Modelling Inernaional Tourism Demand and Uncerainy in Maldives and Seychelles: A Porfolio Approach By Riaz Shareef 1 and Michael McAleer 2 1 School of Accouning, Finance and Economics, Edih Cowan Universiy 2 School of Economics and Commerce, Universiy of Wesern Ausralia School of Accouning, Finance and Economics & FIMARC Working Paper Series Edih Cowan Universiy July 06 Working Paper 0605 Correspondence auhor: Riaz Shareef School of Accouning, Finance and Economics Faculy of Business and Law Edih Cowan Universiy 100 Joondalup Drive Joondalup WA 6027 Ausralia Tel: +61 8 6304 5870 Fax: +61 8 6304 5271 Email: r.shareef@ecu.edu.au
Absrac: Maldives and Seychelles in he Indian Ocean are small island ourism economies (SITEs), boh of which have relaively small populaions, erriorial sizes, land area and narrow producive bases. The wo SITEs are surrounded by vas ocean and have an overwhelming reliance on inernaional ourism for economic developmen. Variaions in inernaional ouris arrivals o hese 2 SITEs have been affeced by unanicipaed oil shocks, naural disasers, crime and global errorism, among ohers. An accurae assessmen of he variaions in inernaional ouris arrivals, paricularly he condiional volailiy, is essenial for policy and markeing purposes. The condiional mean and condiional variance of he weekly inernaional ouris arrivals o Maldives and Seychelles from 1 January 1994 o 31 December 2003 for he 5 main ouris source counries are modelled. Mulivariae models of uncerainy are esimaed and esed. An assessmen and inerpreaion of he esimaes are made for policy makers and our operaors o reach opimal decisions on he basis of a porfolio approach o inernaional ourism demand. The paper assesses 4 ses of counry spillover effecs beween Maldives and Seychelles, namely: (i) he own counry effecs for Maldives and Seychelles; (ii) he counry spillover effecs from he remaining four counries wihin each of Maldives and Seychelles; (iii) he own counry spillover effecs beween Maldives and Seychelles; and (iv) he cross-counry spillover effecs beween Maldives and Seychelles. The empirical resuls for boh Maldives and Seychelles are discussed in erms of each of hese componens. Keywords: Small island ourism economies; Weekly inernaional ouris arrivals; Uncerainy; Condiional volailiy; Counry spillover effecs Acknowledgemens: The firs auhor wishes o acknowledge he financial suppor of he School of Accouning, Finance and Economics, Edih Cowan Universiy. The second auhor is mos graeful for he financial suppor of he Ausralian Research Council.
1. Inroducion The sovereign archipelagos of Maldives and Seychelles in he Indian Ocean are small island ourism economies (SITEs), boh of which have small populaions and are geographically isolaed from he res of he world. These wo SITEs vary profoundly in heir erriorial size, oal land area, prospecs for self-reliance in economic developmen, and an overwhelming reliance on ourism as a source of expors. As a resul of imevarying effecs, such as oil shocks, naural disasers, ehnic conflics, crime, and he hrea of global errorism, among ohers, here have been dramaic changes in he arrivals of inernaional ouriss o hese wo counries. An examinaion of hese variaions in inernaional ourism demand, paricularly he condiional volailiy (or uncerainy) in inernaional ouris arrivals o Maldives and Seychelles, is essenial for policy analysis and ourism markeing purposes. Since Maldives and Seychelles are boh siuaed in he Indian Ocean and approximaely on he same laiudinal level, hey share similar geographical and environmenal characerisics. In Maldives, here are relaively many and smaller coral-based islands, as compared wih relaively few and bigger islands in Seychelles. These islands were formed millions of years ago when he earh's crus moved over a ho spo below i, causing submarine volcanoes o erup. A large proporion of ouriss visi Maldives on package ours and spend heir enire vacaion on one island and in one resor accommodaion, engaging mainly in scuba diving and waer spors. In Seychelles, here are relaively more free independen ravellers, here is self-caering ouris accommodaion, and ouriss evenly disribue heir vacaion in diving, waer spors and ouring he counry. Overall, here is some conras beween he ypes of ourism producs offered in hese 2 SITEs. However, due o he many similariies in he ourism producs offered by hese 2 SITEs, i is imporan for policy makers in hese 2 SITEs, and our operaors who sell holidays o hese 2 SITEs, o make an accurae assessmen of how he variaions in inernaional ouris arrivals from a paricular source counry o Maldives affec inernaional ouris arrivals from he same source counry o Seychelles. Such mulivariae analysis in inernaional ourism demand does no seem o have been underaken o dae. 1
