CALCULATING THE EFFECT OF EMPLOYEE STOCK OPTIONS ON DILUTED EPS

Transcription

1 MPRA Munich Personal RePEc Archive CALCULATING THE EFFECT OF EMPLOYEE STOCK OPTIONS ON DILUTED EPS Warrick Boyd van Zyl Universiy of he Wiwaersrand 17. December 2007 Online a hp://mpra.ub.uni-muenchen.de/6322/ MPRA Paper No. 6322, posed 17. December :00 UTC

2 CALCULATING THE EFFECT OF EMPLOYEE STOCK OPTIONS ON DILUTED EPS Warrick van Zyl Lecurer School of Accounancy Universiy of he Wiwaersrand Johannesburg 2000 Souh Africa

3 Absrac CALCULATING THE EFFECT OF EMPLOYEE STOCK OPTIONS ON DILUTED EPS This paper focuses on how o calculae dilued earnings per share (DEPS) when a firm has ousanding employee sock opions (ESOs). Three possible mehods are described and compared. The firs is he curren Inernaional Accouning Sandard 33 Earnings Per Share (IAS 33) approach which is based on he inrinsic value of he ESOs. The second mehod, advocaed by Core e al. (2002), is very similar o ha of IAS 33 bu insead of he inrinsic value uses he fair value of he ousanding opions. This paper derives an alernaive mehod which adjuss he earnings for he year by he change in fair value of he ousanding ESOs, wih no adjusmen o he denominaor in he DEPS calculaion. The hree mehods are compared using a simple firm. The earnings adjusmen mehod bes describes he change in economic value of he curren shareholders, he fair value is more useful in predicing fuure profis, and he inrinsic value mehod appear o provide no addiional informaion o ha already conained in he oher wo measures. The earnings adjusmen mehod has a furher advanage in ha i provides an idenical resul a a DEPS level o ha which would have been obained if he ESOs were cash-seled and reaed as liabiliies in erms of IFRS 2. Thus using his mehod will improve comparabiliy as cash-seled and equiy-seled opions have a very similar economic effec on curren shareholders.

4 CALCULATING THE EFFECT OF EMPLOYEE STOCK OPTIONS ON DILUTED EPS Inroducion Employee sock opion accouning has undergone significan changes in recen years. For he firs ime, mos firms around he globe are required o accoun for he expense associaed wih employee sock opions (ESOs). Whils his has been a sep in he righ direcion he sandard seers seem o have overlooked a relaed bu equally imporan issue; how should dilued earnings per share (DEPS) be calculaed when he firm has ousanding share opions. Much has been wrien on he appropriae accouning reamen of ESOs in he income saemen and balance shee. This paper will focus on he effecs of ousanding ESOs on DEPS. Generally Acceped Accouning Pracice for Employee Sock Opions One of he more significan recen developmens in generally acceped accouning pracice (GAAP) has been he inroducion of comprehensive rules for he accouning of share-based paymens, and more specifically, employee sock opions. Inernaional Financial Reporing Sandard 2 Share-based Paymens (IFRS 2) (IASB 2003b) and is U.S. equivalen Saemen of Financial Reporing Sandard No. 123 (revised) (SFAS No. 123) (FASB 2004) have cerainly improved accouning convenion in his area. The original SFAS No. 123 was issued in 1995, bu is inroducion was accompanied by such conroversy ha applicaion was made volunary. Unusually for an accouning developmen, he accouning of employee sock opions even araced he aenion of he legislaors. Pressure o inroduce an accouning saemen on ESOs originaed from a US senaor in he early 1990 s, and he Financial Accouning Sandards Board (FASB) was forced o make he applicaion of he original SFAS No. 123 volunary because i was hreaened by legislaion from anoher US senaor 1. The large corporae governance scandals in he early 2000 s again changed he mood of he business communiy, and he inroducion of IFRS 2 and SFAS No. 123 (revised) in 2003 and 2004 respecively was generally welcomed. IFRS 2 inroduced accouning rules in an area ha was largely unregulaed before. Whils IFRS 2 covers all ypes of share-based ransacions, i is ESO s ha arac mos of he aenion and will be he mos affeced by he new sandard. ESOs were largely ignored in he accouning records of mos companies across he globe, especially if hose ESOs were issued wih a srike or exercise price ha was below or equal o he curren share price. IFRS 2 now requires ha all share-based paymens be accouned for iniially a fair value. The sandard hen disinguishes beween cash-seled and equiy-seled ransacions. Those ransacions ha will be seled in cash are considered liabiliies wihin he broader accouning concepual framework, and ransacions ha are seled by he company issuing shares are considered equiy. The disincion is imporan because equiy iems are never remeasured, whils liabiliies ofen are. Thus he sandard requires ha equiy-seled ransacions be accouned for a fair value on he iniial gran dae, and never hereafer adjused. The obligaion incurred in cash-seled ransacions is refleced as a liabiliy in he balance shee, and remeasured a each reporing dae o reflec changes in he value of he underlying shares. This requiremen is somewha unusual, as i resuls in differen accouning reamens for ransacions ha are economically idenical; wheher he ransacion is seled in cash or shares would generally be considered equivalen under general economic or finance heory. When applied o employee sock opions, he following reamen resuls. The iniial gran dae fair value of opions ha are equiy-seled is recognised as an expense in he income saemen over he vesing period of he opions. The iniial gran dae fair value of cash-seled opions is also recognised over he vesing period as an expense, bu he cumulaive expense is also adjused o reflec changes in he value of he expeced selemen amoun due o change in he underlying share price. This means ha he oal expense recognised for equiy-seled opions will be he iniial gran dae value, whereas for cash-seled opions i will be he final amoun paid in selemen of he opion obligaion. For he remainder of his paper he erm ESO will refer o equiy-seled employee sock opions and share appreciaion righs (SARs) will refer o cash-seled employee sock opions. 1 For a deailed accoun of he process leading up o he inroducion of he original SFAS No. 123 see Dechow e al. (1996). 1

