Goodwill Write-Downs, SFAS No 121 and the
Adoption of SFAS No 142
DISSERTATION
Benjamin Segal New York University
Trang 2UMI Number: 3102527 Copyright 2003 by Segal, Benjamin All rights reserved ® UMI UMI Microform 3102527 Copyright 2003 by ProQuest Information and Learning Company
All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code
ProQuest Information and Learning Company 300 North Zeeb Road
P.O Box 1346
Trang 3REV TAMIM SEbAL
L 3
Hereby swear and affirm that no part of the dissertation has been heretofore published, and/or copyrighted in the United States of America, except previously published work and the passages quoted from other published sources; that J am the sole author and proprietor of said dissertation except where a section is clearly labeled as a jointly authored article, that the dissertation contains no matter which, if published, will be libelous or otherwise injurious, or infringe in any way, the copyright of any other party;
and that I will defend, indemnify and hold harmless New York University against all
suits and proceedings which may be brought, and against all claims which may be made against New York University by reason of the publication of said dissertation
` Aro 6€; 2504
Trang 5ACKNOWLEDGMENTS
I am grateful for the advice, guidance, and criticism from my dissertation committee members, William Greene, Joshua Livnat (Chair), Stephen Ryan, and Paul Zarowin, as well as for their interest and insights Useful comments by Ron Lazer and Shai Levi are
Trang 6ABSTRACT
This study examines goodwill write-downs under SFAS No 121 and the effects of SFAS No 142 It contrasts goodwill write-down announcements made before and after this new rule’s provisions came into effect Similar to SFAS No 121 for impairments in general, SFAS No 142 was intended to reduce managerial discretion and to enhance financial reporting for goodwill impairment in specific Two sources of motivation for write- downs are economic impairment and reporting incentives I find significant evidence of reporting incentives playing a role both under SFAS No 121 and under SFAS No 142 In spite of the provisions of SFAS No 142, my results show only partial systematic differences in the reporting incentive variables for goodwill write-downs taken before and after SFAS No 142 I find no such differences for the economic motivation
variables Results of markets tests show a negative market reaction to such write-downs under both rules, and evidence of delayed recognition However, additional tests show no systematic differences in the market reaction between write-downs taken before and after SFAS No 142 There is little evidence that SFAS No 142 has resulted in an
Trang 7TABLE OF CONTENTS FY0.9/690)400)9.6).0061S 1201010578 iii F0 00 iv su vill List of Appendices che TH eeeeeeeeee X I0 0):9)9)6096009)00152 1 2 INSTITUTIONAL SETTING cece cee tecteeesetscsesecesecaeceetseseesensassensenenees 5
2.1 Asset Impairment and Reporting ÍncennfIVeS - - sen 6
2.2 The Signiflcance of Goodwll lmpairmei - chen 8
2.3 A Brief History ofGoodwill Impairment Rules uc nen re 8 2.4 SFASINGO 142— SD€CIÍICS HH HH HH TH Hư Hư thiện 11
3 PRIORLITIERATURE Ặ Sen Ha HH ke 14
3.1 Asset Impairment and Earnings Managemen . cscceeeeerrerrrrreee 14 3.2 _ Asset Impairment and Market EÍÍecfS c.- cành Ha Hhưhaa re 15
4 FORMULATION OF HYPOTHESES AND RESEARCH DESIGN 18
Trang 85 SAMPLE SELECTION AND EMPIRICAL RESULTS 28
›'`'— 28
5.2 Empirical Resul(s vn Hữ HH HH 1H HH 1 11 1g kg 29
li S057 29 An 1 2 32
Comparison of SFAS No 121 and SFAS No 142 eeheeeerre 35
6 SUMMARY AND CONCLUSIONS chua irie 37
