Cost and demand functions of electricity in Gambia from 1982 to 2007

Transcription

1 Afican Jounal of Business Managemen Vol. 5(4), pp , 8 Febuay, 0 Available online a hp:// ISSN Academic Jounals Full Lengh Reseach Pape Cos and demand funcions of eleciciy in Gambia fom 98 o 007 Bukhai M. S. Sillah Depamen of Economics, College of Business Adminisaion, King Saud Univesiy, Riyadh, Saudi Aabia. bsillah73@yahoo.com. Acceped 3 Novembe, 00 This pape agues ha an eleciciy demand should be esimaed simulaneously wih he supply. I hen esimaes he demand fo and he supply of he eleciciy in he Gambia using educed fom egessions and veco eo coecion mehods. The pape finds ha sysems of simulaneous equaions canno be simplified o educed fom egessions o saisfy he saisical equiemens, bu ahe he heoeical modeling equiemens deemine he choice of he saisical model. The veco eo coecion mehod incopoaing he heoeical esicions of he model is found o bee fi he daa han he educed fom egessions. Fom he esimaion esuls of his mehod, he eleciciy demand is found o be pice elasic and income elasic. The eleciciy demand is found o shink if he company chages an aveage pice highe han.3 imes of he pe capia GDP gowh ae; and since he demand is pice elasic, inceasing he eleciciy pice will esul in falling evenues fo he company. The eleciciy indusy, which hee efes o he naional eleciciy company, exhibis diseconomies of scale. The indusy is found inefficien, and failing o innovae and accumulae knowledge o enable i o expand oupu wih falling aveage uni cos. Wih he cuen opeaion, he expansion of oupu could be undeaken only wih inceasing aveage uni cos and hence inceasing eleciciy pice. Key wods: Eleciciy geneaion, economic gowh. INTRODUCTION Eleciciy is he enegy and engine of economic gowh. I is he enegy ha powes he indusial poducion. I is cosly and ime consuming o se up an eleciciy geneaing sysem. Bu once i is in place, i is expeced o expeience deceasing aveage coss as he oupu expands. Ove ime, he sysem is also expeced o innovae and make use of advances in echnology and knowledge. These leaning and expeiences being gained on he poducion shall enable he sysem o expand and poduce bee oupu han peviously due o he exisence of economies of scale and leaning effecs. Gambia has neve enjoyed an adequae supply of eleciciy in is hisoy; unme demand and consan losses have been he chaaceisic of he eleciciy geneaion. Befoe independence, eleciciy was known only o some govenmen headquaes. Afe he independence in 965, he govenmen copoaized a depamen o fom Public Uiliies Copoaion, and i opeaed he eleciciy indusy fom when he govenmen decided o pivaize is opeaions and mainenance o Managemen Sevices (MSG) Gambia Limied. The conac of MSG was eminaed in 995, and GUC was ansfomed o Naional Wae and Eleciciy Company (NAWEC) as a limied liabiliy company. I has seven povincial powe saions and coves 30% of Geae Banjul aea. Bu he supply is sill fa sho of he demand, which is esimaed o ange fom 5 o 30mw, and he pefomance is unsaisfacoy as i coninues o lose 40% of is poducion as unmeeed consumpion; and he eleciciy aiff a $0.8 pe kwh is exemely cosly fo he Gambian sandad, which has an aveage monhly eaning of $40. Thus, he company has no opion bu o lean and innovae o impove he pefomance. Does he cos sucue of NAWEC ell of economies of scale, leaning and innovaion? The answe o his quesion is one aemp his pape will make. The ohe aemp is o analyze he sucue of demand fo eleciciy in he Gambia. Gambians inceasingly puchase elecical appliances o hp://wow.gm/afica/gambia/aicle/007/7//he-enegy-seco-eleciciy- LPG-and-enewable-enegy.

