Macroeconomic Factors Influencing Foreign Remittances: The Case of Bangladesh

Transcription

1 Macroeconomic Facors Influencing Foreign Remiances: The Case of Bangladesh Mohammad Bayezid ALI Deparmen of Finance Jagannah Universiy Dhaka, Bangladesh ABSTRACT KEY WORDS JEL CODES This paper examines he influence of few seleced macroeconomic facors which are assumed o have influence on he flow of foreign remiances in Bangladesh. Yearly daa beween 1976 and 2011 have incorporaed and a mulivariae ime series regression model have been designed o esimae he absolue relaionship beween remiances and macroeconomic deerminans. I has found ha professional and unskilled migran have very low negaive bu significan regression coefficiens bu semi-skilled migrans have posiive significan coefficien. Among oher macroeconomic facors deposi ineres raes and foreign exchange raes are also found o have saisically significan coefficiens. Tes of Granger causaliy reveals ha foreign remiances have bi-direcional causaliy wih semi-skilled and un-skilled migrans and have uni-direcional causaliy from professional migrans o remiances and remiances o DSE all share price index. However, exensive invesigaion is very essenial o idenify oher macroeconomic facors ha migh cause he flow of foreign remiances in Bangladesh. Foreign Remiance, Migran, Macroeconomic Deerminans, Hos Counry, Granger Causaliy E24, E44, E66 1. Inroducion From he middle of 1970s Bangladesh has sen nearly 6.7 million manpower o more han 140 counries of he world. The ravel of his large number of domesic manpower o foreign counry has conribued o reduce he pressure of unemploymen in one hand, bu also open he gae of foreign exchange earnings on he oher. In addiion, hese Bangladeshi migran also conribue o pu in moion of he wheel of naional economy of Bangladesh. In , Bangladesh has earned close o 11 billion U.S dollars which is 11 percen of Naional Income and percen of oal expor earnings in Bangladesh. Ou of many sources of foreign remiances, remiance from Middle Eas counries consiues percen 1. These foreign remiances accelerae he growh of economic developmen, reduce he pressure of unemploymen, boos he capaciy o make essenial impor paymen and also help o increase consumpion, savings, and invesmen boh a micro and macro level. The dependen family members of he migran can fulfill heir basic need, can use he amoun o mee heir educaional expenses, purchase of land, houses and incur medical expenses by he use of foreign remiances. This amoun also increases 1 According o Sudy Repor iled, The Use of Remiances: Fuure Guidelines, Research Division of Bangladesh Bank, Principal Branch, Dhaka. 217

2 he social recogniion and sandard of living of he dependen family members of he migran in heir home counry. However, i is very obvious ha foreign remiances play a vial role boh a micro and macro level of he economy herefore i is very significan o idenify he facors ha can cause he inflow and growh of foreign remiances in Bangladesh. In his case, successful idenificaion also helps o formulae appropriae policy decision from he par of Bangladesh Governmen so ha he counry can reap he bes earnings from he inflow of foreign remiances Objecives The main objecive of his sudy is o idenify he esimae he influence of macroeconomic variables o he inflow of foreign remiances in Bangladesh. However, he broad objecives of his sudy are lised below: 1. Idenify he shor run relaive associaion beween foreign remiances and macroeconomic variables. 2. Esimae he exen of influence of differen macroeconomic variables over he inflow of foreign remiances. 