This paper is prepared f or staff use and is not for quotation. The views are those of the author and not neces- sarily those of the Bank INTERNATIONAL BANK FOR RECCOSTRUCTION AND DEVELOPNf'JT INTERNATIONAL DEVELOPIMNT ASSOCIATION Economic Staff Working Paper No.101 EFFECTIVE PROTECTION: A SUMMARY APPRAISAL March 22, 1971 In this paper, prepared at the request of the organizers of the Conference on Effective Protec-tion (Geneva, December 18-20, 1970), the author provides a summary appraisal of ef- fective protection. In the first part of the paper, he con- siders the reasons for the introduction of the concept and its advantages over that of nominal protection. He further indicates the uses of the effective protection measure and discusses the problems of its application. Finally, sug- gestions are made for future research on the subtject. In writing the paper, the author has drawn on his work on effective protection carried out at the Bank. Much of this work is incorporated in The Struc-ture of 1lrotect,ion in Developing Countries to be published for the Bank by the Johns Hopkins University Press in June 1971. Prepared by: Bela Balassa Consultant EFFECTIVE PRO1'ECTIC(N: A SUMM&RY APPRAISAL Bela Balassa The literature on effective protection has mushroomed since the con- cept received attention only five years ago (Johnson, 1965; Balassa, 1965; and Corden, 1966). A large nwnber of empirical studies have been undertaken while theoretical contributions have considered conceptual and methodological issues. (The bibliography prepared by David Evans provides a complete list of writings on the subject). This Conference has brought together people working on various aspects of effective protection. A number of interesting papers have been presented and there has been inuch useful discussion. The differences in the views of the participants may have narrowed somewhat as a result of the discussions, but they have by no means become identical. Nevertheless, I hope that xrV pre- sentat,ion will. in the main be acceptable to those present. In the following, I will examine the reasons for the introduction of the effective protection concept and note its advantages over that of nomi- nal protection. I will further indicate the possible uses of the concept and discuss the problems of measurement in a general as well as in a partial equi- librium framework. Finally, suggestions will be made for future research on the subject. Nominal vs. L3ffecti-% Protoction Economists traditionally directed their attention to trade in final commodities as if all stages of production were undertaken domestically, and considered the effects of tariffs (nominal rates of protection) on such trade. In the presence of trade in intermediate products, however, nominal rates will not appropriately indicate the extent of protection since decisions will be -2 - affected by the protection of their processing activity rather than of bhe product itself. At the, same time, trade ih. intermodiaWe products accounts for the overwhelming part of irnternational exchange; if we consider machi- nery aa an input, four-fifths of world trade takes place in such products. The introduction of the concept of eff'ective protection reflected a .dissatisfaction on the part of its initiators with models of international tra,de that do not allow for trade in intermediate products. But while this concept is relatively new to economic discussions, it has long been knoi.n to businessmen who are aware of the fact that the protection of their process- ing activity is affected by nominal rates on the product anid on its inputs as well as by the share of the processing mrargin (value added) i,nl the pro- duat price. Explanationa based on nomninal rates of protection on the pro- duct itself have restricted atteiibion to the first os' these three elements, to. the neglect of the other two. The question arises, however, if interconmnodity differences in tar- iffs on mlaterial inputs and in the share of value added in the product price substantially affect the conclusions on the structure of protection or whe- ther this is adequately represented by nominal r4tes on the product itself. Cohn (1969) suggest:s th-t the latter ig the case on the grounds that the estimates reported in his paper generally show a high rank correlation bet- ween nominal and effective rates of individual industries in particular coun- tries.. I do not find Cohen's conclusion acceptable for several reasons. To pegin with, it appears that similarity and dissimilarity in the ranlcing Qf industries by nominal and effective rates of protgction depends on4 the de- gree of aggregation of the irdustrial classification and on the dispersion of nominal rates of protection. In a study of effective protection in seven countries, it has been found that the Spearman rank correlation co- -3 - efficient reaches .95 in countries with a highly aggregated input-output table (Brazil) or a small degree of dispersion of tariffs (Norway). By cointrast, among countries with more disaggregated input-output tables anid greater variability in tariff rates, the rAnk correlation coefficient was only .59 in Pakistan, .71h in Malaya, and .63 or .82 in the Philippines de- pending on whether tariffs or price observations were used in the calcula- tions (Balassa, 1971, ch.3). The intercountry variation of rank correlation coefficionts indicates the usefulness of estimating effective rates of protection for each country, since without such calculations we would not know the relabionship between r.J,.dnal and effective rates in particular countries. And, even in countries where rank correlation