The effects of underwriter reputation on pre-IPO earnings management and post-IPO operating performance.

Yong, Sun ; Kyung, Joo Lee ; Li, Diane 等

INTRODUCTION

The information asymmetries between IPO issuers and outside investors are considerable. Under this condition, IPO issuers may seek to increase their offering proceeds through manipulations of reportable earnings before going public. A stream of prior studies shows that IPO issuers employ opportunistic earnings management before issuing IPOs to obtain some gains, including enhancing initial firm values (e.g. Schipper 1989; Chaney and Lewis 1995; Teoh et al. 1998a, 1998b; Ducharme et al. 2001). For example, Teoh et al. (1998b) find evidence that IPO firms, on average, have high positive issue-year earnings and discretionary accruals, followed by poor long-run earnings and negative discretionary accruals.

However, prior studies on IPO earnings management have largely overlooked the potential roles played by underwriters. Ducharme et al. (2001) include both underwriter reputation and discretionary accruals as explanatory variables of IPO initial firm value, but they have not examined the relation between underwriter reputation and IPO accruals. Related studies on seasoned equity offerings have found the evidence of a negative relation between auditor quality and earnings management (Zhou and Elders 2004), and an inverse relation between underwriter quality and issuers' accruals (Jo, Kim, and Park 2007).

The purpose of this study is twofold. The first purpose is to document the relation between underwriter reputation and IPO earnings management. We argue that the certification role played by IPO underwriters has a restraining effect on opportunistic earnings management by IPO issuers. The second purpose is to examine a related issue that has not yet been studied in the literature, namely, the relation between underwriter reputation and post-issue operating performance of IPO firms given the presence of earnings management. We argue that underwriters have strong incentives to continue supplying monitoring to the firms they take public such that underwriter reputation is positively related to post-issue operating performance of IPO firms, even after the initial earnings management is taken into consideration.

Using a sample of 369 IPO issuers between 1997 and 2002, we find empirical results supporting our hypotheses. The results show that IPOs underwritten by high-reputation underwriters have less initial discretionary accruals. We also find that post-issue operating performance of IPO issuers are positively related to underwriter reputation. These results suggest that the certification and monitoring roles by underwriters not only restrain opportunistic earnings management, but also enhance post-IPO operating performance of the issuers.

The remainder of this paper is organized as follows. In the next section, we review previous literature and develop hypotheses based on the argument of the certification and monitoring roles played by underwriters. Sample selection and measurement of variables are described in section three. The empirical results are presented in section four. Our conclusions appear in final section.

HYPOTHESIS DEVELOPMENT

Underwriter certification and pre-IPO earnings management

Investment bankers play many roles in the underwriting of security issues including production and certification of information, provision of interim capital, and/or supplying distribution and marketing skills. Certification requires the underwriters to bear the liability imposed by the Securities Act of 1933 for ensuring the fairness of the offer price. The role of underwriter certification in reducing information asymmetries and mitigating the adverse selection faced by outside investors has been extensively studied in the context of IPOs.

In a typical model, in return for fees from the issuing firms, investment bankers produce and certify information about the firms that they underwrite. High-prestige investment bankers can have more stringent standards for certification of IPO firm value and can produce superior information about the firms that they underwrite. IPO issuers can signal favorable private information about their own values by choosing reputable underwriters. On the other hand, investors use an investment banker's reputation to revise their estimates of issuing-firm value. Thus, high-reputation investment bankers will represent less risky and higher-quality IPOs, and the use of a high-reputation underwriter is a positive signal about IPO firm value.

Since investment bankers are frequent participants in the equity market, they acquire reputation capital that enables them to act as credible certifiers of information. Chemmanur and Fulghieri (1994) find that high-prestige investment bankers, with valuable reputation capital at risk and superior information regarding the issuing firm's prospects, can credibly certify the value of the issues they underwrite.

The certification role of the underwriter has been investigated more specifically in papers that have examined the relationship between underwriter reputation and IPO underpricing. In general these studies argue that high-prestige underwriters are able to more fully price issues, reducing the level of underpricing. For example, Logue (1973), Beatty and Ritter (1986), Titman and Trueman (1986), Carter and Manaster (1990), and Carter, Dark, and Singh (1998) find that IPOs managed by more reputable underwriters are associated with less short-run underpricing. The empirical consensus is that IPO underwriters have performed their certification role in general, driven by the desire to protect their hard earned reputation capital.

Managers exercise some discretion in computing earnings without violating generally accepted accounting principles. It is possible that firms use discretionary accounting choices to manage earnings disclosures around the time of certain types of events. In view of the well-established correlation between earnings and share prices, earnings management activity seems particularly plausible around the time of unseasoned stock issues. According to this opportunism hypothesis, some firms opportunistically manipulate earnings upward before going public and investors are led to make overly optimistic expectations regarding future earnings of the issuers. Thus, issuing firms would be able to obtain a higher share price for their stock issue than they otherwise would. This view of IPO earnings management emphasizes the incentives that entrepreneurs, venture capitalists, and managers have to maximize issue proceeds, given the number of shares offered.

