do exchange listings mislead investors by bloated

Do exchange listings mislead investors by bloated earnings?

Abstract:

By examining firms switching trading markets from the NASDAQ or AMEX to the NYSE

and from the NASDAQ to the AMEX during 1986–2005, the empirical results show that,

switching firms, on average, have high positive earnings and abnormal accruals in the year

preceding the exchange listing announcement, followed by poor long-run earnings and

insignificant abnormal accruals. Though there is no evidence of a reversal of the stock price

effects of the earnings management at the listing announcements, the pre-listing abnormal

accruals are negatively associated with the post-listing performance, with the association

driven mainly by those firms with poor earnings in both current and future periods. In

particular, we uncover that switching firms with poor current earnings adopt more

income-increasing accruals when they expect to have good future earnings, whereas those with

good current earnings report more income-decreasing accruals when they expect to experience

poor future earnings. The results taken together suggest that abnormal accruals before

exchange listings are associated with managers’opportunistic behavior and the pre-listing

earnings management partially explains the “post-listing puzzle”, supporting the “managerial

opportunism”hypothesis. Additionally, we also find some evidence indicating that sample

firms with good performance in both current and future periods tend to make a strategic

income-decreasing reporting choice under the concerns of avoiding competition and/or

mitigating adverse political attention.

JEL classification: G14, M41

Keywords: Exchange listings; Discretionary accruals; Managerial opportunism; Signaling

2

1. Introduction and overview

Market participants are likely to respond to news of exchange listing because of its

perceived impact on the firm. Although there is little theoretical justification to believe listing

affects all firms equally, prior empirical studies have documented that the average firm

experiences significantly positive excess returns prior to listing, while abnormal returns

following listing on the exchange are typically negative.1 In addition, Papaioannou et al. (2003)

examine the operating performance around the listing change and find evidence of a pattern of

pre-listing improvement and post-listing deterioration in operating performance. The favorable

pre-listing performance is largely ascribed to the information signaling effect. A common

explanation for this information-based effect is related to the motives of management for

changing trading marketplace, including the desire to gain prestige and visibility (Van Horne,

1970; Merton, 1987; Kadlec and McConell, 1994), to convey confidence in the future

prospects of the firm (Ying et al., 1977; McConell and Sanger, 1984; Baker and Edelman,

1991), and to improve stock liquidity (Christie and Huang, 1993; Kadlec and McConell, 1994).

However, the poor post-listing performance phenomenon has been referred to as the

“post-listing puzzle”. McConnell and Sanger (1987) examine several explanations for the

post-listing puzzle, including issuance of new stocks immediately after listing, insider dumping

their stocks after listing, the study period, outlier observations, the original trading locale of the

newly listed stocks, the peculiarities of the exchange where the stock becomes listed, and the

correction of an overreaction that occurs on the impending listing announcement, but no full

explanation can account for the drift. Dharen and Ikenberry (1995) and Webb (1999) suggest

that poor post-listing performance appears related to managers timing their listing decisions in

which firms applying for listing as their performance peaks and subsequently deteriorates.

However, consistent with the pseudo market timing hypothesis by Schultz (2003), Cheng (2005)

finds no underperformance of new listings in the post-listing period by using calendar-time

analyses.2 The debate related to post-listing drift and its explanation thereby remains unsettled.

1 See Van Horne (1970), Ying et al. (1977), Grammatikos and Papaioannou (1986), Sanger and McConnell (1986),McConnell and Sanger (1987), Baker and Edelman (1991), Edelman and Baker (1994), Dharan and Ikenberry(1995), Webb (1999), and Jain and Kim (2006), etc.2 Schultz (2003) argues that managerial actions that coincide with peaks in their firms’stock prices do notnecessarily mean that the managers can forecast the overall market, but instead they can tell when their firms’stock prices are high, and by taking actions at these times, there can be the appearance of foresight.

3

The inconclusive findings spur our interest in questions of whether managers conduct

unusually aggressive earnings management via income-increasing adjustments around the time

of the exchange listing and whether such the accrual behavior can explain a portion of the

post-listing puzzle. If switching firms manage earnings upward before the listing change and

the market fails to perceive that the earnings management represents transitory increases in

earnings, these firms will be overpriced due to investor optimism around the listing

announcement. Subsequently, when earnings management reverses and switching firms report

declined earnings in the post-listing period, the disappointed investors would revise their

valuation downwardly.

Most of the studies on management reporting behavior around firm-specific events,

including management buyout (DeAngelo, 1986; Perry and Williams, 1994), IPOs (Aharoney

et al., 1993; Teoh et al., 1998a), SEOs (Teoh et al., 1998b; Rangan, 1998), and stock repurchase

(Louis and White, 2007; Gong et al., 2008), have interpret the evidence of earnings

management as support of managerial opportunism. However, given that an exchange listing

announcement may serve as an expression of managers’confidence in the future prospects of

the firm (Ying et al., 1977; Baker and Edelman, 1991), managers of switching firms might

have incentives to credibly use discretionary accruals, the proxy of earnings management, to

signal private information rather than to mislead investors. Consistent with this expectation,

prior studies indicate that managers use their reporting discretion to signal their private

information through accruals.3 For example, Louis and Robinson (2005) suggest that combing

the accrual information with other corroborating signals may enhance the credibility of

accruals as a means of signaling managerial optimism rather than managerial opportunism. As

accounting research suggests that accrual manipulation is not an unmitigated evil; within limits,

it promotes efficient decisions.

It is, therefore, interesting to explore the impact of exchange listing decisions on managers’

motivation for the accrual behavior and ascertain whether the information on the discretionary

accruals provides any insight into the valuation of firms changing trading marketplace. The

purpose of this study is to investigate firms’reporting behavior before switching stock

exchanges and the effect of pre-listing discretionary accruals on the market reaction to listing

3 See Watts and Zimmerman (1986), Subramanyam (1996), Demski (1998), and Arya et al. (2003).

4

announcements and on post-listing abnormal stock returns. By examining the association of

pre-listing accruals with short-term stock returns around exchange listing announcements and

long-term post-listing performance, we hope to observe whether managers’reporting behavior

before listing is priced at the listing announcements and whether the accrual pattern before

listing change is driven by managerial optimism or by managerial opportunism as well as the

potential explanation of the post-listing drift.

Based on firms switching trading markets from the NASDAQ or AMEX to the NYSE and

from the NASDAQ to the AMEX during 1986–2005, the empirical results show that switching

firms typically overstate their earnings by reporting significantly positive discretionary

accruals in the year preceding an exchange listing announcement, subsequently followed by

poor long-run post-listing earnings and insignificant abnormal accruals. Consistent with prior

studies, the average switching firm also experiences significantly positive listing

announcement abnormal returns. Though there is no evidence of a reversal of the stock price

effects of the earnings management at the listing announcements, the pre-listing abnormal

accruals are negatively associated with the post-listing performance, with the association

driven mainly by those firms with poor earnings in both current and future periods. The inverse

relation is robust with respect to an extensive set of controls, alternative accrual measure using

cash flow data and testing methods. Overall, the results suggest that the extant evidence of

post-listing puzzle by switching firms is partly attributable to the reversal of the price effects of

pre-listing earnings management, supporting the managerial opportunism hypothesis.

More specifically, the study further suggests that managerial consideration of both current

and expected future earnings affects their reporting behavior around exchange listings. The

evidence of pre-listing upward earnings management is observed especially for firms with both

poor current and future earnings and for those with poor current and good future earnings. For

firms with good performance in both current and future periods and those with good current

and poor future performance, we find evidence of downward earnings management prior to the

exchange listing announcements. Interestingly, the findings on managers’discretionary accrual

behavior reveal that the incentive to inflate (deflate) earnings prior to exchange listing

increases (decreases) in anticipation of good future performance by managers. Regression

results conditional on subsamples with different combinations of current and future expected

earnings indicate no significant relation between pre-listing earnings management and stock

5

returns over the listing announcement period but a significantly negative relation between prior

discretionary accruals and the cross-sectional variation in post-listing earnings and stock

returns for firms with poor performance in both current and future periods. Since the effect of

opportunistic earnings management has to reverse over time, this evidence is consistent with

the conjecture that managers of these firms are likely to use their reporting discretion to

opportunistically mange earnings upward prior to exchange listings, providing support for the

managerial opportunism hypothesis accordingly.

As argued by Fudenberg and Tirole (1995) that managers prefer a smooth profile of earnings

streams over time via career concerns, and thus have incentives to increase current earnings in

bad times by borrowing against future earnings. Likewise, managers have incentives to

decrease current earnings in good times and transfer them to future periods when future

earnings are expected to be poor. Unlike opportunistic earnings management, the smoothing of

earnings across time apparently should have little reversal impact. In this regard, the failure to

find a significant association between pre-listing earnings management and stock returns

around and following the exchange listing announcements is consistent with the earnings

smoothing argument. This result is held for switching firms with good current and poor future

earnings and those with poor current and good future earnings.

With respect to sample firms with good earnings in both current and future periods, we also

find no significant association between pre-listing earnings management and stock returns

around and following the exchange listing announcements. Given that the average firm with

good performance in both current and future periods is observed to report income-decreasing

accruals prior to the listing announcement, it is inconsistent with the signaling hypothesis as if

managers who intend to signal favorable private information have little incentives to manage

earnings downward prior to exchange listings. Therefore, our extended tests find some

evidence indicating that these firms tend to make a strategic income-decreasing choice under

the concerns of avoiding competition in the industry and/or mitigating adverse political

attention. This is consistent with the conjecture that managers of switching firms have the

countervailing incentive to deflate earnings prior to the exchange listings for reasons other than

the smoothing motive.

This study differs from the previous research in two ways. First, it extends prior research on

6

exchange listings by providing new insights about the value implication of managers’reporting

behavior before the listing change. Second, by investigating the effects of managers

considering current and expected future earnings on their reporting behavior around exchange

listings and such accounting choices on the listing announcement period returns as well as

long-run post-listing stock performance, it further identifies conditions when managers are

expected to have incentives to make a particular accounting choice (e.g., income-increasing or

income-decreasing) in the case of exchange listings. In this way, this paper contributes to the

existing literature on the informational role of managerial reporting discretion around specific

corporate events. Moreover, besides managers’income-increasing discretionary accrual

behavior, this paper in a way attempts to find out counterevidence by taking account of

managers’expectations of future earnings for reporting negative discretionary accruals in the

exchange listing setting.

The remainder of the paper proceeds as follows. Section 2 discusses the research questions

and the empirical hypotheses. Section 3 then describes the data and methodology, while

Section 4 presents the empirical results. Finally, Section 5 summarizes the main conclusions

and implications.

2. Hypotheses development

2.1 Managerial motivations for earnings management around exchange listings

We propose two different hypotheses to explain the possible motivations for managers to

manage earnings surrounding exchange listings: the managerial opportunism hypothesis and

the complementary signaling hypothesis. These hypotheses imply same pre-listing financial

reporting behavior by managers, but have different implications for the relation between

pre-listing abnormal discretionary accruals and abnormal stock returns around exchange listing

announcements. They also suggest different associations between abnormal discretionary

accruals prior to exchange listings and long-run abnormal stock returns following exchange

listings. Table 1 presents a summary of the predictions of the various hypotheses discussed

below.

[Insert Table 1 Here]

2.1.1. Managerial opportunism hypothesis

7

The exchange listing process offers managers both motivation and opportunities to manage

earnings. As Merton (1987) contends that the publicity associated with exchange listing

reaches some investors who were previously unaware of the stock and investors only have

incomplete information, there is high information asymmetry between investors and managers

at the time of the exchange switching. Because exchange listing undergoes a transition between

two segmented markets, managers thus may engage in earnings management to report higher

earnings to attract investors from the new market. Given the self-interest motive, managers

may use their discretion in accrual accounting to mislead investors (Rangan, 1998; Teoh et al.,

1998b). If this is the case, abnormal accruals before listing are associated with management’s

opportunistic behavior. This managerial opportunism hypothesis indicates that investors will

fail to recognize such earnings manipulations and naively extrapolate pre-listing earnings

management at the listing announcements, and the market will misvalue the switching firm’s

equity. An insignificant relation between pre-listing abnormal accruals and the market reaction

to the listing announcements would be anticipated. Subsequent to the listing, the reversal of the

price effects of earnings management would induce investors to revise their assessment of prior

bloated earnings and consequently, a negative correlation between pre-listing abnormal

accruals and post-listing abnormal stock returns should be observed.