This paper models he condiional mean and condiional variance of he weekly inernaional ouris arrivals o Maldives and Seychelles from 1 January 1994 o 31 December 2003 from he 5 main ouris source counries. A common consan condiional correlaion model, namely he symmeric vecor auoregressive moving averagegeneralized auoregressive condiional heeroscedasicy (VARMA-GARCH) model of Ling and McAleer (2003), is esimaed and esed. An assessmen and inerpreaion of he esimaes is made o enable policy makers and our operaors o reach opimal decisions on he basis of a porfolio approach o inernaional ourism demand. The paper also makes an assessmen of he counry spillover effecs beween Maldives and Seychelles. There are four ses of effecs ha need o be considered: (i) he own counry effecs for Maldives and Seychelles; (ii) he counry spillover effecs from he remaining four counries wihin each of Maldives and Seychelles; (iii) he own counry spillover effecs beween Maldives and Seychelles; and (iv) he cross-counry spillover effecs beween Maldives and Seychelles. The empirical resuls for boh Maldives and Seychelles will be discussed in erms of each of hese componens. The srucure of he paper is as follows. In Secion 2 an overview of he Maldivian and Seychellois economies are presened. A discussion of he VARMA-GARCH model is given in Secion 3. This is followed by an assessmen of he characerisics of weekly inernaional ouris arrivals daa for Maldives and Seychelles in Secion 4. An empirical examinaion and he implicaions of he resuls for policy and markeing purposes are suggesed in Secion 5. Some concluding remarks are given in Secion 6. 2. Overview of he Maldivian and Seychellois Economies 2.1. Maldives The Republic of Maldives was a former Briish proecorae, which became independen in 1965. I is an archipelago in he Indian Ocean, comprising 1,192 islands, of which 199 are inhabied and 87 designaed as ouris resor islands. The Exclusive Economic Zone of Maldives is 859,000 square kilomeres, and he aggregaed land area is roughly 290 square 2
kilomeres. The oal populaion of Maldives was 298,842 in he 2005 census, and is esimaed o have grown a 1.69 percen per annum over he period 2000 o 2005. In spie of he small size, limied naural resource base, small populaion and remoeness, Maldives has shown an impressive economic growh record, averaging over 8% per annum during he 9 years preceding he December 2004 sunami. This growh rae enabled Maldives o aain an esimaed per capia GDP of US$2,401 in 2004, which is considerably above average for a small island developing counry, which has an average per capia GDP of US$1,500. The engine of growh in Maldives has been he ourism indusry, accouning for one-fifh of GDP, a hird of fiscal revenue, and wo-hirds of gross foreign exchange earnings in recen years. The ourism indusry of Maldives is unique because i is based on he one-island, oneresor concep. Owing o his paricular feaure, he counry has become one of he mos popular holiday desinaions in he world. The firs ouriss o Maldives arrived from Ialy in 1972, comprising wriers, phoographers and journaliss. Since hen Maldives ourism has been growing rapidly, and in 2004 oal inernaional ouris arrivals reached 616,716. Neverheless, due o he December 2004 sunami he number of inernaional ouris arrivals declined by 36% o 395,320. Despie his caasrophic inciden, he ourism indusry has recovered subsanially by reaching he pre-sunami capaciy uilizaion raes wihou having o revise he price of he ourism produc offered. Tourism in Maldives is seasonal and is peak ouris season coincides wih he European winer monhs. Over 80% of ouriss o Maldives are Europeans, and he bigges emerging ourism marke is Russia. The main ourism source counries are Ialy, Germany, UK, Japan, France, Swizerland, Ausria, Neherlands, Spain, and Russia. While here are relaively few scheduled flighs o Maldives, he majoriy of inbound air raffic is non sop charer flighs from major European ciies. The fisheries secor remains he larges secor in erms of employmen, accouning for abou one-quarer of he labour force. I is sill an imporan source of foreign exchange earnings. Due o he high saliniy conen in he soil, agriculure coninues o play a minor role. The governmen, which employs abou 20 percen of he labour force, plays a 3