Trang 9TABLE 12 — Multivariate Analysis — SFAS No 142 ceherrirree 60 TABLE 13 — Descriptive Statistics - SFAS No l42 cccinereerrrrre 61 TABLE 13 - Descriptive Statistics — SFAS No 142 (ConfL) ccecererie 62 TABLE 14 — Multivariate Analysis — SFAS No 142.0 cc cecceeeseeeeceneeenteneeneeneeentens 63 TABLE 15 — Descriptive Statistlcs — SFAS No 142 ceeerrrrrrrre 64 TABLE 15 - Descriptive Statistics — SFAS No 142 (Conf,) cceieirrie 65 TABLE 16 - Multivariate Analysis - SFAS No 142 ceirreerrdie 66 TABLE 17 — Descriptive Statistlcs - SFAS No 142 eeccenhrrrdre 67 TABLE 18 — Multivariate Analysis — SFAS No 142 eeieerdeirieire 68 TABLE 19 — Multivariate Analysis — Determinants Comparison cccceeeeee 69 TABLE 20 — Multivariate Analysis - Market Reacton Comparison - 71 TABLE 21 — Multivariate Analysis - Market Reaction Comparison (Cont.) 72 TABLE 22 — Multivariate Analysis — Timeliness ComparISOn «ccsiceieằ 73
"se 74
Trang 10List of Tables
TABLE 1 - Summary of S†U1€S nh HH Hà Hà kg HH ghi 40 IV 10052207 =0 7a e 46 TABLE 3 — Descriptive Statistics — SFAS No l2Ì à.2eeeHuhHeHhrrde 47 TABLE 3 - Descriptive Statistics — SFAS No 121 (Cont.) cceeiieeeeree Ag TABLE 4 — Multivariate Analysis —SFAS No 121 occ ccsscceseeeneeenereneessaseees 49 TABLE Š — Descriptive Statistics — SFAS No l2Ì cehhhhrerrreererrire 50 TABLE 5 - Descriptive Statisties — SFAS No 121 (Cont.) eeieieee 51 TABLE 6 - Multivariate Analysis — SFAS No 12Ï che 52 TABLE 7 - Descriptive Statistiles — SFAS No Í2Ì cha Hee 33 TABLE 7 - Descriptive Statistics — SFEAS No 121 (Con(.) iee 54 TABLE 8§ - Multivariate Analysis — SFAS No 12Ì se 55 TABLE 9 - Descriptive Statistlcs — SFEAS No 12Ì che 56 TABLE 10 - Multivariate Analysis — SFAS No 12Ì ccceeeeeerirrererrririe 57 TABLE 11 — Descriptive Statistics — SFAS No l42 eehererie 58 TABLE 11 —- Descriptive Statistics - SFAS No 142 (Cont.) ceeeeiee 59 TABLE 12 — Multivariate Analysis - SFAS No 142 ceeerererrrire 60 TABLE 13 — Descriptive Statistics — SFAS No l42 eerrrrrrrriririie 61
TABLE 13 —- Descriptive Statistics - SFAS No 142 (Cont.) re 62
Trang 131 INTRODUCTION
This study examines goodwill write-downs under SFAS No 121 and the effects of SFAS
No 142 on such write-downs I test their underlying motivation and the market reaction By examining goodwill write-down announcements made before this new rule and comparing them with write-down announcements made under its provisions, I try to
assess the effects and the efficacy of SFAS No.142 in enhancing the reporting for
goodwill impairments
The discretion afforded management in impairment decisions and in their amounts prompted the regulatory changes of SFAS No 121 and its successor for goodwill, SFAS No 142 These rules were intended to reduce managerial discretion and to enhance financial reporting for asset impairment SFAS No 142 has eliminated the amortization of goodwill in favor of impairment testing and, if necessary, write-downs It also has replaced the impairment treatment that was normally governed by SFAS No 121, adding many changes to the impairment testing process
In the short period since the issuance of SFAS No 142, the market has witnessed
goodwill write-downs of unprecedented magnitude” It is questionable whether these accounting entries are valuation-relevant On the other hand, the magnitude of such write-downs and their impact on firms’ financial reports are overwhelming I test the market reaction to such write-downs and examine whether it has changed due to the new
' See the later reference to APB No 17 Throughout the paper, write-downs taken in the period before
Trang 14rule My work also provides evidence on the extent to which economic impairment and reporting incentives are the motivation behind goodwill write-downs Economic
impairment is proxied by variables that capture poor historical firm or industry
performance Reporting incentives are proxied by variables that capture management’s interests with respect to compensation or financial reporting.”
