POLICY RESEARCH WORKING PAPER 1280 The Economic Imp act Expon controls can transfer significant profits from raw of Export Controls materials producers to the procesSIn industries, causing significant net losses to an An Application economy and a to iNIongolian Cashmere net decrease in export and Romanian Wood Products earnings. Wendy E. Takacs The World Bank Policy Reseach Deparmnent Trade Policy Division ' March 1994 | POLICY RESEARCH WORKING PAPER 1280 Summary findings Countries sometimes use export controls on raw She finds that (under reasonabie assumptions about materials to encourage domestic processing. The elasticities of supply) export controls can transfer motivation is usually to assure raw materia!s at low significant profits from the raw materials producers to prices for domestic industries, although exports arc the processing industries, causing significant net losses to sometimes controlled in an attempt to increase export the economy and a substantial net decrease in export earnings (by promoting exports of higher value-added earnings. processed goods rather than raw materials). Quantitative export controls will be even more The problem is, export controls hurt raw material distortive if processing industries have any monopsony producers and cause economic distortions that result in (single-buyer) power. This is quite likely in developing net losses to the country. The impact of raw material countries with small industrial bases - or in economies export controls on total export earnings is ambiguous: in transition, where central planning has left a legacy of the decline in raw material exports when production is very large firms in highly concentrated industries. discouraged by lower prices may outweigh the effect of With monopsony power in the processing industry, increased exports of processed goods. both output and exports of final products can be reduce Takacs develops a simple partial equilibrium model of by quantitative export controls on raw material inputs. export controls on raw materials to investigate the The quantitative control bestows effective monopsony impact of export restrictions and to estimate the power on the processing firm and encourages it to potential magnitude of the transfers bet-:en groups and exploit this monopsony power by reducing output. If the the net costs of the export-control regimes. raw materiAls could be freely exported, processors would Her estimates of the magnitude of transfers and costs not be able to effectively exercise monopsony power. of export controls on raw cashmere (in Mongolia) and wood products (in Romania) indicate that the transfers and costs may be substantial. Thispaper-aproduct of theTrade Policy Division, Policy Research Department-is asynthesis of background material prepared for the joint UNDP/World Bank Trade Expansion Program which provides technical and policy advice to countries that want to reform their trade regimes. Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 2;,433. Please contact Minerva Patefia, room N10-013, extension 37947 (27 pages). March 1994. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of Ideas about development issues. An objective of the series is to get the findings out quickly, even if the iresentations are less than fully polished. The papers ca"y the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries. Produced by the Policy Research Dissemination Center The Economic Impact of Export Controls: An Application to Mongolian Cashmere and Romanian Wood Products by Wendy E. Takacsi Department of Economics University of Maryland Baltimore County Baltimore, Md. 21228 USA and Trade Policy Division Policy Research Department The World Bank 1818 H Street, N. W. Washington, D. C. 20433 USA I. Introduction Export controls on raw materials have been used by countries to encourage domestic processing activites.2 Such controls can take the form of complete prohibitions on exports, export quotas, export licensing requirements, or export taxes. They are usually motivated by a desire to assure raw materials at low prices for domestic processing industries, but other stated purposes include preventing exporters from selling without a letter of credit from the importer or selling at prices considered too low. Export controls on raw material are also jusfied as a method of increasing export eaings by promoting exports of higher value-added goods rather than raw materials. Export controls on raw material or inrmediate inputs increase the effective rate of protection to the processing industries by lowering input costs. But they create costly distortions and efficiency losses and they benefit owners of processing facilities at the expense of raw materials producers. The purpose of this paper is to develop a simple partial equiiibrium model of export controls on raw materials to investigate the impact of the restrictions and generate rough estimates of the potential magnitude of the transfers between groups and net costs of the regimes. Section II develops a model that links the markets for material inputs and final goods under the assumption of competitive markets for inputs and outputs, indicates the nature of the transfers and net costs generated by export restrictions on inputs, and indicates a method for calculatng estimates of the size of the transfers and costs. Section m applies the model to two case studies: export licensing requirements for raw cashmere in Mongolia anl export prohibitions and quotas for wood products in Rom-ii. Section IV then extends the analysis to a monopsonistic market structure in the final goods industry. A monopsonistic processor of raw materials may be a relevant market structure for developing countries that lack a broad industrial base and for transitional economies in which central planning led to high industry conentation, and thus a single processor or small mnmber of buyers of certain raw material inputs. 