2.1.2. Complementary signaling hypothesis

An exchange listing may serve as an expression of managers’confidence in the future

prospects of the firm (Ying et al., 1977; Baker and Edelman, 1991). Listing has information

content because of the independent evaluation and approval by an exchange. The market,

therefore, may perceive the exchange listing as a seal of approval that has positive value. A

valid signal conveys useful information and is costly enough to prevent dishonest signals.

Firms are unlikely to use the listing change as a mechanism for generating false signals

because of the costs associated with the exchange listing. By communicating their favorable

private information, managers reduce information asymmetry between outside investors and

management, and thus raise stock prices to higher equilibrium values. Under this scenario,

managers who are optimistic about their firms’prospects appear likely to have incentives to use

discretionary accruals as a supplement to exchange listing announcements rather than as a

manipulation device to deceive the market. This complementary signaling hypothesis suggests

that pre-listing abnormal accruals and listing announcement period returns would be positively

8

correlated (Louis and Robinson, 2005). Moreover, a positive or an insignificant relation

between pre-listing abnormal accruals and long-run post-listing performance would be

observed. If abnormal accruals before listing are priced incompletely by the market at the time

of listing announcements, there would be a positive relation between pre-listing abnormal

accruals and post-listing performance. Alternatively, if abnormal accruals before listing are

completely captured by the market at the listing announcements, there would be no significant

association of pre-listing abnormal accruals with post-listing performance.

2.2 The effect of managerial consideration of current and future expected earnings

Theoretically, specific corporate events (e.g., exchange listings) should not affect all firms

equally with respect to managerial reporting behavior of discretionary accruals. Consistent

with this view, Fudenberg and Tirole’s (1995) analytical model justifies that managerial

incentives for earnings management depend on their consideration of both current and future

relative performance. As the discretionary accruals usually reverse in a future period, the

impact of earnings management is often transient. In this environment, Fudenberg and Tirole

(1995) suggest that managers choose current discretionary accruals partially in anticipation of

future earnings. Consequently, managers’incentives to manipulate earnings might differ for

firms with different combinations of current and future relative performance. Such a concern is

particularly of importance for the exchange listing case since these firms tend to perform well

before the listing but realize poor post-issue stock and operating performance.

To further explain managers’motive for earnings management around exchange listings and

distinguish between the managerial opportunism and the complementary signaling hypotheses,

a separate examination of pre-listing managerial reporting choice for firms with different

combinations of current and future relative performance is required. To this end, switching

firms are partitioned into 4 subgroups, including firms with both poor current and future

performance (Cp, Fp), firms with poor current and good future performance (Cp, Fg), firms

with good current and poor future performance (Cg, Fp) and firms with both good current and

future performance (Cg, Fg). If, as we expect, managers’consideration of current and future

performance affect their intentions to manage earnings around the exchange switching, we

would observe discernible patterns of the abnormal accrual behavior for firms with different

combinations of current and future relative performance.

9

In the setting of exchange listings, we especially focus on firms where both current and

expected future performance are poor (Cp, Fp) and those where both current and expected

future performance are good (Cg, Fg). Based on the agency frictions perspective and

managerial myopia theory,4 we argue that managers of switching firms have strong incentives

to opportunistically report income-increasing discretionary accruals to distract investors’

attention if both current and expected future performance are poor (Cp, Fp). Information

asymmetries between managers and outside investors underpin the incentives for managers in

these firms to dress up reported earnings in bad times to fulfill the exchange listing process,

even though doing so exacerbates the poor future performance. To the extent that managers are

believed to behave myopically, we argue that managers of these firms make an

income-decreasing reporting choice either for the benefit of existing shareholders and/or for

their own benefit.5

By contrast, we expect that managers appear likely to have incentives to use discretionary

accruals as a supplement device of communicating favorable private information about

exchange listings’prospects if both current and expected future performance are good (Cg, Fg)

under the signaling framework. Given that exchange listings would attract enhanced scrutiny at

the time from market monitors such as analysts, underwriters, auditors, boards, the press and

the other parties to the news, as well as enhanced regulatory scrutiny. There is a risk of

subsequent detection of prior earnings management, and hence litigation and regulatory action,

because earnings management can only ‘‘borrow’’ earnings from other periods: earnings

inflation causes subsequent earnings deflation. Because of such litigation and/or reputation risk

associated with inflated earnings and false information, firms are unlikely to signal untruthfully.

Rather, firms are more likely to use reporting discretion to convey their optimism when they

are confident that future performance will meet the expectations raised by the discretionary

accruals.

Additionally, according to Fudenberg and Tirole’s (1995) analytical model, earnings are

smoothed to mitigate the effects of transitory cash flows and adjust reported earnings towards a

4 While the term managerial myopia may conjure up images of managerial "opportunism," managerial myopia,according to Stein (1989), refers to the desire to achieve a high current stock price by inflating current earnings atthe expense of longer-term cash flows or earnings.5 Stein (1989) and Bhojraj and Libby (2005) document that, controlling for agency frictions, managers will makemore myopic project choices in response to increased stock market pressure.

10

more stable trend among those firms with poor current and good future performance (Cp, Fg)

and those with good current and poor future performance (Cg, Fp) via career concerns.

Following Fudenberg and Tirole’s forethought, we expect that managers have indirect

incentives to manage earnings up for (Cp, Fg) firms and down for (Cg, Fp) firms prior to

exchange listings by contributing to, among other things, their job security and future

compensation.

By taking into account of managers’anticipation of future earnings, we formalize

managerial accrual choices before exchange listings as the two-by-two outcome in response to

relative performance depicted in Table 2. The columns partition current relative performance

and the rows partition expected relative performance. Based on the above reasoning, we make

predictions about the signs of discretionary accruals for firms with different combinations of

current and expected future earnings. Switching firms in cell (Cp, Fp) are expected to have an

incentive to overstate earnings to go through the listing screening and in turn mislead investors,

while firms in cell (Cg, Fg) are expected to signal favorable information via positive

discretionary accruals. For firms in cells (Cg, Fp) and (Cp, Fg), we expect that managers of

these firms have incentives to use income-decreasing and income-increasing discretionary

accruals respectively to smooth earnings streams around exchange listings under the concerns

about job security.


3. Data and methodology

3.1 Sample selection and data description

We examine managerial reporting behavior based on listings of Nasdaq stocks on the Amex

or the NYSE and listings of Amex stocks on the NYSE from CRSP daily file database between

January 1986 and December 2005.6 We choose this period primarily because it includes most

of the exchange-listing firms, thus providing an adequate sample size for testing. We include

ordinary common stocks only (CRSP share codes 10 or 11) and exclude real estate investment

trusts (REITs), closed-end funds, and American Depositary Receipts. We also exclude

financial firms, because these firms are subject to different regulatory requirements and have

6 It is rare for a company to move from NYSE to AMEX, or from a national exchange to NASDAQ.

11

distinctly different financial characteristics. To be included, the firm must have necessary data

on stock price and financial statement items, including annual earnings, book value and share

information, with positive values of total assets, total liabilities, stockholders’equity, and sales.

We obtain financial data from Compustat and stock price and returns data from CRSP.

Analysts’ forecasts and institutional holdings are obtained from IBES and CDA Spectrum. To

reduce the effects of a few extreme values, we winsorize each variable at the 1% and 99%

percentiles. Our results are in general robust to winsorization.

The final sample consists of 542 firms, including 341 (63%) from NASDAQ to NYSE, 65

(12%) from NASDAQ to AMEX, and 136 (25%) from AMEX to NYSE. Table 3 lists the

number of firms by the listing year in Panel A and by industry in Panel B. Table 1 shows that

the number of firms switching stock exchanges is increasing in the late 1990s and the majority

of the switching firms come from manufacturing and services industries, which constitute

about 69% of the total sample.


Table 4 summarizes the characteristics of sample firms as of the year prior to the listing

change. Table 4 shows that NASDAQ firms moving to the NYSE have an average market

value of $608.70 millions, total assets of $494.55 millions, M/B of 2.08 and expected earnings

(analysts’one-year-ahead earning forecasts) of 0.90, higher than that of the AMEX-to-NYSE

firms ($538.02 millions, $455.05 millions, 1.82 and 0.88). Among the switching sample, the

NASDAQ-to-AMEX firms have a much lower market value ($72.92 millions), total assets

($122.32 millions), M/B (1.32) and expected earnings (0.45), consistent with the normal

findings on exchange listings. In addition, NASDAQ-to-NYSE firms have the highest mean

(median) number of analyst following and mean (median) percentage of institutional

stockholdings, consistent with the notion that institutional investors and analysts prefer large

visible firms.


3.2 Measuring earnings management

As Subramanyam (1996) points out that discretionary accruals are on average priced by the

12

market, this paper therefore uses them to measure managerial discretion over the amount of

earnings adjustments. We use a modified Jones (1991) model to measure discretionary accruals

and derive abnormal accruals from the difference between actual and expected accruals.

Expected accruals are estimated through a cross-sectional adaptation of the modified-Jones

model and after adjusting for changes in the firm-specific economic conditions which influence

accruals independently of earnings management:

)/(̂]/)[(̂)/1(̂)/( 1,1,21,,,11,01,, titititititititi TAPPEbTAARSALESbTAbTAACCE (1)

where tiSALES , is the yearly change in sales, tiAR , is the yearly change in accounts

receivable, tiPPE , is property, plant, and equipment, and 1, tiTA is total assets at the

beginning of the year. For each sample firm, the model parameters 0̂b , 1̂b , and 2̂b are

estimated cross-sectionally from the regression for event years -3 to +3 using all firm-year

observations in the same 2-digit SIC industry as the sample firm (but excluding the sample

firm):

tititititititititi TAPPEbTAARSALESbTAbTAACC ,1,1,21,,,11,01,, )/(]/)[()/1(/ (2)

where tiACC , is the actual accruals measured as the change in non-cash current assets minus

the change in current liabilities excluding the current portion of long-term debt, minus

depreciation and amortization (Dechow et al., 1995), and 0b , 1b , and 2b are the

cross-sectional estimates of 0̂b , 1̂b , and 2̂b from the modified-Jones model. To ensure that

the estimated coefficients created are not biased, we require there be at least ten two-digit SIC

code peers.

According to Field et al. (2001), the discretionary accruals approach without adjusting for

performance may cause serious inference problems related to incentives for earnings

management. Moreover, Dechow et al. (1995) indicate that accrual-based models for detecting

earnings management may result in biased results when applied to samples of firms with

extreme financial performance. Adjusting for performance is hence adequate and beneficial in

this sense. Such a concern is particularly of importance for firms switching exchanges since

they tend to perform well before listing.

13

Following Kothari et al. (2005), we also examine sample firms’ discretionary accruals

adjusted for performance-matched firms’ discretionary accruals on the basis of prior-year

return on assets (ROA) and industry. We define the performance-matched modified-Jones

model discretionary accrual for each sample firm as the modified-Jones model discretionary

accruals minus the matched firm’s modified-Jones model discretionary accruals. By relying on

such matching-firm research design, firms identified as having abnormally high or low levels

of discretionary accruals are more likely to manage earnings than the comparison sample. In

other words, this approach provides additional controls for what is considered normal earnings

management, and therefore the earnings management labeled by this approach would be

viewed as abnormal earnings management. A measure of discretionary accruals using data

from the statement of cash flows and based on Ball and Shivakumar (2008) nonlinear Jones

model will be examined for robust testing in Section 4.6.

Following Lie (2001, 2005), we choose matching firms that closely resemble the sample

firms in industry classification (two-digit SIC code), level of performance in year –1 (OP–1),

change in performance in year –1 (△OP–1), and market-to-book ratio in year –1 (M/B–1).

Specifically, we identify matching firms with the following characteristics: (1) a level of

operating performance between 80 percent and 120 percent of the sample firm’s level of

operating performance in year –1; (2) a change in operating performance between 80 percent

and 120 percent of the sample firm’s change in operating performance from year –2 to year–1;

and (3) a market-to-book ratio between 80 percent and 120 percent of the sample firm’s

market-to-book ratio in year–1. From this initial sample of matching firms, we select the firm

that minimizes the following function:

|OP−1, sample firm−OP−1, matching firm | +|△OP−1, sample firm−△OP−1, matching firm |

+ |M/B−1, sample firm−M/B−1, matching firm |. (3)

If we cannot find a firm meeting condition (a), then we search for firms with a level of

operating performance within ±0.01 of the level of operating performance of the sample firm.