3. Idenify any uni-direcional or bi-direcional causal relaion beween foreign remiances and seleced macroeconomic variables Deerminans of Migran Remiances Remiances have micro and macroeconomic deerminans. The main microeconomic deerminans are ones such as he characerisics of migrans, heir siuaion relaive o he home counry and relaive o heir families, income of heir family, and educaion level of he migrans. Macroeconomic deerminans are composed of deerminans such as wages in he hos counries, inflaion, exchange raes, economic condiions in boh home and desinaion counries. The macroeconomic effecs of migrans remiances became an imporan issue of analysis as a developmen ool. On he oher hand, he researchers sared o analyze more and more he macroeconomic deerminans of migrans remiances o see if counries of origin can use macroeconomic ools o arac more remiances Reasons for Supplying Remiances The migran s remiances firs depend on he levels of migrans savings in heir hos counry. Migrans can decide o remi or use heir savings in he hos counry. Empirically, many facors deermine his willingness o remi like he migran s family siuaion, he qualificaion level of he migran or duraion of say (OECD 2006). In he microeconomic heoreical lieraure, migrans are supposed o remi for individual reasons or wihin family arrangemens. a) Individual alruisic Individual moives are mainly pure or impure alruism (enlighen selfishness) and exchange moives beween he migrans and his recipien family in he counry of origin. In he case of pure alruisic moives, he uiliy of he migran depends only on he amoun of remiances he sends. In he case of impure alruisic moives, migrans send money back home in order o conribue o he income of heir families lef behind. Then he uiliy of he migran depends also on he income of his family in he counry of origin. The amoun of remiances should increase wih he migrans income, and decrease wih he amoun of he domesic income of he family. The duraion of say should have a negaive impac on he remiances because i is supposed ha he 218

3 aachmen o he family weakens gradually. Family unificaion has also he same effec as here are less people lef behind o look afer. b) Family arrangemen In he case of family arrangemens, we can disinguish hree ypes of moivaions: exchange, insurance and invesmen moives. These moives can also be called as empered alruism or enlighened self-ineres (Lucas and Sark, 1985). 1. Exchange moives The migran ransfers o he whole family in exchange of services offered for he welfare (i.e. healh, educaion ec.). Migrans are supposed o remi even if he family revenue increases because he qualiy of services heir remiances can buy increases. They can expec a beer consumpion, educaion, healh for heir children under he proecion of he large family recipien. 2. Insurance Insurance moive is based on inra-familial arrangemen agains income volailiy. I is a conracual arrangemen beween he migran and his family. In he rural areas of mos developing counries, where financial and assurance markes are incomplee, he revenues are subjec o risks such as drough, price flucuaions ec. o diversify he risk of rural income volailiy, families can decide o allocae some members o urban or foreign migraion. Alhough urban and foreign jobs are also subjeced o risks, hese risks are independen from he agriculural income variaions. A he beginning of he conac, family pays he migraion coss in exchange of fuure remiances. In he case of hese ypes of family conracs, remiances can flow o he family in case of agriculural income drops and o he migran in case of unemploymen. (Rapopor and Dcoquire, 2006). These kinds of arrangemens can also be seen wihin a village bu family is he mos frequen conex of such arrangemens. This can be explained by he muual alruism beween family members. However, as we are in a conracual agreemen conex, bargaining srengh of wo paries plays a role in he amoun of remiances. A high income level in he family increases is bargaining power. In he presence of alruisic moives, i is expeced ha lower-income households receive more remiances. Wihin a bargaining model, he reverse can be expeced because he bargaining srengh of he lower-income household would be smaller (Lucas and Sark, 1985). 3. Invesmen The hird one is he invesmen moives. In his case, he migran ransfers wih he objecive o ge a reurn on he family invesmen in he home counry for him and for his children like inheriance or sraegic behavior. The migrans can decide o inves heir savings in heir home counry as well as in heir hos counry. If he main moivaion o remi is o inves in he home counry, we can say ha invesmen moive dominaes he remiing decision of migrans. In his case, he migran calculaes his poenial reurn in his home counry relaive o his poenial reurn in he hos counry. The macroeconomic sabiliy in he home and hos counries and he ineres rae differenials deermine he remiing decision of he migran. Consequenly, he remiing decision is aken in he microeconomic level. A his sage of decision making, microeconomic facors like he family siuaion and he family income in he home counry, he duraion of say, he income level in he hos counry deermines he 219

