Measure and interpretation of effective protection in the presence of high capital costs : evidence from India

Policy Resewch 1 WOVRKING PAPERS Industry and Finance Operations Country Department IV Asia Regional Office The World Bank March 1992 WPS 873 Measure and Interpretation of Effective Protection in the Presence of High Capital Costs Evidence from India Francois M. Ettori The traditional measure of effective protection based on gross value added does not fully reflect the incentives of the protection structure when the domestic price of capital goods differs substantially from the international price. In particular, reform- ing India's trade policies and reducing its protection rates would be meaningless - even damaging - if India does not first reduce its high protection on capital goods. Policy RescarchWorkingPapcrsdisserninatethe findings ofwork in pwgrcss and encourage the exchange of ideas amniwgBank staff and allothers interested in developrntissucs.Thcsepapcrs, distributcJ by thc Research Advisory Staff,carry thcnamesofthcauthors, reflect only theirviews,and should bc used and cited accordingly. The rindings, intcrprstations,and conclusions am the authors'own. Thcy should not be attributed to the World Bank, its Board of Dirmctors, its management, or any of its mcmber countrics.  Policy Research| Industry and Finance Operations WPS 873 'T'his paper- a product of the Industry and Finance Operations Division, Country Department IV (India), Asia Regional Office - is derived from Policy Research Working Paper 433 (August 1990), which was part of a larger effort in the department to undertake a comprehensive review of India's trade regime and policies and to make recommendations for liberalization of its trade policies. Copies of the paper are available free from the World Bank, 1818 H Strcet NW, Washington, DC 20433. Please contact Francois Ettori, room H4-085, extension 32340 (March 1992, 24 pages). A striking feature of India's protective structure effective protection rate" (NEPR). The relevance has been high tariffs and protection on capital of these notions and their magnitude are tested goods, which limit industrial competitiveness on a sample of 60 industrial projects in India. and export potential, and distort industrial incentives as indicated in "effective protection The paper confirms the finding in a previous rates" (EPRs). Bank review of India's industrial sector that effective protective rates averaged about 40 The distortions introduced by high capital percent in the sector, with large variations and investment costs resulting from high levels between the industrial subsectors and within of protection were corrected in India's analysis each subsector. by introducing the notion of "corrected effective protection rates" (CEPRs). In theory, EPRs Using NEPRs, the paper shows that on computed on the basis of value added net of average the amount of effective protection depreciation could be made immune from capital available from India's protective structure is just cost distortions, provided that depreciation enough to compensate for the high cost of allowances are computed on economically investment that results from heavy protection of meaningful grounds and that EPRs based on net capital goods. Most projects have, in effect, value added ar-e available. But in India as in negative NEPRs, so they are at a disadvantage many developing countries, available EPRs are compared to foreign competitors. based on gross value added. The need to account for the substantial capital cost distortions led to Finally, the paper argues that reforming the use of a substitute tool, the CEPR. India's trade policies and reducing its protection rates would be mcaniingless - even damaging The paper provides a brief refresher, and - if India does not first reduce protection on geometrical interpretations, on the definition of capital goods. When warranted, the nominal F.PR and its limited interpretation as a measure protection rate for capital goods should be of the scope for inefficiency or extra profit slashed to the lowest possible level above the resulting from protection. It introduces the shadow premium for foreign exchange. notions and fornulae for the CEPR and the "net ThePolicy Research Working PaperSerics disseminates the findingsof work under way in thc Bank. Anobjectiveof the series is to ,et these findings out quickly, even if prcsentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the Policy Rcscarch Dissemination Center  Table of Contents Introduction 1 Basic Definitions and Concepts I An Interpretation of the Effective Protection Rate 3 Correctel Protection under a Distorted Price of Fixed Capital 5 A Case Study : India 7 Illustration of the Quantity Effects of X-inefficiencies 12 Conclusions 13 Bibliography Annexes  MEASURE AND INTERPRETATION OF EFFECTIVE PROTECTION IN THE PRESENCE OF HIGH CAPITAL COSTS : THE EVIDENCE FROH INDIA Introductin 1. This paper is derived from a previous Working Paper [61 prepared within a comprehensive review of India's trade regime and protection policies. One striking feature df India's protection structure has been the very high tariffs and protection on capital goods, thereby harming industrial competitiveness (and export potential) and distorting industrial incentives as they are indicated by Effective Protection Rates (EPRs). 