dominan role in he economy, boh in he producion process and hrough is regulaion of he economy. 2.2. Seychelles Since independence from he UK in 1976, per capia oupu in his Indian Ocean archipelago has expanded roughly seven imes, from US$1,000 per capia in 1976 o US$7,600 oday. GDP growh in 2001 was 3.3 per cen. Growh has been led by he ourism secor, which accouns for abou 13 per cen of GDP, employs abou 30 per cen of he labour force, and provides more han 70 per cen of foreign currency earnings. The vulnerabiliy of he ouris secor was illusraed by he sharp drop in 1991-92, mainly due o he Gulf War. Alhough he indusry has rebounded, he governmen recognizes he coninuing need for upgrading he secor in he face of siff inernaional compeiion. Touris arrivals, which are one of he main indicaors of vialiy in he secor, grew by 4.1 per cen in 2000. A srong markeing effor by he Seychelles Tourism Markeing Auhoriy and he inroducion of several new five-sar hoels seems o have spurred he growh. Officials hoped ha he planned new hoels and expanded airline service o he island would help offse he possibiliy of reduced global ravel following he evens of 11 Sepember 2001. In 2003, ourism earnings accouned for US$680 million and 122,000 visiors, comprising 82 per cen from UK, Ialy, France, Germany and Swizerland. Any decline in ourism quickly ranslaes ino a fall in GDP, a decline in foreign exchange receips, and budgeary difficulies. However, he counry s economy is exremely vulnerable o exernal shocks. Seychelles no only depends on ourism, bu i impors more han 90 per cen of is oal primary and secondary producion inpus. The manufacuring and consrucion secors, including indusrial fishing, accouned for abou 28.8 per cen of GDP. The public secor, comprising governmen and sae-owned enerprises, dominaes he economy in erms of employmen (wo-hirds of he labour force) and gross revenue. Public consumpion absorbs over one-hird of he gross GDP. Indusrial fishing in Seychelles, noably una fishing, is an increasingly significan facor in he economy. Recen changes in he climae have grealy affeced he una indusry due o widespread mobiliy of una schools. 4
In 1995, Seychelles saw he privaizaion of he Seychelles Tuna Canning Facory, 60 per cen of which was purchased by he American food company, Heinz Inc. Oher indusrial aciviies are limied o small scale manufacuring, paricularly agro-processing and impor subsiuion. Despie aemps o improve is agriculural base and emphasize locally manufacured producs and indigenous maerials, Seychelles coninues o impor 90 per cen of consumpion goods. The excepions are some fruis and vegeables, fish, poulry, pork, beer, cigarees, pain, and a few locally-made plasic iems. 3. Models of Condiional Volailiy The empirical analysis presened in his paper is based on Engle s (1982) developmen of ime-varying volailiy (or uncerainy), using he auoregressive condiional heeroscedasiciy (ARCH) model, and subsequen developmens associaed wih he ARCH family of models (see, for example, he review by Li, Ling and McAleer (2002)). Numerous heoreical developmens have been suggesed by Wong and Li (1997), Hoi, Chan and McAleer (2002), and Ling and McAleer (2002a, 2002b, 2003). In McAleer (2005), an exensive comparison of univariae and mulivariae condiional volailiy models, including a discussion of he regulariy condiions required for sensible empirical pracice, is presened. A common consan condiional correlaion model is he symmeric VARMA-GARCH model of Ling and McAleer (2003). This model allows he analysis of volailiy spillovers of inernaional ouris arrivals from a common ouris source counry across Maldives and Seychelles, and is esimaed using weekly inernaional ouris arrivals. Consider he following specificaion for weekly inernaional ouris arrivals, single ouris source counry, for eiher Maldives or Seychelles: y, from a ( ) y = E y I + ε, = 1,..., n 1 ε = Dη (1) 5
