Standard Deviation Divided by Average Net Income - DOC by hji80317

VIEWS: 114 PAGES: 48

More Info
									         Earnings Smoothing, Governance and Liquidity: International Evidence




                                       Ryan LaFond
                            Massachusetts Institute of Technology

                                         Mark Lang
                                 University of North Carolina

                                      Hollis A. Skaife
                                   University of Wisconsin


                                        January 2008



                      CAUTION: REVISION IN PROGRESS!!!




We thank the seminar participants at University of British Columbia, Carnegie Mellon
University, University of Chicago, University of Connecticut, University of North Carolina
Global Issues in Accounting Conference, Ohio State University, University of Pennsylvania, and
University of Texas Issues in Financial Reporting Conference for their comments.
         Earnings Smoothing, Governance and Liquidity: International Evidence


                                         Abstract
We examine the relation between earnings smoothing, governance and liquidity for a
sample of non-U.S. firms. We divide smoothing into innate and excess components, and
find that excess smoothing is increasing in incentives to smooth (greater tax-book
conformity, concentrated ownership, related party transactions and weak overall
governance) and decreasing in oversight (investor protection, analyst following, ADR
listing, Big-5 auditor and international accounting standards). Given the potential for
smoothing to affect transparency, we examine the relation between smoothing and
investors’ willingness to transact in the stock as reflected in liquidity. After controlling
for other liquidity determinants, we find that firms with greater levels of excess
smoothing experience lower liquidity as evidenced by higher bid-ask spreads, greater
frequency of zero returns days and lower trading volume. Further, the effect of excess
smoothing on liquidity and transactions costs is greatest in contexts where incentives are
strongest and oversight is weakest. In contrast, results for innate smoothing suggest that
innate smoothing is positively correlated with liquidity. Taken together, our results
suggest that smoothing is affected by firms’ governance environments and that excess
smoothing reduces investors’ willingness to transact in the stock, particularly when firms’
incentives to manage earnings are strong and oversight is weak.
       Earnings Smoothing, Governance and Liquidity: International Evidence



I. Introduction

In this paper, we investigate the relation between the earnings smoothing characteristics

of accruals, transparency and liquidity in the firm’s stock. Ideally, accruals increase the

transparency of reported accounting data relative to cash flows by better reflecting the

underlying economics of the firm. However, because accruals have a significant

discretionary component, their opportunistic application may decrease transparency. One

potential economic consequence of transparency is that it will affect investors’

willingness to transact in the firm’s shares. Prior research suggests that reduced

transparency will tend to result in lower liquidity, raising transactions costs and

increasing the firm’s cost of capital (Bekaert et al. 2006).

        We estimate smoothing based on two measures from the prior literature: the

variability of net income relative to cash flows (Leuz et al. 2003, Francis et al. 2004), and

the correlation between cash flows and accruals (Lang et al., 2006, Barth et al, 2006).

We begin by dividing smoothing into innate and excess components following the

approach in Francis et al. 2004 because much of the smoothness in earnings is a function

of firms’ inherent operating characteristics and the natural role of accruals.1 For

example, the smoothing properties of accruals relative to cash flows naturally vary with

differences in firm characteristics such as the firm’s industry, operating cycle, growth,

size and the inherent variability of the operating environment. Innate accruals have the


1
 We use the terms “innate smoothing” and “excess smoothing” for convenience. We estimate innate
smoothing based on predicted values from a regression of our smoothing measures on firms’ operating
characteristics and industries in the spirit of the Jones (1991) model of discretionary accruals. Excess
smoothing is smoothing beyond that which would be predicted based on the firms’ innate characteristics.
potential to be associated with increased transparency to the extent that they reflect the

underlying operating characteristics of the firm. Accruals that excessively smooth

earnings, on the other hand, are more likely to reflect managerial discretion, potentially

decreasing transparency.2

        To assess whether excess earnings smoothing reflects managerial discretion, we

analyze its relation with various incentive and oversight measures. We consider three

types of measures, (1) internal, firm-specific incentives, as reflected in concentrated

ownership and the alignment of tax and financial reporting, (2) external oversight, such

as investor protection, cross listing on US markets, Big-5 auditor, and nonlocal

accounting standards, and (3) market oversight, as reflected in analyst following. After

controlling for firm operating characteristics, we find, consistent with predictions, that

excess earnings smoothing tends to be more pronounced when managers have greater

incentives to smooth (concentrated managerial ownership and greater tax-book

conformity) and face fewer impediments (weaker investor protection, less analyst

coverage and for firms not cross listed on US markets, with a non-Big-5 auditor and

following local GAAP).

        We also investigate whether a broad governance measure based on ratings

provided by Governance Metrics International (GMI), which reflects not only the

assessment of a firm’s ownership structure, but also captures a firm’s governance related

to board structure, financial information quality, and shareholder rights, is related to

excess smoothing. In addition, we explore the relation between smoothing and the GMI

2
  It is possible that managers sometimes use their discretion to smooth earnings to increase the
informativeness of earnings. However, our evidence suggests that, on average, excess smoothing behaves
as though it decreases transparency. In particular, excess smoothing is more pronounced when managers
have greater incentives to create opacity and oversight is weaker. More directly, transactions costs are
higher and liquidity is lower when there is excess smoothing.


                                                                                                           4
ranking for related party transactions, because related party transactions provide

particular opportunities for expropriation and, therefore, particular incentives for opacity.

While the sample size drops substantially using the GMI data as we are limited to the

largest firms where governance is likely to be less of an issue, our results are consistent in

suggesting that excess smoothing is lower for firms with strong governance, especially

those with fewer related party transactions.

         Given that excess smoothing appears to be predictably correlated with incentives

and impediments for managers to smooth, we examine whether excess smoothing has

effects on trading in a firm’s shares. In particular, if excess smoothing creates opacity, it

should affect investors’ willingness to trade. As argued in Lesmond (2005), investors

will be hesitant to trade if there are concerns over the adequacy of information available

to them and bid-ask spreads will increase, increasing transactions costs.3                 Bekaert et al.

(2006) develops a model to illustrate that liquidity can affect expected returns even under

full market integration.

         Following research like Lesmond (1999) and Bekaert et al. (2006), we use two

measures of liquidity: bid-ask spreads and the proportion of zero-return days. Bid-ask

spreads reflect transactions costs, which affect investors’ willingness to trade in a firms

shares, while zero-return days reflect the frequency with which investors transact. We

examine the relation between excess smoothing, innate smoothing, and these measures of




3
  The disclosure literature contains numerous studies suggesting that transparency can affect investors’
willingness to transact in capital markets. See, for example, Verrecchia (2001) for an overview of the
literature.


                                                                                                            5
liquidity because both measures have advantages and disadvantages as an empirical

proxy for liquidity in international markets.4

        Our results provide evidence of a relation between innate smoothing, excess

smoothing and liquidity. In particular, our evidence suggests that firms with greater

evidence of excess smoothing experience significantly lower liquidity as reflected in

more frequent zero-return days and higher bid-ask spreads, controlling for other factors.

In contrast, our results suggest that, if anything, greater innate smoothing is associated

with lower information asymmetry and lower transactions costs, as reflected in lower bid-

ask spreads and fewer zero return days. This result is striking because it suggests that

innate aspects of smoothing behave differently than the portion that is more likely to be

influenced by managerial discretion. In other words, firms for which earnings are

naturally smooth because of inherent firm characteristics (e.g., operating cycle, growth,

size, profitability and sales variability) are associated with greater transparency and

liquidity. Only excess, discretionary smoothing appears to reduce transparency.

        Finally, we examine the interaction between excess smoothing and the

governance environment of the firm. We expect that smoothing has the greatest potential

to reduce liquidity and increase transactions costs in situations in which incentives are

strong and oversight is particularly weak. Our results suggest that, while excess

smoothing reduces liquidity and increases transactions costs for the sample in general, the

effects are most pronounced in environments where managers have greater incentives to

smooth earnings and face less oversight (i.e., for firms that are not cross listed on the US

4
 As discussed later, we also conduct the analysis using volume as a measure, with similar results.
Research such as Lesmond (2005) and Bekaert, Harvey and Lundblad (2006) suggests that, in international
settings, volume tends to be a relatively weak proxy because it is not be computed consistently across
exchanges, and does not behave like a priced liquidity factor or correlate highly with other liquidity
measures.


                                                                                                      6
markets, are not followed by analysts, do not file under IFRS, are not audited by a big

five auditor and have highly concentrated ownership).

         A potential question is why managers would choose to smooth if it could have

negative effects on investors’ willingness to hold shares. In some cases, it may result

from managers creating opacity for personal gain. However, it is important to note that

smoothing, even if it reduces investors’ willingness to trade, may be optimal for the firm.

For example, in many countries stakeholders other than shareholders create incentives to

report smoother earnings. Taxing authorities, for instance, may create incentives to

smooth earnings because higher profits typically attract higher tax rates and losses may

not provide full tax benefits. Similarly, firms may smooth earnings to reduce perceived

risk and attract lower interest rates on debt, reduce pressure from labor or limit political

costs. Our results suggest that a firm’s tendency to undertake those types of activities is

particularly pronounced when financial reporting oversight is weak, few analysts follow

the firm and ownership is more concentrated. While smoothing may be optimal in those

situations, our results suggest a tradeoff in that excessive smoothing can reduce

transparency, resulting in lower liquidity.5

         Our results make several contributions to the literature.6 First, and most

importantly, we focus on an economic consequence of earnings smoothing. As noted

above, managers face tradeoffs, especially in international contexts, in applying


5
  Desai and Dharmapala (2006) make a similar point in reference to tax avoidance and earnings
management. They observe that earnings management to reduce taxes can create opacity that also reduces
equity holders’ ability to assess firm performance.
6
  Our measures of earnings smoothing can also be interpreted in the context of timely loss (and gain)
recognition. As discussed in Ball and Shivakumar (2005, 2006), timely loss recognition will tend to result
in a less negative correlation between accruals and cash flows because periods of poor cash flows also
typically indicate likely decreases in the present value of future cash flows. As a consequence, timely loss
recognition will be reflected in negative accruals in periods of poor cash flows, attenuating the natural
negative relation between accruals and cash flows and creating greater volatility in the earnings stream.