If we cannot find a firm meeting condition (b), then we search for firms with a change in

operating performance within ±0.01 of the change in operating performance of the sample firm.

If we cannot find a firm meeting condition (c), then we search for firms with a market-to-book

ratio within ±0.1 of the market-to-book ratio of the sample firm.

14

If we do not find any firms that meet these criteria, we repeat the process first for matching

firms with the same one-digit SIC code as the sample firms, and then for all firms

independently of their SIC code. If we still do not find any matching firms, we choose the

matching firm that minimizes equation (3) independently of the filters. Lie (2001) shows that

this performance-adjusted benchmark yields more powerful test statistics than do other

benchmarks. To reduce confounding effects, we only consider matching firms that do not

announce a listing change during the year of the event and during the 3 years before and after

the event.

3.3 Methodology

To investigate whether firms report positive discretionary accruals prior to exchange listing

announcements, we first study the time-series profile of accruals behavior from fiscal years -3

to +3 relative to the fiscal year of the listing change (year 0). We next examine the relation

between pre-listing abnormal accruals and the price effect at the listing announcements for all

sample firms:

.

/

2005

19877

654321

YearOwnershipnalInstitutio

FollowingAnalystDregulatedSizeLeverageMBDACAR(4)

The dependent variable, CAR, is the three-day (1, +1) announcement-period market-adjusted

return, where day 0 is the event date and the market return is proxied by the CRSP

value-weighted return.7 DA is discretionary accruals in year 1 based on the modified Jones or

performance-matched modified Jones models. B/M is the ratio of book-to-market value of

equity in year 1. Leverage is the total liabilities divided by total assets in year 1. Size is the

natural log of market value of common stock in year 1. Dregularted is a dummy equal to one

for firms classified in regulated industries (SIC codes 40-49). Analyst following is the natural

log of the average number of analysts’ forecasts included in the monthly consensus during year

1.8 Institutional ownership is the average proportion of shares held by institutional investors

at the end of year 1. Year is a switching year dummy.

7 We also estimate three-day announcement-period abnormal returns by using the standard market modelprocedure with the parameters estimated for the period 200 days to 60 days before the announcement. The resultsare similar.8 Firms not covered by IBES are assumed to have zero analyst coverage (as in Chang et al., 2009).

15

Besides the discretionary accruals, we also control for various firm characteristics that may

affect investors’perception of exchange listing announcements. These factors include growth

opportunities, financial health, and the richness of information environment. Switching firms

normally exchange-list to exploit growth opportunities (Dharen and Ikenberry, 1995), to

upgrade debt capacity by relaxing financing constraints (Yuan et al., 2009), or to improve their

information environment by reducing the shadow cost of not knowing about their security by

investors (Merton, 1987; Kadlec and McConell, 1994). Following prior studies, we measure

growth opportunities as the ratio of book value of equity to market value of equity in year -1,

financial health as the ratio of debt to total assets in year -1, and the richness of information

environment as firm size, analyst following and institutional ownership in year -1. In addition,

if a firm is regulated, then it is more difficult for earnings information impounded in stock

prices since regulation constrains the determination of accounting numbers (Warfield et al.,

1995). We thus employ a dummy variable, Dregularted, to measure the impact. We also

include a set of year control dummies to account for business cycle effects.

We also examine Equation (4) for four subgroups of firms with different combinations of

current and future expected performance. We follow DeFond and Park (1997) and measure

good or poor performance relative to the sample median by industry and fiscal year. Current

premanaged earnings are used as the current performance and are measured as current period

earnings minus discretionary accruals. Here we treat year 1 as the current period because

prior studies note that firms tend to report higher discretionary accruals prior to the corporate

event. Median analysts’ forecasts of one-year-ahead earnings in year 1 from the IBES

database are used to proxy for managements’ expectations of future earnings performance.

Finally, we examine the relation between pre-listing abnormal accruals and post-listing

long-run stock performance:

2005

198798

7654321 /

YearPLBHAROwnershipnalInstitutio

FollowingAnalystDregulatedLbhrvSizeLeverageMBDALBHAR

(5)

The dependent variable, LBHAR, is the natural logarithm of one plus three-year buy-and-hold

raw returns (BHARs) of switching firms minus the natural logarithm of one plus three-year

16

BHARs of matching firms. 9 The computation of BHARs begins the day after the

announcement of exchange listing and continues through the three-year period (756 trading

days) following the announcement.10 Lbhrv is the value-weighted buy-and-hold market returns

over the three-year period following the exchange listing announcement. PLBHAR is the

buy-and-hold abnormal returns over a one-year period preceding the listing announcement.

The other variables are defined as above. The control variables in Equation (5) follow those

suggested by other studies (e.g., Ritter, 1991; Teoh et al., 1998a; Morsfield and Tan, 2006).

We will also examine Equation (5) for all sample firms and for four subgroups of firms with

different combinations of current and future expected performance.

As before, we control for firm characteristics that have been shown to be related to

post-event stock performance in previous studies when examining the relation between

pre-listing discretionary accruals and post-listing long-run stock returns. Omitting these factors

would bias the statistical inference of earnings management. For example, Dharan and

Ikenberry (1995) find that poor post-listing performance is more severe for smaller firms and

those not widely held by institutional investors. Fama and French (1992) and Lakonishok et al.

(1994) show that stock returns tend to be associated with firm size and book-to-market ratios.

In addition, the existing literature shows that analysts tend to cover firms with a better

information environment, and firms that miss analyst forecasts usually suffer significant

declines in their stock price (e.g., Lang and Lundholm, 1996; Francis et al., 1998; Bushman et

al., 2005). As prior studies indicate that switching firms typically experience superior stock

performance before exchange listings,11 pre-listing stock returns are included to control for the

mean reversion effect of prior run-ups. Similarly, we include a regulated industry dummy to

capture the regulatory effect and use leverage to measure the influence of financial condition.

Also, contemporaneous market returns and a set of year dummies are included to control for

the aggregate market trend and business cycle effects.

4. Empirical results

9 The method of matching firm-adjusted buy-and-hold abnormal returns to estimate long-run stock performanceis widely used in previous studies (e.g., Lakonishok et al., 1994; Chan et al., 2001) and found to be attractive incomparison to other techniques (e.g., Barber and Lyon, 1997; Kothari and Warner, 1997). Our choice of thematching firm is based on the criteria described in the previous section.10 If a firm is delisted, its stock returns are set to zero for the rest of the period.11 Also see note 1.

17

4.1 Operating performance and abnormal accruals around the exchange listings

To investigate whether firms overstate earnings around exchange listings, we initially

analyze the abnormal net income of sample firms in the three years before and the three years

after the exchange listing announcement. We then examine the time series profile of accruals to

evaluate whether they correspond to the pattern of earnings. Panel A of Table 5 reports both

unadjusted and industry-adjusted median annual operating performance and the associated sign

tests. Unadjusted performance is simply the operating performance for the firms that announce

changing stock exchanges. Industry-adjusted performance is the unadjusted performance minus

that of the median firm in the same industry (2-digit SIC code). We report the median

performance as they are not likely to be influenced by extreme observations, although the

results are similar if we use the mean performance. The number of observations varies because

of data availability.


The results of unadjusted asset-scaled net income indicate improving pre-listing

performance but deteriorating post-listing performance. The median grows from 6.3% in year

-3 to a peak of 7.5% in year -1, then declines to 4.9% by year +3. The industry-adjusted net

income measure shows a similar pattern. The median industry-adjusted asset-scaled net income

varies from 0.8% in year -3 to 4.1% in year -1, and decline to -1% by year +3. The pattern of

pre-listing increases in earnings, followed by subsequent earnings declines, is in line with the

findings of Papaioannou et al. (2003). This earnings profile is consistent with managers

borrowing income from future periods to enhance earnings immediately around exchange

listings. This finding, however, can also be interpreted as managers timing exchange listings

following periods of unusually good financial performance, reflecting long-run mean reversion

in earnings (Fama and French, 2000).

We next examine the corresponding pattern of accruals. Panel B of Table 5 presents the

median discretionary accruals estimated using both the modified-Jones and

performance-matched modified-Jones models. The profile of discretionary accruals based on

the modified-Jones model shows that discretionary accruals are significantly positive in year -1

at a 0.8% of total assets, reflecting management of discretionary accruals before an exchange

listing. To remove potential bias in the modified-Jones model for high-performance firms, the

18

discretionary accruals of performance-matched control firm are subtracted from the switching

firm’s discretionary accruals. Similarly, performance-matched discretionary accruals rise to a

peak in year -1 and decrease following the listing, which reveals that relative to the comparable

non-switching firms, switching firms conduct abnormal accrual manipulations to deliberately

overstate earnings prior to the exchange listings. The evidence of the significantly positive

abnormal accruals observed in the year immediately prior to the exchange listing

announcements, combined with the improvements in earnings performance reported earlier, is

consistent with switching firms employing accruals to deliberately overstate earnings before

exchange listings.

4.2 Pre-listing abnormal accruals partitioned by different combinations of current and

future relative performance

To examine whether the behavior of discretionary accruals preceding the listing change is

associated with managerial consideration of both current and future expected earnings

delineated in Table 2, we partition the whole sample into four subsmaples in Table 6, where

Wilcoxon signed-rank test is used to test the hypothesis that the medians are equal to zero.

From Table 6, we find that pre-listing positive median discretionary accruals for the whole

sample are mainly driven by both (Cp, Fp) and (Cp, Fg) groups, i.e., by firms with poor

performance in both current and future periods and those with poor current and good future

performance. The median former (latter) reports statistically significant abnormal accruals of

about 8.7% (9.4%) of total assets based on the performance-matched modified Jones model. In

contrast, firms with good performance in both current and future periods (Cg, Fg) and those

with good current and poor future performance (Cg, Fp) report significantly negative

discretionary accruals. The median former (latter) reports significant abnormal accruals of

about -4.8% (-3.4%) of total assets based on the performance-matched modified Jones model.

As for the accrual measure based on the modified Jones model, we find similar reporting

behavior for firms with different combinations of current and future relative performance.


Remarkably, we find that switching firms with poor current earnings adopt more


19


poor future earnings. This evidence is consistent with the intuition from Fudenberg and Tirole

(1995) that managers choose current discretionary accruals partially in anticipation of future

earnings. Among these four groups of firms, the patterns of managed accruals of (Cp, Fp), (Cp,

Fg) and (Cg, Fp) firms concur with our predictions depicted in Table 2. However, (Cg, Fg)

firms are found to report significantly negative discretionary accruals instead of positive

discretionary accruals, which is inconsistent with our predictions and thus does not support

managers’signaling intent around exchange listing.

An alternative argument is that managers of (Cg, Fg) firms might have an incentive to

strategically report income-decreasing discretionary accruals to avoid competition or potential

political costs associated with high profitability and market dominance. With the improved

visibility associated with exchange listings (Merton, 1987; Kadlec and McConell, 1994, Baker

et al., 1999), switching firms with highly visible earnings and market dominance may face

enhanced market interest and regulatory scrutiny, which in turn heighten the market pressure

on management. In this situation, such firms may have incentives to understate profits to avoid

adverse political attention (Watts and Zimmerman, 1978) or to avoid increased competition in

their industry as pointed out by Ronen and Yaari (2008) that financial reporting contains

information that is useful for rivals’ decision making. Consistent with this view, Raith (2003)

shows that changes in competition affect managerial incentives when expected profitability and

value of managerial efforts are higher and thus stronger managerial incentives are provided to

reduce costs. We will test this interpretation in Section 4.5.

4.3 Association between pre-listing abnormal accruals and the market reaction to exchange

listing announcements

We study the relation between pre-listing accruals and abnormal returns around exchange

listing announcements by first examining the market reaction to the listing announcements and

differences in market response of high and low accrual firms. We group the sample into high

and low accrual firms based on whether the level of pre-listing discretionary accruals of a firm

is above the median discretionary accruals in year -1. Consistent with prior studies, we find a

significant positive market reaction over the three days around the listing announcement from

Panel A of Table 7. The mean (median) abnormal return is 1.04% (0.75%) for the sample used

20

in the modified Jones model and the mean (median) market-adjusted return is 0.98% (0.65%)

for the sample used in the performance-matched modified Jones model, both significant at the

0.01 level. The results are robust to using the CRSP equally-weighted return as the market

return proxy. For the differences in the listing announcement abnormal returns of high and low

accrual firms, the results in Panels A and B of Table 7 show that, relative to high accrual firms,

the mean and median positive listing announcement abnormal returns are higher for low

accrual firms, implying that the managed accruals somewhat influence investors’valuation at

the exchange listing announcements.


