Does Product Market Competition Improve the Quality of Information Disclosure of Listed Companies? Empirical Evidence Based on China A-Share Market

Transcription

1 2017 4h Inernaonal Conference on Modern Economc Technology and Managemen (ICMETM 2017) ISBN: Does Produc Marke Compeon Improve he Qualy of Informaon Dsclosure of Lsed Companes? Emprcal Evdence Based on Chna A-Share Marke Huanyu Zhou* Chna Merchans Group Posdocoral Research Saon, Shenzhen, Guangdong, Chna ABSTRACT: Ths paper examnes how he produc marke compeon under Chnese marke condons affecs he qualy of nformaon dsclosure of lsed companes. Based on he daa of Chna A-share lsed companes from 2001 o 2014, we fnd ha he mprovemen of produc marke compeon measured from he ndusry level and corporae level has a sgnfcan negave mpac on he qualy of nformaon dsclosure of lsed companes. Afer furher dsngushng he earnngs managemen acves ha can change he qualy of nformaon dsclosure, he resuls confrm ha he ncrease of produc marke compeon a he ndusry level wll make he lsed companes end o posve earnngs managemen, whle he ncrease a he corporae level wll lead o more negave earnngs managemen. In addon, he nonlnear model confrms ha boh of hem have an nvered U-shaped relaonshp, bu here s only a sngle negave mpac of produc marke compeon on he qualy of nformaon dsclosure under he sample condons of Chnese lsed companes. Keywords: produc marke compeon; nformaon dsclosure; earnngs managemen 1 INTRODUCTION Wheher he facors a he managemen level wll affec he company s nformaon dsclosure behavor s a ho ssue n he feld of corporae governance. As a good exernal varable ha has a remendous mpac on he company s sae of operaon, produc marke compeon affecs he decson-makng of managers nformaon dsclosure. The relaed research has an mporan gudng sgnfcance on he reasonable adjusmen of he produc srucure of naonal economy, mprovemen of he capal marke supervson sysem and company s nernal governance mechansm. Based on he emprcal evdence of Chna A-share marke, hs paper examnes he acual mpac of Chna s produc marke compeon on he company nformaon dsclosure behavor, and compares wh he smlares and dfferences of measuremen mehods of produc marke compeon a dfferen levels n he mpac, and dscusses he possble causes of generang he above emprcal resuls hrough combnaon wh he nonlnear model. There are hree pars of *Correspondng auhor: huanyu1025@126.com possble nnovaon: (1) Keep a foohold n he crossng feld of ndusral economcs and corporae governance, arrange and heorecally consruc a reasonable pah of produc marke compeon affecng he qualy of nformaon dsclosure; (2) conclude he measuremen ndex of produc marke compeon, and carry ou nfluence research from ndusry and corporae level; (3) dfferenae and furher verfy he earnngs managemen acves ha change he qualy of corporae nformaon dsclosure, and use he non-lnear model o reasonably explan emprcal conclusons of produc marke compeon affecng he qualy of nformaon dsclosure n (2). 2 THEORETICAL ANALYSIS AND RESEARCH HYPOTHESES Produc marke compeon belongs o he nfluencng facors of he qualy of corporae nformaon dsclosure a he managemen level, and manly acs on he qualy of nformaon dsclosure hrough wo mechansms. Vewng from he nformaon mechansm, he assumpon ha he company reduces capal coss by 53

