Asian Journal of Empirical Research

Transcription

1 . Asian Journal of Empirial Researh journal homepage: IMPACT OF INBOUND TOURISM IN KENYAN ECONOMY: AN ANALYSIS USING A COMPUTABLE GENERAL EQUILIBRIUM MODEL Mohamed KARIM 1 Eri Thouamou NJOYA 2 ABSTRACT In both developing and developed ountries tourism is often regarded as an eonomi ativity of immense signifiane reating thousands of jobs. However, alulations dealing with the eonomi nature of tourism are often derived from input-output models, whih largely overstate its effets on employment by assuming linear responses and highly elasti supplies of goods, servies and labor. The purpose of this study is to develop and apply a omputable general euilibrium approah to estimate the effets of growth in tourism spending on the Kenyan eonomy as a whole and on partiular setors within it. The results indiate that the eonomi benefits from tourism expansion in Kenya are small. The paper onludes with a disussion of the poliy impliations and researh limitations. Keywords: Tourism setor, Computable general euilibrium model INTRODUCTION In the past two deades an inreasing number of researhers have sought to determine the impat of supply and demand shoks in one setor on the eonomy as a whole. Domesti or international shoks suh as the outbreak of SARS or the terrorist attaks of September 11, 2001 adversely affet industries suh as air transport, tourism and the eonomy as a whole. This indiates a need to understand the nature of the impat of shoks and poliy hanges in order to gain greater insight into the workings of suh hanges and determine ways of minimizing their adverse effets. However, muh of the researh with referene to developing ountry up to now has been desriptive in nature or has relied on input-output (I-O) analysis. The major objetive of this study is to develop and applied Computable general euilibrium (CGE) models to investigate the effets of a range of alternative poliies or exogenous tourist expenditure shoks. Despite the existene of varied tourist attrations, omprising warm weather, tropial beahes, abundant wildlife in natural 1 University of Mohammed V-Souissi, Rabat, Moroo. mokarim@voila.fr 2 University of Applied Sienes Bremen. Eri.Thouamou- Njoya@hs-bremen.de 911

2 habitats, seni beauty and a geographially diverse landsape, this potential has in many ases not been fully exploited. However, in reent years, investment in tourism infrastruture and publi health standards in most developing ountries (DCs) has improved. On the other hand, a range of fators suh as higher disretionary inomes, smaller family sizes, hanging demographis in many Northern ountries are having a huge impat on tourism demand. Many attempts to explain the linkages between tourism and eonomi growth have been made. Development theorists ontend that inreased servies export (suh as tourism) may ontribute to eonomi diversifiation and to eonomi growth. The net soial benefit of tourism growth, that is the redution of poverty and its effets on inome distribution, is an important and relatively unexplored aspet of tourism in Kenya. Eonomi models of researh in tourism are dominated by the impat of tourism measured in terms of its ontribution to gross national produt, employment and inome generation. As a private setor led, outward oriented industry, the uestion is whether tourism an ontribute to Kenya s urgent need for pro-poor growth - an important area into whih this paper will delve. The goal of this paper is to make use of general euilibrium adjustment mehanisms in answering the following uestion: will the expansion of the tourism setor in Kenya advane or retard the broader development goal of poverty alleviation? TOURISM AND THE KENYAN ECONOMY The Kenyan eonomy has undergone a strutural transformation sine the ountry s independene in The maroeonomi performane of the Kenyan eonomy over the years is best understood in the ontext of external shoks and internal hallenges that the eonomy has had to adjust to. There has been a gradual deline of the share of agriulture in overall GDP from 36.6% in the first deade after independene to about 22.2% in The share of manufaturing has only grown slowly and atually aounts for about 16.4% of GDP. The servie setor aounts for 64.6% of GDP, with the key sub-setors being transport and ommuniation, wholesale and retail trade, and hotels and restaurants. Sine 2000, Kenya s GDP growth has improved and remained strong. In 2011, real GDP grew by an estimated 5.7% due to improved performane aross all the key setors. Inflation has been more volatile, with 2011 inflation measured at 7.2%, up from 3.9% in The total population of Kenya was last reported over 40 million in 2010 (aording to a World Bank report released in 2011) from 8.1 million in The vast majority of people (79% of total population) lives in rural areas and relies on agriulture for most of its inome. The poverty rate is measured at about 48%.Kenya offers varied tourist attrations, omprising tropial beahes, abundant wildlife in natural habitats, seni beauty and a geographially diverse landsape. The most popular tourist attrations in Kenya are the wildlife and the beahes. Aording to the World Travel and Tourism Counil (WTTC) (2012), the diret ontribution of Travel and Tourism to GDP was KES167.6 billion (5.7% of GDP) in 2011 and 13.7% at full impat level and is foreast 912