A priori, the opportunistic earnings management of IPO issuers and the certification of underwriters appear at odd with each other. If high levels of abnormal accruals reflect deceptive accounting, we expect the related IPOs to be shunned by investment bankers that have significant reputation capital at stake. (1) Thus, we expect a negative relation between underwriter reputation and IPO earnings management. That is, when high-reputation underwriters are involved, IPO issuers become voluntarily or involuntarily less aggressive with earnings management. IPO issuers with minimal incentives for earnings management would select high-reputation underwriters to enhance underwriter certification, thereby signaling favorable information to outside investors. Our hypothesis is consistent with the implications of the results of Zhou and Elders (2004) and Jo, Kim, and Park (2007) on seasoned equity offerings.

The negative relation between underwriter reputation and earnings management can also be inferred from the underwriter's monitoring function. Block and Hoff (1999) suggest that underwriters conduct due-diligence investigations to ensure proper information disclosure by issuers and prevent potential legal liabilities. High-reputation underwriters have more resources and more expertise and are therefore more likely to perform higher-quality monitoring in the underwriting process. Thus, high-reputation underwriters are less likely associated with aggressive IPO earnings management. These arguments lead to the following hypothesis:

Hypothesis 1: There is a negative relation between underwriter reputation and the issuer's earnings management before an IPO.

Underwriter monitoring and post-IPO operating performance

The certification role of underwriters ends at the IPO, but the monitoring function continues in the post-IPO period (Stoughton and Zechner (1998) and Jain and Kini (1999)). In general, when new securities are issued, issuing firms carefully examine the investment bankers' track record as part of their lead underwriter selection process. Apart from factors such as pricing and marketing, issuers look to other performance areas such as post-issue price stability, market-making, analyst-coverage, and the ability to underwrite subsequent offerings or conclude corporate restructuring activities. Given the lucrative future opportunities, IPO underwriters have strong incentives to remain engaged in the affairs of the firms they take public and to ensure that managers are following value enhancing strategies. Thus, monitoring by underwriters has the potential to improve post-IPO operating performance.

Prior studies suggest that investment bankers play a valuable monitoring function in ensuring managers to follow a value maximizing path. For example, Easterbrook (1984) suggests that when a firm issues new securities its activities are scrutinized by an investment banker or some similar intermediary acting as a monitor for the collective interests of investors of the new securities. Hansen and Torregrosa (1992) suggest that underwriter monitoring improves the issuing firm's performance and reduces agency costs, thereby enhancing firm value. Stoughton and Zechner (1998) argue that given the active and continuing nature of the relationship between investment bankers and institutional investors, they work together in monitoring the affairs of IPO firms. More specifically, Jain and Kini (1999) find that underwriter monitoring is positively related to post-IPO operating and investment performance.

High-prestige underwriters, given their considerable resources, are more likely to supply long-term monitoring in order to continue the business relationships with their clients. Jain and Kini (1999) find that about 75% of lead underwriters assign at least one analyst to track the company they take public. In addition, the presence of institutional investors in the new issues market also promote underwriter monitoring. As implied in Stoughton and Zechner (1998), given the active and continuing nature of the relationship between investment bankers and institutional investors, high-prestige underwriters have strong incentives to work with the institutional investment community in monitoring the affairs of IPO issuers. Thus, high-reputation underwriters are more likely associated with improvements in post-issue operating performance. These arguments lead to our second hypothesis.

Hypothesis 2: There is a positive relation between underwriter reputation and the issuer's operating performance after an IPO.

SAMPLE SELECTION AND DATA

Sample selection

Our initial sample of IPO issuers is obtained from the IPO database of Hoovers Incorporated. The sample period is from April 1996 to December 2004. Several selection criteria are applied sequentially. First, financial institutions and utility firms are excluded. Also, the sample excludes ADRs, firms with offer price less than one dollar and firms with offer size less than one million dollars. Finally, relevant data availability in COMPUSTAT data files over the period of six years surrounding each IPO (i.e., t = [-2, .0, ..3]) is required. These selection criteria yield the final sample of 369 IPO issuers.

Information regarding reputation of the IPO underwriters is based on the reputation rankings of Carter and Manaster (1990), and updated according to the information on the website of Jay Ritter. We further classify the underwriters into three groups. (2) If an underwriter's reputation rank is greater than or equal to 9.0, the underwriter is in the high-reputation group; if the reputation rank is between 7.1 and 8.1, the underwriter is in the medium-reputation group; and if the reputation rank is less than or equal to 7.0, the underwriter is in the low-reputation group. There are 196 IPOs in the high-reputation group, 136 in the medium-reputation group, and 37 in the low-reputation group.