To further explore whether investors assess switching firms’accruals behavior, we examine

the link between pre-listing earnings management and the abnormal returns over the listing

announcements. Table 8 reports the regression results conditional on full sample and

subsamples with different combinations of current and future relative performance. The

t-statistics for cross-sectional test in Table 6 and elsewhere in this paper are based on White’s

heteroskedasticity consistent estimator for standard errors. The coefficients on both measures

of discretionary accruals are insignificant for full sample and four subsamples, suggesting that

investors fail to understand the pre-listing discretionary accruals at the listing announcement.

The results show that, irrespective of the proxy used to measure earnings management, there is

no significant relation between market price reaction to the listing announcements and prior

earnings management. This is inconsistent with either the notion that investors rationally infer

earnings management from exchange listing announcements and undo the effects of earnings

management at the time of exchange listing announcements, or the viewpoint that manager use

discretionary accruals in conjunction with the exchange listing signal to convey their private

information to the market. To some extent, this evidence appears to concur with the point that

investors are naïve and fail to decipher the earnings management signal at the time of exchange

listings.


4.4 Association between pre-listing abnormal accruals and post-listing stock performance

Given the insignificant relation between pre-listing discretionary accruals and the market

21

response to the listing announcement presented in the previous section, we then examine the

association of long-run post-listing stock returns with prior discretionary accruals based on the

managerial opportunism argument. If abnormal accruals before listing are associated with

managers’opportunistic behavior, then the pre-listing abnormal accruals should be negatively

correlated with the post-listing stock performance since the subsequent reversal of the price

effects of earnings management would induce investors to revise their assessment of prior

bloated earnings. Table 9 reports the results of the regressions of the long-run post-listing stock

performance on prior discretionary accruals conditional on full sample and subsamples with

different combinations of current and future relative performance.


The coefficients on both measures of discretionary accruals are significantly negative for full

sample and the subsample of (Cp, Fp) firms, suggesting that the long-run stock performance is,

at least, partially attributable to the reversal of the effects of pre-listing earnings management

for the average firm and that the reversal of prior earnings management effects are driven

especially by firms with both poor current and future performance. The evidence is consistent

with the prediction of the managerial opportunism argument. For other three subsamples,

however, we again find no significant association between pre-listing discretionary accruals

and post-listing stock performance. As we contend earlier, the failure to find a discernible

association between pre-listing earnings management and stock returns around and following

the exchange listing announcements is not explained by the arguments of managerial

opportunism and signaling, but might be consistent with alternative conjectures, such as the

“earnings smoothing”or the“political cost/competition”concern.

According to Fudenberg and Tirole’s (1995) theory, managers prefer a smooth profile of

earnings streams over time and thus have incentives to increase current earnings in bad times

by borrowing against future earnings. In contrast, managers have incentives to decrease current

earnings in good times and save them for possible use in the future when future earnings are

expected to be poor. Unlike opportunistic earnings management, the smoothing of earnings

across time apparently should have little reversal impact. In this vein, a finding of insignificant

association between pre-listing earnings management and stock returns around and following

the exchange listing announcements in connection with the distinguishable behavior of

22

pre-listing discretionary accruals would be consistent with the “earnings smoothing”argument

by Fudenberg and Tirole (1995). Therefore, the results reported in Tables 6 and 7, together

with the evidence on the discretionary accrual behavior in Table 5, for subsamples of firms

with good current and poor future earnings (Cg, Fp) and those with poor current and good

future earnings (Cp, Fg) offer support for this“earnings smoothing”intent.

4.5 Extended test of the abnormal accrual behavior of (Cg, Fg) firms

With respect to sample firms with good earnings in both current and future periods (Cg, Fg),

we also find no significant association between pre-listing earnings management and stock

returns around and following the exchange listing announcements. As reported in Table 4 that

the average (Cg, Fg) firm is observed to report income-decreasing accruals prior to the listing

announcement, it is inconsistent with the signaling hypothesis as if managers who intend to

signal favorable private information have little incentives to manage earnings downward prior

to exchange listings. An advanced alternative to the signaling intent is that managers of (Cg, Fg)

firms might have an incentive to strategically manage earnings downward to avoid competition

or to avoid adverse potential political attention associated with high profitability and market

dominance. To verify this conjecture, we analyze the competition and political cost

characteristics for this subsample relative to other three subgroups.

Following prior studies (e.g., Harris, 1998; Karuna, 2007), we investigate several

dimensions of competition, including the level of concentration (Herfindahl index), product

substitutability, market size, entry costs, and the number of firms in an industry. The

Herfindahl index measures industry concentration by summing the squared fraction of sales of

the firms in the industry. The degree of product substitutability is measured by the price–cost

margin, which is computed as industry sales to operating costs for firms in a given industry.12

Market size reflects the density of consumers in a market or industry. We measure an

industry’s market size by the amount of industry sales. The minimum level of investment

needed to enter an industry proxies for entry costs in that industry, and is measured by the

weighted average gross value of the cost of property, plant and equipment for firms in that

industry, weighted by each firm’s market share in that industry. The number of firms in an

12 Hence, the smaller the price–cost margin, the greater the intensity of price competition due to highersubstitutability.

23

industry represents the firm number in the same industry as the switching firm. In general,

lower concentration, greater product substitutability, greater market size, lower entry costs and

more number of competitors in an industry reflect greater competition.

Besides competition, we also follow prior research (e.g., Watts and Zimmerman, 1978) and

analyze proxies for potential political costs, including size, tax rate, institutional ownership and

analyst following. Tax rate is calculated by income taxes divided by net income before income

taxes. Size, institutional ownership and analyst following are defined in the same way as before.

Firms with greater size, tax rate, institutional ownership and analyst following are believed to

face severer adverse political attention. All the proxy variables of competition and political

costs are measured at the end of the year prior to the switching.


The results in Panels A and B of Table 10 demonstrate that compared with other three

subsamples, the subsample of (Cg, Fg) firms has significantly greater average competition

characteristics in lower Herfindahl index, entry costs and more number of firms in that industry

and is confronted by significantly higher political costs with larger firm size, institutional

ownership and analyst following. The findings, at some point, manifest that the more

competitive environment and/or severer adverse political attention is faced with in the industry

by (Cg, Fg) firms, mangers of these firms are more likely to report more income-decreasing

discretionary accruals before the listing.


To further inspect the influence of the degree of competition and political attention on

managers’income-decreasing reporting choice, we rank (Cg, Fg) firms by their per-listing

discretionary accruals to examine the within-group differences in their average competition and

political cost characteristics. Similar to the results found in Table 8, the analysis in Panels A

and B of Table 11 indicates the aggressive quartile with the most negative discretionary

accruals has greater average competition characteristics in lower Herfindahl index, entry costs

and more number of firms in that industry, and face higher average political costs

characteristics in larger firm size when compared with the conservative quartile with the least

negative discretionary accruals. It thus suggests that firms report more income-decreasing

24

discretionary accruals when industry competition is greater and/or adverse political attention is

severer. Though the differences between both extreme quartiles of (Cg, Fg) firms are mostly

insignificant, the tenor of the results obtained is unaffected.

Overall, the extended tests find some evidence indicating that (Cg, Fg) firms tend to make a

strategic income-decreasing reporting choice under the concern of avoiding competition in the

industry and/or mitigating adverse political attention. Hence, this finding lends support to the

conjecture that managers of switching firms have the countervailing incentive to deflate

earnings prior to the exchange listings for reasons other than the smoothing motive.

4.6 Sensitivity analysis

To determine whether the tests could be biased by the methodology employed, this study

considers the sensitivity of the results to the variations in testing implementation. First, we also

measure total accruals based on changes in cash flow statement data as a robustness check.

Total accruals are estimated using the Jones model and a piecewise linear variant suggested by

Ball and Shivakumar (2006, 2008):

ititititit

itititititititit

TAOCFDOCFDOCF

TAOCFTAPPETASALESTACFACC

)/(

)/()/()/(/_

154

131121101 (5)

where ACC_CF is total accruals, which are defined as earnings taken from the cash flow

statement (Compustat #123) minus cash flows from operations, also taken from the cash flow

statement (Compustat #308); OCF is operating cash flows taken from the cash flow statement;

and DOCF takes the value 1 if OCF<0, and 0 otherwise.

Likewise, model parameters are estimated separately for each sample firm from a

cross-section of all non-switching firms in its 2-digit SIC with data for event years -3 to +3.

Only industry-years with at least 10 observations are considered.

)]/(̂ˆ)/(̂)/(̂)/(ˆ[/__

154

131121101

itititit

ititititititititit

TAOCFDOCFDOCF

TAOCFTAPPETASALESTACFACCCFDAC

(6)

Discretionary accruals (DAC_CF) are computed as the difference between total accruals

(ACC_CF) and estimated nondiscretionary accruals (^ above parameters denote estimates). We

use this discretionary accrual measure to replicate the above analyses. As reported in the last

part of Tables 3 through 11, the results obtained share a similar pattern with those generated

25

using both the modified-Jones and performance-matched modified-Jones models. Therefore,

none of the inferences about managers’reporting behavior and the informational role of

pre-listing abnormal accruals in explaining stock returns around and following the listing

announcements for the full sample and subgroups is changed by the use of cash flow statement

data.


Secondly, apart from long-run stock performance, we also examine the relation between

pre-listing discretionary accruals and post-listing operating performance changes. As the

results from Table 3 show that the time series profile of earnings reflects long-run mean

reversion in operating performance, we employ the matching-firm adjusted asset-scaled net

income (performance-adjusted ROA) to remove the mean reversion effect. The matching

process is as discussed in Section 3.2.1. The change in performance-adjusted ROA of issuing

firms is measured relative to year 0 (the fiscal year of the listing). Table 12 lists the regression

results. Consistent with the results reported in Table 8 there is a significant relation between the

pre-listing discretionary accruals and the changes in industry-adjusted operating performance

for the average firm. For sub-sample of firms with both poor current and future earnings, we

also find a significantly negative association between pre-listing abnormal accruals and

post-listing operating performance changes, suggesting that reported decline in post-listing

operating performance is due, at least in part, to pre-listing earnings management. This

evidence thus provides additional support to the conjecture that the observed upward earnings

management prior to the exchange listing for this group of firms is largely associated with

managerial opportunistic intend. For the other three types of sub-samples, we again find no

consistent relation between prior discretionary accruals and the changes in industry-adjusted

operating performance, similar to the regression results in Table 8.

Thirdly, as a final robustness test of the return results in Table 8, we use the Fama-MacBeth

panel procedure to re-examine the relation between pre-listing abnormal accruals and

post-listing stock returns. Following Shivakumar (2000), we run cross-sectional regressions

explaining monthly returns from April 1986 through November 2008 with the following lagged

variables: (a) the two accrual measures described earlier, (b) the book-to-market ratio, and (c)

the natural log of market value of equity. The time series of the coefficients from these monthly

26

regressions are analyzed to estimate the relation between prior abnormal accruals and the

abnormal returns following the exchange listings. Table 13 lists the averages of the coefficients

from the monthly cross-sectional regressions conditional on subsamples with different

combinations of current and future relative performance, together with the t-statistics computed

as the mean coefficient estimate divided by its time-series standard error.


Among four subsamples, the average coefficients on three measures of discretionary

accruals are significantly negative for firms with poor performance in both current and future

periods (Cp, Fp), implying that switching firms with higher levels of pre-listing discretionary

accruals perform significantly worse after the listing. The insignificant coefficients on three

measures of discretionary accruals for other three subsamples indicate that pre-listing

discretionary accruals do not predict post-listing returns above and beyond book-to-market

ratio and size. These findings reinforce the results reported in Table 8.

5. Summary and conclusions

This study examine whether managers exploit discretionary accruals prior to exchange

listings in conjunction with exchange listing announcements as a credible means of signaling

favorable private information or as a convenient way to opportunistically mislead investors.

Overall, the evidence in this study indicates that switching firms overstate earnings before

exchange listings on average, while investors fail to unravel this earnings management at the

time of listing announcements. This evidence is consistent with prior studies that find investors

failing to completely undo the stock price effect of earnings management around corporate

events such as equity offering and stock-for-stock mergers (Teoh et al., 1998b; Louis, 2004).