3 to rise by 4.3% in Travel and Tourism diretly supported 313,500 jobs (4.8% of total employment) and 11.9% at full impat level. A fous on the sub-saharan ountry shows that the travel and tourism setor ontributed diretly to about 2.6% of total GDP (USD33.5bn) and 2.4% of total employment (5,265,000 jobs) in International tourist arrivals generally doubled sine 1990 and are still expeted to grow steadily at least for the next deade. Most overseas visitors to Kenya ome from Europe and Ameria, with Europe aounting for over 70% of the ountry s visitors (Ministry of Tourism, 2012). Tourism is a key part of the ountry s eonomi strategy. Tourism has been reognized as one of the setors that will drive eonomi growth towards ahievement of Vision Key drivers in the tourism setor through whih the government aims to ahieve Vision 2030 omprise the following: repositioning of the Coast iruit; opening underutilized parks and providing nihe produts. The strategy aims at making Kenya one of the top 10 long haul tourist destinations, offering diverse and high end experienes by 2012 to a target of five million tourists. The first National Tourism Poliy of Kenya was formulated under Sessional paper No. 8 of 1969, entitled Tourism Development in Kenya. That poliy set growth targets and spelt out strategies on how the government and private setors would develop tourism so that it beomes one of the Kenyan s leading eonomi ativities. In 2002, the Ministry of Tourism and Wildlife initiated the proess of developing a omprehensive tourism poliy and legislation. Figure 1: Tourists Arrivals (thousands) Soure: Kenya Tourism Board ( It should be noted that pakage holiday market now dominates demand. The number of tourist arrivals has inreased in reent years as an be seen in Figure-1. Tourism arrivals peaked in 2007, 3 Kenya Vision 2030 is the ountry s new development blueprint overing the period 2008 to It envisages that Kenya will beome a globally ompetitive and prosperous middle-inome ountry within the next two deades. 913

4 but dropped in 2008 to almost 33 % of the 2007 value as a result of post-eletion violene in Deember Through the 2000s, tourism arrivals grew by an average of 10% per annum. THE ECONOMIC IMPACT OF TOURISM Until reently, measurement of the eonomi impat of tourism has relied on input-output modeling. Input-output models an be used to assess the value added and inter-industries relationship attributable to tourism at the ountry level (Kweka et al. 2003; Arher, 1995; Arher and Flether, 1996; Heng and Low, 1990; Seow, 1981 and Khan et al., 1990) and to examine the impat of tourism in a provine and ity setting (West, 1993; DBEDT, 2002; Frehtling and Horvath, 1999; Finn and Erdem, 1995). Table-1 reports the multiplier effets of seleted applied I- O studies for developing ountries. Table 1: Seleted applied I-O models for developing ountries Eonomy Authors Main Findings Egypt Singapore Seyhelles Tonamy and Swinsoe (2000) Heng and Low (1990) Arher and Flether (1996) Diret tourism jobs onstitute 5.7% of national employment and 12.6% if indiret and indued jobs are inluded. Tourism ontributes over 10% to national GDP The inome impat of one Singapore dollar of tourist expenditure is estimated at S$0.77. Employment multipliers are relatively high (i.e. in 1986, 22 full time euivalent employees per million dollars of tourist expenditure). Tourism expenditure Impats vary by visitor ountry of origin so that higher spending tourists have a greater eonomi impat. Tourism ontributes approximately 24% to GDP. However, despite their general euilibrium struture, I-O models do not pay expliit attention to the effets of tourism on fator inomes or inome distribution. Input-output models assume that wages and pries do not hange regardless of the level of prodution. Thus, I-O analyses do not take aount of resoure onstraints and rowding out effets. Due to their assumptions, I-O models may give misleading results. To address this shortoming, omputable general euilibrium (CGE) models have been widely used in reent years to estimate the eonomi effets of inreases or dereases in tourism demand (Adams and Parmenter, 1995; Zhou et al., 1997; Dwyer et al. 2000, 2003; Blake et al. 2003a, 2003b, 2006a, 2006b, 2008; Sugiyarto et al. 2003; Narayan, 2004; Madden and Thapa, 2000; Gooroohurn and Milner, 2005; Gooroohurn and Sinlair, 2005; Kweka, 2004; Polo and Valle, 2007, 2008; Wattanakuljarus and Coxhead, 2008). These models have proven to be an adeuate tool for understanding the strengths, diretion and hannels of tourism s impat on a speifi setor or the eonomy as a whole. Table-2 shows the key findings of seleted empirial CGE analyses. 914

