industry is an indicator variable equal to one if the company operated in one of the industries classified as risky in Bajaj et al. (2000). Age is
the number of years separating inception from the announcement of the IPO. Blockholder is equal to 1 if there is at least one independent
blockholder and zero otherwise.
VARIABLES
(1a)
(1b)
Two-stage least square
First year total
Volatility
return
(2a)
(2b)
Three-stage least square
First year excess
Volatility
return
(3a)
(3b)
Three-stage least square
First year excess
Volatility
return
ln_ROL
Rate-on-line
Mkt1year
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
-23.40**
(11.81)
0.932***
(0.317)
0.930**
(0.472)
0.0820
(0.0643)
25.15**
(11.57)
-1.432**
(0.651)
-0.0760
(0.173)
-0.199
(0.157)
0.334
(0.334)
-0.0319
(0.112)
-0.00696
(0.0218)
DebtRatio
Age
Blockholder
Constant
-22.85*
(12.14)
12.62***
(3.734)
-26.37**
(12.26)
12.84***
(3.709)
-0.0537***
(0.0207)
2.292
(3.616)
-0.327
(0.210)
-0.126**
(0.0549)
0.176***
(0.0485)
0.0875
(0.0657)
25.55**
(11.67)
-1.406**
(0.661)
-0.0855
(0.177)
-0.245
(0.158)
0.360
(0.335)
-0.0250
(0.115)
-0.00730
(0.0222)
-0.0536***
(0.0207)
2.212
(3.613)
-0.321
(0.210)
-0.127**
(0.0549)
0.175***
(0.0485)
0.0979
(0.0702)
17.30
(12.65)
-0.676
(0.705)
-0.0373
(0.172)
-0.252
(0.169)
0.106
(0.359)
0.0327
(0.125)
-0.0195
(0.0230)
-0.0547***
(0.0206)
2.823
(3.608)
-0.370*
(0.210)
-0.133**
(0.0545)
0.176***
(0.0483)
-0.205***
(0.0646)
-0.00352
(0.0326)
-0.00105*
(0.000573)
-0.0568
(0.0409)
RiskyIndustry
Year Dummy
12.65***
(3.734)
-0.204***
(0.0642)
-0.000502
(0.0324)
-0.00103*
(0.000569)
-0.0540
(0.0405)
-0.207***
(0.0648)
-0.000680
(0.0327)
-0.00105*
(0.000575)
-0.0516
(0.0410)
YES
YES
YES
YES
NO
YES
-1.374
(1.297)
1.652***
(0.414)
-1.436
(1.329)
1.647***
(0.413)
-1.780
(1.417)
1.687***
(0.412)
Observations
83
83
83
83
83
83
R-squared
0.367
0.776
0.303
0.776
0.130
0.773
Coefficients are reported with their standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table VIII Panel A. Simultaneous regressions for the first year total return and volatility, and the first year excess return and
volatility controling for the decision to purchase insurance or not
We evaluate simultaneously the impact of price per unit of coverage on the firms' stock market return in the first year following the IPO. The
dependent variables are the first year total return and the first year volatility. The two main variables of interest are Rate-on-line, calculated
as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that both measures should be negatively related
to the first year return and positively to the first year volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price
multiplied by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is
computed as the ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if
the firm is an income trust. US_Sales is the percentage of sales carried out in the US. Independence is the proportion of board members
that are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the
market value of equity plus the book value of liability, divided by the book value of assets. DebtRatio is the ratio of total liabilities to the
market value of equity at the time of the IPO. Risky_Industry is an indicator variable equal to one if the company operated in one of the
industries classified as risky in Bajaj et al. (2000). Age is the number of years separating inception from the announcement of the IPO.
