Intraindustry information transfers associated with certain unfavorable audit opinion announcements

... current research is w hether certain unfavorable audit opinion announcem ents, specifically the "subject to" qualified audit opinion and the "going con­ cern" audit opinion, contain useful inform... qualified audit opinion has been replaced by unqualified opinions w ith special language T he unqualified audit opinions with special language receiving attention in this study are those with "going... "subject to" qualified audit opinion or the "going concern" audit opinion This hypothesis is stated as follows: Hn: "Subject to" qualified and "going concern" audit opinions are not associated w ith

INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type o f computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IN T R A -IN D U ST R Y IN FO R M A TIO N TR A N SFER S A SSO C IA T E D W ITH C ER TA IN U N FA V O R A B LE A U D IT O PIN IO N A N N O U N C E M E N T S By M ark Schaub A Dissertation Subm itted to the Faculty o f M ississippi State U niversity in Partial Fulfillm ent o f the R equirem ents for the Degree o f D octor o f B usiness A dm inistration in Business A dm inistration in the D epartm ent o f Finance and Econom ics in the C ollege o f B usiness and Industry M ississippi State. M ississippi M ay 1998 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. UMI Number: 9829804 UMI Microform 9829804 Copyright 1998, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . IN T R A -IN D U S T R Y IN FO R M A TIO N T R A N S F E R S A S S O C IA T E D W ITH C E R T A IN U N F A V O R A B L E A U D IT O P IN IO N A N N O U N C E M E N T S By Mark Schaub Approved: R(D A T :H I-M T A :B C y c :M V IB V :E x :E A :AC A R (- u h C O R R ) (10) R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 92 C om petitive E ffect: ± + + - + + + - + *C A R t- U) = R(D A T ; HI: In TA:BCyc: MVIBV: E x; E A ; a C A R ^ A): C O RR) (1!) where: = the abnorm al return to the rival firm during the event w indow , DAT = the book debt to total assets ratio o f the rival firm, HI = the value o f the Herfindahl Index for the industry. In TA = the natural log o f the total assets o f the firm. BC yc = 1 during econom ic contractionary periods at the event date. 0 otherw ise, M V IB V = the m arket to book value o f equity ratio o f the firm. Ex 1 i f rival firm i is traded on the N A S D A Q , 0 is the rival firm is traded on the N Y S E or the A M EX . EA 1 i f the announcing firm concurrently announced earnings. 0 otherw ise. AC'AR (-I.D = the cum ulative abnorm al return to the announcing firm d uring the event window, and C O RR - the returns correlation betw een the announcing firm and its rival. The signs over each variable show the expected direction o f the relationship be­ tw een each variable and the respective intra-industry inform ation effect. V ariables w hich are denoted with ± o v e r them are those for which am biguity exists as to the direc­ tional relationship betw een those variables and the respective effect represented. The R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 93 reasons for selecting the above m entioned contributing factors, the m anner in w hich they are com puted and the justification for the expected directional relationship betw een each contributing factor and the com petitive and/or contagion effect are discussed in the next sections. A lso, the problem o f m ulticollinearity is addressed later in the chapter. Leverage T he variable used to m easure the degree to w hich a rival firm is levered is the debt-to-total assets ratio. A lternative m easures o f leverage w ould give consistent results as the debt-to-total assets ratio, since these m easures provide the exact sam e inform a­ tion. The alternatives include the equity m ultiplier, w hich is com puted as total assets divided by equity, and the debt-to-equity ratio, com puted as total debt divided by total equity. The relationship am ong these variables is as follow s: (1) the equity m ultiplier is equal to one plus the debt-to-equity ratio, and (2) the debt-to-total assets ratio is equal to the debt-to-equity ratio divided by the equity m ultiplier. Since know ing the value o f any o f these ratios given the value o f one o f them is m erely an algebraic transform ation, the overall results are insensitive to the m easure used (num erical regression coefficient m agnitudes may change, but not the sign or the significance). The debt-to-total assets ratio is chosen only because o f the sim plicity o f its inter­ pretation (the percent o f the firm s assets w hich are financed w ith debt). T his variable w as retrieved for each rival firm from the C O M P U S T A T database for 1996. Leverage, as a m easure, is o f im portance to this study because equity absorbs the change in the m arket value o f total assets. For this reason, the m ore highly levered a firm, the greater R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . as a percentage is the change in the m arket value o f the firm 's equity resulting from a change in the value o f the firm 's assets. Lang and Stulz [29] hypothesized that since the contagion effect affects a firm ’s total value (via the reduction in share price), the fall in the value o f the rival firm s' eq­ uity will partially depend on how highly levered the rivals are. Such is the case because leverage increases the elasticity o f the value o f equity with respect to the total value o f the firm and the present value o f direct bankruptcy costs, since a decline in the value o f the firm increases the probability o f bankruptcy. Therefore, the degree to w hich the ri­ val firm is levered can be expected to have a larger negative im pact on the rival firm 's percentage change in equity value when the adverse audit opinion announcem ent results in a contagion effect because the greater the degree o f leverage to a negatively affected firm, the greater the negative m agnitude o f its percent equity adjustm ent. T he case o f the effect o f leverage when a com petitive effect occurs is m ore com ­ plex because, while the greater the leverage, the greater elasticity o f the equity value to cash flows should increase the positive stock price reaction by rivals, at the sam e tim e, too m uch leverage m ay reduce the ability o f a rival to take advantage o f their new co m ­ petitive position. Bolton and Scharfstein [35] m odel the abilities o f firm s w ith little debt to "prey" on firm s w hich are m ore highly levered and. thus, have less flexibility to respond to the new favorable m arket conditions (essentially, the transfer o f future cash flows from the announcing firm to its rivals). Therefore, since leverage m ay potentially produce opposing equity adjustm ents, the am ount o f leverage can be expected to have R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 95 an am biguous effect on the net percentage change in equity value when a com petitive effect exists. This w as found to be the case by Lang and Stulz [29] based on their inves­ tigation into intra-industry effects arising from bankruptcy announcem ents. Degree o f C om petition The H erfindahl Index is a m easure o f how com petitive an industry is. It is co m ­ puted as the sum o f the squared levels or percentages o f industry sales held by each firm in the industry. Essentially, the higher the value o f the Herfindahl Index, the m ore co n ­ centrated the sales in that industry. O ligopolistic industries, in w hich a few firm s a c ­ count for a m ajority o f industry sales, are characterized by high H erfindahl Index values. Low Herfindahl Index values are associated with less concentrated industries, which tend to have a large num ber o f firms, each with a sm all percentage m arket share. Essentially, firms in industries w ith high Herfindahl Indexes fight for m arket share, m aking the firm s inter-dependent upon one another. Industries in w hich sales are d is­ persed am ong m any sm all firm s have little com petition for m arket share, causing the firms to be m ore independent o f each other. T his variable w as com puted using sales figures reported for the industry rivals o f each announcing firm by C O M PU ST A T . w hether or not the firm was included in the re­ gression. The rivals are defined as those firms in the CRSP data set w hich had the sam e SIC code as the announcing firm and had data available for the event period. The sales figures for each firm w ere sum m ed to obtain an industry total. Then the percentage o f R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 96 sales p er firm w ere com puted, squared and added together to obtain the value o f the H erfindahl Index. The degree o f com petition, as m easured by the Herfindahl Index, is expected to im pact the com petitive effect o f a bankruptcy (or, in this case, a m ore probable b ank­ ruptcy) announcem ent. A ccording to Lang and Stulz [29], an im perfectly com petitive (i.e., not perfect com petition) industry should have a higher Herfindahl Index, w hich should result in a greater stock price reaction for non-announcing firms to events w hich are firm -specific to the announcing firm. However, the industry concentration m ay not im pact returns to industry rivals w henever the announcing firm provides a release o f new . industry-w ide inform ation (the contagion effect), since there is no transfer o f prof­ itability from the announcing firm to its rivals. Therefore, the degree o f industry con­ centration should positively im pact the com petitive effect, but have an am biguous im pact w hen a contagion effect is present. Size o f Rival Firms D ifferent variables m ay capture the size effect o f a rival firm. In this study, the natural log o f total assets is used as a proxy for firm size, although sales or m arket capi­ talization may also have been used. The value o f total assets for each rival firm is re­ trieved from the C O M P U S T A T data source. The natural logarithm o f total assets is then taken as a m eans o f norm alizing the variable. The justification for using a variable to proxy for the size o f rival firm s is explained below. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 97 Studies by A tiase [36] and G rant [37] both show that the m arket reaction to earnings announcem ents differs for firm s o f different size. Both theorize that the size o f a firm will capture inform ation effects because m ore sources o f external inform ation are associated with firms o f larger size as com pared to those firms w hich are sm aller. In keeping with this theoretical association betw een size and available public inform ation, the size o f a nonannouncing firm should also be a significant determ inant o f the m ag n i­ tude o f abnorm al returns to those firm s w henever the m arket reacts to new inform ation provided by unfavorable audit opinion announcem ents. The study perform ed by G raham and K ing [38] has show n that the size o f a firm can be used as a representative m easure o f that firm 's "inform ation environm ent" w hen dealing with intra-industry inform ational transfers. In their study, sm aller firm s are be­ lieved to have a greater reaction to a contagion effect because less inform ation is g en er­ ally available about sm aller firm s to investors. A lso, in the case o f com petitive effects, larger firm s are believed to be in a better position to take advantage o f new o p p o rtu n i­ ties to gobble m arket share than sm aller firms. For this reason, the larger firm s should experience greater positive abnorm al returns than that o f sm aller firm s if the unfav o r­ able audit opinion announcem ent brings about a com petitive effect, but less o f an ab n o r­ mal negative return than that o f sm aller firm s if there is a contagion effect. T herefore, in this study, an a p rio ri expectation o f a positive relationship betw een the size o f a firm and the com petitive and contagion effects is set forth, in keeping w ith the G raham and King [38], B annister [30], and K ohers and R ajagopal [39] studies. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 98 B usiness Cycles B usiness cycles are expected to play a significant part in the m agnitudes o f con­ tagion and com petitive effects. T he only econom ic contraction o f relevance to this study is a recession w hich began in July o f 1990 and ended in M arch o f 1991 (U.S. D e­ partm ent o f C om m erce [40]). Therefore, the "going concern" sam ple w hich encom ­ passes 1989 through 1996 will have som e unfavorable audit opinion announcem ents w hich occur during a recession, but the "qualified opinion" sam ple w hich encom passes 1984 through 1988 does not contain announcem ents affected by periods o f econom ic contraction. The variable used to distinguish the overall effect o f business cycles on the com petitive and contagion effect is a dum m y variable w hich takes the value o f one w hen the unfavorable audit opinion announcem ent occurred during the 1990-91 reces­ sion and zero otherw ise. H istory has show n that the m ajor adjustm ents in the stock m arket occur ju s t fol­ low ing periods o f w idely enjoyed upsw ings in the m arket as a w hole as the econom y expands. Essentially, business cycles set an atm osphere for stock price responses to new inform ation. Since investors w ith "bad new s on the brain" tend to be m ore cautious during econom ic dow nsw ings, these negative expectations should already be built into stock prices. For this reason, the m agnitude o f the negative stock price changes associ­ ated w ith a contagion effect should be less when the econom y is experiencing an eco­ nom ic contraction. R ajagopal [41] explains this result as the effects o f a g reater surprise to the m arket w hen a firm receives an unfavorable audit opinion during econom ic R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 99 expansions than during econom ic contractions. Rollers and R ajagopal [39] find this re­ lationship to hold. A lso, the m agnitude o f a com petitive effect should be less w henever the econom y is in recession, reflecting the increased risk aversion o f investors during the contractionary period. The a p rio ri expectation, therefore, is that a positive relation­ ship exists betw een tim es when the econom y is contracting and the contagion effect and a negative relationship between the com petitive effect and a recessionary econom y due to larger surprise effect associated w ith "bad" new s during econom ic expansions. Grow th O ptions S tudies done by G aver and G aver [42] and Barclay and Sm ith [43 ] have indi­ cated that firm s with higher m arket-to-book value ratios tend to have a larger set o f grow th options. A long with these growth options also com es a higher separation o f in­ form ation betw een m anagers and owners. Because o f the higher inform ational asym ­ m etries associated with higher m arket-to-book value ratios, the m agnitudes o f resulting abnorm al returns from a com petitive effect or a contagion effect should be greater for those firm s. Hence, a priori, an expected positive relationship exists betw een the m arket-to-book equity values and the com petitive effect and a negative relationship be­ tw een the m arket-to-book equity values and the contagion effect. These expectations are in accordance w ith the w ork done by R ajagopal [41], w ho also used the m arket-tobook equity value ratio as a proxy for intangible grow th options. O n the other hand, one can also argue that a firm w ith m ore grow th options will suffer less in the case o f a contagion effect than firm s w ith less grow th options due to R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 100 the ability to expand into areas not specifically affected by the "reason" for the unfavor­ able audit opinion. These unique sources o f grow th m ay actually bring about a positive relationship betw een the contagion effect and the m arket-to-book value o f equity. A l­ though this particular reasoning is not docum ented in the literature, it could m ake am ­ biguous the actual relationship betw een grow th options and contagion effects associated w ith unfavorable audit opinion announcem ents. The variable used in the regression as the m arket-to-book value ratio o f equity was obtained from C O M P U S T A T data sources. The fiscal year-end value o f this vari­ able w as used for m ost rivals. W henever the fiscal year-end value w as not available, the last quarterly value o f the variable available was used. This substitution was m ade for less than 3% o f the rivals in the com plete sam ple which had all other C O M P U ST A T variables available for the regressions. E xchange Effects In a study by G rant [37]. a larger average stock price reaction to earnings an ­ nouncem ents for firm s traded in the over-the-counter m arket (such as the N A SD A Q ) as com pared to those traded on one o f the m ajor exchanges (specifically the A m erican Stock Exchange or the N ew Y ork Stock Exchange) is found to exist. In the G raves et al. [44] study, the authors also show that greater inform ational asym m etries are associ­ ated w ith firm s traded on the over-the-counter m arket as com pared to those traded on the m ajor exchanges. R ajagopal [41 ] follow s the theoretical reasoning that, not only are the firm s in the O TC m arket typically "younger" than those on the m ajor exchanges. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 101 resulting in less inform ation available for those firms, but these firm s tend also to have few er insiders, allow ing them to better retain inform ation about the true state and op p o r­ tunities o f the com pany. For this reason, a greater com petitive or contagion effect, as m easured by abnorm al stock returns, should result for firm s trading on the N A S D A Q as opposed to the A m erican Stock Exchange or the N ew York Stock Exchange. C onse­ quently, a p rio r i, a positive relationship betw een the com petitive effect and a N A S D A Q -traded firm and a negative relationship betw een the contagion effect and a N A S D A Q -traded firm is expected. The variable used to control for the "exchange effect" associated w ith com peti­ tive or contagion effects is a dum m y variable. This variable takes on the value o f one if the co m peting firm 's stock is listed in the over-the-counter m arket (the N A SD A Q ). The dum m y is set to zero in the event the rival firm 's equity is traded on one o f the two m a­ jo r exchanges (N Y SE o r A M EX ). C oncurrent Earnings Releases It is not unusual for the release o f a firm 's audit opinion to coincide w ith the re ­ lease o f that firm 's financial perform ance for the fiscal year. In this situation, the im pact o f the audit announcem ent m ay be com pounded by the earnings announcem ent, particu­ larly if the firm operated at a loss. For this reason, a dum m y variable is used to extract the effect o f a concurrent earnings announcem ent. This variable w ill take on a value o f one w henever the unfavorable audit opinion announcem ent is coupled w ith an an­ nouncem ent o f earnings, o r zero otherw ise. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 102 In D opuch et al. [2], the authors show that significant negative abnorm al stock returns are specifically associated w ith "subject to" qualified audit opinion announce­ m ents by the m edia for the announcing firm. They also have show n that the differences betw een the negative abnorm al returns experienced by firm s w hich included earnings releases w ith the audit opinion announcem ent and those o f the firm s w hich did not in­ clude earnings releases with the announcem ent o f their "subject to" qualified audit o pin­ ion w as negative and significantly different than zero. This finding im plies that som e o f the negative abnorm al returns were generated by the concurrent release o f earnings w ith the announcem ent o f the audit opinion qualification. For this reason, this study extracts the effect o f earnings releases w hich coincide with the "subject to" qualified and "going concern" audit opinions. A lthough the im pact o f earnings releases on the com petitive or contagion effect is a function o f w hether those earnings are positive or negative, the findings o f D opuch et al. [2] w ould suggest the earnings announcem ents release infor­ m ation about losses rather than gains to the announcing firm. B ased on this assum ption, an accom panying negative earnings announcem ent should increase the m agnitude o f the contagion effect as well as the com petitive effect. T he positive abnorm al returns to ri­ vals should increase when a com petitive effect dom inates because rivals are able to cap­ ture the com pounded loss o f m arket share by the announcing firm. A lso, during contagion, the negative abnorm al returns o f rivals should be greater in m agnitude due to the com pounded "bad" news effect o f the announcem ent resulting from releasing earn­ ings as well as the audit opinion results. Hence, the a p rio ri expected outcom e o f the R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 103 earnings release coefficient is negative for the contagion effect and positive for the com ­ petitive effect. Effect o f A nnouncing Firm 's A bnorm al Returns In determ ining the m agnitude o f the contagion or com petitive effect on the in­ dustry rivals o f a firm announcing a "subject to" qualified audit opinion o r a "going con­ cern" audit opinion, the abnorm al return to the announcing firm may play a significant role. In this study, the variable used as the abnorm al return to the announcing firm is the three-day cum ulative abnorm al return com puted over the event w indow beginning one day before and ending one day after the announcem ent date. A greater reduction in equity value experienced by the announcing firm m ay be due to the level o f expectation about how the audit opinion w ould turn out. O ne w ould expect totally unanticipated audit qualification announcem ents to be associated with larger negative abnorm al re­ turns for the announcing firm , due to the "surprise" effect o f the new inform ation. If this new inform ation is o f im portance to rivals and is equally unanticipated, the m agni­ tude o f the abnorm al returns associated with the com petitive or the contagion effect should be increased. This is in keeping with the findings o f Firth [45]. w here the re­ sponse to "surprise" dividend changes had a significant im pact on the responses o f other firm s to that announcem ent. The "surprise" effect was proxied by the abnorm al returns experienced by the announcing firm. Also. R ajagopal [41] finds the strength o f the sig­ nal as proxied by the abnorm al returns to the announcing firm as a significant co n trib u ­ tor in explaining the intra-industry effects o f bond dow ngrades. In keeping w ith this R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 104 logic, the greater the surprise brought about by unfavorable audit opinion announce­ m ents, the greater should be the negative abnorm al return to the announcing firm and the inform ation transfer effect to rival firms (if one exists). Therefore, the a p rio ri ex­ pectations are that the com petitive effect should be negatively related to larger abnorm al returns experienced by the announcing firm, since larger negative price m ovem ents as percentages by the announcing firm are expected to cause larger positive price m ove­ m ents for com petitors. Also, the contagion effect should be positively related to the ab­ norm al returns experienced by the announcing firm, since larger negative price m ove­ m ents by the announcing firm should result in larger negative m ovem ents in the returns o f its com petitors. S im ilarities Between Earnings o f A nnouncing Firm and Rivals It has been stated earlier that industries m ay actually be "split" betw een conta­ gion and com petitive effects. A ccordingly, som e firm s in the industry are very sim ilar to the announcing firm , w hile other rivals are not. Lang and S tulz [29] proxied the sim ilarity o f cash flow s betw een the rivals and the announcing firm w ith the correlation coefficient o f returns betw een the announcing firm and its industry rivals. They found significantly larger abnorm al returns for firms w ith higher return correlations w henever a contagion effect w as present. H ow ever the positive abnorm al returns for industry ri­ vals w ith low return correlations w ith the announcing firm s did not test significant. Hence, in keeping w ith the theoretical background provided by Lang and Stulz [29], the a p rio ri expectation o f this study is that there is a negative relationship betw een the R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 105 abnorm al returns resulting from a contagion effect and the correlations o f earnings o f ri­ vals w ith the announcing firm. H ow ever, since correlations betw een the earnings o f firm s proxy the degree to w hich firm s are alike, higher correlations o f earnings are asso­ ciated w ith firm s w hich are m ore directly com petitive. For this reason, the a p rio ri ex­ pectation o f the relationship betw een the correlations o f earnings betw een announcing firm s and its rivals and the com petitive effect is positive. C R SP data (specifically the daily returns) is used to com pute the correlation co­ efficients betw een the returns o f the announcing firms and their industry rivals. These correlations are com puted over the estim ation period for the regression coefficients used in com puting the abnorm al returns to each announcing firm and each rival. T his period begins 299 days before the announcem ent date and ends 50 days before the announce­ m ent date. The Problem o f M ulticollinearity M uiticollinearity occurs w hen a near linear relationship exists betw een tw o or m ore o f the regressors. W hen collinearity is present, the individual effects o f each inde­ pendent variable cannot be partialed out. causing variables w hich w ould norm ally be significant to be insignificant (the inflated variances produce inefficient tests), since the highly correlated variables cancel each other out. M ulticollinearity is not a problem if the R: (predictive ability) is all that m atters to the analyst. H ow ever, for m ost studies in w hich inferences are to be m ade, m ulticollinearity m ust be dealt w ith. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. M ulticollinearity can be detected via 4 different m ethods. These m ethods are as follows: (1) Inspecting the correlation m atrix for highly correlated independent variables is one m ethod. The problem w ith using the correlation as a m easure o f collinearity is that the degree o f correlation defined to be a problem is a subjective decision. This m ethod also has a w eakness in not being able to detect com plex collinearity, w hich is a collinearity problem betw een 3 o r m ore independent variables. (2) The variance inflation factor is another m easure used to detect collinearity. The variance inflation factor (V IF) for a given regressor is com puted by taking the reciprocal o f I m inus the R -S quared value resulting from a regression o f the regressor in question against all other regressors in the m odel. For exam ple, if the grow th options variable w ere used as the dependent variable in a regression run upon all o f the other stated vari­ ables and resulted in a R -S quared value o f .9. then the value o f the variance inflation factor w ould be 10. Like the correlation m atrix, this m ethod is subject to the subjectiv­ ity o f the analyst, w ith som e saying the VIF should be less than 5 and others saying it should be less than 10. A lthough the VIF can detect com plex collinearity problem s, an ­ o th er inherent w eakness o f the VIF m easure is the inability to pinpoint exactly w here the collinearity problem is (i.e., w hat variables are too highly correlated with each other). (3) T he tolerance factor reported for each predictor variable gives the proportion o f variance in the each predictor variable w hich is not accounted for by the other predictor R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 107 variables. V alues o f the tolerance factor lie on the interval [0,1 ]. Larger values o f the tolerance factor im ply large am ounts o f unique or specific variance for each explanatory' variable. This, in turn, indicates a relatively sm all am ount o f m ulticollinearity. T he rule o f thum b is that a tolerance factor greater than o r equal to 0.1 is acceptable for con­ cluding that m ulticollinearity is not a problem . A lthough this m ethod can detect co m ­ plex collinearity, it does not specifically pinpoint w hich variables are the problem . (4) C ondition indexes (C l) tend to give the best results in testing for collinearity since a Cl w hich is greater than 30, associated w ith variance decom position proportions w hich are greater than .5 for two or m ore independent variables, show s that the collinearity ex­ ists. as w ell as w hich precise variables are the problem variables. Since it checks re la­ tionships betw een all variables sim ultaneously, com plex collinearity (which affects m ore than two predictor variables) can be detected using this method. There are also several rem edies for collinearity. The first o f w hich was already m en tio n ed -th at is, do nothing if the purpose o f the m odel is strictly to predict. O th er m ethods include dropping one or m ore o f the problem independent variables, centering the data by subtracting the m ean from each variable, standardizing the independent vari­ ables, using ridge regression and using factor analysis. If m ulticollinearity is d e te r­ m ined to be a problem in this study, several regressions will be run. These regressions will differ in the predictors allow ed to rem ain in each run (specifically predictors am ong w hich no m ulticollinearity exists am ong any tw o or more). R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 108 Summary T his chapter described in detail the sam ple selection process and the m ethodol­ ogy for com puting and testing the abnorm al returns to announcing firm s and rivals re­ sulting from unfavorable audit opinion announcem ents. Included are discussions ju stify in g how and w hy the techniques are to be used, as well as review s o f o ther tech­ niques w hich will not be used. A concise sum m ary o f the research process utilized is presented below. (1) (2) The announcing firm s are selected from the Wall Street Jo u rn a l Index and the B usiness Wire over the period o f 1984-1988 for firm s receiving "qualified" opinions and 1989-1996 for firm s receiving "going concern" audit opinions. The C U SIPS for the announcing firms are retrieved from C O M P U ST A T . at w hich tim e the firm s w hich are not traded on the N A SD A Q . N Y SE o r A M EX w ere rem oved. (3) The sam ple o f rival firm s is retrieved from C RSP based on the SIC code o f the announcing firm. (4) The abnorm al returns are com puted via the single index m odel using C R S P data for the announcing and rival firms. (5) The abnorm al returns are standardized and tested by industry to determ ine w hich effect dom inates, w hether contagion, com petitive or neither. (6) The industries from both tim e periods that have significant positive cum ulative standardized abnorm al returns over the event w indow (com petitive effect) are grouped, at w hich tim e a regression is estim ated to explain the variance in the cum ulative abnorm al percentage returns to the rivals o f the announcing firm s over the event w indow . (7) T he sam e procedure is perform ed on the industries in w hich a significant contagion effect is found to exist. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 109 C hapter HI also presented the a p rio ri expectations surrounding the cum ulative abnorm al returns for the industries (rivals) as a w hole, the announcing firm s and the in­ dependent variables (contributing factors) used to explain the variance in cum ulative ab­ norm al returns to the rivals o f industries w hich have significant com petitive or contagion effects. Included in this chapter is a description o f the sam ple and how it breaks dow n by three-digit SIC codes. Finally, the chapter concludes w ith a discussion o f m ulticoliinearity, how to test for it, and w hat is done if it is found to be a problem . Citations [1] Dodd. P., N. D opuch. R. W. H olthausen. and R. W. Leftw ich. "Q ualified A udit O pinions and Stock Prices: Inform ation C ontent. A nnouncem ent Dates and C oncurrent D isclosures." Journal o f A ccounting a n d Econom ics. 6, 1984, pp. 3-38. [2] D opuch, N., R. W. H olthausen. and R. W. Leftw ich. "A bnorm al R eturns A sso­ ciated W ith M edia D isclosures o f'S u b je c t to' Q ualified A udit O pinions."Jour­ nal o f A ccounting a n d E conom ics, 8, 1986. pp. 93-117. [3] Jones. F. L. "The Inform ation C ontent o f the A uditor's G oing C oncern E valua­ tion." [4] Jo u rn a l o f A ccounting a n d P ublic Policy. 15. 1996. pp. 1-27. Elliott. J. "'Subject to' A udit O pinions and A bnorm al S ecurity R e tu m s-O u tcom es and A m biguities." Journal o f A ccounting Research. 20. 1982, pp. 617-638. [5] Loudder. M. L., I. K. K hurana, R. B. Saw yers. C. C ordery. C. Johnson. J. Lowe. a n d R . W underle. "The Inform ation C ontent o f A udit Q ualifications." A u d it­ ing. 11, 1992. pp. 69-82. [6] B row n. S. J.. and J. B. W arner. "M easuring Security Price Perform ance." J o u r­ nal o f F inancial E conom ics. 8. 1980, pp. 205-258. [7] B row n, S. J.. and J. B. W arner. "U sing D aily Stock R eturns: T he C ase o f Event Studies." Jo u rn a l o f F inancial Econom ics. 12, 1985, pp. 3-32. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 110 [8] Reilly, F. K., and K. Hatfield. "Investor E xperience W ith N ew Stock Issues." F inancial A nalysts Journal, S eptem ber-O ctober 1969, pp. 73-80. [9] Ibbotson, R. "Price Perform ance o f C om m on Stock N ew Issues." Jo u rn a l o f F inancial E conom ics, 2, 1975. pp. 235-272. [10] Ibbotson, R.. and J. F. Jaffe. '"H ot Issue'M arkets." Jo u rn a l o f Finance. 30, 1975. pp. 1027-1042. [11] Levis. M. "The Long-R un Perform ance o f Initial Public O fferings: The UK E xperience 1980-1988." F inancial M anagem ent. Spring 1993. pp. 28-41. [12] Sharpe, W. F. "C apital A sset Prices: A T heory o f M arket E quilibrium U nder C onditions o f Risk." Journal o f Finance, 19. 1964, pp. 425-442. [13] Lintner, J. "The V aluation o f Risk A ssets and the Selection o f Risky Invest­ m ents in Stock Portfolios and Capital B udgets." R eview o f E conom ic a n d S ta tis­ tics. 47, 1965. pp. 13-37. [14] Black. F. "C apital M arket E quilibrium W ith R estricted B orrow ing." Jo u rn a l o f Business. 45. 1972, pp. 444-455. [15] Roll. R. "A C ritique o f the A sset Pricing T heory's T e s ts -P a rt 1: O n Past and Potential T estability o f the Theory." Journal o f F inancial Econom ics, 4. 1977. pp. 129-176. [16] Roll. R.. and S. A. Ross. "On the C ross-Sectional R elation B etw een Expected R eturns and Betas." Journal o f Finance, 49. 1994. pp. 101-121. [17] Fam a. E. F., and K. R. French. "The C ross-Section o f Expected Stock R eturns." Jo u rn a l o f Finance, 47, 1992. pp. 427-465. [18] B aginski. S. P. "Intra-Industry Inform ation T ransfers A ssociated w ith M anage­ m ent F orecasts o f Earnings." Jo u rn a l o f A ccounting Research. 25. 1987. pp. 196-216. [19] Strong. N. "M odeling A bnorm al Returns: A R eview A rticle." Jo u rn a l o f B usi­ ness, F inance & A ccounting. 19. 1992, pp. 533-553. [20] M ikkelson. W. H. "C onvertible C alls and Security R eturns." Jo u rn a l o f F inan­ cial E conom ics. 9. 1981. pp. 237-264. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. [21 ] A lii, K... S. T hapa, and K. Yung. "Stock Price D ynam ics in O verlapped M arket Segm ents: Intra and Inter-Industry C ontagion Effects." Jo u rn a l o f Business, Finance & A ccounting, 21, 1994, pp. 1059-1070. [22] Patell, J. N. "Corporate Forecasts o f Earnings Per S hare and Stock Price B ehav­ ior: Em pirical Tests." Journal o f A ccounting Research, 14, 1976, pp. 246-276. [23] Peterson, P. P. "Event Studies: A R eview o f Issues and M ethodology." Q uar­ terly Journal o f Business a n d Econom ics, 28, 1989. pp. 36-66. [24] D odd, P., and J. B. W arner. "On C orporate G overnance: A S tudy o f Proxy C ontests." J o u rn a l o f F inancial Econom ics. 11, 1983. pp. 401-438. [25] M ikkelson, W. H „ and M. M. Parch. "Stock Price Effects and C osts o f S econ­ dary D istributions." Journal o f F inancial Econom ics, 14, 1985, pp. 165-194. [26] M oore, N. H.. D. R. Peterson, and P. P. Peterson. " S h e lf R egistrations and Shareholder W ealth: A C om parison o f S helf and Traditional Equity O fferings." Journal o f Finance, 41, 1986. pp. 451-464. [27] M asulis. R. W. "The Effect o f Capital Structure C hange on Security Prices." Journal o f F inancial Econom ics, 8, 1980. pp. 139-178. [28] D ann. L. Y.. and W. H. M ikkelson. "C onvertible Debt Issuance. C apital S truc­ ture C hange and Financing-R elated Inform ation." Jo u rn a l o f F inancial E co­ nom ics, 13, 1984. pp. 157-186. [29] Lang. L. H„ and R. M. Stulz. "C ontagion and C om petitive Intra-Industry Effects o f B ankruptcy A nnouncem ents: An Em pirical A nalysis." J o u rn a l o f F inancial Econom ics, 32, 1992. pp. 45-60. [30] B annister. J. W. "Earnings Signals and Inter-Firm Inform ation Transfers." Journal o f Business, Finance & A ccounting, 21, 1994, pp. 1127-1149. [31 ] Firth, M. "The Im pact o f E arnings A nnouncem ents on the Share Price B ehavior o f Sim ilar Type Firm s." E conom ic Journal. 86. 1976, pp. 296-306. [32] Foster. G. "Intra-Industry Inform ation Transfers A ssociated W ith Earnings Releases." J o u rn a l o f A ccounting a n d E conom ics. 3, 1981. pp. 201-232. [33] C linch, G. J.. and N. A. Sinclair. "Intra-Industry Inform ation R eleases." J o u r­ nal o f A ccounting a n d Econom ics. 9. 1987, pp. 89-106. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 112 [34] C heng, L. T. W „ and J. E. M cDonald. "Industry Structure and R ipple Effects o f B ankruptcy A nnouncem ents." Financial Review , 31, 1996, pp. 783-807. [35] Bolton, P., and D. Scharfstein. "A Theory o f Predation Based on A gency Prob­ lems in Financial C ontracting." A m erican Econom ic Review , 80, 1990. pp. 93-106. [36] A tiase. R. ''P redisclosure Inform ation, Firm C apitalization, and Security Price B ehavior A round Earnings A nnouncem ents." Jo u rn a l o f A cco u n tin g Research, 23, 1985, pp. 21-36. [37] Grant, E. B. "M arket Im plications o f D ifferential A m ounts o f Interim Inform a­ tion." Jo u rn a l o f A ccounting Research, 18. 1980, pp. 255-268. [38] G raham , R. C., and R. D. King. "Industry Inform ation T ransfers: T he Effect o f Inform ation Environm ent." Journal o f Business, Finance & A ccounting, 23. 1996. pp. 1289-1306. [39] Kohers, T., and S. R ajagopal. "The Intra-Industry Effects o f C anceled Debt O fferings." W orking paper. 1997. [40] U.S. D epartm ent o f C om m erce. Survey o f C urrent Business, O ctober 1994. [41 ] R ajagopal. S. "B ond D ow ngrades and Intra-Industry C ontagion and C om peti­ tive Effects: The C ase o f Industrials." D issertation, M ississippi State U niver­ sity, M ay 1997. [42] G aver. J. J.. and K. M. Gaver. "Additional Evidence on the A ssociation Be­ tw een the Investm ent O pportunity Set and C orporate Financing, D ividend and C om pensation Policies." Journal o f A ccounting a n d E conom ics, 16, 1993. pp. 125-160. [43] B arclay, M. J., and C. W. Smith. "The M aturity Structure o f C orporate Debt." Jo u rn a l o f Finance, 50, 1995. pp. 609-631. [44] G raves, J. A., S. P. H egde. and R. E. M iller. "Trading M echanism s and the C om ponents o f the B id-A sk Spread." Journal o f Finance, 49. 1994. pp. 1471-1488. [45] Firth. M. "D ividend C hanges. A bnorm al Returns, and Intra-Industry Firm V alu­ ations." Jo u rn a l o f F inancial a n d Q uantitative A nalysis, 31. 1996. pp. 189-211. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. CHAPTER IV RESULTS A N D IMPLICATIONS In this chapter, the results o f the m ethods described in C hapter III are reported. First, the outcom e o f the hypotheses testing for significant abnorm al returns for firms w hich announced an unfavorable audit opinion is presented. Second, the results o f the hypotheses test to determ ine which effect, com petitive or contagion, dom inates in each industry is discussed. Third, the results o f the tw o regressions (one for those industries exh ib iting significant com petitive effects and one for those industries exhibiting signifi­ cant contagion effects) are reported and com pared to the a p rio ri expectations m ade in C h apter III. Finally, inferences are m ade from the regression results and the im plica­ tions o f the entire analysis from start to finish are presented. This chapter also discusses the co ntributions m ade by this study to the existing literature on the subject. Announcing Firms' Abnormal Returns T able 3-1 and Table 3-2 in the previous chapter presented the break dow n by in­ d ustry o f the "going concern" and "qualified opinion" sam ples. The "going concern" sam ple contains 36 announcing firm s representing 19 different three-digit-SIC code de­ fined industries. The "qualified opinion" sam ple contains 43 announcing firm s from 24 113 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 114 industries. T hese two sam ples w ere tested independently to determ ine the overall reac­ tion o f investors to the adverse inform ation com m unicated by unfavorable audit opinion announcem ents and to note any differences in investor reactions from w hen the audit o p inion w as qualified (pre-1989) to w hen the audit opinion becam e unqualified w ith special language (p o st-1988). H ypothesis and Expectations Based on the outcom e o f studies done by Dopuch et al. [1 1 and Jones [2J, firm s w hich received a "subject to" qualified o r "going concern" unqualified audit opinion m ade public through announcem ents in the business m edia were expected to have an adverse stock price reaction to the bad news. B elow is the specific hypothesis w hich w as tested: H0: "Subject to" qualified and "going concern" audit opinions are not associated w ith significant negative changes in the stock returns o f the announcing firm. H,: "Subject to" qualified and "going concern" audit opinions are associated w ith significant negative changes in the stock returns o f the announcing firm. T he outcom e o f the hypothesis test for both sam ples ("going concern" and "qualified opinion") w as based on the standardized cum ulative abnorm al returns to each announcing firm over the three-day event w indow . This event w indow began the day before the "announcem ent date" and ended the day after. These results are presented in T able 4-1 and T able 4-2 respectively. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 115 Table 4-1 A V E R A G E A B N O R M A L R ETURNS TO F IR M S R EC EIV IN G M ED IA C O V E R A G E OF T H E IR "SU B JE C T TO" Q U A LIFIED A U D IT O PIN IO N S FRO M 1984-1988 (EV EN T W IN D O W IN B O LD ) Day R elative to A nnouncem ent Date: A verage Abnorm al Return: T -value: N um ber o f O bservations: (%) - 10 0.2320 0.932 42 -9 0.3346 0.183 42 -8 -1.1202 -0.816 43 -7 -1.6095 -1.503 * 43 -6 1.8899 1.193 43 -5 0.6161 1.138 43 -4 0.8481 -0.147 43 -1.528 * 43 - j -0.9025 . 2 0.0124 -0.691 43 - 1 1.5725 2.707 43 0 -7.0152 + 1 0.8030 (-1,1) -4.6398 + 2 0.9841 1.399 43 + 3 -0.3174 0.012 43 +4 0.4544 0.064 43 + 5 -0.3454 0.238 43 +6 -0.6654 -1.008 43 + 7 0.6012 0.231 43 +8 0.4409 1.566 43 +9 -0.9425 + 10 0.3789 -10.436 *** 1.456 -6.274 *** -2.057 ** 0.643 43 43 43 43 43 *** S ignificant at the 1% level under a one-tailed test. ** S ignificant at the 5% level under a one-tailed test. * S ignificant at the 10% level under a one-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 116 Table 4-2 A V E R A G E A B N O R M A L R ETURNS TO FIRM S R EC EIV IN G M ED IA C O V ER A G E O F T H E IR "G O IN G C O N C ER N " U N Q U A LIFIED A U D IT O PIN IO N S FRO M 1989-1996 (EV EN T W IN D O W IN BOLD) Day R elative to A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (%) - 10 -0.8125 -0.786 36 -9 -0.8834 -1.411 * 36 -8 1.9759 1.401 36 -7 -0.1457 -0.556 36 -6 2.0345 1.255 36 -5 1.2788 1.057 36 -4 0.7938 -0.459 36 -3 -1.4479 -1.733 ** 36 . 2 -1.5901 0.132 34 - 1 2.1206 -1.081 31 0 -25.1997 + 1 9.3106 (-1,1) -13.7685 + 2 1.6121 3.372 36 + 3 5.9782 2.906 36 +4 1.4502 3.916 36 + 5 3.1891 1.638 36 +6 -2.3287 + 7 0.4613 + 8 -0.7698 +9 0.8859 1.621 36 + 10 1.1902 0.798 36 -26.923 *** 6.704 -21.299 *** -3.558 *** 1.304 -2.211 ** 36 36 36 36 36 36 *** S ignificant at the 1% level under a one-tailed test. ** S ignificant at the 5% level under a one-tailed test. * Significant at the 10% level under a one-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 117 A bnorm al Return Test R esults for A nnouncing Firms T able 4-1 gives the results o f the hypothesis tests for the "subject to" qualified audit opinion sam ple. As expected, the average cum ulative abnorm al returns for the event w indow , w hich is denoted as (-1.1), is negative and highly significant. On aver­ age, the firm s announcing the receipt o f a "subject to" qualified audit opinion in the fi­ nancial m edia suffered an equity loss o f 7.02% on the announcem ent day and 4.64% o v er the period defined as the event w indow . T here appears to be som e leakage because there are significant losses on day three and day seven prior to the event w indow . Also, the inform ation dissem ination process w as efficient and effective because there are no significant losses after the event w indow , w ith the exception o f one day (the 9th day af­ ter). This significant abnorm al return m ay be attributable to follow up announcem ents. The results o f the hypothesis tests for the "going concern" unqualified audit opinion sam ple are given in Table 4-2. A gain the a p rio ri expectations concerning the negative average cum ulative abnorm al returns for the event w indow are substantiated as the t-value o f -21.299 reveals these average returns o f -13.77% to be highly significant. T he average abnorm al returns for the announcem ent date o f -25.20% is even m ore sig­ nificant. with a t-value o f -26.923. Interestingly enough, the average abnorm al retum sfor the first few days after the announcem ent date are large and positive. This im plies an initial overreaction on the part o f investors to the unfavorable audit opinion an ­ nouncem ent. follow ed by a subsequent adjustm ent back to the true equilibrium value o f the securities. This theory (the overreaction hypothesis), w hich is discussed in further R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 118 detail in the next section, is presented in studies by D eB ondt and T haler [3.5], and H ow e [4], S ignificant negative average abnorm al returns to firm s announcing "going con­ cern" unqualified audit opinions in the m ajor financial press also occur on the third and ninth day before the official announcem ent date, as well as the sixth and eighth day after the announcem ent date. The returns before the announcem ent date m ay be evidence o f an inform ation leakage. The significant negative returns on the days follow ing the "m arket readjustm ent" may be the result o f subsequent stories released in the financial press, o r even a further readjustm ent. Intuitively, the announcing firms were expected to encounter a negative equity adjustm ent as a result o f the announcem ent o f the unfavorable audit outcom e. The data were checked to determ ine w hether an announcing firm experienced a positive abnor­ m al return on the announcem ent date o f the audit opinion. O f the 79 announcing firms, five encountered non-negative returns on the announcem ent date, four o f which had no stock price change on the announcem ent date. Im plications o f Test Results for A nnouncing Firm s As show n in Table 4-1 and Table 4-2. the test rejects the null hypothesis in favor o f the alternative hypothesis. T he relevant hypothesis specifically states that '"subject to' qualified and 'going concern' audit opinions are associated with significant negative changes in the stock returns o f the announcing firm w hen the audit opinion results are R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 119 released in the m ajo r financial press." This hypothesis was form ally tested by D opuch et al. [1 ] and Jo n es [2] for "subject to" qualified au d it opinion announcem ents released by the m ajor financial m edia. D opuch e t al. [ 1] attribute these results to a m arket reaction to the new inform a­ tion contained in the auditor's opinion released by the independent auditors o f the an­ nouncing firm w hen the opinion is m ade public in the financial press. The D opuch et al. [1 ] study found the "subject to" qualified audit opinion announcem ents to produce a 4.70% significant abnorm al average loss to the announcing firm over the (-1.1) event w in d o w for the sam ple period o f 1970 through 1979. For the sam e event w indow . Jo n es [2] concluded that the "subject to" qualified audit opinion announcem ent released in the m ajor b usiness press netted the announcing firm a significant abnorm al average loss o f 4.2% for the exam ination period o f m id -1979 through 1988. These results are sim ilar to the significant abnorm al average loss o f 4.64% found by this study for the testing period o f 1984 through 1988. using the sam e event w indow o f (-1.1). This is the first study to perform a sim ilar test for "going concern" unqualified au d it opinions o v e r a testing period o f 1989 through 1996. The audit opinion change im plem ented in 1989 required that "going concern" audit opinions now be classified as unqualified o p in io n s with special language rather than "subject to" qualified audit opin­ ions. The results o f this study, nam ely a significant average abnorm al loss o f 13.77% to the firm s w hich h ad a "going concern" audit opinion receiving attention in the m ajor fi­ nancial m edia, essentially show that the unfavorable audit opinion announcem ents R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. surrounding "going concern" uncertainties reveal the sam e type o f new , adverse infor­ m ation about the announcing firm. This inform ation brings about a significant m arket reaction, regardless o f w hat the audit opinion is officially called. T he positive adjustm ent in the equity value o f the announcing firm s suggested by the positive returns for the period o f tim e from the first to the fifth day after the offi­ cial "announcem ent date" indicate an initial overreaction by the m arket to the negative new s surrounding the released unfavorable audit opinions. The overreaction hypothesis deals w ith an "overshooting" o f the true equilibrium stock price as the m arket initially reacts to econom ic news. A fterw ards, gradual adjustm ents in the equity value occur un­ til the stock price reaches its true equilibrium value. This hypothesis w as first presented by W erner D eB ondt and R ichard T haler [3] and suggests an arbitrage opportunity for investors. B row n. Tinic and H arlow [6] explain that "large" stock price m ovem ents based on new inform ation released into financial m arkets are follow ed, on average, by "large" stock price adjustm ents in the opposite direction. The overreaction studies (D eB ondt and T haler [3.5], and Howe [4]) also suggest that the larger the initial "sur­ prise" created by the econom ic event, the larger the initial stock price adjustm ent and the larger the subsequent stock price readjustm ent. The findings o f a different study done by B row n and H arlow [7] suggest that the m arket overreaction hypothesis is true w hen the initial reaction is negative (that is. large negative stock price m ovem ents re­ sulting from the release o f "bad" new s are subsequently adjusted upw ard), how ever their findings contradict those o f D eB ondt and Thaler [5] w hen the initial reaction is positive. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 121 In sum m ary, this study has provided additional evidence that "subject to" quali­ fied and "going concern" audit opinions introduce new inform ation into the m arket about the announcing firm w hen the audit results are m ade public via a press release in the m ajor financial m edia. These findings are as expected and confirm the findings o f earlier studies exam ining "subject to" qualified audit opinion announcem ents perform ed by D opuch et al. [ 1] and Jones [2]. The findings surrounding the "going concern" audit opinions (after the opinion change initiated in 1989) are essentially new and m ay con­ firm the existence o f an overreaction phenom enon m ade apparent by increased price volatility inherent in the m arket reaction to the sam e opinions in the late 1980s and early to m id-1990s. T he significant negative average abnorm al returns associated w ith the m edia announcem ent o f a "subject to" qualified audit opinion for the exam ination pe­ riod o f 1984 through 1988 were 4.64% for the announcing firm, with a t-value o f -6.274 w hen cum ulated over the traditional three-day window. A fter taking the five-day "read­ ju stm en t period" into account (i.e.. cum ulating the returns over eight days rather than three), the significant negative abnorm al returns to the firm s announcing a "going con­ cern" audit opinion from 1989 through 1996 average 3.87% w ith a t-value o f -13.025. Hence, after adjusting for the overreaction. the resulting negative abnorm al returns ap­ pear to be o f sim ilar m agnitude for both audit opinion sam ples. Rival Firms' Abnormal Returns T he m ajor task o f this study is the detection o f significant abnorm al returns to the rival firm s o f those com panies announcing "subject to" qualified o r "going concern" R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 122 au d it opinions in the m ajor financial press. If the industry rivals o f an announcing firm have significant negative returns, then a contagion effect has occurred as a result o f the announcem ent. A contagion effect results w hen announcem ents about a firm specifi­ cally affected negatively by an econom ic event also affect other firm s in the industry in a sim ilar m anner. These econom ic events are seen by the m arket to be industry-w ide phen o m ena w hich are expected to also adversely affect the earnings o f other firm s in the industry. The contagion effect does not im ply that all firm s in the industry suffer nega­ tive equity adjustm ents, but that the average equity adjustm ent o f the industry as a w hole is negative and statistically significant. If the average industry-w ide equity adjustm ent is significant and positive, then a com petitive effect dom inates that industry. A com petitive effect results w hen the m ar­ ket interprets the econom ic event to be firm -specific. This loss o f earnings and m arket share suffered by the announcing firm results in a transfer o f earnings and m arket share to the industry rivals o f the announcing firm. As w ith the contagion effect, this positive equity adjustm ent is not necessarily enjoyed by all firm s in the industry, but dom inates, on average, the industry as a whole (o f course, it excludes the announcing firm ). In som e industries, although individual firm s may gain or lose as a result o f the new inform ation released, neither effect dom inates. In those cases, there are no signifi­ cant abnorm al returns experienced by the industry as a w hole (again, excluding the an­ n ouncing firm). The expectations, results and im plications associated w ith the co n tag ion and com petitive effects are discussed in the follow ing sections. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. Inform ation Transfer Expectations And H ypothesis The transfer o f inform ation resulting from an announcem ent w hich m akes avail­ ab le new inform ation to financial m arkets may result in equity adjustm ents for firm s in the sam e industry. The a p rio ri expectations o f this study are that announcem ents o f "subject to" qualified audit opinions and "going concern" audit o pinions contain infor­ m ation relevant to the com petitors o f the firms m aking such announcem ents and will re­ su lt in significant average industry-w ide equity adjustm ents in som e cases. T hese ex­ p ectations are based on studies w hich have found com petitive and/or contagion effects to exist in industries w here bankruptcy or earnings announcem ents have been m ade and w hich, subsequently, have resulted in significant industry-w ide stock price adjustm ents on average. Firth [8], F oster [9] and C linch and S inclair [10] have uncovered signifi­ can t inform ation transfers as evidenced by stock price adjustm ents to industry rivals fol­ low ing earnings announcem ents. Likew ise, Lang and Stulz [11] have found these inform ation transfers to exist as a result o f bankruptcy announcem ents. S im ilar infor­ m ation is expected to be released by an unfavorable audit opinion, w hich provides infor­ m atio n about the future expected earnings and increased probability o f bankruptcy for the announcing firm . A lso, as w ith the other inform ational transfer studies m entioned, one cannot be sure w hich effect (contagion or com petitive) will dom inate any specific industry for any given announcem ent that does not specifically clarify the reason for the co m pany's troubles. C onsequently, the dom inant effect is m ainly a function o f the per­ cep tio ns o f m arket participants at the tim e o f the announcem ent an d is based on how R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 124 that particular inform ation released affects the future earnings potential o f the industry rivals. T he form al hypothesis tested by industry for the existence o f com petitive or co n ­ tagion effects w as stated in C hapter III as follows: H0: H ,: "Subject to" qualified and "going concern" audit opinions w hich receive attention in the financial press do not elicit a significant stock price reaction from the announcing firm 's industry rivals. "Subject to" qualified and "going concern" audit opinions w hich receive attention in the financial press elicit a significant stock price reaction from the announcing firm 's industry rivals. A p r io r i, the null hypothesis was expected to be rejected for som e industries, providing evidence that a contagion or a com petitive effect results from the announcem ents o f "subject to" qualified or "going concern" audit opinions. These tw o-tailed tests are based on the cum ulative standardized average abnorm al returns to the rivals in each in­ dustry represented by the announcing firm. The cum ulative abnorm al returns for rivals w ere estim ated over the sam e event w indow o f (-1,1) used to report the abnorm al re­ turns experienced by the announcing firms. A bnorm al R eturn Test R esults for Industry R ivals The test results for both sam ples ("subject to" qualified and "going concern" audit opinions) are presented on an industry-by-industry basis in T able 4-3 and Table 4-4. T he "going concern" sam ple (Table 4-3) includes 18 different industries for w hich m ore than 3 rival firm s had data available on the C R S P database. O f these 18 R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 125 Table 4-3 A V ER A G E A B N O R M A L R ETU R N S TO RIVALS O F FIR M S RECEIV IN G M EDIA C O V E R A G E OF TH EIR "G O IN G C O N C E R N " U N Q U A LIFIED A U D IT O PIN IO N S O V ER THE EV EN T W IN D O W O F (-1,1) A verage A bnorm al N um ber o f Rival Firm s: Returns: SIC: Industry: 131 Oil & G as Extraction 0.00*** 238 283 Drugs & Pharm aceutical 0.22 578 357 C om puter and Office Equipm ent 1.26*** 255 367 Electronic C om ponents & A ccessories -2.26*** 168 381 Search & N avigation Equipm ent -2.96 384 Surgical, M edical & Dental Instrum ents 0.46** 96 491 Electric Services 0.29** 71 495 Sanitary Services -1.13 25 504 Professional & C om m ercial Equipm ent-w holesale 0.89 30 581 Eating & Drinking Places -0.31*** 616 M ortgage Bankers & C orrespondents 0.02 15 655 Land Subdividers & D evelopers 1.49 25 679 Investors, nec 1.43*** 737 C om puter Program m ing, D ata Process -0.09 738 M iscellaneous Business Services 0.73** 18 808 H om e Health Care Services 5.18*** 18 809 M iscellaneous Hom e & A llied Services, nec 2.77** 17 873 Research, D evelopm ent & T esting Services 1.61** 34 (% ) 8 102 224 118 *** Significant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 126 Table 4-4 A V E R A G E A B N O R M A L RETURNS TO RIV A LS O F FIR M S R EC EIV IN G M ED IA C O V ER A G E OF TH EIR "SU B JE C T TO" Q U A LIFIED A U D IT O PIN IO N S O V ER THE EVEN T W IN D O W O F (-1,1) A verage A bnorm al R eturns in (% ): N um ber o f Rival Firms: 152 SIC: Industry: 131 Oil & G as Extraction 0.84 138 Oil & G as Field Services 0.77** 245 M obile Hom es -1.73* 283 Drugs & Pharm aceutical -0.92** 307 Rubber & M iscellaneous Plastics -0.31 14 331 Steel W orks & Blast Furnaces -1.17 17 353 C onstruction, M ining, O il & Gas Equipm ent 0.97* 357 C om puter and O ffice Equipm ent 0.90** 366 C om m unication Equipm ent 1.80*** 367 Electronic C om ponents & Accessories -0.77 41 372 A ircraft, Ship Building _2 |3 * * * 26 382 Lab A pparatus, Electrical M easure & Test -0.36 386 P hotographic Equipm ent & Supply -2.53** 11 491 Electric Services 0.12** 366 493 Electric & O ther Services Com bined -0.65*** jj 495 Sanitary Services -0.28 13 506 Electrical G oods-W holesale 6.39*** 13 661 Real Estate Developers 2.74 7 679 Investors, nec -0.01 87 701 H otels and M otels 737 C om puter Program m ing, Data Process -0.38 140 781 M otion Picture Production -1.61 14 806 H ospitals 0.79 8 43 11 131 18 132 61 100 1.42 *** Significant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission . industries, 11 report a significant abnorm al stock price reaction over the event window . T w o industries (S IC s 367 and 581) show significant negative equity adjustm ents, sug­ gestin g the existence o f a contagion effect, while the o th er nine industries (S IC s 131, 357, 3 8 4 ,4 9 1 , 679, 738, 808, 809 and 873) encountered significant positive equity ad ­ ju stm e n ts, im plying a dom inating com petitive effect. T he industry revealing the largest n egative average abnorm al return (SIC 381) did not test significant, even though the re­ turn w as -2.96% . The absence o f statistical significance m ay be the result o f either a low n um ber o f observations (8) and/or the standardization procedure which gives more w eight to changes in stock prices where the returns are less volatile and less w eight to equity adjustm ents to firms w ith highly volatile returns. The H om e Health C are Services industry (SIC 808) show ed the largest positive average abnorm al return o f 5.18% , w hich is statistically significant at the 1% level. O f the 11 industries giving evidence o f the existence o f com petitive o r contagion effects re­ sulting from the "going concern" audit opinion announcem ents. 6 proved statistically sig n ificant at a level o f .01. T able 4-4 reports the results for the "subject to" sam ple. T his sam ple included 23 d ifferent industries for w hich more than 3 rival firm s had data available on the CRSP database. O f these 23 industries, 11 exhibited a statistically significant response to the unfavorable audit opinion announcem ent over the event w indow . Six o f those indus­ tries (S IC s 138, 353, 357. 366. 491. and 506) had dom inating com petitive effects, while the o th er five industries (S IC s 245. 283. 372. 386, and 493) show ed contagion effects R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 128 that dom inated as a result o f the econom ic event. The largest average negative equity adjustm ent to an industry was -2.53% in the Photographic Equipm ent and S upply indus­ try (SIC 386), w hile the largest positive average industry-w ide equity adjustm ent was 6.39% in the Electrical G oods-W holesale industry. Again, som e large average negative and positive stock price reactions occurred w hich did not test significant due to either a low num ber o f rivals o r the standardization process. The standardization process tests standardized average abnorm al returns as com puted in Equation #7 in C hapter III rather than actual returns for significance. This process elim inates inefficiencies in statistical tests w hich result from unequal variances am ong returns o f the sam pled firms. In looking at the overall results from both sam ples. 22 out o f 41 industries tested significant for average abnorm al industry-w ide equity adjustm ents resulting from an an ­ n ouncem ent o f a "subject to" qualified or "going concern" audit opinion by the an ­ nouncing firm. O f these 22 industries, 15 industries w ere dom inated by a positive equity adjustm ent (com petitive effect) and 7 w ere dom inated by a negative average eq ­ uity adjustm ent (contagion effect). T he two sam ples share tw o industries w hich have sim ilar dom inant com petitive effects in both sam ples: C om puter and O ffice E quipm ent (S IC 357) and Electric Services (SIC 491). A n appendix to this study presents the re­ su lts o f the daily abnorm al returns tested for significance for each industry by sam ple. Im plications o f T est R esults for N on-A nnouncing Firm s Inform ation transfer studies seek to determ ine the existence o f interdependencies am ong firm s' stock returns. O f course, announcem ents affecting the m arket as a w hole. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 129 such as interest rate changes in the econom y, will im pact ju st about every security in the m arket. H ow ever, inform ation transfer studies specifically investigate relationships at m ore o f a m icro level than a m acro level by m easuring the im pact o f an announcem ent by a single firm on its rivals, suppliers or dem anders o f its products. In this study, the inform ation transfer centers on the specific inform ation content o f a "subject to" quali­ fied or "going concern" audit opinion for the announcing firm and for the industry rivals o f the announcing firm. The im pact o f the unfavorable audit opinion announcem ent on the announcing firm has already been docum ented in the literature by D opuch et al. [ 1] and Jones [2], The m ain contribution o f this study is the m easurable im pact o f the unfa­ vorable audit opinion announcem ents on the rivals o f the announcing firm. T he o u t­ com e o f the hypotheses tests perform ed on the rivals o f the announcing firm s receiving "subject to" qualified and "going concern" audit opinions has answ ered the m ain re­ search question proposed; that is. unfavorable audit opinion announcem ents have a sig­ nificant im pact on the announcing firm's rivals. In C hapter III. the a priori expectations about the results o f the hypotheses tests em ployed to determ ine the existence o f significant average abnorm al returns to rivals as a result o f unfavorable audit opinion announcem ents were presented. These ex pecta­ tions em phasized the inability to m ake predictions about w hich effect (com petitive or contagion) w ould dom inate an industry without a specific release o f inform ation regard­ ing the cause o f the com pany's troubles, but they did suggest that som e industries in the sam ple w ould provide evidence o f the existence o f significant average abnorm al returns R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 130 to rival firm s o f those m aking unfavorable audit opinion announcem ents. T hese ex p ec­ tations w ere based on research focusing on the existence o f inform ational transfers from bankruptcy announcem ents (see Lang and Stulz [11]) and earnings announcem ents (see B annister [12]). Finally, the expectations o f the results o f the hypotheses testing o f an ­ nouncing firm s' abnorm al returns and the hypotheses testing o f rival firms' abnorm al re­ turns differ in that the literature presents a clear theoretical basis for anticipating the d irection and m agnitude o f abnorm al returns for the announcing firms (see D opuch et al. [1] and Jones [2]); however, the literature available for predicting the results o f tests o f rivals' abnorm al returns is not clear at all. Lang and Stulz [11] found that 7 out o f 41 industries show ed statistically significant evidence o f com petitive or contagion effects associated w ith bankruptcy announcem ents. They provided evidence that inform ational transfers did exist, but provide no a priori m eans o f estim ating w hich industries w ill be affected and to w hat degree. Sim ilarly, this study sim ply expects a significant reaction to exist in som e industries w hich could go either w ay (contagion or com petitive). As stated in C hapter III, "C ontagion and/or com petitive effects are expected to exist in dif­ fering m agnitudes in different industries." O bviously, the announcem ent m ay contain inform ation w hich enables m arket participants to determ ine beforehand w hich effect w ill dom inate. For exam ple, if. in the m edia press release, the reason given for the un ­ favorable audit opinion is that earnings are substantially less because substitute goods are relatively cheaper than the good produced in that industry, then the reduction in de­ m and for the good should affect all producers o f that good and a contagion effect should R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. dom inate. If, on the other hand, the announcem ent reveals that the specific com pany in q u estion has becom e less profitable due to a large turnover in key m anagers, then a com petitive effect should ensue. T he hypotheses tests results for the significance o f abnorm al average returns to industry rivals o v er the event w indow substantiates the expectations o f this study. The im plications o f the hypotheses tests and results are straightforw ard: ( I ) the m arket gleans inform ation about rival firms from announcem ents revealing the issuance o f un­ favorable audit opinions; and (2) based on how the m arket interprets the inform ation, significant industry-w ide equity adjustm ents occur m ore often than not w hen the audit o pinion announcem ent contains inform ation about a firm 's ability to continue as a going concern. Out o f the 41 industries represented in both sam ples. 19 reported negative ab ­ norm al returns o f w hich 7. or 37% . w ere statistically significant, and 22 show ed positive abnorm al returns, w ith 15. o r 68% , testing significant. A bout 68% o f the industries w hich produced significant average stock price reactions in this study resulted in a tran sfer o f profitability from the announcing firm to its rivals (com petitive effect). The results are near-evenly distributed as to w hich direction the abnorm al returns w ent for all announcem ents (ignoring significance). Som e intuitive explanations o f the results m ay be offered. O ne w ould expect certain industries to exhibit com petitive effects and others to exhibit contagion effects m erely by the nature o f the industry itself. A reasonable expectation w ould be for in­ d u stries that provide services, equipm ent and supplies to experience a com petitive R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 132 effect. T he reason for this expectation is that custom ers w hich rely on such com panies for inputs o r service w ould desire solid, long-term relationships w ith these type firms. F or exam ple, if a custom er is purchasing a product such as com puter equipm ent, that custo m er w ould depend on the firm to stay in business to m eet their servicing needs. T he sam e w ould hold true for firm s providing services and necessary production inputs. T his reasoning is borne out by the follow ing industries that experienced a dom inating com petitive effect as a result o f the audit opinion announcem ent: (1) Oil and G as Ex­ traction, (2) Oil and G as Field Services. (3) C onstruction. M ining, and Oil and Gas E quipm ent, (4) C om puter and O ffice Equipm ent, (5) C om m unication Equipm ent, (6) S urgical. M edical and D ental Instrum ents, (7) Electric S ervices, (8) Electrical G o ods-w holesale. (9) Investors, not elsew here classified (nec). (10) M iscellaneous B usiness S ervices. (11) H om e Health Care Services, (12) M iscellaneous H om e and A l­ lied S ervices, nec, and (13) R esearch, D evelopm ent and T esting Services. Every indus­ try ex p eriencing a com petitive effect falls into the services, equipm ent, and supplies categories except the Investors, nec industry. A pparently, financial m arket participants expected custom ers o f these announcing firms to find a m ore dependable firm with w hich to establish long-term relations. A con tag io n effect dom inates an industry w hen the m arket interprets the unfa­ vorable audit opinion to reveal inform ation about a factor affecting the industry as a w hole. E xam ples o f inform ation that w ould produce a contagion effect w ould be in­ creases in resource prices o r input costs, legal settlem ents that cause investors to R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 133 reassess the risk o f other firm s in the industry, or new technologically advanced su b sti­ tutes for the good produced in the industry. The sam pled industries that experienced a dom inant contagion effect as a result o f the unfavorable audit opinion announcem ents are as follows: (1) M obile Hom es, (2) D rugs and Pharm aceutical. (3) Electronic C om ­ ponents and A ccessories, (4) A ircraft and Ship Building, (5) Photographic Equipm ent and Supply, (6) Electric and O ther Services C om bined, and (7) Eating and D rinking Places. Intuitive explanations for the industries affected by contagion are m ore difficult than those experiencing com petitive effects because contagion results from a co m m u n i­ cated industry-side problem . The M obile H om e industry m ay have experienced reduc­ tions in profitability as consum ers purchased perm anent dw ellings during the econom ic expansion. The Drugs and Pharm aceutical industry may have been affected by a legal settlem ent, causing financial m arket participants to reassess the riskiness o f the industry. T he A ircraft and Shipbuilding and Electronic C om ponents and A ccessories industries m ay have experienced lost contracts due to defense spending cuts by one or m ore na­ tions. The Photographic Equipm ent and Supply industry m ay have succum bed to inform ation-based substitutes, such as the internet. W hatever the reason, the m arket ap­ pears to have interpreted the unfavorable audit outcom e o f the announcing firm to re­ lease inform ation adversely affecting these industries. In sum m ary, this study has provided evidence to suggest that both "subject to" qualified and "going concern" audit opinions introduce new inform ation into the m arket about the industry rivals o f the announcing firm w hen the audit results are m ade public R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 134 by the m ajor financial press. The evidence o f the existence o f this "new" inform ation is m easured by the significant average stock price reactions o f rival firm s during the event w indow period. Som e o f the industries represented exhibited significant positive stock price reactions on average (com petitive effect) w hile other industries exhibited signifi­ cant negative stock price reactions on average (contagion effect). These findings are as expected and represent an addition to the existing Finance and A ccounting literature on the subjects o f "A udit O pinion Inform ation Content" and "Intra-Industry Inform ation Transfers." Explaining the Abnormal Returns to Rivals In the previous section, the cum ulative average abnorm al returns for rivals o f firms announcing unfavorable audit opinions were tested for statistical significance. O ut o f the 41 industries represented. 