2 mprovng he qualy of nformaon dsclosure and hen eases he plgh of producon and operaon brough by produc marke compeon ofen fals o be esablshed. Because he melness of nformaon dsclosure ncreases he volaly of he sock, may cause an ncrease n he capal cos. In addon, f dsclosure self leads o more nformaon asymmery, ha s, nvesors do no beleve he auhency of he nformaon dsclosed, he qualy of nformaon dsclosure wll no longer become a facor of judgng he corporae value, and hen he capal cos wll be furher ncreased (Km and Verreccha, 1994). Under he envronmen of ferce compeon, he company wll choose o dsclose nvald nformaon, n order o avod predaory hreas from rvals (Verreccha 1983; Gerner e al. 1988). In oher words, he huge produc marke compeon pressure of he company grealy reduces he ransparency of corporae earnngs nformaon. The purpose of aggressve earnngs managemen (hgher earnngs managemen level) s acually o lm he flow of nformaon o poenal compeors and keep he company from possble compeon. Vewng from he governance mechansm, produc marke compeon has been regarded as a possble exernal governance mechansm for a longer perod of me. In case of effecve developmen, he neress of managers and shareholders wll be effecvely ed ogeher. The more nense he produc marke compeon, he lower he prvae benefs conrolled by he managemen level (Guadalup and Gonzalez, 2010). As a punshmen mechansm, he company s produc compeon wll resrc he manager s behavor (Har, 1983), hereby sgnfcanly reducng he earnngs managemen level. However, he produc marke compeon has a negave mpac on he company s own profably and ably o buffer rsks. Due o he objecve requremens of mananng relavely sable performance and avodng rsks n he operaon, he company has an ncenve o reduce he qualy of nformaon dsclosure. In he explanaon of Har (1983) model, Scharfsen (1988) poned ou ha, f he margnal uly of ncome level s srcly posve a he managemen level, produc marke fals o subsue corporae governance o a ceran exen, bu may weaken manager s ncenve, and fal o change he manager s lazy sae. In addon, he manager s aenon o career may mprove as he produc marke compeon pressure ncreases, resulng n managemen opporunsm and shor-sghed behavor (Fama, 1980; Narayanan, 1985; Roemberg and Scharfsen, 1990; Hermanln and Wesbach, 2012) and he researches show ha compeon can aggravae moral hazard problems (Mlgrom and Robers, 1992) and ncrease managemen escape behavor (Scharfsen, 1988). When he produc compeon of he ndusry n whch he company s locaed n s ferce, he buffer brough by excess profs n he form of Rcardo s ren owned by he company s relavely small (Peeraf, 1993), and s relavely more dffcul for he company o pass he cos mpac on o consumers. When he overall prof of he ndusry decreases, companes n relave monopolzng ndusres and compeve ndusres may be more engaged n earnngs managemen acves, n order o gan he rus and help of nvesors and credors, ensure normal operaon of he company, hus maxmzng he shor-erm value of he company (Shlefer, 2004). In he ndusres wh a hgh concenraon, each company occupes a hgher proporon of gross oupu of he ndusry; relave o he companes n he ndusres wh a relave low concenraon, he nformaon dsclosed by hem may provde more nformaon of he fuure produc demands n he ndusry n whch he company s locaed. Smlarly, he companes wh a sronger prcng ably (a huge produc marke power) may have a relavely lower movaon for earnngs managemen, and such companes ofen have more share prce nformaon conen (Peress, 2010) and hgher sock lqudy (Kale And Loon, 2011), he company s real earnngs performance becomes more sable, earnngs managemen and oher acves reducng he qualy of nformaon dsclosure are even more unnecessary. Hypohess 1: When he company confrons greaer produc marke compeon, ha s, he degree of concenraon of he ndusry n whch he company s locaed reduces or he company s prcng ably s weakened, he qualy of nformaon dsclosure wll be lower. An ncrease n he level of ndusral compeon wll reduce he corporae profs. Vewng from he resrcon mechansm, he company wll face hgher lqudy rsks n he fuure (Hou and Robnson, 2006). A declne n he corporae profs s ofen regarded as a sgnal of declne n he ndusral compeon, whch may lead o greaer punshmen on he sock marke (Markaran and Sanalo, 2014), hus mparng managers prvae ren (n case of srcer supervson, vsble and nvsble ncome wll be reduced, and even dsmssal, ec.). To avod hs suaon, managers ofen ake measures o dsor he company s real performance. Researches show ha, wll become an nsrumen o dsor economc performance when earnngs managemen coss are lower and are no easly found by exernal sakeholders (Bagnol and Was, 2000). Compared o real earnngs managemen, manpulave, coss of accrued profs are lower, and s dffcul for sock nvesors, fnancal analyss and audors o accuraely denfy he rue mpac of accrued earnngs managemen. For managers, he posve earnngs managemen s an effecve ool o proec own neress. When he produc marke compeon a he corporae level ncreases, he margnal prof of he company has rsks of declne. If he unmanaged surplus reflecs a hgher demand of he marke on he company s producs n he fuure, may ncrease he parcpaon of compeors and endanger he neress of 54

3 he company. In order o avod he fnancal repors dsclosed o pass more nformaon on he company s fuure busness developmen sae o he ousde world (Chen Jun and Xu Yude, 2011), preven poenal enrans of he ndusry, elmnae possble predaory rsks, and consoldae her own ndusry saus, he company can ncrease he negave earnngs managemen behavor and creae an lluson ha he performance of man busness fals o reach expecaons. Hypohess 2: An ncrease n he produc marke compeon a he ndusry level, or a decrease n he concenraon of he ndusry n whch he company s locaed wll make he earnngs managemen acves ha can reduce he qualy of corporae nformaon dsclosure end o posve earnngs managemen. Hypohess 3: An ncrease n he produc marke compeon a he corporae level, or weakenng he marke prcng ably of he company wll make he earnngs managemen acves ha can reduce he qualy of corporae nformaon dsclosure end o negave earnngs managemen. 3 INDEX MEASUREMENT AND DATA SELECTION 3.1 Measuremen of produc marke compeon Based on he domesc and nernaonal researches on he mpac of produc marke compeon on he company, hs paper selecs he ndusry concenraon (HHI, Wang Xongyuan, Lu Yan, 2008; Wu Haomn, e al., 2012) and produc-cos margn (Daa e al., 2013) for measuremen, whch reflecs he marke compeon suaon from he ner-ndusry compeon and compeon n he ndusres. 3.2 Measuremen of qualy of nformaon dsclosure The qualy of nformaon dsclosure s ofen confused wh he nformaon ransparency, and he curren wo measuremen mehods of nformaon dsclosure rang and scholar s self-bul ndex scorng are ofen dffcul o be ransplaned or conan unrelaed nfluencng facors. Ths paper selecs earnngs managemen (manpulave accrued prof DA, Dechow e al., 1995) ha can more accuraely descrbe he qualy of nformaon dsclosure for measuremen, and he nformaon dsclosure rang daa of Shenzhen Sock Exchange, Huon and oher people s mehods 1 n 2009 are used for robusness es. 3.3 Oher conrol varables To conrol he mpac of he company s characersc facor on he qualy of nformaon dsclosure, hs paper selecs varables, company scale logasse (logarhm of year-end oal asses, Bhushan, 1989), book-o-marke rao MB (logarhm of prce/book value rao, Chen Yng, 2008), corporae growh Growh (oal asses growh rae), leverage rao leverage (non-curren lables / oal asses, Trueman and Tman, 1988), company ra volaly volaly (daly reurn rae of corporae sock - correspondng o daly reurn rae of Chna A-share marke, Mohanram and Rajgopal, 2009); o conrol he mpac of he organzaon and managemen level on he qualy of nformaon dsclosure under he corporae governance srucure, hs paper selecs he naure of propery rgh SOE (Lu Zhengfe, e al., 2012), shareholdng proporon of conrollng shareholders Frs (shareholdng proporon of he larges conrollng shareholder of he company), sze of board of drecors Board (number of board of drecors), proporon of ndependen drecors ndrecor, dualy Dual (servng as charman and general manager s 1, oherwse 0), shareholdng proporon of senor execuve Mshare, execuve salary Salary (logarhm of a sum of op hree execuve salary wh Un 10,000), shareholdng proporon of nsuonal nvesors Insu (See Table 1). 