Table 1 provides distribution of IPOs by calendar year and underwriter reputation. Two points are worth noting from Table 1. First, almost a half of total IPOs occurred during the bubble period (1999-2000). Second, more than half of sample firms employ high-reputation underwriters. There are only 37 firms (10%) with low-reputation underwriters. Although not shown in table, our sample represents 40 industries (2-digit SIC). However, as typical in IPOs, sample firms are highly concentrated in a few industries, such as computer hardware and software (39.3%) and chemical products (10.8%).

Measurement of Variables

Earnings Management

The proxy for earnings management is measured by discretionary accruals, which are obtained relative to expected benchmark accruals (nondiscretionary accruals) based on firm and industry characteristics. We use cross-sectional modified Jones model to estimate discretionary accruals of each IPO firm (Jones, 1991; Dechow et al., 1995; Teoh et al., 1998a). (3)

For each IPO firm, we find at least ten industry-matched firms with the same three-digit SIC code. If we are unable to find ten industry-matched firms with the same three-digit SIC code, we use industry-matched firms with the same two-digit SIC code. For each IPO firm j, we run the following cross-sectional regression model:

[TAC.sub.iy]/[TA.sub.iy-1] = [[alpha].sub.ij-1][1/ [TA.sub.iy-1]] + [[alpha].sub.1j][([DELTA][REV.sub.iy] - [DELTA][REC.sub.iy])/ [TA.sub.iy-1] + [[alpha].sub.2j][[PPE.sub.iy]/ [sub.Taiy-1]] + [[epsilon].sub.iy] (1)

where,

[TAC.sub.iy] = total accruals (net income before extraordinary items minus cash flow from operations) in year y for the ith firm in the industry group matched with offering firm j.

[TA.sub.iy] = total assets in year y for the ith firm in the industry group matched with offering firm j.

[DELTA][REV.sub.iy] = change in revenues in year y for the ith firm in the industry group matched with offering firm j.

[DELTA][REC.sub.iy] = change in accounts receivable in year y for the ith firm in the industry group matched with offering firm j.

[PPE.sub.iy] = gross property, plant, and equipment in year y for the ith firm in the industry group matched with offering firm j.

Using estimated coefficients from regression model (1), discretionary accruals (DAC) for the issuing firm j in year y are then estimated by subtracting nondiscretionary accruals (NAC) from total accruals (TAC) as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Post-IPO Operating Performance

Post-IPO operating performance was measured by industry-adjusted operating return on assets (OROA), which is defined as operating income before taxes and depreciation divided by total assets. (4) Three years data (t=[1,2,3]) after IPO are averaged to obtain the measure. (5) To get industry-adjusted OROA, we subtract from each firm's raw OROA the median of a group of firms with matched 3-digit SIC code. If there were insufficient firms (less than 10) in the industry, we use 2-digit SIC to find the matched companies.

Other Variables

Issuer age (AGE): company age from initial founding to IPO years.

Offer size (OS): offer price x number of shares offering.

Sales growth (SG): percentage change in sales up to the year prior to IPO.

Pre-IPO operating performance (PREOP): operating performance in the year before IPO (t=-1). Operating performance is measured by industry-adjusted OROA.

Standard deviation of stock returns (SD): standard deviation of daily returns from day 6 to day 255 after IPO.

Leverage ratio (LEV): debt to equity ratio in year t-1, as measured by the ratio of book value of equity to total asset.

Table 2 presents descriptive statistics for selected variables. Mean and median values of these variables are reported for each underwriter reputation group as well as for total sample. On average, the issuers are 13 years old since their founding, have about $395 million in total assets, are offering $148 million, and show 24.4% of leverage ratio.

More importantly, sample firms exhibit some differences in their characteristics. First of all, issuers employing high-reputation underwriters are larger in size, as measured by both total assets and offer sizes, than those with medium and low-reputation underwriters. Also, high-reputation underwriters are more (less) likely to underwrite firms with higher sales growth (operating performance before IPO). However, there are no differences in AGE, SD and LEV among underwriter reputation groups.

EMPIRICAL RESULTS

IPO earnings management and underwriter certification: Hypothesis 1

Table 3 presents the results of comparing discretionary accruals across three subgroups of underwriter reputation (high, medium and low), and corresponding test statistics and p-values. For overall comparison, Kruskal-Wallis [chi square] statistic of 6.929 indicates that there are significant ([alpha] < 0.05) differences in discretionary accruals across underwriter reputation subgroups.

Pairwise comparison results and corresponding Wilcoxon z-statistics show that for IPO issuers employing high-reputation underwriters, median discretionary accruals are smaller than those employing low-reputation underwriters (-0.056 vs. 0.002) and the difference is statistically significant ([alpha] < 0.01). Also, median discretionary accruals show a significant difference ([alpha] < 0.10) between medium- and low- reputation groups. However, there is no significant difference between high- and medium-reputation groups.

