Report No. 566a r uuri Caribbean Regional Study Volume Vib 566 A Study of the Contribution of Tourism vol. 9 +- t-he A r%fibaT. Ec or, o L'~U LiIC /-Ai iu16uCII L%-"1'J I ly' June 9, 1975 Latin America and the Caribbean Region 1No0t for Public Use Document of the International Bank for Reconstruction and Development International Development Association This report was prepared for official use only by the Bank Group. It may not be published, quoted or cited without Bank Group authorization. The Bank Group does not accept responsibility for the accuracy or completeness of the report. CURRENCY EQUIVALENTS Exchange Rates Effective May, 1974 US$1.00 = 20O0 BDS$ (Barbados) = 1.71 B$ (Belize) = 2.18 G$ (Guyana) = 0O91 J$ (Jamaica) = 2.00 TT$ (Trinidad & Tobago) 4' = 2000 EC$ (Eastern Caribbean ) 1 BDS$ = 0Q50 US$ 1 $ = o58 US$ 1 G$ = O046 US$ 1 J$ = 1010 US$ 1 TT$ = 050 US$ 1 EC$ = 0.50 US$ Note: As of July 15, 1974, the currencies of Barbidos, Belize, Eastern Caribbean, Guyana and Trinidad and Tobago were linked with sterling. Jamaica maintains a parity relationship with the US$. 1/ Antigua, Dominica, Grenada, Montserrat, St. Kitts-Nevis- Anguilla, St. Lucia, and St. Vincent. lT AT D V LT Nl Page No. SU'miRY AND COUNCLUCIONS ..... ..................... i I. ACCO-AuNTING PRICES FvIR TOURISM P-KOJECTS IN ANTIGUA ............................. 1 II. ANALYSIS OF THE IMPACT OF THE TOURISM SECTOR ON THE ANTIGUANI ECONOMY ......... 12 III. ECONOMIC POLICY IN ANTIGUA WITI PARTICULAR REFERENCE TO TOURISM .... ............... 30 APPENDICES 1. INFRASTRUCTURE REQUIREMENTS IN RELATION TO TOURISM ................................ 1 - 16 SUMMARY AND CONCLUSIONS 1. Antigua is an island in the Caribbean with a population under 70,000; and aNP nrnhnobly nder US$30 millinn nr ahnot SALnn00 per capita. It is, however, an independent political and economic unit. 2. In the early 1950's Antigua was heavily dependent on sugar and tkh eSIA.A -, 1 4iv4-n waa evtremnea low. The" tha f4irt osi4nvi4 ficant 4tos2m development on the island took place, a residential club for well-to-do A^Mricar.s. At th*-e same time+-,-- ther -a ak- 4oo ir.-#- eirton which eased ~UL~L L.LLO* IL. LLA ~ ~ ..U&e .& -C C UJ.JL ~& =L11.rL C&6L . .LJIWLAJA..LA %AO unemployment and provided a flow of remittance income, but raised consumption standLlarlds ar,.d muay have dleprivedl A.t-iga ofC som-e ofisms4-.epii. S L8LLdi.~ LULU I1V U~L .LV=A AL%.rLLCL WA LU J . OWU L L WIU LAU 0 6L citizens. 3. From the late fifties to late sixties there was a boom in hotel con.structLonL. Ihe economy becarue geareu to grow-h iLn tourism. Sugar pro- duction and export diminished rapidly and ceased completely in the seventies. Sea island cotton productiorn also fell. Mieanwhile, wages rose excepLionally fast. During this period Antigua attained effective internal self-government and indigenous political maturation was bound up with unionization. 4. The sixties also saw a considerable curtailment in emigration opportunities but until the late sixties the economic effects of this were offset by the boom in tourism and construction. At about the same time, Antigua was beset by various other problems. The expansion of tourism had put a severe strain on the infrastructure and there was now more competition for tourism from other Caribbean islands. In the general recession in tourism to the Caribbean, tourism to Antigua fell significantly, from 223,662 bednights in 1968 to 172,797 in 1971. The continuing recession in the tourist industry led to a sharp decline in hotel employment: at the same time, construction of new hotel rooms virtually ceased. For a while government infrastructure projects partially maintained overall employment in construction but by the early seventies infrastructure investment was tapering off. Meanwhile, emigration possibilities were further restricted and unemployment reached extremely high levels. At the same time, the government faced a chronic budgetary situation because of accumulated short-term debts in connection with various infrastructure projects and also the unsuccessful attempt to support the sugar industry. In view of its high level of unemployment and its fiscal problems, Antigua is at present in rather a difficult economic position. S. During the course of its development in the island, tourism has been a considerable stimulus to entrepreneurship and the acquisition of skill. Apart from managers, the Antigua hotel industry now employs almost no expatriates. Taxis and car-rental businesses have offered scope for Antiguan entrepreneurs and so did the construction industry. 6. As regards the social impact of tourism, most of it has been very similar to that which would have been created by any other leading sector and the resultant general economic development. Granted the social and political structure of Antigua, unionization of such a sector would probably have been - ii - inevitable with a sharp rise in wages and Iiigher general living costs. Sugar would have been abandoned, the society would have become more urban and un- employment would have become a major social issue. Social effects specific to the intrinsic nature of tourism are probably of less significance, such as a minor health problem with non-indigenous strains of influenza. Tourism has not seriously exacerbated race relations, and although anti-white feeling can be encountered, the general situation is better than in most large U.S. cities. On the positive side, tourism has stimulated demand for locally- grown produce and for local entertainment groups and craft-work. The high pay available to mechanics and building workers in recent years has also helped to encourage these industries. 7. Hotels are the core of the Antigua tourist industry: in summer 1973 there were 33 hotels with 1,140 rooms on the island. A typical room rate for the 1972/73 season was about US$70 per day for a double room, Modified American Plan. Bed occupancy rates averaged-over the year well under 30%. Direct hotel employment reaches a maximum of about 1 iob per room when a hotel is full. Direct government revenue from tourism includes a ECS 2.00 hotel tax per bednight and an embarkation tax of EC$ 5.00. During 1973 a 3% tax on most of a guest's hotel bill was also imposed. In 1972 there were 72,328 tourist arrivals, the highest figure ever_ Ilowever, bednights were only 177,962, as against 223,662 in 1968, the peak year. The number of cruise visitors has been rising ranidlv and Antigua is also an imnortant yachting center. 8. Tn the 1972/73 season. gross tourist expenditure in Antigua was between EC$ 25 million and EC$ 30 million. Directly and indirectly tourism probably contributes about half of the GD(P Antigua's balance nf navments is now almost totally dependent on tourism; and although direct hotel employ- ment is small, tourism directly or indirectly finances virtually all t-le high income employment on the island. Unemployment and under-employment are ex- tremely high. Because of inappropriate nt pricing and borrowing policies for infrastructure investment, the growth of tourism bas in fact caLusedu a serious budgetary crisis. 9. ~~IfA presentL trendus in Aniu continue, hotel occupancv rattes willl continue to be very low and the island will continue to offer-mainly luxurv or semi-luxcury hlotels. Hig wage rates and lablor troulIles are 1ikeyto- pe- vent much increase in industrial investment. The employment and budgetary situation will remain quiLt UdLifcu'Lt andu may deteriorate. The resuLt could be increasing social and political tensions and this could adversely affect tourism. Rather more probably, social, economic and political affairs mav continue in an unsettled state but some tourism is likely to be sustained. 10. A major objective of this present study is to develop a methodology for estimating the contribution of tourism to the economy through the use oi accounting prices. Estimating accounting prices for the Antiguan economy is easy for traded commodities, because the economy is very open and has no quantitative trade barriers. Ilowever, severe data problems arise in con- nection with nontraded goods. There are also serious difficulties in making the value judgments necessary to estimate shadow prices. - iii - 11. The object of the method is to reduce all costs and benefits to one numeraire which is taken to be foreign exchange in the hands of the Antiguan Government. Direct and indirect inflows and outflows of foreign exchange can therefore be taken at face value. But in order to arrive at these it is necessary in several cases to deduct government taxes and subsidies which. in general, are not real costs or benefits. Furthermore, the object of the method is not to calculate the impact of projects on the foreign balance but to calculate the impact on welfare. Therefore, it is necessary to value incomes created by the projects, e.g., for workers previously unemployed, in terms of the numeraire. Because of data problems, an iterative method was .wed to annronimate the general enuillbrium valltip- nf Ohdonw prinPR- 19. Acrointing nrires for tradeid rommoditipe w workea oit An thp basis of estimated accounting ratios for wholesale and retail margins. The accounting ratio for surplus labor was based upon unemployment and under- employment levels in Antigua, an estimate of leisure preference, and relative valu s _:ar "W 4 4 .mentmand _concsumpt4on. TF wasQ assume that th1 1et pa o *-___-- _--F t_- ---A - r-* - t-s -*- -f professionals is adequately reflected in market prices. The accounting rate of: interest is estimate -A be^ thelile 141- l -1 t yg-of interest at which the Antigua Government could now obtain finance on world capital markets, taking ir.to accon.t th-e gove,,.r..ent s budgetary fut. 4ACies and tbLe scarci ty of savings. Because up-to-date cost breakdowns are not available, it is difficult to estimLLate accoL.t.ir.g price ratios for nontradeLd ---ods a.d serviLces. "lowever estimates were made on the basis of an inter-industry transaction matrix for th-e Antiguar. econ-omy ir. !963. L..LL~ t1.L1L.LU4I LIU y LI 2 Uv..J 13J. louriLsm 'as beer. estiLMtd4LtU LU g e 5U0. Uf GJDJ taking ful account of indirect benefits. A very substantial proportion of the indirect value added by tourism would be lost xi tourism were to cease. The concept of a "tourist multiplier," however, probably has relevance in Antigua only in the off-peak season, when local industries are working below full capacity. At that time it is between 0.75 and 1.0, but falls to 0.2 for marginal tourism expen"LLtures durinLg peak seasor,. 