5 One of he issues ha has been highlighed by he objecions o IFRS 2 is he relaionship beween ESOs and DEPS. The rules for calculaing DEPS are conained in Inernaional Accouning Sandard 33 Earnings Per Share (IAS 33) (IASB 2003a). IAS 33 is again similar o is U.S. counerpar Saemen of Financial Accouning Sandards No. 128 Earnings Per Share (SFAS No. 128) (FASB 1997). IAS 33 requires use of he reasury sock mehod. The reasury sock mehod calculaes he diluive effecs of he ousanding opions on he curren ordinary shareholders by increasing he number of shares used in he earnings per share (EPS) calculaion. The addiional number of shares is deermined by assuming immediae exercise of he opions and ha he firm uses he exercise proceeds o repurchase is own sock. The difference beween he shares issued and he shares repurchased is hen added o he denominaor in he EPS calculaion. In he case of ESOs, he value of he services sill o be received from employees mus be added o he exercise price before he reasury sock calculaion is performed. Many criics mainained ha recognising an ESO expense and adjusing DEPS for he same ESOs would be double accouning (Bodie e al. 2003; Michaels and Waers 2004). There is acually a double effec on shareholders; his is well know and is simply explained by Bodie e al. (2003) and in he examples ha follow laer in his paper. I does, however, emphasise ha he wo issues need o be considered ogeher. Minor changes were made o IAS 33 when IFRS 2 was issued, bu in essence he reasury sock mehod was reained. In comparison o IFRS 2, he provisions of IAS 33 seem archaic and in need of revision. Why should we assume immediae exercise of he opions, when finance heory indicaes ha i is rarely opimal o exercise opions early? The reasury sock approach revolves around he difference beween exercise price and he curren share price, which is normally referred o as he inrinsic value. Why does he sandard use inrinsic value of he ESOs raher han he fair value? Why is he value of services sill o be received added o he exercise price when hey describe compleely differen economic conceps? Why is he denominaor in he earnings per share calculaion adjused raher han numeraor, or boh? By reference o finance and accouning heory, his paper will aemp o answer hese quesions, and criicise he curren GAAP approach o accouning for he effecs of ESOs on DEPS. The specific quesion raised in his paper is how bes o calculae dilued earnings per share for a firm which has ousanding ESOs. Prior research on DEPS The curren GAAP for measuring DEPS has been carried over largely unalered from Accouning Principles Board (APB) Opinion 15 (AICPA 1969) which was issued in The inroducion of his sandard predaed he developmen of opion pricing models in he 1970 s which quickly made he provisions of APB Opinion 15 seem archaic and redundan. These problems were soon idenified by he researchers. Vigeland (1982) and Bierman (1986) poin ou ha APB Opinion 15 ignores he probabiliy of he opion being exercised. Jerris (1992) uses he probabiliy of he insrumen being convered or exercised o calculae his EPS numbers, which he finds have a closer associaion wih sock reurn residuals han he APB Opinion 15 figures. The reasury sock mehod also uses he exercise price raher han he presen value of he exercise price, which undersaes he diluive effecs of he opions. I also ignores he presen value of he dividends foregone (Barlev 1984). All of hese issues, probabiliy of conversion, presen value effecs on he exercise price and dividends foregone are capured in he fair value of he opions by he opion pricing models. More recenly, Caser e al. (2006) have criicised he SFAS No. 128 approach on a more basic level: The curren definiion of dilued EPS describes a long, complex compuaional process, and i defines dilued EPS as being he resul ha one obains a he end of he compuaion. Using he jargon of curren accouning debaes, his definiion is compleely rule-bound and essenially lacks any concepual basis. I also seems oddly ou of sep wih he oher definiions ha FASB has provided for he accouning profession. Research ino he informaion conen of he various EPS figures has suppored he view ha he presen requiremens fail o capure he diluive effecs of poenial fuure shares. Sco and Wier (2000) and Miller e al. (1987) find ha DEPS has a weaker associaion wih sock reurns han basic EPS, suggesing measuremen error in he calculaion of DEPS. This is in conras wih he findings of Jennings e al. (1997) who find ha DEPS explains more of he variaion in sock prices han basic EPS. Rice (1978) 2