                                                                                                           7
discretion. The consequences of managerial discretion on transparency and liquidity are

likely to be particularly strong in international settings where incentives to smooth

earnings are stronger, financial reporting oversight is weaker and liquidity issues are

more pronounced.

        There is little existing evidence on the relation between transparency, transactions

costs, liquidity and trading activity in international settings. Bhattacharya et al. (2003)

investigates, at the country level, the association between country-wide aggressive loss

recognition, loss avoidance and smoothing, and cost of equity capital and turnover. They

provide mixed evidence on the relation between earnings attributes, turnover and cost of

capital depending on the measure of cost of capital and the earnings attribute employed.

Our analysis differs from theirs in several ways. First, because our analysis focuses on

across-firm comparisons, we abstract from cross-country differences that may affect

equity markets and are able to control for a variety of other firm-specific factors that

likely affect earnings attributes and equity markets. Our liquidity analysis includes

country fixed effects, so cross country differences are naturally controlled. Also, our

results suggest a differential effect for the components of smoothing. We find that only

excess smoothing appears associated with reduced transparency, while inherent

smoothing appears to increase transparency. Finally, our results illustrate that the effects

of smoothing on liquidity and transactions costs tend to be particularly pronounced in

environments in which managers have particularly strong incentives to create opacity and

there is relatively little oversight.

        Further, our results are consistent for both measures of trading costs (bid-ask

spreads) and measures of liquidity (zero return days). Measures like bid-ask spreads and




                                                                                              8
zero returns days focus more explicitly on the potential effects of information assymetry

on transactions costs. Results in Lesmond (2005) suggest that the correlation between

transactions cost measures such as bid-ask spreads and zero returns days is relatively low.

Further, Bekaert, Harvey and Lundblad (2006) find that liquidity measures such as zero

returns days (and, to a lesser extent, bid-ask spreads) behave like priced risk factors in

international contexts.

       While we do not attempt to directly measure cost of capital, our results suggest

that a mechanism by which earnings smoothing might affect cost of capital is through its

effect on liquidity. While liquidity is difficult to measure, especially in international

contexts, our results are consistent across our measures, suggesting a robust relation

between excess smoothing and liquidity. Our results indicate that excess smoothing is

associated with higher transactions costs and less trading. While it is always difficult to

draw strong inferences on causality, our results suggest that one potential consequence of

smoothing is a reduction in transparency as reflected in increased transactions costs.

Because reduced transparency is likely to affect investors’ willingness to hold a stock and

the required expected return on the stock, our results provide evidence on a potential cost

of earnings smoothing.

       Second, we differentiate between innate and excess earnings smoothing. By their

very nature, accruals affect the variability of earnings relative to cash flows, and that

relation can affect the information environment of the firm. Therefore, it is potentially

important to separate out the smoothing effects of accruals that occur naturally in the

firm’s operating environment from those that reflect managerial discretion. The fact that

our measures of excess smoothing are correlated with variables that likely reflect




                                                                                              9
incentives for, and limits on, earnings management, and behave differently in the

liquidity tests than do our innate smoothing measures provides greater confidence that

our results do not reflect omitted correlated variables.

         Further, our results potentially help bridge the gap between the notion that

smoothing can convey information about the underlying economics of the firm and the

notion that smoothing can create opacity. Perspectives on smoothing differ in the

literature, with research such as Dichev and Tang (2005) suggesting that smoothing is a

function of matching and can enhance transparency while research such as Leuz, Nanda

and Wysocki (2003) argues that opportunistic smoothing reduces transparency. Our

results suggest that both factors may be at work in practice, with innate smoothing

increasing transparency and excess smoothing increasing opacity. Our results, which are

based on smoothing by non-U.S. firms, suggest that the former is representative of firms

with strong governance and the latter effect may be particularly pronounced in settings

where incentives to smooth earnings tend to be relatively strong and oversight relatively

weak.

         Finally, our earnings smoothing analysis is at the individual firm level. Research

such as Leuz et al. (2003) provides evidence of effects of earnings smoothing at the

country level with a focus on across country variation driven by things like differences in

the value of control rights and investor protection environment.7 While there are

advantages in that approach, we posit and provide evidence that there is also substantial

within-country variation in incentives to smooth and the consequences of innate and

excess smoothing for liquidity. Our analysis complements prior research by

7
  Burgstahler, Hail and Leuz (2006) document that earnings management in general is greater for private
than for public firms and that strong legal systems result in less earnings management for both public and
private firms.


                                                                                                         10
incorporating a menu of firm-specific determinants of smoothing and provides consistent

evidence of a link between governance-related factors and smoothing.

         In the next section, we discuss the methodology and data. Section III presents the

results, followed by conclusions in Section IV.



II. Methodology

We quantify earnings smoothing using two measures common in the literature. The first

earnings smoothing measure (SMTH1) captures the volatility of earnings relative to the

volatility of cash flows (Leuz et al. 2003, Francis, LaFond, Olsson and Schipper 2004).

Specifically, SMTH1 is the standard deviation of net income before extraordinary items

divided by the standard deviation of cash flow from operations, where net income before

extraordinary items and cash flow from operations are scaled by average total assets. We

calculate the standard deviations using rolling time intervals requiring a minimum of

three and a maximum of five years of data. Cash flow from operations is equal to net

income before extraordinary items minus accruals, where accruals are defined as the

change in current assets minus the change in current liabilities minus the change in cash

plus the change in current debt in current liabilities minus depreciation and amortization

expense. SMTH1 is multiplied by negative one so that larger values, i.e., values closer to

zero, represent more smooth earnings.

         The second earnings smoothing measure (SMTH2) is equal to the correlation

between the cash flow from operations scaled by total assets and total accruals scaled by

total assets (Lang et al., 2006, Barth et al., 2006).8 SMTH2 is multiplied by negative one


8
  Leuz et al. (2003) and Bhattacharya et al. (2003) calculate the correlation-based measure using the change
in cash flows from operations and the change in total accruals, whereas our correlation measure is based on


                                                                                                         11
so that firms with larger SMTH2 values (i.e., values closer to one) represent firms with

more smoothed earnings.

III.I Sample

        Table 1 presents the number of firm-year observations and descriptive statistics

on SMTH1 and SMTH2 for the 21 sample countries: Australia, Austria, Belgium,

Canada, Denmark, Finland, France, Germany, Greece, Hong Kong, Ireland, Italy, Japan,

the Netherlands, New Zealand, Norway, Singapore, Spain, Sweden, Switzerland, and the

U.K. We select these 21 countries because they have relatively well-developed capital

markets and managers face diverse incentives to smooth earnings because of the

differences in governance attributes across firms domiciled in these countries.

Accounting and market data are collected from Datastream Advanced (a collaboration of

market statistics from Datastream and accounting data from WorldScope) over the 1994-

2005 time period. We require firm-year observations to have the necessary income

statement and balance sheet data to calculate cash flows, accruals, and operating

characteristic variables.

        Table 1 highlights that, on average, firms domiciled in Greece, Austria, and Italy

report the most smoothed earnings whereas Norwegian, Swedish, and Canadian firms

report the least smoothed earnings. The descriptive statistics on SMTH1 and SMTH2

and rank ordering of countries are generally consistent with those reported in Leuz et al.

(2003) and Bhattacharya et al. (2003), respectively.

III.2 Determinants of Smoothing

III.2.a Innate Controls


the level. We draw identical inferences when defining SMTH2 based on changes; however, the sample
sizes are smaller due to the additional data requirements of the change measures.


                                                                                                    12
        We base our innate measures of smoothing on variables identified in prior

research such as Dechow and Dichev (2002) and, especially, Francis et al (2004). As in

those papers, we are interested in the types of variables that are likely to affect the

inherent smoothing properties of accruals. Our goal is to develop a model for the

“expected” level of accruals smoothing for a firm, recognizing that the characteristics of

accruals levels are naturally a function of factors like the firm’s size, growth, operating

characteristics and industry. By the nature of the accounting system, much of the

smoothing effect of accruals reflects the underlying nature of a firm’s operating

environment and provides information on the volatility of the underlying fundamentals.

As a result, firms with operating characteristics that naturally lead to earnings-smoothing

accruals may have fewer informational issues.9

        We include LNTOTASS, log of total assets, as a measure of firm size, to reflect

the scale and likely diversification of the firm. OPCYLE, measured as log of days of

accounts receivable plus inventories, captures the length of the firm’s operating cycle and

OPLEV, measured as net property, plant and equipment divided by total assets captures

capital intensity. STD_SALES is the standard deviation of sales and measures the

volatility of a firm’s underlying operating environment. BM is the ratio of book value to

market value of equity and is intended to reflect the extent of intangible assets and

expected earnings growth. %LOSS is the proportion of years that a firm experiences

losses over the last five years since loss firms likely have different accruals properties.

LEV is total debt divided by total assets, because financing is likely correlated with




9
 This is the notion inherent in papers like Dechow and Dichev (2002) and Francis et al (2004) in which
earnings-smoothing accruals can increase the informativeness of earnings and reduce cost of capital.


                                                                                                         13
earnings attributes.10 AVECFO is average cash from operations divided by total assets

over the last five years and reflects the notion that a firm’s general level of profitability

likely affects its accruals attributes. Finally, we include indicator variables for a firm’s

industry since the properties of accruals are likely to depend on a firm’s industry.



