Measuring the Risk and Performance in Plantation Sector Using CAPM Based Jensen s Alpha

Transcription

1 Measurng the Rsk and Perormance n Plantaton Sector Usng CAPM Based Jensen s Alpha D.A.I. Dayaratne D.G Dharmaratne SA Hars Department o Accountancy and Fnance Sabaragamuwa Unversty, Belhuloya Abstract Ths paper evaluates the company perormance usng the CAPM Based Jensen s Alpha. The CAPM o Sharpe (1964), Lntner (1965) and Mossn s a wdely used model n modern Fnance to estmate cost o equty and company perormance. We carred out our study or ten plantaton companes lsted on Colombo Stock Exchange (CSE). We used cost o equty that s calculated usng CAPM to determne the Economc Value Added (EVA).The EVA measures whether the companes have created shareholders value durng the estmatng perod. We selected the sample perod o 2000 to 2005 years and we appled the monthly endng prces o common stocks o each company or the regresson. The monthly endng prces o All Share Prce Index (ASPI) are used as the market proxy. To estmate the beta, we appled market model. It was ound that almost all the companes have created value or ther shareholders durng the study perod. To measure the market perormance, we calculated Jensen s alpha or each company and accordng to Jensen s alpha we ound that the market perormance s not satsactory n most plantaton companes. These results are mportant or Corporate Managers undertakng rsk calculatons, or und managers makng nvestment decson and, amongst others, or nvestors who wsh to assess value o ther nvestments. 1. Introducton The Captal Asset Prcng Model (CAPM) o Sharpe (1964) and Lntner (1965) has receved consderable attenton n nancal studes. In ts smplest orm, the CAPM predcts that the excess return o a stock should be proportonal to the market premum. The proportonalty actor s known as the systematc rsk or beta o an asset 1. Early emprcal studes on the CAPM such as Black et al. (1972) and Fama and MacBeth (1973) were supportve o the mplcatons o the model. That s, the average return o hgh beta stocks was hgher than the average return o low beta stocks. An early study by Levy (1977) showed that the analyst used a shorter tme horzon then, the beta estmates were based. Fama (1980, 1981) provded evdence that the power o macroeconomc varables n explanng the stock prces ncreased wth ncreasng tme length. The model gves us a precse predcton o the relatonshp that we should observe between the rsks o an asset and ts expected return. Ths relatonshp 1 See Campbell (2000) Cochrane (1999) or a survey o the recent developments n the nance lterature n general and asset prcng n partcular Sabaragamuwa Unversty Journal, vol 6, no. 1,pp

2 serves two vtal unctons. Frst, t provdes a benchmark rate o return or evaluatng possble nvestments. For example, we are analyzng securtes, we mght be nterested n whether the expected return we orecast or a stock s more or less than ts ar return gven ts rsk. Second, the model helps us to make an educated guess as to the expected return on assets that have not yet been traded n the market place. Estmaton o expected return or cost o equty or ndvdual stocks s central to many nancal decsons such as those relatng to portolo management, captal budgetng, and perormance evaluaton. The two man alternatves avalable or ths purpose are a sngle- actor model (or Captal Assets Prcng Model (CAPM) and the three-actor model suggested by Fama and French (1992). Despte a large body o evdence n the academc lterature n avour o the Fama and French model, or estmaton o portolo returns, practtoners seem to preer CAPM or estmatng cost o equty (see, or example, Bruner, Eades, Harrs, & Hggns, 1998; Graham & Harvey, 2001). Snce the development o the CAPM n the early 60s, many tests, especally n the USA, have been perormed n order to measure how well the model stands n the presence o real le condtons. In ths paper, we use CAPM and Jensen s alpha to evaluate the company perormance n Sr Lankan market context. The objectves o ths study are nvestgatng the company specc rsk and market rsk, measurng the perormance o the plantaton companes and nvestgatng whether the companes have created value or the shareholders. 2. Issues n Research Desgn As a result o the Sze eect, the book value and market value o common stock s not postvely correlated n developng markets when compared to nancally developed countres (Banz, 1981). Ths s a contradctng stuaton, because the CAPM developed wth the mark rms s hgher than that o larger rms. Stattman (1980), and Lansten (1985) nd that average returns on US stocks are postvely related to the rato o a rm s book value o common equty, (BE) to ts market value, (ME). Further Chan, Hamao, and Lakonsshok (1991) nd that Book-to- Market equty, BE/ME, also has a strong role n explanng the cross-secton o average returns on Japanese stocks. Basu (1983) shows that earnngs-prce ratos (E/P) help to explan the cross-secton n tests that also nclude Sze, and Market beta. In addton to ths, Ball (1978) argues that E/P s a catchall proxy or unnamed actors n expected returns. As emprcal evdence s largely nconsstent wth the Captal Assets Model n a developed market scenaro, t s useul to examne the nature o these relatonshps n an emergng stock market such as Sr Lankan. Further, there have been no publshed studes on ths aspect n the Sr Lanka market. 3. The CAPM and the Real World In lmted ways, portolo theory and the CAPM have become accepted tools n the practtoner communty. Many nvestment proessonals thnk about the dstncton between rm-specc and systematc rsk and are comortable wth the D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 69