2 Sillah 85 consume he enegy poduced by NAWEC, o by chemical baeies, geneaos and sola panels. While he indusial consumes ofen se up sand-by geneaos o complemen he NAWEC supply. I is no economical fo evey individual o opeae his /he own eleciciy geneaing sysem. if he consumes in he Gambia inceasingly demand o plan o demand high consumpion of enegy, i will be leaned ha hey ae willing o pay fo he enegy; and NAWEC, povided ha is cos sucue exhibis economies of scale, should be in posiion o incease oupu and o ake up he demand ha is inceasingly offeed by boh he households and he indusial consumes. LITERATURE REVIEW Seveal sudies have examined he chaaceisics and he deeminans of eleciciy demand. The sudies ea he demand as a funcion of income, own pice and ohe vaiables assumed o be elevan. The ohe vaiables can ange fom household size o plasma display panel TV s, Yoo e al. (007) and some include climae condiions, Hondoyannis (004). Eleciciy is found o be a basic necessiy of living, Walke (979); Ubongu (985); Silk and Jouz (997); Naayan and Smyh (005); Naayan, Smyh and Pasad (007); Louw, Conadie, Howells and Dekenah (008). Holedahl and Jouz (004) found he eleciciy demand o be uniay elasic in esponse o income changes; while in Geece eleciciy is a luxuy, Hondoyannis (004). Own pice is found insignifican, Ubongu (985) and Ziamba (008), which could be due o he pice disoions and pice measuemen eo ofen pevailing in eleciciy make. The eleciciy demand is found o be pice inelasic - Walke (979); Holedahl and Jouz (004); Hondoyannis (004); Naayan and Smyh (005); Yoo e al. (007); Naayan, Smyh and Pasad (007) - which is ypical fo eleciciy, given he fac ha i is has no close subsiues in sho un. Holedahl and Jouz (004); Hondoyannis (004) find he demand o be inelasic in boh he sho and he long uns; while Naayan e al. (007) has found i o be elasic in he long, which gives hope ha in he long un consumes ae able o adjus hei consumpion of eleciciy and swich o ohe enegy alenaives. These sudies have implicily assumed he eleciciy supply consan, which can be undesood in he coss secional daa, bu quesionable in he ime seies daa. Tha is, hey claim o obseve idenifiable and sable eleciciy demand, Hondoyannis (004), o shifing eleciciy demand, Silk and Jouz (997). Analyzing demand sepaaely assuming away is inedependency wih he supply could lead o biased esuls and conclusions. Ohe sudies of eleciciy make focused on he supply side assuming he demand consan, McDonald and Schaenhol (00); Abo (006) and Kahouli-Bahmi (009). The cuen pape aemps o fill in his gap by analyzing demand and supply of eleciciy simulaneously. The pape has pesened he model, modifying ha of Fische and Kaysen (96), in a way ha boh esimaes he effecs of leaning and scale, and educes he omied vaiable bias ha ofen aises when esimaing he leaning cuves. Kameshen and Poe (004) also use simulaneous equaion appoach, bu hei model does no incopoae leaning effecs. In addiion, he pape pionees his sudy in he conex of he Gambia, whee no such sudy has been undeaken. Thus, i povides evidence base, which is of high value o he policy makes in he couny. Modeling he demand fo and he cos of eleciciy Eleciciy demand funcion Demand fo eleciciy is a deived demand. This demand is fo he sevices of elecical machines, and duable elecical appliances. This demand changes when he use of hese machines and appliances changes o when he socks of hese machines and appliances vay hough new puchases, eiemens o eooling. This pape uses Fische and Kaysen (96) model o esimae he household eleciciy demand. The household elecical appliances ae named whie goods. The households demand eleciciy due o hei demand fo he sevices of he vaious socks of he whie goods. The socks of he whie goods ae measued in ems of he oal kilowa hous ha could be