III.2.a Discretionary Determinants

        We base our measure of excess smoothing on the residuals from the regression of

our smoothing measures on the innate controls. To the extent that accruals characteristics

also reflect managerial discretion, there should be predictable correlations between

excess smoothing, managerial incentives to smooth earnings and institutional constraints

on those incentives. Ball et al. (2000) suggests that country-specific institutional factors

contribute to a manager’s set of reporting incentives. The first governance variable we

consider is the antidirector rights index of LaPorta et al. (1998) label RIGHTS. Prior

research documents that firms domiciled in countries with weak investor protection

report smoother earnings (Leuz et al. 2003), which is typically interpreted as suggesting

that, in countries with weak investor protection, managers face greater incentives and

have the ability to smooth earnings to conceal opportunistic behavior. We add RIGHTS

to our smoothing model as a proxy for greater constraints on managers’ ability to smooth

earnings when there is strong investor protection.11




10
   Leverage could also affect incentives to smooth earnings. Results are not sensitive to exclusion of LEV
or to including it as determinants of discretionary smoothing in the liquidity tests.
11
   As a sensitivity test, we replace RIGHTS with LAW; a categorical variable coded one for firms
domiciled in common law countries and zero otherwise (La Porta et al. 1997). We do not include both
measures in our model because the Pearson (Spearman) correlation between LAW and RIGHTS is 0.64
(0.76). When we substitute LAW for RIGHTS the results of the analysis are consistent with those reported
in the tables.


                                                                                                       14
         Another institutional factor posited to contribute to managers' reporting incentives

is the alignment between tax and financial reporting (Alford et al. 1993; Ali and Hwang

2000; Kasanen et al. 1996). In some countries, tax-reporting rules permit managers to

use similar accounting methods for tax reporting and financial reporting, i.e., there is a

high degree of tax-book conformity.12 In general, it is optimal for managers to smooth

earnings for tax reporting in order to minimize the likelihood of large tax payments or to

avoid tax losses that may provide reduced benefits to the firm. We posit that when there

is a high alignment between tax reporting and financial reporting, the incentives

managers face to smooth earnings for taxes will carry over to smoother accounting

earnings. Thus, the second governance attribute studied is TXBKCONFORM, which is

coded one for firms domiciled in countries that have high tax-book conformity, and zero

otherwise. Our measure of TXBKCONFORM is based on Ashbaugh and LaFond (2003)

and details of its construction by sample country are provided in the Appendix.

         As noted above, we posit that there is likely to be substantial within-country

variation in managers’ incentives to smooth earnings, and there are factors that can

mitigate or enhance managers’ ability to act on those incentives. First, the effective

regulatory environment of firms may vary depending on whether they list on US

exchanges. The scrutiny of regulators over a firm’s financial reporting affects the quality

of firm’s financial information (Securities and Exchange Commission [SEC] 2000). The

US regulatory environment is considered one of the most demanding in the world and,


12
  In many countries, the explicit link between tax and financial accounting in the consolidated accounts has
been loosened over time. For example, consolidated reports may be prepared under International Financial
Reporting Standards (IFRS) for financial reporting purposes, while parent-entity reports are the basis for
tax reporting. However, managers may still align the choices made in the financial statements with those in
the tax books to reduce scrutiny and ease record keeping. To the extent the misclassification of tax-book
conformity adds noise, the noise would bias against finding any significant relation between
TXBKCONFORM and smoothing.


                                                                                                         15
therefore, managers of U.S. foreign issuers are likely to have less of a tendency to

manage earnings because they fall under the jurisdiction of the SEC.13 Thus, we posit

that the incentives to smooth earnings are less for U.S. foreign issuers relative to firms

not listed with the SEC and use ADR, coded one for sample firms that are U.S. foreign

issuers and zero otherwise, to capture the disincentives to smooth earnings.

         Oversight by informational intermediaries may affect firms’ incentives to smooth

earnings. We use analyst following (ANALYST) as a proxy for the demand for

transparent financial information by capital market participants.14 Analysts depend on

relevant financial reports, as well as other pieces of information, to develop their

forecasts of firms’ future earnings and stock recommendations. In this context, analysts

may serve as a proxy for increased capital market monitoring of managers’ financial

reporting thereby potentially mitigating earnings management. Thus, we expect a

negative relation between ANALYST and the two smoothing measures.

         Further, we expect the level of auditing to affect managers’ ability to smooth

earnings. Because larger auditing firms are likely to have greater resources and greater

legal and reputational exposure, we expect attestation by a Big-5 auditing firm to be

associated with less discretionary smoothing. Thus, we expect a negative relation

between BIG5 and smoothing.


13
   While firms trading in US markets are not required to report local accounts that comply with US GAAP,
Pownall and Schipper (1999), Ashbaugh and Olsson (2002) and Lang, Ready and Wilson (2006) suggest
that non-U.S. firms required to prepare U.S. GAAP financial information choose alternatives under IFRS or
their domestic standards that are closer U.S. GAAP.
14
   Analysts may potentially increase incentives for managers to smooth earnings to meet analysts’ earnings
expectations, or may be attracted to firms that smooth less for other reasons. That is less of a concern with
other measures such as regulatory environment, tax-book conformity, cross listing and ownership structure
since those measures are less likely to be caused by smoothing. Research on US firms in Yu (2006)
investigates the endogeneity of analyst following and suggests that analyst following serves primarily as a
source of capital market oversight and that analyst following appears to mitigate earnings management.
Our results are robust to exclusion of analyst following from our analyses.


                                                                                                          16
       Also, we expect accounting standards to affect the ability to smooth earnings.

While our ADR variable captures cross listing on US markets, many firms adopt US

GAAP or IFRS without cross listing. Research such as Bradshaw and Miller (2007) and

Barth et al. (2007) suggests that firms adopting IFRS or US GAAP are subjecting

themselves to more restrictive accounting standards and, therefore, have less flexibility to

manage earnings. As a result, we expect a negative relation between adoption of IFRS or

US GAAP (INTGAAP) and smoothing.

       The third firm-specific governance attribute that we consider is insider ownership.

We use the percent of shares that are closely held, %CLHLD, to proxy for insider

ownership (Himmelberg, Hubbard and Love, 2002; Lins and Warnock, 2004). While, in

theory, more concentrated ownership could result in increased monitoring of managers’

discretionary accounting practices, research like Lang, Lins and Miller (2003), Leuz, Lins

and Warnock (2005) and Leuz (2006) suggest that more concentrated ownership is

associated with increased agency issues internationally. As a consequence, we predict a

positive relation between %CLHLD and the smoothing measures.

       In summary, we model earnings smoothness as a function of a firm’s operating

characteristics and governance attributes. The OLS regression model with industry fixed

effects used to test the relation between smoothing and governance is as follows:


SMTHt   0  1 LNTOTASSt   2 LEVt   3 BM t   4 STD _ SALESt   5 % LOSSt
          6 OPCYCLEt   7 SGt   8OPLEVt   9 AVECFOt



         10 RIGHTSt  11TAXCONFORMt  12 ADRt  13 ANALYST  14 BIG5t
                                                               t
                                                                                      (1)

         15 INTGAAP  16 %CLHLDt
                    t




         a 1 a INDi   t
             60




                                                                                            17
SMTH is set equal to SMTH1or SMTH2 and the definitions of the operating
characteristic variables are as follows: LNTOTASS is equal to the natural log of total
assets measured in US dollars; LEV is equal to total debt divided by total assets; BM is
equal to book value of common equity divided by market value of equity; STD_SALES
is the standard deviation of sales scaled by total assets calculated requiring a minimum of
three and maximum of five fiscal years; %LOSS is the proportion of years that a firm
reports negative earnings, calculated requiring a minimum of three and maximum of five
fiscal years; OPCYCLE is the natural log of the operating cycle measured in days,
defined as 365*(average accounts receivable /sales)+365*(average inventory/cost of
goods sold); SG is the average sales growth over the past three to five years; OPLEV is
net property, plant and equipment over total assets; AVECFO is equal to the average cash
flow from operations divided by average total assets over the past three to five fiscal
years. All other variables are as previously defined.



III.3 Descriptive Statistics

        Table 2 reports the descriptive statistics on the two smoothing measures as well as

the operating characteristic variables used to explain differences in the innate portion of

earnings smoothing. The mean (median) values of SMTH1 and SMTH2 after pooling all

firm-year observations are -0.634 (-0.513) and 0.721 (0.911) respectively. A typical

sample firm is relatively large (median LNTOTASS 12.317 USD), has a significant

amount of debt in its capital structure (median LEV 0.213) and is a fairly mature firm as

captured by a median BM value of 0.689. The median sample firm has sales volatility of

0.100, reports losses infrequently, has an operating cycle of about 139 days and has

experienced 5.8% sales growth over recent years. The average firm has 26.7% of its

assets invested in net property, plant, and equipment.

        Table 2 also reports the descriptive statistics on the five governance attributes we

predict will affect managers’ incentives to report smooth earnings. The average firm in

our sample has an investor’s right score of 4.000 out of six, where six represents the

strongest investor protection. The mean TXBKCONFORM value of 0.586 indicates that



                                                                                          18
58.6% of our sample firms are from countries with a high degree of conformity between

tax and financial reporting. The descriptive statistics indicate that just over 3% of the

observations are associated with firms trading ADRs in the U.S.

         The mean and median analyst following indicate that our sample firms are, on

average, followed by relatively few analysts. We also see that, for the median firm,

44.6% of shares are closely held, suggesting that there are potential agency issues that

create greater incentives to smooth. Of our sample firms, 35.2% are audited by Big-5

firms and 7.1% prepare their financial statements under nonlocal GAAP (US GAAP or

IFRS).



IV. Results

IV.1 Incentives to Smooth

Our first general hypothesis is that earnings smoothing is affected by managers’

incentives to smooth and that there exist firm-specific and institutional governance

attributes that enhance or mitigate managerial incentives to smooth earnings.

