The Impact of Insider Trading on Stock Price and Volume-Perspectives from Buyers and Sellers

Transcription

1 Roger C. Y. Chen, Chao-Shi Wang, & Chen-Yu Chen The Impac of Insider Trading on Sock Price and Volume-Perspecives from Buyers and Sellers (Received Mar 27, 2009; Firs Revision Jun 6, 2009; Second Revision Aug 3, 2009; Acceped Oc 20, 2009) Inroducion Insider rading refers o he pracice of corporae insiders abusing non-public informaion o engage in price manipulaion. Several scholars have poined ou ha insiders wih access o preferenial informaion have beer iming while making heir buy and sell decisions (Givoly and Palmon 985; Seyhun 986; Lin and Howe 990; Sivakumar and Waymire 994). Their ypically higher reurns (Finnery 976), lead o doubs abou marke fairness. Fairness is closely relaed o healhy marke growh. Invesors who perceive an unfair marke and lack informaion will less likely inves. This may resul in impeded marke growh, increasing volailiy risks and furher affecing he financing of hones companies. Insider rading also implies speculaion, causing he risk-averer o rerea from he marke, and increasing sock volailiy, hus impairing sabiliy. Meulbroek (992), Cornell and Sirri (992) and Aggarwal and Wu (2004) saed ha insider rading leads o sronger volailiy. Regulaions and acions aken are mean o proec invesors, for insance corporae governance, is aimed o reduce informaion asymmery beween invesors and corporae insiders. The key o invesigaing his subjec is he secluded naure of insider rading. Alhough he Securiies and Exchange Ac mandaes direcors, supervisors, execuives and large shareholders o declare radings on heir own socks, supervision insiues canno readily verify he occurrence of insider rading. This hinders researchers from deecing insider rading and furher invesigaing he impac on even sock and inside raders as persons oher han he four insider ideniies. Deecing insider rading from a change of insiders shareholding saus would be difficul for general invesors, as i only suggess company performance. This sudy proposes discovering he insider rading paern hrough observing he impac of insider rading on Roger C. Y. Chen is Professor of Graduae Insiue of Business Managemen, aional Kaohsiung Firs Universiy of Science and Technology. roger@nkfus.edu.w Chao-Shi Wang is Assisan Professor of he Deparmen of Business Adminisraion, aional Taichung Insiue of Technology. miogila@ni.edu.w Chen-Yu Chen is Assisan Professor of he Deparmen of Finance, Chang Jung Chrisian Universiy. roger@mail.cjcu.edu.w The auhors graefully acknowledge he financial suppor of he aional Science Council. (SC H MY3) even sock price and rading volume. Thus, his sudy manually collecs he manipulaion day from auhenic insider rading cases. Referring o he cour verdic, he beginning day of insider rading is defined as he even day. Previous sudies have failed o fully consider insider rading facors and have underesimaed even influences as even socks sar o reflec informaion conen before announcemen. The esimaion period measures he heoreical reurn of even sock prior o cerain incidens. Hence, if we assume insider rading o inerfere wih he esimaion period, i will no conform o he fundamenal heoreical reurn, leading o measuremen bias of abnormal reurn. Only hrough esablishing an objecive even day can sock volume and price be beer and reasonably observed. This sudy gahers verified insider rading cases according o cour verdics as daa for sudy. The cases were divided ino wo caegories - insider purchase and insider sales, according o informaion conen o probe price and volume reacion a hree differen sages of preceding, processing and pos-announcemen. This work aims o provide a full and reasonable illusraion of insider rading behavior. Mehodology Even Sudy Mehod and Measuremen of Abnormal Reurn This sudy adops he even sudy mehod o examine insider-rading ransacions. We analyze changes of price and volume of even sock a he hree sages o discover more concree evidence of insider rading. The purpose for using he even sudy mehod is o invesigae how cerain evens influence sock performance. This work ess abnormal reurn o show he reacion level, or difference beween acual reurn and expeced reurn. The even sudy mehod refers o expeced reurn as he expeced performance prior o evens, and selecs a cerain period o esablish a reurn expecaion model, namely he esimaion period. We uilize he marke model o measure he expeced reurn of individual even socks. To obain he average even effec, we measure he average abnormal reurn of he populaion (equaion ), and conduc a es o examine if he value is significanly differen from 0 (equaion 2). The measuremen of average cumulaive abnormal reurn (equaion 3) and is es saisics is mean o observe he Managemen Review Vol. 30 (Jan 20), 49-52

2 following changes of abnormal reurn wihin a period afer he even (equaion 4). 50 Managemen Review, January 20