6 analysed sock price paerns around he ime DEPS 2 figures were firs disclosed, and finds ha he informaion conained in DEPS was value-relevan. Huson e al. (2001) compare earnings response coefficiens for firms wih varying numbers of diluive insrumens ousanding, and find ha coefficien is smaller for firms wih large numbers of diluive insrumens. Their findings also sugges ha curren DEPS measures do no adequaely reflec fuure diluion. Taken ogeher, hese sudies sugges ha he informaion conained in DEPS is useful, bu is perhaps no correcly measured. Core e al. (2002) direcly address he issue of DEPS measuremen. They derive heir DEPS calculaion by assuming a simple linear relaionship beween share price and earnings. Their DEPS calculaion is similar o he reasury sock mehod, excep ha hey use he fair value of he opions o calculae he adjusmen o he denominaor. Their findings sugges ha he SFAS No. 128 mehod significanly undersaes he diluive effecs of he opions. This paper adds o he work of Core e al. (2002) by deriving an alernaive way o adjus he earnings per share calculaion for he diluive effecs of ESOs. I will be argued ha his mehod has heoreical and pracical advanages over he Core e al. mehod. Deb vs. Equiy An imporan assumpion made hus far is ha equiy-seled ESOs form par of he firm s equiy, and are no liabiliies. This is he approach used in IFRS 2, which follows from he concepual framework definiion of a liabiliy. Liabiliies are defined as: a presen obligaion of he eniy arising from pas evens, he selemen of which is expeced o resul in an ouflow from he eniy of resources embodying economic benefis. IFRS 2 does no classify equiy-seled ransacions as liabiliies because i does no consider he ransfer of shares o be an ouflow of resources embodying economic benefis 3. Many wriers have objeced o his approach (AAA_FASC 2004; Balsam 1994; Kirschenheier e al. 2004; Landsman e al. 2006; Ohlson and Penman 2005). The crux of heir argumens is ha ESOs are economically equivalen o SARs. SARs are accouned for as liabiliies because hey are an obligaion o ransfer cash or oher asses, and are herefore remeasured a each reporing period. The argumen is ha economically equivalen ransacions should be accouned for in he same manner, and hus ESOs should be shown as a liabiliy and remeasured. The American Accouning Associaion s (AAA) Financial Accouning Sandards Commiee (FASC) (AAA_FASC 2004) goes furher o sugges ha his problem indicaes ha he concepual framework needs revision. The fac ha ESOs are economically similar o SARs while he accouning is quie differen cerainly does sugges problems wih he curren accouning rules, bu i does no follow ha ESOs should be accouned for in he same manner as SARs. In fac, while here are many similariies here is one imporan economic difference beween he wo. The obligaion o deliver cash or oher income generaing asses of he business is very differen o he obligaion o deliver shares. Transferring cash or oher asses reduces he eniy s abiliy o generae fuure economic benefis, and a worse may comprise he exisence of he eniy if i resuls in liquidiy or solvency problems. This is no he case wih he obligaion o deliver shares. Issuing more shares obviously dilues he ineress of curren shareholders, bu his is no always derimenal o curren shareholders, and canno resul in he demise of he enire firm. This is one of reasons why accounans go o grea lenghs o reflec liabiliies separaely from equiy on he balance shee. The only ime i makes sense o show ESOs as a liabiliy is if one akes a proprieary raher han eniy view, which is of course he perspecive adoped by hose ha argue for remeasuremen of he ESOs. Which of he eniy or proprieary perspecive is more correc is a debae which will coninue for many years, bu he curren concepual framework is based on he eniy concep. The framework defines equiy as he difference beween asses and liabiliies, raher han he economic ineress accruing o a paricular ype of proprieary holding. This is also consisen wih he curren 2 Rice acually esed he impac of fully dilued EPS informaion, bu fully dilued EPS is subsanially he same as he curren dilued EPS figure. 3 See BC98 of IFRS 2. 3

7 general approach of financial reporing, which is o provide financial informaion on he performance of he eniy, raher han focus on paricular sakeholders. Wha is clear, however, is ha he DEPS calculaion mus ake a proprieary view, as i is clearly reporing he earnings aribuable o ordinary shareholders. Thus while i is argued ha an eniy perspecive is more appropriae for he balance shee and income saemen, i is sill valid o expec pariy beween ESOs and SARs a a DEPS level. Curren rules for dilued earnings per share The requiremens regarding he diluive effecs of opions are conained in IAS 33 p. 45, which saes; For he purpose of calculaing dilued earnings per share, an eniy shall assume he exercise of diluive opions and warrans of he eniy. The assumed proceeds from hese insrumens shall be regarded as having been received from he issue of ordinary shares a he average marke price of ordinary shares during he period. The difference beween he number of ordinary shares issued and he number of ordinary shares ha would have been issued a he average marke price of ordinary shares during he period shall be reaed as an issue of ordinary shares for no consideraion. In oher words, DEPS = Earnings / (Number of ordinary shares + D), where D n o no X P = (1) and D is he diluive effec of he opions which mus be added o he denominaor in calculaing dilued EPS n o is he number of opions ousanding X is he exercise price of he opions P is he average marke price of he exising ordinary shares equaion (1) can be rewrien as his can be simplified o ( Pn n X ) D = 1 o o P (2) D no ( P X ) P = (3) D hus represens he oal inrinsic value of he opions expressed as a proporion of he curren marke price of he ordinary shares. The opions are reduced o an equivalen number of ordinary shares, which is hen added o he exising ordinary shares o calculae he DEPS. Opions ou of he money are ignored. This approach is based on he reasury sock mehod, firs advocaed by APB Opinion No. 15. This mehod sipulaed he diluive effec of he opions can bes be described by calculaing he effec on EPS of he following dummy ransacion; assume he exercise of he opions a year end and use he proceeds from he exercise o repurchase ordinary shares on he marke. The ne increase in he number of shares in issue mus hen be used o adjus he denominaor in he DEPS calc. Clearly he firs sep in his process is flawed, as he opion holders are unlikely o exercise heir opions early, as his would resul in hem losing he ime value of he opions. Thus a more correc applicaion of he reasury sock mehod would be o assume repurchase of he opions a fair value (inrinsic value plus ime value), and hen assume ha he company issues shares a marke value o generae he cash needed o purchase he opions. This approach is more realisic as he opion holders will recover he full marke value of heir opions. This will resul in n o F/P shares being issued, where F is he fair value per opion, 4