5 Table 2: Seleted applied Tourism-CGE models for developing ountries Eonomy Authors Main Findings Tanzania Kweka (2004) Tourism has a substantial positive impat on GDP, total welfare, exports and tax revenue. A 20% inrease in tourism demand results in an inrease in real GDP (at fator ost) of 0.1%. Urban areas will benefit more from tourism expansion than rural areas unless governments invest in improving infrastruture under this senario the distributional impat of tourism expansion disproportionally benefits the rural areas. Tourism growth amplifies the positive impats of globalization on Indonesia prodution and welfare. Globalization, i.e. a 20% redution in the Sugiyarto et al. tariffs on imported ommodities ombined with a 20% redution in (2002) indiret taxation levied on domesti ommodities lead to an inrease in demand by foreign tourists by 10%. Mauritius Thailand Gooroohurn and Milner (2005) Wattanakuljarus and Coxhead (2008) The tourism setors are undertaxed. Taxing tourism-related setors would generate an additional unit of tax revenue and inrease soial welfare in the proess. The authors examine the effets of pieemeal, marginal reforms of the indiret taxes (prodution and sales tax) and, onluding that the additional welfare ost of raising extra revenues from an already existing tax while holding other taxes onstant, is lower for sales tax simulations than for the prodution tax simulations, for all setors. Tourism expansion generates foreign exhange and raises household inomes, but worsens their distribution. Tourism promotion is not a pro-poor strategy beause tourism setors are not espeially labor-intensive, and their expansion brings about a real appreiation that undermines profitability and redues employment in tradable setors, notably agriulture, from whih the poor derive a substantial fration of their inome. Previous appliations of CGE modeling to the Kenyan eonomy were not onerned with tourism. During the 1980s several authors used CGE models to study the impat of eonomi reforms on the distribution of inome. The pioneers in this area in Kenya were Dervis et al., (1982) and Gunning (1983). MMahon (1990) examined the effets of unilateral tariff redution in a dual eonomy (Kenya) using a dynami CGE model on inome distribution, onluding that the trikle-down effets does not take effet sine the poorer lasses do not onsume imported goods or use them in prodution.njugunakaringi and Siriwardana (2001, 2003) applied CGE modeling to analyze the effets of maroeonomi stabilization and strutural adjustment poliies implemented by Kenya in response to two major terms of trade shoks in the 1970s, namely, the oil prie shok and the offee export boom. They suggest that fisal austerity through raising indiret taxes and trade liberalization supported by foreign aid inflows ahieve the best overall outomes. More reently, Balistreri et al. (2009) employed a 55 setor small open eonomy CGE model of the Kenyan eonomy to assess the impat of servies liberalization on both domesti and multinational servie providers. They onluded that redution of the barriers against potential providers would improve the produtivity of labor and apital and ould provide very substantial gains to the Kenyan eonomy. 915

6 MODEL OVERVIEW The model struture follows losely Robinson et al., (1999). The model builds on that of Dervis et al. (1982), whih involve speifiation of a CGE model in terms of non-linear algebrai euations and addressing them diretly with numerial solution tehniues. The model is neolassial in struture. Its main features involve profit maximization by produers, utility maximization by households, mobility of labor, and ompetitive markets. It an be desribed as a stati and singleountry CGE model extended to inorporate international tourism. The model is disaggregated into two households (urban and rural), two fators (labor and apital), and eight ativities and assoiated ommodities.cobb-douglas funtions are used for both produer tehnology and the utility funtions from whih household onsumption demands are derived (see Appendix B). Exported and domestially sold ommodities are assumed to be differentiated by market, with the relationship between them represented by a onstant elastiity of transformation (CET) funtion. Prie ratio and elastiities of transformation determined the level of output exported and sold domestially. Households and produers do not diretly onsume or use imported ommodities but instead use a so-alled Armington s omposite ommodity, whih omprises imports and the orresponding domesti ommodities. The substitution between imports and domesti ommodities is desribed by a CES funtion. Inome to enterprises omes from the share of distributed fator inomes aruing to enterprises and real transfer from the government. Their inomes are used for diret taxes, savings, and transfers to other institutions. As opposed to households, enterprises do not onsume. The government is disaggregated into a ore government aount and different tax aounts, one for eah tax type. The government ollets taxes and reeives transfers from other institutions. The government uses this inome to purhase ommodities for its onsumption and for transfers to other institutions as well as savings. The demand for ommodities by government for onsumption is defined in terms of fixed proportions. Transfer payments between the rest of the world and domesti institutions are all fixed in foreign urreny.the final institution in the model is the representative tourist. Total tourism demand for ommodities is derived from the assumption that all tourists are homogeneous, whereby Kenya faes a downward-sloping demand urve for its tourism exports. There is a representative tourist aounting for the onsumption of a ertain uantity of a omposite good and servie at an aggregated prie level (PQ ). Analogous to household demand, tourism demand is obtained by maximizing the utility funtion of the representative tourist funtion to its budget onstraint. A Cobb-Douglas demand funtion is used to give tourism exports. The demand funtion an be formulated as follows: Ctou PQ tou Ytou (1) 916

7 WhereCtou C is the uantity of ommodity C onsumed by tourist, tou the share of ommodity C in tourism onsumption and Ytou the total expenditure (revenue) of inbound tourist, whih is defined as follows: Ytou. Vtou (2) Where represents the per apita onsumption of tourist ( 1 in the base year) and Vtou the total number of tourist arrival. The household welfare hange is aptured through the Hiksian Euivalent Variation (EV), whih is one measure of welfare ommonly used in the literature. Using hanges in utility level evaluated in monetary terms (i.e. the minimum expenditure level), we ompute the hange needed to ahieve new euilibrium utilities. The welfare hange indiator (EV) is defined as the amount of money neessary to get the new level of utility. The expression of euivalent variation is given below: EV h ep( P, U ) ep( P, U ) (3) h h Where the expenditure funtion ep ( P, U) indiates the minimum expenditure level P*Q(H) that satisfies the given utility U under the prie vetor P. EV h 1 0 U h U h 0 Yh (4) U 0 h where 0 U, 1 h U and 0 h Y are the benhmark utility, the new level utility and the benhmark inome h level of household group h, respetively. From the euivalent variation euation, it is lear that tourism expansion affets household welfare through the effets on pries and onsumption. DATA The model follows the SAM (Soial Aounting Matrix) disaggregation of fators, ativities, ommodities and institutions. The database of the model is the Kenyan SAM for 2003, jointly developed by the Kenya Institute for Publi Poliy Researh and Analysis (KIPPRA) and the International Food Poliy Researh Institute (IFPRI). The struture of Kenyan Maro SAM is given in Table-1.A(see Appendix). The original miro SAM is disaggregated aross 50 setors (22 agriulture, 18 industry and 10 servies). However, for this analysis the original SAM has been adjusted in several ways (i.e. 1 agriulture, 1 manufature and 6 servies). The presene of tourists in the eonomy neessitates an additional demand omponent in the SAM. No detailed onsumption pattern of tourists in Kenya is available. The only survey available is from the World 917