         Table 3 reports the regression results of estimating the earnings smoothing model

using SMTH1 and then SMTH2. Significance levels are based on Fama-MacBeth (1973)

t-statistics to control for potential cross correlation in residuals. Because of the reduction

in sample size due to data limitations on %CLHLD, we table the results of estimating

equation (1) with and without the closely held incentive attribute. However, we discuss

the results of both analyses together because there are no significant differences in the

findings.




                                                                                            19
       Results for out innate controls are generally as expected given the nature of

accruals and the results in Francis et al. (2004). Regardless of smoothing measure, the

results indicate that firms that are larger have more earnings-smoothing accruals,

consistent with the notion of increased stability and diversification for larger firms.

Similarly, firms with higher BM ratios generally have smoother earnings, reflecting

lower levels of intangible assets and lower expected growth rates for those firms, as do

firms with higher LEV, reflecting the notion that firms with more stable fundamentals

have higher levels of debt financing. Consistent with expectations, firms that have more

volatile sales and more frequent losses have more volatile earnings. In addition, firms

with longer operating cycles and lower operating leverage tend to have smoother earnings

relative to cash flows. One potentially surprising result is that the smoothing effect of

accruals is greater for firms with relatively high sales growth, but that result is

conditional on sales variability, firm size and operating cycle. The effect of profitability

on smoothing depends on the earnings specification.

       Turning to our primary incentive variables of interest, all variables enter into the

regression significantly and as predicted. We find a negative relation between RIGHTS

and earnings smoothing indicating that managers report relatively smoother earnings in

countries with weak investor protection. This finding is consistent with prior research

that examines the relation between earnings smoothing and investor protection (Leuz et

al., 2003). Consistent with expectations, the positive and significant coefficient on

TXBKCONFORM indicates that managers are more likely to smooth earnings when

reporting financial information that is more closely linked to taxes. This finding suggests




                                                                                            20
that managers in high tax book conformity countries face incentives to smooth earnings

for tax purposes.

       As expected based on research like Lang, Raedy and Wilson (2006) and Leuz

(2006), we find a negative relation between ADR and earnings smoothing. This finding

indicates that managers of non-U.S. firms that face the regulatory oversight of the SEC

and restricted accounting measurement choices under U.S. GAAP have less of a tendency

to smooth earnings. The results also indicate that higher analyst following is associated

with less smoothing as the coefficient on ANALYST is negative and highly significant.

This finding suggests that capital market monitoring plays an important role in

diminishing managers’ incentives to smooth earnings. Also consistent with increased

oversight reducing earnings management, we find a strong negative association between

the presence of a Big-5 auditor and earnings smoothing. Finally, smoothing tends to be

less pronounced for firms that report under IFRS or US GAAP in their local accounts,

incremental to the effect of cross listing.

       Consistent with the notion that concentrated ownership can create incentives for

opacity, the results also suggest that non-U.S. firms’ ownership structures are associated

with earnings smoothing. Specifically, we find, after controlling for innate operating

characteristics, firms with more closely held shares engage in relatively more earnings

smoothing.

       Taken together, the results reported in Table 3 suggest that concentrated

ownership and tax book conformity encourage earnings smoothing, but the incentives to

smooth are mitigated in the presence of strong investor protection, regulatory oversight

over financial reporting, and monitoring by capital market participants. These results are




                                                                                           21
important because they suggest that, after controlling for innate determinants, excess

smoothing is correlated with incentive variables as would be expected if excess

smoothing reflects managerial discretion.

IV.2 Overall Internal Governance

In the preceding analyses, we examine the effects of internal and external incentives and

oversight to draw inferences on whether governance mitigates the propensity for

managers to smooth earnings. However, measures like ownership concentration have

potentially countervailing effects and a broader range of factors could be important to

incentives. An alternative approach is to consider firms’ governance structures as a

whole based on a more general governance index. To examine governance more

generally, we replace %CLHLD with an overall measure of governance (GOVSCORE)

that is equal to the governance rating of the firm as reported in Governance Metrics

International (GMI). The GMI score reflects not only the assessment of a firm’s

ownership structure, but also captures a firm’s governance related to board structure,

financial information quality, and firm level shareholder rights. The GMI data are only

available for the fiscal 2004 and 2005 reporting periods and only available for 1,122

firms, substantially reducing the sample size and limiting the analysis to large, widely-

followed firms. However, an advantage of the measure is that it includes a wider range

of factors and reflects more judgment in assessing the likely effects of governance

differences in practice

       Within the GMI scoring system, one particular category of interest is the scrutiny

and disclosure of related party transactions (RELATEDPARTY), where higher values

indicate a greater existence and less scrutiny being placed on related party transactions.




                                                                                             22
If managers engage in related party transactions to expropriate the firm's resources, then

they have incentives to manage earnings to mask such expropriation (Gordon and Henry

2005). When there is no or little scrutiny over related party transactions, the manager has

greater incentives to expropriate firm resources and smooth earnings to create opacity.

Thus, we predict a positive relation between RELATEDPARTY and SMTH.

       To test whether internal governance as a whole affects managers’ incentives to

smooth earnings and whether related party transactions have incremental smoothing

effects, we estimate the following OLS regression:


SMTHt   0  1 LNTOTASSt   2 LEVt   3 BM t   4 STD _ SALESt   5 % LOSSt
          6 OPCYCLEt   7 SGt   8OPLEVt   9 AVECFOt

         10 ADRt  11 ANALYST  12 BIG5t
                                t
                                                                                    (2)
         13 INTGAAP 14GOVSCOREt  15 RELATEDPAR t
                                                   TY
         16YR04t   t

where all variables are previously defined.

       Table 4 presents the results of estimating equation (2). As noted above, the

sample size is substantially reduced because we are estimating equation (2) for only two

years and the GMI data only cover the largest, most widely held companies. The

explanatory power of the operating characteristic and governance variables for SMTH

using the smaller sample is lower than when the model is estimated using the full sample,

perhaps reflecting the fact that the GMI firms tend to be larger and more homogeneous.

While the signs are generally consistent with the earlier models, the significance of the

estimated coefficients on the innate determinants of smoothing is reduced consistent with

the smaller sample size and greater homogeneity.




                                                                                            23
       In terms of our primary variables of interest, after controlling for innate operating

characteristics we find evidence of a negative relation between SMTH and GOVSCORE

in both models, although it is only significant in the SMTH2 model, suggesting that firms

with stronger internal governance engage in less earnings smoothing. Further, the

evidence suggests that a lack of scrutiny and disclosure of related party transactions

increases managers’ incentives to smooth earnings as reflected in a positive relation

between SMTH and RELATEDPARTY in both models.

        While the inferences from the analyses presented in Table 4 should be viewed

with caution due to the sample and time period limitations, the overall results presented

in Table 4 are consistent with those in the preceding analyses and support the notion that

smoothing is more pronounced when governance is weak.

IV.3 Smoothing and Liquidity

The preceding analyses provide evidence that excess smoothing is higher for firms where

governance and oversight are weaker and incentives to manage earnings are stronger. If

so, excess smoothing has the potential to affect transparency and, ultimately, the

willingness of investors to transact in a stock. Our second general hypothesis is that

excess smoothing is expected to have negative capital market consequences in terms of

reduced liquidity. In particular, if excess earnings smoothing is associated with reduced

information available to market participants, they will be less willing to transact in a

firm’s stock because of potential information asymmetries and resulting higher

transaction costs. We use the predicted value of equation (1) estimated using only the

operating characteristic variables as a measure of innate smoothing (INNATE_SMTH1 or

INNATE_SMTH2). We define the difference between a firm’s predicted value and




                                                                                            24
reported smoothing measure (i.e., SMTH1 or SMTH2) as excess smoothing

(EXCESS_SMTH1 or EXCESS _SMTH2, respectively), which we posit is affected by

managers’ incentives to smooth earnings and will result in reduced transparency and less

liquidity.

        To test our second hypothesis, we estimate the following OLS model with

industry and country fixed effects:

             LIQUIDITY   0  1 LNMVEt   2 LOG( PRC) t   3 BM t   4 LOSS
                      t

                            5 STD _ RETt
                                                                                   (3)
                            6 RINNATE _ SMTHt   7 REXCESS _ SMTHt
                           a 1  a INDi  b1  b COUNTRYi   t
                              60              20




Where LIQUIDITY is set equal to one of the three proxies for liquidity defined below,
LNMVE is equal to the natural log of market value of equity at the fiscal year end,
measured in US dollars; LOGPRC is the natural log of the firm’s share price as of the
fiscal year end, measured in US dollars; LOSS is equal to one if net income before
extraordinary items is negative, zero otherwise; ROA is equal to net income before
extraordinary items divided by average total assets. All other variables are as previously
defined.


        We use two measures of liquidity. First, we consider the bid-ask spread

(BID_ASK_SPRD), measured as the average bid-ask spread over the fiscal year, where

the bid-ask spread is calculated as (ASK-BID)/((ASK+BID)/2). As noted in research like

Glosten and Milgrom (1985), information asymmetry can lead to increased bid-ask

spreads and reduced share prices (Amihud and Mendelson, 1996). In their international

study, Lesmond et al. (1999) argue that a scarcity of information will increase

information asymmetry and, hence, the bid-ask spread. If excessive smoothing results in

less useful financial information, we predict a positive relation between

BID_ASK_SPRD and excess smoothing.



                                                                                         25
         Our second proxy for liquidity is the proportion of zero return days. As discussed

in Bekaert, Harvey and Lundblad (2006), an advantage of using the zero return measure

in an international setting is that stock prices are widely available and measured

consistently across markets relative to other measures such as volume or bid-ask

spreads.15 Lesmond et al. (1999) argues that a manifestation of high transaction costs

will be infrequent trading reflected in days without price movements. Bekaert, Harvey

and Lundblad (2006) apply the zero return measure in international contexts and find that

the measure predicts future returns and behaves like a priced returns factor. Lesmond

(2005) provides evidence that zero returns are a better proxy for liquidity than is volume

in international settings. Ashbaugh-Skaife, Gassen and LaFond (2006) provide evidence

that a zero return metric is a summary measure of the extent to which firm-specific

information is impounded in share price. Lesmond (2005) demonstrates that more

traditional measures of transactions costs such as bid-ask spreads, where available, tend

to be correlated with zero return days.