3 use o beta to measure systematc rsk.. Stll the nuances o the CAPM are not nearly as well establshed n the communty. For example, the compensaton o portolo managers s not based on alphas calculated relatve to the securty market lne. What can be made o ths? New ways o thnkng about the world (that s, new models or theores) dsplace old ones when the old models become ether ntolerably nconsstent wth data or when the new model s demonstrably more consstent wth avalable data. For example, when Coperncus overthrew the age-old bele that the earth s xed n the centre o the unverse and that the starts orbt about t n crcular motons, t took many years beore astronomers and navgators replaced old astronomcal tables wth superor ones based on hs theory. The old tools t the data avalable rom astronomcal observaton wth sucent precson to suce or the needs o the tme. To some extent, the slowness wth whch the CAPM has permeated daly practce n the money management ndustry also has to do wth ts precson n ttng data that s precsely explanng varaton n rates o return across assets. The CAPM was rst publshed by Sharpe n the Journal o Fnance n 1994 and took the world o nance by storm. Douglas (1969) was the rst to cast doubt on the emprcal content o the model. Douglas ound damnng evdence on two counts. Frst, contrary to the predctons o the theory, non-systematc rsk dd seem to explan average returns. Seconds, the estmated securty market lne was too shallow, that s ts nept was greater than the rsk ree rate, mplyng that deensve stocks (β< 1) tended to have postve alphas, whle aggressve stocks(wth β >1) tended to have negatve alphas. Four years later, Mller and Scholars (1972) publshed a paper demonstratng ormdable statstcal problems that hnder a straghtorward test lke that o Douglas. They estmated the potental error that may have resulted rom each step o Douglas s procedure and sure, enough, they were able to ratonalze hs ndngs. But Mller and Scholar s explanaton does not tsel provde postve evdence that the CAPM s vald. Later studes most notably those o Black, Jensen and scholars (1972), and Fama and MacBeth (1973), used procedures desgned to address the varous econometrc problems. The most mportant o these was to test the CAPM usng cleverly constructed portolos to dmnsh the statstcal nose resultng rom rm specc rsk. But even these eorts could not establsh the valdty o the CAPM. Whle all ths accumulatng evdence aganst the CAPM remaned largely wthn the vory towers o academa, Roll s (1977) paper enttled A Crtque o Captal Assets Prcng Tests shook the practtoner world as well. Roll argued that snce the true market portolo can never be observed, the CAPM s necessarly untestable. The publcty o the new classc Roll s crtque resulted n popular artcles such as Is Beta Dead? That eectvely slowed the permeaton o portolo theory through the world o nance. Ths s qute ronc snce, although Roll s absolutely correct on theoretcal grounds, some tests suggest that the error ntroduced by usng a broad market ndex as proxy or the true, unobserved market portolo s perhaps the lesser o the problems nvolved n testng the CAPM. 70 Measurng the Rsk and Perormance