8 Bank (2010). Aording to WTTC (2011) foreign visitor exports as a perentage of total exports aounted for about 17% of total exports. The expenditure ategories are uite aggregated and they are illustrated in Table-3. Besides being muh aggregated, the expenditure ategories do not ompare exatly with the I-O table of the setors lassifiation and onseuently some amendment is needed. Aommodation, inland transport and exursions and park fees are uite straightforward and are alloated to the hotel and restaurants, transport and ommuniation and the other servies setor respetively. Food and beverage, out-of-poket expenditure and misellaneous are uite problemati. The latter is so beause it is undefined. Food and beverages an atually remain in hotels and restaurants, other manufaturing or in wholesale and retail trade. Part of out-of-poket expenditure will go to the wholesale and retail trade setor but the rest an go to any other setors. Food and beverage is thus alloated to wholesale and retail trade and other manufaturing. Table 3: Expenditures of inbound tourists in Kenya, 2007 Expenditure Categories Premium Wildlife Wildlife Safari Safari Beah (All Inlusive) $/day % of Total $/day % of Total $/day % of Total Aommodation 33,35 18,1 168,3 46,6 36,85 20,3 Food/beverage 36,65 19,9 83,44 23,1 18,81 10,4 Exursions and park fees 40,71 22,1 22,98 6,4 5 2,8 Inland transport 50,36 27,4 51,62 14,3 13,35 7,4 Out-of-poket expenditure 16 8,7 35 9,7 41,43 22,9 Misellaneous 6,84 3,70 0,00 0,00 65,83 36,30 Total expenditure/bed night 183, , , Average length of stay (nights) Soure: World Bank (2010) Five setors are identified as related to tourism as follows: Hotel and restaurant (44%), transport and ommuniation (2%), retail and wholesale trade (2%), manufaturing (0.2%), and other servies (1%). Their ratio, measured as the proportion of inbound tourism demand out of the total, is given in brakets. These alulations are based on statistis provided by the Kenya National Bureau of Statistis 4, whih estimates tourism revenues at 2% of GDP at market pries for the year 2003 (KSH25.8bn). Aording to World Bank (2010) studies, the total in-ountry expenditure of, for example beah pakage in Kenya represents 51.7% of total expenditure; 36.7% of whih onstitute publi setor harges

9 SIMULATION RESULTS We simulate a 10% inrease in tourism demand by foreign tourist. The simulation uantifies hanges in prodution in all industries, hanges in employment, earnings, pries and all other variables in the model. Setors of the eonomy that are losely related to tourism would inrease output as the result of the inrease in expenditure but there would be some ontration of other setors. Table-4 shows the maroeonomi effets of a 10% inrease in all tourism demand. A 10% inrease in tourism demand is shown to inrease GDP by KSH117,713 million. The GDP inrease is euivalent to 0.12% of GDP. In addition to inreasing GDP, the inrease in tourism demand is shown to inrease government revenues by 50 million KSH. There is a % appreiation of the real exhange rate, and slight inreases in labor demand. An eonomi rationale for promoting tourism by developing ountries is the improvement of the trade balane by inreasing export earnings. The simulated 10% inrease in tourist expenditure results in an inrease in total exports (0.008%) whih outweighs the inrease in total import (0.001%), resulting in an improvement in balane payment. Table 4: Maroeonomi effets of simulations Effets of additional tourism growth Base year Perentage Value value hange GDP at market pries (from spending side) (Millions KSH) Private Consumption (Millions KSH) Investment (Millions KSH) Government Consumption (Millions KSH) Total Export (Millions KSH) Total Import (Millions KSH) Domesti Output (Millions KSH) Labor Demand (Millions KSH) Exhange Rate (Index) The expansion of tourism is projeted to have impliations on other industries. Table-5 ontains output projetions for 8 setors aggregated from the 50 setors distinguished in the 2003 Kenyan SAM database. The results demonstrate that, at the setoral level, there will be losers as well as gainers from an expansion in inbound tourism. The industry level expansion patterns in Table-5 show the largest expansion in the setor that sells a larger proportion of their output to foreign tourists, that is hotel and restaurant servies. The setor that ontrats the most is the manufaturing 919

10 setor. Partiularly poor growth prospets are projeted for the onstrution setor and other servies. Table 5: Setoral effets of simulations - Domesti output (Millions KSH) Effets of additional tourism growth Base year Perentage Value value hange Agriulture Manufature Publi Utilities Constrution Wholesale and Retail Trade Hotel and Restaurant Servie Transport and Communiation Servies Other Servies Table-6 shows the effets that tourism demand shok has on labor demand. The results indiate that the effets of inreasing inbound tourism on employment losely math with the effets on domesti output. That is, the industries with large domesti output effets will generate large labor demand effets. In the simulation results of this study, the largest effets are on the restaurant and hotel servies setor. Effets on the restaurant and hotel servies setor trigger an inrease in domesti output of about 0.03% and an inrease in the labor demand of about 0.05%. Table 6: Setoral effets of simulations - Labor demand Effets of additional tourism growth Base year Perentage Value value hange Agriulture Manufature Publi Utilities Constrution Wholesale and Retail Trade Hotel and Restaurant Servies Transport and Communiation Servies Other Servies