However, the regression results show that the long-term underperformance of switching firms

reported in the extant literature is partly attributable to the reversal of the effects of prior

upward earnings management. These findings are consistent with the managerial opportunism

intent and suggest that earnings management before exchange listings is more likely indicative

of managerial opportunism than the signal of conveying private information by managers or

the switching firms’ response to investors’ expectations.13

13 Shivakumar (2000) formalizes and empirically tests the rational expectations argument proposed by Ericksonand Wang (1999) for earnings management and refers to this argument as the “managerial response”hypothesis.

27

Additionally, by taking into account of managerial consideration of both current and future

expected earnings, we uncover that switching firms with poor current earnings adopt more



poor future earnings. Specifically, we observe that the post-listing reversal of the effects of

pre-listing upward earnings management wherein abnormal accruals are negatively related to

post-listing long-run stock returns is driven largely by firms with poor earnings in both current

and future periods (Cp, Fp). For firms with good performance in both current and future

periods (Cg, Fg), however, we find some evidence indicating that these firms tend to

strategically report income-decreasing discretionary accruals under the concern of avoiding

competition in the industry and/or mitigating adverse political attention. Consistent with the

predictions of Fudenberg and Tirole (1995), switching firms with good current and poor future

earnings (Cg, Fp) and those with poor current and good future earnings (Cp, Fg) are found to

exhibit the propensity to smooth earnings streams across time.

The results are important because, in spite of extensive investigation of the post-listing

underperformance, its causes remain a puzzle. Our findings suggest that the long-term

underperformance of switching firms reported in the extant literature is partly attributable to

the reversal of the effects of prior earnings management. In addition, by analyzing the effects

of managerial consideration of current and expected future earnings on their reporting behavior

around exchange listings, our findings have important implications for when managers are

expected to have incentives to make a particular accounting choice (e.g., income-increasing or

income-decreasing) around corporate events. Moreover, besides the income-increasing

reporting behavior observed for the average switching firm, this study also reveals that

managers of switching firms have the countervailing incentive to deflate earnings prior to the

exchange listings for reasons of avoiding competition in the industry and/or adverse political

attention as well as income smoothing.

Shivakumar suggests that earnings management before SEOs is not intended to mislead investors, but is insteadthe markets anticipate it and adjust for the anticipated earnings management at offering announcements.

28

References

Aharoney, J., Lin, C., Loeb, M., 1993. Initial public offerings, accounting choices and earnings

management. Contemporary Accounting Research 10, 61-81.

Arya, A., Glover, J., Sunder, S., 2003. Are unmanaged earnings always better for shareholders?

Accounting Horizons 17, 111-116.

Baker, H.K., Edelman, R.B., 1991. Valuation implications of AMEX listings: a joint test of the

liquidity-signaling hypotheses. Quarterly Journal of Business and Economics 21, 87-109.

Ball, R., Shivakumar, L., 2008. Earnings quality at initial public offerings. Journal of

Accounting and Economics 45, 324–349.

Barber, R., Lyon, J., 1997. Detecting long-run abnormal stock returns: the empirical power and

specifications of test statistics. Journal of Financial Economics 43, 341–372.

Bhojraj, S., Libby, R., 2005. Capital market pressure, disclosure frequency-induced

earnings/cash flow conflict, and managerial myopia. Accounting Review 80, 1-20.

Brennan, M., Copeland, T., 1988. Stock splits, stock prices and transaction cost. Journal of

Financial Economics 22, 83-102.

Bushman, R., Piotroski, J., Smith, A., 2005. Insider trading restrictions and analysts’ incentives

to follow firms. Journal of Finance 60, 35–66.

Chan, L., Lakonishok, J., Sougiannis, T., 2001. The stock market valuation of research and

development expenditures. Journal of Finance 56, 2431-2456.

Cheng, Y., 2005. Post-listing underperformance: Is it really bad to move trading locations?

Journal of Corporate Finance 12, 97-120.

Christie, W.G., Huang, R.D., 1993. Market structure and liquidity: A transactions data study of

exchange listings. Journal of Financial Intermediation 3, 300-326.

DeAngelo, L., 1986. Accounting numbers as market valuation substitutes: a study of

management buyouts of public stockholders. The Accounting Review 61, 400-420.

Dechow, P.M., 1994. Accounting earnings and cash flows as measures of firm performance-

The role of accounting accruals. Journal of Accounting and Economics 18, 3-42.

Dechow, P.M., Sloan, R., Sweeney, A., 1995. Detecting earnings management. The

Accounting Review 70, 193-225.

DeFond, M.L., Park, C.W., 1997. Smoothing income in anticipation of future earnings. Journal

of Accounting and Economics 23, 115–139.

29

Demski, J., 1998. Performance measure manipulation. Contemporary Accounting Research 15,

261-285.

Dharan, B.G., Ikenberry, D.L., 1995. The long-run negative drift of post-listing stock returns.

Journal of Finance 50, 1547-1574.

Erickson, M., Wang, S-W., 1999. Earnings management by acquiring firms in stock for stock

mergers. Journal of Accounting and Economics 27, 149–176.

Edelman, R.B., Baker, H.K., 1994. The postlisting returns anomaly revisited. Quarterly Journal

of Business and Economics 33, 54-68.

Fama, E., French, K., 1992. The cross-section of expected stock returns. Journal of Finance 47,

427-465.

Fama, E., French, K., 2000. Forecasting profitability and earnings. Journal of Business 73,

161-176.

Fields, T., Lys, T., Vincent, L., 2001. Empirical research on accounting choice. Journal of

Accounting and Economics 31, 255-307.

Francis, J., Hanna, D., Philbrick, D., 1998. Management communications with securities

analysts. Journal of Accounting and Economics 24, 363–394.

Fudenberg, K., Tirole, J., 1995. A theory of income and dividend smoothing based on

incumbency rents. Journal of Political Economy 103, 75–93.

Gong, G., Louis, H., Sun, A.X., 2008. Earnings management and firm performance following

open-market repurchases. Journal of Finance 63, 947-986.

Grammatikos, T., Papaioannou, G.J., 1986. The information value of listing on the New York

Stock Exchange. Financial Review 21, 485-499.

Grinblatt, B., Masulis, R., Titman, S., 1984. The valuation effects of stock splits and stock

dividends. Journal of Financial Economics 13, 467-490.

Harris, M., 1998. The association between competition and managers’ business segment

reporting decisions. Journal of Accounting Research 36, 111–128.

Healy, P.M., Palepu, K.G., 1993. The effect of firms’ financial disclosure strategies on stock

prices. Accounting Horizons 7, 1–11.

Ikenberry, D., Rankine, G., Stice, E., 1996. What do stock splits really signal? Journal of

Financial and Quantitative Analysis 31, 357-375.

Jain, P. K., Kim, J., 2006. Investor recognition, liquidity, and exchange listings in the reformed

30

markets. Financial Management 35, 21–42.

Jones, J., 1991. Earnings management during import relief investigations. Journal of

Accounting Research 29, 193-228.

Kadlec, B.G.., McConnell, J.J., 1994. The effect of market segmentation and illiquidity on asset

prices: Evidence from exchange listings. Journal of Finance 49, 611-636.

Karuna, C., 2007. Industry product market competition and managerial incentives. Journal of

Accounting and Economics 43, 275–297.

Kothari, S.P., Leone, A.J., Wasley, C.E., 2005. Performance matched discretionary accrual

measures. Journal of Accounting and Economics 39, 163-197.

Kothari, S., Warner, J., 1997. Measuring long-horizon security price performance. Journal of

Financial Economics 43, 301–339.

Lakonishok, J., Lev, B., 1987. Stock splits and stock dividends, why, who, and when. Journal

of Finance 32,913-932.

Lakonishok, J., Shleifer, A., Vishny, R.W., 1994. Contrarian investment, extrapolation, and risk.


Lang, M., Lundholm, R., 1996. Corporate disclosure policy and analyst behavior. The

Accounting Review 1 71, 467–492.

Lie, E., 2001. Detecting abnormal operating performance: Revisited. Financial Management 30,

77-91.

Lie, E., 2005. Operating performance following open market share repurchase announcements.

Journal of Accounting and Economics 39, 411-436.

Louis, H., 2004. Earnings management and the market performance of acquiring firms. Journal

of Financial Economics 74, 121-148.

Louis, H., Robinson, D., 2005. Do managers credibly use accruals to signal private information?

Evidence from the pricing of discretionary accruals around stock splits. Journal of

Accounting and Economics 39, 361-380.

Louis, H., White, H., 2007. Do managers intentionally use repurchase tender offers to signal

private information? Evidence from firm financial reporting behavior. Journal of

Financial Economics 85, 205–233.

McConnell, J.J., Sanger, G.C., 1984. A trading strategy for new listing on the NYSE. Financial

Analyst Journal 40, 34-48.

31

McConnell, J.J., Sanger, G.C., 1987. The puzzle in post-listing common stock returns. Journal

of Finance 42, 119-140.

McNichols, M., Dravid, A., 1990. Stock dividends, stock splits, and signaling. Journal of

Finance 45, 857-880.

Merton, R.C., 1987. A simple model of capital market equilibrium with incomplete information.


Morsfield, S.G., Tan, C.E.L., 2006. Do venture capitalists influence the decision to manage

earnings in initial public offerings. Accounting Review 81, 1119-1150.

Papaioannou, G.P., Travlos, N.G., Viswanathan, K. G., 2003. The operating performance of

firms that switch their stock listings. Journal of Financial Research, 469-486.

Perry, S., Williams, T., 1994. Earnings management preceding management buyout offers.

Journal of Accounting and Economics 20, 293-316.

Raith, M., 2003. Competition, risk and managerial incentives. American Economic Review 93,

1425–1436.

Rangan, S., 1998. Earnings before seasoned equity offerings: are they overstated? Journal of

Financial Economics 50, 101-122.

Ritter, J., 1991. The long-run performance of initial public offerings. Journal of Finance 42,

3-27.

Ronen, J., Yaari, V., 2008. Earnings Management: emerging insights in theory, practice, and

research. Springer-Verlag.

Sanger, G.C., McConnell, J.J., 1986. Stock exchange listings, firm value, and security market

efficiency: the impact of NASDAQ. Journal of Financial and Quantitative Analysis 21,

1-25.

Schutlz, P., 2003. Pseudo market timing and the long-run underperformance of IPOs. Journal

of Finace 58, 483-517.

Shivakumar, L., 2000. Do firms mislead investors by overstating earnings before seasoned

equity offerings? Journal of Accounting and Economics 29, 339-371.

Stein, J.C., 1989. Efficient capital markets, inefficient firms: A model of myopic corporate

behavior. The Quarterly Journal of Economics 104: 655-669.

Subramanyam, K., 1996. The pricing of discretionary accruals. Journal of Accounting and

Economics 22, 249-281.

32

Teoh, S., Welch, I., Wong, T., 1998a. Earnings management and the long-run market

performance of initial public offerings. Journal of Finance 53, 1935-1974.

Teoh, S., Welch, I., Wong, T., 1998b. Earnings management and the post-issue

underperformance of seasoned equity offerings. Journal of Financial Economics 50,

63-100.

Van Horne, J.C., 1970. New listings and their price behavior. Journal of Finance 25, 783-794.

Warfield, T.D., Wild, J.J., Wild, K.L., 1995. Managerial ownership, accounting choice, and

informativeness of earnings. Journal of Accounting and Economics 20, 61-91.

Watts, R., Zimmerman, J., 1978. Towards a positive theory of the determination of accountingstandards. The Accounting Review 53, 112-134.

Watts, R., Zimmerman, J., 1986. Positive accounting theory. Prentice Hall, Englewood Cliffs,

NJ.

Webb, G.P., 1999. Evidence of managerial timing: the case of exchange listings. Journal of

Financial Research 22, 247-263.

Ying, L.K.W., Lewellen, W.G., Schlarbaum, G.G., Lease, R.C., 1977. Stock exchange listings

and securities returns. Journal of Financial and Quantitative Analysis 12, 415-432.

Yuan, C-C., Baker, H.K., Chou, L-C., Lu, B-W., 2009. Does switching from NASDAQ to the

NYSE affect investment–cash flow sensitivity? Journal of Business Research 62,

1007–1012.