2. The distortions introduced by the high capital and investment costs re- sulting from the protection levels on capital goods were corrected in India's analysis by introducing the notion of Corrected Effective Protection Rates (CEPRs). In theory, EPRs computed on the basis of Value Added (VA) net of de- preciation could be made immune from capital cost distortions, provided that depreciation allowances are computed on economically meaningful grounds and that EPRs based on net VA are available. Because in India, as in many other LDCs, available EPRs are based on gross VA, the need to account for the subs- tantial capital cost distortions led the analysis to use a substitute tool, the CEPR (defined in para. 14 below). Despite the formulae and equations deve- loped herein, the CEPR is not a theoretical development or addition to the abundant litterature on Effective Protection. More specifically, the CEPR notion is a pragmatic one, and does not pretend to substitute to th.- equili- brium EPRs associated to a general- or partial-equilibrium model and to an equilibrium exchange rate. It is meant, more simply, to provide the prac- titioner with a correcting tool for interpreting, under static conditions in a country, EPR estimates as indicators of the relative incentives between industries in that country, and to interpret EPRs from an angle somewhat different from the traditional theoretical angle. 3. With these limitations and modest objectives in mind, the paper presents three main themes. First, it provides a brief refresher on the definition of EPR and its lim4ted interpretation as a measure of the scope for inefficiency or extra-profit resulting from protection, with a simple geometrical repre- sentation. In the second part (and Annex 1), the paper introduces the notions and formulae of CEPR and Net Effective Protection (NEPR), also with a geome- trical representation. Finally, the paper tests the relevance of these notions, and assesses their magnitude, in the case of India on the basis of a sample of some 60 industrial projects. Finally, a brief conclusion summarizes the major findings and draws some tentative conclusions. Basic definitions and conce2ts 4. The concept and definitions of effective protection (cf. [21, [3] and [41) were originally prompted by the desirability, when analyzing the amount of incentives provided to industry by a given structure and level of tariff nominal protection, to net out from the nominal protection granted to the output of an industrial transformation process the additional costs charged to inputs on account of the protection enjoyed by these inputs. 5. This net effect of protection on inputs and output is logically measured  -2- on the difference between the value of output and that of inputs (including non-tradeables), i.e. the gross Value Added before depreciation. The Effective Protection Coefficient (EPC) is defined as: EPC - VAd/VAw , shown to be - NPCi + (NPCo - NPCi)/a (1), where: - VAd and VAw are the Value Added measured respectively with domestic and international border prices (non-tradeable inputs are in both cases valued with domestic prices); - "a" is the Value Added to output ratio in international prices (a - VAw/POw - 1 - (PIw * q)/POw, where POw and PIw are the respective international prices of output and input, and q is the gj_xe4 quantity of input per unit of output; a is by definition smaller than 1); and - NPCo and NPCi are the respective Nominal Protection Coefficients of output and inputs.' NPCs in turn are defined to be the ratio between the domestic (ex-factory, before indirect taxes) price of a tradeable good and its international border price.2 The formula indicates clearly that any difference in nominal protection coefficients between inputs and output is amplified by the factor 1/a into an effective protection coefficient different from NPCI or NPCo. 6. This mechanism is illustrated by Graph 1. The horizontal axis is measured Pnces Graph 1 in terms of NPCs, and the vertical axis measures the international prices Pw and , domestic prices Pd of input and output. P- _-- For NPC - 1 (i.e. no protection), the _ _ vertical bar displays the composition of POw betweer. its Value Added VAw and its V' 'VAd input cost CIw - q * PIw at international POw _ prices. Point A indicates the input cost at domestic prices corresponding to NPCi, , . A and point D the output domestic price POd VAw corresponding to NPCo. The vertical. bar , Cld for NPCo displays rhe components of POd, c.- ,-w i.e. the Value Adde 'lkd and the input _ ___ - cost CId - q * PId a :tomestic prices. 