11 As expeted, a 10% inrease in tourist expenditure impats on welfare and domesti onsumption output. The simulated positive tourism shok results in a 0.3% inrease in rural household onsumption and in a 0.02% inrease in welfare (Table-7). Urban household s onsumption and welfare on the ontrary drop by 0.05% and 0.06 respetively. Private onsumption inreases by 0.18%. Yet the rural household groups are the main benefiiaries of an international tourism inrease. As an be seen from Table-6, a 10% inrease in tourism results in an inrease in the domesti onsumption of agriultural ommodities (0.285%), a derease in the domesti onsumption of manufature and servie produts of 0.005% and 0.03% respetively. From the previous, it an be onluded that tourism growth in Kenya is pro-agriulture. Table 7: Results in % hange in welfare and household onsumption Eonomi indiator Welfare (EV) - Rural Household - Urban Household - Net effet Consumption - Rural Household o Agriulture o Manufature o Publi Utilities o Wholesale and Retail Trade o Hotel and Restaurant Servie o Transport and Communiation Servies o Other Servies - Urban Household o Agriulture o Manufature o Publi Utilities o Wholesale and Retail Trade o Hotel and Restaurant Servie o Transport and Communiation Servies o Other Servies Perentage hange from benhmark o 0.29 o o o o o o o o o o o o o Prie effets are of partiular interest in the CGE model. The 10% inrease in foreign demand leads to inreases in pries, on average, just under 0.2%, whih redues the growth in tourism onsumption to around 0.04%. Inreasing eonomi ativity reated by tourism expansion inreases real wage rates by 0.8%. 921

12 SENSITITY ANALYSIS The elastiity parameters for this study have been obtained from existing studies on Kenya, values assumed in CGE models for other developing ountries and guesstimates. Considering the unertainties assoiated with the elastiity parameter of Kenya, sensitivity analysis is used to demonstrate the robustness of simulation results by varying parameters that may signifiantly affet the results. By inreasing or dereasing the value of the Armington onstant elastiity of substitution (CES) and onstant elastiity of transformation (CET), we examine the range over whih output hanges. We define a higher-elastiity ase with 20% higher values and a lowerelastiity ase with 20% lower value for those parameters. To evaluate the robustness of the simulation results, we set the following two riteria: (a) whether the signs of the setoral output hanges are unhanged in all ases and (b) whether the ordering of the output hanges among setors is maintained in all ases.the results of the sensitivity analysis shown in Table-7 indiate that the simulation results satisfy riterion (b) but not riterion (a). More preisely, while the output of manufature would always be affeted in the same diretion in the different assumed elastiity values, the output of agriulture would inrease in the base-line and higher-elastiity ases but would derease in the lower-elastiity ase. Table 7: Impat of different elastiity values on setoral output Output of: Elastiity of substitution/transformation Baseline ase Higher-elastiity Lower-elastiity ase ase Agriulture Manufature Publi Utilities Constrution Wholesale and Retail Trade Hotel and Restaurant Servies Transport and Communiation Servies Other Servies Unit: hanges from the base run in % CONCLUING REMARKS AND RECOMMENDATIONS In this study we have applied the CGE model to examine the impliations on shoks in tourism expenditures on outputs, inome distribution and national welfare. In doing so, this study analyses the supply side of the tourism setor in Kenya and reviews the literature on tourism using I-O and CGE with a speial fous on developing ountries. The analysis indiates that Kenya is endowed with tourism-attration potentials. To date, most studies of the eonomi impat of tourism have relied on input-output analyses. These studies found that tourism expansion has the potential to ontribute to inreased eonomi growth of developing ountries by generating additional 922