Whether and to what extent SFAS No 142 enhances the reporting for goodwill impairment, and whether the market views such write-downs differently, are open
questions If criticism of SFAS No 121’s flexible provisions is justified and if SFAS No 142 reduces discretion with its detailed and specified provisions, then I expect the
following: I should find not only significant evidence of reporting incentives at work in the SFAS No 121 regime, but also significantly less evidence of reporting incentives under the new SFAS No 142 Similarly, I should find an increased association with economic impairment for the new rule
In addition, the market reaction may reveal whether these write-downs are viewed
differently I expect the market reaction to goodwill write-downs to be negative Thus, if SFAS No 142’s detailed provisions effectively reduce discretion (compared with the lax provisions of SFAS No 121), then I expect the market reaction to be even more negative Finally, I examine the association of goodwill write-downs with lagged returns for delayed impairment recognition SFAS No.142 will probably improve the timeliness of
* A few examples from the Wall Street Journal are described in § 2
Trang 15goodwill impairments in the future, due to its specific scheduled testing guidelines In the short term, however, its more sensitive impairment trigger may prompt write-downs associated with substantially earlier periods (“catching up’) I therefore expect to find evidence of association with larger lags under SFAS No 142
The analyses produce several interesting results I find that both types of variables, economic and incentive, are important Reporting incentive variables are significant both under SFAS No 121 and under SFAS No 142 I do find some differences between the regimes — providing evidence that the new rule (at least in its initial adoption) has
reduced reporting flexibility in some respects but not in others Results from market tests
show a significant negative reaction to goodwill write-downs under both the previous and
the new regulatory regime, and the reaction under SFAS No 142 is smaller rather than
larger in magnitude This is inconsistent with the notion that these amounts better reflect
economic impairment This difference in the negative market reaction, however, is not
significant Results for the write-down timeliness tests show some weak evidence of the
delay in recognition of goodwill impairment under both SFAS No 121 and 142, with some indication of an increased delay under the latter
As a whole, these findings strengthen the evidence and extend the literature on the
association between discretionary write-downs and reporting incentives under SFAS No
121 The findings also provide evidence on the effects and the efficacy of SFAS No.142 I conclude that despite minor indications of a reduced association with reporting
Trang 16goodwill impairments Goodwill write-down characteristics have not significantly
changed, and continue to be significantly associated with reporting incentives
Trang 172 INSTITUTIONAL SETTING
In general, an asset is said to be impaired if its recoverable value falls below its book
value.’ Upon recognition of such impairment, a company decreases (writes down) the book value of the asset by the difference and records the expense in its income statement for that period.” The accounting for write-downs of goodwill, part of long-lived assets, has long been a target of interest, criticism, and consequently legislation Prior literature has identified an increase in the magnitude and in the occurrence of asset write-downs over the past few decades (see, e.g., Strong and Meyer [1987], Elliott and Hanna [1996]) Such asset impairments can average 20% of total assets, with goodwill accounting for the lion's share (Francis et al [1996], Henning et al [2002]) Goodwill write-downs are part of what the literature refers to as "discretionary write-offs" — situations in which little authoritative guidance on accounting exists or the discretion afforded to managers is large.° A few anecdotal examples of SFAS No 142 write-downs follow:
AOL Posts Record Loss on Charge, Lowers Earnings Outlook for Year
The nation's largest media company reported a record quarterly loss of $54.24 billion for the first quarter, or $12.25 a share, largely because of a $54 billion charge to write down the value of assets The charge, which had been expected, reflects new accounting
rules on goodwill and marks one of the largest write-offs in U.S corporate history
The Wall Street Journal, April 24, 2002
4 Pair market values and discounted versus undiscounted cash flows are some of the changes in SFAS No 142
> One of the exceptions to income statement expensing is the write-down of long-term equity investments ° The literature generally looks at write-downs of long-lived assets Current assets such as inventories and
Trang 18USA Networks Inc said its first-quarter loss widened, due in part to a $310.6 million write-down of "impaired" goodwill related to its Citysearch online listing and Precision Response customer-service divisions
The Wall Street Journal, April 25, 2002
Aetna Inc showed the first solid signs of a turnaround with its first-quarter results, cheering investors, who overlooked a $2.97 billion charge due to an accounting change
The Wall Street Journal, April 26, 2002 The uses of write-downs to manage earnings and the shortcomings of previous rules governing asset impairment have been discussed in prior literature (e.g., Kinney and Trezevant [1997], Rees et al [1996], Riedl [2002}) The difficulty in valuing long-lived assets, the impact their write-down has on a firm’s financials, and the large discretion involved in making the impairment decision make such write-downs likely candidates for earnings management and an interesting setting in which to examine firm and market
behavior This would seem to hold even more for intangible assets and, specifically, for
goodwill
2.1 Asset Impairment and Reporting Incentives