4 Fama and French (1990) completed a study that dealt the CAPM an even harsher blow. It clamed that once you control or a set o wdely ollowed characterstcs o the rm, such as the rato o market value to book value, the rm s beta does not contrbute anythng to the predcton o uture returns. The Economst and the New York Tmes pcked up ths tme the pece even beore t was publshed n the Journal o Fnance. Ths latest crtque became the central topc o an academc conerence and a slew o studes. None o these has been publshed yet, but the gst o the emergng results s that Fama and French s conclusons are hampered by subtle problems n statstcal technque. The latest studes, employng the most powerul technques to date, show that systematc rsk does help explan rates o return. Ths s what keeps the CAPM alve and useul n the world o economcs regulaton. 4. Compettve Market Structure and the CAPM One o the drectons o research n the Asset Prcng Models (APM) s to develop a model that can explan better the prce behavor o securtes n a securty market. The CAPM as developed by Sharpe (1964), Lntner (1965) and Mossn (1966), was the rst ormal step n ths drecton. Then, Ross (1977) developed a more generalzed model, the Arbtrate Prcng Theory (APT) model. Though the CAPM and the APT are desred under derent sets o assumptons, both provde one undamental result - a lnear relatonshp between Expected Return E(R) and a measure o systematc Rsk. The lnear relaton between expected return and systematc rsk n the CAPM and the APT s manly because o the compettve market structure. Further generalzaton o the asset prcng models wth the compettve market structure assumpton wll be a trval exercse, as t would always result n a lnear relaton. 5. Objectves o the Research Ths study ams to acheve several objectves coupled wth the man objectve Prmary Objectve To measure the market perormance o the Plantaton Sector Secondary Objectves To measure the company specc rsk and market rsk To examne shareholders value Compare the perormance o the companes 6. Methodology A model descrbng the relatonshp between rsks and expected return s used n the prcng o rsky securtes. CAPM says that the expected return o a securty or a portolo equals the rate on a rsk-ree securty plus a rsk premum. I ths expected return does not meet or beat the requred return, then the nvestment should not be undertaken. Estmatng expected return s crucal or many nancal decsons, such as nvestment decsons, captal budgetng decsons and perormance evaluaton usng measures such as EVA. From two recent surveys, Bruner, Eades, Harrs, and Hggns (1998) and Graham and Harvey (2001), the Captal Asset Prcng D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 71

5 Model (CAPM) was ound to be the model most avoured by practtoners or dong ths. Academcs also commonly base estmates o expected return on CAPM. The reason or ths wdespread use o CAPM s probably ts apparent ease o mplementaton. The CAPM equaton or the Securty Market Lne s gven as (1) E( R ) = R + β [ E( R ) R ] m Where, E ( R )= The expected return on securty R = The rsk-ree rate = β The systematc rsk dened as E( R ) m = (, ) CovR R m 2 σ The expected return on market portolo 6.1 Beta Estmaton The Captal Asset Prcng Model s essentally the reducton o Modern portolo theory nto a sngle actor model- wth that the sngle actor beng called Beta. Instead o a matrx o co -varances between all securtes n the market, there s only one covarance coecent, beta the covarance between a securty and the market. The standard method o estmatng beta s to regress hstorcal returns o company s stocks aganst the return on the market or the same perod (sample perod s ve years and the number o observaton s 60). In Sr Lankan market we use the 2 All Share Prce Index (ASPI) as the market return. The ormula s: R = α + β ( ) + ε R m Where, α = The ntercept o the regresson Co varance( R, R β = solpe= M 2 σ m ) (2) 2 The ASPI measures the movement o share prces o all lsted companes. The ASPI s based on market captalzaton. Weghtng o shares s conducted n proporton to the ssued ordnary captal o the lsted companes, valued at current market prce (.e. market captalzaton). 72 Measurng the Rsk and Perormance

