IMPACT OF EXCHANGE RATE VOLATILITY ON KENYA S TEA EXPORTS

Transcription

1 Inernaonal Journal of Economcs, Commerce and Managemen Uned Kngdom Vol. II, Issue 12, Dec 2014 hp://jecm.co.uk/ ISSN IMPACT OF EXCHANGE RATE VOLATILITY ON KENYA S TEA EXPORTS Reuben Ruo Unversy of Kabanga, Deparmen of Economcs and Fnance, Kenya Alala Ondek Unversy of Kabanga, Deparmen of Economcs and Fnance, Kenya ondekalala@yahoo.com Absrac Ths paper sough o nvesgae he exen o whch exchange rae volaly affecs performance of ea expors. The objecve was o deermne he mpac of exchange rae volaly on ea expors, he conrbuon of ea expors earnng o Kenya s economy and draw polcy recommendaons emanang from emprcal fndngs for enhancng ea expors. Ths paper nvesgaed he behavor of exchange rae volaly on performance of ea expors usng me seres. Johansen and Julus Mulvarae conegraon echnque was appled o annual me seres daa for he perod of n order o recognze he shor run and long run behavor of he varables n he sudy. Conegraon and error correcon echnque (ECM) developed by Engle and Granger was used. Dckey fuller (DF) and Augmened Dckey Fuller (ADF) un roo es for saonary was employed n hs sudy. Phllps Perron (pp) on frs dfference was adoped o es saonary n her frs dfference and esng conegraon feasbly. The daa was sourced from cenral bank of Kenya, Kenya Naonal Bureau of Sascs, Tea Board of Kenya and he Inernaonal fnancal sascs of Inernaonal Moneary Fund (IMF). The resuls ndcae ha exchange rae volaly negavely affecs performance of ea expors n he counry. Ths paper recommends perodc monorng of he exchange rae so as o reduce s mpac and drawng of fscal and moneary polcy ha wll make exchange rae manageable. Keywords: Exchange rae volaly, Tea expors, Kenya, conergraon, mulvarae, fnancal sasc s Lcensed under Creave Common Page 1

2 Reuben & Alala INTRODUCTION Kenya reles heavly on he agrculural secor as he base for economc growh, employmen creaon and foregn exchange generaon. Expors of agrculural producs are of val mporance o he economes of less developed counres. Problems of employmen, balance of paymens, ncome dsrbuon, and prce and ncome sablzaon are hus drecly relaed o he agrculural expor secor. Ten percen of Kenya s populaon depends drecly or ndrecly on he Tea secor n erms of employmen and ncome. Kenya s ea s sold a Mombasa ea aucon under he auspces of he Eas Afrcan ea rade assocaon (EATTA) whose member shp consss of brokers, buyers, producers ware house men and packers. Over 84 % of Kenya s ea s sold a he aucon. In he world eas are offered a he aucon by brokers on behalf of he producers. They nve bds n US Dollars per klogram and knocked o he hghes bdder, ea bough from he aucon s manly expored. The exchange rae n Kenya shllng agans he US Dollar wll deermne he ea expor earnngs a ha parcular me. Agrculure secor performed well from ndependence accounng for 40% of GDP n Is performance n he recen year has been declnng enormously (24% of GDP n 2004) (sascal Absrac 2004). Despe he downward rend, he secor plays a leadng role n he economy. In 2004 agrculural secor n Kenya provded employmen for over 80% of he oal populaon and accouned for over 70% of foregn exchange earnngs manly from ea, coffee and horculure (Economc survey 2005). Tea ndusry s he larges employer n he agrculural secor and almos 10% of Kenya s populaon depends drecly or ndrecly on he ea ndusry. Tea and he Kenyan economy Despe he flucuaon of growh raes of he agrculural secor, ea has been conrbung favourably o he Kenyan economy n erms of value of expors whch are mporan n generang foregn exchange. Tea conrbues more han 20% of oal foregn exchange earnngs. To show hs, n 2002 ea conrbued 26% of he oal agrculural commody expor. In Kenyan economy he exchange rae s deermned by he Cenral bank of Kenya. Tea proceeds are n erms of US Dollars he exchange rae regme ha wll be appled wll deermne he foregn exchange earned from ea expors, hough oher facors lke foregn ncome of mporng counry, expor prces relave o world non fuel prmary commody prces of ea expors, ncome of mporng counres affecs ea expors. These are some of he challenges he ea secor expors have been facng and ha led o he need for hs sudy. Lcensed under Creave Common Page 2