2 II. Raw material export controls with perfectly competdve markets Suppose that a raw mal hiput, I (such as raw cashmere) is used as in input into the production of a final product, F (such as cashmere sweaters). Suppose also that the country under consideration is small, so that it cannot influence the price of either the raw material or the final product in world markets.3 To ease graphical presentation and link tie markets for Inputs and outputs, define units of input so that one unit of Input is required to produce one unit of output.4 The perfectly competlt!ve processing industry is willing to process more inputs into output the higher the value-added per unit In processing: QS . S (PPr (1) where Qs is the quantity of final product supplied, PF is the price of the final good, PI is the price of the raw material input and PFPI is the value-added per unit. Suppose that demand for the final good within the country depends upon the price: s d = D(PP) (2) Net exports of the finished product wiUl be the difference between the quantity supplied and the quantity demanded: XF QP O QF (3) Suppose that the raw material is produced by a perfectly competitive input industry and that the quantiy of the raw material input supplied depends on its price: Qf = S1 (Pr) (4) Suppose that the raw material is not demanded in the domestic market except as an input to 3 the final goods processing industry. This assumption, along with the definition of the 'units' of input, implies that: ox' - Or' ~~~~~~(5) Exports of the raw material will be the difference between the quantity supplied and the quantity demanded as inputs by the domestic industrr: x- ' QU - Qz - St (P1) - ( FPP ) ( 6 ) The equilibrium in this model under free trade is shown in Figure 1. If there were no trade restrictions, the domestic prices of both the final good and the raw material would be equal to .ieir prevailing world market price. The supply curve of processing as a function of the value-added is shown in the upper panel of Figure 1 by the curve S(V). The height of the curve Is the value-added per unit of final good produced, V. Tho processing industry's supply curve of the final product under free trade, Sp*, would lie above S(V) by the raw mateial Input costs per unit, Pf. The demand curve for the product is Dp. The upper panel of figure 1 shows the market equfilibrum for the final product under free trade. If the world market price of the final good were Pf, the world market price of the input were Pf, and there were no trade rerictions, the quantity demanded would be Dp*, the quantity supplied by the processing Industry would be QP*, and Qp*-Dp* (= Xp*) would be exported. Ihe market for the raw material is shown ia the lower panel of figure 1. The units along the horizonta axis measure the quantity of "packcages" of inputs, where each 'package' represents the inputs needed for one unit of output The price measured along the verdcal axis is likewise the price per 'package' of inputs. The domestic raw materis industry supply curve is shown by SI. If the price of the raw materia in the world market were PT, the domestic raw materials industry would 4 produce an output QI*. At this price of the raw material, the domestic fnal goods processing industry supply curve would be Sp*, the output of final goods would be Qp. and the demand for raw matedals on the part of the domestic processing Industry would be D*. ITe quantity exported woulti be XI* (=Qi*-DI*). Export earings from finl goods exports would be PTXp* and export earnings for raw materis exports would be PTXi*. Suppose that the govrnment now controls exports of the raw matera by imposing an export quota that allows no more than a maximum amoulat to be exported6. Suppose this system is enforced through an export licensing system. In terms of the model, the domestic price of the raw material will no longer necessarily equal the world market price because producers are not free to export. A biding quota restriction implies that the quantity of exports is given, and the domestic price of the raw material input would be determined by equation (6) as the price at which the difference between the quanity of raw matrial supplied and the quaty demanded would equal the maximum amount allowed to be exported. Given that the quantity demanded depends upon the output of the processing Idustry, and given that the position of !he processing industry supply curve depends upon the price of the raw materid, the new equilibrium would be determined jointly in both markets. As the raw material input price falls due to the export restrction, input costs of the processing industry fil and the processing industry supply curve shifts downward, or to the right. At a constan final product price, output of the final product (and therefore the quantity of the raw material demanded by the processing industry) increases. Figure 2 illustrates the new equilibrium after an export restriction liming raw marl exports to Xi is Imposed. The price of the Ilput will sede at P?, where the quantity demnded (equal to the quantity of final good supplied at the world market price P'y and Input price P? plus XI) equals the quantity of input supplied along S1. The decline in the equilibrium raw material price 5 will reduce raw material production from Q% to QIn. n the final goods mnarkt, output Inereases from Qp to Qp'. The raw matewal eport reticdon hurts producers of the raw material ad discourages raw material production vad exports, but increases profits of the procesing Industry, as well as production and exports of the fiMlW good. lbe gains to the procesing istry and the losses to the raw material produces, as well u the net efficiency losses from the restriction, can be Identfied and esimated. In the market for raw matria, producer profits fail becaus of fahing prices. 