6 6.2 Estmatng the rsk premum The selected companes are hgh lqud, requently traded and prot makng companes. Thereore as a proxy or determnng the market rsk premum the dvdend growth model s appled. Assumng that the stock market prces the securtes correctly, ths method employs the ollowng equty valuaton model to generate an expected return or the market; then the rsk ree rate s deducted rom the expected return to arrve at the equty rsk premum. The same method was appled by Gunesekara(2004) to calculate the rsk premum. D V = E( R ) g Solvng the above equaton we can derve the ollowng ormula D E( R ) = + ( V g) 6.3 Evaluatng Company Perormance V The requred rate o return determned by CAPM provdes a market-based measure o the return requred by shareholders or nvestng n the rm. Ths method s consstent wth Gunasekarage (2004). Ths s the cost o equty captal o the rm that can be used as the benchmark rate or evaluatng perormance o nvestment proposals. The return on equty (ROE) can be compared wth the CAPM-based requred return on equty to determne whether company managers have worked or the best nterest o shareholders by nvestng n value creatng nvestment projects. For ths purpose the equty Economc Value Added (EVA) s appled. It can be calculated as ollows. Equty EVA = (ROE- Ke) (Equty Invested n the Frm) The non-tradtonal method o measurng EVA s the operatng prot approach. As dened by Stewart (1991), EVA s net operatng prot mnus approprate charge or the opportunty cost o all captal n an enterprse. In ths paper we use tradtonal method o accountng prot to measure the EVA. The evaluaton crteron s that the EVA s a postve gure, the company s creatng more wealth or nvestors and t s negatve, t s destroyng the shareholders wealth. To evaluate the return perormance o the companes, the ollowng ormula s used by rearrangng the CAPM R R R I = R + β ( Rm R ) (a) = R +βrm βr (b) = R ( 1 β ) + βrm (c ) α + β Usng the equaton (c ) wth the regresson equaton ( R = ( Rm )) the comparson o α n the regresson equaton wth R ( 1 β ) n the rearranged D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 73

7 verson o the CAPM equaton provdes the measure o perormance o shares relatve to CAPM. Ths measure s called as Jensen s alpha and t s calculated as: The evaluaton crtera are I the Jensen s alpha s postve the shares have perormed well and t s negatve the shares have perormed badly Assumpton o the Model (CAPM) The CAPM can be used to nd prces o rsky assets. The theory o (Sharpe, 1964) predcts a lnear relaton between rsk and return o the rm: E( R ) = R + β Jensen' s. Alpha [ E( R ) R ] m = α [ R (1 β )] (5) In the above equaton, or a gven rsky asset., β s the senstvty o the return o assets to movements o the return o the market, and t s dened as the normalzed covarance between the return o the rsky asset and the return o the market portolo. Hgh values o β ndcate a rsker asset and low values o β ndcate a more secure asset (n the lmt, the return o asset wll tend to the rsk-ree rate as β approaches zero). Ths model apples n markets wth perect normaton where all nvestors are utlty maxmzers and have smlar expectatons about the mean and standard devaton o the return o every rsky asset. An asset wth zero β yelds a rskree rate at whch every nvestor can borrow or lend. Also, n ths world, there s a portolo where every asset n the economy s ncluded, proportonal to ts market value, and by denton, ts β s 1.0. Eq. (1) and the above assumpton ndcate that every asset n a gven market wll adjust ts prce untl the expected return adjusted or rsk generates a return equal to the return predcted by Eq. (1). In other words, every asset must le on the securty market lne (a lne that wll ntercept the vertcal axs at the rsk-ree rate and wll have a slope equal to the rsk premum n the market). Note here that every nvestor s concerned only wth the systematc rsk, whch s wth the rsk o the market as a whole, because the unque rsk s dversed away by a wellbalanced portolo. For ths reason, β s the only concern nvestors have when they value securtes. 6.5 Data and the sample We selected plantaton sector to estmate the CAPM and to measure ts perormance. The plantaton sector s a newly lsted sector n the Colombo Stock Exchange (CSE) and t s a growng ndustry, whch contrbutes a consderable porton to the Gross Domestc Product (GDP). The employment generaton o 74 Measurng the Rsk and Perormance