3 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom Problem saemen An analyss of Kenyan expor performance shows ha he counry has been able o expand s expor volume o compensae for losses due o deerorang erms of rade. Though 26% of expor earnng sll accrue from expor of ea, here has been a declne n he value of domesc ea expor from Kenya, snce he year 2000, despe aggressve marke campagns by ea frms. Ths s arbued manly o he srengh of Kenyan shllng agans he US Dollar. In oher words, he apprecaon of he Kenyan shllng wll lead o low ea earnngs from expors. Low foregn exchange earnngs consrans mporaon of val raw maerals whch affecs ea qualy. In such a suaon, and gven ha ea s one of he man mansay of Kenya economy, an ncrease n ea earnngs s expeced o conrbue sgnfcanly o he allevaon of economc problems (unemploymen, balance of paymen, budge defc, unequal ncome dsrbuon, deb ssues). The mpac of exchange rae volaly on agrculural expors s well documened n leraure (Arze 1995); Arze, e.al. (2000, 2004); Baen, e. Al. (1984); and Cannon, e. Al.(2005), Accordng o a sudy by Adub (1999) exchange rae volaly has a sgnfcan negave effec on agrculural expors, Tea expor markes are volale and sgnfcanly nfluenced by many facors, expor prces ncome of mporng counres, and of parcular mporance he exchange rae volaly. Each of hese facors can have adverse effecs on ea expors. In hs sudy he auhor zeroed on ea expors before and afer economc lberalzaon snce many sudes on exchange volaly were done afer economc lberalzaon, he ncluson of he perod before economc lberalzaon s expeced o enrch he sudy. The queson ha begs answers, s o wha exend have hese facors especally exchange rae volaly nfluenced he volume, value of earnngs from ea expors. Objecves of he sudy The overall am of hs sudy was o examne he mpac of exchange rae volaly on he performance of Kenya s ea expors. The knowledge of he mpac of such volaly of exchange rae wll faclae n he deermnaon of an exchange rae regme ha favors ea expors and by exenson ncrease foregn exchange earnngs. Specfc objecves 1. To deermne he exsence of a funconal relaonshp beween he expors on one hand and a se of ndependen facors on he oher hand among hem beng exchange rae. 2. To deermne wheher here s a long run relaonshp beween exchange rae flucuaons and performance of ea expors. 3. To esmae he mpac of exchange rae volaly on performance of ea expors. 4. To draw polcy recommendaons for enhancng ea expors. Lcensed under Creave Common Page 3

4 Reuben & Alala Jusfcaon of he Sudy There are much wdenng dspares n ncome dsrbuon beween he rural and urban areas whch s a worryng phenomenon, and mus be adequaely deal wh. More han 60% of ea grown n Kenya s grown by small scale farmers who form he bulk of producers n he counry whch fall n he rural areas, (he naonal ea commee, bullen of sascs, 2000). Those wh small acreages can generae reasonable earnngs from ea. Tea s labor nensve and employmen can be acheved hrough producon of ea. These employmen opporunes wll depend on ea expors. Exchange raes flucuaon has negave effecs on agrculural expors. The more volale he exchange rae changes, he lower he ncome earnngs of farmers, whch subsequenly leads o a declne n oupu producon and a reducon n expor rade. An apprecaon of local currency decreases expors earnngs and vce versa. The mplcaon s ha moneary auhores should adop a mechansm ha wll lead o he sably of he exchange rae. Emprcal resuls of hs sudy wll enable us deermne he bes exchange rae regme favourable o he ea secor. The sudy covered Kenya s ea expors n relaon o monhly exchange raes from 1970 o 2008 usng Secondary daa, from CBK, IFS, KNBS and TBK. LITERATURE REVIEW Expors of agrculural producs have played a val role n economc growh of many developng counres. However he srucural adjusmen programmes of 1980 s dsruped he posve rend of foregn exchange earnngs derved from hese crops (Nzok, 2002). Ths secon surveyed he heorecal and emprcal ssues relevan o he objecves rased. Despe he predcon of purchasng power pary heory, s a well documened fac ha he real exchange rae beween counres flucuaes over me. Real flucuaon n relave prces of counres has varous mpacs on her economes and rade flows. A posve effec of a real deprecaon has been found a he frm and secoral level bu he effec a he aggregae level has been found o be posve and negave n dfferen counres (Marana Colacell, 2008). A larger number of recen sudes focus on effecs of exchange rae volaly Arze (1995), Arze e al (2004). The 2004 sudy by, Arze Osang and Sloje (2004) nvesgaed he mpac of real exchange rae volaly on he expor flows of egh Lan Amercan counres over he quarerly perod The resuls show ha ncreases n he volaly of he real effecve exchange rae exer a sgnfcan negave effec upon expor demand n boh he shor run and long run. The long run elasces range from a low of 0.10 n he Domncan Republc o a hgh of 0.69 n Venezuela, mplyng ha exchange rae volaly exers a sgnfcan adverse long run effec on expor volume. Lcensed under Creave Common Page 4