where y = ( y1,..., ym )' measures weekly inernaional ouris arrivals o Maldives and Seychelles; η = ( η1,..., ηm )' is a sequence of independenly and idenically disribued (iid) random vecors ha is obained from sandardising he shocks o weekly inernaional ouris arrivals, ε, using he sandardisaion 1/ 2 1/ 2 D = diag( h1,..., hm ), where h is condiioned on (ha is, deermined by) hisorical daa, as discussed below; I is he hisorical informaion a ime ha is available o ouriss, ouris service providers and policy makers; m ( = 10) is he number of weekly daa series, namely weekly inernaional ouris arrivals from 5 main ouris source counries common o Maldives and Seychelles, respecively; and = 1,...,522 weekly observaions during he period January 1994 o December 2003. Bollerslev s (1990) consan condiional correlaion (CCC) GARCH model assumes ha he condiional variance of he shocks o he 10 daa series i, i = 1,..., m, follows a univariae GARCH(r,s) process, ha is, h r s 2 i = ωi + αilεi l + βilhi l l= 1 l= 1 (2) where α il represens he ARCH effecs, or he shor run persisence of shocks (namely, an indicaion of he srengh of he shocks in he shor run) o ourism growh, and represens he GARCH effecs, or he conribuion of such shocks o long run persisence (namely, an indicaion of he srengh of he shocks in he long run). This model assumes he independence of condiional variances, and hence no spillovers in volailiy, across he 10 daa series. I is imporan o noe ha Γ is he marix of consan condiional correlaions of sandardized shocks o ourism growh, wih he ypical elemen of Γ being given by ρij = ρ ji for i, j = 1,..., m. Therefore, mulivariae effecs across he 10 daa series are deermined solely hrough he consan condiional correlaion marix. β il 6
As an exension of (2) o incorporae he effecs of shocks across he weekly inernaional ouris arrivals from a common ouris source counry o Maldives and Seychelles, and hence spillover effecs in uncerainy across he 10 daa series, i is necessary o define h i on he basis of pas informaion from ε i, ε j, h i and h for i, j = 1,..., m, i j. Thus, he mulivariae asymmeric VARMA-GARCH model of Ling and McAleer (2003) is defined by (3) and (4). Equaion (3) gives he mulivariae condiional mean, while he mulivariae condiional variance is given in (4): j Φ( L)( Y µ ) = Ψ ( L) ε (3) ε = Dη r s (4) = + lε + l l l= 1 l= 1 H W A B H where 1/ 2 1/ 2 D = diag( h1,..., hm ), H = ( h1,..., hm )', ε = ( ε,..., ε )' 2 2 1 m, and A l, and B l are marices wih ypical elemens α ij and β ij, respecively. The condiional mean in (3) is expressed as an ARMA process. However, for purposes of his sudy, he condiional mean for a weekly inernaional ouris arrivals series, i, is given as an AR(1) or AR(2) process. Therefore he condiional mean is esimaed such ha yi = θ0i + θ1 i yi 1 + θ2i yi 2 + εi, where θ 0i is he consan of he AR(1) process and θ 1i and θ 2i are he auoregressive coefficiens. The order of auoregression is deermined hough he Akaike Informaion Crierion and he Schwarz Bayesian Informaion Crierion. 4. Daa Characerisics In his paper, weekly inernaional ouris arrivals daa, provided by he Minisry of Tourism of Maldives and Naional Saisical Bureau of Seychelles for he five main European ouris source counries during he period 1 January 1994 o 31 December 2003, are examined. As shown in Table 1, he five main European ouris source counries for Maldives in descending order are Ialy, Germany, UK France and Swizerland, and 7
accouns for 65% of oal inernaional ouris arrivals during he sample period. For Seychelles in he same order he five main ouris source counries are France, Germany, Ialy, UK and Swizerland, which consiue 66% of oal inernaional ouris arrivals during he same period. An iniial assessmen of he respecive series for uni roo es for saionariy using he Phillips-Perron procedure, wih runcaed lags of order 5 for each of he en series in levels, rejecs he null hypohesis ha here is a uni roo in he series a he 1% level of significance. Visual examinaion of he en daa series reveals ha here is srong seasonaliy presened in European ouris arrivals o Maldives and