33

Table 1. Summary of predictions

This table summarizes the predictions of the various hypotheses regarding managers’pre-listing financialreporting behavior and the signs of the relations between pre-listing abnormal discretionary accruals and abnormalstock returns around and following listing announcements.

HypothesisManagerial pre-listing

reporting choice

Association betweenpre-listing discretionary

accruals and listingannouncement returns

Association betweenpre-listing discretionaryaccruals and post-listing

stock returns

Managerial opportunism Income-increasingdiscretionary accruals

None Negative

Complementary signaling Income-increasingdiscretionary accruals

Positive None or positive

Table 2. Pre-listing discretionary accrual behavior partitioned by current and future relativeearnings performance

This table summarizes managerial accrual choices prior to exchange listings. The columns partition currentrelative performance and the rows partition expected relative performance.

Current relative performancePoor current premanaged

earnings (Cp)Good current premanaged

Earnings (Cg)

Poor expectedfuture earnings

(Fp)

(1)Income-increasing discretionary

accruals in year -1Managerial opportunism

(2)Income-decreasing discretionary

accruals in year -1Income smoothing

Expectedfuture

relativeperformance Good expected

future earnings(Fg)


accruals in year -1Income smoothing


accruals in year -1Complementary signaling

Table 3. Distribution of the sample during 1986-2005

This table lists the number of firms that changed their trading locations during 1986-2005 by the listing type,listing year and industry. We include ordinary common stocks only and exclude real estate investment trusts(REITs), closed-end funds, American Depositary Receipts, and financial firms. To be included, the firm must havenecessary data on stock price and financial statement items, including annual earnings, book value and shareinformation, with positive values of total assets, total liabilities,stockholders’ equity, and sales.

Panel A: Sample Distribution by Year

Year Moving fromAMEX to NYSE

Moving fromNASDAQ to

NYSE


AMEXTotal % of

sample

1986 0 2 0 2 0.371987 1 1 0 2 0.371988 0 7 2 9 1.661989 6 13 5 24 4.431990 12 17 3 32 5.901991 13 20 7 40 7.381992 11 17 3 31 5.721993 12 16 3 31 5.721994 14 17 3 34 6.271995 7 35 4 46 8.491996 13 37 5 55 10.151997 5 41 4 50 9.231998 10 40 7 57 10.521999 7 14 12 33 6.092000 1 16 6 23 4.242001 8 14 1 23 4.242002 3 23 0 26 4.802003 4 5 0 9 1.662004 6 4 0 10 1.852005 3 2 0 5 0.92

1986-2005 136 341 65 542 100.00Panel B: Sample Distribution by Industry

SIC Industry Moving fromAMEX to NYSE


NYSE


AMEXTotal % of

sample

01-09 Agriculture,Forestry, Fishing 0 3 0 3 0.55

10-14 Mining 9 29 5 43 7.9315-17 Construction 2 6 0 8 1.4820-39 Manufacturing 86 157 31 274 50.55

40-49 Transportation &Public Utilities 9 36 8 53 9.78

50-51 Wholesale Trade 8 13 1 22 4.0652-59 Retail Trade 8 29 4 41 7.5670-89 Services 14 68 16 98 18.08Total 136 341 65 542 100.00

36

Table 4. Descriptive statistics of sample firms around the listing during 1986-2005

This table shows selected characteristics of the 546 firms that switched their stock listing from 1986 to 2005. Allvariable are reported for the year preceding exchange listing. Total accrual is net income before extraordinaryitems minus operating cash flows, scaled by lagged assets. Asset is the book value of total assets. Market value isthe market capitalization. Leverage is the total liabilities divided by total assets. B/M is the ratio of book value tomarket value of common stock. Expected earning is the I/B/E/S forecast for earnings per share for the listingyear. Institutional ownership is the shares owned by institutions divided by total outstanding shares of switchingfirms. Analyst following is the number of financial analysts making the one-year-ahead forecast for the switchingfirms. Figures in square brackets are medians.

Variables All firms Moving fromAMEX to NYSE


NYSE


AMEX0.04 0.05 0.04 0.03

Total accruals[0.02] [0.03] [0.02] [0.01]

440.00 455.05 494.55 122.32Assets ($millions)

[227.92] [205.77] [300.26] [55.12]526.71 538.02 608.70 72.92

Market value ($millions)[258.32] [199.96] [329.01] [38.63]

0.28 0.26 0.27 0.35Leverage

[0.26] [0.22] [0.26] [0.32]0.53 0.55 0.48 0.76

B/M[0.44] [0.50] [0.42] [0.63]0.84 0.88 0.90 0.45

Expected earnings[0.75] [0.77] [0.80] [0.50]0.43 0.43 0.47 0.24

Institutional ownership[0.41] [0.37] [0.45] [0.21]5.32 4.54 6.25 2.11

Analyst following[4.00] [4.00] [5.00] [1.00]

37

Table 5. Operating performance and discretionary accruals of switching firms

This table presents the time-series profile of operating performance and discretionary accruals of switching firmsfrom three year before switching to three year after switching. Unadjusted net income is defined as net incomebefore extraordinary items scaled by lagged assets. The industry-adjusted net income is the unadjusted net incomeminus that of the median firm in the same industry (2-digit SIC code). Three measures of discretionary accruals(DA) are based on (1) the modified-Jones model, (2) the performance-matched modified-Jones model, and (3)Ball and Shivakumar (2008) nonlinear Jones model. The p-values for the Wilcoxon z-statistics are reported inparentheses.“*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a1% significance level.

Year -3 -2 -1 0 1 2 3Panel A: Time-Series Distribution of Operating Performance

Obs. 445 484 542 511 480 458 4181. Unadjusted Net IncomeMedian 0.063 0.067 0.075 0.074 0.061 0.054 0.049p-value (0.00)*** (0.00)*** (0.00)*** (0.00)*** (0.00)*** (0.00)*** (0.00)***2. Industry-adjusted Net IncomeMedian 0.008 0.031 0.041 0.035 0.018 0.008 -0.010p-value (0.00)*** (0.05)** (0.00)*** (0.45) (0.04)** (0.08)* (0.00)****

Panel B: Time-Series Distribution of Discretionary Accruals1. Discretionary Accruals based on Modified-Jones Model

Obs. 445 484 542 511 480 458 418Median 0.005 0.006 0.008 0.003 0.003 0.006 0.002p-value (0.11) (0.21) (0.00)*** (0.41) (0.49) (0.20) (0.16)2. Performance-matched Discretionary Accruals based on Modified-Jones Model

Obs. 364 438 459 424 386 346 301Median 0.002 0.006 0.027 0.001 0.011 -0.008 0.005p-value (0.65) (0.11) (0.00)*** (0.30) (0.14) (0.85) (0.39)3. Discretionary Accruals based on B&S’s Nonlinear Jones Model

Obs. 303 352 458 423 429 391 355Median 0.012 0.015 0.022 0.019 0.015 0.017 0.010p-value (0.04)** (0.01)*** (0.00)*** (0.00)*** (0.00)*** (0.00)*** (0.10)*

38

Table 6. Discretionary accruals of switching firms in year -1 stratified by different combinationsof current and expected future performance

This table presents pre-listing discretionary accruals of switching firms stratified by different combinations ofcurrent and expected future performance. Three measures of discretionary accruals (DA) are based on (1) themodified-Jones model, (2) the performance-matched modified-Jones model, and (3) Ball and Shivakumar (2008)nonlinear Jones model. (Cp,Fp) indicates firms with poor current and future performance. (Cp,Fg) indicates firmswith poor current performance and good future performance. (Cg,Fp) indicates firms with good currentperformance and poor future performance. (Cg,Fg) indicates firms with good current and future performance.Good and poor performance is measured relative to the sample median by industry and fiscal year. Currentpremanaged earnings are used as the current performance and are measured as current period earnings minusdiscretionary accruals. Here we treat year -1 as the current period. Median analysts’ forecastsof one-year-aheadearnings from the I/B/E/S database are used to proxy for managements’ expectations of future earningsperformance. The t-statistics are reported in parentheses. Wilcoxon z-statistic is used to test the median and thep-values are reported in brackets. “*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a 1% significance level.

(Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Panel A: Discretionary Accruals based on Modified-Jones Model

Obs. 140 131 132 139Median 0.057 0.092 -0.059 -0.039p-value [0.00]*** [0.00]*** [0.00]*** [0.00]***

Panel B: Performance-matched Discretionary Accruals based on Modified-Jones ModelObs. 122 108 119 110

Median 0.087 0.094 -0.048 -0.034p-value [0.00]*** [0.00]*** [0.00]*** [0.00]***

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones ModelObs. 118 111 111 118

Median 0.041 0.057 -0.021 -0.010p-value [0.00]*** [0.00]*** [0.00]*** [0.05]**

39

Table 7. Abnormal returns around the listing announcements

This table presents the value-weighted market index adjusted cumulative abnormal returns over the listingannouncement period (t-1 to t+1, where t=0 is the listing date). The sample is stratified by the discretionaryaccruals in year -1. Three measures of discretionary accruals (DA) are based on (1) the modified-Jones model, (2)the performance-matched modified-Jones model, and (3) Ball and Shivakumar (2008) nonlinear Jones model.High (low) DA indicates the discretionary accruals measure is higher (lower) than the median. The t-statistics arereported in parentheses. Wilcoxon z-statistic is used to test the median and p-values are reported in brackets. “*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a 1% significancelevel.

All Firms High DA Low DA Diff.Panel A: Stratified by DA based on Modified-Jones Model

Obs. 542 271 271CAR (%) 1.04 0.90 1.18 -0.28t-statistic (4.97)*** (2.96)*** (4.10)*** (-0.67)Median 0.75 0.72 0.76 -0.04p-value [0.00]*** [0.00]*** [0.00]*** [0.75]

Panel B: Stratified by DA based on Performance-matched Modified-Jones ModelObs. 459 230 229

CAR (%) 0.98 0.45 1.51 -1.06t-statistic (4.39)*** (1.53) (4.57)*** (-2.38)**Median 0.65 0.02 1.25 -1.23p-value [0.00]*** [0.24] [0.00]*** [0.01]***

Panel C: Stratified by DA based on B&S’s Nonlinear Jones ModelObs. 458 229 229

CAR (%) 1.02 1.02 1.03 -0.01t-statistic (4.54)*** (3.14)*** (3.26)*** (-0.03)Median 0.75 0.85 0.55 0.31p-value [0.00]*** [0.00]*** [0.00]*** [0.88]

40

Table 8. Regression results of short-term abnormal returns on discretionary accruals

This table presents regression results of short-term CAR on discretionary accruals. The following equation isestimated,

2005

19877

654321 /

YearOwnershipnalInstitutio

FollowingAnalystDregulatedSizeLeverageMBDACAR

The dependent variable (CAR) is value-weighted market index adjusted cumulative abnormal returns over thelisting announcement period (t-1 to t+1, where t=0 is the listing date). Three measures of discretionary accruals(DA) are based on (1) the modified-Jones model, (2) the performance-matched modified-Jones model, and (3)Ball and Shivakumar (2008) nonlinear Jones model. ((Cp,Fp) indicates firms with poor current and futureperformance. (Cp,Fg) indicates firms with poor current performance and good future performance. (Cg,Fp)indicates firms with good current performance and poor future performance. (Cg,Fg) indicates firms with goodcurrent and future performance. Good and poor performance is measured relative to the sample median byindustry and fiscal year. Current premanaged earnings are used as the current performance and are measured ascurrent period earnings minus discretionary accruals. Here we treat year -1 as the current period. Median analysts’ forecasts of one-year-ahead earnings from the I/B/E/S database are used to proxy for managements’ expectations of future earnings performance. B/M is the ratio of book value divided by market value of common stock in year-1. Leverage is the total liabilities divided by total assets in year -1. Size is the natural logarithm of market valueof common stock in year -1. Dregulated is a dummy equal to one for firms classified in the regulated industry.Analyst following is the natural logarithm of the number of analysts making the one-year-ahead forecast forswitching firms in year -1. Institutional ownership is the shares owned by institution to total outstanding shares ofswitching firms in year -1. Year is used as a switching year dummy variable. The t-statistics are in parentheses.“*” represents a 10% significance level; “**” represents a 5% significance level.