In short, discretionary accruals of IPOs underwritten by high- and medium-reputation underwriters are significantly lower than those of IPOs underwritten by low-reputation underwriters. These results indicate that IPO issuers hiring low- reputation underwriters are more likely to adopt aggressive earnings management policies than those hiring high- or medium-reputation underwriters, which is consistent with the prediction of the underwriter certification hypothesis. (6)

Results in Table 3 are based on unvariate tests, which ignore potential effects of other variables on the degree of earnings management. As an attempt to investigate if these results hold after controlling for other factors related to issuer characteristics and earnings management, we estimate the following regression model:

[DAC.sub.i] = [[beta].sub.0] + [[beta].sub.1][OS.sub.1] + [[beta].sub.2][SG.sub.i(t-1)] + [[beta].sub.3][PREOP.sub.i(t- 1)] + [[beta].sub.4][LEV.sub.i(t-1)] + [[beta].sub.5][RP.sub.i] + [epsilon] (2)

where,

[DAC.sub.i] = discretionary accruals for ith firm in year t-1.

[OS.sub.i] = natural logarithm of offer size for ith firm.

[SG.sub.i(t-1)] = sales growth for ith firm in year t-1.

[PREOP.sub.i(t-1)] = industry-adjusted operating return on assets for ith firm in year t-1.

[LEV.sub.i(t-1)] = debt to equity ratio for ith firm in year t-1.

[RP.sub.i] = underwriter reputation for ith firm, measured by the rankings of Carter and Manaster (1990), and updated according to the information in Jay Ritter's website.

Our hypothesis 1 predicts that [[beta].sub.5] is negative since the issuers employing high-reputation underwriters are likely to have less earnings management.

Table 4 presents the results from estimating the regression model (2). (7) The results are essentially the same as those from univariate analyses. The regression coefficient of RP ($5) has predicted sign (negative) and is statistically significant ([alpha] < 0.10). Overall, these results lend strong support to our hypothesis that underwriter reputation is negatively related to pre-IPO earnings management, even after controlling for other variables.

Post-IPO operating performance and underwriter monitoring: Hypothesis 2

In order to examine the effect of underwriter reputation on post-IPO operating performance after controlling for pre-IPO earnings management and other factors, we estimate the following regression model:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

where,

[POSTOP.sub.i] = average industry-adjusted operating return on assets for ith firm between years t=1 and t=3.

[DAC.sub.i(t-1)] = discretionary accruals for ith firm in year t-1.

[SG.sub.i(t-1)] = sales growth for ith firm in year t-1.

[PREOP.sub.i(t-1)] = industry-adjusted operating return on assets for ith firm in year t-1.

[AGE.sub.i(t-1)] = age of ith firm in the year t-1.

[SD.sub.i] = standard deviation of daily returns for ith firm from day 6 to day 255 after IPO.

[RP.sub.i] = underwriter reputation for ith firm, measured by the rankings of Carter and Manaster (1990), and updated according to the information in Jay Ritter's website.

Since underwriters have incentives to continue providing monitoring services to the firms they take public, the issuers employing high-reputation underwriters with better monitoring capabilities are likely to have better operating performance after IPO. Hence, it is predicted that [[beta].sub.6] is positive.

Table 5 presents the results from estimating the regression model (3). First of all, the coefficient of DAC is positive but insignificant. This indicates that pre-IPO earnings management has no effect on post-IPO operating earnings. More importantly, the regression coefficient of RP ([[beta].sub.6]) has predicted sign (positive) and is statistically significant ([alpha] < 0.10). This result suggests that underwriter reputation is positively related to post-IPO operating earnings, even after controlling for other variables. This supports our hypothesis 2.

Robustness tests

IPO earnings management and the choice of the lead underwriter could be mutually related. IPO issuers with aggressive earnings management may deliberately avoid high-prestige underwriters if they think the underwriters would monitor their accruals management. Likewise, high-prestige underwriters may also choose to avoid IPO issuers with aggressive earnings management given their reputation capital at stake.

To handle this potential endogeneity problem, we use an instrumental variable two-stage least squares (2SLS) regression approach. In the first stage, we estimate the following regression model: (8)

[RP.sub.i] = [[alpha].sub.0] + [[alpha].sub.1] [DAC.sub.i(t-1)] + [[alpha].sub.2] Ln([TA.sub.i(t-1))] + [[alpha].sub.3] [Ln([TA.sub.i(t-1))].sup.2] + [member of] (4)

In the above models, RP is the actual reputation ranking of the underwriter. The regression coefficients from each model are then applied to our IPO sample to find the estimated reputation, ER, of each underwriter. We then estimate the following second stage regression model:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

where, ER is the estimated reputation obtained using the estimates of model (4). Other variables are same as those used in model (3).