Seychelles, where he peak ouris season overlaps wih he European winer. Furhermore, European ouris arrivals in Maldives show ha here are srong posiive rends, owing o he expansion of capaciy in he ourism indusry of Maldives. However, in he case of Seychelles, here are no srong rends presen in he daa. 5. Empirical Analysis The univariae VARMA(p,q)-GARCH(1,1) model is used o esimae he spillover effecs of weekly inernaional ouris arrivals over he period 1994-2003, for he five main European ouris source counries from and wihin Maldives and Seychelles. Tables 2 and 3 presen he empirical resuls for he differen condiional means, and also displays he spillover effecs for he respecive ime series. All he esimaes in his paper are obained using he EViews 4.1 economeric sofware package. The Bernd, Hall, Hall and Hausman (BHHH) (1974) algorihm has been used in mos cases, bu he Marquard algorihm is used when he BHHH algorihm does no converge. Several differen ses of iniial values have been used in each case, bu do no lead o subsanial differences in he esimaes. The asympoic and robus -raios (see Bollerslev and Wooldridge (1992) for he derivaion of he robus sandard errors) for he Quasi-Maximum Likelihood Esimaes (QMLE) are repored in Tables 2 and 3. There are 3 enries for each esimae, namely he coefficien (in bold), he Bollerslev-Wooldridge (1992) robus -raio, and he asympoic -raio. In general, he robus -raios are smaller in absolue value han heir asympoic counerpars. 8
In examining he counry spillover effecs beween Maldives and Seychelles, here are four ses of effecs ha need o be considered: (i) he own counry effecs for Maldives and Seychelles; (ii) he counry spillover effecs from he remaining four counries wihin each of Maldives and Seychelles; (iii) he own counry spillover effecs beween Maldives and Seychelles; and (iv) he cross-counry spillover effecs beween Maldives and Seychelles. The empirical resuls for boh Maldives and Seychelles will be discussed in erms of each of hese componens. 5.1. Maldives (i) Own-counry effecs The magniudes of he long run own-counry effecs are greaer han he shor run counry effecs. The shor run and long run own counry effecs of he 5 main European ouris source counries o Maldives are generally very reasonable and are saisically significan, excep for he long run own counry effec of Ialy. (ii) Counry spillover effecs from four counries There is lile evidence o sugges ha here are counry spillover effecs from he remaining four counries wihin Maldives. However, he esimaes are generally reasonable. (iii) Own-counry spillover effecs The own-counry spillover effecs of weekly ouris arrivals from he same 5 source counries in Seychelles ha affec ouris arrivals in Maldives are mixed, and here are some unreasonable esimaes. However, he shor run own counry spillover effec of German ouris arrivals in Seychelles is wice ha of he own effec of German ouris arrivals o Maldives. Conversely, he shor run cross counry spillover effec of French ouris arrivals in Seychelles is half ha of he own counry effec of French ouris arrivals 9
o Maldives. These resuls are indicaive of he srong influence of German and French ouris arrivals o hese wo SITEs in he shor erm. (iv) Cross-counry spillover effecs There is evidence o sugges ha here are cross counry spillover effecs from Seychelles o Maldives, and vice-versa. Overall, he spillover effecs from Seychelles o Maldives are greaer han he spillover effecs from Maldives o Seychelles. 5.2. Seychelles (i) Own-counry effecs In he case of Seychelles, he absolue values of he long run own counry effecs are greaer han he shor run own counry effecs. Moreover, he shor run and long run own counry effecs of he 5 main European ouris source counries o Seychelles are also saisfacory and saisically significan. (ii) Counry spillover effecs from four counries In Seychelles here are no many counry spillover effecs from he remaining four counries. Neverheless, he esimaes are of a reasonable order of magniude. (iii) Own-counry spillover effecs The own counry spillover effecs of weekly ouris arrivals from he same 5 source counries in Maldives ha affec ouris arrivals in he Seychelles are mixed, and he orders of magniude of some of he esimaes are unsaisfacory. However, he shor run and long run own counry spillover effec of German and Briish ouris arrivals in Seychelles is saisically significan. Furhermore, he long run own counry spillover effec of Swizerland is also saisically significan. These resuls are indicaive of he srong influence of German and French ouris arrivals, boh in he shor and long run, and Briish ouris arrivals in he long run, o hese wo SITEs. 10