Blockholder is equal to 1 if there is at least one independent blockholder and zero otherwise. MillsRatio measures the potential selection bias
for companies that decided to purchase D&O insurance
VARIABLES
ln_ROL
(1)
(2)
Two-stage least square
First year total
Volatility
return
-0.160**
(0.0744)
0.0614**
(0.0243)
Rate-on-line
Mkt1year
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
0.944***
(0.321)
0.909*
(0.469)
0.0841
(0.0645)
24.85**
(11.67)
-1.404**
(0.666)
-0.0729
(0.178)
-0.207
(0.165)
0.347
(0.347)
-0.0161
(0.112)
-0.00680
(0.0216)
DebtRatio
-0.0304
(0.272)
YES
-2.422*
(1.357)
Age
Blockholder
MillsRatio
Year Dummy
Constant
-0.0576***
(0.0217)
2.126
(3.916)
-0.322
(0.230)
-0.123**
(0.0605)
0.180***
(0.0527)
-0.207***
(0.0756)
-0.00656
(0.0350)
-0.00105*
(0.000591)
-0.0490
(0.0425)
0.00372
(0.0976)
RiskyIndustry
(3)
(4)
Three-stage least square
First year total
Volatility
return
(5)
(6)
Three-stage least square
First year
Volatility
excess return
-0.161**
(0.0744)
-23.45**
(11.81)
0.939***
(0.323)
0.929**
(0.472)
0.0807
(0.0649)
24.86**
(11.70)
-1.409**
(0.668)
-0.0692
(0.179)
-0.207
(0.165)
0.321
(0.349)
-0.0326
(0.112)
-0.00693
(0.0218)
0.0615**
(0.0243)
12.70***
(3.748)
-0.0529**
(0.0212)
2.487
(3.818)
-0.339
(0.224)
-0.129**
(0.0591)
0.179***
(0.0513)
-0.0404
(0.273)
-0.201***
(0.0738)
-0.00217
(0.0342)
-0.00105*
(0.000576)
-0.0579
(0.0417)
0.0156
(0.0952)
YES
YES
2.127***
(0.454)
-1.325
(1.339)
(7)
(8)
Three-stage least square
First year
Volatility
excess return
-23.53**
(11.83)
0.899*
(0.467)
0.0850
(0.0644)
25.16**
(11.53)
-1.415**
(0.663)
-0.0741
(0.178)
-0.205
(0.164)
0.336
(0.338)
-0.0159
(0.112)
-0.00695
(0.0216)
-0.0576***
(0.0217)
2.151
(3.914)
-0.323
(0.230)
-0.123**
(0.0605)
0.180***
(0.0527)
0.918*
(0.470)
0.0817
(0.0648)
25.21**
(11.56)
-1.421**
(0.665)
-0.0704
(0.179)
-0.205
(0.165)
0.309
(0.339)
-0.0322
(0.112)
-0.00707
(0.0217)
12.71***
(3.748)
-0.0529**
(0.0212)
2.513
(3.816)
-0.341
(0.224)
-0.129**
(0.0590)
0.179***
(0.0513)
-0.0386
(0.268)
-0.206***
(0.0754)
-0.00653
(0.0349)
-0.00106*
(0.000589)
-0.0496
(0.0424)
0.00427
(0.0975)
-0.0494
(0.269)
-0.200***
(0.0736)
-0.00217
(0.0341)
-0.00105*
(0.000575)
-0.0585
(0.0416)
0.0162
(0.0952)
YES
YES
YES
YES
YES
1.630***
(0.437)
-2.437*
(1.357)
2.126***
(0.454)
-1.337
(1.340)
1.629***
(0.437)
Observations
83
83
83
83
83
83
83
83
R-squared
0.372
0.764
0.367
0.776
0.344
0.764
0.339
0.776
Coefficients are reported with their standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table IX Panel A. Regressions for the first year total return, excess return and volatility for the subset of firms that do not have the
rate-on-line information in the IPO prospectus
We evaluate the impact of price per unit of coverage on the firms' stock market return and volatility in the first year following the IPO,
selecting only firms that do not provide enough information to calculate the rate-on-line in the IPO prospectus. The dependent variables are
the first year total return, the first year excess return and the first year volatility.The two main variables of interest are Rate-on-line, calculated
as the premium-to-coverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that both measures should be negatively related
to the first year return and excess return, and positively to the first year volatility. All model are OLS regressions except for models 7 and 8
that use a three-stage least square approach.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO. LnMVE_IPO is the log of the firm's market value of equity at the time of the IPO (issue price
multiplied by number of outstanding shares). IPOfeerat is the ratio of the IPO fees paid per million dollar of market value of equity. Float is
computed as the ratio of the number of shares issued over the total number of shares outstanding. ITCE is a dummy variable equal to 1 if
the firm is an income trust. US_Sales is the percentage of sales carried out in the US. Independence is the proportion of board members
that are classified as independent. Duality is equal to 1 if the CEO is also the Chairman of the board and zero otherwise. Growth is the
market value of equity plus the book value of liability, divided by the book value of assets. DebtRatio is the ratio of total liabilities to the
market value of equity at the time of the IPO. Risky_Industry is an indicator variable equal to one if the company operated in one of the
industries classified as risky in Bajaj et al. (2000). Age is the number of years separating inception from the announcement of the IPO.
Blockholder is equal to 1 if there is at least one independent blockholder and zero otherwise.