consumed if he appliances ae employed a hei nomal ae, Fische and Kaysen (96). This enails knowing he kilowas hou ha could be nomally consumed by each ype of he whie goods and hen add up ove he vaious whie goods. The summaion of he kilowas hou consumpion of he vaious Whie goods opeaed by household i gives us he sock of whie goods opeaed by household i. Leing W i o be he oal appliance sock fo household i a ime. Household i s demand fo eleciciy will depend on he aes of use of he sock. This elaionship is specified by Fische and kaysen (96) as follows: q = u w () i i i Whee, q = acual enegy consumpion of household i a ime, i u = ae of use of he whie goods socks by household I, i w i = oal socks of he whie goods and u is pice of i hypohesized o depend on pe capia income, I and he i he eleciciy, p i. Thus Equaion () is wien as, q i p = w () i I β i i

3 86 Af. J. Bus. Manage. This is a funcional fom of eleciciy demand poposed by Fische and Kaysen (96). and β ae pice and income elasiciy of demand, especively, ha is, he demand fo eleciciy depends on he pice of eleciciy, he household income and he socks of whie goods. I is a muliplicaive demand funcion ha shows ha p i yβ is i an index which when muliplied by he oal socks w i deemines he level of acual eleciciy consumed by household i. The model is specified fo coss-secional daa esimaion; while his pape aemps o esimae he eleciciy demand in he Gambia ove ime. Taking his ino accoun and aking he naual log of he vaiables we ge Equaion 3, q = p + β I + w (3) The socks of whie goods gow ove ime, and Fische and Kaysen posulae ha hey gow a a consan ae of γ pe cen pe yea. Tha is w = ex( γ ), o w w w = γ Thus, lagging Equaion (3) by one peiod, we ge Equaion 4; q = p + β I + w (4) and subacing i fom Equaion (3), we ge, q o, q q = w = γ + p w + β( I + I + ( p I ) p Equaion (5) is a fis diffeence opeao, and assuming υ is independenly and idenically disibued wih mean zeo and vaiance, he equaion can be esimaed using OLS. The pice of eleciciy poses a measuemen challenge. Pice is ofen offeed as pice blocks o he consumes; no one pice exiss. The blocks ae also faily consan ove ime causing he pice vaiable o be a consan, which can be confused wih he inecep em of he equaion. The sudy uses ime seies economeics, he obsevaions span ove a long peiod; hus, pices changes ae fequenly obseved. I also uses he q q = γ + ( p p ) + β( I I ) o + υ aveage aiff, P fo he secos in he economy o eplace p in he above equaion. Equaion (6) can be esimaed using OLS. I is a fis (5) ) (5) diffeence opeao ha gives he sho un muliplies of he household demand fo eleciciy. Upholding he assumpion on he pice seing elaion; hen he acual eleciciy consumpion q depends on he uni aveage cos and he pe capia income y. in he long un, q is q = A + A p + A + υ (6) 0 I Using he Koyck appoach of esimaing a long un equaion model and assuming ha he adjusmen pocess owads he equilibium follows his fom, ( q ) ( q = b y I ), whee b is he adjusmen coefficien; hen boh he sho un and long un muliplies can be esimaed and deived especively as follows: q o, = ba0 + ( b) q + ba y + ε + ba p q 0 + β q + β p + β3 (7) = β I + ε Whee β 0 = BA 0 ; β = (-b); β = ba ; β 3 = ba A s ae he long un muliplies and ba s ae he sho un muliplies. As ae deived afe he esimaion of Equaion (7). Leaning and cos funcions Leaning cuve expesses he elaionship beween he uni aveage coss and he cumulaive oupu. If a company innovaes, and is wokfoce accumulaes expeiences, he oupu will expand moe han befoe a he same given cos. The cumulaive oupu, which capues advances in knowledge, echnology and expeiences, will have negaive elaionship wih he uni aveage cos. This is specified, Bend (99), as follows: c i = c + ni + ui (8) Whee u is assumed o be independenly and idenically disibued wih mean zeo and vaiance. c i is he uni aveage cos fo Company I, which is NAWEC in his case, is ime seies obsevaions, and c is he iniial uni aveage cos and n i is he cumulaive oupu up o bu no including ime. Assuming he poducion of eleciciy follows a Cobb- Douglas funcion, following Bend (99), he sudy deives he uni aveage cos funcion ha conains infomaion on advances in echnology, economies of scale and euns o scale as follows:

4 Sillah 87 i. Naional Wae and Eleciciy Company (NAWEC) employs only wo inpus, labo and capial, which ae denoed hee as and especively. Labo inpus consis of all human esouces ha go ino poducing and faciliaing he poducion of eleciciy. Capial inpus consis of all non-human esouces ha go ino he poducion of eleciciy oupu. ii. Y is he eleciciy oupu, which is poduced using he echnology A fo combining and. is he echnology elasiciy of oupu. iii. The poducion funcion is, Y = A (9) whee; and ae inpu elasiciy of oupu, and + = indicaes he euns o scale. Fo NAWEC o have economies of scale, i should have inceasing euns o scale, f. iv. The inpu pices ae P and P fo and especively; hen budge consain of NAWEC wih C as he oal budge is, C = P + P (0) The poblem of NAWEC is o maximize Equaion (9) subjec o Equaion (0). Tha is, Max Y = A Subjec o: C = P + P Suppessing ime subscips fo simpliciy, he poblem is educed o maximizing a Langage funcion of; [ P + P C] Max L(,, λ) = A λ Assuming he soluion is unique, he fis condiions ae, L x = = A λp 0 () L x = = P + P C = 0 = A λp 0 () L λ (3) Dividing Equaion () by Equaion (), he sudy obains,. o = (4) P P =.. P P Subsiuing Equaion (4) in he oupu funcion, Equaion (9), he sudy obains, o = A = / A P P Y / / Y / / P / P (5) Subsiuing he value fo in equaion (5) ino equaion (4), he sudy obains he value fo, / Y / / / / / P P = A (6) And subsiuing he values fo and in Equaions (5) and (6) especively ino he budge Equaion (3), he sudy obains he following cos funcion: C / / / / = KA Y P P (7) Whee; / K = / / + / Taking he naual log of equaion (7) and adding he eo em u, hen he cos funcion o be esimaed using OSL is, C = β + β A + β Y + β P 0 + β P 4 + u 3 (8) Whee, β 0 = K, β = /, β = /, β 3 = /, β 4 = / K is he consan em, and A is he echnology. Time vaiable can be used o poxy fo he echnology, o fom he leaning cuve A is he cumulaive oupu vaiable and hus n can eplace A in Equaion (8). Bu he appeaance of inpu pices as egessos can complicae he esimaion esuls. Oupu is a egesso in he cos funcion; cos funcions ae adiionally defined o be a funcion of oupu. To make he cos funcion as a funcion of only oupu vaiables, following Bend (99), who assumes ha some pice index is a funcion of he inpu pices. Hee, he sudy assumes ha he consume pice index is a funcion of he inpu pices; hus, CPI = / P + / P So ha eal cos of he eleciciy C C =, and CPI C = C CPI ; hus, by subsiuing he values fo C and CPI, he sudy obains he eal cos o be, C = K + / A + / Y + / P + / P + u / P / P

5 88 Af. J. Bus. Manage. The pice vaiables will cancel ou. The vaiable A which epesens advances in knowledge and echnology can be eplaced wih he vaiable n fom he leaning cuve, whee n epesens he cumulaive oupu and capues he leaning, expeiences and advances in echnology. A and n ae diffeen measues of he same vaiable, and he sudy assumes ha A = n; and he above eal cos vaiable will look as below: C = K + / n + / Y + u (9) Fom he oal eal cos equaion (9), he eal aveage cos of he eleciciy is deived as he oal eal cos divided by he oupu: =, and c C Y Y c C =, which means, fom Equaion (9) ha he eal aveage cos is, c = K + / n + ( ) / Y + u, and OLS can be used o esimae his equaion as, c = K + A n + A Y + u (0) The sudy could no find daa on oal coss of eleciciy poducion, which ae ofen mixed wih ha of he wae and seweage, since he same company povides he hee