8 ths sector s also much hgher than n the other sectors. Thereore several stakeholders such as current shareholders, prospectve nvestors, und managers, bankers and employees wll get benets rom ths paper. We gathered monthly endng prces o all the plantaton companes orm the CD o data lbrary, whch s ssued by the CSE. To avod problems whch result rom thn tradng and to keep the analyss manageable, only the stocks, whch are traded more than 75% o the market open days n the estmaton perod, are ncluded. In addton to that we consder the Market Captalzaton, ROE, EPS and regular avalablty o data to nalze the sample. Accordng to the above crtera and avalablty o data throughout the sample perod, the ollowng companes are selected or the analyss. Agalawatte (AGLA) Plantaton Ltd. Balangoga (Bala) Plantaton Ltd. Hapugastenne (HAPU) Plantaton Ltd. Kegalle (KGAL) Plantaton Ltd. Kotagala (KOTA) Plantaton Ltd. Kelan Valley (KELA) Plantaton Ltd. Kahawatte (KAHA) Plantaton Ltd. Talawakelle (TALA) Plantaton Ltd. Udapussellawa(UDA) Plantaton Ltd. Watawala (WATA) Plantaton Ltd. The underlnng theory or CAPM s qute specc n ts recommendaton o ndex; t speces that a value-weghted ndex consstng o all the assets n the world should be used. Snce only a small racton o assets n the world trade on stock exchanges, t s mpossble to construct such an ndex, so a proxy must be used nstead. Thereore the most commonly used proxy s the value-weghted All Share Prce Index (ASPI) n Sr Lanka. Ths s consstent wth the recent studes o Jan and Paula (2005). They have used Standard and Poor s Composte Index as the market proxy. As the rsk-ree rate, the government 12 month Treasury bll rate s appled. The reason or usng ths rate s that ts rsk s zero and companes are valued on the bass o long-term uture cash lows generated by them. 7. Emprcal Results Table 1: Summary Statstcs o Companes Company Me Varance an AGAL BALA HAPU KGAL KOTA KELA KAHA TALA UDA WATA Std. Devaton D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 75

9 We run the regresson on monthly endng prces o each o the ollowng companes wth the market ndex or the perod o 2000/20001 to 2005/2006. Thereore the number o observaton n the regresson s 60. Table 1 shows the output summary o Mean, Varance and Std. devaton o ten companes and Table2 shows the same values or the resduals. Table 2: Summary Statstcs o Resduals Company AGAL BALA HAPU KGAL KOTA KELA KAHA TALA UDA WATA Mea n E E E E E E E E E-18 Varance Std. Devaton Systematc Rsk and Unsystematc Rsk The calculaton o systematc and unsystematc rsk s shown only or the Agalawatte plantaton and the calculatons or other companes are shown n Table 3 that shows the systematc and unsystematc rsk o all the sample companes. 8.1 Agalawatte Plantaton The output o regresson model can be explaned as ollows. The ntercept o the regresson s and the slope o the coecent s the regresson equaton or the AGLA s wrtten as ollows. R AGLA = ( RM ) 76 Measurng the Rsk and Perormance

10 The slope coecent, whch represents the beta o the rm, ndcates that the shares n ths company are tmes as rsky as the market ndex, whch has a beta o 01. The company s beta s not statstcally sgncant. The return o AGLA s not sgncantly nluenced by the return on the market ndex. The beta coecent has a standard error o The true beta o the company can take values between and wth a 95% level o condence. 2 The R o the model s Ths ndcates that 0.57% o the varaton o the AGLA return s explaned by the varaton n the return o market ndex. In other words 0.57% o the rsk o the company comes rom market sources whch s known as the systematc rsk whle the rest 99.43% can be attrbutable to rm specc actors that s known as unsystematc rsk.. Total rsk o the company can be calculated as ollows. β σ = Systematc rsk = (0.1083) (0.0056) σ (Resdual mean square) = Unsystematc rsk 2 ( e ) Total rsk = systematc rsk + unsystematc rsk 2 σ AGLA = = Table 3: Systematc and Unsystematc Rsk o the Companes = Company AGAL BALA HAPU KGAL KOTA KELA KAHA TALA UDA WATA Systematc Rsk 2 ( β α 2 ) Unsystematc 2 Rsk ( α ) Accordng to Table 3, unsystematc rsk components s hgher than the systematc rsk component n most companes. In other words company specc actors are mostly aected or the luctuaton o the market prces o shares n plantaton sector than the market actors. As ar as the plantaton sector s concerned, the perormance s closely related to the weather actors such as ran and drought prevalng n the country. On the other hand, the plantaton sector s badly nluenced by the unon actons o Ceylon Workers Congress (CWC). These actors can be attrbutable to the hgh unsystematc rsk o these companes. D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 77