coefficients are high, there are substantial differ- ences in the rankings of several industries by nominal and effective rates and the disaggregation of the aata would farther increase these differences. Nor can we predict absolute magnitudes of effective rates on the basis of information on nominal rates in a particular country, since the statisti- cal relationship between levels of nominal and effective rates is imuch weaker than that between theilr rankings and it also varies greatly from country to country (Guis inger and Schydlowsk-y, 1970). Yet, information on the values taken by the effecbive rates of protection is necessary to evaluate the structure as well as the cost of protection. Tiiis will permit us, for ex- ample, to gauge the escalation of protective ra-tes and to detect cases of negative value added at world market prices. Travis takes a different taclc in questioning the usefulness of the concept of effective protection. He submits that the concept is superflu- ous, since we cail definie all products as end-products -- as if each of them were produced by a single enterprise from the ground up -- in whiich case nominal and effective rates will be identical (Travis, 1968). Bult the equality of nominal and effective rates holds only in the absence of non- competitive imports; such imports create a difference betweeni the price of the product anU the sum of direct and indirect domestic valiie added so that the nfominal and the effective rates of protection will not be the same (Balassa,. Guisinger, and Schydlowsky, 1970). Nor can one follow Travis in arguing that noncompetitive imports should be regarded as factors of production, since one would then indiscriminately group together domesbic factors alnd foreign products. For example, in the case of copper products, tariffs would be said to "probect" doimstic fabrication as well as foreign copper producers. Such a formulation is of little interest even if the price of copper is affected by the tariff and it will have no economic meaning if foreign prices are determined exogeneously. It,should be added that the scope of noncompetitive imports depends on the extent and the strvicture of protection; comTiodities that are produced do- mestically under protection may be imported under free trade. More generally, in the presence of trade in intermediate produc-ts, comparative advantage should be defined in terms of productive activitiees rather than final products. In- ste!.d of considering the comparative advantage of a country in clothing, for example, we have to indicate separately its advantages and disadvantages with respect to the production of raw cotton, cottQn cloth, and clothingl imanufac- turing. Under free trade conditions, a,co ntry may then imrport cotton cloth while exporting raw cotton and clothing. iIn turux, protection will interfere wiith the international division of the production process along the lines of comparative advantage. To indicate the effects of protection on the structure of production, we need therefore a value added concept that the theory of ef- fective protection provides. Uses of Effective Rates The last statement brings me to the uses of the effective protection mea- sure. These uses can be cla^s;sified according to whether they pertain to the realm of positive or normative economics. As regards the former, one should first note the informational advantages of the effective protection measure in indicating the distribution of the subsidy provided by nominal rates of protec- tion among various stages of fabrication. More generally, this measure pro- vides additional infornation on the structure of protection in individual coun- tries. Next, and most appropriately for a Conference held on the premises of GATT, the effective protection measure finds uses in tariff negotiation3. It enables countries better to appraise the "offers" of others., especially in cases when these involve unequal reductions of duties on products at differ- ent levels of fabrication. Thus, for example, despite a reduction in the tar- iff on a final product, its effective protection may increase if the tariff on its major input is reduced to a nuch greater extent. Sucn instances have in fact been observed following the Kennedy Round of tariff negotiations (Balassa, 1968). The effective protection measure has further been used to gauge the discriminatory effects of the system of protection in developed nations on the imports of processed goods from developing countries (Johnson, 1967, ch.6; Balassa, 1967), and to explore the implications of granting tariff preferences through unequal reductions in tariff rates (Johnson, 1969). Related applica- tions include appraising the possibilities of trade deflection in a free trade area whiere participating countries retain the freedom of setting tariffs on imports from non-member countries. Some of these applications have been ex- tended to cases where nontariff barriers are employed (Baldwin, 1970). If appropriate assumlptions are nmade in regard to the relevant elas- ticities, one can also estimate char?,es in imports that, are due to the in- position of, or reductions in9 tariffs and otlher protective measures. The resulting changes in imports will depend on the producl,ion effects associa- ted with effective rates and the consumtption effecbs associated withl nominal rates; they have been estimated for the case of the elimination of tariffs by industrial countries (Balassa, 1965). Similar calculations can be made for exports, and both sets of results need to be ad.just;ed for the dif- ference in the exchange rates wnder the two situations being compared (Basevi, 1968). Under certain assumptions, the effective protectionl measure can be used to gauge the relative incentives protection provides to particular in- dustries. Corden (1966) Nas first concerned with the resource-pull and the resource--push effects of protectioni which pertain to changes in net outputs while Ethier (1971) has shown the linkage between effective rates and gross outputs. While the allocation of resources among industries is said to depend on relative effective rates, the effective rate of protection adjusted for the difference between the actual exchange rate and that obtainable under free trade will indicate the extent of net protection of a particular indus- try. Further interest attaches to the relative incentives provided to import substituting and export activities in protected industries. Relative incen- tives to these activities will depend on the prices received in domestic and in foreign markets and on the.prices paid Lor inputs used in producing for the two markets; their joint effects can be expressed by a modified version of the effective protection measure (Balassa, 1971, ch.l). -7- In a similar fashiorą, one can analy7,e the relative incentives pro- vided to domestic and to fcoreign investients in a particular industry and the costs and benefits of foreign investment.' This last point brings me to the normative applications of the effective protection measure, of which its use in estimating the cost of protection should first be noted. Under certain assumptions, eff'ective rates will express the domestic cost of saving (or earning) foreign exchange through import substitution (or exporting). This formulation pertains to the direct cost of processing (Bal- assa and Schydlowsky, 1968). Other writers have suggested that direct and indirect costs be combined (Bruno, 1963; Krueger, 1966); this amounts to tak- ing averages of effec tive rates of protection at various stages of fabrication, weighted by the contribution of direct and indirect value added to output un- der free trade conditions. If the former concept is adopted and we adjust foa: the d:ifforcrice bc.twccn the exchange raLo un-ider protection and under free tradf, effective rates can be utsed to measure the producti.on cost of protection to th'e national economy (Bergsman, 1Q70; Balassa, 1971, ch.4). Normative uses of effective rates further include the establishment of policy norms. Thus, it has been shown t.hat if a country wants to have a manu- facturing sector of a certain size or a certain proportion of its resources devoted to manufacturing, the tariff policy to be applied should be formula- ted in terms of effective rates of protection (Bertrand, 1970a). Effective rates have in fact been used in the World Bank in formulating guidelines f'or industrial polic,y by the developing coumtries (Balassa, 1970a) and these guidelines have been applied in advising several countries. The effective protection concept has also been applied in the World Bank in appraising investment projects. This involves estimating the effec- tive rate of protection Aecessary for tl-he viability of the project and jud- ging its desirabilit.y by reference to policy norms pertaining to the counitry where the investment would take place. In a similar fashion, requesbs for protection on the part of existinig industries can be evaluated by applying the effective protection measure. Estimation in a General Equilibriim Frawowork In describing tJhe uses of the effective p:rot;ection m:.asure, no refer- ence has been made to the problems that beset the actual mcasuremwnit of ef- fective rates. Ideally, effective rates of protection should be estimated in a general equilibriumn frJ:niiework so as to take accourlb of cha.nges in fac- tor prices associated with the imposition of protec cive Teazvies. The onl.y effort to estimate effective rates in a general eiuilibrium model has been made by Evars (1970). He found a rank correlation of .63 between the effec- tive rates thus measured and the esti-mated changes in resource flows in a model without growth constraints for particular industries which is relevant for our discussion. In turn, the rank correlation copfficient between effec- tive rates measured, respectively, in a general and in a partial equilibrium model is .52. These results cannot be used, however, to derive conclusions on the appropriateness or inappropriateness of effective rates as an indicator of resource allocation or on the existence of substantial differences between estimates of effective rates measured in a general and in a partial equili- brium framework. To begiin with, effective raties are supposed to indicate the resource-pull and resource-push effects of protection in Marshallian long-run under ceteris paribus assptions after all adjustments in capacity have been made. By contrast, Evans has used a medium term model that per- mits the expansion of capacity but does not accomnodate reductions in it. Correspondingly, the solution of his model gives the results that inidustries which can cover variable costs under free trade would continue to operate at existing output levels. Indeed., there is no change in activity levels in nearly one-half of the industries, and this,s in turn, lowers the correlation between effective protec.tion and changes in activity levels. The correlation wou:Ld presumably increase if the time-span of the model were long enough to permit the depreciation of equipment, in which case several of the industries in question would show a decline in oultput. The correlation between effective rates of protection measured in a general equilibrium model and resource