13 employment for the poor or inreasing tax olletion. However, CGE models are now inreasingly being used in tourism eonomis analysis and those applied to developing ountries found that tourism expansion may have an eonomi ost. The CGE models onsist of a set of euations that haraterize the prodution, onsumption, trade and government ativities of the eonomy. The CGE models enjoy an advantage over I-O models in that they take into aount the interrelationships between tourism, other setors in the eonomy and onsumers, and have the ability of inorporating endogenous prie determination mehanisms. Using CGE simulations we analyzed the effets of an inrease of 10% in foreign tourist arrivals. The results show that, overall, the effets of tourism expansion are benefiial but entail osts for other setors and for the urban households group. The analysis has shown that a small proportion of the effets of an inrease in tourism demand would be aompanied by an inrease in pries. Rural households, whih onstitute 77.8 % of the total population (2010) of whih 49% are onsidered poor, will benefit most from tourism growth. Inbound tourism inreases the output of agriultural produts, dereases its pries and inreases employment. Agriulture is a major setor from whih rural households derive a substantial fration of their inome (36%). Moreover, these groups spend a large proportion (53%) of their inome on agriultural produts. These results indiate a strong linkage between the agriulture industry and the tourism industry. This finding is in agreement with Summary s (1987) findings whih, omparing the Kenyan tourism industry to other setors of the eonomy, established that forward linkages were high in agriulture and that the import ontent of the tourism industry was low. One of the more signifiant findings to emerge from this study is that the net benefit to Kenyan from additional tourism is ambiguous. These findings seem to be onsistent with other researhes. These authors found that in destinations where tourism is relatively less labor intensive than agriulture and whose tourism produts are mainly intensive user of natural environment (e.g. Mauritius, South Afria and Zimbabwe), inbound tourism growth will lead to an ambiguous net benefit on national welfare. Of ourse, Afrian tourism produts (large sale resorts, national parks, safaris, golf tourism, adventure tourism, et.) are mainly land-intensive. However, some findings of this study do not support some previous researhes whih highlighted that the tourism output expands at the expense of the agriultural output. It is diffiult to explain this result, but it might be related to the highly aggregated nature of the agriultural setor in the model or the hoie of funtional forms and parameters. In terms of poliy impliations, one the main issues that emerges from these findings is that when deiding on tourism development strategy, poliy makers should give due onsideration to the overall eonomi development. Moreover, they should paid attention to the whole range of distortions that affet the ongoing development of the tourism setor. With regard to the uestion whether inbound tourism growth will advane or retard the broader development goal of poverty 923

14 alleviation, the findings show that unless governments implement omplementary strategies aiming at mitigating the osts of tourism expansion, eonomi development and poverty alleviation will not be attained. This researh is part of an ongoing researh designed to develop uantitative information on the ontribution of tourism in Kenya using CGE models. More researh on this topi needs to be undertaken before the assoiation between tourism growth and welfare is more learly understood. One of the weaknesses of this study is the hoie of the funtional forms, whih assume a Cobb- Douglas prodution and utility funtion. Alternative funtional forms suh as a CES prodution funtion or a Stone-Geary utility funtion (generally preferable sine it allows for subsistene onsumption expenditures) may be preferred. Another weakness of this study is the high level of aggregation of data onerning the agriulture and manufature setors, the fator markets and household ategories as well as the tourism industry. Detailed data on tourism expenditures are needed to improve the understanding of the impat of tourism shoks on different setors and institutions. Furthermore, the results would be more useful to tourism poliy makers if these parameter values were empirially estimated. This is an important issue for future researh. Aknowledgements We would like to thank Nobuhiro Hosoe and Manfred Wiebelt for their help in building the model and Hans Martin Niemeier, Isabella Odhiambo, Ageliki Anagnostou, Faith Miyandazi and for reading the paper and providing onstrutive omments. REFERENCES Adams, P. D. and B. R. Parmenter (1995). An applied general euilibrium analysis of the eonomi effets of tourism in a uite small, uite open eonomy. Applied Eonomis, Vol. 27, pp Arher, B. (1995). Importane of Tourism for the Eonomy of Bermuda, Annals of Tourism Researh Vol. 22, pp Arher, B. H. and J. Flether (1996). The Eonomi Impat of Tourism in the Seyhelles. Annals of Tourism Researh, Vol. 23, No. 1, pp Balistreri, E., T. F. Rutherford and D. Tarr (2009). Modeling servies liberalization: The ase of Kenya. Eonomi Modelling, Vol. 26, No. 3, pp Blake, A., Sinlair M. T. and Sugiyarto (2003a). Quantifying the impat of foot and mouth disease on tourism and the UK eonomy. Tourism Eonomis, Vol. 9, No. 4, pp Blake, A. and M. T. Sinlair (2003b). Tourism Crisis Management: US Response to September 11. Annals of Tourism Researh, Vol. 30, No. 4, pp

15 Blake, A., R. Durbarry, J., L. Eugenio-Martin., N. Gooroohurn, B., Hay, J. and J. Lennon, (2006a). Integrating foreasting and CGE models: The ase of tourism in Sotland. Tourism Management, Vol. 27, pp Blake, A., J. Gillham and M. T. Sinlair (2006b). CGE tourism analysis and poliy modeling. In L. Dwyer and P. Forsyth (Eds) International Handbook on the Eonomis of Tourism, Cheltenham, Edward Elgar Publishing. Blake, A., J. S. Arbahe., M. T. Sinlair and V. Teles (2008). Tourism and Poverty Relief. Annals of Tourism Researh, Vol. 35, No. 1, pp Clemente, P. A and E. Valle (2007). An Assessment of the Weight of Tourism in the Baleari Islands, CRE Working Papers (Douments de treballdel CRE), Centre de Reera Eonòmia (UIB "Sa Nostra"). Dbedt(2002) The Hawaii Input-Output Study: 1997 Benhmark Report, Honolulu: Researh and Eonomi Analysis Division, State of Hawaii. Dervis K., De Melo, J. and S. Robinson (1982). General euilibrium models for development poliy. Cambridge, MA: Cambridge University Press. Dwyer, L., P. Forsyth, J. Madden and R. Spurr (2000). Eonomi Impats of Inbound Tourism under Different Assumptions Regarding the Maroeonomy. Current Issues in Tourism, Vol. 3, No. 4, pp Dwyer, L., P. Forsyth, R. Spurr and T. Ho (2003). Contribution of tourism by origin market to a state eonomy: A multi-regional general euilibrium analysis. Tourism Eonomis, Vol. 9, No. 2, pp Finn, A. and T. Erdem (1995). The eonomi impat in the ase of West Edmonton Mall. Tourism Management, Vol. 16, No. 5, pp Frehtling, D. C. and E. Horvath (1999). Estimating the Multiplier Effets of Tourism Expenditures on a Loal Eonomy through a Regional Input-Output Model. Journal of Travel Researh, Vol. 37, pp Gooroohurn, N. and M. T. Sinlair (2005). Eonomis of Tourism Taxation: Evidene from Mauritius. Annals of Tourism Researh, Vol. 32, No. 2, pp Gooroohurn, N. and C. Milner (2005). Assessing Indiret Tax Reform in a Tourism-Dependent Developing Country. World Development, Vol. 33, No. 7, pp Gunning, W. J. (1983). Inome Distribution and Growth: A Simulation Model for Kenya. In D. G. Greene (prinipal author), Kenya: Growth and Strutural Change, 2 vols. Washington, D. C: World Bank, pp Heng, T. M. and L. Low (1990). Eonomi Impat of Tourism in Singapore. Annals of Tourism Researh, Vol. 17, pp Karingi, S. N. and M. Siriwardana (2001). Strutural Adjustment Poliies and the Kenyan Eonomy: A CGE Model Analysis. Afrian Development Review, Vol. 13, pp