(iv) Cross-counry spillover effecs The esimaes for he cross-counry spillover effecs are mixed, and i is reasonable o sugges ha here is lile or no spillover effec from Maldives o Seychelles. In general, he spillover effecs from Maldives o Seychelles are greaer han he own counry spillover effecs of Seychelles. 6. Conclusion Maldives and Seychelles are SITEs in he Indian Ocean wih very similar climaic characerisics. Tourism is he principal economic aciviy as a proporion of heir expors, and hence is he key foreign exchange earner in hese wo economies. There are many similariies in he ourism producs offered by hese 2 SITEs. Variaions in inernaional ouris arrivals due o exogenous shocks ha are beyond he conrol of hese wo economies have serious ramificaions for every secor in hese wo SITEs. An examinaion of he variaions in inernaional ourism demand, paricularly he condiional volailiy (or uncerainy) in inernaional ouris arrivals o Maldives and Seychelles, is essenial for policy analysis and ourism markeing purposes. I is imporan for policy makers in hese 2 SITEs and our operaors who sell holidays o hese 2 SITEs o make an accurae assessmen abou how he variaions in inernaional ouris arrivals from a paricular source counry o Maldives affec inernaional ouris arrivals from he same source counry o Seychelles. Such a mulivariae analysis in inernaional ourism demand does no seem o have been underaken o dae. The paper assessed he counry spillover effecs of weekly inernaional ouris arrivals beween Maldives and Seychelles in erms of he own counry effecs, he counry spillover effecs from he remaining four counries, he own counry spillover effecs, and he crosscounry spillover effecs. Of he four counry spillover effecs, he mos imporan resuls for policy and markeing purposes are he own-counry and cross-counry spillover effecs. The empirical resuls indicaed ha here was srong influence of German and French weekly ouris arrivals, 11
boh in he shor and long run, and paricularly weekly Briish ouris arrivals in he long run, o hese wo SITEs. The esimaes for he cross-counry spillover effecs were mixed and were reasonable, which sugges ha here were few or no spillover effecs from Maldives o Seychelles. Overall, he spillover effecs from Seychelles o Maldives were greaer han he spillover effecs from Maldives o Seychelles. This suggess ha variaions in weekly inernaional ouris arrivals from Ialy, Germany, UK, Japan, France, and Swizerland o Seychelles affec variaions in inernaional ouris arrivals o Maldives. However, he resuls do no indicae he direcions in which he variaions of weekly inernaional ouris arrivals occur. For such an assessmen, his research will be exended in fuure o incorporae he mulivariae asymmeric VARMA-AGARCH model of Hoi, Chan and McAleer (2002). 12
References Bernd, E.K., B.H. Hall, R.E. Hall and J.A. Hausman (1974), Esimaion and inference in nonlinear srucural models, Annals of Economic and Social Measuremen, 3, 653-665. Bollerslev, T. (1990), Modelling he coherence in shor-run nominal exchange rae: A mulivariae generalized ARCH approach, Review of Economics and Saisics, 72, 498-505. Bollerslev, T. and J.M. Wooldridge (1992), Quasi-maximum likelihood esimaion and inference in dynamic models wih ime-varying covariances, Economeric Reviews, 11, 143-173. Engle, R.F. (1982), Auoregressive condiional heeroscedasiciy wih esimaes of he variance of Unied Kingdom inflaion, Economerica, 50, 987-1007. Hoi, S., F. Chan and M. McAleer (2002), Srucure and asympoic heory for mulivariae asymmeric volailiy: Empirical evidence for counry risk raings, paper presened o he Ausralasian Meeing of he Economeric Sociey, Brisbane, July 2002. Hoi, S., M. McAleer and R. Shareef (2005), Modelling counry risk and uncerainy in small island ourism economies, Tourism Economics, 11(2), 159-183. Hoi, S., M. McAleer and R. Shareef (2005), Modelling mulivariae volailiy in