flows, as well as that between effec- tive rates estimated in a general and in a partial equilibrium framework, are further affected by the assumptions on maximizing behavior, the form of the consumption and investment functions, the supply and productivity of labor., and prospective export demand. While such assumptions are necessary for the ten-year proLection Evans made by the use of his medium-term model, the model does not answer the question on the effects of eliminating protection under ceteris paribus assumptions. Estimation in a Partial Equilibrium Framework Apart from Evans' work, effective rates of protection have been esti- mated in a partial equilibrium framework under the assumptions of zero sub- stitution elasticity between material inputs and primary factors, unchanged factor prices, constant returns to scale, infinite foreign elasticities of demand (for exports) and supply (of imports), no transportation costs, and pure competition. The critics of the concept and nmeasurement of effective protection have raised doubts concerning the validity of these assumptions and have noted the consequences of reroving some of themn. Thus, it has been shown that, retaining the partial equilibrium framework., substitution between primary -10- factors, taken as a unit, and.intermediate inputs, as well as among the inter- mediate inputs themselves, will lead to an overestimation of effective rates of protection calculated from domestic (post-protection) input,-output coef- ficients and an underestimation of effective rates calculat.ed from 'free-trade (pre-protectior) coefficients (Leith, 1968; Grubel and Lloyd, 1971). The mag- nitude of this bias may vary from industry to ind-ustry, thoreby affecting t,he ranking of industries by effective rates. The direction of the bias due to subst;itution between primary factors and intermediate inputs is inldeternminate if substitution elasticities differ between individual primary factors, on the one hand, arid intermediate inputs, on the other. In this even Luality, the interinduisl,ry movemenrt of resources will be affected by differences in substitution elasticitics among pairs of primary factors and intermediate imports and in relative factor intensities among ind-ustries (lRamaswami and Srinivasan, 1968). By contrast, substitution among primary factors does not give rise to bias (Corden, 1969). The importance of the substitution issue depends on the magnitude of substitution elasticities and on interindustry differences in these elastici- ties. Empirical studies give evidence of little substitution betwqeen primary factors and intermediate inputs in response to price changes (Theil-Tilanus, 1964), anc a calculation based on French data has shown a substitution elas- ticity of less than 0.1 between value added and raw materials (Balassa- Guisinger-Schydlowsky, 1970). Furthermore, in the comparative study, cited earlier, estimates of effective rates derived by the use of domestic and free trade coefficients do not show the existence of a bias due to substi- tution; nor is the ranking of industry groups by effective rates sensitive to the choice of the input-output coefficients (Balassa, 1971, ch. 3). Factor-Price aimd Product Price L.ffects These results lead to the conlclusion that, in a partial equilibrium context, input substitution is not likely to modify the results to any con- siderable degree. In general equilibrium, however, factor prices will also vary so that protection will affect particular activities not only through changes in product prices but also through changes in factor prices. Factor- price effects are accentuated in the presence of substitution between prim- ary factors and interniediate inputs (Tan, 1970) and certain definitional problems will also result (Ethier, 1971). But let us first Consider factor- price effects inthe absence of input substitution. It is easy to show that, in a three-commodity model, the effects of protection on particular industries may not be appropriately indicated by the effective protection measure even if substitution elasticities among in- pubs are zero. Thus, commodity A, having a lower effective rate than commo- dity B, may still cnjoy greater protection if it is complementary in factor use with unprotected commodity C and hence benefits from a protection-induced decline in the prices of the primary factors it uses intensively. The error possibilities due to the neglect of protection-induced chan- ges in fac-tor prices will depend on the magnitude of these changes relative to changes in the prices of products and their imaterial inputs. The critics have conducted the discussion in the framework of a Heckscher-Ohlin model where international cost differences are explained by differences in the relative prices of the factors of production. It is in the context of such a model that Travis claimed central importance for factoral protection (1968); also, the conclusions of his latest paper (1970) crucially depend on the as- sumption that, apart from a scalar that applies uniformly to all industries, production functions are identical internationally. -1 2- Retaining the Heckscher-Ohlin model, the relative importarnce of the product-price and factor-price effects of protection will depend on inter- commodity differences in factor intensities as well as in norinal rates of protection. While the former has received much atten-tion in -the literature, the latter has been neglected, presumably because much of the discussion has been conducted in the framework of three-commodity models (RiaLiasw˘ami.