0 1 NPCi NPCO NPC 1 For the sake of simplification, only one input is considered in the definitior. of "a". The argument is easily generalized to the case of several inputs, including non-tradeables. In such case, NPCi would simply be an avera- ge of the NPCis of each input, weighted by the input values consumed per unit. 2 In a protection regime without Quantitative Restrictions (QRs) which permits unconstrained competition from imports, domestic prices of tradeables align themselves with the import prices after tariff duty. Thus nominal pro- tection coefficients are equal to 1 plus the tariff rate. Otherwise, domestic ex-factory prices (before indirect taxes) are the result of various effects of QRs and tariffs relative to the degree of domestic competition in the industry and of the supply/demand balance generated by the regulatory policies. In such cases, the realized nominal protection is best captured by the observed ratio of the domestic ex-factory price to the international border price (CIF). It is this latter definition of the NPC which is used throughout this note.  - 3 - It is clear from the graph's geometry that: EPC - VAd/VAw - AB/VAw + BC/VAw - NPCi/l + CD * POw/l * l/VAw hence thei formula (1) above, since CD - NPCo - NPCi. 3 7. Whenever EPC Is greater than 1, i.e. a positive Effective Protection Rate EPR - EPC-1, the additional Value Added generated by the industry in excess of VAw can be used to remunerate/pay the production factors of labor and capital above the remunerations they receive in VAw at international prices. This reallocation of Value Added can take several forms, as follows: - production factors are consumed or remunerated in excess of their consumption or remuneration under international prices (i.e., undez a free- trade regime), due t;o policy or operational inefficiencies; - if production factors are used and priced efficiently by international standards, the remuneration of capital above the fixed costs of capital, , the accounting profit itself, is increased by the amount VAd - VAw which then represents a "rent" granted by the positive effective protection; and - any combination of these two cases. The causes for inefficiency in the first case above (excluding the extreme ca- se of excessive input consumption, where VAd and EPC are artifi-ially reduced) can be a combination of: (i) x-inefficiency in the use of labor (excess labor, low productivity, high regulated wages,...) or in the use of capital (unecono- mic size of plant below MES, capacity under-utilization,...); and 'ii) alloca- tion distortions created by unappropriate policies (e.g., pricing policies). 8. To distinguish between the cases of inefficiency (operation or policy based) and those of extra-profit (protection rent), it is necessary to analyze one step further the breakdown of VAw and VAd between their different compo- nents and to compare their respective values in both cases. Value Added comprises three main components: VA - L + FK + P , where: L is the total Labor cost; FK is the fixed cost of capital (depreciation, and eventually interest on term debt if any); and P is the profit (gross).4 An interpretation of the Effective Protection Rate 9. The first two cases evoked in para. 7 of reallocation of Value Added between its components are illustrated by Graphs 2-A and 2-B. In both cases, 3 By definition, EPC - VAd/VAw - (NPCo.POw - NPCi.PIw.q)/(POw - PIw.q). Algebraic manipulation gives EPC - [NPCo - NPCi.(l-a)]/a , hence formula (l). 4 Gross profit P in turn can be split between income tax and net profit. In case of a rent extracted from positive effective protection, part of the rent can be appropriated by the government through a higher tax T. However, because the role of T is peripheral to the argument developed in this note, gross profit P will here be used in preference to net profit. Cf. Annex 1.  - 4 - the situation with protection (L.e., with NPCs different from 2.) is de facto compared to the situation where all NPCs are equal to 1, that is the free- trade regime. The reference comparator can be either a foreign competitor ope- rating under international prices (e.g., ln Hong-Kong) or the same industry in the same country after an hypothetical trade reform to a free-trade regime. In either case, international prices establish the reference basis and values for VA, th< output and input prices and the profit-return on investment. Pd greater than Pw is generally interpreted to represent a case of extra-profit or protection rent (graph 2-B). When Ld is greater than Lw, x-inefficiency in the use of labor is probable. If FKd is greater than FKw, it can be due to x- inefficiency in the use of capital per unit of output (uneconomic plant size, under-utilization of capacity, i.e. quantity effect)5, or to higher prices paid for fixed capital (price effect, cf. next section). Graph 2 Prices Prices Od LA Pd Pd FKd _______________________ ~NPC 0 NPCz NPC.' 0 I NPC, NPCo A - Case of inefficient use of production factors H - Case of extra profit from protection 10. For these reasons, the Effective Prctection Rate (EPR), the difference of EPC to 1, can be interpreted to be a measure of the scope for inefficiency or extra-profit granted by the difference VAw - VAd result.ng from protection on input and output. For instance, an industry with an EPR of +30% can pay or remunerate its labor and capitul 30% more than a competitor operating under international prices, or it could axtract a profit (after paying the labor cost and the fixed cost of capltal) substantlally above that of the competi- tor, depending on the cost of capital and the resulting distribution of Value Added between its components. 