3 loss in profits Is usually identified as a los in wduceu surplus equa to area adge in the lower panel of figure 2. Tho height of the raw material supply curve shows the marginal cost of producing each extra unit. When price faUs to Pq, reveues decrease by area adrsge but costs decrease by only gdre, so net Income declines by area adge. Part of this loss (area abfe=area hijkl consists of a ransfer from raw materials producers to the processing industry in the form of higher profits due to lower Input costs. Another part (area bcg) represe a transfer from raw materials producers to the recipients of export licenses, who are able to buy the raw materal on the domestic market at P? and sell abroad at Pl. The remainder (area cdg) is an efficiency loss because the extra units of raw matDrl Ql*-', which would have been eported to earn Pf in the absence of the export restriction, would only have cost an amount equal to the height of the supply curve to produce. With the export restrction, these extra units of the raw material will not be ptoduced and sold abroad for more than the cost of producing them. Area cdg thus represents a net effidiency loss to the country. These transfers and efficiency losses can in principle be measured. The loss to producers of the raw materil equas area adge: 6 Area adge - dP1 Qr + 2 dQ dPz ' dP QOr +x 2 Ox dP1 dPz dPr Qr ( I + 21 Of AI) (7) * dP2. Q (12 t ,e dP V-.!Lx(1+ 1 SP) Pz 2+ 2 PzV Oxpl.+1 X2 Xp Z 2 x r whare cl (dQ/dP1)(PI/Q1) h the elascity of upply of the raw material, VI- PIQI the value of raw materi producdon, dP1 (.P1P-P;?). dQ[ = (QQ)ad w = dPI/PI. rhe efficieny loss area cdg will be equal to the last team of equaton (7). The part of the loss to raw material producers that is transferred to export license recirients Is area bcgf: Area bcgf = dP1 X1 (6) -if r where Vf is the value of raw material exports. In the market for the processed good, the processing industry gains area hild. The industry marginal cost curve shifts downward from Sp* to SF' which reduces costs for each unit previously produced and therefore increases profits by area hnkl. Profts also increase by the difference between revenue and cosu for the exta units produced, area nik. From the point of view of the country, however, the increase in domestic processing of the raw material creates an efficiency loss equal to area min because the artificially low raw mateials price encourages the processing industry to expand output beyond the point at which the price in the world market equals the true cost of producing the final good, including the opportunity costs of exporting the raw material. Since SF* Is parallel to Sp', but lies below It by (P -P;), area min = area ljL This part of the transfer from the raw 7 material producers is not a gain to the processing Industry, but i simply lost due t3 higher proeeing costs. Again, these transfers and costs can be measured. Tho gain to the processing industry is area hikl: Area hiki -Q, dP ^ 2 dO, dP* uPzQ.A d Pz d s PzF dPP (9) QFPz~ 2 et v PP* PZ PZ -lvs-~ 1 ; 4X/ 2.4 where v is the value-added per Uuit of final product, ec = (dQSIdv)(v/Q#) Is the elasticity of supply of processing with respect to the value-added per unit, dQp=(Q#-QF), =PI/Pp is the ratio of raw material cost to the prce of the fial product, and VF = PpQF is the value of fnal good produced. Note that with the fimn' product price unchanged, dvw-dP1. Ihe efficiency loss due to higher-cost domesdc prcessing (area ijk) equas the second term in equation 9. I Information on prices and quantities of inputs and outputs are available, the second step in equation 9 can be used for calcuons, but the last step is more suitable when the only information available is from an input- output table for the economy. The welfare impact aid net efficiency losses are the appropriate criteria for judging the beneficial or detrimental effects of the export restrictions. .rt so much emphasis has been placed on using raw materil export controls to encourage higher-value-added exports and therefore export earnings that we investigate in detail the impact of the raw material export controls on total export revenue. Foreign exchange earnings from final goods exports inerease by P'F(Qp'-QF*). Raw material exports fal from XI* to Xe, and export ening faU by P'Y(X1*-X1). Export earings from exports of final goods increase, but earnings from raw material exports decrease because the quantity 8 exported decasea by the amolnt diverted to the domestic industry, D1'-D1 (-QF'.pF), plus the decrese In production Q*-Q1' due to the lower raw material price. The not impact of the raw material export restriction on told export eanWs is ambiguous The change In total export earnings due to the export control on raw mateials would be the difference between the increase in the value of final good exports ad the decrease In the value of raw matrial exports. Given that XI-Qf -Qy and dQ? - dQl, the change the value of raw material exports would be: d(P'fXzl - Pz(dQj' - dQ°z] -pf es r dpz _ ' - vt Pz dvl px ~ 'r (10) AlI, Q dPz 1A VdP1. m P[erQz +4V, * (1'i)erV1z *4V,+ epg V, ! (1+ ) a Tho first term captures the decline in raw materisl exports due to the Sfll In producdon; the second captures the decline in exports due to the diversion to domestic processing. Holding domestic demand for the final good constant, and allowing for free exportaton of the processed good at an unchanged world market price, the export restriction on raw mateials would increase exports of the fnal good.7 The maitude of the increase in procesed good export earnings would be: P" dX, = Pr e ?,F dPr pw PzUe dPt 11 V P V, 5, b I 9 lherefore total export ears would change by: dvx Vp4 1*x-[ (1.)edVw * 4 .t4) (14 e2):] + v, at x U.