4 remiances. Once he remiing decision is aken, macroeconomic facors deermine hoe much he migrans will remi for he insurance reasons, or how will he migran use is savings for he invesmen reasons. 2. Lieraure Review The lieraure on foreign remiances ends o be characerized by boh microeconomic and macroeconomic approaches (Rapopor and Docquier, 2003). A he micro analyic level, he emphasis usually is on an individual s moivaion o remi. The general moivaion, which have been highlighed are he alruisic moive (Lucas and Sark, 1985), he self, ineres moive, and he implici family conrac or loan repaymen (Durand e al, 1996; Haderi e al, 1999; Djajic, 1986; Quibaria, 1996; Russell, 1986; Taylor e al, 1996; Alleyne and Francis, 2003) and he implici family conrac as co-insurance (Massey and Parado, 1998; Soalimano, 2003). The macroeconomic lieraure has uilized many of hese moivaions o hypohesize he relaionships ha migh explain he growh in remiances. The various moives also deermine he uses o which remiances are pu. As a resul, while remiances can have a srong posiive impac on he curren accoun of he home counry s balance of paymens, hey may also have adverse economic effecs. On he posiive side, remiances conribue o household income and may help o alleviae povery. Naional income may also be enhanced due o increased foreign exchange inflows and he availabiliy of an increased pool of naional savings. I has been suggesed ha hese flow help o lif he balance of paymens consrain which many remiance receiving counries face. On he negaive side remiances may lead o inflaion hrough an increase in he money supply and an increase in demand beyond he home counries capaciy o mee his demand domesically. The balance of paymens may also be adversely affeced as a resul of increased impor demand due o he expanded money supply and any demonsraion effec, which may impac on hose households no in receip if remiances. Some auhors have poined o problems of he so-called Duch disease (Buch e al., 2002). The consumpion gap beween households ha receive remiances and hose ha do no can also be a significan conribuor o income inequaliy. Remiances may creae a dependency syndrome among he recipien populaion, wih aendan labor marke effecs. Among he sudies ha have examined he deerminans of remiances and/or ransfers a he macroeconomic level are Swami (1981), El-Saaka and McNaab (1999), Campbell (2003) and Haque e al. (1994). Typical of mos of hese analyses are single equaion models, which aemp o idenify he deermens of remiances. In he El-Saaka and McNaab (1999) sudy, i was concluded ha he key variables are economic aciviy in he hos and home counries, domesic prices, he unofficial exchange rae premium, and he differenial beween home and hos counries ineres raes. The expecaion is ha here is a posiive relaionship beween foreign income and remiances. In case of domesic income, he relaionship could be posiive or negaive. If a more depressed income in he receiving or home counry were accompanied by larger flow of remiances he relaionship would be negaive. This assumpion is in line wih he alruisic moive of he remier, who would be ineresed in mainaining he welfare of he family back home. Bu an increase in he income of he home counry may be accompanied by an increase in he value of remiances ransferred. This may reflec a real invesmen moive as rise in income may lead o opimism abou he performance of he domesic economy. Boh ypes of relaionships have been repored in he lieraure (Hysnbeggasi and Pozo, 2002). 220