Panel A: Discretionary Accruals Based on Modified-Jones ModelAll firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)

Intercept 0.078 0.021 0.061 0.005 0.150(2.20)** (0.36) (1.11) (0.09) (2.64)***

DA 0.012 0.003 0.047 0.004 0.035(1.06) (0.13) (1.49) (0.11) (1.07)

B/M -0.002 0.003 -0.012 0.002 -0.002(-2.36)** (0.59) (-1.00) (0.13) (-1.57)

Leverage -0.004 0.003 -0.002 0.014 -0.006(-0.43) (0.14) (-0.10) (0.58) (-0.25)

Size 0.002 0.006 -0.004 0.012 -0.007(0.97) (1.29) (-0.62) (2.08)** (-1.27)

Dregulated 0.008 0.008 0.003 0.010 0.013(1.09) (0.54) (0.19) (0.67) (0.63)

Analyst following 0.000 -0.008 0.004 -0.008 0.012(-0.09) (-1.28) (0.41) (-1.17) (1.68)*

Institutional ownership -0.018 -0.016 -0.006 -0.039 -0.013(-1.58) (-0.74) (-0.20) (-1.30) (-0.52)

Year Yes Yes Yes Yes YesObs. 542 140 131 132 139

41

Table 8 (continued)Panel B: Performance-matched Discretionary Accruals Based on Modified-Jones Model

All firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Intercept 0.068 0.026 0.099 0.005 0.067

(1.92)* (0.45) (1.76)* (0.09) (1.18)DA -0.021 -0.005 0.001 -0.061 0.039

(-1.58) (-0.19) (0.02) (-1.25) (0.93)B/M 0.003 -0.004 -0.008 0.008 0.032

(0.77) (-0.59) (-0.63) (1.13) (2.20)**Leverage -0.003 -0.027 -0.032 0.017 0.021

(-0.30) (-1.18) (-1.19) (0.67) (0.71)Size 0.005 0.006 -0.007 0.009 0.008

(1.79)* (1.20) (-1.05) (1.38) (1.21)Dregulated 0.010 0.019 -0.010 0.002 0.037

(1.21) (1.17) (-0.48) (0.10) (1.37)Analyst following -0.002 -0.009 0.003 -0.004 0.004

(-0.69) (-1.35) (0.32) (-0.44) (0.54)Institutional ownership -0.024 -0.018 0.006 -0.032 -0.044

(-1.96)** (-0.70) (0.16) (-1.22) (-1.50)Year Yes Yes Yes Yes YesObs. 459 122 108 119 110

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones ModelAll firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)

Intercept 0.053 0.006 0.083 -0.033 0.036(1.64) (0.09) (1.45) (-0.69) (0.73)

DA 0.004 -0.057 0.031 0.019 -0.020(0.22) (-1.30) (1.09) (0.39) (-0.63)

B/M 0.000 0.000 -0.007 -0.009 0.011(0.06) (0.00) (-0.55) (-0.50) (0.68)

Leverage -0.003 -0.004 0.009 0.017 -0.019(-0.29) (-0.17) (0.35) (0.64) (-0.74)

Size 0.001 0.007 -0.006 0.006 -0.005(0.31) (1.30) (-0.93) (0.84) (-0.88)

Dregulated 0.006 0.012 0.000 0.008 0.016(0.78) (0.76) (0.01) (0.49) (0.63)

Analyst following 0.002 -0.009 0.014 -0.009 0.017(0.59) (-1.11) (1.47) (-1.10) (2.17)**

Institutional ownership -0.023 -0.017 -0.030 -0.025 -0.039(-1.78)* (-0.69) (-0.92) (-0.76) (-1.36)


42

Table 9. Regression results of long-run stock performance on pre-listing discretionary accruals

This table presents regression results of long-run post-listing stock performance on pre-listing discretionaryaccruals and control variables. The following equation is estimated:

2005

198798

7654321 /


FollowingAnalystDregulatedLbhrvSizeLeverageMBDALBHAR

The dependent variable (LBHAR) is the natural logarithm of one plus three-year buy-and-hold raw returns ofswitching firms minus the natural logarithm of one plus three-year buy-and-hold returns for matching firm. Threemeasures of discretionary accruals (DA) are based on (1) the modified-Jones model, (2) the performance-matchedmodified-Jones model, and (3) Ball and Shivakumar (2008) nonlinear Jones model. (Cp,Fp) indicates firms withpoor current and future performance. (Cp,Fg) indicates firms with poor current performance and good futureperformance. (Cg,Fp) indicates firms with good current performance and poor future performance. (Cg,Fg)indicates firms with good current and future performance. Good and poor performance is measured relative to thesample median by industry and fiscal year. Current premanaged earnings are used as the current performance andare measured as current period earnings minus discretionary accruals. Here we treat year -1 as the current period.Median analysts’ forecasts of one-year-ahead earnings from the I/B/E/S database are used to proxy formanagements’ expectations of future earnings performance. B/M is the ratio of book value divided by marketvalue of common stock in year -1. Leverage is the total liabilities divided by total assets in year -1. Size is thenatural logarithm of market value of common stock in year -1. Lbhrv is the value-weighted buy-and-hold marketreturns and is measured as the natural logarithm of one plus the value-weighted buy-and-hold market returns overthe three-year period following the exchange listing announcement. Dregulated is a dummy equal to one for firmsclassified in the regulated industry. Analyst following is the natural logarithm of the number of analysts makingthe forecast for switching firms in year -1. Institutional ownership is the shares owned by institution to totaloutstanding shares of switching firms in year -1. PLBHAR is the value-weighted market index adjustedbuy-and-hold abnormal returns over a one-year period preceding the listing announcement. Year is used as aswitching year dummy variable. The t-statistics are in parentheses. “*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a 1% significance level.

43

Table 9 (continued)Panel A: Discretionary Accruals based on Modified-Jones Model

All firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Intercept 1.384 0.862 -0.012 0.011 2.892

(1.71)* (0.63) (-0.01) (0.01) (2.57)**DA -0.866 -1.584 -0.846 -0.757 0.851

(-3.35)*** (-2.40)** (-1.30) (-0.98) (1.27)B/M -0.037 -0.221 -0.654 -0.305 -0.021

(-1.74)* (-1.87)* (-2.60)** (-0.75) (-0.91)Leverage 0.117 -0.535 -0.251 1.566 0.527

(0.50) (-1.07) (-0.49) (2.66)*** (1.10)Size 0.049 -0.200 -0.035 0.086 0.140

(0.89) (-1.63) (-0.27) (0.65) (1.32)Lbhrv -0.196 0.646 -0.614 -1.060 0.062

(-0.48) (0.69) (-0.81) (-1.02) (0.08)Dregulated -0.068 -0.050 0.030 -0.422 -0.310

(-0.41) (-0.15) (0.08) (-1.23) (-0.74)Analyst Following -0.099 -0.027 -0.027 -0.032 -0.269

(-1.32) (-0.17) (-0.15) (-0.20) (-1.91)*Institutional Ownership 0.564 1.323 0.773 0.215 0.169

(2.13)** (2.50)** (1.21) (0.31) (0.34)PLBHAR 0.011 -0.031 0.072 0.058 0.051

(0.10) (-0.14) (0.29) (0.21) (0.21)Year Yes Yes Yes Yes YesObs. 542 140 131 132 139

44


All firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Intercept 1.389 0.387 -0.806 -0.448 3.288

(1.73)* (0.29) (-0.67) (-0.35) (2.63)***DA -0.862 -1.410 -0.517 -1.722 1.526

(-2.86)*** (-2.29)** (-0.65) (-1.63) (1.62)B/M -0.241 -0.353 -0.432 -0.050 -0.242

(-2.66)*** (-2.10)** (-1.52) (-0.33) (-0.67)Leverage 0.035 0.213 0.290 -0.092 0.402

(0.14) (0.40) (0.49) (-0.17) (0.59)Size 0.062 -0.043 0.035 0.050 -0.013

(1.03) (-0.36) (0.24) (0.36) (-0.09)Lbhrv -0.018 1.162 -0.587 -0.596 1.769

(-0.04) (1.22) (-0.69) (-0.56) (1.55)Dregulated 0.129 -0.333 -0.304 0.216 0.266

(0.71) (-0.89) (-0.73) (0.60) (0.44)Analyst Following -0.106 -0.108 0.109 -0.089 -0.121

(-1.30) (-0.66) (0.51) (-0.50) (-0.69)Institutional Ownership 0.695 1.085 0.185 1.188 0.517

(2.49)** (1.82)* (0.24) (2.11)** (0.80)PLBHAR -0.031 -0.405 0.260 0.217 0.026

(-0.26) (-1.87)* (0.79) (0.80) (0.09)Year Yes Yes Yes Yes YesObs. 459 122 108 119 110

45

Table 9 (continued)Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones Model

All firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Intercept -0.448 0.311 1.006 0.335 -2.756

(-0.63) (0.23) (0.73) (0.20) (-2.00)**DA -0.741 -3.158 -0.252 -1.401 0.392

(-2.13)** (-2.50)** (-0.35) (-1.13) (0.51)B/M -0.252 -0.182 -0.260 -0.199 -0.476

(-2.71)*** (-1.44) (-0.83) (-0.45) (-1.58)Leverage 0.012 -0.247 -0.763 0.408 0.780

(0.05) (-0.43) (-1.38) (0.65) (1.42)Size 0.000 0.024 -0.116 -0.171 0.217

(0.00) (0.18) (-0.84) (-1.04) (1.75)*Lbhrv -0.037 0.328 0.353 -0.085 -0.390

(-0.08) (0.31) (0.43) (-0.07) (-0.40)Dregulated -0.038 -0.265 0.650 0.033 -0.796

(-0.22) (-0.79) (1.69)* (0.08) (-1.80)*Analyst Following -0.075 -0.116 0.021 0.055 -0.389

(-0.92) (-0.61) (0.11) (0.30) (-2.38)**Institutional Ownership 0.453 0.547 0.493 0.852 0.705

(1.59) (0.88) (0.74) (1.21) (1.14)PLBHAR -0.022 -0.166 0.580 0.009 -0.303

(-0.19) (-0.73) (1.98)** (0.03) (-1.15)Year Yes Yes Yes Yes YesObs. 458 118 111 111 118

46

Table 10. Average competition and political cost characteristics of switching firm

This table presents average industry competition and political cost characteristics of switching firms. Theswitching firms are divided into five groups by the level of pre-listing discretionary accruals. Three measures ofdiscretionary accruals (DA) are based on (1) the modified-Jones model, (2) the performance-matchedmodified-Jones model, and (3) Ball and Shivakumar (2008) nonlinear Jones model. Herfindahl index is measuredby summing the squared fraction of sales of the firms in the industry. DIFF is computed as industry sales tooperating costs for firms in a given industry. Market size is the natural logarithm of industry sales. Entry cost ismeasured by the weighted average gross value of the cost of property, plant and equipment for firms in thatindustry, weighted by each firm’s market share in that industry. Number of industry firms is the number of firms inthe same industry as the switching firm. Size is the market value of equity. Institutional ownership is the sharesowned by institutions to total outstanding shares of switching firms. Analyst following is the number of financialanalysts making the forecast for the switching firms. Tax rate is calculated by income taxes divided by net incomebefore income taxes. All the variables are measured at the end of the year prior to the switching. The t-statisticsare in parentheses. “*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a 1% significance level.