- (1+X)*J -x (1+w) efv Total export earinpgs wil increase if equadon (12) is positive but decrease if It s negaive. The raw material export control is more likely to increas export earningp the greater the value-added in processing (the smaller is 0), the greater the elasticity of supply of the processing industry, and the smaller the elasticity of supply of the raw material, and the larger the processing industry relative to raw material production. Ii. Applcation to Mongolan Cashmere and Romanlan Wood Export Controh This section applies the model in an attempt to quantify the impacts of two examles of export controls on raw materials: the export licesing requirements for raw cashmere in Mongolia and the export prohibitions and quotas on timber and other basic lumber products In Romania. Mongolia has imposed export licensing requirements on a number of raw materials. As of August, 1992, these included live animals, wool, raw cashmere, camel wool, sheepskin, goatskin, and timber.8 Sufficient data were available to apply the model to raw cashmere. The values of the variables that were used in equations 7 through 12 to calculate the magnitude of the transfers and net costs associated with the export licensing arrangements, along with an explanation of the sources, appear in Table 1. The gains, losses, and net costs were estimated assuming varying degrees of responsiveness of both raw material and final good production to changes in prices. The calculations presented use 10 combinations of elasticity ranges of from 0.5 to 1 for raw cashmere and 1 to 3 for garment production.9 The results appear in Table 2. Of particular note is the result that the export controls on raw cashmere appear to reduce total export earnings.10 Also noteworthy is the magnitude of the transfers. Because a large amount of raw cashmere is exported, the impact of even the modest 16.6 percent price differental between the world market and domestic price imposes large losses on herdsmen and allows large gains to those able to export. The estimated losses of $5.5 to $7 million dollars would represent $5 to $7 per person to the approximately 1 million people in the countryside. The rents generated from exporting are about $4 million. As of March, 1993, Romania imposed export prohibitions or quotas on a relatively large number of raw material and intermediate inputs. These included wood products that are inputs into the furniture industry, stardng with logs, timber and firewood, up through lumber, plywood, particle board, and veneer.11 Data on quantities and values of production and exports of aU the various products were not available, but the 1990 Romanian input-output table separately identifies the woodworking and furniture industries. Estimates of the impact of the export controls were based on the input-output table, under the assumption that the woodworking sector represented production of inputs used in the furniture sector. Equations 7 through 12 were used to calculate estimates of the magnitude of the transfers and net costs associated with these export restrictions. The values for the variables in those equations, and an explanation of the sources and reasoning behind them, appear in Table 3. The resulting estimated impacts of the wood products export controls appear in Table 4. Romania currentiy imposes restrictions on cutting of timber to protect the environment and allow rebuilding of the timber stock. These restrictions can be interpreted as implying a zero elasticity of supply of timber (and therefore wood products) in the short run. However, in the long it run additional planting and harvesting of timber would be possible, so estimates are presented using both a short-run scenario, in which the elasticity of supply of raw material inputs is assumed equal to zero, and the elasticity of supply of value-added in furniture production Is assumed equal to one, and alternative long-run scenarios, in which the elasticity of supply of raw material input Is assumed to be positive. To illustrate the sensitivity of the results to assumptions about the elasticities of supply of wood product materlI inputs and furniture, the results of two alternative 'long-run scenarios are presented, one assuming that both elasticities of supply equal one, and a second assuming a low elastdcity of supply of wood products of 0.5, but a higher elasticity of supply of furniture production of S. The results indicate that the export controls on wood may impose severe losses on the wood input Iustries. ho esatimated decreases in proft range from about 8 to 10 bilion lei per year. If there is a nonzero elasticity of supply of wood, there are also efficiency losses In the wood market, esimated here between 1 and 2 billion lei per year. The gain In tem of Increased profitability of the furnitue industry is about 1.5 to 2.3 billion lei, while the etmated efficieny losses from increased furniture production amont to 200 million to approximaely I billion lei Ibe result of the wood inputs export controls are ambiguous and depend crucially on the elasicities of supply In the two Industies. In the short-run scenario In which wood inpu production cannot respond to price changes, the export controls increase total export earnings by approximately 2 billion lei. However, in the first long-run scenario, auming equal elasticities of supply of wood products and furniture, the discouraging effect of the export controls on wood production and the diverion of wood eWport to furniture production decreases wood input exports by almost 14 billion lel, whie furitre eworts increase by a much smaller amnt. The net result is a decrease in total export eamings from wood products and furmIture of over 10 billion lei. 12 Raw material export controls with monopsonlstlc