11. The standard interpretation of EPRs as a measure of the scope for inef- ficiency or extra-profit assumes implicitely that behind the difference bet- ween FKd and FKw there is only a quantity effect but no price effect, and that the price of capital is approximately similar or constant under the two situa- tions being compared (free-trade versus protection). This implicit assumption is largely correct in most cases and countries, where the industrial policies 5 In subsectors characterized by slgnificant economies of scale (gene- rally fluid-processitg industries such as chemicals), the investment cost per unit of output increases by about 25X each time the capacity is reduced by half. If capacity C under-utilized, the unit capital cost is inversely rela- ted to the capacity utilization rate. In both cases, x-inefficiency In selec- tion or operation of the production process generates a FKd greater than the FKw of the efficient international price-maker in the considered industry.  -5- are designed to keep costs of Lnvestment and capital goods clove to interna- tional prices, in order to avoid loading production costs with locked-in financial a'nd fiscal charges over the llfe of industrial projects. In all developed cou,.trie9 and practically all developing countr:es, import tariffs and nominal pzotection applied to capital goods are quite low, much lower than tariffs and protection applLed to other tradeables, and capital goods are of- ten exempt from import duties and domestic taxes. For instance, in Brazil which has highly protected its domestic industry, tariff collection rates on machinery in the mid-80L were in the 11-17% range. In Korea, they were about 9% for domestic use, and negligible for ezport production. Even in Pakistan (which has the second highest overall tariff collection rate after India), tariff collection rates on machinery were 15% in 1987/88. Corrected Protection under a distQrted price of fixed capital6 12. By contrast, India is almost unique in levying high tariffs and taxes on capital goods. Its average tariff collection rate on machinery was about 70% in 1987/88, and this high protection in favor of its domestic capital goods industry (coupled with strict QRs) has led to NPCs averaging about 1.40 for domestic capital goAds. As a result, investment costs of industrial projects in India are, ceteris p4rLbus, about 50% highez on average than investment costs of comparable projects at international prices 16]. 13. Under a structure of high protectlon and prices for capital goods, higher investment costs entail impllcitely that: (i) Value Added generated by an industrial project normally include a larger amount of capital remuneration (depreciation, interest, and profit as return on equity); and (ii) thus domes- tic Value Added VAd exceed Value Added at international prices VAw, with a resulting positive EPR, even if the domestic firm is efficiently operated by international standards. A high price of capital goods and investment, charac- terlzed by an average NPC for capital goods (NPCk) substantially above 1, generates in the value added at domestic prices a fixed capital cost FRd which is approximately proportionate to NPCk ( depreciation and interest charges are, ceteris paribus, a fixed proportion of the total investment cost). Assu- ming (for the sake of simplicity) that the labor cost L would not substan- tially decrease, at least in the short/medium term, if the project were to operate under free-trade, a hlgh NPCk requires a high FKd which thus cuts down profit Pd, in particular when NPCk is higher than the output NPCo. This mechanism is illustrated by the Pes Graph 3 . graph 3 where FKd, the projection of . FK on the vertical line NPCk, is Pd proportionate to NPCk. When NPCk .- increases (horizontal arrow), FKd increases, as well as FKd's pro- jection on the VAd vertical line. P The portion of VAd other than Ld .- .-- : (fixed) and FKd shrinks (vertical - .- - .-...-.- arrow), up to a point where the ,.: . ' resldual room left for profit can -__ ______.'