5 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom Cameron, Khangre and Pors (2005) nvesgaed he effecs of exchange rae varably on Uganda s ropcal freshwaer fsh expors. The emprcal evdence suggess ha Uganda s expors of fsh were negavely and sgnfcanly correlaed wh exchange rae volaly. Smlarly, Vergl (2002) nvesgaed he mpac of real exchange rae volaly on he expor flows of Turkey o he Uned Saes and s hree major radng parners n he European Unon for he perod 1990: : 12. The exchange rae volaly measures are negave for all counres and are sgnfcan a 1 % level for France and Germany and a 10 % for he U.S. The resuls obaned provde evdence ha he real exchange rae volaly has a sgnfcan long run effec on real expors. Frey (2005) also fnds sgnfcanly negave coeffcens for he exchange rae uncerany measure n he case of Canada, he Uned Kngdom and he Uned Saes. In he case of Canadan pork and lve swne expors, Fabosa (2002) fnds ha he volaly of he exchange rae has a negave mpac Kpu Kande (2005) suded he movemens of he Kenya shllng exchange rae and he argued ha he Kshs exchange rae has gone hrough varous cycles. The shllng deprecaed by 78.2% n January 1995, o Ocober 2000 followed by a perod of relave sably n Ocober 2000 o November Recenly however, he shllng has experenced a srong apprecaon. The large swngs n he shllng exchange rae are also assocaed wh varyng degrees of volaly. Volaly was hghes durng he perod jus afer lberalzaon, ha s, January 1995 o Ocober 2000 and lowes n he perod from Ocober 2000 o November Snce hen, however, he shllng experenced prolonged apprecaon n nomnal and real erms up o he end of The flucuaon n he exchange rae has araced publc aenon especally from exporers who have argued ha he srenghenng shllng s erodng her compeveness. Esmang he aggregae and ndvdual agrculural expor supply, Fosu (1992) noed ha RER of a domesc currency does no nfluence he economy s agrculural expors drecly; nsead, nfluences agrculural expor hrough s effecs on he ncenve srucure. In hs sudy a 5% level, he lagged expor varable urned ou o be he only sgnfcan varable n he aggregae mode. RESEARCH METHODOLOGY Research Desgn Ths sudy used a heorecal and emprcal approach based on case sudy and hsorcal desgns n Kenya. Ths sudy used he expor demand frame work o deermne he mpac of foregn rae volaly on he performance of ea expors. Lcensed under Creave Common Page 5

6 Reuben & Alala Model specfcaon In analyzng he commody expors he researchers adoped he radonal expor frame work ha was pu forward by Goldsen and Khan (1978) and has been used n several sudes for example, Chowdhury (1993), Arze (1995), Arze, Osong and Sloje (2000) and Kpu Moses (2008). Ths expor demand framework posulaes a long-run relaonshp beween expors, foregn economc acvy, relave prces and exchange rae volaly. The above frame s gven as follows: ln X Where 1 lny 2 ln P 3V X = Tea expors (1) Y = foregn ncome proxed by he ndusral producon ndex of ndusral counres. P = expor prces relave o world non fuel prmary commody prces and V = an error erm. = measure of rsk or uncerany gven by he 12- monhs movng average of he sandard devaon ( ) of absolue changes n he real effecve exchange rae. Several sudes have such a me varyng measure of exchange rae varably o accoun for perods of hgh and low exchange rae volaly (see Chowdhury, 1993:Lasrapes and koray, 1989; Kenen and Rodrk 1986 and Arze e al 2000) and defne volaly as he measure of rsk or uncerany gven by he 12- monh movng average of he sandard devaon of absolue changes n he real effecve exchange rae. The followng s he formulae for fndng volaly: V Where and m 1 2 ( ) ln R 1 ln R 2 V m (2) s exchange rae volaly, m (=12 n hs case) s he order of he movng average, R s exchange rae a a gven pon n me. uncerany. The role of exchange rae volaly s o consder he currency movemen effecs hrough The above expor demand frame work was modfed n hs sudy so as o nclude dummy varable (du) n he analyss whch are perceved o nfluence he dependen varable, before and afer lberalzaon. Lcensed under Creave Common Page 6