procsor The analysis so far has assumed that the markets for both inputs and outputs are competitive, that is, contain a large mnmber of both buyers and sellers. In many cases of raw material export controls, the assumption of a large mumber of producers of the raw material may not diverge too much from reality, but in many cases there may be one or only a few processors, which creates a vmonopsonistic (single buyer) market structure in the raw material marcet. Whether a monopsony Is able to exercise monopsony power depends crucially on the intenatonal trade policy in effect. If the raw materials processed by these firms can be freely exported, then these processors may not be able to effectively exercise monopsony power. They will have to compete with potential exporters of the raw mterals and will be forced to pay the world market prices for their inputs. On the other hand, if the inputs cannot be exported or if exports are limited to predetermined quantities, then the processors will be able to control the price of the raw materials through their purchasing decisions. A relevant quesdon may very well be whether a monopsonistic market structure changes any of the basic conclusions above with respect to the impact of export controls of the raw material. The dmination of prices, production, and exports of raw materials and processed products in the case of a monopsonistic market structure in the absenc of any export restriction is illustrated in Figure 3. As in the previous figures, the upper panel represents the market and production conditions for the final product and the lower panel the market and supply conditons for the raw material WpuL In the upper panel D represents the demand curve for the final product and MC(V) represents the marginal cost for factors of production oaU the raw matial Input. Ihe extra cost of an extra 'unt of raw matal input must be added vertically to MC(V) to obtain the marginal cost of production curve MC. If the world market price of the input is Pt', and exporters are free to export at this price, the processor will also have to pay Pt', and the marginal cost curve for the processor will be MC as shown in the upper panel of figure 3. If the world market price of 13 the flnal product Is P'F, then the processor would maximize profits by producig at an output level of * 12 The quanty of the final good exported would be XF* the difference between Qp and DF*. The value of export earnings from final product exports would be the area abcd In flure 3. In the input market, the processor would demand D1* unit of input, the producers would supply Ql at the price Pf, so the quantity exported would be Xi* (=Q1*-D*). Earnings from exports of the raw material would be P'XI *, equal to area efgh. Compare tdis outcome with the resulting prices, production and export earnings If exports of the raw material are limited by some form of quantative export restriction, such as a restrictive licensing system, an export quota, or an export ban or embargo. The analysis is developed in the form of an export ban because it Is the simplest to illustrate and Mongolia has banned exports of certain raw materials at tmes in the past. The results woud be similar with other forms of quantitative export restrictions.13 A ban on exports of the raw material input implies that the single processor faces the upward sloping supply curve of the raw material input SI. Ihe greater the output of the processor, the more raw material purchased, and the higher the resulting market price of the raw mateial. Because the processor would drive up the market price for aUl unit of the inputs used, not just the last unit bought, the extra cost of an extra unit of input to the processor will exceed the market price. A curve showing the extra cost of extra units of input, labeled MCI in the lower panel of Figure 4, can be derived from the input supply curve SI.14 After the restriction on the exportation of thc input, the marginal cost of the final good would be MC'. It is derived by adding MC, to MC(V). The monopsonistic processor can still sell in the world market at P'Y, so the processor's profits would be maximized at Qp', the output level at which the world market price equals the marginal cost of production. In the case illustrated in Figure 4, this would be at a lo output level than without the export ban. The processor reduces output because lowering output reduces the demand for the input 14 and reduces his cost for the raw material Inputs. As shown in figure 4, when only one profit maximizing processor buys a raw material on the domestic markt, an export ban on the raw matial eliminates exports of the raw material, and can decrease production and exports of the final good. In this case, IgtaI export earings from both the fnal product and the raw materiS must decrease. Figure 4 illustrates the case In which total export earnings fail, but this result is not unambiguous. If exports of the raw material are large relative to production of the final good, output and exports of the fil good may increae. i As In the competitive industry case analyzed first in this appendix, the export ban will reduce real income or profits of the raw material producers and increase the profits of the processing industry. Revenues of the raw material producers fall from efgh to ijkh. Revenue from raw material exports wiU fall by lfgk. Of this reduction in export revenue, 14 represents a net loss to the country because it is the difference between export revenues lost and the incremental cost of producing the quantity that would have been exported in the absence of the export ban. Area elji represents a transfer from raw materials producers to the processing firm. CONCLUSIONS Export controls on raw materials reduce raw material prices to domestic processing industties