15 tested significant for dom inating com petitive ef­ fects and 7 tested significant for dom inating contagion effects. In C hapter III, a regres­ sion m odel w as presented that w as used in an attem pt to explain the cum ulative abnor­ m al returns over the event w indow to the rival firms. T h is regression w as estim ated on those rivals in industries w here com petitive effects dom inated and again on those rivals in industries where contagion effects dom inated. In the next sections, the results o f the regression outcom e are presented. At this point, it should be noted that a regression w as not run on the abnorm al returns to announcing firm s. T he m ain reason for not running a regression on the R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 135 abnorm al returns to firm s receiving audit opinions is that these returns have been ade­ q u ately addressed in previous research in the existing literature. Furtherm ore the sam ­ ple used in this study only contained firms w hich had their unfavorable audit opinion receive attention in the m ajor financial press. R esearch show s little or no abnorm al stock price reaction to qualified audit opinions around the 10-K and annual statem ent release dates (C h o w and Rice [13) and Dodd el al. [14]). D opuch el al. [1] and Jones [2] show that the m ajority o f significant negative abnorm al returns accrue specifically to those firm s w hich had their opinion m ade public by the financial press. Jones [2] tests a com bination o f firm s w hich had official announcem ents in the financial m edia with those that did not isolate this effect. Based on the outcom e o f these studies, those that did not concentrate on firm s which were given attention in the financial m edia (C how and R ice [13] and D odd et al. [14]) and those that did (D opuch el al. [1] and Jones [2]), o bviously the m ajor contributor to the significance o f the abnorm al returns to announc­ ing firm s is the announcem ent itself. The other significant contributors to those returns is the existence o f a concurrent earnings announcem ent, as show n by the D opuch e t al. [1] study, and a logistical regression variable estim ating the probability o f a firm receiv­ ing a "going concern" qualification, as show n by the Jones [2] study. The earnings an­ nouncem ent variable has been included in explaining the significant abnorm al returns to rivals, seeing o th er studies, such as B annister [12], show the effect o f earnings an­ nouncem ents on rival firm s. Because this study only uses those firm s w hich had an ­ nouncem ents in the m ajor financial press, a regression to explain the returns to R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 136 an nouncing firm s becom es redundant and w ould serve only to sidetrack this research from the specific research question at hand: "Do announcem ents o f unfavorable audit o p inions provide inform ation about the rivals o f the announcing firm ?" Sam ple: C om petitive Industries and C ontagion Industries Table 4-3 and T able 4-4 presented the results o f the tests o f significance o f aver­ age abnorm al returns to industry rivals. In those tables, 41 industries w ere represented w hich included 3.511 rival firm s. T he com petitive industry sam ple and contagion in­ dustry sam ple w ere based on the inform ation presented in the above m entioned tables. For a rival firm to be used in the regression, the variables m entioned in C hapter III. to be used as independent predictors, had to be available for that com pany. These vari­ ables w ere com puted based on C O M P U ST A T data and C R S P data, m eaning rivals had to have all inform ation available for the desired tim e periods in both data sets in order to be used as an observation in the regression. The follow ing variables w ere selected as explanatory (independent) variables: (1) the earnings correlation betw een the announcing firm and each rival firm : (2) the natural log o f total assets for each rival firm : (3) the m arket-to-book value o f equity for each rival firm: (4) the H erfindahl Index o f the industry: (5) the debt-to-total assets ratio for each rival firm; (6) the cum ulative abnorm al return to the announcing firm over the event w indow ; (7) the exchange on w hich the rival firm 's equity is traded; R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. m 1 (8) the business cycle during w hich the announcem ent o f the unfavorable audit opinion w as m ade; and (9) the existence o f a concurrent earnings announcem ent along w ith the audit opinion announcem ent. The regression equation applied was given as Equation #9 in C hapter III. D etails o f the sam ple com positions for the com petitive and contagion effects used to estim ate the re­ gressions are given in T able 4-5 and Table 4-6 below. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 138 Table 4-5 SA M PLE O F IN D U STRIES USED IN T H E R EG R ESSIO N IN W HICH A SIG N IFIC A N T C O M PE T IT IV E E F FE C T D O M IN A TED N um ber o f A nnouncing N um ber o f Rival F irm s: SIC: Industry: 131 Oil & G as Extraction 2 223 138 Oil & G as Field Services 2 36 353 C onstruction, M ining,, Oil & G as Equipm ent 3 15 357 C om puter and O ffice Equipm ent 5 352 366 C om m unication Equipm ent 1 54 384 Surgical, M edical & Dental Instrum ents 3 89 491 Electric Services 6 379 506 Electrical G oods-W holesale 1 10 679 Investors, nec 2 82 738 M iscellaneous Business Services 1 16 808 Hom e H ealth Care Services 1 17 809 M iscellaneous Hom e & Allied Services, nec 1 15 873 R esearch. D evelopm ent & Testing Services 1 30 Totals: 13 3-D igit-SIC -D efined Industries Firms: 29 firms 1,318 firm s R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 139 Table 4-6 SA M PL E O F IN D U STR IES USED IN THE R EG R ESSIO N IN W H IC H A SIG N IFIC A N T C O N TA G IO N EFFECT D O M IN A TED N um ber o f A nnouncing N um ber o f Rival Firm s: SIC: Industry: 245 M obile H om es 1 10 283 D rugs & Pharm aceutical 2 114 367 Electronic C om ponents & A ccessories 3 167 372 A ircraft. Ship B uilding 3 17 386 Photographic Equipm ent & Supply 1 9 493 Electric & O ther Services C om bined 1 30 581 Eating & D rinking Places -n Totals: 7 3-D igit-SIC -D efined Industries 14 firms Firms: 95 442 firm s A Priori Expectations: C om petitive Effect As can be seen in Table 4-5. the com petitive effect sam ple contains a little m ore than 75% o f the total observations (firm s) associated with industries w hich show ed sig­ nificant average abnorm al returns during the event window . The contagion effect there­ fore accounts for the other 25 % o f the observations. The regression w hich w as estim ated to explain the com petitive effect used the cum ulative abnorm al returns to the rival firm over the event w indow as the dependent variable. A review o f the a p rio ri expectations m ade about the independent variables for the regression estim ated on those rivals in industries encountering a significant com petitive effect outlined in C h ap ter III is presented in the follow ing paragraphs. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 140 In the event a com petitive effect dom inated an industry, the variable used as a m easure o f how levered the rival firm was at the tim e o f the announcem ent was the book debt-to-total assets ratio. This variable was expected to have an am biguous effect because, w hile greater changes in equity values are associated w ith changes in asset val­ ues w hen the debt ratio is higher, those firms which already w ere too leveraged were not expected to be in a position to take full advantage o f the transfer o f m arket share brought about by a new com petitive environm ent in the industry. T herefore, the overall effect o f leverage on the abnorm al returns experienced by rival firm s could not be spe­ cifically anticipated. The H erfindahl Index, which m easures the level o f com petition w ithin an indus­ try, w as expected to have a positive effect on the returns to rival firm s. Industries in w hich few firm s account for the m ajority o f m arket share tend to have higher values for the H erfindahl Index. There should be a greater transfer o f profitability in industries w hich have higher levels o f sales concentration because the effects accrue to fewer firms. The size o f the rival firm, proxied by the natural log o f total assets, was expected to contribute positively to the rival firm's positive abnorm al returns. Firm s w hich are larger are expected to be in a better position to increase m arket share due to better ac­ cess to the necessary funding required for growth. A lso, sm aller firm s are expected to have less access to grow th funding. This would suggest a higher transfer o f w ealth to R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 141 larger firm s than sm aller firm s, ceteris paribus, should a com petitive effect dom inate an industry, The dum m y variable used to account for the effects o f contractionary periods in the econom y took on a value o f 1 during recessions. B ecause investors are expected to be m ore risk-averse during econom ic contractions, this variable was expected to be negatively related to the positive abnorm al returns experienced by industry rivals. A proxy for a firm 's grow th options is the m arket-to-book-value o f equity ratio. G aver and G aver [15] and B arclay and Sm ith [16] indicated that the higher degree o f in­ form ational asym m etry associated w ith firms with m ore grow th options causes larger changes in equity w hen a com petitive effect dom inates an industry. For this reason, a positive relationship w as expected betw een firm s w ith m ore grow th options and the positive change in equity value brought on by a new com petitive condition in the industry. The dum m y variable capturing the effects o f the exchange on w hich a rival firm ’s stock is traded took on the value o f 1 w henever that firm 's equity w as traded on the N A S D A Q and 0 if that firm 's equity was traded on the N Y S E or A M EX . T he greater inform ational asym m etries associated w ith firm s traded on the N A S D A Q w as expected to result in a larger percentage change in equity value w hen the com petitive ef­ fect d om inates an industry. A concurrent earnings announcem ent m ade by firm s experiencing doubts about th eir ability to continue as a going concern w as expected to be negative, as seen by the R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 142 effects o f these earnings announcem ents on announcing firm s in the D opuch et al. [ 1] study. F or this reason, larger losses o f m arket share to the announcing firm were ex ­ pected to be associated w ith larger gains to rivals. Therefore, w hen the earnings an ­ nouncem ent dum m y takes on the value o f 1. m eaning that the earnings w ere released at the tim e o f the audit opinion, the positive equity changes to rivals w ere expected to be greater. The cum ulative abnorm al return to firms announcing an unfavorable audit o p in ­ ion has tested negative and significant. For this reason, the larger the m agnitude o f these losses to announcing firms, the larger the gains should be to rivals. H ence, one a s­ sum es a negative relationship between the negative returns to announcing firm s and the positive returns to rivals. The correlation betw een returns o f the announcing firm and those o f rivals is a m easure o f the sim ilarity o f the cash flows o f those firms. Rival firm s w hich have earn ­ ings stream s m ore sim ilar to the announcing firm were expected to have sim ilar assets, and thus, higher gains w hen a firm -specific event caused a loss o f m arket share for the announcing firm. C onsequently, higher return correlations betw een rivals and ann o u n c­ ing firms w ere expected to have a positive im pact on the equity adjustm ent o f rivals. R egression R esults for the C om petitive Effect The regression outlined in Equation #9 in C hapter III w as executed on the rivals abnorm al returns in those industries w hich exhibited significant com petitive effects. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 143 B ecause there w ere no announcem ents which occurred during an econom ic contraction, the variable used to extract the effects o f recessions on the abnorm al returns to rivals w as om itted in the regression. Therefore, the regression used to explain the com petitive effect utilized 8 o f the 9 explanatory variables discussed above. As seen in T able 4-5, the sam ple o f rival firms included in the regression esti­ m ated on the firm s in industries exhibiting significant com petitive effects contained 1,318 observations and covered 13 3-digit-SIC -defined industries. The regression was not significant. O f the eight variables used in the regression, only the variable used to m easure industry concentration (the Herfindahl Index) was significant. For this reason, inferences will not be m ade for the regression estim ated on the com petitive effect sam ­ ple o f rival firm s. Furtherm ore, tests for m ulticollinearity, heteroskedasticity and nor­ m ality indicated no severe problem s. The regression results are reported in Table 4-7 below. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 144 Table 4 -7 R EG R ESSIO N R ESU LTS EX PLA IN IN G A V ER A G E IN D U STRY A B N O R M A L R ETU R N S FOR C O M P E T IT IV E E FFEC TS A SSO C IA TED W ITH U N F A V O R A B L E A U D IT O PIN IO N A N N O U N C E M E N T S O V ER TH E P E R IO D 1 9 8 4 - 1996 C A R . 1.1 = a + P i ( A / J W ) + P 2(HI) + p 3(ln TA) + p 4(£W 7) + p f (EA) +pfi( £ v ) + p 7 (cvi/?,;) + p 8( C 0 M ) + s R egression Variables: C oefficient: t-value: -0.0044 -0.51 G row th O ptions 0.0000 0.85 Industry C oncentration 0.0276 2.19** Intercept Firm Size -0.0001 -0.13 Firm Leverage 0.0001 0.70 E arnings A nnouncem ent 0.0021 0.57 Exchange 0.0005 0.11 -0.0251 -1.36 0.0133 0.68 A nnouncing Firm 's A bnorm al Return R eturns C orrelation M odel F-value = 1.011 R: = 0.006 Adj. R: = 0.000 N = 1318 *** S ignificant at the 1% level ** S ignificant at the 5% level * S ig n ificant at the 10% level R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 145 A P riori Expectations: C ontagion Effect A contagion effect occurs in an industry w henever the reason for the "bad" news being announced by the announcing firm is determ ined by the m arket to be an event w hich will produce industry-w ide repercussions. This results in a significant average negative abnorm al return w hich dom inates in the industry returns over the event w in­ dow . As w ith the com petitive effect exam ined above, the regression w hich was esti­ m ated to explain the contagion effect used the cum ulative abnorm al returns to the rival firm over the event w indow as the dependent variable. The a p rio ri expectations m ade about the independent variables for the regression estim ated on those rivals in industries encountering a significant contagion effect outlined in C hapter III are review ed in the follow ing paragraphs. The book debt-to-total assets ratio, used to m easure the degree to w hich a rival firm is leveraged, was expected to be negatively related to the cum ulative abnorm al re­ turn to rival firm s w hen a contagion effect dom inates an industry. Equity absorbs a de­ cline in the value o f assets o f a firm. Since low er values o f equity are associated with higher debt ratios, the negative percentage change in equity value will be greater for m ore leveraged firm s w hen a contagion effect reduces asset values in an industry. The industry concentration, as m easured by the H erfindahl Index, w as expected to have an am biguous effect on the returns to rival firm s in the event o f a contagion ef­ fect. The degree to w hich com petition exists in an industry w as not expected to have a defined effect o n rivals in the event o f contagion effects because there is no transfer o f R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 146 profitability from o n e firm to others. For exam ple, a decrease in dem and for a product is expected to affect m ost firm s in the industry in the sam e m anner, w hether the industry is highly com petitive or is not com petitive at all. These a priori expectations w ere in accordance with th o se outlined by Lang and Stulz [11] in a study exam ining the effects o f bankruptcy announcem ents on industry rivals. T he size o f th e rival firm was expected to reduce the m agnitude o f the equity re­ valuation o f that firm in the event o f a contagion effect. This is in accordance w ith the G raham and King [ 17] study which assum es less inform ation to be available for sm aller firm s than for larger firms. Therefore, the sm aller a firm , the larger the uncertainty ab out its ability to overcom e adverse industry conditions and the larger the percentage reduction in its eq u ity value. Conversely, the larger the firm, the greater will be its abil­ ity to w ithstand a setback and the sm aller should be the percentage equity adjustm ent re­ su ltin g from a release o f "bad" news. C ontractionary periods in the econom y were expected to produce an adjustm ent to equity values as th e outlook o f investors was dam pened by general econom ic co n d i­ tions. This view is supported by R ajagopal [18]. For this reason, the negative expecta­ tions w hich are already built into stock prices during econom ic contractions should lessen the "surprise" effect o f "bad" news. Therefore, econom ic contractions w ere ex ­ pected to be associated with sm aller negative equity adjustm ents w hen a contagion ef­ fect dom inated an industry. Since the dum m y value o f 1 is assigned during tim es o f R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 147 recessions in the econom y, there should be a positive relationship betw een this variable and the contagion effect. In keeping w ith G aver and G aver [15] and Barclay and Sm ith [16], higher m arket-to-book equity ratios are indicative o f higher degrees o f inform ational asym m e­ try associated w ith a firm. For this reason, a negative relationship w as expected to exist betw een firm s w ith m ore grow th options and the negative change in equity values asso­ ciated w ith contagion effects. On the other hand, it w as argued that firm s w ith m ore grow th options m ay in fact experience sm aller negative equity adjustm ents in the event o f contagion because these firm s m ay have the ability to expand into areas not affected by the event responsible for negative industry-w ide equity adjustm ents. T his view , how ever, is not supported by the literature. G reater inform ational asym m etries are associated w ith firm s traded on the N A SD A Q as com pared to those traded on the NYSE or the A M EX . F or this reason, firm s which are traded over-the-counter were expected to experience larger negative percentage changes in equity value when the contagion effect dom inates an industry. Since the dum m y variable used to extract the effects o f the exchange on w hich the rival firm is traded is set to I w hen the rival is traded on the N A SD A Q , there w as expected to be a negative relationship betw een this variable and the abnorm al returns to rivals. As m entioned in the section review ing the a p rio ri expectations associated with com petitive effects, a concurrent earnings announcem ent m ade by firm s experiencing doubts about their ability to continue as a going concern is expected to be negative, as R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 148 seen by the effects o f these earnings announcem ents on announcing firms in the Dopuch et al. [ 1] study. On this prem ise, and based on the results presented by D opuch et al. [1 ], the earnings announcem ent dum m y should be negatively related to the negative ab ­ norm al returns experienced by rival firms in industries suffering contagion effects. T herefore, when the earnings announcem ent dum m y takes on the value o f 1, m eaning that the earnings were released at the time o f the audit opinion, the negative percentage equity value changes to rivals were expected to be greater. Earlier in this study, announcing firms w hich have their unfavorable audit opin­ ions m ade public in the financial press were show n to experience significant negative stock price changes over the event window. A larger change in equity value to the an­ nouncing firm is associated with a greater elem ent o f "surprise" surrounding the an­ nouncem ent. For this reason, the larger the m agnitude o f these losses to announcing firm s, the larger should be the equity adjustm ent to rivals as well, w henever the infor­ m ation released exposes new industry-w ide adverse econom ic conditions. Hence, there is a positive relationship expected between the negative returns to announcing firm s and the negative returns to rivals. The earnings correlation betw een the announcing firm and each industry rival w as anticipated to have a negative relationship w ith the contagion effect. Recall that ri­ val firm s w hich have earnings stream s m ore sim ilar to the announcing firm (higher cor­ relations betw een returns) should have sim ilar assets. For this reason, higher R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 149 correlations between returns w ere expected to be associated with larger negative equity adjustm ents to rivals. R egression Results for the C ontagion Effect The regression estim ated to explain the contagion effect brought about by m edia announcem ents o f "subject to" qualified and "going concern" audit opinions is based on E quation #9 in C hapter III. The sam ple used for the regression consists o f the group o f rival firm s and industries presented in Table 4-6. There were 442 firms in this sam ple representing 7 different 3-digit-S lC -defined industries. The final regression m odel contains six explanatory variables and the overall m odel w as significant at the .01 level. The significant variables in the m odel include: (1) the m arket-to-book equity value ratio, which is a proxy for a rival firm ’s grow th o p ­ tions; (2) the H erfindahl index value, which m easured the degree o f com petition in the industry; (3) the natural log o f total assets, a proxy for firm size; (4) the degree o f lever­ age o f the rival firm as m easured by its book debt-to-total assets ratio; (5) the dum m y variable used to extract the effects o f concurrent earnings announcem ents m ade at the tim e o f the unfavorable audit opinion announcem ent; and (6) the dum m y variable used to extract the effect o f the business cycle, which w as set to 1 during tim es o f econom ic contraction. A lso, the intercept term was significant. In addition, tests for m ulticollinearity. heteroskedasticity and norm ality provided no evidence o f severe problem s. The regression results are show n in Table 4-8 and are explained below. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 150 T able 4-8 R EG R ESSIO N R ESU LTS EX PLA IN IN G A V E R A G E IN D U STR Y A B N O R M A L RETU R N S FOR C O N TA G IO N EFFEC TS A SSO C IA T E D W ITH U N F A V O R A B L E A U D IT OPIN IO N A N N O U N C E M E N T S O V E R TH E PER IO D 1 9 8 4 -1 9 9 6 C A R -U = a + p i ( MV/ BV) + P 2(/W) + p 3(In TA) + Pi ( DAT) + p 5( £ 4 ) + p fi(£C yc) + e R egression Variables: Coefficient: t-value: Intercept -0.0342 -3.06*** G row th O ptions 0.0012 1.84* Industry C oncentration 0.0634 2.21** F irm Size 0.0041 2.20** Firm Leverage -0.0003 -2.00** E arnings A nnouncem ent -0.0302 -1.74* B usiness C ycle 0.0365 2.12** M odel F-value = 2.96*** R2 = 0.04 A dj. R2 = 0.03 N = 442 *** S ignificant at the 1% level ** S ignificant at the 5% level * S ignificant at the 10% level R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 151 T he intercept term has a negative coefficient o f -0.0342 and is significant. T heoretically, the intercept term is the average value o f the dependent variable (cum ula­ tive abnorm al returns to rivals in this study) w hen all o f the independent variables are set to zero. Its significance m ay be due to the low R 2 value. T he R2 m easure specifi­ cally show s the reduction in estim ation error attained by using the estim ated m odel as a prediction tool to estim ate the value o f the dependent variable rather than the m ean o f the dependent variable. The next significant variable is the m arket-to-book value o f equity ratio w hich proxied as a m easure o f grow th options for the rival firm . It is significant at the .10 level and has a positive m agnitudinal influence on the cum ulative abnorm al returns to the rival firm. This m eans that firm s with m ore grow th options tend to have low er negative abnorm al returns w hen a contagion effect dom inates an industry. A p rio ri. a n egative effect w as anticipated because asym m etrical inform ation is considered to be g reater for firm s w ith m ore grow th options, as set forth by G aver and G aver [15] and B arclay and Sm ith [16]. H ow ever, the positive coefficient w ould suggest that firm s w ith m ore grow th options suffer less in the event o f contagion due to an ability to d iv er­ sify outside o f the problem areas in the industry, or because those firm s are already bet­ ter diversified. The H erfindahl Index, w hich quantifies the com petitive nature o f the industry, w as significant at the .05 level and had a positive (m agnitude reducing) effect on the ri­ val firm 's cum ulative negative abnorm al returns. The a p rio ri expectations concerning R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 152 the effects o f this variable were that industry concentration w ould have an am biguous effect, due to a lack o f literature to explain its anticipated effect one way or the other. A possible explanation for the positive, negative abnorm al return reducing effect o f the variable is that perfectly com petitive firms' pricing policies are dictated by the m arket. H ow ever, oligopolies are able to charge different prices and com pete with advertising and product differentiation. For this reason, on average, all firm s in a m arket w ith low industry sales concentration should be equally affected, w hereas the negative effect upon firm s' equity values in industries which engage in fierce com petition for m arket share should be less pronounced, seeing these firm s have m ore control over pricing and output decisions. Furtherm ore, perfectly com petitive firm s seek to operate at m inim um costs, since they can sell all output at the m arket price. O ligopolies, on the other hand, operate with excess capacity and not at m inim um costs. This w ould put oligopolistic firm s in a better position to adjust production costs, if need be. to rem ain profitable. The positive value o f the regression coefficient w ould suggest that there was som e transfer o f profitability even though a contagion effect dom inated the industry. In this light, contagion m ay be explained as the value o f rivals that cannot be attributed to a transfer o f profitability. T he variable proxying firm size (the natural log o f total assets) is also significant at a .05 level. The effect o f this variable on the cum ulative negative abnorm al returns to rival firm s is positive, as anticipated. Larger firms have less asym m etrical inform ation associated w ith them and also have superior access to funds during em ergencies than do R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 153 sm aller firms, m aking them less risky. For this reason, the sm aller firm s encounter greater negative percentage equity adjustm ents during contagion than larger firm s. This reasoning follow s the basic risk-retum relationship. The debt-to-total assets ratio was used as a m easure o f the financial leverage utilized by a rival firm. This variable was significant at a .05 alpha level and had a negative im pact on the cum ulative abnorm al returns experienced by a rival firm w hen the contagion effect dom inated an industry. The sm all regression coefficient estim ate o f -.0003 reflects the fact that C O M PU ST A T reports these values as percentages (i.e., 50 rather than .50). The inherent m eaning o f a negative im pact is that larger negative aver­ age abnorm al returns are experienced by firm s having a higher degree o f leverage. T his im pact was expected due to the greater elasticity o f equity value associated w ith higher degrees o f leverage and the increased costs o f financial distress resulting from contagion effects as sum m arized by Lang and S tulz [11]. The earnings announcem ent dum m y variable took on a value o f 1 w henever a concurrent release o f earnings w as m ade w ith the release o f the auditor's opinion. It is significant at the .10 level and has a negative, com pounding effect on the cum ulative negative abnorm al returns to the rival firm s w hen the contagion effect dom inates the in­ dustry. The study perform ed by D opuch et al. [1] found a significant 3.1% additional loss to accrue to firm s announcing "subject to" qualified audit opinions in the financial press w hen those announcem ents also included an earnings disclosure. For this reason, earnings announcem ents coinciding w ith the unfavorable audit opinion announcem ents R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 154 (w hich m ainly reflect doubts as to w hether a firm will rem ain in existence) are assum ed to be negative. A p r io r i, the com pounded effect o f negative earnings releases and unfa­ vorable audit opinion announcem ents was expected to m agnify the negative cum ulative abnorm al returns to rival firms. This im pact has been substantiated by the outcom e o f this study. F inally, the dum m y variable used to extract the effects o f an econom y in reces­ sion w as significant at the .05 alpha level and had a positive coefficient. B ecause the dum m y w as set to 1 during tim es o f econom ic contraction, the m agnitudinal im pact on the cum ulative negative abnorm al returns was reduced during recessions. This is in keeping w ith the a p r io r i expectations where the market w as expected to have already incorporated recessionary expectations into security prices once the econom ic dow nturn began. F or this reason, the unfavorable results o f the audit w as m uch less o f a surprise than it w ould have been during an expansion, and therefore resulted in less o f a reduc­ tion in equity value than it w ould have during an expansion. R ajagopal [18] finds a sim ilar result in his investigation into the im pact o f bond-rating dow ngrades on industry rivals. T here w ere three variables originally included in the regression w hich did not significantly contribute to the abnorm al returns experienced by the rivals. These are the abnorm al returns to the announcing firm, the exchange dum m y variable and the earn­ ings correlation betw een the announcing firm and the rival firm. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 155 Summary T he em pirical results o f the hypotheses tests w ere presented first in this chapter. The results o f the hypothesis test for significant negative average abnorm al returns over the event w indow o f one day before up to and through one day after the announcem ent date for the firm s announcing "subject to" qualified audit opinion results in the m ajor fi­ nancial m edia w ere sim ilar to the results o f the sam e tests perform ed by other studies (specifically D opuch et al. [1 ] and Jones [2]). That is, the abnorm al returns over the event period w ere negative and highly significant using a one-tailed t-test. This portion o f the study has received substantial prior attention in the literature, the only difference being the sam ple period from w hich the "subject to" qualified audit opinions were selected. S im ilar hypotheses tests were perform ed to determ ine the existence and signifi­ cance o f cum ulative negative abnorm al returns to firms receiving "going concern" spe­ cial language added to their unqualified audit opinion m ade public in the financial m e­ dia. As expected, these firm s encountered highly significant negative equity adjusm ents on average as a result o f the announcem ent over the event w indow beginning one day before and ending one day after the announcem ent date. T hese results represent an ad­ dition to the existing literature in that the im pact o f such announcem ents on firm s re­ ceiving "going concern" special language has not been investigated since the audit o pin­ ion change initiated in 1989. The sam ple period over w hich these firm s w ere draw n was from 1989-1996. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 156 N ext, the industry rivals o f those firm s receiving m edia coverage o f th eir unfa­ vorable audit opinions w ere tested for stock price reactions to the announcem ent over the event w indow . Industries in w hich positive cum ulative standardized abnorm al re­ turns tested significant using a tw o-tailed t-test w ere concluded to have had a dom inant com petitive effect as a result o f the announcem ent. These am ounted to 15 o f the 41 in­ dustries tested for inform ational transfer effects. The industries w hich tested signifi­ cantly for abnorm al negative average returns w ere concluded to have had a contagion effect dom inate as a result o f the unfavorable audit opinion announcem ent. These am ounted to 7 o f the 41 industries tested for significant abnorm al returns resulting from intra-industry inform ation transfers associated w ith unfavorable audit opinion an ­ nouncem ents. O ver h a lf o f the industries tested had significant stock price reactions, suggesting that audit opinions contain inform ation relevant not only to the audited firm , but also to industry rivals. These results w ere as expected and represent an addition to the literature concerning inform ational transfers and the inform ation content o f "subject to" qualified and "going concern" audit opinions. Finally, the industries receiving significant equity adjustm ents on average w ere broken dow n into those w here the contagion effect dom inated and those w here the com ­ petitive effect dom inated. A regression was estim ated on each sam ple using firm specific and industry-specific variables to explain the cum ulative abnorm al returns ex ­ perienced by industry rivals. The regression executed on the com petitive sam ple was not significant and contained only the H erfindahl Index as a significant regressor. The R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 157 regression estim ated on the contagion sam ple was significant and contained six signifi­ cant explanatory variables. M ost o f the regressor coefficients supported the a p rio ri ex ­ pectations presented in C hapter III. Citations [1J Dopuch, N., R. W. H olthausen. and R. W. Leftw ich. "A bnorm al R eturns A ssociated W ith M edia D isclosures o f'S u b je c t to' Q ualified A udit O pinions." Journal o f A ccounting and E conom ics, 8, 1986, pp. 93-117. [2] Jones. F. L. "The Inform ation C ontent o f the A uditor's G oing C oncern Evalua­ tion." Journal o f A ccounting a n d Public Policy, 15, 1996. pp. 1-27. [3] D eBondt, W „ and R. Thaler. "D oes the Stock M arket O verreact?" Jo u rn a l o f Finance, 40, 1985, pp. 793-805. [4] Howe, J. "Evidence on Stock M arket O verreaction." F inancial A n a lysts J o u r­ nal, 42, 1986. pp. 74-77. [5] D eB ondt, W „ and R. Thaler. "Further Evidence on Investor O verreaction and Stock M arket Seasonality." Jo u rn a l o f Finance, 42, 1987, pp. 557-581. [6] Brown. K.. S. Tinic. and V. H arlow . "Risk A version. U ncertain Inform ation, and M arket Efficiency." Journal o f F inancial Econom ics, 22. 1988. pp. 355-385. [7] Brown, K., and V. H arlow. "M arket O verreaction: M agnitude and Intensity." Journal o f Portfolio M anagem ent, 14, 1988. pp. 6-13. [8] Firth, M. "The Im pact o f Earnings A nnouncem ents on the Share Price B ehavior o f Sim ilar Type Firm s." E conom ic Journal, 86. 1976. pp. 296-306. [9] Foster, G. "Intra-Industry Inform ation T ransfers A ssociated W ith Earnings R e­ leases." Journal o f A ccounting a n d Econom ics, 3. 1981. pp. 201-232. [10] C linch, G. J., and N. A. Sinclair. "Intra-Industry Inform ation R eleases." J o u r­ nal o f A ccounting a n d Econom ics, 9. 1987, pp. 89-106. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 158 [1 1J Lang, L. H „ and R. M. Stulz. "C ontagion and C om petitive Intra-Industry E ffects o f B ankruptcy A nnouncem ents: An Em pirical A nalysis." Jo u rn a l o f F inancial E conom ics. 32, 1992, pp. 45-60. [12] B annister, J. W . "Earnings Signals and Inter-Firm Inform ation Transfers." Jo u rn a l o f Business, Finance & A ccounting, 21, 1994, pp. 1127-1149. [13] C how , C. W .. and S. J. Rice. "Q ualified A udit O pinions and Share P rices-A n Investigation." A uditing. 1. 1982, pp. 35-53. [14] D odd, P., N. D opuch, R. W. Holthausen, and R. W. Leftw ich. "Q ualified A udit O pinions and Stock Prices: Inform ation Content. A nnouncem ent D ates and C oncurrent D isclosures." Journal o f A ccounting a n d Econom ics. 6, 1984, pp. 3-38. [15] G aver, J. J.. an d K. M. Gaver. "Additional Evidence on the A ssociation B etw een the Investm ent Opportunity Set and C orporate Financing. D ividend and C om pensation Policies." Journal o f A ccounting a n d Econom ics. 16. 1993, pp. 125-160. [16] B arclay, M. J., and C. W. Smith. "The M aturity Structure o f C orporate Debt." Jo u rn a l o f F inance. 50, 1995. pp. 6 0 9 -6 3 1. [17] G raham . R. C ., and R. D. King. "Industry Inform ation Transfers: T he E ffect o f Inform ation E nvironm ent." Journal o f Business, Finance & A ccounting. 23. 1996. pp. 1289-1306. [18] R ajagopal, S. "B ond D ow ngrades and Intra-Industry C ontagion and C om peti­ tive Effects: T he C ase o f Industrials." D issertation. M ississippi State U niver­ sity. M ay 1997. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. CHAPTER V SU M M A R Y , IMPLICATIONS, A N D CO N C LU SIO N S T his chapter concludes the study o f the im pact o f unfavorable audit opinions on rivals o f the announcing firm. The first section review s the m ethodology used in this research, the outcom e o f the statistical tests used to identify abnorm al returns, and the results o f the regression m odels estim ated to explain the variation in the abnorm al re­ turns to rivals. S ection tw o sum m arizes the conclusions, im plications and contributions o f the study. Finally, section three suggests extensions o f the study. In their decisions about capital allocation, investors utilize publicly available in­ form ation. including inform ation provided in the financial statem ents o f corporations. For this reason, the S ecurities and Exchange C om m ission requires all publicly held co r­ porations to subm it to annual independent audits o f their financial statem ents. These audits serve tw o prim ary purposes: (1) to provide financial statem ent users w ith infor­ m ation concerning the fairness and accuracy o f the financial statem ents, and (2) to e x ­ press an opinion regarding the health o f the com pany and its viability as a going concern. A udit opinions expressing doubts concerning a firm 's continued existence pro­ vide "new " inform ation about those firm s to m arket participants. 159 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 160 Studies perform ed by Dopuch et al. [1 ] and Jones [2] provide evidence o f the in­ form ation content o f "subject to" qualified and pre-1989 "going concern" audit o p in ­ ions. These studies show that the m easurable im pact o f these particular unfavorable audit opinions on the equity values o f the announcing firms is pronounced and m ore easily captured w hen the audit outcom e is revealed to the investing public via a story in the m ajor financial press such as the W all Street Journal. Still, the literature is lacking in investigations into this phenom enon for "going concern" audit opinions m ade public by the financial press follow ing the 1989 change in audit opinions definitions. A d d i­ tionally, no research appears to exist that exam ines the possibility o f intra-industry in­ form ational transfers associated with audit opinion announcem ents. F illing these voids in this area o f research is the prim ary contribution o f this study to the literature. The purpose o f this final chapter is fourfold: (1) to present an overview o f the research m ethodology and results o f the statistical tests o f the hypotheses relevant to this study; (2) to present the interpretation o f the findings o f this study; (3) to discuss the im plica­ tions, conclusions, and contributions o f this study; and (4) to suggest extensions to the com pleted research. Overview of the Methodology and Results This project investigated the existence o f intra-industry inform ation transfers re­ sulting from audit opinion announcem ents in the m ajor financial press. As needed in all event studies, particular care w as taken in defining the appropriate date to capture the announcem ent effect. T he w idely-regarded single index m odel w as used for the R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 161 abnorm al return-generating process. T he follow ing sections provide a b rief review o f the sam pling process and data analysis, followed by a presentation o f the results o f the hypotheses tests and the estim ated regressions. R eview o f the Sam ple and D ata A nalysis The sam ple o f announcing firms included those w ith "subject to" qualified audit opinion announcem ents for the period from 1984 through 1988 and "going concern" audit opinion announcem ents for the period from 1989 through 1996. appearing in the W all Street Jo u rn a l Index o r the Business Wire. Based on this sam ple o f announcing firms, associated sam ples o f industry rivals were retrieved from the C R SP database based on four-digit SIC codes. Because C RSP data is required for the com putation and testing o f the abnorm al returns o f both announcing and non-announcing firm s, the sam ­ ples were lim ited to firm s w ith equity traded on the NY SE, the A M EX and the N A SD A Q . Standard event study m ethodology was em ployed to com pute and test the stan­ dardized abnorm al returns o f the announcing firms for an event w indow beginning the day before and ending the day after the audit opinion announcem ent date. T he sam ple o f announcing firm s used to perform these statistical tests w as separated into tw o cate­ gories: (1) "subject to" qualified opinions issued before 1989. and (2) "going concern" opinions issued after the audit opinion changes were initiated in 1989. To determ ine w hether financial m arkets obtained inform ation about rivals through the announcem ent o f the firm receiving the unfavorable audit opinion, com parable tests for the sam e event R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 162 w indow w ere perform ed for the rivals o f the announcing firms. A fter industries were identified according to which (if any) inform ation effect dom inated, regression analyses were perform ed to identify variables that contribute to explaining the variation o f abnor­ mal returns for the industry rivals. Two sam ples were used to estim ate the regression equations: (1) the sam ple o f rival firm s in industries in which a significant negative ab­ norm al average return dom inated and (2) the sam ple o f rival firms in industries in which a significant positive abnorm al average return dom inated. Test R esults for A bnorm al R eturns o f A nnouncing Firms T he average standardized abnorm al returns o f announcing firm s w ere exam ined to d eterm ine their significance, direction, and m agnitude over the 3-day event windowbeginning one day before and ending one day after the audit opinion announcem ent date. T his procedure entailed the use o f a one-tailed t-test because, based on studies by D opuch et al. [1] and Jones [2], these firm s w ere expected to experience significant negative average abnorm al returns as a result o f the announcem ent. The findings were conclusive, with detection o f large, statistically significant abnorm al negative returns during the event w indow for the firm s announcing "subject to" qualified audit opinions from 1984 through 1988 and for firm s announcing "going concern" unqualified audit opinions w ith special language from 1989 through 1996. The findings also revealed that the largest negative abnorm al returns for both sam ples occurred o n the announce­ m ent day. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 163 The significant negative abnorm al returns w ithin the event w indow for firm s an­ nouncing "subject to" qualified audit opinions were consistent in sign and m agnitude w ith results reported in the related study by D opuch et al. [ 1]. Signs o f leakage were detected for the third and ninth days before the announcem ent. T he analysis also re­ vealed one occurrence o f significant negative abnorm al returns on the ninth day after the announcem ent. This exception m ay be attributable to possible follow -up announcem ents. The tests for negative abnorm al returns o f firms announcing "going concern" audit opinions revealed a large, significant negative equity adjustm ent during the event w indow for the announcing firm s. These returns were larger in m agnitude and were m ore significant than the abnorm al negative returns associated w ith the "subject to" sam ple o f announcing firms. A lso, the five-day period follow ing the announcem ent date w as characterized by positive abnorm al average returns for these announcing firms. T his finding is attributed to an initial overreaction by investors to the "bad" new s con­ cerning the audit outcom e, follow ed by a subsequent upw ard adjustm ent. Significant negative abnorm al returns w ere also identified on days -9, -3. 6, and 8 relative to the an­ nouncem ent date. T he findings for the third and ninth days before the announcem ent date m ay represent leakage. Those for the sixth and eighth days after the announcem ent date m ay have resulted from equity readjustm ents follow ing investor overreaction, or the effects o f subsequent announcem ents. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 164 T est R esults for A bnorm al R eturns o f Rival Firm s T he rival firm s w ere tested for abnorm al returns w ith the sam e statistical tech­ niques used for the announcing firms. These rivals w ere separated by three-digit-SIC classifications and then m erged to test each industry for significant abnorm al returns during the event w indow . O f the 41 industry sam ples tested, 22 resulted in significant abnorm al average returns. A tw o-tailed t-test w as used to test these abnorm al returns for significance because the intent was to determ ine w hich effec t-co m p etitiv e or conta­ g io n - dom inated each industry. O f the 22 industries w ith statistically significant abnor­ mal returns, 15 show ed evidence o f a dom inant com petitive effect and 7 revealed a dom inant contagion effect surrounding the announcem ent date o f the unfavorable audit opinion. B ased on these results, one may conclude that the "subject to" qualified and "going concern" audit opinion announcem ents tend to contain useful inform ation about industry rivals. R egression R esults In an effort to identify significant contributors to the com petitive effect and the contagion effect existing am ong the rival firms, two regression m odels w ere estim ated. The industries w ere first grouped according to the significant effect that dom inated each. Then, to explain the variation in the cum ulative abnorm al returns for the rivals during the event w indow , the regression m odel w as applied to each group. B ased on the findings o f previous intra-industry inform ation transfer studies, the follow ing w ere used as potential explanatory variables: R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 165 (1) return correlations betw een the announcing firm and each rival firm (w hich proxied the sim ilarity o f assets am ong the firms); (2) natural log o f total assets for each rival firm (w hich captured the siz e effect o f the rival firm s); (3) m arket-to-book value o f equity for each rival firm (a m easure ac tin g as a proxy for grow th options o f the rival firm ); (4) H erfindahl Index o f the industry (w hich captured the effect o f industry sales concentration); (5) debt-to-total assets ratio for each rival firm (a m easure o f financial leverage); (6) cum ulative abnorm al return for the announcing firm during the e v e n t w indow (a proxy for "surprise" effects related to the announcem ent); (7) stock exchange the rival firm is traded on (which extracts the effects o f inform a­ tional asym m etry associated with the firm ); (8) business cycle during w hich the announcem ent o f the unfavorable au d it opinion w as m ade; and (9) existence o f an earnings announcem ent concurrent w ith the audit o p in io n an­ nouncem ent (w hich captures the "com pounding" effect o f tw o sim u ltan eo u s adverse announcem ents). The regression perform ed on the industry rivals from industries in w h ich the com petitive effect dom inated w as statistically insignificant. The H erfindahl Index w as the only one o f the nine variables that was statistically significant. N o inferences could be m ade based on the insignificant m odel. The regression estim ated on rivals in industries dom inated by a c o n tag io n effect w as statistically significant, w ith an R2 value o f .04. Six o f the nine su g g ested explana­ tory variables contributed significantly to the negative abnorm al returns ex p e rien ce d by rivals. The three variables that provided no explanatory pow er for the cu m u lativ e R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 166 abnorm al negative returns o f the industry rivals are: (1) the returns correlation betw een the rival and the announcing firm; (2) the exchange dum m y variable; and (3) the cum u­ lative abnorm al returns experienced by the announcing firm during the event window . Four variables that w ere statistically significant and positively related to the cum ulative negative abnorm al returns o f the rivals (reducing the m agnitude o f the negative return) are: (1) the m arket-to-book value o f equity; (2) the Herfindahl Index; (3) the natural log o f total assets; and (4) the business cycle dum m y variable. The other tw o significant variables w ith a negative (i.e., m agnitude increasing) effect on the cum ulative abnorm al returns to the rivals are the debt-to-total assets ratio and the earnings announcem ent dum m y. T hese earnings announcem ents were expected to be predom inantly negative, based on related findings o f Dopuch et al. [1]. Implications, Conclusions, and Contributions of the Study The purpose o f this section is to provide direct responses to the research ques­ tions that prom pted the study. In answ ering these questions, the m ain focus is on the im plications, conclusions, and contributions o f this study. These four questions, as stated in C hapter I, are as follows: 1. Do "subject to" audit opinion qualifications or "going concern" special language convey inform ation about the rivals o f the announcing firm w hen the inform ation is m ade public by the financial press? 