3

4 AR sa AR i, () i CAR sa ( AR ) d 2 AR d ( CAR ) AR i Sˆ 2 i 2 d AR i 2 (2) (3) d i, Sˆ i (4) d d where AR is he average abnormal reurn of he even sock on day ; AR i, is he abnormal reurn of even sock i on day ; is he sample size; sa( AR ) is he es saisics of AR ; S ˆi2 is he residual variance of even sock i; CAR is he average cumulaive abnormal reurn of even sock from d o d 2 ; sa( CAR ) is he es saisics of CAR. Measuremen of Abnormal Trading Volume We employ he abnormal volume raio in equaion (5) and es saisics in equaion (6) developed by Harris and Gurel (986) as a benchmark o deermine abnormaliy wihin he rading volume. When he indicaor appears o be >0 or <0, i shows ha here could be unusual increases or decreases in he rading volume. EVi, AVR ( ) (5) V sa i ( AVR ) i AVR i Sˆ 2 AVR, i (6) where AVR is he average abnormal volume raio of even sock on day ; is he sample size; EV i, is he acual volume raio of even sock i on day, where he acual volume of even sock i on day divided by he acual volume of marke porfolio on day ; V i sands for he normal volume rae of even sock i, where average volume of even sock i wihin he esimaion period is divided by he average porfolio volume wihin he esimaion period; sa( AVR ) is he es saisic of AVR ; Esimaion Period 2 and SˆAVR, i is he abnormal volume raio variance of even sock i wihin he esimaion period. The Period of Sudy on Insider Trading Deermining he even period and he esimaion period is crucial for conducing he even sudy mehod, as he es would be less effecive if he degree of influence is no accouned for. As figure illusraes, suppose ha he even period is A. If we ignore he insider rading facor, hen he esimaed period will have o include he period afer he beginning day of insider rading, as shown in B. In his case, when he even sock begins o reflec informaion conen before he announcemen, i hinders us from deecing he acual effec of even on he announcemen day. We can only discover some price volume changes a he ending period of insider rading in Area C. The impac of he even on sock is seriously underesimaed. If we assume ha he esimaed period is inerruped by insider rading behavior, hen i does no conform o he heoreical expeced reurn. This effec will lead o measuremen bias of abnormal reurn and furher disor he empirical sudy oucomes. everheless, deecing insider rading remained arduous, and he daa was no provided by he Securiy Exchange Commission. We manually colleced relaed insider rading daa and deermined he insider-rading period from a more objecive perspecive. We believe ha direc observaion advanageously avoids including he insider-rading day ino he esimaion period. This work conduced more concree observaion on price volume variaion, o rack and explain he behavior model of insider rading. The curren sudy furher divided he beginning and ending day verified by he cour ino hree sages. These include he preceding sage (sage I: 20 rading days before insider rading), he processing sage (sage II: period from he sar o he end of insider rading, depending on buying insider or selling insider) and he pos-announcemen sage (sage III: 60 rading days afer he end of he insider rading day). This research also compares and discusses price volume reacions during he hree sages. The esimaion period runs from 50 days o 3 days preceding insider rading (20 rading days). The invesigaion has become more horough and complee as we compared he price volume reacions a hree differen sages. FIGURE Illusraion of Insider Trading Even Sudy Even Period Modified Esimaion Period B C 2 3 Announcemen 4 Day Insider Trading Period Day A τ τ 2 Begin Day End Day Day 50 Managemen Review, January 20 Sage I Sage II Sage III