and encourage domestic processing, but hurt raw material producers and cause economic distortions that result in net losses to the country. Exports of raw materials are sometimes controlled in an attempt to increase export earnings by promoting exports of higher value-added processed goods rather than raw materials. However, the impact of raw materil export controls on total export earnings is ambiguous; the decline in raw materia exports when production is discouraged by lower prices may outweigh the effect of increased processed goods exports. Estimates of the magnitude of the transfers and costs of export controls on raw cashmere in 15 Mongolia and wood products in Romania indicate that the transfers and costs involved may be substantial. These estimates are designed to be illustrative rather than definitive, because no esdmates of actual supply elasticities in the raw material and processing industries in these countries are available, so the estimates are based on a range of assumed values for these elasticities. Despite these caveats the calculations do indicate that in both these eases the export controls, under reasonable assumptions of elasticities of supply, may transfer significant amounts of profits from the raw material producers to the processing industries, cause significant net losses, and result in a substantial decrease in export earnings. The impact of quantitative controls on exports will be even more distortive if processing industries have some degree of monopsony power, which is quite likely in developing countries with small industrial bases, or economies in transition where central planning has left a legacy of very large firms in highly concentrated industries. WitlA monopsony power in the processing industry, both output and exports of final products may be reduced by quaniative export controls on raw material inputs, because the quantitative control bestows effective monopsony power on the processing firm and encourages it to exploit this monopsony power by reducing output. 16 ENDNOTES 1. This paper is based on research for the UNDP/World Bank Trade Expansion Program reports for Mongolia and Romania. The author thanks Jaime de Melo and Arvind Panagariya for useful comments on a previous draft. The views are those of the author, not necessarily those of the United Nations or of the World Bank. 2. In England Elizabeth I banned the exportation of vwool fleeces in the 16th century to encourage the textile industry. Other modern examples include Uruguayan prohibitions on exports of raw hides. Many countries currently undertaking the transition from central planning to a market economy also control raw material exports. 3. For Romanian wood products this assumption is reasonable. In the case of cashmere, however, Mongolia may have the ability to influence world market prices. Published estimates of Mongolia's share of world raw cashmere production range between 10 percent (Browne, 1990) and 2025 percent (Economist Intelligence Unit, 1991). Mongolia exports both raw cashmere and knitted cashmere products, so higher world market prices for raw cashmere increase export earnings directly and also indirectly by increasing the costs of competing firms in the market for knitted cashmere products. For an analysis of the potential welfare gains from exploiting market power in world raw material markets when a country also exports the final product, see Jones and Spencer (1989). Including market power in world markets in the model would be a useful extension of this paper. 4. This approach is based on Corden (1973, pp. 30-35). Suppose that the processing industry uses et inputs to produce one unit of output. Then each "package" of cm inputs woid be a "unit" of inputs. For example, if 10 ounces of raw cashmere were used to produce a sweater, then a 10- ounce ball of raw cashmere would be one "unit" of input. If the price of an ounce of cashmere were $1.00, then the price of a "unit" of input would $10.00. 5. Income is omitted as a determinant of demand because of the partial equilibrium assumption. 6. The limit case of a complete ban on exports of the raw material would be analyzed as a zero quota. 7. The increase in exports may be underestimated in the partial equilibrium framework used here if income feedback effects are important. Lower real incomes due to the efficiency losses could reduce domestic demand for cashmere products, which are presumably luxury goods. On the other hand, lower income may reduce imports and therefore exports through balance of payments or exchange rate effects. 8. This list is based on information from the License Bureau, Ministry of Trade and Industry. Exports of other products also were subject to license: scrap copper, scrap iron, scrap aluminum, scrap aluminum alloy, scrap steel, meat, wheat, children's clothes and shoes, and imported 17 9. Elasticities of supply of processing with respect to value-added of up to 10 were tried, but elasticities of over 3.4 resulted in such large estimated impacts on processing that output would have become negative In the absence of the export controls. The large impact on processing stems in part from the high ratio of the price of raw cashmere input relative to the price of finished garments. Raw cashmere cost Is over half the price of the finished garment, so a drop in the price of raw cashmere at constant final good prices represents a larger percentage increase in the value-added per unit. 10. With a smaU elasticity of supply of raw material and a large elasticity of supply of processing (e!=0.5;e;= 10), export earnings increased by a relatively modest $80,608, achieved at an estimated cost of $718,186 in deadweight efficiency losses. This case is not reported because the estimates at these high elasticity levels indicated that output would decrease so much that it would become negative, which would result in nonsensical calculations of transfers. 