inernaional ourism demand and counry risk for Cyprus and Mala, unpublished paper, School of Economics and Commerce, Universiy of Wesern Ausralia. Li, W.K., S. Ling and M. McAleer (2002), Recen heoreical resuls for ime series models wih GARCH errors, Journal of Economic Surveys, 16, 245-269. Reprined in M. McAleer and L. Oxley (eds.), Conribuions o Financial Economerics: Theoreical and Pracical Issues, Blackwell, Oxford, 2002, pp. 9-33. Ling, S. and M. McAleer (2002a), Necessary and sufficien momen condiions for he GARCH(r,s) and asymmeric power GARCH(r,s) models, Economeric Theory, 18, 722-729. Ling, S. and M. McAleer (2002b), Saionariy and he exisence of momens of a family of GARCH processes, Journal of Economerics, 106, 109-117. Ling, S. and M. McAleer (2003), Asympoic heory for a vecor ARMA-GARCH model, Economeric Theory, 19, 278-308. McAleer, M. (2005), Auomaed inference and learning in modeling financial volailiy, Economeric Theory, 21, 232-261. 13
Shareef, R. and S. Hoi (2005), Small island ourism economies and counry risk raings, Mahemaics and Compuers in Simulaion, 68(5-6), 557-570. Shareef, R. and M. McAleer (2005), Modelling inernaional ourism demand and volailiy in small island ourism economies, Inernaional Journal of Tourism Research, 7, 313-333. Wong, H. and W.K. Li (1997), On a mulivariae condiional heeroscedasiciy model, Biomerika, 4, 111-123. 14
Figure 1: Weekly European Touris Arrivals o Maldives and Seychelles, 1994-2003 1600 French Arrivals o he Maldives 2000 French Arrivals o Seychelles 1200 1600 1200 800 800 400 400 0 94 95 96 97 98 99 00 01 02 03 0 94 95 96 97 98 99 00 01 02 03 2500 German Arrivals o he Maldives 1400 German Arrivals o Seychelles 1200 2000 1000 800 1500 600 1000 400 200 500 94 95 96 97 98 99 00 01 02 03 0 94 95 96 97 98 99 00 01 02 03 6000 Ialian Arrivals o he Maldives 1200 Ialian Arrivals o Seychelles 5000 1000 4000 800 3000 600 2000 400 1000 200 0 94 95 96 97 98 99 00 01 02 03 0 94 95 96 97 98 99 00 01 02 03 1200 Swiss Arrivals o he Maldives 300 Swiss Arrivals o Seychelles 1000 250 800 200 600 150 400 100 200 50 0 94 95 96 97 98 99 00 01 02 03 0 94 95 96 97 98 99 00 01 02 03 3000 2500 Briish Arrivals o he Maldives 1200 Briish Arrivals o Seychelles 1000 2000 800 1500 600 1000 400 500 200 0 94 95 96 97 98 99 00 01 02 03 0 94 95 96 97 98 99 00 01 02 03 15
Table 1: Composiion of Touriss o Maldives and Seychelles, 1994-2003 Touris Source MALDIVES Head Coun Proporion Touris Source SEYCHELLES Head Coun Proporion 1 Ialy 983,433 20.85 France 255,379 20.29 2 Germany 803,420 17.03 Germany 185,286 14.72 3 UK 717,492 15.21 Ialy 177,795 14.12 4 Japan 428,313 9.08 UK 172,757 13.72 5 France 284,794 6.04 Swizerland 51,075 4.06 6 Swizerland 266,497 5.65 Souh Africa 48,302 3.84 7 Ausria 131,383 2.78 Spain 36,460 2.90 8 Neher. 66,650 1.41 Scandinavia 31,815 2.53 9 Spain 57,051 1.21 Reunion 28,715 2.28 10 Russia 67,071 1.42 Mauriius 26,070 2.07 Toal 4,717,744 80.68 Toal 1,258,857 80.52 16
Touris Source Condiional Mean Table 2: Spillover Effecs wihin Maldives and from Seychelles C O N D I T I O N A L V A R I A N C E Own Effecs Spillovers wihin Maldives Spillovers from Seychelles France AR(1) AR(2) FR FR GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M FR_S FR_S GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S 0.71 0.21-2,105.63 0.08 0.83-4.7E-03-0.01-4.7E-03-4.7E-03 0.02 0.23 0.01-0.02 0.01-0.02-0.02 0.01 0.21-0.17 0.26 2.44 0.01-0.01 13.51 4.12-0.46 3.20 27.12-1.98-1.78-0.38-0.09 0.42 2.03 1.02-4.25 8.53-7.22-8.18 0.75 4.27-4.03 1.32 0.56 0.45-0.51 8.66 2.37-0.54 1.58 8.44-1.06-11.05-0.32-0.08 0.30 0.78 0.70-0.44 2.27-21.30-0.38 0.15 2.73-2.84 1.00 0.60 0.40-0.39 Germany AR(1) AR(2) GR GR FR_M FR_M IT_M IT_M SW_M SW_M UK_M UK_M FR_S FR_S GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S 0.50 0.07-28,368.38 0.12 0.68 0.18-0.43-0.01-4.7E-03 0.11 0.09-0.01-0.37 0.10-0.05 0.25 0.06 0.50-0.41-1.05 47.73 0.12 0.32 7.99 1.12-3.44 2.56 7.92 0.86-1.63-0.92-0.15 0.39 0.18-0.20-1.69 1.37-0.62 0.99 0.25 2.02-2.25-0.68 45.70 0.43 0.55 8.40 1.08-1.30 2.09 8.77 1.82-3.25-1.64-0.11 0.56 0.23-0.80-5.68 1.59-1.29 1.88 0.40 2.43-2.03-0.98 2.26 1.06 0.83 Ialy AR(1) AR(2) IT IT FR_M FR_M GR_M GR_M SW_M SW_M UK_M UK_M FR_S FR_S GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S 