3.4 Daa Ths paper selecs he annual daa of all Chna A-share lsed companes from 2001 o 2014 as samples, and screens accordng o he followng crera: (1) consderng he avalably of he gong-concern daa of he company, selec he companes landed Chna A-share n 2001 and before; (2) remove he samples of lsed companes classfed no ST, PT, and fnancal nsurance ndusry n ha year. The corporae naure, governance srucure and fnancal daa are from Wnd Info Fnancal Termnal. The daa of he company s sock radng saus are from CSMAR daabase. The daa of he company s hsorcal dynamc ndusry belongs are from he ndusry classfcaon and ndusry change daabase of RESSET lsed companes. To avod he mpac of exreme values, Wnsorze processng s gven o he connuous varables conaned n he daa a 1% and 99% of quanle level. Afer dsngushng he manufacurng ndusry by wo levels of codes, he research samples n hs paper are nvolved n 43 ndusres. Accordng o he sascs of he defnon of ndusry concenraon n Table 1, he ndusry concenraon n Chna from 2001 o 2014 has been mananed a a hgher level of abou 0.16; he dfference n he subdvded ndusry concenraon s larger, bu he ner-ndusry dfference s sgnfcanly shrunk around 2007, whch may be relaed o he reshuffle and mergers and acqusons whn he ndusres caused by he fnancal crss. 1 Huon e al. (2009) Take Dsclosure, = ( DA, -1 + DA, -2 + DA, -3 ) as he proxy varable of nformaon dsclosure qualy. 55

4 Table 1. Defnon of core varables. Varable Code Measuremen mehod Produc marke compeon Qualy of nformaon dsclosure Indusry concenraon (ndusry level) Prce - cos margn (corporae level) Manpulave accrued prof HHI PCM absda n n 2, x s he operang ncome of lsed HHI = ( x / x ) = 1 = 1 company n ha year; n s he number of companes n he ndusry PCM = ( Sales -COGS - SG & A) / Sales, Sales s he operang ncome of he company n ha year; COGS s he correspondng operang coss; SG & A s he correspondng sales and managemen coss n he accounng subjec To oban by usng amended Jones model, ake he absolue value: TAcc ^ ^ ^, 1 Sales, Re c, PPE, DA, = ( α + β1 + β2 ) Asse Asse Asse Asse, 1, 1, 1, 1 Reverse produc marke compeon Reverse produc marke compeon Reverse qualy of nformaon dsclosure Table 2. Descrpve sascs of varables. Varables Observed Mean Sandard Mnmum Maxmum value value devaon value value DA AbsDA HHI PCM LogAsse Growh leverage MB volaly Frs Board ndrecor Insu Mshare Salary EMPIRICAL TEST AND RESULT ANALYSIS 4.1 Produc marke compeon and qualy of nformaon dsclosure Model I s used o emprcally es he relaonshps beween he level of corporae compeon n he ndusry and he qualy of corporae nformaon dsclosure. Among hem, comconrol s a conrol varable reflecng he characerscs of he company; govconrol s a conrol varable reflecng he corporae governance srucure; PMC reflecs he produc marke compeon of he company. Here, he proxy varable a he ndusry level HHI s frs subsued no for analyss, and specfc varables have been descrbed above. The model conrols he annual effec and ndusry effec. absda = α + β PMC + γ comconrol,,, + δ Govconrol + ndusry + year + ε,, ---I Table 3. es resuls of qualy of nformaon dsclosure of sub-samples wh dfferen ndusry concenraon. Group Observed absda mean Sandard devaon value value HHI > medan HHI < medan es of wo samples Dff=mean(HHI > medan) -(HHI < medan) H : dff! = 0 Pr( T > ) = H : dff < 0 Pr(T<)= All he esmaon resuls n hs paper are obaned by usng he panel fxed effec model based on Hausman es, and by usng he cluserng robus sandard errors wh he