Market Valuation of Employee and Director Bonuses for Profitable High-Tech Firms: The Case of Taiwan

Transcription

1 Marke Valuaion of Employee and Direcor Bonuses for Profiable High-Tech Firms: The Case of Taiwan David M. Chen, Professor, Graduae Insiue of Finance, Fu Jen Caholic Universiy, Taiwan Li-Ling Yang, Insrucor, Deparmen of Finance, China Insiue of Technology; Ph.D. suden, College of Managemen, Fu Jen Caholic Universiy, Taiwan ABSTRACT This sudy examines wheher invesors incorporae he effec of employee and direcor bonuses ino firm marke valuaion for a sample of 9 profiable high-ech firms during in Taiwan. We use he Ohlson (995) equiy valuaion model, and compare difference diluion raio model s explanaion power. We find a negaive correlaion beween he value of CSBs and Tobin s bu no significan, i may be he rade-off beween diluive effecs and incenive effors. The resuls sugges ha invesors appear o value hese firms CSBs as an expense bu as an inangible asse on he firm s equiy value. Keywords: employee bonuses; equiy valuaion; equiy diluion; earning diluion. INTRODUCTION Taiwan, as an emerging marke, has a srong sanding in he global elecronics indusry. Employee cash and sock bonuses (CSBs) have been a unique incenive sysem o high-ech indusry in Taiwan. One reason for he prevalence of employee bonuses is heir favorable accouning reamen. Under Taiwan Generally Acceped Accouning Principles (Taiwan s GAAP), employee bonuses and direcor compensaion are regarded as disribuion of earnings, meaning ha employees do no have o pay any cos, and companies do no have o recognize expenses for he bonus shares issued o employees, in oher words, i does no a all affec he boom line. In January 007, he Minisry of Economic Affairs issued a Leer required companies o recognize CSBs as expenses changing his value o earnings. The high-ech indusries me wih srong opposiion he recogniion of expenses for bonus shares o employees because i grealy reduces annual earnings and causes a de-raing of firm valuaions in he capial markes. This sudy examines wheher invesors incorporae he effec of CSBs ino firm marke valuaion. Employee Sock Opions (ESOs), an inernaional sysem similar o bonus shares for employees, is recognized as expenses a fair value during he enures of employees. Therefore, when Taiwanese companies would like o be lised in he US, hey mus follow US GAAP requiremens regarding ESOs by expensing a number of shares, as expeced issued o employees. There have been srange occasions ha companies repor earnings based on Taiwan s GAAP, bu are sill able o disribue dividends and bonuses even hough hey are in he defic as based on US GAAP. Riding he idal wave of liberalizaion and globalizaion, local accouning and valuaion professions were jus abou o ge ready o embrace global convergence a he dawn of he new millennium. In 008, Taiwan s GAAP began o follow Inernaional Financial Reporing Sandards (IFRS) by recognizing bonuses o employees as expenses, raher han earnings disribuions, he pracice of he pas. This is a major change in he accouning sysem in Taiwan. When he world is moving oward IFRS, i is imporan o examine he impac of differen accouning reamens on equiy valuaion. Compensaing employees wih bonus shares raher han cash can be aracive o firm, since bonus shares provide long-erm incenives and migh reduce agency problem. Therefore, CSBs value affecs share prices in wo opposing ways: i dilues he value of he firm s ousanding shares (he diluion effec ), and ye provides incenives o For Taiwanese firms, employee compensaion generally conains hree componens: a basic salary, a year-end bonus, and employees bonus (cash/or sock bonus), direcor compensaion only cash. An average 76% firms gran employee large amouns of CSBs. The Journal of Inernaional Managemen Sudies, Volume 4, Number, February,