11 8.2. Evaluatng Perormance We measure the perormance o the companes only or the nancal year 2003/2004. To generate the expected return, we need rsk ree rate n Aprl 2003 and the equty premum or the market. The 12-month Treasury bll Rate at the begnnng o Aprl 2003 was We adopt the mpled equty rsk premum approach to nd the rsk premum or the Sr Lankan Market. Durng the ve-year perod the market dvdend yeld was 6.5% to 3.2%. Thereore average annual growth rate can be computed wth the ollowng ormula. g = The dvdend growth rate s negatve durng ths perod. The expected dvdend growth n 2004 was [ 0.032( ) ] % The market captalzaton at the begnnng o 2003 was 194 bllon. Thereore the expected dvdend or the market or the year 2003 was ( 194x 0.027) Rs bllon. Applyng the ollowng equaton, we receve the expected return or the market X ( ) ( R ) = = 10.5% 194 E M Thus, the equty rsk premum or the market was (-10.5%-9.18%) = -20.3%. Now we can calculate the requred rate o return or the begnnng o the 2003/2004 perod or each company as ollows. E ( ) = ( )( 0.203) = 11.9% R AGLA E ( R BALA ) = ( 0.203) = 2.37% E ( ) = ( 0.203) = 0.72% R HAPU E ( ) = ( 0.203) = 10.9% R KGAL E ( ) = ( 0.203) = 9.0% R KOTA E ( R KELA ) = ( 0.203) = 11.3% E ( R KAHA ) = ( 0.203) = 8.2% E ( R TALA ) = ( 0.203) = 2.7% E ( ) = ( 0.203) = 3.2% R UDA 78 Measurng the Rsk and Perormance

12 E ( ) = ( 0.203) = 10.2% R WATA Accordng to the above calculaton we can observe that the requred rate o return s negatve n all companes except Agalawatte. The major reason or ths s the average annual growth rate has become a negatve gure due to the declnng trend o the market dvded yeld. But n the case o Agalawatte, the beta s more than 1.0. As a result, ts requred rate o return has become a postve gure n ths company. To evaluate the perormance o companes n the nancal year 2003/2004 and durng the estmaton perod o the market model, we calculate equty EVA and Jensen s alpha respectvely or each company. The EVA s calculated usng the Eq.4 as under: Equty EVA = (ROE-Ke) (Equty Invested n the Frm) The EVA and the Jensen s alpha calculaton s shown n the text only or the Agalawatte plantaton. We have summarzed EVA and the Jensen s alpha or other companes n table Agalawatte Plantaton The ROE o Agalawatte or the year 2003/2004 s 12.06% and the cost o equty as calculated by usng CAPM s 11.9%.Equty nvested durng the year amounts to Rs.891 mllon. Now Equty EVA can be calculated as ollows. EVA AGLA = ( )(891) = 1. 42mllon The EVA o Watawala s 1.42 mllon whch means that the company has created wealth o nvestors by 1.42 mllon durng the perod. From the regresson output, the ntercept (α) s equal to ; the average rsk ree rate or the perod s Now the monthly Jensen s alpha or the company s calculated as ollows [ ( 1 ( ) )] = % MonthlyJensen' s Alpha= = Ths should be converted to the annual rate or the study and t s done as ollows 12 Annual Jensen' s Alpha= [( ) ] 1= 0.07= 7.4% The company has outperormed the market durng the estmaton perod, generatng an annual excess return o 7.4% to ts shareholders. Table 4: Perormance o Companes based on Jensen s Alpha and Shareholders wealth Company Name AGAL BALA HAPU KGAL EVA In Mllon (Rs.) Jensen s Alpha % Market Perormance Satsactory Not Satsactory Not Satsactory Not Satsactory D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 79