company as cluserng varables. The coeffcen of varable HHI n Table 4 (1) s negave and sgnfcan a he confdence level of 1%, ndcang ha he fercer he compeon n he ndusry, he lower he qualy of he corporae nformaon dsclosure, and he produc marke compeon fals o mprove he of lsed companes on Chna s capal marke, confrmng hypohess 1. Dvdng by he medan of ndusry concenraon n each year medan, HHI > medan s lower han he ndusral compeon group, HHI < medan s hgher han he ndusral compeon group. Accordng o Table 3, he mean value of he qualy of nformaon dsclosure n a low ndusral compeon group s sgnfcanly hgher han ha of he hgh ndusral compeon group, also confrmng hypohess 1. Table 4 (1) reflecs ha he coeffcen of explanaory varable of he corporae governance srucure s negave, ndcang ha he organzaon managemen level has a sgnfcan posve mpac on he qualy of corporae nformaon dsclosure, whch s conssen wh he relevan conclusons of Y Zhhong, e al. (2010). However, he correspondng coeffcen afer furher addng HHI and s cross mulply em s nsgnfcan, ndcang ha here s no complemenary or alernave relaonshp beween produc marke compeon and corporae governance srucure n erms of he mpac on he qualy of corporae nformaon dsclosure. The proxy varable reflecng he marke compeon of he company as an ndvdual a he corporae level PCM s subsued no as a proxy varable of PMC, and model I s reused for esmaon. As shown n Table 4 (2), here s no change n he drecon and sgnfcance of he coeffcen of varable PCM. The larger PCM, he greaer he company s produc power, he greaer he prcng ably, he easer o oban monopoly ren, he lower he produc marke compeon for he company, he resuls show ha he companes wh a hgh barganng ably on Chnese capal marke have a lle dffculy n mananng he orgnal profably, lle rsks of prof uncerany and weaker movaon for manpulang profs o reduce he qualy of nformaon dsclosure n case 56

5 Table 4. Panel model esmaon resuls of produc marke compeon affecng he qualy of nformaon dsclosure. Varable (1) (2) (3) (4) (5) (6) (7) absda absda Posve_DA Negave_DA Posve_DA Negave_DA absda HHI *** *** 0.037* ** (-4.12) (-3.72) (1.78) (-2.52) HHI (0.83) PCM *** *** (-8.09) (0.34) (11.56) (8) absda *** (-5.26) PCM *** (3.03) Frs ** ** e (0.38) (1.13) (2.10) (1.24) (2.09) (-0.01) (0.37) (0.97) Board *** *** *** 0.047*** *** 0.041** *** *** (-3.72) (-3.50) (-2.63) (2.86) (-2.58) (2.54) (-3.73) (-3.44) ndrecor ** ** *** ** ** ** (-2.28) (-2.17) (-0.14) (2.71) (-0.15) (2.38) (-2.29) (-2.16) Insu * * ** ** * * (-1.86) (-1.75) (-2.10) (1.07) (-2.20) (1.05) (-1.87) (-1.81) Mshare e e e e e e e e-4 (-0.25) (0.18) (0.07) (0.91) (0.04) (0.35) (-0.25) (0.03) Salary ** * 0.019*** *** ** * (-2.53) (-1.37) (1.70) (4.60) (1.56) (2.98) (-2.54) (-1.66) logasse *** ** *** *** *** ** (-3.35) (-2.20) (-0.25) (5.47) (-0.28) (4.55) (-3.27) (-2.33) Growh 0.034*** 0.036*** 0.047*** *** 0.047*** *** *** *** (16.37) (17.37) (16.92) (-4.41) (16.70) (-6.14) (16.36) (16.89) leverage e (1.11) (0.07) (0.16) (-0.32) (0.13) (1.39) (1.10) (-0.08) MB 0.019*** 0.025*** 0.027*** *** ** *** *** (4.83) (6.26) (4.71) (-0.48) (4.65) (-2.53) (4.86) (5.71) volaly 0.001*** e-3*** e *** 0.298e *** *** *** (3.64) (2.32) (0.78) (-5.33) (0.78) (-3.73) (3.65) (2.67) SOE (0.86) (0.32) (-1.25) (-1.52) (-1.25) (-1.04) (0.86) (0.40) Dual e e e e e e-2 (0.02) (0.13) (0.07) (-1.09) (0.19) (-1.11) (0.01) (0.11) Indusry Effec conrol conrol conrol conrol conrol conrol conrol conrol Tme Effec conrol conrol conrol conrol conrol conrol conrol conrol N R of marke shock. I furher confrms Hypohess 1 a he corporae level. 