2 employees o increase he firm s sock price (he incenive effec ). Aboody (996) shows ha here is a negaive correlaion beween he value of ESOs schemes and firm share prices, his suggess ha he diluive effecs of ESOs schemes are greaer han he incenive effecs. However, Bell e al. (00) indicaes ha he markes do no devaluae he companies who recognize more expenses associaed wih ESOs schemes. As soon as he employee bonus shares in Taiwan are recognized as expenses, here is currenly a dispue on wheher or no he value of CSBs really affecs equiy valuaion. Accordingly, we expec a negaive (posiive) correlaion beween CSBs and firms equiy marke value if he diluion is larger (smaller) han he expeced impac on fuure performance. To examine his issue, wheher invesors incorporae he effec of a firm s ousanding CSBs ino equiy marke valuaion. This sudy is conduced by esimaing he CSBs for each firm. Subsequenly, we es he associaion beween CSBs and firm s equiy marke value hrough a cross-secional regression by applying Ohlson (995) equiy valuaion model. The sample consiss of 9 high-ech firms lised on he Taiwan Sock Exchange during Firm in Taiwan are required by law o pay an annual bonus (cash and/or sock) equal o a percenage of ne income. The empirical resuls show a negaive correlaion beween he value of CSBs and firm s equiy marke value. The remainder of his paper is organized as follows: Secion is lieraure review; Secion 3 explains he research design ogeher wih sample selecion crieria and descripive; Secion 4 presens empirical resuls; Secion 5 concludes he paper. BACKGROUND AND LITERATURE REVIEW Many companies in he U.S. have Employee Sock Opion Plans (ESOPs). According o APB No.5, companies were required o calculae he value of ESOs based on he inrinsic value mehod (by subracing marke value wih subscripion value). Under such rules, here was no inrinsic value because he subscripion price equaled o he marke price of he issue dae. Companies were no required o recognize any compensaion expenses, as ESOs were reaed as valueless financial producs, which resuled in a lack of reliabiliy of financial reporing, and allowed an opporuniy for senior execuives o gain personal profis. A he same ime, shareholders were no able o undersand wha is going on. Due o heavy criicism, he FASB issued he Exposure Draf, as based on he fair value mehod, a he beginning of 993. I sipulaed ha a debi should be made as prepaid compensaion expenses, based on he fair value on he exercise dae of opions, and a corresponding credi should be made as capial reserve ESOs. The amorizaion of he prepaid expenses from he issue dae o he exercise dae was recognized as relaed compensaion expenses. The fair value mehod creaed huge compensaion expenses, which in urn, had effec on he boom line. Meanwhile, he diluive effecs of increases in he number of ousanding shares led o a reducion of EPS, and all financial raios relaed o asses and equiy. For example, boh ROA and ROE dropped as a resul, and hus, his draf was opposed as never before. Therefore, he FASB issued SFAS No.3 on Ocober 995, which encouraged companies o use he fair value mehod insead of requiring hem o do so. I also allowed companies o coninue o use he inrinsic value mehod o evaluae compensaion expenses by following APB No.5. However, if companies choose o coninue o observe APB No.5, hey were required o disclose boh he pro-forma ne income and EPS under he fair value mehod. SFAS No.3 is adoped by mos companies in he U.S., and hus, is he fundamenal principle o rea ESOs as operaing expenses, which is he accouning procedure ha Taiwanese companies mus follow in he fuure, as he cos of bonus employee shares. Such a cos shall be recognized as expenses. Inernaional sudies on ESOs compare differen evaluaion mehods of compensaion expenses, and he marke views on such expenses. Aboody (996) used modified opion pricing model, 478 companies sample, whose number of ousanding ESOs opions exceeded 5% of he oal number of ousanding ordinary shares from 980 ~990, and found ha here was negaive correlaion beween he value of ousanding ESOs and firm share prices; Huson and Sco e al. (00) research by observing 63,656 sampled companies in Compusa from 970 o 995, he resul suppored ESOs had diluive effecs on ERC (earnings reacive coefficien); Aboody e al. (ABK 004) sudy found ha invesors regarded he ESOs expenses under SFAS No.3 as firm expenses, and negaive marke evaluaions by samples consising of S&P 500, S&P 400, and S&P 600 from ; Bell e al. (BLMY 00) used he valuaion models of Ohlson (995, 999) and FO (999) o examine 85 compuer sofware companies wih profi poenials. They compared 90 The Journal of Inernaional Managemen Sudies, Volume 4, Number, February, 009