8 and n o and P are defined as above. This mehod will herein afer be referred o as he reasury opion mehod. Formally he reasury opion mehod can be expressed as D nof P = (4) where D is again he diluive effec of he opions which mus be added o he denominaor when calculaing DEPS. When one compares he reasury opion mehod wih he curren IAS 33 reamen given in equaion 3, i can be seen ha he inrinsic value of he opions, P X, has been replaced by F, he curren fair value. This makes inuiive sense as he curren fair value of he opions is more likely o reflec he diluive effecs of he opions han he inrinsic value. The reasury opion mehod will also recognise he diluive effecs of opions, even when hey are ou of he money. Of course when he reasury sock mehod was firs advocaed fair value informaion was no commonly used in financial reporing. Recen years have seen a growing use of fair value informaion, and i hus makes sense o inroduce fair value informaion ino he calculaion of DEPS as well. Ineresingly, Core e al. (2002) derive a DEPS calculaion which is exacly he same as he reasury opion mehod described here. They derive heir model by using a formal, heoreical approach which assumes ha he oal firm value is a funcion of economic earnings and ha his value mus be allocaed beween curren shareholders and ousanding opions. Alernaive approach o calculaing DEPS There is no clear reason why he reasury sock (or reasury opion) mehod is he bes mehod o use. In paricular, why should only he denominaor be adjused for he DEPS calculaion? Are here oher valid approaches ha adjus he numeraor? This secion will derive an alernaive model for calculaing DEPS which is based on Ohlson s residual income model (Ohlson 1995). Ohlson provides a heoreical framework which can be used o link equiy values, as undersood by finance heory, o accouning informaion. Similarly o Ohlson, he following noaion is used: P = Toal marke value of equiy a dae d = Ne dividends a dae b = Book value of equiy a dae r = cos of capial R = 1 + r E = The expecaion operaor condiional on dae informaion x = Earnings for he period (-1,) x a = x rb -1 (residual or abnormal income) Ohlson describes he fair value of he company s ousanding sock as being equal o he accouning book value plus fuure expeced abnormal earnings. P = b + R τ = 1 τ E a ( x ) + τ EPS can be considered a measuremen of change in value of equiy ΔP = x d E 1 a τ a a ( x ) + R [ E ( x ) E ( x )] τ = 1 For ease of exposiion, bu wihou resricing he generaliy of he resuls, from his poin onwards he erm + τ 1 + τ 5

9 E 1 is assumed o be zero. a τ a a ( x ) + R [ E ( x ) E ( x )] τ = 1 + τ 1 + τ ΔP = x d To reduce his o per share informaion, each side mus be divided by he number of shares in issue (n s ). ΔP n s = x d n s There are wo problems which mus be overcome before using his model. i) The clean surplus resricion, which underpins he Ohlson model, will no normally be me on a per share basis. ii) The model will also only work on a oal value basis if he issue and buying of shares are value-irrelevan ransacions and GAAP measures capial conribuions a marke value (Ohlson 2005). The firs problem can be overcome by defining EPS ΔBVPS + DPS (Ohlson 2005) where BVPS = Book value per share DPS = Dividends per share Whils his relaionship will no always hold under curren GAAP, i faciliaes he analysis ha follows, as he objecive is o deermine he correc way o calculae EPS in he presence of ESOs. The second problem can be overcome by resricing he analysis o ESOs and hen assuming ha he issue of ESOs are value-irrelevan ransacions. This is a fair assumpion in an efficien marke. Curren GAAP for ESOs requires ha he issue of ESOs be recorded a gran dae fair value. The clean surplus relaionship on a per share level should hus hold in he conex of ESOs. If he company issues opions on is own shares he Ohlson relaionship becomes τ S ( aa + O = b + R E x+ τ ) (Hess and Lüders 2001). τ = 1 This formula differs from he Ohlson (1995) model in ha i) he oal value of he firm is described as he value of curren shareholders (S ) plus he value of ousanding ESOs (O ), and ii) x aa is he abnormal earnings afer accouning for he issue of ESOs as an expense based on he fair value on gran dae (which is wha IFRS 2 does). Following he same devoluion as before we have ΔS + ΔO n s = x d n s To calculae he DEPS, as i relaes o he ordinary shares currenly ousanding, we mus ransform he equaion as follows: ΔS n s = x d n ΔO s Imporanly, and in conras o he IAS 33 mehod, his indicaes ha earnings mus be adjused by he change in he value of he ESOs. This approach is no more complicaed han he reasury opion mehod 6