Trang 19for substantial discretion and flexibility with respect to the magnitude and timing of impairment of long-lived assets The combination of magnitude and discretion has led researchers to believe that firms may make strategic use of these write-downs (see, €.g Kinney and Trezavant [1997], Zucca and Campbell [1992]) Management may recognize a larger impairment than is warranted, or recognize it prematurely Conversely, although write-downs are always negative, management may take a smaller (than warranted) impairment, delay taking the write-down, or even avoid it altogether, giving itself control over the accounting impact Some of the incentives for such behavior may be driven by compensation, or even by the desire to reduce the company’s cost of capital Thus, for example, when managers’ compensation is tied to earnings and has an upper bound (cap) or a threshold (Murphy [1999]), they may engage in income smoothing or an earnings bath, potentially improving their future performance and compensation Alternatively, managers may attempt to convey information or simply delay negative market reactions with such behavior.’ In fact, Strong and Meyer [1987] find that a change in senior management is the most important determinant of a write-down decision This phenomenon has not eluded the regulators, as is evident from a speech by Lynn E
Turner, then Chief Accountant at the SEC, given on May 18, 2001:
" _ In certain instances, the staff has noted that the triggering event resulting in an impairment charge appears to be a change in senior management The staff has
questioned the timing and appropriateness of impairment charges recorded at the same time as a change in senior management In particular, the staff has inquired as to the
Trang 20
underlying changes in the business and its economics, when those changes occurred, and whether those changes have been appropriately disclosed in MD&A when they occurred, ona timely basis "
2.2 The Significance of Goodwill Impairment
Goodwill seems to be a major item in many asset write-downs For example, Francis et al [1996] state: “/G/oodwill write-offs are among the most substantial write-offs in our sample they are the largest write-offs in terms of dollar per share and as a percentage of total assets these write-offs reduced goodwill by two thirds ” (p 121) Goodwill is a very large item on many balance sheets of companies that engage in acquisitions, and its impairment may have a substantial impact on the financial reports In a recent study, Henning et al [2002] find in their sample an average goodwill balance of close to $200 million, representing about 20% of the firms’ total reported assets They report an average goodwill write-down of $95 million, representing close to 32% of the original transaction purchase price
23 A Brief History of Goodwill Impairment Rules
Trang 21consistent manner.’ APB Opinion No 17, Intangible Assets, issued in 1970, addressed the accounting for goodwill It viewed goodwill as associated with the enterprise as a whole Accordingly, impairment testing would be done only at the enterprise level In addition, it did not provide guidance for determining when or how to measure
impairment This broad definition allowed for substantial discretion in deciding on write- downs of long-lived assets and on their amounts As stated in SFAS No 121:
“Accounting standards generally have not addressed when impairment losses should be
recognized or how impairment losses should be measured As a result, practice has been
diverse.”
SFAS No 121
SFAS No 121 was intended in part to address the problems with long-lived asset
impairment (including goodwill) It provided a clearer (though still general) prompt for review of assets — whenever events or changes in circumstances indicated that their catrying amount might not be recoverable The standard addressed goodwill that related to specific assets Nevertheless, there is some indication that it was not very successful In fact, the wording of SFAS No 121 leaves management's discretion as the deciding factor in determining:
5 Another undesirable effect was that preparers of financial reports relied on pieces of regulation that were not intended for that purpose or that did not cover the entire subject FASB Statement of Financial
Accounting Concepts No 5, Recognition and Measurement in Financial Statements of Business
Enterprises (Con 5), issued in 1984, gave some general guidance with respect to loss recognition with the
reductions of an asset’s future benefits Yet statements of concepts were not intended to establish a
reporting standard, but rather a basis for detailed standards Thus, Con 5 was too vague and general and led to inconsistent accounting treatment
Trang 22a Whether a write-down should be considered — “ be reviewed for impairment whenever events or changes in circumstances indicate 19
b Whether a write-down is necessary and how much it should be, through
management’s own estimate of future cash flows and eventual asset disposition Recent research on SFAS No 121 seems to support the notion that the rule did not succeed in enhancing the reporting for long-lived asset impairments in general (Ried!
[2002]) As stated in SFAS No 142: “Previous standards provided little guidance about
how to determine and measure goodwill impairment; as a result, the accounting for goodwill impairments was not consistent and not comparable and yielded information of questionable usefulness.”
SFAS No 142
The new SFAS No 142 has drastically changed the timing of the test for impairment and its mechanics It is more specific on three main issues: when to test for impairment, what level of reporting unit to test, and how to determine the amount of the impairment While
SFAS No 121 required testing only in certain circumstances, SFAS No 142 calls for
goodwill to be tested for impairment at least annually, and under certain enumerated circumstances The new standard also requires that the test be performed at the same time every year, blocking any attempt to time the test so as to affect its results The new impairment testing is to be applied at the reporting unit level This lower level of testing
reduces the possibility of overlooking impaired goodwill due to setting off between
business units (essentially, unrecognized “internally developed” goodwill from one part
© Page 1, summary, second paragraph The statement goes on to describe a few examples of such events
Trang 23of the business can offset impaired goodwill from another) The trigger of the new SFAS No 142 is more sensitive (compared to undiscounted cash flows in SFAS No 121) Thus, impairment amounts are expected to be large upon adoption due to the “catching up” of reporting, but smaller and more frequent in subsequent periods, thanks to the annual (or even more frequent) review requirement SFAS No 142 gives specific
guidance and a methodology for determining goodwill impairment, by estimating the fair value of a reporting unit It also calls for professional valuation experts for most of its application
2.4 SEAS No 142 — Specifics