13 KOTA KELA KAHA TALA UDA WATA Not Satsactory Not Satsactory Not Satsactory Not Satsactory Not Satsactory Not Satsactory Table 4 shows the calculated EVA (n mllon) and the Jensen s alpha or the sample companes. The postve EVA states that the company has created value or the shareholders durng the perod. All our sample companes have created vale or the shareholders durng the year 2003/2004.It means that the companes create value or owners only when ther operatng ncome exceeds the cost o captal employed. Another crucal attrbute o the EVA s that t ntegrates three mportant management unctons: captal budgetng, perormance apprasal and ncentve compensaton. Thereore ths study gves a comprehensve pcture about these companes or the stakeholders, partcularly or the managers n makng crucal manageral decsons. However EVA s just one dmenson o a corporate perormance measure. Other actors such as long term sustanable growth o company are equally mportant. The conceptual oundaton o EVA, one o mportant shareholder value measures, s based on the resdual ncome concept. 9. Rsk Analyss o Companes In nvestors perspectve, the analyss o rsk s very mportant as ther nvestment decsons are hghly nluenced by the degree o rsk assocated wth the nvestment. The degree o rsk s classed n to three categores as β 1 hgh rsky stocks, β < 1>0 average rsky nvestments and β< 1 low rsky nvestment. The table 5 shows the beta coecent o each company and proposed potental nvestors or each company based on the degree o rsk. Table 5: Classcaton o Companes based on rsk Company Name Beta Degree o Rsk Proposed Investors AGAL BALA HAPU KGAL KOTA KELA KAHA TALA UDA WATA Low rsky Average rsky Average rsky Hgh rsky Average rsky Hgh rsky Average rsky Average rsky Average rsky Average rsky Deensve Moderate Moderate Aggressve Moderate Aggressve Moderate Moderate Moderate Moderate Accordng to Table 4 almost all the companes have crated wealth or the shareholders durng the perod whch ndcates that the management o these companes have nvested excess unds n postve NPV projects. An mportant attrbute o economc value added s that the present value o an nvestment s annual EVA stream equals the nvestment s NPV. Ths makes t possble to talk about nvestment apprasal n terms o EVA rather than NPV provded. However n almost all the companes, the Jensen s alpha gets negatve values except 80 Measurng the Rsk and Perormance

14 Agalawatte whch ndcates that these companes have badly perormed n the market 10. Concluson Based on the results obtaned n ths study, we conclude that the unsystematc rsk component s hgher than the systematc rsk n the plantaton sector whch means that the luctuaton o the market prce o stocks s mostly nluenced by company specc actors such as weather condton, producton capacty and CWC unon actons. I the companes dversy ther nvestment they can reduce the unsystematc rsk component. However the beta o most companes s less than 1.0 whch s the market rsk. Thereore t can also be concluded that the rsk o nvestng n the plantaton sector s low rsky as compared to market rsk.. Another mportant ndng o ths paper s that the management o these companes has worked or the best nterest o the shareholders by creatng postve EVA. In other words, the companes have undertaken postve NPV projects durng the perod. However EVA s just one dmenson o corporate perormance measure. Other actors such as long term sustanable growth o company should be consdered. Postve EVA and postve NPV represent the ncrease o shareholders wealth. But accordng to the Jensen s alpha, almost all companes have perormed badly durng ths perod. Reerences Banz, R.(1981), The Relatonshps between Returns and Market Value and Return or NYSE Common Stock: Further Evdence?. Journal o Fnancal Economcs, 9(1), pp.3-18 Eugene Fama and James MacBeth Rsk, Return and Equlbrum: Emprcal Tests. Journal o Poltcal Economy Fama, E., & French, K Common rsk actors n the returns on stocks and bonds. Journal o Fnancal Economcs, 33(1),3-56 Graham, J.R., & Harvey, C.R The theory and practce o corporate nance: evdence rom the eld. Journal o Fnancal Economcs, 60(1-2), Gunasekarage, A., Estmatng CAPM n practce and Evaluatng Company Perormance. Journal o the Insttute o Chartered Accountants o Sr Lanka 40, Jan.B. & Paula Peare Estmaton o Expected Return: CAPM vs Fama and French. Internatonal Revew o Fnancal Analyss 14, Lntner, J., The valuaton o rsk assets and the selecton o rsky nvestments n stock portolos and captal budgets. Revew o Economc Statstcs 47, Maxmlano and Gonzalez, F., 2001.CAPM perormance n the Caracas Stock Exchange 10, D.A.I. Dayaratne, D.G Dharmaratne, S.A. Hars 81

15 Mossn, J., Equlbrum n a captal asset market, Econometrca 22, Sharpe, w., Captal asset prces: a theory o market equlbrum under condtons o rsk. Journal o Fnance 19, Stewart, G.B The Quest or Value: A Gude or Senor Managers, New York : Harper Busness. 82 Measurng the Rsk and Perormance