Table 6 presents the 2SLS regression results. The second column of Table 6 reports the results of estimating regression model (4). Firm size, as measured by total asset, has significantly positive relationship with underwriter reputation, indicating that larger firms tend to choose high-reputation underwriters. The fourth column of Table 6 presents the results of estimating the stage II regression model (5), which includes the instrumental variable ER. The results show that while the coefficient of SG (PREOP) is significantly negative (positive), the coefficient of DAC is insignificant. That is, with the endogeneity between IPO earnings management and the choice of the underwriter considered, pre-IPO accruals do not affect post-issue operating performance of IPO firms.

Most importantly, the coefficient of ER is positive and statistically significant ([alpha] < 0.05), which is consistent with the results from simple regression (3). This significantly positive relation between underwriter reputation and post-IPO operating performance implies that underwriters are not disheartened by the initial earnings management of IPO issuers. IPO underwriters keep themselves engaged in the affairs of their clients because of the lucrative business relationships. The monitoring that high-reputation underwriters continue to supply to the firms they take public is value-increasing. This finding supports our second hypothesis and is consistent with Stoughton and Zechner (1998) and Jain and Kini (1999).

In our sample, the high- and medium- reputation underwriter groups have observations several times that of the low-reputation underwriters. To avoid our results being driven by this factor, we apply a weighted least squares (WLS) approach to the instrumental variable two-stage regression model. The weight applied to each observation is equal to the inverse of the number of observations in each underwriter-reputation group. In this manner, each group receives equal weight in the estimation.

Table 7 presents the results of estimating regression model (4) and (5) using two-stage WLS regression approach. The results again show that underwriter reputation has significantly positive impact on the post-IPO operating performances of IPO issuers. However, the coefficient of DAC is again insignificant. This suggests that the presence of effective certification and monitoring by underwriters has not only restrained the opportunistic initial earnings management, but also resulted in improvements of post-issue operating performance for IPO firms.

CONCLUSION

The purpose of this study is to investigate the effects of underwriter reputation on IPO issuers' pre-issue earnings management and post-issue operating performance. We predict that the certification role played by underwriters has a restraining effect on opportunistic earnings management by IPO issuers. We also hypothesize that underwriter reputation is positively related to post-issue operating performance of IPO firms based on the argument that underwriters have strong incentives to continue supplying monitoring to the firms they take public.

Using a sample of 369 IPOs between 1997 and 2002, we find the empirical results supporting our hypotheses. Specifically, our results can be summarized as follows. First, IPO issuers underwritten by high- and medium-reputation underwriters on average have discretionary accruals that are significantly less than those associated with low- reputation underwriters. Second, underwriter reputation is negatively related to pre-IPO earnings management, even after controlling for other variables. Third, underwriter reputation has positive impact on the issuers' post-IPO operating earnings, even after controlling for other variables. Finally, pre-issue earnings management is not related to post-issue operating performance for the IPO issuers

For robustness tests, we consider the possibility that IPO earnings management and the choice of the underwriter are endogenously determined. Using an instrumental variable two-stage least squares (2SLS) regression approach, we find the results that there is a positive relation between underwriter reputation and post-IPO operating performance. We also control for the unequal number of IPOs underwritten by each reputation-group by performing a weighted least squares regression. Our results remain the same.

REFERENCES

Beatty, Randolph P. and Jay R. Ritter. 1986. "Investment Banking, Reputation, and the Underpricing of Initial Public Offerings." Journal of Financial Economics 15: 213-32.

Block, D.J. and J. Hoff. 1999. "Underwriter Due Diligence in Securities Offerings." Cadwalader, Wickersham and Taft LLP.

Carter, Richard B, Frederick H. Dark, and Ajai K. Singh. 1998. "Underwriter Reputation, Initial Returns, and the Long-Run Performance of IPO Stocks." Journal of Finance 53: 285-311.

Carter, Richard B. and Steven Manaster. 1990. "Initial Public Offerings and Underwriter Reputation." Journal of Finance 45:1045-67.

Chaney, Paul K. and Craig M. Lewis. 1995. "Earnings Management and Firm Valuation under Asymmetric Information." Journal of Corporate Finance 1: 319-45.

Chemmanur, Thomas J, and Paolo Fulghieri. 1994. "Investment Bank Reputation, Information Production, and Financial Intermediation." Journal of Finance 49: 57-79.

Ducharme, Larry L. 1994. "IPOs: Private Information and Earnings Management." Ph.D. Dissertation, University of Washington.

Ducharme, Larry L., Paul H. Malatesta, and Stephan E. Sefcik. 2001. "Earnings Management: IPO Valuation and Subsequent Performance." Journal of Accounting, Auditing and Finance 16: 369-96.