7 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom Hence, followng s he modfed model adoped n he sudy ln X 1 lny 2 ln P 3V du.(3) In hs sudy he Johansen mulvarae approach nroduced by Johansen (1988) s appled n order o esablsh he exsence of a long run or equlbrum relaonshp and he maxmum egenvalue sasc s used o es for he exsence of conegraon. Upon confrmaon of exsence of conegraon among he varables, an Error Correcon Model (ECM) s esmaed o capure he shor-run dynamcs. ECM correcs for dsequlbrum or s a means of reconclng he shor run behavor of an economc varable wh s long run behavor and s as follows. ln X 0 1ECT 1 ln P V ln X du. ln Y (4) Equaon (4) = Esmaed wh a fve lag srucure for volaly and foregn ncome proxe and one lag perod for he res of he varables all he varables, a consan erm and one perod lagged error correcon erm denoes frs dfferences X = expors ECT 1 = The lagged error correcon erm and s he resdual from he conegrang regresson. If here exs wo conegrang vecors, he wo error correcon erms wll be ncluded n he above error correcon model. ECT~I(0), capures he adjusmen owards he equlbrum. 0 = consan erm or nercep long-run 1 = he proporon of he dsequlbrum n he expor demand n one perod correced n he nex perod. = coeffcen of lagged expor demand = coeffcen of lagged foregn ncome proxed by he ndusral producon ndex of ndusral counres. = coeffcen of lagged expor prces relave o world non fuel prmary commody prces ω s he coeffcen of he dummy varable = coeffcen of lagged measure of rsk or uncerany gven by he 12- monhs movng average of he sandard devaon ( ) of absolue changes n he real effecve exchange rae. Lcensed under Creave Common Page 7

8 Reuben & Alala Du s he dummy varable ha capures for he perod before lberalzaon and afer lberalzaon. Where 0 before and 1 afer lberalzaon. The coeffcens,,,, ω show how he average speed of expor demand adjusmen may dffer dependng on wheher adjusmens s n responses o lag expor demand, foregn ncome proxed by he ndusral producon ndex of ndusral counres, expor prces relave o world non fuel prmary commody prces, measure of rsk or uncerany gven by he 12- monhs movng average of he sandard devaon ( ) of absolue changes n he real effecve exchange rae shocks. If for example he coeffcens > > > > δ mples a faser response of expor demand o lagged foregn ncome proxed by he ndusral producon ndex of ndusral counres han lagged expor demand o lagged expor prces relave o world non fuel prmary commody prces and lagged measure of rsk or uncerany gven by he 12- monhs movng average of he sandard devaon ( ) of absolue changes n he real effecve exchange rae and nflaon rae. Conegraon analyss Conegraon ess are conduced n case of non-saonary of he seres o ensure long run relaonshps. The long run equlbrum relaonshp among he varables was esed va Johansen (1988) and Johansen and Juselus (1990) approaches. The mehod s superor o he Engle-Granger (1987) wo-sep procedure n he esmaon of boh long-run relaonshps and Error Correcon Models (ECM), as s applcable n a mulvarae case ha mgh be lnked by more han one conegrang vecor. The Johansen and Juselus approach also deermnes he number of conegrang vecors and provdes esmaes of hese vecors ogeher wh esmaes of he adjusmen parameers. The es for he number of sgnfcan characersc roos of a marx s found va he Trace Sasc Tes. Afer deermnng he long run relaonshp beween exchange rae volaly and he explanaory varables, he shor run dynamcs of he relaonshps are examned. The conegraon regresson (Engle and Granger, 1987) s carred ou where he resdual obaned from he equaon of he lnear I (1) seres s aken as he vald error-correcon erm whch s hen bul no an error-correcon model (ECM). Before carryng ou he conegraon ess, we wll frs carry ou he un roo ess of he me seres properes of he concerned varables oulned n he model below: ln X 1 lny 2 ln P 3V du (5) Lcensed under Creave Common Page 8

9 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom In summary, he modelng sraegy o be adoped n hs sudy nvolves hree seps: a) Deermne he order of negraon of he varables by employng Dckey- Fuller (DF), Augmened Dckey-Fuller (ADF) and Phllps-Perron (1988) unroo ess; b) f he varables are negraed of he same order, we wll apply he Johansen Juselus (1990, 1992, 1994) maxmum lkelhood mehod of conegraon o oban he number of conegrang vecor(s); and c) If he varables are conegraed, we can specfy an error correcon model and esmae usng sandard mehods and dagnosc ess. The Daa Ths sudy employed secondary daa from varous sources, Expor volumes, prces of ea and real exchange raes were obaned from Cenral Bank of Kenya (CBK) expor volumes are n ones whle expor prces are n US Dollars. The real exchange raes are compued hrough a weghng process. The oher source of daa was he Inernaonal Fnancal Sascs (IFS) of he IMF. world non fuel commody prces whch ogeher wh expor prces were used o derve relave prces. Tme seres daa was analyzed usng mcrof4.0 ANALYSIS & FINDINGS The Augmened Dckey-Fuller es for un roo Before underakng he Granger-ype causaly es specfed under equaon (1), he sudy conduced a formal es o confrm he me seres properes. The researchers employed he augmened Dckey-Fuller (ADF) un roo procedure o es he level of negraon for he varables concerned. The varables n he seres where found o have a leas un roo beween hem (see able 1). The null hypohess of he seres beng non-saonary s acceped n levels, however from he resuls from he able s evden ha here s exs a un roo for he daa observed under he sudy. Lcensed under Creave Common Page 9