2. D oes this inform ation bring about a positive stock price reaction for industry rivals (called a com petitive effect) o f certain industries or a negative stock price reaction for rivals (called a contagion effect) for other specific industries? R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 167 3. W hich industry-w ide or firm -specific factors account for the particular reaction o f rivals as evidenced by their stock price m ovem ents? 4. W hat can be inferred from these significant stock price m ovem ents? Inform ation C ontent o f A udit O pinions S ignificant stock price reactions by announcing firms and rival firm s were d e­ tected in som e industries w hen the unfavorable audit opinions were publicized by the financial press. The existing literature has already identified the negative im pact on eq­ uity values o f the announcing firm s associated w ith a "subject to" qualified audit o pin­ ion announcem ent. This study contributes additional inform ation to the existing literature by revealing evidence o f a sim ilar reaction by investors in response to "going concern" unqualified audit opinion announcem ents. The m ost im portant contribution o f this study is the finding that, in m any cases, the rivals o f the announcing firm also incur significant abnorm al returns as a result o f unfavorable audit opinion announcem ents. This m easurable reaction im plies that the participants in financial m arkets glean, from these audit opinions, new inform ation about industry conditions for rivals as well as for the announcing firms. On this basis, one m ay conclude that audit opinions, w hich reveal inform ation regarding the ability o f a firm to continue as a going concern, have an inform ation content for announcing firm s and for their com petitors that investors utilize in their capital allocation decisions. In­ tuitively. the findings suggest that auditors m ust have access to privileged inform ation (not accessible to investors) that, when released, reduces the inform ational asym m etries in financial m arkets. The study perform ed by M utchler [3], w hich w as sum m arized in R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 168 the literature review, provided evidence that financial ratios play a sm all role in the d e­ cision by an auditor to issue a "going concern" opinion. T herefore, the actual issuance o f an unfavorable audit opinion is based in part on inside inform ation to w hich financial an alysts have little or no access. C onsequently, unfavorable audit opinions reveal new inform ation about the announcing firm s and their industry rivals that m ay not necessar­ ily be obtained solely through issuance o f the financial statem ents. The results o f this study, therefore, add to the literature concerned w ith issues o f inform ational content o f auditors' opinions and types o f announcem ents that result in intra-industry inform ational transfers to non-announcing firms. C om petitive Versus C ontagion Effects: Explanations and Inferences W hen rival firm s experience significant positive abnorm al equity adjustm ents resulting from the release o f new inform ation about an industry from an announcing firm , the inform ation com m unicated m ay result in a com petitive effect dom inating that industry. Likew ise, w hen significant negative abnorm al returns dom inate an industry based on an announcem ent, a contagion effect occurs. O f 41 industries exam ined for in­ form ational transfers resulting from unfavorable audit opinion announcem ents, 15 show ed statistically significant com petitive effects and 7 experienced significant nega­ tive equity adjustm ents or contagion effects. A com petitive effect is essentially a transfer o f profitability from one o r m ore firm s in the industry to the others. T hus, one m ay infer that an announcem ent resulting in a com petitive effect for the industry reveals inform ation that adversely affects the R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 169 announcing firm , but not the industry rivals. For exam ple, a firm that is expected to bankrupt because o f som e firm -specific factor will obviously lose its custom ers to co m p eti­ tors. The follow ing industries experienced a dom inating com petitive effect as a result o f the audit opinion announcem ent: (1) O il and Gas E xtraction, (2) Oil and G as Field Services, (3) C onstruction, M ining, and Oil and Gas E quipm ent, (4) C om puter and O f­ fice Equipm ent. (5) C om m unication Equipm ent. (6) S urgical, M edical and Dental In­ strum ents, (7) Electric Services, (8) Electrical G oods-w holesale, (9) Investors, not else­ w here classified (nee), (10) M iscellaneous Business Services. (11) H om e H ealth Care S ervices, (12) M iscellaneous H om e and A llied Services, nec, and (13) R esearch. Devel­ o p m ent and Testing Services. O f these thirteen industries. Oil and G as Extraction and E lectric Services experienced a consistent com petitive effect in both sam ples ("subject to" and "going concern" opinions). An apparently com m on factor am ong these indus­ tries is that m ost o f them are characterized by custom ers w ishing to establish and m ain­ tain long-term business relationships. However, m aking an inference on this basis is not supported by the literature. For exam ple. Lang and S tulz [4] found som e o f these sam e industries to experience a contagion effect as a result o f bankruptcy announcem ents, w h ich com m unicate sim ilar inform ation to that o f "going concern" uncertainties. A contagion effect dom inates an industry w hen the m arket interprets the unfa­ vorable audit o pinion to reveal inform ation about a factor affecting the industry' as a w hole. For exam ple, a firm that is troubled by an increase in resource prices o r input R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. costs is obviously affected by an industry-w ide concern. W hen this type o f inform ation is com m unicated as the reason for "going concern" uncertainties about a firm, the m ar­ ket will react accordingly, realizing that industry rivals will also soon experience a de­ crease in profitability. The sam pled industries that experienced a dom inant contagion effect as a result o f the unfavorable audit opinion announcem ents are as follows: (1) M obile H om es, (2) Drugs and Pharm aceutical. (3) Electronic C om ponents and A ccesso­ ries, (4) A ircraft and Ship B uilding, (5) Photographic Equipm ent and Supply. (6) Elec­ tric and O ther Services C om bined, and (7) Eating and D rinking Places. These indus­ tries appear to have even less in com m on than those experiencing com petitive effects. T herefore, one m ay conclude that the announcing firm s com m unicated an industry-w ide problem that had not been detected by the m arket before the announcem ent. Influence o f Industry-W ide and Firm -Specific Factors T he regression analysis perform ed on the industry rivals experiencing a dom i­ nant com petitive effect provided little relevant inform ation. Inferences cannot be m ade from the regression m odel, regarding those industries in w hich the com petitive effect dom inated. Im plications can be draw n, how ever, regarding the effects o f industry-w ide and firm -specific factors on the negative abnorm al returns experienced by rivals in in­ dustries experiencing a dom inating contagion effect. T he regression m odel for these firm s w as statistically significant and contains six significant contributors to the cum u­ lative abnorm al returns o f the rival firms. T he three variables that did not test signifi­ cant nevertheless provide additional valuable inform ation. The conclusions and im pli­ R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 171 cations o f the regression analysis for the rival firm s negatively affected by the unfavor­ able audit opinion announcem ent are discussed in the next few paragraphs. The variables specific to the rival firm s include: (1) grow th options (m arket-tobook value o f equity ratio); (2) firm size (natural log o f total assets); (3) firm leverage (book debt-to-total assets ratio); (4) the stock exchange dum m y (1 if traded on the N A SD A Q ; 0, otherw ise); and (5) the correlation o f the rival firm 's stock returns with those o f the announcing firm. O f these, the particular stock exchange on w hich the rival firm 's equity was traded had no im pact on the negative abnorm al returns to the rival firm. This result w ould im ply that investors did not distinguish which equities w ere sold as a result o f the new inform ation received about the industry. The inform ational asym m etries norm ally associated w ith N A S D A Q -traded firms did not com e into play in this study as it did in R ajagopal's [5] exam ination o f contagion effects associated with bond dow ngrades or in the G raves et al. [6] investigation into the larger stock price re­ actions associated with equity trade utilizing a bid-ask spread rather than auction trad­ ing. The earnings correlation betw een the announcing firm and the rival firm also did not affect the negative return o f the rival. Perhaps investors used a different m easure to d eterm ine sim ilarity o f firm s' assets. Lang A nd S tulz [4] did, how ever, find the earn­ ings correlation to be a m ajor contributor to the contagion effect associated w ith bank­ ruptcy announcem ents. A lso, Foster [7] and C linch and Sinclair [8] actually used earnings sim ilarity m easures to define industry rivals in exam ining the effects o f earn­ ings announcem ents on industry rivals. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 172 T he rival firm 's grow th options, firm size, and firm leverage significantly con­ tribute to the equity adjustm ent incurred by the rival firm. Firm s w ith m ore grow th op­ tions experienced less o f a reduction in equity value than did those w ith few er grow th options. These results could signify that the firm s with higher m arket-to-book equity values had the ability to diversify outside o f the problem areas in the industry or that they already had diversified. Those firms that had fewer grow th options m ay have had "all their eggs in one basket," resulting in greater losses w hen the basket broke. These findings contradict those o f G aver and G aver [9] and Barclay and Sm ith [10]. Larger rival firm s had less negative equity adjustm ent than sm aller rival firms. A ccording to the findings o f G raham and King [11], the inform ation about larger firms typically is m ore readily available than that about sm aller firms. For this reason, the positive coefficient for firm size w ould im ply that the larger firm s are less affected be­ cause the inform ation w as already available to investors. A lso, larger firm s m ay be fi­ nancially m ore capable o f enduring hard tim es in the industry. T he degree to w hich the rival firm w as levered w as also a significant contributor to its negative return. This variable had a negative coefficient, indicating that a m ore highly levered firm has m ore risky cash flows, increasing the present value o f bank­ ruptcy costs for the firm . Further, firm s that are m ore highly levered have a greater elasticity o f equity value w ith respect to changes in asset values. B olton and Scharfstein [12] m odeled the ability o f firm s w ith less debt to "prey" on those w ith higher leverage. The reasoning o f this phenom ena m ay also be applied to the contagion effect in that R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 173 o utside firm s relying on the industry for supplies or production inputs are expected to be m otivated to establish relations w ith firm s m ore likely to rem ain in business. T he rem aining variables used in the regression m odel are those that characterize the conditions in the industry and the inform ation available about the industry. These include: (1) the Herfindahl Index; (2) the dum m y extracting the effect o f a concurrent earnings announcem ent by the firm receiving the unfavorable audit opinion; (3) the dum m y variable set to 1 when the announcem ent coincides w ith an econom ic contrac­ tion; and (4) the abnormal return o f the announcing firm . All these variables signifi­ can tly im pacted the returns o f rival firm s except the abnorm al return for the announcing firm. The im pact o f this variable m ay have been affected by the m arket overreaction observed in the "going concern" sam ple's abnorm al returns to the announcing firms. The event w indow encom passed the trading days from the day before the announcem ent to the day follow ing the announcem ent. The positive abnorm al returns experienced by the announcing firm s for the day follow ing the announcem ent date m ay have sen t m ixed sig n als to the m arket regarding h o w negatively the industry-w ide inform ation actually affected the announcing firm (and thus others firms). The industry concentration variable, as m easured by the H erfindahl Index, pro­ vided inform ation about the sales dispersion in the industry. This variable tested sig­ nificant, w ith a positive coefficient, revealing that rivals in industries w ith m ore w idely d ispersed sales experienced greater negative returns. R ivals in industries w ith m ore co n centrated sales experienced less reduction in equity value. T his variable w as R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 174 expected to have a positive im pact on equity value w hen a com petitive effect dom inated because firms in these industries are in a better position to capture m arket share from a distressed com petitor. A lthough the contagion effect dom inated in this sam ple, one m ay conclude that the reduction in equity value caused by the contagion effect w as par­ tially offset by the ability o f rival firms to gain m arket share at the expense o f the trou­ bled firm . A lso, firm s in industries with higher sales concentration have m ore flexible cost and output conditions than those in industries w ith m ore sales dispersion. The earnings announcem ent dum m y variable was used to capture the additional effect o f a release o f negative earnings inform ation concurrent with the unfavorable audit opinion announcem ent. T he negative coefficient for this variable suggests a com ­ pounded effect o f "bad" new s follow ing "bad" new s, further decreasing the equity val­ ues o f rival firms. The D opuch et al. [1 ] study found a sim ilar com pounded m arket reaction for the announcing firm occurring when an earnings release coincided w ith the announcem ent o f the unfavorable audit opinion. The final variable reflects inform ation that already had been incorporated into the equity values o f all industries in the econom y. The dum m y used to m easure the ef­ fects o f recessions in the econom y had a positive coefficient. This dum m y w as set to 1 for tim es o f econom ic contraction. Econom ic contractions are associated w ith sm aller negative equity adjustm ents in the case o f contagion because unfavorable audit opinions and any other release o f "bad" new s are m uch less a surprise during recessions than dur­ ing expansions. For this reason, unfavorable audit opinion announcem ents had a greater R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 175 negative effect on returns o f rivals during econom ic expansions and offset som e o f the negative returns o f rivals during econom ic contractions. This finding is sim ilar to that o f R ajagopal [5]. An interesting observation regarding the effects o f business cycles on the equity values o f rivals follow ing an unfavorable audit opinion announcem ent w as m ade before the regression m odels were estim ated. N one o f the rival-sam ple observations in w hich a com petitive effect dom inated occurred during an econom ic contraction. H ow ever, three announcem ents w ere m ade during recessionary periods in the econom y, all associated with contagion effects. O ne m ay speculate that although an econom ic contraction re­ sulted in less m agnitude o f change in the equity value o f rivals w hen contagion effects dom inated an industry, unfavorable announcem ents m ade during these dow nsw ings in the econom y were never interpreted as good news for the rivals in the particular sam ple used in this study. On this basis, the general interpretation o f the m arket appears to be negative w hen "unfavorable" new s about a firm is released during a recession, possibly reflecting the general attitude o f investors during econom ic dow nsw ings. Suggested Extensions of the Study The prim ary objective o f this study was to determ ine the existence o f intra­ industry inform ational transfers associated w ith announcem ents o f unfavorable audit opinions. T he analyses have show n that unfavorable audit opinions receiving m edia coverage provide new inform ation about com petitors in the sam e industry. R efinem ents R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 176 and ex tensions o f the study are indicated to provide better understanding about these au d it o pinion effects. O ne possible im provem ent would be to include a variable developed by D opuch et al. [13] as an independent variable in the regression seeking to explain the abnorm al returns o f industry rivals. This variable, which is created using a logistical regression on certain balance sheet and m arket-related m easures, is used to predict the probability o f a firm 's receiving a "going concern" opinion. A sim ilar variable was used by Jones [2] and tested significant in a regression attem pting to explain the abnorm al returns o f firm s announcing "going concern" opinions. A nother refinem ent that m ay be m ade to this study is the use o f standardized ab­ norm al returns as the dependent variable. This m ethodology is based on the use o f these returns to determ ine the significance o f abnorm al returns as w ell as to define w hich intra-industry inform ational transfer effect dom inates. G raham and King [11] used stan­ dardized cum ulative abnorm al returns as the dependent variable in their intra-industry inform ation transfer study determ ining the effects o f a firm's inform ation environm ent on its abnorm al returns relative to earnings announcem ents. U sing an alternative m ethod for identifying industry rivals is another possible extension o f this research. This study used C R SP-defm ed SIC codes to identify indus­ try rivals. H ow ever, the m anner in w hich industry rivals were defined m ay m eet w ith o p p osition by those w ho believe the C O M P U ST A T database to have SIC assignm ents su p erior to those used by CRSP. A lthough this procedure m ay be debatable w ith no one R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 177 actually being ab le to prove beyond doubt which database better defines industry rivals, the results o f such an endeavor could be interesting and potentially inform ative. A nother extension o f this study m ay be to determ ine w hether firm s that do not receive m edia coverage o f their "going concern" unqualified audit opinion w ith special language m ay also release inform ation relevant to industry rivals. A lthough, in the ab ­ sence o f m edia coverage, announcem ent dates become difficult to define, inform ation nevertheless should be m ore efficiently dissem inated in the p o st-1988 era than it w as in the ex am ination periods o f audit opinion studies that included firm s receiving "subject to" q u alified audit opinions not made public in the financial press. Finally, an expansion o f the literature exam ining the im pact o f audit o p in io n s on industry rivals m ay be accom plished by using the variable developed by D opuch at al. [13] to predict the probability o f a firm receiving a "going concern" qualification. Using this variable, a study can be perform ed to predict the audit outcom e. B ased on the pre­ dictions, portfolios m ay be built that short-sell the equities o f announcing firm s and/or the rivals in industries in w hich contagion effects are expected to dom inate. T h e returns for those portfolios can be tested for significance on the announcem ent date o f the ac­ tual au dit opinion results for those firm s expected to receive unfavorable audit opinions. T hese refinem ents and extensions m ay contribute to understanding the m anner in w hich equity m arkets operate. Some studies, such as Foster [7] and G raham and K ing [11], purposely selected only industries that had 12 or few er rivals because the ef­ fects o f inform ation transfers are more adequately captured in m ore concentrated R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 178 industries. U sing that restriction and using C O M PU STA T to define industry rivals for this study m ay have provided further evidence o f inform ational transfers and a greater depth o f inferences to be drawn. Citations [1] D opuch, N.. R. W. H olthausen, and R. W. Leftw ich. "A bnorm al R eturns A sso­ ciated W ith M edia D isclosures o f'S u b je c t to' Q ualified A udit O pinions." Jo u r­ nal o f A ccounting a n d E conom ics, 8, 1986, pp. 93-117. [2] Jones. F. L. "The Inform ation C ontent o f the A uditor's G oing C oncern E valua­ tion." Jo u rn a l o f A ccounting a n d Public Policy, 15, 1996. pp. 1-27. [3] M utchler, J. F. "A uditor's Perceptions o f the G oing C oncern O pinion D ecision." Auditing. 3. 1984, pp. 17-29. [4] Lang, L. H., and R. M. Stulz. "C ontagion and C om petitive Intra-Industry E f­ fects o f B ankruptcy A nnouncem ents: An Em pirical A nalysis." Jo u rn a l o f F i­ nancial Econom ics. 32, 1992, pp. 45-60. [5] R ajagopal, S. "Bond D ow ngrades and Intra-Industry C ontagion and C o m p eti­ tive Effects: The Case o f Industrials." D issertation. M ississippi State U niver­ sity, M ay 1997. [6] G raves, J. A ., S. P. Hegde, and R. E. M iller. "T rading M echanism s and the C om ponents o f the B id-A sk Spread." Journal o f F inance, 49, 1994. pp. 1471-1488. [7] Foster, G. "Intra-Industry Inform ation Transfers A ssociated W ith E arnings R e­ leases." Jo u rn a l o f A ccounting a n d Econom ics, 3. 1981, pp. 201-232. [8] C linch, G. J.. and N. A. Sinclair. "Intra-Industry Inform ation R eleases." J o u r­ nal o f A cco u n tin g an d Econom ics. 9. 1987, pp. 89-106. [9] G aver, J. J., and K. M. Gaver. "Additional Evidence on the A ssociation B etw een the Investm ent O pportunity Set and C orporate Financing. D ividend and C om pensation Policies." Jo u rn a l o f A ccounting a n d E conom ics, 16. 1993, pp. 125-160. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 179 [10] Barclay, M. J., and C. W. Sm ith. "The M aturity Structure o f C orporate Debt." Journal o f Finance, 50, 1995, pp. 609-631. [11] G raham , R. C., and R. D. King. "Industry Inform ation Transfers: The Effect o f Inform ation Environm ent." Journal o f Business, Finance a n d A ccounting, 23, 1996, pp. 1289-1306. [12] B olton, P., and D. Scharfstein. "A Theory o f Predation B ased on A gency Prob­ lem s in Financial C ontracting." Am erican Econom ic Review , 80, 1990. pp. 93-106. [13] D opuch, N., R. W. H olthausen, and R. W. Leftwich. "Predicting A udit Q uali­ fications W ith Financial And M arket Variables." The A ccounting Review , 62, 1987, pp. 431-454. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. BIB LIO G R A PH Y A harony, J.. and I. Swary. "C ontagion Effects o f Bank Failures: Evidences From C apital M arkets." Journal o f Business. 56, 1983, pp. 305-322. A jinkya, B., and M. Gift. "Corporate M angers' Earnings Forecasts and Sym m etrical A djustm ents o f M arket Expectations." Journal o f A ccounting Research, 22, 1984, pp. 425-44. A khigbe, A ., and J. M adura. "Intra-Industry Effects o f V oluntary C orporate L iquida­ tions." Jo u rn a l o f Business. Finance & A ccounting, 23, 1996, pp. 915-930. A lii, K., S. Thapa, and K. Yung. "Stock Price D ynam ics in O verlapped M arket Seg­ m ents: Intra and Inter-Industry C ontagion Effects." Journal o f Business. F i­ nance & Accounting, 21, 1994, pp. 1059-1070. A m erican Institute o f C ertified Public A ccountants. The A uditor's C onsideration When a Q uestion A rises A bout an Entity's C ontinued Existence. S A S No. 34. N ew York: A IC PA , 1981. A m erican Institute o f C ertified Public A ccountants. Reports on A u d ited Financial Statem ents, S A S No. 58. N ew York: A ICPA , 1988. A m erican Institute o f C ertified Public A ccountants. The A uditor's C onsideration o f an Entity's A bility to C ontinue as a G oing Concern, S A S No. 59. N ew York: A IC PA . 1988. A squith, P., and D. M ullins. "Equity Issues and O ffering D ilution." Journal o f F inan­ cial Econom ics, 15, 1986, pp. 61-89. A tiase, R. "Predisclosure Inform ation. Firm C apitalization, and Security Price B ehavior A round E arnings A nnouncem ents." Journal o f A ccounting Research. 23. 1985, pp. 21-36. B aginski. S. P. "Intra-Industry Inform ation Transfers A ssociated w ith M anagem ent Forecasts o f Earnings." Journal o f A ccounting Research, 25. 1987. pp. 196-216. 180 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 181 Ball, R., and P. Brown. "An Em pirical E valuation o f A ccounting Incom e N um bers." J o u rn a l o f A ccounting R esearch, 6, 1968, pp. 159-178. B annister, J. W. "Earnings Signals and Inter-Firm Inform ation Transfers." Jo u rn a l o f B usiness, Finance & Accounting, 21, 1994, pp. 1127-1149. B arclay, M. J., and C. W. Sm ith. "The M aturity S tructure o f C orporate Debt." J o u rn a l o f F inance, 50, 1995, pp. 609-631. Beaver, W. H. F inancial Reporting: A n A cco u n tin g Revolution. Engelw ood C liffs, NJ: P rentice H all. 1981. B erton. L., and B. Ingersoll. "Rep. Dingell to Take A im s at A ccountants. SEC in H earings on Profession's Role as a W atchdog." Wall Street Journal. 