-Srinivascan, 1968; Bertrand-Vanek, 1971). Yet, interconmodity differences in nominal rates assume inmortance in a multi-product world. At one extreme., take the case wlien identical. ba iffs are levied on all importables whereas exportables are not subject to taxes or subsidies. We are then back to the traditional two-commodity model whore -- assuniing giveil world market prices, incomplete specializ ationi, and liniear homogeoneous produc tion functions -- protection-induced changes in relative factor prices are greater in magnitude than changes in relative product prices (Jones, 1965). In turn, under cete paribus assumptions as regar ds factor-price effects, the greater is the variability of tariffs, the greater will be the relative importance of the product-price effects of protection. It may be suggested then that in countries such as Denm-ark or Norway where the dispersion of tariffs i,s small, the factor-price effects of protec- tion can conceivably dominate the product-price eff'ects. However, in the larger industrial nations, and especially in developing countries, where a wide variety of tariffs and othe:? restrictions apply, it may be surmised that the effects of protection on output and input prices tend to outweigh its ef- fects on factor prices. Thus, for example, if effective rates of protection estimated in a partial equilibrium framework are ::O percent on commodity A and 30 percent on commodity B, factor-price effects are unlikely to reverse the relative levels of protection as measured by these rates. -13 - The conclusions regarding the importance of the product-price effects of protection are stren Jthened if we introduce inti3rnational differences in efficiency a la Ricardo that are emphasized by practitioners of effective protection. This would mean that, rather than protecting factors of produc- tion, countries tend to protect industries which have high costs because of the use of small-scale production methods (due to differences in market size), the application of inferior technology and organization (due to differences in techmical and organizational knowledge), and the prevalence of "X-ineffi- ciencies" (due to the failure to minimize costs for the technology applied). Empirical studies relating to the Heckscher-Ohlin theory have not been successful in explaining international specialization (Leontief, 1954 and 1956); empirical tests of the Ricardian theory generally gave better results (MacDou- gall, 1961; Balassa, 1963). The inportance of the variability of intercountry differences in efficiency across industries has also been shown in several com- parative studies (Arrow, Chenery, Minhas, and Solow, 1961; Clague, 1967). Ne- vertheless, further empirical research is needed to indicate the importance of the omission of factor price effects. Numerical examples are of little inter- est for the problem at hand, since they indicate the possibility rather than the probability bf factor-price effects outweighing product price effects. Introducing input substitution in a general equilibrium framework will, however, give rise to problems in defining value added and the effective rate of protection. Ethier (1971) suggests that in such a model one cannot derive an unambiguous measure of resource use (Cordents value-added product concept), and that the relevant definition will have to be couched in terms of marginal value added, leading to a rather complicated formula which might be difficult to measure emp.i.rically, However, Jones (1971) has shown that the value-added product concept has a meaning even with substitution and that the customary -14- definition can be applied. It should be added that if substitut-lion elastici- ties are low, as suggesbed above, input substitut:`.on will not create mrrLch difI- ficulty even if factor price effects are taken into account. The "Linearity" Problem The problem remains that, if all the listed asswnptions were futlfilled, under free trade conditions the number of commodities produced domestically cannot exceed the number of the factors of production (the "linearity" prob- lem). Under protectioni, the same condiltion applies to commodities that are not subject bo prohibitive tariffs; i.e. the numnber of cormmodities exceeding the nwuber of factors should become nontraded or nonproduced (Alderson, 1970). The existence of prohibitive tariffs makes it necessary to distinguish between potential anid realized protection. Potential protection, estimated from tariff observations, indicates the "cushion" available to producers, while realized protection, estimated frorn price comparisons, shows the extent of pro- tection actually utilized. It should be added, however, that realized protec- tion will also depend on domnestic demand conditions. A further consideration is that, for commodities which are produced under protection but not under free trade, the effective rate of protection will not indicate the resource-pull and resource-push effects of protection even in the absence of factor price and substitution effects. For example, in one industry a high rate of effective protection might be necessary to in- duce domestic production while in another low protection will be sufficient for this purpose. This will not affect, however, the usefulress of the ef- fective rate as a measure of the cost of protection. Under the stated assump- tiions, effective rates will indicate the excess o:' domestic costs over costs G,. the world market. The introduction of non-constant returns to scale or less than infinite Icreign elasticities will allow for the production of commodities exceeding the numnber of factors under free trade. But this will entail modifying sonme -15- of the assurrptions underlying