10 Reuben & Alala Table 1 Un roo ess for resduals *************************************************************************************************************************************************************************************************************************** Based on OLS regresson of LX on: C LY LP LV DU 39 observaons used for esmaon from 1970 o 2008 ************************************************************************************************************************************************************************************************************************** Tes Sasc LL AIC SBC HQC DF ADF(1) ADF(2) ADF(3) ADF(4) ***************************************************************************************************************************************************************************************************************************** 95% crcal value for he Dckey-Fuller sasc = LL = Maxmzed log-lkelhood AIC = Akake Informaon Creron SBC = Schwarz Bayesan Creron HQC = Hannan-Qunn Creron The Sably Tes In nerpreng he foregong long-run model was mplcly assumed ha he sample coeffcens remaned sable hroughou he perod. Inference drawn on he srengh of he full sample esmae mgh be nvald f happens ha he coeffcens were ndeed no sable. The plo of he CUSUM es and CUSUM of Squares es (Brown e al, 1975) show ha no errors were sascally sgnfcan over he sudy perod. Insably would have been shown by movemen of he resdue plo ousde he crcal lnes n any of he wo graphs below. Fgure 1 Cumulave Sum of Recursve Resdual The sragh lnes represen crcal bounds a 5% sgnfcance level Lcensed under Creave Common Page 10

11 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom 1.5 Fgure Plo of 2 Cumulave Sum Sum of Squares of Squares of Recursve Recursve Resduals Resdual The sragh lnes represen crcal bounds a 5% sgnfcance level From he above sably es, we conclude ha he sably of he long-run model s remarkable consderng he large number of mporan reforms underaken durng he 1980s and 1990s. Ths also ndcaes ha he model s well specfed. Conegraon Analyss The researchers proceeded o conduc he mulvarae conegraon es applyng he Johansen and Juselus (1990) maxmum lkelhood esmaon procedure. As he selecon of he correc order of ARDL s mporan n hs ype of examnaon, and gven he naure of he sample sze, lag order selecon by eher he Akake nformaon crera (AIC), or by he Schwarz Bayesan crera (SC) s recommended (Pesaran, 1997), hs sudy employed he Akake nformaon crera. The resuls from he conegraon analyss (Table 2) show ha when one lag s used (one lag s suffcen for daa usng annual observaon), he null hypohess of no conegraon (r<= 2) beween varables (exps-v ), s rejeced a 5 per cen or 10 per cen usng eher he race es or maxmum egenvalue es. Ths provdes evdence on he exsence of a leas one conegrang vecor n he model and herefore I conclude ha he varables exhb a long-run assocaon beween hem. Lcensed under Creave Common Page 11

12 Reuben & Alala Table 2 Conegraon resuls ****************************************************************************************************************************************************************************************************************************************************** Conegraon wh unresrced nerceps and resrced rends n he VAR Conegraon LR Tes Based on Maxmal Egenvalue of he Sochasc Marx ****************************************************************************************************************************************************************************************************************************************************** 39 observaons from 1970 o Order of VAR = 3. Ls of varables ncluded n he conegrang vecor: X P Y V Trend Ls of I(0) varables ncluded n he VAR du Ls of egenvalues n descendng order: ****************************************************************************************************************************************************************************************************************************************************** Null Alernave Sasc 95% Crcal Value 90% Crcal Value r = 0 r = r<= 1 r = r<= 2 r = r<= 3 r = r<= 4 r = ****************************************************************************************************************************************************************************************************************************************************** Use he above able o deermne r (he number of conegrang vecors). Conegraon wh unresrced nerceps and resrced rends n he VAR Conegraon LR Tes Based on Trace of he Sochasc Marx ****************************************************************************************************************************************************************************************************************************************************** 39 observaons from 1970 o Order of VAR = 3. Ls of varables ncluded n he conegrang vecor: X P Y V Trend Ls of I(0) varables ncluded n he VAR: OE DU Ls of egenvalues n descendng order: ****************************************************************************************************************************************************************************************************************************************************** Null Alernave Sasc 95% Crcal Value 90% Crcal Value r = 0 r>= r<= 1 r>= r<= 2 r>= r<= 3 r>= r<= 4 r = ****************************************************************************************************************************************************************************************************************************************************** Lcensed under Creave Common Page 12