14. Even at low occupancy rates, tourism in Antigua is a socially pro- fitable activity at the margin. The principal reason for this is that the capital investment in hotels is made by foreigners and as far as tne economy is concerned, this direct capital has zero opportunity cost. The costs consist of the investment in infrastructure that the government makes for the purposes of supporting tourism, and repatriated profits; the benefits consist of the employment of underemployed labor and the net tax revenue accruing to the government. However, perhaps the most significant analytical conclusion is that the social profitability of increased occupancy rates would be very high in Antigua, especially during the off--season. 15. Antigua has no control over its exchange rate and being a monetary board economy cannot run a deficit on its foreign balance. Lack of inter- national competitiveness therefore manifests itself in the low profitability of export industries and in widespread unemployment. Trade unions in Antigua - iv - are powerful and exert upward pressure on money wage rates. Labor productiv- ity in Antigua is low by international standards but is growing slowly. Seen against -his background, the high rate of unemployment and the lack of diversity in the industrial structure are not really surprising. 16. Any sustained economic progress in Antigua is probably conditional upon some control over the growth of money incomes and this means some control over the activities of trade unions. During the off-season; nrices charged by Antiguan hotels are patently too high. Antigua does not possess the monoponlv nnwr to hp able to rharge such nrires= This apnlies not nnlv tn hotel rates, but also to the prices of other tourist services such as taxi- transnprt. Th nprnholm nf hntpl ratps i4 not whnllv the result of the pace of money wage increase but the latter is undoubtedly a major contributary factor. What is perhaps even more important is that the pace of money wage increases prevents the diversification of the productive structure in Antigua through the development of new, interr.ationally compet4 tive 4iA7nd present conditions, it is difficult to imagine a successful export drive in A-. -igua basedAon low=wage labor. The result -s that excessive dependence on tourism becomes self-perpetuating. 17. An unfortunate side effect of tourism is that it prevents moderation of money wage in.reases bDecause it sets thLe star.dards whichL other sectors follow. There is some truth in this though the same argument would apply to any leauding sector. Thle proble-r. ir. Antigua was r.ot the advent ofL tourism. buut ~LI~ .L~aU.LIL~ .L LLI jJLJU.LLL .L L'LIL1h. L,U a JL.LLA UV 1I UL UU .LI LL that the response to it was unplanned. The government should institute a more re alisi ico e polic EIUwL m WL LIZ corr V-sponiA ±Lng. Ly hL±'L Lh A C_ LIL LLIR hotels. 18. Any reduction in Antigua's relative dependence on Tourism in the immediate future seemis to bUe out of the question b ut possibilities exist f'or the longer-term future. If agriculture were better developed, there could be a fruitful symbiotic relationship with tourism. As for manufacturing industry, the problem for Antigua is one of skills and of wage levels, of which the latter is the more immediately relevant. Skills can only be developed over time in the process of work; if wages are too high relative to those elsewhere, new industries cannot get started. 19. Even with far-sighted planning, the expansion of tourism beyond a certain point would create problems. Diversification is not easy for reasons outlined earlier but it is extremely important that the government should make every effort to achieve it. 20. The immediate problem in tourism policy is how to improve occupancy rates. Part of the answer would be to control the rate of increase in money wages. Even if this is taken to be outside government control, there still exists the possibility of lowering off-season prices further. Promotion is equally important. 21. An important issue for future policy concerns the optimal size of hotels. Economies of scale certainly exist in hotel operation. On the other hand, small hotels can use family labor. They also offer social benefits. They are generally less capital-intensive, have local ownership which saves repatriation of profits. create a class of neonle experienced in hotel manage- ment and involve the local people to a greater extent in the tourism business. 22. Another issue in tourism policy is whether to encourage luxury touris or middle-inonm. tourism. The tancyilo Aanto of nf the middle-income tourist is the saving on infrastructure costs. Middle-income tourists might alon a mle it easier to mairntaln high occupan.cy rates during the summer. On the other hand, with a few luxury tourists the sheer pressure of~ nnumbers ounaldAl be S-.1e or l thok~ rsn 1 -4 m, l.- A --4 - .,l.1 a i -a ma.,e to of ~ ~ ~ ~~--- -.mbr -ol esalr r the whole, it m be ----rableto.veo some extent in the direction of middle-income tourists. 23. No rational decisions on tourism policy are possible without a calcuAlati4o.. of thbe cost of iEnfrastructure requ4rements. Y.owever, the Aat-a basis for such a calculation hardly exists at present. This situation must b-e remedied. Thlere is also an urgent needu for uniformi>ty in tax ir.centive u ~ L ~UAU.. LLu I ~ aL~J~1 LLL L. L.LLP UA. LLUL .A.) LLL L.& ~L L .~1L policy and for ensuring that profitable hotels contribute to the public _ __ _Te r _ k___t____c L w ~L14 - __ L _._1____L_ _______ _ L _ S purse. LI dL1U Wl:Le L tle pofablity Uo LIUL%o' ca11 ue ±iicrerUseu', their coni- tribution to the economy will hinge on whether a substantial proportion of tnese proiIts can be taxed. 24. The soundest basis for estimates of future infrastructure require- ments in connection with tourism in Antigua would be a careful analysis of experience in the recent past. Unfortunately, this approach is severely hampered by the quality of the statistical data available. 25. It is interesting to note the different implications for infra- structure investment of luxury and middle-income tourism. For any given number of rooms, a modest hotel requires substantially less indirect infra- structure investment. 26. Very little information is available on the demands made by tourism on port and airport facilities, the costs to the economy of supplying these facilities and the revenue generated. However, it appears that the port makes a loss, while the airport is profitable. The sound financial position of the airport may be due to transit traffic as much as to local tourism, however. I. ACCOUNTING PRICES FOR TOURISM PROJECTS IN ANTIGUA 1. Estimating accounting prices for the Antiguan economy is in some respects easy and in others extremely difficult. It is relatively easy to estimate the accounting prices for traded commodities because it is a very open economy with no quantitative controls on trade and a fairly uniform tariff and tax structure. The difficulties, however, are two-fold. There is the difficulty of obtaining information of the breakdowns, especially of nontraded goods, into cost components. Antigua has not had any national income data to speak of since 1968; the financial accounts for the public utilities are in a very poor state; there is no hard information on the employment situation and virtually no information on the size distribution of income. There are also difficulties of a more philosophical nature in estimating shadow wages and the accounting rate of interest in the economy in which the obiectives of Government policy are difficult to discern. 2. The obiect of the analysis is to reduce all costs and benefits to one numeraire which is taken to be foreign exchange in the hands of the Antiguan Government. Movements of foreign exchange can. therefore, be taken at face value. But in order to arrive at these it is necessary in several cases to allow for the existence of Government taxes and subsidies which. in general, are not real costs or benefits. Furthermore, the purpose of the exercise is not to calculate the impact of Droiects on the forei2n balance but to calculate the impact on welfare. So, it is necessary to value incomes crentetd bv the nroiects. sav for workers nreviouslv unemnloved- in terms of the numeraire. We now proceed to explain how accounting prices were derived fnr Antigua. The mnet act-pntah1e way of doing so iS to r.nl lAate a net of prices which is mutually consistent using a general equilibrium method of solution. The data, however, was i-ahle tn support this approach, so an approximation to equilibrium was derived by making successive iterations ,urtil a roughnly cor.sister.t st of pr4tac were Tproduced 3. The internal prrice of a traded commodity differs from its rc.if= price by the extent of port handling charges, import duty, excise tax, whole- -~l ..#A ..a4v4 ,,a,.n ."a 'rr va~.7-" ^io.- *ha-i " ,1.