(1)
VARIABLES
ln_ROL
FirstDayReturn
lnMVE_IPO
IPOfeerat
Float
ITCE
US_sales
Independence
Duality
Growth
(3)
(4)
First year total First year total
First year
First year
return
return
excess return excess return
-0.165**
(0.0735)
Rate-on-line
Mkt1year
(2)
0.585*
(0.309)
0.994
(0.735)
0.128*
(0.0709)
23.49
(21.75)
-1.227
(1.229)
-0.111
(0.221)
-0.188
(0.195)
0.116
(0.367)
-0.244*
(0.127)
0.0129
(0.0334)
-0.184**
(0.0718)
-22.61*
(11.77)
0.544*
(0.313)
1.066
(0.761)
0.120*
(0.0708)
22.95
(21.46)
-1.228
(1.206)
-0.116
(0.222)
-0.184
(0.199)
0.0998
(0.369)
-0.254*
(0.128)
0.00761
(0.0358)
Volatility
Volatility
-25.06**
(11.52)
0.926
(0.670)
0.146**
(0.0700)
27.04
(21.14)
-1.390
(1.194)
-0.117
(0.212)
-0.143
(0.200)
0.0593
(0.360)
-0.252*
(0.129)
0.0133
(0.0329)
0.999
(0.691)
0.138*
(0.0707)
26.80
(20.90)
-1.409
(1.172)
-0.124
(0.212)
-0.135
(0.205)
0.0335
(0.362)
-0.264**
(0.131)
0.00735
(0.0355)
RiskyIndustry
Age
Blockholder
Constant
(6)
0.0779**
(0.0369)
DebtRatio
Year Dummy
(5)
(7)
(8)
Three-stage least square
First year
Volatility
excess return
-0.170**
(0.0790)
0.0692***
(0.0258)
13.67***
(4.593)
-0.0526
(0.0376)
2.076
(3.460)
-0.313*
(0.186)
-0.105
(0.0706)
0.203***
(0.0698)
-0.0498
(0.0369)
2.364
(3.327)
-0.327*
(0.178)
-0.107*
(0.0607)
0.195***
(0.0638)
-0.153*
(0.0821)
-0.0128
(0.0416)
-0.000915*
(0.000474)
-0.0572
(0.0366)
-0.152*
(0.0795)
-0.00297
(0.0382)
-0.00101**
(0.000487)
-0.0647*
(0.0337)
0.669**
(0.323)
1.025**
(0.508)
0.131**
(0.0655)
23.37**
(11.32)
-1.231*
(0.652)
-0.118
(0.179)
-0.181
(0.161)
0.218
(0.365)
-0.230*
(0.129)
0.00614
(0.0344)
-0.0557***
(0.0215)
2.089
(3.505)
-0.328*
(0.196)
-0.112**
(0.0565)
0.226***
(0.0522)
-0.163**
(0.0709)
-0.0313
(0.0356)
-0.000820
(0.000580)
YES
YES
YES
YES
YES
YES
YES
YES
-2.964*
(1.505)
-1.764
(1.379)
-3.387**
(1.499)
-2.081
(1.403)
2.091***
(0.720)
1.554**
(0.707)
-3.100**
(1.405)
2.051***
(0.455)
Observations
68
68
68
68
71
71
68
68
R-squared
0.370
0.362
0.392
0.382
0.773
0.788
0.368
0.769
Coefficients are reported with their standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table IX Panel B. OLS regression that measures the firms' first year return statistics as a function of the information unknown at the
time of the IPO, for the subset of firms that do not have the rate-on-line information in the IPO prospectus
We evaluate the impact of price per unit of coverage on the firms' stock market return and volatility in the first year following the IPO, selecting
only firms that do not provide enough information to calculate the rate-on-line in the IPO prospectus. The dependent variable is one of the five
return distribution statistics used in the paper (total first year return, first year excess return, first year volatility, first year quasi-Sharpe ratio and
first year excess quasi-Sharpe ratio). The two main variables of interest are Rate-on-line, calculated as the premium-to-coverage ratio, and
ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively related to the return and Sharpe ratio
measures, and positively related to volatility.
FirstDayReturn is our control variable for underpricing and is computed as the first day return on the close of the first trading day. Mkt1Year is
the market return during the first year of the IPO. The Year_Dummy variables are there to take into account everything else that occured in the
year of the IPO.