sevices, and hee ae no clea sepaae cos accounings fo each sevice. Since i is a egulaed monopolis company, is pice will be popoional o is aveage cos, specifically, p = ρc, Whee, p is he aveage pice fo all he consumes a ime, and ρ is he consan of popoionaliy. Taking he naual log of his elaion and solving fo aveage cos, he sudy obains C = p ρ, and by subsiuing in Equaion (0), he sudy obains he following model: p = A0 + A n + A Y + u Whee, A0 = K + ρ, A = /, A = If euns o scale ae inceasing, will be geae han ; if euns o scale ae deceasing, will be lowe han, and if he euns o scale ae consan, will be and A will no be significanly diffeen fom zeo. Afe esimaing Equaion (), he euns o scale and economies of scale can be compued as follows: Reuns o scale,, ES, A ES =. A + = A +, while he economies of scale, This complees he mahemaical modeling of he cos of and he demand fo he eleciciy. In he nex secion, he pape discusses he naue and he souces of daa fo he esimaion of Equaion (7), he demand funcion, and Equaion (), he supply funcion. These wo funcions fom a sysem of equaions, q = β + q + p + I + β β β ε (7) 0 3 Quaniy demanded of eleciciy = f (las peiod quaniy demanded, pice, income) p = A + A n + A Y + u () 0 Pice of eleciciy = f (cumulaive oupu, cuen oupu). Spanos (990) saes ha he idenificaion and simulaneiy poblems associaed wih supply-demand model aises because available daa efe o quaniies ansaced and coesponding pices ove ime". Bu in Equaion () y is no he quaniy ansaced, he quaniy ansaced is q, which is acually poduced and puchased. Wheeas y is he oal oupu poduced ha includes he quaniy puchased and he unmeeed oupu loss including own consumpion. Thus, o ea he idenificaion and simulaneiy poblems in he model, Equaion () s cuen oupu, y, is eplaced wih q, he acual ansaced quaniy plus he unmeeed poducion. This modifies Equaion () as: p = A + A n + A ( q + um ) + u () 0 The educed foms ha esul afe solving Equaions () and (7) ogehe fo p and q values ae esimaed and examined fo he idenificaion of he sucual paamees. Then, i employs he VEC mehod o complemen he educed fom mehod (Spanos, 990). Daa Thee ae fou vaiables in his pape on which annual ime seies daa ae colleced fom 98 o 007. These ae consume pice index which is used o find eal pe capia income of GDP. Real pe capia GDP is used as a poxy fo he income vaiable in he eleciciy demand funcion. The sudy also collecs daa on he acual oal consumpion of eleciciy (eleciciy consumpion by households, govenmen and fims). NAWEC ofen has hee main sale pices, esidenial pice, business pice and hoel pice; he sudy aveaged hese pices o find he mean pice, p. The daa ae souced fom IMF couny saisical appendices of he Gambia and he annual epos 983/984 and 984/985 of he Gambia Uiliies Copoaion. RESULTS AND ANALYSES Table pesens he esimaion oupu of he educed fom fo he quaniy demanded, and he Table pesens ha of he pice vaiable. The educed fom esimaes, which ae deived as he esul of simulaneous soluion of he Equaions () and (7), do no fi he undelying daa

6 Sillah 89 Table. Reduced fom esimaes fo he dependen vaiable: LQ. Vaiable Coefficien Sd. Eo -Saisic Pobabiliy C LQ(-) LI LN LUM R Mean dependen va.064 Adjused R S.D. dependen va Log likelihood F-saisic Dubin-Wason sa.84 Pob(F-saisic) Whie Heeo Tes nr.38 Pobabiliy Ramsey RESET Tes Pobabiliy BG Seial coelaion LM es nr.55 Pobabiliy Table. Reduced fom esimaes fo he dependen vaiable: LP. Vaiable Coefficien Sd. eo -saisic Pob. C LQ(-) LI LN LUM R Mean dependen va Adjused R S.D. dependen va Log likelihood F-saisic Dubin-Wason sa Pob (F-saisic) Ramsey RESET Tes Pobabiliy Whie Heeo Tes nr Pobabiliy BG Seial coelaion LM es nr Pobabiliy on he eleciciy demand and supply in he Gambia. The coefficiens ae mosly insignifican, hough hee appea no seial coelaion o heeoscdasciy