the measurement of effective protection. It has been shonm that, if calcu.lations are made from domestic input-output co- efficients, effective rates of protection measured at existing output levels will be overstated in increasing-cost industries, and understated in decreas- ing-cost industries (Corden, 1971, ch.6). The opposite conclusion holds if we use free trade coefficienbs; we will now overstate effective protection in decreasing-cost indust,ries and understate it in increasing-cost industries. An attempt has also been made to adjust effective rates of protection for less than infinite foreign elasticities (Balassa, 1971, Appendix A), This question would, however, need further study., and attention should also be given to the interaction of tari.tfs and other protective measures in trading countries (Travis., 1970). Estimates would further need. to be made for commodities that would switch from the import to the export category or vice versa. It shoiild be addDd that, apart from the problem of input substitution, all the described difficulties in estimating effective rates of protection ap- ply to nominal rates as well. They are found in models of protection incor- porating only final goods and are in no way affected by -the introduction of trade in iLtermediate products. At the same time, nominal rates have the additional shortcoming of not allowing for the effects of input tariffs and the share of value added in the product price on the protection of process- ing activities. Moreover, although the inclusion of intermediate products in the model gives rise to the substitution issue, this seemns to be a small price to pay for making the model more realistic and, at any rate, available evidence indi- cates that the practical importance of substitiution between primary factors and intermediate inputs is rather small. It follows that the choice is not between effective and nominial rates of protection but we rather need to im- prove the conceptual framework and estimiiation of effective rates. Befor e turning to areas of future research, however, note should be, taken of pos- sible alternative concepts of effective rates. Alternative Conc3pts of Effective Rates Effective rates of protection have originally been formulated as the percentage diff erence of value added per unit of output tunder protection anid under free trade (Johnson., 1965; Balassa, 1965; Corden, 1966). Subsequently, it has been noted that, in the presence of input substitution, changes in value added per unit of output may be due to a clhange in bhe quantity of pri- mary factors per unit of out1put as well as to a chaiige in returns to primarL,y factors. To ensure comparability with nominal rates that represenL- a price change rather than a quantity change, it has been suggested that the concept of the effective rate of protection be limited to the second of the two ele- ments (Leith, 1968; Corden, 1969). Other writers have considered the implications of removing the assump.- tion of fixed factor supplies. Basevi (1966) has calculated the effective rate of protection with respect to labor, assuming that capital is mobile internationally. In turn, Schydlowsky (1967) has suggested that in a c.Dvelop ing country with unlimi-ted labor supply, the incentive effects of protection are indicated by the effective rate of protection to capital. The choice among these measures will depend on the purpose at hand. If we use effective rates to measure the cost of protection, the relevant concept will relate to value added. If, however, we are interested in the incentive effects of protection, effective rates should pertain to the scarce factors of production. Now, if we remove the nor..profit assumptions of pure competition, we may usefully reformulate the concept of effective protection in terms of rents to the entrepreneur. While it may be assumed that there -17- are no such rents under free trade, by affecting the prices of the product and of purchased inputs (including labor and capital), protection will cre- ate rents and resource flows may take place in response to rent differen- tials among industries. We have noted that, in estimating net effective rates and the cost of probection, there is need to adjust for the difference in the exchange rate under protection and under free trade. This amounts to taking the free trade situation as the norm and it corresponds to the maximization of econiomic welfare in a small country that takes world market prices as given. In turn, if foreign el.asticities are less than infinite and we exclvde the possibility of retaliation, the optimum tariff situation will provide the appropriate norm f'or estimating net effective rates and the cost of protec- tion. Wle may also extend. the welfare optimization concept inherent in the determination of the optimum tariff regime to posbulate an optimum that takes account of governmental preferences for industry or for public sector activities (Bertrand, 1970b) and draw up guidelines in terms of effective protection that ensure maximizing the particular objective function. More generally, the optimal structure of protection will depend on the objective function chosen. It should be added that, in the presence of substitution between pri- mary factors and intermediate inputs, different concepts of the effective rate of protection apply, depending on whether we are concerned with chanlges in primary factor use or in gross outputs (Ethier, 1971). This conclusion fol- lows since, if elasticities of input substitution differ from one commodity to another, changes in the amount of intermediate inputs used in the produc- tioIn process