(1)
VARIABLES
ln_ROL
(3)
First year
First year First year
excess
total return total return
return
-0.227***
(0.0791)
Rate-on-line
Mkt1Year
FirstDayReturn
Year Dummy
Constant
(2)
(4)
First year
excess
return
-0.237***
(0.0765)
(5)
(6)
(7)
(8)
First year
volatility
First year
volatility
Sharpe
ratio
Sharpe
ratio
0.0672
(0.0445)
-30.45***
(10.78)
-0.483**
(0.241)
0.633*
(0.355)
0.931**
(0.464)
-28.92**
(11.36)
0.611*
(0.365)
0.844*
(0.505)
11.67
(7.815)
0.951**
(0.430)
0.863*
(0.463)
YES
YES
YES
YES
YES
YES
YES
-1.081**
(0.483)
0.305
(0.209)
-1.142**
(0.471)
0.305
(0.208)
0.862***
(0.258)
0.439***
(0.0540)
-2.021
(1.343)
2.082*
(1.057)
(9)
Excess
Sharpe
ratio
(10)
Excess
Sharpe
ratio
-0.565**
(0.242)
-62.62
(38.53)
1.991*
(1.064)
-75.61**
(35.67)
3.627*
(1.890)
3.781*
(1.907)
YES
YES
YES
0.909*
(0.472)
-2.811**
(1.360)
0.611
(0.512)
Observations
75
75
75
75
78
78
76
76
74
74
R-squared
0.183
0.159
0.225
0.201
0.449
0.460
0.132
0.120
0.201
0.188
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the 1%, 5%, and 10% levels,
respectively.
Table IX Panel C. OLS regression that measures the firms' first year return net of all the information that
should be incorporated in the first day
We evaluate the impact of price per unit of coverage on the firms' stock market return in the first year following the IPO,
net of the return on the first day (one year return minus first day return in the top half, one year return divided by one
day return in the bottom half). The two main variables of interest are Rate-on-line, calculated as the premium-tocoverage ratio, and ln_ROL, the log of the Rate-on-line. We hypothesize that these two measures should be negatively
related to the return. Regression Specifications 1 through 4 use all firms for which we have a rate-on-line in the
management proxy of the first year post IPO. Regression Specifications 5 through 8 use only firms that did not provide
the rate-on-line information in their IPO prospectus.
Mkt1Year is the market return during the first year of the IPO. The Year_Dummy variables are there to take into
account everything else that occured in the year of the IPO.
VARIABLES
ln_ROL
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
First year First year First year First year First year First year First year First year
net of first net of first net of first net of first net of first net of first net of first net of first
day return day return day return day return day return day return day return day return
-0.189**
(0.0724)
Rate-on-line
Mkt1Year
0.632**
(0.285)
-0.220***
(0.0716)
-0.0253**
(0.0107)
0.634**
(0.294)
Year Dummy
Constant
Observations
R-squared
VARIABLES
ln_ROL
Observations
R-squared
YES
YES
0.406
(0.295)
-0.0210*
(0.0105)
0.393
(0.299)
0.636*
(0.354)
-0.0294**
(0.0113)
0.617
(0.372)
YES
YES
0.305
(0.209)
-1.142**
(0.471)
0.305
(0.208)
-2.021
(1.343)
0.909*
(0.472)
-2.811**
(1.360)
0.611
(0.512)
91
0.107
91
0.089
91
0.188
91
0.169
75
0.073
75
0.052
75
0.155
75
0.129
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
First year First year First year First year First year First year First year First year
net of first net of first net of first net of first net of first net of first net of first net of first
day return day return day return day return day return day return day return day return
-0.282**
(0.139)
0.779**
(0.336)
-0.306**
(0.135)
-0.0338**
(0.0162)
0.765**
(0.342)
Year Dummy
Constant
0.881**
(0.381)
-0.0301***
(0.0101)
0.894**
(0.409)
-0.229***
(0.0776)
-1.081**
(0.483)
Rate-on-line
Mkt1Year
-0.172**
(0.0705)
-0.295*
(0.173)
1.136**
(0.506)
-0.0367**
(0.0145)
1.151**
(0.554)
YES
YES
0.583
(0.374)
-0.351**
(0.173)
-0.0317*
(0.0189)
0.541
(0.365)
1.015
(0.609)
-0.0379**
(0.0177)
0.967
(0.637)
YES
YES
1.508***
(0.261)
1.258***
(0.141)
1.663***
(0.340)
1.418***
(0.243)
1.565***
(0.318)
1.278***
(0.158)
1.844***
(0.470)
1.529***
(0.354)
91
0.141
91
0.099
91
0.225
91
0.183
75
0.122
75
0.068
75
0.224
75
0.161
Coefficients are reported with their robust standard deviation in parentheses. ***, **, and * indicate significance at the
1%, 5%, and 10% levels, respectively.