poblems o ende he -aios uneliable. The sucual slope coefficien paamees ae ove idenified, wheeas he sucual consan paamees canno be idenified. Thee explanaions can be given fo he esuls of he wo ables. The explanaoy vaiables have been found o be highly coelaed; he coelaion coefficiens exceed 90% beween he vaiables. No vaiable can be dopped as explained in ealie, he explanaoy vaiables have been heoeically inoduced, and hence hey ae elevan fo he model, and limied daa consain as ofen is he case in he developing counies could no allow us o expand he obsevaions. Anohe explanaion lies in he pesence of he lagged quaniy demanded vaiable as an explanaoy vaiable, which came abou as a esul of eleciciy demand modeling following Fische and Kaysen (96) leading o he fac ha no all he explanaoy vaiables in he educed foms ae exogenous, which violae he assumpions of educed fom egessions. Finally, he educed fom egessions in Tables and 3 uses he level of vaiables in he esimaion, i can be seen in Table ha he vaiables of he model ae mosly fis diffeence saionay; hus, he elaionships esimaed in he educed fom egessions ae spuious. The lae wo explanaions canno be avoided when esimaing demand and supply funcions wih he leaning effecs, and hey have endeed he educed fom esimaion uneliable. Thus, a sysem of simulaneous equaions can be easily educed o some egession equaions and hen esimaed and solved fo he sucual paamees, when acually some exogenous vaiables ae heoeically ielevan fo some endogenous vaiables, such as he case in his pape. The cumulaive poducion is ielevan fo he eleciciy demand; likewise, he income does no have o appea in he esimaion of he supply

7 90 Af. J. Bus. Manage. Table 3. Eo coecion esuls fo demand and cos funcions of eleciciy. Coinegaion esicions: B(, )=, B(, )=,B(, 5)=0, B(, )=B(, 5) B(, 4)=0, B(, 3)=0, A(5, )=0, A(4, )=0, A(3, )=0 Resicions idenify all coinegaing vecos LR es fo binding esicions (ank = ): Chi (5) Pobabiliy 0.07 Sandad eos in ( ) and - saisics in [ ] Coinegaing Eq: CoinEq CoinEq LQ(-) ( ) [.7564] LP(-) (5.3545) [ ] LI(-) (6.66) [ ] LN(-) (.3649) [-.5585] LUM(-) ( ) [.7564] C (5.358) (5.758) [ ] [ ] Eo coecion: D(LQ) D(LP) D(LI) D(LN) D(LUM) CoinEq (0.0047) ( ) ( ) ( ) ( ) [.950] [ ] [ NA] [ NA] [ NA] CoinEq (0.0685) (0.063) (0.009) (0.006) (0.045) [ ] [-.9448] [-.565] [ ] [ ] funcion of he eleciciy, he educed fom egessions ignoe hese facs. To emedy his siuaion, he sudy inoduces VAR fo he model and hen incopoae he esicions ha in he demand esimaion cumulaive oupu and he unmeeed oupu ae ielevan, and in he esimaion of he supply funcion he income vaiable does have o appea, and fuhe esicing ha he co- \efficien esimae of quaniy vaiable and ha of he unmeeed oupu ae equal in he supply funcion, which comes abou as a esul of eplacing he oal poducion, y, wih acual quaniy puchased, q, plus he unmeeed ou-\pu, um. The model is expeced o poduce wo coinegaing equaions, one o be idenified as he demand funcion and he ohe o be idenified as he cos funcion of he eleciciy. The following esicions ae imposed o idenify hese equaions:. In he co-inegaion equaion fo he demand funcion, he co-inegaing vecos ae nomalized by he coinegaing coefficien of he quaniy puchased; he pe

8 Sillah 9 capia GDP is eaed an exogenous vaiable, while he cumulaive eleciciy oupu and he unmeeed eleciciy oupu ae eaed ielevan.. To idenify he second co-inegaing equaion as he cos funcion of he eleciciy, he co-inegaing vecos ae nomalized by he coefficien of he aveage pice; he pe capia GDP is eaed ielevan, and coefficien of he quaniy puchased is equaed o he coefficien of he unmeeed oupu o fi he Equaion () modeling. These esicions poduce wo co-inegaing equaions, on wih no end and inecep and he ohe has inecep bu no end. The esicions in he model wih no end and no inecep ae