-- ^f m t-rnAaA e'o,mnfiA' ti, saea.d retail _:gir_s To work --- the- -cco1tig pMice ofataddcmmdt it is therefore necessary to know the value of wholesale and retail margins inr terms of the numera4ire. The 4issIon was of the opir4.i that t on-ly practicable way to proceed was to estimate the accounting ratios for whole- sa.le ar.d retail margir.s, tLo work out thte accour.ting prices of traded nan-A .odities based on that and then to check the estimates by looking at the cost structure oL' distribution. If the estimates were not credible, JLit W-ould bUe necessary to rework the calculations. As it turned out the estimates were rougly cor- rect. We asseeu, on the basiLs oL sucii LnLotua'Lon aS was available, tha port-handling charges, wholesale and retail margins were 3%, 15% and 20% respectively and that UthirL rspec;ive accounting ratios were U.8, U0. and 0.7. Import duties were calculated by discussions with the Customs' auth- orities. Wnen there were two rates, preferential and non-preferential, a weighted average of the two was used. Where the duty was specific, the ad valorem equivalent was calculated with the help of the Customs' authoritLes. We now demonstrate using imported Scotch whisky for illustrative purposes, how the accounting price for traded comnodities was calculated. Table 1 Item Value Accounting Ratio Accounting Value Scotch Whisky 179.4 x 136.53 Retail Margin 29.9 0.7 20.93 ..olesale Margin 19.5 0.8 15.6 Port Charges 3 0.8 2.4 Consumption tax 6 0 0 Inport duty 24 0 0 Foreign Exchange Cost 100 1 100 Starting with the notional value of a bottle of whisky of 100 we add on addi- tional costs to arrive at the final retal prive. TlUina estim.ated accounting ratios for the various items, the accounting price of whisky was calculated. The accounting ratio x for whisky is equal in the above example to 3 = 0.8. 179.4 In tle following table we present the accouting prices flor certain traded commodities, worked out in the manner explained above. Table 2 Accounting Ratios for Commodities Meat (fresh frozen) 0.924 Meat (canned) 0.92 Dairy products 0.84 Fresh fish 0.907 Canned fish 0.91 Fruits & vegetables 0.89 Cigarettes 0.62 Footwear 0.81 Clothing 0.75 Vehicles 0.73 - 3 - L, In~T acl-4mnl4atin th scia4l cost- of labor 4n Anti,,ua the fol1o--.- elements are of crucial importance: (a) There is known to be a considerable amount of unemployment and underem.ployment, though there has been no survey of the employment situation since 1951. The alternative marginal pFroAuct of la-or is tberefore certainly cons4Aerably b-elow ~.t n. I j .Lu'. ,o L A. .1 C . 1 LA 1* j4L~ ala .L L)J. W unionized wages. (b) On the other hand, the supply price of labor is very high. q.his is due lo several factors. Tlough -le marginal product .LL.L~ £L LU~ A L& LIAL LLI LJ6i LLL1 UI L J±ld L L1U of labor is low, incomes are higher than that even for the un- andU underemployed Uecause of rer.';Ltances from reW Lat-Lves abroad and because of support from those who are employed. aILtgUuanS WUU' aiso appear to VaUe leisure highly a[llU to exhibit a marked disinclination to work at certain kinds of jobs. hne question then arises whether the alternative marginal product of labor is the appropriate shadow price, i' tne supply price of labor happens to be much higher. This question cannot be answered without a prior value judgment concerning whether individual attitudes to work are to be respected in making decisions from a social or national point of view. (c) Thnere is the further question of the relative valuation of investment and consumption. The use of a shadow wage rate lower than unionized wages would encourage employment. But given that the wages that in fact have to be paid are unionized wages, higher employment reduces investible surplus in that this exceeds marginal product. The link is quite obvious in the case of a publicly owned project; it is less obvious but nevertheless important even in a private project because the government can tax private profits and because capitalists' propensity to save and invest is, in general, higher than that of workers. Now, to the extent that invest- ment is more valuable than present consumption, higher emplov- ment at unionized wages has undesirable effects. The optimum employment (and shadow wage) therefore depends on the trade-off between consumption and investment. 5. The formula given below brings in all these coTsiderations. Suppose W = the unionized wage rate f is the "consumption conversion factor" to convert consumption into foreign exchange c is consumption out of w m is the alternative marginal product of labor -4 i is the supply price of labor in consumption units 2 is the vailuation society puts on effort s is the relative price of savings in terms of consumption w* is the shadow wage rate w* tc {(,c-m) - (l-m Thus the shadow wage rate is equal to consumption out of wages minLus the benefit from the increased consumption plus the cost of the increased individual effort involved as valued by society. We can now put some rough numbers. w is about $40 per week which is the minimum unionized wage for unskilled labor in Antigua. The marginal product of labor (m) is taken to be a quarter of w about $10. To put m = o would be too extreme. There is a certain amount of part-time work in agriculture; there is also house- hold work by women which has a value, even though national income statisti- cians choose to ignore it. c is consumption out of wages which is taken to be $38, assuming a propensity to save of 5%. Though m is only $10, the in- come of an unemployed Antiguan is likely to be more than that. it could be $20 as a result of family support. Nevertheless, the employment of an extra man would increase total consumption by roughly c-m or $28. The reason is that the newly employed man presumably loses his previous financial support. So his consumption goes up by S18, the consumption of those pre- viously supporting him (or those they now choose to support) goes up by 520, but there is a product loss of $10. So the net increase in consumption in the economy is c-m = $28. The supply price of labor in Antigua (1) is much higher than m and should in our opinion be put as high as S30 in consumption units. Discussions with informed people suggested that it was qulite im- possible to hire Antiguans for wage employment for anythin,g lower than that. 6. This leaves the estimation of f, s and . f is simply the factor required to convert the above quantities into our numeraire which is foreian exchange. We estimated it hy using consumer budget data for 1959 given by Carleen O'Loughlin. 1/ No more recent figures were available and different classes,of consumers were not distinguished. But we had to malce do with what information there was. Table 3 gives our calculation of the conversion factor for consumers' expenditure is general. The accounting ratios for traded goods were calculated as explained above. For non-traded goods and labor services, we had to estimate and then recheck when the accounting ratios for these were calctulated. We had to perform a couple of interations to achieve rough consistency. Thus the accounting ratio for consumers' ex- npnd- t,rp turns nut to be about n 8L4 somewhat below the tvniral acconnting 1/ Carleen O'Loughlin, National Income Statistics 1954-64, Institute of Socinl and Economic Repserrh (F CGrihhben)j 1965. Accounting Ratio for Consumers' Fcpenditure (f) Col.1 Col.2 Col. 3 Col.2 x Col.3 It_m L"rei hcW vicountin Ratio Weiged .A.R. Meat .051 .92 .°47 Al.. Al. Dairy rruuu.b nefn 01 nLA Cereals .105 *90 .l9k Fruits & vegetiables .075 .89 .06 Sugar & confectionary *91. Alcoholic beverages .038 .80 .030 Tobacco & cigarettes *017 .011 Other foods .032 .89 .028 Footwear .033 .81 .027 Apparel *089 .75 . 067 Household goods .058 085 .V49 Personal goods *035 .85 .030 Personal & Prof.services .60 .75 .045 Transport .047 .8 .037 Rents .129 .8 .103 Purchases fram Uovt. .041 .8 .033 Purchases from overseas .043 i .043 ¶Ibtal. 1.000 0.84 -6- at1o Lor tralelU goods, adlu somewha;t -a zLbove the typiLca'L ratio flor non=traLdedu goods. Incidentally, we may note here that the weighted ratio of c.i.f. __l_ _ , ____J ___ C^ X_ 1._ - = __ _ _-=_C_______ J _- -3 _- A_ _: -. _ - -_ priLces to tax aU tar'LL' X nclusi±ve p fUr XmpotLed gUoUb LL rA1L.LiUd 1 1967 was approximately 0.86, the weight being the proportion of the value of each category of import in the import bill. We now come to the estimation of the more ethereal parametersch and s. At one extreme, we could putcv = 1 which means that the Government regaros eacn individual's supply price of labor as the social cost of employing him. This is an individualistic view- point. If an individual does not want to work ror anytning under $30, that means that $30 is required to compensate him for the various disutilities that he thinks attach to working. If so, then that is the welfare cost of employing him. At the other extreme we could putO( = 0 which would imply that the Government attaches zero weight to the individual's views regarding the cost of effort. It considers the material product lost to be a cost but chooses to ignore the extra effort which the individual dislikes. There are other reasons why it might choose to ignore the high supply price of labor. The unemployed may prefer to remain in that state; but the government may be concerned about developing work habits in the population; it might also be concerned about the social and political effects of having a large section of the population unemployed. If it gives some positive weight to the individual's supply price but less than the individual himself does, then OCa- 71.This is what we think is the most typical situation. We now come to s, the relative price of investment in terms of consumption. This is a complex question. The need for the government to raise more revenue is quite plain given the serious debt crisis that it now faces. There is another way of approaching the matter. Consumption per capita in Antigua has been growing at the order of approximately 1-1/2% p.a. over the last 15 years (though in the immediate past, there has been stagnation). It would not be too far- fetched to suppose that a 1% increase in per capita consumption reduces the marginal utility of consumption by 2%. Then, the consumption rate of interest (which equals the growth rate of per capita consumption times the elasticity of marginal utility of consumption) is about 3% even, ignoring pure time dis- count. Inclusive of a time discount, we could put it at 4%. If we think that projects can be found which can yield 8% of consumption in perpetuity then the value of investment in terms of consumption is 2. Whether such opportunities exist is difficult to say. Our own position would be that if hotel occupancy rates could