205(34), 1985, p. 4. Black, F. "Capital M arket Equilibrium W ith R estricted B orrow ing." Journal o f B usi­ ness, 45, 1 9 7 2 ,p p .444-455. B olton, P., and D. Scharfstein. "A Theory o f P redation Based on A gency P roblem s in Financial C ontracting." Am erican E conom ic R eview , 80, 1990, pp. 93-106. B ow en, R., R. C astania. and L. Daley. "Intra-Industry Effects o f the A ccident at Three M ile Island." Journal o f Financial a n d Q uantitative Analysis, 18. 1983. pp. 87-111. B row n. H. G. "Should 'G oing C oncern' Be A uditor’s H eadache." A ccounting Today. 3. 1989. p. 4. B row n, K.. and V. Harlow. "M arket O verreaction: M agnitude and Intensity." Jo u rn a l o f P ortfolio M anagem ent. 14, 1988. pp. 6-13. B row n, K., S. Tinic, and V. Harlow. "Risk A version, U ncertain Inform ation, and M arket Efficiency." Journal o f F inancial Econom ics. 22, 1988, pp. 355-385. B row n, S. J., and J. B. W arner. "M easuring Security Price Perform ance." Jo u rn a l o f F inancial Econom ics, 8, 1980, pp. 205-258. B row n, S. J.. and J. B. W arner. "Using D aily Stock R eturns: The C ase o f Event S tudies." Jo u rn a l o f F inancial E conom ics, 12, 1985. pp. 3-32. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 182 B runer, R. F., and J. M. S im m s. "The International Debt C rises and B ank S ecurity R eturns in 1982." J o u rn a l o f Money, C redit a n d B anking, 19, 1987, pp. 46-55. C heng, L. T. W „ and J. E. M cD onald. "Industry Structure and R ipple Effects o f B ankruptcy A nnouncem ents." F inancial Review, 31, 1996, pp. 783-807. C how , C. W., and S. J. Rice. "Q ualified A udit O pinions and Share P ric e s-A n Investigation." A u d itin g , 1, 1982, pp. 35-53. C itron, D. B., and R. J. Taffler. "The A udit R eport U nder G oing C oncern U ncertainties: A n Em pirical A nalysis." Jo u rn a l o f A ccounting a n d B usiness R esearch, 22, 1992, pp. 337-345. C linch, G. J., and N. A. Sinclair. "Intra-Industry Inform ation R eleases." Jo u rn a l o f A ccounting a n d E conom ics, 9, 1987, pp. 89-106. C ornell, B., and A. Shapiro. "The R eaction o f Bank Stock Prices to the International D ebt Crises." Jo u rn a l o f B anking a n d Finance, 10. 1986, pp. 55-73. Dann. L. Y., and W. H. M ikkelson. "C onvertible Debt Issuance, C apital S tructure C hange and F inancing-R elated Inform ation." Jo u rn a l o f F inancial E conom ics, 13. 1984, pp. 157-186. DeBondt, W.. and R. Thaler. "D oes the Stock M arket O verreact?" J o u rn a l o f F inance, 40, 1985, pp. 793-805. D eBondt, W „ and R. Thaler. "Further Evidence on Investor O verreaction and S tock M arket Seasonality." Jo u rn a l o f Finance. 42, 1987, pp. 557-581. D odd, P., and J. B. W arner. "O n C orporate G overnance: A Study o f Proxy C ontests." Jo u rn a l o f F inancial E conom ics, 11, 1983, pp. 401-438. D odd, P., N. Dopuch, R. W. H olthausen. and R. W. Leftw ich. "Q ualified A udit O pinions and Stock Prices: Inform ation C ontent, A nnouncem ent D ates and C oncurrent D isclosures." Jo urnal o f A ccounting a n d E conom ics. 6. 1984, pp. 3-38. D opuch, N ., R. W. H olthausen, and R. W. Leftw ich. "A bnorm al R eturns A ssociated W ith M edia D isclosures o f'S u b je c t to' Q ualified A udit O pinions." J o u rn a l o f A ccounting a n d E conom ics, 8. 1986, pp. 93-117. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 183 D opuch, N ., R. W. H olthausen, and R. W. Leftw ich. "Predicting A udit Q ualifications W ith Financial A nd M arket V ariables." The A ccounting Review . 62, 1987. pp. 431-454. Eckbo, B. E. "V aluation Effects o f C orporate D ebt O fferings." Jo u rn a l o f F inancial E conom ics, 15, 1986, pp. 119-151. E lliott, J. '"Subject to' A udit Opinions and A bnorm al Security R etu m s-O u tco m es and A m biguities." Journal o f A ccounting Research, 20, 1982, pp. 617-638. Fam a, E. F., and K. R. French. "The C ross-Section o f Expected Stock Returns." Jo u rn a l o f Finance, 47. 1992, pp. 427-465. Firth, M. "The Im pact o f Earnings A nnouncem ents on the Share Price B ehavior o f S im ilar T ype Firm s." Econom ic Journal, 86, 1976, pp. 296-306. Firth, M. "D ividend C hanges, Abnorm al R eturns, and Intra-Industry Firm V aluations." Jo u rn a l o f F inancial a n d Q uantitative A nalysis, 31. 1996, pp. 189-211. Foster, G. "Intra-Industry Inform ation Transfers A ssociated W ith E arnings R eleases." Jo u rn a l o f A ccounting a n d Econom ics, 3, 1981, pp. 201-232. G aver, J. J., and K. M. Gaver. "Additional Evidence on the A ssociation B etw een the Investm ent O pportunity Set and C orporate Financing, D ividend and C om pensa­ tion Policies." Journal o f Accounting a n d Econom ics, 16, 1993, pp. 125-160. Gay, G. D., S. G. Tim m e, and K. Yung. "Bank Failure and C ontagion Effects: Evi­ dence From H ong Kong." The Journal o f F inancial Research, 14. 1991. pp. 153-168. G raham , R. C., and R. D. King. "Industry Inform ation Transfers: The Effect o f Inform ation Environm ent." Journal o f Business, Finance & A ccounting, 23. 1996. p p . 1289-1306. G rant. E. B. "M arket Im plications o f D ifferential A m ounts o f Interim Inform ation." Jo u rn a l o f A ccounting Research. 18. 1980, pp. 255-268. G raves. J. A.. S. P. H egde. and R. E. M iller. "Trading M echanism s and the C o m ­ ponents o f the B id-A sk Spread." Journal o f Finance. 49, 1994. pp. 1471-1488. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 184 Han, J. C „ and J. J. Wild. "U nexpected Earnings and Intra-Industry Inform ation Transfers: Further Evidence." Journal o f A ccounting R esearch, 28, 1990. pp. 211-219. Hill, J., and T. Schneew eis. "The Effect o f Three M ile Island on Electric Utility Stock Prices: A Note." Jo u rn a l o f Finance, 38, 1983, pp. 1285-1292. Howe, J. "Evidence on Stock M arket O verreaction." Financial A nalysts Journal, 42. 1986, pp. 74-77. Ibbotson, R. "Price Perform ance o f C om m on Stock N ew Issues." Journal o f F inancial Econom ics, 2, 1975, pp. 235-272. Ibbotson, R „ and J. F. Jaffe. '"Hot Issue' M arkets." Journal o f Finance, 30, 1975, pp. 1027-1042. Jacklin, C. J.. and S. B hattacharya. "D istinguishing Panics and Inform ation-B ased Bank Runs: W elfare and Policy Im plication." Journal o f Political E conom y, 96, 1988, pp. 568-580. Jam es, C. "Som e Evidence o f the Uniqueness o f B ank Loans." Journal o f F inancial E conom ics. 19, 1987, pp. 217-235. Jensen, M „ and W. Pugh. "V aluation Effects o f C anceled D ebt O fferings." Journal o f F inancial a n d Q uantitative A nalysis, 26, 1991, pp. 425-431. Jones, F. L. "The Inform ation Content o f the A uditor's G oing C oncern Evaluation." Jo u rn a l o f A ccounting and Public Policy, 15, 1996, pp. 1-27. K ohers, T „ and S. R ajagopal. "The Intra-Industry Effects o f C anceled Debt O fferings." W orking paper, 1997. Lam y, R., and G. R. Thom pson. "Penn Square, Problem Loans and Insolvency Risk." Journal o f Financial Research. 9, 1986, pp. 103-111. Lang, L. H., and R. M. Stulz. "C ontagion and C om petitive Intra-Industry Effects o f B ankruptcy A nnouncem ents: An Em pirical A nalysis." Journal o f F inancial Econom ics. 32. 1992. pp. 45-60. Lev, B., and J. Ohlson. "M arket-B ased Em pirical R esearch in A ccounting: a R eview . Interpretation, and Extension." Journal o f A ccounting Research, 20, 1982. pp. 249-329. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 185 Levis, M . "The Long-R un Perform ance o f Initial Public O fferings: T he UK Experience 1980-1988." F inancial M anagem ent, Spring 1993, pp. 28-41. Lintner, J. "The V aluation o f R isk A ssets and the Selection o f Risky Investm ents in Stock P ortfolios and C apital Budgets." Review o f Econom ic a n d Statistics, 47. 1965, pp. 13-37. Loudder, M. L., I. K. K hurana, R. B. Sawyers, C. C ordery, C. Johnson, J. Lowe, and R. W underle. "The Inform ation C ontent o f Audit Q ualifications." A uditing, 11, 1992, pp. 69-82. M ann, A. "Special R e p o rt-A S B Public M eeting on 'Subject to' O pinion." Jo u rn a l o f A ccountancy, 154, 1982, pp. 10-14. M asulis, R. W. "The Effect o f C apital Structure C hange on Security Prices." Jo u rn a l o f F inancial E conom ics, 8. 1980, pp. 139-178. M enon, K., and K. B. Schw artz. "A n Em pirical Investigation o f A udit Q ualification D ecisions in the Presence o f Going C oncern U ncertainties." C ontem porary A ccounting R esearch, 3. 1987. pp. 302-315. M ikkelson, W. H. "C onvertible C alls and Security R eturns." Journal o f Financial E conom ics, 9. 1981, pp. 237-264. M ikkelson, W. H.. and M. M. Parch. "Stock Price Effects and Costs o f Secondary D istributions." Jo u rn a l o f Financial Econom ics, 14, 1985, pp. 165-194. M ikkelson, W. H.. and M. M. Parch. "V aluation Effects o f Security O fferings and the Issuance Process." Journal o f F inancial Econom ics, 15, 1986. pp. 31-60. M ikkelson, W. H.. and M . M. Parch. "W ithdraw n Security O fferings." Jo u rn a l o f F inancial a n d Q uantitative Analysis, 23, 1988, pp. 119-133. M oore, N. H., D. R. Peterson, and P. P. Peterson. "S helf R egistrations and Shareholder W ealth: A C om parison o f S h elf and Traditional E quity O fferings." Jo u rn a l o f F inance, 41, 1986, pp. 451-464. M utchler, J. F. "A uditor's Perceptions o f the G oing C oncern O pinion D ecision." A uditing, 3 .1 9 8 4 , pp. 17-29. M utchler, J. F. "A M ultivariate A nalysis o f the A uditor's G oing C oncern O pinion D ecision." J o u rn a l o f A ccounting Research, 23, 1985, pp. 668-682. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 186 O ’Hara, M ., and W. Shaw . "D eposit Insurance and W ealth Effects: The V alue o f Being 'T oo B ig To Fail'." Journal o f Finance, 45, 1990, pp. 1587-1600. O lsen, C., and J. R. Dietrich. "V ertical Inform ation Transfers: T he A ssociation B etw een R etailers' Sales A nnouncem ents and S uppliers' Security R eturns." Jo u rn a l o f A ccounting Research (S upplem ent 1985), pp. 144-166. Patell, J. N. "C orporate Forecasts o f Earnings Per Share and S tock Price Behavior: Em pirical Tests." Journal o f A ccounting R esearch, 14, 1976, pp. 246-276. P eterson, P. P. "E vent Studies: A R eview o f Issues and M ethodology." Q uarterly Jo u rn a l o f B usiness a n d Econom ics, 28. 1989, pp. 36-66. R ajagopal, S. "B ond D ow ngrades and Intra-Industry C ontagion and C om petitive Effects: T he C ase o f Industrials." D issertation, M ississippi State U niversity. M ay 1997. R eilly, F. K.. and K. Hatfield. "Investor E xperience W ith N ew Stock Issues." F inan­ cial A nalysts Journal, S eptem ber-O ctober 1969, pp. 73-80. R oll, R. "A C ritique o f the A sset Pricing T heory's T ests--P art 1: O n Past and Potential T estability o f the Theory." Journal o f F inancial E conom ics, A, 1977. pp. 129-176. R oll, R.. and S. A. R oss. "On the C ross-S ectional R elation B etw een Expected R eturns and B etas." Jo u rn a l o f Finance, 49, 1994, pp. 101-121. S chipper, K. "C om m entary on Inform ation Transfers." A cco u n tin g H orizons, Dec. 1990, pp. 97-108. Sharpe, W. F. "C apital A sset Prices: A T heory o f M arket E quilibrium U nder C ondi­ tions o f Risk." Journal o f Finance, 19. 1964. pp. 425-442. S m irlock, M ., and H. K aufold. "Bank Foreign L ending, M andatory D isclosure R ules, and the R eaction o f B ank Stock Prices to the M exican D ebt Crisis." Jo u rn a l o f B usiness, 60, 1987, pp. 347-364. S trong, N. "M odeling A bnorm al R eturns: A R eview A rticle." Jo u rn a l o f Business. F inance & A ccounting, 19, 1992. pp. 533-553. U.S. D epartm ent o f C om m erce. Survey o f C urrent B usiness. O ctober 1994. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 187 W all, L. D „ and D. R. Peterson. "The Effect o f C ontinental Illinois' Failure on the Financial Perform ance o f O ther Banks." Jo u rn a l o f M onetary E conom ics, 26, 1990, pp. 77-99. W arner, J. B. "B ankruptcy and the Pricing o f Risky Debt." Jo u rn a l o f F inancial E conom ics, M ay 1977, pp. 239-276. W aym ire, G. "A dditional Evidence on the Inform ation C ontent o f M anagem ent Earnings Forecasts." Jo urnal o f A ccounting Research, 22, 1984, pp. 703-18. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . APPENDIX ABNO RM AL RETURNS TO RIVALS B Y IND U STR Y A N D SAMPLE T his appendix is included to m aintain the readability o f C hapter IV by separating the 41 tables show ing the results by industry and sam ple o f the hypothesis test for significant average abnorm al returns run on the rivals o f each announcing firm. T hese tables are included because they reinforce the m ajor contribution o f this study; nam ely that contagion and com petitive effects may exist for industry' rivals as a result o f one firm in the industry m aking an unfavorable audit opinion announcem ent in the m ajor financial m edia. 188 R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 189 T able A -l A V E R A G E A B N O R M A L RETU RN S T O R IV A LS OF FIRM S A N N O U N C IN G A G O ING C O N C ER N O PIN IO N : SIC 131 D ay R elative To Date: A verage Abnorm al Return: (% ) - 10 0.0403 -0.689 239 -9 -0.0846 -0.764 239 -8 -0.0145 -1.889 * 239 -7 -0.5197 -2.778 *** 239 -6 -0.3076 -1.919 * 239 -5 -0.1821 -1.056 239 -4 0.1478 1.902 * 239 -3 0.7714 3.928 *** 238 . 2 -0.6239 -1.568 238 - 1 -0.1075 -0.332 237 0 0.1069 + 1 -0.0031 A n nouncem ent (-1,1) -0.00363 T-value: N um ber o f O bservations: 2.066 ** 238 1.454 238 3.188 *** 238 + 2 0.4461 1.744 * 238 + 3 0.1471 2.235 ** 238 +4 0.8861 6.099 *** 238 + 5 0.3351 2.563 ** 238 + 6 0.0314 -0.272 238 + 7 0.1562 -0.843 238 + 8 0.0225 -0.029 238 + 9 0.1451 1.996 ** 238 + 10 0.3383 3.021 *** 238 *** S ignificant at the 1% level under a tw o-tailed test. ** S ig nificant at the 5% level under a tw o-tailed test. * S ig n ificant at the 10% level under a tw o-tailed test. a T he S A R for this observation is positive, indicating a com petitive effect dom inates. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 190 T able A-2 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS OF FIR M S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 283 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: - 10 -0.2108 -2.126 ** 579 -9 -0.5314 -2.433 ** 579 -8 0.0425 0.135 579 -7 0.0828 0.147 579 -6 0.1047 0.024 579 -5 -0.2111 -2.064 ** 578 -4 -0.3615 -1.935 * 578 - N um ber o f T-value: O bservations: (% ) j 0.0081 0.551 578 . 2 -0.0987 -0.381 578 - 1 -0.2024 -0.877 578 0 0.4926 1.473 578 + 1 -0.0661 0.031 578 (-KD 0.2243 0.626 578 + 2 -0.0434 -0.696 578 + 3 -0.2983 -1.705 * 578 +4 -0.3683 -1.719 * 578 + 5 0.1478 0.061 578 +6 -0.2407 -0.736 578 + 7 -0.2063 -1.594 578 + 8 -0.1016 -0.998 578 +9 0.3518 + 10 -0.1279 2.124 ** -0.165 578 577 *** Significant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 191 Table A-3 A V ER A G E A B N O R M A L R ETU R N S TO RIV A LS O F FIR M S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 357 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: - 10 0.2851 0.106 254 -9 0.3787 1.937 * 254 -8 0.3545 1.024 254 -7 0.5692 1.721 * 254 -6 -0.0831 0.355 255 -5 0.1991 1.689 * 255 -4 0.4531 1.193 255 -3 -0.2453 -0.955 255 . 2 0.3343 1.601 255 - 1 0.6765 1.088 255 0 0.0809 1.212 255 + 1 0.5022 0.855 255 (-K 1) 1.2597 3.155 *** 255 + 2 0.0969 + 3 N um ber o f T-value: O bservations: (%) -0.516 255 -1.0374 -3.542 *** 255 +4 0.8289 2.973 *** 254 + 5 -0.8484 -3.173 *** 254 +6 0.6608 2.066 ** 254 + 7 0.0071 + 8 0.7371 2.647 *** 254 + 9 0.3503 2.528 ** 255 + 10 0.3394 0.171 255 -0.255 255 *** Significant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 192 Table A-4 A V ER A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 367 Day R elative To A verage A bnorm al A nnouncem ent Return: Date: (% ) - 10 -0.0485 -0.509 166 -9 0.1745 0.357 166 -8 0.2771 0.352 166 N um ber o f T-value: O bservations: -3.466 *** -7 -1.1006 -6 0.2579 1.465 167 -5 0.1587 -0.894 167 -4 0.5487 1.279 167 -3 -0.1876 -0.878 167 1.763 * 167 . 2 1.6453 - 1 -0.4694 -0.822 168 0 -0.1431 -0.192 168 + 1 -1.6507 -2.928 *** 168 (-1.1) -2.2632 -3.942 *** 168 +2 0.4347 0.932 168 +3 -0.0812 -0.422 168 +4 0.2564 0.606 168 +5 0.4313 0.424 168 +6 1.1711 2.567 ** 168 +7 -0.5965 -1.053 167 168 *** +8 0.7796 3.061 +9 0.8601 1.472 168 + 10 0.1679 -0.162 168 168 *** S ignificant at the 1% level under a tw o-tailed test ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 193 Table A-5 A V E R A G E A B N O R M A L RETU RN S TO RIVALS O F FIRM S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 381 Day R elative To A nnouncem ent Date: A verage Abnorm al Return: T-value: N um ber o f O bservations: (% ) - 10 -0.9001 -1.139 8 -9 -0.9011 -0.281 8 -8 0.4205 -0.283 8 -7 0.8251 0.231 8 -6 0.7844 0.371 8 - 5 -1.8787 -1.484 8 -4 0.3301 -0.223 8 -3 0.5777 0.525 8 . 2 0.5505 0.561 8 - 1 -1.7054 -1.109 8 0 -1.1803 -0.303 8 + 1 -0.0712 -0.144 8 (-1.1) -2.9569 -1.556 8 + 2 -0.3386 -0.045 8 + 3 0.3926 0.461 8 +4 -0.3568 -0.437 8 + 5 0.9489 0.825 8 +6 -1.7379 -1.144 8 +7 1.8921 +8 -0.2016 -0.529 8 +9 -1.7716 -1.135 8 + 10 1.2033 0.525 8 2.758 ** 8 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 194 Table A-6 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS O F FIRM S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 384 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: - 10 0.2282 0.146 95 -9 1.1619 3.056 *** 95 -8 -0.5273 -0.678 95 -7 -0.7129 -2.247 ** 95 -6 0.0949 0.288 95 -5 0.4265 0.505 95 -4 0.0041 -0.367 95 -3 0.0083 -0.401 95 . 2 0.2431 0.639 96 - 1 0.1402 0.606 96 0 0.1167 0.941 96 + 1 0.2056 0.461 96 (-1.1) 0.4625 2.008 ** 96 +2 0.0417 0.114 96 + 3 0.4021 1.305 96 +4 0.3644 0.131 96 + 5 -0.7929 -2.311 ** 96 + 6 -0.3658 -0.609 96 + 7 0.3897 0.883 96 + 8 0.1143 -0.132 96 + 9 0.3061 1.097 96 + 10 -0.7647 -1.747 * 96 T-value: N um ber o f O bservations: (%) *** S ignificant at the 1% level under a tw o-tailed test ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 195 Table A-7 A V E R A G E A B N O R M A L R ETU R N S TO RIV A LS OF FIRM S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 491 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: - 10 -0.4766 -3.883 *** 71 -9 -0.0721 -0.969 71 -8 -0.6297 -3.507 *** 71 -7 0.5481 3.548 *** 71 -6 -0.4541 -2.695 *** 71 -5 -0.2481 -1.878 * 71 -4 -0.4134 -2.811 *** 71 -3 -0.0932 -0.437 71 . 2 -0.0176 -0.335 71 - 1 0.0012 0.014 71 0 0.0552 0.646 71 + 1 0.2308 2.001 ** 71 (- L D 0.2872 2.661 ** 71 +2 0.0992 1.035 71 + 3 -0.2254 -0.864 71 +4 -0.0041 -0.331 71 + 5 -0.9405 -7.707 *** 71 +6 -0.4946 -3.823 *** 71 + 7 -0.1571 -0.232 71 + 8 0.2115 0.831 71 + 9 -0.0303 -1.621 71 + 10 -0.2401 -1.777 * 71 N um ber o f T-value: O bservations: (% ) *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 196 Table A-8 A V E R A G E A B N O R M A L RETU RN S TO RIV A LS O F FIR M S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 495 Day R elative To A verage Abnorm al A nnouncem ent Date: Return: - 10 0.7441 0.662 25 -9 -1.7514 -1.499 25 -8 -0.0896 0.091 25 -7 2.1918 1.589 25 -6 0.0905 0.255 25 -5 0.1848 0.174 25 -4 -0.8484 -1.213 25 -3 0.5279 1.324 25 . 2 1.1483 0.324 25 - 1 0.6139 0.281 25 0 -0.6795 -1.056 25 + 1 -1.0647 -0.768 25 (-1.1) -1.1303 -1.544 25 +2 2.3142 +3 -0.4352 -0.523 25 +4 -0.6267 -0.616 25 +5 -1.4126 -1.412 25 +6 1.4193 1.595 25 + 7 -0.1466 0.552 25 +8 2.3547 2.375 ** 25 +9 -2.2938 -2.202 ** 25 + 10 0.0392 T-value: N um ber o f O bservations: (% ) 1.882 * 0.404 25 25 *** S ignificant at the 1% level under a tw o-tailed test ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 197 Table A-9 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS O F FIRM S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 504 D ay R elative To A nnouncem ent Date: A verage A bnorm al Return: (%) - 10 -2.1429 -9 1.3936 -8 -1.8131 -7 1.8536 0.988 30 -6 -1.6812 -1.539 30 -5 0.0539 0.687 30 -4 -0.5882 -1.075 30 -3 2.8184 . 2 -1.4795 -1.233 30 - 1 -0.3984 -0.279 30 0 0.4007 0.642 30 + 1 0.8864 0.943 30 (-1,1) 0.8887 1.306 30 +2 -0.7173 -1.756 * 30 + 3 -0.1478 0.381 30 +4 0.2563 -0.639 30 + 5 -1.8621 -1.711 * 30 +6 -0.4922 -0.243 30 + 7 1.7251 1.401 30 + 8 -2.4414 +9 + 10 T-value: -3.642 *** N um ber o f O bservations: 30 1.406 30 -1.992 * 30 2.311 ** 30 -3.516 *** 30 0.0779 -0.291 30 -0.0376 0.899 30 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 198 Table A -10 A V E R A G E A B N O R M A L R ETU R N S TO RIV A LS OF FIRM S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 581 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (% ) - 10 -0.4971 -0.481 103 -9 -0.6643 -2.852 *** 103 -8 -0.2156 -0.061 103 -7 0.0566 0.596 102 -6 -0.5738 - 5 -0.1737 0.153 102 -4 -0.2192 -1.265 102 -3 -0.2605 -0.793 102 . 2 0.0638 -0.395 102 - 1 0.6349 -0.327 102 0 -1.0352 -3.297 *** 102 + 1 0.0856 (-1.1) -0.3147 -2.745 *** 102 +2 -0.2548 -0.898 102 + 3 1.1131 1.579 102 +4 -0.8571 + 5 -0.1669 1.067 102 +6 -0.2809 -0.846 102 + 7 0.7514 1.441 102 + 8 -0.2546 1.204 102 +9 0.2191 -0.937 102 + 10 -0.2333 1.062 102 -1.886 * 0.879 -2.552 ** 102 102 102 *** Significant at the 1% level under a tw o-tailed test. ** Significant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 199 Table A -l 1 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS OF FIR M S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 616 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (%) - 10 0.5155 0.618 15 -9 0.8343 1.181 15 -8 0.5415 0.546 15 -7 0.3043 -0.243 15 -6 1.1256 1.412 15 -5 0.4621 0.488 15 -4 -0.7215 -0.795 15 -3 1.0769 1.276 15 . 2 0.0917 -0.511 15 - 1 -1.2337 -1.781 * 15 0 0.4378 0.524 15 + 1 0.8166 0.996 15 (-1-1) 0.0208 -0.261 15 +2 -0.0444 -0.108 15 + 3 0.4936 0.499 15 +4 -1.9505 -3.258 *** 15 + 5 -0.3952 -0.209 15 +6 -0.4621 -0.957 15 + 7 0.5222 0.491 15 + 8 0.5662 0.871 15 +9 0.1881 1.025 15 + 10 0.3384 0.405 15 *** S ignificant at the 1% level under a tw o-tailed test ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 200 Table A -12 A V E R A G E A B N O R M A L R E T U R N S TO RIVALS OF FIRMS A N N O U N C IN G A GOING C O N C E R N O PIN IO N : SIC 655 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (%) - 10 2.9518 3.859 *** 25 -9 0.9846 1.265 25 -8 0.6651 1.105 25 -7 0.7011 -0.374 25 -6 0.7861 -0.008 25 -5 0.5728 -4 -0.8763 -0.735 25 - 3 -0.5681 0.296 25 . 2 -0.2383 -1.178 25 - 1 0.1021 -1.143 25 0 1.4498 + 1 -0.0616 -0.699 25 (- U D 1.4902 0.193 25 + 2 -0.0324 0.026 25 + 3 -1.4715 -1.194 25 + 4 0.0956 -1.218 25 + 5 0.9736 2.166 ** 25 + 6 0.3181 0.031 25 + 7 -0.3591 -0.184 25 + 8 0.6055 0.789 25 + 9 0.2965 0.074 25 + 10 -0.1406 -0.164 25 2.057 * 2.036 * 25 25 *** S ignificant at the 1% level un d er a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 201 Table A - 13 A V ER A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A G O IN G C O N C E R N O PINION: SIC 679 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N u m b er o f O bservations: (% ) - 10 -0.0747 -0.101 225 -9 -0.0885 1.062 226 -8 0.7925 2.619 *** 222 -7 0.2197 1.541 225 -6 0.9482 4.419 *** 226 -5 -0.4813 0.257 224 -4 1.1225 4.654 *** 223 -3 1.8486 6.364 *** 222 2 -0.3134 - 1 0.8524 4.071 *** 224 0 0.6616 2.995 *** 224 + 1 -0.0798 2.458 ** 224 (-1.1) 1.4342 9.524 *** 224 +2 -0.7623 -0.934 224 +3 -0.0426 -0.499 225 +4 0.0618 0.333 222 +5 1.5869 2.715 *** 218 +6 1.5274 1.874 *♦ 218 + 7 0.6557 1.881 ** 218 + 8 -0.6315 -3.157 *** 218 +9 0.1041 -2.624 *** 218 + 10 -1.4039 -4.712 *** 218 . -1.048 224 *** Significant at the 1% level under a tw o-tailed test, ** Significant at the 5% level under a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 202 Table A-14 A V E R A G E A B N O R M A L RETU RN S TO RIVALS OF FIRM S A N N O U N C IN G A G O IN G CO N CERN O PIN IO N : SIC 737 Day R elative To A verage Abnorm al A nnouncem ent Return: Date: (%) - 10 -0.4627 -0.943 116 -9 0.0136 -0.072 118 -8 0.1068 0.601 118 -7 0.2118 0.757 118 -6 0.7135 2.422 ** 118 -5 0.4433 0.524 118 -4 0.5094 1.342 118 -3 -0.6532 -1.759 * 118 . 2 -0.3581 -0.894 118 - 1 0.0277 0.671 118 0 0.2119 0.801 118 + 1 -0.3307 -0.878 118 (-1.1) -0.0911 0.593 118 +2 0.0715 -0.063 118 + 3 0.0561 0.091 118 +4 0.5435 1.573 118 + 5 0.1223 0.249 118 +6 0.9548 1.358 118 + 7 -0.4792 -1.002 118 + 8 -0.4059 -0.639 118 +9 0.5432 1.212 118 + 10 0.7015 1.111 118 T-value: N um ber o f O bservations: *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 203 Table A - 15 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A G O IN G C O N C ER N OPINION: SIC 738 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (%) - 10 -0.7481 -2.722 ** 18 -9 -1.2686 -2.049 * 18 -8 4.8591 5.678 *** 18 -7 -3.3876 -5.093 *** 18 -6 3.0711 5.107 *** 18 -5 0.7652 1.995 * 18 -4 -2.0563 -1.142 18 -3 1.9209 0.654 18 . 