10 and requires no addiional esimaes. In a sense, he change in he value of he ESOs represens a shif of value from he curren shareholders o he fuure shareholders. This ransfer in value is capured by reducing he earnings aribuable o he curren shareholders by he change in value of he ESOs. Careful observers would have noiced ha his will resul in a DEPS figure ha is similar o ha which would be obained if he opions were cash-seled. The only difference is ha he accouning for SARs would only reflec a par of he change in fair value of he opions. Curren GAAP for SARs only requires ha a porion of he change in fair value be recognised in he curren period. This porion is equal o he proporion of SARs expensed versus he oal value of SARs o be expensed. The fac remains, however, ha his approach resuls in a similar answer o wha would be obained wih cash-seled opions, which is useful as he wo ransacions are economically similar. The wo reamens can be made idenical by including he unrecognised change in SARs value in he earnings figure used for he DEPS calculaion. This will beer describe he ransfer in value from he curren shareholders o he SARs recipiens. The mehod of calculaing DEPS described above, by which earnings is reduced by he change in he fair value of he ousanding opions, raher han any adjusmen o he denominaor, will herein afer be referred o as he earnings adjusmen mehod. The analyses ha follow will hus compare he hree mehods described hus far; he reasury sock mehod (he curren IFRS 2 mehod), he reasury opion mehod (or Core e al. mehod) and he earnings adjusmen mehod. Comparison wih Core e al. and Landsman e al. I would be useful a his sage o compare he mehod described above wih hose of Core e al. (2002) and Landsman e al. (2006) Core e al. derive heir DEPS calculaion by equaing oal firm value o he sum of curren ordinary shares value and ousanding opions value. They hen immediaely ranspose heir equaion o express he per share value of ordinary shares as a funcion of oal firm value, he number of ordinary shares ousanding, he number of opions ousanding and he fair value of he opions and he ordinary shares. Firm value is hen assumed o be a muliple of accouning earnings, which hen allows hem o express per share dilued earnings as a funcion of accouning earnings in erms of he following formula: DEPS = E [ n + n ( F P) ] s o / where E is he accouning earnings n s is he number of shares ousanding n o is he number of opions ousanding F is he fair value of he ousanding opions P is he marke price of he exising ordinary shares This is similar o he approach used o derive he earnings adjusmen mehod above, excep ha Core e al. reduce heir equaion o per share informaion a a much earlier sage in he derivaion process, which forces hem o assume a simple linear relaionship beween earnings and share price. The approach used here allows for a more complex, albei sill linear, relaionship beween earnings and share value as described by Ohlson (1995). The Core e al. calculaion also focuses on allocaing earnings beween he various equiy paricipans, which may be more useful in predicing fuure aribuable earnings. The earnings adjusmen mehod describes he shifs in value beween he various equiy paricipans, and should mirror changes in he share price. The Core e al. mehod has he advanage of being relaively simple and easy o undersand. I is also a naural progression from he reasury sock mehod and changing IAS 33 o require his mehod would correspond wih oher shifs in accouning sandards o require he use of fair value informaion. Landsman e al. (2006), like some of he oher auhors who have argued ha ESOs should be reaed as liabiliies, assume a proprieary view. They hen demonsrae, also by using he Ohlson (1995) model, ha he mos correc way o accoun for ESOs is o recognise an asse and liabiliy on gran dae. The asse represens hose fuure services ha will be received from he employees, in much he same way ha an 7