3 he accouning reamens under APB No.5, SFAS No.3, and he Exposure Draf of share-based compensaion schemes in order o undersand how invesors evaluae companies under he hree differen recogniion mehods. Their empirical sudy found he opposie from Aboody; markes do no value companies less jus because companies recognize more ESOs expenses. The resuls are idenical under all accouning reamens. Meanwhile, if companies adop he Exposure Draf mehod, hey can enhance marke evaluaion, in oher words, invesors believe ha ESOs expenses have posiive incenive effecs because hey creae inangible asses for companies. The differen conclusion from Aboody (996, 00) is probably due o he selecion of samples 3. Aboody (004) examined knowledge-based and profiable companies wih insrumenal variables and sill found a negaive correlaion 4. RESEARCH DESIGN AND SAMPLE. The Relaion Beween Equiy Marke Value and CSBs Ohlson(995)shows ha marke value of equiy equals book value of equiy plus he presen value of expeced fuure abnormal earnings, expressed as follows: P = BVE + ( ROE RE Le P :marke value of equiy during period BVE :book value of equiy during period :assessed risk coefficien during period ROE :reurn of equiy during period R E :cos of equiy during period ) Under he concep of fully loaded earnings, all iems ha affec changes of shareholders equiy. The model can be simplified as follows: P = BVE + NI Boh ends of he equaion are divided wih he book value of equiy 5, hen he funcion of marke value can be derived. P NI = +, BVE BVE = 0 + ROE To invesigae wheher invesors incorporae he effec of a firm s ousanding CSBs ino equiy marke valuaion. Equaion (), Tobin s, defined as marke value of equiy divided by book value of equiy, is used as he dependen variable o assess he model s explanaory power. 0 + ROE + CSB + ε = () i, : equiy marke value of firmi during period ROE, : reurn of equiy of firm i during period i CSB : cash and sock bonus shares o employees of firmi during period i 3 Aboody samples conained a wider range of companies, profiable or no, and his approach is differen from sampling knowledge-based and profiable companies. I is possible ha in a knowledge-based economy, ESOs is reaed as an inangible asse, bu no in oher indusries. See Bell e al.,(00),p Aboody (004) commens ha he sample of Bell (00) is neiher exensive nor represenaive. See Aboody e al., (004),p Book value will be a beer approximae of replacemen cos if more iems of asses and liabiliies are o be measured in fair value insead of hisorical cos. The Journal of Inernaional Managemen Sudies, Volume 4, Number, February, 009 9

4 I is expeced ha is significanly posiive, and if invesors regard bonus shares o employees as firm expenses, hen coefficien, should be negaive.. CSBs Effec and Diluion Raio Measuremen Firs, we conduced by esimaing he CSBs for each firm. Subsequenly, we examine he associaion beween CSBs and equiy marke value hrough a cross-secional price-level regression. The pracice of he pas, diluion raio was adoped o explain he CSBs effec o equiy marke value. For example, shareholders equiy diluion raio (see Model A) and profi diluion raio (see Model B) based on face value ( De ; Dp ) or on marke value ( De ; Dp ). The shorcoming in he former, ha based- fair value approach, he higher he marke price he more underesimae diluive effecs on shareholders equiy; hen in he laer, ha based- marke value approach, marke value simulaneously are used o calculae dependen variable and explanaory variables, i will creae quaniaive errors. In order o miigae he spurious relaions ha can resul from using conemporaneous daa. This paper forward sugges adoping number of bonus shares raio (see Model C) wihou consideraion shares prices. To invesigae his issue, we will separaely compare and assess hese models s explanaory power. Model A : Shareholders Equiy Diluion Raio (De); MA-: based on face value De ; MA-: based on marke value De BN * F + C De = E where BN: number of bonus shares F: face value C: bonus cash E: shareholders equiy M: marke value (ex-righ prices) ; BN * M + C De = E Model B: Profi Diluion Raio (Dp); MB-:based on face value Dp ; MB-: based on marke value Dp BN * F + C Dp = P where P: afer-ax profi ; BN * M + C Dp = P Model C: number of bonus shares Raio (Dn) BN C Dn = + TN E where TN: Toal Number of ousanding shares 3. Samples & Descripive Saisics Employee cash and sock bonuses, direcor compensaion and ex-righ adjused share prices are rerieved from Taiwan Economic Journal (TEJ) daa base for he year 00 o 007. Sample firms consis of all Taiwan Sock Exchange lised elecronics companies. These companies all repored pre-ax earnings, and have a sysem in place o issue bonus shares o employees, resuling in a sample of 9 firms or 0 observaions over he 7-year sample period. 9 The Journal of Inernaional Managemen Sudies, Volume 4, Number, February, 009