5 The relaionship beween domesic price level and remiance flows is also uncerain. High raes of domesic inflaion may signal ha domesic economic policy lacks credibiliy. Remiers who are encouraged by he invesmen moive may wihhold remiances, who are moivaed by alruism may wish o send greaer values o mainain he welfare of households in he receiving counry. I is expeced ha higher unofficial exchange rae premium will be accompanied by lower values of remiances recorded as remiers may be more inclined o ransfer funds hrough unofficial channels. The differences beween he domesic and foreign ineres raes are used o capure porfolio effecs. In he El-Saaka and McNaab sudy, his relaionship was negaive implying ha remiers would be beer off keeping heir funds in he hos counry. Haung and Vargas-Silva (2005) ries o deermine wheher he hos and/or home counry macroeconomic condiions are he ones affecing remiances and hey find ha hos counry economic condiions seems o be he mos imporan facor driving remiances. Gupa (2005) analyze he macroeconomic facors ha migh explain he dynamics of remiances in India. His economeric analysis shows ha mos of he macroeconomic facors are insignifican in explaining he behavior of remiances around he rend over ime. One of he variables found o have effec on remiance behavior is he earnings of he migrans. On he oher hand he source counry economic aciviy also has an effec on remiances. For he Indian case, remiances are higher during periods of low economic growh. Aydas O.T. e al. (2005) has examined he facors ha deermine workers remiances in Turkey beween 1964 and Two differen regression models have been formulaed where workers remiances receip per workers and workers remiances flow are used as dependen variable. Differen facors like black marke premium, sock or workers abroad, per capial income, real overvaluaion, domesic inflaion, domesic growh and a dummy variable represening he period of miliary adminisraion has been used as independen variables. They have found ha black marke premium, real overvaluaion, per capia income, has significan relaion wih workers remiances receip per worker. Whereas black money premium, domesic inflaion, ineres rae differenial, miliary regime, growh, per capia income of boh domesic and hos counry are found o be significan wih workers remiance flow. Abdel-Rahman A.M.M, (2006) invesigaes he possible deerminans of foreign workers remiances in he Kingdom of Saudi Arabia (KSA). Here foreign remiances per worker has been used as dependen variable and real GDP, wages per worker, reurns and pariy condiion plus come composie indices peraining o socio-economic facors and o risk indicaors has been used as independen variables. The sudy reveals ha per capia GDP aciviy variable has a posiive relaionship o he levels of per worker remiances; wages also urn ou o be a significan deerminan facor. Differenial reurns variables have he expeced inverse relaionship wih he dependen variable, while he resul relaing o he various poliical, economic and financial risk variables indicae ha he variables measuring he degree of gov. sabiliy and he degree of law and order, have a significan impac on remiances Trend of Foreign Remiances in Bangladesh In case of foreign exchange earnings, Bangladesh holds an imporan posiion in he World Ranking. According o World Bank Ranking of Migraion and Remiances 2010, India holds he 1 s posiion followed by China, Mexico, Philippines, France, Germany, Bangladesh, Belgium, Spain, Nigeria, Pakisan, and. Srilanka. In his World Ranking of foreign remiances, he poion of 221