11 asse would be raised if he firm made a cash advance o employees in lieu of services which will be delivered over an exended period (someimes called a resrain of rade paymen). The liabiliy represens he fuure claims he ESO holders have agains he firm, which mus reduce he firm value aribuable o he ordinary shareholders. The ESO asse would probably be amorised over he vesing period of he ESOs, which will give an idenical resul o he IFRS 2 mehod from an income saemen perspecive. The ESO asse and ESO liabiliy would also perfecly offse each oher on gran dae, which will also give he same ne asse value as he IFRS 2 mehod. The wo will differ pos gran dae, however, as he ESO liabiliy under he Landsman e al. mehod mus be adjused for changes in he fair value of he ousanding ESOs. Again his will resul in a similar effec o ne asse value o ha which would be obained if he opions were cash-seled raher han equiy-seled 4. Landsman e al. focus on he oal value of he ordinary shares, raher han he per share value, bu i is easy o exend heir heoreical analysis o a per share level. Their approach o calculaing DEPS would hus probably be he same as ha suggesed in his paper. The poin of deparure would be ha hey sugges ha a proprieary perspecive should be used in compiling accouning informaion, whereas his paper has argued for reenion of he eniy approach in preparing he income saemen and balance shee, bu agrees ha a proprieary view mus be used for he DEPS calculaion. I is ineresing o noe ha none of he analyses indicae ha he DEPS needs o be adjused for he unexpensed ESO gran dae value. The curren IAS 33 mehod of accouning for ESOs requires ha he unrecognised ESO expense be added o he exercise price when calculaing DEPS 5. I is assumed ha he sandard seers believed ha he employees pay more han he exercise price for heir opion, being he exercise price plus heir fuure services. None of he heoreical analysis sugges ha his is necessary (Core e al. 2002; Hess and Lüders 2001; Landsman e al. 2006; Ohlson and Penman 2005). The Landsman e al. (2006) approach provides he bes explanaion for his. Their mehod would require recogniion of he oal gran dae value of he ESOs graned as an asse, which mus hen be amorised. This asse is hen offse by he recogniion of a liabiliy. This is essenially wha IFRS 2 does, excep ha i does no acually show he asse and liabiliy on he balance shee. Provided he economic value of he asse is equal o he value of he liabiliy, which i should be in an efficien marke, here is no diluion of curren ordinary shareholders on gran dae. The ineress of ordinary shareholders may be dilued a a laer sage if he value of he liabiliy increases, which will occur if he value of he ESOs increases. The diluion is hus no caused by he iniial gran dae value of he opions, bu raher by subsequen changes in ha value. Thus he oal gran dae value need no play any role in he calculaion of DEPS. The examples ha follow will also help o emphasise his poin. Examples The differences beween he differen approaches o accouning for he diluive effecs of ESOs can bes be undersood by means of a series of simple examples. The firs example ses he scene and describes a simple firm wih known fuure income. The firm hen issues shares as compensaion o employees o creae addiional income which is exacly equal o he economic value of he shares issued. I is useful o begin he examples wih shares raher han opions o help esablish a base case scenario, and inroduce he complicaions associaed wih opions and uncerainy in he subsequen examples. This firs example also helps o address an issue raised by many criics of IFRS 2, being ha reflecing he issue of shares as a cos in he income saemen and reducing he earnings per share by increasing he number of shares in he denominaor resuls in double-couning. As can be seen in he example here is a win effec; o accuraely reflec he economic posiion of he eniy he value of he shares mus be shown in he income saemen, and he earnings per share mus be reduced by increasing he number of shares used in he calculaion 6. 4 Again, he resul will no be exacly he same as SARs accouning because under IFRS 2 only a porion of he change in he fair value of he ousanding SARs is accouned for. 5 IAS 33 p. 47A 6 This example is similar o ha used by Bodie e al. (2003). 8

12 Example 1 Firm A begins in 2007 wih en shareholders, no liabiliies and an asse of CU cash. The cash earns a reurn of 10% p.a. Ineres raes are consan hroughou. Ignore axes. No. of shareholders 10 Cash asse 100 Reurn on cash deposi 10% A he beginning of 2007, Firm A issues wo shares o employees, in reurn for which hey will provide services o he firm. The expense recognised in erms of IFRS 2 will be he fair value of he shares a he gran dae, being CU 20. Assume ha he ransacion wih he employees is value-irrelevan, hus in his case he firm will earn addiional benefis of CU 22 in This is CU 20 o compensae for he value of he shares a gran dae, plus CU 2 o compensae for he fac ha had he shares been issued for cash of CU 20 a he beginning of 2007 CU 2 ineres would have been earned on his cash. Saed differenly, he benefis mus equal he end of year value of he shares. Assume furher ha he benefis of CU 22 are realised in cash. Thus he income saemen of Firm A will reflec: CU Ineres received (100 x 10%) 10 Income from services 22 IFRS 2 expense (20) Ne income 12 EPS (12/12) 1 Noe ha he original shareholders are no beer or worse off han hey would have been had he ransacion wih he employees no occurred. Their EPS is CU 1 and NAV per share a he end of 2007 CU 11 8 regardless. This resul is achieved only because he cos of he services is refleced in he earnings and he diluive effec of he new shares is incorporaed ino EPS. Example 2 Amending example 1, Firm A now issues wo opions o employees (insead of shares) a he beginning of 2007 in reurn for which hey will provide services o he firm. The opions enile he employees o purchase shares in he business a he end of 2007 for CU 10 per share. Assume again ha he ransacion wih employees is value-irrelevan, hus he employees will generae benefis of CU 2. As Firm A is a very simple business, he payous from he opions are known wih cerainy a he beginning of 2007 (assuming consan ineres raes). The value of he opions on expiry wih hus be CU 1 per opion ([( ) / 12] 10). Discouned back o he beginning of he year, he gran dae value of he wo opions is CU 1,82 (2/1,1). As in he firs example, i can be seen ha he benefis derived from he employees mus equal he end of year fair value of he opions received. 7 Consisen wih Inernaional Financial Reporing Sandards, he symbol CU is used o refer o currency unis. 8 [( ) / 12] wih he ransacion wih he employees and (110 / 10) wihou. 9