may outweigh changes in the3 use of primary factors. Directions of Plture Research Various efforts have been made to explore the implications of effec- tive protection in a general equilibrium framework. Corde-n (1969) has pro- vided a geometrical model of two products and two factors; iiuffin (1969, 1970a) and Balassa (1970b) have used an ariithmetical model to deal w.ith the three-commodity case and Ruffin (1970b) has generalized the argumlTent to the case of n commodities. These writers assun- constant input coefficients wlhile Tan (1970) and Ethier (1970) have considered the pro)blems related -to input substitution and factor price changes in a general equilibriwum model. Further work on effective protection in a general equilibriuiim control is desirable, both to indicate the limitatioins of partial equilibrium analy- sis and to prepare the ground for estimation in a general equilibrium frame- work. Nevertheless, efforts in this direction will be constrained by the state of the art in hand:ling general equilibriun models, so that estimation in a partial equilibriiu framework cannot be foregone. As regards the latter, attention should be focused on improving the calculations and on indicating the sensitivity ot the results to the asswunp- tions made. Questions of particular importance are averaging procedures (Basevi, 1970; Tumlir and Till, 1970), the treatment cf nontraded inputs (Corden, 1971, ch.7), and the estimation of the effect:s of protection on the exchange rate, together with the implications of less than infinite foreign elasticities (Balassa, 1971, Appendix A; Corden, 1971, ch.5). There is fur- ther need to test the sensitivity of the results to the assunptions made in regard to substitution elasticities and to establish likely values of these elasticit ies. Emrpirical tests are also needed to indicate the relationsh4ip between effective rates of protection and resource flows, Furthermore, one needs to -19- explain how the excess of domestic over world marlcet value added shovi by the estimates of effective protection is absorbed. As noted earlier, this may involve above-norma:. przfits and wages or excess costs due to a variety of causes, including allocative inefficiencies, inferior technology and or- ga:aiation, the application of small-scale production methods in a protected market and "X-inefficiencies'. Finally, it appears desirable to broaden the analysis to incorporate other incentive measures that affect relative prices and the allocation of' resources. In a research projecb under way at the World Bank, in addition to instruments of protection, consideration is given to credit, tax, and expenditure measures. This involves reformalating the effective pro- tection concept in terms of the net incentives provided by governmental mea- sures. Correspondingly, i.t is necessary to establish norms for each of the incentives. In the case of protective measures, the relevant norm is the free trade situation or an optimun tariff regime; for credit preferences it is the interest rate obtainable on a free credit market; for taxes it is a uni- form value added tax, with adjustments made for governmental preferences ex- pressed in the tax systemn; for goverrnent expenditures it is a budgetary sys- tem that is neutral with effects on individual activities. Ideally, these norms should be established in conjunction with each other since e.g. a move to free trade will affect the equilibrium interest rate, In practice, how- ever, the limitation of our knowledge in handling general equilibrium systems would make it necessary to consider them individually, at least for the time being. -20- References ANDERS0N, James E., "General Equilibrium and the Effective Rate of Protection", Journal of Political E2onmy, Part I, July/ August 1970, pp.717-24. ARROW, K., CHENERY, H.B. 1]THAS, B., and SOLOW,, R.M., "Capital- Labor Substitution and Economic Efficiency", Review of Econo- mics and Statistics, August 1961, pp.225-50. BALASSA, B., "An Empirical Dlemonstration of Classical Comparative Cost Theory",Review o' Econiomics and Statistics, August 1963, pp. 231-38. BAIASSA, B., "Tariff Protection in Industrial Coinitries: An Evalua- tion", Journal of Political December 1965, pp.573-94. BALASSA, B., "The Impact of the industrial Countries' Tariff Struc- ture on their Imports of Manufactures from Less-Developed Areas", Economica, November 1967, PP.372-83. BALASSA, B.., "The Structure of Protection in Indus Drial Countries ,and its Effects on the Exports of Pr.ocessed GooC1s from Develop- ing Countries" in The ]Kenned Round: Estimated Effects on Tar- iff Barriers, Geneva, UCTADJ, 19 8, pp.19-2071 (cited as 196bb. EBALASSA, B., Industrial Prot=ection i=n Deveoin Curies, Report No EDC-175, Washington, International Bank for Reconstruction and Development, June 1970 (cited as 1970a). BALASSA, B., "Tariffs, Intermediate Goods, and Domnestic Protection: Comment: and "Further Comrmaent", Amer:ican Economic Review., Decemr- ber 1970, pp.959-63 and 968-69 (cited as 197Ob). BALASSA, B., The Strulctiure of Protection in Developing Conttries, Baltimore, The Johns Hopkins University Press, 1971. BALASSA, B. and SCHYDIDLOSKY, D., "Effective Tariffs, Domestic Cost of Foreign Exchange and the Equilibrium Exchange Rate"., Journal of Political Economy, May/June 1968, pp.348-60. BALASSA, B., GUISINGERg S., and SCHYDLO',;SKY, D, "iThe 'Effec- tive Rates of Protection and the Question of Labor Pro- tection in the United States: A Comnent"l , Journal of Political Ecfonom Septemnber/October 1970, pp.1150662. BALDWIN, R. E., Nontariff Distortions of International Trade Wash- ington, The Brookings