The new rule claims to help users to understand and evaluate the performance of those assets arising from past investments and to assess future profitability It also claims to
"better reflect the underlying economics of those assets." It seems reasonable to assume that the stricter and more detailed guidance of SFAS No 142 will enhance the reporting for goodwill impairment Timely recognition of impairment amounts should also make financial reporting more uniform, more comparable, and more informative Yet, as the following discussion shows, some factors may belie, or at least weaken, the new rule’s promises
Trang 24the same goodwill to a profitable or undervalued unit (with internally generated goodwill that can provide a cushion) will guarantee that no impairment will have to be recognized Second, the amount of impairment is recognized in financial reports as an effect of a change in accounting principles.” Thus, the negative effect of large or even excessive write-downs is pushed below the line and labeled as an accounting change, encouraging higher write-down amounts The rationale behind this argument is that write-downs of a significant magnitude lower the book value of the company and will improve its future performance ratios like ROA and ROE (assuming positive net income) The negative
impact of such write-downs on net income would arguably be reduced or even eliminated if their effect is recorded “below the line” (as the effect of a change in accounting
principles) In addition, companies were afforded six months from the initial adoption of the new rule to determine whether any impairment was needed, and then had another six months to announce the actual amount
Discretion in the application of the measurement procedure: When performing the fair value measurement of a reporting unit, management has a wide array of techniques from which to choose Management may come up with the estimates used in the calculation
and set the assumptions underlying the valuation (SFAS No 142, paragraphs 23-25)
In summary, the new SFAS No 142 may very well not enhance the reporting for goodwill impairment in either the short or the long run The discretion described above may lead to amounts that do not reflect the economic reality of the firm The new rule
Trang 25
may provide management with justification for write down amounts that could not otherwise be taken In this scenario, management may appear to adhere strictly to the
letter of the law (rather than to its spirit) and use that to reach different impairment
amounts.'? With respect to the adoption period, the write-downs may be inflated and make such “transitional period write-downs” misrepresent the underlying economics of
the firm Thus, the effects of SFAS No 142 on the characteristics of impairments of
goodwill currently on the balance sheets, as well as goodwill that will result from future acquisitions, are not clear
Trang 26
3 PRIOR LITERATURE
In reviewing the write-down literature, it may be helpful to use the framework described
in Figure I The major determinants of asset impairment and the market reaction are its
initial reason or driver (adverse economic conditions), its strategic nature (in the market’s eye), its timeliness (with respect to the adverse economic conditions), and the extent to
which the market expected the write-down amount For the most part, prior literature has
used entire write-down amounts to test for the market's reaction, implicitly assuming that the expected write-down amount was zero Table 1 summarizes some of the main
literature on write-downs relevant to this paper A more detailed discussion of the most pertinent articles follows
3.1 Asset Impairment and Earnings Management
Trang 27type of asset written off Incentives play a small role in inventory and PP&E write-down decisions, but a substantial role in more discretionary items like goodwill write-downs and restructuring charges Specifically, they find that performance measures that proxy for "smoothing" or "earnings bath" behaviors are significant in explaining goodwill
write-downs
3.2 Asset Impairment and Market Effects
The results of studies of the market effects of asset write-downs are generally mixed For
example, Strong and Meyer [1987] report a positive average cumulative abnormal return (CAR) before the announcement period, while finding a negative average CAR for the announcement period itself Their results, however, reversed over periods following the announcement They interpret their results as consistent with an anticipated write-down and restructuring charges that may be perceived as insufficient They also find what they believe is evidence that "the bigger the bath [is], the better." Elliott and Shaw [1988], on the other hand, conclude that the opposite is true and that write-downs occur in periods of sustained economic difficulty Zucca and Campbell [1992] find no significant market reaction in the 120 days surrounding the write-down announcement They also find that firms with larger ratios of write-downs to total assets experienced lower financial strength in subsequent periods, refuting the notion that firms that clean their books outperform later
Trang 28are not timely (finding a higher correlation of the write-down amounts with lagged returns)
Prior research generally has not looked at or tested for specific types of write-down announcements to examine their effect Francis et al [1996] are an exception; they find that the market reaction to write-downs is overall negative, but it is conditional on the type of asset written down They observe a negative reaction to inventory write-downs
(perceived as declines in economic value), but a positive reaction to restructuring charges (perceived as signals for future performance) They do not observe a significant market reaction to goodwill write-offs (page 133) In addition, information content studies typically use the entire write-down amount, implicitly assuming the expected write-down amount is zero The extent to which the write down amount is expected and anticipated by the market is important not only for the model specification and for measurement, but also for interpreting these results
The above discussion leads to the following questions: To what extent are goodwill write-downs driven by economic factors rather than reporting incentives? What effect (if any) has SFAS No 142 had on the characteristics of such write-downs? Has SFAS No 142 (so far) enhanced the reporting for goodwill impairments? How does the market
view goodwill impairments? What is the effect of the unexpected portion of the write-
down?