13 Thus he income saemen of Firm A will reflec: CU Ineres received 10,00 Income from services 2,00 IFRS 2 expense (1,82) Ne income 10,18 Calculaing he DEPS by using he earnings adjusmen mehod gives: DEPS 10,18 0,18 = 10 = 1 The 0,18 is he change in value of he opions over he year (2 1,82). Because his example deals wih a single period he DEPS figure for he oher wo mehods is he same, i.e. CU 1 9. Example 3 I is hus useful o exend he example o muliple periods o illusrae some of he differences beween he mehods. If we assume ha he opions ves and expire afer wo years, he equilibrium occurs where each opion is iniially worh CU 1, The employees mus provide benefis of CU 2,08 11 in he firs year (1,74 + 1,74 x 2 x 10%) and CU 2,12 in he second year (1,74 + 1,74 x 2 x 1,1 x 10%). The IFRS 2 expense will be 1,74 each year. Y1 Y2 CU CU Ineres received 10,00 11,00 Income from services 2,08 2,12 IFRS 2 expense (1,74) (1,74) Ne income EPS Earnings adjusmen mehod 1,00 1,10 Treasury sock mehod 1,02 1,10 Treasury opion mehod 1,00 1,10 Share price 11,00 12,10 Opion value 1,91 2,10 The earnings adjusmen mehod and he reasury opion mehod boh give he same EPS figures, whils he reasury sock mehod undersaes he diluive effecs in he firs year because i uses he inrinsic value raher han he fair value of he opions. Example 4 Example 4 uses he same informaion as Example 3, bu inroduces uncerainy. I is assumed ha invesors are risk neural so ha he expeced rae of reurn does no change 9 On mauriy he fair value and he inrinsic value of he opions will be he same, and he DEPS is 10,18 / [ 10 + (2 / 11)]. 10 [(10 x 1,1 2 ) 10] / 1, Some errors are induced by rounding o wo decimals. 10

14 from he previous examples. Circumsances change a he end of year 1 such ha he earnings for year 2 are equally likely o be CU -18,62 or CU 41,38. Y1 Y2(1) Y2(2) E(Y2) CU CU CU CU Ineres received 10,00-19,00 41,00 11,00 Income from services 2,08 2,12 2,12 2,12 IFRS 2 expense 1,74 1,74 1,74 1,74 Ne income 10,35-18,62 41,38 11,38 Earnings adjusmen mehod 0,964-1,444 3,636 1,096 Treasury sock mehod 1,017-1,862 3,893 1,101 Treasury opion mehod 0,997-1,862 3,893 1,096 Share price 10,96 9,52 14,60 12,06 Opion value 2,09-4,60 2,30 Y2(1) and Y2(2) represen he wo possible sae realisaions, and E(Y2) is he expeced oucome for year 2. As is known from opion heory, he value of he opion a he end of year 1 will increase because of he inroducion of uncerainy. The opion can now be calculaed as being worh CU 2,09 a he end of year 1 and CU 0 or CU 4,60 depending on which sae is realised a he end of year 2. The resuls of year 1 are as before, excep ha he earnings per share have dropped o reflec he shif in value from he ordinary shareholders o he opion holders. I can be seen ha he earnings adjusmen mehod exacly capures he increase in wealh of he ordinary shareholders, which neiher of he oher wo mehods do. Ineresingly, however, he reasury opion mehod gives he bes predicor of year 2 earnings, because CU 0,997 x 1,1 is exacly he expeced EPS in year 2. Thus i seems ha he earnings adjusmen mehod beer reflecs he change in economic wealh of he curren ordinary shareholders, whils he reasury opion mehod is more useful in deermining fuure earnings. The reasury sock coninues o undersae he diluive effecs of he opions and does no appear o conain any useful informaion beyond ha already refleced in he oher wo figures. Example 5 Example 5 changes Example 4 slighly by inroducing he uncerainy a he beginning of year 1 raher ha a he end. This will change he iniial gran dae opion value, and he benefis derived from he employees are adjused so ha he ransacion coninues o be value-irrelevan a he gran dae. Y1 Y2(1) Y2(2) E(Y2) CU CU CU CU Ineres received 10,00-19,00 41,00 11,00 Income from services 2,30 2,34 2,34 2,34 IFRS2 expense 1,92 1,92 1,92 1,92 Ne income 10,38-18,58 41,42 11,42 Earnings adjusmen mehod 1,000-1,436 3,636 1,100 Treasury sock mehod 1,020-1,858 3,895 1,104 Treasury opion mehod 1,000-1,858 3,895 1,100 Share price 11,00 9,56 14,64 12,10 Opion value 2,11-4,64 2,32 As here is no shock o he sysem as in Example 4, boh he earnings adjusmen and he reasury opion mehod provide he same informaion, and accuraely convey he change 11

15 in economic wealh of he ordinary shareholders. The reasury sock mehod oversaes he DEPS in year 1 and in he expeced DEPS for year 2. Conclusion IAS 33 needs revision. The reasury sock mehod prescribed by IAS 33 wen ou of dae in he 1970 s, and he inroducion of fair value accouning in many oher areas of financial reporing mean ha using fair value informaion in he calculaion of DEPS should no pose any new pracical problems. The new IAS 33 needs o begin wih a clear objecive. This objecive mus be o accuraely convey he change in economic value of one ordinary share. This necessiaes a proprieary view, bu his need no influence he res of he accouning informaion, and he income saemen and balance shee should coninue o reflec he eniy perspecive. This paper has described an earnings adjusmen mehod for calculaing dilued earnings per share. This mehod reduces he earnings used in he earnings per share calculaion by he change in he value of he ousanding employee sock opions. There is no adjusmen made o he number of shares in issue, as is currenly prescribed by IAS 33. The heoreical analysis and examples ha follow sugges ha he earnings adjusmen mehod described in his paper beer reflecs he changes in economic value aribuable o ordinary shareholders. Fuure research should deermine wheher his can be confirmed empirically. The earnings adjusmen mehod has he added advanage of giving a similar resul o ha of cash-seled opions a a DEPS level, which should go some way o appeasing hose criics ha believe ha equiy-seled employee sock opions should be reaed as liabiliies and remeasured a each reporing period. 12