6 Bangladesh is sevenh. On he oher hand, Bangladesh holds he second posiion ou of six differen counries in Souh Asia. According o his ranking of remiances in Souh Asia 2010, India is he op remiances earner counry which is followed by Bangladesh, Pakisan, Srilanka, Nepal, and Maldives. On he oher hand he maximum migran people of Bangladesh lives in he Middle Eas counries like- Saudi Arabia, UAE, Qaar, Oman, and souh-eas counries like Malaysia and Singapore. According o he saisical record 2010 of Bangladesh Bureau of Manpower, Employmen and Training, he highes number of Bangladeshi migran lives in UAE followed by Oman, Singapore, Bahrain, Libya, Qaar, Saudi Arabia ec. All he naionalized and privae commercial banks are srongly conribuing o ransfer remiances from differen pars of he world. Hare commercial banks are considered as he major source for ransferring remiances o Bangladesh in legal process. According o he saisical daa of Foreign Exchange Division of Bangladesh Bank, percen foreign remiances has been colleced hrough privae commercial banks; percen by naionalized banks; 1.02 percen by specialized banks, and 1.92 percen by foreign banks. Table. 1 World Ranking of Foreign Remiances No. Name of he Counry Amoun of Foreign Remiances (Bil. USD) Number of Migrans (Mil.) 1 India China Mexico Philippines France Germany Bangladesh Belgium Spain Nigeria Pakisan Srilanka Source: worlbank.org/migraionandremiances, The role of foreign remiances in Bangladesh can be easily visualizes hrough he rend of foreign remiances in Bangladesh over ime. Figure 1 presen he bar diagram of foreign remiances beween 1976 and The size of he bar simply reveals ha he flow and growh of foreign remiances from 1976 o 1986 was no significan. Bu he growh of remiances gradually increases from 1988 o I is also found ha a sharp rise in foreign remiances akes place in 2003 which may due o favorable policy decision in manpower expor and good poliical and diplomaic relaion wih manpower exporing counries. Finally, we can conclude ha he flow and growh of foreign remiances in Bangladesh has been found beween 2004 and

7 3. Mehodology of Research This sudy ries o idenify he influence of macroeconomic variables o he flow of foreign remiances in Bangladesh. For his reason, foreign remiances (REMIT) has been considered as dependen variables and few seleced macroeconomic independen variables are - number of workers in professional caegory (PROF), skilled caegory (SKILD), semi-skilled caegory (SSKILD), un-skilled caegory (USKILD), consumer price index (CPI) (i.e. as a measure of inflaion), deposi ineres rae (DIR), Dhaka Sock Exchange (DSE) all share price index (DSI), foreign exchange rae agains USD (EXR) and GDP differenial beween Bangladesh and he average of en differen counries (GDPDIF). In calculaing GDPDIF, we have used simple average of GDP in en differen counries like KSA, UAE, UK, Kuwai, U.S.A, Qaar, Oman, Singapore, Bahrain, and Malaysia and hen find he difference beween GDP in Bangladesh and he simple average GDP in en differen counries. A he beginning, relaion associaion beween dependen and independen variables are esimaed hrough he use of correlaion esimaes. If hey are found o be significanly associaed wih REMIT hen all he variables will be esed for heir Saionariy. In his case wo popular uni roo es like Augmened Dickey-Fuller (ADF) Tes and Phillips-Perron (P-P) Tes will be employed. Based on he esimaes of uni roo es, a Mulivariae Time Series Regression Model based on Sandard OLS formula will be designed according o bes suied specificaion o measure he impac of seleced macroeconomic variables on he flow of foreign remiances in Bangladesh. In addiion residual diagnosic es was applied by examining Breusch-Godfrey Serial Correlaion LM Tes. The possible mulicollineariy has been examined by Tolerance Esimaes, Variance Inflaion Facor (VIF) and Collineariy Diagnosic Condiion Index. Regression coefficiens were esed for heir saisical significance a 5 percen by Wald Tes. A he end he causaliy relaionship beween REMIT and oher macroeconomic variables are compued by using he mehodology based on Granger (1969) Sample Size and Daa Sources This sudy incorporaes yearly daa from 1976 o 2011 (i.e. 36 observaions in each variable). In his case, foreign remiance (REMIT), number of workers in foreign counry in professional caegory (PROF), skilled caegory (SKILD), semi-skilled caegory (SSKILD), un-skilled caegory (USKILD), consumer price index (CPI), 3 o 6 monhs commercial banks deposi ineres rae (DIR), DSE all share price index (DSI) and foreign exchange rae agains USD (EXR) daa has been colleced from Bangladesh Economic review 2011 published by Minisry of Finance, Bangladesh. 223