Trang 304, FORMULATION OF HYPOTHESES AND RESEARCH DESIGN
The questions raised above yield a useful categorization of the research questions under four major economic issues: the characteristics of goodwill write-downs, the
characteristics of write-down firms, the association between write-downs and changes in a firm's market value, and the timeliness of recording these write-downs
4.1 Determinants of Goodwill Write-Downs Hypotheses 1 2
Sections 2.3 and 2.4 discussed in detail the major differences between SFAS No 121 and
SFAS No 142 and their respective shortcomings The large discretion afforded by the
old rule, SFAS No 121, is expected to result in write-downs that are more strongly associated with reporting incentive variables'* than those taken under SFAS No 142 Similarly, the detailed methodology and increased scrutiny called for by SFAS No 142 are expected to result in write-downs that are more strongly associated with economic variables
The following hypotheses assess the extent to which write-downs under the different rules are driven by economic factors (proxied by poor historical firm performance and
‘4 1t may be that "bath" and smoothing considerations are less relevant under SFAS No 142 because write- downs upon adoption are below the line and most likely would not affect compensation (Gaver and Gaver
[1998])
Trang 31declining industry trends) rather than by reporting incentives (the effect such write- downs have on the financial reports):
Hi: The association b/w goodwill write-downs and economic factors is lower under the SFAS No 121 regime than under the SFAS No 142 regime H2: The association b/w goodwill write-downs and reporting incentives is higher
under the SFAS No 121 regime than under the SFAS No 142 regime
Poor historical firm performance and declining industry trends will be used to proxy for economic factors (e.g., changes in sales and in industry ROA).) Incentives and
disincentives for write-down decisions include leverage (debt covenants may constrain
the size of the possible write-down), changes in top management, and earnings bath or smoothing behavior.'° A description of the variables and additional explanations are given in the notes below and on the following page The first equation estimates the regression coefficients on each subgroup (SFAS No 121 and SFAS No 142) separately, while the second contains an interaction term to determine whether any differences between the groups are significant
1Š Beonomic Proxies: These variables proxy for poor historical firm performance and declining industry
trends
Trang 32()
WD, = Oy + ay
+a_,-Debt! Assets , +a
ASALES, pt Oy" AINDROA, 1+ % 4 ‘GOOD pt oq: POOR, 11 Be" SIZE, ri AM +a ‘AManagement,, 6 t (2) WD =a +a_-ASALES +a.-AINDROA, +a,-AMV_ +a, -AManagement it 0 1 it 2 it 3 it 4 it +a_-Debt/ Assets : +a 1 5 6 7 8 F142: : L : ` :
+ [By + By ASA ES +B, AINDROA +B, AMY +B, AManagement
‘GOOD +a,,-POOR_ +a,°SIZE_
lí it it
Debt / A -GO -POOR - SIZE
+B, e ssets +Be OD + Ba it + Be Ss alt ei
The write-down amount is expressed as a positive number throughout the paper, so the expected coefficient sign is negative for ASALES, AINDROA, and AMV (negative changes in these variables due to economic weakness are expected to yield higher write- downs) The expected coefficient sign is also negative for Debt/Assets (the amount of the write-down may be limited by debt covenants) The expected coefficient sign is positive for AManagement and GOOD (higher unexpected earnings could lead to a higher write-
" Notes: Definitions of Variables
Similar to Francis, Hanna and Vincent (1996), Riedl (2002), and others
F142 Dummy variable that equals 1 if impairment.is taken under SFAS No 142, zero otherwise WD Total pre-tax write-down amount divided by beginning of period assets
AINDROA Change in industry median ROA over the two years prior to the quarter of the write-down
announcement
ASALES Change in sales over the two years prior to the quarter of the write-down
announcement, scaled by beginning of quarter total assets
AManagement Equals | if the firm had a change in CEO in the year of the write-down announcement or or it, zero otherwise
Debt/Assets Beginning of period total long-term and short-term debt divided by total assets
GOOD Equals (post-write-down) unexpected earnings if positive, zero otherwise (calculated as the
difference between current and four-quarters-ago earnings, divided by beginning of quarter
assets)
POOR Equals (pre-write-down) unexpected earnings if negative, zero otherwise (calculated as the difference between current and four-quarters-ago earnings, divided by beginning of quarter assets)
Trang 33down amount, under a smoothing argument) and negative for POOR (unexpectedly lower earnings should lead to a higher write-down amount, under a big-bath argument) Thus, for H1, since the coefficients for the three economic proxies (ASales AIndroa, and AMV) are expected to be negative, the expected sign for the interaction coefficients is expected to be negative, if indeed write-downs under SFAS No 142 are more associated with economic determinants Similarly, for H2, the expected sign for the interaction
coefficients is expected to be positive for Debt/Asset and POOR, but negative for
AManagement and GOOD
4.2 The Market Reaction to Goodwill Write-Downs
Francis et al [1996], in their tests of the market reaction to goodwill impairment, find a positive but not significant response This may be due to the measurement error
introduced by using the entire amount of the write-down (the expected amount is implicitly zero) Additionally, it may be attributed to the difficulty in controlling for