13 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom Havng esablshed hs, he sudy hen proceeded o esmae an error-correcon model based on equaon (4) o nvesgae b-dreconal causaly beween expors and v Table 3 Error Correcon Represenaon for he Seleced ARDL Model ARDL(2) seleced based on Akake Informaon Creron *************************************************************************************************************************************************************************************************************************************************************************** Dependen varable s dx 39 observaons used for esmaon from 1970 o 2008 *************************************************************************************************************************************************************************************************************************************************************************** Regressor Coeffcen Sandard Error T-Rao[Prob] dx [.000] dc [.070] Dp [.096] dy [.003] dv [.008] ddu [.017] ecm(-1) [.003] *************************************************************************************************************************************************************************************************************************************************************************** ecm = X -3.87E+0.7*C *P *Y *V *du *************************************************************************************************************************************************************************************************************************************************************************** R-Squared R-Bar-Squared S.E. of Regresson F-sa. F( 7, 34) [.000] Mean of Dependen Varable S.D. of Dependen Varable Resdual Sum of Squares 2.18E+08 Equaon Log-lkelhood Akake Info. Creron Schwarz Bayesan Creron DW-sasc *************************************************************************************************************************************************************************************************************************************************************************** R-Squared and R-Bar-Squared measures refer o he dependen varable The lagged error correcon erm s negave, sgnfcan and he coeffcen s less han uny. Ths means ha he error correcon model s well specfed and also confrms our earler fndngs on he Conegraon of he varables. The es sascs are sasfacory. The goodness-of-f varable (R 2 ) show ha he exogenous varables accoun for 66% of he varaons n expors n he shor run. The DW sasc s 2.02 and larger han R 2, mplyng ha he regresson s no spurous. From he dagnosc es sascs, he null hypoheses of he absence of resdual auocorrelaon, normaly, msspecfcaon and heeroskedascy n he resdual canno be rejeced. Lcensed under Creave Common Page 13

14 Reuben & Alala As he varables are expressed n logarhmc form, he coeffcens are nerpreed as elasces. The error-correcon erm (ecm) s negave as expeced, and sgnfcan (hgh absolue -sasc). The srong sgnfcance renforces he argumen of he model varables beng conegraed. The adjusmen of he model o he prevous year s dsequlbrum s 82.1%. The regresson resuls should be nerpreed cauously. The mmedae prce effec (deprecaon of real exchange rae) s sascally nsgnfcan whle he same varable lagged fve perods s sgnfcan bu negave. Ths could be explaned by he fac ha adjusmens o prce response n he shor run are no lkely o be consderable. Noneheless, expor prces (boh curren and lagged) have a posve nfluence on volume of ea expors n he shor run. Surprsngly, foregn ncome has an unexpeced negave sgn. Ths could be explaned by he shfng markes for he Kenyan commodes, especally wh he rsng economc negraon and he declne n expors o he European Unon. Lberalzaon s found o have posve mpac on ea expors. The Granger causaly es Table 5 provdes he resuls of he error correcon model ha would be used for causaly es beween varables. The esmaed resuls show ha b-dreconal Granger causaly exss beween performance of ea expors and exchange rae volaly based on he F-sascs for he jon sgnfcance of he auoregressve erms. Ths observaon suppors prevous fnds by M C Kpu (2008), and Prof Ndungu e al (2006). Smlarly, he null hypohess exchange rae volaly does no Granger cause expors change s rejeced, The resuls shows ha he null hypohess Granger no-causaly from exchange rae volaly o expors growh can be rejeced a 10 percen level of sgnfcance based on he sascal sgnfcance of he error correcon erm. Fnally, he resuls ndcae a negave relaonshp beween exchange rae volaly and ea expors alhough hs s no sgnfcan when curren values are used, mplyng ha ea expors are no prce elasc. However he same varable lagged fve perods gves sascally sgnfcan resuls. These fndngs are n lne wh Prof Ndungu e al 2006 and Kpu (2008) who also found ou he same resuls. Dscusson of fndngs The sudy was amed a fndng ou he relaonshp beween exchange rae flucuaons and he performance of ea expors n Kenya. Ths sudy was movaed by he Qus for knowledge and he need o draw polcy recommendaon ha can be adoped. Lcensed under Creave Common Page 14