11. Order No. 120 of 31 July 1992 prohibited the export of logs, rafters, lumber, railway sleepers, Christmas fir trees, firewood, wood for cellulose, fiberboard, timber, wooden paUets and veneers, and imposed the following quotas: Produt Ouant Beech-tree plywood 50,000 m3 Panels 1,300,000 m:2 Beech-tree parquet 500,000 m2 Chipboard 800,000 m2 Timber and semifabs 300,000 mn2 of resinous woods, beech and softwoods Door and windowframes 1,000,000 m2 12. This analysis presumes that the behavior of the processing firm is to attempt to maximize profits. In a transitional economy such as Mongolia, it is not clear what the objective of firms that are still state-owned actually is. As these firms become privatized and thus presumably responsive to the shareholders' desire for high dividends or growth in share value, firms wiUl presumably shift their objectives to profit maximization. 13. It is important to note that discouraging exports through an export ta will not give the monopsonistic processor control over the domestic market price. The export tax wil lower the input price, but exporters wiUl stil be free to export as long as they pay the tax, so price wiUl fahl no lower than the world market price mims the tax. The export tax system will also yield revenue for the government, rather than profits in the form of quota rents to exporters who are able to obtain licenses. 14. The marginal cost curve of the input wiUl lie half-way between the supply curve SI and thb vertical price axis. See Ferguson and Maurice (1979) or virally any other intemediate level microeconomic theory textbook for an explanation of the analysis of monopsony. 18 15. MC' -uust cross MC at the output level at which MCI crossed a horizontal line of height P1W. This poit (shown as point a) lies to the I& of Q* in the case shown in figure A.4.4). Because MC' must be steeper than MC, the quantity produced by the processor (and therefore the quantity exported with unchanged demand condition) must be lower. If the supply cunre had been much further to the right (passing through b, for example) then MC, would have passed through point c, MC' would have crossed MC at point d with a steeper slope, and output and exports of the final good could increase. In this case the value of the extra exports of final good would have to be balanced with the elimination of exports of the raw material to determine whether total export eamings decrease or increase. Large exports of the raw material relative to domestic production of the tinal good Increases the probability that there will be an increase in output because point a will be at a larger output level. 19 REFERENCES Browne, R.J. (1990). Cashmere Goat Notes Guilford, N.S.W.: Australian Cashmere Growers Association. Corden, W. Max (1973). The nheorv of Protection New York: Oxford University Press. Economist Intelligence Unit (1991). Luxiu Eibers EIU Special Report No. 2633 London: Economist Intelligence Unit. Ferguson, C.E. and S.C. Maurice (1978). Economic Analysis: Theory and plicatoig Homewood, In.: Richard D. Irwin. Jaakko-Poyry (1992). RomaWa-Private Sector Development and Enterurise Reform Project: Study on the Wood-based Industries of Romania. Stockholm (November 6) Jones, Ronald, and Barbara Spencer (1989). 'Raw Materials, Processing Activities, and Protectionism" Canadian Journal of Economics 22:469-486. 20 Table 1 Variables Used in Calculations of the Impact of Mongolian Cashmere Export Licensing Requirements P w $25.89/"package" Data provided by the large Gobi factory indicate that the factory took in 900 tons of raw cashmere, approximately 450 tons of which was exported as greasy or dehaired cashmere, approximately 80 tons processed for export as tops, and the remaining 370 tons used to produce garments. This would imply that 370,000 kilos of unprocessed cashmere was processed into 250,300 garments. The "package" of unprocessed cashmere input per piece produced, on average, would be 1.48 kilos. The unit value of unprocessed cashmere exports in 1992 was $17.49/kilo: 1657.53 metric tons valued at US$28,986,430 (Ministry of Trade and Industry). Thus the world market price of a "package" of input would be US$25.89. PF $35.05/piece Unit value of exports of cashmere garments in 1992: 142,190 pullovers were exported valued at US$7,664,460; US$341,970 of "other garments" were exported, but no quantity figures were available. Data provided by the large Gobi factory indicated that they had produced approximately 95,800 other garments in 1992, and approximately 90 percent of their output is exported. This would imply exports of 86,220 other garments. Thus an estimated 228,410 garments were exported valued at US$8,006,430 .XI 1,119,953 Exports of unprocessed cashmere in 1992 were 1,657,530 kilos, equivalent to 1,119,953 "packages" of inputs. Qp 275,300 Large Gobi factory output of approximately 250,300 pieces, plus estimated 25,000 of small Gobi factory. Q, 1,395,253 Raw material outpu. measured in "packages" of inputs equals finished product production plus exports of raw material dP, $3.69 Data provided by the Gobi factory indicate that the price of unprocessed cashmere in 1992 was 600 tugriks/kilo (at the official exchange rate of 40 tugrik/US$, $15.00/kilo) for unprocessed cashmere. This would imply a domestic price of $32.40 per "package" of input, $3.69 below the export price. ir .166 Calculated as dP,/(Pw,-dPj) 10 0.633 Calculated as (Pw,-dPO)/PwF E,M 0.5 to 3 5 Elasticities of supply of unprocessed cashmere and cashmere garments, respectively. 