additional information in the write-down announcement Thus, it is difficult to predict the
direction and magnitude of the market reaction that follows goodwill impairment announcements It is more likely, however, that goodwill write-downs are perceived as negative news, reflecting the loss of economic value due to devaluation of assets | expect that after controlling for positive information in the write-down itself (i.e., if it is associated with a strategic action) and in the announcement, the actual goodwill
Trang 34information as early as two yeas in advance Thus, the expected association may be insignificant or zero
Hypothesis 3
By the FASB’s own admission (as stated in the notes to SFAS No 142), impairments under SFAS No 121 are potentially more discretionary (or more driven by reporting incentives) and less economic The opposite should hold for SFAS No 142 Prior literature has identified a difference between the market reaction to asset write-downs ranging from negative (e.g., for inventory) to positive (e.g., for restructuring charges) The direction and magnitude of the market reaction have generally been attributed to the nature of the write-down and the information it conveys Thus, restructuring charges are perceived as more discretionary They are perceived as strategic in nature, and as conveying information on potential improved future performance Similarly, due to the additional flexibility afforded under SFAS No 121,'7 I expect goodwill write-downs to be associated with a less negative market reaction than those under SFAS No 142
(which should be more strongly associated with economic variables) To control for other information in the announcement, I use an earnings surprise measure, change in earnings I construct an additional variable, PositiveNews, that captures other information in the
7 All this depends on the way the market perceives these discretionary write-downs and on their strategic nature and timeliness Theoretically, the market reaction to such goodwill impairments may be positive if they are perceived as signaling It may be positive or negative if they are perceived as earnings
Trang 35announcement — mainly whether the write-down was strategic or operational in nature, or whether a concurrent earnings/sales forecast was made.'®
H3: The market reaction to goodwill write-downs under SFAS No 121 is less negative than under SFAS No 142
(3)
R,=7, +7, EWD, + 7,-RWD, +7, -AEARN, + 7, + PositiveNews, +7, ‘Size; + HU,
(4)
R,=7,+7,:EWD, +7,-RWD, +7,-MEARN, +7, + PositiveNevs, + y, Size, + F142-[6, +0,-EWD, +0, -RWD, +0, -AEARN, +, : PositiveNevs, + 0, - Size, |+ y;
ũ In general, prior studies did not control for the expected portion of the write-down, implicitly assuming that the entire written down amount was a surprise I attempt to
18 DositiveNews captures write-downs that were considered strategic (e.g., closing an operating unit, restructuring) and write-downs that were announced with a concurrent forecast or other major positive events I have also coded negative concurrent news, but the number of such observations is negligible H Notes: Definitions of Varlables
F142 Dummy variable that equals 1 if impairment is taken under SFAS No.142, zero otherwise
Rj Compound excess return for firm i over the event period (measured as the difference
between the realized return and the S&P 500 return over the three days surrounding the
announcement date)
EWD* Amount of the expected write-down measured using a market value metric (scaled by
market value at beginning of period) — see details in Appendix A
RWD* Residual write-down — the difference between the actual write-down amount and the
expected write-down amount (scaled by the market value at the beginning of the period)
PositiveNews Dummy variable that equals 1 if the announcement contains positive information
beyond earnings, zero otherwise
AEarnings Pre-write-down earnings surprise in the announcement (calculated as the difference between the announced earnings and four-quarters-ago earnings)
Trang 36improve on that point by estimating the expected goodwill impairment amount and controlling for it in the market tests Thus, for H3, the coefficient for the expected write-
down, EWD, is expected to be zero I attempt to control for other information in the
announcement (other than in earnings) with the variable PositiveNews This dummy variable codes any major additional positive information conveyed in the announcement, including the nature of the write-down (whether or not the write-down has strategic or operational implications), forecast revisions, or major new contracts Its coefficient sign is expected to be positive, and more so for SFAS No 121 The reason for this
expectation is that with respect to the strategic portion of this variable, write-downs under SFAS No 142 should, in theory, be less voluntary and should represent (more) real value loss for the firm, with less potential for future benefits I therefore also anticipate a lower occurrence of such strategic characteristics for write-downs under
SFAS No 142 The coefficient sign for the residual write-down, RWD, which is the
unexpected portion of the write-down, should be negative I expect the difference
between the coefficients (the interaction coefficient) to be negative if indeed write-downs
Trang 374.3 Timeliness
Hypothesis 4