Friedlan, J. M. 1994. "Accounting Choices by Issuers of Initial Public Offerings." Contemporary Accounting Research 11: 1-33.

Hansen,R. and P. Torregrosa. 1992. "Underwriter Compensation and Corporate Monitor." Journal of Finance 47: 1537-1555.

Jain, B. and O. Kini. 1999. "On Investment Banker Monitoring in the New Issues Market." Journal of Banking and Finance 23: 49-84.

Jo, H., Y. Kim and M. Park 2007. "Underwriter Choice and Earnings Management: Evidence from Seasoned Equity Offerings." Review of Accounting Studies 12: 23-59.

Jones, J. 1991. "Earnings Management during Import Relief Investigations." Journal of Accounting Research 29: 193-228.

Schipper, K. 1989. "Commentary on Earnings Management." Accounting Horizons 3: 91-102.

Stoughton, N. and J. Zechner. 1998. "IPO-mechansisms, Monitoring and Ownership Structure." Journal of Financial Economics 49: 45-77.

Teoh, Siew Hong, I. Welch, and T. J. Wong. 1998a. "Earnings Management and the Underperformance of Seasoned Equity Offerings." Journal of Financial Economics 50: 63-99.

Teoh, Siew Hong, I. Welch, and T. J. Wong. 1998b. "Earnings Management and the Long-term Market Performance of Initial Public Offerings." Journal of Finance 53: 1935-1974.

Titman, Sheridan and Brett Trueman. 1986. "Information Quality and the Valuation of New Issues." Journal of Accounting and Economics 8: 159-72

Zhou, J. and R.Elder. 2004. "Audit Quality and Earnings Management by Seasoned Equity Offering Firms." Asia-Pacific Journal of Accounting and Economics 11: 95-120.

Yong Sun, Southwestern University of Finance and Economics, China

Kyung Joo Lee, University of Maryland-Eastern Shore

Diane Li, University of Maryland-Eastern Shore

John Jongdae Jin, California State University-San Bernardino

ENDNOTES

(1) Another view of earnings management emphasizes the liabilities arising from false earnings signals. These include explicit legal expenses and implicit costs due to a damaged firm reputation. It is argued that the burdens impel stock issuers to signal validly. Thus investors are informed, but not deceived. Even if this view is correct, we argue that high-prestige underwriters will distance themselves from firms with aggressive earnings management because there would be undesirable effects if the underwritten firms are likely to keep reporting continuously declining performance when accruals revert in later reporting periods.

(2) A rational for the cut-off points used to classify underwriters is based on the mean (8.17) and median (9.10) rankings for our sample (see Table 2).

(3) Cross-sectional method is used because a time series approach is not possible for IPOs. The cross-sectional approach has an additional advantage in that it incorporates changes in accruals resulting from changes in economic conditions for the industry as a whole. Since the cross-sectional regression is re-estimated each year, specific year changes in economic conditions affecting expected accruals are filtered out. Moreover, the common practice by underwriters of comparing market prices and financial information of similar firms when pricing IPO equity further shows the importance of extracting industry-wide economic conditions from abnormal accruals.

(4) As additional measures of post-issue operating performance, we use the industry-adjusted operating cash flow return on assets and the industry-adjusted return on assets. The main results remain the same.

(5) Using average performance over three years rather than annual performance can smooth out temporal fluctuations due to distortions arising from accrual accounting by the IPO firm.

(6) These results are also are also consistent with the implication that either high-reputation underwriters have deliberately avoided IPO issuers with high pre-IPO discretionary accruals, or they tend to avoid each other.

(7) We also estimate the regression model (2) using RP as a categorical variable (i.e., assign 3, 2, 1 to high-, medium- and low-reputation group, respectively). The estimation results are qualitatively the same.

(8) We have also estimated underwriter reputation using two different models, which include additional variables (SD; SD and AGE). The results remain basically the same.

Table 1: IPO Distribution by Year and Underwriter Reputation

 Underwriter Reputation (1,2)

 All High Medium Low

Year N % N % N % N %

1997 69 18.70 20 28.99 42 60.87 7 10.14
1998 56 15.18 30 53.57 15 26.79 11 19.64
1999 56 15.18 36 64.29 13 23.21 7 12.50
2000 123 33.33 77 62.60 38 30.89 8 6.50
2001 36 9.76 18 50.00 16 44.44 2 5.56
2002 29 7.86 15 51.72 12 41.38 2 6.90
Total 369 100 196 53.12 136 36.86 37 10.03

(1) Underwriter reputation is based on the rankings of Carter and
Manaster (1990), and updated according to the information in Jay
Ritter's website.

(2) Underwriters are classified into three groups: high-reputation,
if the ranking is greater than or equal to 9; medium-reputation, if
ranking is between 7.1 and 8.1; low-reputation, if ranking is less
than or equal to 7.