be increased by suitable promotion and price policies, such projects could certainly be found in the tourist sector. There is good reason to suppose that even in agriculture if the government invested wisely in infrastructure and research, there are suitable prospects for livestock and vegetables. As for industry, there seems no reason why Antigua should not be able to have labor-intensive international sub-contracting if the rate of growth of money wages can be controlled. All these are big "ifs" but planning for development must assume that the government will try to pursue productive policies. The case for putting s 2 is quite overwhelming in the short-run, as a simple logical conseauence of the government's budgetary difficulties, and the drying up of foreign investment. The case for having s considerably above 1 for a substantial length of time is based on the idea that not only does the Antigua Government want to ride out its debt crisis but is also interested in the growth of the Antieuan economy. -7- 7. Using these figures, we have calculated - for different assumptions concerning s and \ using the formula given above.w Table 4 cL = 0 1/2 L 1 s = 1 0.21 0.42 0.63 s = 2 0.5 0.61 0.71 s = 3 0.6 0.67 0.74 We think it would be unwise to choose an accounting ratio for unskilled labor of less than 0.75 in the short-run and even in the long-run, of less than 0.6. In fact, in our calculation we have used 0.6 as the ratio, to be as favorable to employment as is possible without being irresponsible and implying too low a shadw wage. More subtle questions may be considered. It can be argued that wshould rise with w because the higher w is, the greater is the increment w in consumption as a result of the employment of an extra man; since the marginal utility of consumption declines with increasing income, the higher the union- ized wage in an industry the greater should be the discouragement to the em- ployment of this extra man because the Government could redistribute the money to a large number of poor people which would result in a larger increase in social welfare. This point can we think be ignored in Antigua because the employed are quite efficient redistributors of income. Another point which might be brought up is that shadow wages should be set low because increased employment simply implies less repatriation of profits. At the moment, re- patriation of profits is not significant in Antigua because tourism is not particularly profitable. Even if it became profitable, there are possibilities of taxing profits. And some repatriation of profit is after all necessary so that foreign investment may continue - otherwise, saving and investment would become even scarcer than they are. We see, therefore, no reason to change our estimates of shadow wage rates made above. 8. We assume that the pay of professionals and managers adequately reflects their marginal products at market prices; indeed their marginal products might be higher than their pay would suggest. Since the latter statement might well be true, we think it approximately correct to use an accounting ratio of 0.84 for their salaries which is the same as the accounting ratio for consumption expenditure, even though they no doubt save more than unskilled workers and are also taxed more. As for skilled labor. the assumD- tion is that the employment of an extra skilled man, involves (though possibly at several removes) the emnlovment of an unslkilled man from the ranks of the underemployed. But the latter would have to be trained and that has a cost which the emnlnvor ean uisiallv not fully recover hv naving lower wages Hence, for skilled and semi-skilled labor was use an accounting ratio of 0.75 (which lies between 0.6 ard 0.84). If the wage and salary bill cannot be divided between that of professionals and skilled workers we use an accounting ratio fo: the two t0gether of 0.8. - 8 - 9. For a- co,-,itry- that relies heavily on foreign borrowin, itsa to reason that rates of return on projects should equal the marginal rate at wh.ichi foreign funds can be borrowed. Even 4if w ar thn4k,i4ng of f-re4n venture capital, a project should be capable of earning some minimum rate of retu-..D elow w-hich. t-h-e suppl-y of foreign ca4pa1 -.ldA Ady -p. n he ps given world inflation, Antigua has been able to borrow at real rates around e The ositouatio...p14C nisa.. V L.,.i.ateA. As a-lrad.o the A ,tiguan debt situation is quite difficult and interest rates are in any case rising sharply on world capi'tal mIarkets. It i s motr u.ieyta .iu ol o ~LL~L~± VIIWULlu .~JL~. kdLV. LA. S. C LUS L.L4. N" Ly 4.IatL &LL..Lgua cou.LU noLw obtain any finance at all below real rates of 6-8% p.a. In the short-term, Llth siLuatioUn coUULU ue eVein wurse if Xay sUIb stantiaL sums VI mlUnILVy calre inIVU±VeU because the more attractive sources of credit would dry up and Antigua would have to turnL to other suppliers. There is another approach to the accounting rate of interest. We have already suggested that the consumption rate of interest is in the region of 4%. We [Iave also suggestedu that thLe value o0 savings and of government money in Antigua is high in relation to consumption, particularly at the moment when DuageLary aifiLculties are consiaeraDle. Assum- ing that these will disappear in due course and that the economy will grow faster, we can expect the scarcity of savings and government money to deciine over time. Since our numeraire is government money, the accounting rate of interest must be above the consumption rate of interest by the extent of the rate at which s expected to decline. ie. ARI CRI + s s Assuming that it will take 20-40 years for the scarcity of savings to disappear, the ARI should be 2-4% above the CRI. i.e. it should be 6-8%. In the short- run, assuming that the present budgetary crisis disappears, s could fall faster and the ARI should perhaps be set even higher than 8% to favor projects yielding quick returns. But we assume conservatively that the ARI should be 6-8% in real terms. 10. Estimating accounting price ratios for nontraded goods and services in Antigua is a task of some difficulty. Cost breakdowns are not available, though it must be admitted that a few more man-weeks would probably have resulted in improved estimates. We therefore had to make do with what information there was in the shape of an inter-industry transaction matrix for the Antiguan economy in 1963 compiled by Carleen O'Loughlin. The matrix showed the gross output'of ten sectors and the division of the gross output of each sector into payments to the other sectors and to four "primary inputs": households, profits, government and the rest of the world (i.e. foreign exchange). The commodity coefficients matrix was inverted and by standard techniques the direct and indirect primary input requirements of each sector were worked out. We present below the total primary input requirements of those sectors which are non- traded. -9 - Table 5 Construction Other Services r_ T%4 -~~~~~~~~~~-4 c~t n.--- -Z &LOI,.L Ai:i Fina.ce ' & ReUL o. Engineering bution Transport Insurance Entertainment Dwellingi Primary Input Households .32600 .12432 .22158 .61802 .40179 .07606 Profits .08687 .06880 .37467 .25269 .51729 .81172 Government .03620 .13402 .14909 .03777 .03128 .05355 Foreign Exchange .55087 .67561 .25462 .09148 .04961 .05865 11. Some comments are required on the shadow pricing of each of these primary inputs to arrive at the accounting ratios for the non-traded goods. Households: In each case, we estimated the proportion of the wage and salary bill going to skilled labor and managers and that going to unskilled labor and applied the accounting prices for these categories worked out above. The proportions had tb be based on conversations with informed people who had some knowledge of the sector and in some cases on pure guesswork. In general, the bias, if any, was such as to overestimate the proportion of the payment to households going to unskilled labor. The highest proportion going to un- skilled labor was 50% in transport, the lowest was 10% in finance and insurance. Profits: When a project demands a nontraded input and if there is full capacity in the industry producing the input, it is assumed that its supply will increase. One can then either include the capital cost of increasing the supply of the nontraded input in the capital cost of the project itself or one can charge the project an "appropriate" price for the input which means a price inclusive of a capital charge. This is the course pursued here. If the profit element is no more than such as to cover "normal profit" on the capital stock at the accounting rate of interest and to cover replacement costs required to maintain capital intact (insofar as these are not already included in the operating costs) then the whole of it is a cost, though mul- tiplied as usual by the standard conversion factor of 0.84. If the profit element is greater than that, the "surplus profit" which accrues to the entrepreneur has to be valued in terms of our numeraire. It has been assumed, on the basis of data given in Carleen O'Loughlin's monograph, that 20% of the profit is saved. The 80% which is consumed is counted entirely as a cost because it accrues to rich people. Of the 20% which is saved, only 20% is counted as a cost (In other words, it is assumed that private saving and invest- ment is worth 0.8 times government saving). In short out of surplus profits - 10 - ('bfl'V\f 0 = 1l I 4 J 3 0/A W . t.1 (0).8/.. = e 16% is a benefit aLdIU 84.