_ vanish altogether. X I N.i NPC. NPCk 6 The issues associated with a distorted price of financial capital (e.g. interest aubsidies) on EPRs have been extensively analyzed in the litterature. They are not the subject of this paper, and are not addressed herein. If  -6- 14. It is generally desirable that the lndustrial and trade policles provide all investors with a minimum return on investment, for a number of reasons (attract foreign investment, encourage domestic livestment, leave to the dom- estic industry enough profit to permit it to lnvesc ln modernizatlon, innova- tion and R&D and thas keep abreast wlth its foreign competitors,...). Again, a good benchmark for the minimum return would be the the return on investment available from international prices under the free-trade regime. Intuitively, VAd should have a minimum level to cover, in addition to the labor cost L, a total remuneration of capital FKd + Pd equivalent to (FKv + Pw) multiplied by the inflating factor NPCk of the Lnvestment cost7. When the magnitude of distortion is slgnificant, i.e. when NPCk is significantly above 1, it is shown in Annex 1 that, for various definitions of the return on investment, this minimum value of VAd corresponds to a unique value of the EPR, denoted CEPR, which ls: CEPR - (1 - s).(NPCk -1) (2), whenever "s" - L/VAw, the labor content of VAw, is moderate (say below 30%). CEPR is called herein the Corrected Effective Protection Rate, because it is the level of EPR just sufficient to compensate for the higher price of capital and earn the return on investment achievable under the free-trade regime (if operating efficiently).8 It measures the Rrice effect of invest- ment cost policies on effective protection. The difference EPR - CEPR - NEPR is called the Net Effective Protection Rate; it indicates the additlonal protection available above (resp. under) the CEPR, and thus the scope for x- inefficiency or extra-profit (resp. the obligation for extra x-efficiency or loss). 15. The VAd assoclated to CEPR corresponds to a level for the output price POd, corresponding in turn to the associated concept of Corrected Nominal Protectlen Rate for the output, denoted CNPRo, which is: CNPRo - a.CEPR + (1 -a).NPRi (3) (cf. Annex 1) The first term of the sum represents the impact of the investment cost on the output production cost, and the second term represents the impact of input costs. The difference NPRo - CNPRo - NNPRo between the actual nominal protection rate of the output and CNPRo is similarly called the Net Nominal Protectlon Rate for the output. A balanced and equitable structure of protection rates for output, inputs and capital goods (thus for NPCo, NPCi and NPCk) leaving no room for x-inefficiencies nor extra-profit would be characterlzed by low or null NEPRs, or equivalently NNPRs. 16. The concept of Corrected Protection (effective or nominal) ls not rhetorical. The case of India, analyzed in the following section, illustrates 7 In all rigor, Value Added comprises also the financial costs of working capital, which ls roughly proportlonate to the average NPCL. But this is negligible when NPCk ls sign!flcantly above 1. 8 Using one unit of VAw as numeraire, formula (2) can be interpreted simply: the additlonal VAd over VAw (CEPR) is required for the remuneration of the addltional investment cost (NPCk - 1) at the rate (1 - s). In [6] and an earlier version of this paper, CEPR was called Compensatory Effective Protectlon Rate, which carrled an undesirable normative connotation.  -7- the substantf%l impact of high investment cost4 policies on the l1vel cf protection required to ensure the profitability of industrial investmew;ts. Furthermore, in countries or cases where investsuerc costs are nor distorted upwards but where other policies (protection, competition) permit the finan- cial viability of plants below Minimum Economic Scales (MES), the unit inves- tment costs will be substantially higher than those of the internationally efficient reference comparator (cf. footnote 5). In such cases, a similar ana- lysis with the price factor NPCk replaced by the quantity factor "x" (cf. Annex 1) would permit to separate the effect of uneconomic capacity choices and underutilization (CEPR) from that of other operational x-inefficiencies (NEPR). Also, distorted investment costs can erect barriers on new entry and competition against incumbents. In all these cases, the interpretations and policy conclusions to be derived from EPR analyses should establish first whether the concept of Corrected Protection is significantly relevant to the case, and if so separate CEPR protection from NEPR protection. A case study: India 17. Some 60 industrial projects (mostly in the engineering and chemical industries) financed in 1988 and 1989 by India's Development Finance Insti- tutions were analyzed in preparation for the World Bank's review of the trade regime of India ([1] and [6j). The available data were extracted from project appraisal reports prepared by Indian DFIs which reported systematically the international price equivalents of output, input and investment domestic prices for the standard computation of Economic Internal Rates of Return (ERRs). Notwithstanding the limitations of the available data, the analysis provided valuable indications of the incentives and disincentives resulting from the protection structure as they are perceived ex-ant by project promoters and financiers in India. The following section summarizes the quantitative results of the analysis and underlines the significance of corrected protection under the set of values taken in India by the principal parameters of the protection structure. 