16 Karingi, S. N. and M. Siriwardana (2003). A CGE Model Analysis of Effets of Adjustment to Terms of Trade Shoks on Agriulture and Inome Distribution In Kenya. Journal of Developing Areas, Vol. 37, pp Khan, H., C. F. Sengand, and W. K. Cheong (1990). Tourism Multiplier Effets on Singapore. Annals of Tourism Researh, Vol. 17, pp Kweka, J., O. Morrissey and A. Blake (2003). The eonomi Potential of Tourism is Tanzania. Journal of International Development, Vol. 15, pp Kweka, J. (2004). Tourism and the Eonomy of Tanzania: A CGE Analysis. Retrieved Marh 31, 2012, CSAE2004.pdf MMahon, G. (1990). Tariff poliy, inome distribution, and long-run strutural adjustment in a dual eonomy: A numerial analysis. Journal of Publi Eonomis, Vol. 42, No. 1, pp Madden, J., and P. Thapa (2000). The Contribution of Tourism to the New South Wales Eonomy. Centre for Regional Eonomi Analysis, University of Tasmania. Ministry of Tourism (Kenya), (2012). Tourism performane overview. in websites. Narayan, P. (2004). Eonomi Impat of Tourism on Fiji's Eonomy: Empirial Evidene from the Computable General Euilibrium Model. Tourism Eonomis, Vol. 10, No. 4, pp Polo, C. and E. Valle (2007). A General Euilibrium Assessment of the Impat of a Fall in Tourism Under Alternative Closure Rules: the Case of the Baleari Islands. International Regional Siene Review, Vol. 31, No. 1, pp Robinson, S., A. Y. Naude, R. H. Ojeda, J. D. Lewis and S. Devarajan (1999). From stylized models: Building Multisetor CGE Models for Poliy Analysis. North Amerian Journal of Eonomis and Finane, No. 10, pp Sahli, M. and J. J. Nowak (2007). Does Inbound Tourism Benefit Developing Countries? A Trade Theoreti Approah. Journal of Travel Researh, Vol. 45, pp Seow, G. (1981). Eonomi Impat Signifiane of Tourism in Singapore. Singapore Eonomi Review, Vol. 26, No. 2, pp Sugiyarto, G., A. Blake, and, M.T. Sinlair (2002). Tourism and Globalization: Eonomi Impat in Indonesia. Annals of Tourism Researh, Vol. 30, No. 3, pp Summary, R. M. (1987). Tourism's ontribution to the eonomy of Kenya. Annals of Tourism Researh, Vol. 14, No. 4, pp Tonamy, S. and A. Swinsoe (2000) The Eonomi Impat of Tourism in Egypt, Working Paper No.40. West, G. R. (1993) Eonomi Signifiane of Tourism in Queensland, Annals of Tourism Researh, Vol.20,pp Wattanakuljarus, A. and Coxhead, I. (2008). Is tourism-based development good for the poor?: A general euilibrium analysis for Thailand. Journal of Poliy Modeling, Vol. 30, No. 6, pp

17 World Bank (2010). Kenya s Tourism: Polishing the Jewel. Retrieved Marh 31, World Tourism Organisation (2001). The Least Developed Countries and International Tourism. in Tourism in the Least Developed Countries, edited by the World Tourism Organization. Madrid: World Tourism Organization. World Travel and Travel Counil (2011). Kenya Eonomi Impat Report. Retrieved Marh 28, 2012, Zhou, D., J. F. Yanagida, U., Chakravorty and L. Ping Sun (1997). Estimating eonomi impats from tourism. Annals of Tourism Researh, Vol. 24, No. 1, pp Appendix A: SAM Kenya Table A. 1: 2003 Kenya Maro Soial Aounting Matrix (Millions of Kenyan Shillings) Appendix B: The euations of the models Indies aa ativities C ommodities CE ( C) exported ommodities CNE ( C) ommodities not in CE CM ( C) imported ommodities C CNM ( C) non imported ommodities ff fators ii institutions (domesti, tourist and rest of the world) 927