+ i-s a cost. TILL it the Ac-.Li J.if surplus profits is 0.84. The only sector in which this procedure was not lo.L'oweu was traLLnpur wLLhre it cX L LreasoulLU±| y ue assuiuueu Lltda Llth suLpi'Us profit accrues to tax-drivers and the like who are not very rich. In that case, it was assumed that only half of the surplus profit is a cost. Government: Normally in cost-benefit analysis government receipts are not treated as costs. In this case, however, a different procedure had to be followed. The reason was that in the 1963 inter-industry table, public utilities were unfortunately lumped in with the government sector. So part of government receipts are simply a return to investment in public utilities. Inspection of the government sector in the inter-industry matrix revealed that roughly 50% of government receipts consisted of import and excise duties and company tax and the remainder was public utility income. Therefore, half the government receipts were counted as a benefit and the rest multiplied by the standard conversion factor of 0.84 as a "normal" return to public utility investment and hence as a cost. This was done because it is perfectly clear that there are no "surplus profits" in public utilities; on the contrary public utilities almost certainly do not break even at accounting prices. Foreign Exchange: This item was of course taken to be a pure cost. Using these methods, the following accounting ratios were obtained for the above nontraded goods and services: Table 6 Sector Accounting Ratio Construction and Engineering 0.87 Distribution 0.83 Transport 0.73 Finance and Insurance 0.865 Other services and entertainment 0.78 Rent of Dwellings 0.8 Perhaps a comment is in order on the accounting ratio for construction which might be thought to be rather high. The reason for its high accounting ratio is that a great deal of construction was connected with the hotel industry so that it benefitted from import duty relief. 12. This brings to an end our calculations on accounting prices. Some remarks are in order. The method used above to calculate accounting prices for the nontraded goods is most unsatisfactory and had to be used only as a matter of last resort because cost breakdowns were not available. It is narticularlv imDortant in evaluating tourism projects to calculate accounting - 11 - ratios for public utilities as accurately as possible and we have very little information on them. Another point that the reader may notice, is the dis- screpancy between the accounting prices of nontraded goods worked out above and those used for the same goods in calculating the conversion factor for consumption expenditure. Complete consistency could only have been assured by a simultaneous solution. After one or two manual iterations, the remaining discrepancies were too minor to affect any results and were therefore ignored. - 12 - II. ANALYSIS OF THE IMPACT OF THE TOURISM SECTOR ON THE ANTIGUAN ECONOMY 13. Sectoral analysis is generally restricted to describing the structure of the industry in question and making various policy suggestions on the basis of an educated judgment. This function is, of course, extremely important and our report does concern itself with this task. Is there anything further that can be attempted, especially at a quantitative level? Such quantitative anal- isis at the sectoral level is difficult in the best of circumstances but no progress at all can be made without sorting out rather carefully the questions that should be answered. Some of these questions are tractable, others are extremely intractable especially if the sector in question is large in rela- tion to the national income. In this chapter, we propose to clarify some of the issues involved in the context of the economy of Antigua. Consider the following questions, all of them related to each other but nevertheless dis- tinct: A. What is the contribution made by tourism to GDP and GNP in Antigua? B. What would be the effect of an autonomous increase in tourist expenditure on GNP in Antigua? What, in other words, is the size of the "tourist multiplier"? C. Would increasing the size of the hotel sector be justified at the margin, taking into account the costs involved? Here one might want to distinguish between hotels owned and financed by private foreign investors and hotels owned by the government with the capital being borrowed from abroad. D. Does some DroDosed Dlan for a large exnansion of the tourism sector - say the kind of plan contained in the Zinder Report 1/ make economic sense? We shall consider each of these questions in turn. What is the contribution made by tourism to GDP and GNP in Antigua? 14. In 1964, the direct value added by the hotels sector constituted about 9/ of GDfP BHut this is rlparlv nn indPrqt-sqtaPmPnt of the cnnt-ribltioin of the tourism sector to GDP for two reasons: (a) Tourists spend money outside hotels and thereby contribute dirpettly to valuetip -atddd in other srcators. 1/ H. Zinder and Associates, The Future of Tourism in the Eastern Caribbean, WnQ.4 nt n r l 9ft - 13 - (b) The sectors in which tourists spend their money directly, themselves make payments to other sectors because they purchase inputs from them. For example, tourists spend money on food and drink in hotels. This does not creat value added in the hotel sector but it creates some value added in the distributive trades sector. The same argument applies to the payments by any sector to another which can be attributed to tourist expenditure. 15. For 1964. Carleen O'Loughlin furnishes us with the following data: 1/ (i) the sectoral composition of direct tourist expenditure, (ii) an inter-industry transactions matrix which tells us the value of sales and navments by each sector to every other sector. From (ii) it is possible to work out by standard input-output techniques, the direct- nnd indirpett value added by everv sctortr- On the nnt tmreaRnnahlh assumption that the proportion of the total value added by each sector which * an e att-ribuhtl tn tonurism is- the Ramp as the nrnnnrtinn nf its s%al1pe which can be attributed to tourists, we have calculated that the direct and indirect contribution nf ton,rism to GDP in Antigua in 1963WA w 22.5%. Tn 1968, direct value added in the hotels sector went up to 14% of GDP. If we make the further issmnption. thst diret-t value adeo by the hotels sector as a proportion of total value added by tourism remained constant over the period then. tota val ue added b- tourism would ha1,,- been arnd 33.6X of GDP in 1968. Even this is an underestimate of the contribution of tourism t%o GDP for the *folIWflo6ing two reasor.s. Fis-ly , i.L +the aIbov c.alculat.ss , value added in the construction and engineering sector attributed to tourism w takenr to be sinpl,, that nrising frorm the currernt expernditures by other sectors on the output of the construction sector (i.e. expenditures for mainrtenance and replacem.ent). Tn actual fact, meach of the value aAAeA ir. ILL~~~~~~~~~~~~LL1L~~~~~~~~~~~~~~&&O1AL~~~~~~~~~~~~~~~~~~~ OALU ~~~~~~~~~~~~~~~~~~~ . i~~~~~~~~~~~~~~I ~~~~~~~~~L.LU~~~~~~~~~~~~~~J. LO~~~~~~~~~~~~~~~~.L ~~~~~~~~~~~~ wJJ. ~~~~~~~~~~~~~~~~~~~ VO.LU~~~~~~~~~~~~~~~~~~~ O~~~~~~~~~f.l. L L . %.I4 bL. the construction and engineering sector is the result of new investment in h-otel building and' related activities. Ir, !968, value all-ed in -le cor.= L I L. J UA.L.L U.LLL IU L L. %A U a~A. L L.L ~~ .L I7UU ~ VLU ~UUu L. L-L&~I.L struction sector was 30% of GDP and that attributable to investment in tourist facilities was, at least, 15% of GDP. This gives us a total of 48.6% of GDP'attributable to tourism secondly, the above calculations have excluded thle contributiosL, thilat touriLsm makes to GDP buy increasir.g tlhIe tax revenue of the government and permitting it to sustain a higher level of expenditure. This could easily add another 5% LtU Lts contriUL.LVL L to . UAll these figures are rather rough. More recently, the importance of tourism-induced construcLion has declined -w`hich, wouLUd redUuce our figure; on thiLe other hLl rand, export agriculture has totally collapsed, raising the proportionate contri- bution of tourism. All in all it is very unlikely tnat tourism generates 1/ See Carleen O'Loughlin, National Income Statistics, Antigua 1954-64, Institute of Social and Economic Research (E. Caribbean) 1965. - 14 - less than 40% to 50% of GDP. There is no reason to believe that it is any diffprpnt as a nrnnnrtinn nf (NP - repatriation of profits is iniignifirnnt. The contribution of tourism to GNP at market prices might be even higher becaue tniiricts npv mnrp in taypq than thev rreiv in sbihQidi es. 16. One might, howaever, be interestod in a smeThat different question. One might ask what is the net contribution of the tourism sector to GNP, net in the sense of over _n above what the releva.t factors of production would have produced anyway in the absence of the tourism sector. Frank 1 A,-, .S_.LU _f E -- ,. . .f. O S.L t- . ..* .x;..} G.. a,, where the tourism sector is much smaller. 1/ In Kenya, if there were no o.uLrsU, LLLt LrLeVlea.t factors of production cou.LU Ue -abUsorblelu in other sectors, albeit at somewhat lower rates of remuneration. It is possible to make an eUucatedU guess to allow fLor Le intra-marginal in.cLIeasbe iL faLc-o prices resulting from tourism itself and to estimate the true opportunity costs of *LLe Ld(L6ufs uL pLuuuCL.LUII t:luij±u)U I1L LUULLbill. 1IL- LLtL CULILLLUULIULL UL tourism then is the value added minus the true opportunity costs of the 'Lactors. In Art'Lgua, it would bue al1most a non.senlse questi'on. If' the tourjsism ceased tomorrow, the effects would be so large that it is almost impossible to put arty vaJlue on fwiat the Lactors of1 production conce-rned woulU earn anyway." All one can say is that a very substantial proportion of the direct aIiU iLnudLrect value added Luy tourism wouLdU bUe=Lost 'i' tour'Lsm were to cease. The "Tourist Multiplier" 17. The multiplier effect of tourist expenditures has been the subject of wild assertions. The Zinder Report, 2/ for example, claims that the tourist multiplier in the Eastern Caribbean is 2.3 i.e. that $1 or additionai tourist expenditure would generate $2.3 of extra national income. This estimate is based on a major analytical error. What the Zinder Report has estimated is the increased local turnover resulting from increased tourist expenditure. In other words, the value or ali the transactions resuiting from an extra tourist dollar have been estimated. But the impact on national income is measured by the value added during the transactions, not by the total value of the transactions themselves. 