5 April 30 share prices are used o measure marke value because i is he official deadline for filing annual repor wih he SFB (Securiies and Fuures Bureau). Sock prices are adjused for sock and cash dividend disribuions. Table provides descripive saisics of all variables. The mean (medium) of marke values (), and ROEs of he sampled companies, are.5, 4.66%, (.5, 3.83%), respecively. The saisical calculaes shareholders equiy diluion raio based on face value derives an average of.3% (.6), as based on marke value repors an average of 3.5% (.3%),.67 imes which of he face value approach. The profis diluion raio based on face value, repors an average of 9.5% (9.5%). If calculaed wih marke value, he average was.45% (8.43%),.36 imes which of he face value approach. The number of bonus shares raio was.85% (.60%). The diluive effec was far greaer when he calculaion was based on marke value, raher han face value, especially in 006 when he sock marke peaked. Table provides he correlaion marix of dependen variables and all independen variables. These have a sronger associaion wih he raio, wih he excepion of De. EMPIRICAL TEST AND RESULTS To examines wheher invesors incorporae he effec of employee and direcor bonuses ino firm marke valuaion, we conduc cross-secional regression for each year in he sample period. Resuls are shown in Table3, Table 4 and Table 5. Table 3 he adjused R in Panel A and Panel B shows ha shareholders equiy diluion raio based on face value and /marke value has significan explanaory power for equiy marke value for all year. The coefficien of De is negaive in four of he seven years for equiy marke value bu is nonsignifican for all years. However, he coefficien of De is significan posiive correlaion for equiy marke value for all years, bu we discredi he resuls wih spurious relaions. The adjused R in Panel A and Panel B of Table 4 shows ha profis diluion raio based on face value and /marke value has significan explanaory power for equiy marke value for all year. The coefficien of Dp is negaive in hree of he seven years for equiy marke value bu is nonsignifican for all years bu whole sample is significan. The coefficien of Dp is significan posiive correlaion for equiy marke value, bu we discredi he resuls as he same wih spurious relaions. Comparaive resuls of number of bonus shares raio Dn are presened in Table 5. The adjused R shows ha bonus shares raio has significan explanaory power for equiy marke value for all year. The coefficien of Dn is negaive in five of he seven years for equiy marke value bu is nonsignifican for all years. CONCLUSIONS In 008, Taiwan s GAAP began o follow IFRS by recognizing bonuses employees as expenses, raher han earnings disribuions. The purpose of our sudy is o examine he associaion beween CSBs and equiy marke value and also conducs ess on differen diluion raios marke invesor s views on bonuses o employees. We analyze a sample of 9 profiable high-ech firms graned CSBs o employee during he period in Taiwan. We propose five differen Panel regressions, if adop equiy/profis diluion raio o calculae CSBs based on marke value approach, he resul shows ha invesors regards CSBs as inangible asse, a posiive on firm value, bu we suspec he resul wih spurious relaions. We propose a modified bonus shares raio approach wihou considerae sock price, he resul as he same based on face value approach, a negaive on firm value, invesors regards CSBs as an expense, esimaing he cos of CSBs ino equiy marke valuaion. Alhough he resul is nonsignifican, i may be he rade-off beween diluive effecs and incenive effors. The paper s sample selecion concenraes on elecron indusry. I is possible ha his paper s resuls canno be applied on non-elecron indusry. Fuure research will focus on he deerminans of bonus shares o employees for long-erm performance, shor-erm ax benefis or oher. The Journal of Inernaional Managemen Sudies, Volume 4, Number, February,

6 Table. Descripive Saisics of Samples in Differen Models of Diluion Raio by Year a Diluion Model whole sample b De De Dp Dp Dn ROE sample 9 c a Numbers in each cell are mean, median and sandard deviaion. Numbers everyhing is shown as percenages, wih he excepion of. b Variable definiions are as follows: De : Shareholders Equiy Diluion Raio Based on Face Value; De :Shareholders Equiy Diluion Raio Based on Marke Value (ex-righ prices); Dp :Profis Diluion Raio Based on Face Value; Dp : Profis Diluion Raio Based on Marke Value (ex-righ prices); Dn : Number of bonus shares Raio; =Tobin s raio, mean equiy marke value; ROE: Reurn of Equiy. c Number of firms wih complee daa available Table. Pearson Correlaion Marix of Dependen and Differen Diluion Raio Variables Variables ROE De De Dp Dp Dn 0.5** 0.4** 0.50** ** 0.40** ROE ** 0.70** -0.6** 0.7** 0.74** ¹ Two-ailed es:**significance a % level;*significance a 5% level. Table 3. Summary Saisics from Regression of Tobin s Raio and Shareholders Equiy Diluion Raio Based on Face Value and /Marke Value by Year a Panel A: = ROE De ε 00.07***(6.48).80(.65) 4.8(0.9) *** ***(9.66).85***(4.74) -8.96(-.4) *** ***(7.44) 3.86***(4.4) -0.39(-0.05) *** ***(5.60) 5.49***(8.44) -9.0(-.58) *** ***(6.) 7.59***(6.83) -0.77(-.) *** ***(7.36) 5.5***(5.6) 5.80(.57) *** ***(3.70) 7.69***(7.08) 6.(0.63) *** Whole Sample 0.66***(3.38) 5.9***(.80) 5.(.5) *** Panel B: ROE De = ε 00.4***(6.95).45(.0) 4.06*(.77) *** ***(0.0) 0.83(.7) 3.4**(.63) *** ***(8.).9**(.) 6.8***(.86) *** 94 The Journal of Inernaional Managemen Sudies, Volume 4, Number, February, 009