         Following Bekaert, Harvey and Lundblad (2006), we define the zero-return metric

(ZR) as the number of zero-return trading days over the fiscal year divided by the total

trading days of the firm’s fiscal year and use it as our second LIQUIDITY measure. If

excess smoothing results in greater transaction costs, we expect a positive relation

between ZR and our EXCESS_SMTH measures.

         The control variables are added to the model for consistency with prior literature

(Lee, Mucklow and Ready 1993; Welker 1995; Chordia, Roll, and Subrahmanyam 2000;

and Ertimur 2004). We transform INNATE_SMTH and EXCESS_SMTH into scaled


15
 We include country fixed effects in the model to control for potential cross country differences in the
measurement of the liquidity variables.


                                                                                                           26
percentile ranks, where values range from zero to one, with higher values representing

greater smoothing. The transformation is necessitated by our earlier definition of SMTH,

which cast SMTH as a non-positive value to facilitate interpretation of the SMTH

results.16

           Table 5 displays the descriptive statistics for the dependent and independent

variables of equation (3). As stated above, data requirements to calculate

BID_ASK_SPREAD reduce the sample size relative to the %ZERORET metric. The

descriptive statistics indicate that sample firms, on average, have zero returns on 36.0%

of the trading days in the year and have a spread of 2.8%.

           Panels A and B of Table 6 display the results of estimating equation (4) using the

two measures of liquidity. For both measures, the signs and significance of the

coefficients on the control variables, in general, allow us to draw similar inferences

across the analyses and are consistent with the prior literature, so we discuss them only

once. The results suggest that larger firms (LNMVE), with lower book-to-market ratios

(BM), trading at a lower price per share (LOG(PRC)), with more frequent losses (LOSS)

and more volatile returns (STD_RET) tend to be more liquid, although the relations are

not always statistically significant.

           Turning to the variables of interest, panel A of Table 6 reports the results of the

bid-ask spread analysis. The results indicate that there is a negative and significant

coefficient on RINNATE_SMTH, suggesting that expected smoothing as a result of firms

operations is associated with reduced information asymmetry. A potential interpretation

is that industries and operating environments where accruals naturally smooth earnings

are also characterized by reduced uncertainty and potential for information asymmetry.
16
     Results are consistent for the raw (unranked) smoothing variables.


                                                                                                 27
In contrast, the excess portion of smoothing is positively associated with bid-ask spreads

suggesting that excessive smoothing is associated with more opaque financial

information that increases transaction costs. Coupled with our previous results, it appears

that firms where managers have incentives to increase opacity and there is relatively little

oversight tend to smooth more aggressively and that the excess smoothing is associated

with reduced transparency reflected in higher transactions costs. Further, the fact that the

two components of smoothing have opposite signs indicates that the effect of smoothing

depends on its source and provides some assurance that our approach for splitting

smoothing into components identifies substantive differences.

       Similar results obtain for the zero-return analysis in table 6, panel B. When we

bifurcate earnings smoothing into innate and excess components, we find a significantly

positive coefficient on the excess portion of earnings smoothing, i.e., REXCESS_SMTH.

This finding suggests that investors are less willing to trade in firms’ shares when

managers report earnings that are excessively smooth relative to underlying cash flows.

We also find a marginally significant negative coefficient on the innate portion of

smoothing, i.e., RINNATE_SMTH, suggesting that the smoothing that comes about as a

result of operating characteristics increases investors’ willingness to transact.

       In summary, the results presented in Table 6 for both of our liquidity measures

and both of our smoothing measures support the notion that excess smoothing is

associated with reduced liquidity and higher transactions costs. Results for the relation

between innate smoothing and bid-ask spreads and zero return days suggest that innate

smoothing is associated with lower transactions costs and enhanced liquidity. Taken




                                                                                            28
together, the results highlight the potential countervailing effects of innate and excess

smoothing.



IV.4 Interaction between Excess Smoothing and Governance

A final question is whether there is an interaction effect between smoothing and

governance. In particular, the potential effects of smoothing on transparency are likely

to be less pronounced for firms that are in otherwise rich information environments with

fewer incentives to manage earnings and higher levels of attestation and regulatory

oversight. For example, we expect that, for, large cross-listed firms with significant

analyst following, Big-5 auditors and dispersed ownership that report under IFRS, the

effect of smoothing on the information environment is unlikely to be as pronounced as

for smaller firms with less average coverage, local auditors and concentrated ownership

that trade only on the local market and file under local GAAP.

       To examine that issue we create a governance composite variable (GOV) under

which firms receive one point for each of the following: (1) if they are cross listed in the

US, (2) if they are followed by more than one analyst, (3) if they report under IFRS or US

GAAP, (4) if they are audited by a Big 5 auditor and (5) if they have closely held

ownership of less than 44%. As a consequence, GOV takes on lower values the more

likely it is the case that the firm faces significant incentives to manage earnings and

relatively weak oversight.

       Table 7 reports results interacting GOV with our excess smoothing measure in our

transaction cost and liquidity regressions. Results for the control variables and innate

smoothing are similar to table 6.




                                                                                            29
       In terms of the variables of interest, excess smoothing enters positively as before,

indicating that firms with greater levels of excess smoothing experience lower liquidity

and higher transactions costs. Further, the coefficient on GOV is positive, indicating,

consistent with expectations, that firms that are more likely to have incentive issues and

reduced oversight experience lower liquidity and higher transactions costs. Finally, the

interaction between GOV and REXCESS is significantly negative, indicating that the

effects of smoothing tend to be less pronounced in environments in which governance is

otherwise strong.

       Taken together, the results suggest that excess smoothing tends to create opacity,

which raises transactions costs and reduces liquidity, and that the relation is strongest in

cases where other governance issues are likely to be most pronounced.



IV.5 Other Analyses

       First, as noted above, we replicate the results using trading volume, number of

shares traded over the fiscal year divided by the number of shares outstanding,

(VOLUME) as a proxy for liquidity. An opaque information environment can lead to

lower trading volume because of higher transaction costs and greater information

asymmetry, so we expect a negative relation between excess smoothing and VOLUME.

However, research such as Lesmond (2005) and Bekaert, Harvey and Lundblad (2006)

suggests that volume tends to be a relatively weak proxy for transactions costs and

liquidity relative to bid-ask spreads and zero returns days in that it may not be computed

consistently across exchanges, and does not behave like a priced liquidity factor or

correlate highly with other liquidity measures. Results for volume are consistent with




                                                                                           30
those for bid-ask spread and zero return days in that excess smoothing tends to reduce

trading volume, particularly in cases where other governance factors are weak.

        Second, we repeat the entire analysis after eliminating Japanese and UK firms

because these two countries add the most firms to our sample, potentially threatening the

external validity of our results. We draw similar inferences from the results after

eliminating Japanese and UK firms. Specifically, we continue to find support for both

hypotheses that excess smoothing is significantly lower when firms have better

governance and excessive smoothing is significantly associated with reduced liquidity.

        Third, in testing for liquidity effects, rather than pooling firm-year observations

from all countries, we estimate the smoothing model within each country. We then use

the firm-specific residuals and predicted values from the within-country estimates to test

the relation between excess and innate smoothing, and measures of liquidity. The results

of these analyses are similar to those reported in the tables. Specifically, we continue to

find that excessive smoothing is significantly associated with lower liquidity regardless

of the liquidity measure used. In addition, we find that innate smoothing, when

significant, is associated with greater liquidity.

        Fourth, in testing our second hypothesis, we replicate the analysis including the

governance variables from the first smoothing analysis as controls. In particular, a

potential concern is that our excess smoothing measures may be capturing the notion that

poor governance in general is associated with greater opacity rather than the effects of

smoothing. We do not include the governance variables as controls in our primary

analysis because it is difficult to disentangle the effects of governance overall from the




                                                                                              31
effects of governance through smoothing. However, our conclusions are robust to

replicating the liquidity analysis including the governance controls.

       Fifth, in testing for liquidity effects, we include controls for the overall level of

accruals and for the absolute value of accruals. In particular, papers like Bhattacharya et

al (2007) and Jayaraman (2007) suggest that, in US contexts, larger accruals generally

may be associated with greater informed trading and higher transactions costs.

Intuitively, that could be the case because extreme accruals may indicate unusual

circumstances for the firm that are associated with greater uncertainty and more

information asymmetry. Extreme accruals should not affect our analysis directly because

our interest is only in the excess component of earnings smoothing and our results

suggest an asymmetric relation between smoothing and liquidity depending on the source

of the smoothing. However, to ensure that our results are not affected by the general

level of accruals overall, we replicate our analysis including the magnitude of accruals

and the absolute value of accruals. Our conclusions are unaffected by inclusion of the

magnitude of accruals and the absolute value of accruals.

       The results of these sensitivity analyses support our overall conclusions that better

governance is associated with reduced smoothing and that excessive smoothing is

associated with reduced liquidity.



V. Conclusion

We examine the relation between earnings management via earnings smoothing,

governance and liquidity. Our evidence suggests that better governance mitigates

earnings smoothing. In particular, earnings smoothing is more prevalent when there is




                                                                                               32
weak investor protection, when fewer analysts follow the firm, when there is a greater

proportion of closely held shares, and when there is less scrutiny over related party

transactions. We also document that excess earnings smoothing has capital market

effects as evidenced by a negative relation between our measure of excess smoothing and

liquidity as measured by the frequency of zero-return days, bid-ask spreads, and share

volume. Our results suggest that firms that excessively smooth earnings likely face lower

liquidity and higher transactions costs, potentially increasing cost of capital. Results are

particularly strong in environments in which excess smoothing is coupled with other

governance issues.