3/ 18. Estimating the impact of an additional dollar of tourist expenditure requires - first that we estimate the primary impact on value added, in other words the first round effect on domestic income and second, that we then pro- ceed to estimate the secondary impact on value added, in other words, the 1/ F. Mitchell, "The Value of Tourism in East Africa," East African Economic De..4ew 1970. 2/ The Zinder Report, op. cit. 3/ The Zinder Report follows up with further erroneous estimates on this h o.o increase in incomes resulting from spending and respending of the primary increase in income, taking due account of leakages from income flow along the way. For any accurate estimation, we have to make realistic assumptions concerning the opportunity cost of the domestic factors employed, the extent of excess capacity in local industries and the magnitude of the leakages from the income flow. We begin by making the most generous assumptions possible for obtaining a high multiplier. Assume (i) that newly employed local factors have zero opportunity cost, (ii) that there is underutilized capacity in all local industries and (iii) that the only leakages from the income arise from imports - in other words that any increase in government revenue and in savings is spent as public consumption or investment expenditure. These are clearly optimistic assumptions. If (i) and (ii) have any relevance at all, it would be during the off-season. 19. To work out the Drimarv impact of increased tourist expenditure. we need to know (a) the sectoral composition of tourist expenditures and (b) the direct and indirect imnort content of the activities toward which the increased tourist expenditure is directed. Carleen O'Loughlin provides us entimates for both these for the year 1963. 1/ The sectoral composition of tourist expenditures is given in the following table: Table 7 Expenditure by Non-Residents per Dollar Spent 1963 Purchases 28 cents Taxis 8 cents Finance, Insurance 2 cents Hotels 43 cents Bars and extra meals 5 cents Entertatinment & professional services 2 cents Direct rent 2 cents Direct wages 10 cents $1.00 1/ Carleen O'Loughlin. op. cit. - 16 - We also need an estimate for the direct and indirect import-content of each of the items in Table 7. This can be obtained by inverting the input-output matrix for 1963. For any sector, the direct and indirect import content is given by multiplying the column vector of inverse coefficients by the row vector of direct import coefficients. (We have already made use of such a calculation in the last chapter for deriving shadow prices). For 1963, Carleen O'Loughlin has, conveniently enough, done this calculation and the results are as follows: 1/ Table 8 Allocation of each $ Spent to Local Incomes, Government and Imports Each $ spent on Local Incomes Government Imports (cents) (cents) (cents) Purchases 19 13 68 Taxis 55 15 30 Hotels 51 10 39 Services & Entertainment 75 10 15 Rent wages 1.00 - - Licenses, fees, etc. - 1.00 If we weight the breakdown of tourist expenditure given in Table 7 by the proportions of local and import coefficients given in Table 8 we arrive at the result that the non-resident dollar is allocated according to the following breakdown: Local income 49%, Government 10%, Imports 41%. 20. For our purposes we can conclude from this that the primary impact of the additional tourist dollar on local incomes (including government revenue) is $0.59. We must now consider the impact of the spending and respending of these incomes. Carleen O'Loughlin provides us with a breakdown by final use of household expenditure in 1963. Using this data, along with the inter- industry transaction matrix. it is apparent that the direct import content of household'expenditure is 46%, assuming all saving to be invested. The Droportion going to government is 19% and the rest goes to local incomes and other local industries. 2/ If Government revenue is assumed to be re-spent aRnd its imnort-content worked out as also that of expenditure going to local industries, we end up with a final share-out of 42% of household expenditure endling *n as Inral value-added and 58% as imports. Since, we are assuming 1/ See Carleen O'Loughlin, op. cit. 2/ See Carleen O'Loughlin, op. cit., p. 14. - 17 - that imnorts nre the only leakage from the income flow; simnle national income analysis tells us that the primary impact on value added must be di- videdby 0.58 to apt the tota1 i*mnct on national inrome of additional ex- penditure. In other words, the multiplier is = 1.725. But this, of course, ic not- tha "trii rict mill tininli r" fnr whn ilrict .vnanditilrca ininrra:cza by $1, the multiplicand in working out the increase in national income is the primary impact on local incomes L1hich 4i 0.59 as argued above. Hence the "tourist multiplier" = 0.59 ( ) which is approximately equal to one. It should be noted that this -ultiQi 4_r -14lie to expendit-re by al tourist- which includes those who stay in apartments, villas, etc. The multiplier for those who stay in resort hotels is o10w er nbecause thei4- expendu has a higher import-content. Table 9 below gives the composition of expenditure by tourists staying in 4eso:t hotels 4n 10.63 thle- Atai 4o agu- 4i f ro Carleen O'Loughlin. 1/ Table 9 Expenditure by non-residents staying in resort hotels per dollar spent. lo t L3 (cents) Expenditure outside hotels Purchases 26 Taxis 12 Finance, Insurance 2 Entertainment & professional services 2 Expenditure in Hotels Bills 51 Bar and extra meals 7 $1.00 Weighting this expenditure by the import-contents worked out from Table 8 we get a primary impact on local incomes from an additional dollar of ex- penditure by tourists siaying in resort hotels of $0.555 and a multiplier therefore of 0.555 ( . -- ) which is slightly less than one. 21. One of the assumptions on which the multiplier given above was based was that imports are the only leakage; government expenditure and 1/ Carleen O'Loughlin, op. cit. investment by residents were taken to be endogenous, the only exogenotis items being investment by foreigners and exports, visible and invisible. This is not a particularly natural Keynesian assumption. The normal Keynesian assumption is that government expenditure and investment are autonomous and taxes, savings and imports are leakages. If we operate on the normal Keynesian assumption, the multiplier would be lower. Consider the case of an increased spending of $1 by tourists in general. From Table 7 and Table 8 we know that the primary impact on local incomes would be $0.49. (In fact the primary impact would be 0.47 if we exclude direct taxation on local incomes.) To get the tourist multiplier, we must multiply this by c m where s, t and m are the proportions of local incomes going in savings, taxes and imports respectively. (We assume that average and marginal proportions are equal.) Data from Carleen O'Loughlin suggest that savings and taxes together account for 18% of household income-use and that the direct and in- direct imports by households (this time leaving out the effects of the respending of government revenue and the re-investment of household savings) account for abQut 46% of income-use. So the tourist multiplier is equal to 0.47 ( )= 0.75. The assumption concernine the endogeneity of government * WeiAiture and investment is not, however, the most critical one made above and could in any case be defended as being fairly realistic in a less developed economy. 22. The more critical assumptions concern the complete unemployment of the indidt-inrl 1orn1 fnrtc,rS Pmnl oved and the nresence of substantial excess capacity in local industries. These assumptions could be seriouslv misleading. It has already been argued in Section B that even local labor has some positive opportunity cost. For unskilled labor we have put it at a qu-rter of the wages i4A 4in f-the uinionizedl c't-nr for szkilled lahor it would normally be even higher. In so far as there is an opportunity cost to the employment of extra labor, the primary and secondaru incrrecp in wagp incomes cannot be counted entirely as an addlition to national income. And -n some sectors this extra employm.ent would be negligj,ble inless there is underutilized capital. For example, even if there exists underutilized capacity in hotels, puchse by_ hotel ---m loca s^urce ma s-loe directly into imports in the case of traded goods. In the case of nontraded goods, such as electricity or water, iLf there is no unutilized capacity, then either capacity must be added to or some other consumer must be deprived at the margLn. In either case, the effect on impLorts is likely - be m,u greater than the fixed input-output model would suggest. Or to put it in different words, ipargial import coefficients are l ikely to be much than average import coefficients. Furthermore, in so far as profits increase, we must consider the repatriated part as a leakage. 23. All this suggests that the concept of a "tour-ist moultiplier" 'has relevance in Antigua only in the tourist of'-peak season which is also likely to be the time when local industries are working below full capacity. Even in this situation, taking account of repatriation of profits and putting some positive opportunity cost on local labor the tourist multiplier is very unlikely to be higher than 0.7 or 0.8, even if we consider imports to be the sole leakage from the income flow. During the peak season, the primary value - 19 - added from an extra dollar of tourist expenditure could be as low as 0.2 because of full capacity in domestic industries and the marginal propensity to import out of increasqd incomes could be as high as 0.8 which would give a multiplier of 0.2 ( )=0.25. We have already begun to put more weight on the multiplie&oncept than it can bear. The idea of tourist ex- penditure on goods and services being met by an expansion of domestic capacity takes us into a long-run situation which the multiplier cannot satisfactorily hendle. Another point, which it can even less satisfactorily, is the extra private foreign investment which might be induced by the rising activity. Such induced effects might be important in some cases, but they take us very far from the multiplier idea. 1/ Social Cost-Benefit Analysis of a Small Increase in the Size of the Hotel Sector 24. We now consider the question of whether a small increase in the size of the hotel sector is desirable from the national point of view. This is a somewhat different exercise from the evaluation of a marginal project. Pro- jects might be very different from each other. A particular kind of project might be socially profitable, another might not be. What we seek to examine here is whether a project which has the average coefficients of the hotel sector would be socially profitable. It is, in effect, a shortcut way of assessing the desirability of expanding the hotel sector. 