21 Table 2 Estimated Impact of Mongolian Cashmere Export Licensing Requirements (US dollars) Elasticity of supply of cashmere 0.5 1 0.5 Elasticity of supply of garments 1 1 3 Loss to cashmere producers 5,362,424 5,576,364 5,362,424 Efficiency loss in cashmere production 213,940 427,881 213,940 Gain to garments industry 870,000 870,000 578,287 Efficiency loss in garments production 145,857 145,857 437,570 Decline in raw material export 5,048,857 8,050,979 9,142,326 Increase in garments export eanings 2,770878 2,770,878 8,312,635 Change in total export earnings -2,277,979 -5,280,101 -829,691 Transfer to export license recipients 4,132,627 4,132,627 4,132,627 Total efficiency loss 236,922 420,371 634,266 Increase in gannents production (units) 79,055 79,055 237,165 22 Table 3 Data for Estimation of Impact of Romanian Wood Export Controls VI 16,959 Value of output of woodworking industry. (millions of lei) Data are from the 1990 Romanian input-output table for the woodworking sector (sector 32) VI 20,868 Value of furniture output (sector 77 of 1990 input-output table) (millions of lei) tvx 230 Value of exports under quota of plywood, laminated board and chipboard, all potential inputs for the furniture industry (millions of lei) 0.24 PI/PF, calculated from ratio of value of inputs of woodworking industry into the furniture industry to value of output of furniture industry from 1990 input-output table. r 0.5 (Pw,-P,)/P,, the percentage difference between the world and domestic price of raw material inputs due to the export controls on raw materials. A study of the Romanian wood-based industries by the Swedish consulting firm Jaakko Poyry (1992) estimated that the prices of wood material input costs for the furniture industry were far below the prices in other countries. Input costs were less than half the costs in the next-least expensive country, Poland. Costs of half the world market price would imply a cost differential as a percentage of the domestic price of 100%. Romanian critics of the Jaakko Poyry study argue that it overestimates the wood price differential, and that 50% would be a better estimate. The 50% figure is used here, but the estimates of the costs and transfers would be even larger than those calculated if the actual price differential is larger. e,F 0 to 5 Assumed elasticities of supply 23 Table 4 Estimated Impact of Romanian Wood Export Controls [hort run Lone run Millions of lei 63=0 es-1 efrO.5 eg- ep eF5 Loss to wood indwutries 8,480 10,599 9,539 (area adge) Efficiency loss in wood industries 0 2,120 1,060 industry (area cdg) Transfer to export license 115 115 115 recipients (area bcgf) Gain to furniture industry 2,306 2,306 1,516 (area hikl) Efficiency loss in furniture industry 198 198 988 (area ijk) Decline in wood exports 1,186 13,905 12,291 Increase in furniture exports 3,294 3,294 16,475 Change in total exports 2,108 -10,610 4,184 24 FIGURE 1: Final good and raw materis marks: no trade Mecons S* F pFW . F F !ADD m . D* ~ D Q, F ~~~~F I ~~~~~~St pW _ _ _ _ _ _ _ _ 0 a I IA. a a h. *. a - a U. 0- 0 - -- 15 .1 .D 's-I C%I I I1J a I-- - - a I I II II I b I I I a I I I I 4 6S1p IL 0.  a a. a 0~~~~~~~~~~~~~~0I 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ bt a IL~~~~~. I n~~~S . Ii O am - Zar 0 S~~~~~~~~~~~~~~~1 4 l- - - l - - - - - - - - -! ° 11 i _ -_-X -- ------1 / x X) \~~~~~~~~~~~~ 100 WI -S~c Policy Research Working Paper Series Contact Title Author Date for paper WPS1264 A Rock and a Hard Place: The Two J. Michael Finger March 1994 M. Patena Faces of U.S. Trade Policy Toward Korea 37947 WPS1265 Parallel Exchange Rates in Miguel A. Kiguel March 1994 R. Luz Developing Countries: Lessons from Stephen A. O'Connell 34303 Eight Case Studies WPS1266 An Efficient Frontier for International Sudhakar Satyanarayan March 1994 D. Gustafson Portfolios with Commodity Assets Fanos Varangis 33732 WPS1267 The Tax Base in Transition: The Case Zeljko Bogetic March 1994 F. Smith of Bulgaria Arye L. Hillman 36072 WPS1268 The Reform of Mechanisms for Eliana La Ferrara March 1994 N. Artis Foreign Exchange Allocation: Theory Gabriel Castillo 38010 and Lessons from Sub-Saharan John Nash Africa WPS1269 Union-Nonunion Wage Differentials Alexis Panagides March 1994 I Conachy in the Developing World: A Case Harry Anthony Patrinos 33669 Study of Mexico WPS1270 How Land-Based Targeting Affects Martin Ravallion March 1994 P. Cook Rural Poverty Binayak Sen 33902 WPS1271 Measuring the Effect of External F. Desmond McCarthy March 1994 M. Divino Shocks and the Policy Response to J. Peter Neary 33739 Them: Empirical Methodology Applied Giovanni Zanalda to the Philippines WPS1272 The Value of Superfund Cleanups: Shreekant Gupta March 1994 A. Maranon Evidence from U.S. Environmental George Van Houtven 39074 Protection Agency Decisions Maureen L. Cropper WPS1273 Desired Fertility and the Impact of Lant H. Pritchett March 1994 P. Cook Population Policies Lawrence H. Summers 33902 WPS1274 The New Traue Theory and Its Asad Alam March 1994 A. Alam Relevance for Developing Countries 87380 WPS1275 Female-Headed Households, Ricardo Barros March 1994 K. Binkley Poverty, and the Welfare of Children Louise Fox 81143 in Urban Brazil WPS1276 Is There Persistence in the Growth Ashoka Mody March 1994 M. Patefla of Manufactured Exports? Evidence Kamil Yilmaz 37947 from Newly Industrializing Countries WPS1277 Private Trader Response to Market Steven Jaffee March 1994 C. Spooner Liberalization in Tanzania's Cashew 32116 Nut Industry Policy Research Working Paper Series Contact Title Author Date for paper WPS1278 Regulation and Commitment in the Ahmed Galal March 1994 B. Moore Development of Telecommunications 38526 in Chile WPS1279 Optimal Hedging Strategy Revisited: Ying Qian March 1994 S. Lipscomb Acknowledging the Existence of Ronald Duncan 33718 Nonstationary Economic Time Series WPS1280 The Economic Impact of Export Wendy E. Takacs March 1994 M. Patefna Controls: An Application to Mongolian 37947 Cashmere and Romanian Wood Products