4.2 Produc marke compeon and earnngs managemen drecon Based on he emprcal evdence of produc marke compeon affecng he qualy of nformaon dsclosure, he nformaon dsclosure behavor of he oal samples s subdvded no wo sub-samples accordng o he posve and negave earnngs managemen, and PMC proxy varables HHI and PCM a he ndusry level and he corporae level are respecvely subsued no he model II for regresson analyss, n whch Posve_DA and Negave_DA are respecvely he posve and negave earnngs managemen carred ou by he company o reduce s own qualy of nformaon dsclosure. Posve DA α β PMC γ comconrol _, = +, +, + δ Govconrol + ndusry + year + ε,, _, = α + β, + γ, Negave DA PMC comconrol + δ Govconrol + ndusry + year + ε,, ---II As can be seen from Table 4 (3) - (6), a he confdence level of 1%, posve earnngs managemen s only sgnfcan on he coeffcen of produc marke compeon a he ndusry level, whle negave earnngs managemen s only sgnfcan on he coeffcen of produc marke compeon a he corporae level. Afer observng he symbols of wo coeffcens, he company ends o posve earnngs managemen n he ferce compeon n he ndusry, whle he company ends o negave earnngs managemen when s own prcng ably s weakened. 4.3 Furher dscusson of non-lnear relaonshps absda = α + β PMC + β PMC + γ comconrol 2,,,, δ Govconrol + ndusry + year + ε,, ---III In prevous leraures, he mpac of produc marke compeon on he qualy of corporae nformaon dsclosure s conroversal. A very mporan reason s ha he marke envronmen where he samples are locaed s no conrolled. A hgh qualy of nformaon dsclosure may reduce he capal cos, and allow he corporae sock o oban a reasonable prce (Skafe e al., 2014; Francs e al., 2004; Dchev e al., 2013), bu hese benefs have a very lle chance o ncrease n 57

6 case of ferce compeon n he produc marke (Guo e al., 2015). To examne he relaonshp beween hem under he ransparency of he specfc marke, he nonlnear model III s used for regresson. The resuls n Table 4 (7) and (8) show ha here s no non-lnear mpac of he produc marke compeon on he qualy of nformaon dsclosure a he ndusry level; here s an nvered U-shaped relaonshp beween hem a he corporae level; when he posve axs of he horzonal axs s closer o he orgnal pon, ndcaes ha he produc marke compeon s greaer; when he vercal axs s more upward, ndcaes ha he qualy of nformaon dsclosure s hgher. The symmery axs n he quadrac curve of model esmaon s locaed a PCM = Accordng o he former descrpve sascs of Chna A-share lsed companes, he maxmum value of PCM s only 0.55, whch s locaed on he lef sde of he symmery axs, ndcang ha he qualy of nformaon dsclosure has a very lle chance of ncrease as he ferce produc marke compeon under he condon of overall buyer s marke and overcapacy n mos ndusres, hus furher confrmng Hypohess 1. In addon, hs paper uses he nformaon dsclosure rang daa of Shenzhen Sock Exchange o replace manpulave accrued prof as a proxy varable for he qualy of nformaon dsclosure, and conducs he robusness es, and he resul s conssen wh he above. 5 CONCLUSION Ths paper researches how produc marke compeon affecs he qualy of corporae nformaon dsclosure based on he daa of Chna A-share lsed companes from 2001 o The auhor classfes he ypes of produc marke compeon n deal and examnes he mpac of produc marke compeon on he qualy of nformaon dsclosure from he perspecves of ndusry level (ndusry concenraon level) and corporae level (produc prcng ably), and expands he mpac conen, and dsngushes beween posve and negave drecons of earnngs managemen acves ha can cause he changes n he qualy of nformaon dsclosure, and furher analyzes he possble non-lnear forms. The resuls show ha he exernal governance mechansm of nformaon dsclosure behavor n he produc marke compeon s nsgnfcan under he marke condons n Chna. There s no obvous movaon for he company o mprove he qualy of nformaon dsclosure n order o reduce he fnancng coss. On he conrary, a larger marke compeon pressure and predaory hreas make he earnngs managemen acves