25. The methodology we adopt is to examine the cost structure of the hotel sector in a particular year and to use that to determine the social average cost of operating the sector. We then look at the social average revenue earned by the sector. This enables us to determine the current social average profit being earned by the sector. These calculations provide a basis from which the social marginal revenue and social marginal cost of expanding the sector can be estimated, thus giving some clue as to the social marginal profitability of doing so. The most recent year for available data was 1967. so we had to be content with using that. Not only will it serve well enough as an illustration of the methodology but we also believe that the calculation continues to have some relevance. Occupancy rates in the hotel sector have not improved since 1967 when the average annual bed occupancy rate waR 32.7Z: in fact. occunancy rates in later years have been rather worse. Such evidence as there is indicates that in real terms, tourist expenditure has not increased sivnificantly. Since suDDlV of hotel beds has gone up, average revenue must now be lower than in 1967. Lower occupancy does reduce average costs though less than nrnnortionately; overall nerformance would be worse than the 1967 figures suggest. 1/ The question of the "tourist multiplier" has been ably dismissed in J.M. Bryden, Tourism and Development. A Case Study of the Commonwealth Caribbean, C.U.P., 1973. His conclusions are along the same lines as ours. He does not however distinguish between the multiplier in the peak season and the multiplier in the off-season, a distinction which we think is crucial. - 20 - 26. {rl;to.n,reS,nncli roz Antigua) could involve higher water use. In the short-run, surface or borehole water may perhaps be avai.lable at relatively Irw crnt In the longer rtin however; the mar,ginal cost of water will depend upon the cost of desalinate6 water, and this is probably too hiol to-n prnmir its ,-reonnrnic iiup fnr irricatinn 1/ I%ste water could be used for irrigation, but for healtlh reasons that would not be appropriate in the case of crops eaten raw (cabbages, etc.). 8. AC f o infrastrcre the aconts available in the Estimates pose more questions than they solve: Table 4 Actual Capital Expenditure Other Clalendrar Recurrent from Capital Year Actual Revenue Expenditure Budget Estimates Expenditure -- ~-EC$ 1968 153,482 113,821 About 1969 154,978 99,218 50,000 p.a. 1970 1.26,786 /a 219,490 in 1970-73 1,889,569 1971 406,015 176,964 Automatic plus emoluments exchange (200,000?) /a Large unexplained drop in recorded radio telephone receipts. 1/ This is even more likely to be the case in view of the recent enormous rise in energy costs, since water-desalination is energy-intensive. APPENDIX 1 Page 7 Fortunately these records are not the only source available, since in 1972 there was a detailed study of telecommunications under Canadian auspicies, by Kalvaitis and Szaszkiewicz. According to this study, the calendar year 1971 revenue and expenditure position should have been as follows: Table 5 Telephone Services: 1971 Accounts Revenue Expenditure EC$ EC$ Received for local salaries 215.386 phone use: 334,569 wages 63,829 materials 33,102 Due to Antigua Govt. depreciation 85,031 on overseas calls: 166.637 interest on TOTAL 501,206 borrowed capital 151.166 In addition, the Government miscellaneous 29,388 should have naid itself about TOTAL 577.902 EC$100,000 for local calls, and an unknown amount for overseas calls. In fact, the cash flow position looked very different. Depreciation is not a cash cost, and Cahle and Wirelzess were paid neithpr interet nor princinal on the debt due to them. Nor were they paid for the technical assistance they suppl eAd Firnally, the Gover..nment kept all the mToney it receIved for overseas calls, without handing over to Cable and Wireless the portion due to them for use of t1heir internationall tDleornmmgniit2t1inn facil-ties. AC a resul,t, the cash flow position would have been as follows: Table 6 Telephone Services: 1971 Cash Flow Cash Received Cash Probably Paid Receive,d for 'A.oc a l SSal1a r i e s 1500 phone use: 334,569 Wages 63,829 Mat eri als3,0 Received for overseas Miscellaneous 25,000 approx. * ~ ~~~(rIn AI.1 TrOTA T QQ9521 T,rnAT 28Q7,00 approx. /1 Technical assistance not paid Net Cash Inflow: About EC$ 600,000. APPENDIX I Page 8 This technique of obtaining extra cash has not been without its drawbacks. The current amount due from the Government to Cable and qireless may possibly be as much as EC$4.8m, according to one source, although this figure does seem somewhat high: EC$3.6m including the cost of the original telephone system would be a low estimate. On the whole, the current level of charges should be sufficient to pay off the amortization of the telephone system recently installed, over the 20 year period of the loan. However, the system is already in urgent need of extension. There is a waiting list of over 800, and probably some further hidden demand. One proposal is to put 3,000 new lines in, (as against 2,000 existing lines). According to the Canadian study, this would cost EC$3,330,000, or about EC$1,100 per line. It may be possible to complete the work at this price, althouglh quotes received so far have been higher. 9. For a new 50 room hotel, a minimum of 3 lines is required. If there are telephones in each room usage mighit go high enough to justify 6 lines. Capital costs for a business line are the same as for a residential line, as far as outside line costs are concerned: however, part of the central exchange costs are higher, because of the greater intensity of use, (in Antigua, about 200 units per month for a business line, as against 70 units per month for a residential line). Accordingly, costs per hotel line could be EC$1,250. Roughly 60% of this would be foreign exchange cost. If tourism increased because of better room occupancy off-season, it might be possiblt to get by without investment in telecommunications, in the shortrun. However, the telephone system is already congested. With the Antiguan Government's recent record for debt-servicing, it may not be easy to secure new long-term supplier's credits for an expansion of the telephone system. Either this will be provided under aid, or a short-term credit. In the latter case, telephone charges are unlikely to cover amortization costs, unless installation costs are increased substantially above current levels (about EC$20 or less). One possible solution would be to make new subscribers buy compulsory telephone bonds, of perhaps EC$500: interest on these would be deducted from telephone bills. Another way of easing the burden on public finances would be a partnership arrangement, e.g. with Cable and Wireless. A"ll-r-Tr 4 Page 9 10. The cost and revenue data available for electricity supply are as f ollows: Table 7 Capital Actual Expenditure Other Calendar Actual Recurrent from Capital Year Revenue Expenditure Budget Estimates Expenditure . . . ..ECS ......................... 1968 1,579,547 1,158R,0 abhnut 2,inoo p.a. 1Hwker equipment loan yr aae n A Ron nnn 1969 N.A. 1,316,580 1970-73: Fairbanks equipment loan 10a7n ') 1it'. 240a ))1.A Gon /1 .i,- ~ A 190,16 2,246,990 S More most5 00 recently 1071 ) )J.. O1t 1 OA IdI~~~~~y ,, w ,, ,dIJ,,. S I I cludiUg lLarge - s expWndiLure. ThusLA .1he electricity departmer.t prJobabLlyY ra-tbherl lo.re LthncovIerL its rVecur Lr.t expenditure, but imposes a significant net burden on government finances for Ue'ut-serv4l.c.Lng. Rates are probably rot uduly 'Low. TLLere are varLous rate schedules for different categories of user, with declining charges for successiLve traLc.es of po-wer cons-ur.ed; a typLcaL bus'Lness user probCably pays a weighted average of slightly over EC$0.08 per KWH, or US4-1/4j. With good management and bill collection, this should cover all costs incluuing depreciation, and provide a return on capital equivalent to standard commercial 'niteresL rates LULor loans %(al tL.hLoughiI .it U-LgIlL bIUL UC tULoLL;LtL LV eL Ve UdLeb on very short-term supplier credits). In fact, however, electricity supply in Antigua has been beset with problems. Installed capacity in work'ng order is about 10MW: according to the Minister of Utilities this would be increased to 20 H4 if otner equipment already in Antigua were put into working condition, at a cost of perhaps EC$3m. Stanley Consultants now control the management of electricity power supplies. The availabie cost data are given below: Table 8 EC$ EC $ Grabbs 1971 Labor 115,081 Cassada 1971 Labor 57,366 Fuel 742,493 Fuel 53,555 857,574 110,921 divided by 33,910,400 kwH = $.025 per kwH divided by 2,665,070 kwH = $.042 per kwH. APPED IXTV 1 Page 10 Both plants are diesel. The marked difference in unit costs is due almost entirely to the spreading of the labor overhead over fewer units of output at Cassada. 