To begin with, the literature has addressed the issue of timing asset impairments strategically to manage earnings (see, e.g., Kinney and Trezevant [1997], Gore et al [1996}) Thus, as some previous evidence has shown, write-downs under SFAS No 121 are expected to show some evidence of delayed recognition SFAS No 142 is expected to significantly improve the timeliness of write-downs thanks to its specific scheduled testing provisions, although this may not be the case during the adoption stage The change in regulation could lead to impairments that otherwise might not have been recognized at this time For one thing, the fact that future cash flows will now be
discounted when determining impairment can trigger previously unrecognized write- downs In addition, goodwill may be subject to increased scrutiny These may lead to "catching up" of the reporting Alternatively, managers may use the adoption period to write-down larger amounts than warranted In such a case, we may find little association with any specific prior period.” Thus, although in both periods I expect to find evidence
of delayed write-downs, I expect it to be more evident for write-downs under the new
rule In this context I will examine the differences between the four quarters preceding the write-down announcement and the four quarters prior to that
Trang 38
H4: Goodwill write-downs are more timely under SFAS No 121 than under SFAS No 142 (during the adoption stage) (5) WD, =ơy +ơi -R4Q„¡ +ø; - RÀO, ; +£, (6) WD , =a, +@,-R40, +0, R40 4-5 + F142 -[B,+ B,-R40,1+ B “R40, 5) + € 4
A significant coefficient for prior periods’ returns will indicate the extent of a delay in recognizing the impairment I consider a one or two-quarter delay as normal and attribute it to the nature of the accounting process, and thus should be captured within the four- month period Longer lags may indicate a more intentional delay in recognition or, in the case of SFAS No 142, a trigger of prior events Prior literature has identified the
phenomenon that write-downs overwhelmingly occur in the fourth quarter for most companies (see, e.g., Kinney and Trezevant [1997]), which by itself may be an indication of some delay (the argument that write-downs are mostly done at the end of the fiscal year to delay recognition or to manage yearly earnings when most information is in is often met with the point that greater scrutiny is applied in the fourth quarter due to the audit) Under this setting, | compare between first and second four-quarter lags
ul Notes: Definitions of Variables
F142; Dummy variable that equals | if impairment is taken under SFAS No 142, zero
otherwise
WD: Write-down amount (scaled by market capitalization at the end of period t-1) Ret4Qt, Rụ The return of firm j over four quarters starting t-1, where t is the quarter of the
Trang 39According to H4 I expect to find a more negative reaction to write-downs in the first lag for SFAS No 121 and in the second lag for SFAS No 142
4.4 Estimating the Expected Impairment
Two models” are used to approximate the expected amount of goodwill impairment warranted by the economic reality The expected amounts are used in the tests for
reporting incentives and the information content of the write-downs.”
The first model is a basic "market-based valuation" with several simplifying
assumptions, aimed at estimating the overall value of the company's goodwill This model is similar to what some analysts use to approximate impairments and does not follow the spirit of the rule.”* In essence, it looks at goodwill at the enterprise level only and uses the decline in the stock price to determine and calculate its potential
impairment
The second method is a variation on the previous market-based method Since the drop in the market price may not provide the best indication for impairment (e.g., in volatile
markets, or when this price drop does not correspond to the value of the reporting unit
with which the goodwill is associated), I use residual income valuation to assess an alternate value of the firm and its goodwill
20 third method used is an "industry-based valuation" in which a determination of the "origin" of the
goodwill is made, and its impairment is assessed by the change in the relevant industry market value index from the time of the original transaction For that purpose, historical purchase transactions were examined Their total amount (value) was multiplied by the decrease in the relevant industry index This amount is the expected impairment of goodwill up to the amount of goodwill recorded in the specific transaction (after amortization) This examination only yielded 29 companies Though this subsample was too small for full analysis, comparison with estimation method B yielded no significant differences (see Appendices B, C) 21 Additional details on the estimation methods can be found in Appendix A
Trang 405 SAMPLE SELECTION AND EMPIRICAL RESULTS
5.1 Data
Table 2 describes the sample selection A search in the Dow Jones Interactive Press Release database for the period from July 2001 to December 2002 for firms that announced goodwill write-downs and impairments yielded over 750 results After deleting announcements not containing write-downs or amounts, firms with no available Compustat data, write-downs previously announced, and non-U.S firms, I was left with 221 announcements Of these, 102 relate to SFAS No 121, and 119 relate to SFAS No.142 For these companies I identified the closest previous quarter ending prior to or
on the date of the announcement and obtained the Compustat data items for that quarter and lagged periods as needed, as well as returns around the announcement date
Information about CEO changes for 2000-2002 was obtained from Execucomp Two hundred seven observations with available data were used in the write-down