15 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom The sudy found he followng. Tha here exss a sgnfcan mpac of exchange rae volaly on performance of ea expors n Kenya, hs effec s also noed o be negave as s depced from he coeffcens of he regresson resuls. A 1% ncrease foregn exchange volaly leads o a decrease of 63% n ea expors. I s also seen ha here exs a srong relaonshp beween performance of ea expors and foregn ncome, he fndng s ha foregn ncomes have sgnfcan effec on performance of Tea expors mplyng ha ea expor growh can been drven by facors whch are beyond he conrol of local polcy makers. Ths mples ha exernal developmens are mporan n nfluencng performance of Tea expors. Though capal nflows are drven by boh push and pull facors, recen nflows o Kenya are consdered o be growh-relaed hus suggesng a role for pull facors. A 1% ncrease n foregn ncome leads o a 44% decrease n performance of ea expors. Hand n hand wh he above s he fndng ha relave prces have go sgnfcan nfluence n performance of ea expors. Ths nfluence s found o be posve where a 1% ncrease n relave prces leads o 60% ncrease n performance of ea expors. The dummy varable for lberalzaon s sascally sgnfcan supporng he vew ha lberalzaon has go effecs on nernaonal rade by creang lnkages herefore promong expors. CONCLUSIONS Ths s a mely sudy, gven he relavely poor performance of he Kenyan ea n he nernaonal marke and he recen seasonal flucuaons n he exchange rae and nflaon raes. Ths paper has herefore made recommendaons and polcy nervenon measures o be adoped n relaon o exchange raes However, Tea expors rends observed ndcae ha Kenya s share of global ea rade has been flucuang over me. Wh many counres mplemenng seps o lberalze busness, economc and marke envronmens o reduce barrers o foregn rade, he amoun of expors hey rade wh should ncrease. Ths sudy nvesgaed he shor-run and long-run causaly relaonshps beween performance ea expors and flucuaons n exchange rae and nflaon rae n Kenya, usng he Auoregressve Dsrbued Lag (ARDL) approach n he examnaon of a Granger ype es of causaly wh an error correcon. Moreover, he esmaon resuls sugges a causal lnkage from v o change n expors performance of Kenya s ea. A negave causal relaon from v o expors performance of ea s observed n he sudy, as ndcaed by he esmaed conegrang Lcensed under Creave Common Page 15

16 Reuben & Alala vecors. Fnally, he sascal sgnfcance of F-sascs for jon sgnfcance of auoregressve erms and/or he error erm mples a srong b-dreconal causaly beween exchange rae volaly and performance ea expors. Fnally, he fndng ha foregn ncomes and relave prces have sgnfcan effecs on performance of Tea expors mples ha ea expor growh could be drven by facors whch are beyond he conrol of local polcy makers. Ths mples ha exernal developmens are mporan n nfluencng performance of Tea expors. Though capal nflows are drven by boh push and pull facors, recen nflows o Kenya are consdered o be growh-relaed hus suggesng a role for pull facors. Ths s however an area for furher research especally he need o deermne wheher he source of exchange rae movemens deermnes s mpac on expors. Furhermore, gven he lmed daa avalable, he recen polcy reforms and rade and nvesmen lberalzaon may have helped o creae a more open rade envronmen whch would have oherwse posve effec on ea expors, however he sub secor have been affeced by he exchange rae flucuaons. The sudy fndngs suppor he vew ha exchange rae volaly affecs ea expors performance n Kenya negavely, however he mpac s nsgnfcan, meanng ha ea expors are no prce elasc, herefore callng for reasonable polcy formulaon o proec hese crucal secor consderng he much benefs accrued from hs secor. Recommendaons The resuls emphasze monorng of exchange rae volaly and adopng approprae moneary and fscal polces o ensure sably n exchange rae. Alhough exporers and polcy makers have been preoccuped wh recen seep exchange rae apprecaon, focus needs o shf o exchange rae volaly and suppor owards reforms.ths wll enable exporers o hedge agans exchange rae rsk such as developng forward and fuures marke whch represens long erm soluon. Polcy mplcaons Ths secon of he sudy draws polcy nervenon ha are amed a cushonng he negave effecs of exchange rae flucuaons n performance of ea expors n he counry as well as he effecs of nflaon and exernal ncomes. There s need for a monorng body ha should be mandaed o carry ou frequen and regular monorng of he flucuaons n exchange rae flucuaons for he purpose of drawng polcy recommendaons. Lcensed under Creave Common Page 16