ejeced a % significance level, is esimaion oupu ae epoed in Annex Table. The esicions of he model wih inecep and no end canno be ejeced a 5% significance level. Thus, he sudy epoduces in Table 3 he esimaion esuls of his model: Demand funcion qˆ = p I (7) Cos funcion: pˆ = n.337 q. 337lum () O pˆ = n.337 y () As he esuls in Table 3 illusae, he eleciciy demand is pice inelasic and income elasic. A % incease in he eleciciy pice leads, holding ohe hings consan, on aveage, o a 3.45% fall in he quaniy demanded of eleciciy; wheeas, a % incease in he income, holding ohe hings consan, on aveage, leads o 39.59% incease in he quaniy demanded of eleciciy. Thus, he eleciciy demand in he Gambia is pice elasic; he evenue will fall if he aveage pice inceases. I is income elasic and i is a luxuy fo he aveage Gambian. The oal eleciciy demand will be expansive and pofi geneaing as long as he pecenage pice incease is less han.3 imes of he pe capia GDP gowh ae. Wih a pojeced pe capia GDP gowh of 5% nex yea, he eleciciy company can incease is aveage eleciciy pice by 6.5% and a posiive incease in he evenue. The company should be howeve able o saisfy he expansion of demand. I should be able o expand is supply, innovae and consequenly educe is aveage uni cos. The income gowh is a majo consain on pofiable pice incease; he pojeced maximum pice incease of 6.5% canno land he company in pofi if i coninues o lose moe han 30% of is poducion o inefficiency. This inefficiency is clealy capued by he esimaed supply funcion. The esimaed faco of euns o scale,, is 0.097, which also gives a faco of economies of scale of minus.097. Thus, he company s opeaion exhibis deceasing euns o scale and diseconomies of scale. This implies ha ove his peiod of sudy, on aveage, he company has no innovaed and lean fom expeience; i has no be able o accumulae any useful knowledge o enable i o expand oupu ove ime and naionwide; he lile ha has been expanded has been a coesponding inceasing aveage coss, as he coefficien of he cumulaive oupu clealy illusaes. Chaging inceasingly high eleciciy pices o ecove he inefficien aveage coss canno be susained as pecenage pice incease exceeding.3 imes of he pe capia GDP gowh ae will esul in shinking eleciciy demand. In fac, he cuen esimaed demand funcion shows ha he company s opeaion is he pice elasic egion, whee pice inceases can only educe he evenue; i can only incease now by chaging lowe eleciciy pices. The company should e-sucue is opeaions and modenize is sysems o minimize he unmeeed poducion, which cuenly sands a 44% of is oal poducion. CONCLUSIONS AND POLICY RECOMMENDATIONS This pape has found ou ha sysems of simulaneous equaions canno be simply solved ino educed fom egessions fo esimaion puposes; he heoeical modeling should define he saisical modeling. The veco eo coecion mehod incopoaing he heoeical esicions has bee fi he daa han he educed fom egessions. The sudy finds he eleciciy demand o pice elasic and income elasic. The pape finds ha fo he eleciciy demand no o shink, he company should no chage an aveage pice highe han.3 imes of he pe capia GDP gowh ae. The demand is elasic, which implies ha company canno incease is evenue by fuhe inceasing he eleciciy pice; only educing he pice can esul in inceased evenue. These ae consains on he demand side, bu he majo consain of he eleciciy indusy lies on he poducion side, which is found o exhibi diseconomies of scale. Due o his inefficiency in he eleciciy poducion, he oupu expansion can be done only wih inceasing aveage coss and hence pices. Policy makes should ake dasic acions o e-sucue and e-enginee he company o hold and evese is inefficien opeaions, which is cuenly esponsible fo 44 % oupu wasage. ACKNOWLEDGEMENT The auho hanks he anonymous eviewes fo hei valuable commens and suggesions.