       The results raise questions for future research. First, our study examined a limited

set of governance attributes. Future research can explore whether alternative governance

attributes reduce or increase firms’ smoothing. Second, we investigated only one set of

capital market consequences – transactions costs, investors’ willingness to trade and

resulting liquidity – that has implications for cost of capital. Future research can explore

other capital market consequences and whether other economic events (e.g., mandatory

dividend payouts) affect incentives and economic consequence of earnings management.




                                                                                           33
                                    APPENDIX
                         Country-wide Governance Attributes


                                                  Limited
Sample            Inventory     Depreciation        Tax          TXBK-           Investor
Country          Conformity     Conformity       Incentives     CONFORM         Protection
Australia            No             No              No             0                4
Austria              Yes           Yes              Yes            1                2
Belgium              Yes           Yes              No             1                0
Canada               No             No              Yes            0                5
Denmark              No             No              No             0                2
Finland              Yes           Yes              Yes            1                3
France               Yes           Yes              Yes            1                3
Germany              Yes           Yes              Yes            1                1
Greece               Yes           Yes              No             1                2
Hong Kong            No             No              Yes            0                5
Ireland              Yes            No              No             0                4
Italy                Yes           Yes              Yes            1                1
Japan                Yes           Yes              No             1                4
Netherlands          No             No              No             0                2
New Zealand          No             No              No             0                4
Norway               No             No              Yes            0                4
Singapore            No             No              No             0                4
Spain                Yes           Yes              No             1                4
Sweden               Yes            No              Yes            1                3
Switzerland          Yes           Yes              Yes            1                2
UK                   No             No              Yes            0                5


The tax book conformity index (TXBKCONFORM) is developed from tax summaries
provided in Corporate Taxes: A Worldwide Summary (Price Waterhouse 1995).
Inventory conformity is noted yes when the inventory method used for tax reporting must
also be used for financial reporting, and no otherwise. Depreciation conformity is noted
yes when tax depreciation must also be recorded for financial reporting, and no
otherwise. Limited tax incentives is noted yes when there are fewer than four tax
incentives identified in the country summary, and no otherwise. Tax authorities that
instill tax incentives such as a research and development tax credit, an employment tax
credit, etc. are less likely to require tax-book conformity because tax credits are used to
manage tax payments. Thus the lack of tax incentives is treated as an indicator of high
tax-book conformity. TXBKCONFORM is coded one when two or three tax conformity
measures are denoted yes. Investor Protection is as reported in La Porta et al. (1998).




                                                                                        34
REFERENCES

Ashbaugh, H., Collins, D., LaFond, R. 2006. The Effects of Corporate Governance on
Firms’ Credit Ratings. Journal of Accounting and Economics, forthcoming.

Ashbaugh, H., LaFond, R. 2004. Reporting Incentives and the Quality of Non-U.S.
Firms’ Working Capital Accruals. Working paper, University of Wisconsin.

Ashbaugh, H., Olsson P. 2002. An Exploratory Study of the Valuation Properties of
Cross-listed Firms’ IAS and US-GAAP Earnings and Book Values. The Accounting
Review 77: 107-126.

Ball, R., L. Shivakumar. 2005. Earnings Quality in UK Private Firms: Comparative Loss
Recognition Timeliness. Journal of Accounting and Economics 39: 83-128.

Ball, R., L. Shivakumar. 2006. The Role of Accruals in Asymmetrically Timely Gain and
Loss Recognition. Journal of Accounting Research 44: 207-242.

Barth, M., Landsman, W., Lang, M. 2007. International Accounting Standards and
Accounting Quality. Forthcoming, Journal of Accounting Research.

Bekaert, G., Harvey C., Lundblad C. 2006. Liquidity and Expected Returns: Lessons
from Emerging Markets. Forthcoming, Review of Financial Studies.

Bhattacharya, N., Desai, H., Venkataraman, K. 2007. Earnings Quality and Information
Assymetry: Evidence from Trading Costs. Working paper, Southern Methodist
University.

Bhattacharya, U., Daouk, H., Welker, M. 2003. The World Price of Earnings
Opacity. The Accounting Review 78: 641-678.

Bradshaw, M., and G. Miller, 2007. "Will Harmonizing Accounting Standards Really
Harmonize Accounting? Evidence from Non-U.S. Firms Adopting U.S. GAAP." Journal
of Accounting, Auditing and Finance (forthcoming).

Breeden, R., 1994. Foreign Companies and U.S. Markets in a Time of Economic
Transformation. Fordham International Law Journal 17.

Burgstahler, D., Hail, L., Leuz, C. 2006. The Importance of Reporting Incentives:
Earnings Management in European Private and Public Firms. The Accounting Review.
October.

Bushman, R. Smith, A. 2001, “Financial Accounting Information and Corporate
Governance,” Journal of Accounting and Economics, Vol. 32, Nos. 1-3, December.

Chordia, T., Roll, R., Subrahmanyam, A. 2000. Co-Movements in Bid-Ask Spreads and
Market Depth. Financial Analysts Journal 56: 23–27.


                                                                                     35
Desai, M., Dharmapala, D. 2006. Earnings Management and Corporate Tax Shelters.
Working paper, Harvard University.

Diamond, D., Verrecchia. 1991. Disclosure, Liquidity and the Cost of Capital. The
Journal of Finance. Vol. XLVI, No. 4: 1325-1359.

Dichev, I., Tang, W., 2005. Matching and the Volatility of Earnings. Working Paper.

Fama, E., MacBeth, J., 1973. Risk, return and equilibrium: empirical tests. Journal of
Political Economy 81: 607-636.

Francis, J., LaFond, R., Olsson, P., Schipper, K., 2004. Costs of Equity and Earnings
Attributes, The Accounting Review 79, 967-1010.

Gordon, E., Henry, E, 2005. "Related Party Transactions and Earnings Management"
working paper.

Haw, I., Hu, B., Hwang, L., Wu, W., 2004. Ultimate Ownership, Income Management
and Extra-Legal Institutions. Journal of Accounting Research 42: 423-462.

Himmelberg, C., Hubbard, R., Love, I. 2002. Investment, Protection, Ownership, and the
Cost of Capital. Working Paper National Bank of Belgium.

Jayaraman, S. 2007. Earnings Volatility, Cash Flow Volatility and Informed Trading,
Working paper.

Jensen, M. C., Meckling W. H. , 1976. Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure, Journal of Financial Economics. 3:305-60.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R. 1998. Law and finance.
Journal of Political Economy 106: 1113-1155.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A. 1999. Corporate ownership around the
world. Journal of Finance 54: 471-517.

Lang, M., Lins, K., Miller, D. 2003. ADRs, analysts, and accuracy: Does cross listing in
the United States improve a firm’s information environment and increase market value?
Journal of Accounting Research 41: 317-345.

Lang, M., Lins, K., Miller, D. 2004. Concentrated Control, Analyst Following, and
Valuation: Do Analysts Matter Most When Investors are Protected Least? Journal of
Accounting Research 42: 581-623.




                                                                                         36
Lang, M., Smith Raedy, J., Wilson, W., 2006, Earnings Management and Cross Listing:
Are Reconciled Earnings Comparable to U.S. Earnings? Journal of Accounting and
Economics 42: 255-283.

Lang, M., Raedy, J., Yetman, M., 2003. How Representative are Firms that are Cross
Listed in the United States? An Analysis of Accounting Quality. Journal of
Accounting Research 41: 363-386.

Larcker, D, Richardson, S., Tuna, I., 2006. How Important is Corporate Governance?
Working paper, Wharton School, University of Pennsylvania.

Lee, C., Mucklow, B., Ready, M., 1993. Spreads, depths, and the impact of earnings
information: an intraday analysis. Review of Financial Studies 6: 345-374.

Lesmond, D., Ogden, J., Trzcinka, J. 1999. A New Estimate of Transaction Costs. The
Review of Financial Studies. 12: 1113-1141.

Lesmond, D. 2005. Liquidity of Emerging Markets. Forthcoming Journal of Financial
Economics. 77: 411-452.

Leuz, C., 2006. Cross Listing, Bonding and Firms’ Reporting Incentives: A Discussion of
Lang, Raedy and Wilson. Journal of Accounting and Economics 42: 284-296.

Leuz, C., Nanda, D.,Wysocki, P. 2003. Earnings Management and Investor Protection:
An International Comparison. Journal of Financial Economics 69: 505-527.

Lins, K , Warnock, F. 2004. Corporate Governance and the Shareholder Base. ECGI -
Finance Working Paper.

Pownell, G., Schipper, K. 1999. Implications for Accounting Research for the SEC’s
Consideration of International Accounting Standards for U.S. Securities Offerings.
Accounting Horizons 13: 259-280.

Shleifer, A., Vishny, R., 1997. A Survey of Corporate Governance. Journal of Finance
Vol. LII, No. 2: 737-783.

Tucker, J., Zarowin, P., 2006. Does Income Smoothing Improve Earnings
Informativeness? The Accounting Review 81: 251-270.

United States Securities and Exchange Commission (SEC). 2000. International
Accounting Standards. February 16, 2000. Release Nos. 33-7801, 34-42430; International
Series Release No. 1215; File No. S7-04-00.

Verrecchia, R., 2001. Essays on Disclosure. Journal of Accounting and Economics 32,
1-3: 97-180.




                                                                                       37
Welker, M., 1995. Disclosure policy, information asymmetry, and liquidity in equity
markets. Contemporary Accounting Research 11, 801-827.

Wysocki, P., 2004. Discussion of Ultimate Ownership, Income Management and Extra-
Legal Institutions. Journal of Accounting Research 42: 463-474.

Yu, F., 2006. Analyst Coverage and Earnings Management. Working Paper, University
of Minnesota.