have a sronger ncenve, whch wll generally resul n he decrease n he qualy of corporae nformaon dsclosure as he compeon n he produc marke. Ths paper beleves ha here are wo major reasons for hs phenomenon: on he one hand, he valdy of Chna s capal marke s low, here s an nsgnfcan posve correlaon beween he qualy of nformaon dsclosure and sock lqudy and corporae valuaon, nvesors concern abou he conen of nformaon dsclosure s low and degree of rus s low, and nvesors are lack of he ably o use he nformaon dsclosure conen for value nvesmen; on he oher hand, wh he reform and openng up and he developmen of marke economy, Chnese marke has gradually ransoned from he relave shorage of produc supply o he general supply exceedng demand of producs, he buyer s marke form s ncreasngly obvous, and mos producs have a relavely low echnology, whch are locaed n he mddle and upper reaches of he ndusry, wh a weak barganng power. Therefore, n he fuure, Chna should mprove he effecveness of he capal marke based on he mprovemen of he nformaon dsclosure mechansm, supervson and punshmen srengh of nformaon dsclosure, and enhancemen of nvesor proecon and culvaon of value nvesmen concep, and also mprove he balance and coordnaon of s ndusral srucure, enrch he scale of produc ypes, mprove he barganng envronmen for small and medum-szed enerprses, so as o acheve a healhy connecon beween he asses and he capals of he company, whch can no only expand he sales scale and mproves he profably, bu also reduce he fnancng coss n he capal marke by shapng he fame and repuaon and feedback he producon. ACKNOWLEDGEMENT Ths paper s suppored by he Major Projec of Phlosophy and Socal Scences Research of Mnsry of Educaon of Chna (14JZD007); Surface Projec of Naonal Naural Scence Foundaon ( ); The Ffh Innovaon Projec of Expermenal Teachng Reform of Lngnan College of Sun Ya-sen Unversy ( ). REFERENCES [1] Andre Shlefer Does compeon desroy ehcal behavor? NBER Workng Paper No [2] Daa, S., Iskandar-Daa, M., Sharma, V Produc marke prcng power, ndusry concenraon and analyss' earnngs forecass. Journal of Bankng and Fnance, 37(8): [3] Davd Scharfsen Produc-marke compeon and manageral slack. The RAND Journal of Economcs, 19(1): [4] Dechow, P., R. Sloan and A. Sweeney Deecng earnngs managemen. The Accounng Revew, 70(2): [5] Holls Ashbaugh Skafe, Joachm Gassen, RyanLaFond Does sock prce synchroncy represen 58

7 frm-specfc nformaon? The nernaonal evdence. MIT Sloan Research Paper, No [6] Jennfer Francs, RoyanLaFond, Per M. Olsson, Kaherne Schpper Cos of equy and earnngs arbues. The Accounng Revew, 79(4): [7] Lla D. Dchev, John R. Graham, Campbell R. Harvey, Shva Rajgopal Earnngs Qualy: Evdence from he Feld. Journal of Accounng and Economcs, 56(2-3): [8] Mark Bagnol, Susan G. Was The effec of relave performance evaluaon on earnngs managemen: a game-heorec approach. Journal of Accounng and Publc Polcy, 19(4-5): [9] Margare A. Peeraf Inra-ndusry srucure and he response oward rvals. Manageral and Decson Economcs, 14(6): [10] Markaran, G., J. Sanalo Produc marke compeon, nformaon and earnngs managemen. Journal of Busness Fnance & Accounng, 41(5-6): [11] ParhaMohanram, Shva Rajgopal Is PIN prced rsk? Journal of Accounng and Economcs, 47(3): [12] Rav Bhushan Frm characerscs and analys followng. Journal of Accounng and Economcs, 11(2-3): [13] Rober E. Verreccha Dscreonary dsclosure. Journal of Accounng and Economcs, 5: [14] Yng Guo, Boochun Jung, and Sunny Yanhua Yang Produc Marke Compeon and Earnngs Qualy: A Non-Lnear Relaonshp. papers.ssrn.com. [15] Olver Km, Rober E. Verreccha Marke lqudy and volume around earnngs announcemens. Journal of Accounng and Economcs, 17(1-2): [16] Davd Scharfsen Produc-marke compeon and manageral slack. The RAND Journal of Economcs, 19(1):