0.95-72,747.83 0.25-4.7E-03 0.30-0.32 0.04-0.22 0.27 2.92 0.03 3.28-0.06-0.17 0.23-0.24 4.13 3.08-2.76-23.41-0.18 1.61 12.60-1.86 2.78 0.03 0.57-0.46 0.33-1.48 0.59 1.76 0.27 3.12-2.29-2.30 0.93-0.68 11.36 2.95-1.00-1.07-2.27 0.86 17.28-2.69 2.58 0.04 1.07-0.41 0.75-3.11 1.13 2.40 0.63 3.52-21.01-9.69 0.83-3.20 4.20 2.71-4.00-2.36-2.47 1.58 Swiz. AR(1) AR(2) SW SW FR_M FR_M GR_M GR_M IT_M IT_M UK_M UK_M FR_S FR_S GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S 0.60 0.22 4,147.21 0.09 0.81-0.06 0.12 0.01-0.03 0.01-0.01 0.01-0.07 0.02-0.02-0.01 0.01 0.08-0.06-0.07-0.19-4.7E-03-4.7E-03 8.52 3.06 0.55 1.80 10.46-1.71 2.28 1.21-1.84 0.64-1.09 1.51-55.38 1.46-1.97-0.81 0.94 1.05-0.87-0.19-0.03-0.08 0.01 11.84 5.00 1.57 3.46 15.90-2.65 3.01 1.85-2.64 2.73-3.46 1.43-4.33 1.37-1.53-1.81 2.01 2.16-1.86-0.35-0.07-0.36 0.02 UK AR(1) AR(2) UK UK FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M FR_S FR_S GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S 0.53 0.41 22,127.04 0.11 0.75 0.15 0.15-0.01-0.03-0.02-4.7E-03 0.17-0.42 0.05-0.06 0.22-0.09 0.13-0.16 1.56-13.33-0.08-0.04 11.23 8.93 1.58 2.90 11.03 1.64 1.02-0.72-1.09-2.29 0.82 1.28-1.46 1.00-2.08 1.41-0.77 1.46-1.95 1.79-0.97-0.88-0.35 11.12 8.63 10.19 3.28 11.16 1.00 0.80-0.60-1.45-1.97 0.82 0.76-1.10 1.06-1.74 1.94-1.08 1.60-2.07 1.28-46.13-1.29-0.34 Noe: The hree enries corresponding o each parameer are heir esimaes (in bold), heir asympoic -raios, and he Bollerslev and Wooldridge (1992) robus -raios, respecively. 17
Touris Source Condiional Mean Table 3: Spillover Effecs wihin Seychelles and from Maldives C O N D I T I O N A L V A R I A N C E Own Effecs Spillovers wihin Seychelles Spillovers from Maldivess France AR(1) AR(2) FR FR GR_S GR_S IT_S IT_S SW_S SW_S UK_S UK_S FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M 0.27 4,810.69 0.21 0.54-0.02 0.16 0.05-0.05 0.05-3.17 0.03-0.01-0.01 0.18-0.01-0.04-0.01-4.7E-03 0.07 0.07 0.11-0.03 4.44 2.68 3.29 5.59-2.74 2.75 1.30-1.39 0.08-0.97 0.39-0.03-0.11 1.65-0.46-1.83-1.20-0.11 0.69 0.33 2.80-0.25 5.06 0.71 3.67 9.51-2.10 2.26 2.59-12.35 0.18-0.50 0.58-0.05-0.16 1.77-2.00-3.57-2.12-0.19 1.17 0.55 4.13-0.34 Germany AR(1) AR(2) GR GR FR_S FR_S IT_S IT_S SW_S SW_S UK_S UK_S FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M 0.17 0.13-3,496.44 0.08 0.64 0.03-0.01 0.07-0.06-0.53 13.21 0.05-0.17 0.04-0.11 0.05-0.07-0.01-4.7E-03 0.09 0.06-0.01-0.09 2.26 1.67-0.74 1.19 4.34 1.50-1.46 1.01-0.92-1.14 3.92 1.07-1.84 0.73-1.41 3.71-4.34-2.98 2.07 0.84 0.22-0.67-0.76 3.41 2.67-0.54 2.21 5.52 1.55-1.02 1.74-1.40-1.55 2.46 0.54-2.49 1.10-1.76 2.43-4.06-5.72 4.06 2.27 0.38-1.66-1.42 Ialy AR(1) AR(2) IT IT FR_S FR_S GR_S GR_S SW_S SW_S UK_S UK_S FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M 0.56-0.11 56.81 0.27 0.31-4.7E-03 0.01 0.04-0.05-0.77 10.30-0.04 0.07 0.08-0.10 0.01-4.7E-03-4.7E-03-4.7E-03-0.04 0.17-4.7E-03 0.02 6.14-1.31 0.00 1.66 1.00-0.11 0.27 0.67-0.72-0.95 0.84-0.64 0.21 0.69-0.45 0.30-0.03 0.04-0.71-0.45 0.49 0.08 0.28 10.26-2.07 0.01 3.40 2.81-0.29 0.39 0.77-1.88-2.57 1.65-3.34 1.02 1.26-1.12 0.58-0.08 0.11-5.99-1.18 1.26 0.17 0.11 Swiz. AR(1) AR(2) SW SW FR_S FR_S GR_S GR_S IT_S IT_S UK_S UK_S FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M 0.26-0.14 697.29 0.05 0.58-4.7E-03-4.7E-03-4.7E-03-4.7E-03-0.01-4.7E-03 0.01-0.01-4.7E-03-0.01-4.7E-03-4.7E-03-4.7E-03-4.7E-03-4.7E-03-0.01-4.7E-03-4.7E-03 4.94-2.80 2.65 1.12 4.05 2.42-1.34-0.34 0.22-3.72 0.10 1.87-1.56 0.02-0.93 1.02-2.83 1.05-0.11 0.52-1.28-0.41 0.31 5.30-3.10 9.35 1.37 77.34 2.18-2.08-2.62 0.98-18.66 0.18 2.59-3.06 0.04-1.73 1.96-4.24 3.89-0.19 1.33-8.15-0.81 1.30 UK AR(1) AR(2) UK UK FR_S FR_S GR_S GR_S IT_S IT_S UK_S UK_S FR_M FR_M GR_M GR_M IT_M IT_M SW_M SW_M UK_M UK_M 0.29 0.10 847.16 0.12 0.74 0.02-0.02-0.01 0.01 0.01-4.7E-03-0.13 1.48-0.01 0.13-4.7E-03 0.01-4.7E-03-4.7E-03-0.05 0.02 0.02-0.08 4.33 1.47 0.18 1.76 7.43 1.32-2.42-0.96 0.98 0.36 0.10-0.75 0.33-0.29 1.91-0.65 0.73 0.60-1.04-1.63 0.29 1.63-1.82 5.33 1.84 0.44 2.69 14.87 4.86-14.06-4.96 1.16 0.50 0.16-0.82 0.80-0.52 3.05-1.78 1.28 1.55-4.24-4.05 0.63 2.70-2.48 Noe: The hree enries corresponding o each parameer are heir esimaes (in bold), heir asympoic -raios, and he Bollerslev and Wooldridge (1992) robus -raios, respecively. 18