Table 2: Descriptive Statistics for Selected Variables

 Total Sample High Reputation

Variables (1) Mean Std Dev Med Mean Med

RP 8.173 1.485 9.100 9.100 9.100
AGE 13.369 19.668 7.000 13.347 5.500
TA 394.664 2917.183 31.072 678.633 49.604
OS 148.057 574.962 60.000 227.944 84.000
SG 4.567 23.562 0.499 7.159 0.547
PREOP -1.380 12.891 -0.023 -1.948 -0.043
SD 0.171 1.537 0.054 0.059 0.057
LEV 0.244 0.321 0.105 0.239 0.080

 Median Reputation Low Reputation

Variables (1) Mean Med Mean Med

RP 7.857 8.100 4.424 5.100
AGE 13.596 8.000 12.649 8.000
TA 87.095 29.776 20.919 12.950
OS 64.934 47.500 30.405 23.000
SG 1.735 0.412 1.244 0.414
PREOP -0.839 -0.014 -0.360 0.052
SD 0.057 0.053 1.187 0.053
LEV 0.262 0.161 0.212 0.102

RP = underwriter reputation based on the rankings of Carter and
Manaster (1990), and updated according to the information in Jay
Ritter's website.

AGE = age of IPO firms (years).

TA = total assets ($ Million) in year t-1.

OS = offer size ($ Million); natural logarithm of OS is used in
regression analyses.

SG = sales growth in year t-1.

PREOP = operating performance as measured by the industry-adjusted
operating return on assets (operating income before taxes and
depreciation divided by total assets) in the year before IPO
(t = -1).

SD = standard deviation of daily returns from day 6 to day 255
after IPO.

LEV = debt to equity ratio in year t-1.

Table 3: Comparisons of Discretionary Accruals across Underwriter
Reputation Groups

Underwriter Mean Std Dev Min 25% 50% 75% Max
Reputation

High -0.707 4.786 -62.967 -0.370 -0.056 0.069 1.012
Medium -0.231 2.431 -26.050 -0.240 -0.050 0.070 7.124
Low 0.398 1.326 -1.381 -0.128 0.002 0.262 6.472

Overall 6.929 (0.031) **
Comparison:
Kruskal-Wallis
[chi square]
statistic
(p-value)

Pairwise High vs. Medium Medium vs. Low High vs. Low
Comparison:
Wilcoxon 0.958 (0.338) 1.916 (0.055) * 2.629 (0.008) ***
z-statistic
(p-value)

***: Significant at [alpha] < 0.01; **: Significant at [alpha] < 0.05;
*: Significant at [alpha] < 0.10

Table 4: Effect of Underwriter Reputation on Pre-IPO Earnings
Management: Regression Analysis (1)

[DAC.sub.i] = [[beta].sub.0] + [[beta].sub.1]O[S.sub.i] +
[[beta].sub.2][SG.sub.i(t-1)] + [[beta].sub.3][PREOP.sub.i(t-1)]
+ [[beta].sub.4][LEV.sub.i(t-1)] + [[beta].sub.5][RP.sub.i] +
[member of]

Independent Variables Coefficients (t-value)

Intercepts 0.989 (3.07) ***
OS -0.079 (1.12)
SG -0.002 (0.88)
preop 0.311 (27.59) ***
LEV -0.089 (0.49)
RP -0.073 (1.65) *
Adj. [R.sup.2] 0.721

DAC = discretionary accruals in year t-1.

OS = natural logarithm of offer size.

SG = sales growth in year t-1.

PREOP = industry-adjusted operating return on assets in year t-1.

LEV = debt to equity ratio in year t-1.

RP = underwriter reputation based on the rankings of Carter and
Manaster (1990), and updated according to the information in
Jay Ritter's website.

***: Significant at [alpha] < 0.01; **: Significant at [alpha]
< 0.05; *: Significant at [alpha] < 0.10

Table 5: Effect of Underwriter Reputation on Post-IPO Operating
Earnings: Regression Analysis (1)

[POSTOP.sub.i] = [[beta].sub.0] + [[beta].sub.1][DAC.sub.i(t-1)]
+ [[beta].sub.2][SG.sub.i(t-1)] + [[beta].sub.3][PREOP.sub.i(t-1)]
+ [[beta].sub.4][AGE.sub.i(t-1)] + [[beta].sub.5][SD.sub.i] +
[[beta].sub.6][RP.sub.i] + [member of]

Independent Variables Coefficients (t-value)

Intercepts 0.0904 (1.22)
DAC 0.0112 (1.17)
SG -0.0012 (2.29) **
PREOP -0.0017 (0.59)
AGE 0.0014 (2.23) **
SD -0.0123 (1.47)
RP 0.0156 (1.76) *
Adj. [R.sup.2] 0.040

(1) POSTOP = average industry-adjusted operating return on assets
(operating income before taxes and depreciation divided by total
assets) between years t = 1 and t = 3.