17 Inernaonal Journal of Economcs, Commerce and Managemen, Uned Kngdom Hand n hand wh he above polcy nervenon s ha here should be n place proper moneary and fscal polces ha help o sablze exchange rae. These polces should be mplemened all hrough ou n he economy so as o lower he rsk. Lmaon of he sudy All he daa were based on secondary daa whch a mes may no a rue represenaon of he prmary daa f no auhencaed. REFERENCES Adub, A.A (1999); Prce exchange rae volaly and Ngera s agrculural rade flow a dynamc analyss. AERC Alemayehu, Geda (1999). Deermnans of aggregae prmary commody expor supply from Afrca: an economerc sudy. Ehopan Journal of Economcs, Volume VI, No. 1, Aprl. Arze, A.C. (1995); The Effecs of Exchange Rae Volaly on U.S. Expors: An Emprcal Invesgaon, Souhern Economc Journal, Vol. 62, No. 1, pp Arze, A.C; T.Osang and D.J. Sloje (2000); Exchange Rae Volaly and Foregn Trade: Evdence from Threen LDC s, Journal of Busness & Economcs Sascs, Vol. 18, No. 1, pp Arze, A.C.; T. Osang and D.J Sloje (2004); Exchange Rae Volaly n Lan Amerca and s Impac on Foregn Trade, College of busness and Technology, Texas A & M Unversy, Texas Baen, S.D. and M.T. Belonga (1984); The Recen Declne n Agrculural Expors: Is he Exchange Rae he Culpr?, Federal Reserve Bank of ST. LOUIS, Ocober. Behar, A. and L. Edwards (2005); Esmang Elasces of Demand and Supply For Souh Afrcan manufacured Expors Usng a Vecor Error Correcon Model, School of Economcs, Unversy of Cape Town. Cameron, S.; D Khangre and D. Pos (2005); Has Exchange Rae Volaly Reduced Ugandan Coffee Expor Earnngs? Bradford Cenre for Inernaonal Developmen (BCID), Unversy of Bradford, Bradford Dd7-IDP, U.K. Chowdhury, A.R (1993); Does exchange rae volaly depress rade flows? Evdence from error correcon models, The revew of economcs and sascs, Vol.75, No.4, pp Engle R.E and Granger, C.W.J (1987) Conegraon and Error Correcon: Represenaon, Esmaon and Tesng Economerca vol 55. Fosu, A. F. (1990). Expor composon and he mpac of expors on economc growh of developng economes. Economc Leers 34, pp IMF (2002); Exchange rae volaly and rade flow-some new evdence, Inernaonal moneary fund, Washngon, USA, May. Fabosa, J.F. (2002); Assessng he mpac of he exchange Rae and Is Volaly on Canadan pork and lve swne o he uned saes and Japan;, workng paper 02-WP 305,cener for agrculural and rural developmen, Lowa sae Unversy, Lowa, June. Frey, R. (2005); Exchange rae volaly and Inernaonal Trade Some GARCH Esmaons sress he mporance of rade dversfcaon, Inernaonal economcs, Rocksock Unversy, Germany Goldsen, M. and M.S Khan (1978); The supply and demand for expor: a Smulaneous Approach Revew of economcs and sascs, pp Gouws, A.R. (2005) A Souh Afrcan Expor Model, Unversy of Preora. Lcensed under Creave Common Page 17

18 Reuben & Alala Johansen, Soren (1988); Sascal Analyss of Conegrang Vecors, Journal of economc Dynamcs and Conrol, pp Johansen, S. and K. Juselus (1990),.Maxmum Lkelhood Esmaon and Inferences on Conegraon: Wh Applcaons o he Demand for Money.. Oxford Bullen of Economcs and Sascs, 52, I Kenen, P. and Rodrk (1986); Measurng and Analyzng he Effecs of Shor erm Volaly n REAL Exchange Raes, The Revew of Economcs and Sascs, 52 pp Kpu C. Moses, (2008) Does Exchange Rae Volaly Harm Expors? Emprcal Evdence from Kenya s Tea and Horculure Expors CSAE Conference, Oxford Unversy. Lasrapes, W.D., AND f. Koray (1990); Exchange Rae Volaly and U.S. Mullaeral Trade Flows, Journal of Macroeconomcs, 12pp Marana Colacell (2008), Expor Responses o Real Exchange Rae Flucuaons: Developmen Saus and Expored Good Effecs. Barnard College, Colomba Unversy. Nzok (2002), Exchange Rae pass-hrough n Kenya. Depermen of Economcs, Unversy of Narob. Pesaran. M.H and Y Shn(1999). An Auoregresve Dsrbued lag Modelng: Approach o Conegraon Analyss, In s sorm,(eds) Economercs and Economc Theory n 20 h cenuary : The Rannar Frsh Cenennal symposum, Cambrdge: Cambrdge Unversy press Vergl, H Exchange Rae Volaly n Turkey and s Effecs on Trade Flows, Journal of Economc and Socal Research 4 (1), Warr, P.G. and F.J Wollmer (2005) The Inernaonal Demand for Thaland s Rce Were, M; A. Geda; S.N. Karng; and N.S. Ndung u (2001). Kenya s exchange rae movemen n a lberalzed envronmen: an emprcal analyss. KIPPRA Dscusson Paper No. 10. Narob: Kenya Insue for Publc Polcy Research and Analyss. Lcensed under Creave Common Page 18