                                                                                      38
                                            TABLE 1
                            Sample and Country-wide Earnings Smoothing


                                     SMTH1                               SMTH2
       Country            Mean       Median      Std Dev      Mean       Median      Std Dev        n
      Australia           -0.729      -0.632      0.537       0.618       0.869       0.515       3,174
       Austria            -0.464      -0.358      0.375       0.826       0.957       0.322        636
       Belgium            -0.631      -0.550      0.454       0.725       0.891       0.374        680
       Canada             -0.799      -0.724      0.546       0.581       0.800       0.495       4,496
      Denmark             -0.567      -0.469      0.382       0.763       0.919       0.347       1,118
       Finland            -0.703      -0.599      0.472       0.697       0.858       0.384        916
        France            -0.603      -0.455      0.523       0.784       0.929       0.338       5,021
      Germany             -0.560      -0.423      0.479       0.783       0.940       0.361       4,951
       Greece             -0.388      -0.295      0.312       0.908       0.971       0.177        912
     Hong Kong            -0.793      -0.686      0.621       0.622       0.842       0.481       2,943
       Ireland            -0.716      -0.674      0.423       0.627       0.828       0.456        478
         Italy            -0.477      -0.365      0.396       0.819       0.955       0.317       1,521
        Japan             -0.540      -0.423      0.449       0.791       0.942       0.347      22,160
     Netherlands          -0.540      -0.384      0.470       0.781       0.948       0.371       1,348
     New Zealand          -0.586      -0.509      0.429       0.750       0.909       0.366        416
       Norway             -0.801      -0.735      0.536       0.610       0.788       0.442        961
      Singapore           -0.680      -0.498      0.634       0.720       0.922       0.426       2,235
        Spain             -0.517      -0.397      0.448       0.816       0.954       0.300        881
       Sweden             -0.802      -0.727      0.562       0.574       0.794       0.506       1,666
     Switzerland          -0.701      -0.505      0.683       0.737       0.917       0.382       1,556
          UK              -0.738      -0.648      0.504       0.626       0.835       0.469      11,741

SMTH1 is defined as the standard deviation of net income before extraordinary items scaled by average
total assets divided by the standard deviation of cash flow from operations scaled by average total assets,
where standard deviations are calculated using a minimum of three and maximum of five years of data.
SMTH1 is multiplied by negative one so that larger values, i.e., values closer to zero represent more
smooth earnings. SMTH2 is defined as the correlation between the cash flow from operations and total
accruals where both measures are scaled by average total assets, where correlations are calculated using a
minimum of three and maximum of five years of data. SMTH2 is multiplied by negative one so that larger
values, i.e., values closer to one represent more smooth earnings. n represents the number of firm-year
observations over 1994 – 2005.




                                                                                                        39
                                              TABLE 2
    Descriptive Statistics on Non-U.S. Firms’ Earnings Smoothing and Operating Characteristics


                      Variable                  Mean             Median             Std Dev

             SMTH1                             -0.634             -0.513             0.510
             SMTH2                              0.721              0.911             0.413
             LNTOTASS                          12.426             12.317             1.893
             LEV                                0.233              0.213             0.171
             BM                                 0.858              0.689             0.700
             STD_SALES                          0.151              0.100             0.159
             %LOSS                              0.233              0.000             0.301
             OPCYCLE                            4.875              4.937             0.682
             SG                                 0.304              0.058             2.357
             OPLEV                              0.291              0.267             0.217
             AVECFO                             0.049              0.057             0.109
             RIGHTS                             3.689              4.000             1.264
             TXBKCONFORM                        0.586              1.000             0.493
             ADR                                0.031              0.000             0.173
             ANALYST                            5.165              2.000             7.376
             BIG5                               0.352              0.000             0.478
             INTGAAP                            0.071              0.000             0.257
             %CLHLDa                            0.447              0.446             0.223

Descriptive statistics are based on all firms having sufficient data over the 1994 – 2005 time period
(n=69,810). SMTH1 is defined as the standard deviation of net income before extraordinary items scaled
by average total assets divided by the standard deviation of cash flow from operations scaled by average
total assets, where standard deviations are calculated using a minimum of three and maximum of five years
of data. SMTH1 is multiplied by negative one so that larger values, i.e., values closer to zero represent
more smooth earnings. SMTH2 is defined as the correlation between cash flow from operations and total
accruals where both measures are scaled by average total assets, where correlations are calculated using a
minimum of three and maximum of five years of data. SMTH2 is multiplied by negative one so that larger
values, i.e., values closer to one represent more smooth earnings. The definition of operating characteristic
variables are as follows: LNTOTASS is equal to the natural log of average total assets measured in US
dollars over the SMTH estimation period; LEV is equal to the average total debt divided by total assets
over the SMTH estimation period; BM is equal the average to book value of common equity divided by
market value of equity over the SMTH estimation period; STD_SALES is the standard deviation of sales
scaled by total assets calculated requiring a minimum of three and maximum of five fiscal years; %LOSS
is the proportion of years that a firm reports negative earnings, calculated requiring a minimum of three and
maximum of five fiscal years; OPCYCLE is the natural log of the average operating cycle measured in
days, defined as 365*(average accounts receivable /sales)+365*(average inventory/cost of goods sold) over
the SMTH estimation period; SG is the average sales growth over the past three to five years; OPLEV the
average is net property, plant and equipment over total assets over the SMTH estimation period;
DIVIDEND is equal to the average cash dividends divided by average total assets over the SMTH
estimation period; AVECFO is equal to the average cash flow from operations divided by total average
total assets over the past three to five fiscal years. Governance variables are defined as: RIGHTS is the
antidirector rights index developed by La Porta et al. (1998) for the country; TXBKCONFORM is equal to
one if there is a high degree of conformity between tax and financial reporting in the country, and zero
otherwise (see Appendix for details); ADR is equal to one if the firm trades in the U.S. during the fiscal
year, and zero otherwise; ANALYST is equal to the average number of analysts making a forecast for
fiscal year t’s earnings over the SMTH estimation period; and %CLHLD is the average proportion of



                                                                                                          40
                                                                                                   a
shares that are closely held as of the end of the fiscal year t over the SMTH estimation period.       Requiring
firms to have closely held ownership data reduces the sample size to 53,553.




                                                                                                              41
                                                            TABLE 3
                                                 Incentives to Smooth Earnings


       Incentives to Smooth - Annual Cross-Sectional Fama-MacBeth Regressions
       of SMTH1 and SMTH2 Regressed on Firm Operating Characteristics and Incentive Attributes

        SMTHt   0  1 LNTOTASSt   2 LEVt   3 BM t   4 STD _ SALESt   5 % LOSSt
                6OPCYCLEt   7 SGt   8OPLEVt   9 AVECFO  10 RIGHTSt
                                                                     t
               11TXBKCONFOR t  12 ADRt  13 ANALYST  14 BIG5
                                      M                            t
                                                                                                             (2)
               15 INTGAAP 16 %CLHLDt
              t  a 1 a INDi   t
                    60




                                               SMTH1                                                 SMTH2
                     parameter                    parameter                   parameter                 parameter
                      estimate           p-value   estimate         p-value    estimate        p-value   estimate         p-value


INTERCEPT                -0.486           0.00      -0.576           0.00      0.737            0.00      0.652            0.00
Innate Characteristics
LNTOTASS                 0.020            0.00      0.023            0.00      0.015            0.00      0.018            0.00
LEV                      0.075            0.00      0.059            0.01      0.073            0.00      0.059            0.00
BM                       0.000            0.87      -0.003           0.24      0.021            0.00      0.021            0.00
STD_SALES                -0.128           0.00      -0.131           0.00      -0.098           0.00      -0.101           0.00
%LOSS                    -0.650           0.00      -0.644           0.00      -0.423           0.00      -0.420           0.00
OPCYCLE                  0.009            0.02      0.013            0.01      0.015            0.00      0.018            0.00
SG                       0.003            0.01      0.004            0.00      0.003            0.06      0.004            0.04
OPLEV                    -0.135           0.00      -0.138           0.00      -0.106           0.00      -0.112           0.00
AVECFO                   -0.161           0.00      -0.145           0.00      0.276            0.00      0.294            0.00
Governance
RIGHTS                   -0.029           0.00      -0.029           0.00      -0.021           0.00      -0.020           0.00
TAXCONFORM               0.039            0.00      0.020            0.02      0.023            0.00      0.011            0.07
ADR                      -0.058           0.00      -0.052           0.00      -0.075           0.00      -0.065           0.00
ANALYST                  -0.044           0.00      -0.039           0.00      -0.030           0.00      -0.026           0.00
BIG5                     -0.023           0.00      -0.023           0.00      -0.026           0.00      -0.026           0.00
INTGAAP                  -0.059           0.00      -0.068           0.00      -0.030           0.00      -0.032           0.00
%CLHLD                                              0.073            0.00                                 0.061            0.00


Industry Dummies                  YES                        YES                        YES                        YES
                 2
 Average Adj R                    0.26                       0.25                       0.25                       0.24
     Average n                    5,816                      4,463                      5,816                      4,463


       All variables are defined in Table 2.