Institution, 1970. -21 - BASE\VI, G., "The U.S. Tariff Structure: Estimate of Effective Rates of Protection o.' U.S. Industries and Indusrrial Labor", Review of Economics and Statistics, May 1966, pp.147-60. BA3EVI , G., 'The Restrictive Effect of the U.S. Tariff and Its Wel- fare Value", American Economic Review, September 1968, pp.840-52. BASEV.I, G., "Aggregation Problems in Measuring Effective Protection", presented at the Conference on Effective Protection, Geneva, December 17-20, 1970. BERGSIMAN, J., "Commercial Policy, Allocative Efficiency, and 'X- Efficiency'", June 1970 (mimeo). BERTRAND, T.J.., "Decision Rules for Effective Protection in Less Developed Economies", December 19'70 (mimeo) (cited as 1970a). B:ERTRND, T.J., "Optimal Tariff Policy Desigried for Government Gain", Decemnber 1970 (miyaeo) (cited as 1970b). BZRTRIAD, T.J. anid VAN!7EK, J., "The Theory of Tariffs, Taxes, and Subsidies: Some Aspects of the Second Best", American Economic Review, September 1971. BRUNO, M., Interdependence, Resource Use and Structural Change in Israel, Jerusalem, Bank of Israel, 1963. CLAGU1E, C., "An Irtiernational Coniarison of Industrial Efficiency: Peru a7id the United States" $Review of Economics and Statistics, November 1967, pp.487-94. COHEiN, B.I., "The. Use of Effective Tariffs", Discussion Paper No.629 Econom:ic Growth Center, Yale University., February 1969. COPIEDN, W.M., "The Structure of a Tariff System and the Effective Protective Rate", Journal of PoliticalEc2no., June 1966, pp. 221-37. CORDE4, W.M., "Efifective Protective Rates in the General Equilibrium Model: A Geometric Note", Oxford Economic Papers, July 1969, pp.135-h1. COLDEN, W.M., The Teory of Protection, Oxford, Clarendon Press, 1971. EHEflR, W. J., "General Equilibrium Theory and the Concept of the Effective Rate of Protection", presented at the Confer- ence on Effective Protection, Geneva, December 17-2?, 1970. -ETHIER., W. J., "Input Substitution and the Concept of the Effective Rate of Protection", Journal of Political Econo 1971 (forth- coming). EVANS, H.D., "E'ffects of Protection in a General Equilibrium Frame- work", presented at the Conference on Effective Protection, Geneva, December 17-20, 1970. -22 - GRUBEL, H.G. and LLOYD, P.J., " Factor Substitution and Effective Tariffs", Review of Economnic Studiep 1971 (forthcoming). GUISlthElR, S. and SCHYDLOISKY, D., "The Empirical Relationship Between Nominal and Effective flates of Protection", presented at the Conference on Effective Protection, Qeneva, December 17-20, 1970. J0HSC(, H.G., "The Theory of Tariff Structure with Special Refer- ence to liorld Trade and Development", Trade and Deve , Geneva, Institut Universitaire de Hautes Etudes Internation4les, 1965. JOHISON, H.G., Economic Policies Toward Less Develo 1d Countries Waohington, The Brookings Institution, 1967. JOHNSCN, H.G.e, "The Theory of Effective Prptection and Preferen- ces", Economica, May 1969, pp.119-38. JONES, R.W., "The Structuire of Simple General Equilibriu.m Models", Journal of Political EconoMw, December 1965, pp.557-72. JONES, R.W. , "Effective Protection and Substitution", Journal of International Economics, 1971 (forthcoming). KRUBER, A0., , "Some, Economric Costs of Exchange Control: The Turkish Case", Journal of Political Ecopnmy, October 1966, pp.466-480. LEITH, J.C.# "Substitution and Supply Ela.sticities in Calculating the Effective Protective Rate', Quarterly Journal of Economics, November 1968, ppe588-601. LEcNTIEF, W.W., "Domestic Production and Foreign Trade: The Ameri- can Capital Posi-tion Re -examined", Economica Internazionl, February 1954, pp.9-38. TECUTIEF, W.W., "Factor Proportions and the Structure of American Trade: Further Theoretical and Empirical Analysis", Review of Economics and Statistics November 1956, pp.386-407. ,$4CDOUGALL, G.D.A. "British and American Exports: A Study Sugges- ted by the Theory of Comparative Costs", EEconomic Journal, Fart I, December 1951, pp.697-98. -23 - RAMASWAMII, V.K. and SRENIVASAN, T.H., "Tariff Structure and Re- source Allocation in the Presence of Factor Substitution", Discussion Papc,r No.33, Indian S-tatistical Institute, New Dehli, July 1968 (mimeo). To appear in "Essays in Honor of Charles P. Kindleberger", Nor-th Holland Publishing Co., 19'71. RUFFIN., R.J., "Tariffs, Intermedia-te Goods, and Domestic Protec- tion", American Econoiic Review, June 1969, pp.261-69. RUFTIN, R.J., "Tariffs, Intermnediate Goods., and Domestic Protec- tion: Reply", American Economic Review December 1970, pp. 961h-67 (cited as 1970a)T RKI"FIN, R.J. , "qWelfare Implications of Effective Protection", present;ed at the Conference on Effective Protection, Geneva, December 17-20, 1970 (cited as 1970b). SCHXDLOWSKY, D., "Effective Rates of Protection and Allocation of Reso-urces in a Competitive EconoPTr' 1967 (mimeo). TAN, A.H.H., "Differential Tariffs, Negative Value-Added and the Theory of Effective Protection", Am.erican Econoic Review, March 1970, pp.107-116. THEIL., H. and TILANUS, C.P... "The Demand for Production Factors and the Price Sensitivity of Input-Output Coefficien'ts", Interna-tional Econoomic Review, Septeember 1964, pp.258-72. TRAVIS., W.P., "T1he Effective Rate of Protection and the Ques- tion of Labor Protection in the United States", Journal of Political Econon, May/June 1968, pp.hh3-61. TRAVIS, W.P., "A Critical Rehabilitation of Effective Protec- tion Theory", presented at the Conference on Effective Pro- tection, Geneva, December 17-20, 1970. TU1MIL:R, J. and TILL, L., "Tariff Averaging in International Comparisons", presented at the Conference on Effective Pro- tection, Geneva, December 17-20, 1970.