18 Table B. 1: PARAMETERS ad a a a at pi prodution funtion effiieny parameter shift parameter for omposite supply (Armington) funtion shift parameter for output transformation (CET) funtion onsumer prie index wts weight of ommodity in the CPI ia a uantity of as intermediate input per unit of ativity a inta a uantity of aggregate intermediate input per ativity unit iva a mps h pwe pwm dtst QG C barinv(c) inv se shry if uantity of value-added per ativity unit share of disposable household inome to savings export prie (foreign urreny) import prie (foreign urreny) uantity of stok hange base-year uantity of government demand exogenous (unsaled) investment demand base-year uantity of private investment demand enterprise saving rate share for domesti institution iin inome of fator f te tm t tr ii ty i Vtou exporttax rate import tariff rate rate of sales tax transfer from institution i to institution i rate of nongovernmental institution inome tax number of tourist value-added share for fator f in ativity a fa h share of ommodity in the onsumption of household h tou share of ommodity in tourism onsumption share parameter for omposite ommodity supply (Armington) funtion t share parameter for output transformation (CET) funtion yield of ommodity per unit of ativity a a 928

19 Armington funtion exponent 1 t CET funtion exponent 1 per apita onsumption of tourist elastiity of substitution for omposite supply (Armington) funtion t elastiity of transformation for output transformation (CET) funtion Table B. 2: VARIABLES Ctou EG EXR FSAV GSAV IADJ PA a inbound tourist's onsumption by setor government expenditures exhange rate (LCU per unit of FCU) foreign savings government savings investment adjustment fator ativity prie PD domesti prie of domesti output PE export prie (domesti urreny) PM import prie (domesti urreny) PQ omposite ommodity prie PVA a value-added prie (fator inome per unit of ativity) PX aggregate produer prie for ommodity QA a uantity (level) of ativity QD uantity sold domestially of domesti output QE uantity of exports 929

20 QF fa uantity demanded of fator f from ativity a QFS f supply of fator f QH h uantity onsumed of ommodity by household h QINT a uantity of ommodity as intermediate input to ativity a QINV uantity of investment demand for ommodity QM uantity of imports of ommodity QQ uantity of goods supplied to domesti market (omposite supply) QX Walras WF f aggregated marketed uantity of domesti output of ommodity dummy variable (zero at euilibrium) average prie of fator f WFDIST f YE YF if YG YI i Ytou UU wage distortion fator for fator f in ativity a enterprise inome transfer of inome to institution I from fator f government revenue inome of domesti nongovernment institution total expenditure of inbound tourist utility (fititious) Table B. 3: EQUATIONS Prie Blok Import prie Export prie PM pwm 1tm ) EXR ( CM PE pwe 1te ) EXR ( CM Absorption PQ QQ PD QD PM QM t Market output value Ativity prie (1) (2) 1 ( CD CM ) (3) PX a PX C QX PD QD PE QE CX (4) PA a A (5) a 930

21 Value-added prie PVA PA PQ ia a A (6) a a C a Prodution and Commodity Blok C-D tehnology: Ativity prodution funtion Fator demand Intermediate demand Output Funtion WF WFDIST f QINT QX fa fa PVAa QAa QF fa QA ad a a f F QF fa fa a A (7) a A and f F (8) a ia a QAa a A and C Composite supply (Armington) funtion a QA C a aa 1 (10) (9) QQ QM 1 QD CM CD (11) Import-domesti demand ratio QM PD QD PM CM CD (12) Composite supply for non-imported outputs imports QQ QD CNM (13) Output transformation (CET) funtion QX Export-domesti supply ratio at QE t t 1 QD t 1 t t CE (14) QE PE QD PD 1 t t 1 1 t Output Transformation for nonexported Commodities QX QD CE (15) CNE (16) Institution Blok Fator inome YF if shry if aa WF WFDIST f fa QF Household onsumption demand for marketed ommodities fa i I and f F (17) h 1 mpsh t y YH h PQ QH 1 h C and h H (18) Investment demand QINV IADJ inv C (19) Government onsumption demand QG GADJ g C (20) 931

22 Government revenue YG tyi YI i EXR trgov, row t PQ QQ shrygov, f ii C trgovent, (21) tm pwm QM EXR te pwe QE EXR CM CE Government expenditures EG PQ QG tri, gov (22) C ii Tourism demand Ctou PQ tou Ytou (23) Tourist Revenue (expenditure) Enterprise revenue Ytou Vtou (24) YE tr ent, i shryent, ap ii Objetive funtion UU srwalras (25) (26) System Constraint Blok Fator market QFfa aa QFS f f F (27) Composite ommodity markets aa QQ QINT QH QG QINV dst CTOU C (28) a hh Current aount balane for rest of the world (in foreign urreny) CM ii CE h pwm QM tr pwe QE tr PQ / EXR Ctou Savings-Investment Balane ii MPS i row, i ii i, row C FSAV (29) tyi YI i ( YG EG) EXR FSAV PQ QINV PQ dst WALRAS 1 (30) C C Prie Normalization CPI PQ wts (31) C 932