8 Bu he GDP difference beween Bangladesh and he simple average GDP among en differen counries are colleced from differen publicaions and Fac Book issued by World Bank. 3.2 Model Specificaion This sudy uses Mulivariae Time Series Regression Analysis compued by Sandard OLS Formula o invesigae he influence of seleced macroeconomic variables (i.e. PROF, SKILD, SSKILD, USKILD, CPI, DIR, DSI, EXR, and GDPDIF) on he flow of foreign remiances in Bangladesh. Because of he exisence of a uni roo in all he variables under sudy (Table 3) we ransformed variables ino firs differenced daa. Then, he firs differenced daa for deposi ineres rae (DIR) and consumer price index (CPI) has been ransformed ino log value. Then he mulivariae regression model is developed by he following specificaion: D(REMIT) α 3 6 β D(GDPDIF( 1)) ε 9 1 β D(PROF( 1)) β D(SKILD( 1)) 1 β D(SSKILD( 2)) β D(USKILD( 2)) β Dlog(DIR( 1)) 4 β Dlog(cpi( 1)) β D(DSI( 1)) β D(EXR( 1)) (1) Table 2 presen he relaive associaion beween REMIT and all he seleced independen macroeconomic variables. I is found ha SKILD, SSKILD, USKILD, CPI, DIR, DSI, EXR, and GDPDIF have posiive and srong associaion wih REMIT. Bu PROF and DIR have negaive and weak associaion. Then we have esed he null hypohesis ha correlaion value is no significan beween REMIT and each of he independen macroeconomic variables. The -saisics wih heir respecive p-values reveals ha, we canno accep null hypohesis of no significan correlaion beween REMIT and oher independen macroeconomic variables a a convenional 5 percen significance level. Table. 2 Esimaing Correlaion beween REMIT and Macroeconomic Variables Variables REMIT Correlaion -Saisic Probabiliy REMIT PROF SKILD SSKILD USKILD CPI DIR DSI EXR GDPDIF Tes of Saionariy As mos of he macroeconomic variables are non-saionary in naure, so i is essenial o make i saionary before hey are be used in he regression esimaes. For his reason, all he seleced macroeconomic variables are esed for heir Saionariy. Here we have used Augmened Dickey-Fuller (ADF) Tes and Phillips-Perron (P-P) Tes o idenify he order of inegraion in each 224

9 macroeconomic variable. Table: 3 presen he esimaes of he order of inegraion in each macroeconomic variable. I is found ha all he variables are non-saionary I(I) process according o ADF es. On he oher hand, he same resul has been found in P-P es excep SKILD. Table. 3 Saionary Tess of Seleced Variables Variables Augmened Dickey-Fuller Tes Phillips-Perron Tes Order of Inegraion I(0) I(I) I(0) I(I) -sa prob -sa prob Adj -sa prob Adj -sa prob ADF PP REMIT I(I) I(I) PROF I(I) I(I) SKLID I(II) I(0) SSKILD I(I) I(I) USKILD I(I) I(I) DIR I(I) I(I) CPI I(I) I(I) DSI I(I) I(I) EXR I(I) I(I) GDPDIF I(I) I(I) 4.1. Mulivariae Time Series Regression Resul Table: 4 presen he mulivariae ime series regression esimaes beween REMIT and oher independen. This able also shows -saisic wih heir p-values agains each regression coefficiens. I has been found ha D(PROF(-1)) and D(SKILD(-1)) have very low negaive regression coefficiens and hey are no saisically significan a 5 percen. D(SSKILD(-2)) has also very low posiive regression coefficiens bu i is saisically significan. D(USKILD(-2)) has very low negaive coefficien and is saisically significan. On he oher hand, DLOG(DIR(-1)) and D(EXR(-1)) have very large posiive and negaive coefficiens respecively and hey are all saisically significan. Finally he coefficiens of DLOG(CPI(-1)), D(DSI(-1)) and D(GDPDIF(-1)) are also no significan in his regression model. The model produces R-Squared value of which implies ha approximaely 74 percen changes in he dependen variable (i.e. REMIT) can be explained by he changes in all he independen macroeconomic variables. Table 4. Tes Esimaes of Mulivariae Regression Analysis Variable Coefficien -Saisic Prob. C * D(PROF(-1)) D(SKILD(-1)) D(SSKILD(-2)) * D(USKILD(-2)) * DLOG(DIR(-1)) ** DLOG(CPI(-1)) D(DSI(-1)) D(EXR(-1)) * D(GDPDIF(-1)) Noe: * and ** represens significan a 5 percen and 10 percen respecively. 225