All firm (Cg, Fg) Others Diff.Panel A: Discretionary Accruals based on Modified-Jones Model

Herfindahl index 0.07 0.06 0.07 -0.01(-1.99)**

DIFF 1.33 1.57 1.24 0.33(1.20)

Market size 68888.80 68029.19 69192.05 -1162.85(-0.08)

Entry cost 4859.03 4172.17 5095.93 -923.76(-0.94)

Number of industry firms 310.61 379.78 286.86 92.92(3.32)***

Size 526.71 705.76 464.96 240.80(2.75)***

Institutional Ownership 0.43 0.47 0.41 0.06(2.52)**

Analyst following 5.32 6.56 4.90 1.66(3.30)***

Tax rate 0.32 0.32 0.32 0.01(-0.22)

47


Herfindahl index 0.07 0.05 0.07 -0.02(-3.16)***

DIFF 1.29 1.22 1.32 -0.10(-1.13)

Market size 64856.69 60129.77 66426.99 -6297.21(-0.44)

Entry cost 4543.16 2969.64 5039.12 -2069.48(-2.79)***

Number of industry firms 302.58 355.21 286.10 69.11(2.47)**

Size 523.01 603.39 497.67 105.72(1.31)

Institutional Ownership 0.43 0.49 0.41 0.07(3.22)***

Analyst following 5.41 6.32 5.13 1.19(2.25)**

Tax rate 0.32 0.33 0.31 0.02(0.89)

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones Model

Herfindahl index 0.06 0.06 0.07 -0.01(-2.29)**

DIFF 1.34 1.67 1.23 0.44(1.38)

Market size 59324.58 59158.18 59384.62 -226.44(-0.01)

Entry cost 4565.80 3965.85 4774.02 -808.17(-0.78)

Number of industry firms 297.29 343.60 281.21 62.39(2.46)**

Size 524.84 681.49 470.48 211.01(2.45)**

Institutional Ownership 0.43 0.49 0.41 0.08(3.33)***

Analyst following 5.36 6.40 5.00 1.40(2.64)***

Tax rate 0.33 0.33 0.33 -0.002(-0.12)

Table 11. Average Competition and political cost characteristics of switching firm with goodcurrent and future performance partitioned by the level of discretionary accruals in year -1

This table presents average industry competition and political cost characteristics of switching firms with goodcurrent and future performance. These firms are divided into four groups by the level of discretionary accruals inyear -1. Three measures of discretionary accruals (DA) are based on (1) the modified-Jones model, (2) theperformance-matched modified-Jones model, and (3) Ball and Shivakumar (2008) nonlinear Jones model.Herfindahl index is measured by summing the squared fraction of sales of the firms in the industry. DIFF iscomputed as industry sales to operating costs for firms in a given industry. Market size is the natural logarithm ofindustry sales. Entry cost is measured by the weighted average gross value of the cost of property, plant andequipment for firms in that industry, weighted by each firm’s market share in that industry. Number of industryfirms is the natural logarithm of the number of firms in the same industry as the switching firm. Size is the marketvalue of equity. Institutional ownership is the shares owned by institutions to total outstanding shares of switchingfirms. Analyst following is the natural logarithm of the number of financial analysts making the forecast for theswitching firms. Tax rate is calculated by income taxes divided by net income before income taxes. All thevariables are measured at the end of the year prior to the switching. The t-statistics are in parentheses. “*” represents a 10% significance level; “**” represents a 5% significance level; “***”represents a 1% significance level.

Quartile 1 2 3 4 DiffDA Conservative Aggressive 1-4

Panel A: Discretionary Accruals based on Modified-Jones Model

Herfindahl index 0.08 0.05 0.06 0.06 0.02(1.18)

DIFF 2.06 1.12 1.21 1.93 0.13(0.12)

Market size 93375.85 24534.94 50262.66 105972.66 -12596.81(-0.26)

Entry cost 5768.24 4701.48 2368.10 3896.47 1871.77(0.71)

Number of industry firms 399.35 345.41 330.37 443.57 -44.22(-0.59)

Size 683.82 704.29 516.15 918.14 -234.32(-0.85)

Institutional Ownership 0.48 0.44 0.49 0.46 0.02(0.34)

Analyst following 6.71 6.69 5.34 7.51 -0.81(-0.52)

Tax rate 0.37 0.37 0.29 0.28 0.09(1.28)

49



DIFF 1.31 1.21 1.14 1.22 0.10(0.78)

Market size 50487.46 67695.54 43519.85 77468.76 -26981.30(-0.70)

Entry cost 2779.01 4179.16 2556.66 2342.17 436.84(0.45)


Size 554.36 675.17 509.98 668.97 -114.61(-0.60)

Institutional Ownership 0.55 0.46 0.50 0.44 0.11(2.25)**

Analyst following 6.78 5.89 5.85 6.75 0.03(0.02)

Tax rate 0.29 0.37 0.34 0.32 -0.03(-0.68)

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones Model


DIFF 2.14 1.35 1.89 1.27 0.87(0.91)

Market size 86877.69 85343.29 23316.46 40697.65 46180.04(1.50)

Entry cost 4244.32 6706.29 3021.37 1869.19 2375.13(2.28)**


Size 582.17 506.74 870.32 769.70 -187.53(-0.95)

Institutional Ownership 0.48 0.45 0.54 0.48 0.002(0.05)

Analyst following 4.34 5.10 8.17 7.97 -3.62(-2.91)***

Tax rate 0.32 0.32 0.33 0.34 -0.01(-0.19)

50

Table 12. Regression results of operating performance changes on discretionary accruals

This table presents OLS regression model of operating performance on discretionary accruals. The followingequation is estimated:

2005

198787

654321 /


FollowingAnalystIndDregulatedSizeLeverageMBDAACROA

The dependent variable (ACROA) is the summary of matching-firm adjusted change in return on assets duringyear 0 to year 3. Three measures of discretionary accruals (DA) are based on (1) the modified-Jones model, (2) theperformance-matched modified-Jones model, and (3) Ball and Shivakumar (2008) nonlinear Jones model. (Cp,Fp)indicates firms with poor current and future performance. (Cp,Fg) indicates firms with poor current performanceand good future performance. (Cg,Fp) indicates firms with good current performance and poor future performance.(Cg,Fg) indicates firms with good current and future performance. Good and poor performance is measuredrelative to the sample median by industry and fiscal year. Current premanaged earnings are used as the currentperformance and are measured as current period earnings minus discretionary accruals. Here we treat year -1 asthe current period. Median analysts’ forecasts of one-year-ahead earnings from the I/B/E/S database are used toproxy for managements’ expectations of future earnings performance. B/M is the ratio of book value divided bymarket value of common stock in year -1. Leverage is the total liabilities divided by total assets in year -1. Size isthe natural logarithm of market value of common stock in year -1. Dregulated is a dummy equal to one for firmsclassified in the regulated industry. Analyst following is the natural logarithm of the number of analysts makingthe one-year-ahead forecast for switching firms in year -1. Institutional ownership is the shares owned byinstitution to total outstanding shares of switching firms in year -1. PLBHAR is the value-weighted market indexadjusted buy-and-hold abnormal returns over a one-year period preceding the listing announcement. Year is usedas a switching year dummy variable. The t-statistics are in parentheses. “*” represents a 10% significance level;“**” represents a 5% significance level; “***”represents a 1% significance level.

Panel A: Discretionary Accruals based on Modified-Jones ModelAll firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)

Intercept -0.211 -0.489 0.027 -0.338 -0.472(-0.72) (-1.24) (0.05) (-0.93) (-1.00)

DA -0.173 -0.488 -0.038 -0.136 -0.242(-1.83)* (-2.56)** (-0.13) (-0.62) (-0.86)

B/M -0.015 -0.067 -0.376 -0.038 0.000(-1.99)** (-1.95)* (-3.35)*** (-0.34) (-0.03)

Leverage -0.213 -0.391 0.030 -0.409 -0.201(-2.54)** (-2.72)*** (0.14) (-2.46)** (-1.01)

Size 0.027 -0.021 -0.033 0.072 0.077(1.33) (-0.61) (-0.59) (1.94)* (1.72)*

Dregulated 0.014 -0.051 -0.048 0.082 0.057(0.24) (-0.52) (-0.29) (0.85) (0.32)

Analyst Following 0.011 0.057 -0.034 -0.024 -0.013(0.40) (1.26) (-0.43) (-0.51) (-0.22)

Institutional Ownership 0.134 -0.154 0.514 0.036 0.415(1.39) (-1.01) (1.82)* (0.19) (1.99)**

PLBHAR -0.034 -0.030 -0.174 -0.092 0.174(-0.86) (-0.48) (-1.57) (-1.23) (1.72)*


51


All firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Intercept -0.150 -0.873 -0.096 -0.093 -0.533

(-0.51) (-2.32)** (-0.18) (-0.23) (-1.08)DA -0.197 -0.432 -0.018 -0.420 -0.541

(-1.77)* (-2.51)** (-0.05) (-1.21) (-1.46)B/M -0.082 -0.007 -0.379 -0.103 -0.011

(-2.45)** (-0.16) (-2.98)*** (-2.08)** (-0.08)Leverage -0.151 -0.313 -0.027 -0.364 0.055

(-1.64) (-2.09)** (-0.10) (-2.01)** (0.21)Size 0.023 0.043 -0.073 0.032 0.075

(1.02) (1.27) (-1.16) (0.71) (1.37)Dregulated 0.036 -0.075 0.157 0.050 0.343

(0.54) (-0.71) (0.85) (0.42) (1.42)Analyst Following 0.010 0.026 0.049 0.033 -0.034

(0.33) (0.56) (0.52) (0.56) (-0.49)Institutional Ownership 0.091 -0.171 0.821 -0.107 0.402

(0.89) (-1.03) (2.53)** (-0.59) (1.57)PLBHAR -0.040 -0.070 -0.010 -0.072 0.043

(-0.91) (-1.16) (-0.07) (-0.81) (0.38)Year Yes Yes Yes Yes YesObs. 459 122 108 119 110

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones ModelAll firms (Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)

Intercept -0.528 0.325 -0.197 -1.723 -0.653(-2.00)** (0.82) (-0.33) (-3.89)*** (-1.19)

DA -0.253 -1.110 -0.181 -0.336 -0.451(-1.97)* (-2.87)*** (-0.58) (-1.01) (-1.45)

B/M -0.062 -0.065 -0.265 0.065 -0.093(-1.79)* (-1.68)* (-1.95)* (0.55) (-0.77)

Leverage -0.158 -0.470 0.210 -0.228 -0.080(-1.72)* (-2.66)*** (0.87) (-1.36) (-0.37)

Size 0.025 -0.019 0.005 0.161 0.026(1.10) (-0.48) (0.08) (3.65)*** (0.52)

Dregulated 0.062 0.135 -0.042 -0.078 0.013(0.96) (1.31) (-0.25) (-0.66) (0.07)

Analyst Following 0.015 0.021 -0.039 -0.029 0.069(0.49) (0.36) (-0.48) (-0.59) (1.04)

Institutional Ownership 0.141 -0.008 0.373 -0.110 0.318(1.34) (-0.04) (1.32) (-0.58) (1.27)

PLBHAR -0.061 0.012 -0.119 -0.069 -0.013(-1.40) (0.17) (-0.93) (-0.83) (-0.12)


52

Table 13. Time-series averages of monthly cross-sectional regressions of returns on discretionaryaccruals and firm characteristics in year -1

This table presents time-series averages of monthly cross-sectional regressions of returns on lagged discretionaryaccruals and firm characteristics, including the book-to-market ratio (B/M) and the natural log of market value ofequity (Size). For each month from April 1986 to November 2008, the following cross-sectional regression is runacross all firms that have switched stock exchanges in the previous 36 months:

SizeMBDAR 321 /

The dependent variable (R) is the monthly stock return of the switching firm. Three measures of discretionaryaccruals (DA) are based on (1) the modified-Jones model, (2) the performance-matched modified-Jones model,and (3) Ball and Shivakumar (2008) nonlinear Jones model. (Cp,Fp) indicates the firms with current premanagedearnings below median earnings and expected earnings below the sample median. (Cp,Fp) indicates firms withpoor current and future performance. (Cp,Fg) indicates firms with poor current performance and good futureperformance. (Cg,Fp) indicates firms with good current performance and poor future performance. (Cg,Fg)indicates firms with good current and future performance. Good and poor performance is measured relative to thesample median by industry and fiscal year. Current premanaged earnings are used as the current performance andare measured as current period earnings minus discretionary accruals. Here we treat year -1 as the current period.Median analysts’ forecasts of one-year-ahead earnings from the I/B/E/S database are used to proxy formanagements’ expectations of future earnings performance. The t-statistic is the mean coefficient estimate dividedby the time-series standard error of the coefficient estimate and is reported in parentheses. “*” represents a 10%significance level; “**” represents a 5% significance level; “***”represents a 1% significance level.

(Cp,Fp) (Cp,Fg) (Cg,Fp) (Cg,Fg)Panel A: Discretionary Accruals based on Modified-Jones Model

Mean -0.075 -0.043 -0.014 -0.009t-statistic (-3.08)*** (-1.62) (-0.92) (-0.35)Panel B: Performance-matched Discretionary Accruals based on Modified-Jones Model

Mean -0.068 -0.048 -0.045 -0.006t-statistic (-2.52)** (-1.39) (-0.26) (-0.32)

Panel C: Discretionary Accruals based on B&S’s Nonlinear Jones ModelMean -0.095 -0.031 -0.030 -0.001

t-statistic (-3.38)*** (-1.44) (-0.99) (-0.04)

do exchange listings mislead investors by bloated

Documents