5 Daa The curren sudy obained empirical daa from he cour verdics of insider rading. Based on he principle of presumpion of innocence, one is innocen unil proven guily. However, insider-rading cases conviced afer he hird rial are rarely found in Taiwan, herefore, we employed he firs rial verdics as samples for sudy. The cour verdics are more auhenic compared o prosecuor indicmens. This sudy individually exraced facual daa from cour verdics and sored ou relevan deails from he daabases of Lawbank and Securiies and he Fuures Invesors Proecion Cener. The deails needed for even sudy such as sock price, rading volume and marke reurn raio were colleced from he Taiwan Economy Journal daabase. From January 993 o December 2008, he disric cour conviced sixy-eigh cases of insider rading lawsuis. Afer eliminaing hose wih insufficien deails, here are fory-nine cases lef. This invesigaion sudied cour verdics, dividing hem ino wo caegories, insider purchase and insider sales. Insider purchase oaled weny-wo cases and insider sales oaled weny-seven cases. The informaion conen of boh samples was also divided ino good news and bad news. Empirical Resuls Sage I: Preceding Period This sudy found ha paricularly a he laer period of he sage, which approaches he beginning day of insider rading, boh insider purchase and sales have significanly posiive reurns. Among insider purchase and sales, insider sales seem o rade more. Emphasizing his sage provides a reference for he reacion of price a cerain sages of insider rading behavior. This can help o find ou if he price and volume of even sock displays a more significan reacion when i moves from he preceding sage ino he processing period. Then, we can clearly observe he variaion of influence ha insider rading can cause on he sock. Sage II: Processing Period We firs deermine he lengh of observaion period according o he ime lengh recorded in he cour verdics. The median of insider purchase and sales are hiry-five days and weny days respecively. Insider sales indicae a shorer manipulaion period. From he price volume reacion a his sage, we can find ha boh insider purchase and sales group reached a saisical significance in heir abnormal volume raio. The increase is higher han he preceding sage. The performance of he wo groups appears o be significanly differen. Where he insider purchase group shows a significanly posiive abnormal reurn, he insider sales group presens a significanly negaive one, wih boh siuaions appearing a he early period of his sage. Thus, we know ha insider purchase even sock is conrolled by he bull marke during he insider rading period, while he bear marke appears sronger in insider sales. When insiders wih privae informaion engage in he marke, rading volume appears o have abnormal changes differen from he esimaion period. The even sock wih inerference of insider purchase and sales appears o have posiive and negaive reacions on performance. Sage III: Announcemen Period When insider purchase and sales ended, cumulaive abnormal reurn remained significanly negaive, while abnormal volume raio appeared significanly posiive. Theoreically, he marke should embrace a posiive aiude owards announcemen of good news. However, when invesors find ha he price volume of sock reflecs informaion conen before he announcemen of good news (sage II), hey suspec insider-rading involvemen. If he price reflecs good news in advance, pursuing he sock will lead o greaer risk for invesors and affec he willingness o purchase. If inside raders have accumulaed posiions prior o announcemen, hey will sell ou heir shares afer he announcemen. When insider purchase is over, even sock appears o have negaive cumulaive abnormal reurn. Theoreically, he marke should have a negaive response o bad news. ormal invesors are in a less favorable posiion if hey canno receive bad news in ime, and can only sell ou afer he announcemen. Even if invesors discover he negaive reacion before he announcemen of bad news, hey migh sill be unable o decide when he negaive effec will end, so hey may sill sell he sock afer he announcemen. Alhough inside raders wih privae informaion can sell shares before he announcemen of bad news, hey sill canno be cerain abou he end of bad news effec. Thus, hey canno make a reverse akeover rade of shares a he early announcemen period. The empirical resuls provide evidence for he above assumpion. When insider sales are over, he even sock coninues o perform poorly, especially in he early period, and always appears o have abnormal increase. Corporae insiders selling heir socks during he processing period suffer relaively minor losses compared o oher invesors. Discussion and Conclusion Measuring and racking he insider rading behavior model is a vial subjec. The securiies marke has been he plaform o bring boh fund supplier and demander ogeher. However, insider rading behavior explois he righs of mos invesors, occupies he marke wih speculaive pracices and affecs marke operaion. This sudy collecs sample insider rading companies from cour verdics and employs he even sudy mehod o invesigae he impac of insider rading on price and volume. To increase he es effec and preserve he validiy of empirical hypohesis, his work employs he direc observaion mehod o avoid measuremen bias of abnormal reurn. Alhough he daa collecion is similar o Meulbroek (992) and Cornell and Sirri (992), we disinguished our sudy by dividing he insider rad- The Impac of Insider Trading 5

6 ing period ino hree sages, he preceding, processing and pos-announcemen sage. This procedure conducs a more complee es and comparaive analysis for even sock a hree sages. Our sudy no only measures price endency bu also focuses on unusual changes in rading volume. We also disinguish he good news and bad news refleced by insider rading o discuss he variaion in reacions of price and volume under differen informaion conen. The empirical resuls describe he behavior model of insider rading, showing ha he price and volume appeared significanly differen a he hree sages. Paricularly, price refleced conen during he processing period, and rading volume showed a significanly unusual phenomena. Invesors should be concerned abou he pos-announcemen sage, as he resul repored ha boh buyer and seller insider rading are conrolled by he bear marke when he insider rading period is over. The resuls of insider rading impac on sock price and volume prove ha insider rading has seriously damaged he fair marke and exploied he righs of less-informed invesors. Securiies supervision auhoriies mus execue heir duy o rack insider-rading behavior. We advise invesors o carefully review price and volume reacions before he announcemen day. If he informaion reflecs conen in advance, we advise invesors no o involve in ransacions so as o avoid risks. REFERECES Aggarwal, Rajesh K. and Guojun Wu (2004), Sock Marke Manipulaion - Theory and Evidence, working paper, Universiy of Michigan. Cornell, Bradford and Erik R. Sirri (992), The Reacion of Invesors and Sock Price o Insider Trading, Journal of Finance, 47(3), Finnery, Joseph E. (976), Insiders and Marke Efficiency, Journal of Finance, 3(4), Givoly, Dan and Dan Palmon (985), Insider Trading and he Exploiaion of Insider Informaion: Some Empirical Evidence, Journal of Business, 58(), Harris, Lawrence E. and Eian Gurel (986), Price and Volume Effecs Associaed wih Changes in he S&P 500 Lis: ew Evidence for he Exisence of Price Pressures, Journal of Finance, 4(4), Lin, Ji-Chai and John S. Howe (990), Insider Trading in he OTC Marke, Journal of Finance, 45(4), Meulbroek, Lisa K. (992), An Empirical Analysis of Illegal Insider Trading, Journal of Finance, 47(5), Seyhun, Hasan. (986), Insiders Profis, Coss of Trading, and Marke Efficiency, Journal of Financial Economics, 6(2), Sivakumar, Kumar and Gregory Waymire (994), Insider Trading Following Maerial ews Evens: Evidence From Earnings, Financial Managemen, 23(), Managemen Review, January 20