10 4.2. Tes of Mulicollineariy The exisence of mulicollineariy is a vial issue in applying Mulivariae Time Series Regression Model. Mulicollineariy is a siuaion when predicor variables in he regression model are highly inercorrelaed. The es for mulicollineariy is significan for he fac ha, i gives abnormal R-Squared value along wih spurious regression coefficien value wih large sandard error esimaes. In his sudy we have esimaed Tolerance Saisic Value, Variance Inflaion Facor and Coefficien Diagnosic Condiion Index for invesigaing he mulicollineariy problem in he regression model. Afer examining he es resul, i is found ha D(PROF(-1)), D(SKILD(-1)), DLOG(DIR(-1)), DLOG(CPI(-1)), D(DSI(-1)), D(EXR(-1)) and D(GDPDIF(-1)) have absoluely no mulicollineariy bu he residual wo variables i.e. D(SSKLID(-2)) and D(USKILD(-2)) have very low mulicollineariy problem (Table 5). Table 5. Tes of Collineariy Diagnosic Model Collineariy Saisics Tolerance Value VIF Condiion Index (Consan) D(PROF(-1)) D(SKILD(-1)) D(SSKILD(-2)) D(USKILD(-2)) DLOG(DIR(-1)) DLOG(CPI(-1)) D(DSI(-1)) D(EXR(-1)) D(GDPDIF(-1)) Tes of Serial Correlaion The presence of serial correlaion is examined by Breusch-Godfrey Serial Correlaion LM Tes. In Table: 6, The F-saisics and Obs*R squared value along wih heir respeced p-value indicaes ha he regression residuals were no serially correlaed a 5 percen significance level. Table 6. Tes of Serial Correlaion Breusch-Godfrey Serial Correlaion LM Tes: F-saisic Prob. F(3,9) Obs*R-squared Prob. Chi-Square(3) Tes of Saisical Significance of Regression Coefficiens Table: 7 presen he es esimaes of Wald es which is used o es he saisical significance of regression coefficiens in he regression model. Under null hypohesis, his es assumes ha he regression coefficiens are joinly equal o zero i.e. joinly regression coefficiens are no saisically significan. According o he value of F-saisics wih is respecive p- value implies ha we canno accep he null hypohesis a 5 percen significance level. Tha implies regression coefficiens are no joinly equal o zero. 226

11 Table 7. Wald Tess of Regression Coefficiens Tes Saisic Value df Probabiliy F-saisic (9, 12) Chi-square Tes of Heeroskedasiciy Heeroskedasiciy es has been used o examine he assumpions of homoskedasiciy in he residuals. According o his assumpion, residuals in he regression model mus have consan variances. If hey are no consan, he siuaion can be defined as heeroskedasiciy. In his sudy, we have esed he homoskedasiciy assumpion by applying Whie s General Heeroskedasiciy (WGH) Tes. Under null hypohesis, i assumes ha residuals are homoskedasic. The F-saisic and Obs*R-squared value wih heir p- values indicae ha we canno rejec null hypohesis of homoskedasiciy in he residuals. Tha means our model is free from heeroskedasiciy issues. Table 8. Tes of Heeroskedasiciy Heeroskedasiciy Tes: Whie F-saisic Prob. F(9,12) Obs*R-squared Prob. Chi-Square(9) Scaled explained SS Prob. Chi-Square(9) Tes of Granger Causaliy Table 9 presen he es esimaes of pair-wise Granger Causaliy beween REMIT and macroeconomic variables under sudy. Based on F-saisic wih heir p- values, i has been found ha PROF granger causes REMIT a 10 percen level which implies ha professional migrans in foreign counries can cause he flow of foreign remiances in Bangladesh. Bi-direcional causaliy has been found beween SSKILD and REMIT, and USKILD and REMIT a 5 percen level which implies ha boh he semi skilled and unskilled migran can cause foreign remiances and a he same ime remiances also causes hem. In addiion uni-direcional causaliy has been found from REMIT o DSI a 5 percen level which means he flow of foreign remiances can cause Dhaka Sock Exchange (DSE) all share price index. Bu his sudy failed o idenify any unidirecional or bidirecional causaliy beween REMIT and oher macroeconomic variables like SKILD, DIR, CPI, EXR and GDPDIF. Table 9. Tes Esimaes of Pair-Wise Granger Causaliy beween Foreign Remiances and Seleced Macroeconomic Variables Direcion of Causaliy F-Saisic Probabiliy PROF REMIT ** REMIT ~ PROF SKILD ~ REMIT REMIT ~ SKILD SSKILD REMIT * REMIT SSKILD * 227