1/ 11. Because of the problems of the past, installed capacity in Antigua could probably be increased at unusually low cost. Better room-occupancy in the low season may just lessen the swing between peak and low demand, and hence necessitate virtually no capital expenditure. New hotels would eventually lead to a need for an increase in installed capacity, (present peak load is about 7.7 NW). More immediately, they would require expenditure on the provision of distribution facilities. For reference one may note that normal capital costs for a power unit suitable for Antigua might be about EC$400 per kw, with as much again to be spent on power distribution. With average bed occupancy of well under 30%, a hotel in Antigua may use 25 kwH per bed-night, or a peak demand of about 3kW per room -- yet annual consumption of electricity could be less than 3,000 kwH per room. Higher occupancy rates could raise annual electricity usage very substantially, although to the extent that the improved occupancy was mainly in the low season, peak demand (and hence need for installed capacity) would not be greatly affected. Electricity usage for new hotels would vary very greatly according to the type of hotel, particularly in regard to requirements for air-conditioning. A hotel room air-conditioner set uses about 1-1/2 kW, and if public rooms are also air-conditioned, this increases power consumption even more. It should be mentioned that another possibility is for hotels to use their own power supplies, which of course avoids the need for capital expenditure by the government. Indeed, some existing hotels operate on just this basis, with their own private power supplv. A danger here is that such hotels will use government electricity for peaking purposes, and thus worsen the load factor. Possibly for this reason, Stanley Consultants (now in charge of electric power in Antigua) wish to legislate private electricity supply systems out of existence. However, in the future, when all available government equipment is in working condition and used to capacity, there could be a case for looking favorably on private supy1V systems. This would be particularly true if sources of government financing were tight, and peaking use of the government electricity supply could be controlled. One might thus economize on reeuirements for public finance at the probable expense of some loss of technical economies of scale. 12. The data available from the Estimates on road expenditure are hardly worth reporting: the Ministry of Works has snent large stums for nurnoses not clearly specified, and it is impossible to allocate these sums between roads and buildings. In particular, it is very diffitilt to allnrate Pxnpnditure for personnel. Detailed figures provided to the Mission specially by the Public. Works Mi-4itry. anparr far tnn lnw, nA noprhlans refer onlv tn materials costs. In fact, recurrent expenditure may have varied from EC$1-1/2 million to EC$3 m,illion a year, probably with sizeable year-to-year fluctuations, Another Public Works figure is total local funds expenditure on roads of I,/ RICcets developments idra. -orld mkets are of course likely to raise fue! costs very considerably. APPENDIX 1 Page 11 EC$4,338,833 over the years 1970-73, or EC$1.! million a year - which could be correct. It is a characteristic of road maintenance expenditure that it is temporarily postponable, so that actual maintenance expenditure in a given year need not correspond to the effect of road usage in that same year. Furthermore, a recurrent expenditure merges easily into capital expenditure. Some of the aid projects - presumably, in theory, capital expenditure - are for "defcrred maintenance!'. The Estimates give about EC$i00,000 a year tor capital expenditure. Aid-funded expenditure, under C.D. &W. was EC$1,383,028 over the years 1970-73, according to Public Works - it would certainly have been EC$200,000 per annum or more. Not too much can be made of this very limited information. Total expenditure on roads has recently probably been of the order of ECS1-1/2 million. Of this sum roughly a third goes on each of the sectors; city roads, village roads, and trunk roads. It is probably trunk road expenditure which is the most essential for the promotion of tourism. Because of the location of hotels and population, most roads serve both tourists and local population. These are a few partial exceptions: for example, the Anchorage Road, built largely for tourism, cost EC$258,000, and there are a few similar cases. 13. A regards road usage, the following data is available on vehicle registration by type of vehicle: APPENDIX 1 Page 12 Table 9 Vehicle Ragistrations Private Car Rentals, Commercial Tractors & Year Total Cars Taxis Buses Vehicles Motorcycles 1955 1235 448 290 107 726 43 1956 1438 488 380 117 801 58 1957 1736 587 430 117 918 75 1958 1957 593 538 122 1137 91 1959 2248 568 792 240 1500 104 1960 2595 655 1130 286 1590 114 1961 2995 842 1232 332 1659 127 1962 3292 907 1432 413 1941 130 1963 3500 975 1530 490 2070 134 1964 3690 1020 1612 536 2112 146 1965 4080 1137 1868 551 2673 136 1966 453L 13L6 2155 648 2679 143 l967 5018 1628 2218 673 3045 136 I196 5400 1858 2895 709 3348 85 1969 6278 2222 2785 770 3849 128 1970l 5A30 2797 2618 786 3880 130 1971 5684 2776 2622 900 5600 133 1972 5817 2815 2850 972 5632 122 APPENDTX 1 Page 13 HLowever, notaL. a:A Lregitere vlicl ae licensed or ir. use. Tn orAer to estimate road usage costs, adjustment needs to be made for: (a) Proportion of registered vehicles which are licensed (b) Annual mileage of different classes of vehicles (c) Road deterioration per mile of usage for each class of vehicle. On tlis basis, renta'l cars or taxLs - bVoth' U of which catdegorULeb a aW -st exclusively used by tourists - would be responsible for 20-30% of all road deterioration in most years. The tourism sector would also ,make some direct use of commercial vehicles, and to a very limited extent of private registra- tion cars. Estimates of shares in road usage costs are giLven below: APPENTDIX 1 Page 14 Table 10 Road Usage Private Car Rentals, Conunercial Tractors & Yaar Cars Taxis Buses Vehicles -ro torcycles v.... .ercert o.' We ged Tl . 1955 27.7 L7.02 6.6 44.9 2.6 1956 26.5 20.6 6.3 43.4 3.1 1957 27.6 20.2 5.5 43.2 3.5 1958 23.9 21.7 4.9 45.8 3.7 1959 17.7 24.7 7.5 46.8 3.2 1960 i7.4 29.9 7.6 42.1 3.0 1961 20.1 29.4 7.9 39.6 3.0 1962 18.8 29.7 8.6 40.2 2.7 1963 18.7 29.4 9.4 39.8 2.6 1964 18.8 29.7 9.9 38.9 2.7 1965 17.9 29.3 8.7 42.0 2.1 1966 19.3 30.9 9.3 38.4 2.1 iy96Y 21.1 28.8 &.7 39.5 '1. d 1968 20.9 32.5 8.0 37.6 1.0 1969 22.8 28.5 7.9 39.4 1.3 1970 27.4 25.6 7.7 38.0 1.3 1971 23.1 21.8 7.5 46.5 1.1 1972 22.7 23.0 7.8 45.4 1.0 APPENDIX 1 Page 15 Of course, some road deterioration would occur due to weather regardless of road usage. 14. As against the costs of road maintenance, road vehicles do provide the Government with revenue. Drivers' licences, and vehicle licences, yielded the following revenue: EC$ 1968 136,200 1969 137,500 1970 n.a. 1971 232,051 In addition, revenue from taxes on petrol may have been perhaps EC$300,000 p.a. Finally, taxes on the importation of new cars could yield EC$200,000-EC$400,000 p.a. according to the number of cars imported in a particular year and their country or origin, (the preferential import duty rate is much below that for other countries). Taxes on road use are presumably a factor in taxi and car rental charges: and tourists also pay a very high fee for a driver's licence - EC$10.00. In theory this is for a year, but in practice most tourists use the licence only for a few days. Any expansion of tourism would lead to extra road use by tourists, and hence extra costs. However, more off-season tourism, and at existing hotels, should require little road-widening or build- ing of new roads: the same could not be said of extra hotels being built. One should also note that there is some acknowledged over-staffing at PWD, for political reasons. Thus social costs of extra road maintenance may be lower than current money costs of existing road maintenance would indicate. Of course, tourism also has income effects, which increase road use. 15. It is possible that to some extent Antigua can choose between enipha- sizing luxury tourism and emnhasizing middle-income tourism. If so; it is interesting to note the different implications for infrastructure investment. For any given ntmber of rooms, a modest hotel, with nO telephones in the rooms. may require only half the number of telephone lines of a luxury hotel. Rooms writ-h .Lwers alone will ulse far l ess .,water than roomQ uwith Lbaths -And hotels whose rooms do not have airconditioning will require far less electric power capacity than hotels which it, hnven nir-lornditioninc, often only a half or a third as much. Thus modest hotels not only require less capital per room directly, but also substa-1 tially less indirectly, i.n i invetmoent. 16. Ve ry i t tle1~ infe% m.*atio4r.w ^antyavailable on then Ar dman A m rlade y touirisQm on port and airport facilities, the costs to the economy of supplying these facin4l tis-4 an. the -ba ,.-.vae nar,ar.tA for tha by these fa cin,-a1o4 n What is fairly clear is that the port, like water, electricity and telephones, operat-es as a loss, utl that h.the airport is profitable. The relevat fl-ures for 1971 are given below: APPENDIX 1 Page 16 Table 11 Profit arAd Ta cS Poi4tion in VUarious PuiJc t1i1ities ($'000) Electricity, Water & Telephones Port Airport Operating ExpenAditure 35Q505 357 PFubliLc Debt Servicing 3,984 837 290 Other 786 Tota'L recurrent Expenditure 8,320 1,342 647 Total Recurrent Revenue 4,070 1,032 2,154 Surplus (deficit) (4,250) ( 310) 1,507 Source: British Development Division in the Caribbean. The sound financial position of the airport is not, however, necessarily to be attributed to tourism to Antigua itself. The main source of airport revenue is landing charges levied on aircraft but the traffic of transit passengers through Collidge is much greater than that of stay-over tourists. What is needed, therefore, is an accurate estimate of the benefits and costs associated with the airport which are to be attributed just to tourists to Antigua. This it was not possible to obtain. Our guess is that the tourist- generated airport activity is not a net cost to the economy. A long-term plan for tourism in Antigua would of course have to look fairly carefuly at the costs of adding to airport capacity at the proper time.