2 -2.3396 -0.432 18 - 1 0.7017 -1.384 18 0 -0.6959 -1.775 * 18 + 1 0.7246 (-1.1) 0.7303 + 2 0.571 18 -2.589 ** 18 -1.9491 -1.713 18 + 3 1.3446 -0.237 18 +4 -1.7477 -1.053 18 + 5 3.1823 2.067 * 18 +6 1.7586 3.169 *** 18 + 7 -0.4615 3.468 *** 18 + 8 -1.4676 -3.431 *** 18 + 9 1.9564 2.187 ** 18 + 10 -2.4725 -2.201 ** 18 *** S ignificant at the 1% level under a tw o-tailed test. ** S ig nificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 204 Table A - 16 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS O F FIRM S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 808 Day R elative To A verage A bnorm al A nnouncem ent D ate: Return: - 10 0.1661 0.454 18 -9 0.7746 0.761 18 -8 -1.1298 -0.611 18 -7 -0.6574 -0.748 18 -6 -2.2911 -1.471 18 -5 -0.2027 -0.667 18 -4 1.7434 1.766 18 -3 -3.4785 -2.664 ** 18 2 -1.1848 -0.465 18 - 1 0.8604 0.561 18 0 3.0005 2.788 ** 18 + 1 1.3206 1.803 * 18 (-1-1) 5.1814 5.152 *** 18 + 2 -0.8908 -0.418 18 + 3 1.7868 1.554 18 +4 0.4578 0.351 18 + 5 -0.0304 0.713 18 + 6 1.6455 1.866 * 18 + 7 1.8491 1.401 18 + 8 -3.8252 +9 2.1379 1.723 18 + 10 -1.5072 -1.523 18 . N um ber o f T-value: O bservations: (% ) -3.203 *** 18 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 205 Table A - 17 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS OF FIR M S A N N O U N C IN G A G O IN G C O N C E R N O PIN IO N : SIC 809 D ay R elative To A nnouncem ent Date: Average Abnorm al Return: - 10 -1.9064 -0.775 17 -9 3.1941 1.162 17 -8 0.9834 1.594 17 -7 -1.8701 -1.725 17 -6 -0.1914 -0.192 17 -5 3.0519 1.403 17 -4 -2.6141 -1.025 17 - j -0.2619 -0.412 17 . 2 1.5825 1.367 17 - 1 0.9941 1.275 17 0 2.8386 0.561 17 + 1 -1.0615 0.293 17 (-U ) 2.7711 2.131 ** 17 + 2 0.0789 0.293 17 + 3 -1.7491 +4 T-value: N um ber o f O bservations: (%) -2.434 ** 17 0.3609 -0.162 17 + 5 2.5867 0.361 17 +6 -2.8203 -1.005 17 + 7 1.1446 0.352 17 + 8 0.3612 0.418 17 +9 3.2521 2.529 ** 17 + 10 -1.2648 -0.557 17 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 206 Table A -18 A V E R A G E A B N O R M A L RETU R N S TO RIVALS O F FIR M S A N N O U N C IN G A G O IN G C O N C ER N O PIN IO N : SIC 873 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (%) - 10 -0.1819 -0.337 34 -9 -0.4981 -0.112 34 -8 1.8053 -7 -1.1458 -1.171 34 -6 0.9924 1.063 34 -5 -0.4638 -0.588 34 -4 -0.3321 -0.863 34 -3 1.2838 1.461 34 . 2 -0.6481 -0.534 34 - 1 -0.3377 -0.306 34 0 0.8874 0.789 34 + 1 1.0619 1.712 * 34 (-1.1) 1.6116 2.195 ** 34 +2 0.2019 -0.138 34 + 3 -1.3868 -1.319 34 +4 0.6692 0.667 34 +5 0.6833 0.666 34 +6 -1.0867 -1.221 34 + 7 0.2383 0.051 34 +8 0.9232 1.142 34 +9 -0.5317 -0.058 34 + 10 -1.2974 -2.137 ** 34 1.739 * 34 *** Significant at the 1% level under a tw o-tailed test, ** Significant at the 5% level under a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 207 Table A -19 A V E R A G E A B N O R M A L RETU RN S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A LIFIED A UD IT O PIN IO N : SIC 131 Day R elative To A nnouncem ent Average A bnorm al Return: Date: (%) - 10 0.1701 0.403 152 -9 -0.1791 -0.647 152 -8 -0.6267 -1.846 * 152 -7 -0.1935 -0.789 152 -6 -0.3461 -0.681 152 -5 0.7891 2.698 ♦** 152 -4 0.1908 0.207 152 -3 0.0125 1.231 152 . 2 -0.0531 0.716 152 - 1 -0.3628 -1.281 152 0 0.1171 1.536 152 + 1 -0.5993 -1.733 * 152 (-1.1) -0.8449 -1.477 152 + 2 0.2271 1.049 152 + 3 0.0241 0.411 152 +4 0.1415 -0.028 152 + 5 0.0524 0.196 152 + 6 0.3628 1.157 152 + 7 0.8699 1.781 * 152 + 8 -0.0495 -0.239 152 + 9 0.4729 0.354 152 + 10 -0.2464 -0.827 153 N um ber o f T-value: O bservations: *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 208 Table A-20 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A L IF IE D A U D IT OPINION: SIC 138 Day R elative To A nnouncem ent A verage A bnorm al Return: Date: (% ) - 10 -0.0158 -9 2.7401 -8 -0.4448 -0.426 43 -7 -1.3025 -1.541 43 -6 0.4139 0.184 43 -5 -0.5809 -0.922 43 -4 -0.6379 -0.822 43 -3 0.5262 0.868 43 . 2 0.6802 0.782 43 - 1 -0 . 0 2 2 2 0.521 43 0 -0.0627 0.227 43 + I 0.8579 1.785 * 43 ( - 1, 1 ) 0.7731 2.532 ** 43 +2 -1.7173 -1.619 43 +3 0.9642 1.561 43 +4 -0.6916 -1.229 43 + 5 1.2171 1.513 43 +6 -0.6801 -0.809 43 + 7 -0.0628 0.235 43 +8 -0.4508 -0.023 43 + 9 -0.5315 -0.775 43 + 10 1.1561 T-value: N um ber o f O bservations: 0.459 43 2.975 *** 43 1.769 * 43 *** Significant at the 1% level un d er a tw o-tailed test. ** S ignificant at the 5% level un d er a tw o-tailed test. * Significant at the 10% level un d er a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 209 Table A-21 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS OF FIR M S A N N O U N C IN G A Q U A L IF IE D A U D IT O PIN IO N : SIC 245 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: N um ber o f T-value: O bservations: (%) - 10 0.5703 0.861 -9 -2.4034 -8 1.6422 1.484 11 -7 1.5144 0.854 11 -6 0.6252 0.806 11 - 5 -0.4597 -0.909 11 -4 -0.6148 -0.844 11 -3 -0.0479 0.694 11 . 2 1.7371 2.176 * 11 - 1 -0.9032 -1.268 11 0 -1.0083 -0.318 11 + 1 0.1865 -0.262 11 ( - 1. 1 ) -1.7251 -1.848 * 11 +2 0.8519 1.119 11 +3 -1.6384 -1.559 11 +4 4.4091 4.675 *** 11 + 5 0.6737 0.564 11 +6 0.8021 0.998 11 + 7 -0.0786 -0.786 11 + 8 0.8231 1.682 11 +9 -0.3385 -1.386 11 + 10 0.5051 0.439 11 -2.607 ** 11 11 *** S ignificant at the 1% level under a tw o-tailed test ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 210 Table A-22 A V E R A G E A B N O R M A L R ETU R N S TO R IVALS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 283 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: N um ber o f T-value: O bservations: (%) - 10 0.1483 0.318 132 -9 -0.4471 -1.405 132 -8 -0.2402 -1.191 132 -7 -0.3123 -0.708 132 -6 0.5031 1.966 ** 132 -5 0.0842 0.571 132 -4 -0.8416 -2.825 *** 132 -3 -0.4375 -1.605 132 . 2 -0.6226 -1.494 132 - 1 -0.3576 -0.648 132 0 -0.1375 -0.643 131 + 1 -0.4255 -1.193 131 ( - 1, 1) -0.9207 -2.484 ** 131 +2 -0.2403 -0.772 131 +3 -0.3437 -1.075 131 +4 -0.6961 -2.614 *** 131 +5 -0.4652 -1.816 * 130 +6 -0.6122 -1.638 130 + 7 0.2291 0.395 130 + 8 -0.1245 -0.826 130 +9 -0.1198 -0.403 130 + 10 -0.4102 - 1. 121 130 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission. 211 Table A-23 A V ER A G E A B N O R M A L RETU RN S TO RIV A LS OF FIR M S A N N O U N C IN G A Q U A LIFIED A UD IT O PIN IO N : SIC 307 Day R elative To A verage Abnorm al A nnouncem ent Return: Date: (%) - 10 0.4518 0.656 14 -9 -2.1341 -1.166 14 -8 -0.4707 -0.578 14 -7 0.1388 0.341 14 -6 -0.5508 -0.398 14 -5 5.4853 -4 -1.6065 -0.746 14 -3 0.1222 0.406 14 . 2 0.4232 0.274 14 -1 -1.0468 -0.746 14 0 0.5886 0.601 14 + 1 0.1599 0.239 14 (-M) -0.2983 0.095 14 +2 0.2844 0.231 14 +3 0.3031 -0.072 14 +4 -1.5847 -1.317 14 + 5 0.1084 0.894 14 +6 -0.4691 -0.545 14 + 7 0.6147 0.659 14 + 8 -0.9781 -0.968 14 +9 -0.7331 -0.381 14 + 10 1.3647 1.085 14 N um ber o f T-value: 2.673 ** O bservations: 14 *** Significant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission . 212 Table A -24 A V E R A G E A B N O R M A L RETURNS T O RIV A LS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 331 Day R elative To A nnouncem ent Date: A verage Abnormal Return: T-value: N um ber O f O bservations: (% ) -10 0.2999 0.103 17 -9 0.1908 -0.195 17 -8 0.6465 0.547 17 -7 0.6871 2.281 ** 17 -6 -0.1735 -1.301 17 -5 0.7981 0.371 17 -4 -0.7465 -0.918 17 -3 -0.9979 -0.931 17 . 2 -0.4363 -0.998 17 - 1 -0.2721 -0.425 17 0 -0.9905 -1.091 17 + 1 0.0948 -0.057 17 (-1 .0 -1.1678 -1.572 17 +2 -0.3264 -0 . 6 6 8 17 + 3 1.2956 +4 -1.3568 -1.675 17 + 5 0.2346 0.248 17 +6 -0.2459 -0.093 17 + 7 1.2201 0.926 17 + 8 0.1751 -0.762 17 +9 -1.1169 -1.318 17 + 10 -0.0968 -0.366 17 1.855 * 17 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 213 Table A-25 A V E R A G E A B N O R M A L R E T U R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A L IF IE D A U D IT O PIN IO N : SIC 353 Day R elative To A nnouncem ent Date: A verage A bnorm al R eturn: - 10 2.8104 3.429 *** 18 -9 1.2914 1.545 18 -8 -1.3909 -2.049 * 18 -7 0.8504 0.981 18 -6 0.7269 1.354 18 -5 0.0548 -0.218 18 -4 0.6166 0.953 18 -3 -0.7574 -0.395 18 . 2 0.3446 0.159 18 - 1 -1.2771 -1.055 18 0 1.4464 1.471 18 + 1 0.8024 1.573 18 (-LD 0.9718 1.989 * 18 + 2 0.4689 0.377 18 + 3 -0.2388 0.206 18 + 4 0 .8184 1.186 18 + 5 -0.3688 -0.317 18 + 6 0.9058 1.305 18 + 7 0.6235 0.796 18 + 8 0.5441 0.381 18 +9 0.3537 0.359 18 + 10 0.9498 1.311 18 T-value: N um ber o f O bservations: (% ) *** S ignificant at the 1% level under a tw o-tailed test ** Significant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 214 Table A-26 A V E R A G E A B N O R M A L RETU RN S TO RIV A LS OF FIR M S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 357 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N u m b er o f O bservations: (% ) - 10 -0.6303 -1.374 132 -9 -0.6083 -1.528 132 -8 -0.3252 -0.653 132 -7 -0.7244 -2.937 *** 132 -6 0.5215 1.431 132 - 5 0.2189 1.607 132 -4 0.3282 0.959 132 -3 -0.3479 -0.736 132 . 2 -0.3521 -1.406 132 - 1 0.1599 0.124 132 0 -0.0542 0.651 132 + I 0.7968 1.689 * 132 ( - 1. 1 ) 0.9025 2.464 ** 132 +2 -0.8149 -2.743 *** 132 + 3 0.3941 1.574 132 +4 0.4576 1.191 132 + 5 -0.2775 -0.472 132 +6 -0.1331 -0.806 132 + 7 0.4621 1.503 132 + 8 -0.3645 -1.028 132 +9 0.3076 0.998 132 + 10 0.6107 1.437 132 *** Significant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 215 T able A-27 A V E R A G E A B N O R M A L R ETU R N S T O RIV A LS O F FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 366 Day R elative To A nnouncem ent Date: A verage Abnorm al Return: N um ber o f T-value: O bservations: (% ) - 10 -0.3397 -9 0.9535 4.095 *** 61 -8 0.2841 0.805 61 -7 1.2818 3.873 *** 61 -6 0.4909 1.746 * 61 -5 -0.1793 -0.741 61 -4 -0.4091 -1.351 61 -3 0.3211 1.659 61 . 2 -1.0838 -1 1.0351 0 0.0775 + 1 0.6866 1.791 * 61 ( - 1 . 1) 1.7991 5.338 *** 61 + 2 -0.3279 -1.441 61 +3 -0.1623 0.071 61 +4 0.7791 1.133 61 + 5 0.2681 0.557 61 +6 0.0851 0.284 61 + 7 0.3381 1.131 61 + 8 0.1997 0.955 61 +9 0.1007 0.081 61 + 10 0.0031 -0.053 61 -0.331 -2.173 ** 3.879 *** -0.331 61 61 61 61 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 216 T able A-28 A V E R A G E A B N O R M A L R E T U R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A L IF IE D A U D IT OPINION: SIC 367 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (% ) - 10 -0.0844 -0.164 41 -9 -0.7277 -1.605 41 -8 -0.6565 -1.714 * 41 -7 -0.3196 -0.907 41 -6 0.2411 1. 021 41 -5 -0.2883 -0.367 41 -4 -0.1583 -0.261 41 -3 -0.5032 -0.929 41 . 2 -0.3001 -1.154 41 -1 -0.6753 -0 . 6 8 6 41 0 -0 . 2 2 2 2 -0.163 41 + 1 0.1283 -0.254 41 (-KD -0.7692 -1.103 41 + 2 -0.3391 -1.045 41 + 3 0.9993 1.896 * 41 + 4 0.0931 1.399 41 + 5 -0.4463 -0.842 41 + 6 -0.5086 -0.959 41 + 7 0.1251 0.243 41 + 8 0.1602 0.726 41 +9 -0.5245 -0.874 41 + 10 0.6301 1.338 41 *** Significant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level u n d er a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 217 T able A -29 A V E R A G E A B N O R M A L R ETU R N S TO RIV A LS OF FIR M S A N N O U N C IN G A Q U A L IFIE D A U D IT O PIN IO N : SIC 372 Day R elative To A nnouncem ent A verage A bnorm al Return: Date: (% ) - 10 2.2961 2.811 *** 27 -9 0.5451 1.929 * 27 -8 0.1468 0.271 27 -7 -1.1997 -1.385 27 -6 1.2099 0.372 27 -5 0.0808 -0.319 27 -4 -0.0155 -0.425 27 -3 0.4043 0.721 26 . 2 -0.3067 0.298 26 - 1 -1.4616 -1.685 26 0 0.1698 0.137 26 + 1 -0.8423 -1.368 26 (- U 1 ) -2.1341 -2.916 *** 26 + 2 1.1498 +3 -1.6432 -2.692 ** 26 +4 2.0811 2.307 ** 26 + 5 -2.5654 -3.918 *** 26 +6 0.8048 2.434 ** 26 + 7 0.1683 0.783 26 + 8 -1.1671 -0 . 8 6 6 27 + 9 -0.1484 0.601 27 + 10 -0.4083 -0.854 27 T-value: 0.992 N um ber o f O bservations: 26 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 218 Table A -30 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS O F FIRM S A N N O U N C IN G A Q U A L IF IE D A U D IT O PIN IO N : SIC 382 D ay R elative To A nnouncem ent A verage A bnorm al Return: Date: (%) - 10 -0.3944 -1.094 100 -9 0.1974 0.098 100 -8 0.9581 3.072 *** 100 -7 0.1676 0.138 100 -6 -0.1287 0.171 100 -5 0.3305 0.962 100 -4 0.5582 1.539 100 -a 0.2078 -0.446 100 . ? -0.6104 -2.092 ** 100 - 1 0.5527 1.063 100 0 -0.7411 -1.387 100 + 1 -0.1735 -1.114 100 (-LD -0.3619 -1.437 100 +2 -0.3251 -0.409 100 + 3 -0.3102 -0.965 100 +4 0.9425 + 5 -0.3116 -1.218 100 +6 0.1551 0.359 100 + 7 -0.2115 -0.701 100 + 8 0.4955 1.301 100 +9 0.5078 1. 221 101 + 10 -0.4665 T-value: 3.015 *** -1.878 * N um ber o f O bservations: 100 100 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 219 Table A-31 A V ER A G E A B N O R M A L R ETU R N S TO RIVALS OF FIR M S A N N O U N C IN G A Q U A L IF IE D A U D IT OPINION: SIC 386 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: N um ber o f T-value: O bservations: (% ) -10 -0.3482 -0.187 11 -9 -0.5964 -0.878 11 -8 0.1803 0.856 11 -7 0.2611 0.959 11 -6 0.6604 1.954 * 11 -5 -0.1369 -0.311 11 -4 -0.5756 -1.065 11 -3 0.1165 0.712 11 . 2 -0.5766 -0.247 11 - 1 0.1347 0.451 11 0 0.0023 -0 . 2 0 1 11 + 1 -2.6698 -3.239 *** ]1 (-LD -2.5328 -2.989 ** 11 +2 -0.2037 -0.643 11 + 3 -0.5722 -1.134 11 +4 0.0036 -0.003 11 + 5 0.4954 1.583 11 +6 0.8931 1.835 11 + 7 0.6825 0.983 || + 8 1.5456 3.207 ** 11 + 9 - 1.1021 -0.301 11 + 10 0.4396 1.528 11 *** Significant at the 1% level under a tw o-tailed test, ** Significant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 220 Table A-32 A V E R A G E A B N O R M A L RETURNS TO R IVALS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 491 Day R elative To A nnouncem ent Date: A verage Abnorm al Return: T-value: N um ber o f O bservations: (%) - 10 -0.1991 -2.822 *** 366 -9 -0.2473 -2.949 *** 366 -8 -0.2333 -3.926 *** 366 -7 -0.3913 -5.921 *** 366 -6 -0.0358 -0.236 366 -5 -0.0549 -1.375 366 -4 -0.2559 -4.354 *** 366 -3 -0.4771 -7.258 *** 366 . 2 -0.0928 - 1. 2 1 1 366 - 1 0.2571 4.864 *** 366 0 -0.1991 -3.158 *** 366 + 1 0.0668 0.785 366 (-LD 0.1247 2.491 ** 366 +2 0.4184 6.615 *** 366 + 3 -0.2095 -4.056 *** 366 +4 -0.2578 -5.094 *** 366 + 5 -0.2284 -2.679 *** 366 + 6 0.1989 3.402 *** 366 + 7 -0.3087 -4.358 *** 366 + 8 -0.0071 0.074 366 + 9 -0.0107 0.272 366 + 10 -0.1714 -2.855 *** 366 *** S ignificant at the 1% level under a tw o-tailed test. ** S ig nificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 221 Table A-33 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A L IFIE D A U D IT O PIN IO N : SIC 493 Day R elative To Average A bnorm al A nnouncem ent Date: Return: (% ) - 10 -0.3492 -1.951 * 33 -9 -0.1006 -0.571 33 -8 -0.0301 -0.135 33 -7 -0.4845 -2.545 ** 33 -6 -0.2861 -1.507 33 -5 0.1424 0.809 33 -4 0.2731 1.581 33 j -0.2536 -1.171 . 2 -0.1962 -0.906 33 - 1 -0.2898 -1.475 33 0 -0.2901 -1.659 n^ JJ + 1 -0.0665 0.201 ^ *■» JJ (-M) -0.6464 -2.933 *** 33 +2 -0.3787 -1.911 * 33 +3 -1.0701 -5.293 *** 33 +4 0.2447 1.325 +5 0.0304 0.502 33 +6 0.3855 1.522 33 + 7 0.0977 0.371 j j + 8 0.1784 1.067 33 + 9 -0.1416 -0.519 33 + 10 -0.1347 -0.892 33 - N um ber o f T-value: O bservations: *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 222 Table A -34 A V E R A G E A B N O R M A L R ETU R N S TO R IV A LS O F FIR M S A N N O U N C IN G A Q U A L IF IE D A U D IT O PIN IO N : SIC 495 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: - 10 -0.3814 -0.159 13 -9 -0.4615 -1.023 13 -8 1.0113 0.929 13 -7 -1.5758 -1.571 13 -6 -0.6645 -0.438 13 -5 0.5195 0.424 13 -4 -0.3697 0.576 13 N um ber o f T-value: O bservations: (% ) - j 1.3036 1.734 13 - 2 0.0551 -0.073 13 - 1 -0.2009 -0.496 13 0 -0.0757 -0.061 13 + 1 -0.0042 -0.776 13 ( - 1, 1 ) -0.2808 -1.333 13 + 2 0.4769 0.686 13 + 3 -0.8226 -0.765 13 +4 -0.0891 0.007 13 +5 -0.3985 -0.657 13 +6 0.3891 0.055 13 + 7 2.2864 2.707 ** 13 + 8 -0.2525 -0.275 13 +9 -0.7982 - 1.101 13 + 10 0.2721 0.649 13 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * Significant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 223 Tabie A-35 A V E R A G E A B N O R M A L RETU RN S T O R IV A L S OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O P IN IO N : SIC 506 Day R elative To A nnouncem ent Date: A verage Abnorm al Return: T-value: N um ber o f O bservations: (%) - 10 0.3487 0.671 13 -9 0.5046 0.285 13 -8 -0.6766 -0.948 13 -7 0.0975 -0.195 13 -6 1.2383 1.249 13 -5 -0.1856 -0.419 13 -4 -1.3891 -0.921 13 -3 0.3321 0.328 13 . 2 1.4098 1.372 13 - 1 1.4635 1.256 13 0 3.6256 2.521 ** 13 + 1 1.3003 0.895 13 ( - 1. 1) 6.3894 4.671 *** 13 +2 -0.2191 -0.463 13 + 3 0.8067 0.843 13 +4 1.4648 0.855 13 + 5 1.4129 0.964 13 +6 2.0941 1.622 13 + 7 -1.2609 -1.054 13 + 8 1.4271 0.974 13 +9 -0.5943 -0.622 13 + 10 0.2092 0.518 13 *** S ignificant at the 1% level under a tw o-tailed test ** S ig nificant at the 5% level under a tw o-tailed test. * S ig n ificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 224 T able A-36 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 661 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: T-value: N um ber o f O bservations: (% ) - 10 1.7684 1.449 8 -9 -0.8125 -0.926 8 -8 -1.8895 -0.617 8 -7 -0.6225 -0.922 7 -6 0.8879 0.891 7 -5 -0.2576 -0.968 7 -4 0.9524 1.432 7 -3 -2.0089 -1.359 7 . 2 1.6917 0.861 7 -1 -1.6515 -1.448 7 0 1.3805 0.678 7 + 1 3.0112 2.088 * 7 ( - 1 . 1) 2.7401 1.318 7 +2 -0.9536 -1.223 7 +3 2.5927 +4 -1.5582 -1.577 7 +■ 5 0.6089 0.878 7 +6 0.4441 0.052 7 + 7 -3.0376 + 8 1.7188 0.791 7 +9 0.8529 0.773 7 + 10 -1.0549 -0.975 7 2.639 ** -2.762 ** 7 7 *** Significant at the 1% level under a tw o-tailed test. ** Significant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 225 Table A-37 A V E R A G E A B N O R M A L R ETURNS TO RIVALS OF FIR M S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 679 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: - 10 0.0714 -9 -1.0258 -8 0.1455 0.671 84 -7 -0.1966 -1.024 84 -6 0.0642 1.102 87 -5 0.5371 3 3 4 4 *** 87 -4 0.0675 0.333 87 -3 0.0491 -0.577 87 . 2 0.0721 -0.133 87 - 1 -0 . 0 2 2 2 0.314 87 0 -0.4301 + 1 0.4502 1.088 87 ( - 1 .D -0 . 0 0 2 1 -1.408 87 +2 0.0373 0.259 87 + 3 -0.3501 -2.075 ** 87 +4 -0.0295 -0.414 87 +5 -0.6356 -3.531 *** 87 +6 -0.0541 -1.282 87 +7 0.2402 +8 0.1092 +9 0.5879 2.531 ** 87 + 10 0.1491 1.261 87 T-value: N um ber o f O bservations: (% ) 0.676 -3.178 *** -2.811 *** 1.974 * -0.485 84 84 87 87 87 *** S ignificant at the 1% level under a tw o-tailed test. ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission. 226 Table A-38 A V E R A G E A B N O R M A L R ETU R N S TO RIV A LS OF FIRM S A N N O U N C IN G A Q U A L IFIE D A U D IT O PIN IO N : SIC 701 Day R elative To A nnouncem ent Date: A verage A bnorm al Return: N um ber o f T-value: O bservations: (%) - 10 -0.6049 -0.857 34 -9 -0.0584 -0.322 34 -8 0.0076 -0.982 34 -7 -0.7235 -1.999 * 33 -6 0.2823 0.909 33 -5 0.1681 0.141 jj -4 0.5275 1.195 -3 -0.4736 -0.771 jj . 2 -0.1293 -0.165 33 - 1 -0.7991 - 1.212 33 0 0.4911 -0.079 33 + 1 0.5471 1.373 33 ( - 1. 1 ) 0.2391 0.081 33 + 2 0.3377 0.424 33 + 3 0.0147 0.166 jj +4 0.2155 0.802 33 + 5 -0.1434 -1.009 33 + 6 1.1952 + 7 -0.5145 + 8 0.7157 1.791 * 33 + 9 0.8278 1.282 33 + 10 -0.3048 -0.181 33 2.387 ** -0.758 33 33 *** S ignificant at the 1% level under a tw o-tailed test. ** S ig nificant at the 5% level under a tw o-tailed test. * S ig nificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of the copyright ow ner. Further reproduction prohibited w ithout p erm ission . 227 Table A-39 A V E R A G E A B N O R M A L R ETU R N S TO R IVALS OF FIR M S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 737 Day R elative To A nnouncem ent Date: A verage Abnorm al Return: T-value: N um ber o f O bservations: (% ) - 10 0.2747 -0.322 139 -9 -0.0574 -0.066 140 -8 0.1471 0.281 140 -7 0.3876 0.581 140 -6 -0.0286 -0.061 140 -5 0.2502 0.204 140 -4 -0.1244 -0.373 140 - j 0.3102 0.384 140 . 2 -0.1049 -0.465 140 - 1 -0.2153 0.839 140 0 -0.4306 -1.154 140 + 1 0.2684 -0.779 140 (-M) -0.3775 -1.094 140 +2 -0.0352 -0.107 140 + 3 -0.0781 -0.445 140 +4 -0.9023 -2.867 *** 140 + 5 -0.4973 -2.424 ** 140 + 6 0.0751 0.453 140 + 7 0.5443 0.964 140 + 8 0.8543 2.053 ** 140 + 9 -0.3569 -0.833 140 + 10 -0.2179 -0.978 140 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of th e copyright ow ner. Further reproduction prohibited w ithout p erm ission . 228 Table A-40 A V E R A G E A B N O R M A L R ETU R N S TO RIV A LS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PIN IO N : SIC 781 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: -1 0 1.1198 1.004 14 -9 -0.6754 -0.249 14 -8 -1.3093 -0.753 14 -7 -2.1973 -1.051 14 -6 0.2221 0.115 14 -5 1.1994 0.569 14 -4 2.0415 1.462 14 -3 -1.0971 -1.098 14 . 2 -0.5002 0.227 14 - 1 -3.1205 -2.285 ** 14 0 4.1578 1.985 * 14 + 1 -2.6396 -1.249 14 (-M) -1.6023 -1.549 14 +2 2.5883 1.018 14 + 3 4.3042 1.262 14 +4 -4.1115 +5 -0.7342 0.117 14 +6 0.4541 0.656 14 +7 2.1803 1.098 14 +8 0.5175 0.615 14 +9 -0.9714 -0.787 14 + 10 -2.7629 -1.625 14 N um ber o f T-value: O bservations: (%) -2.163 ** 14 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. 229 Table A-41 A V E R A G E A B N O R M A L R ETU R N S TO RIVALS OF FIRM S A N N O U N C IN G A Q U A LIFIED A U D IT O PINION: SIC 806 Day R elative To A verage A bnorm al A nnouncem ent Date: Return: - 10 2.4148 1.306 8 -9 0.6597 0.723 8 -8 1.3423 1.077 8 -7 -0.9378 -0.437 8 -6 0.7123 0.704 8 -5 -0.2357 -0.257 8 -4 -1.0481 -0.586 8 -3 2.3221 . 2 0.5475 -0.291 8 - 1 0.2901 0.315 8 0 0.1626 0.543 8 + 1 0.3347 0.039 8 (-1.1) 0.7873 0.898 8 +2 1.4723 1.154 8 + 3 0.0242 -0.128 8 +4 0.2377 0.118 8 + 5 0.4229 0.361 8 +6 -3.8652 -2.334 ♦* 8 + 7 -0.8531 -0.764 8 + 8 -0.0262 -0.167 8 +9 0.5591 0.725 8 + 10 1.8881 0.898 8 N um ber o f T-value: O bservations: (% ) 1.952 * 8 *** S ignificant at the 1% level under a tw o-tailed test, ** S ignificant at the 5% level under a tw o-tailed test. * S ignificant at the 10% level under a tw o-tailed test. R ep ro d u ced with p erm ission of the copyright ow ner. Further reproduction prohibited w ithout p erm ission . IMAGE EVALUATION TEST TARGET (Q A -3 ) A 1.0 E 1- 123 L. lii 2.2 U U£ 2.0 l.l 1.8 11.25 1.4 1.6 150mm A P P L IE D A INA4GE . Inc 1653 East Main Street Rochester, NY 14609 USA • ^ = ~— Phone: 716/48245300 -= ~-= = Fax: 716/288-5989 ■= = r - 0 1993. Applied Image. Inc. All Pig Ms Reserved R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout p erm ission. [...]... this project reveals the public opinion that a m edia announcem ent o f audit results contains useful in­ form ation about the announcing firm O f interest to the current research is w hether certain unfavorable audit opinion announcem ents, specifically the "subject to" qualified audit opinion and the "going con­ cern" audit opinion, contain useful inform ation applicable to rivals o f the announcing... announcem ent o f unfavorable audit opinions on the firms receiving those opinions W ith these findings as a foundation, the objectives o f this study are as follow s: (1) to determ ine the effects o f "subject to" qualified and "going concern" audit opin­ ions on the firms m aking such announcem ents in the business financial press; (2) to reveal the im pact o f these unfavorable audit opinion announcem... firm results near the announcem ent date o f a "subject to" qualified audit opinion or the "going concern" audit opinion This hypothesis is stated as follows: Hn: "Subject to" qualified and "going concern" audit opinions are not associated w ith significant negative changes in the stock returns o f the announcing firm R ep ro d u ced with p erm ission o f th e copyright ow ner Further reproduction prohibited... "subject to" qualified audit opinion and "go­ ing concern" audit opinion announcem ents have provided evidence o f a negative stock price reaction by the announcing firm (D opuch et al [1] and Jones [6]) H ow ever, no published research appears to exist that determ ines w hether these unfavorable audit opinion announcem ents affect the stock returns o f industry rivals B ecause audit o p in ­ ions provide... ascertain the ability, significance, and directional im pact o f certain industry and firm -level characteristics in explaining the abnorm al stock price m ovem ents o f industry' rivals brought about by unfavorable audit opinion announcem ents In m eeting these objectives, the present study fills an im portant void in the pre­ vious research by exam ining the announcem ent effects o f unfavorable audit. .. projects is their pointing out that "subject to" qualified and "going concern" audit opinions do not necessarily im ­ pact the firm s receiving these unfavorable opinions unless the outcom es o f the audits are m ade public through the financial press The Effects o f Inform ation Transfers Inform ation transfer studies deal with the resulting stock price adjustm ents w ithin and am ong industries when... aining variation O ne R ep ro d u ced with p erm ission o f the copyright ow ner Further reproduction prohibited w ithout p erm ission such factor m ay be the audit opinion outcom e, w hich som etim es accom panies earnings releases R esearch about audit opinions and their effect on stock prices has becom e popu­ lar T hese studies attem pt to determ ine if audit opinions contain inform ation useful... to identify how unfavorable audit opinion an ­ nouncem ents affect the industry rivals o f announcing firms R ep ro d u ced with p erm ission o f the copyright ow ner Further reproduction prohibited w ithout p erm ission q Objectives of the Study The literature has already established the existence o f intra-industry inform a­ tional transfers attributable to the announcem ent o f certain econom ic... ents; (2) qualified opinions, w hich explain that the financial statem ents are in accordance with G A A P and fairly present the current condition o f the firm "except for" a specific item or "subject to" the unknow n outcom e o f som e event; (3) disclaim er o f opinion, w hich is a refusal by th e auditor to express an opinion on the financial statem ents; and (4) an adverse opinion, w hich declares... ced with p erm ission o f th e copyright ow ner Further reproduction prohibited w ithout perm ission statem ents included a qualified audit opinion The research by D opuch et al [1 ] differs from the other studies in that it focuses on returns o f firm s w hen the "subject to" q ualifi­ cation o f the audit opinion draw s attention from the financial press Thus, this project reveals the public opinion