                                                                                                                       42
                                               TABLE 4
                               Overall Governance and Incentives to Smooth

Incentives to Smooth – Pooled Cross Sectional Regressions, firm cluster standard errors
of SMTH1 and SMTH2 Regressed on Firm Operating Characteristics and Incentive Attributes

SMTHt   0  1 LNTOTASSt   2 LEVt   3 BM t   4 STD _ SALESt   5 % LOSSt
        6OPCYCLEt   7 SGt   8OPLEVt   9 AVECFO       t
       10 ADRt  11 ANALYST  12 BIG5  13 INTGAAP 14GOVSCOREt
                                 t
       15 RELATEDPAR t  15YR04t   t
                         TY


                                                SMTH1                    SMTH2
                                         parameter                parameter
                                          estimate  p-value        estimate  p-value


                INTERCEPT                 -0.568           0.00      0.447           0.01
                Innate Characteristics
                LNTOTASS                  0.020            0.09      0.026           0.00
                LEV                       0.023            0.79      0.103           0.15
                BM                        -0.014           0.71      0.022           0.49
                STD_SALES                 0.054            0.62      0.005           0.96
                %LOSS                     -0.786           0.00     -0.684           0.00
                OPCYCLE                   0.000            0.98      0.026           0.08
                SG                        0.004            0.39      0.004           0.15
                OPLEV                     -0.115           0.06     -0.105           0.04
                AVECFO                    -0.689           0.00      0.138           0.50
                Governance
                ADR                       -0.022           0.42      0.003           0.87
                ANALYST                   -0.055           0.01     -0.049           0.01
                BIG5                      0.011            0.68      0.017           0.46
                INTGAAP                   -0.010           0.71     -0.021           0.31
                GOVSCORE                  -0.009           0.11     -0.015           0.00
                RELATEDPARTY              0.010            0.01      0.005           0.09


                     Year Dummies                  YES                       YES
                                2
                        Adj R                      0.13                      0.17
                           n                       2,049                     2,049

This table reports the results of a pooled cross-sectional regression using data from fiscal 2004 and 2005
using 2,071 firm-year observations for 1,122 firms. The smaller sample sizes are due to the requirement
that firms be followed by Governance Metrics International (GMI). GOVSCORE is GMI’s global overall
rating, where higher values represent stronger governance; RELATEDPARTY is equal to the GMI
assessment of related party transactions where higher values indicate a greater existence and less scrutiny
being placed on related party transactions. All other definitions are provided in Table 2.




                                                                                                        43
                                                TABLE 5
                       Descriptive Statistics for Market Consequences Variables



                Variable                  n             Mean            Median          Std Dev

        %ZERORET                       69,721           0.360             0.273          0.260
        BIDASK                         23,698           0.028             0.014          0.035
        LNMVE                          69,721          11.862            11.703          1.958
        LOG(PRC)                       69,721           1.465             1.497          2.007
        BM                             69,721           0.858             0.689          0.700
        LOSS                           69,721           0.242             0.000          0.428
        STD_RET                        69,721           0.128             0.102          0.170

Variable definitions are as follows: %ZERORET is equal to the percent of days in fiscal year t for which
the stock price does not change; BIDASK is equal to the average bid ask spread over the fiscal year, where
the bid ask spread is equal to (ASK-BID)/((ASK+BID)/2)); LNMVE is equal to the natural log of market
value of equity at the fiscal year end, measured in US dollars; LOG(PRC) is the natural log of firm’s share
price as of the fiscal year end, measured in US dollars; LOSS is equal to one if net income before
extraordinary items is negative, and zero otherwise; STD_RET is the standard deviation of monthly returns
over the past three to five years.




                                                                                                        44
                                               TABLE 6
                                    Market Consequences of Smoothing

     Panel A: Annual Cross Sectional Fama-MacBeth Regressions of the Bid Ask Spread (BIDASK) Regressed
     on Expected and Excess Smoothing (2001-2005 time period)


       Log( BIDASKt )   0  1 LNMVEt   2 LN ( PRC) t   3 BM t   4 LOSS   5 STD _ RETt
                        6 RINNATE _ SMTHt   7 REXCESS_ SMTHt  a 1  a INDi
                                                                                60


                          b1  bCOUNTRYi   t
                              20



                                                   SMTH1                              SMTH2
                                            parameter                           parameter
                                             estimate  p-value                   estimate   p-value
INTERCEPT                                    5.956            0.00                5.957            0.00
LNMVE                                        -0.394           0.00                -0.394           0.00
LOG(PRC)                                     0.012            0.25                0.013            0.21
BM                                           0.004            0.33                0.007            0.15
LOSS                                         -0.015           0.25                -0.017           0.21
STD_RET                                      -0.002           0.97                -0.004           0.94


RINNATE_SMTH                                 -0.270           0.00                -0.263           0.00
REXCESS_SMTH                                 0.075            0.00                0.063            0.00


Industry Dummies                                      YES                                  YES
Country Dummies                                       YES                                  YES
                 2
 Average Adj R                                        0.77                                 0.77
     Average n                                        4,740                                4,740




                                                                                                    45
     TABLE 6 Continued
     Panel B: Annual Cross Sectional Fama-MacBeth Regressions of the Zero Return Metric (ZR) Regressed
     on Expected and Excess Smoothing


       ZRt   0  1 LNMVE t   2 LN ( PRC ) t   3 BM t   4 LOSS   5 STD _ RETt
              6 RINNATE _ SMTH t   7 REXCESS _ SMTH t  a 1  a INDi  b1  b COUNTRY i   t
                                                                         60               20




                                                       SMTH1                                 SMTH2
                                               parameter                             parameter
                                                estimate  p-value                     estimate     p-value
INTERCEPT                                        5.299           0.00                  5.450              0.00
LNMVE                                           -0.520           0.00                  -0.519             0.00
LOG(PRC)                                         0.123           0.00                  0.122              0.00
BM                                               0.040           0.00                  0.044              0.00
LOSS                                            -0.174           0.00                  -0.169             0.00
STD_RET                                         -1.254           0.00                  -1.247             0.00


RINNATE_SMTH                                    -0.100           0.04                  -0.133             0.14
REXCESS _SMTH                                    0.046           0.00                  0.059              0.00


Industry Dummies                                         YES                                      YES
Country Dummies                                          YES                                      YES
 Average Adj R2                                          0.63                                     0.63
     Average n                                           5,810                                    5,810


     Each panel reports the mean results of estimating annual cross-sectional regressions over Panels A and C
     1994 – 2005 and 2001-2005 Panel B, where p-values are based on the time-series standard errors of the
     coefficient estimates. ZR is equal to log(%ZERORET /(1-%ZERORET)), where %ZERORET is equal to
     the percent of days in fiscal year t for which the stock price does not change. RINNATE_SMTH is equal
     to the scaled percentile rank of INNATE_SMTH, where INNATE_SMTH is equal to the predicted value
     from the earnings smoothing model. REXCESS_SMTH is equal to the scaled percentile rank of
     EXCESS_SMTH, where EXCESS_SMTH is equal to the percentile rank residual value from the earning
     smoothing model. All other variables are defined in Table 5.




                                                                                                             46
                                           TABLE 7
                                Market Consequences of Smoothing

Panel A: Annual Cross Sectional Fama-MacBeth Regressions of the Bid Ask Spread (BIDASK) Regressed
on Expected and Excess Smoothing and Governance (2001-2005 time period)

 Log( BIDASKt )   0  1 LNMVEt   2 LN ( PRC) t   3 BM t   4 LOSS   5 STD _ RETt
                  6 RINNATE _ SMTHt   7 REXCESS _ SMTHt   8GOV
                  9 GOV * REXCESS _ SMTHt  a 1  a INDi  b1  bCOUNTRYi   t
                                                       60               20



                                             SMTH1                    SMTH2
                                      parameter                parameter
                                       estimate  p-value        estimate  p-value
               INTERCEPT               5.708            0.00    5.702            0.00
               LNMVE                   -0.344           0.00    -0.344           0.00
               LOG(PRC)                -0.002           0.78    -0.002           0.83
               BM                      0.004            0.41    0.007            0.21
               LOSS                    0.008            0.51    0.007            0.59
               STD_RET                 0.001            0.99    -0.001           0.99
               RINNATE_SMTH            -0.264           0.00    -0.260           0.00
               REXCESS_SMTH            0.067            0.01    0.070            0.03
               GOV                     -0.142           0.00    -0.138           0.00
               REXCESS_SMTH*
               GOV                     -0.009           0.09    -0.017           0.04


                Industry Dummies                YES                      YES
                Country Dummies                 YES                      YES
                                  2
                  Average Adj R                 0.78                     0.78
                    Average n                   4,740                    4,740




                                                                                              47
TABLE 7 Continued
Panel B: Annual Cross Sectional Fama-MacBeth Regressions of the Zero Return Metric (ZR) Regressed
on Expected and Excess Smoothing

        ZRt   0  1 LNMVEt   2 LN ( PRC) t   3 BM t   4 LOSS   5 STD _ RETt
              6 RINNATE _ SMTHt   7 REXCESS_ SMTHt   8GOV
              9GOV * REXCESS_ SMTHt  a 1  a INDi  b1  bCOUNTRYi   t
                                                   60               20



                                               SMTH1                     SMTH2
                                        parameter                 parameter
                                         estimate  p-value         estimate  p-value
                 INTERCEPT                5.395           0.00      5.328           0.00
                 LNMVE                   -0.433           0.00     -0.433           0.00
                 LOG(PRC)                 0.119           0.00      0.119           0.00
                 BM                       0.020           0.10      0.023           0.07
                 LOSS                    -0.121           0.00     -0.119           0.00
                 STD_RET                 -0.943           0.00     -0.936           0.00
                 RINNATE_SMTH            -0.159           0.00     -0.129           0.00
                 REXCESS_SMTH             0.049           0.09      0.139           0.00
                 GOV                     -0.247           0.00     -0.222           0.00
                 REXCESS_SMTH*
                 GOV                     -0.030           0.00     -0.054           0.00


                  Industry Dummies                YES                       YES
                  Country Dummies                 YES                       YES
                                   2
                   Average Adj R                  0.63                      0.63
                      Average n                   5,810                     5,810

GOV is equal to a governance composite, where firms receive one point for each of the following, 1 if they
trade and ADR in the US, 1 if there are followed by more than one analyst, 1 if they report under either
IFRS or US GAAP, 1 if they are audited by a big five auditor and 1 if there closely held ownership is less
than 44 percent. Both analysts following and closely held ownership are based on the sample medians




                                                                                                        48

								
To top