18. In India, the factor representing the ratio of investment costs (per unit of output) in domestic to world prices is substantially above 1 for seve- ral reasons. First, collected teriff duties on imported capital goods are high, averaging some 70% over the sample (cf. [6]). Only electronics industry machinery enioys a lower tariff duty of 35%, while other industries in the sample pay an average tariff duty of 80% on their imported equipment. Second- ly, locally procured capital goods, generally representing a large share of total equipment, carry purchase prices which average 40% above international prices. Thirdly, other goods and materials used in investment projects, cement especially, are charged substantial excise and other taxes which increase further the financial costs of investment above international costs.9 19. As per its objectives, the India review focussed exclusively on the effect of protection and pricing policies, especially for capital goods, on the competitiveness and the structure of production costs in industry. For this reason, the analysis considered only the price factor NPCk in the assessment and determination of effective protection and of the associated CEPRs, leaving aside the quantity factors due to x-inefficiencies in the selection and operations of project capacities. The estimate of NPCk is the 9 Another reason, relevant essentially to the quantity factor "x", is that the overall level of protection has encouraged the entry of many projects with uneconomic plant sizes below MES and low utilization. Cf. para. 29.  - 8 - ratio of tha financial investment cost to the same adjusted for tariff diltlies on imported equipment and for the high prices of local equipment (the first two price fActors described in the previous paragraph).10 20. As noted in para. 16 and Annex 1, the simple formula desin4ing the CEPR is valid provided that the share "s" of labor costs in VAw is relatively small. Parameter "s" was estimated for each sample project by adding to the direct labor cost the "overheads" costs (which comprise not only non-direct labor but also certain services inputs such as telephone,...). The estimated values of "s' represent thus an upper limit of the labor content of Value Added VAw. It is cle .- that "s" is indeed small in these projects, as shown by the following table which summarizes the basic parameters of the samp'e: TABLE 1: BASIC PARAMETERS OF THE PROJECT SaMPLE /a SubsectrqMs Sample share oX) ICOR Factor s Factor NPCk Nc.proi. VAw (X Heavy Chemicals 17 20 3.2 12 1.36 Synthetic Yarns 7 15 3.1 12 1.54 Basic Steel Goods 12 7 2.4 19 1.46 Heavy Miscellns 9 13 4.1 21 1.42 Avera3.e Intermdtes 45 55 3.3 15 1.43 Light Chemicals 5 3 1.6 19 1.33 Food Industries 7 6 1.9 16 1.36 Electronics 17 12 1.3 25 1.25 Other Engineering 26 24 1.5 17 1.45 Avge Final Goods 55 45 1.5 20 1.38 Overall Average 100 100 2.5 17 1.42 /a All parameters in this table are weighted averages. Source: Annex 2 21. The sample data confirm that: (i) investment costs in India are substantially higher than international costs, by about 42X on account of the price of capital goods; and (ii) selected processes and technologies are capital-intensive with a low share "s" of labor costs in the value added VAw (17X on average, with a maximum of 25X in electronics). The data reveal also the presence in t-v sample of two distinct groups of industries: (i) the first group consists oi heavy industries producing intermediates and inputs, with ICORs generally above 3 and a low "s" averaging 15X; and (ii) the second group comprises less capital-intensive industries producing final goods, with ICORs around 1.5 and a slightly higher "s" averaging 20X. 10 A second estimate NPCk2 is the ratio between the financial investment cost to the leconomic" invastment cost (used for ERR computation), which captures not only the effect of capital goods prices but also the price distortions of other goods and factors (e.g., cement). This second estimate represento an upper limit of t i cumulative effects of all forms of investment taxation in India. Results corresponding to NPCk2 are given in the Annexes.  -9- 22. The relktive magnitudes of NPCk and 's" justify the application of the CEPR formul& of para.14. Table 2 below presents both the resulting CEPRs and the actual Effective Protection Rates (EPR), as well as the NEPRs. These results indicate large inter-sectoral variations in EPRs and NEPRs. Sub- sectoral CEPRs (adequately weighted by VAw) range between 15% for electronics to 30-40% for most other subsectors and up to 45% in synthetic yarns. Furthermore, though the overall average NEPR Is null (within an error it- gin of + 10%), the large inter-sectoral variations of NEPRs indicate unequal incentives from one subsector to the other. The only subsector where the actual EPR is close to CEPR is Heavy Chemicals. Positive NEPRs are significant (25%) in Basic Steel Products and Synthetic Yarns, and substantial (47%) in Heavy Miscellaneous industries (paper, tyras). Nezative NEPRs are tiodest in electronics (7%), significant in food industries (22%), substantial in light chemicals (34%) and large in engineering industries (52%). TABLE 2: CORRECTED AND ACTUAL EFFECTIVE PROTECTION BY SUBSECTOR/a Distributn(%) Actual of Projects Subsector CEPR (X) EPR (%) NEPR(%) with +/- NEPR Heavy Chemicals 28 37 9 60/40 Light Chemicals 28 -6 -34 0/100 Synthetic Yarns 44 70 26 50/50 Basic Steel Products 37 60 23 57/43 Electronics 14 7 -7 60/40 Other Engineering 40 -12 -52 0/100 Food Industries 30 8 -22 50/50 Ileavy Miscellaneous 30 77 47 80/20 Overall 30 30 0 42/58 /a CEPRs, NEPRs and Actual EPRs in this table are averages weighted by VAw. Source: Annex 3 23. Indian policy-makers have traditionally tried to adjust nominal protection levels to their perceived need of each industry, or product-group, through a multiplicity of ad-hoc tariffs and exemptions. The large inter- sectoral variations in NEPR reflect the ad-hocism and inadequacy of such protection policies. This is further illustrated by the intra-sectoral discrepancies within each subsector. Only in light chemicals and engineering are all the projects subject to the same incentive (negative NEPR), as indicated by the last column of Table 2. In practically every other subsector, the projects are distributed in approximately equal shares between those benefitting from a positive NEPR and those subject to a negative NEPR, which can be construed to be the result of a quasi-random prccess. The structure of Effective Protection in India provides thus very heterogeneous incentives. 24. The use of the Corrected Nominal Protection rates (CNPRs, cf. para.15) permits to estimate the respective impacts of capital costs and input costs on  - 10 - the production costs and competitiveness of Indian industries. The results are summarized in Table 3 below which presents the subsectoral CNPRs along with the actual Nominal Protection Rates (NPRs) and the Net nominal protection rates (NNPRs). An important conclusion emerges from the table. Surcharges on capital and investment costs related to the CEPR (first column of the table) increase domestic prices on average by 12% of output's international prices. Extra price and cost of inputs increase domestic prices on average by 31%. The overall impact is that the sample projects require normally a Corrected Nominal Protection rate (CNPR) of 43% on average, s:y about 45%, two-thirds of which to offset the impact of high input costs and une-third to offset the impact of high capital and investment costs. TABLE 3: CORRECTED AND ACTUAL NOMINAL PROTECTION BY SUBSECTOR/a (In % of Output Value at International Prices) Distributn(%) VA/CEPR. Input Cost Actual of Projects Subsector Impact Impact CNPR NPR NNPR with +/- NNPR Heavy Chemicals 10 46 56 59 3 60/40 Light Chemicals 15 19 34 16 -18 0/100 Synthetic Yarns 17 46 63 65 3 50/50 Basic Steel Goods 10 50 60 62 2 57/43 Electronics 3 31 34 32 -2 60/40 Other Engineering 19 24 44 19 -25 0/100 Food Industries 1 23 24 26 2 50/50 Miscellaneous 6 31 37 46 9 80/20 Overall 12 31 43 36 -7 42/58 /a CNPRs, NNPRs and Actual NPRs in this table are averages weighted by Output in world prices. Source: Annex 4 25. The concordance between actual and compensatory NPRs is higher than for effective protection. The actual NPR is close to CNPR for Heavy Chemicals, Synthetic Yarns, Basic steel products, Electronics and Food industries. Positive NNPR is negligible (about 9%) in the Miscellaneous industries (dominated in the sample by Tyres). On the other hand, negative NNPRs are significant in engineering industries (25%) and in Light Chemicals (18%). Finally, the intra-sectoral discrepancies observed within each subsector for effective protection (cf. para. 23) are equally applicable to the nominal protection, as summarized in the last column of the table above. Except for light chemicals and engineering industries where all projects are subject to negative NNPR, the projects in each other subsector are distributed in approximately equal shares between positive and negative NNPRs. On a product- wise basis (cf. Annex 4), it is important to remark that subsectors producit., intermediates and inputs for other subsectors (heavy chemicals, synthetic yarns, basic steel products) show quite higher NPRs and CNPRs (averaging about  - 1U - 62%) than the other subsectors producing final goods which have NPRs and CNPRs averaging about 32%. 26. It should be noted that the average actual NPRs and CNPRs, of 35% and 45% respectively, are substantially lower than the average tariff collection rates of 60-70% for 1987/88, which in turn are much lower than the average official customs tariffs of 130-140% (cf.