16 References American Accouning Associaion (AAA) Financial Accouning Sandards Commiee (FASC) Evaluaion of he IASB's Proposed Accouning and Disclosure Requiremens for Share-Based Paymen. Accouning Horizons vol. 18 no. 1 (March 2004): American Insiue of Cerified Public Accounans, Accouning Principles Board Earnings Per Share. Opinion No. 15 AICPA. Balsam, S Exending he Mehod of Accouning for Sock Appreciaion Righs o Employee Sock Opions. Accouning Horizons vol. 8 no. 4 (December 1994): Barlev, B Theory, Pragmaism and Conservaism in Reflecing he Effecs of Warrans on Dilued EPS. Abacus vol. 20 no. 1:1-15. Bierman, H Common Sock Equivalens, Earnings per Share, and Sock Valuaion. Journal of Accouning, Audiing and Finance (Winer 1986): Bodie, Z., R. S. Kaplan, and R. C. Meron For he Las Time: Sock Opions Are an Expense. Harvard Business Review (March 2003): Caser, A. B., R. J. Elson, and L. G. Weld Is Dilued EPS Becoming More Ar han Fac? CPA Journal vol. 76 no. 9 (Sepember 2006): Core, J. E., W. R. Guay, and S. P. Kohari The economic diluion of employee sock opions: Dilued EPS for Valuaion and Financial Reporing The Accouning Review vol. 77 no. 3 (July 2002): Dechow, P. M., A. P. Huon, and R. G. Sloan Economic Consequences of Accouning for Sock- Based Compensaion. Journal of Accouning Research vol. 34:1-20. Financial Accouning Sandards Board (FASB) Earnings Per Share. Saemen of Financial Accouning Sandards No Norwalk, CT: FASB. Financial Accouning Sandards Board (FASB) Share-based Paymen. Saemen of Financial Accouning Sandards No. 123 (revised). Norwalk, CT: FASB. Hess, D., and E. Lüders Accouning for Sock-Based Compensaion: An Exended Clean Surplus Relaion. Discussion Paper 01-42, Cener of Finance and Economerics, Universiy of Konsanz [cied Dec 2006]. Available from fp://fp.zew.de/pub/zew-docs/dp/dp0142.pdf. Huson, M. R., T. W. Sco, and H. A. Wier Earnings Diluion and he Explanaory Power of Earnings for Reurns. The Accouning Review vol. 76 no. 4 (Ocober 2001): Inernaional Accouning Sandards Board (IASB). 2003a. Earnings Per Share. Inernaional Accouning Sandard 33. London: IASB. Inernaional Accouning Sandards Board (IASB). 2003b. Share-based Paymen. Inernaional Financial Reporing Sandard 2. London: IASB. Jennings, R., M. J. LeClere, and R. B. Thompson Evidence of he Usefulness of Alernaive Earnings per Share Measures. Financial Analyss Journal (November/December 1997): Jerris, S. I Opion-Based EPS Measures: An Alernaive o Primary and Fully-Dilued EPS. Journal of Business Finance and Accouning vol. 19 no. 3 (April 1992): Kirschenheier, M., R. Mahur, and J. K. Thomas Accouning for Employee Sock Opions. Accouning Horizons vol. 18 no. 2 (June 2004): Landsman, W. R., K. V. Peasnell, P. F. Pope, and S. Yeh Which approach o accouning for employee sock opions bes reflecs marke pricing? Review of Accouning Sudies 11: Michaels, A., and R. Waers EDS Backracks on Expensing Opions. Financial Times, March 26, 2004, p. 29. Millar, J. A., T. Nunhirapakorn, and S. Courenay A Noe on he Informaion Conen of Primary and Fully Dilued Earnings Per Share. Financial Analyss Journal (Sepember-Ocober 1987): Ohlson, J. A Earnings, Book Values, and Dividends in Equiy Valuaion. Conemporary Accouning Research vol. 11 no.2 (Spring): On Accouning-Based Valuaion Formulae. Review of Accouning Sudies 10: Ohlson, J. A., and S. H. Penman Deb vs. Equiy: Accouning for Claims Coningen on Firms' Common Sock Performance wih Paricular Aenion o Employee Compensaion Opions. Whie Paper No. 1, Cener for Excellence in Accouning and Securiy Analysis, Columbia Universiy. Rice, S. J The Informaion Conen of Fully Dilued Earnings per Share. The Accouning Review vol. 53 no. 2 (April):

17 Sco, T. W., and H. A. Wier On Consrucing an EPS Measure: An Assessmen of he Properies of Diluion. Conemporary Accouning Research vol. 17 no. 2 (Summer 2000): Vigeland, R. L Diluion of Earnings per Share in an Opion Pricing Framework. The Accouning Review vol. 57 no. 2 (April 1982):