9 9 Af. J. Bus. Manage. REFERENCES Fische FM, Kaysen C (96). "a sudy in economeics: he demand fo eleciciy in he Unied Saes." Amsedam, Noh-Holland Holedahl P, Jouz FL (004). "Residenial eleciciy demand in Taiwan", Enegy Econ., 6: 0 4. Hondoyannis G (004). "Esimaing esidenial demand fo eleciciy in Geece." Enegy Econ., 6: Kameshen DR, Poe DV (004). "The demand fo esidenial, indusial and oal eleciciy " Enegy Econ., 6: Louw K, Conadie B, Howells M, Dekenah M (008). "Deeminans of eleciciy demand fo newly elecified low-income Afican households." Enegy Pol., 36: Naayan PK, Smyh R (005). "The esidenial demand fo eleciciy in Ausalia: an applicaion of he bounds esing appoach o coinegaion." Enegy Pol., 33: Naayan PK, Smyh R, Pasad A (007). "Eleciciy consumpion in G7 counies: a panel co-inegaion analysis of esidenial demand elasiciies." Enegy Pol., 35: Silk JI, Jouz F (997). "Sho and long un elasiciies in U.S. esidenial eleciciy demand: a co-inegaion appoach." Enegy Econ., 9: Spanos A (990). "The simulaneous equaions model evisied: saisical adequacy and idenificaion." J. Econ., 44: Ubongu RE (985). "Demand fo eleciciy in Nigeia: some empiical findings." Sociol. Econ. Plan. Sci., 9(5): Walke JM (979). "The esidenial demand fo eleciciy." Resou. Enegy., : Yoo S, Lee J, Kwak S (007), "esimaion of esidenial eleciciy demand funcion in Seoul by coecion fo sample selecion bias." Enegy Policy, 35: Ziamba E (008), "he demand fo esidenial eleciciy in Souh Afica." Enegy Pol., 36:

10 Sillah 93 ANNE Table. Veco eo coecion esuls fo demand and supply funcions of eleciciy (no end and no inecep). Veco eo coecion esimaes Sandad eos in ( ) and -saisics in [ ] Coinegaion Resicions: B(, )=, B(, )=, B(, 4)=0, B(, 5)=0 B(, 3)=0, A(3, )=0, A(4, )=0, A(5, )=0 B(, )=B(, 5) Convegence achieved afe 99 ieaions. Resicions idenify all coinegaing vecos LR es fo binding esicions (ank = ): Chi-squae(5) Pobabiliy Coinegaing Eq: CoinEq CoinEq LQ(-) (0.8388) [.9403] LP(-) (0.059) [ ] LI(-) (0.04) [ ] LN(-) (.7495) [-.063] LUM(-) (0.8388) [.9403] Eo coecion: D(LQ) D(LP) D(LI) D(LN) D(LUM) CoinEq (0.704) (0.3338) ( ) ( ) ( ) [ ] [ ] [ NA] [ NA] [ NA] CoinEq (0.0967) (0.060) (0.0045) (0.0030) (0.057) [-0.78] [ ] [ ] [ 0.60] [-.569]