DAC = discretionary accruals in year t-1.

OS = natural logarithm of offer size.

SG = sales growth in year t-1.

PREOP = industry-adjusted operating return on assets in year t-1.

AGE = age of IPO firms (years).

SD = standard deviation of daily returns from day 6 to day 255
after IPO.

RP = underwriter reputation based on the rankings of Carter and
Manaster (1990), and updated according to the information in
Jay Ritter's website.

***: Significant at [alpha] < 0.01; **: Significant at [alpha]
< 0.05; *: Significant at [alpha] < 0.10

Table 6: The Effect of Underwriter Reputation on Post-IPO Operating
Performance: Two Stage Least Square (2SLS) Regression Analysis

Stage I: [RP.sub.i] = [[alpha].sub0] + [[alpha].sub1][DAC.sub.i(t-1)]
+ [[alpha].sub.2] Ln([TA.sub.i(t-1)]) + [[alpha].sub3]
[[Ln([TA.sub.i(t-1)])].sup.2] + [member of]

Stage II: [POSTOP.sub.i] = [[beta].sub.0] + [[beta].sub.1]
[DAC.sub.i(t-1)] + [[beta].sub.2][SG.sub.i(t-1)] + [[beta].sub.3]
[PREOP.sub.i(t-1)] + [[beta].sub.4][AGE.sub.i(t-1)] +
[[beta].sub.5][SD.sub.i] + [[beta].sub.6][ER.sub.i] + [member of]

Stage I Regression (1)

Variables Coefficients

Intercept 6.0624 ***
DAC -0.0239
Ln(TA) 0.7824 ***
[[Ln(TA)].sup.2] -0.0525 ***
Adj. [R.sup.2] 0.16

Stage II Regression (2)

Variables Coefficients

Intercept -0.5840 *
DAC -0.0000
SG -0.0009 **
PREOP 0.0435 *
AGE 0.0004
SD 0.0038
ER 0.0799 **
Adj. [R.sup.2] 0.14

Stage I Regression:

RP = underwriter reputation.

DAC = discretionary accruals in the year t1

TA = total assets in the year t=-1.

Stage II Regression:

POSTOP = average industry-adjusted operating return on assets
(operating income before taxes and depreciation divided by
total assets) between years t = 1 and t = 3.

SG = sales growth in year t-1.

PREOP = industry-adjusted operating return on assets in year t-1.

AGE = age of IPO firms (years).

SD = standard deviation of daily returns from day 6 to day 255
after IPO.

ER = instrumental variable (estimated reputation of underwriter)
from the first stage regression.

***: Significant at [alpha] < 0.01; **: Significant at [alpha]
< 0.05; *: Significant at [alpha] < 0.10

Table 7: The Effect of Underwriter Reputation on Post-IPO Operating
Performance: Weighted Least Square (WLS) Regression Analysis

Stage I: [RP.sub.i] = [[alpha].sub0] + [[alpha].sub1][DAC.sub.i(t-1)]
+ [[alpha].sub.2]Ln([TA.sub.i(t-1)]) + [[alpha].sub3]
[[Ln([TA.sub.i(t-1)])].sup.2] + [member of]

Stage II: [POSTOP.sub.i] = [[beta].sub.0] + [[beta].sub.1]
[DAC.sub.i(t-1)] + [[beta].sub.2][SG.sub.i(t-1)] + [[beta].sub.3]
[PREOP.sub.i(t-1)] + [[beta].sub.4][AGE.sub.i(t-1)] +
[[beta].sub.5][SD.sub.i] + [[beta].sub.6][ER.sub.i] + [member of]

Stage I Regression (1)

Variables Coefficients

Intercept 6.0624 ***
DAC -0.0239
Ln(TA) 0.7824 ***
[[Ln(TA)].sup.2] -0.0525 ***
Adj. [R.sup.2] 0.16

Stage II Regression (2)

Variables Coefficients

Intercept -0.0007 *
DAC 0.0023
SG -0.0002
PREOP 0.0011 **
AGE 0.0015
SD -0.0029
ER 0.0148 *
Adj. [R.sup.2] 0.10

Stage I Regression:

RP = underwriter reputation.

DAC = discretionary accruals in the year t-1.

TA = total assets in the year t=-1.

Stage II Regression:

POSTOP = average industry-adjusted operating return on assets
(operating income before taxes and depreciation divided by total
assets) between years t = 1 and t = 3.

SG = sales growth in year t-1.

PREOP = industry-adjusted operating return on assets in year t-1.

AGE = age of IPO firms (years).

SD = standard deviation of daily returns from day 6 to day 255
after IPO.

ER = instrumental variable (estimated reputation of underwriter)
from the first stage regression.

***: Significant at [alpha] < 0.01; **: Significant at [alpha]
< 0.05; *: Significant at [alpha] < 0.10