7 ***(6.7) 3.0***(4.73) 4.94***(3.49) *** ***(8.66).5(.6) 6.58***(7.0) *** ***(8.33) 4.34***(5.3) 7.89***(4.50) *** ***(5.67) 3.5***(3.48) 7.76***(6.3) *** Whole Sample 0.75***(5.9) 3.65***(9.) 6.4***(7.36) *** a Number in parenheses are -saisics(f-saisics for adjused R²). Two-ailed es: ***significance a % level, ** significance a 5% level and *significance a 0% level. Table 4. Summary Saisics from Regression of Tobin s Raio and Profis Diluion Raio Based on Face Value and /Marke Value by Year a Panel A: = ROE Dp ε 00.05***(3.87) 3.9***(.89) 0.4(0.) *** ***(5.7).33***(5.75) 0.8(.6) *** ***(6.08) 3.83***(5.9) -0.0(-0.03) *** ***(3.99) 4.77***(.) 0.44(0.9) *** ***(3.96) 6.6***(9.40) -0.0(0.0) *** ***(3.5) 7.03***(0.0).30(.6) *** (.63) 8.3***(.06).7(.05) *** Whole Sample 0.57***(8.6) 5.8***(9.5) 0.9**(.06) *** Panel B : = ROE Dp ε 00.03***(5.7) 3.06***(.79) 0.9(0.7) *** ***(7.56).89***(5.0) 0.55***(4.07) *** ***(6.0) 3.4***(4.77).00***(.98) *** ***(4.6) 3.95***(8.64).09***(3.68) *** ***(4.53) 5.75***(8.45).47***(4.67) *** ***(5.) 5.94***(8.90).64***(4.87) *** **(.65) 7.4***(9.40).3***(3.47) *** Whole Sample 0.53***(0.4) 5.3***(7.8).0***(7.) *** a Number in parenheses are -saisics(f-saisics for adjused R²). Two-ailed es: ***significance a % level, ** significance a 5% level and *significance a 0% level. Table 5. Summary Saisics from Regression of Tobin s Raio and Number of Bonus Shares Raio by Year a Panel A: = ROE Dn ε 00.07***(6.46).88*(.93).68(0.8) *** ***(9.57).65***(4.45) -4.(-0.99) *** ***(7.44) 3.94***(4.3) -0.97(-0.6) *** ***(5.58) 4.99***(7.64) -.4(-0.57) *** ***(6.03) 6.6***(6.4) -0.07(-0.0) *** ***(7.40) 5.95***(5.68) 7.59(.0) *** ***(3.77) 8.6***(7.7) -0.68(-0.0) *** Whole Sample 0.66***(3.39) 5.40***(.34).8(0.98) *** a Number in parenheses are -saisics(f-saisics for adjused R²). Two-ailed es: ***significance a % level, ** significance a 5% level and *significance a 0% level. REFERENCE Aboody David (996), Marke valuaion of employee sock opions, Journal of Accouning and Economics,, pp Aboody David, Barh Mary E. and Kasznik Ron (004), SFAS No.3 Sock-Based Compensaion Expense and Equiy Marke Values, The Accouning Review, Vol.79, No., pp The Journal of Inernaional Managemen Sudies, Volume 4, Number, February,

8 Bell Timohy B.e al. (BLMY 00), The Valuaion Implicaions of Employee Sock Opion Accouning for Profiable Compuer Sofware Firms, The Accouning Review,Vol.77, No.4, pp Dhiman,Chowdhury and Zahirul,Hoque (998), Profi Sharing and Corporae Performance: Some Evidence from Bangladesh, The Inernaional Journal of Accouning,Vol.33, No.4, pp Frye, Melissa B. (004), Equiy-Based Compensaion for Employees: Firm Performance and Deerminans, The Journal of Financial Research, Vol. XXVII, No., pp Huson and Sco e al.(00), Earnings Diluion and he Explanaory Power of Earnings for Reurns, The Accouning Review, Vol.76,No.4, pp The Journal of Inernaional Managemen Sudies, Volume 4, Number, February, 009