12 Direcion of Causaliy F-Saisic Probabiliy USKILD REMIT * REMIT USKILD * DIR ~ REMIT REMIT ~ DIR CPI ~ REMIT REMIT ~ CPI DSI ~ REMIT REMIT DSI * EXR ~ REMIT REMIT ~ EXR GDPDIF ~ REMIT REMIT ~ GDPDIF Noe: (i)* and ** implies significan a 5 percen and 10 percen respecively (ii) (~) implies lack of any causal relaionship. ( ) shows he direcion of causal relaionship. 5. Conclusions This sudy ries o idenify macroeconomic deerminans of he flow of foreign remiances in Bangladesh. Few macroeconomic variables are incorporaed o deermine heir influence on he amoun of foreign remiances. Mulivariae ime series regression model has been used and he es resul reveals ha semi skilled and unskilled migran have very small bu significan regression coefficiens. Deposi ineres rae and foreign exchange rae of he home counry are also found o be significan in he regression model. On he oher hand, Granger Causaliy es reveal ha professional migran causes foreign remiances and a he same ime he amoun of foreign remiances also causes DSE all share price index. However, semi-skilled and un-skilled migran have bi-direcional causaliy wih he amoun of foreign remiances in Bangladesh. This sudy has obvious limiaions of no incorporaing few macroeconomic variables like Ineres rae differenials, unemploymen rae of he hos counry, economic sabiliy and per capia GDP of he home counry, diplomaic relaion wih he hos counry ec. if hey could be incorporaed, we could expec o have beer resul in he fuure. References 1. Abdel-Rahman A.M.M., (2006), The Deerminans of Foreign Worker Remiances in he Kingdom of Saudi Arabia, Journal of King Saud Universiy, Adminisraive Science, Vol. 18, No. 2, pp Bascom, Wilber, (1990), Remiance Inflows and Economic Developmen in Seleced Anglophone Caribbean Counries, Working Paper No. 58 of he commission for he Sudy of Inernaional Migraion and Cooperaive Economic Developmen, Washingon D.C. 3. Brown, Richard P.C. (1997), Esimaing Remiance Funcions for Pacific Island Migrans, World Developmen, Vol. 25, pp Buch, Caludia, M. e al. (2003), Workers Remiances and Capial Flows, Kiel Insiue of World Economics Duesernbrooker Weg Kiel (Germany), Kiel Working Paper No Campbell, Trevor, (2003), Deerminans of Foreign Inflows from Curren Transfer ino Barbados, Savings and Developmen, Issues 2/ Durnad, J. e al, (1996), Inernaional Migraion and Developmen in Mexican Communiies, Demography, Vol. 33, No. 32, pp El-Saaka, M.I.T, and Rober MvNaab, (1999), The Macroeconomic Deerminans of Migran Remiances, World Developmen, Vol. 27, pp

13 8. Glysos. P, (1988), Remiances in Temporary migraion: A Theoreical Model and Is Tesing wih Greek- German Experience, Welwirschafliches Archive Vol. 124, pp Gupa P., (2005), Macroeconomic Deerminans of Remiances: Evidence from India, Inernaional Moneary Fund, Working Paper No. WP/05/ Haung P., Vargas-Silva, (2005), Macroeconomic Deerminans of Workers Remiances: Hos Vs Home Counry s Economic Condiion, Econpapers: Inernaional Finance, econpa.wusl.edu 11. Lucas, R.E.B, and O. Sark, (1985), Moivaions o Remi: Evidence from Boswana, Journal of Poliical Economy, Vol. 93, no. 5, pp Messy, D.S, and.a. Parrado, (1998), Inernaional Migraion and Business Formaion in Mexico, Social Service Quarerly, Vol. 79, No. 1, pp Osman T. A. e al., (2003), Deerminans of Workers Remiances: The Case of Turkey, Emerging Markes Finance and Trade, M.E. Sharpe. Inc. Vol. 43, No. 3, pp Piorine, B., (1997), A Theory of Remiances as an Implici Family Loan Arrangemen, World Developmen, Vol. 25, No. 5, pp Rapopor Hillel and Fredrick Docquire, (2003), The Economics of Migran Remiances, (Chaper o appear in he Hand Book on he Economics of Reciprociy, Giving and Alruism, Forhcoming a Norh Holland). 16. Solimano, Andres, (2003), Workers Remiances o he Andean Region: Mechanisms, Coss and Developmen Impac, ECLAC, Prepared for he Mulilaeral Invesmen Fund- IDB s Conference on Remiances and Developmen, Quio-Ecuador. 17. Swamy, G., (1981), Inernaional Migran Worker s Remiances: Issues and Prospecs, World Developmen, Inernaional Migraion, Vol. 30, pp